Ritual Dynamics and the Science of Ritual: I: Grammars and Morphologies of Ritual Practices in Asia

Ritual Dynamics and the Science of Ritual General Editor Axel Michaels Editorial Board Michael Bergunder, Jörg Gengnage...

Author: Axel Michaels

155 downloads 2580 Views 236MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Ritual Dynamics and the Science of Ritual General Editor Axel Michaels Editorial Board Michael Bergunder, Jörg Gengnagel, Alexandra Heidle, Bernd Schneidmüller, and Udo Simon

I

2010

Harrassowitz Verlag · Wiesbaden

Grammars and Morphologies of Ritual Practices in Asia Including an E-Book-Version in PDF-Format on CD-ROM Section I Grammar and Morphology of Ritual Edited by Axel Michaels and Anand Mishra Section II Ritual Discourse, Ritual Performance in China and Japan Edited by Lucia Dolce, Gil Raz, and Katja Triplett

2010

Harrassowitz Verlag · Wiesbaden

Publication of this volume has been made possible by the generous funding of the Deutsche Forschungsgemeinschaft. Cover: Girl at the Ihi-Ritual, Kathmandu valley, Nepal. Photo: Niels Gutschow

Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.d-nb.de abrufbar. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the internet at http://dnb.d-nb.de.

For further information about our publishing program consult our website http://www.harrassowitz-verlag.de © Otto Harrassowitz GmbH & Co. KG, Wiesbaden 2010 This work, including all of its parts, is protected by copyright. Any use beyond the limits of copyright law without the permission of the publisher is forbidden and subject to penalty. This applies particularly to reproductions, translations, microfilms and storage and processing in electronic systems. Printed on permanent/durable paper. Printing and binding: Memminger MedienCentrum AG Printed in Germany ISBN 978-3-447-06201-5

Preface to the Series “Ritual Dynamcis and the Science of Ritual” The present publication is the first of five volumes of the proceedings of the conference “Ritual Dynamics and the Science of Ritual” that was held from 29 September to 2 October 2008 in Heidelberg, and organised by the Collaborative Research Centre “Ritual Dynamics” or SFB 619. This research programme was established in 2002 by the German Research Council (Deutsche Forschungsgemeinschaft), and has since then focused on a number of crucial questions in ritual studies, for example, how are rituals created and how do they come to disappear? How is it that they are transformed or transferred from one religion, region, or time period to another? Who invents rituals? Who are the agents of rituals? Ritual dynamics can be observed on several levels. Looking at the level of the dynamics of history, ritual dynamics become apparent in the tension between the poles of old versus new, traditional versus modern, and regional versus global. On the level of social dynamics, they arise from the tension emanating from the opposition of priestly versus laicistic (for example Brahman/anti-Brahman), individual versus collective, and also from the tensions between majority and minority cultures (in the diaspora). On the structural dynamic level, they result from a change of media, for example from script to image and internet, from script to performance, from norm to event, from prescription to realisation, or from description to construction. The Heidelberg collaborative research centre is interested in such ritual movements and processes, and structural as well as functional changes. It has now become apparent that change within and through rituals is as evident as the continuity that can be observed in many ritual practices. The ritual dynamics project has contributed significantly to this rethinking of ritual theory. To mention only a few major points: 1. Ritual criticism is not an exclusive feature of modernity, but has always been part of ritual performance. In other words: criticising the “old”, including conventional rituals, is just as much a sign of modernity or modernisation as the creation of new rituals or the rearranging of “old” action patterns. 2. The thesis that rituals degenerate or disappear in the modern Western world is unsustainable. Quite to the contrary, we learn that society plays with rituals, that rituals are part of a certain modern attitude and lifestyle, that they even persevere far away from the actual contexts of action (such as the internet). The global “experience-driven society” rediscovers and constantly reinvents rituals by combining old and new patterns, or by adopting them from different cultures. 3. The dynamics of rituals lead to many more renewals and modifications than hitherto assumed, since form is, in particular, related to variance, if not, in fact, determining it. Therefore, rituals are by no means limited to the function of excluding alternatives or abridging contingency and liberties. Instead, they must be understood as propositions

VI

The Editorial Board for alternatives and as symbolic actions, thus as highly productive elements for social interaction and the creation of meaning.

4. Due to their transcultural and transnational reception, and thanks to new media, traditional rituals are no longer bound to particular traditions, but hold an almost inexhaustible and increasingly global potential for the invention of new ritual practices. Dynamics in ritual practice is therefore the rule rather than the exception. Hence the new and novel is an important, if not the most significant, indicator of rituals. Rituals are not stereotyped and invariant events; rather, if seen in their context and historical dimension, they are the cause of social and cultural change. These and many other ideas were discussed during the Conference in more than twenty panels. They have been condensed into five volumes with 16 sections. The first volume “Grammars and Morphologies of Ritual Practices in Asia” focuses on grammars and morphologies of ritual practices in Asia (section 1), as well as on ritual discourse and ritual performance in China and Japan (section 2). Volume II “Body, Performance, Agency and Experience” covers the notions of body, performance, agency, and experience, focusing on ritual and agency (section 1), ritual, performance, and event (section 2), the body and food in ritual (section 3), and the varieties of ritual experience (section 4). The third volume “State, Power and Violence” contains contributions on state, power, and violence, with chapters on ritual and violence (section 1), rituals of power and consent (section 2), usurping rituals (section 3), and the state and ritual in India (section 4). Volume IV “Reflexivity, Media and Visuality” covers the aspects of reflexivity and discourse on ritual (section 1), ritual and media (section 2), ritual and visuality (section 3), and ritual design (section 4). The fifth and final volume “Transfer and Spaces” contains articles on ritual transfer (section 1) and ritualized space and objects of sancrosanctity (section 2). We would like to express our sincerest thanks to the German Research Foundation, DFG, for generously funding the SFB 619 as well as for providing the financial support for hosting one of the largest conferences in humanities and cultural studies ever held at Heidelberg University, and for publishing the conference proceedings. The editorial work would not have been possible without the tireless support of Sibylle Zerr for the printer’s copy, and Elvira Bijedic, Douglas Fear, Chris Allen, Anne Moßner and Raffaella Cengia for proofreading. The Editorial Board: Michael Bergunder Jörg Gengnagel Alexandra Heidle Axel Michaels Bernd Schneidmüller Udo Simon Heidelberg, April 2010

Table of Contents The Editorial Board Preface to the Series “Ritual Dynamcis and the Science of Ritual”

V

Section I: Grammar and Morphology of Ritual Edited by Axel Michaels and Anand Mishra Axel Michaels and Anand Mishra Introduction

3

Axel Michaels The Grammar of Rituals

7

Jan E. Houben Formal Structure and Self-referential Loops in Vedic Ritual

29

Olga Serbaeva Saraogi When to Kill Means to Liberate: Two Types of Ritual Actionsin Vidyāpīṭha Texts

65

Anand Mishra On the Possibilities of a Pāṇinian Paradigm for a Rule-based Description of Rituals

85

Axel Michaels (in collaboration with Johanna Buss) The Dynamics of Ritual Formality: The Morphology of Newar Death Rituals

99

Cezary Galewicz Inscribing Scripture Through Ritual: On the Ritual Cycle of Trisandhā

117

Martin Gaenszle Grammar in Ritual Speech: The Use of Binomials in Rai Invocations

141

Johannes Bronkhorst Ritual, Holophrastic Utterances, and the Symbolic Mind

159

J.C. Heesterman The Development and Impact of Ancient Indian Ritual

203

VIII

Table of Contents

Annette Wilke Basic Categories of a Syntactical Approach to Rituals: Arguments for a “Unitary Ritual View” and the Paraśurāma-Kalpasūtra as “Test-Case”

215

Timothy Lubin Ritual Self-Discipline as a Response to the Human Condition: Toward a Semiotics of Ritual Indices

263

Rich Freeman Pedagogy and Practice: The Meta-pragmatics of Tantric Rites in Kerala

275

Frederick M. Smith Historical Symmetry and Ritual Asymmetry: The Interrelations Between Vedic Ritual and Temple Construction in Modern India

307

Frits Staal A Theory of Ritual: The Indo-Iranian Fire Offering

331

Section II: Ritual Discourse, Ritual Performance in China and Japan Edited by Lucia Dolce, Gil Raz and Katja Triplett Lucia Dolce, Gil Raz and Katja Triplett Introduction

355

Michael Puett Ritualization as Domestication: Ritual Theory from Classical China

359

Joachim Gentz “Living in the Same House”: Ritual Principles in Early Chinese Reflections on Mourning Garments

371

Christian Meyer Interpretations of Confucian Ritual (“li”) in Chinese Scholarly Discussions in the Eleventh Century

397

Grammars and Morphologies of Ritual Practices in Asia

IX

Gil Raz Ritual Theory in Medieval Daoism

407

Julius Tsai 蔡南亭 Mutation or Permutation? A Ritual Debate in Tang-Song Daoism

425

Lucia Dolce The Contested Space of Buddhist Public Rituals: The shunie of Tōdaiji

433

Fumi Ouchi Buddhist Liturgical Chanting in Japan: Vocalisation and the Practice of Attaining Buddhahood

459

Katja Triplett Esoteric Buddhist Eye-healing Rituals in Japan and the Promotion of Benefits

485

M.A. Butler Ritual Specialists and Collective Agency in Song Dynasty China

501

Paul S. Atkins The Stages of Seppuku: Performing Self-Execution in Premodern Japan

23

Poul Andersen The Life of Images in Daoist Ritual

535

Benedetta Lomi The Iconography of Ritual: Images, Texts and Beliefs in the Batō Kannon Fire Offering

557

Abstracts

579

Section I: Grammar and Morphology of Ritual Edited by Axel Michaels and Anand Mishra

Axel Michaels and Anand Mishra

Introduction The dynamic character of rituals points towards the necessity of the existence of some regularity or structure in them. Any talk about variations, change or transformation assumes the existence of some structure through which this change becomes visible; which provides the necessary coordinates through which the trajectory of change can be envisaged. Ritual dynamics, like any other dynamic process, is an interplay of ritual structures and variations; structures being necessary, although not sufficient, to comprehend the totality of this phenomenon. A comprehensive study of the dynamics of rituals thus calls for investigating and specifying those underlying frameworks that enable, propel and give contour to this dynamism. Without such a framework there would be no dynamism but only directionless chaos. One of the future challenges of ritual scientists, therefore, is to evolve an acceptable model or a meta-framework for representation of ritual structures. As a grammar provides the requisite meta-level framework to stipulate linguistic structures, similarly the meta-level framework(s) for representation of ritual structures can be termed as “grammar(s) of rituals”. Representation of ritual structures can be attempted in terms of their characterising features or “parameters”. It is in this sense, that we can talk of “morphologies of rituals”, where one specifies ritual structures as a complex combinatorics of the concrete instances of ritual parameters, like the specification of linguistic utterances by means of combinations of constituting units. The first part of the present volume is an outcome of the inter-disciplinary conference on the science of rituals in Heidelberg (Sept. 2008) where the panel “Grammar and Morphology of Ritual” sought to initiate a systematic attempt to respond to the challenge of providing a scientific framework for comprehending ritual structures and variations. It focuses on the characterising features or “parameters” of rituals, as well as the possibilities of discerning certain structures which could further be specified through a kind of grammar. Here again, it is only metaphorically that one can talk of a “grammar” of rituals, because strictly speaking rituals are not identical to a language, although they often involve language (which might in themselves reveal a specific grammatical structure: e.g. Martin Gaenszle’s contribution). Yet there are two reasons why it makes sense to pursue this figure of speech. On the one hand, because it has long been customary in linguistics to include non-verbal

4


communication in grammars, and on the other because scholars are forever attempting to analyse the rules of rituals in linguistic terms. Moreover, we intend to use the term “grammar” in the general sense of specification of the elements and structure of an area of knowledge or skill. An overview of the attempts made thus far is summarised by Axel Michaels, who raises certain theoretical questions pertaining to the characterisation of ritual structures as well as their variations, suggests the possible ways of its realisation and points out its relevance for ritual studies. An important aspect while investigating the ritual structures is whether (or not) rituals in general and ritual components in particular have meaning content. Jan Houben argues and shows that this at least is the case with Vedic nihnava ritual. The constituent units of the nihnava refer to each other in a meaningful manner. This intra-reference renders the whole ritual as self-referential. He, thus, demonstrates the self-referentiality of these rituals and posits that meaning conditions are necessary to describe formal ritual structures. The relationship between ritual structure and its meaning is further discussed by Olga Serbaeva Saraogi, who, taking the examples of two śaiva tantric rituals first illustrates that they are structurally collateral and then hypothesizes the possibility of their having similar meaning on the basis of this structural parallelism. Rituals, especially liturgical rituals, are relatively structured and rule based actions. This idea is also expressed in the Sanskrit word for (life-cycle) rituals – saṃskāra – which, analogous to the expression saṃskṛta for the holy and refined language, indicates the notion of perfection contained in it. This implies that rituals are not haphazard or idiosyncratic, but specific and in many cases specified or named actions. A grammar of ritual could then be considered as those (rule-based) specifications which comprehend the structure and also the variance of rituals. This could be attempted by identifying the characterising features of a ritual ascertaining the degree of regularity and relationships between these features, and specifying paradigms or “rules of grammar” through which these regularities and relationships can be grasped. This approach is elaborated by Anand Mishra in his contribution, where he attempts to provide a Pāṇinian framework for a rule based description of ritual actions. In their analysis of the death rituals performed by a Newar Jyāpu (farmer) family from Bhaktapur and a Newar Rājopādhyāya Brahman priest originally from Patan, Michaels/Buss deal with the theoretical aspects of ritual formality, especially the question how and to what extent ritual sequences can be transformed, left out, added and transposed, and how the priest creates using these means his own ritual. Johannes Bronkhorst raises the fundamental philosophical questions about the origin of language and symbolic representations and posits that the essential differ-

Introduction

5

ence between the ritual and normal activity is that the former is holistic and the latter guided by symbolic representation. This relationship is similar to the one between ritual and normal linguistic utterances, the former being holistic expressions and the latter exemplifying symbolic reference. In his article about trisandhā rituals associated with recitation of vedic mantras, Cezary Galewicz takes up an example of rendering the utterances of linguistic expressions through a structured ritual action. Martin Gaenszle, by comparing the grammar of Rai ritual language in east Nepal with the ordinary language demonstrates that the grammar of ritual speech has distinct characteristics. He thus suggests that the question perhaps is not whether ritual has grammar-like rules or not but rather what kind of grammar-like rules distinguish it from ordinary action. Inspired by the Sanskrit grammarian Bhartṛhari, Annette Wilke suggests a less formal but holistic approach in her proposal to include four syntactical categories – sequence, relation, expression and imagination – for ritual analysis. This, she claims will take into account the aesthetic message of ritual as well as its auto-reflexive and auto-poetic effects. The aesthetic effects and visual attractiveness of ritual performances are also emphasised by Frederick M. Smith, who while analysing the mutual dependency of two unrelated but simultaneously performed rituals, considers this to be instrumental in ritual efficacy. Rich Freeman puts forward arguments and examples for a culturally nuanced model of semiotic pragmatics rather than models based on the traditional grammatical analysis in terms of phonology, morphology, syntax and semantics to describe rituals. The ritual action, like a language, becomes meaningful in a pragmatic context and in turn changes and affects the context as well. It is this to and fro interaction which can explain a ritual (as also a linguistic) expression. He tries to prove this by showing how the ritual and textual sources reciprocate in rituals as learned and performed by the students in a ritual school in Kerala. The strategic use of gestures and objects with iconic or indexical value are important factors to establish the plausibility of the claim of rituals to make real what is represented. This aspect of the analysis of ritual signification is the subject of Timothy Lubin’s contribution. On the basis of the indo-iranian fire offering, Frits Staal proposes a clear demarcation of ritual versus ordinary activity. The dynamics of rituals, according to him, is a controlled and categorical change as compared to social changes. There are rules and “rules about rules” or meta-rules which are prescribed in ritual texts and can correspondingly be interpreted as special and general principles. In his insightful analysis of śrauta rituals and its relationship with sacrifice, Jan Heesterman underpins the absence of contest and uncertainty in śrauta rituals in contradistinction to the uncertainty and open-endedness of sacrifice. He views

6


śrauta system as a fundamental reform of the institution of sacrifice, where the ritualists restructured the agonistic pattern of sacrifice, replacing “rules of the game” with the “game of the rules”. His illustration of the systematic construction of śrauta rituals, as arrangements of various formats of sacrifice according to increasing complexity, points towards the possibilities and potential of a structural analysis of such ritual activities. Analysing a ritual through its salient and characterising features or what we may call as ritual-parameters is not new and have been one of the general methods employed to study rituals since a long time. A standardised framework for the representation of these parameters, however, still needs to be evolved. Only such a framework which is based upon logically consistent methods can fulfill the requirements of developing a science of rituals. It will not only allow us to identify, represent and analyse the static as well as dynamic nature of rituals, but further make it possible to ask the next theoretical questions about the guiding principles of the dynamics of rituals and finally about the functionality of the propelling forces of this dynamics. This way of representing rituals through ritual-parameters will further enable us to represent ritual concepts on computer by means of an “ontology of rituals”, which is an explicit specification of a conceptualisation. Thus, ritual-objects, persons participating in a ritual, important concepts, symbolism, purpose, even meaning associated with a ritual and other such ritual-parameters together with the relationships existing between them will provide the conceptual entities for an ontology of rituals. Development of a computer enabled research platform for the analysis of rituals using the ritual knowledge base of this ontology would certainly prove to be a valuable asset towards establishing the “science of rituals”.

Axel Michaels

The Grammar of Rituals1 Although rituals often involve language, strictly speaking they are not themselves a language. So basically any talk of a grammar of rituals is meaningless, or only possible on a metaphorical level. And yet there are two reasons why it makes sense to pursue this figure of speech. On the one hand because it has long since been customary in linguistics to include non-verbal communication in grammars, and on the other because people are forever attempting to analyse the rules of rituals in linguistic terms. The aim of this paper is to bring together these two approaches and come up with some preliminary thoughts on a “grammar of ritual”. If Ritual Studies or Ritology2 are ever to establish themselves in their own right, on a broader basis than simply more or less arbitrary phenomenological3 similarities and associations, it is high time not only to define and delineate “ritual” more precisely,4 but also to determine the possibilities of formal and functional rules for rituals in such a way that they can be fitted into a kind of universal grammar of rituals. Why “universal”? Because all societies and cultures have rituals – even when isolated members of these groupings do not practice any rituals or are unaware of doing so. The question as to how far rituals must also be religious rituals depends admittedly on the hotly debated definition of the term “religion”. But given the wide usage of the word “ritual” it seems sensible for the moment not to draw the boundaries too tightly.

1. Ritual and Speech The first move to conceive of ritual as a language came from Franz Boas, who remarked in The Mind of Primitive Man (1911) that in its unconscious elements, ritual resembled language. But Franz Boas only hinted at these parallels and did not go into them any further. Susan K. Langer, on the other hand, picked up on his remarks in her major work Philosophy in a New Key (1942) and spoke of a basic 1 Slightly revised English translation of Michaels 2007. 2 On 27.4.2006 Ron Grimes gave his inaugural address in Nijmegen of the first chair of “Ritual Studies”. 3 See Michaels 2001. 4 See Michaels 1999 and 2003.

8

Axel Michaels

need for symbolism that finds expression in, among other things ritual, which to her mind is the “language of religion”: “Magic, then, is not a method but a language; it is part and parcel of that greater phenomenon, ritual, which is the language of religion. Ritual is a symbolic transformation of experiences that no other medium can adequately express.”5 Edmund Leach expanded the concept and described rituals as “cultural sets of behaviour [that] function as language”.6 For him “all customary behaviour [is] a form of speech, a mode of communicating information”.7 In his book Culture and Communication he adds: “I ... assume that all the various non-verbal dimensions of culture, such as styles in clothing, village lay-out, architecture, furniture, food, cooking, music, physical gestures, postural attitudes and so on are organised in patterned sets so as to incorporate coded information in a manner analogous to the sounds and words and sentences of a natural language.”8 For Leach the communicative function of rituals arises from their expressiveness. At the same time, Leach also sees clear differences between verbal and non-verbal communication.9 Thus unlike verbal language, the possibilities of non-verbal communication generating new comprehensible forms are, in his view, limited by customs and conventions. He also surmises that the syntax of non-verbal “language” must have a much simpler structure than written and verbal language. As Werlen has rightfully noted, the first point does not however hold, because less formalised or weakly liturgical rituals do in fact evince a wide range of variation.10 Similarly, the generative nature of languages also applies to their use in rituals, regardless of how strongly codified or standardised they may be. A decisive impetus regarding the communicative function of rituals was given by John L. Austin and John R. Searle and their speech act theory. The theory enabled actions to be understood as performative expressions and words were thus able to actually do something. From this point on it was no longer simply a matter of grasping language as a reference system, by which meanings and ideas can be exchanged or conveyed, but also of charting the communicative aspect of situations revolving around speech, language and writing. (That Austin and Searle were heavily criticised by Jacques Derrida, who wished to reassert the primacy of the 5 6 7 8 9 10

Langer 1992: 49. Leach 1968–72: 523. Ibid. Leach 1976: 10. Leach 1976: 19. Werlen 1984: 35.

The Grammar of Rituals

9

written word or indeed the sign function vis-à-vis the speech act, is more related to the philosophy of language than to grammatological aspects). The relationship between the verbal and non-verbal aspects of acts could no longer be resolved through speech act theory, which led to the question of a “grammar” of rituals being posed anew, particularly because the demand for an analysis of the structural features of the non-verbal side of actions, the pragmatics, was emphasised. All of these and other approaches were, however, more programmatic than proto-grammatical, which is to say that not even an attempt to glean the structural rules of rituals by comparative means and to develop them into a universal grammar were apparent. The analogies made between language(s) and ritual(s) relied chiefly on ascertaining formalism and regularity, and many were only basically interested in the functional aspects. In fact even the performative and communicative theories of ritual were essentially functionalistic theories.11 Other works on the relationship between ritual and language focused chiefly on the semiotics of rituals12 and the aspect of language in rituals13, and less on the question of rituals as language. This will be the sole concern in the following, and not a metaphorical use of the term “grammar”, in keeping with Iwar Werlen, who writes: “The metaphorical talk of grammar and language leads … to ambiguities: if one wishes to regard the rite as a structured constant that is a language, one should view grammar more in its narrower, structural, linguistic or generative, transformational sense. Otherwise all talk of ritual as language is reduced simply to the semiotic nature of rituals.”14 So what was missing and by and large continues to be so is a corresponding specification of the “morphology”, “syntax”, “semantics” and “pragmatics” of rituals. Alone Frits Staal and Lawson & McCauley have taken preliminary steps in this direction by concentrating on the syntax of rituals. But since Staal simultaneously reduced rituals to syntax and denied them any meaning, and Lawson & McCauley concentrated solely on religious rituals, this quasi-linguistic field in their investigations was mostly overlooked or never touched on again. There have been other reasons why no agreement has been reached as yet on a “grammar” of rituals. One is that the complex of actions termed a “ritual” is viewed too heterogeneously. Another is that the recent term “ritual” is seen as too woolly – “[There is] the widest possible disagreement as to how the word ritual should be understood”15 – at times as misleading, and even as useless. Further,

11 12 13 14 15

Michaels 1999; Lawson & McCauley 1990: 56. E.g. Wiedenmann 1991: 201ff.; von Ins 2001. See for instance Werlen 1984; Klein 1987; Grainger 1974; Harth 2005. Werlen 1984: 35, my translation. Leach 1968–72: 526.

10

Axel Michaels

difficulties in demarcation between Religious Studies and Linguistics may have led to mutual reservations about this matter. Certainly, non-verbal communication obeys different criteria to language, and it is hard if not impossible to pinpoint and delineate actions in the same clear way as words and sentences. This problem, which could only lead to a general theory of action by the inclusion of cognitive theories, will only be touched on here and not tackled any further. This would first require additional findings as to whether rituals genuinely constitute a unique form of action. But the answer to this question depends crucially on whether one can name enough mutually distinguishable criteria. And these criteria for their part are nothing other than the elements of a possible grammar of rituals. So the following will be concerned with compiling a list of just such criteria, in analogy to linguistic models, i.e. making a start at mapping out the morphology, syntax, pragmatics and semantics. Whether this will lead us to developing a deep structure or a deep grammar of rituals is anything but certain. But there is a reasonable suspicion, which has been expressed often enough, that ritual actions may well conform to a universal structure.

2. “Morphology” or the Building Blocks of Rituals Rituals are pieced together – from repeatable sections and sequencess of actions, episodes, sub-rituals or rites, and much more. To this come special ritual utensils, times, places, decor and such like. So it is not a question of single actions but of complexes of actions, i.e. lots of actions with at times lots of actors, which may extend over a long period. I shall refer to the smallest units of rituals as “ritual elements”. It may be difficult at times to distinguish these elements, but the fact that rituals can be taught and have a design shows that this is not a fundamental problem.16 Because instruction in rituals consists in nothing else than the composition of ritual elements. Even if in some cases it is a difficult and often hotly debated question as to what constitute the smallest or self-contained units, “the ‘atoms’, as it were” and the “smallest conveyors of meaning in ritual behaviour”17, ritual manuals and instructions are often nothing more than lists of brief ritual acts, ritual utensils, times, places, and decor. The discreteness of the ritual elements contrasts strongly, as Claude Lévi-Strauss has noted, with myths, which tend to make clear boundaries difficult: “The fluidity of the real is such that it constantly tends to escape through the mesh of the grid that mythic thought has placed over it so as to bring out only its most contrasting features. Ritual, by fragmenting operations and re16 Cf. Gladigow 2004: 76; Grimes 2000. 17 Lang 1998: 444–5.


11

peating them unwearyingly in infinite detail, takes upon itself the laborious task of patching up hole and stopping gaps. (…) This desperate, and inevitably unsuccessful, attempt to re-establish the continuity of lived experience, segmented through the schematism by which mythic speculation has replaced it, is the essence of ritual.”18 The individual steps have also been termed the building bricks of rituals. Michael Oppitz, for instance, uses this metaphor of building and analyses it along the lines of a “construction plan”: – Rituals are pieced together – from individual units – The units of rituals are prefabricated structural elements – The prefabricated structural elements are mobile and transposable – The transposition of structural elements follow a design – The prefabricated parts pieced together according to the design produce a recognisable product – the ritual in question. “The prefabricated structural elements or units from which rituals are made are the building blocks with which they are composed. They come simultaneously from a number of levels: a material level in the form of certain objects required to conduct the ritual at a set time and at a designated place; a linguistic level of ready formulated utterances such as prayers, magical formulae or recited myths; an acoustic level in a broader sense, with musical and other sonic means of expression; and a kinetic level with special actions, movements and gestures.”19 Here Oppitz assigns to the material level such things as the individual objects that either are manufactured for the ritual in question, or receive a specific meaning from it. The acoustic level includes various rhythms and sequences of melodies produced by ritual instruments. The building blocks on the linguistic level include, according to Oppitz, “isolated words, phrases and turns of speech or whole text sequences from different genres, with vocabularies that may come from everyday speech or that even occur within the ritual framework (secret languages, ritual languages, special archaisms).”20 The building blocks on the kinetic level may include gestures, postural attitudes, dance figures, or even whole sequences of actions. It is clearly an important prerequisite for a ritual “morphology” that the individual ritual elements are delineated – just like building blocks. Burkhard Gladigow among others pointed to this when he termed the elements that form the lexicon of 18 Lévi-Strauss 1981: 674–675. 19 Oppitz 1999: 73, my translation. 20 Ibid.

12

Axel Michaels

a ritual “ritemes”.21 Only then can the ritual elements become distinguishable and recognisable and thus repeatable. The parts of a ritual must be discrete and easily perused so as to allow the possibility of notation or scripts, ritual manuals and liturgical compendia. And this goes hand in hand with the possibility of abbreviations, citations, transpositions or rearrangements, or in short an analysis of syntactical structure (see section 3 below). The ability to demarcate the ritual elements and the ritual as a whole is achieved through authority, age, orthodoxy or extraordinariness. The determination of or limitations placed on the meanings of the ritual elements arise through reference to precisely these constants. An everyday action can be sanctified and made into a special, extraordinary, canonical, traditional or in fact ritual action by regarding it as the repetition of an action that has been authorised since time immemorial by (religious) specialists (e.g. priests), ancestors, or deities, or by reference to either these persons or to especially sanctified values. Since (almost) all ritual operations are actions that also appear in non-ritual contexts, these criteria refer to the mode of action. Thus the action of “pouring water over a statue” becomes the ritual act of “blessing or consecrating a statue” by referring to the aforementioned constants. Were it not for such distinctions, the discussion about rituals would be meaningless. And yet the question remains as to what are the smallest or the self-contained units of a ritual. In keeping with the possible generativity of rituals and thus their potential grammaticality, it is necessary, although probably not possible, to compile a finite quantity of ritual elements – the ritual inventory – in such a way that, together with an also finite number of formal rules, it produces an infinite number of rituals and thus allows new rituals as well as deviations within the ritual involved. Only then will we have a formalism that does not rule out variants and dynamics, but that permits them in a circumscribed manner focused on the criteria we have named: “authority, age, orthodoxy or extraordinariness”.22 At the same time, this stops the use of the concept of ritual getting out of hand. In the case of a phenomenological “morphology” of rituals, among other things one would now work out which are the smallest units or elements that can be repeated and transposed, but not the spontaneous or arbitrary components (although they also constitute part of the rituals in which they appear). So in most cases we would be dealing with elements that are premeditated, prefabricated, archived and recorded in scripts or ritual manuals. The possibility of repeating rituals or ritual elements can also relate to the language used in the ritual and with that, as Lévi-Strauss has correctly remarked, be excessive and boundless: 21 Gladigow 2004: 59. 22 Essentially these criteria correspond with those I termed “religio” in earlier definitions of ritual: see Michaels 1999, 2003 and 2003a.


13

“(…) at the cost of considerable verbal expenditure, it goes in for a riot of repetition: the same formula, or formulae similar in syntax or assonance, are repeated at short intervals, and are only operative, as it were, by the dozen; the same formula must be repeated a great many times running, or alternatively, a sentence containing a very slight meaning is sandwiched, and almost concealed, between accumulations of identical and meaningless formulae.”23 More specifically, the ritual elements can refer to the following realms. The square brackets contain examples from Hindu rites de passage for initiations (Sanskrit upanayana, vratabandha: see Michaels 2004:71–158): – Agency: choice and appointment of ritual specialists [Brahmins, barbers, cooks], invitation of the ritual participants – Body: festive or new garments [men: suites, women: red saris], fancy costumes, masks, veils, changing or taking off shoes, putting on jewels and finery (necklaces, rings, feathers, make-up), headwear (hats, crowns, laurel wreaths), changes in hair (tonsures, new hairdos, wigs), ablutions [snāna, samavārtana], anointment [abhiṣeka] or injury to body (wounds, tatoos, circumcision), out-of-the-ordinary movements (change of location, processions, circumambulations [pradakṣiṇā], practiced steps, parades, escorts) – Language and gestures: formulaic resolution to perform a ritual [saṃkalpa: cf. Michaels 2000 and 2005], (increased) literacy (minutes, documents, certificates, ritual texts [Vratabandhapaddhati], myths), ceremonious speeches in restricted codes (addresses, singing, reciting texts, soliciting blessings, prayers, supplications; new names [nāmakaraṇa], [use of mantras in ritual actions]) or silence – Decor: decorated, purified arena (temple, place of sacrifice, sacred sites [yajña, maṇḍapa]), flowers, fire, incense, light, candles; acoustic means: music, drums or bells – Framing: setting special times (leisure time, working hours, evenings, determined according to horoscope or astral sciences) – Material: involvement of special offerings or utensils: paraphernalia, ritual objects and instruments; food for shared meals, hospitality or offerings. All of these ritual elements can be transposed and may appear in different contexts, but they are also distinguished by authority (Mantra), age (“Vedic”), orthodoxy (led or conducted by Brahminical priests) and extraordinariness (determina-

23 Lévi-Strauss 1981: 673.

14

Axel Michaels

tion of the point in time by astral sciences, marking out the sacred arena, and the like). The problem with any morphology of ritual is that if units of action are taken as the smallest building blocks, this does not correspond to the morphemes of a language, but at most to the sentences and thus more to a syntactic structure. This is a point not lost on Lawson & McCauley: “Returning, then, to the analogy with linguistics, it is the action that is the analogue of the sentence (which is the fundamental unit of linguistic analysis).”24 Presumably it will be difficult if not impossible to determine the smallest units for every ritual. What does seem likely, though, is that we can work out a basic stock of delimitable (“citable”, as it were) ritual elements within a more or less closed cultural framework using rituals that are liturgically standardised, or that extend over a lengthy period of time that has to be determined. This would have to be performed at the beginning of a ritual analysis. It was done for instance in exemplary fashion by Yasuhiro Tsuchiyama when he demonstrated that the chief significance of the old Indian consecration rite (abhiṣeka) in the Vedic sources lies in the transfer of power and glory (varcas) to the offerer.25 The term also retained this significance when it was transferred to the consecration of statues. In more recent ritual texts, the Gṛhyasūtras, abhiṣeka came however to be mingled with the ritual bath (snāna) and consequently was no longer distinguishable as a Vedic ritual element in its own right.

3. “Syntax” or the Composition of Rituals While the “ritual morphology” remains largely on the surface and in the realm of phenomenology, the search for connections within the ritual elements leads almost inevitably to a deep structure, assuming the hypothesis that there is a universal mode of action for rituals. Evidence for such a deep structure could be produced empirically, such as by developmental studies with children and youngsters, or by transcultural comparisons. This work can already be seen in its inception, albeit only a very small scale. The psychologist Melanie Gail Jacobs, for instance, showed in her dissertation in 2005 that only on attaining school age are children able to understand rituals as rituals, and to regard it as necessary to perform their like in certain situations. For this they must have a grasp of authority, age and causality. Likewise Lawson & McCauley assume an intuitive competence, an implicit or “tacit knowledge” for the ability to recognise rituals as such: “Although this knowledge for some individuals may prove to be exclusively tacit, they demonstrate their knowledge, nonetheless, through their intuitions 24 Lawson & McCauley 1990: 84. 25 Tsuchiyama 2005: 51–94.


15

about the form of rituals and their successful participation in them. Participants in rituals who are unable to formulate explicitly even a single rule that governs their ritual system still have many, if not most, of the requisite intuitions about ritual form. (Similarly, many native speakers cannot state even a single rule of their grammars.)”26 This ritual competence, as I shall call it, is acquired during childhood without any explicit instructions, like a sense of tact. And the same applies to the acquisition of language and grammar.27 Only after these ritual elements, as such polyvalent and in this way meaningless, are related to one another does the ritual emerge as a whole. Each ritual element, each detail must be understood as part of an entire ritual. But the entire ritual consists solely of the syntactical connection of the ritual elements. Syntax (Greek σύνταξις) means “putting in order, together”, and in a ritual the ritual elements are put together into larger units and sequences, partly, it may be presumed according to universal patterns and rules. To speak with Oppitz: the syntax of rituals is their “construction plan”.28 The first more rigorous attempt to chart the syntax of rituals was undertaken by the Indologist Frits Staal.29 He went so far as to claim that ritual should be studied not as religion, but as syntax without semantics and semiotics (see section 5 below), which is to say in purely formal terms. In fact Staal simply arrives at an analysis of the various possibilities for sequencing ritual elements; on numerous occasions he refers to theories of composition in music. Taking the old Indian fire ritual (Agnicayana), he demonstrates various recurring methods for arranging ritual elements. To begin with he divides the complex ritual into smaller units (A, B, C ...) so as then to tease out a number of structures. One of these is embedding or framing (“embedding”: B → ABA); this syntactical structure can recursively yield similar structures AABAA, AAABAAA etc. Thus the construction of a layer of the Agnicayana altar (A) is embedded in two ritual sequences: a) pravargya, i.e. offerings of hot milk to the Aśvin deities (= B), and b) upasad (a ritual battle against demons = C). A customary sequence would thus be BCABC, but since BC also forms a unit say B’ in this case we have the basic structure B’AB’. However, these ritual elements might also appear as follows: pravargya – (upasad – altar layer – upasad) – pravargya (BCACB); but once again, according to Staal, CAC form a unit say A’, so the whole can be reduced once again to BA’B. Staal presents further “syntactical” structures along these lines: e.g. inversion (ABA, BAB), insertion (BC → BAC), mirroring (ABA1, 26 27 28 29

Lawson & McCauley 1990: 77; cf. also 2002. Lawson & McCauley 1990: 78. Oppitz 1999; see also Kohl 2006. Staal 1989.

16

Axel Michaels

AABA1A2) and serial sequences (A1A2A3...). But despite the limited range of patterns in these ritual sequences, Staal considers the complex Vedic ritual to be unlimited in its ritual possibilities.30 Staal’s syntactical theory of rituals is based on the theory of their meaninglessness and is influenced by generative transformational grammar. His point of departure is that rituals do not reflect meanings and point to something, just as words do not simply reproduce reality. Hans H. Penner has shown, however, that Staal’s understanding of language and ritual is unsatisfactory.31 He quite rightly points out that “language as we all know is composed of signs, and all linguistic signs have phonological, syntactic and semantic components”.32 So if rituals are supposed to have a syntax, they must also have semantics because the two cannot be separated from one another. Syntax means the combination of signs, and signs are always pointers to something; that is what gives them their meaning. “Staal ... does not argue that rituals are not semiological systems. On the contrary, he argues that rituals have a syntax, but they are meaningless. Given the ... evidence from linguistics, Staal’s position is simply wrong”.33 With greater rigour even than Staal, E. Thomas Lawson and Robert N. McCauley have developed a syntax of rituals that comes the closest to date to a “grammar” of rituals, particularly since they make precisely this claim while likewise drawing on Chomsky.34 To begin with they note “striking similarities between speaker-listeners’ knowledge of their languages and participants’ knowledge of their religious ritual systems”: “Both languages and religious ritual systems are examples of what we call ‘symbolic-cultural systems’”,35 and in their book Rethinking Religion they go on the assumption “that a system of rules much like the grammar of a natural language can account for the patterned structure of religious ritual”.36 What they actually develop, as they freely admit, is more a theory of ritual competence than of ritual behaviour.37 Their abstract and complex theory, which can only briefly be summarised here, is made up of four parts:38 (1) a system for representing actions (“Action Representation System, ARS”), containing the underlying syntactic components and the formative rules that allow symbolic categories to be realised by action elements; (2) religious concepts (“Religious Conceptual Scheme”) penetrate the ARS, 30 31 32 33 34 35 36 37 38

Staal 1989: 91. Penner 1985. Ibid.: 9. Ibid.: 11. Lawson & McCauley 1990: 78. McCauley & Lawson 2002: 4. Lawson & McCauley 1990: 56. Ibid.: 77. See fig. 1 or Lawson & McCauley 1990: 103


17

charging the ARS as it were with semantic and cultural components; (3) a set of “universal principles of religious ritual”, such as a principle of “superhuman agency” or corresponding agents (culturally postulated superhuman agents, CPSagents),39 which underlies all religious rituals, or a principle according to which religious rituals always demand an object;40 and (4) a kind of self-control within the system (“Assessment of the Status of the Represented Religious Action”), in which the Object Agency Filter – a filter that makes rituals into universals but not innate actions – also plays a part. Fig. 1: Lawson & McCauley’s Ritual Theory

39 In their most recent publications Lawson & McCauley have altered their terminology: the CPS-agents are now called CI-agents, which is to say “agents possessing counter-intuitive properties”. The reason for this is given in McCauley & Lawson 2007: fn. 7: “We abandon that usage [of CPS], since cultures are not the sort of things that postulate anything, so far as we can tell.” 40 Prayer according to Lawson & McCauley is a religious actions but not a ritual (McCauley & Lawson 2002: 13).

18

Axel Michaels

With this scheme, Lawson & McCauley dissect rituals into their component parts and analyse them in structural “formation trees”, which are openly based on linguistic sentence analyses. Apart from the possibility of empirically testing and formalising rituals, this theory is particularly interesting due to the question it poses about the formative rules of a ritual. A number of examples can be named of just such formative, with some reservations “syntactical” rules, with which one may glean the dependence of the parts on the whole or on one another:41 – Repetition: also including imitation – Reduplication: the doubling of ritual elements, partly interrupted by other ritual sequences, is a common structural element. – Seriality: ritual elements recurring in sequences that can also spread to other rituals. – Substitution: the replacement of one ritual element by another viewed as equal in value (cf. also McCauley/Lawson n.d.: 16). – Option: the optional or alternative employment of a number of ritual elements viewed as equal in value. – Transformation: the temporal staggering or interpolation of ritual elements. – Fusion: the merging of two or more different ritual elements. – Reduction: abbreviations of the combinations of ritual elements. – Omission: the elision of stipulated ritual elements, more the rule than the exception. – Transfer: transferring ritual elements to another ritual, e.g. as a ritual quote. – Intermissions: staged interruptions in the succession of ritual sequences. – Framing: the emphatic commencement and emphatic end of a ritual. – “Principle of superhuman immediacy”: Lawson/McCauley state that this is a universal principle according to which those ritual elements in which “superhuman agency” is demonstrably involved are more central than those in which this agent is not the direct protagonist. These and other “syntactical” structures are internalised by the ritual participants like the syntax of a language. Even if individuals are unable perhaps to explain the rules of the rituals, they nevertheless know when something has gone astray or is wrong.42 As is echoed by McCauley & Lawson: “Although the performers of the ritual may not be able to explain the rules, they do know what is to be done when 41 For examples from the Indian ritual context see Michaels & Buss, this volume. 42 Cf. Hüsken 2006.


19

the rules are broken. The parallel between ritual and language at this point is quite striking.”43 Just as was the case with morphology, the metaphor of syntax for rituals soon reaches its limits. Because if ritual elements are already defined essentially as “sentences”, the syntax in the rituals basically revolves round textual linguistics, which is to say the relationships between the sentences.

4. “Pragmatics” or the Agency of Rituals If rituals can be grasped under certain premises as a kind of special language, then they also require not only ritual competence but also communicative competence. This means that factors must be included such as their aptness to a situation, the relationship between the ritual participants and observers, and the cultural context. Moreover, for religious rituals it is significant that the addressee is mostly a supernatural being, so that the communicative situation occurs with answers and responses that are not directly perceivable, or only so to the participants. Of the greatest importance to a ritual are the performative utterances that go to make it up, so that it is no longer simply a communiqué but also an action. “The crucial point is that religious rituals (despite their various unusual qualities) are actions too”.44 Pertinent here are for instance those (speech) acts in which everyday actions are turned into ritual actions by explicit explanations. When for instance an express declaration is made to remain awake the whole night long, a vigil becomes a ritual action. But anyone who remains awake without this express declaration does not usually benefit from the vigil. Humphrey & Laidlaw have also stated that this “ritual (self-)commitment” is an important criterion of rituals.45 Almost every Hindu ritual requires for instance a ceremonial resolution to undertake the ritual (saṃkalpa), and this must meet certain formal criteria.46 Thus for instance the resolution must be uttered consciously and verbally before the actual ritual begins. In addition the time and place of the ritual has to be named precisely: continent, land, district, village, river, etc.; era, year, month, lunar and solar day, hour and minute, etc. The ritual participant is given as it were a sacred location and thus removed from everyday life. It is the saṃkalpa that first makes an action a ritual action. For this reason it is normally accompanied by a change in the style of speech. In a ritual, flowers become ritual flowers, water becomes ritual water, rice ritual rice, a rock a divine throne. All this is largely accomplished by the change from vernacular, say, to Sanskrit. 43 44 45 46

McCauley & Lawson n.d.: 11. Ibid.: 9. Humphrey & Laidlaw 1994. The following draws partly on Michaels 2000; see also Michaels 2005.

20

Axel Michaels

With these formal structures, the saṃkalpa reveals a pragmatical and syntactical congruence with speech acts, as formulated by John L. Austin (1962) and John R. Searle (1969). According to them, a speech act is a verbal and performative utterance that effects something in the moment it is carried out, even if it relates to future actions. As Searle47 has explained with reference to promises, an illocutionary act of this kind has to fulfil certain non-linguistic criteria apart from a certain linguistic form. For instance, the promised action must not yet have been performed, the listener must value the action, and the action would not simply be done regardless. Promising one will give someone else a clip round the ear when one has already done so is not a particularly successful promise. By not merely asking whether sentences are true but also whether they prove successful, Searle introduced a decisive turn in the analysis of language and with that of ritual. Much the same is true of a promissory act, such as an oath, vow, pledge, troth, a solemn declaration or assurance, a curse, truth spells, a confession of faith (credo), or indeed saṃkalpas. In most cases promissory speech acts are connected with religious affirmations. With regard to rituals, I subsume promissory acts and signals together under the term intentio solemnis and refer by that to every act that makes an everyday action a ritual one. Put differently and more simply: spontaneous, chance, arbitrary actions are not rituals, while actions performed with an intentio solemnis can be. The concept of agency is also special when it comes to the pragmatics of ritu48 als , because many rituals also assume the involvement of supernatural agency, and groups or “corporative subjects (families, clans, brotherhoods, ritual circles, military units, etc.) can also appear in ritual contexts as ‘agents’”.49 Agency can in that case be described as a “transformative power to act” or “competence for action”. In the field of pragmatics I would also finally include the gestural, decorative and acoustic parts of rituals. For the musical realm the following units can more or less be applied and, after suitably adapted, can also be used for dance in rituals or other realms of ritual pragmatics:50 – Embodiment: ritual music must not only be heard but also embedded within a field of action. The occasion and the space (temple, procession etc.) are fixed. But often it is not unimportant who plays the music, because it is not a question of music alone. Rather, specific persons have to play specific pieces of music on specific instruments at specific places.

47 48 49 50

Searle 1969: 55–71. Here in more depth Krüger & Nijhawan & Stavrianopoulou 2005. Ibid.: 6. Michaels 2005a.


21

– Formalism: repetition is a part of ritual, which lives from imitation and mimesis. The rhythm, especially of the percussion instruments, draws the (body of the) ritual participant into the rhythm of the ritual. But this rhythm should not merely be grasped in acoustic terms, but also as a specific rhythmicity produced by the number, kind, sequence and repetition of the ritual sequences. This creates a certain sensitivity to rituals – both rhythmic and moral. – Framing: rituals or ritual sequences are often announced by bells, gongs, cymbals, tympani, or even simply by tapping a wine glass. In this way one knows not only when a ritual begins, but also when it ends or a decisive change is about to occur. – Transformation and effect: music in rituals does not merely create a sound, it also has an effect – and not merely aesthetic, but also performative. It meets with an inner response. The listeners get drawn in to the ritual music, not least through their rhythmic body movements. This music is directed to the community, which communicates by means of joint action (and not only through language or images). Repertoire and instruments are often considered to be blessed or even to have been given by the gods. They promise fortune, success, or the pleasure of the gods and thus the efficacy of the ritual actions. – Elevation: ritual music transports into another time and another world, and is thus regarded as sounds from that world. In addition to the religious or transcendentalising principles in illo loco and in illo tempore one could add something like in illo sonore. Ritual music is sacred or mythical music, so what applies to mythical spaces and times also applies to it.51 Thus ritual music and ritual are not complementaries, but a whole, such that music is not only in time and space but also beyond. It is not something in the ritual or during the ritual, but part of the overall power of the ritual.

5. “Semantics” or the Meaning of Meaninglessness Frits Staal in particular forwarded the thesis of the meaninglessness of rituals. And Humphrey & Laidlaw52 as well as McCauley & Lawson53 have expressed much the same: “We think that much about religious rituals’ forms are overwhelmingly independent of meaning.” Staal refers among other things to Van Gennep’s observa51 See in this context Hübner 1985, as well as Michaels 1994: 314–346 and 2005a (with additional evidence). 52 Humphrey & Laidlaw 1994. 53 McCauley & Lawson n.d.: 9.

22

Axel Michaels

tion that “the same rite, remaining absolutely the same, can change its meaning depending on the position it is given in a ceremony, or on whether it is part of one ceremony or another. The aspersion rite ... is a fecundity rite in marriage ceremonies, but an expulsion rite in separation ceremonies.”54 Staal concludes from this that “it is only a small step from ‘changing meaning’ to: ‘no intrinsic meaning’ and ‘structural meaning’, and from there to: ‘no meaning’”.55 Staal does not deny that rituals may have “useful side effects”, such as pleasure, the allaying of fears, social unity or changes in status, but these should not be mistaken for the goal or as functions.56 If, according to Staal’s theory, rituals fulfil a specific purpose, other means that can fulfil this even better must be permissible. But mostly this is not the case. It is not a matter of indifference how light is produced; it will be set down that it must be from a candle or a blazing torch, which cannot simply be replaced by a fluorescent tube. So in rituals it is not the goal that determines the means, but vice versa. For this reason rituals could have a number of meanings, but are as such meaningless: “The meaninglessness of it explains the variety of meaning attached to it.”57 According to Staal, it is religion that first gave rituals their meanings. In his view, religion is the equivalent of ritual plus meaning: “The chief provider of meaning being religion, ritual became involved with religion and, through this association, meaningful.”58 Staal nevertheless regards ritual as a kind of anti-religion and as an autonomous practice with its own rules. This is why for him ritual should not be studied as a religion but as syntax without semantics. So Staal wants to study ritual completely “in itself and for itself”. With this he contradicts the idea that rituals are symbolic systems of reference that point to a different (e.g. higher) reality. Rituals are indeed fabricated, because they are not only instrumental operations tailored to a specific purpose (e.g. encountering a god, allaying fears, or heightening the feeling of community), or expressive (e.g. ludic or theatrical) actions, but also – irrespective of the motives connected with them – are also intended as stagings of a freedom from purpose and intention. There are ethological and neuropsychological indications that rituals are often stagings of immutability precisely because a primary or archetypal state is associated with them: what was good once should remain so. 54 Staal 1989: 128. 55 Ibid.: 134 56 Staal 1979: 11. See also Goody: “By ritual we refer to a category of standardized behaviour (custom) in which the relationship between means and ends is not ‘intrinsic’, i.e. either irrational or non-rational.” (Goody 1961: 159). 57 Staal 1979: 12. 58 Staal 1989: 137.


23

This also allows us to say that for this reason, many rituals are rigid, stereotypical and largely unchangeable because they are directed against change and changeability. This is why they are removed to some extent from human amenability. Rituals preserve what is old, which cannot be touched. In this way acquired knowledge and abilities are also preserved, without their usefulness forever having to be legitimated anew; they can be more or less summoned on demand and are available as techniques in whatever new context might call for them. Inasmuch as people in (religious) rituals stage timelessness and constancy, they counter in some respects the uncertainties of the new and the future, of contingency and the transition from life to death. Unlike Staal, I do not view the meaninglessness of rituals negatively, but rather as an expression of this staging of constancy. For Staal, rituals are a regression to ancient, prehistoric times, to the behaviour of animals or little children, and so he shares the view of many that rituals obstruct progress. But to my mind there must be some benefit in preserving rituals and their meaninglessness. And I see this as lying in the way they allow people to deal with both the potential for change and the necessity of preservation, with revolution and tradition. This possibility of overthrowing everything without simultaneously forgetting all one has learnt from the past is exclusive to human beings. It is hard to learn something new, but even more so to forget the old. There where people consider progress unnecessary, because nothing can be better than it already is, where they have no wish for no change because everything is already good enough, where they recoup, as Kurt Hübner says, the old and original by repeating it, there especially ritual is the means for choice: it works of its own accord, ex opere operato, by dint of its own non-referential effectiveness, but not because it has a meaning or remains limited to the one that people attach to it cognitively.59 Which is why I say once again60 and almost ritualistically: rituals are meaningless, but that is not without its own meaning.

6. Conclusions As we have seen, the possibilities for a grammar of rituals, if too narrowly conceived, are limited. The difficulties as regard the “morphology” of rituals lie chiefly in determining the smallest units of a ritual. It is easier to demonstrate these in highly liturgical and thus formalised rituals than in expressively ritualised performances, but it will scarcely be possible to draw up a universal morphology. Similarly the study of the “syntax” of rituals will result at most in possibilities of teasing out comparable structures, while a deep structure of some kind will be difficult if not impossible to reveal. The “pragmatics” of rituals, on the other hand, 59 Hübner 1985: 142. 60 See Michaels 1999: 40–45, 2003a and 2006, on which this section has drawn.

24

Axel Michaels

constitutes an area of its own, in some ways a genuine field of ritual studies. It will be important to tie in here with semiotic theories, and in particular with theories of agency and performance. The “semantics” of rituals is the most controversial topic, not least because religious convictions often play a part here. Independent of the question of whether rituals are meaningless or not, it can nevertheless be said that as a rule, rituals are polysemantic (and in this sense “meaningless”). It will be decisive for the development of ritology or ritual studies in the closer sense that common structures are elicited from as many rituals as possible. This will only be possible once conventions have been developed for the notation of rituals and ritual elements. On the road to this goal, the analogy between a “grammar” of rituals and a grammar of language is useful. But for the moment the main concern is to describe rituals as precisely as possible in their relevant context, and to record them with the means provided by other media, i.e. to segment them and present them in their all manifold pragmatic and semiotic, as well as culturally and socially significant connections. This, however, would mean that we would have to develop a standard for describing rituals. It seems possible to develop a “condensed ritual description” that will be pre-processed using computational linguistic techniques although this means that we would have to prefer an analysis that is strictly limited to linguistic approaches. This is the approach being followed by an ongoing project in “Ritual Dynamics” where we do not analyse anything beyond language for even a nonverbal action depends on somebody who describes this action. We intend to apply methods of shallow linguistic processing and general linguistic and ontological tagging using semantic lexicon resources, i.e. role-semantic lexica together with semantic relations for time and space. The ontology-based semantic annotation of rituals must be employed to evolve a refined model for the formal representation of ritual structures. Here then logical methods of ontology-based reasoning will be employed in order to detect and formally represent generalisations over ritual descriptions, and to capture possible variations across rituals. We will explore the use of the models obtained to detect trans-cultural variants of rituals. Conceptually annotated ritual descriptions from different cultures can be investigated automatically to bring to bear differences and similarities in the underlying structure of rituals, such as: variations in event sequences, the conceptual types of ritual participants, or the inventory and characteristics of ritual markers. Comparative quantitative analyses can be used to capture the specificity, and thus characteristics of ritual actions: by quantifying, for example, the strength of selectional preferences of placing events (e.g. how often is rice the object of a placing event) in ordinary text, as opposed to ritual descriptions. I am not saying that this is the only way to proceed in developing a grammar of ritual, but it would be worthwhile to help with understanding the rules by which rituals are governed. For a grammar of ritual would be nothing other than a set of


25

rules for rituals, and Wittgenstein’s famous remark needs no more than a slight modification in this context (as here, inside the square brackets): “To obey a rule, to make a report, to given (give?) an order, to play a game of chess [or to do (enact?) a ritual] are customs (uses, institutions) [and rituals]. To understand a sentence [or a ritual element] means to understand a [ritual] language. To understand a [ritual] language means to be master of a technique.”61

61 Wittgenstein 1922: 199.

26

Axel Michaels

References Austin, John L. 1962. How to Do Things With Words. Cambridge (Mass.) Gladigow Burkhard (2004): “Sequenzierung von Riten und die Ordnung der Rituale“. In: Michael Stausberg (ed.). Zoroastrian Rituals in Context. Leiden, Boston: Brill: 57–76. Goody, Jack 1961. “Religion and ritual: the definitional problem”. British Journal of Sociology 12/1: 142–164. Grainger, Roger 1974. The Language of the Rite. London: Darton, Longman & Todd. Grimes, Ronald L. 2000. Deeply into the Bone: Re-inventing Rites of Passage. Berkeley et al.: University of California Press. Göhring, Heinz 1967. “Generative Grammatik und Kulturanthropologie”. Anthropos 62: 802–814. Harth, Dietrich 2005. “Rituale, Texte, Diskurse. Eine formtheoretische Betrachtung”. In: Burckhard Dücker & Hubert Roeder (eds.). Text und Ritual – Kulturwissenschaftliche Essays und Analysen von Sesostris bis Dada. Heidelberg: Synchron: 19–48. Hübner, Kurt 1985. Die Wahrheit des Mythos. Munich: C.H. Beck. Humphrey, Caroline & James Laidlaw 1994. The Archetypal Actions of Ritual. A theory of ritual illustrated by the Jain rite of worship. Oxford: Clarendon Press. Hüsken, Ute (ed.) 2006. Getting it Wrong? Mistakes, Failure and the Dynamics of Ritual. Leiden: Brill. von Ins, Jürg 2001. Der Rhythmus des Rituals. Grundlagen einer ethnologischen Ritualsemiotik, entwickelt am Beispiel des Ndëpp der Lebu (Senegal). Berlin: Reimer. Jacobs, Melanie Gail 2005. Transforming reality with ritual: Children’s understanding of ritual grammar and causality, PhD Diss. University of Pittburgh. http://etd. library.pitt. edu/ETD/available/etd-05062005-154010/unrestricted/mgjacobs_etd2005.4.pdf (15.4.2006). Klein, Wolfgang (ed.) 1987. Sprache und Ritual. Göttingen: Vandenhoek. Kohl, Karl-Heinz 2006. “Die Syntax von Ritualen”. In: Helwig Schmidt-Glintzer (ed.). Liturgie, Ritual, Frömmigkeit und die Dynamik symbolischer Ordnungen. Wiesbaden: Harrassowitz: 203–126. Krüger, Oliver & Michael Nijhawan & Eftychia Stavrianopoulou 2005. “Ritual” und “Agency”– Legitimation und Reflexvität ritueller Handlungen. Heidelberg, SFB 619 (= Forum Ritualdynamik, Nr. 14). http://archiv.ub.uni-heidelberg.de/volltextserver/ volltexte/2005/5785/pdf/AgencyAug05.pdf Lang, Bernhard (1998) “Ritual/Ritus”. In: Hubert Cancik & Burkhard Gladigow & KarlHeinz Kohl (eds.). Handbuch religionswissenschaftlicher Grundbegriffe, vol. 4. Stuttgart: Kohlhammer: 442–458. Langer, Susan K. 1992. Philosophy in a New Key. Cambridge: Harvard University Press. Lawson, E.Thomas 1976. “Ritual as language”. Religion 6: 123–139. Lawson, E.Thomas & Robert McCauley 1990. Rethinking Religion. Connecting Cognition and Culture. Cambridge: Cambridge University Press. — 2002. “The Cognitive Representation of Religious Ritual Form: A Theory of Participants’ Competence with Religious Ritual Systems”. In: Ilkka Pyysiäinen & Veikko


27

Anttonen (eds.). Current approaches in the cognitive science of religion. New York: Continuum: 153–176. Leach, Edmund 1968–72. “Ritual”. In: David L. Sills (ed.). International Encyclopedia of the Social Sciences, vol 13. New York: McMillan: 523. — 1976. Cultur and Communication. The logic by which symbols are connected. Cambridge: Cambridge University Press. Lévi-Strauss, Claude 1981. Mythologiques Volume Four. The Naked Man. Chicago: The University of Chicago Press. McCauley, Robert & E. Thomas Lawson 2002. Bringing Ritual to Mind. Psychological Foundations of Cultural Forms. Cambridge: Cambridge University Press. — 2007. “Cognition, Religious Ritual, and Archaeology”, http://www.scribd.com/ doc/16502688/Cognition-Religious-Ritual-And-Archaeology-McCauley-Lawson-2007. Michaels, Axel 1999. “‘Le rituel pour le rituel?’ oder Wie sinnlos sind Rituale?”. In: Corinna Carduff & Joanna Pfaff-Czarnecka (eds.). Rituale heute. Berlin: Reimer Verlag: 23–48. — 2000. “Ex opere operato: Zur Intentionalität promissorischer Akte in Ritualen”. In: Klaus Peter Köpping & Ursula Rao (eds.). Im Rausch des Rituals. Gestaltung und Transformation der Wirklichkeit in körperlicher Performanz. Münster: Lit-Verlag: 104–123. — 2001. “Die Religionsphänomenologie ist tot – Es lebe die Religionsphänomenologie (Nachwort)”. In: Axel Michaels & Daria Pezzoli-Olgiati & Fritz Stolz (eds.). Noch eine Chance für die Religionsphänomenologie? Bern: Peter Lang: 489–492. — 2003. “Zur Dynamik von Ritualkomplexen”. In: Dietrich Harth & Axel Michaels (eds.). Forum Ritualdynamik – Diskussionsbeiträge des SFB 619 der Ruprecht-KarlsUniversität Heidelberg, Heft 3. http://www.ub.uni-heidelberg.de/archiv/4583 (09 Oct 2009). — 2003a. “Inflation der Rituale? Grenzen eines vieldeutigen Begriffs”. Humanismus aktuell 13: 25–36. — 2004. Hinduism. Past and Present. Princeton: Princeton University Press. — 2005. “Saṃkalpa: The Beginnings of a Ritual”. In: Jörg Gengnagel & Ute Hüsken & Srilata Raman (eds.). Words and Deeds: Hindu and Buddhist Rituals in South Asia. Wiesbaden: Harrassowitz Verlag: 45–64. — 2005a. “Rituelle Klangräume”. In: Annette Landau & Claudia Emmenegger (eds.). Musik und Raum: Dimensionen im Gespräch. Zürich: Chronos: 33–44. — 2006. “Ritual and Meaning”. In: Jens Kreinath & Michael Stausberg & Jan Snoek (eds.). Theorizing Rituals, vol. 1. Leiden: Brill: 247–261. Oppitz, Michael 1999. “Montageplan von Ritualen”. In: Corina Caduff & Joanna PfaffCzarnecka (eds.). Rituale heute. Berlin: Reimer: 73–95. Penner, Hans H. 1985. “Language, Ritual, and Meaning”. Numen 32/1: 1–16. Searle, John 1969. Speech Acts: An Essay in the Philosophy of Language. London: Cambridge University Press. Staal, Frits 1979. “The Meaninglessness of Ritual”. Numen 26: 2–22.

28

Axel Michaels

— 1989. Rules without Meaning. Ritual, Mantras and the Human Sciences. New York: Peter Lang. Tsuchiyama, Yasuhiro 2005. “Abhiṣeka in the Vedic and post-Vedic Rituals”. In: Shingo Einoo & Jun Takashima (eds.). From Material to Deity. Indian Rituals of Consecration. New Delhi: Manohar: 51–94 (Japanese Studies on South Asia. 4). Werlen, Iwar 1984. Ritual und Sprache. Tübingen: Gunter Narr. Wiedenmann, Rainer E. 1991. Ritual und Sinntransformation. Berlin: Duncker & Humboldt. Wittgenstein. Ludwig Tractatus Logico-philosophicus, 1922. Transl. by C. K. Ogden (1922), prepared with assistance from G. E. Moore, F. P. Ramsey, and Wittgen-stein. London: Routledge & Kegan Paul.

Jan E.M. Houben

Formal Structure and Self-referential Loops in Vedic Ritual “If you continue to think in terms of the lower levels – individual ants – then you miss the forest for the trees. That’s just too microscopic a level, and when you think microscopically, you’re bound to miss some large scale features.” Dr. Anteater in Hofstadter 1999: 319

1. Introduction: Vedic Ritual and Ritual Theory Over several millennia, the Vedic ritual system has been of great direct and indirect importance in the cultural history of the Indian subcontinent and neighbouring areas. The study and criticism of Vedic ritual were at the basis of the development, more than two millennia ago, of disciplines such as grammar, phonology, and metrics;1 through the centuries and in fact till the present day Vedic ritual has provided symbols of social status and political power;2 it still provides a large number of prayers and ritual models in modern Hinduism,3 and it provided starting points for antagonism to Buddhism and Jainism, both of which retained several of its (ritual)

1 A few pointers, still valuable: Winternitz 1908: 229-231; 1922: 380ff.; Scharfe 1977: 77–87; in the emergence of linguistic disciplines problems of interpretation of sacred texts were a driving force: Houben 1997. 2 Currency in the form of coins commemorating the performance of a royal horse sacrifice or Aśvamedha, as in the case of the Gupta Aśvamedha coins (cf. Altekar 1954: 38–49; Raven 1994: 38 with note 2.95), is a most tangible illustration of the “currency” of these symbols of social status and political power. 3 As I could observe this applies, for instance, to modernised Hindu rituals in Mangal Karyalayas (ceremonial halls for marriage, Upanayana, etc.) and for rituals performed by female priests in Maharashtra (cf. Manjul 1997). “Change and continuity” (cf. Gonda 1965) are a perennial theme in the study of cultural and religious phenomena in India from the Vedic period onwards. Einoo in Einoo & Takashima (eds.) 2005 emphasises a Vedic-Hindu discontinuity in ritual which, however, presupposes continuity as well (cf. Houben 2007).

30

Jan E.M. Houben

structures, ideas, and presuppositions even when offering their criticism.4 As a phenomenon of extraordinary dimensions and unique configurations, it poses a major challenge both to indologists and to ritual scientists. After Hubert’s and Mauss’ study of the nature and function of sacrifice at the end of the nineteenth century,5 Frits Staal is the first scholar who made an extensive study of Vedic ritual and formulated on that basis a comprehensive theory of ritual that pretends to be valid beyond the confines of Vedic ritual itself.6 In the present study we will be especially interested in the formal structure of Vedic ritual, and start with an investigation of the Vedic Nihnava-rite, which was used by Staal as a major example of a fundamental meaninglessness of ritual in general and of Vedic ritual in particular. From the point of view of the history of the study of Vedic ritual, it is only natural to take Staal’s theory of ritual as a starting point. Also from a systematic point of view, in order to focus on the formal side of ritual, it is useful to take Staal’s supposed “meaninglessness” as starting point and to see next what minimal “meaning” we have to accept in an analysis of the formal structure of the ritual.7 Taking Staal’s “meaninglessness” as starting point places us from the outset in a theoretical framework conducive to the investigation of formal structure in ritual, which is only one of several aspects of ritual explored in several decades of ritual studies.8 This formal structure of ritual as understood normally 4 Both early Buddhist texts (e.g. Aṅguttaranikāya 2 and 4) and early Jaina texts (e.g. Uttarādhyayana-Sūtra chapter 25) engage in explaining who is a “real” Brahmin and what a “real” offering (yajña) is; in other words, they promulgate their own system partly by reinterpreting crucial terms in the Brahminical system. Also in terms of monastic rules there are important affinities and apparently close historical ties between Brahmanism, Buddhism and Jainism (cf. Bronkhorst 1993 and 1998; the argument of Jacobi is discussed in Houben 1999: 132 note 48; see now also Bronkhorst 2007). Moreover, it is well known that on a popular level many Vedic ideas continued in Buddhism and Jainism, for instance with regard to the Vedic god Indra who is not denied but given a new, significantly lower position in a new cosmology (for Indra and his heaven in early popular Buddhism cf. van Kooij 1989 and the references and illustrations given there). 5 Hubert & Mauss 1899. 6 Staal 1979; 1989. 7 We limit ourselves initially to the dimension of what Rappaport (1999) called the “canonical message” and leave out performative aspects which will have our attention in the second part of our argument. 8 In the recently growing interest in ritual studies, marked by the foundation of a specialised journal, the Journal of Ritual Studies in 1987, and the establishment of a major, multidisciplinary research project at Heidelberg University devoted to the “dynamics of ritual” in 2002, it has become more clear than ever that “[t]here is just no such thing as a theory-free description of ritual” (Snoek 2004: 98; cf. also Kreinath 2004). An interest in the formal structure of units and sub-units in rituals may be shared by several theoretical approaches, but others may tend to “look away” from such formal structure and focus on various other aspects. The emphasis on the formal structure of ritual of those whom Staal mentions as his predecessors, Henri Hubert and Marcel Mauss (1899; 1964), thus contrasted with an ap-

Formal Structure and Self-referential Loops in Vedic Ritual

31

pertains in fact to what can be called, as we will see below, the canonical dimension,9 which is the dimension for which it makes sense to ask whether or not, or to what extent, there is scope for a “grammatical” or “morphological” approach.10 We will then analyse self-referential loops that may occur within this formal structure of ritual. Practical and theoretical implications of the presence of self-referential loops will have our attention at the end.

2. Organisational Levels in Vedic Ritual According to Staal, “Ritual is pure activity, without meaning or goal”.11 Moreover, according to him, to “say that ritual is for its own sake is to say that it is meaningless, without function, aim or goal, or also that it constitutes its own aim or goal”.12 Nor are rituals to be understood in terms of the beliefs of those who perform them: “If ritual is prior to belief, as it happens to be in the scheme of evolution, it must be interpreted in different terms”.13 Staal argues that participants, ritual authorities and scholars can verbalise the meaning of rituals only in a very unsatisfactory manner or in mutually contradictory ways. “In the development of our concepts and theories of ritual it is only a small step from ‘changing meaning’ to: ‘no intrinsic meaning’ and ‘structural meaning’, and from there to: ‘no meaning’”.14 This argu-

9

10

11 12 13 14

proach such as the one of Frazer (1894; 1900) that emphasised the interpretation and attributed or supposed “original” meanings of rituals which were fitted into an evolutionary scheme from primitive to modern (a scheme that in outline goes back to Tylor 1871). For an overview of theories of ritual over around a century Bell 1997 provides a convenient starting point. The recent study by J.A.M. Snoek of the Indian Zoroastrian initiation ceremony “Navjote” (2004) is a brilliant example of what can be expected if a “classical” description of a ritual (in this case J.J. Modi’s description of the Navjote) is confronted with “modern” ritual theories (in this case, those of van Gennep, Eliade, Snoek and Bloch). The analysis of other formal structures underlying or linked to ritual may be useful as well, depending on the specific ritual under investigation. One could think of the structure of social relations or kinship relations – basically an aspect of the performance dimension of ritual (see below) – which were revealed as crucial, much more than any structure of concrete ritual acts amenable to a “syntactic” or “grammatical” approach, in the case of the Naven of the Iatmul on Papua New Guinea (Houseman & Severi 1994; 1998). This applies already to Hubert & Mauss. The main reason why Evans-Pritchard (ethnographer and functionalist anthropologist) in the early 1960s is interested in getting the work of Hubert & Mauss translated into English is not for its conclusions, which he finds “rather lame”, but because it presents “a study of the structure, or one might almost say the grammar, of the sacrificial rite” (E.E. Evans-Pritchard in Hubert and Mauss 1964, Foreword; my emphasis). On the theoretical background of the central challenge of “Grammar and Morphology of Ritual” see Michaels 2007. Staal 1989: 131. Ibid.: 131–132. Ibid.: 131. Ibid.: 134.

32

Jan E.M. Houben

ment is only another way of saying that in the approach to ritual advocated by Staal we should look exclusively at the ritual form and should completely neglect dimensions of ritual meaning.15 However, even if various ritualists may provide diverging interpretations, or even if it turns out to be impossible to verbalise satisfactorily the “meaning” of a complex ritual performance as a whole, elements and details may be indisputably “significant” in some sense. So much is admitted by Staal: “It might be more proper and fruitful to ask specific questions about the meaning of particular rites. Some such questions do receive specific answers, on which participants and scholars agree. The Yajamāna must keep his hands closed ‘like a child in the womb of its mother, ready to be reborn’.

15 Since its first presentation in 1979, Staal’s theory of meaningless ritual has had a reception that can be characterised as stormy. Critical reactions and reviews appear even on the basis of only the 1979 article (Penner 1985). Others discuss Staal’s view on meaningless ritual together with his 1983 monumental description of the Agnicayana ritual studied on the basis of a performance in Kerala (Knipe 1986; Schechner 1987b). The proceedings of a symposium (with, for instance, Strenski 1991) on Staal’s Rules Without Meaning (1989) appear in vol. 21 of Religion. Several critics of Staal try to formulate, each in their own diverging ways, why ritual is in fact meaningful. Bodewitz shows that Staal’s “ingenious argumentation” as presented in his 1979 article “does not prove the meaninglessness of this ritual” (1990: 9). Witzel (1992) demonstrates in detail the various dimensions of meaning (Vedic, indological, anthropological) of a Tantric Agnihotra ritual in Nepal of which he gives a structural analysis not in terms of Staal’s rules and trees but in terms of “frames” or “boxes.” In his 1979 article, Staal quoted Isadora Duncan “If I could tell you what it meant, there would be no point in dancing it” (1979 note 11). Far from being an indication of the meaninglessness of her dance, this rather points to its meaningfulness, as I observed in 1991 (Houben 1991: 9 note 13). Heesterman can accept Staal’s meaningless ritual as a characterisation not of all ritual of all times but at least of Vedic ritual as it is known to us on the basis of late classical texts whose performative tradition continues till today. “Vedic ritual [is] known to us as a closed, unchanging, and meaningless structure, a separate realm cut off from the lived-in world” (1993: 227). Houseman & Severi discuss Staal’s position, together with the theory of Lawson & McCauley (1990), as a major representative of “Syntactic schemes” in the study of rituals (1998: 183–192). Humphrey & Laidlaw (1994) arrive on the basis of Indian but non-Vedic material (Jain ritual worship) at a position somewhat similar to the one of Staal: ritual is basically meaningless. The observations of Verpoorten (1991: 169) on Sāyaṇa-Mādhava’s understanding of the contrast between daily life versus ritual indicate that a simple pragmatic relationship between act and function, and hence between intention and act, is often conspicuously absent in ritual, which would be in line both with Staal’s position and with the position of Humphrey & Laidlaw on meaningless ritual. After the 1990s, Staal’s theory has acquired the status of a position that is both widely known (among anthropologists, scholars in textual and religious studies, India and Asia specialists) and clear and unmistakable in its extremism, but which no one accepts in its full-fledged form. See also note 17.


33

The fire altar has the shape of a bird because fire, as well as Soma, were fetched from heaven by a bird. The priests do not go south if they can help it for the southern direction is inauspicious.”16 Staal’s denial of meaning apparently concerns the meanings attributed to the units of some level – that is, large complex rituals as a whole – while units of a lower level or lower order are in some cases accepted to have generally acknowledged “meanings”. Other ritual details are presented by Staal as an example of his main thesis of the meaninglessness of ritual.17 There are first of all Vedic utterances: here Staal mentions only very briefly the praiṣas (priestly commands within the context of the ritual) and their undeniable pragmatic meanings, before going on to elaborate on the meaninglessness of small elements, the stobhas (i.e. insertions that do not derive from the text of the source verse), in Sāmavedic chant.18 Apart from these small meaningless elements within Vedic chant that is otherwise based on (meaningful) Vedic verses, Staal’s main example of a meaningless rite or ritual episode is the so-called Nihnava-rite or nihnavanam which forms part of the Soma-sacrifice.19 Staal describes the Nihnava as follows: “This mysterious rite is performed twice daily, in the morning and afternoon, during several consecutive days. The ritual patron and several of his priests place their hands on a bundle of grass, which lies to the west of the altar. The position of the hands is different in the two surviving traditions […]”20 [next follow details on the placing of the hands, J.H.]. Elsewhere, the reader is informed how a young performer of Vedic ritual learned the Nihnava from his father. Remarkable about “this case of ritual trans16 Staal 1989: 116. 17 An overview of arguments regarding the meaninglessness or meaningfulness of rituals (see also note 15) appears in Axel Michaels’ “Ritual and Meaning,” in which the author argues that: “the meaninglessness of rituals only concerns the invariability of prescribed actions and the polysemy of rituals (that is, the multiplicity of meanings). Apart from that, rituals have a great variety of meanings and functions. The tradition of commentaries demonstrates the history of the meaning that was attached to rituals” (Michaels 2006a: 261; cf. also Michaels 2006b). What is clear, even from Michaels’ cursory overview of ritual theories (and others such as Bell 1997), is that not only ritual, but also ritual theory is a field that deserves detailed study of its dynamics, if not for its own sake then because “[t]here is just no such thing as a theoryfree description of ritual” (Snoek 2004: 98). 18 Staal 1989: 199–208. 19 For the sake of my argument I give here a very brief account of the Nihnava; a more elaborate one giving a more detailed overview of the relevant sources and in which the problem of the hand gestures is analysed, I presented at the World Sanskrit Conference in Helsinki (2003); it is to appear in the proceedings of the Vedic section (Houben forthcoming). 20 Staal 1989: xv.

34

Jan E.M. Houben

mission,” according to Staal, is that “it disregards not only meaning, but does not even use language.”21 If we regard the rejection of meaning as the negative side of Staal’s theoretical approach, there is also a positive side to it, viz. the study of ritual form on its own terms.22 However, is it possible to study the ritual forms properly and satisfactorily without any consideration to meanings attributed to them? This may have been Staal’s ideal, in accordance with a trend – now largely abandoned – in twentiethcentury linguistics to approach language without any reference to meaning.23 The problem in this approach is that the relation between (a) the complete rituals, which according to Staal are entirely meaningless, and (b) the units of lower organisational levels, which, as admitted by Staal, are occasionally meaningful, remains entirely unclear. The very determination of basic units remains unfounded if meaning is consistently neglected. As Lawson and McCauley aptly observed, Staal does not give us a motivated account of what his “ritual sequence” is a sequence of.24 Indeed, Staal takes those ritual episodes as units which are accepted as such by the tradition, not without often implicit but still traceable considerations of meaning and function. Let us accept for the moment that the detail about the placing of the hands in the Nihnava is “meaningless” (and is transmitted without reference to its meaning). Can, then, the Nihnava as a whole be regarded as equally “meaningless”? The rite has, to begin with, a name. It is true that translators of ritual texts have often rendered this name, nihnava or nihnavana, in a vague or too general way.25 But dictionaries do record and illustrate a quite specific meaning for the verb nihnu

21 Ibid.: 187. 22 For instance, Staal 1980; 1982; 1983. 23 Leonard Bloomfield evinced a “rather pessimistic attitude towards semantics” (Robins 1979: 208) and “Bloomfieldian” linguists from the 1930s to ca. 1957 preferred to concentrate on the formal analysis of language (Robins 1979: 209). In the article which Frits Staal wrote together with Paul Kiparsky (Kiparsky & Staal 1969), semantics occupied an important place in their model of Pāṇini’s Aṣṭādhyāyī as a generative grammar. Parallel with Chomsky gradually evading a “semantic” interpretation of his “deep structure” in language (proposed by, among others, Chomsky’s former student Lakoff 1971), Frits Staal gradually lost his interest in semantics – a sequel to his “Indian Semantics I” (1966) was never published – and, as far as ritual is concerned, averse to attempts to make a scientific study of meaning. For semantics and other aspects of modern linguistic theory at the background of the study of Indian grammar see Houben 2009a. 24 Lawson & McCauley 1990: 59. 25 Eggeling, for instance, in his translation of the Śatapatha-Brāhmaṇa, translates nihnuvate as “to make amends”. Caland translates atha nihnuvate (ĀpŚS 11.1.12) as “Dann halten sie Abbitte”. Brough 1950 arrives at the following meaning for the verb ni-hnu: “to conceal one’s actions from a person, so as to avoid arousing his anger.”


35

and derived nouns such as nihnava, nihnavana and nihnuti;26 the verb may be translated as “to deny”, to “disavow”, “dissimulate”;27 nihnava or nihnavana may therefore be variously paraphrased as: – to distract the attention from a wrong act by means of nice words or gestures; – to hide a big crime with a small nice act; – to beat about the bush. The name Nihnava or nihnavana thus says something about the meaning or expected function of this rite. Second, its sense is explained in the Śatapatha-Brāhmaṇa (ŚB 3.4.3.19–21) and elsewhere.28 The explanation presupposes familiarity with the sequence of the various episodes of the Soma-sacrifice. Here, the Nihnava is preceded by the Tānūnaptra, in which a “contract” is established between all the participating priests and the sacrificer, and the Āpyāyana, in which the Soma-plant is sprinkled with hot water. In order to proceed from the collectively performed Tānūnaptra to the Āpyāyana in which the Adhvaryu makes each of the other priests and the sacrificer sprinkle the Soma (kept on the “throne” installed just south of the offering fire), each priest and the sacrificer must walk around the fires and move southward to the throne. Because the participants in the sacrifice had to turn their back to the fires and go southward in the immediately preceding rite of the Āpyāyana, they next, in the Nihnavana, disavow this undesirable but contextually necessary act by pronouncing a prayer to heaven and earth while placing their hands in a certain position near the ground in the Vedi, keeping the hands on or under a grass-bundle, the prastara. The Taittirīya-Saṁhitā (TS 6.2.2.6) speaks of “falling away from this world” rather than turning the back to the fires and going south as the problem that is to be taken care of, but the end result and the necessity for going through the Nihnavana are the same. Third, the accompanying prayer is a very general formulation consisting of a request for wealth and a respectful greeting to Heaven and Earth:29 26 Cf. MW, s.v. ni-hnu. 27 Staal’s paraphrase of nihnava as “hiding rite” is misleading, as it could suggest that the bundle of grass is hidden in this act. It would correspond to Haug’s interpretation and translation (“concealment”) (1863: 58), which was justly criticised by Weber in his review (1865: 221): “Und obschon Haug hinzufügt: the concealment is done in the manner expressed in the translation as I myself witnessed it, so verhält sich die Sache, den Ritualtexten nach, denn doch erheblich anders, als er angiebt. […] (bei der vormittägigen upasad) legen sie (die Priester) ihre Hände, die rechte aufwärts gerichtet, auf den prastara nieder – bei der nachmittägigen die linke aufwärts – , und entziehen sich (durch diesen demüthigen namaskāra allen üblen Folgen dessen, dass sie vorher nach Süden gegangen sind […]).” 28 Cf. Houben forthcoming. 29 TS 1.2.11c éṣṭā rāyaḥ préṣé bhágāyartaṁ ṛtavādíbhyo námo divé námaḥ pṛthivyái.

36

Jan E.M. Houben “Wished towards us are riches (rāyaḥ), exceedingly, for food (iṣé), for wellbeing (bhágāya), (and wished is) the divine order (ṛtaṁ) for those promulgating the divine order; obeisance to heaven (námo divé), obeisance to the earth (námaḥ pṛthivyái).”

As such it suits and supports the meaning and function attributed to it in the ŚB and TS. Fig. 1: Tānūnaptra

Fig. 1–3: Three plans show (1) the position of the priests at the time of the Tānūnaptra (performed by all priests together); (2) how, after this, each priest walks behind the Āhavanīya-fire to the Rājāsandī in order to perform the rite of sprinkling the Soma-plant (Āpyāyana); and (3) how each walks then in a curve around the Dakṣiṇāgni and Gārhapatya-fire in order to perform, one by one, the Nihnava at the point where formerly the Tānūnaptra was done.30

R. = Rājāsandī, “throne” on which “king” Soma is kept; S. = Saṁrāḍāsandī, “throne” on which the “emperor”, i.e., the Pravargya-pot, is kept;

30 For a more detailed plan and brief discussion of the sacrificial area of a Soma-sacrifice (with a Pravargya) see Houben 1991: 135–137.

Formal Structure and Self-referential Loops in Vedic Ritual Fig. 2: Āpyāyana

Fig. 3: Nihnava

37

38

Jan E.M. Houben

Staal himself mentioned earlier that the inauspicious nature of the southern direction was a detail on which participants and scholars will agree.31 Hence, finally, the “disavowal” with the prayer and gesture make sense – if we take adjacent ritual episodes into account. And even if the young ritualist in Staal’s example is not told so explicitly, he may very well understand it by himself from the structure of the ritual – perhaps only when he goes through it in practice – just as any scholar may do who is willing to open his eyes for it. The point is easily missed, however, if we only read descriptions of the ritual and never observe its performance: see pictures 1 and 2, taken at the beginning of the Tānūnaptra and during the performance of the Nihnava (simultaneously with another priest’s Āpyāyana) within an Atyagiṣṭoma performed in 2001 in Barsi, India.

Image 1: The priests gather for the Tānūnaptra ceremony. From video by the author

Image 2: One priest (left, kneeling) performs the Nihnava at the Vedi; on the right the Āpyāyana is performed. From video by the author

As we will see, however, although this is a very well-established meaning which must be and must have been valid for a considerable number of performing ritualists, it is not the only nor necessarily the most original meaning attributed to the Nihnava. If we consistently apply Staal’s approach to the Nihnava it would have, on account of its meaninglessness in Staal’s broad sense, neither meaning nor any function that would go beyond its own performance. Hence, it would have no relation with preceding or following ritual episodes except for accidental contingency. This is not just an incidental problem, as the same would apply to other episodes, and we would have a series of utterly unrelated episodes. Even each individual episode 31 Staal 1989: 116, cited above.


39

would be, in turn, an adventitious grouping of smaller ritual acts. Schematically, this situation can be represented as in Figure 4.

Fig. 4: Complete meaninglessness: the Nihnava and other ritual episodes are without any meaning or function and the ritual consists of utterly disconnected elements. The situation is different if we accept the meaning and function attributed to the Nihnava in the Śatapatha-Brāhmaṇa and the Taittirīya-Saṁhitā. The Nihnava then acquires a specific relation with the neighbouring rites. To be more precise, since it disavows the undesirable but unavoidable act of turning the back to the fires and going south, it is exclusively dependent on the preceding Āpyāyana rite. As it supposedly averts a negative aspect of the Āpyāyana rite, it possesses or is supposed to possess what can be called a virtual causality. It is an appendix to the Āpyāyanarite, where it was necessary to do this inauspicious act. It has nothing to do with the rites that follow immediately, for instance the Subrahmaṇyā invocation. Nor does it have any contribution to make to the receiving and installing of the Soma which is the major theme of the surrounding ritual episodes. Thus, if we do take into account meaning and function, and only then, we perceive that the units can have specific functions and relations with regard to other units so that they form a specific structure which would be entirely absent if there were absolutely no meaning.

Fig. 5: The Nihnava is meant to disavow the necessary but inauspicious and possibly dangerous act in the immediately preceding episode: going to the south.

40

Jan E.M. Houben

Since authoritative interpretative texts such as the Śatapatha-Brāhmaṇa and the Taittirīya-Saṁhitā attribute to the Nihnava the interpretation of a “disavowal” after having gone in an inauspicious direction (south or away from this world) we may infer that generations of learned ritualists of these traditions accepted this as the canonical meaning of the rite, which, moreover, suits the context and the accompanying mantra. We can thus accept the performative reality of this interpretation for a sufficient “mass” of performers (just as de Saussure’s sign needs a sufficient mass of speakers for its survival over time). The Brāhmaṇa passages reflect, and on account of their status prescriptively encourage, an interpretation that must have retained its validity among generations of performers of the rite. This does not mean that it is the only or necessarily the most original meaning attributed to the Nihnava. In the Maitrāyaṇī-Saṁhitā we first find a passage (MS 3.8.2, 93.18ff.32) which is very similar to the relevant passage in the TaittirīyaSaṁhitā. Next, the Maitrāyaṇī-Saṁhitā (MS 3.8.2, 94.333) gives an alternative explanation according to which the Nihnava or “Disavowal” would not concern a wrong or inauspicious act that has already taken place – viz. going south or away from this world – but a wrong act that is still to take place, viz. the pressing or “killing” of the Soma-plant, addressed at this stage of the ritual as “king Soma”. We note that there are episodes called Nihnava in other ritual contexts,34 where they frequently occur before and not after the wrong or inauspicious act concerned. According to the mentioned Maitrāyaṇī-Saṁhitā interpretation, the priests would ask Heaven’s and Earth’s forgiveness not for going south but for something that is still to take place: the terrible deed of “killing” king Soma, that is, of pressing the Soma-plant. This interpretation suits the mantra as well as the name Nihnava not less than the first interpretation. It has, however, consequences for the structure of the complex ritual. The surface structure, the sequence of episodes, remains exactly identical, but the underlying structure becomes quite different:35 the Nihnava is now related to one of the main actions of the Soma ritual. With this interpretation the Nihnava is directly linked to the action of pressing or “killing” the Soma-stalks in order to extract their juice.

32 Maitrāyaṇī-Saṁhitā 3.8.2, 93.18ff.: yánti vā etè 'smāl lokād yé sómam āpyāyáyitum udásthuḥ párāñco hí yántīśvarāḥ prámetor yád āhéṣṭā rāyā éṣṭā vāmāni préṣé bhágāyéti ténāsmāl lokān náiti ténāsmím̐l loké dhṛtā[-táḥ] 33 anáyor (viz. of heaven and earth) vā eṣá gárbha ābhyām eṣá āvṛścate yáḥ sómam̐ hanti yád āha námo divé námaḥ pṛthivyā íty ābhyām evá námo 'kar. 34 For instance within the Sākamedha, in an offering to the Fathers (TS 1.6.9.8). See further Brough (1950) for a discussion of this and other cases. 35 What becomes visible here is that the development of Vedic ritual over a long stretch of time has resulted in different “‘blueprints’ overlaying each other” (Heesterman 1993: 243).


41

Fig. 6: Following a second interpretation offered in the Maitrāyaṇī-saṁhitā: the Nihnava is meant to disavow the necessary but inauspicious and possibly dangerous act of pressing or “killing” the Soma in a subsequent part of the ritual. In other words, if we apply Staal’s meaninglessness (“To say that ritual is for its own sake is to say that it is meaningless, without function, aim or goal, or also that it constitutes its own aim or goal”)36 consistently, not only the ritual as a whole but also the Nihnava and other episodes would have neither any meaning nor any function that would go beyond their own performance. Hence it would not be possible that any of these episodes would have any relation with preceding and/or following ritual episodes except for accidental contingency. In the Soma ritual according to the classical sources, an episode taking place just before another episode that concerns the pressing of Soma may itself be concerned with quite something else, for instance the food of the sacrificer. If each episode is strictly for its own sake, however, it would be impossible that one episode contributes to one superordinate unit and the next to another one, as we would have a series of unrelated episodes; even each individual episode would be an adventitious grouping of smaller acts. Only if the episodes are allowed to have meanings and functions that go beyond their own performance do we have a basis for arriving at larger structural units within the ritual, allowing also for some overlapping at the level of sequential, concrete acts. One may point to the partly parallel case of language, where, even without trying to understand and interpret a sample, we will have insufficient clues to decide on issues like word boundary and word groups, unless we take into account semantics. Thus, as can also be deduced from Staal’s admission cited above that “Some such questions [viz. on meaning] do receive specific answers, on which participants and scholars agree”, the relation between ritual act and meaning may often be arbitrary but appears nevertheless as tenacious. A simple pragmatic relationship between act and function, and hence between intention and act, is often 36 Staal 1989: 131–132.

42

Jan E.M. Houben

conspicuously absent in ritual,37 whether it is to distinguish the world of ritual from daily life or for any other reason. As in the case of de Saussure’s “speaking mass” transmitting a language where words have an arbitrary but tenacious relation with meanings, the stability in the relation between ritual acts and meanings will be determined by the “mass” of persons that is engaged in the performance and by the quality of the transmission of the ritual. Just as in language we distinguish, in one and the same string of (tenaciously meaningful) linguistic units, the distinct organisational levels of (a) phonemes, (b) morphemes, (c) words and (d) sentences, similarly in a complex ritual such as Vedic ritual we have to distinguish, in one and the same string of (tenaciously meaningful) ritual episodes, distinct organisational levels. As is clear from the example of the Nihnava, Staal’s perception and acceptance of meaning at the level of some specific ritual acts but not at the level of the ritual derives from a confusion of different organisational levels in ritual – levels that have so far hardly been identified let alone studied. Nor is it decided that language rather than, for instance, biology, will be the most appropriate analogy in the study of ritual. Although we will not elaborate this analogy here one could with equal justification compare the organisational levels in ritual with those current in biology: (a) genes, (b) cells, (c) organs and (d) organisms.38 In our investigation of the Nihnava, an episode or sub-ritual in the Soma-sacrifice, we have thus seen the following: first, we are forced to accept some minimal meaning in order to distinguish units within the entire sacrifice and to avoid that the Soma-sacrifice becomes an amorphous mass of atomistic ritual elements. Second, in the case of the Nihnava we have an episode that refers to another ritual element (part of another ritual episode), according to one interpretation to a preceding one, according to another interpretation to a following one. Third, in Vedic rituals such as the Soma-sacrifice, we have to accept different organisational levels, starting from (a) elementary units or elements, to (b) ritual episodes, (c) more complex units consisting of several episodes and (d) the ritual as a whole. Finally, we may attribute to the Nihnava a virtual causality, as it is supposed to accomplish “something”, in this case with regard to a preceding or following episode in the ritual. From the point of view of the Nihnava as ritual episode the causality is directed outward (towards a ritual element in another episode), but on account of this reference by the Nihnava the ritual (Soma-sacrifice) as a whole is 37 Cf. Verpoorten 1991: 169 on Sāyaṇa-Mādhava’s understanding of the contrast between daily life and ritual, and Humphrey & Laidlaw 1994 on the shift in intentionality between the act in daily life and the ritualised act in Jaina ritual. 38 As is well known, the scientific analysis of language and that of biological organisms have mutually inspired each other in the early stages of development of the disciplines of, respectively, linguistics and biology (cf. the term “morphology” current in both disciplines); see for instance Picardi 1977.


43

rendered self-referential. In this regard, we can call the Nihnava a self-referentiality episode (episode leading to self-referentiality of the ritual as a whole). This ritualinternal referentiality and the resulting self-referentiality of the ritual (as a whole) constitute a minimal meaningfulness that is to be accepted beyond the aspect of meaning that allows us to distinguish units (in our case we could manage with the traditionally accepted units39) in order to maintain the ritual as a structured whole: for this we do not have to leave the world of the ritual and need not get entangled in any extra-ritual reality.

3. Ritual Self-referentialities The Nihnava is by no means the only ritual episode with self-referential characteristics: general corrective offerings are regular parts of common rituals such as the Iṣṭis40 as well as of the Soma-sacrifice.41 The offering for Agni Sviṣṭakṛt (for “Agni who makes a good offering”) is a frequently recurring episode as well, and it makes a perfect candidate for being considered not only an episode contributing to the self-referentiality of the ritual as a whole, but as a directly self-referential episode: the offering would logically include itself as the object that is to be rendered “good” by its execution. Vedic rituals and the texts connected with them from the Rg-veda onwards are in fact quite rich not only in recursive patterns42 but also in instances of self-reference and circularity in several dimensions: (a) self-referentiality of the language: speech, poetry, poetic creation reflecting on itself: “At certain moments, the author seems to abandon his topic or rather superimposes on top of it another topic, which consists of making a curve back to himself. [...] he pays 39 Here modern scholars of Vedic ritual profit from the work done in ancient Indian ritual science as reflected in, for instance, the Śrauta-sūtras. In a ritual tradition for which no “own” science of ritual is available (as it is absent or not transmitted), the “sequentialisation” and the analysis of cross-ritual links (“interrituality”) are far from trivial: cf. Gladigow 2004. 40 Cf. Hillebrandt 1879 and 1897: 113–114. 41 Cf. Caland & Henry 1906–1907, vol. 2: 387f. on the Mindāhuti. 42 Both Staal and Witzel have been, each in their own way, fascinated by recursive patterns; Staal (1989: 85-99; 1993) wanted to formalise them through rewriting rules and inverted trees (inspired by generative grammar of the 1970s–1980s), Witzel (1992) in terms of a structure of frames within frames. Next to the principle of recursive patterns – which from an evolutionary perspective points to the accretion of elements at the beginning and end of an existing ritual – there is the diametrically opposed principle of “succeeding ‘blueprints’ overlaying each other” (Heesterman 1993: 243) which points to the conscious design by subsequent “authors” (and “interpreters”) of a ritual who form and reform it and whose involvement pertains to the entire structure of a ritual, not just to its beginning and end. A major pattern in the Soma sacrifice in which many elements find a natural place is, for instance, the reception of the gods as guests (Thieme 1957); not always in harmony with this is the contest pattern (Heesterman 1993: 43–44, 52).

44

Jan E.M. Houben

tribute to poetic inspiration [...]”;43 (b) instrument-circularity: adoration of the fires (Agnihotra) which are instrumental in, among other things, the transport of offerings to the gods, adoration of the pot from which an offering is made (the Pravargya44), etc.; (c) self-reference of ritual acts: either ritual acts refer to other ritual acts or to a larger unit that includes themselves: corrections, confirmations of ritual acts (cf. the Nihnava discussed here), advertisements and case histories of the ritual transmitted together with the rules for its performance; (d) self-reference for the sake of transmission: encouragements by the personified ritual system to transmit it carefully (see below, note 65). One major type of self-referentiality I have not yet mentioned as it is better placed in a separate category is that of the performing agent. To explain this we have to refer to the comprehensive theory of ritual propounded by Roy A. Rappaport (1999). Rappaport based himself on entirely different material from Staal: ethnographic fieldwork done among the Tsembaga (sub-clan of the) Maring in New Guinea. Like Staal, he emphasises the formal nature of ritual activity in his basic definition. However, when he elaborates his basic definition he does not abstract from the semantic dimension; on the contrary, he develops a concept of meaning that is greatly expanded in several dimensions, so that it includes what others label as semantics, semiotics and pragmatics. Rappaport defines the term ritual as “the performance of more or less invariant sequences of formal acts and utterances not entirely encoded by the performers”.45 In the remaining chapters of his work, Rappaport argues that this definition “logically entails the establishment of convention, the sealing of social contract, the construction of [...] integrated conventional orders [...], the investment of whatever it encodes with morality” as well as “the construction of time and eternity [...]”.46 Rappaport’s theory is not based on Vedic ritual but is very well applicable to it and is able to highlight unexpected dimensions as I briefly demonstrated with regard to its capacity to 43 My translation of Renou 1961: 15–16: “A de certains moments, l’auteur semble abondonner son propos ou plutôt y superposer un autre, qui consiste à faire un retour sur soi-même.” Renou (ibid.) continues: “Il évoque alors son propre ouvrage, qui, d’accompagnant, deviendra un élément actif; il salue l’inspiration poétique, laisse entrevoir les vertus du langage” (He evokes then his own work, which, in the process, will become an active element; he pays tribute to poetic inspiration, points to the virtues of the language). A related but still distinct self-referential pattern has been highlighted by Gonda (1975: 66–67): “The poets are moreover deeply convinced of the existence of an interplay of factors, reciprocity or rather cyclical process with regard to inspiration and the suprahuman power inherent in inspired poetry. [...] Strengthened by the hymns of the poets the gods [...] will continue furthering their inspirations.” In the poet who becomes conscious of himself as poet it is also the poetic tradition that becomes conscious of itself as poetic tradition. Cf. further Gonda 1963, passim. 44 Cf. Houben 2000a; 2000b; 2009b. 45 Rappaport 1999: 24. 46 Ibid.: 27.


45

undo “historicity” and to construct “time out of time” and “eternity”.47 For the participant in the ritual this construction of “time and eternity” contrasts with his experience of mundane time and historicity. The participant leaves this mundane time when he enters upon a ritual which takes him to a transcendent time and eternity; he comes back to the mundane time when the ritual is over, not without, however, taking with him something of the status of “eternity” and the transcendent. Parallel to this aspect of Rappaport’s comprehensive theory, Maurice Bloch (1992) proposed an account of (various types of) rituals which accepts, like Rappaport, that a shift in “times” is important, from mundane to transcendent and back, but which emphasises that this shift involves a double “violence”: first a form of “violence” against the vital aspect of the participant who leaves mundane time, next, in the form of a “rebounding violence” that comes with the participants who return to the world after they have, in the central (liminal) part in the ritual, appropriated knowledge and acquired a non-mundane, spiritual status. In one of its milder forms, the “rebounding violence” consists in the consumption of food after or in the latter part of the ritual. This appropriation of a new vitality after the central part of the ritual may under certain circumstances extend into aggression to neighbours and even to expansionist wars, hence Bloch’s claim that this theory can explain also political violence.48 In these types of rituals self-referentiality of the performer or performers is crucial. The performer concerned in this self-referentiality, however, does not remain the same throughout the ritual, but is transformed either in a (more or less) reversible or in an irreversible way with regard to basically two variables, (a) vitality and (b) spirituality or knowledge. In Rappaport’s theory of formal-ritual-plus-rich-meaning, rituals transmit messages, of which two types are distinguished: (a) canonical messages (which derive from the invariant aspect of what is encoded by others than the performers); and (b) what Rappaport calls self-referential messages (messages transmitted both to the performers themselves and to others and that provide information on the participants’ own current physical, psychic, economic, and/or social status). It is the former, the canonical messages, which represent universal orders transcending concrete time and space. It is this dimension also that has mostly been the subject of continued efforts of philologists and classical indologists and that has sparked their interest in mythological and ritual studies. But it is through the necessary capacity to transmit self-referential messages that rituals are interwoven in the social

47 Houben 2002. 48 The “rebounding violence” is first of all directed to other species and expresses itself in the consumption of meat; it “may flow over to other human groups, resulting in conquering wars, which makes these rituals of political significance” (Snoek 2004: 96).

46

Jan E.M. Houben

and political history of a country or area.49 It is the dimension of self-referential messages that links up best with a performative approach to ritual50 but also with an approach to ritual as social practice.51 For several reasons52 it is better to refer to the two major types of messages distinguished by Rappaport as (a) the canonical dimension and (b) the performance dimension. Staal’s interest in the formal side of ritual evidently concerns the canonical dimension, but avoids the performance dimension. Staal did study the performance and even filmed it (together with Robert Gardner) but apparently only in order to have a better understanding of the canonical dimension. Other ritual studies specialists such as Richard Schechner are mainly interested in the performance dimension of ritual – as Schechner wrote in 1986: 360 “Ritual is performance” – and attribute only marginal significance to the canonical dimension. An amusing but hardly fruitful discussion thus emerged in the 1980s on the occasion of the publication of Staal’s Agni in 1983.53 The scholar who succeeded in giving due recognition both to Staal’s and to Schechner’s dimension within a comprehensive theory of ritual is Roy A. Rappaport. Making use of his framework our observations on self-referentiality in Vedic ritual can now be summarised as follows: Ritual: I.

performance (Schechner’s) dimension: there is always self-referentiality of the performing agent;

II. canonical (Staal’s) dimension: its structure is open to formal analysis; there may be self-reference of ritual acts, self-reference of the ritual in its transmission, instrument circularity, self-reference of ritual poetry – next to episodes which concern intra-ritual events such as the production, libation and consumption of the Soma-drink, the restricted meals of the fasting sacrificer; 49 For instance, a ritual may require the sacrifice of “many” animals of a certain sort. The sacrificer may then decide to sacrifice as many animals as he can afford, depending on economic power and social status. To compare the very abstract entities of power and status of two persons would be very difficult, but expressed in ritual in the form of the number of animals sacrificed – a self-referential message of the performer – these entities would become comparable. In another context the performer may want to show his political power or sovereignty to himself, his subjects and his neighbours, as in the case of an Aśvamedha. An increased attention for the performance dimension even in domains which are traditionally more interested in the canonical dimension suits the recent “cultural turn” in the humanities (and South Asian studies) – away from essentialism and ahistoricity – see Michaels 2004, 2005. 50 Tambiah 1979; Schechner 1987a; 1993. 51 Bourdieu 1977; Bell 1992; 1997: 76–83. 52 One reason is that, as we have seen, self-referentiality may also occur within the canonical dimension and is therefore not suitable to characterise the other dimension. 53 Schechner 1986; 1987b; Staal 1987; 1989: 250–251.


47

and episodes that refer to extra-ritual events, prayers for (real-life) success and well-being, for rain, cows, etc. The need to distinguish different organisational levels, which we demonstrated above, pertains to the canonical dimension. Both in language and in biology, which have been identified as domains that can provide useful extended metaphors in the study of ritual (we may now add: in the study of the canonical dimension of ritual), there are limiting cases in which some or all levels collapse or are conflated: the one-phoneme-word-sentence and the one-cell-organism. In the case of ritual we have the isolated, thinly ritualised act54 which is nevertheless evocative on account of what is “not entirely encoded by the performers”,55 and which, as a “single act” ritual where we find a collapse of levels that are to be distinguished in rituals with a developed canonical dimension, parallels the one-phoneme-word-sentence and the one-cell organism. Bloch’s “rebounding violence”, which he sought to demonstrate in rituals with a developed canonical dimension, can now be seen as the result of the interplay of these two dimensions, the performance dimension and the canonical dimension, when the performer transits from mundane to ritually created time and back to mundane time. The amount of virtual causality or “rebounding violence” generated for the performer depends, naturally, on the intensity of his moral and/or substantial engagement, or on the amount of his investment in the ritual performance which pertains to the performance dimension. On the other hand, however, the force with which the canonical dimension is able to create “time out of time” or “eternity” and to impress these on the performer has implications for the capacity to generate virtual causality or “rebounding violence” for the performer. A reinforcement of the canonical dimension will then equally contribute to the generation of virtual causality or “rebounding violence”. If we abstract from the performing agent, to whose performance there is always an aspect of self-referentiality or “showing off”, we are left with the canonical dimension, not because we would believe, with Staal, that it is the only dimension of importance, but because we want to continue with an exploration of cases of self54 See, for instance, Schechner’s reference (1993: 57) to the “cosmopolitan” dance (in a style that is “not traditionally Chinese”) of two students in Tiananmen Square, spring 1989. Rituals in which the performative dimension dominates or in which the canonical dimension is not much developed need not be of short duration and can have their own complexities, as was demonstrated in the case of the Naven (Houseman & Severi 1998). The weak or virtually absent canonical dimension leaves such rituals quite vulnerable; the Naven as Bateson observed it in the 1930s (one type of which presupposing a successful headhunt by a young member of the tribe) has more or less died out in modern, partly Christianised Iatmul society. 55 Rappaport 1999: 24 quoted above.

48

Jan E.M. Houben

referentiality which are characteristic for Vedic ritual. It is within what we can now see as the formal structure of the canonical dimension that we found our remarkable self-referential loop involving the Nihnava; and it is in this dimension that we have to place the “instrument-circularity”, “self-reference of ritual acts”, “self-reference of ritual poetry”, etc., mentioned above. The Nihnava, under both interpretations, we may regard as a self-referentiality episode of the yajñéna-yajñám type. This refers to the oldest attestation of this principle in the form of a verse of the Rg-veda, viz. RV 1.164.50ab = 10.90.16ab: yajñéna yajñám ayajanta devā́s tā́ni dhármāṇi prathamā́ny āsan “With ritual worship the gods worshipped ritual worship: these institutions were the foundational ones.” Whatever the ritual or religious significance of this statement, there is also a logical problem. It is like pulling oneself out of the swamp by one’s own hair. Or it is like the sentence which says of itself that it is true, which is, from the point of view of its logic, hardly less problematic than its more spectacular and more famous counterpart, the sentence which says of itself that it is false. Is it because ritual studies usually remain within the domain of religious studies that this problem has not yet been addressed from a logical point of view?56

4. yajñéna-yajñám: Bad Logic or Coil Generator? With regard to Vedic ritual we have found that it is a formal system with different organisational levels; that its units are associated with meanings, semantic interpretations and functions or pragmatic aims; that minimal meaningfulness is to be accepted to demarcate the functional units, and to allow intra-ritual reference. Now, as in “this sentence is false/true”, intra-ritual reference can become self-reference, depending on the interplay of intra-ritual reference and the appropriate organisational levels. Self-reference within a formal system, or within a language that we would like to be the basis of a formal system, has been a problem for logicians ancient and modern, east and west. Philitas of Cos is said to have died from excessive reflections on the ‘liar paradox’ according to his epitaph, recorded in the work of Athenaeus (Deipn. 401e): “Stranger! I am Philitas. Among logical propositions the ‘liar’ (pseudómenos), and studies of riddles late in the evening, have killed me.” Bertrand Russell, and later on Rudolf Carnap and Alfred Tarski, pos56 I am aware of one other researcher of rituals, G. Bateson, who stumbled on logical problems – different from our problem – in his ritual, the Naven of the Iatmul on Papua New Guinea (a ritual in which the performative dimension is dominant), and tried to solve them with reference to Bertrand Russell’s theory of logical types (Russell 1908). Following the theory of Bertrand Russell, Bateson distinguished a hierarchy of messages: messages, metamessages, meta-meta-messages; this, in turn, led him to propose the concepts of “deuterolearning” and “double bind” (cf. Bateson 1972) which are still of importance not so much in ethnography but rather in behavioural science and psychiatry.


49

tulated a hierarchy of formal languages to avoid the problem of paradoxes. According to Russell: “The arguments for the necessity of a hierarchy of languages … are derived from the paradoxes; their applicability to the words ‘true’ and ‘false’ is derived from the paradox of the liar.”57 On the Indian side, the grammarianphilosopher Bhartṛhari (fifth century C.E.) was evidently aware of the problem of “the liar”; analysing it quite differently from Russell and other modern logicians, he identified the crucial parameter that turns harmless colloquial expressions into an inescapable paradox so that his own statements can stay clear of contradiction (virodha) and infinite regress (anavasthā).58 In recent decades, however, the aim of logicians interested in “the liar” is no longer to evade and avoid paradoxes in the way Russell and others tried, but to strengthen them.59 Next to the classical “Epimenides the Cretan says: All Cretans are liars” we now have a strengthened paradox in “This sentence is false” which abstracts from the human agent (speaker). The self-reference of the strengthened paradox “This sentence is true/false” involves a crossing of organisational levels (which Russell and others tried to “forbid”) as it depends on the capacity of the word “this” to refer to something else, which is in our case the sentence “This sentence is true/false”. In ritual the situation is similar though far more complex. Depending on whether we give the levels of words and the sentence as separate or conflated, the self-reference of “This sentence is true/false” can be represented in different ways.

Fig. 7: Representations of the self-reference of “This sentence is true/false”. 57 Russell 1950: 62. 58 Houben 2001. 59 Cf. Barwise & Etchemendy 1987.

50

Jan E.M. Houben

Although our motives were not the same as those of the logicians who created the strengthened paradoxes by abstracting from the speaker, we have in a similar way abstracted from the performing agent in Vedic ritual. In our self-referring sentence we have a word, “this”, referring to “something” else, namely the sentence in which “this” occurs. In the case of our rituals, we have a ritual episode E which refers to another ritual episode or to the ritual as a whole which includes that ritual episode. Ritual episode E may be a Nihnava, a Sarvaprāyaścitta or an offering to Agni Sviṣṭakṛt. For the sake of simplicity, let us assume that the ritual episode E expresses (or has the function) “this ritual is well done”. In that case we may represent the state of affairs in three different ways, parallel to our three representations of the self-referential sentence, as in Figure 8.

Fig. 8: Representations of a ritual that is rendered self-referential by episode E. What does it mean if Vedic ritual has episodes that are directly or indirectly selfreferential? Were the designers of Vedic ritual bad logicians? If we revive here Sylvain Lévi’s60 comparison between Vedic ritual (sacrifice) and “electricity” (as understood at the end of the nineteenth century) – 60 The importance of Sylvain Lévi for Hubert and Mauss and eventually for Staal’s theory of ritual has been highlighted by Ivan Strenski in an article in which he also points out (1997: 532) that Sylvain Lévi defends the position – original in his time – that ritual is not so much a means to understand religion, but that “the nature of the religion revealed in the Brāhmaṇas is constituted by sacrificial ritual”: “So potent is the sacrifice,” says Strenski paraphrasing Sylvain Lévi’s explanation, “that even if gods are relevant, those very gods are ‘born’ from sacrifice, are ‘products’ of it.” Thus, according to Strenksi (ibid.), “Sylvain Lévi in effect argues for what I have earlier termed causal ritualism.” In a similar vein, Louis Renou, according to Strenski, observed about Sylvain Lévi that for him “ritual dominates mythology” (Strenski, 1997: 537 note 55, with reference to Renou 1937: xxiii; cf. Renou 1936); but to this we should add that it is not certain to what extent this is Sylvain Lévi’s


51

“Sacrifice is a learned and complicated combination of ritual acts and sacred speech, or rather it is the impalpable and irresistible power which is released when they are brought together, like electricity is born from elements put into contact.”61 – are directly self-referential episodes then not like fatal short circuits? or do they form the coil of a dynamo or electric generator? As we have seen, although this self-reference that disregards the distinction between organisational levels was a major problem for Russell and logicians up to around the 1970s, other perspectives on the same situations of self-reference were proposed in subsequent decades. One remarkable approach that is not easily categorisable in a single discipline or domain – mathematical logic, formal semantics, visual art, music – is the one presented by Douglas R. Hofstadter in his Gödel, Escher, Bach (1979; I will refer to the 20th Anniversary Edition, 1999). Inspired by the work of the mathematician Kurt Gödel and his criticism on the Principia Mathematica of Russell and Whitehead, Hofstadter observed the following regarding “formal systems”: “meaning cannot be kept out of formal systems when sufficiently complex isomorphisms arise. Meaning comes in despite one’s best efforts to keep symbols meaningless!”62 Thanks to scholars such as Hubert & Mauss (1899; 1964) and Staal (1979; 1989) we have already started to look at ritual as a formal system, which is apparently an appropriate representation of at least one of its major dimensions: the canonical one. It therefore seems permissible to apply Hofstadter’s insight also to ritual, even if this was not among the formal systems taken into account by him. In the case of ritual, and particularly in the case of Vedic ritual, we have seen that even if we try our utmost to see it as an entirely meaningless formal system, we are bound to accept, counter to Staal’s theory, its meaningfulness at least with minimal meanings that allow the distinction of sub-units and intra-ritual reference. With regard to selfreference within formal systems, Hofstadter has other insights to contribute that may be applicable to the dimension of formal structure in ritual.

well-considered position, as the statements are made by him when he explains his understanding of the Vedic Brāhmaṇa texts (Lévi 1898; 2003). 61 My translation of Lévi 1898: 77: “Le sacrifice est une combinaison savante et compliquée d’actes rituels et de paroles sacrées, ou plutôt il est la puissance impalpable et irrésistible qui se dégage de leur rapprochement, comme le fluide électrique naît des éléments mis en contact.” 62 Hofstadter 1999: P-3.

52

Jan E.M. Houben “For the French [in the 1920s–1930s, J.H.], the enemy was Germany; for Russell, it was self-reference. Russell believed that for a mathematical system to be able to talk about itself in any way whatsoever was the kiss of death, for self-reference would – so he thought – necessarily open the door to self-contradiction, and thereby all of mathematics crashing to the ground. In order to forestall this dire fate, he invented an elaborate (and infinite) hierarchy of levels, all sealed off from each other in such a manner as to definitively – so he thought – block the dreaded virus of self-reference from infecting the fragile system. It took a couple of decades, but eventually the young Austrian logician Kurt Gödel realized that Russell and Whitehead’s mathematical Maginot Line against self-reference could be most deftly circumvented (just as the Germans in World War II would soon wind up deftly sidestepping the real Maginot Line), and that self-reference not only had lurked from Day One in Principia Mathematica, but in fact plagued poor PM in a totally unremovable manner. Moreover, as Gödel made brutally clear, this thorough riddling of the system was not due to some weakness in PM, but quite to the contrary, it was due to its strength. Any similar system would have exactly the same ‘defect’. [...] Something very strange thus emerges from the Gödelian loop: the revelation of the causal power of meaning in a rule-bound but meaning-free universe.”63

Hofstadter goes one step further – if not too far – when he states: “this is where my analogy to brains and selves comes back in, suggesting that the twisted loop of selfhood trapped inside an inanimate bulb called a ‘brain’ also has causal power”; according to this suggestion, “an ‘I’ comes about [...] via a kind of vortex whereby patterns in a brain mirror the brain’s mirroring of the world, and eventually mirror themselves, whereupon the vortex of ‘I’ becomes a real, causal entity”.64 This is a thesis that will be difficult to either prove or falsify and it does, moreover, place us on a side-track away from our subject, Vedic ritual. Nor will it help us much to attribute an “I” to the Vedic ritual system, even if this is what the tradition itself does when it encourages its own careful transmission in a remarkable self-referen-

63 Hofstadter 1999: P-5, P-6. 64 Hofstadter 1999: P-6; cf. also Hofstadter 2007.


53

tial loop: the Vedas, the collections of utterances implying ritual acts, are made to address the Brahmin, knower of the Veda, thus:65 “protect me, I – the knowledge of the Vedas – am your treasure; do not proclaim me to the envious, the unstraightforward, the undisciplined; this way I can become strong. [...] To protect your treasure, O Brahmin, proclaim me only to the one whom you know to be pure, focused, intelligent, following the life-rules of a Vedic student, and who is never hostile towards you.” More interesting than a putative “I” of the Vedic ritual system – more interesting also than the question whether or not Hofstadter should be followed in his final step – is his remark on the “causal power” of meaning in a self-referral formal system. One of the “works” which rituals do according to Rappaport’s theory is to take the participant from his daily historicity to liturgical time and eternity, and finally back to his daily historicity. Bloch illustrated how this movement involves two “violences”: one directed towards the participant at the beginning of the ritual, and one directed outwards, e.g. in the form of the consumption or even conquest of food, when the participant comes back to daily life towards the end or after the ritual. An explanation of where this “rebounding violence” or, in more neutral terms, virtual causality, comes from is, however, largely wanting. Could Hofstadter’s “causal power” have a role to play in this? Just as Snoek (2004) added the Navjote Zoroastrian initiation to rituals that can profitably be seen in the light of Bloch’s theory, it would seem that we can also add the Vedic Soma-sacrifice, with its evident phases of giving up “native” vitality during the Dīkṣā and gaining it back from an outside source at the end in the form of a stylised meal (which at an earlier stage was real and not entirely unfunctional66). However, whereas the shifting between “times” and the self-referentiality of the ritual performer who under65 [...] gópāya mā śévadhiṣ ṭe 'hám asmi / ásūyakāyā́nṛjave 'yatā́ya ná mā brūyā vīryávatī tathā́ syām // [...] yás te vidyā́ḥ śucím apramáttaṁ médhāvinaṁ bráhmacaryopápannam / yás te ná druhyét katámac canā́ha tásmai mā brūyā nidhípāya brahman [read thus instead of bráhman] // These verses have been quoted in the Nirukta 2.4 (ed. Svarup 1927), and similar ones appear in the Saṁhitā-Upaniṣad-Brāhmaṇa 3 (ed. Sharma 1983: 55) and elsewhere. The interpretation of the speaking vidyā “knowledge” explicitly as “knowledge consisting of the Vedas” is found, e.g. on Saṁhitā-Upaniṣad-Brāhmaṇa 3.9 vidyeti – [...] iyaṁ caturvedasvarūpā vidyā. 66 Cf. Heesterman 1993 for food in classical Vedic ritual and in pre-classical Vedic ritual as proposed. The older situation of a “real” meal is reflected in RV 1.162.5 hótādhvaryúr ā́vayā agnimindhó grāvagrābhá utá śáṁstā súvipraḥ / téna yajñéna svàraṁkṛtena svìṣṭena vakṣáṇā ā́ pṛṇadhvam // “The Hotar, Adhvaryu, the Āvayāj, the one who lights the fire (Agnimindha), the one who handles the pressing stone (Grāvagrābha) and the well-inspired Śaṁstar: fill your bellies with this well-prepared, well-offered offering”; and cf. RV 1.162.12. For other aspects of this sacrificial meal see Thieme 1957 on “guest reception”in Vedic ritual; and Malamoud 1989 and 1996 on the central place of “cooking” in Vedic ritual which is thought to entail a positive transformation of the world (loka-pakti) for the performer.

54

Jan E.M. Houben

goes this shifting and is transformed in the process are similar to Vedic ritual67 and the rituals described by Bloch, the strong presence of self-referentiality within the canonical dimension may at present be regarded as – further investigation is obviously required – rather unique for Vedic ritual.68 We have here not studied self-referentiality in the performance dimension which has its own share to contribute to ritual “causal power”, or rather, if we follow Hofstadter’s line of thought, to an extension of the “causal power” already inherent in the performer’s self-referent “vortex of ‘I’.” Self-referentiality in the performance dimension is characteristic for all rituals surveyed by Rappaport – from a sociological perspective rather than a perspective of formal semantics – and by others and we are happy to leave the exploration of this aspect to a specialist in ritual studies with sufficiently broad orientation. Self-referentiality within the canonical dimension, however, should be expected to contribute to special properties of a ritual system beyond the “normal” ones. From an overview of their work we can infer, in any case, that the ancient authors of Vedic rituals appreciated various types of self-referential loops in their compositions. In this they were not different from the Vedic authors whose poetry, as we have seen, is characterised by remarkable self-referential patterns. Which work is done, then, by these self-referential episodes in the canonical dimension of Vedic ritual? First, since the generation of virtual causality which is “normal” for more or less “all” rituals depends on the interplay of both the performance and the canonical dimension, a reinforcement of the canonical dimension will contribute to the generation of “normal” virtual causality or “rebounding violence”. Second, the self-referential episodes in the canonical dimension contribute to an independent momentum within the ritual that remains quite aloof from extra-ritual reality but that, like the flywheel or drivewheel of a motor, can drag along other cycles which do have real-life involvement. Specific interventions by the ritual specialists or priests – rather than the ritual patron – can direct the momentum of the flywheel towards a concrete aim which normally pertains to the success and prosperity of the patron who has invited the priests and bears the expenses. Just as the flywheel can provide the energy to pull a car forwards but also backwards, the ritual can be employed for the patron’s progress but also, in a theoretical or in any case extreme case referred to in ancient sources, for the patron’s (self-wished) death.69 67 The transformation is most visible in the fact that the sacrificer obtains a “title” which he carries in “mundane life”, for instance “Soma-yājin”, “Vājapeya-yājin” etc. On the shift in identity within the process of Vedic ritual see also Houben 2002. 68 It is still to be investigated to what extent this characteristic is carried over to ritual systems that have a direct relation with Vedic ritual, for instance Hindu ritual. 69 The so-called Śunaskarṇa Agniṣṭoma is a relatively simple modification of the standard Soma-sacrifice (Agniṣṭoma). The sacrificer’s death is brought about by certain details in the


55

Third and finally, there is a conspicuous characteristic of Vedic ritual over time: its extraordinary stability, solidity, resilience. The well-known “mobility” of the Vedic ritual system (in contrast with Hindu ritual that centres around geographically fixed places such as temples, mountains, etc.) points to a non-sedentary way of life for those who established and transmitted it, from ca. the mid-second millennium B.C.E. onwards. A non-sedentary life of Vedic people is further reflected in words of the language. In Vedic texts the word grā́ma, for instance, which is ‘village’ in classical Sanskrit, means ‘caravan’ or, in the words of Wilhelm Rau, firstly, a “train of herdsmen roaming about with cattle, ox-carts and chariots in quest of fresh pastures and booty”; and secondly, “a temporary camp of such a train, sometimes used for a few days only and sometimes for a few months at the most.”70 Agriculture was practised;71 but this does not imply a sedentary existence as the soil and climate of north-western India make it possible to raise and harvest a crop within a few months. For the Vedic people cattle was of major importance, as is clear from the constant references to cows, bulls and oxen in Vedic hymns. The Vedic people thus appear as non-sedentary agro-pastoralists.72 It must have been to a large extent or at least partly on account of the extraordinary self-referentiality within the canonical dimension that the Vedic ritual system has pushed into oblivion other systems that must have been in vogue at the time it started its expansion, from its early position in the north-west, throughout the Indian subcontinent. Overstaying its natural period in which it progressed and knew creative development in an environment conducive to non-sedentary agro-pastoralism, Vedic ritual witnessed a remarkable come-back in the Aśvamedha or horse sacrifice of Puṣyamitra Śuṅga (second century B.C.E.) after it had almost succumbed under the pressures of emerging religions such as Jainism and Buddhism (from ca. the sixth century B.C.E. onwards) which were in tune with the current conditions of recently urbanised society.73

70 71 72

73

performance, especially in the chanting of the Sāmavedins. If the sacrificer does not find the death he desires he may fast to death after the ceremony, or give up his project of dying (LāṭyŚS 8.8.40–41). Cf. Houben 1999: 123. Rau 1997: 203. Ibid.: 205. Cf. J.C. Heesterman’s observations, e.g. 1993: 111ff., and 128f.: “The temporary nature of the place of sacrifice fits in with the general mobility [...] of the fire and its vihāra. This is well illustrated by the so-called yātsattras or ‘moving sessions’ [...] It is this mobile pattern that is preserved, in telescoped form and drawn together within the confines of the single devayajana, by the basic paradigm of the soma ritual, the agniṣṭoma.” Cf. Gadgil & Guha 1992.

56

Jan E.M. Houben

5. Conclusion: Formal Structure and Enginery of Vedic Causal Ritual The Vedic Nihnava-rite, an episode within the Soma-ritual that has been used as a major example to illustrate a supposed fundamental meaninglessness of ritual, is solidly meaningful if we properly distinguish different organisational levels in ritual – levels which are to some extent comparable to the levels of (a) phoneme-sequences, (b) morphemes, (c) words and (d) sentences in language; and to the levels of (a) genes, (b) cells, (c) organs and (d) organisms in biology. The distinction not only of these organisational levels, but also of (1) the canonical dimension in which these levels appear and (2) the performance dimension, would be vital in any project of a “grammar of ritual”; both distinctions have till now been largely overlooked. In a further analysis, the Nihnava turns out to presuppose intra-ritual reference, which renders the ritual as a whole self-referential. Together with the meaning that is needed to identify distinct units within the ritual, intra-ritual reference and ritual self-reference (within the canonical dimension) form the minimal meaningfulness to be accepted in a formal representation of ritual that abstracts from collateral aspects of meaning. With reference to ritual theories proposed by Frits Staal, Roy A. Rappaport and Maurice Bloch, we have briefly reviewed different cases of ritual self-reference, especially in the canonical dimension, not so much from the perspective of the extended metaphor of language, nor from the perspective of the extended metaphor of biology, but rather from the perspective of formal systems. These cases include forms of “level-crossing self-referential loops” of which we argue that they are crucial both for the minimal meaningfulness of ritual and for its becoming an entity with virtual causality (beyond the virtual causality generated in the performance dimension). In the presence of suitable agents (performers of the ritual) this virtual causality can have a wide range of real-life results that vary from transformations in identity and changes of social role, to increased consumption and political violence. Both Vedic ritual as described in classical sources and the Vedic poetry that is at its heart are rich in self-referentialities in the canonical dimension which form part of the basic enginery of Vedic causal ritual. Over time, self-referentialities in the canonical dimension provide extraordinary solidity and resilience to the Vedic ritual system which have allowed it to largely overstay its naturally progressive period in the expanding agro-pastoral society of early Vedic people (roughly before the middle of the first millennium B.C.E.) until far beyond India’s “second urbanisation” from ca. the sixth century B.C.E. onwards, in fact – though margin-


57

ally and not without modifications and transformations in several respects – till the present day.74

74 I wish to express my sincere gratitude to Prof. Max Sparreboom and the staff of the International Institute for Asian studies, Leiden, for hosting me as visiting senior fellow in 2008– 2009.

58

Jan E.M. Houben

Abreviations LāṭyŚS = Lāṭyāyana-śrauta-sūtra. MS = Maitrāyaṇī-saṁhitā. MW = M. Monier-William’s Sanskrit-English Dictionary, Oxford 1899. RV = Rg-veda (-saṁhitā). ŚB = Śatapatha-Brāhmaṇa (Mādhyandina). TS = Taittirīya-saṁhitā.

References Altekar, A.S. 1954. Catalogue of the Gupta Gold Coins in the Bayana Hoard. Bombay: Numismatic Society of India. Barwise, Jon & John Etchemendy 1987. The Liar: An Essay on Truth and Circularity. New York: Oxford University Press. Bateson, Gregory 1972. Steps to an Ecology of Mind. London: Jason Aronson. Bell, Catherine 1992. Ritual Theory, Ritual Practice. New York: Oxford University Press. — 1997. Ritual: Perspectives and Dimensions. New York: Oxford University Press. Bloch, Maurice 1992. Prey into Hunter: The Politics of Religious Experience. Cambridge: Cambridge University Press. Bodewitz, Henk W. 1990. The Jyotiṣṭoma Ritual: Jaiminīya Brāhmaṇa I, 66–364. Leiden: E.J. Brill. Bourdieu, Pierre 1977. Outline of a Theory of Practice. Cambridge: Cambridge University Press. Bronkhorst, Johannes 19932 [1986]. The Two Traditions of Meditation in Ancient India. Revised Edition. Delhi: Motilal Banarsidass. — 19982 [1993]. The Two Sources of Indian Asceticism. With updated introduction. Delhi: Motilal Banarsidass. — 2007. Greater Māgadha: Studies in the Culture of Early India. Leiden: E.J. Brill. Brough, John 1950. “The Meaning of Ni√hnu in the Brāhmaṇas”. In: Vishva Bandhu (ed.). Siddhabhāratī or Rosary of Indology. Presenting 108 Original Papers on Indological Subjects in Honour of the 60th Birthday of Dr. Siddheshwar Varma. Part One. Hoshiarpur: V.V.R. Institute: 126–130. Also in: Brough 1996: 74–78. — 1996. Collected Papers. Edited by Minoru Hara & J.C. Wright. London: School of Oriental and African Studies, University of London. Caland, Wilhelm & Victor Henry 1906–1907. L’Agniṣṭoma. Description complète de la forme normale du Sacrifice de Soma dans le culte védique. Tomes 1–2. Paris: Ernest Leroux. Einoo, Shingo & Jun Takashima (eds.) 2005. From Material to Deity: Indian Rituals of Consecration. New Delhi: Manohar. Frazer, James George 1894. The Golden Bough: A Study in Comparative Religion. 2 vols. London: MacMillan.


59

— 19002. The Golden Bough: A Study of Magic and Religion. 3 vols. Revised and enlarged edition. London: MacMillan. Gadgil, Madhav & Ramachandra Guha 1992. This Fissured Land: An Ecological History of India. Delhi: Oxford University Press. Gladigow, Burkhard 2004. “Sequenzierung von Riten und die Ordnung der Rituale”. In: Michael Stausberg (ed.). Zoroastrian Rituals in Context. Leiden: E.J. Brill: 57–76. Gonda, Jan 1963. The Vision of the Vedic Poets. The Hague: Mouton & Co. — 1965. Change and Continuity in Indian Religion. The Hague: Mouton & Co. — 1975. Vedic Literature: Saṃhitās and Brāhmaṇas Wiesbaden: Otto Harrassowitz (A History of Indian Literature 1/1). Haug, Martin 1863. The Aitareya Brahmanam of the Rigveda Containing the Earliest Speculations of the Brahmans on the Meaning of the Sacrificial Prayers and on the Origin, Performance, and Sense of the Rites of the Vedic Religion. Edited, Translated, and Explained. 2 vols. London: Trübner. Heesterman, J.C. 1993. The Broken World of Sacrifice: An Essay in Ancient Indian Ritual. Chicago: Chicago University Press. Hillebrandt, A. 1879. Das altindische Neu- und Vollmondsopfer in seiner einfachsten Form mit Benutzung handschriftlichen Quellen dargestellt. Jena: Gustav Fischer. — 1897. Ritual-Literatur. Vedische Opfer und Zauber. Strassburg: Karl J. Trübner. Hofstadter, Douglas R. 1999² [1979]. Gödel, Escher, Bach: An Eternal Golden Braid. A Metaphorical Fugue on Minds and Machines in the Spirit of Lewis Carroll. 20th Anniversary Edition. With a new preface by the author. New York: Basic Books. — 2007. I Am a Strange Loop. New York: Basic Books. Houben, Jan E.M. 1991. The Pravargya Brāhmana of the Taittirīya Āranyaka: an ancient commentary on the Pravargya ritual. Delhi: Motilal Banarsidass. — 1997. “The Sanskrit Tradition”. In: Wout Jac. van Bekkum et al. (eds.). The Emergence of Semantics in Four Linguistic Traditions: Hebrew, Sanskrit, Greek, Arabic. Amsterdam: John Benjamins: 49–145 (Studies in the History of the Language Sciences 82). — 1999. “To Kill or not to Kill the Sacrificial Animal: (yajña-paśu)? Arguments and Perspectives in Brahminical Ethical Philosophy”. In: Jan E.M. Houben & Karel R. van Kooij (eds.). Violence Denied: Violence, Non-Violence and the Rationalization of Violence in South Asian Cultural History. Leiden: E.J. Brill: 105–183. — 2000a. “On the Earliest Attestable Forms of the Pravargya Ritual: Rg-Vedic References to the Gharma-Pravargya, especially in the Atri-Family Book (Book 5)”. Indo-Iranian Journal 43: 1–25. — 2000b. “The Ritual Pragmatics of a Vedic Hymn: The ‘Riddle Hymn’ (Rgveda 1.164) and the Pravargya-Ritual”. Journal of the American Oriental Society 120/4: 499–536. — 2001. “Paradoxe et perspectivisme dans la philosophie de langage de Bhartṛhari: langage, pensée et réalité”. Bulletin d’Études Indiennes 19: 173–199. — 2002. “The Brahmin Intellectual: History, Ritual and ‘Time out of Time’”. Journal of Indian Philosophy 30/5: 463–479.

60

Jan E.M. Houben

— 2007. Review of Einoo & Takashima 2005. Journal of the Economic and Social History of the Orient 50/1: 77–80. — 2009a. “Pāṇini’s Grammar and its Computerization: A Construction Grammar Approach”. In: Amba Kulkarni & Gérard Huet (eds.). Sanskrit Computational Linguistics. LNCS Sublibrary: SL 7 – Artificial Intelligence. Proceedings of the Third International Symposium. Hyderabad, India (15-17 January 2009). Heidelberg: Springer-Verlag: 6–25. — 2009b. “Transmission sans écriture dans l’Inde ancienne: énigme et structure rituelle”. In: Gérard Colas & Gerdi Gerschheimer (eds.). Écrire et transmettre en Inde classique. Paris : 81–105. — (forthcoming). “Organizational Levels in Vedic Ritual: The Problem of the Meaning of the Nihnavana-Rite”. In: Asko Parpola (ed.). The Proceedings of the Vedic Section of the 12th World Sanskrit Conference, Helsinki 2003. Houseman, Michael & Carlo Severi 1994. Naven ou le donner à voir: essai d’interprétation de l’action rituelle. Paris: Ed. de la Maison de Sciences de l’Homme. — 1998. Naven or the Other Self: A Relational Approach to Ritual Action. Leiden: E.J. Brill. Hubert, Henri & Marcel Mauss 1899. “Essai sur la nature et fonction du sacrifice”. Année sociologique 2: 29–138. — 1964 [1899]. Sacrifice; Its Nature and Function. Translated by W.D. Halls. Foreword by E.E. Evans-Pritchard. Chicago: University of Chicago. Humphrey, Caroline & James Laidlaw 1994. The Archetypal Actions of Ritual. A Theory of Ritual Illustrated by the Jain Rite of Worship. Oxford: Clarendon Press. Kiparsky, Paul & Frits Staal 1969. “Syntactic and Semantic Relations in Pāṇini”. Foundations of Language 5: 83–117. Knipe, David 1986. Review of Staal 1983. Journal of Asian Studies 45: 355–358. Kooij, Karel R. van 1989. “Indra’s Heaven in Buddhist Art”. In: Debala Mitra & Gouriswar Bhattacharya (eds.). Nālinī Kānta Śatavārṣikī: Dr. N.K. Bhattasali Centenary Volume (1888–1988): Studies in Art and Archeology of Bihar and Bengal. Delhi: Sri Satguru Publications: 97–117 (Sri Garib Dass Oriental Series 83). Kreinath, Jens 2004. “Meta-Theoretical Parameters for the Analysis and Comparison of Two Recent Approaches to the Study of the Yasna”. In: Michael Stausberg (ed.). Zoroastrian Rituals in Context. Leiden: E.J. Brill: 99–136. Lakoff, George 1971. “On Generative Semantics”. In: Danny D. Steinberg & Leon A. Jakobovits (eds.). Semantics: An Interdisciplinary Reader in Philosophy, Linguistics and Psychology. Cambridge: Cambridge University Press: 232–296. Lawson, Thomas E. & Robert N. McCauley. 1990. Rethinking Religion. Connecting Cognition and Culture. Cambridge: Cambridge University Press. Lévi, Sylvain 1898. La doctrine du sacrifice dans les Brāhmaṇas. Paris: Ernest Leroux (Bibliothèque de l’École des Hautes Études Sciences Religieuses 11). — 2003. La doctrine du sacrifice dans les Brāhmaṇas. Turnhout: Brepols (Bibliothèque de l’École des Hautes Études Sciences Religieuses 118). [Réédition de Lévi 1898 avec Préface de Louis Renou (pour la réédition de 1966) et Postface de Charles Malamoud.]


61

Malamoud, Charles 1989. Cuire le Monde : Rite et Pensée dans l’Inde ancienne. Paris : Editions La Decouverte. — 1996. Cooking the Worls: Ritual and Thought in Ancient India. (Transl. from the French by David White). Delhi: Oxford University Press. Manjul, V.L. 1997. “The Hitherto Forbidden Realm”. Manushi 99: 38. Michaels, Axel 2004. “Indology and the Cultural Turn”. In: Angelika Malinar & Johannes Beltz & Heiko Freese (eds.). Text and Context in the History, Literature and Religion of Orissa. Delhi: Manohar: 457–481. — 2005. “General Preface to the ‘Heidelberg Studies in South Asian Rituals’”. In: Jörg Gengnagel & Ute Hüsken & Srilata Raman (eds.). Words and Deeds: Hindu and Buddhist Rituals in South Asia. Wiesbaden: Harrassowitz Verlag: 7–13. — 2006a. “Ritual and Meaning”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing Ritual. Vol. 1: Issues, Topics, Approaches, Concepts. Leiden: E.J. Brill: 247–264 (Studies in the History of Religions 114/1). — 2006b. “The meaning of the meaningless of rituals”. In: Kreinath, Jens & Jan Snoek & Michael Stausberg. Theorizing Rituals, Vol. 1: Issues, Topics, Approaches, Concepts. Leiden: E.J. Brill: 247-261. — 2007 “‘How do you do?’ – Vorüberlegungen zu einer Grammatik der Rituale”. In: Heinrich Schmidinger & Clemens Sedmak (eds.). Der Mensch ein “animal symbolicum”? Sprache – Dialog – Ritual. Darmstadt: Wissenschaftliche Buchgesellschaft: 239–258. Penner, Hans H. 1985. “Language, Ritual and Meaning”. Numen 32: 1–16. Picardi, Eva 1977. “Some Problems of Classification in Linguistics and Biology, 1800– 1830”. Historiographia Linguistica 4: 31–57. Rappaport, Roy A. 1999. Ritual and Religion in the Making of Humanity. Cambridge: Cambridge University Press. Rau, Wilhelm 1997. “The Earliest Literary Evidence for Permanent Vedic Settlements”. In: Michael Witzel (ed.). Inside the Texts – Beyond the Texts: New Approaches to the Study of the Vedas. Cambridge: Department of Sanskrit and Indian Studies, Harvard University: 203–206 (Harvard Oriental Series, Opera Minora 2). Raven, Ellen 1994. Gupta Gold Coins with a Garuḍa-Banner: Samudragupta – Skandagupta. Groningen: Egbert Forsten. Renou, Louis 1936. “Sylvain Lévi et son œuvre scientifique”. Journal asiatique 228: 1–59. — 1937 [1996]. “Sylvain Lévi et son œuvre scientifique”. In: Jacques Bacot (ed.). Mémorial Sylvain Lévi. Paris: Paul Hartmann: xi–li. [Re-published in: Eli Franco (series ed.). Mémorial Sylvain Lévi. Delhi: Motilal Banarsidass (Landmarks in Indology 1).] — 1961. Études védiques et Pāṇinéennes. Tome 9. Paris: E. de Boccard. Robins, R.H. 19792. A Short History of Linguistics. London: Longman. Russell, Bertrand 1908. “Mathematical Logic as Based on the Theory of Types”. American Journal of Mathematics 30: 222–262. — 1950 [1992]. An Inquiry into Meaning and Truth. London: George Allen & Unwin [Routledge].

62

Jan E.M. Houben

Scharfe, Hartmut 1977. Grammatical Literature. Wiesbaden: Harrassowitz (A History of Indian Literature 5/2). Schechner, Richard 1986. “Wrestling against Time: The Performance Aspect of Agni”. Journal of Asian Studies 45: 359–363. — 1987a. “The Future of Ritual”. Journal of Ritual Studies 1: 5–33. — 1987b. “A ‘Vedic Ritual’ in Quotation Marks”. Journal of Asian Studies 46: 108–110. — 1993. The Future of Ritual: Writings on Culture and Performance. London: Routledge. Snoek, Jan A.M. 2004. “‘Initiation’ in Theory and in Zoroastrianism”. In: Michael Stausberg (ed.). Zoroastrian Rituals in Context. Leiden: E.J. Brill: 77–98. Staal, Frits 1966. “Indian Semantics I”. Journal of the American Oriental Society 86: 304– 311. — 1979. “The Meaninglessness of Ritual”. Numen 26/1: 1–22. — 1980. “Ritual Syntax”. In: Masatoshi Nagatomi et al. (eds.). Sanskrit and Indian Studies: Essays in Honor of Daniel H.H. Ingalls. Dordrecht: Reidel: 119–142. — 1982. The Science of Ritual. Poona: Bhandarkar Oriental Research Institute. — 1983. “Ritual Structure”. In: Staal et al. 1983: 127–134. — et al. (eds.) 1983. AGNI. The Vedic Ritual of the Fire Altar. Vols. 1–2. Berkeley: Asian Humanities Press. — 1987. “Professor Schechner’s Passion for Goats”. Journal of Asian Studies 46: 105–108. — 1989. Rules without Meaning: Ritual, Mantras and the Human Sciences. New York: Peter Lang. — 1993. “From Meaning to Trees”. Journal of Ritual Studies 7: 11–32. Strenski, Ivan 1991. “What’s Rite? Evolution, Exchange and the Big Picture”. Religion 21: 219–225. — 1997. “The Social and Intellectual Origins of Hubert and Mauss’s Theory of Ritual Sacrifice”. In: Dick van der Meij (ed.). India and Beyond: Essays in Honour of Frits Staal. London: Kegan Paul: 511–537. Tambiah, S.J. 1979. “A Performative Approach to Ritual”. Proceedings of the British Academy 65: 113–169. Thieme, Paul 1957. “Vorzarathustrisches bei den Zarathustriern und bei Zarathustra”. Zeitschrift der Deutschen Morgenländischen Gesellschaft, 107: 67–104. Tylor, E.B. 1871. Primitive Culture: Researches into the Development of Mythology, Philosophy, Religion, Art and Custom. London: Murray. Verpoorten, Jean-Marie 1991. “Condition profane et condition sacrificielle selon SāyaṇaMādhava”. Acta Orientalia Belgica VI (Humana condicio – La condition humaine): 151–170. Weber, Albrecht 1865. “Über Haug’s Aitareya-Brāhmaṇa (a. aus der Bombay Native Opinion; b. aus der Saturday-Review; c. vom Herausgeber)”. Indische Studien 9: 177– 380. Winternitz, Moritz 1908. Geschichte der indischen Literatur. Vol. 1: Einleitung; der Veda; die volkstümlichen Epen und die Purāṇas. Leipzig: Amelangs Verlag. — 1922. Geschichte der indischen Literatur. Vol. 3: Die Kunstdichtung; die wissenschaftliche Literatur; neuindische Literatur; Nachträge. Leipzig: Amelangs Verlag.


63

Witzel, Michael 1992. “Meaningful Ritual. Structure, Development and Interpretation of the Tantric Agnihotra Ritual of Nepal.” In: A.W. van den Hoek & D.H.A. Kolff & M.S. Oort (eds.). Ritual, State and History in South Asia. Essays in Honour of J.C. Heesterman. Leiden: E.J. Brill: 774–827.

Olga Serbaeva Saraogi

When to Kill Means to Liberate Structure and Meaning of Two Types of Ritual Actions in Vidyāpīṭha Texts Being intrigued by the idea of Axel Michaels to attempt a rule-based description of rituals, I wanted to know how śaiva tantric material could contribute to such an ambitious project.1 The aim of this article is to discover the relation between structure and meaning of the ritual on the basis of two kinds of ritual actions that appear in śaiva tantric texts that were circulating in India approximately between the sixth and the eleventh centuries. The two chosen types of rituals have three features in common: (1) they cannot be easily observed and they will be addressed on the basis of textual sources only; (2) they show preferences for non-verbal communication, and (3) both rituals are marked by belief in direct contact with “superhuman agency”, which makes them emotionally intense. The first of the two ritual actions is a part of śaiva initiation (dīkṣā), the ritual essential for and highly respected by śaiva tantric tradition.2 It is a strictly codified cultural action through which the disciple is liberated from his (or her) bondage by the guru. In Vidyāpīṭha texts, the initiation takes the form of a direct experience of a particular state of consciousness provoked by the guru who is believed to be acting from inside the very body of the disciple. The disciple, in a possession-like state, may have visionary experiences or may spontaneously fold his body into various positions encoded by tradition, or he may just convulse and fall on the ground. This sort of initiation has been discussed in the articles of Alexis Sanderson and those of other scholars.3

1 Michaels in this volume: 15–36. This article was written with a generous support of UFSP “Asien und Europa”, University of Zurich, to whom I would like to express my gratitude. 2 General lines of śaiva dīkṣā of the Saiddhāntika type, which are also applicable to more radical śaiva traditions, can be found in Brunner 1963. 3 Published and unpublished works of Sanderson, especially 2007: 231–442 and 551–582; Wallis 2008; Takashima 1992; Padoux 1999 and 1990; Lidke 2005; Smith 2006.

66


The second kind of ritual action to be discussed here is much less known,4 and appears to be an imaginary one as it is performed by human or not so human illogical and wild beings, called yoginīs in śaiva tantric texts.5 Fear, and not respect, appears to be the predominant emotion here. The ritual action is also sudden and wild. The briefest picture of such a ritual action can be found in the Rājataraṅgiṇī, “The Chronicle of the Kings of Kashmir”, which narrates the life and death of the king Baka:6 “There the king passed sixty-three years and thirteen days as ruler of the earth. Then a certain sorceress [yogeśvarī], Bhaṭṭā by name, having assumed the appearance of a lovely woman, approached the king one evening. Losing his sense over her various captivating words, he joyfully accepted an invitation to view the wonders of [her] sacrificial feast [yāgotsavamāhātmyam]. Then when in the morning the sovereign came to that place followed by his hundred sons and grandsons, she made of him a sacrificial offering to the ‘circle of the goddesses’ [devīcakropahāratām]. To this day there is seen on a rock the double impression of her knees, showing [where], on attaining by that act supernatural power, she has risen to the sky.”7 4 It is interesting to note that the stories about yoginīs (the term is not used) draining blood or killing appear in the writings of Sleeman 1844: 90–96; Crooke 2003 throughout; Knowles 1888: 197–203 and 211–227, 412f, as early as the middle and the end of the nineteenth century. However, these stories were not related to tantric practices by these authors as the term “tantric” meant at that period a very restricted set of śākta rituals. More details will be provided on this subject in my Habilitation thesis. 5 For more information on yoginīs, see Serbaeva 2006 and (forthcoming)b. 6 Stein 1961: 49–50: Book 1, ch. 22, v. 330–336. 7 RT 1.22.330–334: tatra triṣaṣṭir varṣāṇāṃ satrayodaśavāsarā | atyavāhyata bhūpena tena pṛthvīṃ praśāsatā || atha yogīśvarī kācid bhaṭṭākhyā rajanīmukhe | kṛtvā kāntākṛtiṃ kāmyām upatasthe viśāṃ patim || tayā manoharais taistair vanair glapitasmṛtiḥ | sa yāgotsavamāhātmyaṃ draṣṭuṃ hṛṣṭo nyamantryata || putrapautraśatopetaḥ prātas ca tatra tato gataḥ | cakravartī tayā ninye devīcakropahāratām || karmaṇā tena siddhāyā vyomākramaṇasūcakam | jānumudrādayaṃ tasyā dṛṣadyadyāpi dṛśyate ||. This passage was first translated into French by Troyer already in 1840. This passage can be seen as a reference to a practice described in TST 7.97–100.1, where a sādhaka is supposed to get the ability to fly immediately after partaking of the flesh of a seven-times born paśu: nātaḥ parataraṃ devi sādhanaṃ sāmudāyikam | ekavīravidhānena viśeṣāt parameśvari || japataḥ prayato mantrī śuddhirāśuḥ prajāyate | devyā śuddhasya suśroṇi yacchanti paśum uttamam || saptajanmodbhavaṃ caiva tajjuṣṭacaruko’thavā | tena jugdhena deveśi tattulyas tu prajāyate || vicared devataiḥ sārddhaṃ kṣaṇād gaganago bhavet |. The translation of the unedited tantric text cannot be anything more than a tentative: “O goddess, there is no practice higher than this in the universe. O Parameśvari, the mantrin [here a person who tries to win over the power of mantras by the practice described in the tantras] intent on the repetition of the mantras, especially by means of the solitary practice [ekavīravidhānena],

When to Kill Means to Liberate

67

This story has another meaning when analyzed with the help of the śaiva tantric texts and I will argue that the king, when killed, was, at the same time, initiated and liberated. In order to do this, it is necessary to restore and compare the sequences of ritual elements in the named two types of ritual actions described in a number of Mantramārgic texts, ranging from the mild Svacchandabhairava tantra to the extreme ritual practices prescribed in the Brahmayāmala and the Jayadrathayāmala.8 The analysis provided in this article shall be limited to the brief, yet emotionally charged ritual sequence, when both a śaiva guru and a yoginī do exactly the same thing, namely, manipulating consciousness from inside the disciple’s (in the case of initiation), or the victim’s body (in the case of the yoginī-related rituals)9. It is important to note that in this type of texts, a person who has not been initiated and a sacrificial animal are named by exactly the same word – paśu. The hypothesis is that since in both cases the sequence of ritual elements is similar, the meaning of the two kinds of ritual actions must be similar too. Which ritual elements, then, are at stake in such a context? As it is presented in the article of Axel Michaels,10 those elements may be practically anything ranging from an object, an act, to even a word. Not being able to include in this analysis such heterogeneous elements as maṇḍalas, mantras, particular places, or the various objects used in the two ritual actions, the focus will thus be on one type of ritual element which appears to be the most important for the chosen context, namely the mudrās. will quickly achieve purification [here purification seems to imply the ‘maturation’ process]. O Beautiful [suśroṇi], the goddesses [devyā, in the absolute majority of cases these are the yoginīs [usually plural] who help the sādhaka to get the victim] will offer/show/hold the supreme victim who has been ‘born seven times’ [saptajanma, the term qualifying the highest sort of victim] in relation to the purification [śuddhasya, though unclear, seems to imply “sādhaka being purified, the yoginīs…”], or [even] the pleasing caru [transgressive transformative substance to be eaten in ritual practice and usually coming from such kinds of victim]. By offering that [victim or caru] [jugdhena as a form of juhana, Edgerton 1985: 244], o goddess, [the sādhaka] will reach the identity [tulya] with that [by “that”, tat, in such kind of expressions, is usually meant ‘the highest state’ or ‘a being representing that state’], he will move together with the devatās [usual term for yoginīs or khecarīs] and immediately become a sky-goer.” 8 I wish to thank Alexis Sanderson for his unpublished editions of a number of these sourcetexts, which he so generously sent me in 2004. For the general classification of śaiva currents see Sanderson 1990. 9 In the case of śaiva initiation, the term used is mostly yojanikā, occasionally āveśa and stobha are also encountered. For the correspondence of such terms in various śaiva traditions, see Dwivedi 1990: 50. In the case of the yoginī-related rituals, besides the ‘extracting of the nectars’ [amṛtākarṣaṇa], it appears to be called melaka or melāpa. The haṭhamelaka in the majority of cases requires a human paśu to be offered to the yoginīs according to JY. 10 Michaels 2008.

68


Mudrā, generally translated as “a sign, a special combination of fingers”, has in śaiva tantric texts a number of different meanings, which can be classified as follows:11 1. An object-mudrā is an object (arm, lotus, severed head, etc.) held by the deity and displayed by the initiated with a particular combination of fingers, used in pūjās, etc. 2. A yoginī-related mudrā is a symbol drawn by a yoginī on her house or the sign of an object-mudrā particularly pleasing to her. Both enable others to recognize her type and ‘family’,12 which is important for sādhakas in order for them not to be killed and so that they may be able to reach the kind of siddhis they are looking for. 3. A “secret sign” is a mudrā serving for communication between the initiated and the yoginīs. This communication is non-verbal and these mudrās correspond to a limited and predefined number of questions and answers. 4. Finally, there are mudrās that are believed to provoke a modified state of consciousness in the practitioner himself or in another person, often without the agreement of the latter. In this case, mudrās are complicated mental acts of concentration accompanied by particular body positions or movements and sometimes breath-control. These mudrās are believed to result in a spectacular mental or even physical response: they allow the practitioner to see the yoginīs or to become Bhairava, or they can make someone execute the will of the initiated, for example, sleep, eat or fall whenever ordered.13 It is this type of “hyp11 A detailed analysis of mudrās, their classifications and meanings in JY, was presented at the 14th World Sanskrit Conference, 1–5 September 2009, Kyoto, Japan. 12 The classification of śaiva yoginīs is discussed in detail in Serbaeva 2006. 13 The clearest example here is a commentary by Kṣemarāja to NT 20.30: “tatstha aśnīyāt pibed gacchet tiṣṭhet supyāt tāvad yāvat samāsāditasakalaceṣṭāphalaḥ sa yogī saṃpadyate tatastam ānayed visṛjen mohayed unmīlayed āpūrayed viśiṣṭaṃ vā sthānaṃ prāpayet ity ādi |” This means that the yogi, through his total control of the aspects [of the victim], can control the victim, for example, can bring him/her to a particular place, make him inhale, open the eyes, can cheat him, make him drink or pour something forth, etc., all these actions being made possible by identification with the victim, thus it apparently works when the yogi himself is eating, drinking, walking, standing, sleeping … I thank Alexis Sanderson for having gone through this difficult chapter with me in July 2007. All mistakes are mine. This quotation can probably be better understood with the help of an image of the practice of a Russian witch given in Kuprin’s Olesya, p. 337-338: “I really did not understand her completely. But, if I am not mistaken, this particular trick consists in the fact that she, going after me step by step, leg-in-leg, and holding her eyes on me steadily, at the same time, tries to copy everything, even the slightest of my movements, so to say, and identifies herself with me. Having gone a few steps in this way, she starts to imagine in her mind a rope at some distance in front of me, hung across the road arshin high from the ground. At the moment when I would have touched with my leg this imaginary rope, Olesya suddenly makes a


69

notizing” or “possessing” mudrā that is used by both a guru performing initiation and a yoginī killing a victim or draining his14 blood.15 Such mudrās appear in chapters dealing with initiation already in the Svacchanda tantra 4 (pāśa, aṅkuśa, saṃhāra, etc.), but it is the later text of the Jayadrathayāmala 4.2 that provides the longest list of them. About two hundred and thirty different mudrās in forty śaiva tantric texts were found, among them about seventy are clearly believed to provoke particular states of consciousness. During the most intimate part of śaiva initiation – yoganikā (joining), or nāḍīsaṃghaṭṭa (joining of the channels) – when the guru links the disciple’s consciousness with the highest tattvas or states, and also in the violent actions of yoginīs, the same special “possessive” mudrās are used.16 The information concerning the actions of the guru is scattered around in chapters 3 and 4 of the Svacchandabhairava tantra, and when assembled it gives the following picture:17 after preliminary rituals, the guru protects the disciple with the kavaca-mantra, he comes out of his own body by exhalation (recaka), and enters the body of the disciple through one of the “doors”.18 Yet what are these doors?

14

15

16 17

18

falling movement, and then, according to her words, even the most solid man would inevitably fall …”. (My translation from Russian). The gender of the victim requires some precision. In the case of initiation, the absolute majority of passages describe it as the transmission of the highest state from a male guru to a male disciple. In the case of the yoginī-related practices, the victim is masculine as well in all cases known to me. Women do not have problems with yoginīs. However, in a third possible case, which shall not be discussed here, namely, when sādhaka copies the procedures of yoginīs in order to drain blood from a victim or to put him or her in control of a victim, the victim can be male as well as female. André Padoux was the first to publish on this question. He studied these sorts of mudrās on the basis of the texts of the Tripurā cult. As for the JY, Sanderson was the pioneer (Padoux 1990: 72): “It appears still more, I am told by Alexis Sanderson, in such tantras as the Jayadrathayāmala, in one of whose ṣaṭkas a large number of kāpālika mudrās are described, their use being to induce possession by the deity or, as this text says, to cause the entering into, or the being possessed by, the particular “flavour” of the deity: it leads to rasāveśa”. The precise reference is JY 3.38.167cd–179ab. See fig.1 (page 88), which is meant to be read from top to bottom. The left part of the table graphically represents the mudrās, mantras, and other means used by the guru in the context of śaiva initiation. This part was reconstructed on the basis of the following references: SVT 3.49–53ab, 3.148–156, 3.165cd–172, 3.179–180, 4.69cd–74, 4.110–112, 4.134–135, 4.161cd–162 and 166, 4.174cd–177, 4.299–315, 4.409–410ab. These should be supplemented by the references from later and more radical tantras: MVT 9.58ab– 60, 9.76–79ab, TST 9.128–130, 9.132cd–148, 9.296–298, 9.300–315. Exhalation [recaka] and the entrance into the heart of the disciple is referred to in: SVT 3.169cd: huṃkāroccārayogena recakena viśedd hṛdi ||. “[The guru] should enter the heart [of the disciple being initiated] by means of utterance of the sound ‘Huṃ’ on exhalation”; 4.69cd: recakena tato gatvā śiṣyadehe viśedd hṛdi || “[The guru] should enter the heart in the disciple's body having gone there [tatas refers here to the act of leaving of one’s own body

70


Few variants are possible: (1) either these are the openings of the central channel, or (2) the organs of the senses, or (3) these can theoretically also be cakras.19 The guru then moves to the heart of the disciple where he cuts out the principle of consciousness (variously named as caitanya, cit, or jīva, or rarely ātman) with the help of a ‘cutting mantra’ (astra).20 He further makes that principle of consciousness and entering the body of the disciple]”; 4.110: śiṣyaṃ saṃprokṣya cāstreṇa tāḍayed astram uccaran | recakenātmano gatvā chindyāttasyāsinā hṛdaḥ || “Having consecrated the disciple, he should ‘beat’ him [tāḍayet] with a ‘weapon’ [astra, it is a particular mantra used in initiation] and uttering ‘weapon’, having gone [out] with one's own principle of consciousness [ātman] by exhalation [recakena], he should pierce his [the disciple’s] heart with the ‘knife/sword’ [asi, might refer to the same ‘weapon’ mantra as well as to a cutting gesture, a mudrā, called kṣurikā.]”; 4.174cd–175: tataḥ śivāmbhasā śiṣyaṃ prokṣya cāstreṇa tāḍayet || tenaiva cāstrabhūtena huṃphaṭkārayutena tu | ātmano recakenaiva śiṣyadehe viśedd hṛdi || “Afterwards [tatas] having consecrated the disciple with the ‘water of Śiva’ he should ‘beat’ him with ‘weapon’, by the same ‘weapon’ joined with the sounds ‘Huṃ’ and ‘Phaṭ’ [he should] with his own principle of consciousness [ātman] on exhalation, enter the heart in the disciple's body”. 19 The dvāras as an expression appear only in the commentary of Kṣemarāja on NT 20.35: sādhyasya mūrdhadvāreṇa niṣkramya […], and in ŪK 2.230ab: tattvanyāyena [-yona] tenaiva dvāreṇa praviśed balāt |. The first option is supported by the TST 16.199–201ab where some sorts of lower yoginīs enter the victim from below, see also KSS 6.6.153–169, BKM 7.1.430–437. The upper opening is brahmarandhra, a most common door to enter and to exit in these sorts of procedures. The second option, however, finds support in SVT 4.409–410ab: śivena sahacāritvād ācāryas tena cocyate | tasya darśana-saṃbhāṣāsparśanāt smaraṇād api || bhavaty evaiśvarī vyāptir na bhavet tadadhogatiḥ |, where the guru is said to be able to provoke a state of pervasion [vyāpti] with a glance, a word, a touch or a remembrance, all related to the organs of senses. The third option appears rarely. In JY some yoginīs are said to enter by the navel and reach the heart, both being important body cakras. Nine doors are mentioned in a vampiric procedure in JY 2.17.886cd, while JY 4.12.41 and 44 name yogadvāra which is brahmarandhra. 20 A variety of ‘weapon’ mantras and mudrās are to be found in śaiva tantric texts. I present here only a small selection: SVT 3.168cd–169ab: śivāmbho ‘streṇa samprokṣya śiṣyasya hṛdayaṃ punaḥ || tāḍayed astrapuṣpeṇa hṛdi citsaṃhṛtā bhavet | “Having consecrated the heart of the disciple with ‘the water of Śiva’ and the ‘weapon’-mantra again, [the guru] should beat him with the flower [‘charged’] with ‘weapon’-mantra, [and his] consciousness [cit] will be absorbed in the heart.”; 3.171ab: hṛtsthaṃ chittvāstrakhaḍgena humphaṭkārāntajātinā | “At the end of [the utterance] of the sounds ‘Huṃ’ and ‘Phaṭ’, having cut off [the principle of consciousness] stationed in the heart by means of the sword of ‘weapon’-mantra …”; 4.70–71ab: oṃkārādi śivaṃ japtvā astramantraṃ phaḍantagam | viśleṣakaraṇaṃ kṛtvā caitanyasya vidhānataḥ || chedayed astramantreṇa kavacenāvaguṇṭhayet | “Having repeated the śaiva mantra starting with ‘Oṃ’ and ending with ‘Phaṭ’, having performed the ‘separation’ [viśleśa] of the consciousness [caitanya] accordingly, [the guru] should cut off [the principle of consciousness of the disciple] by means of the ‘weapon’-mantra and protect it with the ‘shield’-mantra [kavaca]”; 4.110cd, cited in note 18; 4.161cd–162ab: prokṣaṇaṃ tāḍanaṃ cheda ākarṣagrahaṇe tathā || dhāmnāpūrya kumbhayitvā chittvātha grāhayet punaḥ | “He [the guru] should grasp again [the principle of


71

one with himself, and this mingled ‘piece’ of consciousness is subsequently visualized like a ‘ball’ (kadambagolaka, golaka), as ‘an undifferentiated form’ (lolībhuta), ‘a unity’ (ekībhūtā).21 With the help of the aṅkuśa-mudrā, representing a hook used to drive elephants, the guru then drags this ‘ball’ up along the central channel until the ‘end of twelve’.22 There he grasps the mingled consciousness and consciousness] having performed the consecration, ‘beating’, cutting, attraction and grasping, having inhaled and held it [while] arresting the breath [kumbhayitvā], in accordance with the rule [dhāmna], and having cut it off”; 4.176ab: astramantreṇa saṃchedya viśeṣāśleṣyāstreṇa karṣayet | “Having cut it off with the ‘weapon’-mantra, and having separated it with ‘weapon’ [probably mudrā] he should attract it”; TST 9.304: nābhihṛtkaṇṭh-tālvante bindunāde ca suvrate | chedayed astrarājena churikā yāthavā priye || “O beloved, [the guru] should cut it [the principle of consciousness] off by means of the ‘king among weapons’ [mantra] and/or churikā [the ‘knife’-mudrā, kṣurikā], in the navel, heart, throat, palate [tālu] and the ‘end’ [the ‘end of twelve’, dvādaśānta, one of the centres of the subtle body believed to be situated twelve fingers above the head]”; and v. 315: kathitā tu mayā bhadre kṣurikā marmachedanī | chedayeta tato marmaṃ yojayet parame pade || “O auspicious [bhadre], the ‘knife’ cutting off the marmas [sensible points, here practically the places where the consciousness is ‘fixed’], has been explained by me. [When] [the guru] cuts off a sensible point, he afterwards [tatas], joins [the consciousness of the disciple] with [lit.; in] the supreme state [parame pade]”. 21 SVT 3.170: nāḍīrandhreṇa gatvā tu caitanyaṃ bhāvayec chiṣoḥ | kadambagolakākāraṃ sphurattārakasaṃnibham || “Having gone through the opening of the nāḍīs [the channels of the subtle body] [the guru] takes care of the consciousness [of the disciple] as if it is of a baby, [he mingles it] into a ‘ball’ [of the size or shape] of a kadamba flower, extremely shining …”; 4.299cd: paścād ātmani saṃyojya lolībhūtaṃ vicintayet || “Afterwards, having joined [his and his disciple’s] principles of consciousness [ātmani], he should visualise it as a ‘unity’ [lolībhūta]”; and 301–302: samānena samākṛṣṭā ekībhūtā bhavanti tāḥ | […] prāṇe samarasīgatāḥ || nāḍayas tu suṣumnāyām ekībhūtā vyavasthitāḥ | tato vai uccaren mantraḥ nāde līnaṃ vicintayet || “Attracted by their common nature [samānena] they [two principles of consciousness] become one [ekībhūtā], […], they reach the ‘same taste’ [state of perfect unity, samarasa] in the vital breath, they are present as one in the channels and in suṣumnā [the central channel] [in particular], then, [the guru] should utter the mantra and imagine how [they] melt in the sound … “. 22 SVT 3.51cd: dvādaśāntaṃ paraṃ nītvā karastho mantravigrahaḥ || “Having guided it [up] to the supreme ‘end of twelve’, the mantra-body is in the hand …”; 3.171cd–172ab: dhāmnā cāṅkuśarūpeṇa karṣec chaktyavadhi kramāt || dvādaśāntaṃ tu saṃgṛhya sampuṭya hṛdayena tu | “Accordingly [dhāmna], by means of [gesture] having the form of the ‘elephant goad’ [a particular mudrā], he [the guru] attracts it [the consciousness] by order [chaktyavadhi is unclear to me], having grasped it at the ‘end of twelve’ and having enveloped it by [his own] heart …”; 4.71cd: aṅkuśena samākṛṣya dvādaśānte tu kārayet || “Having attracted [the principle of consciousness] by means of aṅkuśa-mudrā he should perform [grasping] at the ‘end of twelve’”; 4.111ab: dhāmnākṛṣya tad ātmānaṃ dvādaśānte nidhāpayet | “Having attracted it accordingly [dhāmnā], he should cause the principle of consciousness [ātman] to be fixed in the ‘end of twelve’”; MVT 9.77ab: ekīkurvañ chanair gacched dvādaśāntam ananyadhīḥ | “[Having] made [them] one he slowly goes to the ‘end of twelve’ and not otherwise …”. TST 9.139ab: nīyate dvādaśānte tu […] “He leads it to the ‘end of twelve’”, etc.

72


returns to his own body, usually upon inhalation (pūraka), making the disciple experience there, in the guru’s body, the highest state. After that, the guru finally restores the principle of consciousness to the disciple with a saṃhāra-mudrā by a reversed procedure.23 The yoginīs are said to act in a very similar way according to the Netra tantra 20 and various vampiric procedures described in the Tantrasadbhāva and the Jayadrathayāmala.24 Having attracted a victim by way of a variety of means described in TST 16,25 the yoginī (or sādhaka doing it to please the yoginīs) fixes him with the mahāpañjara-mudrā.26 Then, according to the Netra, by means of repeated movements of ‘catching and releasing’, ‘going to and fro’,27 ‘helādolā’ in JY,28 she 23 SVT 3.172cd: saṃhāramudrayā yojyaṃ sūtre nāḍīprakalpite || “[The guru] joins [the principle of consciousness] to the rope [sūtra] of the subtle channel [nāḍīprakalpite, represented in initiation also by a real rope considered to be the externalisation of that channel] by means of the gesture of reabsorption [saṃhāra-mudrā]”; 3.179cd: huṃphaṭkārāntayogenāgṛhya saṃhāramudrayā || “Having grasped it by means of a procedure related to the sounds ‘Huṃ’ and ‘Phaṭ’, he, by means of the gesture of reabsorption…”; 4.72–73ab: tatrasthaḥ pudgalo grāhya saṃpuṭyaiva dhruveṇa tu | saṃhāramudrayā samyak pūrakeṇa viśedd hṛdi || saṃkṣubhya sarasīkṛtya recayet pudgalaṃ punaḥ | “Having suddenly grasped the principle of person [pudgala] stationed there and having enveloped it, [the guru] enters [his own] heart by means of the gesture of reabsorption and on inhalation [pūraka]” accordingly, having shaken up [saṃkṣubhya, unclear] and having melted [lit. having made it with rasa] [the principle of consciousness], he should exhale the ‘person’ again”, and 74cd: saṃhāramudrayoddhṛtya śiṣyasya hṛdi yojayet || “Having extracted it by the gesture of reabsorption, [the guru] joins it in the heart of the disciple.” 24 See the right side of the table in fig.1 (p. 88). 25 These means include, for example, standing on the head naked, marking a “magical territory” in various ways, TST 16.186cd ff, 204 ff, 208, marking a victim by offering him something, like the “lotus” which is in reality a human hand in KSS 14.4.15–61, BKM 13.1.189–196, or a human hand or bone in TST 16.212 ff, or performing a transgressive act such as licking her own urine from a human skull, for details see Serbaeva 2006 and (forthcoming)b. For the question of gender, see note 14. 26 JY 2.17.680cd–681ab: … paśuḥ [paśum, Serbaeva] paśyed agrasthaṃ śaktivedhanāt | taṃ mudrayed vajravṛtyā mahāpañjarayogataḥ ||. “He [the sādhaka] will see a victim standing in front of him because of his being pervaded by the power, he ‘fixes/seals’ him [the victim] by means of the procedure of ‘the great tent’ and ‘the indestructible covering’”. He uses the saṃhāra-mudrā to drain the blood later in the passage. 27 This movement appears already in SVT in the context of initiation, see SVT 3.151: grahaṇākarṣaṇārthaṃ tu gṛhṇan muñcan punaḥ punaḥ | dīkṣākāle yataś caivaṃ tadarthaṃ nāḍisaṃ[g]hatiḥ || “In order to perform ‘grasping’ and ‘attraction’, [one should do] again and again [the procedure of] ‘seizing and releasing’”; JY 4.3.448cd: gamāgamāniyogena … “By means of [the procedure of] ‘coming and going’ …”; JY 4.31.30: nābhihṛtpadmamadhye tu amṛtākṛṣṇakarṣayet | hṛtpadmakaṇṭhakūpe tu gamāgatanane’tra ca || “He [the sādhaka] extracts the nectars in the centres [lit. lotuses] of navel and heart, in the heart and the throat by means of ‘coming and going’”. 28 JY 2.8.27cd, 2.9.52ab, 2.9.56ab, 2.17.606cd, 2.23.71.cd, 3.20.187cd, 4.57.13ab. The term as such does not appear in any Sanskrit dictionary, however, it can be found in Bengali, see


73

forges a contact with the victim with whom she establishes a strong feeling of identity. Having entered the body of the victim through one of the ‘doors’ and having settled in some place in the ‘central channel’, usually the ‘end of twelve’, she sucks the victim’s blood.29 However, if she is a type of yoginī that kills the victim,30 she goes straight to the heart, and after having cut out the principle of consciousness (jīva) with the help of the ‘knife-mudrā’ (kṣurikā),31 she goes on to envelop and melt it. Forster 1802: 339, where it means “to reel, stagger, toss as a boat”, while in Marwari it means “to rock a cradle”, according to Pradeep Saraogi. All meanings refer to repetitive movements in opposite directions. 29 The references to such a vampiric practice appear everywhere in the texts. A small selection includes BY 99 where the actions of yoginīs are discussed; TST 3.210.1, which lists amṛtākṛṣṭivijñānaṃ among siddhis, and links Cāmuṇḍā to such practices in chapter 17. TST 16 contains important references throughout, and it provides a sort of “religious” (for the yoginīs) explication of such practices: in order to go higher in the cosmic hierarchy, the yoginīs attract their victims by use of ruses and drink their blood. The killing of a victim constitutes for the yoginīs one of the possible ways of attaining a higher level, see v. 170: yogena tapasā vāpi svamantrasya ca darśanāt | yānti yogeśvarīsaṃsthāṃ tathā janmapaśor balāt || “They go to the domain of Yogeśvarī [here a supreme goddess], by yoga, tapas, realization of their mantra and killing a victim born a number of times [for this very purpose].” (Translated in Serbaeva 2006). Another chapter of the same text, TST 27, describes how the extraction of blood is to be done in practical terms, see v. 42–74 (numeration of verses of M. Dyczkowski): first, with a particular mantra given in v. 42–43ab, with which the practitioner should by means of upgoing breath go out [of one’s own body]: udānena prayogena guhyadvāre viśed bahiḥ || 43cd, he or she further reaches the brahmarandhra of a victim and attracts the five elements: brahmasthānaṃ tataḥ prāptaḥ pañcabhūtādi saṃharet | 44ab. Being stationed in ‘the end of twelve’, the yogi or yoginī visualises a flow of blood from the victim, which he or she extracts by means of moving in the central channel: dvādaśāntasthitaṃ bhānyas tatpaśuṃ marmabheditam || 45 śravantaṃ cintayed yogī raktaughaṃ raktayā saha | suṣumṇācārayogena ayānenā’pakarṣayet || 46. This practice allows to achieve all sorts of siddhis, v. 47– 48 and v. 67–73. 30 The information on yoginīs killing victims appears in śaiva tantric texts as soon as yoginīs appear there themselves. There are code-words for paśu in chapter 15, appended to the earlier material of the SVT, see SVT 15.12cd etc. The Yāmalas (especially BY 99 in the edition of Shaman Hatley), and the texts of early Trika abound in description of sacrificial practices, see TST 7, 16, and 21. NT 20 is entirely dedicated to the subject. The practices described concern not only the yoginīs, but also sādhakas who imitate them and it is even clearer in later texts that these are sādhakas who offer paśus or extract blood in JY 2.1.167–172, 2.9.50–60, 2.11.82–90, 2.17.602–635, 677–714, 880–891; 2.23.62–77; 3.8.304–328; 3.26.106–108; 3.27.47–50; 3.28.16–17, 39–46, 54–62, 91–92; 4.31.28–33. 31 It is still a mantra in JY 1.13.18ab, but becomes a mudrā in JY 4.1.51ab and 4.2.381–384. The other mudrās used in such contexts are the same ones as used in initiation: the pāśa- and aṅkuśa-mudrās. These, however, are used in JY for a number of different purposes ranging from draining blood to putting victims under control and killing, see, for example, JY 2.8.27cd and 32cd, 2.9.56cd, 3.20.193cd, 3.23.104ab.

74


Netra tantra 20 is a unique source explaining in detail what is going on inside the body of the victim whose ‘soul’ is being ‘extracted’ and it is important to present it here in detail.32 We shall try to follow chronologically the sequence of action of a yoginī. First she identifies herself with the highest state, represented by Bhairava in this tantra, which appears to be the basis of her power allowing her to enter the body of another. “One who having obtained that which is supreme, non-manifest, stable, unmovable and perpetual, by way of yoga, is to enter the bodies of others, and having become [one with] the supreme being he envelops the soul by [his own] soul by his power.”33 Next, the Netra describes a series of procedures by which the yoginī achieves ‘oneness’ with the victim, having entered his body. These include ‘binding’ the vital energy of another through the practice of 'repetitive movement' technically called ‘ball’ (golaka), being stationed in the body of the victim as if ‘non-manifest’. This ‘ball’ practice is to be performed for a period of time of one hundred mātras (mātrāśataṃ), which allows the yoginī to catch hold of the vital energy of the victim.34 In the mind of Kṣemarāja this part of the procedure – from leaving one’s own body to entering the body of another – is encoded by krodhinī and karaṅkiṇīmudrās.35 The next part, which would come under ‘enveloping’ in the context of śaiva initiation, here includes the penetration of the ‘sense organs’ of the victim by those of the yoginī and the isolation of the 'subtle body' of the victim that becomes ‘hidden’ inside the ‘subtle body’ of the yoginī: “… [he or she], having established strong ideas of identity with the other, first having entered the senses of that [person] with one’s own ‘sense organs’ 32 NT 20.27cd–35 with a commentary by Kṣemarāja. The passage makes explicit that not only a yoginī or other supernatural being may perform the practice but human yogis too (NT is marked in bold, the commentary is in normal). 33 NT 20.28-29ab: yattatparamam avyaktaṃ śāśvataṃ hy acalaṃ dhruvam | tatprāpya yogamārgeṇa praviśya paradehataḥ || paro bhūtvā svaśaktyā tu jīvaṃ jīvena veṣṭayet |, the unity of the power with the supreme consciousness is further underlined by Kṣemarāja: Tatprāpya paraṃ cinmayaṃ balaṃ … 34 Commentary of Kṣemarāja to verses 28-29ab: … golakābhyāsāsādita-samasta-tadrasopalam-bhaḥ svadeha evāvikāsasthityā prāṇākarṣāpakarṣābhyām asvatantrīkṛtaprāṇabalo yogī pāda-śākhābrahmarandhrata …, further in the commentary golaka seems to refer also to the means of transport between two bodies, the form which the ‘principle of consciousness’ takes to come out of the body: Ekatareṇa pathā golakasthityaiva parahṛdayaṃ praviśya paro bhūtveti … 35 See note 42.


75

and then, having made the body of the other one’s own through ordered ‘moving forward and backward’ of one’s own vibration, … having enveloped by his/her own vital energy, which is the power of ‘opened’ śakti, ‘the soul’, which is the ‘subtle body’ [puryaṣṭakaṃ] of another, by one’s own soul which is one’s own ‘subtle body’.”36 The Netra suggests the following stages: first ‘oneness’ with the victim allows entrance, ‘sameness’ with it bestows a certain degree of control, while when a yoginī resorts to the ‘supreme cause’ (Bhairava) being inside the victim's body, it apparently allows her to control the victim completely: “Having entered the place of the heart and as a result of entering below or above, [and] having achieved ‘oneness’ [with the other], [s]he should exercise there ‘sameness’, having resorted to the ‘supreme cause’, [s]he should exercise the ‘state of autonomy’.”37 Although the practices allegedly allowing someone to put another being under their control are very popular in śaiva tantras, and Kṣemarāja gives examples of how the victim carries out orders, the main idea here appears to be the fact that the independent power of the being should be ‘paralyzed’ or even ‘broken’ in order to proceed: “He breaks the power by the power of pervading one’s own essence. Having covered up one by another, he performs the cutting off of the śakti [of another]. ‘He breaks’ means he strikes in order to separate that power of vital energy by one’s own śakti, having stationed in the process pervasion, which is vīrya [something extremely powerful here] because of the power of pervasion.”38 The Netra tantra says that the rays of the victim’s sensory organs will further be, in the process of ‘melting’, replaced by those of the yoginī or other possessive being, who will further drag up all melted essence, seize it, and eventually return to his or her own body: 36 Comm. to v. 28-29ab: … svendriyaśaktibhistadindriyākramaṇapūrvamātmaśakti-svīkṛtasya parasya śarīraṃ svaparispanda-pravṛttinivṛttikrameṇātmaśarīrīkṛtya jīvaṃ parapuryaṣṭa-kaṃ jīvena svapuryaṣṭakena … svaprāṇena veṣṭayedākramet ||. 37 NT 20.29cd-30: ākramya taṃ hṛdisthaṃ vai adhaūrdhvapraveśataḥ || ekībhāvaṃ samāsādya samatvaṃ tatra cābhyaset | paraṃ kāraṇamāśritya svatantratvaṃ tadābhyaset ||. Kṣemarāja glosses the ‘supreme cause’ as a reference to the śākta-immersion: Paraṃ kāraṇaṃ śāktaṃ balam āśritya tadgraheṇāvaṣṭabhya |. 38 NT 20.31 and comm. to v. 31ab: vyāpakena svarūpeṇa śaktyā śaktiṃ tu dārayet | tadetatsaṃpuṭīkṛtya śakticchedaṃ tu kārayet ||31 samāveśabalād vīryabhūtāṃ vyāpakatāmāsthāya svaśaktyā tatprāṇaśaktiṃ dārayet chedārthamākṣipet ||, Kṣemarāja further refers to the kṣurikā-mantra as a means to cut off the power of the other.

76

Olga Serbaeva Saraogi “Then, O goddess, the knower of yoga, having become that being [sattvam] having the nature of power, he heats up as if by the sun of consciousness which has the form of nature of its own nature, he liquefies the rays stationed in the other in the manner of the sun with its rays. … This is to say that [s]he, having become stationed in the body of another, makes liquid the rays of the eyes and other organs by means of melting or liquefying them with the rays of his own eyes and other organs of sense, in the same manner as the sun melts or liquefies the rays of the moon.”39

The process of melting is represented by the lelihānā-mudrā,40 and the yoginī now proceeds to assemble all the melted organs of the victim: “Having joined all these elements related to the object, to the external and internal organs of senses, etc., of the other, which are all ‘liquid’ by means of the process of melting explained, all these he should join in him[/her]self and make ‘one’. So, … having grasped it with one’s own mind, the yogi then should enter everywhere in order, suddenly grasping all these, he will quickly become stationed in his own body [ātmastham]. ‘This’ means the already referred to ‘having grasped everything’, according to the āgamas, he goes out of ‘the city’ of the body [of another] by means of his mind, by the door of the head of the victim and enters his own body. Having entered there by his power, he should quickly bring into his own body all the seized and previously melted elements.”41 This looks very close to what the guru is performing in initiation and indeed, in some texts, both ritual actions are encoded by the same set of five mudrās: karaṅkiṇī, used upon coming out of the body, krodhanā, used to enter the body of the other, lelihānā, used to melt the senses of the victim, khecarī, employed to depart with that essence back to one’s own body and, finally, bhairavī, which is a sort of 39 NT 20.32-33ab: Śaktirūpaṃ tato devi sattvamāsthāya yogavit | svasattvasattārūpeṇa citsūryatvena tāpayet ||32 drāvayet tu parastho hi raśmīn raśmibhir arkavat |. Kṣemarāja glosses ‘having the nature of śakti’ [śaktirūpaṃ] as ‘having the vital energy as the predominant aspect’ [prāṇaśaktipradhānaṃ]. Rest of the commentary: Tatastadīyān cakṣurādiraśmīn pare tatraiva paradehe sthitaḥ sandīptaiḥ svaiścakṣurādibhi raśmibhiḥ sūrya iva somaraś-mīn drāvayed vilāpayet ||. Alexis Sanderson specifies: the objects of the senses [moon] are melted by the senses [sun], personal communication, July 2007. 40 See note 42. 41 NT 20.34d–35, introduction and commentary: Parasaṃbandhīni viṣayabhūta-bahirantaḥkaraṇāni proktavilāpanato rasabhūtāni drutatvamāptāni sarvāṇi yojayed ātmany ekīkuryāt | Atha … gṛhītvā tatsvacetasā ||34 praviśet tu tadā yogī puramākramya sarvataḥ | drutaṃ gṛhītaṃ tatsarvaṃ kṣipram ātmastham ānayet ||35 Tad iti pūrvoktaṃ sarvaṃ gṛhītvā āgamoktadṛṣṭyā sādhyasya mūrdhadvāreṇa niṣkramya cetasā puraṃ śarīraṃ praviśet | tatas tadākramya śaktyā’sya yat pūrvaṃ drutaṃ vilāpitaṃ saṃgṛhītam tat sarvaṃ kṣipram ātmaniṣṭhaṃ kuryāt ||.


77

non-dual state upon coming back into one’s own heart. This similarity of initiation and the action of yoginīs is particularly well expressed by Kṣemarāja in his commentary to the Netra tantra 20, yet the same ideas can be found in Jayadrathayāmala 1.17, 2.17, Ūrmikaulārṇava 2, etc.42 I would argue, however, that the sequence of these five mudrās seems to be applied to both initiation- and the yoginī-related material “artificially”. The proof of this is found in the numerous instances when the same mudrās are neither related to the procedures including the entrances into the bodies of others, nor appear as a sequence of five.43 The main differences between the initiation and the violent actions of yoginīs lie in the fact that, firstly, the mantras in the texts describing initiation have a tendency to become mudrās when the yoginīs’ aggressive yoga is being described. This is the case with various cutting mantras in more ancient texts, which become mudrās in JY. The second difference is that the yoginīs usually do not give back the principle of consciousness of the victim, thus killing and thereby, liberating him. The last phase, liberation, which is supposed to be achieved by these two procedures differing in such key aspects as actors and general context, requires some precision since it does not appear clearly from the sequence of mudrās. In śaiva tantric texts, liberation up to the time of Utpaladeva and Abhinavagupta is, as a rule, post mortem. Although we find in the texts prior to Abhinavagupta some 42 These mudrās are also mentioned in later Krama texts, but they refer instead to states of consciousness linked to the five kinds of yoginīs. See NT 20.36ab, introduction and commentary: Taditthaṃ yogī tatkṣaṇād ānayejjīvaṃ mudrāmantraprayogataḥ | Mudrā’tra prathamaṃ svaśarīraniḥ-saraṇasamaye karaṅkiṇī taccharīrākramaṇe krodhanā tadraśmyādivilāpane lelihānā tatpurād niḥsaraṇe khecarī svahṛtprāptau bhairavī |. In this way the yogi, “in this moment takes control of the ‘soul’ by means of mudrā and mantra”. The mudrās are the following: while departing from one’s own body – karaṅkiṇī, while entering the body of the other – krodhanā, while melting the rays of that – lelihānā, while departing from that body – khecarī, [and] upon attaining [back] one’s own heart – bhairavī. ŪK 2.100–101ab: saiṣā karaṃkiṇī proktā bhairavī sā varānane | lelihānā tu sā jñeyā krodharājeśvarī tu sā || pāṇimudreti vikhyātā bhoginī bhogamokṣadā |. 43 This is the case with JY 4.2 where all five appear in a broken sequence, each having its own functions, and JY 4.80.70–98, where these five are given in sequence but preceded by dantura-mudrā, having again a meaning independent from each other. Even Abhinavagupta in his TĀ 32.5 does not provide a sequence used here by Kṣemarāja. All this allows us to suppose that such a sequence was taken by Kṣemarāja (or someone else closely before him) from Krama texts and “read into” the Kaula texts in order to provide a valid and nondualistic explanation of the apparent contradictions concerning initiation and the actions of the yoginīs. The other proof is the fact that in JY, for example, there is a large number of mudrās that can be used to kill a victim, or to put it under control or to drain its blood, where the whole procedure is encoded by just one single mudrā. The examples include kapāla- in JY 4.2.10–15 and 3.32.170cd–172ab; pheṭkārikā- JY 4.2.573–583, 2.17.604–610, 697–682, 2.19.37–47; vikaṭa- JY 1.24.51–53, etc.

78


jīvanmukti or other occurrences expressing ‘liberation while alive’, the idea is far from important.44 The śaiva initiation guarantees the liberation, but it will take its full shape only after death. Death and liberation are, therefore, closely interrelated. Moreover, the descent of Śiva’s power (śaktipāta) occurring during the initiation can have various degrees, the highest being a direct merging into Śiva, with all impurities and limitations being destroyed. Apparently, the physical body cannot support such a violent descent of grace, and so, from the human point of view, it is a death. Therefore, by killing a victim the yoginīs also directly “activate” for him the higher state. “It is liberation and not killing”, argues Kṣemarāja, supporting the opinion of Bhairava in his commentary to the Netra tantra 20. Kṣemarāja glosses the actions of yoginīs: “they kill” [upahāranti] as “they liberate”. He also explicitly identifies the initiation and the actions of the yoginīs by giving the list of the five named mudrās (krodhinī, karaṅkiṇī, lelihānā, khecarī and bhairavī) used in both kinds of ritual actions.45 The Ūrmikaulārṇava 2 also links the five named mudrās both to the context of initiation and of killing a paśu.46 44 See Serbaeva (forthcoming)a. 45 The Netra tantra 20 is a dialogue between the goddess and Bhairava in which the goddess wonders why yoginīs and the like are so cruel, why they drain blood and why they kill victims. Bhairava explains that it is in accordance with his will and that the beings were created only for this, v. 4–7. Bhairava states that once killed, the souls are liberated from sins and thus it is grace, v. 8. The immolated victims also achieve “going upwards”, v. 9, on which Kṣemarāja comments: “the ‘going upwards’ means the liberation [through] rebirth in the states such as ‘pure knowledge’ and others”. These killed beings become in fact one with Śiva, v. 15. 46 ŪK also provides the clearest link between the extraction of blood or the soul and initiation. In ch. 2, v. 207cd–215, Bhairava describes the initiation, which includes the piercing [vedha] of the six centres (navel, heart, throat, bindu, head and the highest, v. 210cd–211ab) and śaktisaṃghāta, described in v. 213cd–214, in which passage the ‘essence’ to be extracted from the disciple’s body is clearly compared to ‘nectar’: yoginīsaṃpradāyena pāramparyakrameṇa tu [kramādi taṃ] || sṛṣṭisaṃ-hārayogena [-saṅghārayogema] paradehaparaṃ viśet | ākarṣayec chiṣyadehāt [-hā] svarūpaṃ cāmṛtopamam [ca-] ||. “[The guru] should enter the body of another [disciple] by means of the practice related to the creation and the reabsorption which is the revelation of the yoginīs, and which came down through initiatory lineage [pāraṃpara]. He should extract the ‘essence’ from the body of the disciple which is like ‘nectar’.” Bhairava further classified the totality of Kaula doctrines into various sorts of attractions and extractions, he names at first five such practices in v. 217cd–218ab: hinnākṛṣṭiḥ [bhinnākṛṣṭi] kalākṛṣṭiḥ [-kṛṣṭi] padmākṛṣṭis tathaiva ca || rasākṛṣṭir jīvanāśaḥ kaulaḥ pañcavidhaḥ kramaḥ |. The first practice refers to cheating and attracting the paśu, the second apparently to the entrance into his heart, the third to vedha when the 'lotuses' are pierced, the fourth to the extraction of juices [rasa], the fifth to the extraction of the soul [jīva], and the sixth, added later, in v. 221ab – to yojanā, the joining of a person with Bhairava, thus the extraction of blood and initiation are essentially similar.


79

The paśu who had suffered such a death is said in numerous tantras to take a higher position, to become either liberated or reborn with some extra capacities. Netra explicitly states that the person obtains liberation and, in the worst case, a higher rebirth if “used for yāga” by the yoginīs.47 We also have statements that a

See in particular ŪK 2.220cd–222: māyājāle mohanaṃ syāt mudritasya paśos tataḥ || jvālanaṃ tu dvītīyena tṛtīyena tu vedhanam | caturthena rasākṛṣṭi pañcame jīvakarṣaṇam || ṣaṣṭamayojanaṃ kuryāt paścād bhairavavigrahe | māyājālena tārasā [tarapā] samākṛṣyāgrataḥ paśoḥ [paśo] || Furthermore, details are given concerning the practice of killing a paśu who is joined to the highest state and thus liberated. This is directly followed by the initiation through a provoked possession [āveśa, v. 237cd] of śāṃbhava-, śākta- and āṇava-type. The first type of pervasion used in initiation, śāmbhava, is the same as killing a victim, v. 229cd–235: athavā yatra tatraiva sthāpito mudritaḥ paśuḥ || tattvanyāyena [-yona] tenaiva dvāreṇa praviśed valāt | śaktivīryeṇa tenaiva tataḥ [tata] svairyaṃ pareṇa tu || kṛtvā [kṛtyā] tṛtīyabījena cchedanaṃ mandasandhiṣu | jīvāmṛtakalākṛṣṭiḥ [-kṛṣṭi] caturthenātmano grahaḥ || kṛtvā nivedanaṃ kuryāt [kuryā] pañcamena vibhopari | athavā pañcakenaiva pañcakaṃ pāśabandhanāt || vedhayitvā samānīya sajīvarasapañcakam | pañca vā bhāgaśaḥ kṛtvā kulapañcakamaṇḍale || nivedya prāśanaṃ kṛtvā yojayitvā paśuṃ pare [pari] | yogibhir mātṛbhiḥ [yogibhimātṛbhi] sārdhaṃ kham utpatati kaulavit || anena kramayogena vedhasaṃkramaṇaṃ param | matsamair gurubhiḥ [matsamai] sākṣāt kartavyaṃ parameśvari ||. To resume, the victim being fixed [mudrita], the guru enters him by one of the ‘doors’, and uses the ‘seed-syllables’ [bījas] of a particularly powerful mantra, one bīja for each action, the third for cutting off, the fourth in order to extract the ‘soul-nectar’, the fifth encodes the liberation of the victim from the bondages through the fact that the five nectars of the victim and his soul are being offered to the maṇḍala of ‘five families’, and when the guru ‘eats’ it, having previously joined the victim with the highest state, the knower of Kula flies up together with the mothers and yoginīs. By the same procedure by the gurus who are equal to me [Śiva] the initiatory procedure [vedhasaṃkramaṇa] should be performed. Bhairava further explains that śākta- and āṇava-types of provoked pervasion. 47 NT 20.14–16 and the commentary of Kṣemarāja: eṣām anugrahārthāya paśūnāṃ tu varānane | mocayanti ca pāpebhyaḥ pāpaughāṃśchedayanti tān ||14 “O beautiful, this [takes place] in order to bestow grace upon the paśus, who are released from sins and for them the accumulation of sins is being cut off.” Kṣemarāja, glossing this passage, equates being killed by a yoginī with cutting off the impurities, which, in its turn, is a state of liberation. The next passage is even more explicit: paśūnām upayuktānāṃ nityam ūrdhvagatir bhavet | te muktāḥ śivabhūtās tu śivaśaktyā śiveritāḥ | nirmalāḥ śivarūpās tu tatprabhāvād bhavanti ca ||15 yathā yogena dīkṣāyāṃ śivatvam upalabhyate | tathā vai yogiyogena śivatvam upayānti te ||16 “The immolated [or ‘joined’] victims will always become ‘going upwards’, those are liberated, [they are] śaiva creatures, by the power of Śiva, they are impelled by Śiva, [they are] pure, having the form of Śiva, they become through that splendour [of yoginīs’ self-identification with the supreme]. As in initiation through yoga one obtains the nature of Śiva, in the same way, through yoga of yoginīs, they will take hold of the nature of Śiva.”

80


person if reborn will have a memory of their past life in the Brahmayāmala 99, which provides details of the yoginīs’ aggressive yoga.48 This allows us now to read the story of king Baka sacrificed to the circle of goddesses in a different and more optimistic light.

Conclusion I have demonstrated that the similarity in the sequence of the ritual elements leads to the possibility of interpreting the two rituals as having a similar meaning even if there is a difference in the main actors, circumstances, or the position each ritual occupies within its own tradition. Another possibility is that in order to give similar meaning to two originally very different rituals, they were filled with similar ritual elements, as seems to be the case with the sequence of the five mudrās, karaṅkiṇī, etc. Taking into account this “artificial meaning” of a ritual, which is constructed by filling the ritual action with a set of particular components, it makes it difficult, on the basis of the material presented here, to see the rituals as something as “natural” as a language. Should then the rituals rather be compared with a language of programming, which is also “artificial” and uses as well the same sequences of elements in order to achieve a desired result when the program is running?

48 The haṭhamelaka, according to the edition and translation of Hatley of BY 99.10–15, suggests that the sādhaka is killed to be reborn in a higher state with the memory of past birth etc., v. 14ab in particular: muktvā deham avāpnoti jātijñaś ca prajāyate |.

When to Kill Means to Liberate Fig. 1: The sequence of actions of guru and yoginī

81

82


Abbreviations ŪK – Ūrmikaulārṇava tantra Ūrmikaulārṇava tantra. Ed. Dyczkowski, Mark. MS 5–5207 (sic. 5–5202); NGMPP: B115/9; folios 83; Newari script. E-text: www.muktabodha.org (04.2008). KSS – Kathāsaritsāgara of Somadeva The Kathāsaritsāgara of Somadevabhaṭṭa. Dvivedī, Durgāprasād & Kāśīnāth Pāṇḍuraṅg Parab (eds.) 1930. Bombay: Pāṇḍuraṅg Jāvajī. JY – Jayadrathayāmala tantra Ṣaṭka 1, NAK 5/4650, NGMPP B122/7, paper, Devanāgarī script. Ṣaṭka 2, NAK 5/4650, NGMPP A153/2, paper, Devanāgarī script. Ṣaṭka 3, NAK 5/1975, NGMPP A152/9, paper, Newari script. Ṣaṭka 4, NAK 1/1468, NGMPP B122/4, paper, Newari script. All four parts are transliterated by Serbaeva Saraogi, O., 2007–2009. TST – Tantrasadbhāva tantra Tantrasadbhāva. Paṭala 16. Critical ed. from the codex unicus by Sanderson, Alexis. 27.08.2004. MS NAK 5–445, NGMPP A44/2, folios 186, palm-leaf, Kuṭila script. Paṭala 16 occupies ff.108r3–118r1. Tantrasadbhāva, partially and provisionally edited by Dyczkowski, Mark S.G. MS K: 1–1985 Śaivatantra 1533, NGMPP A188/22; folios 132; Devanāgarī script. MS Kh: 1–363 Śaivatantra, NGMPP A44/1; folios 140, Newari. MS G: 5–445 Śaivatantra 185, NGMPP A44/2, folios 210; Newari. E-text: www.muktabodha.org (2005). TĀ – Tantrāloka of Abhinavagupta 1918–1938. The Tantrāloka of Abhinavagupta. With Commentary by Rājānaka Jayaratha. Mukund Rām Śāstrī (Vol. 1) & Madhusudan Kaul Śāstrī (eds.) (KSTS 23, 28, 30, 34, 35, 39, 41, 47, 59, 52, 57, 58). Allahabad: The Indian Press/Bombay: Tatva-Vivecaka Press. E-text: www.muktabodha.org (2005). NS – Niśisaṃcāra tantra Niśisaṃcāra tantra, Ed. Sanderson, Alexis. Oxford, 9 October 2004. MS NAK 1–1606 Śaivatantra 102, NGMPP B26/25, folios 48; palm-leaf; Nepalese Kuṭila script. NT – Netra tantra 1939. The Netra Tantram, with Commentary by Kṣemarāja. Madhusudan Kaul Śāstrī (ed.). Vol. 2. (KSTS 61). Bombay: Tatva-Vivecaka Press. E-text: www.muktabodha.org (2005). BKM – Bṛhatkathāmañjarī of Kṣemendra 1982. Śrīkśemendraviracitā bṛhatkathāmañjarī. Sivdatta P. & Kashinath Pandurang Parab (eds.). Delhi: Panini. BY – Brahmayāmala Hatley, Shaman 2007. The Brahmayāmalatantra and Early Śaiva Cult of Yoginīs. Pennsylvania: University of Pennsylvania. Unpublished Ph.D. thesis including edition of selected chapters. MVT – Mālinīvijayottara tantra


83

1922. Śrī Mālinīvijayottara Tantram. Madhusudan Kaul Śāstrī (ed.). Srinagar (Kashmir series of texts and studies 37). Vasudeva, Somadeva 2004. The Yoga of the Mālinīvijayottaratantra. Chapters 1–4, 7, 11–17. Critical ed., tr. and notes. Institut Français de Pondichéry, Ecole Française d’Extrême-Orient. E-text: http://www.sub.uni-goettingen.de/ebene_1/fiindolo/gretil. htm (04.2008) RT – Rājataraṅgiṇī of Kalhaṇa Stein, M.A. 1961. Rājataraṅgiṇī of Kalhaṇa. A Chronicle of the Kings of Kashmir. Delhi: Motilal Banarsidass. VB – Vijñānabhairava tantra 1918. Vijñānabhairava with Commentary Partly by Kṣemarāja and Partly by Śivopādhyāyā. Mukunda Rāma Śāstrī (ed.) (KSTS 8). Bombay: Tatva-Vivecaka Press. E-text: www.muktabodha.org (04.2008). SVT – Svacchandabhairava tantra 1921–1935. The Svacchanda-Tantram with Commentary by Kshema Rāja. Madhusudan Kaul Śāstrī (ed.). (KSTS 31, 38, 44, 48, 51, 53, 56). Bombay: Nirnaya Sagar. E-text: www.muktabodha.org (04.2008).

References Brunner, Helène 1963. Somaśambhupaddhati. (Part 3). Pondichéry: Institut Français d’Indologie. Crooke, William 2003. An Introduction to the Popular Religion and Folklore of Northern India. Delhi: Low Price Publications. Dwivedi, R.C. 1990. Kāśmīr kī śaiv paramparā. New Delhi: Neśnal Pabliśiṅg Hāus. Edgerton, Franklin 1985. Buddhist Hybrid Sanskrit Grammar and Dictionary. Vol. 2. Delhi et al.: Motilal Banarsidass. Forster, H.P. 1802. Vocabulary, in Two Parts of Bongalee and English and Vice Versa. (Part 2). Calcutta: P. Ferris, Post Press. Knowles, J.H. 1888. Folk-Tales of Kashmir. Edinburgh, London: Trubner. Kuprin, A. I. 1971. Olesya. In: Collected Works (Sobranie sochinenii). Vol. 2. Moscow: Hudojestvennaya Literatura: 311-381. Lidke, Jeffrey S. 2005. “Interpreting Across Mystical Boundaries: An Analysis of Samadhi in the Trika-Kaula Tradition”. In: Knut A. Jakobson (ed.). Theory and Practice of Yoga. Essays in Honour of Gerald James Larson. Leiden, Boston: Brill: 143–180. Michaels, Axel 2008 “‘How Do You Do?’ Some Preliminary Thoughts on a Grammar of Ritual”. Paper distributed to the participants of the panel “Grammar and Morphology of Rituals”. Padoux, André 1990. “The Body in Tantric Ritual: The Case of the Mudrās”. In: Teun Goudriaan (ed.). The Sanskrit Tradition and Tantrism. Panels of the VIIth World Sanskrit Conference. Leiden, New York: Brill: 65–77.

84


— 1999. “Transe, possession ou absorption mystique? L’āveśa selon quelques textes tantriques Cachmiriens.” In: Jackie Assayag & Gilles Tarabout (eds.). La Possession en Asie du Sud. Parole, corps, territoire. (Collection Puruṣārtha 21). Paris: Editions de l’école des Hautes études en Sciences Sociales: 133–147. Sanderson, Alexis 1990. “Śaivism and the Tantric Traditions”. In: Friedhelm Hardy (ed.). The World’s Religions. The Religions of Asia. London: Routledge: 128–172. — 2007. “The Śaiva Exegesis of Kashmir”. In: Dominic Goodall & André Padoux (eds.). Mélanges tantriques à la mémoire d’Hélène Brunner. Tantric Studies in Memory of Hélène Brunner. Pondichéry: Institut Francais de Pondichéry: 231–442; 551–582. Serbaeva Saraogi, Olga 2006. Yoginīs in Śaiva Purāṇas and Tantras: Their role in transformative experiences in a historical and comparative perspective. Unpublished Ph.D. thesis (presented at the University of Lausanne). — (forthcoming)a. “Liberation In Life and After Death in Early Śaiva Mantramārgic Texts: The Problem of Jīvanmukti”. In: Welten Süd- und Zentralasiens. — (forthcoming)b. “Yoga from Yoginīs’ Point of View”. In: The Proceedings of the 13th World Sanskrit Conference, Edinburgh, Scotland, July 2006. Sleeman, W.H. 1844. Rambles and Recollection of an Indian Official. London: J. Hatchard and Son. Smith, Frederick M. 2006. The Self Possessed. Deity and Spirit Possession in South Asian Literature and Civilization. New York: Columbia University Press. Takashima, Jun 1992. “Dīkṣā in the Tantrāloka”.東洋文化研究所紀要 (Tōyō bunka kenkyūsho kiyō) (Memoirs of the Institute of Oriental Culture 119): 45–84. Troyer, M.A. 1840. Radjatarangini. Histoire des Rois du Kachmir. Vol. 2. Paris: L’imprimerie Royale. Wallis, Christopher 2008. “The Descent of Power: Possession, Mysticism and Initiation in the Śaiva Theology of Abhinavagupta”. Journal of Indian Philosophy 34: 247–295.

Anand Mishra

On the Possibilities of a Pāṇinian Paradigm for a Rule-based Description of Rituals 1. Introduction Rituals are special actions. At least this can be asserted without any loss of generality, even if the question as to what exactly is special about them continues to be a topic of active research among the ritual scientists. The attempt to identify, stipulate and comprehend the distinguishing characteristics of ritual actions has given rise to a number of ritual theories depending upon the nature of rituals examined, method of examination as well as the differnt approaches taken by ritual experts.1 A formal approach, which attempts to explore rituals by examining their structure, saw substantial progress in the last few decades. Thus, Frits Staal’s analysis of vedic agnicayana rituals lead him to propose that ritual actions are performed for their own sake, are in fact an elaborate rule based affair, which can be described through rules similar to syntactical rules of a transformative generative grammar, and have no meaning.2 Staal also hypothesized that language and grammar can have rituals as their precursor. The theory of meaninglessness of rituals attracted a spate of criticism, notably by Penner who suggested to liberate oneself from referential theories of meaning and rejected Staals arguments based on the non-existence of references towards which ritual actions point to.3 Based on their examination of Jain pūjā, Humphrey and Laidlaw strengthened the argument that the difference between acting in a ritualized manner in contrast to an unritualized way can be decided by identifying, among other factors, the stipulated and archetypal character of ritualized actions.4 A Montageplan of rituals was proposed by Oppitz where he suggested that rituals are a conglomeration of individual parts and these parts are composed of constitutive elements. Further, these constitutive elements can be moved and replaced based upon a scheme, thus giving rise to a ritual.5 1 2 3 4 5

For an overview of different approaches towards a theory of rituals, see Michaels 1999. Staal 1979. Penner 1985. Humphrey & Laidlaw 1994. Oppitz 1999.

86

Anand Mishra

Focussing primarily on the structure of ritual actions, Gladigow further pointed to the existence of ritual elements (Riten) which can be identified, replaced, revised, cited (rituelle Zitate), repeated, abbreviated (rituelle Abbreviaturen), interrelated (Interritualität) and learned. These ritual elements form ritual sequences of varying complexities where they combine to constitute a ritual complex.6 While demonstrating the composition of hinduistic initiation rituals (upanayana, vratabandha) out of constitutive elements (Bausteine), Michaels stipulated five criteria which distinguish a ritual from other non-ritual actions:7 (i) causal transformation (causa transitionis) (ii) formal decision (solemnis intentio)8 (iii) formal criteria of ritual actions (actiones formaliter ritorum) (iv) modal criteria for ritual actions (actiones modaliter ritorum) and (v) change of status (novae classificationes; transitio vitae) This brief overview of some of the landmarks of the development of a formal theory of ritual is meant to indicate that they refer to, in some form or other, the following propositions: – Rituals are special actions. – The distinguishing character of special ritualized actions can be represented in terms of certain attributes and criteria. – The ritual actions are also not unrelated. – The relatedness of ritual actions can be comprehended through associated attributes and can be formulated in terms of rules complying certain criteria. A parallel set of propositions can be made for languages. – Languages are special utterances. – The distinguishing character of linguistic utterances can be represented in terms of certain attributes and criteria. – These utterences are also not unrelated. – The relatedness of linguistic utterances can be comprehended through associated attributes and can be formulated in terms of rules of a grammar. In this article, I propose to follow the above propositions in a parallel manner, taking the example of Sanskrit for linguistic utterances and hinduistic pūjā for ri-

6 Gladigow 2004. 7 Michaels 1999. 8 Michaels 2005.

A Pāṇinian Paradigm for a Rule-based Description of Rituals

87

tual actions.9 Despite the historical (and as suggested by Frits Staal, also the methodological) order, at least in indian context, is that rituals precede languages10, I choose to follow the reverse order. That is, I first state the linguistic model and then accordingly suggest a conceptually similar approach for description of rituals. This is because, we have a highly evolved grammatical process and accordingly a clearly specified grammar for Sanskrit – the Aṣṭādhyāyī of Pāṇini – providing us a good choice to view rituals in the context of Pāṇinian paradigm.11

2. A Framework for Description of Rituals If rituals are special actions and if we wish to undertake a systematic study of their peculiar characteristics, then it is incumbent to evolve a framework which facilitates an adequate representation of these ritual charateristics as well as a formulation of those instances of interconnectivity, which makes them different from other non-ritual activities. Such a framework for description of ritual actions is what I intend to explore. For this, I take the pāṇinian paradigm and attempt to improve upon my modelling of the pāṇinian process12 to incorporate the features within the broader context of ritual actions, as compared to the linguistic utterances in case of Aṣṭādhyāyī of Pāṇini. The framework is meant for a computer assisted analysis of ritual phenomena motivated by the pāṇinian way of dealing with a similar challenge.13 There is a certain amount of similarity between languages and rituals which justifies the choice of grammar of a language as our source of guidance. Both, linguistic utterances and ritual actions, are some special subset of a bigger collection. For example, from among all the possible combinations of every possible soundunits, only certain special combinations of a few selected sound-units constitute the set of valid linguistic expressions of a spoken language. A corresponding statement can be made for rituals. Not every action is a ritual action but only certain combinations of some selected activities constitute the set of ritual actions.

9 For my examples, I choose some of the daily (nitya) and occasional (naimittika) rituals from a ritual hand-book Nityakarma-Pūjāprakāśa (Miśra & Miśra 2003) fairly popular in North India, which is a compilation of rituals in the tradition of Pāraskara-Gṛhyasūtra. 10 Staal 1989. 11 For an introduction to the Pāṇinian system, see Introduction in Katre 1989. 12 Mishra 2009. 13 An ontology-based representation, incorporating domain-specific concept hierarchies is being attempted by the ongoing subproject “Ontology modelling for ritual structure research” of the SFB 619. For more information see: http://www.ritualdynamik.de/

88

Anand Mishra

2.1 Language: A Subset Consisting of Special Utterances Let us say, if we represent the set of all possible combinations of any kind of sound-units by the letter U and the set of utterances of a language (e.g. Sanskrit) by the letter L, then L ⊂U It is clear that both these sets contain infinite elements. But there exists at least one element, which belongs to the set U but not to the set L. That is to say, L is a proper subset of U. In other words, a spoken language is a special subset of utterable sound combinations. 2.2 Rituals: A Subset Consisting of Special Actions Consider now that the letter H represents all possible combinations of every possible human actions and the letter R represents those actions which we would call as ritual actions (say for example hinduistic rituals). Then, a statement similar to the one above can be made, R⊂H Again, both the sets R as well as H contain infinite number of elements. But there exist at least one element in H, which does not belong to R i.e. R is a proper subset of H. In other words, hinduistic rituals are special actions. A number of other observations can be made as regards the set of linguistic utterances L and that of rituals R. For example, these sets are not static, i.e. new elements get included in it over a period of time, or the elements of these sets cause certain effects on us i.e. the efficacy of a linguistic utterance or some ritual action. However, the feature which is of interest to us is that these sets can be described through rules of grammar.

3. Grammar – its Function and Scope Without entering into the debate about the role, function and scope of grammar in general and pāṇinian grammar in particular, it is pertinent to specify it in our present context.14 In our case, the purpose of a grammar is, first and foremost, to describe a particular set in a compact, concise and unambiguous manner through a collection of rules. Consider for example the set: E = {2, 4, 6, 8, 10, 12,... } This set can be described by endlessly counting the appropriate numbers or by a rule which says: The set E contains any positive number which is divisible by 2. 14 See Bhattacharya 1966 & Houben 1999.


89

Describing a language through a grammar implies specifying the collection of rules which describe the set of linguistic utterances of that language. In other words, providing a mechanism to decide whether a particular utterance is a part of the collection of the linguistic utterances of a language. Thus, Aṣṭādhyāyī of Pāṇini is a rule based mechanism to decide whether a given utterance can be considered as a member of the set L. Similarly, a grammar of a group of ritual actions R can be considered as a rule based mechanism to decide whether a given action can be considered as a member of the set R. 3.1 Specification of Grammar The specification of the grammatical process of Pāṇini can be summarized as consisting of three very basic and general steps: 1. Prescription of the fundamental components which constitute the language 2. Characterization of these fundamental components by assigning them a number of attributes 3. Specification of the grammatical operations based on the fundamental components and their attributes The fundamental components are phonemes (like a, u, m etc.), morphemes (like bhū, ŚaP, tiP etc.) and lexemes (like aja etc.).15 The grammatical corpus of Aṣṭādhyāyī together with Śivasūtras, Dhātupāṭha, Gaṇapāṭha and Uṇādisūtras can be seen as a big collection of these fundamental components. The analytical process, beginning perhaps with the padapāṭha of vedic mantras, is not recorded in terms of any rules of analysis, but the grammar has at its disposal these fundamental components.16 A similar approach for ritual actions would imply that a corresponding set of fundamental components for rituals needs to be discovered. The normal definitions of phoneme as “the smallest segmental unit of sound employed to form meaningful contrasts between utterances” or morpheme as “the smallest linguistic unit that has semantic meaning” are of little help. It is very difficult if not impossible to decide upon the smallest ritual unit which can be distinguished from other such units or for that matter, a smallest meaningful ritual unit.17 This problem of defining the socalled ritemes can be avoided once we evolve an appropriate framework for representing these units.18

15 For writing pāṇinian terms, I follow the general conventions followed in Katre 1989 with the exception that I write all attributes in CAPITALS. 16 Deshpande 1990. 17 See Michaels 2009 (in the present volume). 18 See section 4 below.

90

Anand Mishra

The fundamental components become grammatically meaningful and useful once they get a grammatical identity. Thus, the morpheme bhū becomes grammatically significant once it is given an attribute DHĀTU. Similarly, the attributes HṚSVA, GUṆA and aC are a few attributes, which impart a grammatical identity to the phoneme a. The suffix tiP is assigned a number of attributes like: PRATYAYA (suffix), PRATHAMA (third person) and EKAVACANA (singular) etc. So, the technical terms (saṃjñā), the it-markers and even sigla (pratyāhāra) can be regarded as attributes which impart a grammatical identity to the fundamental components. We can write the above observation using the following notation. Consider the representation of the suffix tiP. It will look like: tiP{PRATYAYA, PRATHAMA, EKAVACANA, ...} Here, a set of attributes is attached with the fundamental component tiP. The itmarker P is also an attribute and we can include it in the set of attributes as P-IT. ti{PRATYAYA, PRATHAMA, EKAVACANA, p-it, ...} This morpheme consists of two phonemes, which have their own phoneme attributes. Thus the second phoneme is a short (HṚSVA) vowel (aC) and the first one a consonant (haL). We can write the phoneme attributes by attaching them to respective phonemes: t{haL, ...}i{HṚSVA, aC, ...} Now, instead of associating the morpheme attributes at the level of a morpheme, if we include them with each of the phonemes of that morpheme, the following will result: t{PRATYAYA, PRATHAMA, EKAVACANA, P-IT, haL, ...}i{PRATYAYA, PRATHAMA, EKAVACANA, P-IT, HṚsva, aC, ...} This can be written in an optically better way as: t{PRATYAYA, PRATHAMA, EKAVACANA, P-IT, haL, ...} i{PRATYAYA, PRATHAMA, EKAVACANA, P-IT, HṚSVA, aC, ...} So now we have attached a number of attributes to each phoneme of a given morpheme. Both phoneme and morpheme attributes are within a single set of attributes. This is permissible as long as we uniquely identify and know a given attribute to be a phoneme or a morpheme attribute. Moreover, applying simple set intersection, we can always win back the list of phonemes corresponding to the morpheme having a given morpheme attribute. For example, we can easily find out that the string of phonemes which have the attribute PRATHAMA are ti. It is to be noticed that the string of phonemes within a morpheme has a definite order and so we represent this as an ordered set or a list. [t{PRATYAYA, PRATHAMA, EKAVACANA, P-IT, haL, ...}, i{PRATYAYA, PRATHAMA, EKAVACANA, P-IT, HṚSVA, aC, ...}] Moreover, we can also include the phonemes within this set of attributes, once we are sure that we can seperate them out if needed. This is assured by assigning a


91

unique key to each of the fundamental constituents as well as attributes in the database. [{t, PRATYAYA, PRATHAMA, EKAVACANA, P-IT, haL, ...}, {i, PRATYAYA, PRATHAMA, EKAVACANA, P-IT, HṚSVA, aC, ...}] Now the only information originally available but not yet included in this representation is the fact that this is all about the morpheme tiP. We do this by adding a unique key for this morpheme, to all the sets which correspond to it. Thus, we have [{tip, t, PRATYAYA, PRATHAMA, EKAVACANA, P-IT, haL, ...}, {tip, i, PRATYAYA, PRATHAMA, EKAVACANA, P-IT, HṚSVA, aC, ...}] The above representation is in terms of a list of certain sets. We call these sets “sound-sets” and define such a sound-set to be a collection of fundamental components and attributes with the restriction that it contains only one element from the set of phonemes. This is an abstract data structure, a representational scheme to differentiate at the elementary level of individual sounds, but at the same time store information corresponding to other (morpheme, word or sentence) levels as well. Thus, although a sound unit corresponds to a phoneme, it records its context as well and therefore, contains more information than just phoneme attributes. A language component can be seen as a list of sound-sets. A sound-set, in turn, as a collection of attributes and fundamental components with only one phoneme key. The formulation of rules can be seen now in terms of these fundamental components and their attributes. The rule operations of Aṣṭādhyāyī, like augmentation (āgama), substitution (ādeśa), elision (lopa), reduplication (dvitva) etc. can be implemented in terms of two basic operations of addition of an attribute to one or more sound-sets and extension of the language component by inclusion of new sound-sets.19 Example of Augmentation (āgama) Consider the language component corresponding to the verbal root bhū. [{bh, bhū, DHĀTU, ...}, {u, bhū, DHĀTU, UDĀTTA, DĪRGHA, aC, ...}] Rule vartamāne laṭ (A 3.2.123) says that the morpheme lAṬ is added after a DHĀTU if the present action is to be expressed. To implement this rule, we first look for the indices of sound-sets which have the attribute DHĀTU and then append the sound-set corresponding to lAṭ after the last index. We get, [{bh, bhū, DHĀTU, ...}, {u, bhū, DHĀTU, UDĀTTA, DĪRGHA, aC, ...}, {l, lAṬ, PRATYAYA, A-IT, Ṭ-IT, ...}] 19 For details, see Mishra 2007.

92

Anand Mishra

Example of Substitution (ādeśa) Consider the language component corresponding to the verbal root ṇīÑ. [{ṇ, ṇīÑ, DHĀTU, Ñ-IT,...}, {i, ṇīÑ, DHĀTU, Ñ-IT, DĪRGHA, aC, ...}] Rule ṇaḥ naḥ (A 6.1.065) says that the initial retroflex ṇ of a DHĀTU is replaced by dental n. To implement this rule we first search the sound sets corresponding to DHĀTU, check whether the first one has a retroflex ṇ and if the conditions are fulfilled, add the attribute δ in that sound-set indicating that this set is replaced and append the sound-set corresponding to n after it. Further we transfer all attributes (except the phoneme attributes) from the ṇ-sound-set to n-sound-set for sthānivadbhāva. We get, [{ṇ, ṇīÑ, DHĀTU, Ñ-IT, δ,...}, {n, ṇīÑ, DHĀTU, Ñ-IT,...}, {i, ṇīÑ, DHĀTU, Ñ-IT, DĪRGHA, aC, ...}] These examples show the manner in which pāṇinian grammatical operations can be implemented while using the representational framework described above.

4. Representing an Element of R The representational framework for pāṇinian grammatical process can be extended for ritual actions. The general grammatical process mentioned in the above section can be reformulated for ritual actions as follows: 1. Prescription of the fundamental components which constitute a ritual 2. Characterization of these fundamental components by assigning them a number of attributes 3. Specification of the grammatical operations based on the fundamental components and their attributes The task then would be to identify the fundamental components of a ritual, to assign attributes to them and specify rules for grammatical operations based upon them. A particular ritual-action can be considered as the basic unit of a ritual and can serve as the unit corresponding to the sound-sets. Let us term them as ritual-actionset. In order to characterize a ritual-action, we need a number of ritual specific attributes. These attributes can be assigned to the ritual-action-sets. These attributes can also be of different types. For example, the action of sipping can have several attributes, depending upon our perspective. To a linguist, attribute types like its agent or instrument etc. is important. To a ritual scientist, apart from these linguistic characteristics, its ritual specific features are also of much value, e.g. the manner in which it is performed, the occasion, its context, purpose, effect etc. All


93

these observations, especially the ritually relevant ones, can be recorded by storing them in pairs: (attributeType, attributeValue). We can now define an attribute as a unit which records some observation about a ritual action in terms of a pair: (attributeType, attributeValue). Consider the ritual action of ‘chanting the mantra oṃ keśavāya namaḥ by the performer (yajamāna)’. The action of chanting can have several attributes: .(action, chanting) .(actor, performer) .(object, mantra:oṃ keśavāya namaḥ)

A ritual-action-set corresponding to a ritual action can then be represented as a collection of attributes, having only one attribute with action as its attributeType. The ritual action of ‘chanting the mantra oṃ keśavāya namaḥ by the performer (yajamāna)’ is represented by the set: c = {(action, chanting), (actor, performer), (object, mantra:oṃ keśavāya namaḥ), ...}

Consider another ritual action of ‘sipping of water by the performer (yajamāna) through the base of his right hand.’ Let us also record the ‘manner’ of sipping, where we notice that the ‘index finger of the right hand touches the base of his right hand thumb, and the palm is sunken like cow’s ear (gokarṇa).’ .(action, sipping) .(object, water) .(instrument, base of right hand of actor) .(manner, index finger of the right hand of actor touches the base of his right hand thumb and his palm is sunken like cow’s ear)

This ritual action can now be recorded as: s = {(action, sipping), (object, water), (instrument, base of right hand of actor), (manner, index finger of the right hand of actor touches the base of his right hand thumb and his palm is sunken like cow’s ear), ...}

A ritual activity will naturally consist of a number of such actions connected with each other depending on several criteria. The representation in terms of an ordered collection or list of a number of ritual-action-sets is not adequate for this. We represent this phenomenon in terms of a graph, where ritual actions constitute the nodes of the graph and the edges correspond to the ritual connection of two actions. The edges can again be seen as a collection of pairs with exactly one pair specifying the two connecting nodes and the rest being attribute pairs defined above. Consider now the ritual activity of ‘chanting the mantra oṃ keśavāya namaḥ by the performer (yajamāna) while sipping water through the base of his right hand in a manner that the index finger of his right hand touches the base of his right hand thumb, and his palm is sunken like cow’s ear (gokarṇa)’.

94

Anand Mishra

It has two actions chanting c and sipping s connected by the edge cs (represented by solid line in the figure), where cs = {(c, s), (temporal-interconnection, simultaneously)}

This ritual activity can now be represented by the ritual-complex which is nothing but a collection of nodes and edges. x1 = {c, s, cs}

Pictorially, it is a graph with nodes as individual actions connected through the edges. The nodes connected to the action nodes through dotted lines represent attribute pairs. Fig. 1: Graphical representation of a ritual-complex

The proposed representation thus seeks to record a ritual by registering the observations about it in terms of attribute pairs, without necessarily compelling one to decide a priori the smallest meaningful unit etc. It also allows one to add as many attributes as required and having various attribute types as necessary, thus theoretically enabling a potential adequate representation of different aspects of a ritual. Furthermore, it is also possible to concieve and coin meta-level attributes for the sake of grammatical rule formulations.

5. Rule-based Description of Rituals Given a representation of ritual actions suggested above, grammatical operations similar to the one in Aṣṭādhyāyī can be applied. A set of such rules may describe the structural aspect of a group of rituals or some other culture specific phenomena. I now present a few examples to show the possibilities of such a rule-based description of rituals. Instead of attempting to describe an entire ritual, I select a few observations and render it in terms of certain rules.


95

The set of rituals I have chosen for the purpose are the daily (nitya) and occasional (naimittika) rituals performed mainly in north India following the Pāraskaragṛhyasūtra of the mādhyandina śākhā of Śukla Yajurveda. The source book is the widely distributed Gītā Press edition of these rituals.20 Observation: A mantra is chanted, either by the priest or by the performer (yajamāna). {(action, chanting),(object, mantra)} ⇒ {(action, chanting), (object, mantra), (actor, priest)} ∨ {(action, chanting), (object, mantra), (actor, performer)}

The above representation shows that if there is a situation where we have a ritual-action-set with the action of chanting of mantras, then it should have the (actor, priest) or (actor, performer) attributes. This is similar to attribute addition based on certain conditions stated again in terms of certain attributes. Observation: Offerings are made by the performer. If the priest offers, ritually speaking it is to be understood as an offering by the performer. {(action, offering)} ⇒ {(action, offering), (actor, performer)} {(action, offering), (actor, priest)} ⇒ {(action, offering), (actor, performer)}

The first part of this rule is again an example of attribute addition. The second part, however, is a substitution of an attribute where the (actor, priest) attribute is substituted by the (actor, performer) attribute. Observation: The nine planetary deities (navagraha-maṇḍala) are placed in a fixed manner and the offerings to these deities are made in a particular order. {(action, offering)} {(recipient: [sūrya, candra, maṅgala, budha, bṛhaspati, śukra, śani, rāhu, ketu] ⇔ [C,NE,E,NW,W,N,S,SE,SW])}

The above representation uses a list depicted through square brackets [ ]. This is for an ordered collection. So the offerings are to be made in the order as mentioned above. Moreover, another list with placement directions is mentioned with the sign ⟺ which shows respective correspondence. This, incidently, is a meta-rule mentioned in the Aṣṭādhyāyī.21 This rule represents the structural constraints of the ritual actions connected with the worship of nine planetary deities. 20 Miśra & Miśra 2003. 21 yathā-saṃkhyam anudeśaḥ samānām (A 1.3.10). If two lists have the same number of elements, then the elements of the following list correspond respectively with the elements of the previous list.

96

Anand Mishra

Observation: The offerings in a pūjā are done in a particular order. {(action, offering), (object, [water, milk, curd, ghee, honey, sugar, perfume, clothes, sacread thread, sandal paste, rice, garland, grass, vermillion, incense, lamp, sweets, fruits, betel nut, gold coin])}

This again represents the sequential ordering of certain actions in a ritual. Observation: For the purpose of purification ācamana is performed, pavitrī is worn by the performer and water is sprinkled with certain mantras. These actions are optional, but if all three are performed, then in this order. {(purpose, purification)} ⇒ {(purpose, purification), (action, sipping), (instrument, base of right hand)} {(purpose, purification)} ⇒ {(purpose, purification), (object, pavitrī)} {(purpose, purification)} ⇒ {(purpose, purification), (action, sprinkling)} ∧ {(action, chanting), (object, mantra: apavitraḥ...)} [{(action, sipping)}, {(object, pavitrī)}, {(action, sprinkling)}]

The first three rules depict that if the purpose of purification is to be achieved, then the respective actions in appropriate manner should be undertaken. The fourth rule shows that in case more than one purificatory action is undertaken, then it should be done in a fixed order. These examples show that it is possible to analyze and describe the structure of rituals in terms of rules specifying the combinatorics of constituent components. The description will depend upon the nature and type of observations made by a scholar and therefore not unique. Nonetheless, a framework based on logical principles and implemented on computer can facilitate the long term goal of developing a grammar of rituals.


97

References Bhattacharya, R.S. 1966. Pāṇinīya Vyākaraṇa kā Anuśīlana. Varanasi: Indological Book House. Deshpande, Madhav M. 1990. “Semantics of Kārakas in Pāṇini: An Exploration of Philosophical and Linguistical Issues”. In: B.K. Matilal & Purusottama Bilimoria (eds.). Sanskrit and Related Studies: Contemporary Researches and Reflections. Delhi: Sri Satguru Publications, 33-57. Gladigow, Burkhard 2004. “Sequenzierung von Riten und die Ordnung der Rituale”. In: Michael Stausberg (ed.). Zorostrian Rituals in Context. Leiden, Boston: Brill (Numen Book Series, Studies in the History of Religions vol. CII), 57-76. Houben, J.E.M. 1999. “Meaning Statements in Pāṇini’s grammar: on the purpose and context of the Aṣṭādhyāyī”. Studien zur Indologie und Iranistik 22: 23-54. Humphrey, Caroline & Laidlaw, James 1994. The Archetypal Actions of Ritual: A theory of ritual illustrated by the Jain rite of worship. Oxford [u.a.] : Oxford Univ. Press. Katre, Sumitra M. 1989. Aṣṭādhyāyī of Pāṇini. Delhi: Motilal Banarsidass. Michaels, Axel 1999. “‘Le rituel pour le rituel’ oder wie sinnlos sind Rituale?”. In: Corina Caduff & Joanna Pfaff-Czarnecka (eds.). Rituale heute: Theorien – Kontroversen – Entwürfe. Berlin: Reimer, 23-47. — 2005. “Saṃkalpa: The Beginnings of a Ritual”. In: Jörg Gengnagel & Ute Hüsken & Srilata Raman (eds.), Words and Deeds: Hindu and Buddhist Rituals in South Asia. Wiesbaden: Harrassowitz Verlag, 45-64. Mishra, Anand 2007. “Simulating the Pāṇinian System of Sanskrit Grammar”. In: Gerard Huet (ed.). Proceedings of the First International Sanskrit Computational Linguistics Symposium. Paris: Rocquencourt: 89-95. — 2009. “Modelling the Grammatical Circle of the Pāṇinian System of Sanskrit Grammar”. In: Amba Kulkarni Amba & Gerard Huet (eds.). Sanskrit Computational Linguistics. Heidelberg: Springer, 40-55. Miśra, Śrī Rām Bhavan Jī & Śrī Lāl Bihārī Jī Miśra 2003 [V.S. 2060]. NityakarmaPūjāprakāśa. Gorakhpur: Gītā Press. Oppitz, Michael 1999. “Montageplan von Ritualen”. In: Corina Caduff & Joanna PfaffCzarnecka (eds.). Rituale heute: Theorien – Kontroversen – Entwürfe. Berlin: Reimer, 73-95. Penner, Hans H. 1985. “Language, Ritual and Meaning”. Numen 32/1: 1-16. Staal, Frits 1979. “The Meaninglessness of Ritual”. Numen 26/1: 2-22. — 1989. Rules without meaning: ritual, mantras and the human sciences. New York: Lang.

Axel Michaels in collaboration with Johanna Buss

The Dynamics of Ritual Formality The Morphology of Newar Death Rituals The following deals with a specific part of death ritual called sapiṇḍīkaraṇa or latyā (“forty-five”) in Nevārī performed by a Newar Jyāpu (farmer) family from Bhaktapur and a Newar Rājopādhyāya Brahman priest originally from Patan. I shall analyse the ritual, in which the departed person is basically transformed from a yet unpacified and helpless deceased (preta) to an ancestor (pitṛ), on three levels: the ritual performance at a certain date and place, the manual or ritual handbook used in situ by the performing priest, and local sapiṇḍīkaraṇa texts. My theoretical concern is ritual formality, especially the question how and to what extent ritual sequences can be transformed, left out, added and transposed, and how the priest creates using these means his own ritual referring to a set of established ritual elements more or less known to his fellow priests and customers, thus using a kind of ritual language in both a stereotype and creative way.1

1. Introduction Ritual action – as Sanskrit equivalents one could count Brahmanic-Sanskritic domestic rituals (saṃskāra, karma, kriyā), sacrifices (homa, yajña, iṣṭi, bali), festivals (utsava), pilgrimages (tīrthayātrā), religious service (pūjā, sevā), or oaths (vrata), – is generally distinguished from everyday action by certain criteria, e.g. embodiment, formality, framing, transformation or transcendence. Among these, formal-

1 The present paper is part of the project “Life Cycle Rituals in Nepal” of the Collaborative Research Centre 619 (Sonderforschungsbereich) “Ritual dynamics” of the German Research Council (DFG). I am grateful to my former student Johanna Buss for clarifying several questions regarding piṇḍa and preta, scrutinizing the tables presented below and submitting bibliographical material. I feel also deeply thankful to Niels Gutschow for providing excellent material from the field-work in Bhaktapur, Christian Bau for filming the ritual according to our instructions and with the consent of the family, as well as Nutan Sharma and Mahendra Sharma for sharing with me their detailed knowledge of the ritual practice.

100


ity, which includes stylized gestures and words, liturgical order, a certain invariance of fixed sequences or restricted codes, ranks first. Thus, Catherine Bell (1997: 139) remarks: “Formality is one of the most frequently cited characteristics of ritual, even though it is certainly not restricted to ritual per se. [...] In general, the more formal a series of movements and activities, the more ritual-like they are apt to seem to us.” And she adds: “One of the most common characteristics of ritual like behaviour is the quality of invariance, usually seen in a disciplined set of actions marked by precise repetition and physical control. For some theorists, this feature is the prime characteristics of ritual behaviour” (Bell 1997: 150). Likewise Roy Rappaport (1999: 24) takes “the term ‘ritual’ to denote “the performance of more or less invariant sequences of formal acts and utterances not entirely encoded by the performers.” For him, “Formality, i.e. adherence to form, is an obvious aspect of all rituals. It is often, but not always, through the perception of their formal characteristics that we recognize events as rituals, or designate them to be such” (1999: 33). Rappaport even states: “Whereas there is no ritual without formality, all that is formal is not ritual” (1999: 37). Indeed, to a certain degree rituals are more or less formal, stereotypical, and repetitive (therefore imitable) as well as – in principle – public and irrevocable.2 This makes them interesting for a grammatical analysis. They are therefore not primarily spontaneous, private, revocable, singular or optional for everyone. Ritual acts are not deliberately rational, “given over to technological routine” (Turner 1967: 19) for they cannot easily be revised to achieve a better or more economical goal. Therefore, formalism does constitute a central criterion in most definitions of ritual and its grammars. However, as is argued in the following, rituals are much less formal than is generally assumed, and the generative or dynamic aspect of rituals should also be given its due consideration. A major part of the ritual formalism is due to liturgy preserved in ritual texts, i.e. in manuals, local handbooks, and other written sources, which can be classified into six categories (in brackets: examples of texts used during the analysis of the concerned latyā ritual): –

Personal handbooks of the priests, normally not printed, but occasionally microfilmed by the Nepal-German Manuscript Preservation Project; such texts are generally written in a mixture of Sanskrit, Nevārī and Nepālī (personal handbook of the domestic priest Mahendra Raj Sharma, used during the ritual; manuscript, no title, size: ca. 5 x 6 cm, 31 foll.).

–

Printed manuals in Sanskrit, often published with a commentary or translation in Nepālī; such texts are used during the rituals by many pūjāris.

2 See Michaels 1999 and 2004: 235f. for a more detailed discussion of these criteria.

The Dynamics of Ritual Formality

101

–

Elaborate ritual texts in Sanskrit, written by learned known scholars and locally published from the Mahendra-Saṃskṛta-Viśvavidyālaya, Tribhuvan Viśvavidyālaya, Nepāla Rājakīya Prajñā-Pratiṣṭhāna or private publishers (Antyakarmapaddhati by Dadhi Rāma Marāsini with a commentary by Rṣi Rāma Śarmā Ghimire, Kathmandu V.S. 2056).

–

Ritual texts in Sanskrit from the great tradition with a pan-Indian distribution (Antyeṣṭipaddhati = Uttaranārāyaṇabhaṭṭī of Nārāyaṇa Bhaṭṭa, Garuḍapurāṇa-Pretakalpa).

–

Documents from private persons or from the Guṭhi Saṃsthāna related to the rituals performed.

–

Grey literature (pamphlets etc.) from ritual organisations.

However, to what extent is formality based on such texts really practised? Or how much freedom of choice and alterations is left? These questions are my theoretical concern, and I want to demonstrate the problems by pointing out a specific part of a specific death ritual, the use of piṇḍas. Methodologically, I focus on a singular ritual: the sapiṇḍīkaraṇa or latyā ritual as performed in Bhaktapur on 22nd August 2002 for the deceased person Rabindra S. (called Rabi) by Mahendra Raj Sharma (pūjāri) and Narain Kumar S. (yajamāna) who is the father (sic!) of the deceased – a point to which I will come back below. I will then try to understand the ritual by using the texts of the priests and locally used or distributed texts of the so-called great tradition. My approach is therefore inductive rather than deductive, my starting point being the actual and singular ritual practice which I do not consider as deviant, but as authentic. What happens in situ is for me not a more or less apt realization or enactment of what is textually prescribed, but a ritual performance in its own right. Given the variety of textual sources as well of ritual practice, I try to understand both the agency in rituals and the function of the texts in contexts. I hold that the agency is not only with the priest (pūjāri) but also with the yajamāna and its family, the spectators of the ritual, and even transhuman agents. From this point of view, I find rituals as well as ritual texts much less formal than they prima facie appear to be.

2. The Gift of Fifteen Balls (Pañcadaśapiṇḍadāna) Basically, Hindu death rituals3 include rituals concerning the dying person and the corpse including cremation and disposal of the bones and ashes as well as offerings and worshipping the deceased including his ritual union with the ancestors, com3 See Gutschow & Michaels 2005 for a detailed description and analysis of Newar death rituals.

102


memorative rites for the ancestors.4 These are lengthy and complex rituals, of which I will concentrate here only on a particular part, namely the offering of third series of sixteen piṇḍas. The subritual is variously called pañcadaśapiṇḍadāna, pañcadaśakalāśrāddha or caturdaśakalāpiṇḍaśrāddha, sucipiṇḍa (śucipiṇḍa), māsikaśrāddha or ṣoḍaśaśrāddha. It is performed after five kumbhakeśvaras have been worshipped (kumbhapūjā), namely – according to the priest – Sūrya, Nārāyaṇa, Sadāśiva, Gaṇapati and Varuṇa, and before the sapiṇḍīpakaraṇa (i.e. latyā) has started. The main steps of the Kumbhapūjā and Pañcadaśapiṇḍadāna ritual as performed and represented in the concerning texts are given in the following table (Table 1). In the left and middle column of Table 1 the rituals steps are given as they have been performed during the ritual of the 22nd of August 2002 along with the handbook (HB5) used by the priest for the ritual.6 This is the so-called “little tradition”. The right column list the ritual steps as given in one Nepalese ritual handbook (AKP) from the so-called “great tradition”. To be sure, this Sanskrit text was neither explicitly nor ostensively used in the actual ritual, nor has it necessarily been consulted in writing the personal handbook of the priest. It must also taken into consideration that this text is more rooted in the Indo-Parbatiyā culture whereas the material of the left and middle column of Table 1 is basically Newar. However, both sources (HB and AKP) deal with the same ritual, and both sides are Brahmanic-Sanskritic.

4 See Evison 1989: 448-469, and Michaels 2004: 133-135 (Table 11) for more detailed lists of the ritual events at death and ancestor rituals. 5 Personal handbook of the priest Mahendra Raj Sharma, Bhaktapur, dated [vikrama] saṃvat 1997 phālguṇa śudi 10 roja 7 (fol. 31r), i.e. 1940 A.D. Nepālī paper, 31 foll., size: 11,5 x 7 cm, 5-7 lines per folio, Devanāgarī script, black ink, occasional underlining with yellow (kuṃkuma) and red (haridrā) colour, some additional remarks on the margins. The manuscript was written by the father of Mahendra Sharma, partly using his memory, partly relying on similar (lost) manuscripts. For an edition and translation of this text see Gutschow/Michaels 2005. 6 At one point the priest also used another personal note-book in which he has written excerpts from the handbook of his grandfather.


103

Table 1: Ritual practice and textual prescriptions of pañcadaśapiṇḍadāna (also called māsikapiṇḍaśrāddha7) 1. Ritual of 22.8.02

2. Personal Handbook (HB) of the priest

3. Antyakarmapaddhati pp. 106-110

Kumbhapūjā Purification of the floor with cow dung. Priest draws maṇḍalas, i.e. piṇḍavedi, kumbhakeśvaravedi and kamala. śālagrāma on maṇḍala. Drawing of kumbhakeśvaravedis. Breaking of kuśa.

Bringing the kumbhas (Nev. alapo) inside the house of the yajamāna. Installation of 5, 7, 9, 11 or 21 kumbhas.

3, and 5 pieces of śāla leaves are placed near the śālagrāma.

Greeting of Śrī-Sūryārgha.

ācamana (3x). arghya to Sūrya. saṃkalpa (amuka gotra amukoddeśa caturdaśakalāpiṇḍaśrāddhe karttuṃ

(Nev.) tāy (dried fried rice) puṣpa. mixed with water is kept on those leaves as naivedya. Installation of the 5 āsana and puṣpa for Sūrya, Nārāyaṇa, Sadāśiva, Gaṇapati kumbhas representing Sūrya, Nārāyaṇa, Gaṇapati, and Varuṇa. Sadāśiva and Varuṇa on the kumbhakeśvaravedi. āsana and puṣpa for deities in An earthen cup (Nev. the courtyard. saliṃ) with banana, peas and pieces of guava is placed on top of each kumbhas. According to the priest, they should have put triphalā (harro, barro and amalā).8 One separate 7 Nota bene: The Table does not list the ritual steps simultaneously; it should thus be read vertically in each column, but not horizontically. 8 Others declare the fruits as (Nev.) hala, behala, ama. Ama is the seed of amalā, a sour fruit from the so-called tree of immortality. Harro (Skt. haritaki) is chebulic myrobalan;

104


1. Ritual of 22.8.02



saliṃ is also placed for a light near the kumbhas. Placing (Nev.) sinhamū (sindūra) in a pot near śālagrāma.

arghya water on feet and hands.

piṇḍapātra with wheat flour, banana and honey (?). Cow milk is added.

candana, yajñopavīta, dhūpa, dīpa, gandha.

yajñopavītas on kumbhas.

kumbhakeśvarastotra (mentioning also trimūrti and Sadāśiva).

Lighting of sukuṇḍā (Nevar lamp) on two rice plates. Pañcadaśapiṇḍadāna (“sucipiṇḍa”)

Ūnamāsikādipañcadaśaśrāddha 9

saṃkalpa with flower, rice and black sesame.

gandha etc.

prayoga: preparation of the ritual arena and materials.

upacāra (worship) of the five kumbhas.

dīpa (Nev. matāpūjā).

pratijñāsaṃkalpa: saṃkalpa for the performance of the complete śrāddha.

yajamāna starts to knead the dough for the piṇḍas.

Siddhikeśvarapūjā with puṣpa, hastārghya, candana, akṣata, yajñopavīta.

āsana: Seat in form of kuśa grass for the deceased.

Outside: two ladies are dhūpa, dīpa, naivedya. preparing a flower garland.

rakṣaṇa: Evil spirits are expelled from the śrāddha

The yajamāna divides the dhūpa etc. dough into 15 similar sized

hastārgha: Water for washing hands is offered to

Terminalia chebula Retz (Combretaceae). Barro (Nev. hala, Skt. bibhītaka) is belleric myrobalan; Terminalia nelleriva Roxb. Amalā (Nev. ãbaī, Skt. āmalakī) is emblica myrobalan; Phylallanthus emblica L. (Euphorbiaceae). Myrobalan is also called the “Indian almond“ (Latin myrobalanus). People say, the pitṛ like these fruits. 9 Full title: ūnamāsikādipañcadaśaśrāddhānāṃ tantreṇa (sāmūhikarūpeṇa) prayogaḥ. This is an abbrevetiated (tantreṇa) form of the ritual of presenting and worship of all Fifteen Balls collectively. For the alternative, the presentation of Fifteen Balls successively, see AKP pp. 102-105.

The Dynamics of Ritual Formality 1. Ritual of 22.8.02


and round shaped pieces.

105

3. Antyakarmapaddhati pp. 106-110 the deceased.

Black sesame is added to the piṇḍa.

Siddhikeśapūjā according to vidhi.

brāhmaṇapūjāna: Worship of the brāhmaṇa with gandha, puṣpa etc.

The priest is directing the yajamāna to place the 15 piṇḍas on the piṇḍavedi and names the piṇḍas as dvitīyakalāpiṇḍa, tétīyakalāpiṇḍa up to pañcadaśakalāpiṇḍa. 14 piṇḍas are placed, the fifteenth is placed silently on top of the 14 other piṇḍas.

Siddhikeśvarastotra.

bhūsvāmine 'nnadāna Offering of rice mixed with black sesame seeds (tila), honey (madhu), ghee and water to the forefathers as lords of earth (bhūsvāmin)

Offerings of water, flower, āsana for 15 piṇḍas kuśa water, yajñopavīta, vastra, candana, flowers, water to the piṇḍas and the piṇḍapātra is kept upside down. Worshiping piṇḍa with akṣata. Touching piṇḍa by yajamāna.

annadāna: Offering of rice with other ingredients to the deceased.

yajamāna places all piṇḍas Preparation of piṇḍas from back into the piṇḍapātra. the dough (Nev. vāle piṇḍa thaye).

madhvādipāṭha Recitation of madhuvātā... and other verses.

Washing hands.

saṃkalpa (amu[ka]nāmne dvitīyakalāmāsikapiṇḍaṃ tasmai upatiṣṭhatām).

Avanejanadāna: Pouring water on the kuśa grass as offering to the deceased.

Pouring water around the piṇḍapātra.

dto. for piṇḍas 3-14.

piṇḍadāna: Offering of 15 piṇḍas in the seize of a coconut to the deceased.

Washing of hands over the copper pot (piṇḍapātra, Nev. kvala).

pratyavanejanadāna Again offering of water on the piṇḍas.

Preparation of piṇḍa part (bhāga).

sūtradāna: Offering of threads as clothes for the deceased on the piṇḍas.

The piṇḍapātra with the piṇḍas and the lamp is taken out.

106


1. Ritual of 22.8.02



upacāra (worship) of the five kumbhas.

Offering of the piṇḍabhāga on the other 14 piṇḍas.

piṇḍārcana: Worship of the piṇḍas with sūtra, gandha, puṣpa etc. The rest of the piṇḍa-rice is scattered next to the piṇḍas.

kumbhas are taken to the shrines of the same gods.

tilodaka.

jalapuṣpākṣatadāna: Offering of jala and akṣata on the annapātra.

A married daughter of the family cleans the floor where the piṇḍavedimaṇḍala has been drawn.

Pouring tilodakārghya over the 14 piṇḍas.

akṣayyodakadāna: Offering of “inexhaustible” water.

Offering of candana, yajñopavīta, tāmkarāja and bhṛṃgārāja flowers.

jaladhāradāna: Pouring a stream of water on the piṇḍas.

gandha etc.

dugdhadhārādāna: Pouring a stream of milk on the piṇḍas.

Lifting of the piṇḍapātra and pouring sesame water over the piṇḍas.

dakṣiṇā: Gift of dakṣiṇā to the brahmins.

Turning the piṇḍapātra upside down.

viṣṇusmaraṇa: Meditation of Viṣṇu.

pitaraśivakumbhastotra. gandha etc. Worship of kumbha deities with puṣpa. Worship of Sidhikeśvara (and?) Agni with puṣpa. Dedication of the caturdaśkalāpiṇḍa as food for the journey in the afterworld. Placing of the piṇḍas in the piṇḍapātra.

The Dynamics of Ritual Formality 1. Ritual of 22.8.02


107


Circumambulation (3x) of piṇḍapātra with water. Recitation. Sending piṇḍas in order to throw them into river and sending for Ganeśapūjā. sapiṇḍīkaraṇa

Table 1 suggest that each column concerns a different ritual. A surprising complexity of ritual action and variations is given which is reached by a number of common ritual methods and techniques: (1) Substitution (pratyāmna, cp. Müller 1992: 23): Within the pañcadaśa-piṇḍadāna, three fruits (triphala) have to be offered to five deities represented by kumbhas. Since the prescribed fruits were not available during the actual ritual, other fruits had to be substituted. Interestingly, the triphalāni are not textually prescribed but by a norm the priests is aware of. Thus, he asks the yajamāna during the ritual: “Do you have triphala? (...) If you have triphala, keep it (there). If not, use three (other) fruits: harro, barro and amalā. ” Similarily, the priests ask several times during the ritual for ingredients and substitutes them ad hoc. Thus, when preparing the paste for the ṭīkās he asks: “Do you have (some) rice flour (Nev. pucuṃ)? If not (use) wheat flour (Nev. chuchuṃ), add some water to it, mix it and use for pasting the ṭīkā on the kalaśa.” It is quite common to substitute leaves for flowers, (rice) akṣata for feeding brāhmaṇas, pilgrimages (tīrthayātrās) for prāyaścitta, coins for sacrificial gifts (dakṣiṇā) etc. Thus, the priest asks for gold and silver pieces: “Have you brought a piece of gold and silver (Nev. luṃ kucā and vah kucā). If not, use dakṣiṇā (i.e. money) for this (purpose). Place one coin down.” (2) Alteration: In texts as well as in practice, often alternatives or options are explicitely mentioned, which leave choice for the priest to modify the ritual according to the situation. Thus, according to the Handbook of the priest the number of necessary kumbhas is given with 5, 7, 9, 11, or 21. Since the early Brāhmaṇa texts we find such alternatives mentioned by for instance referring to other schools or regional practice. (3) Shifting, postponement or interpolation: In Nārāyaṇabhaṭṭa’s Antyeṣṭipaddhati (Müller 1989: 154), the māsikapiṇḍas have to be removed immediately. In the ritual, however, the māsikapiṇḍas are kept aside and thrown into the river

108

Axel Michaels in collaboration with Johanna Buss after sapinḍīkaraṇa (i.e. together with tripiṇḍa) and after śayyādāna. This is instructed by the priest himself: “Throw this piṇḍa (out later). (...) It can be kept here (for a moment). Draw a circle (maṇḍala) by hand. (...) Collect these all (left overs of the śrāddha) from the flour and put it on the pot (piṇḍapātra). (...) Go there and wash your hand. (...) Now, one person shall come in front to go for worshiping (to the outside shrine). (...) A daughter or a son-in-law is okay (to do this). This (Sūrya kumbha) is for the Sun-god (Sūrya). (...) This (Nārāyaṇa kumbha) is for the God Nārāyaṇa. (...) Where is the (shrine of) God Nārāyaṇa? (...) This is to (offer to) a well (Nev. tūṃ) (or) a water fountain (hiṭī). Is there a well outside (for Varuṇa kumbha)? (...) It (i.e. the Nārāyaṇa kumbha) is allowed to offer it to Nārāyaṇa after offering (this Sūrya kumbha) to Sūrya. Take both of them (together).”

(4) Omission: At several levels, certain prescribed ritual actions are neither mentioned in the texts nor performed. Thus, the feeding of the brāhmaṇas is neither performed in the ritual of 22nd of August, 2002, nor mentioned in the Handbook of the priest; it might have been substituted by throwing rice (akṣata). The same is true for the yajñopavīta, which is not at all mentioned in the Antyakarmapaddhati though being an essential part both of upacāra and dāna elements. Generally speaking, not only texts lack mentioning what in the ritual happens, but also the ritualists leave out elements, which even according to their personal handbooks should have been ritually carried out. Thus, a Siddhikeśvarapūjā should have been performed according to the Handbook, but was not so during the ritual of 22.8.02. Another example is the position of the yajñopavīta: according to the AKP at several stages the yajamāna has to shift the thread from his left to the right shoulder and vice versa. This practice is completely abandoned in the performance of the latyā and also not mentioned in the hand book. (5) Fusion: Through the introduction of kumbha deities śrāddha and pūjā are combined. Thus, the kumbhapūjā has become a kind of pañcāyatanapūjā, although traditionally there should not be any devapūjā in the house of the deceased as long as sapiṇḍīkaraṇa has not been performed. This is particularily noticeable during the kumbhapūjā, which, in the Antyeṣṭipaddhati, is declared as water purified by three gods (i.e. trimūrti) and presented to the preta for his way to Dharmarāja or Yama. However, in the ritual of 22.8.02, the kumbhas are worshipped as seats of five deities who should protect the ritual and the preta. Thus, a dāna for the preta is merged with an upacāra for deities. As KlausWerner Müller (1992: 88f.) aptly remarked, any Hindu death ritual is a constant mixture of śrāddha, pūjā, dāna and homa. The ritual specialists skillfully play with such ritual bricks. A viṣṇuśrāddha, for example, is basically a viṣṇupūjā. However, it is mostly, but not even always denoted śrāddha simply by


109

being incorporated in an ancestor ritual. Similarily, a homa – basically a rite for gods – can easily become a rite for ancestors. (6) Reduction (tantreṇa) or abbreviation: On (Nev.) du bye(ṃ)kegu, the tenth day after death, ten piṇḍas are three times prepared in order to constitute the new body of the deceased person. This is a combination or reduction of what – according to most texts – should happen on the ten consecutive days after death. However, as the priest says, the chief mourner is so impure that he cannot go to the river without meeting others and thereby polluting them. Therefore this ritual should now be performed on a singular day. Similarily, the sapiṇḍīkaraṇa and the ābdikaśrāddha, which according to textual traditions should be performed one year after the death are nowadays mostly performed much earlier, on the 45th day or even on the 12th or 13th day. (7) Re-duplication: A common technique is to repeat ritual elements, e.g. ācamana in the beginning of the kumbhapūjā. (8) Invention: Sometimes the priests more or less invent new ritual parts or interpret them in their own way. Thus, in the actual ritual a piṇḍa is offered for unknown pretas. Moreover, during the sapiṇḍīkaraṇa priest introduced a kind of shadow piṇḍas saying to the yajamāna: “Continue to make (also) three pieces (of dough) of this (smaller) size. (...) According to my father, they are also a kind of pitṛ. (But) they are not offered after spelling their name. They are (for) the older ancestors than these ones (he points to the tripiṇḍa).” According to ĀpastambaGS and HiraṇyakeśinGS a fourth piṇḍa should be given to the ancestor in the fourth generation, i.e. preceeding the deceased great-grand-father without saying anything (Caland 1893: 7). The Matsyapuṛāṇa (18.29) also mentions that “ancestors from the fourth (i.e. the father, grand-father and greatgrand-father of the great-grand-father of (the) performer) are entitled only to the wipings of the articles of food (sticking to the hand of the performer of śrāddha). Similarly Manu (III.216) calls the three paternal ancestors after the paternal great-grand-father who benefit from the wipings lepabhāgin (see Kane 1973: IV:482f.). All these forms of variance are means to adapt ritual liturgy to the actual situation depending on the macro-, meso- or micro-level of comparison. Adaption of rituals and thus their variability rather than their strict formality has always been accepted by the Dharmaśāstrins, priests and theologians since rituals often had to be reduced and modified in times of distress (āpad) or adapted to the specific circumstances of region and time (deśakālānucāra). Moreover, they often could also be internalized (mānasa) if they, for various reasons, could or should not be performed at all. However, the variability of rituals has their limits since there are always some core elements, which can never be exchanged or substituted. They make for the

110


particular type of the ritual. Thus, in all four sources of the Pañcadaśapiṇḍadāna, especially saṃkalpa, piṇḍa, pūjā, dāna or homa, appear. Among these only the piṇḍas are specific for death rituals, whereas saṃkalpa (Michaels 2004a), pūjā, dāna and homa are also to be found in other rituals. But even the piṇḍas and other core elements are neither prepared nor used in a fixed and homogeneous way.

3. The Monthly Balls (māsikapiṇḍa) Piṇḍas are offered several times during Hindu death rituals. They are offered not only to the deceased and the ancestors, but also to gods and the clergants of Yama. The number of piṇḍas varies as well generally and according to the highly influential Garuḍapurāṇasāroddhāra three sets of sixteen piṇḍas have to be offered. With reference to the gradually decreasing impurity the first set is called malinaṃ ṣoḍaśaṃ, the second madhyaṃ ṣoḍaśam and the third uttamaṃ ṣoḍaśam (GPS 12.66 ff.) The first six piṇḍas of the first set are given on the way to and at the cremation ground, the other ten of the first set are given during the first ten days. They are meant to create a transcendental body for the deceased. Fifteen piṇḍas of the second set are offered to the gods and one only is offered to the preta. The final set of sixteen, with which I am dealing here, are also called māsikaśrāddhas or māsikapiṇḍas. The name refers to the monthly offerings for the deceased during the first year after his death before the sapiṇḍīkaraṇa. Apparently the sapiṇḍīkaraṇa was previously to be held eleven, twelve or thirteen days after the death. However, when the māsikaśrāddhas were included in the death rituals, the sapiṇḍīkaraṇa was performed after a year, but this period came again to be shortened to twelve days or so which made it necessary to ritually reduce the māsikaśrāddhas as well.10 According to the tradition, the 16 monthly piṇḍas are meant to feed the deceased during his one year long journey to Yamas world, passing 16 different cities, where he eats the piṇḍas and receives the gifts given along: 1. Saumya, 2. Sauripura, 3. Nagendrabhavana, 4. Gandharva, 5. Śailagrāma, 6. Krauñca, 7. Krūrapura, 8. Vicitrabhavana, 9. Bahvāpada, 10. Duhkhada, 11. Nānākrandapura, 12. Sutaptabhavana, 13. Raudra, 14. Payovarṣaṇa, 15. Śītāṭhya, 16. Bahubhīti, and finally Dharmabhavana, the City of Yama.11 The number 16 includes the twelve months and four additional points of time. The third set of 16 piṇḍas ist often offered at the 11th or 12th day in advance for the following year. In the special case of the latyā the timings are different: the sapiṇḍīkaraṇa is performed at the 45th day (= traipakṣe) after death, which is according to the handbooks another possible point of time for the sapiṇḍīkaraṇa.12 10 See Abegg 1921: 174 n. 1. 11 See GPS I.59. 12 See Müller 1989: 173 (AKP 4.45), GPS XIII.28.


111

The different handbooks do not agree about the exact classification of the piṇḍas and śrāddhas, and not all of them mention these three sets, but there is certainly a consciousness with the priests of the importance of this three sets. Thus, the author of the Nepalese Antyakarmapaddhati gives the complete set of rules for the death ritual including the first set of sixteen piṇḍas and the third set but omits the middle set, which usually seems not to be offered in the Nepalese ritual. The author of the AKP discusses this matter in the appendix and cites the GPS and then gives the rules for the offering of the middle set in the appendix. In Evisons table (Evison 1989: 448–469) the māsikapiṇḍas are not specially named, she generally mentions the offering of piṇḍas during the first twelve days. Table 2 presents the third set of piṇḍas (i.e. the māsikapiṇḍas) according to different sources.

112


Table 2: Māsikapiṇḍas13 māsikapiṇḍas as offered māsikapiṇḍas according to in the ritual of August 22, HB (foll. 4r-6v): 2002

māsikapiṇḍas according to AKP (p. 14):

[1. madhyama ṣoḍaśa14]

[1. prathama?]15

ādyamāsikaśrāddha; ādyābdikaśrāddha

2. dvitīya (1)16 3. tṛtīya (2) 4. caturtha (3) 5. paṃcama (4) 6. saṣṭa (5) 7. saptama (6) 8. aṣṭama (7) 9. navama (8) 10. daśama (9) 11. ekādaśa (10) 12. dvādaśa (11) 13. trayodaśa (12) 14. caturdaśa (13) 15. pañcadaśa (14) – śyeṣapiṇḍa19 (15)29

2. dvitīya (unmāsa?) 3. tṛtīya (tripakṣa) 4. caturtha (II.)18 5. paṃcama (III.) 6. saṣṭa (IV.) 7. saptama (V.) 8. aṣṭa (unaṣāna before VI.) 9. navama (VI.) 10. dasama (VII.) 11. ekādaśa (VIII.) 12. dvādasa (IX.) 13. trayodasa (X.) 14. caturdasa (XI.) 15. –

1. ūnamāsika (29)17 2. dvitīya (30) 3. traipakṣika (45) 4. tṛtīya (60) 5. caturtha (90) 6. pañcama (120) 7. ṣaṣṭha (150) 8. ūnaṣāṭ (179) 9. saptama (180) 10. aṣṭama (210) 11. navama (240) 12. daśama (270) 13. ekadaśa (300) 14. dvādaśa (330) 15. ūnābdika (359) 16. ābdikaśrāddha (360)

[16. separately offered [16. ṣoḍaśakalāstotra: fol. before the sapiṇḍīkaraṇa 9b] (16)]

13 The ordinalia (dvitīya, tṛtīya etc.) refer to different time points of the one-year-journey of the preta as well as to 16 cities in the under world which the preta has to pass on his way to Yama. 14 In this case, offered on the 11th day after death: these 16 piṇḍas count as the first piṇḍa in the third series. 15 Probably to be offered on the 11th day after death; the text has kalāprathamaparyyaṃtaṃ unmāsādikapiṇḍakaṃ. It is unclear whether unmāsādika (sc. ūnamāsika) has to be counted as the first or second piṇḍa. 16 Numbers in brackets refer to the piṇḍas as placed on the maṇḍala in the ritual itself. 17 Numbers in brackets refer to the days after death when the māsikapiṇḍas have to be offered. 18 Numbers in brackets refer to the months of the one-year journey to Dharmarāja. 19 Placed silently on top of the other 14 piṇḍas for three earlier generations or for an unknown preta.


113

Texts and the practice again differ considerably. Thus, in the ritual from 22nd of August, 2002, the first piṇḍa is not prepared in the ritual but mentally taken from the second set of piṇḍas prepared on the 11th day after death. Moreover, a piṇḍa for an unknown preta or for three previous generations is added without giving it a number. This is not a spontaneous invention of the priest but common practice in Newar Pañcadaśapiṇḍadāna rituals. Also, the sixteenth piṇḍa is offered separately before the sapiṇḍīkaraṇa in the ritual, and this ball, also called vikalapiṇḍa, is sometimes dedicated to all unpacified family members. Thus, even in a core element of death rituals, i.e. the piṇḍas, ritual formality is limited and overestimated. How, then, comes it that formality is seen as such a central category of rituals.

4. Conclusion The variability and complexity of Hindu death rituals have, of course, to a great extent to do with the liminal status of the deceased being somewhere between preta and pitṛ, between unpacified being, ghost and ancestor. The piṇḍas are food – for the deceased, for deities, for the messengers of Yama, or for the crows – but they are also identified with the ancestors and the deceased. This is especially evident, when the priest instructs the yajamāna to knead the piṇḍas with great attention. Several times he asks him to be careful and not to hurt the deceased or ancestors: “Do not hit to the piṇḍa while offering them the coins. It is something like hurting him.” Thus, the piṇḍas are food, but also represent the preta, pitṛ or viśve devāḥ. Taking rituals, texts and the commentary of the priests together, there is no formal ritual preparation and dedication of the piṇḍas, but only the consciousness that a number (mostly 16) piṇḍas have to be offered during the ritual preceeding the sapiṇḍīkaraṇa. This is the “plot” which makes the ritual a latyā ritual. All other action is also to be found in other rituals. Moreover, the offering of piṇḍas is as meaningful as it is meaningless, but this opens a discussion I better avoid.20 The agency of the priest in rituals seems to be much higher than it is generally assumed. This holds at least true for Brahmanic Hindu domestic rituals in a nonBrahmanic community. This becomes clear in the suprising substitution during the ritual of 22nd of August, 2002: the substitution of the father of the preta for his son. For, it is the father of the deceased who performs the death ritual, although the dead son had himself a son. During the ritual, when the father kneads the piṇḍas, the priest says: “It is (still) not allowed to uplift the piṇḍa from the ground... Is there a son (born from the dead person)? [The father agrees] (...) However, if he has a son, it is not necessary to make (mix) all (piṇḍas) one. If he had not have a son, (in such a case) all (the 20 See Michaels 2006 for a detailed discussion of this point.

114


piṇḍas) shall be made one... Don’t worry that it (the piṇḍa) will be broken... Now this way, move it round... Make (it) little thinner upside. Now fold it on like this. Now, again move it round. Keep pressing in this way (demonstration). (Be careful) it is going to be broken. It is not allowed to break the piṇḍa”. To be sure, the son mentioned by the yajamāna was born shortly before his own son died. He therefore was too young to perform the death rituals. In most if not all texts21 it is explicitly forbidden for a father to perform the antyeṣṭi and śrāddha rituals for his son since that would imply that the succession of the ancestors was broken down. However, the agency of the priest in the ritual is remarkable. He can alter the presumptive rules, according to the situation although since Vedic times the ritual succession of father and son is well established. To sum up: The great formal variety within both ritual practice and ritual texts has to do with the fact that ritual action has to be understood as a kind of grammar which makes up for a creative use of its basic elements and structures. Neither the symbolism nor the meaning of rituals can explain this possibility of variation. To be true, there is a relation between a core action or plot of the ritual (pradhāna) and sub-rituals (aṅga), but the formality is not fixed by a limited and restricted procedure and protocol or liturgy, but by using more or less deliberately a known set of ritual elements, certain basic structures of action and decorum from various sources. This holds at least true for the rituals I have taken into consideration and observed. It might be more strict with complex (temple) rituals, which involve a great number of priests and other ritual specialists who control each other. It might be different in purely Brahmanic or spiritual traditions. However, for many domestic rituals in non-Brahmanic communities of South Asia it becomes now clear that the varieties and the possibilities for change are much greater than the restrictions, established rules and limitations. In other words, ritual formality refers to a kind of floating inventory of ritual action and decorum, which has to be bound and mixed by several agents to more or less meaningful action. Thus, it is not the formality but the variability in the use of forms that explains the dynamics of rituals. From this point of view priests and other ritual specialists of domestic rituals of South Asia are to be graded as ritual artists rather than administrators of rituals.

21 See Kane 1973: 257, GPS XI.19.


115

References I. Texts AKP

AP

GP

GPS

HB PM

Antyakarmapaddhati by Dadhi Rāma Marāsini with a commentary by Ṛṣi Rāma Śarmā Ghimire. Kathmandu: Mahendra-Saṃskṛta-Viśvavidyālaya, V.S. 2056. – XVIII, 285 pp. (Mahendra-Saṃskṛta-Viśvavidyālaya-Granthamālā, vol. 38) Antyeṣṭipaddhati (= Uttaranārāyaṇabhaṭṭī) of Nārāyaṇa Bhaṭṭa. Ed. by Vāsudeva Paṇāśīkar. Bombay: Nirṇaya Sāgara Press, 1915 German Transl. by Klaus Werner Müller, Das brahmanische Totenritual nach der Antyeṣṭipaddhati des Nārāyaṇabhaṭṭa, Stuttgart: Franz Steiner Verlag, 1992 (Beiträge zur Südasienforschung, Südasien-Institut, Universität Heidelberg, vol. 151). Garuḍapurāṇa, Uttarakhaṇḍa (GarPur-UKh) (Pretakalpa). Ed. by J. Vidyāsāgara. Calcutta 1890 [Summary in Abegg 1921: 8–27]. Garuḍa Purāṇa. Trsl. and annotated by a board of scholars. 3 vols. Delhi 1978–1980. Garuḍapurāṇa-Sāroddhāra: German transl. by Emil Abegg, Der Pretakalpa des Garuḍa-Purāṇa (Naunidhirāma’s Sāroddhāra). Eine Darstellung des hinduistischen Totenkultes und Jenseitsglaubens. Berlin und Leipzig: Vereinigung Wissenschaftlicher Verleger (Walter de Gruyter & Co), 1921. Personal handbook of the Mahendra Raj Sharma; manuscript, no title, size: ca. 11,5 x 7 cm, 31 foll. Antyapaddhati (Pretamañjarī) with a Nepāli commentary (Nepālī bhāṣā-ṭīkā sahita), no author. 3rd. ed. Vārāṇasī: Jaya Nepāla Prakāśana, 1999.

II. Secondary Literature Abegg, Emil: see I. Texts, GPS. Bell, Catherine 1997. Ritual. Perspectives and Dimensions. New York, Oxford: Oxford University Press. Caland, Willem 1893. Altindischer Ahnencult. Leiden: E.J. Brill. Evison, Gillian 1989. Indian Death Rituals. The Enactment of Ambivalence. Oxford (unpubl. Ph.D. thesis). Gutschow, Niels & A. Michaels 2005 (with contributions by Johanna Buss and Nutan Sharma and a film on DVD by Christian Bau). Handling Death. The Dynamics of Death and Ancestor Rituals among the Newars in Bhaktapur, Nepal. Wiesbaden: Harrassowitz. (Ethno-Indology – Studies in South Asian Rituals; vol. 3). Kane, Pandurang Vaman 1973. History of Dharmaśāstra. Vol. IV. Poona: Bhandarkar Oriental Research Institute.

116


Michaels, Axel 1999. “‘Le rituel pour le rituel?’ oder Wie sinnlos sind Rituale?” In: Corinna Carduff und Joanna Pfaff-Czarnecka (eds.), Rituale heute. Berlin: Reimer: 23-48. — 2004. Hinduism. Past and Present. Princeton: Princeton University Press. — 2004a. “Saṃkalpa: the Beginnings of Rituals”. In: J. Gengnagel, U. Huesken and S. Müller (eds.), Text and Context of Rituals in South Asia. Wiesbaden: Harrassowitz (Ethno-Indology – Studies in South Asian Rituals ; vol. 1): 45–64. — 2006. “Ritual and Meaning”. In: J. Kreinath, M. Stausberg and J. Snoek (eds.), Theorizing Ritual. Leiden: Brill: 247–263. Müller, Klaus-Werner 1992: see I. Texts, AP. Rappaport, Roy 1999. Ritual and Religion in the Making of Humanity. Cambridge: Cambridge University Press. Shastri, Dakshina Ranjan 1963. Origin and Development of the Rituals of Ancestors Worship in India. Calcutta: Bookland. Turner, Victor 1967. Forest of Symbols. Kingport Press.

Cezary Galewicz

Inscribing Scripture Through Ritual On the Ritual Cycle of Trisandhā Je désigne par écriture l’activité concrète qui consiste sur un espace propre, la page, à construire un texte qui a pouvoir sur l’extériorité dont il a d’abord été isolé.1 To inscribe may entail more than just to write down. A cultural gesture of ceremonial inscribing may have a binding social force, or a powerful communicative appeal. All the more so if put into a ritual frame, subjected to a liturgical order. Inscribing scripture in the course of a specific ritual entails an operation of selfreflection on the part of the community and a play upon a particular relation between its identity and the scriptural corpus expressed through ritual action. Rather than addressing an otherwise most interesting question of whether a general “grammar of rituals” is capable of being articulated I am going to focus in this short case study on some aspects of possible analogies between text, or a particular textual practice, and ritual. This calls for a slightly different angle of thinking but addresses a problem of the fruitful use of analogy between two structures of two systems of representation. The reason for that is the specific form of this ritual which stands half way between the solemn, public and domestic rites, escaping clear assessment. Whether classed as meaningful or meaningless, understanding a ritual requires a “reading” of its structure, the dynamic of its proceedings as well as its relationship to the broader social context. All of that seems to be possible only after identifying basic building blocks of the ritual in question and the algorithm of their sequence. The latter does not have – as the present example shows – to be simply linear. Since the case related here is one of a rare ritual, a one-time chance to be seen, only repeated analysis of its video-recorded proceedings allowed for a sorting-out of its basic building units and a syntax-like pattern of their combination in sequences. The only way of representing this analysis (beside a short video 1 Michel de Certeau, l’Économie scripturaire.

118

Cezary Galewicz

screened during the conference panel) proved to be that of a graphic scheme proposed below in Fig.1. The figure represents in a graphic pattern a morphology of the basic ritual units of Trisandhā put to an organising algorithm of sequence, repetition and variation of level and modality. Because of the reasons that become clear below the pattern refers at the same time almost exclusively to the verbal component of the ritual under study. For a topic such as “grammar of rituals” the case of Trisandhā seems to present a rather liminal example: its core appears to be composed of unitary blocks made just of linguistic utterances, disciplined as they are by their canonic form, but not necessarily meant to be understood in their primary semantic message neither by the participants nor the audience. At the same time the unitary blocks are put to an organising discipline of an algorithm that may be seen as a sort of ritual syntax. Trisandhā2 drew rather little, if any, attention from the academic world, save for a few remarks,3 and one, rather fragmentary description,4 perhaps because of the few textual sources regulating its procedure and limited accessibility to its actual practice.5 Nevertheless, Trisandhā appears to be linked to important early Indian concepts of scripture, scriptural canon and the collective management of its resources, and as such deserves more systematic attention.6 One such link concerns the related concepts of svādhyāya and brahmayajña – Vedic memorisation/study and a sacrifice said to be performed with Vedic scripture as substance matter. It has been noted that the presumably original idea of an independent rite of brahmayajña-svādhyāya – conceived of in Taittirīya Āraṇyaka II as a “sacrifice of scripture” – proved to be largely redefined in its later appropriations.7 According to the early Indian ideology of sacrificial ritual in the form systematised by the exegetical school of Mīmāṃsā, and true to the evidence of the actual practice of the solemn Vedic sacrifice called śrauta (itself a derivation of śruti – the name for the “scriptural” tradition), only selected parts of the canonical collection of the Ṛgveda find immediate ritual application in the Vedic śrauta system. Those parts which are of no śrauta ritual use are left outside the training system for preparing śrauta performers, and thus run the risk of being slowly erased from the collective memory. Although memorised at an early age, with hardly any prescribed application in the traditional śrauta perspective, they are 2 This study is primarily based on the field research carried out between 2000 and 2008 and it remains complementary to my earlier presentations of other rituals of Kerala Nampūtiri Brahmins (Galewicz 2003, 2004 and 2005). 3 Pisharoti 1928, Staal 1961, Galewicz 2004. 4 Govindan Namboodiri 2002. 5 Cf. Pisharoti 1928: 706. 6 The terms “scripture” and “scriptural canon” and “inscribing scripture” shall be used in the course of this short study with all due reservations required for the concept of a canonical corpus preserved through predominantly oral tradition. 7 Malamoud 1977: 60–61.

Inscribing Scripture Through Ritual

119

bound to be eventually forgotten. To guard against this situation, an old imperative of ritualised daily recitation of svādhyāya seemed a natural ally of the otherwise endangered textual tradition of the Ṛgveda. Yet it was not the Ṛgvedic circles where the idea of svādhyāya appears to have been formulated first. In a range of brahmanical prescriptive literature of the post-Vedic period one can find references, albeit not infrequently contradictory, to brahmayajñasvādhyāya as an independent rite in itself, believed to be one of the five so-called “great sacrifices” (mahāyajña).8 It appears to have been conceived of as a continuous sequence of everyday recitation of the Veda resuming every new day from the place where it was left the day before. In that respect it may be of interest to recall a distinction between “svādhyāya of apprenticeship” and “svādhyāya of brahmayajña”.9 While the first belongs to the stage of life characterised as brahmacarya (roughly: chastity), the second comes closer to that of the householder’s realm of duties. The difference also bears on the form of recitation: in the former case it is to be traisvarya, i.e. respecting all the proper accentuation of three Vedic svaras, in the latter it may (it is recommended) be ekaśruti (monotone).10 Though a ritual engaging extensive Vedic recitations, Trisandhā does not belong to the multivedic śrauta fold. Its basic textual systematisation (Trisandhāpaṟipāṭi) refers in the very beginning to the idea of svādhyāya and the recognised duty of Vedic study by the quotation from Taittirīya Āraṇyaka II.15.7. Thus it seems to draw from the strand of anti-śrauta ritualism initiated in the later Vedic period and systematised in the idea of the five great sacrifices (pañcamahāyajña) of Taittirīya Āraṇyaka. But in what sense can svādhyāya be a sacrifice? What is being sacrificed and to whom? we may ask. According to C. Malamoud: “il ne s’agit pas de rendre culte aux ṛṣi au moyen du texte vedique, mais rendre culte aux Veda au moyen de la récitation. Ce qui implique que le mot svādhyāya est ici un nom d’action, et non un synonyme de Veda: svādhyāyaśabdas ca kriyāśabdo nātra vedavacano yathā svādhyāyo ‘dhyetavyaḥ.”11 Other traditional authorities hold that brahmayajña-svādhyāya is a sacrifice for the brahman in contradistinction to the rest of the five mahāyajñas. While the original concept of brahmayajña-svādhyāya developed in a number of texts of post-Vedic prescriptive literature into a widely acknowledged duty for every true brahmin, later schools of thought made specific use of it. Some of the Mīmāṃsā thinkers took the isolated formulation of svādhyāyo ‘dhyetavyaḥ to represent a religiously sanctioned injunction for memorising one’s family Veda. Thus, 8 See for instance Āśvalāyanagṛhyasūtra III.1, Manusmṛti III.60, 70. 9 Malamoud 1977: 52. 10 The prescriptive literature does not state unanimously whether during the solemn śrauta rituals Vedic texts should be recited in their trasvarya or bhāṣika form. Or, maybe in a still different one? (cf. Malamoud 1977: 55, fn.2), though most often it is taken to be the latter case. 11 Malamoud 1977: 57, fn. 5.

120

Cezary Galewicz

they dissociated the idea of svādhyāya from that of brahmayajña and its individual touch. For Mīmāṃsā the act of svādhyāya was no longer an individual, solitary experience of the scriptural text but a reproduction of a collective identity of a community of practice defined by belonging to a particular Vedic śākhā. Later Mīmāṃsā tended to understand svādhyāya as twofold: a duty of memorising the Veda by a young brahmacārin and a duty of transmission of the Vedic heritage to a pupil through the act of adhyāpana or “teaching”. In the course of this reformulation the original idea seems to have been largely lost or distorted.12 A discussion on the actual scope of this basic Vedic vidhi (injunction) reverberates throughout the works of the adherents of the Mīmāṃsā school of thought for over a millennium from Śabara through Kumārila up to the times of Sāyaṇa in the fourteenth century C.E. who touched upon the problem from his own eclectic positions as a Vedic commentator.13 Almost none of the Mīmāṃsā authors proper, however, seems to have considered it important to comment upon brahmayajña-svādhyāya as independent ritual pertaining to later stages of life. In particular no reflection can be traced considering the brahmayajña-svādhyāya as an important form of prolonging the Vedic study of the brahmacārin into adult age and the situation of the householder.14 The present short study concerns the ritual practice of Trisandhā that supplies, in my mind, this missing link and testifies to an instance of a genuine regional appropriation of the original idea expressed in Taittirīya Āraṇyaka.15 At the same time it offers important insight into the nature of the mutual relationship between the scriptural canon of the Veda and historically and regionally distinct communities of its guardians, custodians and users. Most of the essential elements making up the ritual cycle that I am going to present here happen to be used also outside its specific frame. They can be seen in other configurations, making part of other, more frequent ritual frames, while remaining part and parcel of a shared liturgical order of the Nampūtiri Ṛgveda. What gives the ritual of Trisandhā its identity is definitely its distinct, genuine pattern of Vedic recitations (re)constructed each time with a set of fixed elements by the participants who must be well versed not only in its vocabulary, but also in its grammar and syntax. This roughly means that the participants of Trisandhā are supposed not only to have memorised enough of the Ṛgvedasaṃhitā texts but also to inte12 Malamoud speaks of “glissement” or “gauchissements” in the process of appropriation and reformulation of the concepts of svādhyāya separated from brahmayajña. See Malamoud 1977: 60. 13 For a discussion of positions of Sāyaṇa against Mīmāṃsā thinkers, see Galewicz 2009a. 14 Cf. Malamoud 1977: 56: “Le brahmayajña, à n’en pas douter, est un sacrifice. Mais il est aussi une des formes par lesquelles se prolonge l’étude.” See also Galewicz 2009a. 15 Cf. Malayāḷam commentary to Trisandhāpaṟipāṭi, p.5: trisandha ekakartṛkam alla. nānakartrkam āṇu. (“Trisandhā is not a single-person affair but a collective enterprise”).


121

riorise the rules that allow for textual operations on the elements of its textual body, such as can be seen in permutations, declinations and conjugations, as well as the options for the overall pattern of daily proceedings to be followed throughout the sessions of ritual recitation. At the same time we must admit that to understand the working of this ritual within its wider social context entails an effort to identify the web of mutual interrelations between the ritual practices of the community and the claims this community lays to its own identity and status against the other rival Nampūtiri Ṛgvedic association,16 as well as within the broader community of Nampūtiris and in competition with other groups aspiring to brahminhood in the region. Hardly any evidence is available as to the antiquity of Trisandhā. To my best knowledge, available written sources are not earlier than the end of the eighteenth century.17 Even so, they are said to record and systematise a much earlier tradition. Its contemporary practitioners appear to still draw from the collective memory rather than from written artifacts regulating its procedure. For a number of reasons to be commented upon below, Trisandhā represents a rather rare example of a ritual, believed to be a variety of upāsana, or offering, adoration and worship at the same time. While an idea of offering appears to form its core, it remains focused on its very verbal component in the shape of the scriptural canon which is believed to be the very substance of the offering, just as it is in its late Vedic paradigm of brahmayajña.18 Held in highest esteem by the members of the community of Nampūtiri Ṛgvedins, the ritual of Trisandhā appears slowly to recede into the shade of oblivion owing to its hermetic character and demanding form. At the same time other Vedic traditions of the Nampūtiris – the spectacular public rituals of śrauta sacrifices – seem to gain momentum through various revivalistic initiatives that attract new audience to their elaborate proceedings.19 Whatever is the reason for this new phenomenon, it seems that both today and in the past some rituals suc16 The families of Nampūtiri Ṛgvedins remain divided into two traditionally rival associations (Mal. yogam). This tradition dates back to the former competing kingdoms of Samūtiri and the king of Cochin, each of the two patronising their own Ṛgvedic school located in Tirunāvāya and Trichur respectively. 17 A manuscript in the Whish collection dated to the second half of the eighteenth century contains a short text called Trisandhālakṣana (Winternitz 1902: 93). 18 See Taittirīya Āraṇyaka II.10.6, 15.7, cf. also Śatapatha Brāhmaṇa 11.5.7.4. In its early form the sacrifice of brahmayajña appears to be conceptualised as a “rite of friendship” (Malamoud 1977: 32) between a brāhmaṇa and his Veda and its aim is conceived as paying off the debt towards the ṛṣis – the authors of the Vedic texts. The eventual remittal of this original dept appears to be possible only after turning oneself into a creditor through thoroughly interiorising the Veda in one’s memory. See Malamoud 1977: 32–33. For a formulation of svādhyāya-vidhi, see Āśvalāyana Gṛhyasūtra III.10.2–3. 19 For a discussion of some aspects of this new revival in śrauta ritualism, see for instance Mahadevan & Staal 2003, Galewicz 2003.

122

Cezary Galewicz

ceeded and some failed to mobilise their own community’s or the wider public’s attention. Whether this should necessarily have any direct relation to their particular morphology, I am not sure. However, in order to identify and articulate their message and describe their power of appeal carefully tailored analogies to language and grammar might be of considerable help. Here, a graphic representation of the pattern of concatenated units of ritual is proposed in an attempt to bring into relief the actual structure and dynamics of the ritual under study. Since Trisandhā remains exceptionally focused on text and recitation, the method of its representation by this very fact tends to remain closer to language and its morphology. Thus the terms used here to this purpose, such as index marking of beginning and end, or index marking repetition, refer basically to operations on text portioned into blocks in order to be manipulated. To call Trisandhā a ritual cycle refers to its prolonged cyclical form requiring a number of concatenated sessions conceived of as strings of three to four adhyāya days (recitation days) each. Trisandhā makes a somewhat strange example of a ritual complex with rather blurred agency and purpose. At first sight, the former apparently belongs to the unspecified collective, though made up of selected individuals, and the latter appears only to serve the need of keeping the community’s memory fresh. From the perspective of traditional Indian classifications, Trisandhā stands probably somewhere in between “public” and domestic rituals. Its immediate concern lies with the part of the Vedic canon transmitted in the tradition of the Ṛgveda as saṃhitā, or a collection of verses, by the families of the Ṛgvedī Nampūtiris. Thus, it is the very basic part of the “scripture”, or rather its recitation competence, that remains within the primary focus of its ritual performance. It is important to keep in mind that, while Nampūtiri Ṛgveda remains an affair of skilful memory and recitation competence, it does not appear in its full shape (as the printed editions mostly do) save for a ritualised recitation ceremony such as Trisandhā.20 The extraordinary events of Trisandhā are rarely seen today, if anyone from outside is permitted to do so at all.21 The last Trisandhā to be held in the place referred 20 Another rite of Ṛgveda recitation called lakṣārcana (“a hundred thousand stanza adoration”) comprises the whole stretch of the Ṛksaṃhitā but only in its single, most basic modality with no display of the full variety of the recitation competence. It is made part of Hindu temple ceremonies and has no independent standing as Trisandhā does. The same is the case with muṟa-japa which remains an affair dominated by the Travancore princely claims to prestige involving other brahmanical groups, and not focused on the specifically Nampūtiri way of representing the Ṛgvedasaṃhitā. 21 The Trisandhā session witnessed by me was organised by the community of the Tirunāvāya yogam in the premises of the once prosperous but now deserted Ṛgvedic school of a maṭham type, by the Bharatapuḻa river in Mallapuram district. It was a session in a cycle that started in May 2004 and, unfortunately, has not been completed as yet and the chances for its continuing are rather slim. Already in 1928 Pisharoti mentions that “… on account of the heavy


123

to in the present description had been organised in 1969.22 A Trisandhā cycle is said to have a religious significance and may take place also in a temple or a monastery-like institution of a maṭham. A temple milieu is not a must, however, for Trisandhā is seen first and foremost as a display of learned skills and knowledge, though this is the case of the sacred knowledge of the Veda and the display itself is highly ritualised. Trisandhā is an event of considerable prestige for a community that formally announces its readiness to perform it and manages to successfully complete it. This means a community of Nampūtiri Ṛgvedins belonging to one of the two traditionally rival associations (yogam) of Ṛgvedī families, that of Tirunāvāya and that of Trichur. The two boast a long tradition of separate Vedic monastery-like schooling institutions (brahmasvam maṭham) and distinct histories dating back to the rivalry of the kingdom of Samūtiri and the kings of Cochin, both competing in patronising their own Ṛgvedic schools. The two associations (and their schools) have been traditionally locked in a never-ending rivalry and developed a series of social institutions focused on ritualised Vedic competition. Trisandhā represents one of them. While the truly agonistic form of the actual proceedings of Trisandhā is rather gone today, the memory of the good old times is still very much present not only in the recollections of the participants but also in the traces it left on the actual pattern of Trisandhā proceedings. The Malayāḷam word trisandha/trisandhā,23 appears to be a cognate of a Sanskrit compound of tri + sam + dhā, expressing the idea of putting together, collating, conjoining [of three basic modes of Ṛgvedic recitation] though no Malayāḷam dictionary known to me seems to contain such an entry. It must derive from the same root dhā with sam suffix from which the Sanskrit term saṃhitā is formed24 and may suggest a culturally cognate work of putting together disjoined elements so that they can be systematically arranged in one corpus organised to predetermined rules and hierarchy. This entails a basic effort of organisation as a token of textual heritage mastering. An approximate English rendering of the name could

expense and the number of day’s sittings involved in conducting it, it is found celebrated not so commonly …” as other Vedic festivities of the Nampūtiris (Pisharoti 1928: 707). 22 After this, there was one more occurrence of Trisandhā elsewhere, namely in Trichur in 1992, which was held at the premises of the triple Trichur maṭham (Vaṭakku, Naṭuvil and Ṭekku). 23 While nowadays usually rendered in Malayāḷam with short “a” at the end (thus in the title of its ephemeral printed edition), the name appears to be known also in the other form and it is thus written in the Vivaraṇa to Trisandhāpaṟipāṭi as well as in the MS. in the Whish collection mentioned above. 24 It is perhaps of some interest in the present context to note that according to the Monier Williams dictionary: saṃdhā + vācam = “to hold or interchange conversation”, and a dialogue-like pattern is indeed to be seen in the architecture of the Trisandhā proceedings.

124

Cezary Galewicz

perhaps be the “triple arrangement” or the “triple conjoining”.25 And that which is to be joined together is said to be the three basic modes of Ṛgvedic recitation, in other words the acknowledged three different basic (prakṛti) forms of the Ṛgvedic text. These three “modes of recitation/basic types of recital” (this rendering is far from being precise), are said to be varga ([continuous within] a group of stanzas) pada (word by word) and krama (step by step).26 Thus, the trisandhā understood to be a complex pattern of recitation is said to have three members (aṅga) and proceed as a form of pārāyaṇa, or “going over [the entire text]”, according to a prescribed procedure based on a sort of algorithm [see Fig.1]. There are three regional models for this lengthy and sacred ceremony identified by distinct patterns of recitations. They are described in detail in the Trisandhāpaṟi-pāṭi. The one staged at Tirunāvāya maṭham was said to require no less than 200 adhyāya days (recitation days) – which, depending on the length of intervals, may take from eight months up to approximately two years to complete, with a succession of recitation sessions of roughly three to four days each.27 They are to cover the bulk of the Ṛgvedasaṃhitā consisting of 64 chapters (adhyāya) each comprising differing numbers of textual units called vargas (groups of stanzas). As mentioned above, it is the only ritual of the Nampūtiris in which the whole of their Ṛgvedic recitation competence finds full display. It thus stands to reason that Trisandhā makes a pretty long and challenging ritual cycle requiring the presence of the best experts, whose number has been dwindling. It is made up with a number of consecutive sessions consisting of ceremonially enacted congregational recitations (vedasya pārāyaṇa) following a pattern of repetitions and change of mode and register. During Trisandhā the community’s collective memory of the textual corpus of Ṛksaṃhitā effectively gets meticulously “scanned” with lacunae patched and doubts cleared. The rules of the proceedings allow no progress if a mistake creeps in. If that happens the whole section of the text under recitation should be repeated until recited correctly.28 25 Nampūtiri Yajurvedins hold a ritual of their own named pañcasandha. Its name seems to express a parallel idea as concerns a concept of a totalising rendering of a scriptural text in one stretch while representing it in a pattern weaving together different modes of recitation. 26 Trisandhāparipāti 2b. For translation of this stanza, see below. 27 The Trisandhā cycle referred to here commenced on 21 May 2004 with the first session lasting three days. I had the first session covered by a professional Nampūtiri cameraman and personally witnessed and documented another one which took place on 28–31 January 2005. The organisers agreed to my exceptional presence at their traditional maṭham premises in Tirunāvāya on three rough conditions: that I didn’t take a bath in the same pond as the participants, that I refrained from touching any of the chanters after they had taken their ritual bath, that I refrained from entering or touching the platform upon which Vedic recitation would take place. 28 A long repetition due to a detected error was actually witnessed by me during the Trisandhā session in January 2005.


125

The rules for the procedure are hardly to be seen in writing, however. The only exception seems to be the Trisandhāpaṟipāṭi29 (written in Sanskrit verse at the beginning of the last century by K.P. Kṛṣnan Bhaṭṭatirippāttu) which is believed to have recorded a much earlier oral tradition. It is supplied with a commentary by Parameśvara Bhārati Swāmiyar who wrote a vivāraṇa on the basic text in Malayāḷam. It is this commentary which refers to the basic text as Trisandhāpaṟipāṭi, a name actually remembered by the insiders of the Nampūtiri Ṛgvedī community but substituted with a simple short title of Trisandha in the, unfortunately, much corrupted Malayāḷam script version printed locally by a small publishing house dedicated to preserving the important treasure texts of the unique Nampūtiri culture. A daily pattern of Trisandhā can be seen as being neatly embedded in a complex of routine brahmanical rites of purification and preparation of the body. It commences early in the morning after a seva, or adoration ceremony, for the guardian deities of the place, said30 to be Gaṇapati, Devī (Sarasvatī) and Vyāsa represented, however, not by idols but by standing brass lamps (nila viḷakku), a typical Keralan custom. Sometimes Ayyappan or – especially in Tirunāvāya – Śāstā (in its form adored in a nearby river island sanctuary) is also revered, the latter one being recognised as a tutelary deity of the community (yogadevatā) and important for understanding the sense of communal identity. Before that, still in the dark, all the participants, the core of whom consist of elderly well-known Ṛgveda virtuosi, one by one take their morning ablution bath in a water tank adjacent to the maṭham premises and execute proper prātar sandhyavandana rites, or salutation to the morning light, as well as due purification and marking of the body. Thus, ritually purified, dressed properly in simple white cloth and with their tied śikhā hair tufts they usually follow with sūryanamaskāra, or salutation to the morning sun. Only then, having completed their morning duties in the way they were once trained to do during the Vedic education days, they slowly mount a wooden platform situated in the courtyard on which vedasya pārāyaṇa or ceremonially “going over” the Vedic text will take place during three daily sessions. While mounting the platform each of them pays respect to the place with a gesture of salutation (right hand touches the ground and then the forehead) and prostrates himself before the guardian deity of the place of Ṛgvedic recitation, represented again by a standing brass 29 Trisandha. The organisers of the Trisandhā meeting-session witnessed by me kept a locally printed version of Trisandhāpaṟipāṭi and handwritten notes concerning one of the three optional versions of the correct procedure to follow. The participants, who hotly disputed details of the procedure, never consulted these texts which seemed to have been kept at hand as pieces of aide-memoire in case something went wrong A manuscript of this work along with another containing a technical manual are said to be deposited with Taikkāṭ māna – one of the traditional hereditary guardian families of the tradition of Nampūtiri Ṛgvedins. The participants in the Trisandhā session did not recall any other textual source. A short manuscript entitled Trisandhālakṣaṇa is in the collection of a London Library (See fn.17 above). 30 Trisandha: 24.

126

Cezary Galewicz

oil-lamp (viḷakku).31 The participants take their seats on the platform according to a pattern of positions which is said to be meaningful and corresponds to the recitation sequence and a certain level of progress in the proceedings of the ritual chanting to follow. One of them sits in a position named as that of a teaching preceptor (ācārya) and faces the others who for this occasion are called the students (śiṣyas). There are usually four to ten śiṣyas sitting along the three remaining sides of a quadrangle. In this way a situation of the Veda transmission from the gurupreceptor to pupils is ceremonially enacted and this overall scheme of repeating the words of the teacher by the pupils is retained throughout the recitations spoken on the platform. Participants will take turns in the position of ācārya for a duration of a full textual unit of one varga. This general rule of the whole procedure appears to match the prescriptions of the svādhyāya injunction which conceives the duty of preserving the text of the Veda as twofold: the duty of learning/interiorising (adhyayana) and the duty of teaching/transmitting (adhyāpana). This is not the case, however, with the additional chanting of vikṛti performed outside the platform, during the communal meal served twice a day (see below). Two other, more sophisticated modes of recitation that happen to be performed during meals are considered not to form a part of trisandhā in its narrow sense.32 On the other hand the same two recitation modes are believed to fit the occasion of ceremonial feasts to the Vedic brāhmanas that are considered as important parts of ritual complexes like Trisandhā in its broader sense. Before any of the two starts, however, or truly before the feast itself commences, a peculiar form of quasi-recitation can be seen that is also considered extraneous to trisandhā proper but most characteristic for ritual recitations of Nampūtiri Ṛgvedins. It is called koṭṭu, which in Malayāḷam seems to mean “drumming”. In fact, it is rather a humming sound produced by reciters chanting a Ṛgvedic stanza with their mouth closed and their nose held between two fingers. Both koṭṭu as well as the two meal-time forms of chanting (named jaṭā and ratha) represent important remnants of the previous agonistic form of this ritual cycle where the guests who were expected to arrive suddenly in order to challenge the host party’s ability to go through Trisandhā had to engage in these demanding exercises of Ṛgvedic virtuosity. As indicated above, a varga, or a group of stanzas, remains the basic measuring unit marking the progress of recitation while the acting teachers and pupils alternate in their positions on the platform and continue with the pārāyaṇa. A rough idea of the scale of the enterprise may be given by a reminder of the fact that the 31 Different answers were given by the Trisandhā participants to the question about the identity of the deity presiding over recitation: either brahman, iśāna, “the almighty one,” or no specific name of a deity was given. 32 I use the word trisandhā (lower case, italic) when referring to the core of Trisandhā in the shape of a particular recitation pattern, and Trisandhā (with initial capital) when referring to the whole ritual complex.


127

collection of Ṛgvedasaṃhitā contains as many as 2,006 vargas. It seems that one simple session of recitation tends to cover one varga recited according to a characteristic pattern of concatenation on three levels corresponding to three modes of recitation with their respective units of repetition (Fig.1). If successfully completed (no errors occurred), each varga ends with the words virāmastāvat spoken together by the preceptor and the students, after which all again prostrate themselves before the lamp representing the deity and, paying respect to the place of Vedic recitation, descend from the platform. Usually three to four recitation sittings are performed daily, the first one starting after sunrise, a second one after the first meal and the third one in the afternoon, when the second meal is over. It appears that an ideal progression is marked with one full adhyāya completed within a three-day long Trisandhā session. In this case a simple calculation shows that it takes sixty-four meeting sessions of three days each for the community to successfully “go over” the whole of the Ṛksaṃhitā. This result agrees with the approximation of two hundred recitation days voiced by a participant in a January 2005 session.33 A few important landmarks of the daily procedure of Trisandhā appear to deserve mentioning. However, before actually starting a Trisandhā session a formal act of demarcating the beginning of Trisandhā takes place. Though usually not admitted as such by the performers themselves, a form of saṃkalpa – or a gesture of ceremonially declaring “intention”34 – marks the operation that transforms a daily procedure of teaching-learning into a ritual action, and the recitation of scriptural text into the substance of the sacrifice.35 The so-called śatadhāram – a mantra of Ṛksaṃhitā 3.26.9 – is traditionally used as an opening maṅgala vākyam, or an auspicious introduction to the trisandhā proper36 and it is mirrored by another auspicious mantra recited as a mark of completing and leaving the ritual circle (see below). While mounting the wooden platform and sitting down among other co-performers, each of the participants mur33 An approximation by Govindan Namboodiri seems not to take into consideration that Trisandhā is rather not performed in one stretch, but at a number of meeting sessions, so while it is more or less true that seven months’ time (200 hundred recitation days) is needed to complete the Ṛksaṃhitā, actually the whole cycle may take more time, up to two years with 2–3 meeting sessions of roughly three days a month. 34 For a discussion of saṃkalpa as intentio solemnis, see Michaels 2005. 35 According to Taittirīya Āraṇyaka II.11 it is the preparation of the body, purification, position, attire and a solemn attitude that make an everyday recitation a ritual of brahmayajña. The solemn declaration of the will to start a Trisandhā is customarily made by the assembly of the yogam on behalf of the community and marks the beginnings of preparations for its actual enactment (personal communication by M.J. Namboodiri). 36 śatádhāram útsam ákṣīyamāṇaṃ vipaścítam pitáraṃ váktvānām | meḻím mádantam pitrór upásthe táṃ rodasī pipṛtaṃ satyavā́cam || (“The spring that fails not with a hundred streamlets / Father inspired of prayers that men should utter / The Sparkler, joyous in his Parents’ bosom / - him, the Truth-speaker, sate ye, earth and Heaven” (tr. Griffith).

128

Cezary Galewicz

murs this auspicious mantra, then touches his ears and the earth in front of him and only then proceeds to join in the chant. The number of acting pupils on the stage fluctuates whilst keeping a minimum of four and thus allowing participants to rest off the platform during the prolonged sittings. The performer who was selected for executing the role of the first ācārya to start the recitation is charged with an introductory pūjā-offering to mark the beginning of the whole cycle. The first ācārya exceptionally sits with his pupils facing West while all recitations which follow feature successive ācāryas facing East and the pupils. The ācārya continues up to the end of the first varga with no help from the śiṣyas. When he successfully completes his part, the śiṣyas take over in a collective way (Mal. caṅkiṭi = “singing together”) and repeat the same thing once (Fig. 1). When they finish, he recites a kuṭuma (Mal. “the lock of hair worn as class distinction”) marker indicating/indexing the end of a varga-unit, or a unit of repetition at the first level of saṃhitā-mode recitation. This again is repeated by the band of students. Next to come is an action-block of so-called pōkkuru, or a repetition of a level unit, in this case one varga by the students, followed again by the “end-marker” (kuṭuma) locating the particular text unit by way of textual coordinates within the corpus of the Ṛksamhitā. It is recited first by the preceptor, then by the students. In that way meta-textual markers get included, and at the same time also remembered, in the bulk of the canonical text corpus of the whole Ṛgvedasaṃhitā. Now, the first round of the saṃhitā-level recitation of a given varga is completed as the first out of three consecutive levels. Next comes the recitation of the same textual portion – the same varga – according to a different mode of recitation, this time the one called pada. This marks the second level of recitation (line 2 in Fig.1). When this is completed, it is followed by a recitation of the same text unit in still another mode, called krama. This, in turn, marks level 3 (line 3 in Fig. 1). While a repetition unit for the first level (saṃhitā) is one varga, it is one anta (a so-called half-stanza) and one ṛk (a full stanza) for level 2 and level 3 respectively. This makes the sequence of alternate recitations a bit more complex in the other two levels. The direction of progress has been represented in Fig. 1 as movement from left to right. The very moment an ācārya starts reciting a second-level, i.e. the pada version of the passage under such a patterned repetition, the śiṣyas start to mirror his words with hand movements of the (hasta)-mudrās – a characteristic method of sound and pitch representation to be seen exclusively among Nampūtiri Ṛgvedins (see image 1).37 In actual practice observed during the 2005 session, whenever any mistake could be detected in the performance of the ācārya or any doubt on his part noticed by the śiṣyas, the latter immediately started to support him with particularly clear 37 The ācārya, however, remained motionless while śiṣyas kept repeating due portions during the Trisandhā session witnessed by me in 2005.


129

display the mudrās proper for the given passage. They demanded repetition of the whole passage if a mistake went on uncorrected. In cases of mistakes considered to be grave ones the whole varga had to be repeated which resulted in additional hours of recitation effort. The pattern of the sequence of Trisandhā shows how cleverly oral tradition handles the problems of efficient memory retrieval through procedures developed to control and confirm the allocation of coordinates for the unique addresses of each and every text unit within the bulk of the community’s memory store. For this to be effected efficient operations need to be reaffirmed for text delimitation, sound assimilation, accent distribution and grammatical analysis necessary for securing a perfect transmission: different modes of recitation entail regular switches between the rules of accentuation and the compound analysis of the pada.38 One of the modes of recitation used regularly during Trisandhā is that of krama. It requires from the reciter that he remain in a state of utmost alert since the elements of a given text passage when repeated are to be presented in a different mode from their first occurrence (this concerns sandhi rules of sound assimilation as well as rules of accentuation). If we take a passage from Ṛksaṃhitā 1.1: agním īḷe puróhitam yajñásya devám ṛtvíjaṃ and ascribe to its six elements a succession of letters “abcdef”, then the three basic (prakṛti) modes of recitation (pāṭha) could be represented as follows: saṃhitā mode “(‘connected speech’ recitation)”: a b c d e f pada mode “(‘word for word’ recitation)”: | a | b | c’ c’’ | d | e | f | krama mode “(‘step by step’ recitation)” can be seen as a combination of the two in a progressive order: | a b | b c | c d | c iti c’ c’’ | d e | e f | Where (|) sign represents suspension of sound assimilation between two adjacent units, and (c’ c’’) stand for two components of the compound represented in saṃhitā mode as (c).39 The particle iti is actually pronounced to mark the operation of resolving the compound (c) into its two components. In fact, in order to get a more comprehensive view of the interrelationship between the basic unit-blocks of Trisandhā proceedings and to better understand what is actually going on during a Trisandhā recitation session, it is not enough to define the modes of recitation and describe the behaviour of the reciters. The most important thing is to realise the recurring pattern behind the sequences of recited passages in terms of the distinct units of text, units of repetition, markers of 38 For more general discussion on memorisation techniques of the Nampūtiri Ṛgvedins, see Galewicz 2009b. 39 The simplified model of representation adopted here for the immediate purpose of the this short study shows only a limited number of otherwise important musical and rhythmical aspects of the actual practice of Nampūtiri style Vedic recitation. For more elaborate models, see Staal 1961: 41–44, or Gray 1959.

130

Cezary Galewicz

beginning and end, modalities of recitation and pattern of concatenation. Fig. 1 represents graphically a reconstructed model of a pattern as witnessed during a Trisandhā session on 18–20 January 2005. It corresponds to the option type referred to in the manual of Trisandhāpaṟipāṭi by the term apāda.40 It is drawn with simple geometrical figures representing the ācārya and the śiṣyas alternating in sequential recitation. The pattern seems to be interiorised within the memories of the reciters who do not need any graphic representation for their agency competence. Most of them expressed a considerable surprise when presented with one drawn by the present author. It seems beyond doubt that to follow this strict pattern is a must and the important sense of correct performance is profoundly associated with a belief in a necessity of respecting its rules. The technical terms, mostly in Malayāḷam, featured in Fig 1 as used by Nampūtiris need, however, a short explanation: varga (a group of stanzas) – a basic text division unit for recitation in saṃhitā mode and a general unit of measure marking the progress of vedapārāyaṇa, i.e. the actual progress on a given day indicated by the number of completed vargas. Each daily recitation starts and ends according to the boundaries of the varga division. The aṣṭaka/adhyāya/varga division system facilitates formal operations on the body of the Vedic text, though these are not the units of sense. antam/anta (functionally same as Sk. ardharca, “half stanza”) – a smaller text-division unit. Most of the stanzas are composed of two antas. It is understood as a half-stanza, though depending on the actual metre of a given stanza, it can be in reality less or more than a half-stanza. If the metre is gāyatrī (3 x 7), first two pādas (quarters) of such a stanza are said to form the first anta, and the remaining third pāda is said to form the second anta. The anta forms a repetition unit for the pada (level 2) recitation. kuṭuma (Sk. cūḍā “tuft of hair”) – the end/limit marker. It refers to the beginning of the next varga if the basis for recitation is one varga (level 1, line 1 in Fig. 1), or to the beginning of the next half-stanza (anta) in level 2 (line 2 in the Fig. 1), or to the beginning of the next stanza (ṛk ) when the unit of repetition is one stanza (line 3 in Fig. 1). pōkkuru (Sk. anuvāda,“repetition”) – indicates a repetition by the pupils of the whole varga in the mode specific to a particular level of recitation.

40 Trisandha. The other two options given by Trisandhāpaṟipāṭi 2 are pṛthaksthāpadaka and pādādya. All three are said to be characteristic of regional variations of Trisandhā reflecting once again the historical rivalry between the two associations of the Nampūtiri Ṛgvedins who tend to mark their distinctive identity by way of alternatives woven into basically the same cultural and religious institutions of their shared tradition.


131

Incidentally, the basic meta-textual marking indices, as well as the sequence, distribution and repetition scheme of the recitation units, appear to follow the very discipline of memorising the Ṛgveda by young brahmacārins that is still to be seen in the vedapāṭhaśālā in Trichur. The same holds true for the general situation of enacting the teaching and memorising of the Veda through ceremonially inscribing it into the memory of the śiṣyas. Thus the scope and number of repetitions follows closely what can be observed in the teaching process at the Vaṭakku Brahmasvam Maṭham of Trichur. Slight alterations to be noticed during Trisandhā might be due to general differences between the Ṛgvedic tradition of the two rival associations (yogam) of the central Kerala Ṛgvedins.41 As already mentioned above, beside the three main modes of recitation put in a characteristic pattern featured in Fig. 1, two other modes occasionally happen to be performed at proper moments during Trisandhā. They are held in utmost respect and proudly displayed as instances of extraordinary skill in mastering recitation techniques. Both are performed by two reciters squatting on the ground and facing each other with their hands pressed against their ears. A peculiar characteristic of these modes of recitation is that they are performed only while other participants are feasting: the recitation called jaṭā (“twisted hair”) occurs during the morning meal, and that of ratha (“chariot”) during the evening meal.42 They are usually preceded by the inarticulate recitation of koṭṭu (Mal. “drumming”)43 executed by alternating participants who hold their nostrils between two fingers while producing a somewhat unusual sound. It happens after the seva to Āyyappan is over and before the meal is served. The other participants execute meticulously the mudrā gestures showing their full control over a given passage. It all appears like no particular passage at all to an outsider since humming inarticulate sound is all that can be heard to an untrained ear. The mode of recitation for koṭṭu is said to be the krama and each subsequent participant who joins in is to recognise the passage recited in this way by his predecessor and accordingly resume the recitation when indicated by him with a finger gesture of alternation. The textual unit for this mode of recitation is one ṛk. Another form of recitation to happen during Trisandhā is that of vāram. This is a kind of testing procedure during which a participant aspiring to the status of a virtuoso reciter is asked to properly recite a passage indicated in a particular way, usually with the help of a set of stones arranged in a peculiar

41 For details concerning the memorisation process at work in Nampūtiri Ṛgveda teaching, see Galewicz 2009b. 42 More on these modes of recitation – see Staal 1961: 45–48, Govindan Namboodiri 2002: 180, Galewicz 2004: 375–376. 43 This mode of (inarticulate) recitation was otherwise explained to me by a Nampūtiri expert as a “control of sound”. It happens to be performed also during other festivals engaging virtuoso Ṛgveda recitation. See Galewicz 2004: 379.

132

Cezary Galewicz

pattern on the ground by a testing party for a tested reciter to decipher.44 But unique to Trisandhā alone are the recitation procedures called mūṭippacca and tripuccha (“triple tail”).45 They are said to be allowed to be performed only on certain calendar days and both represent a complex variety of sequential recitations executed by several reciters to a particular pattern. They are rarely performed today as they represent a difficult art of textual competence and are known only to selected individuals. All the modes of recitation are proudly employed as a display of absolute mastery over the textual body of the Ṛksaṃhitā. So too are the intricate patterns of ritual procedure engaging the operations on text and its different units and modalities. In those times when Trisandhā was an event of truly competitive character (which is not the case any more) these most difficult and demanding modes of recitation would be put as a test to a “visiting party” made of members of the other of the two rival schools of Vedic recitation. They might suddenly appear at the location of the Trisandhā organised by their opponents in order to challenge their claim to superiority in the art of representing the Ṛgveda. Trisandhā is said to come to its close with pronouncing the Ṛksaṃhitā 10.185, i.e. the three gāyatrī verses46 considered to be an auspicious mantra for someone to leave. It is believed to be especially used by the Vedic teacher (ācārya) voicing his benediction to his student (śiṣya) when the latter is about to leave after having completed his Vedic studies. The same triplet is said to guarantee that the student taking his leave will “face no fear” whatsoever.47 In this way not only the foundation experience of completing one’s Veda study gets recreated with the completion of the Trisandhā cycle, but it is also the unique moment of one’s guru benediction which is renewed with the help of the sound of the Ṛgvedic triplet pronounced by the participant who happens to recreate the role of the ācārya as the last one to lead the śiṣyas across the veritable ocean of the Ṛksaṃhitā. Judging by the looks of the participants, one is tempted to admit that a person who has successfully completed a Trisandhā session may indeed expect to experience “no fear” since – as it is believed – his Veda will protect him.48

44 See Galewicz 2004: 377. 45 See Govindan Namboodiri 2002: 178. 46 mahi trīṇām avo ‘stu dyukṣam mitrasyāryamṇaḥ | durādharṣaṃ varuṇasya ||1|| nahi teṣām amā cana nādhvasu vāraṇeṣu | īśe ripur adhaśaṃsaḥ ||2|| yasmai putraso aditeḥ pra jīvase martyāya | jyotir yacchanty ajasraṃ ||3|| 47 śiṣyam ācāryo ‘nena sūktenābhimantrayati (Sāyaṇa on Ṛksaṃhitā 10.185 (Müller 1983 vol. 4: 509)). For the motif of the accomplished student leaving his preceptor to live “with no fear,” see Āśvalāyana Gṛhyasūtra 3.10.7–8. 48 Cf. an interesting passage that crowns the introduction to Sāyaṇa’s bhāṣya on the Ṛksaṃhitā – it warns the students who do not act properly upon the words of the Guru: such students will not be protected by the Vedic knowledge:…tān adhāmāñ ṡiṣyāṃs tacchrutaṃ gurūpadiṣṭaṃ vedavākyaṃ na pālayati (Müller 1983 I: 22).


133

Let me reflect briefly on a possible reason and aim behind this sort of ritual of ceremonial recitation of the whole of the Ṛksaṃhitā. A first remark, already made above, is that Trisandhā while engaging in a total representation of the textual integrity of the Ṛksaṃhitā derives from the idea opposing the Vedic śrauta ritualism which for its part makes a limited use only of selected portions of the Ṛksaṃhitā. The modification procedures applied to Ṛgvedic texts in śrauta rituals are also generally different from those to be seen in the Trisandhā. The ritual recitation of the whole of the Ṛgveda in a specific way seems to represent one of the important forms of marking the identity of the community of the Ṛgvedins among other Nampūtiri Brahmins, as well as the identity of particular groupings (yogam) within the community of the Ṛgvedins. On the other hand, a complex pattern of recitations and different text modifications employed during Trisandhā exemplify perhaps the old conviction that a Vedic text can be made truly efficacious only when put into ritualised practice. In this situation the means of transforming the elements of the textual repertoire in the community’s memory remain within the hands of this community. In this context Trisandhā seems to continue, albeit in a local modification, the trend initiated by the Taittirīya Āraṇyaka. This is at least what seems to be hinted at by Trisandhāpaṟipāṭi 2.49 The Trisandhā ritual cycle is a totalising experience: it is designed to constitute a sacrificial offering of the whole of the collection of the Ṛgvedasaṃhitā in a ritual cycle holding the attention of the community for long months. There is no other rite among Nampūtiris concerned with recitation of the whole of this scriptural corpus in its full potential in one stretch. It is at the same time an efficient means of retrieving, scanning and (re)inscribing the whole of the textual body of the Ṛksaṃhitā into the collective memory of the community for whose identity its preservation in perfect oral shape is of primary concern. Through this ritual procedure the community shows its inner potential for the efficient management of its most important cultural heritage. Thanks to this potential, successive parts of the systematically organised text are first retrieved from memory, then made to undergo a process resembling scanning and refining through a careful and controlled transmission, group repetition, repacking and fixing in the collective memory. This is effected through a systematic following of a concatenated pattern of repetitions, mode

49 svādhyāyādhyayanaśruter avagataṃ vedasya pārāyaṇam tattadvargapadakramānvitam idaṃ trīṇyākhyaṃ etat punaḥ | mukhyaikokti bahu dvijānuvacanāvṛtyā trisandhābhidham pādādyaṃ ca pṛthaksthapādakam apādaṃ ca tridhā tatsmṛtaṃ ||2| “The going over the Veda known from the heard tradition to be the recitation of svādhyāya Following a triple pattern of varga, pada and krama in which Words of a single headman get repeated by many twice-born ones is called Trisandhā It is remembered as triple-form: pādādya, pṛthaksthapādaka and apāda.” (Trisandha).

134

Cezary Galewicz

changing, marking of the textual turning points and borderlines, recreating the rules for sound modification, etc. What then should be imagined as a result of a successfully completed Trisandhā? If we take it to seriously represent a variety of brahmayajña-svādhyāya, then, being one of the “five great sacrifices” (pañcamahāyajña), it offers for its participants a powerful ancient means of requiting the original debt against the ṛṣis while acting as a substitute for all other sacrifices.50 As basically true to the early understanding of the need of paying off the original debt against the ṛsis as authors of the Veda, Trisandhā testifies to the ongoing tradition of understanding this need to be worked out through the conjoined effort of the teacher-pupil pair.51 And the actual enactment of the situation of the Veda transmission from the mouth of the teacher to the ears of the pupils makes the warp of the Trisandhā festival as it has been recorded in written sources as well as how it is still practised among contemporary Ṛgvedins of Nampūtiri community of Kerala. On the other hand, we must keep in mind that all the festival proceedings are structured clearly in such a way so as to hold the performing participants all the time under the scrutinising eye of co-participants (in the original model also their rivals), ready at every moment to correct or (originally) to question the very eligibility of a given reciter to perform at the recitation festivals like that of Trisandhā or Anyōnyam. Moreover, the very act of such questioning itself amounts to a challenge, and this equals a risk of being put to a test since the scheme of the original Trisandhā entails the reversal of the roles: the visiting party will organise their own Trisandhā one day to be visited by the adherents of their rival school. Like many other rituals Trisandhā escapes one exclusive classification in terms of function. A way of representing the inner idea of the ritual manipulation of the text units and modes of recitation proposed here in Fig. 1 is but a humble step towards better understanding of the way some rituals work – at least in the eyes of an outsider looking for an inner logic that might explain away each and every conundrum of the ritual. The graphic figure proved not to appeal to the participants whom I showed it to. In this matter I am reminded of a more general aim of anthropological attitude summarised simply some time ago by Jonathan Smith: “Anthropology has become largely an enterprise of ‘decipherment’, attempting to ‘decode’ an encrypted message from ‘another’ with the firm conviction that, because it is human, it will be intelligible once it is ‘broken.’”52 With the passing of time I am also less and less convinced that I have “broken” a code by articulating a graphic pattern of Trisandhā recitations. The participants of the Trisandhā session witnessed by me a 50 See Malamoud 1977: 191. 51 Cf. Malamoud 1977: 32: “La dette envers le Veda, ou les ṛṣi, est payée par l’effort conjoint du maître et de l’élève dans le processus de l’apprentissage.” 52 Smith 2004: 261. Cf. more general remarks on the topic and criticism of performance theory “with its underlying interpretive-textual paradigm” in Bell 1992: 39–46.


135

few years ago came from all sorts of different walks of life, with teachers, insurance agents, architects and medical doctors performing on the stage together while switching time and again in their roles of teachers and students. For them Trisandhā remained something extraordinary and specifically theirs, opposing any finalising grasp of a drawing hand. In contradistinction to other ritual occasions where Ṛgvedic mantras get recited with some specific aim to fulfil, the least being the consecration of manual acts of offering, during Trisandhā the Ṛgvedic text appears not to be voiced with an aim of communicating anything to be immediately grasped from the words, but in order to be offered in a sacrifice. As such it is an ideal sacrifice paralleling the old concept of “a sacrifice offered by gods with sacrifice as substance” (yajñéna yajñám ayajanta devāḥ (Ṛksaṃhitā 10.90.16c)). It should be stressed again at the end that Trisandhā does not constitute a single and separate case. There are more festivals, during which the community exhibits its dexterity in mastering the recourses of its treasure-trove textual competence and readiness to transmit the Veda to the next generation. There are more traditionally accepted and time-honoured social institutions during which personal and collective textual competence is put to the test and display. The yearly calendar of the Nampūtiris has been dotted with other ritualised occasions53 for expertise in showing the art of Vedic chanting. It is around these events that individuals and groups tend to strengthen their own identity. The case of Trisandhā and similar rituals seems to testify to an ever present disquietude caused by the possibility of the Veda being lost or corrupted and which amounts to the same danger of losing one’s identity. The community inscribing the Veda in its memory deserves a comparison to a “living library”. Notwithstanding rare exceptions, most of the members of this living “resource-pool” of Vedic tradition, so to say, do not burden themselves with the meaning in the form of the semantic value of the stored data (particular texts and procedures). Or, perhaps one should rather say, that the meaning of respective areas of this library situates itself for the insiders on a different level of perception of the whole structure. It has more to do with appreciating the technical perfection of the text-retrieval and cross-reference competence of the reciters and the knowledge of the rules for proper implementation of the texts. This reflects a rather old Indian cultural phenomenon of the sacred text credited with a meaning described as adṛṣṭa – the term being interpreted either as “unseen” and “secret” or “invisible” and “non-existent”. It is ritual use of the sacred text that is to give the opportunity of apprehending its essentially active nature. According to Indian grammarians the vikṛti, or modified Vedic recitation, is the way to express the unseen purpose (adṛṣṭa artha) of the Veda. Per53 Such is, among others, the festival of Anyōnyam. Similar function is fulfilled partially also by the royal festival of Muṟajapam. Nampūtiri Yajurvedins also hold their own festivals.

136

Cezary Galewicz

sonal study (svādhyāya) may constitute a ritual in itself according to authoritative opinions, but still, it is not necessarily focused on the semantic value of the scripture and, in the case of Nampūtiri Ṛgvedins, it is certainly fused with a sense of mastering the “data resource” of their Veda through scholarly patterns of different access to the same resource pool of the scripture. As in the case of other rare collections of precious books of antiquity, from the Middle Ages or modern times, the members and holders of this “living library” proudly display (though not to everyone) their treasures, which in their eyes most aptly reflect, not to say constitute, their own identity and status, being often a reservoir of social integrity and political power. The working of this living resourcepool is still only partially known and awaits further research enhanced by comparative perspective. It is only natural that the inner beauty of the structures and symmetries of the resources of such a “living library” appeals to the mind and aesthetic taste of the connoisseurs who appreciate the exhibition of its patterns, cadences and interrelations.54 The practice of such a display remains a strong enhancement of a collective communal identity negotiating its place and status in a rapidly changing social reality of South India. In this practice the image of a skilful virtuoso reciter of the Ṛgveda, committed to a collective enterprise of reinscribing the treasure text into the collective memory, remains still a powerful but endangered cultural paradigm of identity that can probably be circumscribed with the words taken from a passage by Roy Rappaport, though addressing a different context: “By drawing himself into a posture to which canonical words give meaning the performer incarnates or embodies canonical form. As he participates in the form or order he incorporates it into himself. His body gives substance to the canon as the canon provides his body with form.” 55

54 Jacob 2001: 209 “La Bible, le Canon Confucéen, …, le Veda, le Coran,… et leurs commentaires sont aussi des bibliothèques… les bibliothèques qui reflètent … un désir de beauté et de perfection formelle, de symétrie et d’ordre mathématique, de cohérence et de signification.” 55 Rappaport 1999:153.


Image 1: Nocturnal session of recitation during Trisandhā: the śiṣya repeat pada mode with mudrās. Photo by Cezary Galewicz

137

138

Cezary Galewicz


139

References Bell, Catherine M. 1992. Ritual Theory, Ritual Practice. New York et al.: Oxford University Press. Bhattatiripatu, K.P.C. Anujanan & Viṣṇu Bhaṭṭatirippāṭu (eds.) 1988. Trisandhāpaṟipāṭī of Kunnattūr Paṭiññāṟēṭattu Kṛṣṇan Bhaṭṭatirippāṭṭu with a Vivaraṇa of Kaliyattu Parameśvarabhārati Svāmiyāl. Kunnaṃkuḷam: Pañcāṅgaṃ Press. (=Trisandha) Galewicz, Cesary 2003. “A Keen Eye on Details. Reviving Ritual Perfection in Trichur Somayaga 2003”. Bulletin d’Etudes Indiennes 21: 239–253. — 2004. “Kaṭavallūr Anyōnyam: a Competition in Vedic Chanting?”. In: Arlo Griffith & Jan E.M. Houben (eds.). The Vedas. Texts, Language & Ritual, Groningen: Egbert Forsten. — 2005. “L’Anyonyam. Un ritual de recitation des texts sacrés au Kerala”. Annales. Histoire, Science sociale (EHESS) 2005/3: 551–572. — 2009a. A Commentator in the Service of the Empire: Sāyaṇa and the Royal Project of Commenting on the Whole of the Veda. Vienna: DeNobili, forthcoming. — 2009b. “Things to Remember, things to Forget”. In: Christian Jacob (ed.). Lieux de Savoirs II, Paris: Albin Michel, forthcoming. Govindan Namboodiri, V. 2002. Śrauta Sacrifices in Kerala. Calicut: Publication Division, University of Calicut (Calicut University Sanskrit Series 13).Calicut University Press. Gray, J.E.B. 1959. “An Analysis of Nambudiri Ṛgvedic Recitation and the Nature of the Vedic Accent.” Bulletin of the School of Oriental and African Studies 22/1: 499–530. Jacob, Christian 2001. “Réunir” In: Luce Giard & Christian Jacob (eds.). Des Alexandries I. Du livre au texte. Paris: Bibliothèque nationale de France: 199–211. Mahadevan, T.P. & Frits Staal 2003. “The turning point in a living tradition”. Electronic Journal of Vedic Studies 10/1a. Malamoud, Charles 1977. Le svādhyāya. Recitation personelle du Veda. Taittirīyaāraṇyaka Livre II. Paris: Institut de Civilisation Indienne. Michaels, Axel 2005. “Saṃkalpa: The Beginnings of a Ritual.” In: Jörg Gengnagel & Ute Hüsken & Srilata Raman (eds.). Words and Deeds: Hindu and Buddhist Rituals in South Asia. Wiesbaden: Harrassowitz Verlag: 45–64. Müller, F. Max (ed.) 1983 [1890]. Rig-veda-samhitā. The Sacred Hymns of the Brāhmans, together with the Commentary of Sāyanāchārya, vols I–V. Varanasi: Chowkamba Sanskrit Series Office (Chowkamba Sanskrit Series 99). Pisharoti, K.R. 1928. “Religion and Philosophy in Kerala.” The Indian Historical Quarterly 4: 702–719. Rappaport, Roy 1999. Ritual and Religion in the Making of Humanity. Cambridge: Cambridge University Press. Smith, Jonathan Z. 2004. Relating Religion. Essays in the Study of Religion. Chicago: University Of Chicago Press.

140

Cezary Galewicz

Staal, Frits 1961. Nambudiri Veda Recitation. S’Gravenhage: Mouton & Co. Winternitz, Moritz (ed.) 1902. A Catalogue of South Indian Sanskrit Manuscripts (especially those of the Whish Collections) Belonging to the Royal Asiatic Society of Great Britain and Ireland. With an Appendix by F.W. Thomas. London: Royal Asiatic Society (Asiatic Society Monographs)

Martin Gaenszle

Grammar in Ritual Speech The Use of Binomials in Rai Invocations The issue under discussion is the idea that ritual action is somehow structured like grammatical sentences. But even if the “grammar of ritual” is taken in a metaphoral sense only – and I agree with Michaels (this volume) that this is the only way to pursue the issue –, it leads to the question what should be taken as the major characteristic, or essence, of “grammaticality”. It is generally held that ritual is “more rule-bound”, “more structured”, “more ordered” than ordinary action. This, however, raises the larger issue concerning the nature of action: is it possible to conclude that ordinary, everyday action is “non-grammatical”, i.e. not according to any rules, chaotic, entirely ad hoc, without structure etc.? Most of us would agree that this is not the case. Ordinary action is more spontaneous, less determined, somehow less constrained, but as social scientists have argued since long,1 it is nevertheless structured – sometimes more and sometimes less. Cognitive anthropologists, for example, have analysed the micro-structures of everyday action in terms of “scripts” and “schemas”, which can be quite complex.2 Thus, the question is perhaps not so much whether ritual has grammar-like rules or not but rather: what kind of grammar-like rules distinguish it from ordinary action. In this paper I will not look at ritual action as in some way grammar-bound, but rather focus on ritual speech as indeed grammar-bound. After all, if one looks for features of a grammar in ritual, it certainly can be found – in a literal sense – on its linguistic level. In order to gauge the distinctive grammaticality of ritual in comparison with ordinary action, I will pursue the question: What are the salient grammatical features and relations in ritual speech, and how are these different from ordinary forms of speaking? Thus I am not contrasting ritual action as grammar-bound versus ordinary action as not grammar-bound, as on the level of speech this would be nonsense: all language has grammar. But if we compare the grammar of ritual language with that of ordinary language it becomes clear that the

1 G.H. Mead, T. Parsons, E. Goffman, J. Habermas etc. 2 E.g. Schank & Abelson 1977.

142

Martin Gaenszle

difference is one of kind. Understanding this difference, I argue, can show us something about the particular properties of ritual in general. I discuss these issues in view of my various ethnographic studies on Rai ritual language in east Nepal. Examples will be taken from Mewahang,3 Chintang4 and Puma,5 which are distinct languages belonging to the Kiranti language family of Tibeto-Burman. In particular, I will focus on the phenomenon of binomials which is the most distinctive feature of the ritual language as a special register of speaking. After taking a closer look at the structure and “productivity” of these ritual nouns, I will deal with the apparent and unique high ratio of nouns to verbs and inquire into the significance of this feature. But before venturing into the linguistically constructed cosmology of Rai ancestral religion, I will first discuss some general views of grammar and try to show how certain linguistic ideologies (i.e. ideas what languages are) have affected the debate about the grammaticality of ritual.

1. Grammar in Linguistic Ideologies The view of what grammar is about is apparently shaped by culturally and historically embedded assumptions. An example would be the distinction between prescriptive grammar and descriptive grammar, the former being part of an agenda for standardization, the latter being the object of a science of language. Whereas the wider concept of language ideologies refers to generally held (popular) beliefs about the properties of a language,6 I will speak of linguistic ideologies as referring to academically established views of how a language works.7 Thus, when speaking about the “grammar of ritual” one has to be aware of what concept of grammar is implied. Often what is actually meant by ritual grammar is mainly its “syntax”, and this again is usually taken as dealing with the proper sequencing of elements, the production of a linear order which is seen as generated by formal algebraic rules. This, I would argue, is a rather reductionist view of grammar, and even of syntax. Many of the grammarians of ritual have been 3 Gaenszle 2002. 4 Gaenszle & Bickel et al. 2005, Rai 2007. 5 See Schackow 2008. The project dealing with Chintang and Puma was funded by the Volkswagen Foundation under the DOBES program (Grant No. II/79 092, 2004–2008, PI B. Bickel). For further information on the “Chintang and Puma Documentation Project” see www.uni-leipzig.de/~ff/cpdp. All files and recordings will be deposited at the DOBES archive (www.mpi.nl/dobes). I am grateful to the funding agency and to my colleagues in Leipzig and Kathmandu who made this work possible. I also wish to thank Balthasar Bickel for his stimulating comments on an earlier version of the paper. 6 E.g. Kroskrity 2000. 7 See Joseph & Taylor 1990.

Grammar in Ritual Speech

143

strongly influenced by certain Chomskyan views of language, i.e. by (early versions of) “generative grammar”8 and its model of parsing linguistic phrase structures.9 Often, however, these ideas have been highly simplified and reduced to a mere “sequentialist” view which hypostatises the tree model.10 In recent decades theories of grammar have immensely diversified, and it has become apparent that there is more to syntax than simply the question of sequencing and parsing. Though generally it is agreed that syntax is about the rules “how you can put words together”, it does not only regulate the sequential order of words and phrases, but also their dependency relations. Whereas, for example, an expression like “man the” is grammatically wrong because it violates the proper linear order, an expression like “the eats” is ungrammatical because the verb requires a noun or pronoun, and therefore this phrase violates the dependency relations.11 Observations like these have led linguists to develop a great variety of grammar theories. For example, new versions of dependency grammar theories were developed, i.e. theories in which the structure of a sentence is determined by the verbal phrase (as the “head” of the structure) and its dependent elements.12 Whereas classical phrase-structure grammars focus primarily on word order and the formalisation of parsing rules, other approaches pay more attention to the content of syntactic relations. A further kind of theory deals with construction grammars which question the inborn character of abstract formal rules and stress the process of learning general meaningful patterns. Based on his work on language acquisition Michael Tomasello observes: “As opposed to linguistic rules conceived as algebraic procedures for combining symbols that do not themselves contribute to meaning, linguistic categories and constructions are themselves meaningful linguistic symbols – since they are nothing other than the pattern in which meaningful linguistic symbols are used to communicate.”13

8 In particular Chomsky 1965. 9 See Staal 1979, 1989; Lawson & McCauley 1990. 10 It may have to be stressed that tree models of syntax were used in linguistics long before Chomsky, who criticised simplistic views of grammar in an early and influential article (Chomsky 1956). 11 See Bickel 2008: 1. He adds that it does not matter whether one represents these patterns by trees, by feature matrices or by prose (ibid.). These are only different forms of representation. 12 This is an approach which ultimately goes back to the French linguist Lucien Tesnière (1959), who uses terms like “actants” for nouns and speaks of the “valence” of verbs. 13 Tomasello 2003: 99. In another context he contrasts this with the “lexical rule approach” (derived from generative grammar), where the syntactic structure is seen as determined by the lexical properties of the verb. “The alternative is a construction grammar approach, in which constructions themselves are symbolic units with meaning […]. In this view, much of the creativity of language comes from fitting specific words into linguistic constructions that

144

Martin Gaenszle

Whatever the particular theoretical approach, these developments have highlighted the complex ways through which a verb (predicate or function) is linked with nouns (arguments) in a particular argument structure. For example, accusative languages distinguish subject and object, marking the latter, whereas ergative languages distinguish agents and patients, marking the former. These relationships have also been called grammatical relations or else syntactic functions.14 The crucial point about these more recent approaches, in which semantic roles (like, for example, “agent”, “theme” and “goal” in ditransitive verbs) play an important part, is that the distinction between syntax and semantics is not seen as absolute, or at least as more dynamic: “Thus argument structure is an interface between the semantics and syntax of predicators (which we may take to be verbs in the general case)... Argument structure encodes lexical information about the number of arguments, their syntactic type, and their hierarchical organization necessary for the mapping to syntactic structure.”15 In this perspective special attention is paid to, among other things, the role of word classes, the way they are marked, and their relative “weight”. Some parts of speech need special focusing (marking by affixes), others can be avoided altogether (“pro-drop”). So one of the implications is that it is not only important which argument positions are filled in a sentence, but also which are not filled. For example, one can say in Nepali: “Lāl Bahādur gayo”, but also just “gayo”. Kiranti languages, such as the ones discussed below, have this characteristic in an extensive way, they are what is sometimes called “pro-drop” languages (the phenomenon of dropping pronouns is als known by the term “zero anaphora”). Accordingly, languages can also be distinguished in terms of what has been called referential density, i.e. the ratio of actual arguments (noun phrases) in a clause to potential argument position.16 Basically this concept measures the density

are non-prototypical for that word on a specific occasion of use […].” (Tomasello 2003: 160f.). 14 “Traditionally, the term ‘grammatical relation’ (GR) refers to the morphosyntactic properties that relate an argument to a clause, as, for example, its subject or its object. Alternative terms are ‘syntactic function’ or ‘syntactic role’, and they highlight the fact that GRs are defined by the way in which arguments are integrated syntactically into a clause, i.e. by functioning as subject, object etc. […] Research over the past three decades has greatly expanded the range of syntactic properties that identify GRs in particular languages, and one of the most important results of this research is that properties often do not converge on a single set of GRs in a language.” (Bickel in press). 15 Bresnan 2001: 304. 16 See Bickel 2003. There is some debate about the proper definition of referential density in linguistic typology. Noonan, e.g., proposes the following one: “The percentage of overtly


145

or explicitness of referential expressions in a language: typically English or German, have a high degree of referential density, whereas “pro-drop” languages such as Kyirong or Puma Rai have a low degree. (This can best be measured by using large corpora of utterances). In a broader sense, the term referential density has also been used to express the narrower notion of noun/verb ratio, or “nouniness”/ “verbiness”, of a language. The following discussion deals with the particularly salient role of nouns in Rai ritual speech. I will try to show that these ritual languages in east Nepal are characterized by a particular type of “doubled” nouns (binomials) which are not only a limited stock of lexical items but can also be produced ad hoc, according to certain rules. Ritual discourse therefore is marked by a kind of inflated focus on nouns (in relation to the use of verbs), which intensifies the referential aspect of speech in a unique way, thus distinguishing it clearly from ordinary ways of speaking.

2. Rai Ritual Speech The ritual languages among the Kiranti groups of east Nepal are distinct registers of speech (sometimes also called sub-languages) which are used in formal ritual contexts in order to address elders or ancestral beings. The performance of this speech form is part of a complex oral tradition which is known by the word muddum or cognate forms (such as mukdum, mundhum, or ridum). This concept comprises the whole body of “ancestral knowledge”, which was transmitted through generations of elders, shamans and priests, and forms the basis of the Kiranti groups’ “traditional way of life”.17 In linguistic terms this special register of speech can be seen as a form of etiquette, a particularly polite form of speaking, which in comparison to ordinary ways of speaking is characterized by a unique lexical inventory and, more generally, a higher degree of formality. Ritual speech genres can be distinguished according to the performative context: most names of rituals simply state the act of the performance (like “Calling the House Deity”, or “Offering to the Hearth Stones”). More general distinctions are made by taking into account the level of performative competence. For example among the Mewahang rituals are grouped into those which can be performed by any knowledgeble person, those which can be done by elders, and those which require initiatory experience.18 Each ritual speech genre (like ceremonial dialogues, lay invocations, sacrificial offerings etc.) has its own terminology, idiomatics and a distinct performative style. But in all expressed arguments to possible [i.e. notional] arguments is referred to as ‘referential density’” (Noonan 2003: 1). 17 Gaenszle 2002: Ch. 1. 18 Gaenszle 2002: 85ff.

146

Martin Gaenszle

cases they are marked by the use of formulae, poetic figures and imagery, prosodic rhythm and structural redundancies. 2.1 Binomials Linguistically the most striking feature is the pervasive parallelism which is found on various levels of ritual discourse: e.g. on the level of syntax, prosody, as well as morphology. As has been pointed out by Roman Jakobson, parallelism is at the base of all poetic language since it can be viewed as the extension of the binary principle of opposition to various levels of expression.19 For the regular or “compulsory” pairing of successive verbal sequences Jakobson coined the term “canonical parallelism”, which is now commonly used as a descriptive category in linguistic anthropology.20 Though usually pairs predominate, there can also be more than two parallel elements which make up a parallelist structure. On the syntactic level we encounter what has been termed “grammatical parallelism”: the parallel construction of clauses, usually through a repetition of verb forms which are linked with a noun phrase. The following are two examples from the Mewahang corpus.21 1. ikka-mi yakhiwa ne: maʔa, risiwa ne: maʔa “We don’t know how to speak, we don’t know how to chant” 2. harimiʔa harilawa hendanumye casummiʔa casumlawa hendanumye “Provide the farmer with soul of fields, provide the planter with soul of rice” Thus, a common feature is the use of a small number of verbs, which recur frequently linked with a variety of nouns. In particular, they are linked with the unique form of nouns, which are themselves an instance of parallelism on the level of the word, namely the noun pairs or “binomials” (which consist of two “limbs”). In the following example the noun pairs rita bhāta, (N “customs and manners”) and sobhāu sila:ka (N/M “polite speech and slokas”) are split up and combined with the verb separately.22 3. rita ne: le:knum, bhāta ne: le:knu(m)-maŋ-le you know the customs, you know the manners, 19 20 21 22

Jakobson 1979: 265. See Fox 1977, 1988. Gaenszle 2002: 47, 207, 222. I will use the following abbreviations for the languages: N “Nepali”, M “Mewahang”, C “Chintang”, P “Puma”, B “Bantawa”. The subsequent asteriks denotes the ritual language register (e.g. M*).


147

kodoŋe:loŋ sobhāu mu:kni, sila:ka mu:kni and in this way you are speaking politely, you are reciting the slokas. As indicated above, one of the most distinctive features of Rai ritual language is the frequent use of such binomials, so let us have a close look at these. The term “binomial” has been suggested by Nick Allen in his programmatic article on Rai ritual language,23 and others (including myself) have taken it up in Himalayan studies.24 Other possible terms which have been in use are “noun pair”, “twin word” (German: Zwillingswörter), or “linked noun”. Not everything occuring in ritual does have such a designation (i.e. “ritual name” or : dopsi niŋ, M., lit. “chanting name”). There are also ordinary nouns which are used in ritual contexts, such as haŋ (“king”,“chief”) or saya (“headsoul”, “life-force”). In other words, the use of such terms is not ruled out, or “ungrammatical”, but if a ritual term does exist it will be mandatory. And, second, there are ritual nouns which consist of only one word (e.g. M* bedachu, “ginger”, ord M* riyamdu, “body”). However, especially those “referents” in ritual discourse which are associated with ancestral origins and the “sacredness” of tradition, do have a binomial designation. Typical Mewahang examples are khapcuni buyoni, “wife” or thakcakhi: be:cakhi: “weaving”. There is a great variety of such binomials, which cannot be discussed in detail here. But roughly, one may distinguish the following types: – Ideal Type: an „irrversible binomial“25 e.g. M* situluŋ thuŋmaluŋ (M. “hearth”) This is the most common type: it usually consists of two three-syllabic limbs, the last syllable of each being identical and carrying the semantic weight. In the example given here, referential meaning is contained in the syllable -luŋ which is the proto-Tibeto-Burman root *luŋ “stone”, “rock”. In a traditional Rai house the hearth consists of three stones on which the cooking vessels are placed. These three stones are also considered to be seats of the ancestors and are given offerings on certain occasions. – “Irregular” (i.e. deviating from this ideal type) trinomials, e.g. poŋwama pocema sariŋwa reversable, e.g. saŋgemma ruwemma o. ruwemma saŋgemma Though most ritual terms are binary, there are cases of more than two nouns being linked. This is particularly the case when the focal referent contained in the final syllable of the noun is specified through an attributive complement, e.g. C* 23 Allen 1978. 24 Also see Strickland 1987. 25 Malkiel 1959.

148

Martin Gaenszle

nawaŋgi-sewa toŋtaiʔma-sewa lasima-sewa (“offering of first fruits, of the descending times, of the concluding rites”). In rarer cases, the otherwise irreversible sequence can be reversed so that the nouns appear in a non-canonical order. All this shows that there is a certain compositional openness and freedom. This is further corroborated by the existence of – Symmetrical compounding (“free binomials”) M* siroŋla bedachu, < siroŋla, “offered rice” + bedachu, “ginger” Complementary elements (which make up a whole) can often be combined in various orders. As noted above, bedachu, “ginger”, can stand as a single noun, but here it is used as the limb of a somewhat “loose” binomial. That other combinations are possible is shown by the occurrence of the (similar) term arawa bedachu, “rice grains and ginger”, where a different term for rice is used. More generally, one can see the binomials as the result of word formation strategies: Through various kinds of compounding binomials are “generated”. As this appears to be a rule-bound process, it is clear that new binomials can easily be created. Empirical data which include Nepali lexemes show that this has indeed been the case. The following types of compounding strategies can be distinguished. The position of the construction’s dominant semantic centre (“head”) is an important criterion. – endocentric (with semantic head included), e.g. – M* cikhimtaŋma puwaŋtaŋma “deity of the house” (< taŋma “head”). – M* bāyupu jāgipu “death spirits” (here the semantic weight is on the two Nepali terms bāyu and jāgitra, kinds of death spirits; -pu seems to derive from an older classifier but it is unclear of what specific meaning). – P* sencihoŋ senmuŋhoŋ “dream king” (< hoŋ “king”, “chief”) – exocentric (without head, main referent is not included), e.g. – M* kharukhom casumkhom “harvest” (< casum “rice” + khomma “to gather”) – P* dumdhit mandhit “affliction” (< dum “word” + dhitma “to flatter”; this refers to the persuasive talk by the spirits) In these cases the referent of the term is allusively described by certain characteristic actions. – copulative or coordinative, e.g. – P* apturoŋ puwaroŋ “bow and arrow” – C* poluwa pakuwa “roasted and boiled”

149


Often complementary elements are simply added up into pairs (see also the examples of symmetrical compounding above. – appositional (the limbs give different descriptions of the referent), e.g. P* wasepkha wamakha “place of origin”, (< -kha “place”, lit. “where the waters flow and disappear” Thus there is ample room for a creative production of terms, and this shows that one has to be careful with general statements about the “archaic” nature of the ritual language. Nevertheless, it is indeed striking that the ritual terminology has a lot of similarity across the different Rai languages. This is most evident in the role of proto-Tibeto-Burman roots already mentioned above (see the Tibeto-Burman root *luŋ). Further examples are given in the table below. Table 1 Mewahang

Chintang

Puma

situluŋ thuŋmaluŋ “hearth stones”

suktuluŋ micaluŋ “hearth stones”

bobiluŋ khaitaluŋ “hearth stones”

bhεgimi lamlumi “priestly assistant”

maŋbopmi taŋmaʔmi “protective being”

lesami suntummi “knowledgeable being”

carima “earth”

heŋkhama makhama “earth”

heŋkhama muwama “earth”

casumri cheberi goduri laŋsuri “plenty of rice & wealth” “godavari flower”

nawari suntumri “ancestral tradition”

Partly this “family resemblance” may have to do with an inherent conservative nature of ritual speech, which often is found is such idioms. But it is also likely that it is linked to the frequent use of the ritual language in multilingual contexts. As members of different subtribes often intermarry it is not rare that priests chant in households where more than one Rai languge is spoken. Thus it is perhaps more the common grammar of compounding than an identity of terms which enable the speakers to understand neighbouring priests. 2.2 Noun-to-verb Ratio In any case, the effect of this doubling of the nouns is an inflation of the “nouniness” of the ritual speech form in relation to its “verbiness”. In fact, this is a clear inversion of the situation in ordinary speech. Rai languages, like Chinese, are

150

Martin Gaenszle

known to have a lower noun-to-verb ratio than English, for example.26 Experiments examining the rhetorical style of ordinary narratives brought out a relatively low “information density” (see discussion below). Thus it is striking, that in the ritual idiom just the opposite can be observed. This ratio of nouns to verbs can be measured when one has a large corpus of speech. Preliminary calculations of the Chintang corpus in Leipzig point to a clear distinction (Bickel personal communication). When measuring the number of nouns in relation to verbs the ratio in a sample of fifteen ritual language recordings was roughly between 2.8 and 3.9. In contrast to this, the value in a comparable number of ordinary language recordings was between 0.6 and 1.1. A more extensive analysis is still pending, but the tendency in this first result is unambiguous. Now, why is there such an emphasis on nouns? Of course, a purely structural approach would be satisfied to say: the inversion of the ordinary creates a space for the extra-ordinary, the “sacredness” of speech. In fact, it would also be possible that a ritual language does just the opposite and makes an equally heavy use of verbs, likewise distinguishing it from ordinary speech. (It might be interesting to see whether such cases are linked to a tendency of a language’s inherent referential density). Thus, one certainly cannot generalize this observation as a common characteristic of all ritual languages. But what is the effect of a pronounced “nouniness” in discursive style? As Sanskritists (who are represented by a majority in this volume) well know, the predominant use of nouns, what has long ago been termed “Nominalstil” by Hermann Jacobi,27 is characteristic of a particular formal and abstract style of writing. In Jacobi’s paper, „Über den nominalen Stil des wissenschaftlichen Sanskrit“, this was Sanskrit scientific language. Though ritual speech is certainly not academic, it is also abstract in the sense that it refers not so much to descriptive but to ideal entities. However, this is not the whole story. The shift in emphasis towards nouns, I would argue, is also a shift towards the focus on sacred objects. As we have seen, not all entities in ritual, are equally marked by binomials. It is interesting that in ritual language there are only few verbs which occur exclusively in these contexts, such as verbs for ritual speaking, ritual presentation etc. (e.g. do:pma, M* “to name”, yameʔma “to present”). Most verbs used in ritual are the same as in ordinary speech, and also their grammatical forms, agreement behaviour etc. is exactly the same. In a few cases, ritual verbs are specially marked by a prefix to denote the ritual context (e.g. yoŋ- in M* yoŋyepma < yepma “to stand”, or walu- in M* waluha:ma < ha:ma “to divide”). This appears to suggest that only some of the ritual activities are qualitatively distinguished form ordinary action. 26 See Stoll & Bickel et al. 27 See Jacobi 1903.


151

In order to understand the inflationary use of ritual nouns, one has to look at the syntactic postions in which they are employed. Which, then, are the prominent syntactic roles of binomial expressions? What is the particular focus expressed by the terms marked in this way? 2.2.1 Person In most cases a ritual addressee has a ritual name and is thus designated by it. All the divine forces, ancestral deities, the hearth etc., have such ritual names. Only in the case of recently deceased relatives the ordinary name is used, but this only occurs in lay invocations. In the following example (Chintang), the addressee is the hearth, which is seen as an ancestral link between the earth and the sky.28 4. peimaluŋ taŋma na alise sumciluŋ peimaluŋ taŋma na a- lis -e sumciluŋ three.hearth.stones three hearth stoneshead PTCL 2- be -PST C* C* C* C C- C -C You have become the lord of the three heavenly stones. heŋkhamluŋ makhamluŋ alise heŋkhamluŋ makhamluŋ a- lis -e earth earth 2- be -PST C* C* C- C -C You have become the stone of the earth.29 Otherwise reference to the persons participating in the ritual is rare (see below). The major focus is on the divine beings. 2.2.2 Context Typically a ritual address to the divinity is followed by references to the particular time and occasion of the ritual performance.

28 In the following glosses abbreviations according to the Leipzig Glossing Rules are used (see www.eva.mpg.de/lingua/resources/glossing-rules.php): EMPH “emphasis”, ERG “ergative”, FOC “focus”, GEN “genitive”, INF “infinitive”, P “patient”, PST “past”, PTCL “participle”, SEQ “sequential”, SIM “simultaneous”, TOP “topic”, 2 “second person”, 3 “third person”, p “plural”. 29 Rai 2007: 145, my revised translation.

152

Martin Gaenszle

5. asi ko labayu dochenu laniʔa doniʔa asi ko lani -ʔa labay -u doni -ʔa dochen -u today EMPH moon -ERG touch -3P moon -ERG touch -3P M M M* -M M -M M* -M M -M 30 Today the moon is right, the time is ripe. Frequently, as in the Chintang triple expression cited above (nawaŋgi-sewa toŋtaiʔma-sewa lasima-sewa “offering of first fruits, of the descending times, of the concluding rites”), the text refers to the season of the year, either the “ascending time” (N ubhauli) or “descending time” (N udhauli). But most prominently, binomials refer to the ritual offerings presented to the ancestors: 2.2.3 The Object of the Offering 6. poluwa pakuwa aca athuŋ poluwa pakuwa a- ci a- thuŋ sacred.fried.meat sacred.cooked 2- eat 2- drink C* C* C- C C C You have eaten the roasted and boiled meat. phurmani sewa akhattu mani sewa a- khatt -u maniparma phurmani acceptance acceptance service 2- take -3P C* C* C/B C- C -C May you accept and take this worship.31

kiŋa kina SEQ C

kha kha FOC C

Similar forms of presentation can be found in abundance in the Mewahang corpus.32 Generally, it is striking that the speaker rarely refers to himself directly, i.e. first person marking is avoided. To a certain degree such avoidance also includes other living speech act particpants who are never referred to by their ordinary individual names. If they are referred to this is only done by using the ritual name of the proto-clan (the samet-group). More generally, in the Kiranti cultural context any direct reference to the person who is in a pleading position tends to be avoided, 30 Gaenszle 2002: 147. 31 Rai 2007: 127. 32 See Gaenszle 2002: 149.

153


as it is considered impolite. Likewise a direct request is wrapped in more indirect words. Though imperative forms can be and are used, they are felt to be too bald. In Chintang subjunctive forms (“may you”) are frequently used instead of imperatives.33 One has the impression that even verbs are turned into words which appear like nouns. There is a clear preference for non-finite forms, such as infinitives. Consider the following Mewahang example.34 7.

chelam hemma chelam hem -ma path open -INF M* M -M Open up the path!

delam delam path M

hemmai hem -ma open -INF M* -M

-i -EMPH -M

khaŋma nu:to yemma nu:to yem -ma nu: khaŋ -ma nu: -to see -INF be.nice -SIM hear -INF be.nice M -M M -M M -M M Look upon us benevolently, hear my words benevolently!

-to -SIM -M

Such a use of simple infinitives (with the suffix -ma) is a common way to express imperatives in an depersonalized manner. In the second line, such infinitives are accompanied by an adverbial expression consisting of a verb (nu:ma, “to be good”) plus simultaneous marker (-to). This latter marker is also often used to express an activity without finite forms. 8. chelam hento chelam hem -to path open -SIM M* M -M Opening the path

delam delam path M*

hento hem -to open -SIM M -M

Such a use of non-finite verb forms, I argue, tends to assimilate the verbal expressions to nouns. In fact it blurs reference to the agent (who is it who opens the path?) and only focuses on the action as such. By thus deemphasizing the agent in 33 Gaenszle & Bickel et al. 2005. 34 Gaenszle 2002: 274ff.

154

Martin Gaenszle

favour of the action the verb refers to a general and more abstract idea beyond the contingencies of the present. This tendency to avoid finite verb forms and employ mechanisms of "quasinominalization" contributes to the fact that syntactical structures in the non-lay texts are often rather simple. The rhythm of the chants, which is usually characterized by the repetition of equal and relatively small prosodic units, creates the impression as if the text consisted of a long enumeration of terms with equal weight. Thus the emphasis on nouns has the effect of constructing a distinct ontology, a world in which reference to ideal entities is prominent, whereas reference to individuals is downplayed. The heavy use of nouns gives a particular density to this world.

3. Density We have seen that binomials are the most prominent feature of ritual speech and rather than being only “dead” lexical baggage from an ancient past, they can be seen as a lexical resource which which may be used to generate new meanings. Sentences in ritual speech require such binomials in certain positions, thus inflating the proportion of nouns in relation to verbs in rituals discourse as a whole. This emphasis on nominal phrases can be seen as a basic characteristic of the argument structure in ritual syntax. But this syntactic constraint has also bearings on the semantic level of ritual texts, it also affects its meanings. What then are the consequences of this high noun ratio for an understanding of ritual language? Apparently this shift of emphasis in ritual speech implies a shift in the ontology of the ritual world. In our case it gives more referential weight to ritual objects and entities, in other words, it induces a focus on ritual “things” which exist on a different plane of reality. Binomials like situluŋ thuŋmaluŋ (“hearth”) refer to both the everyday hearth which is used for cooking as well as the “seat of the ancestors” consisting of a trinity of stones which are associated with narratives of the ancestral past. Thus binomials increase the referential depth of the ritual words. But they not only refer to a pre-existing reality, rather they contribute to creating it. At this point, I find it useful to take inspiration from literary studies of fiction. I am aware that novels and religious cosmologies have a different ontological status for the people, and it is not my aim to make any judgement on the truth of such discourses. All I want to do is consider ritual texts as contributing to “world-making” in similar ways. (As anthropologists we are used to proceed like phenomenologists and bracket the issue of truth values). Thomas Pavel in his book called “fictional worlds” argues for making use of a “possible-world-philosophy” which acknowledges the possibilitiy of “secondary


155

universes” which are structured similarly to (though differently from) the primary universe. Basically all literary texts similarly construct such fictional worlds through linguistic means and thereby they give different emphases. “Not unlike natural languages, in which a given syntactic category, say the subject slot in a sentence like ‘— sneezed’, can be filled with segments of various lengths, from single words (he, she, John) to interminable noun phrases including numerous embedded clauses, textual manifestations are subject to expansion and contraction.”35 Forms of discourse thus differ not only in terms of their different ontologies but also in more formal terms according to their textual compactness, or what he calls referential density: “the relationship between world dimension and text dimension determines what might be called the referential density of a text”.36 The term referential density has also been employed in linguistics, though the meaning is not quite the same.37 Basically what is at stake is the “information density” a text displays. Taking up ideas by Marshal McLuhan linguists have distinguished between relatively “hot” discourse and relatively “cold” discourse: the former (like e.g. movies) appeals to the senses and is more detailed and elaborate, the latter (like e.g. writing) is more terse, and requires more active involvement by the recipient. Giving a more complex nominal designation to an entity makes it also more formal, it endows it with more detailed and deeper meaning, dignity, and weight. In fact, in should be added, the details may not always be clearly understandable but rather remain enigmatic. But this only enhances the depth of this kind of language.38 A discursive world populated with such over-coded and often enigmatic beings is therefore denser, more authoritative, and – perhaps – more powerful. As Maskarinec observes discussing shamanic chants and mantras in Western Nepal, such speech forms demonstrate “the primacy of language over a world in which the more names something has, the more fixed in place, the more real, the easier to order and manipulate it becomes.”39 This kind of ontological shift in the ritual worlds, which we found constructed in ritual speech, can also be found in ritual action in general, where ritual objects, ritual situations, or else ritual activities, are similarly “expanded and contracted”, sometimes “heated”, sometimes “cooled”. So, perhaps at least in this sense there is 35 36 37 38 39

Pavel 1986: 94. Pavel 1986: 101. See Bickel 2003. Porzig 1925. I am grateful to Jan Houben for this reference. Maskarinec 1995: 180f.

156

Martin Gaenszle

some similarity between the grammar of ritual speech and the grammar of ritual action.


157

References Allen, Nicholas J. 1978. “Sewala Puja Bintila Puja: Notes on Thulung Ritual Language.” Kailash 6/4: 237–256. Bickel, Balthasar 2003. “Referential Density in Discourse and Syntactic Typology”. Language 79: 708–736. — 2008. “Aspects of Kiranti Syntax: Grammatical Relations” [Paper presented at the Central Dept. of Linguistics Tribhuvan University, Kirtipur, August 14, 2008, Handout]. www.uni-leipzig.de/~bickel/research/presentations/kirantiGR_ktm2008. pdf (23 October 2009). — in press. “Grammatical Relations Typology”. In: Jae Jung Song (ed.). The Oxford Handbook of Language Typology. Oxford: Oxford University Press. Bresnan, Joan 2001. Lexical-Functional Syntax. Oxford: Blackwell. Chomsky, Noam 1956. “Three Models for the Description of Language.” I.R.E. Transactions of Information theory 2: 113–124. Chomsky, Noam 1965. Aspects of the Theory of Syntax. Cambridge: The MIT Press. Fox, James J. 1977. “Roman Jakobson and the Comparative Study of Parallelism”. In: Daniel Armstrong et al. (eds.). Roman Jakobson: Echoes of His Scholarship. Lisse: Peter de Ridder Press: 59–90. — (ed.) 1988. To Speak in Pairs: Essays on the Ritual Languages of Eastern Indonesia. Cambridge: Cambridge University Press. Gaenszle, Martin 2000. Origins and Migrations: Kinship, Mythology and Ethnic Identity among the Mewahang Rai of East Nepal. Kathmandu: Mandala Book Point & The Mountain Institute. — 2002. Ancestral Voices: Oral Ritual Texts and Their Social Contexts among the Mewahang Rai in East Nepal. Münster, Hamburg, London: LIT Verlag. — & Balthasar Bickel et al. 2005. “Worshiping the King God: A Preliminary Analysis of Chintang Ritual Language in the Invocation of Rajdeo”. In: Yogendra Yadav et al. (eds.). Contemporary Issues in Nepalese Linguistics. Kathmandu: Linguistic Society of Nepal: 33–47. Institut für Linguistik, Universität Leipzig (n.d.). “The Chintang and Puma Documentation Project (CPDP)”. http://www.uni-leipzig.de/~ff/cpdp/index.html (23 October 2009). Jacobi, Hermann 1903. „Über den nominalen Stil des wissenschaftlichen Sanskrits“. Indogermanische Forschungen 14: 236–251. Jakobson, Roman 1979. Poetik: Ausgewählte Aufsätze (1921–1971). Frankfurt: Suhrkamp. Joseph, John Earl & Talbot T. Taylor (eds.) 1990. Ideologies of Language. London: Routledge. Kroskrity, Paul V. (ed.) 2000. Regimes of Language: Ideologies, Polities & Identities. Santa Fe: School of American Research Press. Lawson, E. Thomas & Robert McCauley 1990. Rethinking Religion: Connecting Cognition and Culture. Cambridge: Cambridge University Press.

158

Martin Gaenszle

Malkiel, Yakov 1959. “Studies in Irreversible Binomials”. Lingua 8: 113–160. Maskarinec, Gregory G. 1995. The Rulings of the Night: An Ethnography of Nepalese Shaman Oral Texts. Madison: The University of Wisconsin Press. McLuhan, Marshal 1964. Understanding Media: The Extensions of Man. New York: New American Library. Michaels, Axel 2010. “The Grammar of Rituals”. This volume: 7–28 Noonan, Michael 2003. “A Crosslinguistic Investigation of Referential Density.” Association for Linguistic Typology at the University of Cagliari. (Handout). Pavel, Thomas G. 1986. Fictional Worlds. Cambridge, London: Harvard University Press. Porzig, Walter 1925. „Die Rätsel im Rigveda. Ein Beitrag zum Kapitel ‚Sondersprachen‘“. In: Eduard Sivers. Germanica: Eduard Sievers Zum 75. Geburtstage. Halle an der Saale: Niemeyer: 646–660. Rai, Ichchha Purna 2007. “An Outline of the Chintang Mundum /Ritual Language”. Unpubl. M.A. Thesis. Kathmandu: Tribhuvan University. Schackow, Diana 2008. “Clause Linkages in Puma (Kiranti)”. Unpubl. M.A. Thesis. University of Leipzig. Schank, Roger & Robert Abelson 1977. Scripts, Goals, and Understanding: An Inquiry into Human Knowledge Structures. New York: Halsted. Staal, Frits 1979. “The Meaninglessness of Ritual”. Numen 26: 2–22. — 1989. Rituals without Meaning: Ritual, Mantras and the Human Sciences. New York etc.: Peter Lang. Stoll, Sabine & Balthasar Bickel & Elena Lieven (in press). Nouns and Verbs in Chintang: Children’s Usage and Surrounding Adult Speech. Leipzig (Ms.). Strickland, Simon S. 1987. “Notes on the Language of the Gurung Pe”. Journal of the Royal Asiatic Society 1: 53–76. Tesnière, Lucien 1959. Éléments de Syntaxe Structurale. Paris: Klicksieck. Tomasello, Michael 2003. Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge (Mass.), London: Harvard University Press.

Johannes Bronkhorst

Ritual, Holophrastic Utterances, and the Symbolic Mind 1. Introduction The evolutionary proximity of human beings to other primates is sufficiently established for studies of “the naked ape” or “the third chimpanzee” to be possible and justified.1 Nevertheless, many of those working in the human sciences insist that there are fundamental differences between humans and the other primates. Something happened in our evolutionary past which led to the rise of a number of features that we do not share with our evolutionary cousins. The most notable of these are language, ritual, religion and culture. What happened that had these momentous consequences? The following pages will propose an answer. Given the nature of the question, it is appropriate to begin with primates different from us.

2. Symbolic Representation In research carried out during the seventies at Georgia State University, two chimpanzees – called Sherman and Austin – participated in an experiment meant to investigate whether chimpanzees could be taught to use language, and to what extent they would do so.2 Since the vocal tract of chimpanzees does not allow them to pronounce the sounds of human language, a different method was employed: the apes were taught to use a special computer keyboard made up of so-called lexigrams. The more specific aim of the experiment was to find out whether they would be able to use lexigrams in combinations. These combinations might then be looked upon as elementary syntactical relationships. The chimps were trained to chain lexigram pairs in simple verb-noun relationships. For example, the lexigram GIVE followed by the lexigram BANANA would result in the giving of a banana. GIVE followed by ORANGE would be similarly successful, but other combinations would not. The following scheme, where « means “successful”, and ´ means “unsuccessful”, illustrates this:

1 These are the titles of two popular books on human behaviour by Desmond Morris (1967) and Jared Diamond (1992) respectively. 2 See Savage-Rumbaugh & Rumbaugh 1978: esp. 279ff.

160

Johannes Bronkhorst

GIVE + BANANA → « GIVE + ORANGE → « BANANA + ORANGE → ´ ORANGE + GIVE → ´ etc. Initially our chimps would have just two “verb” lexigrams and four food or drink lexigrams to choose from. But the mere learning of a number of lexigram pairs was not sufficient for the chimps to understand the general system (viz. “verb” followed by “noun” leads to the appropriate result). They could learn individual associations between specific pairs and their result, but for a long time they failed to grasp the system of relationships of which these correlations were a part. In the end they did grasp the system, but at the cost of thousands of trials in which they had to find out what combinations of lexigrams led to no result whatsoever. Once they had grasped the system, they had crossed the symbolic threshold. They had succeeded in using “words” as symbols in the special sense used here: as signs that do not only refer to “their” objects, but also to each other. For a correct appreciation of what follows it is essential to understand that the word symbol is not used here in its usual sense and that it should on no account be confused with it. I borrow the word in this specific sense – i.e. signs that do not only refer to “their” objects, but also to each other – from Terrence Deacon, who in turn borrowed it from the philosopher Charles Sanders Peirce. Symbols of this kind play a role in what we will call symbolic reference. Even the highly condensed presentation of long and elaborate experiments given above shows how much is required to acquire symbolic reference. To cite Deacon: “What the animals had learned was not only a set of specific associations between lexigrams and objects or events. They had also learned a set of logical relationships between the lexigrams, relationships of exclusion and inclusion. More importantly, these lexigram-lexigram relationships formed a complete system in which each allowable or forbidden co-occurrence of lexigrams in the same string (and therefore each allowable or forbidden substitution of one lexigram for another) was defined. They had discovered that the relationship that a lexigram has to an object is a function of the relationship it has to other lexigrams, not just a function of the correlated appearance of both lexigram and object. This is the essence of a symbolic relationship.”3

3 Deacon 1997: 86.

Ritual, Holophrastic Utterances, and the Symbolic Mind

161

What Sherman and Austin learned with enormous and prolonged exertion we humans learn in our childhood, apparently with much less effort.4 Where the two chimpanzees took a long time to learn to pay attention, not just to the desired object of their activity, but also to the signs and to their relationships with each other, we learn to do so almost automatically. Indeed, all normal human children learn the language of their care-takers. Yet the complexities of those languages are much greater than those of the simple sets of “words”, two in each set, which Sherman and Austin managed to master. All human children cross the symbolic threshold at an early age, without realising it. They are able do so because evolution has equipped us for this task, primarily by giving us a much enlarged prefrontal cortex.5 Let us stay somewhat longer with our two chimpanzees. Initially, they had learned the combination GIVE + BANANA as an indivisible whole, which gave rise to a pleasant result, viz. that a banana was given to them. After thousands of trials and errors, they had mastered the system behind it. They had learned, for example, that the lexigram GIVE could be used with a following ORANGE so as to lead to a different result, viz. that an orange would be given to them. Meanwhile they had learned that these two outcomes had something in common, and that this common element was somehow represented by the lexigram GIVE. Similarly, other combinations had taught them that the lexigram BANANA was associated with the element banana in various activities. In other words, Sherman and Austin were in the process of creating representations corresponding to elements of objective reality. These representations were the result of the overlap of events: the representation “give” resulted from the overlap of “give banana” and “give orange”. Human beings appear to arrive at their representations in a similar, though much more complex manner. These few reflections show that the learning of language facilitates the formation of representations. These representations correspond initially to the shared parts of different linguistic utterances, i.e. primarily words. However, these same reflections show that this process, once begun, does not necessarily stop here. We can easily imagine a situation where our chimpanzees wish to receive an object for which there is no lexigram on their keyboard, say a piece of chocolate. In that case they might create a symbolic representation, chocolate, which they would know how to deal with syntactically if only there were a lexigram CHOCOLATE, for example in GIVE + CHOCOLATE. In other words, once the learning of language has initiated the capacity of creating symbolic representations, the animal may be in a position to create new symbolic representations for which there are no words. 4 Only in extremely “unfavorable” external circumstances, such as those experienced by socalled feral children, may human infants not succeed in acquiring language. On feral children, see Candland 1993; Newton 2002; Strivay 2006. 5 See Deacon 1997: 145–318 (Part Two: Brain).

162

Johannes Bronkhorst

This accounts for the human capacity not only to create new words, but also to have symbolic representations for which there are no words.6 Consider in this connection another ape, a bonobo called Kanzi. Kanzi, for reasons that will not be discussed at present, acquired language to a degree far superior to Sherman and Austin. Strikingly, his linguistic skills subsequently facilitated other kinds of learning, even in domains that were not directly associated with language. Indeed, “his understanding encompassed all manner of novel events and even of metaphor. His understanding of language informed his interpretation of real world events and his broadened capacity to interpret and appropriately classify real world events informed his linguistic comprehension in a boot strapping effect. An example of this was the ease with which Kanzi learned to flake stone tools given a modicum of both visual and verbal instruction. Similar attempts by other apes required long and arduous conditioning and shaping regiments.”7 Crossing the symbolic threshold, it appears, involves more than being able to learn to use language. It opens up a world of representations which, by the boot strapping effect mentioned in this passage, extends well beyond the representations covered by the words of one’s language.8 It is tempting to connect these reflections with what neurobiologists tell us about human consciousness. Consider the following passages from a recent book by Gerald M. Edelman, a neuroscientist: “By its very nature, the conscious process embeds representation in a degenerate,9 context-dependent web: there are many ways in which individual neural circuits, synaptic populations, varying environmental signals, and previous history can lead to the same meaning. [...] There is no single circuit activity or code that corresponds to a given conscious ‘representation’. A neuron may contribute to that ‘representation’ at 6 Symbolic representations must be distinguished from concepts. An animal of prey may have a concept of the animals that it hunts, in the sense that it will know and recognise them. Without symbolic representation, however, it cannot think about those animals the way those endowed with symbolic representation can. 7 Savage-Rumbaugh & Fields & Taglialatela 2000: 916. 8 It is for this reason hardly surprising that Kanzi is reported to have made, all on his own, four new “words”, standing for “banana”, “juice”, “grapes” and “yes” (New Scientist, 2 January 2003). 9 On p. 43 Edelman explains: “Degeneracy is the ability of structurally different elements of a system to perform the same function or yield the same output.”


163

one moment, and in the next have no contribution to make. The same is true of context-dependent interactions with the environment. A shift of context can change the qualia that are parts of a representation, or even recompose some qualia and still keep that representation. [...] [D]epending on input, environment, body state, and other contexts, different core states can underlie a particular representation. The interactions are relational and have the properties of polymorphous sets. These are sets, like Ludwig Wittgenstein’s ‘games’, that are defined neither by singly necessary nor by jointly sufficient conditions.”10 Elsewhere Edelman draws attention to the hippocampus, a neural structure in the brain which is “necessary for episodic memory, the long-time memory of sequential events, the brain’s ‘narrative’”. He comments: “Higher-order consciousness rests in part on episodic memory, and in the absence of such memory coherent semantic activity would not be likely to develop.”11 It seems probable then that the overlap of episodic memories and other mental events plays an important role in the creation of representations in the mind, just as for Sherman and Austin the intersection of the useful units of communication GIVE + BANANA and GIVE + ORANGE led to the representation give. Symbolic reference greatly increases the number of representations thus created. It seems that Sherman and Austin had not analysed the representation give out of the numerous situations themselves in which they had been given various things. Only the exercises with the lexigrams taught them to do so. We may assume that the same happens to human children when learning to speak. Symbolic reference makes, in this way, an almost limitless multiplication of representations possible, as well as their combination in countless ways. Indeed, worlds of imagination can now be created that take us away from our immediate impulses and experience. The results are multiple. Symbolic representation, to begin with, permits us to think, and speculate, about our own past and future and, what is more, to think about ourselves as characters in numerous scenarios. In this way, we can think about ourselves the way we think about others. This is what Deacon refers to in the following passage: “Consciousness of self in this way implicitly includes consciousness of other selves, and other consciousnesses can only be represented through the virtual reference created by symbols. The self that is the source of one’s experience of intentionality, the self that is judged by itself as well as by others for its moral choices, the self that worries about its impending departure from the 10 Edelman 2004: 105f. 11 Ibid.: 99.

164

Johannes Bronkhorst world, this self is a symbolic self. It is a final irony that it is the virtual, not actual, reference that symbols provide, which gives rise to this experience of self. This most undeniably real experience is a virtual reality.”12

Symbolic representation influences our experience.13 We will return to this below. It also allows us to be objective with regard to ourselves, which in turn is behind our tendency to judge ourselves the way we judge others (at least to some extent). This in turn allows for empathy on a wider scale than might otherwise be possible. Without symbolic representation there would be no detachment from immediate arousal and compulsion, no possibility to judge oneself the way we judge others, no place for moral choices, no developed forms of empathy, no ordinary sense of self, and much else. Symbolic representation does not stop at self-representation. The immediate arousal and compulsion mentioned earlier, as well, are objectified and find a place in the “outside world” in the form of values and institutions. The result is as described by Roy A. Rappaport, who uses the word symbol approximately in the same way as we do:14 “The epochal significance of the symbol for the world beyond the species in which it appeared did not become apparent for many millennia – perhaps hundreds of millennia – after it had emerged. But earlier effects of language and even proto-language upon the lifeways of the hominids in its possession must soon have become enormous. [...] [L]anguage permits thought and communication to escape from the solid actualities of here and now to discover other realms, for instance, those of the possible, the plausible, the desirable, [...] [However,] [l]anguage does not merely permit such thought but both requires it and makes it inevitable. Humanity is a species that lives and can only live in terms of meanings it itself must invent. These meanings and understandings not only reflect or approximate an independently existing world but participate in its very construction. The worlds in which humans live are not fully constituted by tectonic, meteorological and organic processes. They are not only made of rocks and trees and oceans, but are also constructed out of symbolically conceived and performatively established 12 Deacon 1997: 452. This self is what Dennett (1991: 418) calls the narrative self; see below. 13 Van Driem, forthcoming: “The idea that language exerts an unfavorable effect on perception itself and blinds us to reality is an old idea already espoused by Bertus Brouwer and Frederik van Eeden.” 14 Cp. Rappaport 1999: 4: “only humans, so far as we know, are possessed of languages composed, first, of lexicons made up of symbols in Peirce’s sense of the word […] or Buchler’s […]: that is, signs related only ‘by law’, i.e. convention, to that which they signify, and second, of grammars, sets of rules for combining symbols into semantically unbounded discourse.”


165

[...] cosmologies, institutions, rules, and values. With language the world comes to be furnished with qualities like good and evil [...]”15 Once again, this description concerns the world experienced with the help of symbolic representation. The world experienced without it, if such a thing is possible, is without these features. The effects of symbolic representation go even further than this. Objective reality is in part social reality. This is what John R. Searle set out to show in his book The Construction of Social Reality (1995). And indeed, it cannot be denied that there is “an objective world of money, property, marriage, governments, elections, football games, cocktail parties and law courts in a world that consists entirely of physical particles in fields of force, and in which some of these particles are organized into systems that are conscious biological beasts, such as ourselves”.16 In other words, “there are portions of the real world, objective facts in the world, that are only facts by human agreement”.17 This social reality, which is real, is yet language-dependent (chapter 3). In institutional reality, language is not used merely to describe the facts but, in an odd way, is partly constitutive of the facts.18 And being language-dependent, it depends on symbolic representation. Largely as a result of symbolic representation, human thought is, to at least some extent, narrative in nature.19 To cite the philosopher Daniel Dennett: “Our human environment contains not just food and shelter, enemies to fight or flee, and conspecifics with whom to mate, but words, words, words. These words are potent elements of our environment that we readily incorporate, ingesting and extruding them, weaving them like spiderwebs into self-protective strings of narrative. Indeed, [...] when we let in these words [...] they tend to take over, creating us out of the raw materials they find in our brains. Our fundamental tactic of self-protection, self-control, and self-definition is not spinning webs or building dams, but telling stories, and more particularly concocting and controlling the story we tell others – and ourselves – about who we are.”20 One of the characteristics of narrative is its so-called ‘chunking’ of experience. As a matter of fact, “[i]t is easier to organize knowledge and behavior if the vast realms of experience are subdivided; indeed, the world would quickly become 15 16 17 18 19 20

Rappaport 1999: 8. Searle 1995: xi–xii. Ibid.: 1. Searle 1999: 115. See, e.g. Turner 1996. Dennett 1991: 417–418.

166

Johannes Bronkhorst

unmanageable if I had to sort through every possible concept and potential course of action at every given moment.”21 To avoid the threatening chaos, the narrative, i.e. symbolic mind sifts through the data of perception and apportions different parts to different narratives. Depending on the “stories” in which “I” figure, certain objects will receive extra attention, others will be neglected. This sifting process is, once again, at least in part the result of symbolic representation. We may sum up what precedes by once more citing Deacon: “Because of our symbolic abilities, we humans have access to a novel higher-order representation system that not only recodes experiences and guides the formation of skills and habits, but also provides a means of representing features of a world that no other creature experiences, the world of the abstract. We do not just live our lives in the physical world and our immediate social group, but also in a world of rules of conduct, beliefs about our histories, and hopes and fears about imagined futures.”22 Here two points have to be emphasised. First, symbolic representation does not only affect the way we think, or the way we communicate, it also affects our cognition and the way we experience the world. Second, the way we experience the world with the help of symbolic representation – in short, symbolic experience – is based on, and cannot exist without, non-symbolic experience. In other words, we have two cognitive styles, one of which (the symbolic one) cannot exist without the other, but not vice versa. Let us look at these two points: 1. Symbolic representation affects the way we experience the world. As symbolic beings we live in a constructed world, which contains many things that the real objective world, which is “outside” and independent of us, does not contain, or which it only contains by human agreement. Among these constructed things, as we have seen, we must count our objectified self (“self”, “soul”), our objectified urges (“values”, “morality”), objects to which a function has been attributed (bank-notes where there are only pieces of paper), and much else. 2. Symbolic experience is rooted in non-symbolic experience and would become seriously dysfunctional without it. Without non-symbolic experience we might lose contact with reality altogether, being locked, without possibility of escape, into a world of imagination that symbolic representation creates for us.23 The young child has pure non-symbolic experience, which allows it to subsequently “cross the symbolic threshold” and make the quantum jump into the world of

21 Herman 2003: 172. 22 Deacon 1997: 423. 23 According to certain sleep researchers, this is what happens in dreams. See Llinás & Paré 1991; Jouvet 1999: 106f.; Jeff Warren 2007: 137f.


167

symbolic experience. But even after this jump it needs non-symbolic experience to anchor the world it constructs into objective fact. The conclusion we are led to draw is that we, normal human adults, experience the world in a double manner: a constructed world of symbolic representation is added on to a world of “raw” experience which underlies and accompanies it. What would non-symbolic experience be like, if it could free itself from symbolic representation? Consider first the following passage from Searle’s book: “From a God’s-eye view, from outside the world, all the features of the world would be intrinsic, including intrinsic relational features such as the feature that people in our culture regard such and such objects as screwdrivers. God could not see screwdrivers, cars, bathtubs, etc., because intrinsically speaking there are no such things. Rather, God would see us treating certain objects as screwdrivers, cars, bathtubs, etc. But from our standpoint, the standpoint of beings who are not gods but are inside the world that includes us as active agents, we need to distinguish those true statements we make that attribute features to the world that exist quite independently of any attitude or stance we take, and those statements that attribute features that exist only relative to our interests, attitudes, stances, purposes, etc.”24 We may not be gods, but our experience of the world without symbolic representation would be close to the one attributed to God in this passage: we would see no screwdrivers, cars, bathtubs, etc., but only the objects that people who do use symbolic representation treat as screwdrivers, cars, bathtubs, etc. On the basis of our earlier reflections, we may add further features. If we could free ourselves from symbolic representation, we would have no “objective” notion of self, we would inhabit a world without values, and our expectations of the future and many of our memories of the past would not affect our present experience. And finally, we would not filter out many of the features of the world that have been neglected because they do not fit into any of our present narratives. Once again it will be interesting to cite the observations of a neurologist. Antonio Damasio deals with the question of consciousness in his book The Feeling of What Happens (1999). He distinguishes between two kinds of consciousness, which he calls core consciousness and extended consciousness. Core consciousness, he argues, is of a non verbal nature. Extended consciousness is based on core consciousness, and cannot exist without it. The reverse is not true: core consciousness can exist without extended consciousness, and this is indeed what happens in certain neurological disorders. Extended consciousness, in the words of Damasio, “goes beyond the here and now of core consciousness, both backward and forward. The here and now is still there, but it is flanked by the past, as much past as you 24 Searle 1995: 12.

168

Johannes Bronkhorst

may need to illuminate the now effectively, and, just as importantly, it is flanked by the anticipated future. The scope of extended consciousness, at its zenith, may span the entire life of an individual, from the cradle to the future, and it can place the world beside it. On any given day, if only you let it fly, extended consciousness can make you a character in an epic novel, and, if only you use it well, it can open wide the doors of creation”.25 About the experience of core consciousness he says: “In a neurologically normal state, we are never completely deprived of extended consciousness. Yet it is not difficult to imagine what a possessor of only core consciousness probably experiences. Just consider what it may be like inside the mind of a one-year-old infant. I suspect objects come to the mind’s stage, are attributed to a core self, and exit as quickly as they enter. Each object is known by a simple self and is clear on its own, but there is no large-scale relation among objects in space or time and no sensible connection between the object and either past or anticipated experience.”26 For our present purposes it is important to remember that extended consciousness cannot exist without core consciousness, so that core consciousness is present in some way in every normal conscious human being. To this may be added that extended consciousness, whatever its neurological basis, is largely shaped by symbolic representation. Damasio’s extended consciousness may therefore be considered as coinciding for a large part with the consciousness that results from symbolic representation. Damasio’s observations confirm in this manner that symbolic experience is based on, and cannot do without, non-symbolic experience. Edelman, whom we met before, distinguishes between what he calls primary consciousness and higher-order consciousness. Even though he thinks that “it is likely that [such primates as chimpanzees] have a form of higher-order consciousness”, he recognises that the acquisition of language makes a major difference. About this he says: “Clearly, one of the largest steps toward the acquisition of true language is the realization that an arbitrary token – a gesture or a word – stands for a thing or an event. When a sufficiently large lexicon of such tokens is subsequently accumulated, higher-order consciousness can greatly expand in range. Associations can be made by metaphor, and with ongoing activity, early metaphor can be transformed into more precise categories or intrapersonal and interpersonal experience. The gift of narrative and an expanded sense of temporal succession then follow. While the remembered present is in fact a reflection of true physical time, higher-order consciousness makes it possible to relate a socially constructed self to past recollections and future 25 Damasio 1999: 195–196. 26 Ibid.: 202.


169

imaginations. The Heraclitean illusion of a point in the present moving from the past into the future is constructed by these means. This illusion, mixed with the sense of a narrative and metaphorical ability, elevates higher-order consciousness to new heights.”27 Note that more is needed than just a “sufficiently large lexicon” of arbitrary tokens that stand for things or events. These tokens – as has become clear from the experiments involving Sherman and Austin – should be recognised and employed as symbols in the sense used here, i.e. as referring not just to “their” things or events, but also to each other, and as constituting a system of logical relationships with each other. In view of what has gone before we can use the image of a web woven by symbolic representation between us and the objects of our perception, a web that separates us from the outside world. This is of course only an image: there is no real web, and there is no real separation from the outside world. In reality, incoming signals are interpreted in the light of the numerous associations which they evoke. In the case of those who master symbolic reference, these associations are richer and far more intricate than for those who don’t.

3. Mysticism In spite of its obvious insufficiencies, I will continue to use the image of a web woven by symbolic representation that situates itself between us and the outside world. Is it possible to push the web aside and experience the world in a more direct manner? Are we condemned to remain separated from the objects of our perception by this artificial construction that has interposed itself between us and the world? Some of the testimonies of people variously referred to as mystics, madmen and meditators suggest that the web sometimes tears, or is torn, whether on purpose or by accident.28 Since William James, mystical experience has often been characterised as being ineffable and as possessing a noetic quality.29 Literal-minded philosophers find the

27 Edelman 2004: 103. 28 Cf. Pyysiäinen 1993: 36: “mystical experiences may count as an exception to the linguistic quality of man’s being-in-the-world”; Staal 1990: 139: “mysticism is characterized by the absence of language. It points to a pre-linguistic state which can be induced by ritual, by recitation, by silent meditation on mantras, or by other means”. On the link of epileptic seizures with “deeply moving spiritual experiences, including a feeling of divine presence and the sense [of being] in direct communion with God”, see Ramachandran & Blakeslee 1998: 179f. 29 Wulff 2000: 400.

170

Johannes Bronkhorst

idea that it is ineffable, and therefore beyond the realm of language, puzzling.30 Less cavilling readers may find this a particularly appropriate manner of describing experience that is no longer co-determined by symbolic representation, and therefore by language. Mystical and related experience is also said to give access to a different, higher reality, or to allow its subject to perceive things as they really are (this is James’s noetic quality). Indeed, a universal effect of mystical experience is said to be “an understanding that what was experienced was more real/important than any prior experience.”31 The second of these two claims corresponds to what we would expect to hear from someone who has succeeded in discarding symbolic representation, if only for a short while and perhaps only in part. Such a person may be expected to experience the world differently, and in a more direct and immediate fashion. For ordinary human experience is always separated from its objects by the web of symbolic representation. Seen this way, the claim that mystical experience gives access to a different and higher reality is, though not strictly true, almost correct. Experience through symbolic representation is indirect, mediate. Experience without it, or with less of it, is direct, immediate. Strictly speaking it is the same “objective” reality that is experienced, but it is experienced so differently that the experiencer may have difficulty recognising it. Indeed, recognition itself involves connecting the present cognition with earlier ones. In non-symbolic experience the link with constructed former and anticipated later experiences is weakened, or absent. (We will see below that all associative links with the past may be interrupted during mystical experience.) The subject is, as a result, easily convinced that he or she is confronted with a different reality, where in actual fact it is the same reality that is experienced differently. Let me add that, if our mystic has to answer the question which of the two realities he or she has experienced is more real, the answer can easily be predicted. The mystical experience is so much more direct and immediate, and so undoubtedly real, that the choice is obvious: the reality experienced in the mystical state is more real than the one of ordinary experience. Other features commonly ascribed to mystical experience are easily explained by the weakening or disappearance of the web of symbolic representation. The ordinary self, for example, will tend to disappear. This does not surprise, given the fact that the ordinary self was a construction of symbolic representation to begin with. The unitary and undifferentiated nature of mystical consciousness emphasised by Stace (1960) and others is another feature that fits in well with the absence of symbolic representation. Let us recall that the symbolic cognitive style divides the world into representations. Without symbolic representation, reality presents itself as undivided, and therefore unitary. 30 See Gellman 2005: § 3.1. 31 Paper 2004: 4.


171

It seems safe to conclude from all this that there are ways to free experience, if only temporarily and perhaps partially, from the web woven by symbolic reference.32 This extraordinary experience will subsequently be interpreted in ways proper to the culture of the person concerned. Such interpretations are, of course, of the greatest significance both to those who have these experiences and to those who study them on the basis of witness reports. However, they do not interest us in the context of the present investigation. We note in passing that this non-symbolic experience (the “mystical” experience) is highly valued and repeatedly sought after by many of those who have had it. There is a notable tendency to ascribe deep significance to it. We will have more to say about this below. Having established that there appear to be ways of discarding the web of symbolic representation, we may give some thought to the question of how this result is obtained. The aim is not to enumerate techniques and methods that are de facto used by mystics and others who succeed in having mystical experiences. Rather, we wish to consider what mechanism we would expect, in the light of our reflections so far, to lead to that result. In order to make progress, we have to abandon the image of a web woven by symbolic reference, and try to understand in a more factual way what happens when a person crosses the symbolic threshold. The case of Sherman and Austin will once again be helpful. Where these two chimpanzees had initially learned the connection between the pair of lexigrams GIVE + BANANA and the giving of a banana that followed, they had come to associate, after extensive supplementary learning, the single lexigram GIVE with a variety of situations and elements, among them other lexigrams. These numerous associations – with other lexigrams, with the order of lexigrams to be respected, with situations in which different things were given – allowed them to use this lexigram in a simple system of symbolic reference. Essentially the same applies to human language users. A vast number of associations allows them to use words the way they do. Many, perhaps most, of these associations are not conscious. But they are there, below the surface.33 Without them, language use would not be possible. What should one do to remove the web woven by symbolic representation? The obvious answer is: Stop the extra associations that constitute symbolic reference.

32 It will be argued below that culture, too, weaves a web of symbolic reference that is added on to, or integrated with, the “linguistic” web. 33 Cp. Deacon 1997: 265–266: “The symbolic reference emerges from a pattern of virtual links between [neurological] tokens, which constitute a sort of parallel realm of associations to those that link those tokens to real sensorimotor experiences and possibilities.”

172

Johannes Bronkhorst

How does one stop innumerable associations, most of which are not even conscious? Focusing the mind presents itself as a plausible method. Focusing the mind is a faculty which humans and many animals possess. Situations that are particularly threatening, to take an example, make us forget almost everything else.34 This faculty is to some extent subject to ordinary conscious control. We then speak of concentration. It can take more extreme forms, in which case the expression absorption is to be preferred.35 Absorption typically accompanies mystical experiences. Absorption, we must assume, reduces the number and perhaps the intensity of associations, including unconscious associations. What remains is an experience in which the associations that are responsible for symbolic representations have been reduced or suppressed: a non-symbolic, mystical experience.36 This explanation of mystical experience has an additional advantage. If focusing the mind can reduce the number and/or intensity of associations, there is no reason to insist that in its strongest forms it affects only those associations that participate in symbolic representation. No symbolic representation is required to associate present with past experience: animals that have not crossed the symbolic threshold can yet learn from past experience, and therefore interpret the present in the light of the past. Mystical experiences that are completely timeless, and therefore without any associative connection with the past, can be understood as resulting from a state of absorption in which all associations with the past, including those that do not contribute to symbolic representation (if such exist), are interrupted. In other words, focusing the mind can serve as an explanation for mystical experiences that are free from symbolic representation, but also for aspects of such experiences that cannot be fully understood by the mere absence of symbolic representation.

34 The filtering of irrelevant information is preceded by activity in the prefrontal cortex and basal ganglia, particularly in the globus pallidus; cf. McNab & Klingberg 2007. 35 “Absorption is the tendency to alter our perceptions and surroundings while in a state of highly focused attention with complete immersion in a central experience at the expense of contextual orientation” (Maldonado & Spiegel 1998: 59). 36 Compare these reflections with the following passage from Pyysiäinen’s book Belief and Beyond (1996): “If we now define the so-called external mystical experience [...] as an experience with sensory input but with no thoughts, it corresponds rather well to [the] description of the receptive mode in which [...] logical thinking and categorization are reduced to a minimum [...] [S]uch experience would mean that syntax and the linguistic aspects of consciousness [...] as well as off-line thinking are turned off. Moreover, also consciousness of one’s own conscious nature [...], on which (together with language) the experience of self is based, supposedly disappears. What remains, is [...] on-line thinking, and an awareness of one’s sensations. The external world does not disappear from one’s consciousness, but is experienced as a here and now continuity with which oneself is coextensive.”


173

The explanation of mystical experience by means of absorption has a further consequence. There can be degrees of absorption, with, as upper limit, the total interruption of all associations. This upper limit is not necessarily reached by all those who have mystical or semi-mystical experiences. It follows that there is a sliding scale connecting “ordinary” and “mystical” experience. A slight reduction of the number and intensity of associative mental connections may be part of everyone’s daily experience, yet may not strike those who undergo it as particularly remarkable: they may not even take notice.

4. Ritual Ritual is often accompanied by formulas, strings of speech sounds. In India the term used is mantra, and I shall occasionally speak of mantras when referring to these kinds of formulas. Frits Staal, in several publications (e.g. 1984; 1985; 2006), has drawn attention to the fact that mantras, when used as mantras, do not express meaning, at least not in the ordinary way. He does not deny that many, though not all, mantras have a form that resembles or is identical with linguistic utterances. He emphasises, however, that mantras, even those that consist of language, are not used as language. Formulas used in ritual settings are not analysed.37 Indeed, often they cannot be analysed. They may be in a language that is no longer used and that is unknown to those who hear or even pronounce the formulas. This is as true of the Latin of the Catholic mass38 as it is of the Vedic Sanskrit used in Brahmanical ritual. Sometimes the formulas are in no known language. This applies to many of the mantras used in Tantric forms of Hinduism and Buddhism. A more familiar example may be abracadabra. Whether or not this formula has a historical etymology, it is and has been unanalysable and “meaningless” for most if not all of its users. Why should ritual formulas be unintelligible, “meaningless”? This question may find its answer in the observation that ritual formulas are, or perhaps have to be, holophrastic. Ordinary holistic phrases – i.e. the formulas we use in daily life – 37 Cp. Rappaport 1999: 151: “It is virtually definitive of ritual speech that it is stereotyped and stylized, composed of specified sequences of words that are often archaic, is repeated under particular, usually well-established circumstances, and great stress is often laid upon its precise enunciation. As Maurice Bloch [...] has emphasized, in contrast to ordinary discourse in which considerable choice is open to speakers at a number of points in any utterance, in ritual formulae the ‘features of juncture’, those components of speech indicating relations among the referents, are immutable. In M. C. Bateson’s [...] terms, ritual utterances are ‘fused’. This is to say that meaning is derived from them as unsegmented wholes, or as wholes only segmented into minimal meaningful units of considerable length, usually much greater length than is the case in ordinary speech.” 38 For an ethnographic description of the Latin mass, see Jucker 2007.

174

Johannes Bronkhorst

do have functions. They may have a literal meaning as well, but this literal meaning plays a reduced role in the formula when used as formula. Sometimes the literal meaning may even be misleading. The phrase “how do you do?” is not a question about the way the person addressed does something. But even though the constituent words are misleading, the phrase is not without function, as anyone who refuses to use it in appropriate circumstances will soon discover. Ritual formulas, then, are like ordinary formulas, but more so. Ritual formulas often do not even pretend to have a literal meaning. Indeed, it seems to be a plus for ritual formulas to be unintelligible. Let us therefore forget about their literal meanings. What remains? Let us recall that ritual formulas, being holistic utterances like formulaic phrases in ordinary language, can still have a function. Ordinary holistic utterances, as recent research has shown, often “seem to be geared towards manipulating the situation in which the speaker finds him- or herself”.39 Nothing prevents us from assuming that ritual formulas do the same. And indeed, it would be easy to provide illustrations of the frequently manipulative intent of the use of ritual formulas. In India, “mantras are understood by the tradition as polyvalent instruments of power”;40 “their function [...] is a direct action, generally a ritual one, or a psychological or mystical one”.41 It seems natural to conclude that ritual formulas do have a function, but unlike ordinary formulas they no longer pretend to have a linguistic meaning as well. Ritual formulas, seen this way, are the perfect holistic utterances. Compared to ritual formulas, formulaic expressions in ordinary life are holistic only in the sense that we do not bother to analyse them. Ritual formulas are different: they are un-analysable. Like ordinary formulaic expressions, they have a function, often a manipulative one. Unlike ordinary formulaic expressions, they perform this function directly, without the intervention of the elements that are designated by the words that occur in ordinary formulas. Let us recall now the experiments in which the two chimpanzees Sherman and Austin learned, with great effort, to master symbolic reference. They started with what we might call holistic phrases like GIVE + BANANA. Once they had crossed the symbolic threshold, they could use the constituents of these phrases independently. They could, for example, use the lexigram BANANA in a new context created by themselves. Our ordinary language use is of that type, though infinitely more complex. We can use words in a virtually limitless number of contexts of our own choosing. However, ritual utterances are different. They are holistic. They correspond to the combined lexigrams GIVE + BANANA before Sherman and Austin had learned to segment this sequence. This suggests that people who opt for the use of holistic phrases – which is what we all do in specific, “ritual” circum39 Wray 2002b: 87. 40 Alper 1989a: 6. 41 Padoux 1989: 302.


175

stances – attempt, by so doing, to step back out of the realm of symbolic representation into the immediacy of non-symbolic experience. To avoid confusion, let me add that not all formulaic expressions are, for that reason, ritual formulas. Quite on the contrary, it seems clear that formulaic expressions are also used in ordinary language, in play, and perhaps elsewhere. The question why certain formulaic expressions are accepted as ritual formulas whereas others are not, is interesting and deserves attention in its own right. It cannot be dealt with in this study. Let us turn to ritual action. Is it possible to maintain that ritual activity relates to ordinary activity in the same way as ritual utterances relate to linguistic utterances?42 Let us recall that ritual utterances are holistic, unlike most utterances used in ordinary language. What is more, normal linguistic utterances exemplify symbolic reference. Is it possible to say the same, or approximately the same, about ritual and ordinary activity? Is it true that ritual activity is holistic in some sense whereas ordinary activity is guided by symbolic representation? I suggest that this last question can be answered with a double yes. Yes, ritual activity is holistic, and yes, ordinary activity is guided by symbolic representation. The holistic nature of ritual activity can be established with relative ease.43 Scholars have often observed that ritual actions are divorced from their usual goals. Some emphasise that they have no meaning, or that they are not communicative.44 This does not mean that rituals never have a specific overall purpose; they often do. 42 This was Rappaport’s (1999: 151) position: “As far as form is concerned, ritual formulae are to ordinary language as ritual postures and gestures are to ordinary instrumental activity.” 43 Cp. Rappaport 1999: 253: “to perform a ritual is not to analyze it. Indeed, the import of performance is exactly the converse of that of an analytic operation.” The correctness of this observation is not necessarily affected by the fact that many rituals consist of a concatenation of “elements of ritual” (Ritualelemente; Michaels 2007: 242). See Michaels 2007: 246: “Das Problem aller Morphologie des Rituals ist: Wenn Handlungseinheiten als kleinste Bausteine des Rituals genommen werden, dann entspricht dies nicht den Morphemen einer Sprache, sondern allenfalls den Sätzen [...]”. Similarly Lawson & McCauley 1990: 84 (cited Michaels 2007: 246): “Returning, then, to the analogy with linguistics, it is the action that is the analogue of the sentence (which is the fundamental unit of linguistic analysis).” It is of particular interest, and very intriguing, that some people resort to semantic etymologising when trying to explain the elements of ritual: “[A]n important part of the Ndembu explanation of symbols rests upon folk etymologizing. The meaning of a given symbol is often, though by no means invariably, derived by Ndembu from the name assigned to it, the sense of which is traced from some primary word, or etymon, often a verb. Scholars have shown that in other Bandu societies this is often a process of fictitious etymologizing, dependent on similarity of sound rather than upon derivation from a common source. Nevertheless, for the people themselves it constitutes part of the ‘explanation’ of a ritual symbol” (Turner 1969: 11). For more on semantic etymologising, see below, Appendix 1. 44 Staal 1990; Humphrey & Laidlaw 1994.

176

Johannes Bronkhorst

But “the set of sequences that compose the ritual are not connected to this goal in the same way as sub-actions connect to sub-goals in ordinary behavior”.45 Let us recall in this context what was said above about the frequently manipulative intent of the use of ritual formulas. Both ritual acts and ritual formulas can be used to reach some end, but both seek to do so in a holistic manner, forgoing the functions which their constituent actions and words (if there are any) would normally accomplish.46 What about ordinary activity? Must we accept that it is guided by symbolic representation? Yes, and the distinction between two cognitive styles introduced above explains why. Our dominant cognitive style is the one determined by symbolic representation.47 It is the cognitive style that allows us to plan our lives and carry out projects of some complexity. It is this cognitive style that gave our ancestors, and us, an edge over competitors. It is the cognitive style on which our well-being, even our life, depends. Ritual activities, seen this way, are, like ritual formulas, a denial of symbolic representation. Like ritual formulas, they are an attempt to step back out of the realm of symbolic representation into the immediacy of non-symbolic experience. Neither ritual activities nor ritual formulas are symbolic in the sense used here. Both concern a “different”, “higher” reality, a reality not touched by symbolic representation. Rituals and ritual formulas have their role to play, not in our ordinary world, but rather in the “higher” reality that is concealed by symbolic representation. Does this mean that ritual activities are carried out, and ritual formulas uttered, with the intention of reaching mystical states? Scholars have pointed out that altered states of consciousness do often accompany ritual.48 It is also known that ritual formulas are used in a number of traditions to evoke altered states of consciousness: the use of mantras in Indian yoga and the so-called dhikr in Islam come 45 Boyer & Liénard 2006: 3. 46 Cp. Rappaport 1999: 390: “It may be suggested [...] that ritual recaptures a state having its ontogenetic origin in the relationship of pre-verbal infants to their mothers. If this is the case the ground of the numinous precedes the development of any awareness of the sacred or the sanctified for, being discursive, that awareness can come only with language.” 47 Cp. Deacon 1997: 257: “In general terms, human information processing should be biased by an excessive reliance on and guidance by the kinds of manipulation that prefrontal circuits impose upon the information they process. We humans should therefore exhibit a ‘cognitive style’ that sets us apart from other species – a pattern of organizing perceptions, actions, and learning that is peculiarly ‘front-heavy’, so to speak.” 48 Cf. Rappaport 1999: 227: “ritually altered consciousness is widespread if not, indeed, culturally universal.” Staal 1990: 139: “mysticism [...] points to a pre-linguistic state which can be induced by ritual [...]” (cited above). According to Michaels (2006: 261), rituals “often create an auratic sphere or arena of timelessness and immortality”. See further Goodman 1988: 34f.


177

to mind. Nor is it surprising that these activities or recitations might bring about such results. We have seen that mental absorption can be a means to reach such states, perhaps the only one. Ritual activities and utterances may be conducive to mental absorption. The fact that the difference between ordinary and mystical experience is not abrupt, so that people can have “weak” or “partial” mystical states, confirms the claim that altered states of consciousness may accompany many if not all rites. In spite of this, I hesitate to look upon the search for altered states of consciousness as the only explanation for ritual in all its manifestations. I rather assume that most if not all human beings, including those who are unfamiliar with altered states of consciousness, are implicitly aware of the fact that the world created by symbolic representation is not the only world there is. This assumption gains in credibility if we recall that even “ordinary”, i.e. symbolic experience is not always to the same extent accompanied by associative mental connections: concentration and absorption reduce them, if ever so little. Even the person least susceptible to mystical experience is likely to undergo the resulting fluctuations in the thickness of the web woven by symbolic reference. In a way we are obliged to make this assumption of an implicit awareness. The use of holistic utterances and holistic activities in ritual in widely separate cultures indicates that there must be such an awareness of the process that leads to ordinary, symbolic cognition, viz. through the division of an originally undivided cognition. Without the assumption of implicit awareness we might have to postulate that ritual utterances and ritual activities have been discovered by chance, presumably because these utterances and these activities gave rise to altered states of consciousness. This postulate does not, as it seems to me, deserve serious consideration. We therefore hold on to the view that implicit awareness of the two cognitive styles common to all normal human beings is behind ritual in its various manifestations. Why should anyone – indeed, almost everyone – wish to act in or on the reality which is believed to be hidden below the surface of our ordinary experience? One reason is no doubt that virtually everyone, and not only mystics, is convinced that such a different reality exists, and that it is more real than ordinary reality. We have already seen that this conviction is almost correct. All humans are in possession of an alternative cognitive style which gives them more direct access to reality than symbolic representation does. All of them “know”, in some way, that their ordinary experience of the world is incomplete. There is another, “higher” reality, which has to be manipulated to secure one’s well-being. It cannot be manipulated with the help of symbolic reference. As a result, the manipulative tools provided by tradition – rituals and ritual formulas – are and have to be free from symbolic reference: they are, and have to be, holistic.

178

Johannes Bronkhorst

However, the implicit knowledge that people appear to possess about the nature of ordinary experience and its relation to a “different”, “higher” reality goes further than this. Symbolic representation is grounded in non-symbolic experience. Symbolic representation allows for deception and dishonesty, because it can create and communicate imaginary worlds and situations. Escape from symbolic representation through ritual is, in a number of cases, an attempt to re-establish reference by grounding it in the real world. Deacon gives in this context the example of the Yanomamö Indians from Venezuela, notorious for the almost constant warfare among themselves.49 Occasionally peace is assured by means of an elaborate ritual known as a “Feast”. Deacon describes it as follows:50 “First, the hosts who wish to make peace prepare a meal. When their guests are due to arrive, dressed as for war and carrying their weapons, the hosts put their weapons away and the men recline on their hammocks waiting for the guests to enter their village. The guests enter, dancing and chanting, and circle around the camp stopping in front of each host. There they ritually threaten them, raising an axe or drawing a bow and arrow. The hosts must remain unmoved, trying to show no fear and no offense at provocative remarks. After this has been repeated for a while (and latent hostilities have not erupted in violence), the roles are reversed. The guests recline in hammocks, their weapons hidden away, while the hosts circle around the camp dancing and ritually threatening their guests. Finally, when it is clear that nothing untoward is likely to happen, they break off and the guests are offered food. Later they may chant together, barter and exchange goods, or even arrange a marriage.”51 It is ritual which protects hosts and guests from the surprise attacks which are otherwise common among this tribe. Grounding in the real world and the accompanying escape from symbolic representation with its possibilities of lies and deceit52 may explain the use of ritual on other occasions, too. A wedding ritual goes beyond the promises that are exchanged. These promises are, by means of ritual, grounded in a reality in which no

49 Deacon 1997: 403f. 50 Cp. Chagnon 1968: 97 ff. For a brief presentation and discussion of the recent allegation of mistreatment of the Yanomamö by this anthropologist, see Benson & Stangroom 2006: 154 ff. 51 Deacon 1997: 404. 52 This theme is taken very seriously by Rappaport in his study of ritual; see Rappaport 1999: 11f. and passim.


179

deception is possible.53 Similar considerations may be applicable to other rituals, but this topic cannot be further explored here.

5. Mythology and Religion Many of the points discussed above are relevant to the realm which we commonly think of as religion. Symbolic reference accounts for more than just language. It also creates a barrier that separates human beings from a more immediate experience of reality. What is more, human beings are in some way aware of this fact. Some succeed in breaking through this barrier and attain a more direct experience of the world; we have called them mystics. Others search for the hidden reality which they believe must exist by various means: we have considered ritual, but there are no doubt others, among them philosophico-religious speculation. The implicit awareness of a more direct access to the world in which we live often takes the shape of a deep conviction that there is a higher reality, different from the ordinary world. What is more, there appears to be an implicit awareness that the world of our ordinary experience has come forth out of that higher reality. This notion may find expression in “creation myths”, stories which speak of the transition from an initial undivided whole subsequently divided. A primordial formless substance takes form, often under the influence of the spoken word. Rappaport gives a number of examples from a variety of cultures.54 Also the idea that the highest reality is an encompassing whole is known to religious thought. The present study will not enter into an analysis of these beliefs. It can, however, be suggested that they contain an element of truth. Our world has indeed arisen out of an earlier one that was not yet divided. And this undivided whole is really still there and underlies all our cognition. But the transition from an undivided to a multiple world did not happen at the beginning of cosmic time. It took place in our childhood, when we, each of us individually, crossed the symbolic threshold. The preceding reflections, if correct, show that symbolic reference, and the symbolic representation which results from it, have a profound effect on the way we experience the world. Symbolic representation is, in an important sense, that which makes us human, that which distinguishes us even from our nearest cousins in the animal realm. It does not just allow us to learn and use language, it provides us with the implicit or explicit conviction that there is a “deeper” or “higher” reality behind the “ordinary” reality which we experience in our everyday lives. This in its turn impels us to carry out certain forms of activity (“rituals”), and induces us 53 This observation is at variance with Deacon’s view according to which marriage is essentially a symbolic relationship the need for which explains the acquisition of language by early humans: “Symbolic culture was a response to a reproductive problem that only symbols could solve: the imperative of representing a social contract.” (1997: 401). 54 Rappaport 1999: 162ff.

180

Johannes Bronkhorst

to think certain kinds of thoughts. In view of all this, it is hard to overestimate the importance of symbolic reference. This in its turn raises the question whether symbolic reference makes itself also felt in other domains of human activity. The next section will argue that it does, and that culture may be one such domain.

6. Culture We have so far considered symbolic representation in connection with language and, in particular, with vocal utterances. There is, however, no reason to believe that the faculty that allows us to create symbolic representation is only active in the presence of vocal utterances. This same faculty allows the deaf to use sign language, which employs gestures rather than vocal utterances. Sign language has syntax like ordinary language, and shares many features with it. It is indeed possible that gestures preceded vocal signs in the historical development of language.55 However, we may go one step further and consider that the faculty which makes symbolic representation possible is not confined to language in any of its forms, whether vocal, gestural, or other (including lexigrams). As pointed out above, the crucial faculty that enables the human infant to learn language is the capacity to pay attention not just to the objects in its environment, but also to features – whether vocal, gestural, or other – that it interprets as signs of those objects. This capacity allows, even obliges, the child to pay attention to further signs that are somehow related to these initial signs, and to the relationships between them. We have seen that the child uses this capacity when learning the language of its caretakers. This capacity is not, however, confined to the signs that make up this language. It also covers everything else that the infant associates, for whatever reason, with the objects of its experience and which it therefore interprets as signs.56 The very capacity that allows the child to learn its language will also make it pay attention to other, non-linguistic, “signs” and look for the ways in which these signs are interrelated. In the case of language, the system of relationships between linguistic signs which it ultimately adopts is the structure that governs the language of the child’s caretakers. In the case of other, non-linguistic “signs”, too, the child is likely to adopt the system of relationships that its caretakers adhere to. These somewhat abstract reflections suggest that the child, at the appropriate age, is not only busy acquiring its language. Alongside language, it explores the relationships that may exist between other potential signs. As in the case of lan55 See Rowan Hooper in New Scientist of 5 May 2007, 6f. 56 Cp. Deacon 1997: 264: “The contributions of prefrontal areas to learning all involve, in one way or another, the analysis of higher-order associative relationships. More specifically, [...] they are necessary for learning associative relationships where one associative learning process must be subordinated to another.”


181

guage, the child will in the end settle on the web of relationships which is the one accepted by those in its surroundings, or one close to it. It follows that the faculty underlying symbolic representation weaves not just one, but two webs that separate us from the outside world. One of these is the web woven by language; we have discussed it at some length. The other is the web woven by the non-linguistic signs that have come to be recognised as such in the culture of the individual concerned. This web of non-linguistic signs can vary from culture to culture. Examples will therefore be culture-specific. As a widely understood, yet simple, and no doubt simplifying, example from the western world we may consider the Christmas tree.57 The Christmas tree is not just a sign that designates something (Christmas, the birth of Christ, or something else again). It is also, perhaps even primarily, an element in a wide web of associations, at least for those who have grown up in cultures where Christmas was celebrated. People who have grown up in other cultures may learn what the Christmas tree stands for, they will, however, not share this wide associative web. A Christmas tree will not “mean” anything to them. In the case of those for whom it “means” something, we may assume that the Christmas tree is treated in a way not dissimilar to the way in which symbols (i.e. Peircean symbols, primarily the words of language) are treated. Like words, the Christmas tree is used to construct a system, a web of associations. (Unlike words, it may here be added, the Christmas tree and other cultural “symbols” are not likely to give rise to the precise representations that the use of words obliges their users to create.) It may be justified to say that this web, or rather the accumulation of webs of which the Christmas-tree-web may be one, is the culture in which the child grows up.58 As in the case of language, the web of culture may affect cognition. Interestingly, there is some evidence to suggest that cultures do indeed affect cognition differentially, that is to say, members of different cultures cognise the world differently.59 57 The first four notes of Beethoven’s Fifth Symphony constitute another example. Marvin Minsky said the following about them: “no one could remember Beethoven’s Fifth Symphony entire, from a single hearing. But neither could one ever hear again those first four notes as just four notes! Once but a tiny scrap of sound, it is now a Known Thing – a locus in the web of all the other things we know, whose meanings and significances depend on each other.” (Cited in Sacks 2007: 211). 58 For a recent discussion of culture, see Plotkin 2007a. 59 See Plotkin 2007: 236ff., with references to various publications (Lillard 1998; Nisbett et al. 2001; 2003; 2005; Siok et al. 2004; Atran 1998; Medin & Atran 2004; Greenfield et al. 2003; Cole 2006). This suggestion is parallel to the hypothesis according to which individual languages are responsible for the way their users understand the world and act in it, the so-called Sapir-Whorf hypothesis; see, e.g. Gleitman & Papafragou 2005. Indeed, some maintain that the two theories cannot be separated: “a language is shaped by its culture, and a culture is given expression in its language, to such an extent that it is impossible to say where one ends and the other begins, i.e. what belongs to language and what to culture” (Grace 1987: 10).

182

Johannes Bronkhorst

The web of culture, like the web of language, can be torn, presumably by the same method, viz. when the person concerned enters into a state of absorption. This suggests that the so-called mystics experience a world that is not only free from all that language has superimposed on it, but also from numerous other superimpositions, which we may globally refer to as cultural superimpositions. It follows from what precedes that the very capacity that enables us to learn and use language and is behind certain behaviours and experiences which are commonly referred to as religious, may also be responsible for human culture.60 This capacity identifies certain objects and sensory inputs as signs of others, and looks for the system of their mutual relationships. A subset of these signs along with their mutual relationships constitutes language. Others constitute networks of connections, systematised and rationalised in ways that may differ from one culture to the next (just as languages vary). These networks of connections may incorporate background assumptions about the nature of the world around us. They may also, as in the case of the Christmas tree, have a primarily emotional character, and give “meaning” to our world. As in the case of language, these networks of associative connections are likely to be deeply anchored in each person. And as in the case of a first language, considerable effort may be required for alternative networks to find a place beside the original ones. Appendix 1: Semantic Etymologies61 Symbolic reference is essential for language learning. It allows us to “extract” words from larger utterances, just as Sherman and Austin had “extracted” GIVE from GIVE + BANANA. However, there is no obvious reason why this analytical process should stop at words. We might expect that human beings are inclined to push the analysis further, so that symbolic reference then goes beyond the level at which it is useful in daily life, and overshoots the mark, so to speak. This is what happens in the phenomenon to be considered in this Appendix. Let us recall that human beings, more than chimpanzees, have the capacity to analyse acoustic and other inputs so as to discover symbols that may be hidden in them. Sherman and Austin had a hard time segmenting the unit GIVE out of the combinations GIVE + BANANA and GIVE + ORANGE. Human children segment countless words out of a linguistic input that contains no clear indications as to where one word ends and the next one begins, even less where one morpheme 60 It goes without saying that the term “religion” as commonly used also covers certain behaviours and experiences that are “cultural” in the sense here employed. In terms of the theory here presented, what is ordinarily understood as religion combines elements that require different explanations: some are “cultural” features, others fall into the category of “attempts to step back out of the realm of symbolic representation”. 61 Many of the examples given in this Appendix have been taken from Bronkhorst 2001.


183

ends and the next one begins. These same children nevertheless succeed in attributing meanings to the results of their segmentations. The question is, do children, and adult speakers for that matter, stop at the commonly recognised words of language, or do they move on beyond words, perhaps even beyond the morphemes that make up words? This question is of some importance. If one thinks that words correspond to preexisting representations, language users will then divide the linguistic input they receive into words that fit those pre-existing representations, and no further, because there are no representations further down the line.62 In this publication we take a different position: many representations are created, often (but not always) on the basis of the words of language. This position leads to the reasonable expectation that the segmentation that humans apply to their linguistic input is not halted once they arrive at some presumably natural representations. We would rather expect that human segmenting extends also beyond the words of language, into the realm of parts of words and perhaps even individual sounds.63 This is indeed what we find. Human segmenting of this kind finds expression in the phenomenon which I will call semantic etymologising. This phenomenon is widespread. A semantic etymology is to be distinguished from a historical etymology. A historical etymology presents the origin or early history of a word; it tells us, for example, that a word in a modern language is derived from another word belonging to an earlier language, or to an earlier stage of the same language. The English word militant, for example, is derived from Latin militans through the intermediary of French militant. And the Hindi pronoun maim ‘I’ is derived from Sanskrit mayâ ‘by me’ through Prakrit mae.64 Semantic etymologies do something different. They connect one word with one or more others which are believed to elucidate its meaning. The god Rudra, for example, has that name according to the Vedic Satapatha Brâhmana (6.1.3.10), because he cried (rud-) in a story that is told about him. Semantic etymologies tell us nothing about the history of a word, but something about its meaning. Semantic etymologies have largely gone out of fashion these days. Most sensible people have serious doubts about the possibility of finding the meaning of one word by comparing it with other, more or less similar words. We tolerate such semantic etymologising from children, who indulge in it quite freely, as Jean Piaget (1925) and others after him have shown. We are less tolerant with respect to adults who do so; the person who analyses the word contentment as concerning 62 See Pinker 2007: 89f. for a discussion of different points of view regarding representations (there called “concepts”) in language and thought. 63 The question of individual speech sounds is tricky, for the notion that speech can be analysed into a sequence of phonemes, and their perception has been argued to be inseparable from alphabetical writing; Warren 1999: 169f. 64 Oberlies 1998: 17.

184

Johannes Bronkhorst

being content with men, or with tea (content-men-t), is diagnosed as schizophrenic by modern investigators, perhaps rightly so.65 And yet semantic etymologies are widespread in all pre-modern cultures. Here are some examples: In the Sumerian myth of Enki and Ninhursag the former is cured when Ninhursag causes deities to be born corresponding to Enki’s sick members: “The correspondence between the sick member and the healing deity rests on the [...] etymologizing of the ancient scribes; the Sumerian word for the sick organ contains at least one syllable in common with the name of the deity. Thus e.g. one of the organs that pained Enki was the ‘mouth’, the Sumerian word for which is ka, and the deity created to alleviate this pain is called Ninkasi; similarly, the goddess born to alleviate the pain of the rib, the Sumerian word for which is ti, is named Ninti, etc.”66 An ancient Egyptian text carved inside two pyramids dating from the twentyforth century B.C.E. “is full of plays on words” such as: “O Atum-Kheprer, [...] thou didst arise (weben) as the ben-bird of the ben-stone in the Ben-House in Heliopolis.”67 Sauneron adds further examples and points out that ‘plays on words’ were considered to give an ‘explanation’ of the world.68 In the Hebrew Bible etymologies are common, especially in connection with names: Adam is linked with adama ‘earth’ (Gen. 2.7); woman, isha, is derived from man, ish (Gen. 2.23); Cain from qaniti ‘I have gotten’ (Gen. 4.1); etc.69 Kirk70 emphasises the use of etymologies in Greek myths and states: “The poets of the Homeric tradition were already intrigued by the resemblance of the name ‘Odysseus’ to the verb odussomai ‘I am angry’. [...] Pytho, the old name for Delphi, is derived [in the Hymn to Apollo, probably late seventh century B.C.E.] from the serpent destroyed there by Apollo and allowed to rot, puthein. [...] Heraclitus the Presocratic philosopher found it significant that one word for a bow resembled the word for ‘life’ (biós and bíos), and Aeschylus related the name of Helen to the idea that she ‘took the ships’ (hele-naus), that of Apollo to apollunai, ‘destroy’, and that of Zeus to zên, ‘live’.”71 Similar efforts at etymologising characterise later Greek antiquity.72

65 66 67 68 69 70 71 72

So Werner & Kaplan (1963: 259), citing a patient of Maria Lorenz (1961: 604). Kramer 1969: 37, n. 13. Wilson 1969: 3. Sauneron 1957: 123f. See further Morenz 1957; Sander-Hansen 1946, esp. 19f. Böhl 1991: 163f. Kirk 1974: 57f. Ibid.: 58. For a study of the etymologies in Homer, see Rank 1951; also Kraus 1987: 31f. For an (incomplete) list of etymologies in Plutarch, see Strobach 1997: 186f.


185

An example from medieval Europe is provided by the secret spiritual organisation of the Fedeli d’Amore, whose representatives were active in France, Italy, and Belgium from the twelfth century onward. They used a hidden language in order to keep their mystery of love secret. Love for them is a soteriological means, and accordingly the word amor ‘love’ is interpreted as a-mor ‘without death’: A senefie en sa partie Sans, et mor senefie mort; Or l’assemblons, s’aurons sans mort.73 Caesarius of Heisterbach (ca. 1170 – ca. 1240) gives an explanation of the word mors ‘death’ in his Dialogue on Miracles:74 “Through the transgression of the first created, death entered into the world. Hence death (mors) received its name from ‘biting’ (morsus). As soon as man bit (momordit) the apple of the forbidden tree, he incurred death and subjected himself as well as his whole posterity to its necessity. Death is also said to have come from ‘bitterness’ (amaritudine), because, as it is said, no pain in this life is more bitter than the separation of body and soul.” Elsewhere he explains the word puer, “boy”: “Puer (‘boy’) signifies purus (‘pure’)”.75 An example from ethnographic records is the following: Among the inhabitants of the Trobriand islands the word vatuvi occurs in a magical formula.76 This word has no grammatical form; it is neither noun nor verb. Malinowski observes: “the real etymological identity of this word will define it as connected with vitawo, or the prefix vitu-, and the word vituvatu, ‘to institute’, ‘to set up’, ‘to direct’, ‘to show’. [It has] also [...] fortuitous, but magically significant associations with vatu, ‘coral boulder’, ‘coral reef’, and the more or less real word va-tuvi, ‘to foment’, ‘to make heal’.”77 All these semantic etymologies illustrate one and the same phenomenon: the search for expressive units beyond the level recognised in the speech community concerned. They remind us that the capacity for symbolic reference was not created for language learning (even though its usefulness in language learning no doubt explains the selective advantage it offered). 73 74 75 76

See Eliade 1986: 112. Cited in Zaleski 1988: 50. Cited in ibid.: 52. Malinowski (1935: vol. 1: 96; vol. 2: 257) describes it as the most important formula in all Omarakana garden magic. 77 Malinowski 1935: vol. 2: 249; cf. 260–261. Regarding the last association, va-tuvi, Malinowski observes (ibid.: 260–261): “As a matter of fact, one or two natives [...] gave me this explanation of the word when commenting upon the spell.” It is not clear whether any native made the association with vatu explicit.

186

Johannes Bronkhorst

The Chinese language, with its monosyllabic words, lends itself less easily to an analysis beyond the level recognised in the speech community. However, its many homonyms invite its users to connect unrelated things that have the same name. Indeed, “Han commentators applied a form of correlative thought in their philological studies, frequently explaining the meaning of obscure characters by sound analogy on the assumption that a phonetic correspondence indicated a semantic relation”.78 “Sometimes highly complex circular shou emblems [symbols of long life or immortality] had incorporated into their design a swastika (pronounced wan), to express by a pun the concept of wan shou, meaning ‘ten thousand years of long life’.” Similarly: “The endless knot [was] interpreted [...] as symbolizing Buddha’s intestines (ch’ang). [...] [S]ince its name, ch’ang, made a pun on the word for long, the whole figure [...] symbolized [to the later Chinese] a long life [...]”79 etc. An example closer to our time is found in the weekly journal Newsweek of July 6, 1987, p. 18: “Hong Kong’s new British governor, Sir David Wilson, bowed to local tradition by changing his Cantonese name, Ngai Tak-ngai, shortly before assuming office last April. Its characters were homophones for the phrase ‘so hypocritical it’s dangerous’; his new moniker, Wai Yik-shun, means ‘guardianship’ and ‘trust’, conjuring up more soothing images to colony residents [...]”. These examples do not illustrate the search for expressive units beyond the level commonly recognised, but something different. They illustrate the search for a shared semantic interpretation of minimal units, even where the speech community assigns different interpretations to them. This search is, of course, similar to the search that allowed Sherman and Austin to arrive at one representation for GIVE, starting from GIVE + BANANA and GIVE + ORANGE. Appendix 2: Symbolic Reference and the Origin of Language The origin of language is increasingly debated. To get an impression of some current opinions, we turn to a recent issue of the journal Lingua, which is dedicated to language evolution.80 One of the contributors, Derek Bickerton, points out that “[t]his is an interdisciplinary game, played by biologists, neurologists, anthropologists, archaeologists, computer scientists, philosophers, and more – as well as [...] by linguists”.81 On the following page he emphasises the distinction between “language evolution” and “changes in languagES”82. The former is part of the biological evolution of humans, the latter belongs to the realm of cultural change.

78 79 80 81 82

Henderson 1984: 19–20. Cammann 1962: 98, 99–100. Lingua 117/3 (2007): 503–604. Bickerton 2007: 510. Ibid. 511.


187

An essential concept in the study of language evolution is protolanguage. Says Bickerton: “The notion that the earliest stages of language evolution involved a largely if not entirely structureless protolanguage [...] is now so widely accepted that the term seems to have passed into the general vocabulary of language evolutionists.”83 This protolanguage is, as indicated, largely if not entirely structureless. What were its constituents? According to Bickerton, “there does not, as yet, exist any compelling reason for rejecting the original concept of a protolanguage as containing a categorially complete, if severely limited vocabulary of items roughly equivalent to modern words, but lacking a sophisticated phonology and any consistent structure”.84 Not everyone accepts this position. A rival one is the theory according to which protolanguage was holophrastic. Bickerton says the following about it: “[T]he most radical proposal with regard to the constituents of protolanguage is that these were holophrastic rather than synthetic. This proposal has been most thoroughly developed by [Alison] Wray [...]. Wray’s proposals support genre continuism, since [...] animal calls are roughly equivalent to holophrases, rather than words. Wray claims that protolanguage simply increased the number of such units to a point where they began to impose an excessive memory load, at which point the holophrases were decomposed on the basis of phonetic similarities. Here we can do no better than quote Wray’s own example: ‘So if, besides tebima meaning give that to her, kumapi meant share this with her, then it might be concluded that ma had the meaning female person + beneficiary.’ ([Wray] 2000: 297)”85 Bickerton does not agree with this proposal, as is clear from his comments: “This leaves out of account the possibility that, although the syllable ma might occur as Wray suggests, it would also occur in a number of holophrases lacking any references to either females or beneficiaries. Not only is this extremely likely, but the only possible alternative is, if anything, even more damaging to Wray’s case. For if ma occurred always and only where female beneficiaries were involved, the holophrastic protolanguage would be a hollow charade, a mere disguise for a medium already fully synthetic. But if ma also occurred where a female + beneficiary was impossible – contexts perhaps as numerous as, or more numerous than, those that can bear such a reading – why would the hearer assume that it referred to a female benefici-

83 Ibid.: 515. 84 Ibid.: 517. 85 Ibid.: 516–517.

188

Johannes Bronkhorst ary in just those cases where such a reading was possible, and how would that hearer account for the other cases?”86

For further arguments against the holophrastic model Bickerton refers to one of his own earlier publications (Bickerton 2003) and to the contribution of Maggie Tallerman in the same issue of Lingua (Tallerman 2007). Alison Wray, the main target of Bickerton’s criticism, is also a contributor to this issue of Lingua, in an article that is co-signed by George W. Grace. The two authors describe their understanding of the first language users in the following words:87 “[W]e must understand that what made the very first language users different from their parents was that they possessed the capacity to identify patterns inside their existing message units and extract (apparently) recurrent material for recombination. [...] The first ‘segmenters’ need not have stood out all that much from those around them, for theirs would have been a marginal activity relative to the general use of holistic forms with agreed functions. Those who could segment out sections from holistic utterances for recombination could do so [...], while the others carried on using what they already knew. The analysis, operating in direct response to interactional need, could thus be naturally very slow, and indeed would need to be, both because the analyticity of the modern speaker would be little challenged by the holistic usages of his pre-modern companions, and because, in the short term, his novel expressions, while meaningful to him, would be impenetrable to the rest, unless they learned them whole. But little by little, under the influence of even one analytic operator and his/her descendants over a number of generations, an initially immutable protolanguage could progressively transform into something more flexible, until a command of the flexibility became advantageous to survival and/or reproduction.”88 This passage is slightly puzzling in that it calls those with the capacity to segment existing units “the first language users”. One would rather have expected that these segmenters be called “the first word users”: through their segmenting they would have arrived where Bickerton’s protolanguage users were all along, viz. in the possession of “a categorially complete, if severely limited vocabulary of items roughly equivalent to modern words, but lacking [...] any consistent structure” (see above). This is no doubt what Bickerton had in mind when he wrote: “All the sub86 bid.: 517. 87 The idea of a holistic protolanguage is not new. As Van Driem (2004) points out, Hugo Schuchardt argued already in 1919 that the first utterance arose from the splitting of a holistic primeval utterance (“Sprachursprung” I & II, Sitzungsberichte der Preussischen Akademie der Wissenschaften 52: 716–720, 863–869). 88 Wray & Grace 2007: 570–571.


189

stantive problems in language evolution – how symbolism got started and fixed, how, when, and why structure emerged, where and to what extent any of this got instantiated in neural tissue – remain to be solved, whether one accepts a holistic account or not.”89 Bickerton concludes from this that “it is more parsimonious to assume that language began as it was to go on – that discrete symbols, whether oral or manual, were there from the beginning”.90 I am not sure whether parsimony settles this issue, even if we were to follow Bickerton in other respects. It is at least conceivable that language evolution passed through various phases, starting from a holophrastic phase, which was then (partly) succeeded by a synthetic phase.91 Alternatively, one might think that the very process whereby segmenting became possible was also responsible for at least some structure. We will return to this issue below. Let us assume, for argument’s sake, that there was once an exclusively holophrastic protolanguage. The first segmenters differed from their parents on account of a new capacity, the capacity to segment holophrastic phrases. This new capacity may have had a biological basis. This, at any rate, is what Wray and Grace assume. Only in this way can “even one analytic operator and his/her descendants over a number of generations” exert the influence referred to in the above passage: if this capacity were cultural, there would be no need to wait a number of generations, and the novel expressions of the first language users would not necessarily be “impenetrable to the rest”. We will see below that this is not the only possible point of view.92 For the time being we will, however, follow Wray and Grace in assuming

89 Bickerton 2003: 87 (emphasis mine, JB). 90 Ibid. Interestingly, Wray (2000: 287) invokes Occam’s razor to argue for the continuity between the holistic communication used by primates and the holistic language we use today. 91 This is Mithen’s position (2005: 260): “This ‘words before grammar’ is the type of language evolution that Bickerton proposed – so we can see that his views are not necessarily wrong, but are simply chronologically misplaced and require a pre-existing history of holistic protolanguage to be feasible.” Wray (2002a) speaks of “the holistic system that I propose preexisted and provided the context for the emergence of the analytic system” (116) and of “the holistic protolanguage that I propose coexisted with – and preceded – the simple analytic system from which our modern grammatical capability developed” (ibid.: 118); see also the following: “there is a place for a simple analytic proto-language of the kind that Bickerton proposes, but not as the sole medium of communication” (ibid.). Arbib (2003: 183; cp. 2005: 108, 118f.) introduces the notion of a prelanguage which, he suggests, was composed of “unitary utterances” and which preceded the discovery of words: “On this view, words in the modern sense co-evolved with syntax through fractionation, a process of discovery and diffusion quite distinct from the formation of a genetic module for grammar.” 92 Citing Dennett (2003: 184–185), but putting language instead of religion, one might say: “If there are any genes for [language], this is, in fact, one of the least interesting and least informative of the Darwinian possibilities.”

190

Johannes Bronkhorst

that the new capacity must be thought of as being based on a genetic mutation, and therefore as part of biological evolution. The above passage from Wray and Grace’s article also presupposes that all members of the group, including those who do not yet possess the capacity to identify segments, are capable of learning new utterances. Only in this way can they “learn whole” expressions that are impenetrable to them. These expressions are not therefore genetically transmitted. This presupposition is not spelled out in Wray and Grace’s article, but seems to be important in view of the fact that it takes for granted an important difference between human protolanguage and primate calls, which are genetically transmitted. The specification that the protolanguage out of which modern language eventually arose was different from primate calls and that it consisted of learned utterances that could be distinguished from each other answers a number of the objections voiced by Tallerman in her contribution to the Lingua issue, notably those that deal with the presumed proximity of the protolanguage to animal calls93 and with its presumed phonological insufficiency.94 We cannot deal with them in detail. Some of her other criticisms are summed up and responded to by Mithen in the following terms: “Although linguists unsympathetic to Wray’s approach, such as Derek Bickerton and Maggie Tallerman, have questioned the feasibility of segmentation, their criticisms have been unsubstantial. Tallerman claims, for instance, that the likelihood of any chance associations arising is remote, but computer simulations have shown that they could easily have happened [...] while Wray herself has explained how the process of segmentation may have been supported by ‘near matches’. Tallerman also claims that holistic utterances are unlikely to have contained multiple phonetic strings with the potential for segmentation because they would have been too short. But one need do no more than consider the lengthy holistic phrases of monkeys, apes, birds, and so forth, to appreciate that those of our human ancestors may

93 Elsewhere in her article Tallerman accepts that “protolanguages (like languages) are culturally transmitted” (2007: 599). 94 See in this connection Wray 2002a: 115: “I have suggested that the holistic cries and gestures of our pre-human ancestors were transformed, over a long period of time, into a phonetically expressed set of holistic message strings [...]” (emphasis mine, JB). Also Arbib 2003: 196: “[T]he lowering of the larynx in humans or pre-human hominids might have served a similar purpose (as in the red deer, viz. deepening the animal’s roar, JB) – without denying that further selection could have exploited the resultant increase in degrees of freedom to increase the flexibility of speech production. [...] [T]his selective advantage would hold even for a species that employed holophrastic utterances devoid of syntax.” (Emphasis mine, JB).


191

have been of considerable length,95 having evolved over millennia and proliferated in number to provide ever greater semantic specificity. Similarly, an argument that Wray’s theory of segmentation has to assume the prior existence of discrete segments, which then invalidates her conception of holistic utterances, also has no foundation. Holistic utterances may have been multi-syllabic but they were – by definition – holistic, with none of the individual or groups of syllables mapping onto a discrete aspect of the utterance’s complex meaning.”96 It is not the aim of this Appendix to enter into an exhaustive discussion of all the arguments for and against the holophrastic model (whether in the form of the assumption that all human protolanguage was holophrastic, or rather that a holophrastic protolanguage gave rise to an analytic protolanguage which in its turn gave rise to language). What counts at this point is that this model is not dead, far from it, it is alive and kicking. This allows us to consider some of its consequences. Wray and Grace observe: “For reasons that may be partly biological, the older the individual becomes, the more likely he or she is to dissect language more than is strictly necessary for effective general communication.”97 Since the capacity of dissecting language does not respond to a specific need, the result may be surprising: “Any biological influences on the balance between formulaicity and compositionality may be limited to the peripheral capacity to open up a form before there is a specific need to do so, thus maintaining a creative edge to one’s engagement with language. Since such a capacity would simply need to exist, rather than achieve any specific goal (such as a complete analysis of the language into atomic particles and rules), we might anticipate finding that its effects are haphazard and idiosyncratic – and indeed we do.

95 Consider in this connection the following: “In a paper with Nobuyuki Kawai, [Tetsuro] Matsuzawa showed that Ai, a young chimpanzee, could remember a sequence of at least five numbers, more than a preschool child; and at a recent symposium in Chicago on ‘The Mind of the Chimpanzee’, he showed how Ai, with further training, had developed powers of working memory beyond that of most human adults. He suggested that ‘our common ancestors might have had immediate memory, but in the course of evolution, they lost this and acquired languagelike skills.’” (Sacks 2007: 159, n. 8; see further Kawai & Matsuzawa 2000, and Cohen 2007). 96 Mithen 2005: 254. 97 Wray & Grace 2007: 561.

192

Johannes Bronkhorst One individual might suddenly wonder what it is that is ‘done’ in how do you do?, while another notices that barking seems to contain the morphemes bar and king. Such inappropriate analyses will capture attention whereas others, that turn out to be rational in terms of the shared perceptions of compositional structure in the language, are simply absorbed.”98

The “peripheral capacity” referred to in this passage is central to the holophrastic model (in either of its two forms). The model predicts this capacity, and would indeed be empty without it. The existence of this capacity in modern humans, on the other hand, does not suffice to prove the correctness of the holophrastic model. The fact that one sometimes wonders what is ‘done’ in how do you do? does not, by itself, prove that human language arose out of a protolanguage (or proto-protolanguage) which was holistic. However, if it could be demonstrated that the capacity to open up forms before there is a specific need to do so is more than a capacity, that it is a proclivity that manifests itself in various ways across human cultures and stages of development, this would constitute an argument in support of the holophrastic model. Such a proclivity would be harder to account for in the synthetic model, for here words are extracted to correspond to pre-existing concepts. The widely attested use of semantic etymologies (discussed in Appendix 1, above) provides such a demonstration. Let us now restate some of the objections that have been voiced against the holophrastic model. Bickerton, in a passage already cited, found fault with Wray’s hypothetical example in which, besides tebima meaning give that to her, and kumapi meaning share this with her, early language users might conclude that ma had the meaning female person + beneficiary. Bickerton’s criticism took the following form: “This leaves out of account the possibility that, although the syllable ma might occur as Wray suggests, it would also occur in a number of holophrases lacking any references to either females or beneficiaries. Not only is this extremely likely, but the only possible alternative is, if anything, even more damaging to Wray’s case. For if ma occurred always and only where female beneficiaries were involved, the holophrastic protolanguage would be a hollow charade, a mere disguise for a medium already fully synthetic. But if ma also occurred where a female + beneficiary was impossible – contexts perhaps as numerous as, or more numerous than, those that can bear such a reading – why would the hearer assume that it referred to a female

98 Ibid.: 561–562.


193

beneficiary in just those cases where such a reading was possible, and how would that hearer account for the other cases?”99 Tallerman discusses the same hypothetical example and comments: “A major problem in this regard is that logically, similar substrings must often occur in two (or more) utterances which do not share any common elements of meaning at least as many times as they occur in two utterances which do share semantic elements. For instance, suppose that a string mabali also contains the ma sequence, but means ‘put that rock down!’. What ensures that ma gets associated with ‘her’?”100 Bickerton and Tallerman are probably not aware of the fact that reflections similar to theirs occur already in one of Plato’s Dialogues, the Cratylus. This Dialogue presents Socrates in discussion with Cratylus, an etymologist. Socrates subjects the procedure of (semantic) etymologising to a thorough analysis. This procedure is not dissimilar, in Socrates’ analysis, to the extraction of ma in the above hypothetical examples. Socrates would call ma a “primary name”; he further maintains that primary names are by their very nature like the things they denote.101 In the course of his analysis he comes to assign certain meanings to various letters of the Greek alphabet: the letter rho is expressive of speed, motion, and hardness; the letter lambda is like smoothness, softness, and other qualities. However, this leads to difficulties in words like sklêrótês which, though containing the letter lambda, means “hardness”.102 Like Bickerton and Tallerman, Socrates considers these difficulties insurmountable: “I myself prefer the theory that names are, so far as possible, like the things named; but really this attractive force of likeness is [...] a poor thing, and we are compelled to employ in addition this commonplace expedient, convention, to establish the correctness of names”.103 Socrates was not, of course, discussing the transition from protolanguage to language. He was concerned with semantic etymologies. Just as Bickerton and Tallerman try to undermine the hypothesis of a holistic protolanguage, Socrates’ arguments should have discouraged semantic etymologising. They did no such thing. Consider the reflections of Isidore of Seville, the sixth-century author of a monumental work called, precisely, Etymologies. About etymologies he says, among other things:104

99 Bickerton 2007: 517. 100 Tallerman 2007: 597. 101 Plato, Cratylus: 169 (Transl. Fowler). 102 Ibid.: 173. 103 Ibid.: 175. 104 Barney et al. 2006: 55 (Etymologies I.xxix).

194

Johannes Bronkhorst “Etymologies of words are furnished either from their rationale (causa), as ‘kings’ (rex, gen. regis) from [...] ‘acting correctly’ (recte agendum); or from their origin, as ‘man’ (homo) because he is from ‘earth’ (humus), or from the contrary, as ‘mud’ (lutum) from ‘washing’ (lavare, ppl. lutus), since mud is not clean, and ‘grove’ (lucus), because, darkened by its shade, it is scarcely ‘lit’ (lucere).”

Isidore should perhaps have followed Socrates’ example and pondered upon the question whether there is some meaning that can be ascribed to the sound sequence hom/hum and that recurs in all formations of which this sequence is part. He might have offered similar reflections with regard to lu and luc. He did not. Difficult cases and counterexamples clearly did not discourage him. Examples of etymologies like these can be cited from widely different cultures and literatures; further examples are provided by children and psychiatric patients. This has been shown in Appendix 1. Strictly speaking they do not extract meaningful segments, at least not explicitly. It is, however, clear that they are based on the similarity or identity of parts of the words compared. In many cases there is no linguistic justification whatsoever for these semantic etymologies. This is stated in so many words in a text from ancient India that deals with etymologising, Yaska’s Nirukta. Ancient India too had a sophisticated tradition of grammar, which obliged scholars to distinguish between the two activities. Yaska’s position in this matter is clear: etymologising is the complement of grammar, it treads where grammar cannot go.105 Consider now Sverker Johansson’s criticism of the idea of a holistic protolanguage:106 “It is not obvious to me [...] why the segmentation process envisaged by Wray [...] would be expected to work. A similar process is certainly present in modern-day language acquisition – children first acquire some stock phrases as unanalyzed wholes, and later figure out their internal structure – but that only works because these stock phrases have an internal structure, given by the grammar of the adults from whom the child acquires them.”107 Johansson assumes (as do Bickerton and Tallerman) that the segmentation process can only work where the material already contains, in a consistent and organised manner, the segments that are being extracted. Socrates would agree, for he refers repeatedly to the name-givers who gave names to things in a long distant past. But not all semantic etymologies are based on this belief. Most semantic etymologising takes place quite independently of it. Semantic etymologising 105 Bronkhorst 1984. 106 Cited by Tallerman 2007: 594. 107 Johansson 2005: 234.


195

“works”, i.e. it is practised, in spite of the fact that the words and phrases dealt with do not have any relevant internal structure. We must conclude that there is a widely attested proclivity to extract segments and assign meanings to them even in cases where the material is inconsistent and resists such segmentation. This proclivity coincides in all essentials with the capacity to segment postulated by the upholders of the theory of a holophrastic protolanguage. It is not limited to situations in which the prior existence of discrete segments has to be assumed. This, then, does away with some of the most serious criticism of the idea of a holophrastic protolangage. Having come this far, let us consider in some detail what the transition from holistic protolanguage users to segmenters implies. It has been suggested that a genetic mutation may be responsible, but this remains vague. Indeed, it amounts to little more than giving a name (“genetic mutation”) to a process which one does not understand. And yet, the process concerned may not escape understanding, and what is more, it is open to question whether a genetic mutation has to be postulated to explain it. The event which turns holophrasts into segmenters may be nothing else than crossing the symbolic threshold. Our hominid ancestors crossed this threshold (phylogenetically) many thousands of years ago, and all human beings who learn to use language do the same (ontogenetically) during their early years. Symbolic reference even seems able to explain at least some of the syntactic rules of language. Deacon argues, to begin with, “that the structure of [Universal Grammar] would not yield to a biological evolutionary solution, nor would it have a neurological explanation”.108 According to him, many structural features of language may derive from semiotic constraints, inherent in the requirements for producing symbolic reference. Indeed, “some major universals of grammar may come for free, so to speak, required by the nature of symbolic communication itself”.109 Deacon is of the view that “it may turn out that many core features of [Universal Grammar] – recursive operations, the minimally diadic structure of sentences, and many subjacency, ordering, and dependency constraints of grammar – trace their origin to [...] indexical [...] requirements [which] apply to all symbolic linguistic relationships”.110 As a result, “theoretical efforts now directed toward showing how innate [Universal Grammar] could have evolved and how it might be instantiated in brains could well turn out to be mere intellectual exercises”.111

108 Deacon 2003a: 90, with a reference to 1997. See also Van Driem 2001: 66ff. (“The splitting of the symbol and the birth of syntax”). 109 Deacon 2003: 138. 110 Ibid.: 133. 111 Ibid.: 139.

196

Johannes Bronkhorst

It would seem, then, that the crossing of the symbolic threshold had a double effect on the forms of communication of our remote ancestors. Symbolic reference enabled them to segment the holophrastic utterances (or other forms of communication, e.g. gestures) they had hitherto employed. On top of that, symbolic reference imposed certain constraints on the combined use of the resulting segments.


197

References Alper, Harvey P. (ed.) 1989. Understanding Mantras. Albany: State University of New York Press. — 1989a. “Introduction”. In: Harvey P. Alper (ed.). Understanding Mantras. Albany: State University of New York Press: 1–14. Arbib, Michael A. 2003. “The Evolving Mirror System: A Neural Basis for Language Readiness”. In: Morton H. Christiansen & Simon Kirby (eds.). Language Evolution. Oxford: Oxford University Press: 182–200. — 2005. “From Monkey-Like Action Recognition to Human Language: An Evolutionary Framework for Neuro-linguistics”. Behavioral and Brain Sciences 28: 105–167. Atran, Scott 1998. “Folk Biology and the Anthropology of Science: Cognitive Universals and Cultural Particulars”. Behavioral and Brain Sciences 21: 547–609. Barney, Stephen A. et al. (eds., trans.) 2006. The Etymologies of Isidore of Seville. Cambridge et al.: Cambridge University Press. Benson, Ophelia & Jeremy Stangroom 2006. Why Truth Matters. London, New York: Continuum. Bickerton, Derek 2003. “Symbol and Structure: A Comprehensive Framework for Language Evolution”. In: Morton H. Christiansen & Simon Kirby (eds.). Language Evolution. Oxford: Oxford University Press: 77–93. — 2007. “Language Evolution: A Brief Guide for Linguists.” Lingua 117/3: 510–526. Böhl, Felix 1991. “Die etymologische Namensdeutung in der rabbinischen Literatur”. In: Jean-Pierre Chambon & Georges Lüdi (eds.). Discours étymologique, Actes du Colloque international organisé à l’occasion du centenaire de la naissance de Walther von Wartburg. Avec la collaboration de Hans-Martin Gauger, Frank Lestringant, Georges Pinault. Tübingen: Max Niemeyer Verlag: 161–172. Boyer, Pascal & Pierre Liénard 2006. “Why Ritualized Behavior? Precaution Systems and Action Parsing in Developmental, Pathological and Cultural Rituals”. Behavioral and Brain Sciences 29: 1–56. Bronkhorst, Johannes 1984. “Nirukta, Unadi Sutra, and Astadhyayi”. Indo-Iranian Journal 27: 1–15. — 2001. “Etymology and Magic: Yaska’s Nirukta, Plato’s Cratylus, and the Riddle of Semantic Etymologies”. Numen 48: 147–203. Cammann, Schuyler 1962. Substance and Symbol in Chinese Toggles. Chinese Belt Toggles from the C. F. Bieber Collection. Philadelphia: University of Pennsylvania Press. Candland, Douglas Keith 1993. Feral Children and Clever Animals. Reflections on Human Nature. New York, Oxford: Oxford University Press. Cardeña, Etzel & Stephen Jay Lyn & Stanley Krippner (eds.) 2000. Varieties of Anomalous Experience: Examining the Scientific Evidence. Washington: American Psychological Association. Chagnon, Napoleon A. 1968. Yanomamö: The Fierce People. New York et al.: Holt, Rinehart & Winston.

198

Johannes Bronkhorst

Christiansen, Morton H. & Simon Kirby (eds.) 2003. Language Evolution. Oxford: Oxford University Press. Cohen, Jon 2007. “The World Through a Chimp’s Eyes”. Science 316: 44–45. Cole, Michael 2006. “Culture and Cognitive Development in Phylogenetic, Historical and Ontogenetic Perspective”. In: William Damon et al. (eds.). Handbook of Child Psychology. Vol. 2: Cognition, Perception and Language. New York, Chichester: Wiley: 636–686. Damasio, Antonio 1999. The Feeling of What Happens. Body, Emotion and the Making of Consciousness. London: Vintage. Deacon, Terrence W. 1997. The Symbolic Species. The Co-Evolution of Language and the Human Brain. Harmondsworth: Penguin Books. — 2003. “Universal Grammar and Semiotic Constraints”. In: Morton H. Christiansen & Simon Kirby (eds.). Language Evolution. Oxford: Oxford University Press: 111–139. — 2003a. “Multilevel Selection in a Complex Adaptive System: The Problem of Language Origins”. In: Bruce H. Weber & David J. Depew (eds.). Evolution and Learning: The Baldwin Effect Reconsidered. Cambridge, London: MIT Press: 81–106. Dennett, Daniel C. 1991. Consciousness Explained. Boston et al.: Little, Brown and Company. — 2003. Freedom Evolves. London: Allen Lane. Diamond, Jared 1992. The Third Chimpanzee: The Evolution and Future of the Human Animal. Toronto: HarperCollins. Edelman, Gerald M. 2004. Wider than the Sky. The Phenomenal Gift of Consciousness. London: Penguin Books. Eliade, Mircea 1986. Histoire des croyances et des idées religieuses. Vol. 3: De Mahomet à l’âge des Réformes. Paris: Payot. Gellman, Jerome. 2005. “Mysticism”. In: Edward N. Zalta (ed.). The Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/archives/spr2005/entries/mysticism/. Gleitman, Lila & Anna Papafragou 2005. “Language and Thought”. In: Keith J. Holyoak & Robert G. Morrison (eds.). The Cambridge Handbook of Thinking and Reasoning. New York: Cambridge University Press: 633–661. Goodman, Felicitas D. 1988. Ecstasy, Ritual, and Alternate Reality. Bloomington, Indianapolis: Indiana University Press. Grace, George W. 1987. The Linguistic Construction of Reality. London et al.: Croom Helm. Greenfield, Patricia M. et al. 2003. “Cultural Pathways through Universal Development”. Annual Review of Psychology 54: 461–490. Henderson, John B. 1984. The Development and Decline of Chinese Cosmology. New York: Columbia University Press. Herman, David. 2003. “Stories as a Tool for Thinking”. In: David Herman (ed.). Narrative Theory and the Cognitive Sciences. Stanford: CSLI Publications: 163–192. Humphrey, Caroline & James Laidlaw 1994. The Archetypal Actions of Ritual. A Theory of Ritual Illustrated by the Jain Rite of Worship. Oxford: Clarendon Press. Isidore of Seville: Etymologies. See Barney et al. 2006.


199

Johansson, Sverker 2005. Origins of Language: Constraints on Hypotheses. Amsterdam, Philadelphia: John Benjamins. Jouvet, Michel 1999. The Paradox of Sleep: The Story of Dreaming. Cambridge: The MIT Press. Jucker, Jean-Luc 2007. “Des marqueurs rituels. Ethnographie d’une messe tridentine”. ethnographiques.org 13 – juin 2007. http://www.ethnographiques.org/2007/Jucker. html. Kawai, Nobuyuki & Tetsuro Matsuzawa 2000. “Numerical Memory Span in a Chimpanzee”. Nature 403: 39–40. Kirk, Geoffrey S. 1974. The Nature of Greek Myths. Harmondsworth: Penguin Books. Knight, Chris & Michael Studdert-Kennedy & James R. Hurford (eds.) 2000. The Evolutionary Emergence of Language. Social Function and the Origins of Linguistic Form. Cambridge University Press. Kramer, Samuel N. 1969. “Sumerian Myths and Epic Tales”. In: James B. Pritchard (ed.). Ancient Near Eastern Texts relating to the Old Testament. Third Edition with Supplement. Princeton: Princeton University Press: 37–59. Kraus, Manfred 1987. Name und Sache: Ein Problem im frühgriechischen Denken. Amsterdam: B. R. Grüner (Studien zur antiken Philosophie 14.) Lawson, E. Thomas & Robert N. McCauley 1990. Rethinking Religion. Connecting Cognition and Culture. Cambridge University Press. Lillard, Angeline. 1998. “Ethnopsychologies: Cultural Variations in Theories of Mind”. Psychological Bulletin 123: 3–32. Llinás, Rodolfo R. & D. Paré 1991. “Of Dreaming and Wakefulness”. Neuroscience 44/3: 521–535. Lorenz, Maria 1961. “Problems Posed by Schizophrenic Language”. Archives of General Psychiatry 4: 603–610. Maldonado, Jose R. & David Spiegel 1998. “Trauma, Dissociation, and Hypnotizability”. In: Douglas Bremner & Charles R. Marmar (eds.). Trauma, Memory, and Dissociation. Washington (D.C.), London: American Psychiatric Press, 57–106. Malinowski, Bronislaw 19662 [1935]. Coral Gardens and Their Magic. Vol. 1: SoilTilling and Agricultural Rites in the Trobriand Islands. Vol. 2: The Language of Magic and Gardening. London: Allen Unwin. McNab, Fiona & Torkel Klingberg 2007. “Prefrontal Cortex and Basal Ganglia Control Access to Working Memory”. Nature Neuroscience 11: 103–107. Medin, Douglas L. & Scott Atran 2004. “The Native Mind: Biological Categorization and Reasoning in Development and across Cultures”. Psychological Review 111: 960–983. Michaels, Axel 2006. “Ritual and Meaning”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing Rituals: Issues, Topics, Approaches, Concepts. Leiden, Boston: Brill: 247–261 (Numen Book Series, Studies in the History of Religions 114/1).

200

Johannes Bronkhorst

— 2007. “‘How do you do?’ Vorüberlegungen zu einer Grammatik der Rituale”. In: Heinrich Schmidinger & Clemens Sedmak (eds.). Der Mensch – ein “animal symbolicum”? Sprache – Dialog – Ritual. Darmstadt: Wissenschaftliche Buchgesellschaft: 239–258. Mithen, Steven 2005. The Singing Neanderthals. The Origin of Music, Language, Mind and Body. London: Phoenix. Morenz, Siegfried 1957. “Wortspiele in Ägypten”. In: Festschrift Johannes Jahn zum 22. November 1957. Leipzig: 23–32. (Reprint: Morenz 1975: 328–342). — 1975. Religion und Geschichte des alten Ägypten. Gesammelte Aufsätze. Köln, Wien: Böhlau. Morris, Desmond 1967. The Naked Ape. London: Cape. Newton, Michael 2002. Savage Girls and Wild Boys. A History of Feral Children. London: Faber and Faber. Nisbett, Richard E. et al. 2001. “Culture and Systems of Thought: Holistic versus Analytic Cognition”. Psychological Review 108: 291–310. — & Takahiko Masuda 2003. “Culture and Point of View”. Proceedings of the National Academy of Sciences of the United States of America 100/19: 11163–11170. — & Yuri Miyamoto 2005. “The Influence of Culture: Holistic versus Analytic Perception”. Trends in Cognitive Sciences 9/10: 467–473. Oberlies, Thomas 1998. Historische Grammatik des Hindi. Die Genese seines morphologischen Systems aus dem Mittel- und Altindischen. Reinbek: Inge Wezler (Indologische Lehrmaterialien 3). Padoux, André 1989. “Mantras – What Are They?”. In: Harvey P. Alper (ed.). Understanding Mantras. Albany: State University of New York Press: 295–318. Paper, Jordan 2004. The Mystic Experience. A Descriptive and Comparative Analysis. Albany: State University of New York Press. Piaget, Jean 1925. “Le réalisme nominal chez l’enfant”. Revue Philosophique de la France et de l’étranger 99: 189–234. Pinker, Steven 2007. The Stuff of Thought. London et al.: Penguin. Plato: Cratylus. With an English Translation by H. N. Fowler. Vol. 4. Cambridge: Harvard University Press 1977 [London: Heinemann 1921] (The Loeb Classical Library 167). Plotkin, Henry 2007. Necessary Knowledge. Oxford: Oxford University Press. — 2007a. “The Power of Culture”. In: Robin I.M. Dunbar & Louise Barrett (eds.). Oxford Handbook of Evolutionary Psychology. Oxford: Oxford University Press: 11–19. Pritchard, James B. (ed.) 1969. Ancient Near Eastern Texts relating to the Old Testament. Third Edition with Supplement. Princeton: Princeton University Press. Pyysiäinen, Ilkka 1993. Beyond Language and Reason. Mysticism in Indian Buddhism. Helsinki: Suomalainen Tiedeakatemia (Annales Academiae Scientiarum Fennicae, Dissertationes Humanarum Litterarum 66). — 1996. Belief and Beyond. Religious Categorization of Reality. Åbo: Åbo Akademi. (Religionsvetenskapliga skrifter 33).


201

Ramachandran, V. S. & Sandra Blakeslee 1998. Phantoms in the Brain. Probing the Mysteries of the Human Mind. New York: William Morrow. Rank, Louis Philippe 1951. Etymologiseering en Verwante Verschijnselen bij Homerus. Utrecht (unpublished dissertation). Rappaport, Roy A. 1999. Ritual and Religion in the Making of Humanity. Cambridge: Cambridge University Press. Sacks, Oliver 2007. Musicophilia. Tales of Music and the Brain. London: Picador. Sander-Hansen, C. E. 1946. “Die phonetischen Wortspiele des ältesten Ägyptischen”. Acta Orientalia 20: 1–22. Sauneron, Serge 1957. Les prêtres de l’ancienne Égypte. Paris: Éditions du Seuil. Savage-Rumbaugh, Sue & Duane M. Rumbaugh 1978. “Symbolization, Language, and Chimpanzees: A Theoretical Reevaluation Based on Initial Language Acquisition Processes in Four Young Pan Troglodytes”. Brain and Language 6: 265–300. — & William Mintz Fields & Jared Taglialatela 2000. “Ape Consciousness – Human Consciousness: A Perspective Informed by Language and Culture”. American Zoologist 40: 910–921. Searle, John R. 1995. The Construction of Social Reality. Harmondsworth: Penguin Books. — 1999. Mind, Language and Society. Doing Philosophy in the Real World. London: Weidenfeld & Nicolson. Siok, Wai Ting et al. 2004. “Biological Abnormality of Impaired Reading Is Constrained by Culture”. Nature 431: 71–76. Staal, Frits 1984. “Ritual, Mantras and the Origin of Language”. In: Amṛtadhārā. Professor R. N. Dandekar Felicitation Volume. Delhi: Ajanta Publications: 403– 425. — 1985. “Mantras and Bird Songs”. Journal of the American Oriental Society 105/3 (Indological Studies Dedicated to Daniel H. H. Ingalls. Special Issue): 549–558. — 1996 [1990]. Ritual and Mantras: Rules without Meaning. Reprint. Delhi: Motilal Banarsidass. — 2006. “Artificial Languages across Sciences and Civilizations”. Journal of Indian Philosophy 34: 89–141. Stace, Walter T. 1960. Mysticism and Philosophy. London: MacMillan & Co. Strivay, Lucienne 2006. Enfants sauvages: Approches anthropologiques. Paris: Gallimard. Strobach, Anika 1997. Plutarch und die Sprachen. Ein Beitrag zur Fremdsprachenproblematik in der Antike. Stuttgart: Franz Steiner (Palingenesia 64) Tallerman, Maggie 2007. “Did Our Ancestors Speak a Holistic Protolanguage?”. Lingua 117/3: 579–604. Turner, Mark 1996. The Literary Mind. The Origins of Thought and Language. Oxford et al.: Oxford University Press. Turner, Victor W. 1969. The Ritual Process. Structure and Anti-Structure. London: Routledge & Kegan Paul. (The Lewis Henry Morgan Lectures 1966)

202

Johannes Bronkhorst

Van Driem, George 2001. Languages of the Himalayas. An Ethnolinguistic Handbook of the Greater Himalayan Region. Vol. 1. Leiden et al.: Brill (Handbook of Oriental Studies, Section Two: India 10/1) — 2004. “Language as Organism: A Brief Introduction to the Leiden Theory of Language Evolution”. In: Ying-chin Lin et al. (eds.). Studies on Sino-Tibetan Languages. Papers in Honor of Professor Hwang-cheng Gong on His Seventieth Birthday. Taipei: Institute of Linguistics, Academia Sinica: 1–9 (Language and Linguistics Monograph Series W–4). — (forthcoming). The Language Organism: Parasite or Mutualist? Warren, Jeff 2007. The Head Trip. Adventures on the Wheel of Consciousness. Oxford: OneWorld Publications. Warren, Richard M. 1999. Auditory Perception. A New Analysis and Synthesis. Cambridge: Cambridge University Press. Werner, Heinz & Bernard Kaplan 1963. Symbol Formation. An Organismic-Developmental Approach to Language and the Expression of Thought. New York, London, Sydney: John Wiley & Sons. Wilson, John A. 1969. “Egyptian Myths, Tales, and Mortuary Texts”. In: James B. Pritchard (ed.). Ancient Near Eastern Texts relating to the Old Testament. Third Edition with Supplement. Princeton (New Jersey): Princeton University Press: 3–36. Wray, Alison 2000. “Holistic Utterances in Protolanguage: The Link from Primates to Humans”. In: Chris Knight & Michael Studdert-Kennedy & James R. Hurford (eds.) 2000. The Evolutionary Emergence of Language. Social Function and the Origins of Linguistic Form. Cambridge University Press: 285–302. — (ed.) 2002. The Transition to Language. Oxford: Oxford University Press. — 2002a. “Dual Processing in Protolanguage: Performance without Competence”. In: Alison Wray (ed.) 2002. The Transition to Language. Oxford: Oxford University Press: 113–137. — 2002b. Formulaic Language and the Lexicon. Cambridge: Cambridge University Press. Wray, Alison & George W. Grace 2007. “The Consequences of Talking to Strangers: Evolutionary Corollaries of Socio-Cultural Influences on Linguistic Form”. Lingua 117/3: 543–578. Wulff, David M. 2000. “Mystical Experience”. In: Etzel Cardeña & Stephen Jay Lyn & Stanley Krippner (eds.) 2000. Varieties of Anomalous Experience: Examining the Scientific Evidence. Washington: American Psychological Association: 397–440. Zaleski, Carol 1988. Otherworld Journeys. Accounts of Near-Death Experience in Medieval and Modern Times. Oxford: Oxford University Press.

J.C. Heesterman

The Development and Impact of Ancient Indian Ritual 1. The most elaborate and arguably oldest corpus of ritual texts forms part of the canonical Vedic scriptures, the hallowed śruti. It consists of prose brāhmaṇas which, together with the later manuals, the śrauta-sūtras, make up the corpus rituale that later theoreticians understood to contain the dharma, the universal order and man’s place in it. This order is marked by the codanā, the summons to sacrifice. Put briefly, the universal order is the ritual of sacrifice. The Vedic ritual, then, may be “meaningless”, as Frits Staal famously proposed1 – a point to which we shall have occasion to come back – but it does have a function, namely sacrifice, seen as the dynamics of the universal order. The primordial act of creation was sacrifice. As the involuted Ṛgvedic verse has it: “with sacrifice the gods sacrificed sacrifice; these were the first ordinances” (dharmāṇi prathamāni).2 Since these first ordinances are the ritual of sacrifice, we are given to understand that śrauta ritual shared in the primordiality of creation. Considered to be, therefore, eternal (nitya) and of non-human origin (apauruṣeya) the authority of the Vedic scriptures is unassailable – though equally unattended to. Or, as Louis Renou put it: “The Veda is a monument to which one tips one’s hat in passing without bothering about it any further.”3 Nonetheless, even though generally disregarded, the Veda and especially the Vedic śrauta ritual, did have, as I shall argue, a powerful formative impact. Its exalted place in Indian civilisations as well as its antiquity could not fail to attract intense academic interest at the time of the expansion of humanities studies in the nineteenth and twentieth centuries. It was the poetics and the mythology of the Ṛgveda and the metaphysical arcana of the Upaniṣads that drew most of the attention. The Vedic corpus rituale, however, though not neglected, was considered of decidedly lower order. It is only recently, now that ritual as such has become an independent field of research in its own right that Vedic ritual has come 1 Staa1 1980. 2 Ṛgveda 1.164.15 (Vishva Bandhu 1963: 1062–1063); 10.90.16 (Vishva Bandhu 1965: 3688); Atharvaveda 7.5.1 (Vishva Bandhu 1961: 879–880). 3 Renou 1960: 2.

204

J.C. Heesterman

into its own in scholarly appreciation. The elaborate and surprisingly well-preserved textual documentation as well as the development of theoretical reflection that may be rated an ancient Indian “science of ritual”, make Vedic ritual an incomparable source for ritual studies.

2. Not only the elaborate texts and their age-old tradition but no less so the rigidity and the transparent systematisation of the ritual invite scholarly attention. These qualities are especially favourable to the “syntactic” approach preferred by Staal instead of the usual semantic one. A brief overview may suffice to illustrate the systematic construction of śrauta ritual. The various formats of sacrifice are arranged according to increasing complexity and, concomitantly, the increasing number of priests. The simplest, basic ritual act is the libation in the sacrificial fire(āhuti), followed by the vegetal sacrifice (iṣṭi), then the animal sacrifice (paśubandha) and finally the intricate paradigm of the soma sacrifice (somayāga), requiring the full complement of seventeen officiating priests. The latter is preceded by the consecration (dīkṣā) of the sacrificer of which the main part is the procuring of the stalks of the soma plant, the preparation of the soma beverage by pressing the stalks, offering in the fire and finally the drinking of the soma. Pressing, offering and drinking the soma is done thrice – morning, midday and evening – accompanied by chants (stotra) and recitations (śastra). The simpler paradigms are integrated into the next more intricate ones, in the way of Russian dolls, and all of them turn up in the soma sacrifice. Then again, there are many types of soma-sacrifices mainly differing in length – from one day to a (fictitious) thousand years containing the whole array of the one-day paradigm in an endless repetition. By means of subtle variations within the structure of acts, mantras, recitations and chants, as well as in certain particular rites (such as royal ones), the otherwise rigid system of rules shows a remarkable flexibility in accommodating various occasions and purposes. The imposing ancient Indian śrauta system of ritual – as complex as it is rigid – tempts us to analyse its structure, taking it as an ideal example of ritual tout court, irrespective of its purpose. The question then is, how this ritual system relates to the institution of sacrifice; or in what way does the ritual represent sacrifice?

3. This question requires that we first take a brief look at the world-wide institution of sacrifice. Generally speaking, sacrifice appears to respond to feelings of insecurity

Development and Impact of Ancient Indian Ritual

205

and anxiety. Man is aware of being surrounded by all kinds of danger, spelling ruin and death. Ultimately sacrifice is concerned with the irresolvable tension between life and death – inextricably connected with each other, while at the same time each other’s absolute negation. Sacrifice cannot, of course, resolve the irresolvable. It does, however, give expression to the contradictory nexus of life and death, so as to make the tension manageable. It may do so by alleviating or channelling the unavoidable anxiety by way of strictly ruled action – action, karma, being a usual term for sacrifice. Consequently sacrifice has its natural place at critical times, such as the turn of the year or season, or in the personal life-cycle, as well as times of incidental threats or in no less critical cases of exceptional good luck. Put briefly, sacrifice is concerned with crisis. In fact, it is itself an evocation of crisis. It may even function as a “pre-emptive strike”; better than waiting for the crisis to hit, one may provoke a crisis one can control, as in the story of the exceptionally fortunate Polycrates who sacrificed his ring – tantamount to his “external soul” – albeit in vain. Usually the sacrificed food or valuables are rationalised – not least by the participants – as a gift to supernatural powers in order to obtain their favour and support. However, there is a significant difference between gift and sacrifice. Whereas the gift only passes from one owner to the other and so remains in circulation, the sacrificed goods are definitively withdrawn from circulation and given over to destruction in the sacrificial fire – as usually in Vedic sacrifice – immersion in water or a watery place or again simply exposed and left unattended to waste away. Even when impressive amounts of food are consumed and valuable gifts distributed, the essence of sacrifice is not gift-giving but the element of loss and destruction. Sacrifice, then, turns on an act of derelictio, abandonment pure and simple. It is this voluntary abandonment, the tyāga (or utsarga) that is the one act in the lengthy and complicated ritual procedure that brings about the sought-after bliss. It is not the gods, the nominal addressees, who produce the hoped for effect but the act of abandonment. Such is the explicit doctrine of the Mimamsa theoreticians.4

4. The matter of abandonment and gifts brings us to another aspect of sacrifice, namely sacrifice as contest.5 The most direct and concrete expression of the harrowing tension of the life-and-death nexus – that is the focus of sacrifice – is the contest for the “goods of life” as against the threat of their loss and of death. The contending parties – who usually maintain relations of exchange (such as commensality and connubium) – alternately take turns as host and guest, challenging each 4 See Biardeau 1964: 87. 5 See Heesterman 1993: 41–44.

206

J.C. Heesterman

other in the arena of sacrifice in games such as chariot racing and playing at dice, which interestingly we find inserted in the paradigms of śrauta rituals, as well as competitive spending of wealth. Essential is the cyclical exchange of positions, when the one-time recipient of his host’s largesse must reciprocate, “sich revanchieren”, on pain of shame and dishonour. This cyclical pattern has no end and by itself may go on forever. By the same token the losing party may hope to recuperate at the next round while conversely the victorious one is honour-bound to stake his winnings again in the recurring sacrificial contest. The agonistic pattern of sacrifice is akin to the potlatch cult among Amerindian tribes of the American Northwest, where competitive gift-giving and sacrificial destruction reached a surprisingly high level of intensity. In fact, such dualistic agonistic patterns are fairly widespread in archaic societies.

5. Viewed against this background sacrifice is deeply ambiguous, if not plainly contradictory. Prodigal distribution of life-affirming wealth we find juxtaposed to equally prodigal loss and destruction, the two opposites – like the contending parties – being united by the contest. When we now look at the Vedic śrauta sacrifice we notice a striking contrast. The śrauta system of ritual leaves no room for contest and uncertainty. Everything is fixed by the rules beforehand. If the ritual involves a chariot race there is no uncertainty about who will win the race. Similarly, where a game of dice is prescribed the ritual makes clear who shall make which move and when, while the outcome is, of course, equally prescribed. Moreover, the texts leave no doubt that the sacrificer will attain his aim in undertaking the sacrifice, even though one does not know when the result will come about. The only thing that might go wrong is a mistake in the execution of the complicated ritual. But for such eventualities there are corrective rites (prāyaścitta) and at the end there is a general corrective rite to compensate for mistakes that passed unnoticed.6 But how then is the śrauta ritual related to sacrifice? Far from fitting in with each other sacrifice and śrauta ritual are at opposite ends from each other. The essential uncertainty and open-endedness of sacrifice is denied by the rigid and closed ritual. Interestingly this strictly closed character of the ritual appears to have embarrassed the ritualists themselves. Thus we learn that it is better to obtain one’s domestic fire from elsewhere rather than drilling it oneself (as is the rule for the sacrificial fire). The latter method is, of course, most meritorious (puṇyatama) but, 6 The compensatory rites form a separate section in the ritual manuals (śrauta-sūtras); for the general compensatory rite see Hillebrandt 1879: 164–l68.


207

being closed to the outside world, it is anardhuka,7 barren. In the case of the domestic fire then, the open-endedness is preferred.

6. Actually the classical śrauta system of ritual the late Vedic texts present bears the signs of a fundamental reform of the institution of sacrifice. It was, it would seem, the endlessness of the cyc1ical recurrence of agonistic sacrifice that does not know a definitive conc1usion but instead offers the prospect of renewed sacrificial contest, involving reckless expenditure and destruction of wealth. It carries with it the permanent threat of collapse, such as India’s great epic, the Mahābhārata, extensively depicts. As Marcel Mauss appositely put it: “The Mahābhārata is the story of a gigantic potlatch”.8 The ritualists’ response to this threat was to rule out the recurrent contest by the simple but drastic means of exc1uding the rival party from the sacrificial arena, which then was no longer an arena but a private place of sacrifice for the now unopposed solitary sacrificer. Hence the chariot race is no longer a contest but a purely “as-if” show.9 And so is the game of dice when it is part of the ritual. What the war- and racing-chariot stands for is still perfectly clear. It is the emblem of the warrior and his prowess. Thus we see the royal sacrificer at the celebration of his consecration driving a chariot in a mock raid conquering a herd of cattle. Exceptionally, in this case we do see a live opponent, playing the role of the owner or guardian of the cattle; but the opponent is no more than an inactive dummy. And, anyway, the scene is outside, at a safe distance from the place of sacrifice. In another case where we find the chariot again, its warlike function is less obvious. This case concerns the standard paradigm of the soma ritual. After having been consecrated the first task of the soma sacrificer is to acquire wealth – mainly cattle to be spent in sacrifice.10 According to the ritual rules he should do so by soliciting voluntary donations either through emissaries or by himself. We are then extensively told which mantras he should utter upon receiving the various kinds of goods collected as well as which in case of loss, depending on the kind of goods lost. Apparently it is a rather circumstantial affair involving extensive travel. These details seem to suggest an expedition of transhumance or ransom-taking, the two not being necessarily different. And indeed, immediately afterwards – but without any connection with the foregoing – we are given the rules for travelling by 7 Khādiragṛhyasūtra 1.5.1–3 (Singh 1934: 37–38); Gobhilagṛhyasūtra 1.1.16–18 (Knauer 1884: 1). 8 Mauss 1990: 55. 9 See Heesterman 1993: 64–68. 10 Āpastambaśrautasūtra 10.18.4–19.5 (Tithe 2004: 526–527); Heesterman 1993: 12.

208

J.C. Heesterman

chariot, including those for crossing a river.11 One need not be overly suspicious to see what is behind the desultory sequence of details regarding the collection of the goods for sacrifice, the circumstantial travel and the use of a chariot. We have already been alerted to their background by the royal cattle raid. Apart from the possible loss of cattle on the way we are not told of any unpleasantness. There is, however, one case of an inimical confrontation, but again completely disconnected from the foregoing rules. The occasion is the sacrificer’s arrival at the place of sacrifice with his fire-drill and the soma stalks he has obtained.12 The brahmayana texts suggest that it is a critical event. The move to the place of sacrifice – here without the benefit of a chariot and without any reference to the foregoing rules – is characterised by these texts as a conquering progress. On arrival the gods Soma and Agni, represented by the soma stalks and the fire-drill, must be pacified by means of a sacrificial victim. Otherwise they would in unison attack the sacrificer. But who else apart from the arriving consecrated sacrificer, who has Soma and Agni with him, is there to be attacked? The point is that the consecrated sacrificer has to double as his own rival, because there is no real rival any more. Of course, the rival’s part can be taken over, as some manuals prescribe, by one of the priestly officiants. But even so the sacrificer is still his own challenger. Incidentally, this double role of the sacrificer may also explain – at least in part – the utter rigidity and closure of the śrauta ritual. Such a strict system of rules is needed to enable the single sacrificer to play the two opposite roles – a feat requiring an accurate “choreography”. Of course, agonistic sacrifice did not lack in ritual rules. But these are of another order. They are the “rules of the game” that do not determine the outcome but are on the contrary meant to keep the outcome open. The śrauta rules, on the other hand exclude risk and uncertainty. They replace the “rules of the game” with their reverse: the “game of the rules”.

7. These examples – the chariot and the arrival at the place of sacrifice – give us a view of the way the ritualists restructured the agonistic pattern of sacrifice. A leading principle appears to be faithfully to preserve the elements of the agonistic pattern, such as the chariot, but to break up the agonistic scenario, to which they belong, into separate parts. Thus the chariot is safely detached from the collection of goods for sacrifice. The movements outside the place of sacrifice are broken into separate parts – those connected with the collection of cattle, the movement by chariot to the place of sacrifice, long distance travel, involving crossing rivers, a

11 Āpastambaśrautasūtra 10.19.6–7 (Tithe 2004: 528), 13–14 (ibid. 529). 12 Āpastambaśrautasūtra 10.29.4 (Tithe 2004: 553). See Blondeau & Schipper 1995: 7–14.


209

separate march to and confrontational arrival – without chariot but with Soma and Agni – at the place of sacrifice. In sum, the elements suggestive of agonistic proceedings have been made either harmless – such as the collection of goods and the confrontation on arrival at the place of sacrifice – or are deprived of their coherence and so rendered meaningless. They are then separately reassembled in a rather desultory sequence. The rejection of the sacrificial contest and the consequent loss of coherence also means the break-down of the cyclical pattern of “revanche” and reversal. Sacrifice was now complete in itself and concluded by the same rite as at the beginning. The cycle was closed within the single sacrifice that was no longer followed by a “return match”. The cyclical concatenation was flattened into a linear sequence of separate parts. The transition from cyclical recurrence to linear sequence was not without importance. The cyclical pattern is, as a matter of course, endless; barring death and destruction it lacks a self-evident conclusion. The linear sequence of the śrauta ritual, on the other hand, can be lengthened or shortened at will. Thus for instance, the sacrifice, in which the royal consecration (rājasūya) is embedded, is a lengthy succession of soma and other sacrifices (in one of which the royal unction, abhiṣeka, is inserted), stretched over a couple of years. It thereby suggests its origin in a pattern of cyclical recurrence. But it can also be shortened to a one-day event, called rājābhiṣeka. Similarly the standard paradigm of the soma sacrifice lasting one day is, after its concluding bath, to be followed by a vegetal iṣṭi (udayanīyā), then still another animal sacrifice and finally, after moving to another place and setting up the sacrificial fires there, a last vegetal iṣṭi (udavasānīyā). Then, at long last the sacrificer returns home. This rather arbitrarily drawn out recurrence of travel and sacrifice – reminiscent of the sacrificer’s itinerary before the actua1 soma sacrifice – unavoidably suggests a truncated cyclical pattern flattened into a linear sequence. A remarkable point in all this is that the śrauta ritual presents us with a “double focus”. It honours the tradition of cyclical agonistic sacrifice by preserving characteristic elements of the agon and, on the other hand, breaks down the agonistic scenario and lines up the disconnected elements. This carefully contrived “double focus” does not appear to be a flaw in the ritualists’ reconstruction of sacrificial ritual but rather a matter of principle.

8. The incisive reform of sacrificial ritual had far-reaching consequences – the removal of the rival sacrifice off from its social context. Śrauta sacrifice is a strictly private affair of the individual sacrificer. Accordingly the śrauta canon has no

210

J.C. Heesterman

sacra publica. Even the royal sacrifices – such as rājasūya and vājapeya – are not public affairs and the king is not different from the commoner sacrificer, both being equal in the eye of the law of ritual. Lifted out of its worldly moorings the sacrificial śrauta ritual was placed in a separate sphere of its own, transcending the lived-in material world. Against this background we can understand the exalted status of the śrauta ritual as the ultimate dharma, the universal law. The absolute ultramundane character is made perfectly clear by the role played by an act of pure dharma, defined as being adrṣṭārtha, lacking any “visible”, i.e. worldly aim. In that sense the śrauta system of ritual is indeed “meaningless”. On the other hand, how was the transcendent dharma to relate to the world? Like the contradictory saying about the constitutional king, the dharma reigns on high but does not govern. We may find the answer to our query in the ritualistic reworking of the myth of the warrior god Indra and the dragon Vṛtra who holds the universe in his grip. This version tells us about the long-lasting sacrificial contest of Prajāpati, the Creator and Lord of Life, who takes the part of Indra, against his riyal Mṛtyu, Death, in the role of Vṛtra.13 Each brings to the contest his own ritual panoply. Mṛtyu’s equipment consists of song, dance and what is prudishly indicated as “what is wantonly done” (yad vṛthācaryate). Prajāpati, however, puts his trust in what we know as the classical śrauta ritual of chant, recitation and acts. The contest remains for a long time undecided till Prajāpati has the decisive vision of the qualitative and numerological equivalence (saṃpad, saṃkhyāna) of his own with Mṛtyu’s ritual activity. This vision enables him to encompass and integrate Mṛtyu’s ritual, thereby conquering him. “Nowadays”, the text triumphantly concludes, “there is no sacrificial contest (saṃsava) any more; the sacrifice is one, Prajāpati is the sacrifice”. What is at stake in this story is the way in which the saṃsava, the contest of Life and Death, is decided. The decisive point is that Mṛtyu and his ritual are not annihilated but integrated by Prajāpati and his classical ritual as they are. Equally significant is the resemblance with the mythical scene of Indra’s victory over Vṛtra. When Indra is about to give the dragon the coup de grâce his defeated enemy pleads: “Do not hit me, you are now what I (was before)”. So they arrive at a compact under which Vṛtra, the power that holds the universe in its grip, enters into Indra.14 Incidentally, viewed in this light, the double role of the soma sacrificer taking over the role of his rival partner appears to be more than just a ritualistic sleight-of-

13 Jaiminīyabrāhmaṇa 2.69–70 (Caland 1970: 143–144). See Heesterman 1993: 53–58. 14 Śatapathabrāhmaṇa 1.6.3.7 (Eggeling 2008); cf. Taittirīyāsaṃhitā 2.4.125–6 (Dharmādhikārī & Sonaṭakke 1970–2007).


211

hand. In fact, it is in consonance with Prajāpati integrating his rival partner, the more so since the śrauta sacrificer is routinely identified with Prajāpati.

9. To sum up briefly, the ritualistic reform of sacrifice – the central institution of archaic society – broke up the common arena that united the rival partners. In its stead it created a gap between the transcendent dharma of ritual, reigning on high, and an unreformed world it is incapable of governing. Yet the story of Prajāpati’s victory by encompassing his rival Mṛtyu, harking back to the primordial conflict of Indra and Vṛtra, stabilises the uneasy relationship of the lived-in order of the world and transcendence. Here we may find the seedbed that produced what Axel Michaels has felicitously called the “identificatory habitus” of Indian thought.15 Its counterpart is the “double focus” of the ritual thought that, arisen from the same source, evokes the ultimate oneness of dharma. Both together, “identificatory habitus” and “double focus”, have taken over the function of the agonistic partnership in the perennial contest, the partners being amalgamated in the one sacrificer who is identified with Prajāpati. Notwithstanding the sharp divide that set the transcendent dharma of ritual apart in a separate sphere of its own, it had a formative impact beyond its scope. The single individual sacrificer had no partner any more with whom to exchange his karma but himself in an unremitting succession of birth and death, pessimistically known as punarmṛtyu, “repetitious death”.16 Hence it is stated in the esoteric, Upanisadic speculation that he who realises that Death is his own self, his ātman, overcomes the feared punarmṛtyu – that is, how Prajāpati overcame Mṛtyu by integrating him into himself. At this point we hit upon what arguably may be rated the most significant heritage of ritualistic thought. This heritage has its origin in the least conspicuous element of sacrifice, the tyāga, the abandonment of the offering to destruction in the sacrificial fire. In terms of worldly experience it is clearly meaningless. It cannot and should not serve any worldly purpose. In other words, it is the act of pure dharma, and exactly for that reason capable of producing its transcendent result. Here again we are confronted with the “double focus” we have come to expect. The hallowed scripture, the śruti, does promise the sacrificer all manner of earthly desires, such as wealth, health, long life, offspring – as is faithfully acknowledged by the Mīmāṃsā theoreticians. But the next statement declares that it is upon 15 Michaels 1998: 19ff. 16 Śatapathabrāhmaṇa 10.6.5.8 (Eggeling 2008); cf. Bṛhadāraṇyaka-upaniṣad 1.2.7 (Swāmī Mādhavānanda 1965: 35–39).

212

J.C. Heesterman

abandoning these desires that the sacrificial acts will be conducive to man’s ultimate purpose (tadutsarge karmaṇi puruṣārthāya). Or as the so-called Forest Teachings of the Mahābhārata succinctly teach: “Perform the ritual and abandon” (kuru karma tyajeti ca).17 Although the element of competition is absent from it, the tyāga, the abandonment to destruction, is akin to the potlatch of destruction. Moreover, the willing abandonment of life by the consecrated or ascetic warrior in what the epic calls the “sacrifice of battle” (raṇayajna) is equally known as tyāga. And even far later, in Rajastāni epic poetry, such a warrior is known as tyagi-vir, a “life renouncing hero”. Here, of course, the “double focus” makes itself known again. The opposite aspect is the sacrifice of “world renunciation” (saṃnyāsa), characterising Indian religiosity.

17 Mīmāṃsādarśana 4.1.3 (Abhyankar 1969–1974); cf. Mahābhārata 3.2.70b (Sukthankar 1942: vol.3, 12).


213

References Abhyankar, Kashinath Vasudev (ed.) 1969–1974. Mīmāṃsādarśānam. Puṇyākhyapattana: Ānandāśrama-mudraṇālaya. Biardeau, Madeleine 1964. Théorie de la connaissance et philosophie de la parole dans le brahmanisme classique. Paris: Mouton. Blondeau, Anne Marie & Kristofer M. Schipper 1995. Essais sur le rituel: Colloque du centenaire de la section des sciences réligieuses de l’École pratique des Hautes Études, vol 3. Louvain et al.: Peeters. Caland, W. (ed. and transl.) 1970 [1919]. Das Jaiminīya-Brāhmaṇa in Auswahl. Reprint. Wiesbaden: Dr. Martin Sändig oHG. Dharmādhikārī, Trivikrama Nārāyaṇa & Nārāyaṇaśarmā Śrīpāda Sonaṭakke (eds.) 1970–2007. Kṛṣṇayajurvedīyā Taittirīya-Saṃhitā. Pune: Tilaka-Mahārāṣṭra-Vidyāpīṭha-ŚākhābhūtaVaidika-Saṃśodhana-Maṇḍala. Eggeling, Julius (transl.) & Maitreyee Deshpande (ed.) 2008. The ŚatapathaBrāhmaṇa: Sanskrit Text with English Translation and Notes. Delhi: New Bharatiya Books. Heesterman, Johannes C.1993. The Broken World of Sacrifice. Chicago: University of Chicago Press. Hillebrandt, Alfred 1879. Das altindische Neu- und Vollmondsopfer in seiner einfachsten Form dargestellt. Jena: G. Fischer. Knauer, Friedrich (ed. and transl.) 1884. Gobhilagṛhyasūtra. Erstes Heft. Dorpat: C. Mattiesen/Leipzig: Commissionsverlag von Simmel & Co. Mauss, Marcel 1990. The Gift: The Form and Reason for Exchange in Archaic Societies. London: Routledge. Michaels, Axel 1998. Der Hinduismus: Geschichte und Gegenwart. Munich: Beck. Renou, Louis 1960. Études védiques et paninéennes, vol. 6. Paris: Boccard (Publications de l’Institut de Civilisation Indienne, 10). Singh, Thakur Udaya Narayana (transl. and ed.) 1934. Khādiragṛhyasūtram. Mozaffarpur: Shastra Publishing House. Staal, Fritz 1980. “The Meaninglessness of Ritual”. Numen 26: 2–22. Sukthankar, Vishnu S. 1942. The Mahābhārata: For the First Time Critically Edited. Poona: Bhandarkar Oriental Research Institute. Swāmī Mādhavānanda (ed. and transl.) 1965. The Bṛhadāraṇyaka Upaniṣad. Calcutta: Advaita Ashrama. Tithe, G.U. (ed. and transl.) 2004. Āpastamba-Śrauta-Sūtra. Delhi: New Bharatiya Book Corporation. Vishva Bandhu et al. (eds.) 1961. Atharvavedaḥ (Śaunakīya). Part II (Kāṇḍas VI–X). Hoshiarpur. Vishveshvaranand Vedic Research Institute (Vishveshvaranand Indological Series 14). — (eds.) 1963. Ṛgvedaḥ. Part II (Maṇḍala I, Sūkta-s 81–191). Hoshiarpur: Vishveshvaranand Vedic Research Institute (Vishveshvaranand Indological Series 20).

214

J.C. Heesterman

— (eds.) 1965. Ṛgvedaḥ. Part VII (Maṇḍala X, 46–191). Hoshiarpur: Vishveshvaranand Vedic Research Institute (Vishveshvaranand Indological Series 25).

Annette Wilke

Basic Categories of a Syntactical Approach to Rituals Arguments for a “Unitary Ritual View” and the Paraśurāma-Kalpasūtra as “Test-Case” In this paper I suggest a fresh approach to questions that have been constantly asked within ritual studies: What makes ritual a ritual? What are the basic categories which must be invariably there? These questions touch upon the question of ritual grammar and syntax. But speaking of grammar and syntax another serious question pops up. After all, grammar is there to form meaningful sentences. Should this linguistic model be applicable to rituals, too? I think it is. A ritual may be compared indeed to a sentence or a poem. The question of meaning (semantic, functional, phatic, emotional, etc.) cannot be escaped. The major difference is only that rituals are generally much more replete with aesthetic stimulation, using not only language, but a large amount of non-verbal means, sensory pageantry, the body and body motion. This makes ritual communication extremely complex, but anything but meaningless. A theory of ritual syntax must account for ritual as highly symbolic action. A syntactical approach to rituals therefore has to tackle the question of how structure and form relate to meaning. I want to explore linguistic models that have a lot to offer to this question, but have not yet been explored within ritual studies. My major source will be the Indian grammarian Bhartṛhari (fifth century A.D.) and his “unitary sentence view” which entails a structural approach to meaning and some basic categories like relation and imagination that help explain ritual as a self-communicating entity and the basic “powers” which are at work. This Bhartṛhari-inspired approach is very much in keeping with Roman Jakobson’s semiotics and what he called “poetic function” which in turn inspired Umberto Eco’s semiotic theory of aesthetic messages and their innate performativity. Since theory cannot do without empirical foundation, my “test case” will be all along the ritual manual Paraśurāma-Kalpasūtra (PKS), a late “left-hand” Tantra source (c. sixteenth century A.D.) of high esteem which belongs (amazingly) to the Śrīvidyā tradition, i.e. a Tantric tradition which is commonly associated with “right-hand” rites. Considering that Tantrics of any denomination have been justly called “super-

216

Annette Wilke

ritualist”,1 the manual promises a fairly comprehensive taxonomy of ritual elements. The analysis of these elements and ritual units of the PKS will furnish my argument for a “unitary ritual view” regarding morphology, syntax, pragmatics and semantics which is inspired by Bhartṛhari. I propose the combination of a structural, semiotic and performative approach to ritual syntax. The agenda is to present a rigorously holistic assessment of ritual and integrate cultural description and formal analysis – i.e. two basic academic approaches to ritual which often have been set apart by their proponents. The aim is not to present a complete set of ritual conventions or an elaborate theory on ritual syntax, but rather basic syntactical categories (sequence, relation, expression and imagination) which are, however, very powerful operational keys for analysing ritual morphology, syntax, semantics and pragmatics. A major argument will be that structure in itself tells already a lot about content and meaning, a self-referential meaning which is intrinsic to the given rite or ritual sequence. Formal aspects like length, singularity and metaphorical density of a given ritual element or increase and decrease of sensory pageantry tell us something about the marked or unmarked position of this part of the ritual and its meaning and significance within the ritual complex. I suggest such structural units are always also meaning units which set in motion and channel perception and the self-generative principle of meaning production. At the outset, some preliminary remarks on methodology may be useful, since any description, method and theory involves implicit and explicit theoretical assumptions. So far, much academic fervour has gone into the search for operational keys to delineate ritual from non-ritual action, and into the question of whether rituals are meaningful or not. The discussion gives me the chance to introduce some major features of the PKS which performatively deals with both. The major objective for this introduction is, however, to discuss some important previous attempts to find a grid of formal and functional rules and categories that make up something like a ritual grammar to which my Bhartṛhari-inspired set of categories will supplement some additional ideas that have not been touched upon so far in the systematic way they deserve to be tackled.

1. What Makes Ritual a Ritual? — From Emic kalpa to Etic Theory and Back The question of defining rituals has been a matter of constant debate. Basically, I do not think that we need a definition of ritual to give ritual studies a sound base. If universal definitions were needed for serious academic study we would have to close down much of our departments of humanities. Just think of terms like “relig1 Sanderson 1988: 662.

Basic Categories of a Syntactical Approach to Rituals

217

ion”, “culture” and “Hinduism”, the meaning of which is the subject of constant debate as well. But I guess most of us would agree that substantial academic work is done in disciplines like indology, cultural anthropology and the study of religions. There is no reason to suppose ritology or ritual studies to be exceptions. The reason is that we have a good deal of common-sense knowledge of what religion or ritual entails. This common-sense knowledge helps us to discern that it would be hard to find a culture or society of past or present where religion as phenomenon and symbolic system which transgresses the mundane while vitally relating to it, is not found. Similarly, we need no definition to discern that all societies and cultures seem to have rituals. The term may be lacking, but the phenomenon is pretty universal. My ritual text-book, for instance, claims to be a Sūtra-treatise on kalpa. Kalpa is a different word from “action”, karma or kriyā. It refers to a particular kind of action. Kalpa covers the semantic field “ritual”, “ceremonial”, “procedure”, “method” and thus perfectly meets a transculturally spread intuition of some most important features ritual action entails. Common-sense knowledge as well as ritual theory will see ritual, especially liturgical rituals, as non-ordinary stylised action, in short some sort of ceremonial. This kind of action is generally applied to mark out non-ordinary circumstances and expectances, and often appeals to supramundane powers who will re-enforce the non-ordinary aura and the sacralising effects of ceremonial action. Further on, ritual action is not up to choice, but rather a prescribed “procedure”, a highly rule-governed activity. Its “methods” must be known. The know-how and skills sometimes have to be learned for years or need special initiation, etc. which in the case of my Kalpasūtra and its predecessors, the Vedic ones, is very much true. In short, the ancient Indian term kalpa demonstrates that we can account for a common-sense knowledge of ritual which is spread in more than one culture. This common sense working in terms like “religion” or “ritual” turns into an academic approach if we develop critical sensitivity for the contexts in which the phenomena appear, and for the discursive and constructivist ways in which religion or a given ritual is ascertained, reproduced and reconstructed in a culture in the run of history and among different groups. We do not need formal rules to describe rituals as precisely as possible and analyse their pragmatic and semiotic cultural and social significance. However, systematics will surely sharpen our analytical approach and move beyond the common-sense intuition or what the cognitivists call ritual competence; the “innate” recognition of ritual form and successful participation.2 In fact, such intuition may sometimes badly fail if we turn to the customs of other cultures. Moreover, in so far as science wants to make general claims and develop typologies, common structures and ideal types, the search for a “ritual grammar” 2 Lawson & McCauley 1990: 77; 2002: 4–10.

218

Annette Wilke

which specifies the elements of ritual structure certainly makes sense. It would be a handy tool for transcultural comparison, and possibly even for ritual notation. The existing ritual theories and debates have already provided good parameters to work with. They have been helpful in delineating ritual from non-ritual action or religious from non-religious rites in a more reflexive way than our usual common sense understanding. By doing so they have been highly fruitful in making more precise our theorising and sharpening our awareness of ritual as specifically marked action. Most importantly for our question, they have already provided a number of formal categories, functional rules and structural elements to establish something grammar-like. Just consider Catherine Bell’s six characteristics of rituallike activities which provide operational keys for distinguishing ritual from normal routine action, namely: formalism, traditionalism, invariance, rule-governance, sacral symbolism and performance.3 This is precisely the abstract formulation of what my term kalpa is all about. This term deliberately points to Vedic traditionalism and orthodoxy, which the PKS claims to attach itself to and at the same time to surpass.4 Bell’s features make a pretty good set of formal categories and structures pertaining in more or less strong degree to any kind of ritualised action. Axel Michaels suggests another helpful list pertaining to ritual pragmatics: intentio solemnis, agency, embodiment, formalism, framing, transformation and effect, elevation or extraordinariness.5 The PKS starts with the declaration (intentio solemnis): “Hence we unfold initiation” (PKS 1.1.). What follows first are, however, not ritual pre/descriptions, but thirty aphorisms (sūtras) on Tantric cosmology, soteriology and ethics (1.2–30), the teaching of which will actually be part of the three-fold initiation rite proper (1.31–43). In this doctrinal introduction, agency is placed into the divine realm; PKS 1.2 traces the Tantric practice and empowerment back to the revelation of the divine agent Parama-Śiva, and the initiatory and daily rites in turn will transform the human agent “in all his limbs” into second Śiva (PKS 10.50). The initial aphorisms are about framing and formalist qualifications: they teach the strict secrecy of the school’s major rituals (mantra practice and “left-hand” body practices, coupled with the yogic body schemes and non-dual contemplation). Only after this highly semantic and cognitively aligned initial part, the very complex and highly formalist initiation rites start, beginning with activities of the guru, and followed by activities of the disciple. Like all initiation rituals these actions perform a socio-religious transformation. They make the disciple a member of the secret Kaula “family” and entitle him to perform the daily rites described in the following chapters. At the same time they perform a spiritual rebirth and a bodily change. 3 Bell 1997: 138–169. 4 Wilke (forthcoming). 5 Michaels 2007: 251–253; see also 2005: 34–35.


219

The guru’s actions mainly consist of visualising practices that transform the disciple’s body into a divine and cosmic one. The disciple is first blindfolded and passive, but after the guru’s mental manipulations are over, he becomes active and the blindfold is removed. He has to step into a ritual diagram (cosmogram), eat a piece of meat soaked in alcohol and touch a limb of the guru’s body which decides about his new name. Invariably, this name will end in the title “lord of bliss” (Ānandanātha) which is a verbalisation of the transformational effects the Tantric rites are claimed to bring about: the corporal enjoyment of the bliss of Brahman, emancipation while living and the becoming of Śiva in all one’s limbs. These procedures may equally well be described by means of another grid of categories suggested by Michaels, namely authority (the highest god Śiva), age (primordial times, predating even the Veda), orthodoxy (the Vedic fold absorbed, supplemented and superseded by Tantric orthodoxy) and extraordinariness (consumption of alcohol, etc. that invert the normal daily social behaviour, attainment of Śivahood).6 Moreover, the PKS fits most perfectly with the formal conditions which must invariably be there according to Thomas Lawson and Robert McCauley to call a religious activity a ritual.7 These are membership condition, change in the religious world and foundational action of a superhuman agent / connection with other rituals in the system. According to the authors the ritual will be most powerful if it is immediately embedded in the archetypal “hypothetical ritual” of the superhuman agent. Such is indeed the case in the PKS. The diverse formal sets of categorisation make a good catalogue to work with. So we have already important operational keys for systematisation which may function as ingredients for a ritual grammar that discerns common ritual structures and conventions for the notation of rituals and ritual elements. It is noteworthy, however, that not only the emic PKS starts with semantic issues, but even the etic formal rules are not entirely without the meaning element. Sacral symbolism (Bell), extraordinariness (Michaels) or CPS-agents (Lawson & McCauley) involve heavy semiotisation and fictionalising “ways of world making”.

2. Are Rituals Meaningful? My answer is quite blank: of course they are. Why should people of all cultures perform ritual actions if they were utterly meaningless to them? Rituals are even the most perfect means to embody meaning and make them social realities. Just consider a wedding and a funeral. These rituals use different sensory pageantry, i.e. aesthetic means that appeal to the senses, emotions and intellect, to express and 6 Michaels 2007: 245. 7 Lawson & McCauley 1990: 84–136; McCauley & Lawson 2002: 13–14, 17–19, 143–145.

220

Annette Wilke

naturalise the culture’s concepts about life and reproduction or death and afterlife. I should add to the PKS-description above that all the ritual procedures involve an orchestrated network of verbal, mental and physical features that are apt to invoke a sensory and affective experience of the semantic claims of cosmisation, empowerment and emancipation. Rituals have semantic as well as functional meaning. They are expressive as well as instrumental. They are cultural “symbolic capital” (Bourdieu). I do not deny that there are reasons why some eminent theorists pretend the meaninglessness of ritual,8 a position Frits Staal most decisively stands for. Ritual is indeed predominantly action, if not only action while it is actually being performed. Meaning often is quite irrelevant while acting out the rite – a sense of duty or the long tradition and custom of fathers and forefathers may be good enough. But at one point in history people learned about the meaning. This meaning may have changed substantially in the course of time. However, invariably there will be some social value and function attached to the given rite. Some sort of code regarding social and cultural significance will be always there. Ritual is like any communication actualising meaning on the basis of a system-rationality (Systemrationalität). The meaning element may not always be semantic in a narrow sense. For ritual agents and participants the cognitive aspect of a religious ritual, for instance, may merely lie in the knowledge that their action is a religious one relating to a sacred cosmos from which can be expected powerful effects. Staal’s strong emphasis on form helped us appreciate ritual as highly rule-governed, stylised action, irrespective of conscious reflection. His great achievement was the innovative, meticulous reconstruction of the ingenious construction plan and ritual form of Vedic rites and mantra performances, or what he calls ritual syntax. He restricted syntax, however, merely to formal structures, excluding even physical sound (the major element of Vedic performance!). Mantra rites became mathematical patterns, while the body and sensory perception had no place in this formal approach, let alone the meaning aspect or emotive associations. Certainly, any ritual will always include a more or less skilful orchestration of verbal and non-verbal media or means of expression. But already this skilfulness and deliberate arrangement which a structural approach is able to discern raises doubts about the meaninglessness of ritual actions. According to Staal ritual differs from language substantially; it is something like the cry of cock-birds or Bach’s fugues. But as Ernst Cassirer (1944) pointed out, humans are “symbolising animals” (animal symbolicum). They will always perceive something “as” something whether they hear a cock-bird’s cry or Bach (Westerners will immediately categorise, for instance, the first as an animal noise and the second as music). This “as” is always shaped by cultural conventions. Neither ritual designers nor performers are free 8 Staal 1979; 1986; 1990; Humphrey & Laidlaw 1994; Michaels 1999; 2006.


221

from it, and it is highly unlikely that the designers have been free from any deliberation, intention or “cultural intuition” while “putting in order” (Greek syntax) their verbal and non-verbal material, i.e. while composing the ritual in this way and not in another way. What each individual performer does with it, is yet another question. It always puzzled me that the attribution of multiple meanings should be taken as an argument and proof that rituals are meaningless. It would rather seem that rituals are particularly meaningful, precisely because they generate such multiple meanings and subjective interpretations. George Thompson made a good point claiming several levels that are at work simultaneously in a given rite – such as cognitive, emotive and phatic features. He is right remarking, “We need to overcome the habit of assuming that a given utterance has only one function at a time”.9 Michaels tried to solve the problem by pointing out the “polysemy of rituals”.10 I think one can clarify this point in a more precise way. We can learn a great deal from Bhartṛhari’s syntactical approach to world and language and from Eco’s literary theory to resolve this point. In keeping with these theories I will suggest a structural, semiotic and performative approach to ritual syntax. The ritual material, if viewed in a syntactical way, will always impose restrictions on meaning production by the way the ritual elements are put together. Non-verbal ritual elements share with ritual language (and even ordinary language) physical, syntactic and symbolising features, and like language itself, or the words of a sentence, non-verbal ritual elements gain their semantic or symbolising power by the composite whole which is more than its parts. If we want to apply the linguistic term “syntax” to rituals at all, which I find indeed fruitful, we must take into account that linguistic signs will always consist of phonological, syntactic and semantic components, and that syntax and semantics cannot be separated from one another.11 So I would very much agree with Staalcritics that it does not make much sense to speak of syntax without the meaning element. On the other hand, Staal (1986) had a good point when he criticised social anthropology and cultural semiotics for having pushed the meaning aspect (sometimes) too far. Whereas formalist-structuralist ritual theories had a tendency to underdetermine meaning, semiotic-symbolist ritual theories showed a tendency to overdetermine meaning and find some “deep-structure” social sense and function that was only known to the researcher and not to the actual ritual participants (whose own religious codings the social anthropologists were rarely interested in). A more or less strong logocentrism pertains to both. I think it is necessary to overcome this logocentrism by expanding it to the corporal, sensuous, imaginative and 9 Thompson 1995: 32. 10 Michaels 2006: 261; 2007: 255. 11 Penner 1985: 9.

222

Annette Wilke

emotional aspects which belong to language and ritual as well. Or to be more precise, I do not think that a semiotic, meaning-oriented understanding of culture can be given up, but meaning has to be expanded beyond the intellect and functional pragmatics. Meaning is also created and inscribed by the senses and bodily movements. This pertains very much to ritual. Theories of performance and performativity resolve much of the contested issue of meaning and meaninglessness as well as other dualisms within ritual debate, such as doing and knowing, language and body, expressive and instrumental action, meaning and function, or textual and functional meaning.12 Rituals first of all communicate themselves. They create situative presence and illocutionary and perlocutionary (culturally bound and specified) expectancies and effects. Acting out a religious rite changes the frame of reference from a profane to a sacred sphere and actually converts, for instance, normal water into sacred water, or alcohol into the “water of immortality” (PKS 3.26–30). Such conversion is often achieved by speech acts which assert or declare water or alcohol to be a sacred substance. The PKS is no exception, making use of such linguistic devices which closely interact with the bodily and sensuous ones. Such networks of homologous forms (Eco) make up the rituals’ semiotics and performativity. As Wade T. Wheelock points out, ritual speech contains the same kind of utterance types as ordinary language, namely assertives, directives, commissives (prompts to a future action), expressives, declarations, etc.13 But there is one important difference according to the author: illocutionary speech acts in the ritual are not informative, but “situating”. Nothing new is communicated, but rather participation in a given situation is brought about. Ritual does of course not stop with language. Even the non-verbal features and gestures have highly communicative capacities and do something in the ritual agents and participants. Stanley Tambiah’s performative approach to ritual accounted for both.14 He plumbed ritual language, but also paved the way for calling the whole performance, including nonverbal means, a “language” that communicates and does something. Tambiah’s approach was practice-oriented and cultural, and he was more interested in formal structures than Wheelock. Tambiah stressed the formal organisation of rituals (such as formality, rigidness, stereotypy, thickness, recursiveness, redundancy, etc.) and the necessity to integrate formal analysis and cultural description. In my assessment this approach still counts among the most fruitful ritual theories. Tambiah clearly favoured semiotics by defining ritual as a culturally constructed multi-media system of symbolic communication that is structured in sequences of words and actions. He saw ritual as an instrument for establishing cosmological networks 12 Sax 2004: 363–368. 13 Wheelock 1980; 1982. 14 Tambiah 1979.


223

(relating not only to religion, but also to law, polity, social organisation, etc.). The “message” to the participants, according to him, is not a specific social one, but a more encompassing meaning related to cosmology which the ritual as conventionalised action helps to make self-evident and unshakable. While only a few rituals communicate cosmology in a very verbal sense (as does the PKS to a significant extent), all rituals will invariably be embedded in a cosmology or culturally specific (and sometimes also clearly milieu-specific) world-construction that is embodied, perpetuated and stabilised by them, but may as well be transformed, recoded and inverted by ritual action. So the PKS does not only reflect Tantric cosmology, but also inverts Vedic cosmology and social practice by recoding Vedic mantras, combining them with Tantric ones, redefining Upaniṣadic nonduality and Vedic moral codes of purity and caste, and offering a new soteriology (considered to be more powerful).

3. Linguistic Theories and Ritual: Bhartṛhari’s Contribution Various linguistic theories (de Saussure, Peirce, Austin, Searle, and Chomsky) have inspired a great variety of influential ritual theories (structuralist-formalist, semiotic-symbolist, performative, and cognitive ones). Eminent representatives (e.g. Staal, Turner, Tambiah, Lawson & McCauley) approached the question of cultural description and formal analysis and particularly the question of function and meaning of rituals in quite diverse and controversial ways. The few attempts towards a “grammar” or “syntax” of rituals (Staal, Lawson & McCauley) have been highly formalistic, based on Chomsky’s generative “universal grammar”. While such abstract formalism may be a powerful explanatory model for discerning formal rules, it has some difficulty in accommodating the dynamic nature of rituals and the dimensions of body, senses, emotions and imagination. I want to elaborate a less formalist model of ritual grammar and syntax based on the “indivisible/unitary sentence view” of the Indian linguist Bhartṛhari who developed in the fifth century an early “universal grammar”. As it is well known Bhartṛhari belongs to the most prominent exponents of Indian grammar and linguistics, and he was as well one of the most eminent and original language philosophers of his time and even beyond his time.15 Discussing him (as limited as it will be in the context of this article), I claim that it is worthwhile to look at other cultures’ analytical categories which might provide helpful tools for our theorising. Grammar in India produced at a very early age something like a morphology or structural theory; people learned to think in structures and to abstract from the material. This is particularly true for Bhartṛhari.

15 Cf. Aklujkar 1993; Houben 1995: 1–26; Kelly 1993.

224

Annette Wilke

Bhartṛhari’s main work and only complete one to survive is the Vākyapadīya (VP), which means literally “Treatise on sentence and word [order]”. The VP was certainly conceived as a linguistic treatise by Bhartṛhari himself, although his Indian recipients generally read it as a work of philosophy, since it starts with metaphysics. In the linguistic part of the VP (primarily book 2), Bhartṛhari devotes himself to questions of syntax, a topic which was considered marginal up to his time and was therefore hardly given any treatment by earlier grammarians. In contrast, Bhartṛhari’s focus on language was entirely syntactical. Like today’s linguistic scholars he recognised that in language it is primarily not words that are understood, but sentences. Just as in the conventional use of language, letters do not have any meaning of their own and only produce words in combination, so it is with words too.16 The meaning can only be determined in connection with other words, i.e. it is the complete context that decides what is actually intended and creates meaning in a strict sense. The sentence is the primary meaningful and “indivisible” linguistic unit, whereas words are only subordinate, secondary linguistic units.17 Basic to Bhartṛhari’s “unitary sentence view” is thus not a fractified approach to single words and phonemes, i.e. primarily not a morphology, but a rigorously holistic approach. He supplemented Indian linguistics by the important insight that single words attain meaning only within sentences. Not singular words, but syntax, accords that a word is understood. It can be understood only by relation. The communicative function of a word therefore lies in the relational whole of the sentence and the meaning-whole expressed by it. I propose this theory can be applied to rituals, too. Not single ritual elements nor even sequences make up the sense of the ritual, but the whole of the rite, which I suggest to be comparable to a sentence. Like in language, the communicative function of a ritual is not found in singular ritual morphemes but in their relation to each other and their relation to the ideas the ritual wants to express. The scope of application is irrespective of whether we want to understand ritual morphemes physically (gestures, garments, decors, linguistic elements, music, artefacts, etc.) or formally (stereotypy, recursiveness, redundancy, reduplication, seriality, transposition, citation, abbreviation, substitution, etc.). Morphology by itself won’t be really helpful. We can make a list of body movement and the senses, particularly the voice, in action, we can list propositional and non-propositional ritual language and sacred formulas, vocal and instrumental music, etc., and extend our list to important elements like ritual space and time, ritual-specific artefacts (flowers, fire, food, etc.), and not forgetting imaginative devices such as guided visualisations and meditations, etc. The list would be long and virtually never ending, nor will there be much chance to get a set which fits to every rite. Even if such a wonderful set 16 See, for instance, VP 2.36 and Vṛtti. 17 VP 2.55ff., 2.143–145, 2.413–431.


225

should ever occur, it would be quite differently used. A syntactical approach has the advantage of asking how the elements fit within a given rite, at which point which element is used, whether the body and all senses are involved, or certain senses excluded, what stimulating elements are prevalent, for instance, speech acts or music. Only the mutually interactive relation of all these elements makes up the ritual and conveys something about its deep structure and its meaning. Similarly, it is not enough to regard only singular formal features, such as recursiveness, etc., as syntactical. Syntax has to do with the relation of elements to each other and to the ritual as a whole, or in other words, syntax has to do with the meaningful combination of structural units. I therefore propose that a syntactical approach will have to see the whole ritual or meaningful ritual units as a sentence. The major departure must be the entire ritual or at least perceivable ritual cluster(s). This also means that the segmentation of structural units will always have to take them simultaneously as meaning units. Here we are with the second point where Bhartṛhari’s theory is of interest. His approach to language was emphatically syntactical precisely for the meaning’s sake. The mysterious way meaning occurs was one of his central questions, and I do not think that we can do without inclusion of the meaning aspect even regarding ritual activity, if we don’t want to tackle it in a rather reductive and castrated way. Bhartṛhari may offer some solutions which add new ideas to our contemporary ritual discussion. Syntax is to him inherently about meaning resolution. One of his most interesting contributions to the ritual debate is his formalised approach towards meaning. As a grammarian he was less interested in contents (semantics and hermeneutics), and more in morphology, semiotics and structures (i.e. in meaningbearing units). Closely related to this was his interest in how meaning resolution takes place at all. The main concern of early grammar traditions was the material of language: letters, etyms and word formation. The sequence of words in a sentence was considered natural and self-evident. With Bhartṛhari we encounter a shift of interest from linguistic atoms to linguistic sequences and structures. Nevertheless his starting point was also the ever-returning fundamental question of Indian linguistics: what links sound and meaning? The major question was now: if language is mainly perceived via structures, the question arises, “how does one recognise structures?” So Bhartṛhari regards the central question in linguistics as, “who or what actually decides whether a context is ‘complete’ and whether a real meaning has been communicated?” Bhartṛhari calls the communicative element determining that a complete unit of sense has been communicated and an actual meaning is present, a sphoṭa.18 This term has been interpreted in a number of different ways due to the – at first sight bewildering – fact that Bhartṛhari sometimes refers to the acoustic aspect as sphoṭa, 18 VP 1.50ff.; see also 1.44–49 and 2.13.

226

Annette Wilke

while at other times to the meaning aspect, i.e. the word as meaning-bearer. Thus, the sphoṭa appears to denote at one time meaning and another time sound, or at one time the universal or class notion and at another time the individual substance. An etymological analysis can, however, clarify what is meant, more or less. In the colloquial sense, the root √sphuṭ from which the term sphoṭa is derived means “to burst, pop open”. So the term can stand for such things as breaking open a nutshell. The figurative (and also more common) sense is “to be (self-) evident”. So a sphoṭa is something that “causes bursting”, “opens” or “makes evident”. The linguistic sphoṭa therefore denotes the immediate resolution of meaning upon hearing a word or sentence. In Bhartṛhari’s system the sphoṭa refers to a function of understanding that has to be postulated, an integral and irreducible Gestalt, which only exists in the consciousness.19 It is important for Bhartṛhari that the information constituting the sphoṭa itself is not in any way contained in the words, nor can it be inferred from them. Only when the integral idea of an all-encompassing sense appears between and behind the words can one say that the sentence has been “understood” and the meaning has revealed itself.20 Bhartṛhari’s postulate is thus a basically integral vision in all of our understanding. He localises linguistic sense not in individual words or phonemes, but in the whole of a linguistic utterance. A sentence is only understood when the sphoṭa occurs, i.e. a holistic form of thought that encompasses the whole and sees more in the whole than the sum of its parts. This is how ritual is conceived as well. We perceive actions and simultaneously some encompassing sense in and behind them. Similar to the linguistic sphoṭa which refers at once to the chains of sound, to the object referred to and to the mental image and cognition, ritual transports images of world order, sacred power, efficiency and so on. I recall that the sphoṭa has for Bhartṛhari a material dimension (language sounds) and an immaterial one (meaning), it is both the physical sign which reveals itself to sensory perception as well as the signifier connoting something beyond the sound material. This means that Bhartṛhari does not strip language from its physical substrate, even when he speaks of concepts and mental forms. Language to him has bodily (vaikharī), discursive (madhyamā) and non-discursive integral “seeing” or “visionary” (paśyantī) aspects.21 In the latter the former are melted into one and virtually present. This visionary level of language corresponds with the sphoṭa. Unlike Europe, India never gave up oral-aural communication systems even after developing written traditions. Bhartṛhari’s language paradigm must be seen in the context of a performance culture and has therefore more affinities to ritual than the linguistic theories of the western scholars mentioned above. So he will conceive sentences as complex forms composed of 19 VP 1.44–66, 1.76–94, 1.106, 1.119, 2.143–145, 2.425–426, 2.440, 446, 3.52–57, 3.160. 20 For a pictorial illustration see VP 1.53–54. 21 VP 1.134 and gloss; see also ed. Rau 1.159–170.


227

clusters and chains of sound patterns that are arranged in a certain way and form a context with each other. The sequence of arrangement may differ, but invariably the sound sequences must be understood as something beyond sensory stimulation that it is understood as a sentence. Likewise we could say: it is only when the collection of gestures, movements, words, flowers, etc. are comprehended as a mutual context that they will be understood as a ritual. According to Bhartṛhari none of the singular elements of the sentence can explain the sphoṭa; they point to the sphoṭa, but the sphoṭa is not in them or immediately dependent on them, it is rather some third space and happens at one blow. Bhartṛhari speaks of an intuition or visionary capacity (pratibhā)22 which is formed by instinct, memory, practice, custom, habit (and includes also clairvoyance, yogic insight, etc.).23 Likewise ritual and ritual meaning will be recognised due to intuition or innate know-how awareness, instinct, tradition, practice and custom. Meaning resolution for Bhartṛhari has always something miraculous. It cannot really rationally be explained why this particular word or this particular sentence produces this particular sense, and yet one is absolutely clear about it (VP 2.144). This is why he postulates (prereflexive) intuition and conceives the sphoṭa as a necessary intermediary holistic link. According to Bhartṛhari, a wholeness mediating between sound and object must exist for several reasons, including this one: a word or a sentence is still comprehensible even if one does not hear the whole word or the whole sentence. The physical sound of the language, or the chain of sounds comprising a word, is only the contour or the outline that indicates or suggests the word idea or “Gestalt” that is actually meant. So too, a comprehensive ritual sequence, such as the diagram-worship in the PKS which is only part of the daily performance, communicates (if the cultural knowledge is there) that we have to do with a Tantric practice of a specific deity and her powers, with a way of ordering space and time, with a vision of the cosmos pervaded by divine dynamic powers, and with a firm conviction that these powers can be transferred to and be incorporated by the ritual agent. This actual word “Gestalt” or ritual “Gestalt”, i.e. the word or ritual in the consciousness, does not possess any physical dimensions, but is and always remains a whole. This whole is only translated within the spoken word into a linear sequence of defined sounds,24 or only manifested within the ritual of linear sequences of relationally defined ritual morphemes. The mutual transfer of word sound and object, or of ritual action and objective, is based on the efficacy of this one integrative form or sphoṭa, which is neither sound nor thing, but rather the unitary (mental) image embodied by and arising with the word and the sentence – or vice versa with the ritual cluster. The presence of a sphoṭa in the 22 VP 1.38, 2.117, 2.143–147, 2.151. 23 VP 2.117, 2.149f., 2.152. 24 VP 1.44–51, see also 1.93–97, 1.110 and the Vṛtti exposition of 1.23 which expounds words (linguistic sounds), meanings, and their relationship.

228

Annette Wilke

sound or in the physical ritual action gives the sensory material the character of a message and transforms it into a sign. Or conversely, the presence of a message in the sensory material requires a certain sensory contour and transforms it into a mental “form”. The sound/the ritual action represents the sphoṭa in the speaking/ritual process, and the sphoṭa represents the object/the meaning. So double transfer takes place interlinking the positions – a) the word/ritual element is essentially connected with the object/the meaning and b) the word/ritual is accidentally allocated to it – in the concept of sphoṭa. This universal process of meaning resolution takes place, irrespective of whether the sentence is correctly understood, misunderstood or interpreted differently at different times. Bhartṛhari thinks first of all of correct understanding. This is the achievement of grammar. But he also accounts for historical and cultural change and contingency. From a very early date he regarded human beings as cultural beings. Bhartṛhari’s theoretical framework is apt to account for both the “self-referentiality” of rituals – i.e. their illocutionary power to create intense situational presences of different kinds (e.g. new status, deities, social belonging, emotional rapture) by stimulating somatic attention, sensory engagement and mental focusing – and for a “reflexive holism” regarding semantics.25 In Bhartṛhari’s system, however, this reflexive holism is grounded in a metaphysical framework. On the one hand, he prefigured the linguistic turn by claiming the indissoluble interlinking of language, thought and reality.26 On the other hand, he remodelled old Vedic ideas about the Brahman by claiming an ontological “language principle” or “word reality” (śabda-tattva), also called “Communicative Brahman” or “Word-Brahman” (śabda-brahman), to be the non-dual, dynamic basis of fractified existence (VP 1.1–2). Bhartṛhari seems to conceive this language principle as a global form of sense or universalised sphoṭa. The unitary sentence view is extended to the whole of reality. The basic idea associated with the śabda-tattva (elaborated in the commentary supposed to be his auto-commentary) seems to be: All our understanding is linguistic and this linguistic source is not static, but dynamic, self-expressive and self-generative. It communicates itself by taking on perceivable forms, i.e. by embodiment. Although the global form of sense is actually one and not many and without time sequence, it can only manifest itself in a plurality of forms and splintered in time and space. Of particular interest for our discussion is that Bhartṛhari conceives his language principle to be replete with an intrinsic inner dynamism and creativity which is productive of world-making. Although the principle is an indivisible one, it inheres “powers” (śakti) that account for its exteriorisation and multiplication, or in Bhartṛhari’s terminology, for its transmutation into plural temporal existence and into the division of experiencer, object of experience 25 Both have also been called for by Lawson & McCauley 1990: 138–169. 26 VP 1.123, ed. Rau 1.131.


229

and act of experience (VP 1.2–4). I mention only those powers which Bhartṛhariresearchers conceived as particularly important ones, namely temporal power or sequence,27 relation or inherence,28 means of expression and imagination.29 As elaborated below, I regard these four linguistic elements also to be basic syntactical categories of ritual and good operational keys to work with analytically. Even Bhartṛhari’s ontological framework can be “translated” into ritual contexts and ritual syntax. The first thing which comes to mind is the ritual designers whose idea is embodied in the ritual and singular ritual elements. Consider a rite de passage, for instance the initiation rite in the PKS. The idea of a complete change from a human frame to a divine, cosmic body and mind must be externalised or embodied so that it becomes manifest, formative and socially effective. This is done by a plurality of ritual actions that are indicative of a ritual code-switching (from the Vedic ritual world to the Tantric ritual world) and of a change from normal social behaviour to a secret society of elite free-thinkers who are allowed to transgress the normal social codes. The adept does something divine by eating a piece of meat soaked in alcohol, etc. This is only possible within the non-Vedic, Śaiva-Śākta-Tantric cosmological context. Or speaking in Bhartṛhari’s terminology it is only possible by the powers inhering in the Śaiva-Śākta-Tantric “global form of sense”, of which each ritual element is a fractified “transmutation” into space and time. Vice-versa, each ritual element points thematically or a-thematically to this Whole of Śaiva-Śākta-Tantric cosmology, habitus and culture. So I suggest that Bhartṛhari’s “global form of sense” may be understood as the larger cultural context (Lawson & McCauley) or the cosmology (Tambiah) in which the ritual is always embedded. But there is also a more formal way to understand Bhartṛhari’s “language principle” in terms of ritual syntax, namely as ritual principle or an ontology of ritual which would intrinsically include self-communicative, -expressive and -generative “powers”, or in modern parlance performativity. In fact, this more formalist understanding is actually closer to Bhartṛhari’s argument of a metatheoretical global form of sense or a universal grammar and syntax which can explain all sentences (= all rituals), and is yet independent of them. I should add that both these “ritual translations”, the cultural one and the formalist one, are to be found in Bhartṛhari’s ontological thesis, and are closely interlinked therein. Regarding culture, his example is the Veda. The Veda is one and yet appears in four branches which again are divided in many different recensions that are recited in again different ways (VP 5 and commentary). The concrete Veda is but a “sign” of the primordial revelation which is beyond empirical perception and exists only in the mind. All later post-Vedic literatures, the combating views of philosophy, the 27 Coward & Raja 1990: 38–39. 28 Bronkhorst 1992: 62. 29 Isayeva 1995: 107–108.

230

Annette Wilke

distinct sciences and arts, the sacred lore, the ritual cultures are pluralist transmutations of the original Veda (VP 6–10). They must be necessarily pluralist, because already the earthly Veda reveals pluralist discontinuity. In Bhartṛhari’s view revelation is dynamic and processional and humans have substantial part in it. It is the more formalist features which I want to explore further, i.e. the inherently dynamic and self-communicative powers like temporal sequence or means of expression that belong to the global form of ritual syntax. The pretention is that we need to envision ritual syntax on a more basic level than has been done so far and introduce more basic categories to get the rules or forces that are at work on that formal structures like repetition and substitution to relate to each other and make up a ritual whole. The four syntactical linguistic “powers” of sequence, relation/ inherence, expression and imagination, substantially constitute what I would call “basic syntactical categories of ritual action”. They do not so much help to delineate ritual from non-ritual action, but they do help in envisaging ritual syntax in a holistic way (not stripped of meaning aspects) and offer a perspective for conceiving how form and content are interlinked and produce creative ways of world making. Although basic indeed, they are of high analytical value, as I want to show below in my analysis of the daily rites in the PKS. But of course, we are always necessarily confronted with transmutations in time and space, when we combine formal analysis and cultural description. The PKS in fact confirms Bhartṛhari’s Veda-argument and the fractification of revelation in space and time; it presents itself as transmuted Veda and contains a pantheon of deities that mirror transfers and transformations of earlier Tantra pantheons. It is not by accident that Bhartṛhari introduces the “power of time” immediately after his proposition of a (timeless) language principle and its inner dynamism (VP 1.1–3). Time is the “wire-puller” of the universe; everything is controlled by “the string of time” (VP 3.9.2–3).

4. Basic Syntactical Categories of Ritual Action and the “Test Case” Paraśurāma-Kalpasūtra 4.1 The Power of Time and Sequence The first mentioned and maybe most prominent power inhering in the śabda-tattva is time (kāla); this power is closely related to sequence which any action, any sentence structure and any form of worldly existence is bound to.30 Bhartṛhari regards kāla as particularly important because of its function in causing the bursting forth of worldly phenomena and making the language principle or the one whole (exist30 VP 1.3 and commentary; see also VP 3.9, particularly 3.9.2–4,9,14–15; Coward & Raja 1990: 38f.


231

ing only in consciousness) appear as many. Time allows manifested, successive sequence and has the capacity to push change, to divide, fragment and segment, and thereby to control and measure objects and activities. Bhartṛhari associates the temporal power with the six temporal modifications or actions seen in the world (something is born, exists, changes, increases, decreases, is destroyed), with control over the running of the world, with differentiating and sequencing things, with phenomenal existence and diversity as such, with cause-and-effect processes, and of course also with syntax and with the verbal tenses and moods. The words of a sentence unfold themselves in a certain order and temporal sequence. The rulegoverned sequential order makes the sentence comprehensible. The sequence itself is not a whole, but a “sign” of the whole. The sign is constituted by individual elements which are shaped by the rules of grammar and syntax. Any ritual will contain this principal force of time and sequence substantially, since any action cannot but unfold in time and space. Ritual action will always have a marked beginning and end (“framing”). It is rule-governed, composed of a sequence of smaller ritual units and sometimes limited to certain time-periods (daily, occasionally, once in a lifetime, etc.) The tenses and moods are the performances, body-movements, recursive repetitions, citations of past ritual actions, actions in hope of immediate or future results, etc. All of these are “signs” of a larger order, the “Whole” of the given rite. Sequence is the major dividing principle of the ritual script PKS. It encompasses ten chapters which reflect a very clear structure. This structure is ordered by sequences of worship (krama) that unfold in time, starting with initiation (ch. 1) and followed by the daily ritual cycles of the school’s major deities Gaṇeśa/Gaṇapati (ch. 2), Lalitā (ch. 3–5), Śyāmā (ch. 6), Vārāhī (ch. 7), Parā (ch. 8) who are worshipped one after the other. The script ends with an integral description of “the mantras of all deities” and special (occasional) rites (ch. 10). Already this simple order of ritual segments reveals something about the content: obviously Lalitā is the most prominent goddess, since three chapters are devoted to her alone, whereas the other divinities have only one chapter each. And indeed, this structural “intuition” is semantically confirmed: all the female deities are related to the highly erotic goddess Lalitā, the supreme sovereign of the “jewel island” (her ritual diagram representing the cosmos). Śyāmā, the goddess of music and dance, is introduced as Lalitā’s minister and major attendant (PKS 6.1–2); the fear-inspiring boar-faced Vārāhī as her commander-in-chief and the fierce judge (7.1), and the “supreme” Parā is introduced as Lalitā’s “heart,” i.e. inner essence (8.1). The elephant-headed god Gaṇeśa is of course the first to be worshipped just like in nonTantric contexts; in the PKS, however, he is tantricised by receiving the very same ritual sequence as the Tantric goddesses and typical Tantric ritual units like nyāsa (placing mantras on the body) or alcoholic Arghya.

232

Annette Wilke

The deity-cycles of worship are literally called “sequences” or “serial procedures” (krama). All the daily ritual cycles or sequences have exactly parallel structures (fig. 1). They mirror and “echo” each other. On the one hand, they certainly reflect the most common Tantric practices of this epoch (fifteenth–sixteenth century) which are compiled in the famous ritual manual. Many of its assertions and ritual descriptions are direct or indirect citations and glosses from earlier and contemporary Tantra text books. On the other hand, the PKS designs a new ritual universe. By means of stereotypy, recursiveness, reduplication, seriality, looping and (ritual) citation the PKS establishes a link between deities which are in other contexts worshipped alone, and thus generates in the ritual agent a sense of their belonging together (the pairing of Śyāmā and Vārāhī is also found in other Śrīvidyā literatures, for which the PKS may have been the initial inspiration). At the same time, the parallel structures contain slight formal changes which mirror and typify the character, role and eminence of the deity of the given cycle. This is reflected also in semantic transpositions, the extent of external and internal worship, and the dismissal rites. A closer look at the ritual form of each deity cycle of the PKS (fig. 1) reveals some interesting manipulations and transmutations within their common sequential structure (sunrise ritual – except for Vārāhī! –, preliminary rites, Arghya rites, pūjā and their subdivisions). Although the Gaṇeśa-worship (ch. 2) contains exactly the same ritual units as Lalitā’s worship (ch. 3–5), his ritual sequence appears in abbreviated form (one aphorism for each unit), whereas the ritual cycle of Lalitā shows most signs of redundancy and elaboration (greatest number of aphorisms for each unit). I should add that the length and style of the “aphorisms” differs substantially. Some are short and terse. Many are highly technical. The mantras are always presented in encoded form which produces a poetic subtext besides the technical information. Other aphorisms are highly narrative and descriptive. Some are very poetic and full of metaphors and metonomies, like for instance the description of the sunrise ritual of Gaṇeśa (PKS 2.2,4) which makes a “good starter” to “tune into” the ritual and arouse emotional participation. Here it is not abbreviation, but aesthetic elaboration of the visualising practice of showering the body with “the water of immortality” and creating a mental image of the deity. Such “metaphorical density” as well as the technical information in abbreviated and encoded form increases in the chapters on Lalitā and the other goddesses. Gaṇeśa remains the most “exoteric” of the group of major deities. This is mirrored formally and rhetorically: his chapter 2 does not contain any encodings (in contrast to PKS 10.17). After the second chapter we encounter increasing complexity. All three chapters of Lalitā worship contain a large amount of encodings, while the mere sequential structure of ritual units mirrors exactly that of Gaṇeśa (see fig. 1, above). This changes in the other goddesses’ cycles (fig. 1, below).


233

Even though the cycles of Śyāmā, Vārāhī, Parā are again perfect reflections or “echoes” of the Gaṇeśa and Lalitā worship, they contain within the “preliminary rites”, “Arghya rites” and “Pūjā-rites” transpositions, rearrangements, permutations, insertions and recursive embeddings or framings which grow in number correlative to the chapters, i.e. succeeding deity-cycles. While in the Śyāmā-chapter there is only minor transposition by inserting the ritual sanctification of the body (B) in between the ritual sanctifications of the place (A), we find in addition to that in the Vārāhī and Parā cycles a growing interface with the Arghya rites – very important ritual units which ritually cosmicise water (C) and alcohol (D) – and the Pūjā rites (E, F) (see PKS 6.6–17, 7.4–17, 8.6–18). If formalised and abstracted we get structural patterns that equal the ones that Staal worked out so beautifully for the Vedic sacrifice. The straight sequence of the ritual units in the Gaṇeśa and Lalitā worship (A, B, C, D, E, F) receives in the later cycles increasingly complex transpositions: ABA, CDEF (Śyāmā), ABCA, BCB, DED, EDEF (Vārāhī) and ABCDABEFA (Parā). The rearranged order may be taken as a sign of the complexity of composition and beauty of form. As it will turn out, the semantics will be a vital reason, too. I presume also practical reasons; a worshipping practice that integrates the ritual units of the different deities into one. However, such an integrative worship may be difficult in actual practice considering the elaborate and exacting ritual elements pertaining to each deity (more elaborately see fig. 2). PKS 10.55 suggests worship spread throughout the day, i.e. each deity is allotted a particular time unit: Lalitā is worshipped in the morning, Śyāmā in the afternoon, Vārāhī at night and Parā in the brāhma-muhūrta shortly before sunrise. This means that pragmatic reasons (integrative practice) are not enough to explain the rearranged sequences. We will have to go into semantics, the deity-specific imaginations and inherences among them, to get a clue. As will be seen in the next two chapters, there is creative interface and often highly artful interplay between formal and semantic transpositions, permutations and discontinuities, and a close interaction of symbolic forms and expressive means.

234

Annette Wilke


235

236

Annette Wilke


237

238

Annette Wilke

4.2 The Power of Relation or Inherence and its Interaction with Sequence Already in the foregoing chapter I have hinted at the power of relation and its creative interaction with sequence. The goddess cycles are closely interrelated by common structures, and at the same time differentiated by increasingly complex permutations of these structures (fig. 1). Semantically, all goddesses are related to Lalitā, and yet have each their very own individual traits. This is already perceivable by the form of their common and yet different sequences. Vice versa there are semantic homologies to the formal structures. This complex network of homologies makes up the “Whole” of the daily rites. It is, as I want to elaborate a bit more, a meaningful whole. Inherence is one of the major components and perhaps even the central one of an “indivisible/unitary ritual view”. Whereas the power of temporal sequence divides, fractifies and measures, the power of inherence and relation unites. Both belong to the “ritual principle”. Regarding Bhartṛhari, Johannes Bronkhorst conceived the samavāya-śakti, “the power of relation” or more literally “the power of inherence”, to be the most important power. Bhartṛhari himself likely saw it this way, too, because the unifying capacity of relation mirrors the śabda-tattva and the sphoṭa. He uses the terms relation (saṃbandha) and inherence (samavāya) synonymously, i.e. views relation as a force that brings about very intimate connection or union (saṃyoga) and inseparable inherence (samavāya).31 Bhartṛhari emphasises the potency of relation as a category of meaning of its own which does not rely on verbalisation.32 It is “the power of the powers”.33 Relation is so fundamental because it assists the other powers; it reveals identity and difference while itself transgressing identity and difference. Bronkhorst quotes VP 3.136–139, wherein it is said: “[...] elle [samavāya-śakti] est un pouvoir qui assiste les [autres] pouvoirs; elle est au-delà de différence et identité, et existe d’une façon tout à fait différente. Selon la tradition [qui vient] des anciens, la relation universelle, qui dépasse les caractéristiques de toutes les catégories, favorise [ce samavāya].”34 Bhartṛhari’s interest in the Saṃbandha chapter35 is to plumb philosophically the relationship of inherence between word and meaning, sound and sense, signifier and signified. Signifier and signified “fit to each other” like the sense organs and their respective sense objects.36 31 32 33 34 35 36

VP 3.3, particularly 3.134, ed. Rau 3.3.6; see also Houben 1995; Bronkhorst 1992: 62–63. VP 3.132, ed. Rau 3.3.4. VP 3.133, ed. Rau 3.3.5. Bronkhorst 1992: 62–63. VP 3.129–216, ed. Rau 3.3.1–88. VP 3.157, ed. Rau 3.3.29.


239

Concerning syntax, relation is again the most decisive principle at work in making an interrelated, homologous whole of the parts. Relation may be called the central sentence principle. It is the power that produces the cohesion and coherence of the sentence, since it guarantees that the parts of a sentence connect with each other. The sentence’s predicate is constituted by the inherent, mutually effective connection of subject, verb, etc. Relation is nowhere verbalised (except indirectly in the genitive case) and yet, it is very much there as the invisible power of networking. The power of relation and inherence is one of the most central and decisive for ritual syntax as well. There is not only simply sequence of certain ritual elements, but there is an intensely interactive relation among them, an inner, inherent and inseparable relation. Such relational inherence must be there. The inherent networking is the power which allows the singular ritual elements mutually to interact, “explain” each other and make of singular actions a ritual whole. It is the very “construction plan” of the rite. And precisely the relational power makes this one ritual into a specific one which differs from other rites. So you will find pretty much the same kind of ritual units in the common South Indian Śrīvidyā worship (or even other Tantric pūjās) as in the PKS. But the common South Indian Śrīvidyā worship won’t contain the “special Arghya” (with alcoholic liquor) or leave out sexual intercourse in the Śakti-pūjā, and it will rarely contain Vārāhī-worship or only in a domesticated, “purified” and highly spiritualised form. Instead you may find a substantial increase of nyāsas, the “placement” of mantras on the body and ritual paraphernalia to make them “divine” and create a sacred ritual universe. In fact, the PKS already incorporates an increased number of nyāsas compared with older texts which contain only the “six-limb-nyāsa”, and in the later PKS tradition the number increases even more. So I do not pretend that each and every unit must be always there. Humphrey & Laidlaw (1994) showed regarding the Jain pūjā that this is not the case at all. But at least one offering (and be it only a mental one) must be there to make a pūjā a pūjā. Tantric pūjās are generally more strict and rule-oriented than the Jain pūjā of lay persons described by Humphrey and Laidlaw. Certain elements, such as the sanctification of the body by placing mantras on it and succeeding diagram worship, or at least mantra practice containing one or more “seed-syllables” (bīja), must be invariably there to make a Tantric pūjā a Tantric pūjā. The number, form and content of the nyāsas and bījas may differ, and like the “root mantras” the diagrams will always be specific to a specific deity. But what is always decisive is their relation to each other and to further units of the given rite, i.e. their “fitting in” with the whole. The presence or absence of certain units or their increase or decrease will often change a ritual substantially. An analysis of the interior “networking” of ritual elements in a ritual or in ritual clusters will always be informative about the formal and semantic “construction plan”.

240

Annette Wilke

Regarding the PKS a look at fig. 1 will reveal that the formal differences are most explicit in the beginning and the end of the deity cycles, i.e. the framing. Śyāmā, for instance, has no explicit dismissal rite, but specific concluding rites (PKS 6.36–37) and separate rules of behaviour at the end of her chapter (6.38–39). Formally, these rules should actually belong to the teaching of Tantric ethics during initiation (or the 10th chapter). They have no ritual function in the Śyāmā-worship at all. They are, however, semantically typical for Śyāmā and “fit in” well with her image. The rules prescribe an attitude of veneration for singers, dancers, musicians, erotic things and prostitutes. PKS 6.35 identifies Śyāmā in a highly encoded language with the (low-caste) goddess Mātaṅgī (who is sometimes associated with cult-prostitution). So the rules, which break the perfect form of sequences, can be explained by the power of inherence and semantic relation which was obviously more important in the construction plan of the PKS author than formal correctness. The most obvious transposition in the ritual sequences is Vārāhī’s time of worship at midnight, in contrast to the other deities whose worship starts at sunrise and mirrors (but tantricises) the Vedic-brahmanical Gāyatrī-worship in structure as well as content (even including a Vedic mantra for Gaṇeśa and a tantricised Gāyatrīmantra for Lalitā, see fig. 2,1). Vārāhī’s midnight worship is very fitting for her “very dark” and terrifying character.37 She is invoked to strike and kill, to drink the blood and sperm of enemies and bring them under control, but protect the Tantric “family” of initiates and bestow upon them success. Vārāhī is the only one whose ritual cycle includes blood offering (rice mixed with blood within the ritual unit of bali-sacrifice). Vārāhī (herself an old Yoginī) reminds us most of the wild goddesses and Yoginīs of the early Kaula Tantra of which the PKS is the heir. But the PKS-pantheon is much more elegant. Characteristically, Vārāhī has a subordinate role and is “royalised” as commander-in-chief of the beautiful and elegant queen of the jewel island. The effects of Vārāhī’s worship are omnipotent controlling and commanding power (7.38). The overall aim of the Vārāhī-ritual seems to be more profane than spiritual, whereas the other ritual cycles are rather spiritual than profane. The latently military language and images in all cycles perhaps reflect royal clients, such as the court of Tanjore who sponsored Umānanda’s PKS-extension, the Nityotsava. As already seen, the Parā cycle has the most complex formal structure (A-BCDA-BEF-A). The reasons are semantic. The sanctification of the place (A) coincides in Parā’s case largely with the sanctification of the body (B) and includes visualisations of the yogic body (cakras), in which the 36 principles of the cosmos are mentally absorbed and re-established (E, F). This merely mental procedure is an internal substitution of the external pūjā of the other deities. Parā as the “heart” or 37 See particularly PKS 7.8, 7.13, 7.26–28, 7.34.


241

supreme form of Lalitā (and at the same time the inner light and consciousness of the worshipper’s own self) has no physical image and diagram. She receives almost mere mental worship in the form of meditation verses and esoteric manipulations of the yogic body. Instead of a geometrical diagram the cosmos itself is her “throne” or seat. In the other cycles the physical and mental creation of the ritual diagram (cosmogram) as the deity’s throne belongs to the “sanctification of the place”. But in Parā’s case the real cosmos is her “diagram”. The cosmos is not graphically reproduced, but conceptualised and abstracted into the 36 cosmic principles (the tattvas of the Sāṃkhya plus the twelve extra principles of the ŚaivaĀgama and Tantra which relate to the more esoteric and interior “first” creation, the universe of language). So Parā shows very distinct features which make her different from the other goddesses. She is the most esoteric and literarily speaking the innermost substance, not only of Lalitā, but also of the worshipper. In her ultimate form she transcends the cosmos and gender difference. As light and reflection she is one with the highest godhead, Parama-Śiva. This is formally symbolised by the permutation and interpenetration of their bīja-sounds. In contrast to the other goddesses, Parā has only one singular seed-syllable (SAUḤ) just like Parama-Śiva (HSAUṂ); Śiva (HA) is contained in Parā’s mantra (SA-AU-Ḥ), just like hers in his (see PKS 10.26–27). Parā is most clearly associated with soteriology and internal yoga. Ritual turns into “gnostic”, non-dual contemplation. Here we find the most explicit cognitive merger of worshipper and deity (which is virtually there in each deity cycle). This also explains why Parā has no dismissal rite. Ritual incorporation of the supreme power Parā means recognising the supreme light as one’s innate limitless nature. Emancipation and freedom while living are brought about, but strictly speaking nothing new is produced. The samavāya power of uniting, which cognitive science would speak about as cognitive blending, is thus very present in the daily rites, formally, semantically and pragmatically on many different levels. I recall the parallel sequential structures and the recursive, almost identical ritual units/morphemes, but also the specific nature and yet inner identity of the goddesses, the non-dual ideology, the participation in the divine powers’ omnipotence, omniscience and control, the mystic merger with the deity and the inherent oneness of ritual agent and deity, etc. A major objective of the “left-handers” seems to be that they seek to unite flesh and spirit, body and mind, and actually do it performatively by their means of expression and their active imaginations. Already the observations above on the goddesses’ profiles have shown clearly how strongly the four basic syntactical categories of sequence, relation, expression and imagination interact. They must be seen as a continuum, and as mirroring echoes of form and content which reinforce each other.

242

Annette Wilke

I wish to add and emphasise that it is not merely form and structure nor the addition of semantics which account for syntax, but most of all relation and inherence, the putting together of ritual elements and activities in a certain way and their connection with ideas, cultural codes and imaginations which justify speaking of syntax at all. This means knowing the construction plan is not yet good enough, we need the knowledge of the contents which are associated with the construction plan and make it a ritual whole. I am not yet speaking of agency. In fact, I leave out the agents, and merely look at the ritual script itself. But the ritual script, in my case a religious text, reveals that there is intrinsic meaning connected with each ritual element. Similar to the post-modern claim of the authorless text and to literary reception theory which stresses the iconic nature of literary performance, I suggest for the time being to leave out the agents and their subjective interpretations and concentrate on the imaginations and meanings and their relation to formal structures which the script itself reveals. The structure marks meaning units. Each and every ritual element is meaning-loaded. What the recipients do with it is quite a different question. But in any case, imagination will be invariably there – imagination of those who designed the ritual (regarding the PKS the legendary author Paraśurāma, introduced as great Kaula master) and the imagination of the recipients who perform the rites. Let’s therefore stick with the legendary author and his construction plan and mental mappings. He constructed a highly ordered ritual universe and used an extreme amount of expressive physical means and creative imagination. 4.3 Expression and Imagination and their Interaction with Sequence and Relation Expressive power or means of expression (abhidā-śakti) and imagination, fantasy or mental constructions (vikalpa, kalpanā) are the set of creative potentials selected by Natalia Isayeva in her discussion of Bhartṛhari.38 She sees the two powers of “expression” and “imagination” as the very key-concepts due to her emphasis on the pulsating dynamism, creativity and world-forming potency of the śabda-brahman, and she correlates the two powers to poetic speech and the creative activity of the poet – the “aesthetic counterpart” of the “power of words” in “the creative urge manifested on the cosmological plane of existence”.39 Isayeva contends: “Creation here is simultaneously regarded as ‘expression’, or even as a ‘manifestation’ of a hidden meaning, since the actual, cosmological aspect of the universe is no more ‘real’ than its mental aspect [...] The Brahman-Word (śabda-brahman) is nothing but a perpetual shimmering, perpetual pulsation 38 Isayeva 1995: 106–120. 39 Ibid.: 107.


243

of meanings, images, things. Elements themselves are interchangeable, they dissolve and melt away, they disappear only to be replaced by new ones; what remains is the energetic charge, which again and again brings along the same structure, the same set of elements, the same pattern of being.”40 “Language ‘thinks itself’, or more precisely ‘utters itself’, it is again and again engaged in its own interplay, in sending itself forth as [...] a perpetual throb which reproduces itself in a fixed, self-organizing pattern [...] This creative activity, which takes place on the universal, cosmic scale, is being replayed and re-created again on the level of separate mental structures of the ‘speaker’ and the ‘listener’.”41 Expression and imagination generate a “net of meanings”, a network of innumerable relations and semantic links. Isayeva’s notes on Bhartṛhari read almost like a perfect description of the PKS ritual expounded above and below. This may be explainable with the actual historic inspiration which the high Hindu Tantra received from Bhartṛhari. But the power of expression and imagination is not restricted to India; the categories can easily be transposed to poetry and ritual in every culture. Expression and imagination are important categories for explaining the selfcommunicating and self-generating potential of poetry as well as of ritual action, i.e. the performativity or performative power which poetry and ritual share and their “throbbing” creation of perfect form and hidden meaning in which inside and outside melt into one. Literary works live from the power of expression, from the choice of words, composition, style, rhetoric, figurative speech, rhyme, etc. by means of which reality is re-created, fictionalised and surmounted. Likewise, there is a “ritual rhetoric” and expressivity which depend on the sensory and verbal means that are applied, the way they are put together, “rhymed”, permutated, differentiated. On this level we find what has previously been called ritual morphology. But the term “expression” indicates that the elements are more than a loose and arbitrary collection. They are not just pieced together, but (many times very skilfully) composed. Their selection and the ways they are combined and related to each other makes them an expressive and meaningful composition, a creative staging and dramatic performance. As already hinted at, the PKS uses a great number of means of expression which form an artful composition: meaningful and (lexical) meaningless sounds, colours, fragrances, movements, ablutions, decorations, garments, highly expressive goddess images, abstract diagrams, yogic body schemes, and not forgetting the pañcamakāra (alcohol, meat, fish, parched grain, sexual intercourse) which affect the sense of touch and bodily awareness even 40 Ibid.: 109–110. 41 Ibid.: 112.

244

Annette Wilke

more than other ritual actions. Ritual compositions of any kind have potentiality and the power to trigger imaginations, appeal to or curb the senses and channel emotions. The kinds and degrees of stimulations or expressive means may differ, but rituals are always multimedia aesthetic events and sometimes even highly artful compositions. The PKS is aesthetically particularly strong and may be called a work of art, a “Gesamtkunstwerk” – both the script and the ritual it describes. This ritual affects to a high degree sensory perception and embodied participation as well as cognitions and emotions. It channels and controls attention and prevents tediousness and boredom by means of visualisations, complicated ritual manipulations, mantra permutations and highly symbolic codings, etc. However, there is need of a further element so that the creation of an artful composition can take place at all and that expressivity can manifest itself and be received by the participants. This element is vikalpa-śakti, the “power of imagination”. Language has the creative power to semiotise and fictionalise reality, and the literary productions will be actively moulded and further developed by the listeners and readers. Each ritual as well enjoys creative, world-making potency and at the same time it affects imagination. The Tantrics even know deliberate active imagination; they make conscious use of it in their rites. But imagination is actually a must in any ritual. It is the power of imagination which transforms ordinary water into holy baptismal water, wine into the blood of Christ, or in my Tantra example, wine into the “nectar of immortality” and into the goddess herself (see PKS 3.28). Imagination can be called a third space which connects (relates!) real-world phenomena and subjective mental space with one another and thereby creates something new (by mutual inherence!). This imagination will often be verbalised, particularly in such central ritual elements as the transubstantiation, “this is my blood”. In the PKS it is said that the pañcamakāra (alcohol, etc.) make the bliss of Brahman manifest (1.12). In such central ritual propositions, the “powers” of sequence, relation/inherence and expressiveness will always be combined with imagination which is in fact their deep structure and final base: It is imagination which allows alcohol consumption (which can go as far as heavy intoxication and delirium) to be perceived as the physical appearance of the deity or of possession by a god, and again it is by imagination alone that the beautiful young woman is the goddess Lalitā with whom the practitioner unites himself mentally, verbally and corporeally. Imagination is a major force that makes ritual meaningful and effective. I recall that the PKS starts with meaning, the school’s cosmology, doctrine, and ethics (PKS 1.1–12, 13–25), and this meaning-based approach is necessary as a plausible foil to unfold the ritual’s pragmatics. So the “rules of initiation” also contain the esoteric teachings on the school’s major practices (mantra rites and pañcamakāra) and their ritual effects: the mantras’ unimaginable power that brings about supreme command, the pañcamakāra’s capacity to make the bliss of Brahman manifest in


245

the body, the attainment of Śivahood (śivatā) and non-dual cosmic awareness. Only after this heavy semiotising of the ritual universe – and creating a corresponding frame of mind – are the three-fold initiation rites (śāṃbhavī, śakti-, mantrīdīkṣā) described, which in turn are the necessary preconditions for the performance of daily ritual. The initiation prefigures the ritual scene wherein we find complex relations of inherence between structure, means of expression and imagination, and a wilful merger of interior and exterior ritual. The three initiation rites involve a particularly high level of visualisations, i.e. of active imagination. Deliberate imagination is consciously made use of in the Tantric ritual in order to make the virtual more real than reality itself and create a physical experience of the mental world. In the śāṃbhavī initiation (PKS 1.35), the teacher mentally establishes on the crown of the disciple’s head the red and white feet of the god and goddess (in their love aspect), (Śiva-)Kāmeśvara and Kāmeśvarī, from which zone oozes the “water of immortality” and cleanses the disciple’s body that is now adorned by the guru. In the śakti initiation (1.36), the teacher visualises to imagine a wave of light (the supreme aspect of the goddess) in the disciple’s yogic body from the mūladhāra (near the genitals) to the brahmarandhra (the thousandpetalled lotus above his head). This procedure makes the disciple’s interior body become like a blazing fire that burns all sins. In the mantrī initiation (1.37–43), the longest and most complex one, the disciple is showered with sanctified, perfumed water while standing in a cosmogram, into which HSAUṂ and the letters of the alphabet are written. After being newly clad in luxurious garments and adorned with a perfumed flower-garland, the disciple’s body is covered with the alphabet sounds (which are known as the primal matter of the universe), i.e. he attains a cosmic body. Thereafter he becomes active, chooses/receives a new name, learns the mantras and receives the secret teachings about the Tantric practices, rules and regulations. In the daily worship which follows, it is the disciple who performs very similar imaginary manipulations like the guru during initiation: at sunrise he visualises his body showered with the “water of immortality” oozing from the crown of his head; during the sanctification rites he burns the old cosmos and his sinful body and replaces it with a pure mantra body; recitation of the alphabet and the bījas HSAUṂ and SAUḤ in the normal or permutated order belong very much to the esoteric veneration of Lalitā.42 Imaginative practices encode reality and convert it into something else, something extraordinary, and at the same time they allow active participation in the sacred cosmos. The Tantrics knew many means and ways to do so. It happens, for instance, by mentally and verbally “placing” the mantra-sounds on the body (sometimes touching the body parts) by means of which the body changes into another one, a divine one, a sounding mantra-body. A very important, typical Tantric pro42 See PKS 3.20; 3.24; 3.30; 4.3; 4.9,13,15; 5.3; 5.6,7; 5.9, 5.13.

246

Annette Wilke

cedure which involves several procedures of active imagination is to actually “create” the deity mentally to make it really “real” before the worship of the physical icon takes place.43 This can be done, for instance, by invoking the deity in the heart, making it part of the body by breathing exercises and cakra-yoga, make it a living being by very intensely visualising the deity’s meditation verses (which are generally highly poetic and linguistically pretentious), or by uttering the deity’s seed sounds in combination with visualisations. The procedure is often very complex. In the PKS all deities are “mentally created” and the most detailed description is in the Lalitā-chapter. It involves the “primordial consciousness” or the mental goddess image being invoked in the breath and the heart-lotus becoming literally transferred to the physical image by means of breathing out into a flower which is thereafter put on the icon or diagram (and removed, “breathed in” and taken back into one’s heart after worship).44 The aim of such creative manipulations is obviously to make the deity a living reality and connect oneself with the divinity not only mentally, but also make it an embodied experience. The ritual aim of the PKS is nothing less than becoming Śiva in “all one’s limbs” (PKS 10.50), i.e. embodying the god and his power physically and attaining emancipation while living (10.82f.). The correct, rule-governed performance of the ritual sequences (starting with initiation), but also the interior relationship of the ritual elements and the high expressivity of the rite(s), are decisive in achieving this aim. All of them will invariably be associated with imagination. This is reflected in just the formal structure of the more finely sequenced ritual units of fig. 2, a synopsis based on the Lalitā-krama. The units give quite detailed information about the means of expressions and imaginations in the script and will be of special interest regarding ritual syntax. As already stated, the PKS relates different deities to one another by means of parallel structures of worship. But more so than the mere sequences (fig. 1), the synopsis of the daily ritual cycles (fig. 2) will reveal that the parallel structures are broken by many blanks (-) or by certain ritual units which appear only for one of the deities or only for some, but not for all. This means, the more concretely we look at rituals, the more fuzzy and messy (or at best the more complex) will be their formal structure and the more need there is for semantic analysis. The blanks (-) in fig. 2 may be there for three reasons. The first reason is that the ritual element is actually there in actual performance, but not mentioned because the practice is known, i.e. part of the oral tradition and Tantric praxis. Such is the case, for instance, with the sanctification of the door (fig. 2, 2.1.1). The more elaborated ritual script Nityotsava, which is based on the PKS, fills in these blanks, i.e. prescribes door-worship for each deity. I have marked such cases with a special 43 See PKS 2.4, 4.1–3, 6.20, 7.19–24, 8.16–18. 44 See PKS 4.1 and Gupta 1979: 149.


247

sign (-.-); the list may be incomplete. However, sometimes even one and the same tradition disagrees about the blanks; sometimes the Nityotsava and the PKS commentator Lakṣmaṇa Rānade fill in blanks where the PKS commentator Rāmeśvara pretends that they must be there. Rāmeśvara is always very faithful to what is literally expressed. This is the case, for instance, in the Śakti worship (fig. 2, 4.4.3). Śakti-pūjā, i.e. the worship of a beautiful girl or young woman as human representative of the goddess, is there in all deity cycles (invariably at the end of the pūjā) except the one for Parā. But it can mean different things like mantra chanting, offering wine, alcohol, food and gifts to the girl/young woman, or bathing and perfuming her/them (without having intercourse), or finally actual sexual intercourse. Rāmeśvara restricts intercourse to the Lalitā cycle, the other authors have different opinions. A second reason for the blanks can be that summary aphorisms state that the worship of this or that deity should be done just like the worship in a previous deity cycle. Or they contain the information in extremely abbreviated form, because there was previously an elaborate description which need not be repeated. Such cases are found, for instance, in 3.1 and 3.2 (Arghya rites); I have indicated them with another sign (-“-). Within the Arghya rites we find most formal and semantic parallels regarding the deity cycles. Moreover, the Kaula-Tantric alcoholic Arghya mirrors exactly the ordinary water Arghya of the (“right-hand”) Āgama-Tantric traditions, i.e. a commonly known structure. But we find some extra ritual units which indicate what makes the Arghya Kaula-Tantric. These are the additional substances wine, meat, fish, and the highly symbolic linguistic features whose “hidden” meanings are typically Tantric imaginations: the grapheme “ī” (in the centre of the Arghya diagram) represents the head, breasts and vulva of the goddess (kāmakalā) and the A-KA-THA-triangle is the segmented alphabet spread around the “ī” in triangular shape (indicative of the vulva of the goddess as womb of the universe of words and meanings, i.e. objects). There is a third reason for the blanks which is the most interesting one because it is purely semantic and has to do with the specific theology and ritual imagination of the deity. Certain ritual units are restricted to only one deity, others are common only to two or to the three goddesses Lalitā, Śyāmā, and Vārāhī, but not to Parā. I would call such cases “structural units of metaphorical density”, because almost invariably they will be formal as well as semantic signals of some theological specificity attributed to the deity/deities in question. Consider this extract from the synopsis (fig. 2):

248

Annette Wilke

Fig. 3: Gaṇeśa Lalitā Śyāmā Varāhī 2.2.3 nyāsa with 8 language goddesses, alphabet 3.2.1 inscribing in Arghya diagram the “ī” (denotive of kāmakalā) 4.1.10 worship of 15 lunar goddesses, vowels 4.3.2 meditating the “parts of love” (kāmakalā)

Parā

–

3.20

6.11

7.7

–

–

–

6.19

–“–

–“–

–

4.9

–

–

–

–

5.16

–

–

–

In the simple table above we learn in a nutshell about Lalitā’s most eminent traits: her association with language and eroticism. The “eight goddesses of language”, who are verbally and mentally placed on the body in the Lalitā cycle, are nothing else but the phonetically ordered Sanskrit alphabet segmented into eight groups (the group of vowels a, ā, i, ī, etc. and the seven groups of consonants ka, ca, ṭa, ta, pa, ya, śa). Both Śyāmā and Vārāhī share this linguistic imagery, but in a “cruder” form. Their nyāsa consists of the unsegmented alphabet sounds (mātṛkā). Another pattern we find in the alcoholic Arghya. All goddesses, except Lalitā, share the graphic symbol of kāmakalā (explicitly mentioned only for Śyāmā), but the actual meditation of the “parts of love” (face, breast and vulva) is prescribed in the Lalitā cycle. The erotic element is most pronounced in Lalitā’s cycle. In her Gāyatrīmantra she is invoked as “the wet one” (PKS 3.5) and her ritual cycle is full of erotic imagery. Lalitā’s “powers”, the lunar days – personalised as Nityā deities – are worshipped with highly sexual epithets (“vulva”, “ever wet”, “full of passion”, etc.), but the same aphorism identifies them also with the fifteen vowels of the alphabet and proposes the sixteenth digit to be the “semen”, the visarga-sound “aḥ”. How to interpret this curious combination of alphabet and eroticism? What have language and sex to do with each other? The hidden meaning is: both are world creative. And with creation the power of time is at work (Nityā deities). So ritual expression in the PKS is highly metonymic. But this imagination makes use of traditional codes. In the Tantra the Sanskrit alphabet is commonly known as “mother” (mātṛkā). There is esoteric inherence of the tattvas and the alphabet. The alphabet is seen as primordial matter of the cosmos. The universe cannot only be thought of and conceptualised in 36 categories (tattvas), but these abstractions can also be heard in the alphabet recitation! It is the esoteric, sonic universe of creative mantra power. Reciting the alphabet (each of the sounds always nasalised, i.e. made “mantric” by means of the anusvāra) therefore leads to omniscience (PKS 10.21).


249

Some of the phonemes are particularly coded like the letter “ī” (the goddess’ “parts of love”) and the sound “aḥ”; the visarga “aḥ” discharges creation (male principle). Lalitā’s śrīcakra-diagram also has erotic and creative associations. It is a cosmogram containing Lalitā’s powers and more esoterically a symbol of the yogic body. At the same time it is coded with the most central, fanciful and emotionally pleasing imagination associated with this goddess: the jewel island. The description of Lalitā’s diagram as jewel island in the nectar ocean of immortality is one of the longest and most poetic aphorisms in the PKS (see 3.10). It is a persuasive and intensely sensuous visualisation of an abstract geometrical diagram and full of “metaphorical density” which is apt to inspire fantasy, aesthetic moods, creative imagination and sensuous-emotional delight. In the midst of a garden with wonderful sweet-smelling, blooming and wish-fulfilling trees and surrounded by fences of rubies, sapphires, pearls, emeralds and ponds of “immortality water”, “bliss” and “reflection”, stands Lalitā’s precious thousand-pillared pavilion covered over and over with jewels and pearls and brilliantly shining in the early morning heat and the lustre of the moon. The secret palace is sealed with the bolt of love-making. The Tantric lineages (āmnāyas) are the entry doors to Lalitā’s inner bed-chamber. Here one encounters the ravishing goddess whose form is constant bliss (see also PKS 4.3,5). The supports of her throne of pearls are the great cosmic gods and her seat is the body of Sadā-Śiva/Kāmeśvara. The jewel island imagery is generally restricted to Lalitā’s śrīcakra (a symbolic cosmogram), but is transferred in the PKS to Śyāmā and Vārāhī too, which makes them Lalitā-like. They share literally the same space, although their diagrams have nothing to do with the jewel island image and their actual diagram worship won’t allude to it (in contrast to Lalitā’s). But the mental image is invoked preceding the worship. Parā, on the other hand, has no diagram at all, but instead has the real cosmos (tattvas) as her throne and seat. The tattvas are in fact conceptualisations that exist only in consciousness, whereas the sonic primal elements, the mātṛkā sounds that link Lalitā, Śyāmā and Vārāhī, can be heard. This way a tight network of inherent relationships among the different goddesses is established by means of formal permutation and semantic transposition. Consider the following extract from fig. 2, most of which has already been discussed. What is interesting here is not the singular segments, but the innumerable relations and semantic links that are generated by the two powers of expression and imagination, i.e. the complex “net of meanings” of the ritual whole:

250

Annette Wilke

Fig. 4: Gaṇeśa Lalitā 2.1.5 establishing the diagram of the deity

Śyāmā Varāhī

Parā

2.7f.

3.9

6.17

7.15

8.18 (tattvas)

2.1.6 visualisations of the “jewel island”

–

3.10

6.14

7.16

–

2.2.1 dissolution of body elements (bhūta)/of the cosmos (tattvas)

–

3.12

6.9

7.6

8.10 (tattvas)

4.1.2 discharging “consciousness” from the nostril and heart lotus

–

4.1

–

–

–

- mental creation of the god/dess image (invok. in the heart)

(2.4)

4.1ff.

- leading the melted cosmos (tattvas) into the heart lotus and perceiving Parā as its reflected unity

–

–

–

–

8.17f. (tattvas)

2.7 (5)

4.4f. (64)

6.20 (16)

7.23 (16)

–

4.1.9 worship of supreme light (prakāśa+vimarśa)

–

–

–

–

8.21

4.1.10 worship of the 15 lunar goddesses (Nityās)

–

4.9

–

–

–

4.1.4 image worship (5, 16 and 64 upacāra)

4.2 worship of the diagram (āvarana-pūjā)

2.7f.

6.20 7.19ff.

8.20f.

5.1-14 6.21ff. 7.26ff.

–

- cosmos in the heart – offered into supreme light

–

–

–

–

8.18, 22 (tattvas)

4.3.2 meditating the “parts of love” (kāmakalā)

–

5.16

–

–

–

4.3.3 visualising the “auspicious heart” (SAUḤ)

–

5.17

–

–

(SAUḤ)

2.9

5.21

6.36

7.36

–

4.4.3 Śakti-pūjā (worship of a human female)

Lalitā and Parā have most ritual units of their own (for Lalitā see also fig. 3). The two “framing” deities, Gaṇeśa and Parā, occupy the most extreme ends, i.e. primarily external and primarily internal worship, while the other goddesses share


251

in both. Śyāmā and Vārāhī are modelled in line with the Lalitā image by coding them with the jewel island symbolism. Parā is set apart (no jewel island symbolism; not receiving image-, diagram- and Śakti-pūjā) and yet related to the other goddesses by her cosmic imagery and practices of visualisation and active imagination. Nowhere will you find exoteric worship in her case, but always ritual manipulations concerning the 36 tattvas (cosmic principles) connected only with her. Note the numbering: from the synoptic point of view, i.e. seen through the lens of the Lalitā Krama which was the model for the table above, Parā’s “ritual sphoṭa” is without sequence just like Bhartṛhari’s linguistic parā, the supreme “visionary” state of language. Parā’s worship activates a “visionary space” without sequence. There is no poetic jewel island nor is there rule-governed diagram-worship. There is another semantic transposition, however. We meet with an interesting metamorphosis of diagram worship and a different form of imagination. The expressive means in Parā’s case are purely cakra-yoga and metaphers of fire and light. Instead of sensuous visualisation or verbal worship of a graphic symbol we find a highly imaginative exercise, the absorption of the cosmic principles in the ritual agent’s and the goddess’ body and consciousness. In the Parā-chapter there is of course a clear ritual sequence. After mentally sealing the physical universe (the 36 tattvas) in the navel region and “burning” it with the fire from the mūladhāra-cakra near the genitals (PKS 8.10), the melted universe is transposed “like heated metal” into the heart-lotus (8.17f.). Thereafter Parā is evoked as that which reflects the essential unity of these outer world elements (8.19). She is visualised and the universe is offered into her, the blazing supreme light who is pure awareness (prakāśa) and reflection (vimarśa) melted in one (8.20-22).. The play with erotic imagery which is very explicit in the Lalitā-chapter is completely metaphorised and spiritualised in the worship of Parā. Her ritual includes, however, alcohol rites, but characteristically no Śakti-pūjā. In contrast, the other goddess cycles will make much more use of the pañcamākara and include cakra-yoga, too (for instance in the nyāsa rites), but not in such a pronounced and exclusive way as Parā. Whereas Lalitā is coded with the fantastic jewel island, erotic imagery and earthly light, Pāra is primarly associated with the tattvas, interior cakra-yoga and the supreme non-dual light. Lalitā is the in-between, the knot which connects Parā and the other deities. She partakes in the nature of both. This becomes clear in the image worship: Gaṇeśa’s physical image and diagram receive five services, Śyāmā’s and Vārāhī’s diagrams sixteen (their image is purely mental), whereas in Lalitā’s case the 64 services for the image are only mental (kalpayāmi, “I imagine”) and verbal (PKS 4.4–5). The worship of the image activates the imagination of the inner core of the jewel island; the ritual encounter is the real-world encounter with the beautiful Lalitā in her innermost bed-chamber. The ritual agent verbally adorns, perfumes and decorates the goddess with costly scents, garments and jewellery, whereafter he sets up the śrīcakra and ascends with the goddess (once more ver-

252

Annette Wilke

bally and mentally) the seat which is the body of Kāmeśvara. Here she/he drinks a glass of liquor of immortality and enjoys bliss and rapture. Lalitā’s worship is followed by the worship of the Nityā goddesses, diagram worship, kāmakalā visualisation and meditation of Parā’s seed sound SAUḤ (Lalitā’s “auspicious heart”). Here we find formal and semantic echoes which link Lalitā to all goddesses (via the kāmakalā), but particularly to Parā (via the “heart,” the luminary imagery and Parā’s seed sound SAUḤ). Lalitā’s sensuous kāmakalā meditation on the world-creating power (PKS 5.16) is immediately followed by the meditation of Parā’s mantra SAUḤ (5.17), connoted with transcendence and liberation. Both these meditations precede real sexual intercourse (5.22), the mimetic and metonymic echo of the world-creating heavenly intercourse of Lalitā-Kāmeśvarī and Śiva-Kāmeśvara (if the human female partner agrees). Whereas Parā is associated with transcendent existence, the supreme light, i.e. awareness, and the seed of liberation SAUḤ (Lalita’s inner essence or “auspicious heart”), Lalitā is associated with a heavenly paradise (jewel island) and pure, untainted creation (the vowels which the Nityās represent), but also with earthly existence, the waning and waxing of the moon, the sequenced alphabet and the rapture of sexual intercourse. But throughout the worship there is constant allusion to the highest, transcendental plane of existence, expressed by the use of the bījas HSAUṂ (Parama-Śiva) and SAUḤ (Parā). While in Parā’s case imagination is completely turned inward into a “visionary space”, and in Lalitā’s case both appealing to the senses and connected to the inner world, imagination is corporeally embodied in the Śakti-pūjā. The human Śaktis (and their partners) become living deities by the power of imagination. So you will find in the Gaṇeśa cycle a boy (baṭuka) and a girl representing Gaṇapati and Siddha-Lakṣmī. They are offered fragrances, flowers, rice, alcohol, meat and fish. The only unambiguous prescription for sexual intercourse occurs in the Lalitā-chapter, but it is very dry and short: “a single Śakti who resembles the goddess” should be satisfied with the pañcamakāra (PKS 5.21). The Śakti representing Śyāmā must be “dark”. Only mantra incantation, alcohol and meat are explicitly mentioned in her case. Vārāhī’s three Śaktis (the horrific goddesses Vārāhī, Krodhinī and Stambhinī) should be “very young and infatuatingly beautiful” and accompanied by a boy representing Caṇḍoccaṇḍa-Bhairava. The group is worshipped, bathed, adorned with fragrances, etc. and thereafter offered “all the substances”. This is the only descriptive passage on the Śakti-pūjā. Otherwise the PKS is not very explicit about it; sexual intercourse in particular is only indirectly hinted at – in striking contrast to the highly descriptive, poetic and lush passages about the jewel island. Nevertheless, the short passages create a suggestive background, and express as well as stimulate imagination and fantasy to think about the rest (which even the commentators disagree about). But I need not repeat how much such subjective fantasy of individual readers and performers is textually and ritually channelled, curbed


253

and controlled. The religious context does not allow ritual intercourse to be conceived as a means of better sex of the Kāma-Sūtra type as Western Tantra imaginations would assume.

5. Concluding Remarks on Theory: Unitary Ritual View and Ritual Poiesis The four syntactical categories have been extremely helpful as analytical keys for decoding a very difficult and bewilderingly complex ritual script. This script is commonly acknowledged as one of the most important Tantric ritual manuals,45 but has received so far no systematic treatment (nor even translation into a European language). The Bhartṛhari-inspired categories have made it possible to tackle it for the first time in an encompassing perspective. The analysis of sequence, relation, expression and imagination has shown how composed in a very strict and deliberate sense the PKS is, and how creative, poetic and symbolic it is at the same time. There is vice versa coding of form and content. The four parameters have made it possible to disclose the tight networking of formal structure and meaning. The analysis of fig. 1 revealed information about sequence and relation and showed the semantic homologies in the formal structures. It became clear how much they form together a meaningful whole and relate the goddesses to each other, while at the same time mirroring each deity’s own individual traits in terms of theology and worship. The artful composition of a macrostructure of mirroring ritual units which include increasing transpositions, rearrangements, insertions, permutations and a few unstructured discontinuities in the successive cycles of worship, establishes a net of relational inherences and generates at the same time a net of formal permutations and semantic transpositions. Fig. 2 allowed a more indepth analysis of expression and imagination and their interaction with sequence and relation. Even blanks turned out to be meaningful in so far as ritual units restricted to one deity or only certain deities almost invariably turned out to be encoded with a specific theology and ritual imagination particular to this/these deities. I have suggested calling them “structural units of metaphorical density”. Most revealing regarding meaning have been those cases where a ritual unit pertains exclusively to one deity. I have presented all cases except for one: the special mantra attached to Vārāhī’s diagram (PKS 7.17). This is the only case where pragmatic reasons were presumably in the forefront for establishing an extra ritual unit. Since the “hundred and thousand corners” of the diagram mentioned in the mantra can hardly be graphically represented, the verbalisation is a practical necessity. But this unit also shows something about the ritual imaginations attached to Vārāhī which are mirrored in her monstrous means of expression. These expressive means are 45 See, for instance, Goudriaan & Gupta 1981: 150–151.

254

Annette Wilke

dark and horrific goddess features (midnight worship, cruel names, etc.) as well as a graphically exploding diagram and a long root mantra of 112 syllables (PKS 10.46). Likewise, the singular ritual units of the other deities have invariably been indicative about the imaginations attached to the specific goddess form. All of them (except for Vārāhī’s diagram-mantra and Śyāmā’s special rules of conduct) are reserved for Lalitā and Parā, the very core deities of the ritual. The formal elements are charged with hidden meaning and a web of interconnections which “burst open” when the textual semantics and cultural codes (e.g. the alphabet as “mother” of the universe) are taken into account. As I have shown Lalitā is established as connecting link between the goddesses by expressive ritual means and semantically loaded symbolic forms and imaginations. The four syntactical categories – sequence, relation, expression and imagination – as analytical tools made clear how tightly their interaction is in the ritual itself. They have been revealing in terms of ritual poiesis and semiosis. They must be seen and analysed as a continuum. If my “test case” allows a general conclusion it would be this: formal ritual structures (such as identical, rearranged and differential ritual units or an increase and decrease of sensory pageantry) will themselves already reveal a lot about content. Formal/structural units are at the same time meaning units. A unitary ritual view will be attentive to the organisation of ritual sequences and their relation to semantics. It will focus the means of expression and imagination and their relationship of inherence as manifestor and manifested. A unitary ritual view will conceive their ritual design as an interactional net of homologies and pay attention to the cultural specific ideas attached to the construction plan. I am not sure whether my observation on the vice versa coding of form and content in the PKS will hold true to such great extent for every ritual. I am sure, however, that their interplay will always be there and invariably there will be intensively interactive relations among the ritual elements. In fact, I did not even know to what extent the PKS would turn out to be such a highly artful composition (even regarding minute details) when starting my analysis, although there was a clear “intuition” from the very start that structural units mirror meaning units and that Bhartṛhari’s syntactical categories would be fruitful in throwing light on this. In fact, they helped me to understand the text more deeply and revealed some new insights which would have gone unnoticed without the structural approach by means of the four parameters. I am confident that any ritual analysis will profit by making use of these categories, because every ritual is a complex composition and will always be based on the powers of sequence, relational mapping, means of expression and culturally specific imaginations. So I claim that the four categories helped not only to understand a difficult script more in-depth, but also introduced a new theoretical approach to ritual syntax, an approach which is rigorously holistic. Bhartṛhari’s grammar provided the major idea that it wouldn’t be enough to gather and analyse ritual morphemes and


255

units, that ritual is (like language) always organised in a holistic way. The singular ritual elements must be seen in their interrelation, how they mutually “explain” each other and form a meaningful whole. None of them alone makes up the message, but together they form a meaning unit, a ritual sphoṭa, which is in them and yet beyond them. As in sentences, the four basic syntactical categories interact in ritual. They function as mirroring echoes of form and content which reinforce each other. They are not only vital powers of the “language principle”, but also of the “ritual principle”. What is gained by approaching ritual with these four categories? I think a lot. Looking at the four “powers” and their interplay in a given rite will a) reveal the ritual’s construction plan, b) explain how the ritual’s performativity works and comes about, c) allow a formalised approach to the question of meaning and d) catch hold of the aesthetic communication which is vital to each ritual. I should not forget, e) they allow ritual to be approached in a formal way which is not stripped of cultural contexts and imagination. The analysis of sequence very much confirmed academic expectations regarding ritual syntax (redundancy, parallelism, recursiveness, transposition, etc.), but sequence coupled with the other powers also helped to go a step further. The set of four allowed the introduction of semantics, performativity and aesthetics into the merely formalist and structuralist discussion. This was possible not least because of the two highly efficient syntactical and synthesising “powers” which have been given too little credit so far: relation/inherence and imagination. It is to a large extent these two powers that account for the ritual Whole and its intrinsic meaning production, i.e. its inherent self-communicative capabilities. More than form and sequence, relation/inherence and imagination proved to be very important operational keys for decoding the ritual construction plan. I have illustrated in detail how they generate networks of meanings, innumerable relations and formal and semantic links in their interaction with the other “powers” and explain the formal permutations and semantic transpositions. My suggestion is that relation/inherence is the central power of a ritual syntax. While sequence fractifies into observable actions, objects and performers, relation unites the different ritual morphemes. Both belong to the “ritual principle”. The complex network of interrelated and interactive ritual units which is established by the power of relation make up the “Whole” of the rite. This whole is a highly meaningful one. Imagination is a major force which accounts for the extraordinariness, meaning and efficacy of ritual. But imagination needs embodiment. These are the many means of aesthetic-sensory and linguistic expression used in ritual. Reciprocally, these expressions need imagination to be viewed as something else (for instance, alcohol as “water of immortality”). Similar to Bhartṛhari (and modern cultural theory), my understanding of the “power of imagination” is not restricted to subjective fantasies. Rather, imagination must be seen as a third space which

256

Annette Wilke

relates real-world phenomena and subjective mental space with each other and thereby creates something new. This something, for instance immortality-wateralcohol, is a collective knowledge and a shared experience. It allows active participation in a sacred cosmos. The means of expression and imagination generate a “net of meanings” which includes body knowledge, sensory awareness and emotions. Expression and imagination are important categories for explaining the selfcommunicating and self-generating potential of ritual and its “world-making” potency. What accounts for ritual syntax is relation and inherence, or more precisely relation as inherence. Relation/inherence is about putting together the ritual elements and activities in a certain way, their mutual interaction and their connection with ideas, cultural codes and imagination. All of this is needed so that the “ritual sphoṭa” can take place and simultaneously permit the grasping of sense data and meaning. My presentation has been a plaidoyer for the necessity to combine a formalist approach with a culturalist one, as well as a plaidoyer for expanding mere formalism and its logocentrism to take into account sensory awareness, bodily knowledge and emotional participation which the means of expression generate. I have deliberately chosen a linguistic model that is less formalist, but would have the great advantage of accounting for the aesthetic message of ritual and its auto-reflexive and auto-poietic effects. The aim was on the one hand to provide basic categories of ritual syntax that would help explain ritual as an instrument of knowledge and transformation, and as a machine that generates and channels perceptions, imaginations and emotions, and at the same time to offer operative keys for formal (structural) analysis. I have been proposing that ritual is comparable to a sentence and putting forward a “unitary” syntactical approach that does not avoid tackling ritual as meaningful, highly symbolic action and analysing the ways its structure, form and content interact. This interaction constitutes in fact the rituals’ aesthetic messages and their capacity for meaning production. The four parameters or “powers” are very basic and at the same time very complete and comprehensive. They contain everything that belongs intrinsically to a ritual. I have initially stated that they won’t be of great help in distinguishing ritual from other actions. But this is not quite true. Their answer would be different. What delineates, according to my basic categories, ritual from other actions will not primarily be a formal element but rather imagination. Imagination overlaps more or less with what Bell called “sacral symbolism” or with Michael’s “extraordinariness”. Even intentio solemnis, transformation and effect (Michaels), are imaginative to a high degree. Imagination is a very powerful category if used not only descriptively, but as an analytical tool. Other criteria which Bhartṛhari’s powers can cover are formalism (Bell, Michaels), rule governance and performance (Bell), as well as framing (Michaels). More difficulties exist regarding traditionalism and invariance (Bell) or authority and orthodoxy (Michaels), if they are not


257

explicitly mentioned in the ritual (script). Both pairs are the only parameters in Bell’s or Michael’s list which are (generally) extrinsic to ritualised action, i.e. ascribed to it. However, they can be accommodated in Bhartṛhari’s set of four as well, in so far as traditionalism and invariance have to do with the “power of time”, while authority and orthodoxy are concerned with relation. Of special interest is the remaining parameter “agency” (Michaels) which is the only one that does fit into my group of four. According to this group, however, ritual itself has agency. I recall that Bhartṛhari views the categories as creative powers. This can be well explained with modern theories of performativity, and vice versa the “powers” explain how performativity comes about. So, many things are gained. Formally, the most important gain is abstraction and minimalisation of complexity, two very important principles for theorising. A strong theory will be one with minimal parameters that explain a lot and provide new insights. The Bhartṛhari-inspired approach allows a new perspective: it throws another light on ritual material and can therefore come to fresh results. Structurally, the categories are very strong because among them there is interior relation. There have so far been few catalogues of ritual criteria which fulfilled this objective. Bell’s parameters, for instance, are good operational tools to work with but are just an incoherent assembly of formal criteria that do not yet make a “grammar of ritual” and even less a syntax. To develop a ritual syntax the very criteria should methodologically reflect syntactical relation. I think Bhartṛhari’s categories are so strong because they do. Although simple they allow complex analysis. There is also vital gain regarding the contents of ritual. The approach provided two new classificatory keys which are able to catch hold of most vital elements that operate in any rite: relation and imagination. These two powers are valid for each and every ritual and are highly efficient. They care for ritual “individuation” by making “the” ritual into “this” concrete and tangible ritual. They allow the analyst to see and systematically focus the inner connections between the singular ritual elements and their relation to the whole of culture. Imagination and relation lend themselves least of all to formalisation. They are the most difficult to be operationalised in a computer-based storage system of ritual elements. There can be no doubt, however, that they must be there in every ritual. So the Indian linguist Bhartṛhari and his basically holistic approach towards language and its communicative ability have a lot to offer. Rituals are creative constructions, and often also highly artistic constructions. Even a “ritual grammar” must somehow account for that. Bhartṛhari’s “universal grammar” and his “powers” offer good avenues. They are compatible with Roman Jakobson’s semiotics and what he called “poetic function”.46 By poetic function he understood the concentration of the poetic message on itself, and the principle of equivalences which 46 Jakobson 1988: 87–96, particularly 94–96.

258

Annette Wilke

governs it. Regarding poetry, we find a network of rhythmic, phonetic, lexicalsemantic and syntactic equivalences which structure what is said in the way that there is mutual motivation of rhyme, phonetics, choice of words and rhetorics, so that the way something is said echoes the communicated meaning. I find this very much applicable to ritual, at least the ritual script which I have analysed. But I dare to repeat the general conclusion made above: rituals are always interactive networks of equivalences of physical forms and non-physical meanings which motivate and reinforce each other. My proposition is that a formal analysis will always uncover both structural units and meaning units. This fact accounts for the possibility to segment ritual elements and understand each element as a necessary part of the symbolic communication of the given rite as a whole. This whole is not only comparable to a theatrical “Gesamtkunstwerk”, but also to a sentence, a poem or an artistic piece of literature. I suggest the language model to be more than a metaphor. Jakobson pointed out the close interplay of sound and meaning, and equally the physical elements of ritual and ritual form may be seen as material and conceptual equivalent, both reinforcing the semantic message that is communicated simultaneously. By claiming this I do not deny some substantial differences between language and ritual action. In contrast to language, ritual action, movement and material means of expression are not informative by themselves. A structural analysis can indeed only reveal formal patterns and functional rules. But such merely formal features actually turn into propositional contents and referential meaning if we include the group’s selfrepresentations, the imaginations connected with the rites and certain ritual elements that are seen as particularly important. Often, we do not have to ask the participants or go into extensive textual studies to discern such meaning. The language used in ritual will tell us all about it. But meaning needs perceptible expression to become a shared social conviction. This is a major aspect ritual action is all about. Another question is how different agents, participants and groups understand the meaning and their performances of ritual action. Each of them may attribute indeed quite different meanings. I have deliberately left out the agency of ritual participants (connected with qualifications, social status, power relations, etc.) and their subjective imaginations. Instead, I have suggested treating ritual akin to the postmodern theory of the “authorless text”. I think it is fair to do so. Bhartṛhari’s thesis that the meaning communicated by the sounds of language is produced neither by the speaker nor the listener, but exists between the two, goes in a similar direction. Rituals are like discourses or canonical literatures, transindividual entities, i.e. interobjective and intersubjective realities. This is why they can be historically effective. The PKS inspired large commentaries and ritual elaborations right up until modern times and is still practised in Kerala. I am quite positive that empirical research would lead to similar observations concerning meaning as those of Staal on Vedic performances. Each PKS-practitioner would presumably produce


259

his own version. Polysemy has also been proposed as a rule. This is somewhat misleading. A given ritual will never produce any kind of meaning whatsoever, but always meaning within a prefigured field. We can learn from literary theory to clarify this issue and account for the more subjective imaginations of individual performers. Much of Umberto Eco’s observations regarding literary works and their “aesthetic messages” are applicable to rituals as well.47 Eco was inspired by Jakobson and gave the theory of poetic function and equivalences a new twist by emphasising performativity. According to Eco literary works do not denote but rather connote. This means there is constant dialectical interaction between form and openness on the level of semantic message, and fidelity and creative initiative on the level of recipients. Literary works are therefore not entities with fixed meaning, but machines of constant meaning production. They stimulate free interpretative choices and creative imagination. However, such freedom of interpretation is restricted, channelled and controlled by certain codes, namely conventionalised schemes, given structures and context relations. So we find on the one hand an offer of empty space to be filled by the recipients’ imagination and on the other hand determining schemata and offers of decoding. The aesthetic message is modelled and patterned by its physical bearers, differential elements, syntactical relations, stylistic repertoires, tropes and not least by ideological expectations, by what Eco called the global connotatum. All of this is received simultaneously while receiving the message. Or in other words: the message consists of a network of homologous forms which makes up its code. There is no break, according to Eco, between “semantic information” and “emotional information”. The auto-reflexive (and inherently ambivalent) message is both a machine that generates emotions as well as an instrument of knowledge. Eco vehemently refuses to distinguish cognitive and emotional content and holds both forms of meaning to be equally important: “If one wishes to delegate the emotional meaning and association to a sphere about which one cannot speak (as for instance Carnap does), one can detect the referential function of language with a certain precision, but one loses the richness of the communication process. Semiotics is not ready to accept such a castration, even if it entails to accept some inaccuracies.”48 Eco’s important argument is that the aesthetic message of a literary work encompasses simultaneously different levels of reality, function and meaning. Since there is homology we get a good deal of information already from formal structures. Eco holds that the more intensive aesthetic expression is, the more complex will likely be the message. All of this is very much applicable to rituals as well. 47 Eco 1994: 108–110, 145–167. 48 Eco 1994: 108.

260

Annette Wilke

And since rituals are generally even more replete with aesthetic stimulation than literary works, for involving besides the verbal a large amount of non-verbal means, decors, sensory pageantry and the body, we can safely conclude that ritual communication is particularly complex. So we may have to think about a ritual poiesis, ritual aesthetics, poetics, rhetorics and performativity instead of/correlative to ritual grammar. Theories of performance and performativity resolve much of the question regarding meaning and meaninglessness. Rituals first of all communicate themselves. They create situative presence and illocutionary and perlocutionary effects. Even leaving out agency, it became clear that we need to look at the functional, pragmatic, semiotic, aesthetic and poietic aspects of the rite and the larger sphere of culture in which the ritual is embedded, in order to make sense of it. We need to take up a suggestion from Tambiah and from Lawson and McCauley, a holistic approach towards ritual action which includes the cultural and group-specific codes and cosmologies. Only by including all these elements can we arrive at deep structures and semantically meaningful contents, and a ritual syntax which deserves this name. Since rituals are creative compositions, ritual syntax will always be about ritual rhetorics and ritual poetics, too. Bhartṛhari’s syntactical categories of sequence, relational inherence, means of expression and imagination can account for all this. They are both operative keys in the ritual and analytical tools for explaining and interpreting ritual in a very thorough and complete way. Being “powers” they naturally account for its innate performativity.


261

References Aklujkar, Ashok 1993. “An Introduction to the Study of Bhartṛhari”. Asiatische Studien 47/1: 37–43. Bell, Catherine 1997. Ritual. Perspectives and Dimensions. New York: Oxford University Press. Bhartṛhari. See Iyer and Rau. Bronkhorst, Johannes 1992. “Études sur Bhartṛhari, 4. L’absolu dans le Vākyapadīya et son lien avec le Madhyamaka”. Asiatische Studien 46/1: 56–80. Cassirer, Ernst 1944. An Essay on Man. An Introduction to a Philosophy of Human Culture. New York: Doubleday Anchor Books. Coward, Harold G. & K. Kunjunni Raja 1990. “Introduction to the Philosophy of the Grammarians”. In: Karl H. Potter & Harold G. Coward & K. Kunjunni Raja (eds.). Encyclopedia of Indian Philosophies. Vol. 5: The Philosophy of the Grammarians. Delhi: Motilal Banarsidass: 3–91. Eco, Umberto 1994. Einführung in die Semiotik. München: Fink. Goudriaan, Teun & Sanjukta Gupta 1981. Hindu Tantric and Śākta Literature. Wiesbaden: Harrassowitz. Gupta, Sanjukta 1979. “Modes of Worship and Meditation”. In: Sanjukta Gupta & Dirk Hoens & Teun Goudriaan. Hindu Tantrism. Leiden: Brill: 121–185. Houben, Jan E.M. 1995. Saṃbandha-Samuddeśa (Chapter on Relation) and Bhartṛhari’s Philosophy of Language. Groningen: Egbert Forsten. Humphrey, Caroline & John Laidlaw 1994. The Archetypal Action of Ritual. A Theory of Ritual Illustrated by the Jain Ritual of Worship. Oxford: Clarendon Press. Isayeva, Natalia 1995. From Early Vedanta to Kashmir Shaivism: Gaudapada, Bhartrhari and Abhinavagupta. Albany: SUNY. Iyer, K.A. Subramania (ed.) 1966, 1973, 1983. Vākyapadīya of Bhartṛhari. 3 Parts in 4 Vols. with Commentaries. Poona: Deccan College. (= Bhartṛhari) Jakobson, Roman 1988 [1965]. “Suche nach dem Wesen der Sprache”. In: Elmar Holenstein (ed.). Semiotik. Ausgewählte Texte 1919–1982. Frankfurt a.M.: Suhrkamp: 87–96. Kelly, John D. 1993. “Meaning and the Limits of Analysis. Bhartṛhari and the Buddhists, and Post-Structuralism”. Asiatische Studien 47/1: 171–194. Lawson, E. Thomas & Robert N. McCauley 1990. Rethinking Religion. Connecting Cognition and Culture. Cambridge: Cambridge University Press. McCauley, Robert & E. Thomas Lawson 2002. Bringing Ritual to Mind. Psychological Foundations of Cultural Forms. Cambridge: Cambridge University Press. Michaels, Axel 1999. “‘Le rituel pour le rituel?’ oder Wie sinnlos sind Rituale?”. In: Corina Caduff & Joanna Pfaff-Czarnecka (eds.). Rituale heute. Berlin: Reimer Verlag: 23–48. — 2005. “Rituelle Klangräume”. In: Annette Landau & Claudia Emmenger (eds.). Musik und Raum: Dimensionen im Gespräch. Zürich: Chronos: 33–44.

262

Annette Wilke

— 2006. “Ritual and Meaning”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing Rituals. Vol. 1: Issues, Topics, Approaches, Concepts. Leiden, Boston: Brill: 247–261 (Studies in the History of Religions – Numen Book Series 114/1). — 2007. “‘How do you Do?’ – Vorüberlegungen zu einer Grammatik der Rituale”. In: Heinrich Schmidinger & Clemens Sedmak (eds.). Der Mensch – ein “animal symbolicum”? Sprache – Dialog – Ritual. Darmstadt: Wissenschaftliche Buchgesellschaft: 239–258. Penner, Hans H. 1985. “Language, Ritual, and Meaning”. Numen 32/1: 1–16. Rānade, Lakṣmana. Sūtratattvavimarśinī. Manuscript No. TR 587.1 and 587.2. Chennai, Adyar: Adyar Library. (= Paraśurāma-Kalpasūtra, commentary) Rau, Wilhelm (ed.) 1977. Bhartṛharis Vākyapadīya. Wiesbaden: Steiner. (German trans. Wilhelm Rau; ed. Oskar von Hinüber 2002). (= Bhartṛhari) Sanderson, Alexis 1988. “Śaivism and the Tantric Traditions”. In: Stewart Sutherland et al. (eds.). The World’s Religions. London: Routledge: 660–704. Sastri, A. Mahadeva & S.Y. Sastri Dave (eds.) 19502 [1923]. Paraśurāma-Kalpasūtra with Rāmeśvara’s Commentary, Reprint 1979. Baroda: Oriental Institute. (= Paraśurāma-Kalpasūtra) Sax, William 2004. “Heilrituale der Dalits im indischen Zentral-Himalaya”. In: Dietrich Harth & Gerrit Jasper Schenk (eds.). Ritualdynamik. Kulturübergreifende Studien zur Theorie und Geschichte rituellen Handelns. Heidelberg: Synchron: 363–379. Staal, Frits 1979. “The Meaninglessness of Ritual”. Numen 26: 2–22. — 1986. “The Sound of Religion“. Numen 33: 33–64, 185–224. — 1990. Rules without Meaning. Ritual, Mantras and the Human Sciences. New York: Peter Lang. Tambiah, Stanley J. 1979. “A Performative Approach to Ritual”. Proceedings of the British Academy 65: 113–169. Thompson, George 1995. “On Mantras and Frits Staal”. Journal of Ritual Studies 9/2: 23–44. Wheelock, Wade T. 1980. “A Taxonomy of the Mantras in the New- and Full-Moon Sacrifice”. History of Religions 19/4: 349–369. — 1982. “The Problem of Ritual Language. From Information to Situation”. Journal of the American Academy of Religion 50/1: 49–71. Wilke, Annette (forthcoming). “Negotiating Tantra and Veda in the Paraśurāma-Kalpa Tradition”. In: Ute Hüsken & Frank Neubert (eds.). Negotiating Ritual. Oxford: Oxford University Press.

Timothy Lubin

Ritual Self-Discipline as a Response to the Human Condition Toward a Semiotics of Ritual Indices 1. The contest between Nature and Nurture for sovereignty over the human psyche is hard to referee, and observers have given the advantage to the one and then the other depending on their differing vantage points and the periodic changes to the rules of play. The nineteenth century brashly asserted the possibility of fully scientific disciplines – an array of Wissenschaften – for the study of society and culture, promising the steady accumulation of positive data that could be analysed to lay bare the fundamental structures of thought and behaviour. This project lives on, though with perhaps less hubris, in the social sciences (psychology, sociology, economics, linguistics).1 The black sheep of the social scientific family, perhaps, was anthropology, which by the middle of the twentieth century, allied with humanistic fields such as phenomenology and literary theory, led the assault on positivism. By now it has become generally accepted that cultural values and individual experience – in fact, all meaning – are socially constructed. The corollaries of this view were that “truth” was contingent, that values like “good” and “bad” were relative, varying from age to age and from place to place, and that the quest to define human nature was quixotic. In the sphere of the history of religions, this implied a break with any theological orientation that asserted the existence of certain truths transcending human consciousness. Theological reflection, scriptural exegesis, and personal piety were analysed “in their own terms”; evolutionary models of religious ideas and institutions came to be regarded as embarrassingly naïve. In its strongest form, this perspective led many observers to insist that distinct traditions are incommensurable; further, even to speak of discrete “religions” was criticised as reductionistic, as 1 The model presented here draws on research supported, at various times, by grants from the American Institute of Indian Studies, the National Endowment for the Humanities, the U.S. Department of Education, and Washington and Lee University, and conducted in part under the auspices of the Institut français de Pondichéry – support I gratefully acknowledge.

264

Timothy Lubin

reifying basically indeterminate, ceaselessly “negotiated” webs of relations, and worse, as endorsing an artificial construct put forward as orthodoxy by a powerful elite. The study of ritual has been subject to the same tendency, no doubt, but the field has harboured a disproportionate number of contrarians – people still influenced by Chomskian linguistic theory, Society for the Scientific Study of Religiontypes – who continue to speak in terms of general features and natural causes of at least some aspects of religiosity. The peculiar features of ritual action – its formulaic and repetitive patterns, the fact that its performers often can give no reason for its performance other than that “it is done”, its outward resemblance to neurotic tics and compulsions, and its susceptibility to being analysed as rule-bound behaviour – all these suggest to some observers that it springs at least in part from a subconscious or nonrational part of the mind, that it is “hard-wired” or instinctual in some way. In other words, a suspicion persists that some ritual activity may by-pass convention and artifice at least on some level, and that the meanings it acquires are not altogether independent of some natural substratum. Even if this were true, we would be ill advised to ignore the meaningfulness of ritual acts: their capacity to denote, to be associated with a signified, whether the function of that signification be to express, to communicate, or to transform something (or someone). Ritual differs from other sorts of activity in that its component acts serve a purpose apparently beyond the reach of physical action. This overreaching quality is figurative, non-literal performance – hence the interpretations of ritual as a special (or original) mode of drama.2 One might say that theatre becomes rite to the degree that it is intended to accomplish invisible purposes beyond those of representation and the evocation of aesthetic response for its own sake. Thus, the analysis of ritual signification can piggy-back on models produced by linguistics, aesthetics, and literary and dramatic theory to a wide extent. But what is left over, still requiring special explanation, is precisely ritual’s further overreach: its claim to make real what is represented – that is, its denial of mere mimesis or fictiveness. If this audacious ambition is to be realised, the ritualist must establish plausibility or verisimilitude. Apart from the aid in this direction afforded by its dramatic and poetic effects, this is ensured (1) in part by an appeal to transcendent authority or its near equivalent, immemorial custom; (2) by the rite’s regularity, by which it mimics the laws of nature and seems to stand above the vagaries of human artifice; and (3) by the strategic use of gestures and objects with iconic or indexical value – i.e. signs that outwardly resemble, or that actually participate in, the phenomenon

2 E.g., Turner 2001; Schechner 2003.

Ritual Self-Discipline as a Response to the Human Condition

265

that the rite is intended to effect. It is this last factor that I wish to discuss here now, using examples from Brahmanical ritual and exegesis.

2. Vedic ritual is extremely complex. What makes it distinctive is its elaborate symbolical system, deciphered in the canonical Brāhmaṇas, and later in the analyses of the Mīmāṃsakas. Like symbols everywhere, the value associated with a Vedic symbol is a matter of convention, posited and elaborated upon by the tradition. I wish to suggest, however, that this symbolism – as contextual and parochial as it generally seems – is an edifice raised on foundation stones that might be deemed intelligible by people from disparate and unconnected cultures. There are ritual signs that are significant independently of a particular cultural context and in fact appear in very similar roles across cultures. In fact, the study of religion today finds itself somewhat at a loss to explain the widespread occurrence (if not universality) of many phenomena ranging from flood myths to relic cults. To observe that all humanity shares certain experiences inherent in the human condition – mortality, sexual reproduction, life on earth, experience of the sun, moon, rain, etc. – is certainly the start of an answer, but it is not enough given the claim that all such experiences are mediated and given shape by individual cognitive habits and the conventions and hermeneutics of the observer’s cultural matrix. Without denying that this particularistic refraction of direct experience is indeed going on all the time, still I suggest that certain individual signs built into complex rituals may be said to point to their own meaning, either directly or indirectly. Some context-sensitivity may be there, but it is not culture-specific context. For example: water points to several distinct but ubiquitous concepts: vitality and growth – in which context it may stand in for other vital fluids: blood, semen, sap, milk, rain quenching of thirst cleansing, purification formlessness or vastness Although these associations come from human experience, they transcend parochialism. Ritual traditions utterly ignorant of each other may appeal to one or more of these basic significations, and someone completely unfamiliar with a particular ritual practice can draw on the same repertory of associations to begin to “decode” it. Insiders will make other, more tradition-specific (purely conventional) associations – associations that may well be perceived as more important or indeed indispensible, and which in some cases may eventually displace such meanings alto-

266

Timothy Lubin

gether – which the outsider will utterly miss. But in general, such a repertory of possible associations will continue to be available as a point of departure.

3. Charles Sanders Peirce, writing about logic in general, and language and languagelike systems in particular, sketched an analysis of different types or functions of signs that many subsequent theorists have found useful. Peirce distinguished between: –

icon = “a sign which exhibits or exemplifies its object”, by the principle of similarity.3

–

index = “a sign which refers to the Object that it denotes by virtue of being really affected by that Object” (i.e. by “a real connection” with it)4

–

symbol = “a sign which refers to the Object by virtue of a law” (i.e. by convention)

Peirce was considering the broadest imaginable range of phenomena, so that his examples of indices ranged from “a man with a rolling gait” (indicating sailor) or “a rap at the door” (a visitor), to instruments such as a barometer or weathercock, which actually point – even the “pointing finger” itself! Other philosophers attempted to refine Peirce’s sometimes obscure classifications, noting that these categories are not mutually exclusive; one might speak of the overlapping iconical, indexical, or symbolical properties of any given sign.5 Arthur Burks emphatically seconded a remark of Peirce’s that there could be no such thing as a pure index, and that all indices must be conventional. In fact, many of Peirce’s examples of indices seem to require a conventional rule to be legible (e.g. the barometer).6 Many subsequent scholars writing on religion and ritual also found this semiotic trichotomy suggestive. Roy Rappaport, Stanley Tambiah, Lawrence Sullivan, and 3 As paraphrased by Burks 1949: 675 (emphasis omitted). 4 Of these three, the index remains the least well defined. Elsewhere, Peirce describes it as “a sign, or representation, which refers to its object not so much because of any similarity or analogy with it, nor because it is associated with general characters which that object happens to possess, as because it is in dynamical (including spatial) connection both with the individual object […] and with the senses or memory of the person for whom it serves as a sign […].” He adds that one of the indices’ “characteristic marks” is “that they direct the attention to their objects by blind compulsion. […] Psychologically, the action of indices depends upon association by contiguity, and not upon association by resemblance or upon intellectual operations [like symbols]” (1931: 2.107–109). 5 Fisch 1978; Giordano 1981; Sebeok 1976. 6 Burks 1949.


267

Malcolm Ruel identified “indexical values” in ritual performance; in their examples, the indices often depend on a conventional association.7 Other examples of ritual, though, may involve a type of pointing that is less culture-specific. Nancy Jay, in her comparative study of sacrificial traditions around the world, proposed that the sharing of flesh by a group “points to” and thus reaffirms the community of participants.8 It might be argued that all human societies would recognise the sharing of food as an indication of belonging to a group. This signification could be recognised quite apart from any question of what sort of group it was, what was eaten, or why. Let’s call that a general signification – one widely apprehensible with minimal contextualisation. Without invoking Peirce, the anthropologist Mary Douglas in 1970 proposed the category of natural symbols. In the intellectual climate of the day, this was seen as either radical or reactionary. One largely appreciative reviewer wrote: “One of our most cherished doctrines, the arbitrariness of symbols, is again up for discussion. Indeed, the title, Natural Symbols, sounds to an old-timer (but not to a very old-timer) like an oxymoron.”9 Douglas argued: “The human body is common to us all. Only our social condition varies. The symbols based on the human body are used to express different social experiences”.10 Since then a cottage industry has sprung up expounding the manifold meanings that get attached to the body – how culture grinds the lens through which we experience even ourselves. Does this drain all the blood out of her argument? I think that what Douglas called “natural symbols” are actually context-independent indexical signs, i.e. signs that in themselves direct attention to certain basic significances (or associations) that seem perspicuous to human beings precisely by virtue of their being human. Particular, second-order significances may be built up within each distinct cultural tradition over time, taking the general significance

7 Tambiah 1979: 119 (Yelle 2003: 80 objects that this characterisation is inconsistent with the conventional character of Austinian “performatives”) and 1984: 336, where Tambiah refers to emblems of Thai Buddhist saints (and of the Buddha) as indexical icons; Sullivan 1986; Ruel saw indexicality in the Nyakyusa “coming out” ritual, in which the planting of a tree “signified a particular occasion and a particular chiefship” and in the Kuria “rite of opening” performed by someone upon reaching elder status, such that the rite “is therefore tied to ritual elderhood” (1987: 108). Rappaport’s approach to the index (1979: 173–193; 1999: 54–68) focused mainly on its expressive or communicative potential, neglecting its pragmatic or performative implications; see the critique in Kreinath 2006: 456–462. 8 Jay 1992: 6–7, citing Peirce 1931: 4.447: the index is “connected with [its object] as a matter of fact.” 9 Silverman 1971: 1295. 10 Douglas 1970: vii.

268

Timothy Lubin

as raw material, as a point of departure. But it is the underlying general significance which provides a basis for cross-cultural comparisons.

4. For example, consider some commonly encountered features of asceticism, purification, and rebirth. I will adumbrate their indexicality, and then illustrate how the general significance of each of these – broadly concerned with overcoming human defects – gets integrated into a culture-specific religious system. Asceticism Ascetical systems everywhere are, among other things, methods for asserting control over those aspects of bodily experience to which most people are inclined to succumb: the urge to eat, to sleep, and to have sex. These urges, of course, are essentially endemic (even if non-conforming individuals might be found). In fact, we might regard these as instincts that we share with other mammals. For us highly self-conscious animals, these urges have corollaries: (A) the satisfaction of these urges brings pleasure; (B) the frustration of them is not just unpleasant, it is fatal: if I do not eat or sleep, I suffer and die; if I do not have sex, I die without issue (which is almost the same thing). Taken together, these corollaries – which rise virtually to the level of universal truths – further correlate carnal pleasure with the weakness, vulnerability, and mortality of the human condition. We might construct a simple syllogism: if A and B, then C: the pleasure we take in satisfying carnal urges confirms our status as weak, mortal beings. The ascetic impulse is a counter-urge, one of the possible responses to recognising the basic indexical significance of these urges. Ascetics attempt to negate the premises: they refuse to take pleasure, and they suppress the appetites as far as they can, hoping thereby to surpass their limitations, to become superhuman. In Peircean terms, if eating and sleeping are taken to indicate helplessness and mortality, fasting and vigil contrariwise indicate fortitude and independence – perhaps even immortality. This line of association is reinforced by other Peircean signs: eating entails defecation, a vivid index of both decay and impurity; sleep is an icon of death (“the big sleep”); sexual climax has even been called “la petite mort”. Purification In all societies, one of the basic uses of water is the removal of dirt. Water then is readily recognisable as an index of cleansing. Dirt, likewise, commonly is taken as an index of any undesirable, compromising accretion. In purification rituals


269

throughout the world, water is mimetically used to remove even metaphorical (spiritual or moral) dirt.11 Rebirth All cultures have recognised that people and animals come into existence by being born. They have also noted that birth is a consequence of sexual union, and that both sexes have a role to play in the conception and development of the foetus in the womb. Gestation and birth, then, constitute an index pointing to other types of cultivation and emergence in a new status, often with its own father and/or mother figure. Life-cycle rites and initiations are typical examples of practices constructed around pivotal indexical signs. These are all examples of indexical signification that can be readily legible in roughly these terms across cultures, independent of local cultural conventions – for which, in other words, all the sufficient context is provided by the interpreter’s being human.

5. I conclude with a brief illustration of how a particular tradition – Vedic rituals of consecration and initiation, as expounded in the masterpiece of canonical Vedic exegesis, the Śatapatha-Brāhmaṇa – constructs a web of extended meanings around the basic, first-order meanings of some core indexical signs. Consecration as Rebirth On the upanayana (initiation into Vedic study, brahmacarya), the ŚatapathaBrāhmaṇa (ŚB) explains: áthainam āha kó nāmāsīti prajāpatir vaí káḥ prājāpatyám evaìnaṃ tát kṛtvópanayate | áthāsya hástaṃ gṛhṇāti | índrasya brahmacāry àsy agnír ācāryàs távāhám ācāryàs távāsāv íti || ŚB 11.5.4.1–2 [The teacher] says to [the student], “Who are you by name?” “Who” indeed means Prajāpati. So making him (a child) of Prajāpati, he initiates him. [2] Then (the teacher) takes his hand, saying, “You are Indra’s student, Agni is your teacher, I am your teacher, O so-and-so.” “Formerly”, it is said, teachers would begin by teaching the Sāvitrī verse only after a year, on the grounds that “foetuses are born by the measure of a year” 11 In analysing rites de passage (of which the rites described here are examples), Edmund Leach invokes Mary Douglas's discussion of the symbolism of dirt and its removal (Leach 1976: 61–71, 77–87).

270

Timothy Lubin

(saṃvatsarásammita, lit. “equal to a year”) (11.5.4.6–11). This ritual gestation is very important; ritual improprieties such as not sitting face to face can even result in a metaphysical birth defect: “He begat him crookedly, so he will be crooked (bulbá)” (bulbáṃ nv ā ayám imam ájījanata bulbó bhaviṣyati, 11.5.4.14). In support of a delay of only three days, the Śatapatha quotes a verse: “Laying his right hand on him, the teacher becomes pregnant; on the third [night], the brahmin is born, together with the sāvitrī [verse]” (ācāryò garbhī bhavati hástam ādhāya dákṣiṇam | tṛtīyasyā̀ṃ sá jāyate sāvitryā sahá brāhmaṇá íti; cf. AV 11.5.3). Finally, the redactor dissents, arguing that just as fire is born instantly (sadyó vā agnír jāyate), so is the new-born brahmin, and so he should be taught at once (ŚB 11.5.4.12). Some say that a teacher who is “pregnant” with a student should abstain from sexual relations, “lest I generate this brahmin out of the ejaculated semen” (néd imám brāhmaṇám, víṣiktād rétaso janáyāni). The Śatapatha also records a dissenting view that makes a distinction between human (physical) birth and divine (spiritual) birth: tád u vā āhuḥ kāmam evá cared | dvayyò vā imāḥ prajā daívyaś caivá manuṣyàś ca tā vā imā manuṣyàḥ prajāḥ prajánanāt prájāyante cándāṃsi vaí daívyaḥ prajās tāni mukható janayate táta etáṃ janayate tásmād u kāmam evá caret || ŚB 11.5.4.17 But they also say, “He may follow his desire. Twofold indeed are these progeny, divine and human. Those human progeny are born from the vagina (prajánanāt). The divine progeny are the metres – he generates them from the mouth, and from there he generates the (student). That is why he may follow his desire.” By this line of reasoning, the student’s rebirth is a birth of an entirely different order, taking place as it does through the medium of speech. Hence, the teacher’s sexual activity is irrelevant; the birth of his student takes place via a different and purer organ.12 This gives the brahmin teacher control over the higher, divine reproductive process, which improves upon the impure, merely biological birth; the male teacher supersedes the mother. Humanity can control the divine world by ritually mastering divine speech. The same image of rebirth is developed even more elaborately in connection with the Soma consecration (dīkṣā). The consecrated one wears garments representing the amnion and chorion, “for a child is born with a caul.” During this period, he should stammer and keep his fists closed, like a baby. He should scratch himself only with a deer’s horn (circuitously explained as the compressed womb of 12 It is said that a person is impure or mortal below the navel, and pure or immortal above the navel; in the consecration, the belt distinguishes between the two regions, Taittirīya-Saṃhitā (TS) 6.1.3.4; cf. ŚB 6.7.1.9–11; 10.1.2.11.


271

Speech herself), lest he cause his own abortion. The fire hall in which he moves is described as his womb. The result is that he is reborn as a god for the duration of the sacrificial ritual. Purification with Water In discussing the vrata consecration for an iṣṭi offering, the Śatapatha comments upon four main factors: (1) purification with water; (2) verbal assumption of the regimen, or assumption of a divine state of truth; (3) fasting; and (4) restrictions on sleeping. These outward, ritual measures are said to signal an inward transition from a human to a divine state, based on the assumption that one must assume a godlike form if one is to approach and serve the gods. All of these elements occur in a much more developed form in the dīkṣā, the consecration for the Soma sacrifice. Vratám úpaiṣyan [...] apá úpaspṛśati tád yád apá upaspṛśáty amedhyó vaí púruṣo yád ánṛtaṃ vádati téna pūtir antarató médhyā vā āpo [...] pavítraṃ vā āpaḥ [...], ŚB 1.1.1.1. “He who is about to enter on the regimen touches water […] because man is unfit for sacrifice (amedhyá). He is inwardly putrid because of the untruth that he speaks. Water is fit for sacrifice (médhya) […] Water is the purifier (pavítra) [...]” Only then, devoid of untruth, is he ready to utter his vow (“Agni, Lord of the Regimen, I will take up the regimen [...]” ágne vratapate vratáṃ cariṣyāmi). Yet divine words (mantras) are presumed to establish actual effects: by announcing the change as fact, the sacrificer “obtains the body of a deity” (devatāśarīro bhavati), as the commentator Mahīdhara says it. In fact, what makes the gods divine is their observance of the vratá of speaking only truth; by doing likewise, one partakes of their glory (yáśas) (1.1.1.5). The transmutation from human to divine is explicitly reversed in the end, when he releases himself from the regimen by uttering: “Now I am who I really am (idám aháṃ yá ev āsmi sò ’smi, Vājasaneyi-Saṃhitā 2.28b).” The purifying touch of water is what makes this verbal self-transformation possible at all. Rules on Eating Another central factor is fasting (ánaśana). The sage Āṣāḍha Sāvayasa is said to have considered the vrata regimen to consist in fasting (ánaśanam evá vratám): indexical self-denial on the fast-day (upavasathá) attracts the gods who know an offering will be made the next morning, so they come to visit (upa-vas-) with the worshipper (ŚB 1.1.1.7). The question remains what the fast should entail: total abstinence from food, or a special diet. The redactor’s view (1.1.1.8) is that, as a

272

Timothy Lubin

ritualised act of hospitality and deference, worship requires that the host (the yajamāna) abstain from eating until his (invisible) guests have partaken. On the other hand, Yājñavalkya, the master of casuistry, finds a paradoxical third option in this either/or situation. Not eating at all would cause one to “make the ancestors one’s gods” (pitṛdevátyo bhavati) – turning the (ritual) world upsidedown, and perhaps implying that one would die. Instead, “he should eat what, though eaten, is not (really) eaten” (sá yád evā̀śitam ánaśitaṃ tád aśnīyād íti), namely, food of a type not used in the offerings (uncultivated foods that grow wild) (1.1.1.9–10). In this way, one can nourish oneself without preempting the gods. From another angle, it is elsewhere proposed that wild things are especially appropriate for the vratá-observer: “He eats of the wild. The wild is power (indriyá), so he puts power in himself” (indriyáṃ vā āraṃyám indriyám evātmán dhatte, TS 1.6.7.3–4 ~ Kāṭhaka-Saṃhitā 32.7:26.8–9). So this particular tradition, guided by its own hermeneutics and values,13 ends up reinterpreting the rule on fasting by redefining what counts as food, by specifying substitute foods, and by offering various ethical rationales for it. In this case, the general signification gets pretty thoroughly obscured, though the underlying notion remains of death-defying abstention as an inversion of mere human weakness, and it gets picked up again in later, professional forms of Indian asceticism, which ensure that considerations of quantity and quality be abandoned by requiring food to be obtained by begging – a practice that begins with the Veda student. Rules on Sleeping Another rule enjoined upon the Veda student is: “Do not sleep” (mā suṣupthāḥ). “Thus he really tells him, ‘do not die’” (ŚB 11.5.4.5). Later versions of the rule qualify it with the word dive (by day), suggesting that the student serves the sun by day (by remaining awake) and the ritual fire by night (by remaining beside the fire and keeping it kindled, but not necessarily by abstaining from sleep altogether). This idea is echoed in explaining the same requirement in connection with the regimen of an iṣṭi-offerer, who must spend the night in one of the fire-hearth enclosures. The Śatapatha explains it as a sign of quasi-divine status: “For he who enters on the regimen approaches the gods, and he sleeps in the midst of those very gods whom he approaches” (devān vā eṣá upāvartate yó vratám upaíti sá yān evòpāvártate téṣām evaìtan mádhye śete). The exegesis of the Śatapatha seems to depart very far from what I have suggested to be the general signification of the key ritual signs in this performance. However, the creativity of the exegete is merely offering various explanations for the presumptions that are already implicit in the ritual. Touching water cleanses the 13 The association between the wild and sanctification appears also in the symbolism of dīkṣā, in which the consecrated worshipper takes on the attributes of a wild animal.


273

putrid stain of untruth in a man, allowing his self-transforming vow to be efficacious. The sacrificer’s abstention from food is interpreted both as self-denial out of respect for his divine guests, and as rejection of ordinary food in favour of a special, restricted diet. Chastity The renunciation of all sexual activity (including involuntary responses like nocturnal emission) is the fixed rule for the Veda student throughout the period of study (indeed, brahmacarya becomes the standard term for sexual continence, even in Buddhism). It is also enjoined on the sacrificer while he is consecrated (by vrata or dīkṣā). In almost all cases, this rule is explained in terms of conserving one’s vital energy for self-transformation, rather than expending it on pleasure and procreation.

6. This cluster of ritual practices – first presented and interpreted in the late Vedic literature as a sequence of progressive consecrations aimed at maximising personal potential through temporary regimens of rigorous self-control – went on to become the basis for the rules of “professional” ascetics as well. The indexical character of many elements of these practices (many more elements than I have had time to cover here) is repeatedly overlaid with theological and cosmological meanings determined by the conventions of the hermeneutics applied. It is evident from these examples that novel, second-order interpretations are continually being devised by canonical tradition or popular opinion, and these interpretations often became more prominent. But here and there, the underlying general or first-order significances of these indices peek out, and they can be indentified precisely because they reappear in analogous ritual practices in other cultures. The indices are merely starting points. The ritual choreography in which they are set is by no means determined by them. Rather, it is the ritualists – both clergy and laity – who play the signs like instruments, and the resulting “music” may be rigidly programmatic or full of improvisation. Then the ritual exegetes, like music critics, wield their authority, as did the virtuosic authors of the Brāhmaṇas. Yet if their interpretations, however finessed, strike a chord, it may be because the underlying drone of the index continues to resonate.

274

Timothy Lubin

References Burks, Arthur W. 1949. “Icon, Index and Symbol”. Philosophy and Phenomenological Research 9: 673–689. Douglas, Mary 1970. Natural Symbols: Explorations in Cosmology. New York: Pantheon. Fisch, Max A. 1978. “Peirce’s General Theory of Signs”. In: Thomas A. Sebeok (ed.). Sight, Sound and Sense. Bloomington: Indiana University Press: 31–70. Gerow, Edwin 1984. “Language and Symbol in Indian Semiotics”. Philosophy East and West 34/3: 245–260. Giordano, M.J. 1981. “Icon and Symbol: A Reappraisal of the Resemblance Debate”. In: John N. Deely & Margot D. Lenhart (eds.). Semiotics 1981. Proceedings of the Sixth Annual Meeting of the Semiotic Society of America. New York: Plenum Press: 29–38. Jay, Nancy 1992. Throughout Your Generations Forever: Sacrifice, Religion, and Paternity. Chicago: The University of Chicago Press. Kreinath, Jens 2006. “Semiotics”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing Rituals. Leiden: Brill: 429–470 (Studies in the History of Religions 114). Leach, Edmund 1976. Culture and Communication: The Logic by Which Symbols Are Connected. Cambridge: Cambridge University Press. Peirce, Charles S. 1931. Collected Papers of Charles Sanders Peirce. Edited by Charles Hartshorne & Paul Weiss. Vol. 2. Cambridge: The Belknap Press of Harvard University Press. Rappaport, Roy A. 1979. Ecology, Meaning and Religion. Richmond: North Atlantic. — 1999. Ritual and Religion in the Making of Humanity. Cambridge: Cambridge University Press. Ruel, Malcolm 1987. “Icons, Indexical Symbols and Metaphorical Action: an Analysis of Two East African Rites”. Journal of Religion in Africa 17/2: 98–112. Schechner, Richard 20032. Performance Theory. London: Routledge. Sebeok, Thomas A. 1976. “Iconicity”. Modern Language Notes 91/6: 1427–1456. Silverman, Martin 1971. “Review of Natural Symbols: Explorations in Cosmology by Mary Douglas”. American Anthropologist 73: 1293–1295. Sullivan, Lawrence E. 1986. “Sound and Senses: Toward a Hermeneutics of Performance”. History of Religions 26/1: 1–33. Tambiah, Stanley J. 1979. “A Performative Approach to Ritual”. Proceedings of the British Academy 65: 113–169. — 1984. The Buddhist Saints of the Forest and the Cult of Amulets. Cambridge: Cambridge University Press. Turner, Victor 2001. From Ritual to Theatre: The Human Seriousness of Play. New York: PAJ Publications. Yelle, Robert 2003. Explaining Mantras: Ritual, Rhetoric, and the Dream of a Natural Language in Hindu Tantra. London: Routledge.

Rich Freeman

Pedagogy and Practice The Meta-pragmatics of Tantric Rites in Kerala The broad goal of this paper is to explore the relation between sequences of ritual performance and related structures of variously textualized language, as both of these are concurrently learned by Brahman youth studying to work as priests in the Hindu temples of Kerala. In terms of our panel theme of relating the structuring of ritual to that of grammar, I believe research into the pedagogical context of how novice practitioners actually learn and rehearse ritual, and their use of language in this process, may prove illuminating. In particular, I am led to take a rather different approach to the problematic of formalizing ritual analysis on the model of linguistic grammar. This approach comes out of a tradition of anthropological linguistics that casts profound doubt on the whole tradition of treating language as a fundamentally mentalist phenomenon, reducible to the sentential level of morpho-synactic-lexical grammar. It argues instead for a reframing of grammar itself under a pragmatics that is continuous with socio-cultural semiotics.1 Aside from these arguments (which I believe to be both true and necessary), it is also hoped that the substantive documentation of a living ritual tradition in conjunction with its use of its own guiding texts will suggest something of the complexity that accompanies any attempt to abstract a formalization on the model of grammar from either ritual practice or the indigenous textualization of this practice. Indeed, the substantive argument of this paper is that ritual performance and its various textual representations form discrete and only indirectly related systems of knowledge, requiring special mediating knowledge to navigate between them. This follows from the observation that it is only after students have achieved a certain level of practical proficiency in the rites that they take up the ritual treatises, along with studying how to bridge pragmatically between the rites as enacted and the ritual system as textually represented. This should therefore serve as a caution against the assumption that ritual treatises (even of the more practical, prayoga, variety) are a descriptive testament, real or intended, of lived ritual practice. But it also suggests that in recognizing the articulation between these relatively discrete 1 Silverstein 1976; 2004.

276

Rich Freeman

trajectories, we can neither dismiss the enacted rituals as a degeneration away from some textual ideal, nor denigrate the texts as useless language games, riding atop a real world of ritual. Both are particular forms of socio-cultural practice, and it is both their relative integrity as discrete forms of practical knowledge, and their modes of articulation with each other that have made the total system robust and historically viable. I shall argue, however, that neither this integrity, nor the principles of articulation can be modeled as “grammatical”, in any conventional sense, and I now turn to a brief overview of the semiotic principles of language pragmatics that I believe we need to adequately deal with these issues.

1. Semiotic Linguistics and Ritual Let me begin by stating an evident ethnographic fact: how thoroughly imbricated the learning and practice of Kerala temple ritual is with language use. There is almost no aspect of learning the rites that is not embedded in verbal description, categorization, instruction, and commentary, aside from the many parts of the rituals that are themselves substantively linguistic – the required mantric recitations accompanying the rites. The modes of this language are both oral and written, in Malayalam and Sanskrit, with an intricate textuality implicated at all levels within and between these modes. By textuality, I refer not merely to those properties of language which allow its inscription into the artifact of a physical text (though I shall deal with such extensively). I refer rather to all those projections of language which provide texture or cohesion in and across utterances and sentences in context (the domain of what some might call “text grammar” were that model not fundamentally misconceived).2 I view these projections as working in a critically indexical semiotic mode, and both their constitution and their engagements with behavior as being fundamentally pragmatic, in both a broadly social and specifically philosophical sense. Since these terms pragmatic and semiotic have a variety of historical usages, it is necessary to specify the scope and interrelation I intend here a bit further. I use pragmatic in the sense that all meaning is always operative only as it is construable in some context. The implication here is that the usual constituent hierarchy of phonology-morphology-syntax-semantics is therefore always occurrent in, and subject to, higher-level pragmatic regimentation.3 Therefore, the whole semanticoreferential or denotational “logic” of the language system as traditionally under2 I use textuality as a shorthand here for the processes of entextualization and contextualization developed in Silverstein and Urban (2003), drawing on the earlier work of Bauman and Briggs (1990), as well as the concept of textual cohesion in the work of Halliday and Hasan (1976). 3 Silverstein 1993.

Pedagogy and Practice

277

stood (sense and reference, denotation, truth-values, etc.), culminating in the “sentence” (where most linguistic analysis stops) comprises just one pragmatic functional domain of language, among the many social purposes that language serves. In hindsight, it is therefore not surprising that no attempt at abstraction or formalization at these lower constituent levels has ever proven adequate for an extensional model even for textuality, narrowly understood, let alone cultural systems of behavior, such as ritual. I believe there are thus trenchant theoretical and methodological reasons why, as Axel Michaels has pointed out in his essay, the bounding or recognition of ritual “units” and “rules” on the model of narrowly linguistic constituents or grammatical regulations can only be – at best, and highly problematically – metaphorical. From the perspective of the actual analysis of language components under such a pragmatic construal, it is the semiotic modes of their operation that necessarily provides the functional architecture of language. Semiotics is understood here not in the Saussurean sense of abstracted systems of signifiers projected onto an equally abstract, logically partitioned space of signifieds, but rather in the Peircean sense of signs, as things that stand for other things, their objects (thus including all the relations of a merely Saussurean semiotics), but always in respect of some contextual ground, and always from the vantage point of some interpretant. While the semiotic modes of iconicity (where there is some formal similarity between sign and object), and symbolism (where the relation is purely a matter of conventional habitus) are important, it is indexicality that grounds the Peircean semiotic most significantly for an anthropology of language and culture. Indexicals are those signs which signify by their spatial, temporal, or causal copresence or contiguity with their objects, in some dynamically conditioned relationship with them. In the language system, these indexicals work both internally to pick out and anchor linguistic signs to other such signs or stretches of language, and externally, to connect those indices with aspects of their surrounding context. So the power of the Peircean semiotic is that the signs of language and of the extra-linguistic world form an analytic continuum, exactly as we experience or enact them in social behavior, such as rituals.4 Furthermore, we may say that while language is pragmatically multifunctional in terms of serving many social and cognitive purposes, the overall mode of language, per se, is pragmatically indexical in that it works to change the context of occurrence by its social enactment. So social action and communication are not, in 4 The characterization of pragmatically semiotic anthropological linguistics here is so broad as to require an extensive bibliography. One can see, however, the following: on pragmatics, as a departure point still within traditional grammar Levinson (1983); on the Peircean semiotic as pragmatics in social scientific thought Mertz and Parmentier (1985); on the relation of this approach to language philosophy, truth-functionality, etc., Lee (1997); and for a recent statement on the relation of this school to culture theory Silverstein (2004).

278

Rich Freeman

this view, phenomenologically or analytically disjoint realms. A little reflection should also therefore make it evident that all language is therefore “performative”, in this sense of saying something counting as doing something, where the special use of this for so-called “speech-acts” merely highlighted most clearly the systematic failure of the semantico-referential reduction to capture the pragmatics of a whole class of linguistic actions. Finally, such a semiotics encompasses a broader understanding of the scope and significance of the poetic functioning of language and of linguistic-contextual imbrications, very much following up on the insights that Jacobson (1960) brought to bear on this issue. When we understand that indexical ties can access and juxtapose previous semantic domains into novel, metricalized, “syntactic” configurations, setting off new and relatively autonomous meanings, we can use this purchase on all sorts of tropes, to appreciate the “poetically” semiotic configurations that ritual effects, as well. Ritual, is, in this understanding, a kind of actional “text”, that is indexically iconic, a dynamically indexical figuration of actions and events that is simultaneously iconic of a hidden, other order which they project, and from which they (at least ideologically) derive. This much has been eloquently argued by Tambiah (1976), in a deservedly famous paper drawing on Jacobson, but it is Silverstein who has further called attention to the poetic nature of such figurations, in their ritual particulars (1981) and in their overall role in the architecture of interactive textuality (1984). While there are further important distinctions to be made in terms of the modes of indexicality, and various strata of meta-levels at which these can operate, this much should suffice for characterizing this alternative model of a socio-culturally pragmatic semiotics, as against all those species of linguistics which excise language from its pragmatic matrix and attempt to establish an autonomous realm of mental-formalist structures, on the model of Chomskian or other such traditionally oriented grammars.

2. The Ethnological Setting The cultural and historical context for this research is the Kerala tradition of institutionalized temple worship, called Tantra in Kerala, but largely corresponding to the Āgamic tradition of worship elsewhere in South India. It represents a Smārta-Brahmanical fusion of the Śaiva-siddhānta and Pāñcarātra-Srī Vaiṣṇava traditions (with a bit of Śākta worship thrown in for good measure), and so has been essentially non-sectarian since at least the fifteenth century. This nonsectarianism at the doctrinal level is reflected socially, as well, since the single and largely integrated caste of the Nambudiri Brahmans have been the undisputed authorities of ritual management and priestly offices of upper-caste temples the length and breadth of Kerala throughout most of its known history.


279

While there is textual evidence of much earlier variability that came together in this synthesis, what is now accepted as the Kerala tradition was brought together and consolidated in the fifteenth century in a Sanskrit compendium called the Tantrasamuccaya5 (hereafter, TS), which was rapidly accepted as authoritative among Nambudiris and the society at large, through the whole region of Kerala. Two commentaries, a shorter one (-Vimarśinī), and a more elaborate one (-Vivaraṇa), were written on this text in the next generation, among the author’s sons or disciples, and a supplementary text, the Śeṣasamuccaya,6 was produced around that time in these same circles. The TS tradition was transformed into the vernacular Malayalam in the eighteenth century, in a work that was part translation, part commentary and handbook, called the Kuḻikkāṭṭu Pacca7 (hereafter KP). While a number of Nambudiri priestly houses produced similar -paccas (vernaculars), this one gained prominence, especially subsequent to its publication from existent manuscripts in 1974. Today it is seen everywhere in the houses (and hands) of practising priests in Kerala. The field setting for this research is a school of Tantric studies, Thantra Vidya Peedhom (to use one of the variant spellings), currently situated near Alwaye, in central Kerala. It was founded in 1972 in Guruvayur, partly under Hindu Nationalist (RSS) auspices, in order to address what was seen as a crisis in the demise of the traditional guru-system of imparting ritual expertise among Nambudiris. The idea was to salvage this system by bringing students together in a semi-modern residential school environment, and have them taught by traditionally trained gurus, as their “teachers” in “classes”. The first chief guru (“Principal”) was a well-known traditional Tantri, Śrī Kalpuḻa Divākaran Nampūtirppāṭu, who had both been brought up under, and had taught for many years in, the dwindling, traditional gurukula system of Nambudiri Tantric practice. He trained the first generation of students at the new school, who now carry on as the teachers of the current generations of students, following his demise. Students from hereditarily entitled Nambudiri families of Tantris (temple ritual authorities) or Śāntis (everyday priests) apply to the school and are accepted after their SSLC (i.e., high-school) graduation, generally around sixteen years of age, for a six- to seven-year course of study. They study in classes, by class-year, using printed materials (such as the KP, a modern edition of the TS, glossed into Malayalam, and a variety of xeroxed hand-outs), and, most especially, their own notebooks which they write as they advance through the years. The practical teaching of the kriyas (rites of worship)8 begins at the outset, with a combination of 5 6 7 8

Ramaswami Sastri 1945–1962. Narayana Pillai 1951. Divākaran Nampūtirippātu 1986. It will be my practice to shift between Malayalam and Sanskrit spellings, depending on the context of my exposition, without further registering the minor differences.

280

Rich Freeman

imitation, rote memorization, note-taking, and recitation of accompanying mantras. This practical teaching was greatly enhanced by the shifting of the school in 1985 into a building adjacent to a working, traditional Kerala temple whose administration the school took over, and which it now operates on behalf of the local, worshipping constituency. From the third year students begin active participation as working assistants and priests, so that by the time they leave the school, they have had years of on-the-job training, both in all manner of daily rites, and the complex yearly festivals. In addition, many senior students are hired out for other temple rites in the region and state, and younger students accompany them as assistants. Students also begin studying general Sanskrit from their first year, and this is gradually integrated into their study of the rites, including memorization of many portions from the TS. A few of the more gifted students use this as a basis to go on to university study and to become Sanskrit teachers or scholars, but most apply their course of study purely as vocational training to become priests in their hereditary or other, public or private temples. I have had close acquaintance with this school from 1993, and have spent many months in residence there over the years, sitting in on classes, interviewing students and teachers, and attending pūjas and festivals. My own dear friend and teacher, the late Śrī K.P.C. Anujan Bhattathiripad, also happened to have studied in his childhood home under Śrī Kalpuḻa Divākaran Nampūtirippāṭu, back when something like the traditional gurukula system was still intact. I was thus able to discuss the traditional system of study at length on a number of occasions with Śrī Bhattathiripad, who was himself a Tantri of great repute, and to compare the system of his youth with the curriculum of the current, school-setting.

3. Architecture of the Ritual-Textual Structures As my introduction suggested, the Kerala ritual system, whether for the learner or the expert, entails a highly stratified system of textualized and embodied knowledge, with multiple mediations (in language and register, behaviors and contextual settings), across its nested domains. To trace out just a few of the sets of behavioral routines and some of the indexical links they entail, I have chosen a small sub-set of rites, from the most elementary order of daily worship (pūja) for exemplification. I will follow this sub-set from the elementary, simplified form, into the more elaborated, advanced pūja. This is more or less how the students (and myself, as a student of the students) encounter and make sense of these elementary units, in the larger architecture of the system as a whole. One rendering of the standard phases of the pūja sequence in Kerala templeworship might be as follows, with roughly literal glosses and brief explanations, following. I will be focused here only on the sub-set of sequences in bold type.


281

Phases of the Kerala Pūja Abhivādya – “reverent salutation” of the god by the worshipper Dehaśuddhi – “purifying the body” in a series of destructive and reconstructive rites Śaṅkhapūraṇa – “filling the conch” with divinized water that will be used in lustrations Ātmārādhana – “worship of the self” as the initial locus of the in-dwelling, generic deity Pīṭhapūja – “worship of the seat” on which the image of the god sits Āvāhana – “invocation” of the particular deity out of the body, into the image Upahāra – “offering” of various kinds of foods and honorary services to the imaged deity Mūrti-pūja – “worship of the image” of the iconographically embodied deity Naivedya – principal food that is offered in “communication” with the deity, who absorbs its essence Prasanna-pūja – “worship of the gratified” god with items offered to royalty Puṣpāñjali – “handfuls of flowers” offered with reverence for final satisfaction Layāṅga – “limb absorption” of the deity’s components, withdrawn back into the body of the worshiper from the image of the deity, as the termination of worship Some of the names given to these phases are quite standard, others somewhat more variable, and the parsings themselves vary somewhat in scope and boundaries depending on context. (An analogy would be perhaps to the underdetermined sense of where paragraph and section breaks might come in a longer written text, and how one titles the sections). Each of these names, however, delimits notional runs of activities – physical, verbal, and imaginary – using implements and substances, engaging or mapped onto the body in various ways, and all breaking down into further constituent sequences that are also named or described. To help me with this study, the students set up a demonstration for me of the simplest pūja to Visṇu, in a vacant room, where I video-taped it. The major implements and vessels, though not filled with their various substances of worship, were used as props, as is sometimes done by the students themselves (Images 1–4). I was thus treated to a kind of imaginary pūja, with the various motions and hand poses (mudras) illustrated, the various fillings, emptyings, and depositing of substances mimed, and the various mantras recited. This kind of laminate of description, recitation, and demonstration, all in a kind of imaginary realtime, is typical of student’s practical classes. Later, I was able to video-tape actual counterparts of the demonstration pūja on the temple premises for comparison (which will be touched on, below). Apropos of what I just noted about the naming and bounding of sequences, what I have segregated as the abhivādya, or “salutation” phase is often included

282

Rich Freeman

under the label deha-śuddhi, as its prelude. This prelude to what I taped and treat here as the deha-śuddhi, included: in addition to the named “salutation” where the worshiper announces his name to the deity, and gives praise (vandana) to the guru, and to the god of inceptive rites, Gaṇapati; the blessing of the plank on which the worshiper sits; the protective sealing of the quadrants with imagined “ramparts of fire” (agni-prākāraṇa); and his own transformation into the primordial deity, Virāj, through mantras, and through the internal manipulations of his life’s breath (prāṇa). This last manipulation of the prāṇa with the syllable oṁ (praṇavaprāṇāyāma), especially, shades readily into the deha-śuddhi, which the student’s own presentation broke down for me as follows: Phases of the simple Deha-śuddhi (bodily purification) 3 vyāpakams (pervasions) 2 sets of aṅga-nyāsams (limb deposits) chandas (meter) mānasa-pūja (mental pūja) mūla-japam (recitation of root [mantra]) chandas (repeat of the above) The student did not give me the whole architecture of the deha-śuddhi at the outset, but rather said simply, “then three vyāpakam, aṅgam, chandas”, and launched into the vyāpakams, literally “pervasions” of the body with the mantric power or essence of the deity. This was executed by waving the hands from the sides of the head, crossing them in front of the face, uncrossing them at the shoulders, bringing them down to the sides and waist, following out the thighs of the crossed legs, down the shanks, and to the ends of the feet, each fluid sweep accompanied by the recitation of the “root-mantra” of the deity being worshiped (in this case, “oṁ namo nārāyaṇāya” for Viṣṇu). This is a named unit in the vernacular (I believe always done in triplicate), called “doing vyāpakam”, which recurs regularly throughout the compendia of rites. It is often accompanied, as here, with the “depositing” (nyāsa) of sets of aṅgams or “limbs” of the deity, on the worshiper’s body, and I believe the vyāpaka is therefore notionally a kind of abbreviated synopsis of these rites of building up the deity on one’s body, since it’s equivalent in the Sanskrit of the TS is sakalī-√kṛ, “making complete” (sa-kalī being literally “with [all] its parts”). Following this, for the demonstration of installing two distinct sets of Viṣṇu’s aṅgas (limbs), the first set of eight, and the second of five, the student simply said, “aṅgam”, and demonstrated all thirteen as a single run of actions, using the appropriate mudras on different sites of his body with the appropriate mantras. While these are clearly paradigmatically different sets at the verbal level (though


283

with overlap for five of the same sites of the body, with their mudras), performatively these hung together for the student as a single behavioral chain, flagged as well by his singular, collective nomination (aṅgam) for it. The chandas (meter) notionally must be intended to provide a certain Vedic authorization for all these Tantrically worshiped gods, since we find the same attached to those many deities that are wholly Tantric in origin and worshiped in private rites of “sorcery” (mantravāda) in other treatises. The nomination of this little rite as chandas is metonymic, since it always includes the serial listing of the Vedic seer-sage (ṛṣi), the Vedic meter (chandas, proper), and an associated deity (devatā), each accompanied by the same mudra of the hand, touching the crown of the head, the upper lip, and heart, for the three respective entities, as they are named. The next rite, mānasa-pūja (mental worship), which had not been formerly announced as part of the sequence, was then simply named, and the student launched into it, again, by simply demonstrating the appropriate mudras, accompanied by the recitation of their respective mantras. This unit, named as such in the Sanskrit commentaries and sub-commentaries to the TS (with variants mānaso yāga, etc.), as well as in the Malayalam manuals and current student’s usage, is clearly modular, in that it recurs in a number of different contexts across the larger sequence of pūja and festival rites of all kinds. It takes the form of an explicit series of speech-acts (which lie on a continuum with mental acts), since each mantra ends, “I imagine/declare” (kalpayāmi). The object of these acts of declarative conception are five offerings (of water, unguent fragrance, flowers, incense smoke, and lamp-light) each declared to contain the essence (-ātmanā) of the five physical elements which are thereby imaginatively offered in worship to the deity. Added to these five mantra-mudra strings is a sixth, the food-offering of naivedyam, as the essence of ambrosia (amṛta), which I will return to shortly. For the final two items in this simplest of deha-śuddhi sequences, my young helper simply announced, “Do the mūla-incantation as much as you can (yathāśakti), and the chandas”, and he then proceeded to sit, barely murmuring, while counting out repetitions of the mūla-mantras by folding the fingers of his righthand into his palm through a couple of cycles. He then went immediately into a repetition of the “chandas” triad, and announced, “Deha-śuddhi is finished; next, śaṅkha-pūraṇam.”9

9 He had shown hesitancy, earlier, as to where the deha-śuddhi began, confirming with a friend, off-stage, as to when he should announce it, after his performance of the abhivādya; the hiatus between the end of this sequence, and the śaṅkha-pūraṇam, however, was obviously more distinct in his own mind.

284

Rich Freeman

4. Some Reflections on the Rites’ “Grammatical” Order This simplest version of the deha-śuddhi performed for me was just the rehearsal of what this student had learned in class, and this was textualized in terms of all its essential elements in the notebooks which students prepare in their classes, and later take with them into their practice. What even this simplest textual inscription shows, is the complex relation between the verbal, physical, and notional orders of activity that make up ritual and its textual inscription. For initially, it is clear that students acquire the verbal and physical sequences of action, learning simultaneously to label or describe them at a meta-level, and to order these, in turn, at metameta-levels. There is a part-wise build up of these linkages from simpler to complex, and from actions to inscribed texts, through the years of their studentship. This “practical” learning itself, however, has a textual legacy behind its classroom teaching and re-inscription, for many elements and partial sequences of this simplest deha-śuddhi are recognizably correlated with parts and sequences from the eighteenth century KP which the teachers use to teach with, and which students study. While I do not believe there is anything in the published tradition that corresponds to the particular abbreviated form for what has now become the standardized simplest pūja, students nevertheless learn to navigate around the KP to find the units and sequences that correspond to their practical tradition. This formal text that is closest to their performance does not therefore provide any clearly iconic ordering to the actual rituals they learn, nor does it provide rules for abbreviating or deriving them in any explicit formulation. Students rather learn all these moves outside the text, on the practical side of their study. So one finds the “grammar” of ritual in the KP neither more or less clear, in terms of units, their sequencing, and rules of combination, than the “grammar” of the behavior itself. What one does find is that certain named units, or meta-designations across the behavioral realm and the textual one are correlated as “the same” (whether overtly or by stipulation and glossing), and that this then allows the pedagogic build up of correspondences both across, and between the two orders. These correspondences, however, are indexical pair-wise projectings to-and-fro, creating meta-level chunkings of only partially labelled units of sequential runs of varying scope, that shift according to pedagogical and practical context. The order is thus ultimately an emergent pragmatic and meta-pragmatic one. This is not to say the texts themselves do not set up their own attempts at formal orders. The whole TS tradition, in fact, attempts to set up Viṣṇu as a “paradigm”, in the sense that whenever there are deity-specific stipulations (as there are for most of the mantras and many constituent rites), those for this god are presented first, and then the various alterations or substitutions for Śiva and the rest of the seven gods to which the TS has attempted to reduce the Kerala pantheon. This often creates difficulties for the text’s expository architecture. It also prompted the writing of a whole supplementary text to the TS, the Śeṣasamuccaya (śeṣa being the “re-


285

mainder”), which treats a whole second set of gods and all their specialties, left out of the TS. In any case, this pedagogically inceptive status given to Viṣṇu no doubt governed the choice of which god the students chose to present me with for their demonstration, but this choice has been itself determined by tradition, and not for any paradigmatic efficiency dictated by the “deep structure” of the ritual. So at this level, the text seems to follow its own order, but it is rather a merely “textual” set of ties governed by its own strings of choices made discontinuously across many of its sections, and not motivated by the kinds of practical choices facing a learner or a practitioner. Taking a unit like the “vyāpakam”, we can see that it has a fairly specific Kerala form as a practice, but that it is underdetermined by the authoritative text of the TS or its subcommentaries. A textual warrant for this comes between the discussion of the deposition of the mantras on the hands (kara-nyāsa) and other depositions, when the text uses the verb vy-√āp (to pervade) in the causative, saying, “using the two hands, from the [deity’s] own root[-mantra] one should cause pervasion in the body thrice, making it complete.”10 The different physical actions used to realize this “pervasion” earlier in the text for the hands, arms, and individual fingers obviously don’t apply here, and students, in fact, evolve their own distinctive styles of realizing this head-to-foot sweep of the hands in practice. It seemed to me that some seem to treat their body more in the manner of tracing a presentational outline, in an iconic fashion, whereas others use their flattened hands as though in greater contact, seemingly treating their bodies as a mass-object. The forming of such units, which may be hypostasizations out of the various strata of texts, or which may emerge in practice and then go in search of a textual warrant, are clearly critical as architectural nodes in the whole perception of a ritual structure. But such units themselves may be subject to fusion or segmentation. The aṅgas of Viṣṇu clearly form two textually discrete sets, but follow on each other so automatically (and with such verbal and physical overlap) that they have fused not just in the performance, but even in the vernacular designation of this particular performance, as the singular collective “aṅgam”. Relatedly, the three elements of the rṣi-chandas-devatā triplet, have been verbally reduced to simply chandas, as a vernacular nomination, and because the mudra extends the two middle fingers of the right-hand to contact the three sites on the body, the vernacular short-hand for this triplicate of mantra-mudra strings is often “touch the chandas” (chandas toṭṭu ...). Finally, mānasa-pūja, while it is a simply learned sequence as a relatively bound unit, suggested by its discrete naming, and grammatical objectification (“do a 10 ābhyāṃ karābhyāṃ nija-mūlatas trir vyāpayya dehe sakalī-karotu...TS 7.13. Note, again, the association of sakalī-√kṛ (to make complete), with this pervasion. While I have not systematically studied all co-occurrences, its seems clear that vyāpaka and “completing” the deity, are matters of degree, depending on the numbers and elaborateness of the “depositions” (nyāsa) of limbs (aṅgas) and mantras according to the appropriate contexts.

286

Rich Freeman

mānasa-pūja”) has a rather complex relation to the earlier texts. The designation in the TS which precipitates such an injunction in the commentaries and vernacular texts seems underdetermined, since it usually says just, “sacrifice/offer with the mind” (dhiyā iṣṭvā/arcayitvā), or some such. The actual mantras are (apparently) an abbreviation of those given for a juncture in the pūja of an image when the five elemental essences are offered as extracted from five gross offerings of the worship; namely, the earlier mentioned water, unguent, flowers, censor smoke, and lamplight. The derivation (uddhāra) of these mantras (“for the purpose of offering the essence [rasa] of water, etc.”, says the commentary, TS 12.166) is given in the last chapter of the TS which is a kind of appendix for what’s been left out of the main exposition. While this particular verse gives the mantras as including the names of five incarnate forms (mūrtis) of the deity being worshipped (Aniruddha, etc., in the case of Viṣṇu), the last half-line is interpreted by the vernacular tradition (through a derivation obscure to me) as indicating that these mūrtis are to be left out of the formula when these are used for the mānasa-pūja. When we turn a bit later to the hand-mudras that accompany these mantras (TS 12.207), however, we find reference to six mudras, which includes a final one to the principal food offering in worship, naivedya. In the mānasa-pūja received in today’s practice, there is indeed another mantra for this additional last item, following the same formula as for the previous gross elements, but now using ambrosia (amṛta) as the subtle essence of naivedya in the formula, and reverting back (clearly with poor fit) to the first hand-mudra for water (since we only have five fingers, but now six “elements”). In addition to these hand positions, however, the mudras are performed on specified sites of the body, and these are found in the seventh, main chapter on pūja of the TS, which again reports that they correlate with the “essence portions” (rasāṁśa) of the gross substances of worship. The final mantra and mudra of contemporary performance then reverts to an earlier verse in the same section, where the naivedya that is offered is conceived as purified, (much in the same manner as the body in the elaborate deha-śuddhi), having its gross elements burned up, and then, most relevant here, converted into amṛta with a special mantra (unrelated to the earlier ones, and deviating from their paradigmatic array of substituted elements). This mantra declares the fall of a rain of amṛta, accompanied by the surabhi mudra. This mudra, however, which, represents the udder of Surabhi, the mythical wish-granting cow, points towards the practitioner’s mouth in the mānasa-pūja, whereas it is supposed to point downwards (again, as described in the twelfth chapter), when used with actually present offerings. I dwell in some detail on all this primarily to illustrate how complexly entangled is the textual, authoritative derivation of what is, in fact a simple set of motions and recitations that are learned readily and easily in performance by students as a standard, modular, and variously recurring little rite. It is not easy for me to tell whether the TS actually intended this rite in its current form by the occasional phra-


287

ses “sacrifice with the mind”, but it is clear that this rite is a kind of condensation of the more elaborate phase of offerings in an external pūja, performed without any such offerings here, essentialized into an abbreviated set of mantras, and turned back on the body of the practitioner. Whatever the case, historically, however, it seems that rather than the text guiding practice, in any pragmatic sense of coherent sequencing, the present practice is used to construct and construe scattered textual fragments and their relations to each other. And it is furthermore these chunkings and namings of the practices, at a meta-pragmatic level, that create the “textuality” or cohesion of the rites, and through them, that of the various text artifacts themselves, both internally, and probably back through their various vernacular and Sanskrit strata.

5. Expanding on the Textual Order While the deha-śuddhi rite is largely confined to strings of mudras and mantras oriented to the body and its imagined transformation, the next phase, the “filling of the conch” (śaṅkha-pūraṇa) is more externally oriented to a set of implements, and to actions performed with and on them. What my video-tapes show are a series of mimed movements of smearing places for the water-vessels (kiṇḍis), filling them with water and transferring it back and forth between them, washing and filling the conch, depositing offerings of flowers and grains into these vessels with mantras, invoking the presence and power of the deities into these by these operations, and thereby charging the water for further lustrations of the site and implements for later use in the worship. Again this was accompanied by first naming the whole rite, as a unit, and then giving a running description of what was being done. The recitational mode and standardization of this description were so consistent that it was clear the students had virtually committed these instructions to memory. This fact is directly registered in the largely identical contents of their notebooks. Again, there is generally conformity with the prose description of the KP’s Malayalam, though the latter is not so practically detailed and complete. The whole rite took about ten minutes to complete, but the corresponding section of the TS comprises only three verses, and is grammatically just one sentence. To give some sense of the conciseness, ellipsis, and potential obscurity of the TS compared to what actually is done and needs to be known, I will look at and compare these verses with the corresponding student notebook, since the latter corresponded very closely to what I saw performed. Part of the obscurity occurs because of the form itself of the TS, for it uses both learned and uncommon Sanskrit vocabulary, and is cast in a variety of literary meters which serve to disorder what would be normal prose order for Sanskrit, or for the syntax of Malayalam. All of this serves, however, to heighten the text’s status as learned śāstra, and therefore authorize the rites it sanctions and mandates, even

288

Rich Freeman

if it doesn’t always illumine them at a practical level. I would indeed argue, despite the TS belonging merely to the genre of ritual digest, that its writing marked a major investment in the symbolic capital of literary form, validated by the subsequent strata of commentarial efforts. Structurally, these three verses of the TS are not as disordered as some, since the verbal clauses (once they are ordered internally around their objects and various complements), chart a series of consecutive actions that do map iconically onto the order of actions intended. Of course, it should be noted that it is not always clear, as it mostly is in the present verses, which objects and complements go with which verbs, and vital objects or complements may need to be construed from context (or the commentary), even for basic intelligibility. To show the structure of this threeverse “sentence”, I simply display the verses, below, underlining the serialized verbal forms. These are gerunds, unmarked for person, number, tense, or mode, so that this information is held in abeyance until one encounters the subject (manuvid, “the knower of mantras”, i.e., the ritual officiant), in the third line, and then the finite verb, here the third singular optative, abhyukṣatu, “he should sprinkle”, at the very end. So the effect would be in English, “The officiant” ... having x’d, having y’d, and having z’d ... “should sprinkle”. The verses thus appear as follows, with the subject in bold type, and the finite verb in bold and underlined. Note that the distribution of the gerunds is not uniform after the first verse, that one cannot predict whether objects or complements will precede or follow their associated gerunds, and that in some cases, they are wanting. All one can say is that the serial occurrence of the gerunds themselves maps the sequential order of the understood actions. TS 7.27-29 kṛtvā go-maya-vāriṇā su-catur-aśraṃ sthaṇḍilaṃ tārato nikṣipyâkṣata-gandha-puṣpakaṃ amuṣmin astra-saṃśodhitaṃ śaṅkhâṅghriṃ praṇidhāya śaṅkhaṃ api tat-tan-mantrato mantra-vic chaṅkhe nyāsya hṛdâtmakena manunā gandha-prasūnâkṣatam sa-kṣâdyena śirōṇunā śuci-jalair āpūrya gandhârtavaṃ nyasyâgny-arka-sudhāṃśu-biṁba-manubhiḥ prārcyâṅghri-śaṅkhôdakaṃ gālinyā śikhayôtprapūya viviṣī-kṛtyâmṛtī-kṛtya tat toyaṃ netra-nirīkṣitam kara-yujā pracchādya varmmâṇunā baddhvâstreṇa diśo’tha tīrtha-manunā tīrthaṃ raver maṇḍalād āvāhyârccita-pīṭhake sva-hṛdayāt devaṃ ca mūlâṇunā āvāhyâṅga-yutaṃ samarccya paramī-kṛtya prajapyâṣṭaśo vardhanyāṃ pratiṣicya kiṃcid amunâtmāntaṃ trir abhyukṣatu


289

To give a more concrete sense of the degree of ellipsis in the TS from the perspective of ritual practice, let us just compare the first couple lines of the first of the above verses with what students actually learn to do, as directed by their notebooks. First, following the practice learned by all Sanskrit students, we must restore the normal prose order (anvaya) of the individual clauses and supply the subject from the third line. Ordering these with complements, modifiers, and objects before the clause terminal verbs, and giving a rough English-ordered gloss, these lines read: manu-vid “the mantra-knower” go-maya-vāriṇā su-catur-aśraṃ sthaṇḍilaṃ kṛtvā “making a good square space with cow-dung-water” amuṣmin tārataḥ akṣata-gandha-puṣpakaṃ nikṣipya “therein depositing whole-grain, unguent, and flowers with an ‘oṁ’” There would seem to be two simple actions here, but this is, at its most generous, merely skeletal. It hardly suffices to adequately capture what the ritualist actually must do according to the same in the student’s notebook and exhibited routine. I now give the corresponding passage from a notebook, with the correspondences to the TS verse underlined and some changed/expanded items in brackets: “Moving the right-hand kiṇḍi (water vessel) and the śaṅkhu (conch) to the front, smear [make] a rectangle [square], divide it in two, smear a circle in the place to put the right-hand kiṇḍi, and a triangle in the place for the śaṅkhu, say an ‘oṁ’, deposit unguent, flowers, and grain, move the righthand kiṇḍi back, put unguent, flowers, and grain in the right-hand kiṇḍi, cover (it) so as to hold it, touching the water with the ring-finger, putting the left-hand on top, intone the mantra, ‘imā āpaḥ ...’ etc. and then one should ‘fortify’ (āpyāyiccu) and ‘sprinkle’ (prōkṣikkuka) thrice.” The TS verses say nothing of the necessary implements or the set-up, or the various operations in constructing the outlined space with its internal divisions and figures for the two ritual vessels (kiṇḍis) and their fillings prior to the washing and filling of the conch. The Sanskrit commentaries give a fuller account, but again, without these particulars, and with only one vessel mentioned. While the full mantra, beginning “imā āpaḥ ...”, is given in the shorter commentary, there is nothing about the peculiar attitude of the hands, with the ring-finger extended into the vessel mentioned. Lastly, there is nothing said of the two peculiar motions, indicated by the denominative vernacular verbs coined from the Sanskrit into priestly Malayalam to name these actions. The first action entails covering the kiṇḍi with the open hand and making swivelling, circular motions over it, and is derived from the

290

Rich Freeman

causative of a Sanskrit verb, ā-√pyai, “to fortify”. The second action is to hold the kiṇḍi in the left hand, pouring a little water from the spout into the right, and then emptying it back in the open top, and is from the root pra-√ukṣ, meaning “to sprinkle”. These actions and their nominations thus represent a kind of lexification into Malayalam verbs of little ritual actions verbally derived from the Sanskrit, but without apparent warrant or mention in the texts of the tradition. A similar kind of linguistic encoding of a more extended ritual sequence occurs in what seems a local coining of a Sanskrit formula. The verbal clause from the second line of the second verse from the TS, cited above, may be translated, thus: “offering to the conch, stand, and water with the mantras of the fire, sun and moon ...” The relevant part of the student notebook reads, Then saying, oṁ vahni-maṇḍalāya namaḥ aṁ sūrya-maṇḍalāya namaḥ uṁ sōma-maṇḍalāya namaḥ with the maṇḍala-triad, worship jalādi-jalāntam, showing the mudras. While the student text first expands on what the fire, sun, and moon mantras are, it ends with a jargon abbreviation that stands for its own modularized rite, “beginning with water (jalādi-), [and] ending with water (-jalāntam).” What this signifies is a specified mode of expansion on the mere "offering" enjoined by the TS. This entails: taking up some sprigs of the tulasi plant, and making a series of offerings into the conch by thrice dipping the sprig into the water vessel (kiṇḍi) and aspersing the conch, reciting the three mantras of the above maṇḍala-triad each time; doing this thrice again with the fragrant unguent (gandham) or sandalwood (candanam) that is ground with water and kept in the small round stand (ōṭam) to the left of the conch; following this with one more aspersion of water, then three offerings of flowers from the flower-tray (pūppālika), followed by another single offering of water, each offering or aspersion, accompanied by intoning the mantras. The normal jalādijalāntam module would have ended here, if it had closed with a final three offerings of water, instead of just the one. In this case, however, because the enjoined module is modified by saying “showing the mudras”, this means that it is expanded by next showing the conch the mudras of the censor smoke (dhūpam), and then of the lamp (dīpam), each accompanied by the maṇḍala-triad, conjoined with the respective mantras especially for dhūpam and dīpam, before a final round of three water offerings, as the “ending with water” (-jalāntam). The vernacular, and possibly more recent realization of the TS tradition, has thus come up with its own linguistic mode of mapping this ritual module, which is verbally iconic in


291

terms of offering water (jala) at the beginning and end, but with two possible levels of ellipsis/expansion in the middle. At the basic, default level, it can incorporate the first three offerings, or by expansion, all five of the basic set we saw above was included in the mānasa-pūja: water, fragrance, flowers, smoke, and lamp. A more textually underdetermined, but more highly systematic and revealing pattern to the whole set of actions and their underlying conception occurs later during the rite of the śaṅkha-pūraṇa. The main deity, Viṣṇu, is invoked out of the worshipper’s body and discharged through the breath into flowers brought up to the nostrils and then deposited into the conch. Once his presence is fixed there, he is worshiped just as the student’s own body was earlier, in the deha-śuddhi, but now as though he were hovering in the air over the conch, facing the worshipper. The earlier three vyāpakams, aṅgams, and chandas are now all done as though transposed onto this imaginary entity, suspended in space before the officiant (cf. two of the aṅgam sets, Images 1–4). Such transposition finds its closest linguistic analogue in deictic projection (as in the routine switching of grammatical person from one reporting frame to another), but such is not a realm that traditional grammars have ever successfully been able to accommodate; this is rather, as with much of what goes on in ritual, in the domain of pragmatics.11 And this means, as I have argued, that there are both higher meta-levels of structuring implied, and indexical linkages connecting them.

6. Ritual Condensation of the Textual Order I turn now to some brief considerations of the final subsection earlier highlighted, that of the ātmārādhana (self-worship), for here we can glean something of these higher meta-levels of pragmatic semiosis. We come here to a consideration of the opposite case of the kind of text-to-practice expansions we have seen earlier in relation to the TS, for with this section we find six full Sanskrit verses (7.30–35), twice what we had for the śaṅkha-pūraṇa, but here they are all reduced in the elementary worship to the following notebook entry: On these places – the brow, throat, right arm, left arm, and heart – in gōpi fashion, with the root [mantra], put on the marks, wash your hand, and with the root mantra, adorn the head (cūṭi) with a flower, wash your hand, and do mānasa-pūjā. Here, merely the most externalized actions of one verse stand, by synecdoche, for the whole of the more complete rite. The gōpi mode refers to the upright sandal-paste marks (ūrdhva-puṇḍra) signifying Viṣṇu, as opposed to the horizontal marks (tiryak-) for Śiva. The choice and mode of contraction, like that of expan11 Levinson 1983: 54–96.

292

Rich Freeman

sion or the earlier case of transposition, are the expression of complex semiotic processes, amenable to a certain amount of formal description, but not on analogy to grammar. These semiotic relations become even more complex once we move to actual pūjas, which I also taped in their simpler and more complex forms, for these entail actual arrays of objects and contingencies of real spaces and contexts, rather than imaginary, idealized projections. Like the pūjas themselves, these sites of worship representing the god can vary in complexity. A simple pūja may use only a small pīṭham (seat/throne) consisting of banana leaves layered with raw rice, on which is placed a coconut, surmounted by a conical pyramid of darbha-grass, all resting on a small diagram (padmam, literally “a lotus”) drawn with colored powders on the floor, and flanked by standing lamps (Image 5). The fuller, more complex pūja will often have a larger, more elaborate padmam (Image 6), when still done on a temporary site, but be directed onto a permanently enshrined image for regular, public temple worship. Most interestingly in the actual performances, all of the verbalizations, both the descriptions of the learning/demonstration mode (not surprisingly), but also the actual utterance of the mantras themselves are rendered silent, with only occasional movement of the lips.12 The entire set of verbal, meta-pragmatic structuring devices, are thus internalized and rendered largely mental, even as the physical spaces, objects, and implements become manifest and enter real manipulations through transformative activities. Perhaps consonant with this, the mudras, as we already saw with the two sets of aṅgams in demonstration, expand into a more elaborate set, but can also fuse into each other, often in fluidly stylized performance flows with marked aesthetic effect. The expanded pūjas are performed at fixed times of the day in temples (the number and elaborateness of pūjas are regularized, but vary according to scale and staffing of the temple), and the expansions apply to each phase of the pūja as given in the initial listing. For instance, the deha-śuddhi will have a greater number and variety of mantric depositions (nyāsas) on the body or the god’s image than those given for the simpler rites. These will entail then both a greater number and variety of (largely silent) mantric recitations, but also, most prominently, a more elaborate set of mudras to accompany these. These can be both impressive physically, but also rendered with more aesthetic flourish, as noted. Furthermore, their variety of representational categories such as “ornaments” (bhūṣaṇa) and “weapons” (āyudha) render the body of the practitioner more fully and evidently iconic of the imagined or sculpted body of the god itself (Images 7 and 8). What is most interesting, however, since this worship in a temple is normally sequestered away from 12 This is the opposite of the neighboring Tamil tradition, with which the Kerala tradition is closely cognate, where priests are expected to intone the mantras out loud.


293

the worshipping public in the inner sanctum, is to consider for whose benefit this iconic presentation of the priest-as-god, coupled with the silent introjection of the mantras, is performed. The obvious answer would seem to be that, phenomenologically speaking, it is for the priest himself, or for the god the priest has become and mirrored onto an image. As I have long argued, I believe this kind of impersonation, coupled with the fact that the god is invoked from the body of the priest himself, shows the continuum of Tantric worship in South India with the ancient cults of spirit possession in that region.13 The śaṅkha-pūraṇa rite is similarly expanded with one extra set of mantric depositions and invocations, compared to the simpler pūja, but I shall close this discussion with the ātmārādhana, since there is a more systematic set of expansions there, as well as occasion to highlight other semiotic modalities engaged, well beyond the complexity of mere grammatical analogy.

7. Leap-frogging through the Narrative Whereas we have seen that the simpler ātmārādhana enjoins only five sites for deposition of the Vaiṣṇava marks with the same, single mūla-mantra, the more complex one requires twelve different sites, with twelve different mantras to different forms of Viṣṇu. Similarly, where there was earlier simply a single rite of adorning the head with a flower with the mūla-mantra, this is now repeated five times with a fairly complex mantra to the “crown” (kirīṭa) of Viṣṇu. Finally, though the student’s notebooks do not record this, the TS enjoins that the mānasapūja is considerably elaborated into a fuller set of pūjas to the deity that is installed in one’s heart. These differences can be summarized in tabular form: Table 1 Simpler ātmārādhana

Expanded ātmārādhana

5 sites for the sandal-paste marks

12 sites for the sandal-paste marks

5 repetitions of mūla-mantra for marking 12 mantras to different forms of Viṣṇu flower put on head once with mūla- flowers put 5 times with “crown” mantra mantra mānasa-pūja as closing rite

mānasa-pūja expanded with other pūjas

What is here given as the expanded version seems the normative one for the TS, and I have not found any textual warrant for the particular reductions of the student practice. In particular, the five sites seem to be those listed for Śiva, in the same 13 Freeman 1998; 1999.

294

Rich Freeman

verse enjoining twelve for Viṣṇu (TS 7.32), and there is nothing suggesting that their respective mūla-mantras may be substituted for the specific five and twelve mantras, respectively, assigned for these gods. Similarly, five floral head-adornings are enjoined, with specifically different mantras for Viṣṇu and Śiva (though, by implication, the mūla-mantra can be used for other gods). Finally, a theoretical argument could be made for mānasa-pūja standing in for the larger set, since, especially in this case, the pūjas are all internally directed to the self or body of the worshiper, and the quintessentially “mental”-pūja seems to be most suited to this function. All these cases suggest, however, that these reductions take place through a variety of thematic or analogic substitutions or ellipses that were probably of an ad hoc and pragmatic nature, however they might now be justified. Even the derivation of the elaborate forms, however, reveals something else about the text-practice interface, since they point to a more pervasive feature of the textuality of the TS itself. One not already deeply immersed in the textual content and architecture of the TS would be hard-pressed to derive the twelve sites and mantras for Viṣṇu's markings, for this passage says merely, “smear on the sites of the mūrti-pañjara with his [the deity’s] own twice-six mantras” (7.32), and it is only the sub-commentary that quotes an earlier passage in the same chapter (7.22) where one might find these sites from a twelve-fold deposition of aṅgas. For the list of forms of deities themselves, however, one would need to go back to the fifth chapter (TS 5.90-91) treating the worship of the largest padmam (or maṇḍala) drawn for Viṣṇu's worship, the cakrābja, in the context of originally installing a temple-image for worship. Even there, these forms are differently ordered, however, and one would need the sub-commentary to understand the reordering of the set implied by the TS’s terse injunction to “deposit the forms starting with Keśava” (7.22). Indeed, the whole attempt to learn about pūja from the TS entails shuttling back and forth between the seventh chapter, statedly devoted to pūjā, and earlier chapters associated with the building of temples and installation therein of divine images. This is because the overall structure of the TS follows a sort of narrative structure. This structure indeed charts out the “biography” of the ideal temple: from choosing its founding priest, to selecting the site, building it, designing and making images, and then executing all the complex rites of installing them. It is only after the images are properly installed that regular rites of pūja are treated, and then the annual festivals, various contingent rites of purification, and then long-term replacements and renovations. The result of this narrative structure is that often complex elements of pūja occur topically in earlier verses, before one reaches the elementary forms, and so, one needs to grasp this architecture of the overall “storyline”, as well as its particulars, for it to serve as either a handbook or textual map of the ritual sequences. The pragmatic structure of the text is thus different from that of the rites themselves, and requires a complex mediation to move back and forth between them. That is no doubt why, at Thantra Vidya Peedhom, the elements and


295

structures of pūja are learned in practical classes before the study of the TS (called ślōkam) is take up in later years of studenthood. One learns the TS, I would argue, in order to establish and validate as socially authorized what one has already learned, practically. The final kirīṭa-mantra of the ātmārādhana, for example, is simply memorized, and nothing practical is gained from later learning it’s versified, contorted mode of “extraction” from the Sanskrit of the TS (12.95–97), other than validating that it is indeed authorized there.

8. The Poetic Pragmatics of Ritual Finally, in addition to a macro-narrative structure which informs the texualization of the TS’s ritual forms and sequences, I want to also point out an internal, higherorder pragmatic feature of many of the ritual arrays themselves: an order which we can only call poetic, in Jacobson’s usage. This feature can be illustrated by presenting the final form of the array of mūrti-pañjara deities of Viṣṇu we have been discussing with their respective mantras, and the sites on the body where these are deposited in the form of unguent markings. Table 2 Mantra

Bodily sites

oṁ keśavāya namaḥ

brow

naṁ nārāyaṇāya namaḥ

belly

moṁ mādhavāya namaḥ

heart

bhaṁ govindāya namaḥ

neck

gaṁ viṣṇave namaḥ

right side

vaṁ madhusūdanāya namaḥ right arm teṁ trivikramāya namaḥ

right side of neck

vāṁ vāmanāya namaḥ

left side

suṁ śrīdharāya namaḥ

left arm

deṁ hṛṣīkeśāya namaḥ

left side of neck

vāṁ padmanābhāya namaḥ back yaṁ dāmodarāya namaḥ

nape of neck

Jacobson’s insight, introduced earlier, was that poetry creates a novel set of paradigmatic equations (the kinds of associative substitutions traditional linguistics thinks of as “semantic”), by juxtaposing these in the matrix of some sort of metric,

296

Rich Freeman

as a novel “syntactic” order. The paradigmatic and syntactic orders of language are thus played off against each other, in a number of dimensions, to generate a novel set of meanings. In the above array, the various names of Viṣṇu’s accompanying, divine forms are set in a series of recurrent matrices of the pattern: (single syllable+ṁ) (name-dative case) (namaḥ) At the most basic level, these names are all associated with each other “poetically” through occupying the same position in the “metric” of these formulae, reinforcing the cultural convention that these are all names of forms of Viṣṇu. Each line further conforms to a speech-act of praise (namaḥ, “praised be ...”) with the requisite dative of the object of praise, i.e., the named deity. An inceptive single syllable with -ṁ is further conventionally coded as a bīja- (“seed”) mantra, signifying in Tantric ideology that it condenses the power of the deity in a generative, sexually imaged potency of divine power. What we thus have is a cumulative series of culturally empowered speech-acts of praise, poetically laminated together, even at the surface level. At the level of ritual behaviors, enacted physically with the simultaneous utterance of each mantra, we have the “deposition” (nyāsa) through smearing of the sandal paste on the indicated succession of sites on the ritual officiant’s body (correlated across the lines of our graphic array). Culturally, there is no doubt that this represents the installation of the divine powers, invoked through their praise, on their respective sites, but there is also something revealed through the meaning associated with the name for this array itself, the mūrti-pañjara, the “cage” or “skeleton” (-pañjara) of “divine forms” (mūrti-). While the TS uses this term in reference to the aggregate of these sites, a knowledge of it’s meaning in the wider Vaiṣṇava context reveals that it represents the divine “retinue” (parivāra) of Viṣṇu, beings who are emanations of his own powers that surround and protect him.14 The temporal order of these mantras further forms a sequence or succession, and this dynamic is poetically charted through the sequence of the bīja-mantras, which, reading their distinctive single-vowel elements downwards in the array (highlighted in bold type) spells out a single, unifying master-mantra to Viṣṇu, “oṁ [the most powerfully comprehensive of all mantras] praise to the Lord (bhagavate) Vāsudeva”! The poetic reading of the sequence is thus that the various retinue deities of Viṣṇu are emanations of and comprise the unity of his higher being. By shifting this reading to the ritual mappings of these energies onto the body, it also further suggests that the body of the officiant similarly comprises, through its divinely aggregated parts, the whole of the god’s identity. This identity 14 This conceptualization is found in the Ahirbudhnyasaṁhitā, 36.61–62 (Ramanujacharya 1916: 351).


297

is further confirmed when it is recalled that the original array of these deities in the TS is treated in the worship of the cakrābja padma, which is unambiguously treated as a locus of worship of the god’s actual physical presence, equivalent to the god’s own iconic image as his body. The overall message this poetic order dynamically unfolds is finally affirmed (with that redundancy typical of the multifunctional, laminated strata of ritual and poetic structures) through the next (and final sequence in the students’ pūja) after this array. This is the aforementioned crown-mudra and -mantra, which demonstrates that the worshipper’s own head now wears the crown of the god, and that the worshiper himself is identified with the god. In an uncharacteristic bit of exegetical exuberance, the usually laconic TS declares, “let him [the worshiper] intone the crown-(kirīṭa-)mantra in order to achieve His [Viṣṇu's] state of being (bhāva)” (7.22). While bhāva (from a root meaning “to be”) can have the extended meaning of a mood or emotive state, it is intended here at the literal level, and the commentary unambiguously glosses this state as “self-identity with Viṣṇu” (viṣṇor tādātmya). Thus the net-effect and culmination of this poetics of ritual, is the kind of semiotic possession by the god, which, as I have written about over the years, illustrates the continuum between ancient South Indian worship through spirit possession, and the Brahmanical appropriation of this power through the authorizing structures of Tantric ritual. What we should finally note with this closing discussion of the poetic order of this particular ritual sequence, is that despite my displaying certain correlative structures in a synoptic graphic array, this sequence is learned, rehearsed, performed, and experienced as verbal and behavioral enactments unfolding in realtime. The textualization draws from this, first as note-taking, and then only subsequently proceeds to build meta-level correlations, partial and pragmatically adjusted, with earlier textual strata. Furthermore, the verbal part itself is eventually suppressed to a silently internalized habitus in practice. The ability to display parts of this as a graphic array, only serves to highlight that the real work, whether performative, exegetic, or poetically interpretive, proceeds by parasatizing such arrays through projecting indexical linkages between and across behavioral, verbal, and cognitive fields. These pragmatic processes are not incidental or supplemental to a pre-existent, underlying set of autonomous “grammatical” structures, but are rather constitutive in their selectively creative invocation and construal.

9. Conclusion: The Cultural Pragmatics of Ritual Orders In conclusion, I believe this material strongly indicates why a culturally nuanced model of semiotic pragmatics is alone adequate to describing and theorizing the “science” of ritual, and not any mere extension or supplementation of grammar, as traditionally understood. Starting from the simple sequences of verbal and physical behaviors which students learn as rituals, we have seen that meta-pragmatic prac-

298

Rich Freeman

tices of describing and naming are critical to what constitutes ritual units and how these are composed, but that these are further subject to fusion, fission, and ellipsis, depending on performative context. The performative utterances (mostly in the form of individual mantras and their arrays) are themselves objects of naming or meta-linguistic description (e.g.s kirīṭa-mantra, mūrti-pañjara, etc.). The coordination of these meta-linguistic entities, with physical actions, bodily or spatial places, and ritual operations, are then subject to further pragmatic and meta-pragmatic regimentation. All of these complexities are redoubled when it comes to textualization, for these various levels of text not only have their own structures, but their own modes of utilization and interpretation, vis-à-vis the ritual structures with which they only partly articulate. So there is a special set of inter-textualized devices and practices appropriate to navigating the texts themselves, which establish the (at best, partial) correlations between textual strata (and languages), and the further mediation between these and the ritual practices that are their object. The systematic differences between these textual levels from each other, and from the actually enacted rituals are readily apparent in instances we have reviewed: in all those cases of ritual expansion, where a few lines of text are elaborated into whole ritual sub-routines; in ritual contraction, where longer textual sequences may undergo essentialization and abbreviation; and in the regular instances of ritual pastiche, where established practice can be correlated with the texts only by leapfrogging backwards and forwards through the textual narrative, weaving together different fragments with extra-textual practices into a synthetic facsimile. Even if parts of routines or subroutines can be graphically mapped into “paradigms” that one would hope show formal analogies to grammatical patterns, these varied pragmatic meta-processes of their invocation and configuration thwart any attempt, in principle, to reduce them to a set of fixed units, constrained by the iterative application of rules. Indeed it is the relatively plastic nature of the textpractice interface, in which the text-artifacts get continually redeployed as a set of pragmatic resources, that both encourages the historical proliferation of textual strata (i.e., commentaries and their vernacular mediations), and alteration of the original texts themselves as they undergo reinscription, editing, redaction, etc. This push and pull is the source of all the truly interesting variation and change that textcritical scholarship can reveal when it is marshalled to a broader historical purpose. A major finding of the current research is therefore to show the extent to which the text-artifact is embedded (and indeed historically crafted) in its own pragmatic culture of interpretation and use, which is far from isomorphic with the ritual practice on which it leans. The work of correlation is meta-indexical and meta-pragmatic in that it creatively brings the discrete contexts of practice and text together, by invoking and forging higher-order, interpretive links between them.


299

Of course to talk of pragmatic and meta-pragmatic correlations is really to invoke those associative systems of culture itself, as we saw when the poetic structures of the ātmārādhana took us to postulated entities of sonic theology (in the “seed” mantras), their “installation” as divinities on the body, and the transformation of that body into the aggregate, personal presence of the deity. These cultural postulates do not derive, in any determinate logical or historical sense, from the constituent signs in which they find expression. Rather, the constituents and structures of the rites have been shaped through the use and re-use to which they have been put by a culture of ritual practice, through time. So while the formal patterning of those rites and constituents can teach us much in terms of the resources of the tradition, there is no science of their combinatorial properties that will generate strings of “grammatically” acceptable rituals for us. For that is not where the “logic” of ritual resides, but rather in the higher orders of culture.

300

Rich Freeman

Image 1: Deposition (nyāsa) of Viṣṇu’s eyes onto the officiant-as-god. Photo by the author

Image 2: Transposition of the same onto imagined conch-as-god. Photo by the author


Image 3: Deposition (nyāsa) of Viṣṇu’s armour on officiant-as-god. Photo by the author

Image 4: Transposition of same onto projected conch-as-god. Photo by the author

301

302

Rich Freeman

Image 5: Śaṅkapūraṇam from actual simpler pūja, showing utensils filled with ingredients, and banana-leaf-grain-and-coconut “pīṭham” set on simple padmam. Photo by the author

Image 6: Set-up for more complex pūja, showing the same “pīṭham” set on a more elaborate padmam, with silk-draped main lamp and auxillary standing lamps for retinue-deities. Photo by the author


Image 7: More elaborately iconic mudras, showing the officiantas-Viṣṇu, with his adornment (bhūṣaṇa) called Puṣṭi. Photo by the author

Image 8: The same, showing the adornment (bhūṣaṇa) called Śrīvatsa. Photo by the author

303

304

Rich Freeman

References Bauman, Richard & Charles Briggs 1990. “Poetics and Performance as Critical Perspectives on Language and Social Life”. Annual Review of Anthropology 19: 59–88. Divākaran Nampūtirippāṭu, K. (ed.) 1986. Kuḻikāṭṭu Pacca (Tantra Grantham). Compiled by D. Subrāya Tantri. Kunnamkuḷam. Kerala: Pañcāṅgam Press. Freeman, John R. 1998. “Formalized Possession among the Tantris and Teyyams of Malabar”. South Asia Research, 18/1: 73–98. — 1999. “Dynamics of the Person in the Worship and Sorcery of Malabar”. In: Jackie Assayag & Gilles Tarabout (eds.). Possession in South Asia: Speech, Body, Territory. Paris: École des Hautes Études en Sciences Sociales: 149–181 (Special Issue of Purushartha 21). Halliday, Michael A.K. & Ruquaiya Hasan 1976. Cohesion in English. London: Longman Group Ltd. (English Language Series 9). Jacobson, Roman 1960. “Concluding Statement: Linguistics and Poetics”. In: Thomas A. Sebeok (ed.). Style in Language. New York, Cambridge: M.I.T. Press: 350–377. Lee, Benjamin 1997. Talking Heads: Language, Metalanguage, and the Semiotics of Subjectivity. Durham: Duke University Press Levinson, Stephen 1983. Pragmatics. Cambridge: Cambridge University Press (Cambridge Textbooks in Linguistics). Mertz, Elizabeth & Richard Parmentier 1985. Semiotic Mediation: Psychological and Sociocultural Perspectives. Orlando: Academic Press. Narayana Pillai, P.K. 1951. The Śeṣasamuccaya. Trivandrum: Government Press (Trivandrum Sanskrit Series 166). Ramanujacarya, M.D. (ed.) 1916. Ahirbudhnya Saṁhitā of the Pāñcarātrāgama. Vol. 1. Supervised by F. Otto Schrader. Madras: Adyar Library and Research Center. Ramaswami Sastri, V.A. et al. (eds.) 1945–1962. The Tantrasamuccaya of Nārāyaṇa with the commentaries Vimarśini of Śaṅkara and Vivaraṇa of Nārāyaṇaśiṣya. Trivandrum, Kerala: Government Press (Trivandrum Sanskrit Series 151, 169, 200). Silverstein, Michael 1976. “Shifters, Linguistic Categories, and Cultural Description”. In: Keith H. Basso & Henry A. Selby (eds.). Meaning in Anthropology. Albuquerque: University of New Mexico Press: 11–55. — 1981. Metaforces of Power in Traditional Oratory. (Unpublished Lecture, Yale University). — 1984. “On the Pragmatic ‘Poetry’ of Prose”. In: Deborah Schiffrin (ed). Meaning, Form and Use in Context. Washington: Georgetown University Press. — 1993. “Metapragmatic Discourse and Metapragmatic Function”. In: John Lucy (ed.). Reflexive Language: Reported Speech and Metapragmatics. New York: Cambridge University Press: 33–58.


305

— & Greg Urban 1996. “The Natural History of Discourse”. In: Michael Silverstein & Greg Urban (eds.). Natural Histories of Discourse. Chicago: University of Chicago Press: 1–17. — 2004. “‘Cultural’ Concepts and the Language-Culture Nexus”. Current Anthropology, 45/5: 621–652. Tambiah, Stanley J. 1985. “A Performative Approach to Ritual” In: Stanley J. Tambiah (ed.). Culture, Thought and Social Action: An Anthropological Perspective. Cambridge: Harvard University Press: 123–166.

Frederick M. Smith

Historical Symmetry and Ritual Asymmetry The Interrelations Between Vedic Ritual and Temple Construction in Modern India Vedic soma Sacrifice (agniṣṭoma) and Hindu Temple Consecration (pratiṣṭhā) In February-March 2007, I observed the simultaneous performance of two major Indian rituals. One was a Vedic soma sacrifice, the second a Hindu temple consecration. These two complicated rituals emanate from two different textual and historical strata in the vast Indian repertory of ritual performance and, in spite of sharing the Sanskrit language as the idiom of their performance, they are distinguished by very different aesthetics, principles of organisation, religious foundations, and the manner in which they relate to their audience. In other words, they featured very different sets of ritual dynamics. I will first describe the rituals and their historical strata, then the conditions under which they were drawn together. I will then discuss a small and surprising rite, an act of professional ritualising, which embodied a unique syntactical function and reveals some of the dynamics that brought them together. I will use these as primary components in a discussion of methodological issues, referring to recent discussions of ritual grammar and syntax. Finally, I’ll consider a perspective on ritual that I hope will contribute to the broader general discussion on ritual grammar, a discussion which has neglected a number of items that I believe have important grammatical and syntactical roles, namely the elements of ritual that define its visual attractiveness. I have often asked what is it about ritual that people enjoy watching? Among other neglected items, I find that prevailing theories fail to address this, and in order to highlight this I require a theoretical apparatus to explain the stratification of the two rituals at hand as well as the rite that drew them together, which struck me as analogous to a narrow passageway connecting two large lakes. Among the aspects of these rituals that I find unquantifiable, unable to assume definite structural loci, are intention, reverence, visual aesthetics and agency. While these contribute to the overall sense and meaning of religious rituals, which is to say to their deep syntax, they are resistant to grammatical identification.

308

Frederick M. Smith

The two rituals I intend to examine, attending particularly to their juxtaposition, are, first, a Vedic śrauta ritual (technically two of them) in which a young man is initiated into a lifetime of ritual performance, and second, the formal opening of a temple dedicated to the Maharashtrian saint known as Shirdi Sai Baba (1838?1918). The first is called agnyādheya, agnyādhāna (or simply ādhāna), “the setting up of the (sacred) fires,” in the lexicon of Vedic rituals.1 This was immediately followed by the performance of a soma sacrifice called agniṣṭoma, the first in the catalogue of such sacrificial rituals.2 The performance of the two of these together is called by the ritualists somapūrvādhāna, “the setting up of the fires immediately preceding a soma sacrifice.” This is a word that I’ve heard for decades, although I am unable to find it in the ritual literature. The temple opening featured primarily the consecration of the image of Shirdi Sai Baba, called prāṇapratiṣṭhā, “the establishment of breath or life-force (prāṇa)” in the image. I emphasise that this was a formal opening, because very little existed of the actual temple when this was performed, except the specific site of the primary image consisting of a freestanding stage and the image itself, which had very recently been placed on it. This was covered by a makeshift roof that was surely replaced after the construction was renewed. However, this structure, which appeared to be hurriedly erected (and entirely lacked security), served the overall ritual purposes well, because the Vedic śrauta ritual,3 which was performed in front of it, has a history of being performed on bare ground. These two major rituals, which assumed the functional pragmatics of building blocks in a larger ritual formation, were performed in the small town of Dodamarg in Sawantwadi district in southern Maharashtra, barely one kilometre from the Goa border. The chief patron or yajamāna of the śrauta ritual was a young man in his late 20s hailing from the best known family of śrauta ritualists in Goa. His name was Ballal Apte, the son of Hari Mahadev Apte of Mapuca, in north central Goa, about an hour and a half from Dodamarg. Hari Apte has performed a large number of soma sacrifices since 1980 (when in fact I observed his own somapūrvādhāna), and is highly respected in the communities of vaidikas in southern Maharashtra and northern and coastal Karnataka. Similarly, Hari Mahadev Apte’s father, Mahadev Hari Apte, performed an agniṣṭoma in 1969, after setting up his Vedic fires in an agnyādhāna in 1963. Although the senior Apte is still alive, he has not performed a soma sacrifice since 1976. The officiating crew for the present somapūrvādhāna, which included experts in the ritual and recitation of the Yajurveda, Ṛgveda, and Sāmaveda, came to Dodamarg from their homes in Gokarna, Honnavar, and Mattur in western and coastal Karnataka. 1 Moody 1981, Krick 1982 2 See Caland & Henri 1906–1907; Śrautakośa 1958. 3 The word śrauta means “Vedic”, a derivative of the word śruti, “that which is heard”, another word for the Vedas.

Historical Symmetry and Ritual Asymmetry

309

Although the Apte family has historically followed in their ritual practice the texts of the Hiraṇyakeśi school of the Taittirīya rescension of the Kṛṣṇa- or Black Yajurveda, they have employed the ritual texts of the Āpastamba school of the Taittirīya rescension of the Kṛṣṇayajurveda ever since the senior Apte commenced his śrauta practice in 1969. This has been entirely for reasons of convenience. Not only are the two schools, Hiraṇyakeśi and Āpastamba, extremely close in their ritual prescriptions, meaning that the Apte family did not venture too far from their own school, but the Hiraṇyakeśi school has practically ceased to exist as an active school of ritual performance. Its adherents, found only in southern Maharashtra, have almost entirely ceased to perform Vedic ritual. Thus, H.M. Apte and his descendants have been unable to field a crew of competent ritualists of the Hiraṇyakeśi school to officiate in their rituals. In addition to this, the ritual texts of the Āpastamba school are particularly lucid and complete. Indeed, the Apte family (and many other śrauta ritualists) closely follow the prayoga or libretto for their bimonthly new and full moon sacrifices composed according to the ritual instructions of the Āpastamba Śrautasūtra by Vāman Śāstrī Kiṃjavadekar, published in his Darśapūrṇamaprakāśa.4 The yāgaśālā or sacrificial arena was constructed on the bare ground about five or six metres in front of the marble stage containing the image of Shirdi Sai Baba, which was sculpted in the pose in which he is virtually always depicted: sitting on a stool with the right leg crossed over the left, and heavily draped in orange silk cloth. The front edge of the stage, facing west, featured two finished staircases on the left and right from the ground to the top of the stage, a height of less than two metres. The general public was forbidden entry into the sacrificial arena, a prohibition that was largely observed. The stage was positioned at the eastern end of the yāgaśālā, where the yūpa, the sacrificial post to which animals to be sacrificed were to be eventually bound, was erected. The eastern direction is the most important in the Vedic manner of thinking, being the direction of the sunrise and of the gods. The site bearing the image of Shirdi Sai Baba, then, extended in an easterly direction beyond the āhavanīya, the eastern fire that receives the preponderance of offerings to the deities. Because the image of Shirdi Sai Baba rested, in its pose of ease and elegance, on a raised platform, it conveyed a clear sense that Shirdi Sai served the sacrifice as a final arbiter and supervisor. This speaks for itself in the process of hierarchical ordering. The area between the yāgaśālā and the stage was public space, even if manoeuvring within it was often difficult because it was constantly used for ritual preparations, and occasionally for ritual itself. Largely this was the provenance of the temple staff, not the staff of the śrauta ritual, the preparations for which were confined almost entirely to the yāgaśālā itself. Although the Vedic ritual complex – first the ādhāna, then the dīkṣā or initiation for the soma 4 Pune: Ānandāśrama Sanskrit Series, no. 93, 1924.

310

Frederick M. Smith

sacrifice, followed by the performance of the pravargya and upasad rituals over three days (van Buitenen 1968, Houben 1991), then the agniṣṭoma itself over a day and a half – kept the staff of Vedic officiants quite busy, the events surrounding the temple consecration, by contrast, were more sporadic and relaxed. They were also timed so that the final acts of the pratiṣṭhā and the sūtya or soma pressing days of the śrauta ritual would occur on a weekend, thus attracting a larger crowd, which translates into a more substantial patronage base. I will shortly say a few words about the patronage of both of these events. It is not essential here to review the details of Vedic śrauta ritual; most of this is available in published form and is not necessary for the arguments I am making here (see, e.g. Kane 1974; Śrautakośa). What may be less well known are the details of the ritual of temple consecration, even if it can be readily surmised that it is a ritual which dates to an era much later than the śrauta sacrifice. The points I wish to make here, however, have less to do with the minutiae of śrauta ritual or temple pratiṣṭhā or pratiṣṭhāpana rites than with the dynamics of their simultaneous performance, shared space and reception, and why they were performed together on this occasion. Part of my objective is to investigate further into a question examined in a series of seminars held at the University of Heidelberg in 2007. The question was “How does ritual work?”, a question that is very basic, the answers to which were particularly elusive. I hope here to offer a positive contribution to this discussion. Considering the theoretical positionings of others, I would like to add a number of fresh thoughts.

Image 1: Offering on main fireplace (āhavanīya). Agniṣṭoma, Dodamarg March 2007. Photo by the author


311

Image 2: Vedic triple fire and altar (tretāgni and vedi). Agniṣṭoma, Dodamarg March 2007. Photo by the author

312

Frederick M. Smith

Image 3: Chief Ṛgveda officiant (hotā) sitting by vedi. Agniṣṭoma, Dodamarg, March 2007. Photo by the author

Several reasons may be cited for the simultaneous performances of these rituals. First, considering the Vedic initiation and soma sacrifice, the sons of Hari Apte were raised with the expectation that they would eventually adopt the lifestyle and practices of Vedic śrauta specialists. This proved to be a perfect occasion for Ballal’s ādhāna for a number of reasons. Importantly, he was in his late 20s and appeared to be enthusiastic about it. Furthermore, he was married and has a son of his own, thus meeting the oft-stated, if not always observed, criteria of being jātaputra kṛṣṇakeśa (has had a son born to him and still has dark hair). Second, he was sufficiently well versed to undertake the ritual in much the way his father and grandfather had. Finally, the opportunity arose to have Ballal’s ādhāna performed in circumstances in which the family was not obligated to meet all the expenses, due to agreements concerning the expenditure of funds raised by the Shirdi Sai Baba temple’s administrative committee. As for the Shirdi Sai Baba temple pratiṣṭhā, it is not an exaggeration to say that the sect that has sprung up around this saint, who died more than ninety years ago,


313

is quite likely the fastest growing Hindu sect in India. It has pretensions to cutting through sectarian and religious divides; pretensions, it appears, because the ritual and (most of the) theology of the sect are distinctly brahmanical in spite of the fact that Shirdi Sai Baba himself might have been born a Muslim (his origins are obscure), and in any case he did not appear to be an observant Hindu, and certainly not a brahmanically observant Hindu.5 Nor is there clear evidence that he advocated any specific form of Hindu practice. Indeed, he seems to have participated in Sufi spirituality as much as characteristically Hindu renunciate practice. Although the sect that bears his name has wrapped itself in distinctly brahmanical garb in contemporary India, it was a long time coming; this was not dominant for the first half century after his death in 1918. Thus, the sect has made claims to spiritual universality and to rejection of caste and religious prerequisites in its participation. The town of Shirdi itself, in northern Maharashtra about two hours from Nasik (120 kms), is built around the Sai Baba temple complex, and is possibly the fastest growing spiritual centre in India, and among its best attended pilgrimage sites, rivalling Nathadwara (in Rajasthan) and Vaishno Devi (in Jammu), if not quite Tirupati (in southern Andhra Pradesh). In a visit to Shirdi in March 2007, shortly after the sacrifice in Dodamarg, I discovered that it is probably the second wealthiest religious pilgrimage centre in India (again, after the Venkateshwara temple in Tirupati). Thus the patronage base has become widespread, accounting for the fact that for the last decade or so, according to my informants in Shirdi, close to a hundred temples every year are constructed across India dedicated to Shirdi Sai Baba. It has gained the reputation of being a camatkāra site, to use the local term, where frequent reports of miracles and other wondrous events are believed to bear directly on the devotee after visiting the site. According to my own informal estimate, no less than ten thousand people visited during each of the three days I was there. None were special days on the ritual calendar. According to my informants among the śrauta ritualists, the planning for the two events began several months prior to the commencement of the soma sacrifice on 28 February 2007. Because Hari Apte and his family are well known in the community of vaidikas in the area, Śrī Apte was informed that a temple dedicated to Shirdi Sai Baba was about to be constructed in the small market town of Dodamarg. Informal discussions commenced, not terribly unlike negotiations for a marriage in India, regarding the performance of a soma sacrifice on the building site. Without much disagreement, the Apte family decided that this was a good occasion in which to initiate the young Ballal into a lifetime of Vedic ritual practice. Regardless of the importance of the financial arrangements forged between the Apte family and the Shirdi Sai Baba temple committee (much of which remains unknown to 5 In spite of the work of Rigopoulos (1993) and Ruhela (2000), an authoritative account of the life of Shirdi Sai Baba has yet to be written.

314

Frederick M. Smith

me), both had their own, somewhat asymmetrical reasons for performing their rituals synchronously. We shall shortly examine these reasons.

The Vedic Ritual and Ritual Analysis It would not be remiss to mention here that one project I’ve always imagined, but which might be too much for one scholar, is a complete analysis, from the point of view of ritual theory, of the Vedic agniṣṭoma, a ritual of mind-boggling complexity and organic elegance that has clearly been spliced together from dozens, if not hundreds, of spare parts.6 It appears to be organised in a way that was intended to flow seamlessly. Indeed, from a temporal distance of three millennia (or more), we can only praise the redactors of this sacrifice for their mastery of ritual symbolism and syntax. But we have not so far fully deciphered it. The attempts by Frits Staal and a few others to ferret out a ritual syntax are praiseworthy, as are the attempts by Stephanie Jamison and others to understand selected parts of the whole,7 but these have all been limited in scope, segmenting the Vedic ritual and examining a few elements of it microscopically.8 It is a daunting task that requires mastery of different disciplines to piece together a single large-scale ritual out of disparate parts, parts that were forged in different fires (literally and figuratively), many emerging from varying intentions. Although we might note that the forging of the agniṣṭoma proved to be a successful job of ritual artistry, an observation that we might also make of a temple prāṇapratiṣṭhā, can we also claim the same of the weaving together of the two rituals performed at Dodamarg in 2007? Ritual studies has been in part marked by attempts to typologise ritual, establish categories, and develop ritual syntax. In order to help interpret the confluence of events at Dodamarg, let us examine a few strands of ritual theory. In a working paper on ritual syntax written for use at a major conference on ritual studies in 2008 at Heidelberg, Axel Michaels (2010) cites Michael Oppitz’s (1999) use of the metaphor of building, invoking the image of a “construction plan.” Oppitz asserts that – – – –

“[r]ituals are pieced together – from individual units The units of rituals are prefabricated structural elements The prefabricated structural elements are mobile and transposable The transposition of structural elements follows a design

6 See the table of contents of Caland and Henri 1906-07 or the table of contents of the Sanskrit (part 2)

and English (part 2) sections of the Śrautakośa on the agniṣṭoma, all of which conveniently break down the agniṣṭoma into its constituent ritual units. 7 See Jamison 1996 for her efforts to understand the symbolic role of the woman in the Vedic sacrifice. 8 See Staal 1979, Michaels 2010, for attempts to argue the case for ritual syntax.


315

– The prefabricated parts pieced together according to the design produce a recognizable product – the ritual in question.” It may be difficult to see the simultaneous performances in Dodamarg as a single seamless ritual, especially if we use as points of comparison other more closely allied rituals that display seamless, if unexpected, harmony in their simultaneous performance. An example within the Vedic ritual corpus is the chanting of two different texts, the Ṛgveda and the Sāmaveda, in the pravargya, an important rite preparatory to an agniṣṭoma. Although the Sāmaveda is largely a musical variant of the Ṛgveda, they were not intended to be chanted and sung simultaneously. However, in the pravargya they are orchestrated to serve the same ritual end, the heating of a ceramic pot called mahāvīra (“great hero”) and, somewhat unexpectedly, are not at all dissonant. In the present case, however, the dissonance was evident to most of the observers, perhaps particularly to those who were familiar with the śrauta ritual (which was less than a handful beyond the performers themselves). The reason for this was because the organisers and ritualists made no attempt at orchestration or choreography. The assumption that their roles were circumscribed by different strata of Indian religious history and religiosity placed them at such an extreme ideological distance that the possibility of choreographing them to flow into each other as a single visual or meaningful unit was not considered. Nevertheless, the viewer was privy to a recognisable ritual product. Thus, although the agniṣṭoma and the temple pratiṣṭhā were not performed with an eye to orchestration or mutual referentiality in the sense that the rituals would refer to each other, they nevertheless offered an overall sense, vaguely grasped by participants and observers alike, of a higher purpose served by their oddly juxtaposed simultaneous performance. A few elements of the agniṣṭoma may be identified to serve as an analogy, namely the three days of performance of the pravargya and upasad prior to the soma pressing day,9 or, more intimately and tightly woven into the soma pressing and offering rites, the performance of the three animal sacrifices in the body of the agniṣṭoma.10 These were unquestionably rituals representing different performative strata of the early and middle Vedic periods that were knitted together effectively in antiquity. While it’s true that the śrauta ritual and the temple consecration are prefabricated units that proved to be mobile, and were in fact planned out in order to follow a design and produce a recognisable product, the two did not employ the same set of actors, which is to say they did not share their staffs of ritualists, although Hari Apte, the father of the yajamāna in the śrauta ritual, assumed an active bridging role in the Shirdi Sai Baba temple consecration.

9 This is discussed by van Buitenen (1968) and Houben (1991). 10 The composite character of the agniṣṭoma is evident from this.

316

Frederick M. Smith

Before we seek to further contrive a grammar to this, let us examine the dynamics of the joint performance. First, the Shirdi Sai Baba temple committee sought the sanction of leading religious figures in the area. The Apte family, as the last remaining śrauta ritualists in Goa, is certainly among them. In addition, Hari Apte and the other members of the family are eminently approachable, a characteristic that is still rather uncommon among śrauta ritualists in India. This friendly and socially relaxed attitude of the Aptes is consonant with the intention of the Shirdi Sai Baba sect, which purportedly embraces members of all communities, at least among devotees, if not among its sectarian leaders or priests. Second, the committee was convinced that the performance of a śrauta ritual on the grounds of their temple would add an aura of sanctity to the temple itself that would remain after the performance was completed and confer enhanced prestige on the consecration. Śrauta ritual is of course an oddity in the Hindu world; it is not well understood, yet it is conferred the greatest inviolability. Its antiquity is itself an argument for its sacredness in the eyes of most practising Hindus. On the other side, the Apte family and some of the team of ritualists (those from Mattur, especially) believed that the exposure they would receive from a well-publicised local event would augment their prestige and enhance their patronage base, which, in today’s politically charged anti-imperial world, is the general public. Although the Apte family operates a well-attended Hanuman temple in Mapusa, they are not wealthy; indeed, they depend on outside patronage to perform their frequent śrauta rituals. At present, four members of their family perform śrauta rituals and maintain the sacred fires (in a large hall in their home). In spite of the relatively high profile of the Apte family in the small world of śrauta ritual, they are well aware that practically the entire audience in Dodamarg, including the Shirdi Sai Baba temple committee, and they themselves, no longer envision their social world or shared values as inscribed in the śrauta ritual to which they have committed their lives. Rather, they see the śrauta ritual as an isolated episteme, totally selfreferential, thus as a closed system resistant to analysis, classification, and deconstruction. They nevertheless see it as a point of entry into an exclusive experience, a liminality based on their unique eligibility (adhikāra) to perform these rituals. In this way, their knowledge, special gestural abilities, and intent provide them with access to a distinctive quality of direct experience (pratyakṣa) that does not rely on a closed system of symbols. As such, this is not an experience that can or should be explained or proven through anumāna or even śabda pramāṇa, which is to say through formal logic or acceptance of the Veda as a final arbiter. Rather, it is an inner world of complex self-referentiality.11 Although I am arguing here that their 11 Such notions of self-referentiality were articulated in papers by Jan Houben, “Formal Structure and Self-Referential Loops in Vedic Ritual”, and Johannes Bronkhorst, “Ritual and Holistic Utterances”, both in this volume.


317

experience of the śrauta ritual is that of a complex epistemic whole in which their agency, based on their adhikāra, knowledge, and intent, is a major determinant, they still recognise that parts of the rituals occur elsewhere, outside the closed circle of this self-referentiality. The performers of this ritual understand, and have long understood, that the śrauta ritual possesses a grammar and a syntax, that it is a ritual warehouse subject to structural rearrangement. The meaning of the śrauta ritual, then, to the Apte family and most other śrauta ritualists, lies in its embodiment; śrauta ritualists are in a sense physically and psychically marked by it. This self-referentiality, then, becomes a form of embodiment, it is what these śrautins experience through the body, through the ritual. Their generally undisguised satisfaction after completing a performance testifies to the fact that their experience is in fact within the realm of embodiment. Its meaning is also social. The self-referentiality of the ritual, and its gateway into an inner world that is off limits to others, does not, and cannot, separate the ritualists from their social environment. In spite of the physical purity that the yajamāna and the other ritualists maintain during the performances, they continue to interact with others during the performances, including, in this case, with the Shirdi Sai Baba temple committee, their families, the large support staff, and even interested parties such as myself. The juxtapositions of their interior and external negotiations are intrinsically based on moral and ethical codes that they manifest and generate in their performances. Their understanding of themselves as moral beings is expressed in terms of their knowledge, ritual skills and social position. Without any sense of inconsistency, they see these as instruments to generate prestige, even as they remain content to locate themselves hierarchically in a secondary position with respect to prevailing cultural arrangements. In Dodamarg, the ritualists understood that the agniṣṭoma would be performed at the feet of Shirdi Sai Baba. In another recent agniṣṭoma that I observed, performed in a public park in London in 1996 (cf. Smith 2000), the yajamāna, Gokulotsavajī Mahārāj of Indore, whose understanding of Vedic ritual and prestige in his community approximated that of the Apte family, also placed the śrauta ritual in a secondary position spiritually and politically. Spiritually, Gokulotsavajī explained lucidly and often, the agniṣṭoma was an aṅga or limb of Kṛṣṇa devotion, while politically it was subservient to the charismatic presence of Morari Bapu, one of the most popular kathāvācaks or religious lecturers in India (see Lutgendorf 1991), who was lecturing to enormous numbers of Hindus just across the street in the same public park. In sum, as the religious position of Vedic śrauta ritual has become superannuated, its social and ritual positionings have also declined.

318

Frederick M. Smith

The Mahābhārata as a Paradigm I have recently completed a translation of the Āśvamedhika parvan of the Mahābhārata, the fourteenth book of the great Sanskrit epic (Smith forthcoming). As the title suggests, the main topic in this parvan is the aśvamedha or horse sacrifice. This particular horse sacrifice was performed by Yudhiṣṭhira, the eldest of the victorious Pāṇḍava brothers after the great (and surrealistically bloody) war concluded, in order to legitimise his kingship. Although we are led to believe that Yudhiṣṭhira’s aśvamedha was performed more or less according to ritual injunction, the sacrifice itself assumes a secondary, even tertiary place in the parvan. This subordination occurs often in the parvan; the reader is constantly reminded of the fact that the actual performance is not particularly unusual, that the routinisation of even this most exotic of śrauta sacrifices carries with it very little to report, and perhaps no small amount of ennui. Repeatedly, the account of the aśvamedha is deflected in favour of events and textual interruptions with greater contemporary moral urgency and public understanding. It is not necessary here to describe this fully, but it will help us understand the relationship between the two strata of ritualising in Dodamarg if we realise that it is consistent with a process that was evident two thousand years ago. If the Āśvamedhika parvan is about anything that is readily identifiable, it’s about the dilemma of performing Vedic sacrifice in a world that has moved beyond it morally, spiritually, and ritually. The dissonance between the aśvamedha and other forms of brahmanical ritual practices was evident even then. At the end of the parvan, for example, it is recommended that seeds be offered into a fire, a common practice in later purāṇic and tantric Hinduism. Similarly, the text constantly exhorts against animal sacrifice, a practice that appears to discredit Vedic ritual in the eyes of early first millennium C.E. brahmans. I understand from many sources, not least my own research into deity and spirit possession in India – both textual and ethnographic – that what brahmans preach is often not practised (Smith 2006). What I saw in Dodamarg resonates with the situation in the Mahābhārata; both exhibit a series of asymmetries between two ritual complexes. These lie in the areas of public recognition and patronage, the extent and kinds of brahmanical education and knowledge required for their performance, the sense of culturally inscribed morality that in part transforms these rituals into independent epistemes, and the manner in which these epistemic units agree with the sense of ritualising or ritual intent (cf. Humphrey and Laidlaw 1994) and with the moral underpinnings, ambiguous though they might be, of these rituals. It is possible to argue for all of these as well in the Mahābhārata. The MBh aside, however, what I saw in the bewilderment of the spectators at Dodamarg, an audience consisting of potential donors, was a mixture of incomprehension and awe. Many were curious and wondered what it all meant. Most, however, came for the temple consecration and paid little attention to the agniṣṭoma. Regardless of whether or not the ritual structure of


319

the agniṣṭoma, which is to say the ritual syntax that Staal and others have recognised and addressed, was comprehended at all, there was an overriding recognition of and concern with agency that was above all a question of morality. The structure, intent, and asymmetries that characterised the simultaneous performance of these two rituals cannot fully account for the problem of recog-nition, or, more accurately, non-recognition, that was apparent in the spectacle of this event. Two other considerations were of overriding importance to the specta-tors: agency and simple visual recognition. It should be clear by now that the audience, supported by the machinations of contemporary culture in India, resona-ted more clearly with the temple pratiṣṭhā and the figure of Shirdi Sai Baba than with the agniṣṭoma and the yajamāna, even though the members of the audience knew, on the whole, as little about the details of the pratiṣṭhā as they did of the agniṣṭoma. In order to understand the effectiveness of this juxtaposition, then, it is necessary to examine, albeit briefly, agency and visual recognition. To whom did the agniṣṭoma belong? Who reaped the benefit of it? Textually, it is the yajamāna, and it appears that this was viscerally understood by the audience. This is one reason it was rejected, and why it really has little place in the public domain except as an aṅga or subsidiary of the temple pratiṣṭhā, which really does have a sense of shared agency and effect (phala). This is perhaps particularly the case in the Shirdi Sai Baba sect, with its openness to non-Hindus and people of all social and economic categories, in spite of its brahmanisation in recent decades.

Vedic Tradition as Local Tradition This amalgam of Vedic tradition and local tradition, even if local in this case is in fact transregional, legitimises the local by placing it within the orbit of Sanskritic culture. Simultaneously, this arrangement confers on the patrons a relevance that is important to legitimising them to the local culture. By making themselves conspicuous in the temple pratiṣṭhā, which we can see from Image 6, the Apte family participated in both levels of ritual. What I found decisive in my studies of deity and spirit possession is that there is nearly always reciprocation between folk and Sanskritic culture in India, or between asymmetrical levels of Sanskritic culture, as we find here. Even if the Apte family appeared to be somewhat intrusive in the pratiṣṭhā, the amalgamation works in part because the patrons legitimised themselves on both levels. Their inaccessibility in the agniṣṭoma was rescued by their intimacy in the pratiṣṭhā, where they set aside the walls around them necessitated by the śrauta restrictions, thus allowing their humanity and personal qualities to attract an audience that was otherwise sacrificed at the altar of classical ritual structure. What makes this ritual arrangement “work”, in the final analysis, is that the śrauta ritual was forced to prostrate before the image of the non-Sanskritic Shirdi

320

Frederick M. Smith

Sai Baba. Although, as already noted, the temple apparatus and ritual of Shirdi Sai has been gradually appropriated by brahmanical forces, there remains an undercurrent of knowledge, a memory trace, that there was something fundamentally nonbrahmanical about him. The visual representation of Shirdi Sai Baba has become standardised. In the dozens of Shirdi Sai Baba temples that are popping up every year across India, he is depicted in the same way: an image carved of white marble, as I mentioned earlier, seated on a stool or chair with his right leg crossed over his left. This standardisation inscribes a ritual environment that ensures reliability. As such, the Shirdi Sai Baba temple in Dodamarg did not need the Vedic śrauta ritual in order for it to “work”; such complicated and grandiose addenda are not required for a pratiṣṭhā in other Shirdi Sai Baba temples. But this one is believed (at least by its local sponsors) to carry more prestige than others. In its inception at least, it absorbed the image of the Vedic yajña. In this case, complementary processes of localisation and Sanskritisation made both the Vedic agniṣṭoma and the inauguration of the Shirdi Sai temple “work”. Image 4 depicts the image of Shirdi Sai Baba after the pratiṣṭhā, in royal garb, overseeing the śrauta ritual, and Image 5 depicts the pūjā apparatus below the stage, between the stone image above and the śrauta ritual behind it. Let us examine one more rite in this complex ritual formation, a rite that appears to have had the effect, probably intentional, of linking together the Vedic agniṣṭoma and the temple pratiṣṭhā. The rite consisted of the construction and offering of a rice flour effigy of Shirdi Sai Baba. The action in this comparatively inconspicuous ritual took place entirely in the area between the stage and the yāgaśālā, an area otherwise used for pūjās that were secondary parts of the larger pratiṣṭhā ceremony, but which served as a bridge between the two primary ritual arenas. Image 7 depicts the effigy, with several rice flour balls (puroḍāśa, offered in the śrauta ritual) and a few flowers, in a wicker basket. It was not part of the regular pratiṣṭhā ceremony, and the ritualists were not eager to explain it to me. Neverthe-less, the presence of puroḍāśas and flowers, two of the most characteristic offerings of śrauta ritual and temple ritual, respectively, strongly suggests that the idea of this rite as linking the two was intentional. This brief rite concluded with the offering into a laukika (ordinary, which is to say non-Vedic) fire, about two metres away, of the puroḍāṣas and the head of the effigy, which was cleanly cut off with a knife by one of the ritualists. This also reinforces the idea of this as a linking rite, with offerings, both Vedic and non-Vedic, into the laukika fire. Of all the ritual events that contributed to the consecration of the new temple, this attracted the least amount of public attention. It was, however, of considerable interest from the perspective of Sanskritic ritual, and indeed of ritual dynamics, first because of its function as a linking rite, and second because it featured a pratinidhi, a ritual substitution of a representation of the figure at the centre of the temple worship: a rice flour effigy of a marble statue of a saint who lived a century


321

ago. As an effigy, it was called a puttalī or mannequin by the ritualists, a term found in Tantric, rather than Vedic, texts.12 I cannot comment on the normativity of this rite at a pratiṣṭhā of Shirdi Sai Baba temples, even if it appears that it could have been unique here because of the presence of the śrauta ritual. As I have already noted, however, the ritual within Shirdi Sai Baba temples has become brahmanically normativised in spite of the stridently unbrahmanical figure of Shirdi Sai Baba himself. Regardless, then, of whether this rite is performed elsewhere, I found it to be synchronous with the Vedic soma sacrifice being performed under the same roof.

Image 4: Shirdi Sai Baba after consecration, Dodamarg, March 2007. Photo by the author

12 For references to this in the Tantrarāja Tantra, Īśānaśivagurudevapaddhati, Khaḍgarāvaṇa, and current practice in Kerala, see Smith 2006: 476, 531ff..

322

Frederick M. Smith

Image 5: Pūjā in front of stage. Photo by the author

Image 6: Apte family at temple pratiṣṭhā (Hari Apte, left centre pointing; Ballal Apte, holding grass bundle; Ballal Apte’s wife). Photo by the author


Image 7: Effigy of Shirdi Sai Baba Photo by the author

323

324

Frederick M. Smith

Visually, the construction of the effigy was quite striking, with flowers and Vedic rice flour balls (puroḍāśa). A great deal of care was taken in creating this tableau, even if it was soon destroyed. I found this to be consistent with the pratiṣṭhā but a signal contrast with the unadorned offerings of the Vedic sacrifice, which lack visual attractiveness to a contemporary audience. Aside from the Vedic ritual officiants and the curious foreigner whom some of the officiants, as well as the father of the yajamāna, have known for more than twenty-five years, and who has made at least part of a career studying Vedic ritual, no one in attendance put forth any effort to try to follow the unfamiliar and visually unadorned and alien proceedings of the agniṣṭoma. In this way, the visual image of the offering of Shirdi Sai Baba was quite arresting, supporting the enterprise of the Vedic ritual both aesthetically and to some extent financially.

The Theory of Equivalence/Equivalents After alluding to notions of the visual attractiveness of these rituals and how crucial this is to performer and observer alike, I would like to offer a few thoughts on the aesthetics of ritual and the importance of visual attractiveness as a component of ritual. I am particularly thinking of the “theory of equivalence” developed by the early twentieth-century photographer Alfred Stieglitz (1864-1946), who was as well known for being married to the American artist Georgia O’Keeffe as he was for his own innovative work in the field of photography. Stieglitz’s theory of equivalence was subsequently refined by one of his protégés, the photographer Minor White (1908-1976). This theory, I believe, should prove useful for deepening our understanding of ritual. I must emphasise that by suggesting this I’m venturing into completely uncharted terrain. Although such theorising quite likely contains dangers of which I am unaware, it may nevertheless help rein in the visual aspect of ritual, an aspect of it that is not very often addressed or theorised in either ritual studies or Indology. I must also emphasise that this is not “visual anthropology”, a subfield (if not a particularly well-trodden one) within anthropology that is concerned with ethnographic filmmaking and the anthropological analysis of film, television and other forms of mass media. Rather, the theory of equivalence as articulated by Stieglitz and White falls within the orbit of artistic conception (although it seems to me that it would behoove theorists of visual anthropology to seriously engage with this theory). Mike Johnston, a photography critic, writes of this theory: “The term is simply the observation (or the assertion) that a picture of some specific thing can evoke other things altogether, in terms of feeling, meaning, associations, or form” (Johnston 2006).13

13 For another attempt to apply Stieglitz’s theory of equivalents in academic discourse, see Hyatt 1992.


325

What I propose to address now, utilising this theory, is the visual aspect of the Vedic agniṣṭoma and the Hindu temple pratiṣṭhā held in Dodamarg. What I am striving to capture is the sensitivity to form and arrangement that these rituals visually engage, which is to say how they inspire an emotional resonance. It is clear that the presence of images of deities (or in this case a saint), a multitude of flowers, and a colourful layout – in other words a preponderance of visual gaiety – contribute greatly to the mood. As a cognate to this, we should once again invoke the Āśvamedhika parvan of the Mahābhārata. Chapters 58 and 87 of this parvan describe the attractiveness of festivals and sacrifices at that time, perhaps two millennia ago. Before we enter into this more deeply, however, let us discuss in greater detail Stieglitz’s theory of equivalence. Equivalence or equivalents was the subject of a great deal of discussion at wellknown New York art galleries such as Gallery 291 during the teens of the last century. It was initially energised by the ideas of the Russian painter Vassily Kandinsky, especially the belief that colours, shapes and lines reflect inner, often emotive “vibrations of the soul.” In his cloud photographs (1922-1931), which he termed Equivalents, Stieglitz emphasised pure abstraction, adhering to the modern concept of equivalence or, as he termed it, equivalents, planting this idea firmly in the vocabulary of photography.14 Declaring that his photographs were “equivalents” of his philosophy of life, he developed a theory which proposed that abstract forms, lines and colours could represent corresponding inner states, emotions and ideas, building on the views of Kandinsky. In a 1963 essay, “Equivalence: The Perennial Trend”, Minor White describes a photographer of an “equivalent” as one who “recognized an object or series of forms that, when photographed, would yield an image with specific suggestive powers that can direct the viewer into a specific and known feeling, state, or place within himself” (see also Gantz ND). Beaumont Newhall, one of the great historians of photography, writes that “White’s goal was to make photographs that extend beyond the subject. Surface appearance, although of secondary importance, is essential, but the image must be transformed into a new event, to be discovered by the viewer” (1982: 281). White hoped that the viewer would share his conceptions of these encoded elements that would persuade the viewer to see the image in these terms. As was the case with Stieglitz, the photographic process for White was in this way constantly embedded in both his personal and spiritual life. As we have seen, the visual field of the audience in this complex of rituals in Dodamarg, which consisted largely of religious, if educated, devotees, saw the unique juxtaposition of ritual elements before them, but personally identified with only one of them, namely the pratiṣṭhā at the new Shirdi Sai Baba temple. This is not to say that they 14 For the photographs themselves, see: http://www.phillipscollection.org/american_art/artwork/StieglitzEquivalent_Series1.htm.

326

Frederick M. Smith

shunned the śrauta ritual; they certainly did not. Indeed, they treated it respectfully; but only one of these two possessed the lines, shapes and colours that were inscribed in their sense of what was personally meaningful in a religious sense. Only the temple pratiṣṭhā possessed the evocative power to inspire their emotional engagement. It is this instant visual evaluation, at least in this case, that proves crucial in generating not just emotional support, but financial support as well. The Vedic agniṣṭoma is ultimately dependent on other more recognisable forms of ritual for its survival in the modern world, and by this I mean visually recognisable. In this way the temple pratiṣṭhā enhanced the agniṣṭoma, at least within the public domain, while the opposite can be asserted in this respect. That said, we cannot deny that a Vedic soma sacrifice is in certain respects picturesque, and it is appealing for that reason – note, for example, the impressive photos in Staal’s Agni volumes (1983) (see also Images 1–3 above). Yet its visual presentation is neither comprehensible nor religiously engaging to practising Hindus of (at least) the last century or so.15 It has a visual feel of inward action, of ritual being performed solely for the benefit of the performers; indeed, this is how it is depicted in the Vedic ritual texts. Above all, as exemplified in the Dodamarg agniṣṭoma, this sense of public alienness or even disaffection is highlighted by its characteristic absence of images of deities, its comparative dearth of colour and action that is comprehensible and effective only to a closed circle of highly educated colleagues. Thus, the aesthetic appreciation of the śrauta ritual is circumscribed by the sacrificial arena and personnel involved in the sacrifice itself (again see Images 1–3 for this). Image 8 depicts the crowd with its back turned to the śrauta sacrifice, facing a member of the Shirdi Sai Baba temple committee who is delivering a lecture on the saint and describing the pratiṣṭhā. As mentioned, the Mahābhārata, which contains so much that is recognisable as beautiful and tragic, as attractive and repellent, in Indian culture, contains the prototype for this visual structure. Two passages from the Āśvamedhika parvan exemplify this. MBh 14.87 describes the scene at Yudhiṣṭhira’s aśvamedha. Although this is clearly apocryphal, it expresses an aesthetic of pomp laden upon ceremony that remains paradigmatic to this day. Indeed, ceremony without pomp is of little value in India today, as it was two millennia ago. Examples of this can be replicated in Indian literature, even elsewhere in the MBh. For example, MBh 14.58 describes a festival on a mountain called Raivataka, that Kṛṣṇa passed on his way to Dvārakā, returning home after the great war, which is even more grandiose than the scene at the aśvamedha. But we will limit ourselves to the description of the aśvamedha. The following account from MBh 14.87 is somewhat edited down. 15 In spite of misunderstanding a number of points of the agnicayana that Staal, et al., make clear (cf. Staal 1983), McCauley and Lawson (2002: 150 ff.) nevertheless note correctly that the agnicayana – and we can add other Vedic śrauta rituals – “neither contains sensory stimulation nor incites emotional arousal that is at all out of the ordinary compared with other rituals in this tradition” (2002: 152).


327

Image 8: Crowd facing stage Photo by the author

At the commencement of the sacrifice, many eloquent disputants articulated their logic, wishing to defeat each other. The kings observed the ultimate injunctive force of this sacrifice designed by Bhīma, joy of the Kurus, as if it were for the king of the gods himself. They saw there the golden archways, and many couches, seats, and luxury items ornamented with gems. These princes saw water and storage pots, eating vessels, saucepans with handles, lids, and covers, none of them not made of gold. They saw sacrificial posts made of wood and ornamented with gold…. Lord, these kings saw a selection of animals born on land and in water, all brought together there. They saw cows and buffalos, including females and those put out to pasture, aquatic animals, predatory animals, birds, those born from embryos, egg-born creatures (besides birds), sweat-born creatures, those produced from digging the earth (insects), and animals from the mountains, shores, and forests. Seeing the entire sacrificial area decked out gaily with animals, herds of cattle, and grain, the princes entered into a state of complete amazement. A rich setting of much fine, delicate food was prepared for the brahmans and commoners. After a hundred thousand brahmans had completed their meal, drums were beaten again

328

Frederick M. Smith

and again, like thunder in the clouds. This thundering went on repeatedly day after day. In this way the sacrifice of the highly intelligent dharma king proceeded. Gifts of food were there, comparable to mountains, O king, and the people saw small rivers of yogurt and lakes of melted ghee. O king, at the great sacrifice of that king. The entirety of Jambūdvīpa was seen in one place, including its many townships. Thousands of communities of people were there. Many kings came, bull of the Bharatas, bringing money, wearing garlands and bright, well-polished earrings, and served the foremost of the twice-born by the hundreds and thousands. This idealised account of the scene of the princes in attendance at Yudhiṣṭhira’s aśvamedha looks, sounds and feels much like a great soma sacrifice today. The Pāṇḍava princes and the common people recognised and enjoyed the festival that inevitably accompanied Vedic ritual of that period, and indeed of all times and places, particularly because, if we are justified in reading the MBh as culturally paradigmatic, the festival opened the proceedings to other forms of ritual and religiosity, and wove together the secular and the religious. In fact, Vedic śrauta ritual by itself was performed only in closed brahman communities. When it was opened up to others, it required other forms of religiosity, other forms of recognisable and comprehensible ritual in order for it to look right and to feel right. By itself, śrauta ritual did not look right two thousand years ago, nor did it look right, at least not exactly right, in 2007.


329

References van Buitenen, J.A.B. 1968. The Pravargya: An Ancient Indian Iconic Ritual. Building Centenary and Silver Jubilee Series, vol. 58; Poona: Deccan College. Caland, Willem, and Victor Henri. 1906–07. L’Agniṣṭoma: Description complète de la forme normale du sacrifice de Soma dans le culte védique. 2 vols, Paris: Ernest Leroux. Gantz, Ryan. ND. “The Transmissions of Minor White.” http://www.sixfoot6.com/ words/essays/minorwhite.htm. Houben, Jan E.M. 1991. The Pravargya Brahmana of the Taittiriya Aranyaka: An Ancient Commentary on the Pravargya Ritual. Delhi: Motilal Banarsidass. Humphrey, Caroline, and James Laidlaw. 1994. The Archetypal Actions of Ritual. A theory of ritual illustrated by the Jain rite of worship. Oxford, Clarendon Press. Hyatt, Kenton S. 1992. “Stieglitz, Martin Buber, and The Equivalent.” History of Photography 16/4: 398–404. Jamison, Stephanie W. 1996. Sacrificed Wife/Sacrificer’s Wife: Women, Ritual, and Hospitality in Ancient India. New York, Oxford: Oxford University Press. Johnston, Mike. 2006. “The Online Photographer.” http://theonlinephotographer. blog spot.com/2006/02/top-ten-number-10_28.html Kane, P.V. 1974. History of Dharmaśāstra, volume 2, part 2. Poona: Bhandarkar Oriental Research Institute. Krick, Hertha. 1982. Das Ritual der Feuergründung (Agnyādheya). Ed. Gerhard Oberhammer. Vienna: Österreichischen Akademie der Wissenschaften. Lutgendorf, Philip. 1991. The Life of a Text: Performing the “Rāmcaritmānas” of Tulsidas. Berkeley and Los Angeles: University of California Press. McCauley, Robert, and E. Thomas Lawson. 2002. Bringing Ritual to Mind. Psychological Foundations of Cultural Forms. Cambridge: Cambridge University Press. Michaels, Axel. 2010. “The Grammar of Rituals.” In this volume: 7–28. Moody, Timothy F. 1981. The Agnyādheya: Establishment of the Sacred fires. Ph.D diss. McMaster University; Hamilton, Ontario, Canada. Newhall, Beaumont. 1982. History of Photography: From 1839 to the Present. Boston: Little, Brown & Co. Oppitz, Michael 1999. “Montageplan von Ritualen”. In: Corinna Caduff & Johanna Pfaff-Czarnecka (eds.). Rituale heute. Theorien, Kontroversen, Entwürfe. Berlin: Reimer: 73–95. Rigopoulos, Antonio. 1993. The Life and Teachings of Sai Baba of Shirdi. Albany: State University of New York Press. Ruby, Jay. 1996. “Visual Anthropology.” In: David Levinson & Melvin Ember (eds.). Encyclopedia of Cultural Anthropology. Vol. 4. New York: Henry Holt and Co: 1345–1351. Ruhela, Satya Pal. 2000. Sri Shirdi Sai Baba.– The Unique Prophet Of Integration. Delhi: Diamond Pocket Books Pvt. Ltd.

330

Frederick M. Smith

Smith, Frederick M. 2000. “Indra Goes West: Report on a Vedic Soma Sacrifice in London in July 1996.” History of Religions 39.3: 247–267. ––– 2006. The Self Possessed: Deity and Spirit Possession in South Asian Literature and Civilization. New York: Columbia University Press. ––– (forthcoming 2010). The Mahābhārata: 14. The Book of the Sacrifice of the Horse. Introduction, Translation, and Annotations. Chicago: University of Chicago Press. Śrautakośa, Sanskrit section, pt. 1, ed. C.G. Kashikar. Poona: Vaidika Saṃśodhana Maṇḍala, 1958. English section, vol. 1, part 1, trans. R.N. Dandekar. Poona: Vaidika Saṃśodhana Maṇḍala, 1958. Sanskrit Section, part 2, ed. C.G. Kashikar. Poona: Vaidika Saṃśodhana Maṇḍala, 1970. English section, vol. 2, parts 1 & 2, trans. R.N. Dandekar. Poona: Vaidika Saṃśodhana Maṇḍala, 1973, 1982. Staal, Frits. 1979. “The Meaninglessness of Ritual.” Numen 26: 2–22. ––– 1983. Agni: The Vedic Ritual of the Fire Altar, vols. 1 and 2. Berkeley: Asian Humanities Press. White, Minor. 1963. “Equivalence: The Perennial Trend.” PSA [Photography Society of America] Journal 29.7: 17–21.

Frits Staal

A Theory of Ritual The Indo-Iranian Fire Offering 1. Introduction The Heidelberg Research Centre on Ritual Dynamics has become a presence in academia and beyond. It is filling gaps in ritual studies that are widening elsewhere and widening fast. One of the most remarkable and least understood aspects of life continues to be studied by disjointed departments of the human and social sciences while ritualisation among the higher animals, which includes the human again, is taken care of by life sciences on the borderline. Heidelberg ritual provides a new focus but Ritual Dynamics proffers an opinion. I quote: “Our focus is on the (re)invention of rituals, transfer and change – which we see as the rule not the exception.” What is a rule and what is an exception is a scientific question. Ritual Dynamics, therefore, needs the assistance of a Science of Ritual. But “science” is a label with variable content in different languages and civilisations: English science, French science, German Wissenschaft, Indic vidyā or śāstra and Japanese gaku do not refer to the same thing. Oldenberg (1919) used it with a twist in a once famous book. In this essay I use the word theory which points at the first of two inseparable areas.1 The first domain is an abstract, underlying area that explores general structures. The second consists of disparate masses of particular happenings and occurrences, facts if you like, though they are apprehended intuitively. There is no harm in apprehension like that and theories must be tested by such facts; but there are no methods or algorithms that describe or explain how theories originate from them. They seem to require an intermediary: the creative minds of human beings who are immersed in facts. One fact is important in our context: later rituals incorporate earlier ones. This remarkable fact about ritual applies to the particular area and period with which we shall be concerned: the northern half of the sub-continent of South Asia, presently Pakistan and northern India; and the period between roughly 1100 and 600 B.C.E., that is, late Vedic. Earlier rites have been incorporated into these later rituals and we shall begin with one of them. We are not interested in this rite for its own sake, but because of the general feature of ritual it illustrates. 1 As it does in India: Matilal 1986; Minkowski 2002: 499 referring to David Pingree 2001: 187.

334

Frits Staal

The paper that follows consists of eight sections. The first four are concerned with the Indo-Iranian Fire Offering, the Oral Transmission of the Vedas, the Four Vedas themselves, and the verbal root śru- which means “to hear”. They provide the background and some essential facts without which the remaining four sections cannot be understood.

2. The Indo-Iranian Fire Offering The frontispiece depicts a fire rite that reflects and celebrates a major discovery of our species: how to make fire.2 “Indo-Iranian” is a linguistic and geographic, not a cultural or anthropological appellation. It refers to a branch of the large family of Indo-European languages that was spoken by nomads or semi-nomads in Central Asia. Languages move freely: they are not bound by genes, ethnicity or race. The speakers of Indo-European were tribals of various origins about which we know little. We are, however, familiar with some of their rituals. During frequent stops, the head of the family built a fire on a flat piece of ground, loosening the soil and demarcating it ritually by a square of twigs. The fire was fed with wood, incense and animal fat, later replaced by butter poured from a wooden ladle called juhū as depicted on the frontispiece which shows another ladle with a smaller head, the sruva, waiting in a dish. These ladles had a future: they accompanied Buddhist fire rites on their way to China, Korea and Japan, starting about a millennium after they had raised questions in the heads of the theorists of ritual. The illustration in the frontispiece does not come from Central Asia but from the Nambudiri brahmans of Kerala, southwest India. Their ancestors or predecessors had incorporated this small offering into the largest of their Vedic fire ceremonies, the Agnicayana or “Piling of Agni”, often simply referred to as Fire or Agni, an Indo-European word for fire that is related to Latin ignis, English ignition, etc. We shall meet with details from this ritual which is more fully described in Staal et al. 1983. The sponsor of the ritual plays a central role, more specific than the phrase “head of the family” that I used a moment ago. He is the patron or Yajamāna on whose behalf the ceremonies are performed. It is a rare privilege. The patron pays his priests or officiants if he can afford it – if he is, for example, a chieftain or king. In recent times it has been organised by the community in some parts of India, including Kerala which is the theatre for our story. I have already mentioned one fact about the development of ritual: few rites disappear. They are maintained over long periods of time for fear of losing something that is not understood but felt to be auspicious or good for some reason or other. That insight or fact may or may not be a feature of the theory of ritual. Wherever it

2 Staal et al. 1983: plate 4.

A Theory of Ritual

335

belongs, the Indo-Iranian fire rite developed several millennia before there was any such theory.

3. The Oral Transmission of the Veda The ritual from which my data and illustrations are taken is a Vedic ritual. The Vedas go back to the second millennium B.C.E. and continue to be transmitted orally together with their rituals. In India writing did not exist until the time of the Buddhist King Aśoka who ruled in the third century B.C.E. Early Indic civilisation was based on oral tradition, in contrast with Egypt, Mesopotamia, China and the monotheistic traditions which used writing much earlier and sometimes located their origin in what became a sacred book. In India, written sources are of secondary importance and so are manuscripts, books and most of the scholarship that derives from them. Even today, no Indian paṇḍita or traditional scholar studies books; he carries his knowledge in his head. The Vedas remind us again and again: “Revere Memory”. How does one remember something for thousands of years? From father to son and teacher to pupil, taking extraordinary care to safeguard the oral transmission of their traditional knowledge. Words are repeated and linked with those that precede and follow. Their sequence in each sentence is marked and so is the sentence end. Permutations are memorised and other word games played. Vedic was an accented language with three main pitches: high, middle and low. Image 1 shows how the teacher of the Vedas teaches his young pupil, pushing his head down when the accent is low (A) and bending it to one side when it is high (B).3 Learning the oral tradition of the Vedas takes four to six years and follows the ceremony of initiation. That ceremony is a life-cycle ritual such as are common elsewhere on the planet. Life-cycle rituals (Übergangsrituale)4 like initiation are characterised by crossing a boundary such as birth, marriage or death. Following Arnold van Gennep (1909), modern sociology and anthropology have been largely confined to these rituals. The theory of ritual that we shall study is fundamentally different. It crosses no boundaries. It belongs to a family of rituals that has been neglected by contemporary scientists or scholars but is an appropriate topic for Ritual Dynamics. Modern Indic and Sanskrit scholars, especially philologists, continue to study Vedic ritual but rarely display a taste or aptitude for theory. Roughly a century ago it did happen, but that was due to a rare encounter of minds. We shall merely be able to nibble at a theory of ritual, but it was the subject of Hubert & Mauss’ Essai sur la nature et la fonction sociale du sacrifice of 1899. Mauss was the pupil of a

3 Staal et al. 1983: plate 15. 4 Michaels 1998: 85–175.

336

Frits Staal

Image 1: Oral Transmission of the Veda

Photo by Adelaide de Menil, copyright by the author

great Sanskrit scholar, Sylvain Lévi, who taught him Vedic ritual in a class in which Mauss was the only pupil. It inspired a few later savants including J.C. Heesterman in the Netherlands where the groundwork was laid by W. Caland. In France, it was done differently by Louis Dumont and in Germany, where Max Weber had provided theory without Sanskrit, Axel Michaels is now combining both.

4. The Four Vedas I have mentioned the Veda or Vedas but we should be a little more precise. The Sanskrit word veda means “insight” or “knowledge” but a major concern of the Vedas is ritual. Knowledge is stressed in some compositions of the Rgveda and in the latest sections of the Vedas, especially the Upaniṣads. There are four Vedas: 1. The RV or Ṛgveda is the Veda of rg or “verse”. Its poetry displays a great variety of meters unlike the later Indic Mahābhārata and Rāmā-yaṇa and other ancient oral epics such as Homer’s that employ only one: the hexameter, monotonous as it is. Rgvedic verses describe the adventures and fights of the tribal communities to which the poets belong, address them as sages or seers (ṛṣi) who bestow wisdom and invoke gods to gain power, wealth and sons.

A Theory of Ritual

337

2. The SV or Sāmaveda consists of chants or songs: the words are largely taken from the RV and the SV might be described as a RV set to music. The native Sāmavedins, “fit their older chants on to the new language on the block”.5 3. The YV or Yajurveda, familiar with both RV and SV, deals primarily with karman in the original meaning of ritual action. Poets and singers are replaced by experts, commentators, scholars and professors. Sāyaṇa, who published extensive commentaries on the Vedas in the fourteenth century A.D., wrote: “RV and SV are the pictures; YV is the wall”.6 It will be illustrated in our final section. The Black YV is a mixture of poetry and prose. The poetry is generally from or similar to the RV and is now called mantra. The prose or brāhmaṇa provides a ritual commentary. The White YV is a reaction against the Black: it returns to the “pure” poetry of the RV and has a separate commentary that deals with ritual. Mantras are recited but their meaning is not known as was noted by Kautsa, a much neglected thinker, grammarian and ritualist who observed: “There is a tradition for mantras to be learnt by heart, but no corresponding tradition to teach and thereby preserve their meaning”.7 4. The AV or Atharvaveda deals with a variety of topics including healing and magic. It contains many verses from the RV but is not relevant to the rituals we shall look at. The Vedas consist of numerous schools which present us with a labyrinth of names that refer to a host of oral compositions: Brāhmaṇas or commentaries, Āraṇyakas or wilderness alternatives8 and Upaniṣads, already mentioned. Another category consists of Śrauta Sūtras, fifteen in number. They include detailed discussions of special cases that embody a theory of ritual. Their composers may have been individuals, groups or committees. Śrauta Sūtras developed primarily in the YV. They exhibit its central importance as a theory and organising power. The difference between YV Brāhmaṇas and Śrauta Sūtras was explained by one of the great authorities on the Vedas and on Vedic ritual, Naoshiro Tsuji: “The former endeavoured to interpret the meaning of mantras and to explain the origin and mysterious significance of ritual proceedings, and in doing so they happened to give, often rather briefly or vaguely, prescriptions as to this or that action of a rite which they presupposed to be well known to the initiated. The Sūtrakāras (composers of the Śrauta Sūtras, F.S.), on the contrary, aimed at a systematic description of each Vedic ritual in its natural sequence.”9 5 6 7 8 9

Menon 2008: 65. Discovered by Renou 1947: 9 and explained in Staal 2008: Chapter 7. Discussion in Staal 2008: 141–145. Malamoud 1989: 95; Witzel 2003: 82. Tsuji 1952: 187.

338

Frits Staal

The composers of the main Śrauta Sūtras belonged to distinct families and lineages and reflect differences in language, locality, period and ritual. They agree with each other that one school recites, chants or performs rites in one way and another in another. They do not quarrel about doctrine. Ritualists are not sectarians as one finds in later Hinduism and what are now customarily referred to as “religions”. They do not kill each other. These features reflect a basic truth about the Vedas: they are not a religion but embody a civilisation. The Śrauta ritual operates with sixteen priests who assist and serve the patron: four groups of four, one for each of the four Vedas. Important in our context is the Adhvaryu, main officiant of the YV and chief organiser and manager of the ritual. His tasks are variously described in different Śrauta compositions of which the oldest and most explicit was composed by Bodhāyana or Baudhāyana in Kosala (now eastern Uttar Pradesh, north-east India) after 800 B.C.E.10 He is typically tolerant of other authorities and makes statements like: Baudhāyana says that one should prefer X; Śāliki says one should not and Aupamanyava presents the alternative Y.

5. śru – “to hear” The term Śrauta comes from a verbal root that is especially productive in an oral tradition and from which many other important terms and concepts are derived: śru- which means “to hear”. Many of the terms have cognates in Greek with parallel phrases too: śrava “fame” corresponds to Greek kleos (κλεος) and the Rgvedic expression akṣiti śravah corresponds to Sappho’s kleos aphthiton (κλεος άφθιτον) or “imperishable fame” which occurs in a verse with a similar structure.11 The past passive participle of śru- is śruta which means “heard”. The similar form śruti means “what is heard” and connotes the tradition of the Vedas. Similarly, śrotra means “organ of hearing” or “ear”, and śrauta “to be heard” refers to the theory of ritual that is the topic of this essay. To provide a more adequate idea of this theory and more context, books like Caland & Henry 1906–1907 (or at least chapters 4–7 in Staal 2008) would be required. I shall in the remainder of this essay confine myself to four topics, two of a more general nature and two more specific and technical: 1. 2. 3. 4.

Ritual versus ordinary activity A dynamic and social, but not a ritual change Rules, metarules and default options Ritual space: the four directions.

10 Caland 1966 [1903]; Witzel 1997b: 316–317. 11 Watkins 1997: 254.

A Theory of Ritual

339

6. Ritual versus Ordinary Activity Most ritual activity depends on the Adhvaryu, chief officiant of the YV. He is the ritualist in charge of the wall on which the RV recitations and SV songs of the others rest as if they were pictures as Sāyaṇa put it. Image 2 illustrates the difference between ordinary and ritual activity.12 During the Agni or Fire ceremony, a large altar will be constructed or “piled” from kilnfired bricks in the shape of a bird facing east. In Image 2, its head is to the right. The shape of the altar has already been demarcated on the ground with the help of string and pegs; precise measures were used as is common in an ordinary activity. To be of ritual value, it must be confirmed ritually and this the Adhvaryu does by wielding his measuring stick as if it were a magic wand. The altar will consist of a thousand bricks, piled up in five layers. These bricks are waiting in the background: large squares as well as triangles and other shapes. It is the beginning of geometry in India, topic of the śulba-sūtra, called after the strings or ropes (śulba) that are seen here. It is the subject of Michaels 1978. Only the patron and his officiants can enter the enclosure. Some are standing on the right. Members of the public are further away. The only woman inside the ritual enclosure is the wife of the patron, but she has her own enclosure and is not seen here.

7. A Dynamic and Social, but not a Ritual Change 7.1 “May these Bricks be Cows for Me”: I Piling up the five layers of the altar takes five days. When it is completed, the patron, on whose behalf the ceremonies are performed, faces the altar. In Image 3 the head of the bird is in front of his knees.13 A few non-ritual bricks and parts of bricks are visible in front: they serve as stepping stones. Outside the enclosure a Nambudiri lady stands under a parasol. Behind her is a rice-field. A farmer is walking to the right. It is April, the fields are dry and it is very hot.

12 Staal et al. 1983: plate 63. 13 Staal et al. 1983: plate 80.

340

Frits Staal

Image 2: Ritual Measure of the Altar Space


A Theory of Ritual

Image 3: “May these Bricks be Cows for me”: I


341

342

Frits Staal

The patron then embarks upon a long recitation. He wishes that his bricks will turn into cows, one, a hundred, ten thousand, etc., all referred to with the help of names in the decimal system as powers of ten. His recitation goes up to ten to the power thirteen (1013), a very large number. The language of the Yajurveda has a name for each of these numbers, unlike most modern languages which have to say: “a hundred hundred thousand million”. Large numbers are common in Indic civilisation. The Jainas went up to ten to the power 23, Buddhists and the Rāmāyaṇa to the power 60. But it should be recalled that the Vedas are an oral tradition. Written expressions like 1013 do not occur there. How the decimal system originated orally is the subject of Divakaran and Bhavare.14 Staal adds zero.15 One and a half millennia later, these numbers were regarded with suspicion in Arabic, Greek and Latin texts.16 The Greeks never liked large numbers. The infinite had been taboo since Pythagoras. Indic civilisation loved the infinite, created infinite series long before Newton and used large numbers to express the age of the universe. These did not result from calculations based upon experiments, but were based upon the Vedic invention of the decimal system. Monotheistic religions did not need a decimal system strictly speaking since they believed in a creation that had been completed in six days. 7.2 “May these Bricks be Cows for Me”: II Rituals seem to change over long periods – but it is not the rites that change. Image 4 shows the same ritual as performed more than a quarter century later by a different patron and different officiants. Image 3, like the frontispiece and other photographs taken by Adelaide de Menil in 1975, depicts a lone patron and a few lookers-on. Rites and recitations remained unknown to the outside world. Image 4 depicts the ritual as performed in 2006. The ritual is the same and the same mantras are recited. Second from left, the patron with his stick is now bald and is being photographed across the bird from a different direction. He is surrounded by others and smiles broadly. The change in society that took place during those few decades was that the ritual had become a public event.17

14 15 16 17

Divakaran and Bhavare (forthcoming) Staal 2010. Burnett 2006. Mahadevan & Staal 2003.

A Theory of Ritual

343

Image 4: “May these Bricks be Cows for me”: II

Copyright: Vaidika Vidya Peedham, Thrissur, Kerala, India, 2006

In 1975, the patron knew his lines. He had no need of outside assistance. In 2006, the patron still belonged to the same community; he knew his lines too. But in Image 4, the Adhvaryu has decided to stand to his left. He does not know the patron well and is nervous that he might make a mistake – a thing the patron would never do. The chubby brahman further to the right has a different concern: he hopes to perform the same ritual in the near future and is fully qualified. Furthest on the right, another Yajurvedin accompanies the recitation with hand gestures (mudrā). Any outsider who looks at Images 3 and 4 would assume that they illustrate different ceremonies and a dynamic change. What they do is depict a development in society, but not in ritual. A historical fact that is relevant to Ritual Dynamics. I do not wish to imply that changes in ritual never happen. A substitute replaced the original Soma, probably a hallucinogenic, some two millennia ago.18 The 1975 performance exhibited a similar change: the eleven goats that used to be sacrificed were replaced by substitutes.19 18 Staal et al. 1983: vol. 1, 186; Staal 1987; 2001. 19 Staal et al. 1983: vol. 1, 303, 608, plate 98A; vol. 2, 456–465 with exhibit of the folding of the substitute inside the back cover.

344

Frits Staal

8. The Theory of Ritual 8.1 Rules and Metarules The Theory of Ritual employs the kind of scientific Sanskrit that was described by Jacobi in 1903 and Hartmann in 1955, but we cannot capture its spirit unless we begin with the difference between rule (sūtra) and metarule (paribhāṣā) that is: a rule about rules. The theory is, if only for that reason, different from anything familiar to modern readers who associate ritual with religion. For one thing, rules and metarules are fully explicit as would be expected from a scientific theory. Sometimes they are formulated in what is almost a formal language which is what they became among the Sanskrit grammarians when they adopted their logical techniques from Vedic ritual.20 We have to scale higher mountains and steeper cliffs than students of ritual in the Euro-Americas have so far climbed. In the case of the Agnicayana piling of bricks, the theory provides the rules that describe which mantras are used for the deposition of each of the bricks and the direction in which it is done. Many rules are of the form: “Brick X is deposited with mantra Y in direction Z.” It is to such rules that metarules apply, e.g.: “Rules like ‘he deposits them toward the east’ imply ‘the agent who does the depositing faces east’.” Two things should be noted here. First of all, “he” refers to the Adhvaryu priest if it is not stated otherwise – a default mechanism to which we shall return in a moment. Second, I have used single quotation marks to express rules and double ones to express metarules. Metarules make a statement about the relationship between rules. I am expressing this in a written language with the help of two kinds of quotation marks although the theory of ritual is composed and expressed in an oral language. How did they do it? As it happens, spoken Sanskrit possesses a quotation mark: iti (literally “thus”). It marks the end of a quotation but not the beginning which has to be construed from the context. For example: Rāma said to Sītā follow me iti. Where is the beginning of the quote? Before “follow”, stupid! The metarule we have discussed is a rather special rule. Others have a more general character, e.g. “Each act is accompanied by one mantra” (ekamantrāṇi karmāṇi)

20 Renou 1941–1942.

A Theory of Ritual

345

It states a general principle to which there are exceptions. There are, for example, silent acts or rites which are unaccompanied by mantras or accompanied by mantras that are articulated within the mouth without being heard. Another general metarule expresses the paradigm of the relation between acts or rites and mantras: “he lets the beginnings of the act coincide with the ends of the mantras” (mantrāntaiḥ karmādīn saṃnipatayet) This metarule (in which “he” is again the Adhvaryu) is exemplified by a large number of offerings or oblations. One officiant, say X, makes an offering of clarified ghee by pouring it into the fire. Another, Y, accompanies it with a mantra. An example is the mantra śrauSAT! which means “Let him hear!” The first part śrauis derived from the verbal root śru- and the second part SAT! is an archaic verbal ending which I have written in capital letters because it is an exclamation made at the top of his voice by officiant Y at the very moment that X makes his oblation. The histories of grammar, logic and the theory of ritual in India and Europe were first compared in Staal.21 On rule and metarule in Europe I shall only add the briefest of sketches. When Aristotle around the middle of the third century B.C.E. observed: “it is not possible to assert and deny the same”, it could have been stated as a metarule about two rules: “X is true” and “X is not true”. Aristotle did not refer to it as such but Andronicus of Rhodes made history in 60 B.C.E. when he called Aristotle’s work in which the third chapter offers numerous formulations of the principle of contradiction: the Metaphysics. The reasons are that Greek metadoes not only mean “beyond” but also “after” and that Andronicus, who was arranging his books, put the Metaphysics after the Physics on his shelf. Two thousand years later, David Hilbert introduced the term meta-mathematics. It referred to what were later called the foundations of mathematics. Another quarter century passed and Alfred Tarski used object- and meta-language. Closely connected with Gödel’s theorems on completeness and consistency, it led to Alan Turing on decidability and the invention of the computer which, like a modern deity, tries to tell us what to do and not to do. 8.2 Default Options Alexandra Heidle has told me that German uses the English word “default” but also German “Standard-Einstellungen”. Default options, as I shall continue to call them, are expressed by metarules. We make use of such options on our keyboard. It types small letters on the screen without an explicit instruction. If we need capitals we have to tell the machine. We have already met with one default option: the use of “he” which refers to the Adhvaryu priest if it is not stated otherwise. 21 Staal 1988: 36–45.

346

Frits Staal

The prehistory of the concept of default was discussed by Albrecht Wezler in the Zeitschrift für vergleichende Sprachforschung in 1972. Neither he nor the predecessors to whom he refers, Renou, Thieme and Ikari, used the term “default”. Other relevant contributions were made by Yagi in 1984 and by Ojihara in publications that were made available in 2007. None of these scholars made the distinction between rules and metarules but Wezler showed that the road was paved by statements such as: “when no special prescription is made, the usual one about the sacrificial fees prevails.”22 The “usual” in that statement is the default. In order to explain how default options are introduced into our theory of ritual, I shall quote from one of the later and therefore more perspicuous Śrauta Sūtras: Āpastamba Śrauta Sūtra 24.1.23–26.23 The following four rules are metarules that refer to oblations, priests and implements, and the last, no. 26, shows that there are degrees of default. I add the Sanskrit because it shows, at least in nos. 24–26, that sūtras are brief as they are supposed to be: “23. When an oblation is enjoined, he uses clarified butter (juhotīti codyamāne sarpirājyam pratīyāt) 24. The officiating priest is the Adhvaryu (adhvaryuḥ kartāram) 25. The implement is the juhū ladle (juhūm pātram) 26. When (the juhū) is already used (and no other implement is specified, the oblation is made) with the Śruva (vyāprtāyāḥ Śruveṇa).” Statement 23 is clear: if a substance different from clarified butter is used, Soma for example, it is stated. The “he” of no. 23 is explained by no. 24: if another officiant performs, it is stated who he is. No. 26 is of interest because it mentions two degrees of default. The juhū and the sruva are the same two ladles that were shown in the frontispiece of the Indo-Iranian offering with which we started.

9. Ritual Space: The Four Directions Ritual is not a celestial matter. It is performed on earth as we shall see and must be located somewhere. Like the square of twigs of the Indo-Iranian fire ritual, ritual space is an area demarcated on the ground that may be used for power or political 22 Wezler 1972: 12. 23 German translation in Caland 1928: 386.

A Theory of Ritual

347

ends. Sāmavedins made the first move as we have seen. Their form of cooperation with RV newcomers did not eliminate tensions between SV and RV which continued to be felt.24 The Yajurvedins exploited the situation with the help of ritual. At the centre of the ritual space in which Śrauta Rituals are performed (part of which is shown in Image 2), they created a separate enclosure, the Sadas. Inside they put Rgvedins and Sāmavedins and separated them from each other. The Sadas is the YV wall on which the pictures of RV and SV hang as Sāyaṇa put it. Image 1 is a schematic presentation of the one direction in which the reciters of the RV recite and look; and of the three directions in which the chanters of the SV chant and look. It shows that the two groups never face each other. They share the same space but do not see eye to eye. The Adhvaryu had put the eggs of different birds in the same nest. Image 5 looks inside the Sadas from the West, where our photographer stood, that is at the “T”, the final letter of “WEST” in the schema of Fig. 1. Fig. 1: Facing Different Directions

24 An example is discussed in Staal 1992: 657.

348

Frits Staal

Image 5: Inside the Sadas


Image 5 looks east. The rays of the sun enter through an opening at the top of the picture on the right. It is a rare photograph by Adelaide de Menil, first published in Staal et al. 1983, plate 103, then in Staal 2004 and the first edition of Staal 2008 where the image was reversed along the North-South axis, causing havoc since ritual directions were the topic at hand. It was corrected in the second and third printings. Corrected also in the present article, the three Sāmavedins sit in front, facing their three directions around a white piece of cloth on which they mark the progression of their chants with the help of sticks. As only a connoisseur will be able to see, Image 5 shows that nine of fifteen chants have been completed. Behind the Sāmavedins, the Rgvedins sit in a row from left to right. Three of the six are visible but they face east. We only see their backs. The relevant fact is that Ṛgvedins and Sāmavedins never face each other. Image 5 illustrates several important points about a theory of ritual. It is a theory that is not perpetuated by books but by the oral tradition which includes, in addition to the performing experts, visiting friends and relatives, young and old. Another important fact, which I have tried to unearth for the benefit of readers who

A Theory of Ritual

349

sit on chairs, is that the photograph portrays participants and visitors who are sitting on the ground which connects with the earth.25 Such sitting is expressed by the verbal root sad- which occurs in upaniṢAD, “sitting close to (the teacher)”, and in SADas, “a place to sit”. It is not primarily a matter of “science” – exeunt Oldenberg and other venerable savants – but there is theory all right. The youngest of the RV reciters turns his face around on his shoulder and is partly visible behind the SV chanter on the left. His recitals are professional but he has not yet mastered the art of sitting still. The SV chanter is on a month’s leave from his job at the Trichur post office. Unlike my postal clerk in California, an Indian expat, he does not imagine that Tokyo is in China though he has forgotten bits and pieces of the vast repertoire of his songs. For him, much water has flowed under the bridge since he was trained by his father who now supports him silently, sitting to his right on the edge of the picture. With his short white beard, earring on the right and dishevelled frontal topknot, he is the grand old man of Agni which he performed in 1956. The topknot itself26 is more clearly visible on the right: it belongs to the one other then existing Nambudiri who performed the ritual. In the left corner and everywhere else, youngsters are jumping around but that does not prevent them from absorbing recitations, chants, rites and even a theory of ritual unless they had imbibed it from the beginning: ab ovo or from the egg.

25 Gonda 1985; Staal 2008: 245. 26 Mahadevan (forthcoming).

350

Frits Staal

References Burnett, Charles 2006. “The Semantics of Indian Numerals in Arabic, Greek and Latin”. Journal of Indian Philosophy 34: 15–30. Caland, Wilhelm 1966 [1903]. Über das rituelle Sūtra des Baudhāyana. Leipzig: Brockhaus. — 1928. Das Śrautasūtra des Āpastamba. Sechzehntes bis vierundzwanztigstes und einunddreissigstes Buch aus dem Sanskrit übersetzt. Amsterdam: Koninklijke Akademie van Wetenschappen. — & Victor Henry 1906–1907. L’Agniṣṭoma. Description complète de la forme normale du sacrifice de Soma dans le culte védique. Vols. 1–2. Paris: Ernest Leroux. Divakaran, P.P. and Bhavare, B. (forthcoming). “Genesis and Early Evolution of Decimal Counting: Evidence from the Rigveda”. Dumont, Louis 1966. Homo Hierarchicus. Essai sur le système des castes. Paris: Gallimard. Gennep, Arnold van 1909. Les rites de passage. Paris: Gallimard. Göhler, Lars (ed.) 2005. Indische Kultur im Kontext: Rituale, Texte und Ideen aus Indien und der Welt. Festschrift für Professor Klaus Mylius. Wiesbaden: Harrassowitz Verlag. Gonda, Jan 1985. The Ritual Functions and Significance of Grasses in the Religion of the Veda. Amsterdam, Oxford, New York: North-Holland Publishing Company (Koninklijke Nederlandse Academie van Wetenschappen, Afdeling Letterkunde 132). Griffiths, Arlo & Jan E.M. Houben (eds.) 2004. The Vedas. Texts, Language and Ritual. Proceedings of the Third International Vedic Workshop, Leiden 2002. Groningen: Egbert Forsten. Hartmann, Peter 1955. Nominale Ausdrucksformen im wissenschaftlichen Sanskrit. Heidelberg: Carl Winter. Heesterman, Johannes C. 1985. The Inner Conflict of Tradition. Essays in Indian Ritual, Kingship and Society. Chicago: The University of Chicago Press. — 1993. The Broken World of Sacrifice. An Essay in Ancient Indian Ritual. Chicago: The University of Chicago Press. Hoek, A.W. van den & D.H.A. Kolff & M.S. Oort 1992. Ritual, State and History in South Asia. Essays in Honour of J.C. Heesterman. Leiden, New York, Köln: E.J. Brill (Memoirs of the Kern Institute, Leiden 5). Hubert, Henri & Marcel Mauss 1899. “Essai sur la nature et la fonction sociale du sacrifice”. L’Année Sociologique 2: 29–138. Jacobi, Hermann 1903. “Über den nominalen Stil des wissenschaftlichen Sanskrits”. Indogermanische Forschungen 14: 236–251. Lévi, Sylvain 1966. La doctrine du sacrifice dans les Brāhmaṇas. Paris: Presses Universitaires de France.

A Theory of Ritual

351

Mahadevan, T.P. & Frits Staal 2003. “The Turning Point in a Living Tradition: Somayāgam 2003”. Electronic Journal of Vedic Studies 10. (Revised Version in Göhler 2005: 365–389). — (forthcoming). Two Studies on the Veda in South India. Malamoud, Charles 1989. Cuire le Monde. Rite en Pensée dans l’Inde Ancienne. Paris: Editions de la Découverte. (English Translation by David White 1996. Cooking the World. Ritual and Thought in Ancient India. Delhi: Oxford University Press). Matilal, Bimal Krishna 1986. Perception. An Essay on Classical Indian Theories of Knowledge. Oxford: Clarendon Press. Menon, Sunil 2008. “Hymn: An Aural History. A Scholar-Gypsy’s Guided Tour through Vedic Country”. Outlook, August 25: 65. Michaels, Axel 1978. Beweisverfahren in der vedischen Sakralgeometrie. Ein Beitrag zur Entstehungsgeschichte der Wissenschaft. Wiesbaden: Franz Steiner. — 1998. Der Hinduismus. Geschichte und Gegenwart. München: C.H. Beck. Minkowski, Christopher 2002. “Astronomers and Their Reasons: Working Paper on Jyotiḥśāstra”. Journal of Indian Philosophy 30: 495–514. Ojihara, Yutaka 2007a. “Mahābhāṣya ad Pāṇini 1.1.56. Un essai de traduction”. In: Toru Yagi& Minoru Hara (eds.). Mémorial OJIHARA Yutaka. Studia Indologica. Tokyo: The Toyo Bunko: 43–67. — 2007b. “À la recherche de la motivation ultime du Pāṇini-sūtra 1.1.62”. In: Toru Yagi & Minoru Hara (eds.). Mémorial OJIHARA Yutaka. Studia Indologica. Tokyo: The Toyo Bunko: 86–100. Oldenberg, Hermann 1919. Vorwissenschaftliche Wissenschaft. Die Weltanschauung der Brāhmaṇa-Texte. Göttingen: Vandenhoeck & Ruprecht. Pingree, David 2001. “Nīlakanṭha’s Planetary Models”. Journal of Indian Philosophy 29: 187–195. Renou, Louis 1941–1942. “Les connexions entre le rituel et la grammaire”. Journal asiatique 233: 105–165. (Republished in Staal 1972: 434–469). — 1947. Les écoles védiques et la formation du Veda. Paris: Imprimerie Nationale. Sharf, Robert H. 2003. “Thinking Through Shingon Ritual”. Journal of the International Association of Buddhist Studies 26/1: 51–93. Staal, Frits (ed.) 1972. A Reader on the Sanskrit Grammarians. Cambridge, London: The MIT Press. (Studies in Linguistics 1). — et al. (eds.) 1983. AGNI. The Vedic Ritual of the Fire Altar. Vols. 1–2. Berkeley: Asian Humanities Press. (Reprint 2001. Delhi: Motilal Banarsidass). — 1987. “Professor Schechner’s Passion for Goats”. Journal of Asian Studies 86: 206– 209. — 1988. Universals: Studies in Indian Logic and Linguistics. Chicago, London: The University of Chicago Press. — 1989. Rules without Meaning: Ritual, Mantras and the Human Sciences. New York, Bern, Frankfurt a.M. et al.: Peter Lang. (Reprint 1996. Ritual and Mantras: Rules without Meaning. Delhi: Motilal Banarsidass).

352

Frits Staal

— 1992. “AGNI 1990. With an Appendix by Harold Arnold”. In: A.W. van den Hoek & D.H.A. Kolff & M.S. Oort 1992. Ritual, State and History in South Asia. Essays in Honour of J.C. Heesterman. Leiden, New York, Köln: E.J. Brill: 650–676 (with illustrations in the back of the volume). — 2001. “How a Psychoactive Substance Becomes a Ritual: The Case of Soma”. Social Research 68/3: 745–778. — 2004. “From prāṅmukham to sarvatomukham: A Thread through the Śrauta Maze”. In: Arlo Griffiths & Jan E.M. Houben (eds.) 2004. The Vedas. Texts, Language and Ritual. Proceedings of the Third International Vedic Workshop, Leiden 2002. Groningen: Egbert Forsten: 521–555. — 2008. Discovering the Vedas. Origins, Mantras, Rituals, Insights. New Delhi: Penguin Books India. Reprinted 2009. — 2010. “On the Origins of Zero.” In: Bhatia Rajendra (ed.). Studies in the History of Mathematics in India. Proceedings of a Seminar held at the Chennai Mathematical Institute, February 2008. New Delhi: Hindustan Book Agency. Tsuji, Naoshiro 1952. On the Relation between Brāhmaṇas and Śrauta Sūtras. (In Japanese with English Summary). Tokyo: The Toyo Bunko. Watkins, Calvert 1997. “The Indo-European Background of Vedic Poetics”. In: Michael Witzel (ed.) Inside the Texts/Beyond the Texts. New Approaches to the Study of the Vedas. Proceedings of the International Vedic Workshop, Harvard University, June 1989. Cambridge: Department of Sanskrit and Indian Studies, Harvard University: 245–256. Weber, Max 1921. Gesammelte Aufsätze zur Religionssoziologie. Tübingen: J.C.B. Mohr. Wezler, Albrecht 1972. “Marginalien zu Pāṇini’s Aṣṭādhyāyī I: sthānin”. Zeitschrift für vergleichende Sprachforschung 86: 7–20. Witzel, Michael (ed.) 1997a. Inside the Texts/Beyond the Texts. New Approaches to the Study of the Vedas. Proceedings of the International Vedic Workshop, Harvard University, June 1989. Cambridge: Department of Sanskrit and Indian Studies, Harvard University (Harvard Oriental Series, Opera Minora 2). — 1997b. “The Development of the Vedic Canon and its Schools: The Social and Political Milieu”. In: Witzel, Michael (ed.). Inside the Texts/Beyond the Texts. New Approaches to the Study of the Vedas. Proceedings of the International Vedic Workshop, Harvard University, June 1989. Cambridge: Department of Sanskrit and Indian Studies, Harvard University: 257–245. — 2003. “Vedas and Upaniṣads”. In: Flood, Gavin D. (ed.). The Blackwell Companion to Hinduism. Oxford et al.: Blackwell Publishing: 68–101. Yagi, Toru 1984. Le Mahābhāṣya ad Pāṇini 6.4.1–19. Paris: Collège de France (Publications de l’Institut de civilisation indienne 50). — & Minoru Hara (eds.) 2007. Mémorial OJIHARA Yutaka. Studia Indologica. Tokyo: The Toyo Bunko. (Reviewed by Oskar von Hinüber 2008. Indo-Iranian Journal 51: 79–80)

Section II: Ritual Discourse, Ritual Performance in China and Japan Edited by Lucia Dolce, Gil Raz and Katja Triplett

Lucia Dolce, Gil Raz and Katja Triplett

Introduction The papers in this section present diverse studies of ritual in China and Japan, ranging in time from the second century BCE to the present, across distinct religious traditions, and following different methodological approaches. Undergirding this diversity is a common assumption that ritual is a fulcrum for understanding debates on meaning, power, agency, and efficacy. Thus, not only must interpretations of ritual be fully contextualized in their cultural matrix, but ritual itself helps to constitute this matrix. At the same time, a cross-cultural perspective across a geographical area with shared ritual and religious history such as East Asia helps identifying alternative modes and models of ritual discourse. Based on new historical and empirical data, these contributions provide new insights into East Asian ritual practices. Several of the contributions in this section were originally presented in two panels, “Debating ritual in China” chaired by Gil Raz, and “Ritual Practices in Japan” chaired by Lucia Dolce and Katja Triplett. These papers were augmented with papers from other panels, which dealt with similar issues. Major common themes in these contributions are indigenous ritual theories, performative features of ritual, and the tension between canonical prescriptions and actual enactment of ritual. A first group of papers explores indigenous ritual theories and debates about the meaning of ritual. Michael Puett examines the Book of Rites (Liji) as one of the earliest sources for indigenous theorizing about ritual. Composed in early China, this text would ultimately become one of the most influential ritual texts in East Asia. The theoretical stance of this text informs the Confucian understanding of ritual as cultural patterning, which defines civilization through its artificiality. Joachim Gentz examines the theoretical assumptions elaborated by later exegetes of the mourning rites codified in the Book of Etiquette and Ceremonial (Yili). Correct mourning ritual is explicitly discussed in terms of maintaining and displaying social and cosmic order. Christian Meyer examines debates among Confucian ritualists of the eleventh century, as they try to define the category li, ritual, a project not unlike our own. In contrast to the Confucian theory that saw ritual as constructed by and constructing human civilization, Daoists perceived their ritual system spontaneously generated in the celestial realms. Gil Raz examines the theoretical underpinning of Daoist ritual. Raz argues that debates on correct practice are not just about ortho-

356


praxy but about ritual theory. Furthermore, the modular structure of Daoist ritual schemes should be seen as an implicit theoretical claim about ritual. Julius Tsai’s examination of tenth century debates about changes in ritual protocols highlights underlying tensions between charisma and lineage, pragmatism and tradition, and human and divine initiative in the construction of ritual. A second group of papers addresses the question of ritual meaning by focusing on the performativity of ritual, that is, how it creates meaning, its force and efficacy. This perspective highlights the layered signifiers that construct and grant authority to a specific practice, based on role (ritual specialist/layman) and on knowledge of the ritual technology and its theoretical or social underpinning. The performative focus shows that ritual is not a unitary performance, but a complex practice composed of several segments with distinct dynamics. The meaning of ritual is therefore layered and stratified. The performative aspect of ritual is especially emphasized in the contributions by Lucia Dolce, Benedetta Lomi, Fumi Ouchi and Katja Triplett, who explore varied forms of Buddhist ritual in Japan. These essays highlight the process of localization of a pan-Asian religious tradition through specific ritual practices. Buddhist rituals in Japan are often divided into exoteric and esoteric (or Tantric) categories. Based on textual sources as well as ethnographic evidence, these papers explore both types of rituals, addressing the historical contexts of the Buddhist liturgies as well as contemporary performances. Dolce discusses one of the oldest public liturgical assembly still performed today, the shunie of Tôdaij, which centres on Buddhist repentance rituals. Lomi and Ouchi focus on two fundamental rituals of the Tantric tradition, a fire ritual (goma) performed for achieving specific worldly benefits (Lomi), and the eighteen-method procedures (jûhachidô), a training practice for the empowerment of Tantric practitioners (Ouchi). Triplett discusses esoteric eye healing rituals in Tsubosakadera, a medium-sized temple near Nara, surveying a liturgical tradition related to and derived from esoteric texts broadly used across denominations. Received scholarship has approached Buddhist rituals in Japan as the patrimony of a single school or lineage. Thus Buddhist specialists have examined rituals in relation to the doctrinal context of a specific school, while historians have looked at the institutional dimensions of ritual as reflections of existent socio-political structures. The contributions here, however, explore the interaction of the Buddhist canonical tradition and local (religious and mythological) traditions as well as the transformations of normative sources, imported from continental China, through the re-interpretation, sometimes along “unorthodox” lines, by Japanese practitioners. The papers are concerned with the strategies of legitimation devised by the ritual actors. Other culture specific practices that are clearly ritualized yet detached from institutional religious traditions are presented in the papers by Marcia Butler and Paul Atkins. Butler discusses divination rites documented in military manuals, and

Introduction

357

performed by either professional diviners or by military commanders. Atkins discusses seppuku, conventionally known as ritualized suicide. Atkins argues that seppuku is neither a ritual nor a form of suicide, but a ritual element that has been combined, at various times, with acts of suicide or execution. These contributions highlight the autonomy of ritual practices and techniques from established religious traditions even when they address matters related to the extra-human realm or to the ethical sphere. Within these broadly conceived explorations of ritual meaning, a number of themes are explored, which we see as fundamental factors in the process of signification that is at work in the conception and performance of ritual: multiple agency, ritual efficacy, visual focus, ritual change, and the corporeal and emotional dimensions of ritual. Dolce investigates the multiple agents of the repentance liturgy of a major Nara temple by drawing attention to the distinct spaces, movements and sounds that mark the performative arenas, and the competing technical skill of different ritual actors that produce ritual signification. Lomi’s study of fire ritual in a small Japanese temple reveals the interaction between Japanese folk practices and Buddhist forms of worship. Katja Triplett examines the competition among cleric elite and non-elite interest groups regarding the promotion of the benefits offered by the temple. Both Dolce and Lomi emphasise the role of the audience, along with the practitioners and the celebrants, in generating meaning. Ouchi also argues that the soteriological goal of a ritual is attained when the performance successfully involves the audience. This relation between ritualist and the beneficiary of the ritual highlights a second important theme examined in several papers, that of ritual efficacy. Marcia Butler draws attention to the ability of occult specialists in the context of warfare, while Dolce examines the effectiveness of ascetic exercises and extravagant fire spectacle. Triplett analyses late medieval and early modern production of narratives that promote the benefits offered by the temple to the pilgrims who visit it. Miracle stories emphasized the efficacy of the ritual for the main object of worship of the temple. This points to the crucial role of the material objects of worship in the construction of ritual meaning. The function of icons in two ritual traditions, the Daoist and the Tantric Buddhist, is explored by Poul Andersen and Lomi. Andersen examines the multivalent Daoist understandings of iconography, showing their variations in different parts of China, at different times, and in different liturgical traditions within Daoism. Lomi discusses the relation between icon and purpose of a ritual, and shows how different interpretations of the central image of a liturgy lead to contrasting constructions of the ritual itself. She suggests a shift from the relation that the celebrant entertains with the icon to the relation that worshippers establish with the icon.

358


Ritual change is a theme in several contributions, whether it addresses changes in the ritual protocol or in the meaning attributed to the ritual. The underlying assumption of our analysis is that the meaning of ritual is redefined every time according to the participants, the geographical and historical circumstances in which the performance takes place, and the use of different media. The contributions thus emphasize discrepancies between the actual performance and the canonical explanations (Ouchi, Lomi) and ritual developments according to socio-economic factors (Lomi); how ritual change forces a renegotiation of authority among different agents that contribute to the construction of the ritual (Tsai), and foregrounds the usually implicit tension between normativity and permutation (Meyer, Lomi). Finally, several essays examine ritual with a particular focus on the embodiment of ritual codes and the corporeal aspects of ritual (Gentz, Dolce, Atkins, Ouchi). A basic assumption in East Asian ritual is that ritual expresses human emotions. Such controlled expression in discussed by Atkins in the context of seppuku, and by Ouchi in her examination of mantra recitation in esoteric Buddhist rites. Similar notions of controlled expression of emotions are inherent in the mourning rites examined by Gentz. Ritual also produces emotions. Ouchi shows how ritual fittings and offerings affect the practitioner both physiologically and psychologically and how the physical and mental stress that Tantric practitioners experience leads them towards distinct experiences closely linked to the soteriological system in which the ritual is enacted. Such analyses of the participatory and experiential aspects of ritual highlight the meaning of ritual communication as a process in which the human body, encompassing both verbal and sensory attributes, plays a crucial role.

Michael Puett

Ritualization as Domestication Ritual Theory from Classical China As we continue to develop our theoretical understandings of ritual, it is important that we also explore the enormous body of ritual theory that has been generated in non-Western cultures throughout the world. Frits Staal has done excellent work in bringing theories of ritual from South Asia into our theoretical frameworks, and, following his example, we need to do more of the same from the many other traditions throughout the world that have developed indigenous ritual theories.1 As several scholars have argued, many of our current theories are implicitly based at least in part upon Christian or more specifically Protestant assumptions.2 Bringing more indigenous theories into our discussions – in other words, taking non-Western traditions seriously from a theoretical perspective and not simply as objects of our analyses – will help us to overcome the potential biases in our current theoretical understandings. This paper will be a small contribution to this larger project by discussing ritual theory from China – one of the cultures that has a lengthy, indigenous tradition of theorizing about ritual.3 I will focus particular attention on the “Li yun” chapter of the Book of Rites (Liji), a work from early China that would ultimately become one of the most influential ritual texts in East Asia. It is important to emphasize at the outset that the theory of the “Li yun” should not be taken to represent “Chinese” assumptions about ritual. As in any tradition that has developed theories on ritual, those theories became widely debated and contested. Thus, the views provided in the “Li yun” chapter are one among many from the classical period in China, and would continue to be hotly debated throughout East Asia over the ensuing two millennia. It is also important to emphasize that the “Li yun” should not be taken as a description of ritual practice of the day. As we will see, the theory espoused here 1 Staal 1989. 2 Sahlins 1996, Asad 1993. 3 For attempts to take early Chinese ritual theory seriously as theory, see Puett 2006; and Seligman et al. 2008: 17–42 and 179–182.

360

Michael Puett

builds upon several aspects of contemporary practice, but it also explicitly opposed other aspects. In all of these respects, the theory of ritual in the “Li yun” is like theories of ritual developed in the West over the past two centuries. It builds on certain aspects of ritual practice, while ignoring or opposing others. It also has underlying normative arguments about ritual. (This is, of course, equally true of ritual theories developed in the West, but at least in early China these normative arguments are more overt and explicit.) As we will see, it will accordingly bring out certain crucial elements of ritual left out by other theories, while missing others. It is, in short, much like other theories of ritual, and our understanding of ritual will be enhanced by bringing it into our discussions.

The Argument of the “Li Yun” Like the other chapters of the Book of Rites, the “Li Yun” presents ritual as a human construction. Ritual is not handed down by divine powers; indeed, ritual defines divine powers just as much as it defines humans. Such a position was a controversial view in early China, but, at least in this aspect, the text was building upon certain aspects of popular practice. One of the interesting and most enduring features of sacrificial practice in China, from the early times to the present, is the notion that human ritual domesticates an otherwise capricious and dangerous world of divine powers and determines the pantheon within which those divine beings operate.4 I have argued elsewhere that many of the chapters of the Book of Rites build upon this aspect of sacrificial practice to develop their theories of ritual.5 The “Li yun” does as well, but the focus on the “Li yun” is primarily on the human side of this human project – the ways in which ritual orders human dispositions and thereby orders the relationships that humans take toward both the divine and natural worlds. Many later practices and theories – notably those of later Daoists – explicitly opposed such a vision of ritual as a human construction and argued on the contrary that only rituals revealed by higher gods should be followed.6 But the “Li yun” is clear on this: ritual was created by humans, and the construction of a proper order is a human project of transforming and organizing the world through ritual. The “Li yun” also argues that, given the transformative power of ritual, humans must construct the world well. Ritual done well creates continuity among disparate phenomena, but, if not done properly, can also create the opposite. This is one of 4 Puett 2002, Jordan 1972, Wolf 1974, Harrel 1974, Watson & Rawski 1988, Weller 1987, Sangren 1987. 5 Puett 2005, 2008. 6 Schipper 1993.

Ritualization and Domestication

361

the reasons, the text argues, that the world humans find themselves in at birth is one of discontinuity – rituals have not been used properly. Thus, the aspects of ritual that can create links among what are experienced as discrete and unrelated phenomena in the world are the primary concern of the authors of the text. This point is also of interest in reflecting on the interplay of cosmology and ritual. Scholars have often read early Chinese as assuming a harmonious, monistic cosmos in which everything was inherently related. I have argued elsewhere that this view was never an assumption in early China.7 And certainly the “Li yun” does not assume it. The “Li yun” on the contrary views the world as being one of discontinuity. It does posit a period of unity in distant antiquity – but this was also a world in which humans lived like animals. The subsequent rise of human civilization broke this unity and allowed humans to thrive. The text is thus calling for a vision of ritual in which a new, superior unity can be constructed – a unity in which humans take the central position. A harmonious, monistic cosmos, the text makes clear, would be the product, not the opening assumption, of human ritual. Indeed, among the various theories that have been developed throughout the world arguing for a strong constructionist vision of ritual, those in the Book of Rites are among the most extreme. With this as a brief introduction, let us turn to the text itself.

The Opening Dialogue The chapter opens with a dialogue between Confucius and his disciple Yan Yan. Confucius reflects on his desire to practice the Great Way as it was practiced in the past, or at least as it was practiced by particular illustrious figures after the dynastic system was created: “The practice of the Great Way and the illustrious figures of the Three Dynasties – these I have not been able to reach. But my intent is to do so.”8 In distant antiquity, according to the narrative, the world was not divided into families: “In the practice of the Great Way, all under Heaven was public. They selected the talented and capable. They spoke sincerely and cultivated peace. Therefore, people did not only treat their own kin as kin, and did not only treat their own sons as sons.”9

7 Puett 2002. 8 Liji, “Li yun” ICS, 9.1/59/23–24. My translations from the Liji here and throughout have been aided greatly by those of James Legge 1885. 9 Liji, “Li yun” ICS, 9.1/59/24.

362

Michael Puett

But this has now been lost: “Now, the Great Way has become obscure. All under Heaven is [divided into] families. Each treats only its own kin as kin, only their own sons as sons.”10 Confucius clearly has in mind here the creation of the dynastic system, in which kingship came to be controlled by a single lineage until it is overthrown by another. Confucius is thus regretting the fact that people have come to only think in terms of their immediate kin, and that the world consists of an endless competition between these kin groups. But the chapter includes a crucial twist: it attributes this loss in part to ritual. “Ritual and propriety are used as the binding. They are used to regulate the ruler and subject, used to build respect between the father and son, used to pacify elder and younger brother, used to harmonize husband and wife, used to set up regulations and standards, used to establish fields and villages, used to honor the courageous and knowledgeable, taking merit as personal. Therefore, schemes manipulating this arose, and because of this arms were taken up.”11 The attempt to bind people back together again using ritual has simply resulted in people scheming to manipulate these ritual links. However, according to Confucius, six figures since the creation of the dynastic system have been able to use ritual effectively: “Yu, Tang, Wen, Wu, Cheng, and the Duke of Zhou were selected because of this. These six rulers were always attentive to ritual, thereby making manifest their propriety, thereby examining their trustworthiness, making manifest when there were transgressions, making the punishments humane and the expositions yielding, showing constancy to the populace. If there were some who were not following this, they would be removed from their position and the populace would take them as dangerous. This was the Lesser Peace.”12 This for Confucius was the Lesser Peace – lesser, that is, than the Great Way practiced in distant antiquity.

10 Liji, “Li yun” ICS, 9.1/59/27–28. 11 Liji, “Li yun” ICS, 9.1/59/28-30. 12 Liji, “Li yun” ICS, 9.1/59/30–32.


363

The Development of Ritual From this opening frame, one might assume that the text will argue against the following of ritual. On the contrary, however, the text presents Confucius as still strongly calling for rituals to be followed: “Yan Yan asked again, ‘Are the rituals of such urgency?’ Confucius said: ‘Rituals are what the former kings used to uphold the way of Heaven and regulate the dispositions (qing) of humans.’”13 When Yan Yan asks Confucius to explain, Confucius provides a narrative of the development of ritual. As the narrative makes clear, the period of the Great Way in distant antiquity – the period when unity prevailed – was also one in which humans lived in caves and nests, and ate berries and drank animal blood for food: “In ancient times, the former kings did not yet have houses. In the winter they lived in caves, in the summer in nests. They did not yet know the transformations of fire. They ate the fruits of grasses and trees, and the meat of birds and animals. They drank their blood and ate their feathers.”14 The subsequent inventions of the sages (including dwellings, fire, and agriculture) lifted humanity out of this state. Confucius, therefore, is not calling for a return to the Great Way (since that would also mean a return to living in caves and drinking blood) but rather a return to those few rulers in more recent times who have been able to use ritual appropriately to create a different kind of unity. And what precisely would this appropriateness mean? It would mean using ritual to re-create the sense of the world as being linked as a single family – the same sense that existed in distant antiquity, only now humans would have this sense while also being part of a complex society practicing agriculture, having a ruler, etc. The remainder of the text is devoted to explaining how this is possible.

The Creation of Continuity Confucius begins by defining the dispositions of humanity: “What are the dispositions (qing) of humans? Happiness, anger, sadness, fear, love, detesting, liking – humans are capable of these seven without study.”15

13 Liji, “Li yun” ICS, 9.2/60/1. 14 Liji, “Li yun” ICS, 9.5/60/14–15. 15 Liji, “Li yun” ICS, 9.23/62/8.

364

Michael Puett

The dispositions, in other words, are the basic emotional qualities that all humans have at birth. Ritual works through an ordering of these dispositions. To explain how rituals do this, Confucius provides a lengthy explanation of the nature of ritual. The text begins by stating that humans have pieces of the rest of the cosmos within themselves: “Thus, humans are [a product of] the powers of Heaven and Earth, the interaction of the yin and yang, the joining of the ghosts and spirits, and the subtle energies of the five phases.”16 Humans are formed through an interaction of pieces of the rest of the cosmos. This is presumably why humans in their primitive state were part of a larger unity. The subsequent invention of rituals was related to the other inventions of the sages that lifted humans from the level of animals. All of these inventions involved a domestication of the natural world, but a domestication undertaken by placing humans in a different relationship with those natural elements: “Thus, when the sagely humans created rules, they necessarily took Heaven and Earth as the basis, took yin and yang as the level, took the four season as the handle, and took the sun and stars as the marker [of time]; the moon was taken as the measure, the ghosts and spirits as the assistants, the five phases as the substance, the rites and propriety as the instruments, the dispositions of humans as the field, and the four efficacious creatures as the domesticates.”17 All of these were elements of the given world within which humans found themselves, but human sages have now appropriated these elements and utilized them in the larger project of human domestication. Indeed, the text even makes an implicit comparison of the domestication of the dispositions with the formation of agriculture by discussing the dispositions as a field. Later, this comparison becomes quite explicit: “Therefore, the sage kings cultivated the handles of propriety and the arrangements of the rites in order to regulate human dispositions. Thus, human dispositions are the field of the sage kings. They cultivated the rites in order to plough it, arrayed propriety in order to plant it, expounded teachings in order to hoe it; took humaneness as the basis in order to gather it; and sowed music in order to pacify it. Therefore, rites are the fruit of propriety.”18

16 Liji, “Li yun” ICS, 9.24/62/15. 17 Liji, “Li yun” ICS, 9.26/62/22–24. 18 Liji, “Li yun” ICS, 9.35/63/25.


365

Rites, in short, involve a domestication of the dispositions just as agriculture and animal husbandry involved a domestication of nature. The rites thus serve to inculcate those domesticated dispositions and to define the ways in which those dispositions should relate to the rest of the cosmos: “Thus the former kings were worried that the rites would not reach to those below. They therefore sacrificed to Di in the suburban sacrifice as a means by which to determine the place of Heaven. They made offerings to the Earth altar in the kingdoms as a means by which to array the benefits of the earth. They offered ancestral sacrifices at the shrines as a means to provide a basis for humaneness. They offered at the mountains and the streams as a means by which to host the ghosts and spirits. They gave the five offerings as a means by which to provide a basis for activities. Therefore, there were ancestral invocators at the shrines, the three dukes at the court, and the three elders at the schools. In front of the king were the ritual specialists and behind were the astronomers; the diviners by crackmaking, the diviners by stalks, drummers, and assistants all stood to the right and left. The king was at the center. His mind was without activity, so as to hold fast to the utmost correctness.”19 The rituals define the relationship and proper dispositions that humans should have to Heaven, to the produce they have taken from the earth, and to the ghosts and spirits. The specialists for each of these relationships would be given a place in the court, with the king standing at the center of these relationships. The rituals also give offices to the spirits, define what can be properly appropriated from the earth, and inculcate proper feelings of filiality through ancestral worship. “Thus, the rites were practiced in the suburbs, and the myriad spirits received offices through them. The rites were practiced at the earth god’s altar, and the hundred goods could be fully appropriated through them. The rites were practiced in the ancestral shrines, and filiality and kindness were submitted through them. The rites were practiced with the five sacrifices, and the correct standards were taken as models through them. Therefore, from the suburban sacrifice, earth god altar, ancestral shrine, mountains and rivers, five sacrifices, propriety was cultivated and the rites were embodied.”20 In short, the entire world comes to be linked through a set of normative relationships embodied in ritual. As such, rituals, although an invention of humans, and initially part of what led to the fragmentation of the world, are nonetheless constitutive of a larger unity: 19 Liji, “Li yun” ICS, 9.28–9.29/63/4–8. 20 Liji, “Li yun” ICS, 9.30/63/10–11.

366

Michael Puett “It is for this reason that the rites are necessarily based in the Great One, which separated and became Heaven and Earth, revolved and became yin and yang, alternated and became the four seasons, were arrayed and became the ghosts and spirits.”21

But the crucial difference with distant antiquity is that now the sage king is the central figure connecting everything into this greater unity: “Thus, the sage forms a triad with Heaven and Earth and connects with the ghosts and spirits so as to control his rule.”22 The sage has inserted himself between Heaven and Earth, connecting them along with the ghosts and spirits and thus asserting his rule. Thus, ritual operates much like agriculture. Prior to the formation of agriculture, humans were linked to the natural world, but linked in the sense of being animals themselves – eating the blood of other animals and often being eaten themselves; gathering berries that would sometimes end up being poisonous; being subject to a dangerous series of weather changes from freezing cold to torrential rains to heat and drought. Once the world was domesticated by humans, however, the wild animals were killed or transformed into animals of service to humanity; the forests were cleared and the grasses domesticated into foodstuffs for humanity; the dangerous series of weather changes became seasonal shifts that operated usefully within an agricultural cycle. The earlier unity was lost, but from another perspective a greater unity was achieved: the rains and sunlight from Heaven and the domesticated produce from the earth became part of a larger unity, organized by humanity and thus with humanity at its center. And ritual operates in the same way. Although ritual, unlike the primitive period of universal sharing, assumes a fragmented world in which people only think in terms of their own kin, rituals nonetheless allow these distinct families to once again be made into a single family. The overall argument of the “Liyun”, therefore, is that ritual allows a type of re-creation of an earlier unity that existed in deep antiquity. But that earlier unity was a primitive world in which humans starved to death for lack of food and died from the elements for a lack of shelter. Now, humans have innovations that allow them to transform the natural world, and they have a ruler to regulate the distinct families and organize them in the world. But, with ritual, the ruler comes to be seen not as an arbitrary form of power but rather as the linchpin of this connected world. As Confucius puts it: “Therefore, as for the sage bearing to take all under Heaven as one family and take the central states as one person, it is not something done overtly. He necessarily knows their dispositions, opens up their sense of propriety, clari21 Liji, “Li yun” ICS, 9.31/63/13–14. 22 Liji, “Li yun” ICS, 9.18/61/26.


367

fies what they feel to be advantageous, and apprehends what they feel to be calamitous. Only then is he capable of enacting it.”23 By covertly working upon the dispositions of the populace, the ruler is able to create a sense of everything under Heaven as being a single family. Everything is now united, but united through the person of the ruler. Thus, if humans find themselves in a world of discontinuity, brought about in part through ritual, those few sages who Confucius reveres were able to use ritual to create continuity – but now a graded, hierarchical continuity in which the cosmos is fully linked, and linked in a way that places the ruler in a position of centrality. In short, humans, along with the rest of the world, have been domesticated and organized into a unified realm, with the ruler as the center. The cosmology of monistic harmony was not an assumption. In the “Li yun” it is rather a product of ritual – a ritual order that is seen as domesticating humans, as well as the rest of the cosmos, in the same way that agriculture and husbandry domesticate plants and animals.

Conclusion I mentioned above that one of the enduring features of sacrificial practice in China has been the notion of the dramatically transformative properties of sacrifice, involving a domestication of the capricious world of the divine. The chapter under discussion here has taken aspects of the workings of this practice – the fact that it results in a commitment to a human construction of the world – but has focused its theory of the domesticating powers of ritual on human dispositions. The political goals of the authors are clear. The text was written in opposition to the forms of extreme state centralization that were emerging in the third century BCE and that ultimately culminated in the formation of the first empires in the late third and second centuries BCE.24 The text is not so much arguing against such strong forms of statecraft as it is arguing that centralization can be masked through the ritual transformation of the dispositions of the populace. But, of course, the same argument concerning domestication and the building of continuity through ritual can equally well be used for practices unrelated to or oppositional to such chilling efforts of building a strong state. This same general theory is very productive for explicating (and has been appropriated historically to

23 Liji, “Li yun” ICS, 9.22/62/5. 24 Puett 2008.

368

Michael Puett

support) the operation of the temple network system in later Chinese history, as well as, obviously enough, sacrificial practice in later Chinese history as well.25 And, perhaps more significantly, it is of use in explicating forms of ritual practice outside of China as well. If we are to start taking theories that arose in China seriously as theory, then we need to use them not simply as providing a powerful lens onto rituals from China but also as a lens onto rituals in general. In the case at hand, we have a theory of ritual focused on domestication, and domestication in the specific sense of building continuity among phenomena otherwise seen as disparate and discrete. This is a vision that opens up a potentially very powerful set of issues to explore in a larger comparative context concerning the workings of ritual. Although there is not room here to explore the full complexities of the theories in the Book of Rites, hopefully this brief discussion of the argument in one of the chapters concerning ritual as domestication and the formation of continuity has at least given a sense of the richness of the theories to be found in the text.

25 On the temple network system, see Schipper 1990, 1977; Dean 2006, 2003, 1998, 1995; Katz 1995. On the logic of the sacrificial system, see Puett 2005.


369

References Asad, Talal 1993. Genealogies of Religion: Discipline and Reasons of Power in Christianity and Islam. Baltimore: Johns Hopkins University Press. Dean, Kenneth 1995. “Multiplicity and Individuation: The Temple Network of the Three in One Religion in Putian and Xianyou”. In: Anon.: Proceedings of the Conference on Temples and Popular Culture. Taipei: Center for Chinese Studies. — 1998. Lord of the Three in One: The Spread of a Cult in Southeast China. Princeton: Princeton University Press. — 2003. “Local Communal Religion in Contemporary South-east China”. China Quarterly 174: 338–358. — 2006. Taoist Ritual and Popular Cults of Southeast China. Princeton: Princeton University Press. Harrel, C. Stevan 1974. “When a God Becomes a Ghost”. In: Arthur P. Wolf (ed.): Religion and Ritual in Chinese Society. Stanford: Stanford University Press: 193– 206. Jordan, David 1972. Gods, Ghosts, and Ancestors: The Folk Religion of a Taiwanese Village. Berkeley: University of California Press. Katz, Paul R. 1995. Demon Hordes and Burning Boats. The Cult of Marshal Wen in Late Imperial Chekiang. Albany (N.Y.): State University of New York Press. — 2007. “Orthopraxy and Heteropraxy beyond the State: Standardizing Ritual in Chinese Society”. Modern China 33/1: 72–90. Liji 1992. Chinese University of Hong Kong, Institute of Chinese Studies, Ancient Chinese Text Concordance Series. Hong Kong: The Commercial Press. Legge, James 1885. Li Chi: Book of Rites. Oxford: Oxford University Press. Puett, Michael 2002. To Become a God: Cosmology, Sacrifice, and Self-Divinization in Early China. Cambridge: Harvard University Asia Center. — 2005. “The Offering of Food and the Creation of Order: The Practice of Sacrifice in Early China”. In: Roel Sterckx (ed.): Of Tripod and Palate: Food, Politics, and Religion in Traditional China. New York: Palgrave MacMillan: 75–95. — 2006. “Innovation as Ritualization: The Fractured Cosmology of Early China”. Cardozo Law Review 28/1: 23–36. — 2008. “Human and Divine Kingship in Early China: Comparative Reflections.” In: Nicole Brisch (ed.): Religion and Power: Divine Kingship in the Ancient World and Beyond. Chicago: The Oriental Institute of the University of Chicago: 199–212. Sahlins, Marshall 1996. “The Sadness of Sweetness; or, The Native Anthropology of Western Cosmology”. Current Anthropology 37/3: 395–428. Sangren, Steven. 1987. History and Magical Power in a Chinese Community. Stanford: Stanford University Press. Schipper, Kristofer M. 1977. “Neighborhood Cult Associations in Traditional Tainan”. In: George W. Skinner (ed.): The City in Late Imperial China. Stanford: Stanford University Press: 651–676.

370

Michael Puett

— 1990. “The Cult of Pao-sheng ta-ti and its Spreading to Taiwan: A Case Study of Fen-hsiang.” In: Eduard B. Vermeer (ed.): Development and Decline in Fukien Province in the 17th and 18th Centuries. Leiden: Brill: 397–416. — 1993. The Taoist Body. Translated by Karen C. Duval. Berkeley: University of California Press. Seligman, Adam et al. 2008. Ritual and its Consequences: An Essay on the Limits of Sincerity. Oxford: Oxford University Press. Staal, Frits 1989. Rules without Meaning: Ritual, Mantras and the Human Sciences. New York: Peter Lang. Watson, James L. & Evelyn S. Rawski. 1988. Death Ritual in Late Imperial and Modern China. Berkeley: University of California Press. Weller, Robert P. 1987. Unities and Diversities in Chinese Religion. Seattle: University of Washington Press. Wolf, Arthur P. 1974. “Gods, Ghosts, and Ancestors”. In: Arthur P. Wolf (ed.) Religion and Ritual in Chinese Society. Stanford: Stanford University Press: 131–182.

Joachim Gentz

“Living in the Same House” Ritual Principles in Early Chinese Reflections on Mourning Garments 1. Introduction 1.1 Multiple Approaches to Ritual Literature In recent years, much scholarly attention has been directed towards ritual practice, ritual theory and ritual literature. This has been motivated by diverging research interests, mostly theoretical in nature. Ritual as an etic analytical concept of theoretical interest has only been developed in Europe from the nineteenth century onward and from then on formed the academic reflection on ritual in Western style universities.1 Within that Western discourse various approaches have been developed out of different research questions and analytic perspectives following the diverse disciplinary interests and personal agendas of individual scholars.2 The application of these theories to ritual contexts in prehistory3 or outside of Europe4 has also been critically reflected in scholarly debates since the 1960s,5 and quite some Western language works on Chinese ritual have been published since.6 In contrast, interest in the practice of ritual has been the main driving force for most indigenous forms of ritual studies, or “ritualistics”.7 Emic theoretical interest 1 Cf. Asad 1993. 2 Boudewijnse 1995. Also Bremmer 1998: esp. 14–24. Three important monographies have been published which have collected (Belliger & Krieger 1998) and systematized (Bell 1997; Kreinath & Snoek & Stausberg 2006) the most important Western theoretical approaches to ritual. 3 Cf. Zipf 2003. 4 Cf. Prohl 2003. 5 Cf. Zipf 2003: 11–12. Kreinath & Snoek & Stausberg 2006 regard the year 1966 as a crucial watershed, cf. ibid.: vol. 2, xiii. 6 Cf. the following selection of Western publications relevant fort his study: Kato 1963; Fehl 1971; Cua 1979; Ames 1988; Jao 1990; Vandermeersch 1990; Hankel 1995; Cook 1997; Pines 2000; Cua 2002; Kline 2002; Hagen 2003; Gentz 2004; 2005; 2006; Jiang 2006; etc. For a rather comprehensive overview on Western publications on Chinese ritual cf. Philip Clarts online bibliographies (2009). 7 For the term “ritualistics”, denoting indigenous forms of ritual studies, see Stausberg 2003. See also Stausberg 2006: xiii.

372

Joachim Gentz

in ritual therefore mainly focuses on questions of the origin and correct transmission, about meaning and correct understanding of ritual practice. Chinese discourses about details of ancient rites and corresponding correct ritual practice find their earliest literary forms in ritual manuals and treatises produced especially in the (Confucian) ru 儒 tradition in the fifth through third centuries BC. Chinese scholarly traditions of ritual studies have their origin in constant discussions on the meaning and correct definition of ritual (li 禮) among the hundred schools of thought as they are reflected in the texts of the pre-Qin masters.8 The earliest more abstract theoretical treatises on ritual can be found in the texts of the Liji9 and the Xunzi.10 These early ritual theories, like their further developments induced by political interest in ritual by the emperors Han Xuandi (73–49 BC), Liang Wudi (464–549 AD) and others,11 expressed in works of commentators like Zheng Xuan 鄭玄 (127–200), through Tang (618–907) and Song (960–1279) scholars debating ritual reforms,12 to the philological expertise and kaozheng 考證 scholarship of jingxue 經學 scholars such as Jiang Yong 江永 (1681–1762), Huang Yizhou 黃以周 (1828–1899), Zhang Taiyan 章太炎 (1869–1936) and others,13 have all been motivated by practical, mostly social and political, reasons.14 In China, Korea and Japan the interest in practical applications of ritual is still relevant for research on ancient Chinese ritual today. Throughout Chinese literature on ritual, mourning rites occupy a central place. This is evident already in early Chinese ritual literature.15 In the The Book of Eti-

8 Cf. Wang Qifa 2006. 9 For an interpretation of ritual theory in the Liji which connects these different aspects in a modern reading according to which early Chinese ritual worked especially well because it refined and extended human dispositions, transformed relationships and centred emotions towards harmony see Michael Puett, “Combining the Ghosts and Spirits, Centering the Realm: Mortuary Ritual and Political Organization in the Ritual Compendia of Early China”, in: John Lagerwey, Marc Kalinowski (eds.), Early Chinese Religion: Part One: Shang Through Han (1250 BC-220 AD), Leiden: Brill, 2008, pp. 695-720, pp. 697-700. See also Michael Puett’s contribution in this volume. 10 Cf. Gentz 2004. 11 Cf. Su Zhihong 1990. 12 Cf. Wechsler 1985; McMullen 1987; Meyer 2008. See also Meyer’s article in this volume. 13 Cf. Zhou He 1998. Also Ding Ding 2003. 14 Cf. the studies by Christian Meyer 2008; Chow 1994; Wu Wanju 1999; Deng Shengguo 2005; Lin Cunyang 2008; etc. 15 The DaDai Liji seems to be an exception because it does not contain a single chapter which solely deals with mourning rites. However, the DaDai Liji is known to consist mainly of pastiches and passages from other texts and to have collected material of secondary importance not incorporated into the Liji. Texts on mourning issues have obviously be regarded as so important that all of them have been included into the collection of the Liji.

“Living in the Same House”

373

quette and Ceremonial,16 the Yili 儀禮,17 the chapters on mourning rites constitute the greatest thematic bloc. Accordingly, a great part of the ritual chapters collected in the Book of Rites,18 the Liji,19 deal partly or as a whole with the topic of mourning rites.20 Connected to this large complex of diverse rites are rules concerning the time and different stages of mourning, the dwelling place during mourning, the food and music used during different stages of the mourning time, the calling back of the soul of the dead, the positioning and preparation of the dead corpse for the first mourning stage before the funeral, the handing over, taking over, positioning and administration of mourning gifts, different ways of behaviour towards mourning guests of different status, the selection and shaping of the grave, forms and times of wailing, detailed actions and formula as well as ritual actors in the different stages of the funeral ceremony, sacrifices, prayers etc.21 Among all the texts on mourning rites one very important, perhaps the most important, part is the highly complex field of mourning garments (sangfu 喪服). Of the three chapters22 in the Yili on death rituals one chapter, the “Shi sang li”, deals with all the above mentioned details. The second deals especially with the yu 虞ritual which consists of a set of appeasement rituals for the soul of the deceased conducted after the funeral. The third one, titled “Sang fu”, is entirely dedicated to questions of mourning garments. It clearly stands out as the most important of all the 17 Yili chapters because it is the only one which “has been extracted and treated as a separate text, both in having a wealth of commentaries written expressly for it

16 Thus the title of the only existing English translation of the book published in 1917 by John Steele in London. 17 The Yili basically contains descriptions of ceremonies of the shi-level, i.e. the lowest level of the aristocracy. It preserves material likely produced in the Eastern Zhou period (1045–256 BC), certain parts of it perhaps dating back as early as the sixth or fifth century BC, and purports to represent the orthodox ritual of the Zhou court. Cf. Boltz 1993: 237. 18 Thus the subtitle of the only existing English translation of the book published in 1885 as The Li Ki by James Legge in Oxford as vol. 27 of Max Mueller’s series “Sacred Books of the East”. Now it is also available online at several urls in the www. 19 The Liji is a collection of 49 chapters on usages, prescriptions, definitions, anecdotes, rules, explanations and discussions of early Chinese ritual. Language, style, age (ca. between fourth and first century BC) and authorship of the chapters are extremely divers; content, approaches and viewpoints vary and are often even contradictory to each other. Cf. Riegel 1993. 20 11 out of 46 chapters deal mainly or in great extent with death rituals: “Tan gong 檀弓”, “Sangfu xiaoji 喪服小記”, “Da zhuan 大傳”, “Za ji 雜記”, “Sang daji 喪大記”, “Ben sang 奔喪”, “Wen sang 問喪”, “Fu wen 服問”, “Jian zhuan 間傳”, “Sannian wen 三年問”, “Sangfu sizhi 喪服四制”. Other chapters also contain passages on death rituals. 21 For a collection and analysis of all these details of the first days of ancient Chinese funeral rituals from different early Chinese sources cf. Hankel 1995. 22 I take the “Ji xi li” as a continuation of the “Shi sang li” and not as an own chapter on death rituals.

374

Joachim Gentz

alone, and in having an early transmission of its own.”23 The book catalogue of the Suishu lists a total of four entries for the Yili and 43 separate titles on Mourning Garments, many of them (especially those first in the list) certainly in a direct commentarial way or in an indirect discursive way related to the Sangfu text. This not only reflects the prominence of mourning garments among all other ritual matters but also the independence and great importance of the Sangfu text vis à vis the other chapters of the Yili. In the Liji, we find a similar special focus on mourning garments. Of the eight chapters which deal with death rituals in the Liji24 four chapters are entirely devoted to questions of mourning garment,25 the other four chapters also deal in great length with that topic, focus in these chapters lies on the garments of the ritual actors and not on those of the guests. Looking at these findings the question arises: Why mourning garments? What gives them such an extraordinary outstanding importance in these texts? Why are mourning garments obviously much more discussed, scholarly so much more interesting and intellectually far more stimulating than all the other fields of ancient rites? In early Chinese ritual literature the great field of mourning rites seems to be divided into such texts dealing with the correct setup and conduct of the mourning rites from the perspective of the ritual experts or actors, and texts dealing with the proper behaviour of the participants in the ritual.26 This might be one of the reasons why there is such abundant material of the Sangfu-genre, mourning garments being one of the very few ritual matters that related to non ritual experts and had to be explained to the common people. The great attention attached to mourning garments thus seems to lie in the fact that no other element of the complex mourning ritual has the same potential of linking and representing so many different social groups and perhaps also other important aspects of ritual in one single field at the same time. To prove this hypothesis, I will in the first part of this paper differentiate various social, psychological, economical, political, ethical, cosmological and religious aspects of mourning which find their united expression in the order of mourning garments. Through an application of various ritualogical perspectives that are evident in other early Chinese texts I aim at demonstrating how integrative this topos

23 Cf. Boltz 1993: 235. 24 “Sangfu xiaoji 喪服小記”, “Sang daji 喪大記”, “Ben sang 奔喪”, “Wen sang 問喪”, “Fu wen 服問”, “Jian zhuan 間傳”, “Sannian wen 三年問”, “Sangfu sizhi 喪服四制”. 25 “Sangfu xiaoji 喪服小記”, “Fu wen 服問”, “Jian zhuan 間傳”, “Sangfu sizhi 喪服四制”. 26 The whole first part of the “Jian zhuan 間傳” chapter is an attempt to reconcile these two realms through relating specific mourning garments with the six realms of mourning: bodily appearance (rongti 容體), sound of the voice (shengyin 聲音), speech (yanyu 言語), food and drink (yinshi 飲食), dwelling place (juchu 局處) and garment (yifu 衣服).


375

of mourning garments is in the way it contains and systematizes all these diverse aspects in one system.27 In the second part I will take this approach further into the extant early texts connected to the Sangfu ritual and investigate how the intellectual integration of these various aspects is negotiated in the successive scholarly interpretations of the texts. The Sangfu chapter of the Yili as the earliest of all these texts and as an authority to which all subsequent texts refer, will in the light of the analysis of the first part be analysed not as a text describing or prescribing actual mourning practice but as the first attempt to interpret and explain it through definitions and general principles. It thereby creates a “discourse” the growth of which will then be further detected in different stages of textual production of a related literature on mourning garments. The analytical focus will thus not lie on a reconstruction of early Chinese mourning practice but on a reconstruction of an interpretative discourse as reflected in a set of interrelated early Chinese texts on mourning garments. Taking the Sangfu as the earliest reference point of this discourse, I will then proceed to analyse the negotiation of the Sangfu system as it find its expression through six later texts on mourning garments which follow the Sangfu chapter of the Yili in a commentarial mode: the appended ji-note to the Sangfu (Sangfu ji 喪服記) and the interlinear commentary Sangfu zhuan (喪服傳) which are both direct commentaries to the Sangfu text. And the Sangfu xiaoji (喪服小記), the Da zhuan (大傳), the Fu wen (服問) and the Sangfu sizhi (喪服四制), which are transmitted as chapters 13, 14, 33 and 46 of the Liji. Chinese commentarial literature on ritual texts has been abundant and very well edited28 and explored,29 especially by Chinese and Japanese scholars, in terms of its rich explanations of ritual terminology and material objects. Scholarship has focused on the descriptions contained in the explanatory notes given in the commentarial literature to reconstruct the rules, objects, and forms of ancient Chinese ritual

27 Cf. Lai 2003. Also Ding Ding 2003: 4–7. 28 One of the earliest big collections of commentaries on the Yili has been compiled in the thirteenth century by Ao Jigong 敖繼公 (fl. 1279–1301): Yili jishuo 儀禮集說 in 17 juan, republished in Beijing 2004 in 24 vols. (Chinese traditional style in three boxes). A little earlier Wei Liaoweng 魏了翁 (1178–1237) had written a lengthy commentary on the Yili: Yili yao yi 儀禮要義 in 50 juan, republished in Beijing 2003 in 24 vols. (Chinese traditional style in four boxes). 29 Cf. the bibliographies by Liu Zhaoyou 2003a; 2003b.

376

Joachim Gentz

in much detail.30 Little attention, however, has been paid to the commentarial literature as a literary genre itself. We therefore possess very few studies on commentaries as composite texts which apart from their explanative single contributions also carry a further historically specific overall meaning as a particular genre. I will therefore address the question of the different interests, approaches and meanings of different ritual commentaries as texts in their own right, yet interrelated in a common and growing discourse. Detecting a development in a set of interrelated and inter-referential discourses I will provide evidence of the development of this genre in Early Han. In order to analyze comparable commentaries I will focus my investigation on the Sangfu-commentarial texts within the ritual books Yili and Liji. The Sangfu chapter is special in that two early texts are directly appended to it: the Sangfu ji, a note added to it, and the interlinear commentary Sangfu zhuan. With a copy excavated from a late Western Han (206 BC–8 AD) tomb discovered 1959 in Wuwei, Gansu, the Sangfu zhuan is special not only in being one of the very few commentaries found in an early tomb, but also as being the earliest extant commentary inserted into a text in an interlinear mode. The Sangfu chapter is further special in that it provides the basis of a whole range of texts on mourning garments which are collected in the Liji and relate to it in a more or less direct way. Building upon such an extensive literature on mourning garments transmitted in the Liji, it is possible to reconstruct a “thick” context of rules, and arguments concerning rules, of mourning garments in these Sangfu-commentarial texts. On the basis of such contextual analysis this paper will demonstrate how commentaries devoted to the sangfu-topic argue differently in different stages of exegesis. By comparing the Sangfu zhuan to commentaries with similar exegetical features, the specific form of exegetical strategy and the specific method of interpretation can be exposed. I will demonstrate how further commentarial and sub-commentarial explanations of later texts build upon the earlier texts to establish new and more differentiated modes of interpretation on mourning garments through the invention of new explanation principles which question the kinship system as the sole guiding principle of social relations and move other aspects of representation in the centre of focus. 30 See, for example the amazingly detailed commentary in 15 vols. on the Yili by Kawahara Juichi 川原寿市: Girei shakkō 儀禮釋攷, Kyōto: 1958-1960 with vols. 7 and 8 entirely on the Sangfu chapter. For the Sangfu in particular see also the detailed monographs by Kageyama Seiichi 影山誠一, Mofuku sōsetsu 喪服總, Tōkyō: Daitō bunka daigaku, 1969 and Mofuku keiden chūso hogi ｢喪服｣經傳注疏補義, Tōkyō: Daitō bunka daigaku, 1984 (repr. of 1964 original). Cf. also the two recent monographs by Lin Suying 林素英, Sangfu zhidu de wenhua yiyi: yi “Yili Sangfu” wei taolun zhongxin 喪服制度的文化意義: 以《儀禮・喪服》為討論中心, Taipei: Wenjin cbs, 2000 and Ding Ding 丁鼎, "Yili Sangfu" kaolun《仪礼・丧服》考论, Beijing: Shehui kexue wenxian cbs, 2003.


377

In a conclusion I will lay out the different exegetical voices of the various commentaries and will reconstruct the dynamics of the development of their exegetical strategies as connected to changing historical usages of the ritual literature. A hypothesis on a relative dating of these texts shall follow from this analysis at the end.

2. Mourning Garments as Integrative Systems of Signs On the basis of ritual discussions in early Chinese texts such as the Mozi, the Mengzi, the Xunzi and the Lunyu, I would like to differentiate eight different ritualogical perspectives on particular aspects that can be reflected and conceptualized through the specific choice of mourning garments. 1. Social relationships: The most obvious aspect expressed through mourning garments is the social relationship between the mourning person and the dead person. The expression of kinship relation is one of the most basic functions of mourning garments. The “Feiru 非儒”-chapter of the Mozi begins with a critique against a specific hierarchy of family relationships as it is expressed through the mourning rules of the Ruists, and clearly notes the contradictions which are caused by the combination of different grading criteria such as close and distant kinship (qinshu 親疏) and honorable and low position (zunpo 尊卑).31 However, since social relationships are defined differently within different social systems such as the professional, the political and the kinship contexts, the contexts of family succession, friendship, gender and age, the mourning garments also reflect the relative grades of the social relationships from these different contexts. To give an example: the ritual regulations have to define garments for the mourning of a ruler as well as garments for the mourning 31.儒者曰：「親親有術，尊賢有等．」言親疏尊卑之異也．其禮曰：「喪父母三年，妻，後子三年，伯父叔父弟兄庶子其 (期)，戚族人五月．」若以親疏為歲月之數，則親者多而疏者少矣，是妻後子與父同也．若以尊卑為歲月數，則是尊其妻子與父母同，而親伯父宗兄而卑子也．逆孰大焉． “When the Ruists say: ‘treatment of relatives has degrees, respect towards worthies has grades’ they speak about the differences between close and distant relatives and between the honored and the humble. Their code of rites says: ‘Mourning for a father or mother should last three years; for a wife or eldest son, three years; for a paternal uncle, brother, or younger son, one year; and for other close relatives, five months.’ If it is closeness and distance that determines the number of years and months of mourning, then it is longer for closer and shorter for distant relatives. This means that wife and eldest son are equally close to the father. If it is the respectable and lower positions that determine the number of years and months of mourning, then this not only means that wife and eldest son are respected to the same degree as are father and mother, but that paternal uncles and brothers are regarded as low in position as the younger sons. What could be more perverse than that!” See also the translation by Watson 1967: 124.

378

Joachim Gentz

of a family member, a successor, a colleague or a friend. Following this differentiation of social systems, different sorts of mourning garments could be designed for each of the basic social systems. This, however, would be too complicated since only ritual specialists would be able to decide which garment is appropriate in a given case, and no one would recognize the meaning of the mourning garments anymore. Moreover, it would further complicate the regulations of cases where several mourning rules would apply to one and the same mourner at the same time. The mourning garments are thus reduced to five basic garments which in the case of the Sangfu-text are further differentiated into eleven garments. These five or eleven garments have to be applied to all existing social systems which means that the social systems become comparable through compatible degrees of social relations as defined through the garments. The mourning garments thus function as a common signifier for social relationships of different social systems. It is obvious, I think, that through the correlation of different degrees of mourning garments with different social relationships also different qualities of social relationships are compared. To take up the case of the father and the ruler again, in the Sangfu-text, the social relationships with the father and the ruler both count as relationships of utmost respect (zhi zun 至尊). Equally, the relationships between a wife or a concubine and her husband count as relationships of utmost respect whereas the one between a husband and his wife counts as one of utmost intimacy (zhi qin 至親) which is ranked two degrees lower in the sangfu system. The compatibility of mourning within the first degree of the sangfu system might thus be defined as the sort of mourning which in any relevant social system is expressed towards the utmost respectable person. We may then conclude that there are several social systems which are regarded as equally important, namely the family system, the political system and the system of family succession. Aspects of gender and age cut through these systems of which they form implicit parts. 2. Emotions: Following the ritual theory of the Xunzi (“Li lun” chapter) many of the ritual texts, especially those on mourning, define the function of rites as the regulation of human feelings or emotions (qing 情).32 From this perspective, the mourning garments are considered as expressions of human feelings of mourning. At the same time, they help to regulate the sometimes prevailing sad feelings of mourners in a social and cultural appropriate way and protect them from the pressures of the social norms of a non-mourning context. According to this perspective, the sangfu-system reflects different emotional degrees of mourning. As far as I am aware, there is no reflection about different qualities of mourning apart from these degrees. The question, for example, whether mourn32 This theory is especially important in reflexive ritual texts on music, poetry and mourning. Cf. Gentz 2001: 300–303.


379

ing for the ruler, for the father, for the heir and for the husband – which in each case requires the first degree of mourning garments – has similar emotional qualities is nowhere posed. However, following the conclusion from the first point above, the emotions would be comparable in being the specific mourning feelings directed towards the most respectable person in equally relevant social systems. Connected to the system of emotions is another important point of the sangfu-system, namely its proceeding through different periods of time. Like the ritual regulations regarding mourning diet, dwelling place or going to work, the mourning periods of the sangfu-system are often grounded and explained in terms of the process of waning mourning emotions. The garments are expressions of grief comparable to the expression of a face, they may be taken as physiognomical expressions of the mourning body represented through ritual garments.33 3. Socio-economical wealth: The quality of mourning garments reflects the social and economic wealth of the person wearing them. By displaying the economic potential of the mourner hidden in the nuances of the quality of the hempen cloth the garment shows not only the social relationship to the deceased person but at the same time the social status of the person wearing it. In the chapter on “Moderation in Funerals” (Jie zang 節葬) the Mozi attacks all too elaborate funerals as harming state and society, and in his chapter “Against Ruists” he criticises all too complex burials as meaningless hypocrisies. In his view, idleness and pride are the reasons for excessive rituals. The vehemence with which he argues against elaborate funerals and the fact that he devotes a whole chapter to this matter show the importance of this issue at that time. 4. Knowledge and education: The correct observation of mourning ritual expresses the mourner’s degree of education and ritual knowledge, hence the degree of her or his personal cultivation is expressed through the way she or he pays attention to the proper and accurate dressing. The Lunyu contains several sayings of Confucius in which he criticises unlearned practices of rites and time and again emphasizes learning as essential for the proper conduct of rites. The way he discusses his observations of other people carrying out rituals clearly indicates that the level of learning of the ritual actor is obvious for him when participating at a ritual. In the same way a number of accounts in the Mengzi, the Mozi and other books take funeral customs as indicative of the level of cultivation or barbarism of people from different countries. 5. Moral and political position: Following particular mourning rites and not others is an expression of the mourner’s position in regard to different conceptions of ritual, moral and political order. Since any ritual code represents a very specific concept of socio-political and moral order, any adoption of its rules is at the 33 Cf. the reflections in the first part of the “Jian zhuan” chapter of the Liji.

380

Joachim Gentz

same time an own positioning. In his “Moderation in Funerals”, the Mozi formulates two of such extreme positions which at the same time represent philosophical and political positions of different schools. The texts which I am going to analyse in the second part of this chapter negotiate such positions. All great scholarly debates on ritual in Chinese history discuss different positions in political affairs.34 6. Value principles: The ritual order of the sangfu-system is strictly hierarchical. The hierarchy is based on a set of value criteria which are either implicit or can be defined as ritual principles. The mourning garments thus also express value principles. This point follows from point five, and the Mozi quote in the footnote above clearly reflects this point. 7. Cosmological order. In many ritual texts analogies between ritual order and cosmological order are taken to explain the correctness of certain rites. In the “Sangfu sizhi” chapter of the Liji the sangfu-system is also explained in cosmological terms. Thus, a cosmological order can also be expressed by the sangfusystem.35 8. Religious effects. There are some passages in early ritual texts such as the Shi yu li 士虞禮 chapter of the Yili which relate the mourning rites to the soul of the deceased.36 In this perspective, the mourning garments have the function of appeasing the wandering soul through proper ritual behaviour. These are eight different aspects which can be reflected and conceptualized through the specific choice of mourning garments. Each of these eight aspects can be inferred from the ritual texts from late Zhanguo and Han times. The system of mourning garments had thus the potential to express social, political, psychological, economic, ethical, ritual and religious discourses at the same time.

3. Discourses on the System of Mourning Garments 3.1 The “Sangfu”–chapter The difficult and highly complicated task of any system of mourning garments is the arrangement of these different aspects into a harmonious all encompassing sangfu-system. The “Sangfu”-chapter of the Yili is the most well known early attempt to establish such an all-enclosing system. It operates with three distinct social systems: the family system, the political system and the system of family suc34 Cf. Christian Meyer’s article in this volume. 35 Cf. Lai 2003: 73–91. See also my analysis of the Mawangdui funeral banner: Gentz 2009. 36 In his chapter “Against Ruists” the Mozi calls Ruists hypocrites because they climb the roofs of their houses to call back the soul of the deceased and look into every mousehole to find the soul although they exactly know that nothing is going to happen.


381

cession, and differentiates eleven sets of mourning garments. The text sometimes reflects the asymmetry of the correlations of kinship relations and grades of mourning garments through commentarial remarks like: this garment is worn in order to “match [the garment of a higher ranking guest with whom one is visiting the funeral]” (bao 報) or: when there is “no one [else] to lead the sacrifices” (wu zhu zhe 無主者), or when someone was “living together” (tong ju 同居) with the deceased. The text thus first displays clear signals that there are cases where some of the garments do not correspond to the proper ranks as defined by the regular social systems. It secondly offers hints as to how to solve these deviations of social rank and mourning rank. The text thus shows that it proposes not a perfect construction of an ideal system of social rank and sangfu correspondences but that it rather reflects a correct ritual practice which yet has to take into account, and be further explained by, principles that lie beyond the frame of the social ranking system. Hence the text first of all informs the reader that part of its message, namely the ritual order conveyed through it, may be based in principles that lie outside the social ranking systems. And secondly, it makes itself into a commentary on a ritual practice and thus adopts an exegetical mode. The Sangfu may thus be taken as an early textual commentary to existing social practices. As Guolong Lai has shown, the Mawangdui Diagram of Mourning Garments, the Sangfu tu 喪服圖 can be regarded as an earlier version in form of a diagram that tried to systematize only kinship relations of the first three grades in a system of mourning garments.37 The more complicated task of systematizing more grades of kinship relationship and also further social systems within a singular sangfu-system, like the Sangfu text does, requires a much more sophisticated ordering frame. Since the combination of distinct social systems in a complex group of mourning guests leads to a greater number of possible cases of mourning relationships, the text has to cover more definitions of mourning garments for the various possible social forms. Because the ranking of social status was not only defined through the correlation of social systems but further depended on elements such as cohabitation, age, family name etc. the possibilities were not calculable in a mathematical way but rather could be further developed in endless casuistic constellations. As a consequence, lacunae were necessarily left in the sys37 Gulong Lai has argued that the Diagram of Mourning Garments, the Sangfu tu 喪服圖 which was excavated from tomb no. 3 in Mawangdui represents an older layer of a mourning system than the Sangfu-text from the Yili. The main difference lays, as he demonstrates, in the missing “differentiations between the principal line and the branches, and hence the eldest son and other sons. Nonkin relations are not included. Moreover, the mourning for the wives of all the male kin (including mothers), minors, and outside relatives are not presented. In addition, all the mourning periods for father, son, grandfather, father’s brother, brother, and so on are one grade shorter than those prescribed in the Yili. This is because these inscriptions describe only the basic mourning system and include only the closest relatives within three generations that are needed to establish a system.” Cf. Lai 2003: 59. He further notes that in contrast to the later Han systems in the Mawangdui system “the women’s mourning is not affected by their martial status” (ibid.: 72).

382

Joachim Gentz

tem and provoked problems that had to be solved through additional notes to the text, additional texts and specifying commentaries. 3.2 The Commentaries on the “Sangfu”–chapter There are different types of commentarial texts to the Sangfu text, some of them directly referring to the text: a note (ji 記) and an interlinear commentary (zhuan 傳), some of them forming a sort of commentarial context like a number of chapters in the Liji in the form of notes (ji 記), commentaries (zhuan 傳), questions (wen 問) and a systematical approach (zhi 制). They represent different exegetical modes. Some of them, however, take up, and further develop, the exegetical hints given by the Sangfu text itself. 3.3 “Sangfu ji 喪服記喪服記” First of all there is a passage at the end of the Sangfu text introduced by the term “ji 記” (note) which seems to function as a title marking a different text which had been added later to the Sangfu text. This note added to the Sangfu text is no exception, 13 out of 17 chapters of the Yili have such “ji 記”-notes at their end. The note of the Yili chapter on capping “Shi guan li” is instructive since it contains a Confucius quote and thus positions itself in a Confucian discursive context. Furthermore, the whole note from the Shi guan li, which in the note-text is formulated as a quote from a text called “Guan yi”, is also fully quoted in the “Jiao te sheng” chapter of the Liji, and one passage of the note is also quoted in the “Guanyi” chapter of the Liji.38 The beginning of the note of the “Ji xi” chapter of the Yili has also a short parallel in the beginning of the “Sang daji” chapter of the Liji.39 If these notes of the Yili chapters represent a coherent commentarial layer and can thus be compared, they seem to have existed as independent free floating text passages before they were then included or appended to other texts. The Sangfu ji-note contains more ritual regulations concerning further special cases of mourning such as cases when someone dies in a foreign country, when two persons die at the same time, when a grave has to be transferred to another place or when an uncapped youth has to lead the ceremony since there is no one else who can do it etc. It also contains more specified descriptions on the detailed production and outlook of certain mourning garments. It mentions the explanatory principle of “matching” (bao 報) but does not introduce new basic exegetical principles. Thus on the one hand it basically continues the Sangfu text and fills in casuistic and regulative gaps of the ritual rules left open by it. On the other hand, however, it expands two of the basic systems of the Sangfu: first: it adds a new social system, namely the system of 38 Cf. Sun Xidan 1989: 702–706, 1412. 39 Cf. Hu Peihui 1993: 1913; Sun Xidan 1989: 1128.


383

friends which is not dealt with in the Sangfu text. Second, it introduces a new class of clothes, the clothes for a gongzi (公子), prince. All in all, the note added to the Sangfu text is basically a commentary in the form of an additional text which continues it and expands it a little further within the explanatory frame provided by the Sangfu text. The fact that the zhuan-commentary which comments on the Sangfu text also comments on the ji-note shows that the note was considered by the zhuan -commentary to be part of, and on the same level of, authority as the Sangfu text. As concerns further texts of the ji-genre. The title of the Liji defines it as a collection of ji 記-notes to rites. It contains several chapters which are defined as notes to different ritual matters. Among those chapters there are three chapters on death rites. One of them, the Sangfu xiaoji (smaller notes to mourning garments), deals entirely with mourning garments. The two other chapters, the Sang daji chapter (greater notes to the mourning procedure) and great parts of the Za ji chapter (miscellaneous notes), generally deal with funeral rites.40 The ji-notes of the general chapters on funeral and mourning rites are collections of descriptive rank oriented prescriptions on every detail of action which should respond to the process of the death of a person. They start at the moment of heavy illness and reach up to the time after the burial. They are ordered chronologically according to the ritual process and do not contain any explanations or exegetical interpretations of the regulations. The smaller ji-notes of the chapter on mourning garments, the Sangfu xiaoji, also contains many of such descriptive notes which, very much like the ji-note-part at the end of the Sangfu text, regulate in detail even more particular cases and special circumstances. Two things are especially interesting about these notes. First, one of the notes which repeats a ritual rule from the Sangfu text indicates that those notes were not especially written as additions to the Sangfu text but possibly rather stem from one or several independent sources of mourning regulations which where compiled in a way as to accomplish the given rules of the basic Sangfu text. The second interesting discovery is that many of the ji-notes are taken 40 It is striking that the chapter exclusively dealing with mourning rites does not contain any rules on mourning garments – only regulations of the garments of the ritual actors (the same is true for the mourning chapters in the Yili. Since the Za ji chapter contains all sort of miscellaneous rites we also find a passage on mourning garments in there). Hence funeral rites and mourning garments seem to have been two entirely separated fields of ritual expertise. Although identical systems of time periods, social ranks and kinship relations (represented through different degrees of positioning, locations, gifts, styles etc.) in the funeral rites are interlocked in the same way as in the field of mourning garments, there are no cross references in between the system of the mourning rites which regulate the dealing with the dead corpse and the social and diplomatic etiquette on the one hand and the system of the mourning garments on the other hand side. The first part of the Jian zhuan chapter of the Liji is a systematical attempt to combine these two realms through connecting specific mourning garments with the six different spheres of mourning: bodily appearance (rongti 容體), sound of the voice (shengyin 聲音), speech (yanyu 言語), drinking and eating (yinshi 飲食), dwelling-place (juchu 局處), dresses and garments (yifu 衣服).

384

Joachim Gentz

up in the explanations of the zhuan-commentary of the Sangfu. Thus we may infer that the Sangfu zhuan drew on a text or a text corpus which also provided the basis for the compilation of the Sangfu xiaoji. Interesting is also that some of the Sangfu xiaoji rules can be found in the Gongyang zhuan and Zuo zhuan,41 and the Da zhuan chapter of the Liji also contains numerous verbatim parallels. Looking at these different ji-texts we may for our context first of all define a text genre that might probably be the earliest first and very basic commentarial genre of ji-notes which these chapters constitute, the basic characteristics of which is the descriptive and additive mode of ritual prescriptions which continue, expand and try to complete the text. 3.4 “Sangfu zhuan 喪服傳” 喪服傳 The second important commentarial genre to the Sangfu text is an interlinear commentary called zhuan. The Sangfu zhuan is the only interlinear commentary that has been excavated from early tombs so far. Of the three versions of the Sangfu text which seem to date from late Early Han, excavated 1959 in a Han tomb (possibly also Wang Mang period) in Mozuizi (磨嘴子) at Wuwei in Gansu,42 one version consisted only of the plain Sangfu-chapter and two versions contained the Sangfu zhuan. This proves first that the zhuan was not original part of the Sangfu text and second that it was written before Zheng Xuan 鄭玄 (127–200) who had been assumed to be the author of the commentary. The Sangfu zhuan is partly written in a dialogical form in which questions are posed which rather than real questions are markers of such words, sentences or text passages which the commentator finds striking and worth commenting upon because they otherwise would escape the reader’s attention. The questions thereby open the exegetical stage for commentarial explorations. Since the text itself provides no hints for the peculiarity of these textual locations the commentarial insertions into the text are based on an authoritative knowledge and not on a formal or structural plausibility which is based in the text itself like we find it, for example, in the exegetical strategy of the Chunqiu commentaries.43 The Sangfu zhuan does not add any new cases but comments on terms, expressions, sentences and text passages of the Sangfu text. This commentary follows different exegetical strategies and seems to be an edition of different older exegetical materials. It contains glosses on terms, explanatory definitions of ritual details and dialogical parts in a question-answer scheme which analyze the meaning of the text. This combination of glosses, explanatory narrative and dialogical analytical 41 Cf. Gongyang zhuan Zhuang1.3, Xi 8.4; Zuo zhuan Xi 8.4. 42 Cf. Gansu sheng bowuguan, Zhongguo kexueyuan kaogu yanjiusuo 2005. Cf. also Wuwei Kankan shakubun 1972. 43 Cf. Gentz 2001.


385

exegesis reminds very much to the Gongyang and Guliang commentaries which also belong to the commentarial genre of zhuan and use similar formulations. The glosses are sometimes formulated in dialogical form and sometimes as propositions. This even differs within the versions of the Sangfu zhuan as discovered in the Wuwei tomb in 1959 and as transmitted in the Yili.44 This might point to the possibility that the glosses were taken over from glossaries as parallels with other texts also suggest.45 The explanative passages might stem from ritual handbooks, notes or lists which consisted of detailed descriptions of ritual tools, instruments and garments. We already mentioned that the Sangfu zhuan has many passages in parallel with the ji-notes from the Sangfu xiaoji which it probably copied from this text or from the same text corpus this text was compiled of. The dialogical questions about the meaning of text passages, however, follow the exegetic strategy which the Sangfu text itself had indicated and which is now further unfolded, developed and extended by the Sangfu zhuan. The Sangfu zhuan thereby establishes the Sangfu text as a system which does not solely convey an order of social rank but which also transmit basic ritual principles which have to be discovered and displayed by the commentary in order to explain the proper order of the text. Apart from the three principles of, first, matching accompanying persons, second, acting as a manager of sacrifice for persons who do not have anyone else to act and, third, cohabitating, established by the Sangfu text in the first place, the commentary further introduces a whole series of new ritual principles. These principles enable the commentary to insert two new layers of meaning into the text. First, it defines passages of deviations between social and ritual rank. The Sangfu zhuan thereby redefines the system of social rank through its claim of correspondence or deviation with the ritual rank. Second, it introduces a whole set of ritual principles and thereby inserts its own ritual theory into the old text. The principles are marked through their repetitious formulations alleging through their repetitive occurrence that the principle indeed is a basic principle of the text. Some of the main principles to which the commentary refers time and again are the following: – The principle of honour or respect or authority (zun 尊). The most respectful mourning garments are worn for the most respectful persons of the different relevant social systems. The authority of the family leader is based on a further

44 In both versions of the “Sang fu (Fu zhuan)” chapter which were excavated from a Han tomb 1959 in Wuwei and include the zhuan commentary the commentary writes at the beginning of the chapter: “‘Frayed’ means ‘it is not hemmed’ 斬者，不習 (with silk radical).” Cf. Wuwei Kankan shakubun 1972: 3 and 41: “Qie 習(with silk radical)” is a variant for „ji 緝”. Cf. also Wang Niansun's subcommentary to that passage in Guangya 廣雅 as quoted in Hanyu da cidian: vol. 9, 1012. In the transmitted version of the Yili, however, the commentary comments the same passage as follows: “What does ‘frayed’ mean? It is not hemmed 斬者何，不緝也.” Cf. Yili, p. 1346. 45 Cf. Gentz 2001: 366–371.

386

– – – – – – – – –

Joachim Gentz

principle, the principle that the family may not be allowed to lapse (bu ke yi jue 不可以絕). The principle of following the mourning garments of an authority or the order of an authority, for example Y does not dare to diminish what an authority X does not diminish (X之所不降．Y亦不敢降也). Somehow connected, the principle of belonging to the same body (yu zun zhe (wei) yiti 與尊者(為)一體). Taking social roles, not kinship relations as criteria (不敢殊也). The principle of being intimate, close (qin 親). The same family name (yi ming fu 以名服). The three following principle for wives (furen you sancong zhi yi 婦人有三從之義): before marriage she obeys her father, after marriage her husband, and when he is dead, her son. The principle of downgrading one rank of a guest who is not directly related to the deceased but only accompanying another guest (congfu 從服). The principle to keep with the people (yu min tong ye 與民同也). Moreover there is the basic principle of analogy of ranks of different relationships: the father is like the ruler, the father is like Heaven, to the child as to his wife, the concubine serves her husband like a wife serves her parents in law (妾之事女君．與婦之事舅姑等) etc.

As basic principles we can thus differentiate the relationships of kinship, social rank and personal intimacy, of age, gender, dwelling place, name, living context, pragmatic context, political needs etc. To summarize this second stage of commentarial development within the Sangfu context, we may thus take the zhuan-commentary as a commentary which in its combination of glosses, explanatory narrative and analytical exegesis tries to attain two main aims: first, it tries to clarify the descriptive text-layer of ritual material and action by explaining terms and ritual material contexts. Second, it aims at giving intellectual reasons for the correctness of the rites by giving general and comprehensive principles as further and deeper foundation and entanglement of the isolated single rites of the Sangfu text. It thereby inserts its own ritual principles into the reading of the text. Although it introduces a series of principles, it does, however, not reflect on its own principles in an abstract way. 喪服小記 3.5 “Sangfu xiaoji 喪服小記” The Sangfu xiaoji chapter which we mentioned in the context of the additive jinote-commentaries is probably the first text to formulate such principles in a more general way within the context of mourning garments. In addition to the ji-notes which in the title of the Sangfu xiaoji lead to its classification as a text of the jigenre the Sangfu xiaoji contains further exegetical layers. In these parts the Sangfu


387

xiaoji differs from the ji-notes in being far more analytical and explanatory. These more analytical commentarial parts can be differentiated into three different basic exegetical modes. First, explanations of mourning regulations in the way the Sangfu zhuan provides them, too, giving exegetic principles as basis of the rules. Second, and here it goes one step beyond the exegetical horizon of the Sangfu zhuan, it reflects on the principles of the ritual order. We find in the Sangfu xiaoji a first systematization of principles which we do not find in the Sangfu zhuan. The Sangfu xiaoji defines a set of four basic principles according to which ranks of mourning garments are differentiated: intimacy 親親, authority 尊尊, age 張張, and gender 男女之別. Although the Sangfu zhuan made use of these basic principles in its interpretation it had not defined them explicitly as a set of principles as the Sangfu xiaoji does. 3.6 “Da zhuan 大傳” 大傳 We find them, however, either in exact parallels or in further elaborated forms in the Da zhuan chapter of the Liji. The Da zhuan chapter takes another step of exegetical abstraction and further de-contextualizes these principles using the same categories as abstract principles in a political context as something whereupon no changes could be enjoined by a ruler. Regarding the art of garments (fushu 服術) it then uses a further elaborated set of six expressions: intimacy 親親, authority 尊尊, family name 名, living in the same house or having moved elsewhere 出入, age 長幼, the system of downgrading according to gender, age or kinship relation 從服.46 To explain the last of these expressions: 從服 it again differentiates six categories47 whereas the Sangfu xiaoji only defines two. I shall not go into further details in order not to confuse the reader. The next important step which the Da zhuan takes within the commentarial development is that it formulates all these ordering principles not only for the context of mourning garments where they originally probably were developed, as the Sangfu xiaoji suggests, but expands them as abstract principles to the whole realm of ritual rules. The Da zhuan therefore is in my view a later more systematized and more sophisticated commentary on ritual ordering principles. 46 Guolong Lai has argued that these six principles should be divided into two sets. Given the many more principles which we found in the Sangfu zhuan I do not think that this is a reasonable differentiation. 47.從服有六．有屬從．有徒從．有從有服而無服．有從無服而有服．有從重而輕．有從輕而重. There are six kinds of relationships in regard to downgrading: kinship relations, social relations, an grade of downgrading which for certain reasons (taboo, hierarchy etc.) can not be followed by another mourner, a grade of downgrading which for certain reasons (taboo, hierarchy etc.) can not be accepted by the direct guest, the downgraded person being in a lighter position, the downgraded person being in a heavier position (because of formal criteria like family name, age, or gender). Cf. Sun Xidan 1989: 912.

388

Joachim Gentz

3.7 “Fu wen 服問服問” An interesting point about this abstraction process of the Da zhuan chapter is that the whole first part of the following Fu wen chapter (asking about garments) is an attempt at re-contextualizing the abstract principles of the Da zhuan by explicitely citing the Da zhuan as “zhuan yue 傳曰” and then giving concrete examples from the context of mourning garments for some of the principles formulated in an abstract way by the Da zhuan. The Fu wen can thus be taken as a re-contextualizing commentary to a de-contextualizing commentary (such as the Da zhuan) which again is a further step of exegetical complexity. Moreover, it also further differentiates and specifies the principles formulated by the Da zhuan and hence further develops them. 喪服小記小記” 3.8 “Sangfu xiaoji 喪服 Let us turn back to the Sangfu xiaoji and look at its third important exegetical layer which then will lead us to the last Sangfu related chapter of the Liji, the Sangfu sizhi. The Sangfu zhuan had already drawn an analogy between father and Heaven. Using this analogy, it argues that a wife has to follow the three following principle (sancong 三從). Now the Sangfu xiaoji further develops this analogy in the realm of Heavenly seasons through connecting the mourning periods with the natural periods of the cosmological cycle. This is a further way to explain rites through analogical reference to another system and constitutes a third exegetical strategy of this complex chapter. 3.9 “Sangfu sizhi 喪服四制” 喪服四制 The Sangfu sizhi which seems to be a patchwork of very different text layers,48 correlates the changing of mourning garments not only to the cosmological principle of the changes of the four seasons (sishi 四時) but also to the four emotional principles of kindness (en 恩), ordering patterns (li 理), regulations (jie 節) and acting according to circumstances (quan 權), which are then further correlated to the four basic Confucian ethical values of humanity (ren 仁), righteousness (yi 義), rituality (li 禮) and wisdom (zhi 知).49 Referring to the principle which also had been formulated in the Sangfu xiaoji and the Da zhuan of honouring the honoured (zun zun 尊尊) it then applies this principle to equate father and ruler as an example for 48 A big part of this last Liji-chapter is identical with a part of the “Ben ming 本命” chapter 39 of the DaDai Liji 大戴禮記. In several passages it quotes the Sangfu zhuan, it also quotes the Shangshu 尚書. Other parts are so heterogeneous that the Qianlong 乾隆 editors contended that it could not have been in the original compilation of the Liji. Cf. Legge 1885: 59. 49 喪有四制．變而從宜．取之四時也．有恩．有理．有節．有權．取之人情也．恩者仁也．理者義也．節者禮也．權者知也．仁義禮知．人道具矣.


389

the realization of righteousness (yi). The Sangfu sizhi thus correlates and thereby combines several different systems of reference, interlocks elements of these different systems with one another and then inserts ritual rules from the Sangfu tradition in this multilayered combined system. It creates new connotations of the Sangfu regulations through basing and interweaving it into the web of the most basic Confucian value systems of ancient China.

4. Conclusion First of all, I have conducted the reader through a jungle of texts in regard to their textual dating and editorial reliability being very well aware that we are moving on thin ice in terms of the reliability of the texts. Second, I have worked biased in assuming a development within the different commentarial layers which I have distinguished in regard to a Sangfu topos as an imagined center of these texts. In fact only few of the texts clearly show a text-commentary relationship as we would expect it from the classical concept of text and commentary. Yet, the multi-voiced discourse on the sangfu system relates to earlier (textual and non textual) conceptions and arguments and therefore first of all can be analysed within the frame of a commentarial thema-rhema relationship. Moreover, if we assume that the texts refer to a common valid set of ritual rules defined in the Sangfu text any debate on the correct interpretation aims at establishing an orthodox version of the rules and thus should be interpreted within the traditional frame of the concept of (canonical) text and commentary. If we accept that a development can be assumed throughout the texts, I will now finally try to propose something like an exegetical development in six stages for the commentarial texts which at least in its distinctions might appear plausible to some readers. 1. A commentary collects further casual evidence in the material field of the text. It thereby enriches, expands and tries to complete the text. 2. A commentary clarifies uncertain terms and descriptions on a linguistic, grammatical and lexicographical basis. It establishes basic and comprehensive principles to explain propositions of the text. 3. A commentary summarizes and reflects basic principles in a set of closed principles. 4. A commentary de-contextualizes the set of principles and defines them into an abstract set of principles which is widely applicable to other realms. 5. A commentary re-contextualizes the set of principles creating a new potential of these principles for being valid beyond the frame of the specific context whereby the context gains a greater validity.

390

Joachim Gentz

6. A commentary interweaves basic principles into cosmological, ethical, philosophical, political or other systems through analogy and thereby fixes them in a complex worldview. There seems within the dynamics of the exegetical development of the Sangfu commentaries to be a movement which more and more focuses on principles which are not explicitly expressed in the text but are rather inserted into the text in a first step, systematized in a second and then connected to other ruling value systems whereby the original text in its single exegetical partitions can be applied to present usage in different contexts. In this special case we can see commentarial and subcommentarial attempts to establish new and more differentiated modes of interpretation on mourning garments through the invention of new principles which more and more disregard the kinship system as the sole guiding principle of social relations. Instead they introduce new value principles into the realm of human relationships. They thereby reflect a new approach towards an all too rigid system of human relationships as based on the kinship system and promote new perspectives on defining relationships on the basis of other systems based on new ritual principles which emerged during the late Zhou period and slowly through the intellectual work of many generations found their way into the reading of the old texts. These new perspectives appear also in other spheres of intellectual activity like the political field, the field of law and others. In an analysis of the exegetical layers of the Chunqiu tradition I was able to demonstrate a similar exegetical development starting from the Gongyang zhuan through the different early chapters of the Chunqiu fanlu to the sub-commentary of He Xiu.50 This gives us some further confidence that the model of exegetical development applied to these undatable texts is plausible and might provide a basis for further explorations into the complex realm of early Chinese ritual texts.

50 Cf. Gentz 2001. Cf. also Gentz 2008.


391

Relevant Liji Chapters Sangfu xiaoji 喪服小記: contains ji-notes as well as systematic series of terms and cosmological associations Da zhuan 大傳: contains abstracted systematic series of terms from the Sangfu xiaoji Za ji 雜記: contains ji-notes to mourning rites, mourning garments and other rites, no explanations Sang daji 喪大記: contains ji-notes to mourning rites, not to mourning garments Wen sang 問喪: contains descriptions on mourning procedure with special emphasis on the emotional states, has questions and explicative answers on ritual details. Fu wen 服問: takes up the abstracted systematic series of terms from the Da zhuan and recontextualizes them for the concrete context of mourning garments Jian zhuan 間傳: The first part connects different variants of sorrow to different forms of appearance as it exists within the different fields of mourning. Sangfu sizhi 喪服四制: combines different cosmological and ethical systems and connects them to rules of mourning garments.

Relevant Yili Chapters Sang fu 喪服 with appended ji 記 note and zhuan-commentary: Sang fu zhuan 喪服傳 Shi sang li 士喪禮 describes details of the mourning process Ji xi li 既夕禮 continuation of Shi sang li Shi yu li 士虞禮 describes appeasement rituals for the deceased soul after the burial

Different Genres of Commentaries note (ji 記) interlinear commentary (zhuan 傳) question (wen 問) systematical approach (zhi 制)

Sangfu xiaoji and Da zhuan Principles intimacy 親親 authority 尊尊 age 張張 gender 男女之別

392

Joachim Gentz

Da zhuan Principles intimacy 親親 authority 尊尊 family name 名 living within the same house or having moved elsewhere 出入 age 長幼 the system of downgrading 從服.

Sangfu sizhi Correlative Systems four seasons (sishi 四時) four emotional principles: kindness (en 恩), ordering patterns (li 理), regulations (jie 節) acting according to circumstances (quan 權) four ethical values: humanity (ren仁), righteousness (yi 義), rituality (li 禮), wisdom (zhi 知)

References Akai Kiyomi 赤井清美 (ed.) 1972. Bui Kankan sekibun 武威漢簡釋文. Tōkyō: 小宮山印刷株式會社. Ames, Roger T. 1988. “The Role of Ritual Action (li) in Confucian Philosophy”. In: International Symposium on Confucianism and the Modern World. Proceedings, Taipei: International Symposium on Confucianism. Taipei: Guo ji Kong xue hui yi da hui mi shu chu: 53–81. Ao Jigong 敖繼公 2004. Yili jishuo 儀禮集說. 24 vols. Beijing: Beijing tushuguan cbs. Asad, Talal 1993. “Toward a Genealogy of the Concept of Ritual”. In: Talal Asad. Genealogies of Religion: Discipline and Reasons of Power in Christianity and Islam. Baltimore, London: John Hopkins University Press: 55–79. Bell, Catherine 1997. Ritual. Perspectives and Dimensions. New York, Oxford: Oxford University Press. Belliger, Andréa & David J. Krieger (eds.) 1998. Ritualtheorien. Ein einführendes Handbuch. Wiesbaden: Westdeutscher Verlag. Boltz, William 1993. “I li 儀禮”. In: Michael Loewe (ed.). Early Chinese Texts: a Bibliographical Guide. Berkeley: The Society for the Study of Early China & The Institute of East Asian Studies, University of California: 234–243 (Early China Special Monograph Series 2). Boudewijnse, H. Barbara 1995. “The Conceptualization of Ritual. A History of Its Problematic Aspects”. Jaarboek voor Liturgieonderzoek 11: 31–56. Bremmer, Jan N. 1998. “‘Religion’, ‘Ritual’ and the Opposition ‘Sacred vs. Profane’. Notes towards a Terminological ‘Genealogy’”. In: Fritz Graf (ed.). Ansichten griechischer Rituale. Stuttgart, Leipzig: Teubner: 9–32.


393

Chow, Kai-wing 1994. The Rise of Confucian Ritualism in Late Imperial China: Ethics, Classics, and Lineage Discourse. Stanford: Stanford University Press. Clart, Philip 2009. “Ritual” [bibliography]. http://www.uni-leipzig.de/~clartp/Ritual.html (4.6.09). — 2009. “Death, Afterlife, Tombs, Ancestral Cult” [bibliography]. http://www.unileipzig.de/~clartp/Death_%26_Ancestral_Cult.html (4.6.09). Cook, Scott 1997. “Xun Zi on Ritual and Music”. Monumenta Serica 45: 1–38. Cua, Antonio S. 1979. “Dimensions of Li (Propriety): Reflections on an Aspect of Hsün Tzu’s Ethics”. Philosophy East and West 29/4: 373–394. — 2002. “The Ethical and the Religious Dimension of Li (Rites)”. The Review of Metaphysics 55/3: 471–519. Deng Shengguo 邓声国 2005. Qing dai “wufu” wenxian gailun 淸代《五服》文献槪论. Beijing: Beijing daxue cbs. Ding Ding 丁鼎 2003. Yili Sangfu kaolun 儀禮喪服考論. Beijing: Shehui kexue wenxian cbs. — 2003. “Yili Sangfu” kaolun《仪礼丧服》考论. Beijing: Shehui kexue wenxian cbs. Fehl, Noah Edward 1971. LI. Rites and Propriety in Literature and Life: A Perspective for a Cultural History of Ancient China. Hong Kong: The Chinese University of Hong Kong. Gansu sheng bowuguan, Zhongguo kexueyuan kaogu yanjiusuo 甘肃省博物馆、中国科学院考古研究所 (ed.) 2005 [1964]. Wuwei Hanjian 武威汉简. Beijing: Zhonghua shuju. [Wenwu, Beijing]. Gentz, Joachim 2001. Das Gongyang zhuan. Auslegung und Kanonisierung der Frühlings- und Herbstannalen (Chunqiu). Wiesbaden: Harrassowitz (Opera sinologica 12). — 2004. “Ritus als Physiognomie. Frühe chinesische Ritentheorien zwischen Kosmologie und Kunst”. In: Dietrich Harth & Gerrit Schenk (eds.). Ritualdynamik: Kulturübergreifende Studien zur Theorie und Geschichte rituellen Handelns. Heidelberg: Synchron: 307–337. — 2005. “Ritual Meaning of Textual Form: Evidence from Early Commentaries of the Historiographical and Ritual Traditions”. In: Martin Kern (ed.). Text and Ritual in Early China. Seattle, London: University of Washington Press: 124–148. — 2006. “‘Ritual’: Related Emic Concepts: China”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing Rituals. Vol. I: Issues, Topics, Approaches, Concepts. Leiden: Brill: 63–67. — 2008. “Language of Heaven, Exegetical Skepticism and the Reinsertion of Religious Concepts in the Gongyang Tradition”. In: John Lagerwey & Marc Kalinowski (eds.). Early Chinese Religion: Part One: Shang Through Han (1250 BC–220 AD). Leiden: Brill: 813–838.

394

Joachim Gentz

— 2009. “Ein Augenblick Unsterblichkeit: das Bildprogramm von Mawangdui Banner und Xiuzhen tu”. In: Roland Altenburger, Martin Lehnert & Andrea Riemenschnitter (eds.). Dem Text ein Freund: Erkundungen des chinesischen Altertums: Robert H. Gassmann gewidmet. Bern: Peter Lang: 145–171. Hagen, Kurtis 2003. “Xunzi and the Nature of Confucian Ritual”. Journal of the American Academy of Religion 71/2: 371–403. Hankel, Bernt 1995. Der Weg in den Sarg: Die ersten Tage des Bestattungsrituals in den konfuzianischen Ritenklassikern. (Ph.D. thesis, Münster). Bad Honnef: Bock + Herchen (Münstersche Sinologische Mitteilungen 4). Hu Peihui 胡培翬 1993. Yili zhengyi 儀禮正義. 3 vols. Nanjing: Jiangsu guji cbs. Jao, Tsung-i 1990. “Le Canon des Rites et quelques théories majeures du ritualisme suivant le Commentaire de Zuo des Annales des Printemps et Automnes”. In: Anne-Marie Blondeau & Kristofer Schipper (eds.). Essais sur le rituel II. Colloque du centenaire de la Section des Sciences Religieuses de l’École Pratique des Hautes Études. Louvain, Paris: Peeters: 27–44. (Bibliothèque de l’École des Hautes Études, Sciences Religieuses 95). Jiang, Tao 2006. “Intimate Authority: the Rule of Ritual in Classical Confucian Political Discourse”. In: Peter Hershock & Roger Ames (eds.). Confucian Cultures of Authority. Albany: State University of New York Press: 21–47. Kageyama Seiichi 影山誠一 1969. Mofuku sōsetsu 喪服總 . Tōkyō: Daitō bunka daigaku. — 1984 [1964]. Mofuku keiden chūso hogi ｢喪服｣經傳注疏補義. Tōkyō: Daitō bunka daigaku. Kato, Joken 1963. “The Meaning of Li”. Philosophical Studies of Japan 4: 79–95. Kawahara Juichi 川原寿市 1958–1960. Girei shakkō 儀禮釋攷. 15 vols. Kyōto: hōyū shoten 朋友书店. Kline, T.C. III (ed.) 2002. Ritual and Religion in the Xunzi. New York: Seven Bridges Press. Kreinath, Jens & Jan Snoek & Michael Stausberg (eds.) 2006. Theorizing Rituals. Classical Topics, Theoretical Approaches, Analytical Concepts, Annotated Bibliography. 2 vols. Leiden: Brill (Numen Book Series 114). Lai, Guolong 2003. “The Diagram of the Mourning System from Mawangdui”. Early China 28: 43–99. Legge, James (trans) 1885. Li Chi. Book of Rites. Oxford: Oxford University Press (The Sacred Books of the East 27–28). Lin Cunyang 林存阳 2008. Sanli guan: Qingdai xueshu yu zhengzhi hudong de lianhuan 三礼馆: 清代学术与政治互动的链环. Beijing: Shehui kexue wenxian cbs. Lin Suying 林素英 2000. Sangfu zhidu de wenhua yiyi: yi “Yili Sangfu” wei taolun zhongxin 喪服制度的文化意義: 以《儀禮・喪服》為討論中心. Taipei: Wenjin cbs. Liu Zhaoyou 劉兆祐 2003a. Sanli zongyi zhushu kao 三禮總義著述考. Taipei: Guoli bianyi guan.


395

— 2003b. Yili zhushu kao 儀禮著述考. Taipei: Guoli bianyi guan. McMullen, David 1987. “Bureaucrats and Cosmology: the ritual code of T’ang China”. In: David Cannadine & Simon Price (eds.). Rituals of Royalty. Power and Ceremonial in Traditional Societies. Cambridge: Cambridge University Press: 181–236. Meyer, Christian 2008. Ritendiskussionen am Hof der nördlichen Song-Dynastie (1034–1093): zwischen Ritengelehrsamkeit, Machtkampf und intellektuellen Bewegungen. St. Augustin: Institut Monumenta Serica. Pines, Yuri 2000. “Disputers of the Li: Breakthroughs in the Concept of Ritual in Preimperial China”. Asia Major 13/1: 1–41. Prohl, Inken 2003. “Vom Ritual zur Performance – neuere Forschungen und Perspektiven”. In: Carola Metzner-Nebelsick (ed.). Rituale in der Vorgeschichte, Antike und Gegenwart. Rahden: Marie Lidorf: 201–210. Puett, Michael 2008. “Combining the Ghosts and Spirits, Centering the Realm: Mortuary Ritual and Political Organization in the Ritual Compendia of Early China”. In: John Lagerwey & Marc Kalinowski (eds.). Early Chinese Religion: Part One: Shang Through Han (1250 BC–220 AD). Leiden: Brill: 695–720. Riegel, Jeffrey Kenneth 1993. “Li chi 禮記”. In: Michael Loewe (ed.). Early Chinese Texts: a Bibliographical Guide. Berkeley: The Society for the Study of Early China & The Institute of East Asian Studies, University of California: 293–297. Stausberg, Michael 2003. “Ritual Orders and Ritologiques. A Terminological Quest for some Neglected fields of Study”. In: Tore Ahlbäck & Björn Dahla (eds.). Ritualistics. Based on Papers Read at the Symposium on Religious Rituals Held at Åbo, Finland, on the July 31–August 2, 2002. Åbo: Donner Institute for Research in Religious and Cultural History: 221–242. — 2006. “Introductory Essay”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing Rituals: Issues, Topics, Approaches, Analytical Concepts, Annotated Bibliography. Vol. 1. Leiden: Brill: xiii–xxv. Steele, John (trans.) 1917. The Yili, or Book of Etiquette and Ceremonial. London: Probsthain & Co. Su Zhihong 蘇志宏 1990. QinHan li yue jiaohua lun 秦漢禮樂教化論. Chengdu: Sichuan renmin cbs. Sun Xidan 孫希旦 1989. Liji jijie 禮記集解. 3 vols. Beijing: Zhonghua (Shisan jing Qingren zhushu). Vandermeersch, Léon 1990. “Ritualisme et juridisme”. In: Anne-Marie Blondeau & Kristofer Schipper (eds.). Essais sur le rituel II. Colloque du centenaire de la Section des Sciences Religieuses de l’École Pratique des Hautes Études. Louvain, Paris: Peeters: 44–56 (Bibliothèque de l’École des Hautes Études, Sciences Religieuses 95). Wang Niansun 王念孫 1992. Subcommentary to that passage in Guangya 廣雅. In: Luo Zhufeng 羅竹風 et al. (eds.). Hanyu da cidian 漢語大詞典. Vol. 9. Shanghai: Hanyu da cidian chubanshe: 1012.

396

Joachim Gentz

Wang Qifa 王啟發 20062. “XianQin zhuzi lun li yu fa 先秦諸子論禮與法”. In: Wang Qifa 王啟發. Lixue sixiang tixi tanyuan 禮學思想體系探源. Zhengzhou: Zhongzhou guji cbs: 113–142. Watson, Burton 1967. “Against Confucians”. In: Burton Watson. Basic Writings of Mo Tzu, Hsün Tzu, and Han Fei Tzu. New York, London: Columbia University Press: 124. Wechsler, Howard 1985. Offerings of Jade and Silk. Ritual and Symbol in the Legitimation of the T’ang Dynasty. New Haven: Yale University Press. Wei Liaoweng 魏了翁 2003. Yili: Yili yao yi 儀禮要義. 24 vols. Beijing: Beijing tushuguan cbs. Wu Wanju 吳萬居 1999. Songdai sanlixue yanjiu 宋代三禮學硏究. Taipei: Guoli bianyi guan. Zhou He 周何 1998. Lixue gailun 禮學概論. Taibei: Sanmin. Zipf, Gabriele 2003. “Formalisierung, Reduzierung, Inszenierung – Zur wissenschaftlichen Konzeption von Ritualen und ihrer Umsetzung in der Interpretation archäologischer (Be-)Funde”. In: Carola Metzner-Nebelsick et al. (eds.). Rituale in der Vorgeschichte, Antike und Gegenwart. Studien zur Vorderasiatischen, Prähistori-schen und Klassischen Archäologie, Ägyptologie, Alten Geschichte, Theologie und Religionswissenschaft. Interdisziplinäre Tagung vom 1.–2. Februar 2002 an der Freien Universität Berlin. Rahden: Verlag Marie Leidorf: 9–16.

Christian Meyer

Interpretations of Confucian Ritual (“li”) in Chinese Scholarly Discussions in the Eleventh Century1 Introduction Subject of this paper is the applied hermeneutics of Confucian ritology in middle Northern Song time (eleventh century) as an example of Chinese reflexivity on ritual. I will first focus on two major preliminary aspects: Firstly, a specialty of this approach is due to the examined material a “post-classical” perspective: Already in antiquity2 the Chinese possessed a sophisticated ritual culture not only in practice, but also of reflecting about it. Much or most of the Chinese Confucian ritual culture was already fixed in that time. The same was the case with the Confucian canonical scriptures, which all go back to ancient time3 and which provided a widely accepted base for discussing ritual. Insofar the case studies of Michael Puett and Joachim Gentz included in this volume about two of the three ritual classics represent a rather still formative and more open stage of Confucian ritual discourse, while that of the eleventh century builds up on a very much established base of classical texts and history of interpretation. This already leads naturally to a first question: Why did Chinese even later discuss – and, as it seems, needed to discuss – ritual? Or how far was there still any space (and any meaning) to discuss it – beyond reproducing the same ideas again and again? Furthermore, another point how our later case differs from the Zhanguo and Han periods is given by the sources: Although the earlier stages are more basic and formative, one explicit advantage of this later time is that we are able to a higher extent to reconstruct concrete discourses with people, whose biographies and circumstances we know – or at least are able to reconstruct – much better. In terms of ritual discourses the two basically different situations of classical and postclassical time also represent very different paradigms of reflecting ritual which could possi1 This paper was originally presented in the panel “Reflexivity and Discourse on Ritual”. 2 For our purpose this formative period may be defined as including the Han dynasty (206 BC to 220 AD). 3 Most of the original canonical scriptures (jing 經) were written until the end of Zhanguo period (or the latest in the early Han), important commentaries – among them some earlier ones which partly became canonical – were composed until the end of the Eastern Han (until 220 AD). See Loewe 1993.

398

Christian Meyer

bly even be found in many other ritual cultures with a relatively long textual tradition. Secondly, there are different ways of interpreting ritual debates. This can even be connected to the question of the relations of “reflexivity” and ritual.4 In an earlier article I have already explicated, how the ritual discourses in the eleventh century can be understood as closely connected to political debates, identity issues and to power struggles.5 This would, of course, already give a partial, meaningful answer also to the question why ritual issues were discussed again and again. In that way reflexivity6 would be understood in the way that ritual matters became representing, for example, power. And indeed ritual debates could be well understood as power struggles or part of them. This is also no surprise as much of the concept of Chinese ritual(ity) (li 禮) was explicitly dealing with power, especially as rituals were conceptualized mainly hierarchically and main parts were closely connected to the emperor and so to politics. Another understanding of reflexivity is however how ritual itself was reflected, either in arguments in debates about concrete cases, or in more abstract philosophical thinking. My interest in this paper is rather in this question, i.e. how Confucian Chinese in the eleventh century argued and thereby reflected about ritual(ity) or the Chinese concept of “li”. I will first introduce shared concepts and arguments about ritual, developed until the Song dynasty as a hermeneutical base of ritual discussions in the eleventh century. Then I will demonstrate, especially referring to one example, how two very different types of ritual reflections, one more rigid and the other more flexible in attitude, were built up on this same base. And finally I will try to identify these two as somehow “necessary” types of interpretation in ritual discourse, which complement and balance each other in the process of adapting ritual to actual situations.

1. Shared Concepts and Arguments About Ritual Developed Until the Song Time as a Hermeneutical Base of Ritual Discussions in the Eleventh Century The textual and hermeneutical base of interpreting Confucian ritual seems to be already largely fixed until Northern Song time. In addition to the three ritual clas4 This paper was originally held in a panel about “ritual and reflexivity”, organized by Udo Simon at the same conference, see volume IV of this publication. 5 Meyer 2010. 6 For a differentiated reflection on the term of reflexivity, also related to ritual, see Stausberg 2006.

Interpretations of Confucian Ritual

399

sics, we also have the other canonical scriptures which contain a huge amount of normative material regarding ritual,7 but also ritual cases recorded in history books which served as precedents. Finally, of course, we have the commentarial tradition, which partly became also canonical or at least highly influential.8 These sources already provided a substantial base of ritual material like rules and cases (precedents). At another level we have, however, something like hermeneutical rules how to interpret classical ritual norms and actual cases: While there was no comprehensive and systematic hermeneutics of understanding [the classics, or] Confucian ritual, there was at least a series of principles well known and applied in the tradition: 1) First of all and most basic: Norms for ritual were to be found normally in the very concrete normative rules from the classics, which often also exemplarily show the ideal behavior or how to act as a sage (a shengren 聖人) like Confucius himself. 2) On the other hand, there is the principle that concrete rites can indeed change or can be changed: According to the Analects of Confucius (Lunyu) rites of one ancient dynasty were different from the previous one (cf. LY 2,23 and 3,14), and “Confucius himself” (as said in the Lunyu) suggested in one saying modifications in detail as long as the basic (ethical-hierarchical) meaning is not changed.9 Therefore, although basically the norms of the later periods like of the Song dynasty were set by the idealized early Zhou-Dynasty,10 not only were changes principally acknowledged as possible, but ritual , although so important for Confucianism (Ru), was seen as something relative to a more basic normative pattern of obligation (yi 義) or the idea of “being appropriate” (yi 宜) to a respective situation.11 This difference is the point where hermeneutically space for interpretation was opened. 3) Thirdly, in order to judge situations in concrete cases discussants would further – in a kind of casuistic approach – refer to positive and negative examples from the past, which have to be interpreted contextually in its historical conditions as 7 The process of canonization was actually not yet completed up to the full number of thirteen books, but aside from the nine accepted classics of this time (jiujing 九經) others like Xunzi or Mengzi were at least accepted as important ancient sources to which it was possible to refer. 8 For the dynamics of the textual ritual discourse within the commentarial tradition see Joachim Gentz’s paper in this volume. 9 The locus classicus is LY 9,3. 10 And the perceived interpretation of it by the imagined “Confucius”. 11 Cf. Legge 1885: Liji, chap. Quli (shang) li cong yi 禮從宜; see also the explanation of yi 義 (obligation) through yi 宜 (appropriate[ness]) in Zhongyong. Besides, “li” here normally translated as “ritual”, could not only have a more concrete, but also a more abstract aspect of meaning as basic principle and norm in contrast to more concrete rituals (or “ceremonies”) (yi 儀).

400

Christian Meyer

precedents reflecting classical norms – and with good or bad fortune of the acting rulers as evidence for “truth” or the appropriate decision. 4) Fourthly, if ritual meant to do “what is appropriate” (yi 宜), which is maybe the most general understanding of ritual (li) in Confucian tradition, there were ways of thinking ritual (li) in more principal (although never fully systematized) terms: a) Ritual (li) was thought as referring to and actually establishing hierarchical sociopolitical order.12 b) Another understanding, not necessarily in contradiction to the first, was the idea that ritual (li) follows “renqing 人情”, here approximately translated as “natural human feelings”. There is, however, a range of possible understandings of this term: It could mean natural feelings like in mourning for the deceased parents (cf. LY 17,21), but also rather “common sense”, “what everybody naturally feels to be right”.13 This interpretation (b) could also merge or go well together with the idea of “changes of the times” (principle 2), that ritual would not only change with the establishment of dynasties, but also, at least in details, with the changing conditions, customs, and feelings of the people. c) Thirdly, there is an equation of the two homophones li 禮 and li 理, one meaning “ritual” as used here before. The other one has been translated as a more naturally given, cosmological “pattern” (or principle) which can and is to be found in any thing or any situation.14

12 See for example Liji (Legge 1885), chap. Yueji 522: 樂者天地之和也, 禮者天地之序也, here together with Musik (yue) explicitly correlated to xu 序 (hierarchical order) and he 和 (harmony). For Cheng Yi’s explication of li (ritual) as even cosmologically conceptualized hierarchical order based on this saying from Yueji see ECJ Yishu 18.225f.: “禮只是一箇序, 樂只是一箇和. […].「禮莫是天地之序? 樂莫是天地之和?」曰:「固是. 天下無一物無禮樂. 且置兩隻椅子, 纔不正便是無序, 無序便乖, 乖便不和.」”. 13 And in a normative sense it could possibly also be understood as “how one should feel” (for Cheng Yi see ECJ Yishu 11.127: 禮者因人情者也, 人情之所宜則義也. 三年之服, 理之至, 義之盡也). 14 This aspect which is important for example for Cheng Yi and was based on the locus classicus “Li, li ye 禮也者, 理也” (Liji, chap. Zhongni yanju) is not further pursued in this article, for more detailed information cf. my book Meyer 2008: esp. 67, 483, 492–493, 513, 529–536.


401

2. Diverging Interpretations of li (Ritual) in Northern Song Dynasty: The “Ritualistic” and the Flexible Mode of Perceiving and Interpreting Ritual In Northern Song times (from 960–1126) we find major ritual discussions which provide us with contributions by major intellectuals of its time, many of them very prominent and famous thinkers, well-known until today. What I want to show here are two very different modes of interpreting classics and reasoning about ritual which represent two views of the role of ritual and probably two kinds of diverging general attitudes and worldviews. The two groups are mainly represented by Ouyang Xiu 歐陽修 (1007–1072) and the Su family on the one hand, and the more “conservatives”, Cheng Yi 程頤 (1033–1107), but also Sima Guang 司馬光 (1019–1086) and Zhang Zai 張載 (1020–1077) on the other hand. I will concentrate on Ouyang and Cheng in the following: First of all, ritual was a matter which within Confucian circles could attract more or less interest according to intellectual conviction and maybe due to personal character of the involved people: While scholars like Ouyang Xiu and Su Shi, in fact, were interested in ritual, too, they had a rather loose affinity to it compared to downright “ritualists” like the so-called Neo-Confucians Cheng Yi or Zhang Zai.15 2.1 Ouyang Xiu’s Position: a Flexible Interpretation of li For the first group a lively and close-to-life expression of correct human relationships performed in ritual seemed to be the most appropriate. Exemplary models could be found in the classics, but did not have to be applied normatively in every detail. Their hermeneutical keyword was “natural human feelings” and “common sense opinion” of the people (renqing). Rituals should follow these natural feelings, express and thereby strengthen them. According to different times and situations rituals could be re-designed “according to the original intention of the Sages” (shengren zhi benyi 聖人之本意), an important general hermeneutic rule of interpreting the classics in general.16 A good example for this was the case of an emperor, Yingzong 英宗 (reigned from 1063–1067), who was adopted from a minor branch of the imperial family as there was no biological son, and so no direct heir to his predecessor. After the death of the previous emperor he succeeded him and also properly fulfilled the role of a pious son to the adoptive father as if he were a biological son – a proper 15 For other genuine ritualists like Li Gou 李覯 (1009–1059) or Lü Dalin 呂大臨 (1046–1092) see Meyer 2008 (Index) or for Li Gou Hsieh 1997. 16 See for example Ouyang Xiu 1975: 34.867, cf. Bol 1992: 195–196, but also Cheng Yi’s writings offer similar quotations as he may have been influenced by discourses of his time which were at least partly dominated by Ouyang as a major opinion leader of his time.

402

Christian Meyer

example of fulfilling mourning rites. However, he was brought up by his natural parents, and had still feelings towards his also already deceased biological father. So he expressed to the government the wish to venerate the biological father in a more particular way. For Ouyang Xiu, who was a vice-chancellor at this time, this was a natural human feeling (renqing) which should be expressed also in ritual forms, although surely not in the same way as to the adoptive father and former emperor, but still differently from dealing with him as if he were a normal, more distant relative. As question how to deal with it in detail most properly and appropriately, it was transmitted to the bureaucracy to be worked out. 2.2 Cheng Yi’s Position: a Systematical and “Ritualistic” Interpretation of li (Ritual) For the other group, that of the more ritualistic, nowadays so-called “Neo-Confucians” ritual, was an important matter which also played its role as educative embodiment of Confucian values in contrast to Buddhist, Daoist and popular religious (e.g. geomantic) practices.17 I will now refer especially to the famous Neo-Confucian Cheng Yi: Although he did not categorically reject adaptations to the time and its changed customs in minor aspects, ritual, as it was created and in details designed by the Sages of the past, meant for him basically the best form to express and intensively exercise Confucian values and attitudes. A strict adherence to the classics and ancient rites (guli 古禮) could help “learners of the way” to attain sagehood.18 It would be exercised in each situation in an attitude of attentiveness (jing 敬). In the state of perfection (as a sage [shengren]) one would act in accord to the norms naturally and spontaneously. Beyond this personal use, correct ritual norms had to be adhered to also in the higher (and more “public”)19 imperial arena as embodied visibility and visible embodiment of li enacted by the emperor as a “father” who could serve as a “role model” for the whole empire.20

17 These had become common even in families of those who were educated in the “Confucian” Classics (the so-called shidafu 士大夫), cf. Ebrey 1984: 219–245. 18 This is found programmatically in Cheng Yi’s early “Treatise on What Yen Tzu Loved to Learn” (ECJ Wenji 8.577–578), translated by Wing-tsit Chan (Chan 1963: 547–550), for the use of guli see for example Songshi jishi benmo (Chen Bangzhan 1981: 45.439; Shao Bo 1997: 20.159–160. For Cheng Yi’s ritualism in general see also my article “Cheng Yi as a Ritualist” (Meyer 2007). 19 Public here in terms not of general or public access, but representability. 20 For the role of the emperor as father (and the empress as mother) of and role model for the people see for example the struggle about the deposing of the Empress Guo, see Songshi jishi benmo (Chen Bangzhan 1981: 25.194): 皇上天下之父, 皇后天下之母.


403

So when it came to the case of the adopted emperor and his wish to venerate his biological father this case was taken as a serious and critical matter by many rather “conservative” Confucians like Cheng Yi: Instead of taking it as a rather private case of the emperor and his personal human feelings (renqing) as suggested by the official question from the government, this other group in the bureaucracy saw a deep potential interference with the proper hierarchical order of the imperial clan system, relevant not only to the order of succession of the throne and the more “religious sphere” of the ancestral shrine, but also to the legitimacy of the dynasty. 2.3 Diverging Suggestions in the Ritual Case of Yingzong (the “Debate About Granting a Title to the Prince of Pu”) The two positions of government and “conservative bureaucrats” resulted in two radically contrarian suggestions: While the former group suggested finally a title “parent” (qin 親), as a biological father was still be identified in the “Mourning garments” (Sangfu 喪服) chapter of the Yili 儀禮 with the term father (fu 父) and also treated as a special case in its casuistic system.21 The second group saw this use of “father” in the classics as rather technical term, which, however, could not be used as a title in a ritual context. Their main argument against it was that as “there is only one sun at the sky”, there could “only be one father”, and one loyalty, as well as there is also “only one emperor”22. Their countersuggestion was to address the deceased biological father in ritual context as Imperial Uncle (huangbo 皇伯). Only this would fit the relation of a son of the former emperor – either natural or adoptive – to his biological father who in this case was a cousin of the former emperor. If adoption was radically thought of as a complete change of the position of the adopted one in the family or clan system, then according to his new position the adopted son had now to be regarded as a nephew to his biological father. This approach was in its way in accord with a perfect system, but indeed very rigid. The idea, however, that a biological father would suddenly be called an “uncle” seemed to the first group not only absurd, but even against “natural human feelings” (renqing). Against most of the bureaucracy, but in accord with the emperor they enforced their final solution.23 Although this case had also strong political implications and indeed caused a strong power struggle which undermined the government’s acceptance and support 21 Steele 1917: vol. 2, 18 t, for the Yili and its commentarial structure see Joachim Gentz’s paper in this volume. 22 Liji, chap. Fangji “tian wu er ri, tu wu er wang, jia wu er zhu, zun wu er shang 天無二日，土無二王，家無二主，尊無二上”; and, as quotation of Confucius, in Liji, chap. “Zengzi wen”. Cf. also Liji chap. “Sangfu sizhi”, DaDai Liji chap. “Ben ming”, Mengzi, chap. “Wanzhang shang” 5A4, Gongyang zhuan Wen 9.1, Kongzi jiayu 50.15 23 In an attempt of making a compromise an earlier suggestion of applying the attribute “huang” (Imperial) to the biological father in combination with addressing him as “parent” was abolished.

404

Christian Meyer

amongst the generation of younger officials,24 it can not simply reduced to that factor of power struggle: The two distinguishable interpretations in fact fitted just to their original respective perception and conception of ritual (as outlined above). Other cases or writings witness a clear consistency of their respective thinking.25

Conclusion I would like to summarize this situation in the following way: There were two modes of ritual hermeneutics, both based in the classics and built on principles which were not necessarily exclusive. However, these two modes clearly clashed in the mentioned case of adoption: Here we find a conflict of the idea of a hierarchical, consistent system with much more complex and dynamic social realities. In the case of adoption human feelings were rather not naturally expressed by the normal idea of mourning grades according to regular social relations in a clan, and it was an open question26 where to draw the line between a special treatment of an biological father and interference with the regular model. Ritual rules – in the version of Cheng Yi and Sima Guang – became rather normative and, if necessary, feelings would rather have to be regulated to a desired norm. On the level of principles there was therefore a conflict of the idea of a strict hierarchy associated with “rituality” (“li”) and the demanding of the feeling of clear loyalty on the one hand clashing with the idealistic principle of ritual as expressing natural human feelings on the other. On the level of attitude and “thinking ritual” this was connected to a conception of ritual as not only systematical, even cosmological principle, but also absolutely normative and strictly to be followed rules (mostly formulated and so found already in the classics) in contrast to the idea of ritual as something, what much more flexibly followed changing circumstances and human customs, could be found in the natural human feelings and had to be adapted even in compromises. If we relate this further to the question of reflexivity in general, we find in the eleventh century of the Chinese Song dynasty an explicit culture of reflecting on ritual and a sort of “applied ritual hermeneutics”. However, this hermeneutical approaches and general conceptions consisted of rather diverse advices which had never fully systematized. They would rather be applied from case to case. Confu24 For an account of the political implications and the related power struggle in the bureaucracy see again my book (Meyer 2008), for a shorter English version see also Meyer 2009. 25 For another clash between these two interpretations see also Cheng’s struggle with Su Shi in 1086 about the mourning rites for the late chancellor Sima Guang (Meyer 2008: 289–297). 26 As mentioned above the Sangfu chapter in the classics already made a compromise and provided rules for the case of adoption, but it concentrated only on the categories of mourning periods and left out the case of formal addressing, thereby still leaving space for later interpretation.


405

cian ritual tradition left – more pragmatically than systematically – space to different interpretations which like in our case, however, sometimes collided. The two outlined opposite perceptions and conceptions of ritual represent a range of possible interpretations from rigorosity to flexibility. The differences may not only have existed due to individual conception and worldview or character (either more rigid or more flexible), and so of “taste”. I wonder if this could not rather be understood also as a very general, intrinsic pattern in regard to ritual as (1) human behavior and rules for it can always follow either a more strict or more flexible mode (in a variety of grades) and so (2) also application of ritual rules in actual situations had always to be negotiated in discourses again and again. Both positions might then emphasize somehow necessary and integral elements of ritual (or li): So while one position is arguing for stability and authenticity of a tradition which is seen as a base of legitimacy and which can not easily be abolished, the other one argues rather for the necessity of flexibility and adaptations as in life circumstances and contexts change. Insofar ritual discourses may not only reflect power struggle (as I hinted to before), but also a struggling for necessary adjustments and still loyalty to the tradition and authenticity. It might also be seen as interesting result of the long-term struggle between the two positions that in following dynasties the more rigid “ritualistic” version of Confucianism, so-called Neo-Confucianism, which with clear rules probably allowed easier orientation and was so able to promise stability, prevailed over the more flexible, open one which relied on case-by-case decisions and vague “common sense” or “people’s feeling” (renqing), and so the first group became the orthodoxy in later times. At the same time it may be remarked, that however in the long run, although this rigorosity was in practice not always followed so strictly (at least not in Cheng Yi’s version), after it had been rigorously emphasized in China’s last dynasty27 it contributed to the image of Confucian inhumanity when revolution took place in modern times in 1911 and Confucianism was finally completely abolished.28

27 See Chow 1994. 28 A major charge against Confucianism by the New Culture movement in the decade after the revolution of 1911 was expressed in the slogan of “Killing the people through ritual” (yi li sharen 以禮殺人).

406

Christian Meyer

References Bol, Peter K. 1992. This Culture of Ours: Intellectual Transitions in ‘T'ang’ and Sung China. Stanford: Stanford University Press. Chen Bangzhan 陳邦瞻 1981. Songshi jishi benmo 宋史紀事本末. Taibei: Liren shuju 里仁書局. Chan, Wing-tsit 1963. A Sourcebook in Chinese Philosophy. Princeton: Princeton University Press. Chow, Kai-wing 周啟榮 1994. The Rise of Confucian Ritualism in Late Imperial China. Ethics, Classics, and Lineage Discourse. Stanford: Stanford University Press. Ebrey, Patricia Buckley 1984. “Conceptions of the Family in the Sung Dynasty”. Journal of Asian Studies 43: 219–245. Er Cheng ji 二程集 (ECJ) [The Collected Works of the Cheng Brothers; Cheng Hao 程顥 & Cheng Yi 程頤] 1981. Beijing: Zhonghua Shuju. Hsieh, Shan-yuan 1979. The Life and Thought of Li Kou, 1009–1059. San Francisco: Chinese Materials Center. Legge, James (trans.) 1885. The Li Ki. Oxford: Clarendon Press (Reprinted in Max Müller (ed.), The Sacred Books of China, The Texts of Confucianism, vols. 27–28, The Sacred Books of the East, Delhi: Motilal Banarsidass, 4th edition 1976. Loewe, Michael (ed.) 1993. Early Chinese Texts. A Bibliographical Guide. Berkeley: The Society for the Study of Early China. Meyer, Christian 2007. “Cheng Yi as a Ritualist”. In: Oriens Extremus 46: 211–230. — 2008. Ritendiskussionen am Hof der nördlichen Song-Dynastie 1034–1093. Zwischen Ritengelehrsamkeit, Machtkampf und intellektuellen Bewegungen. (Originally Ph.D. Thesis, University of Erlangen–Nürnberg 2003). Nettetal: Steyler Verlagsbuchhandlung. (Monumenta Serica Monograph Series 58). (with Online Index: http://downloads.steyler-missionare.de/18/1759/1/69942989465125140410.pdf). — 2010. “Negotiating Rites in Imperial China: The Case of Northern Song Court Ritual Debates from 1032 to 1093”. In: Ute Hüsken & Frank Neubert (eds.). Negotiating Rites. Oxford et al.: Oxford University Press (Oxford Ritual Studies Series). Ouyang Xiu 歐陽修 1975. Xin Tangshu 新唐書. Beijing: Zhonghua shuju. Shao Bo 邵博 19972 [1983]. Shaoshi wenjian houlu 邵氏聞見後錄. Edited by Liu Dequan 劉德權 & Li Jianxiong 李劍雄. Beijing: Zhonghua shuju. Stausberg, Michael 2006. “Reflexivity”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing Rituals. Vol. 1: Issues, Topics, Approaches, Concepts. Leiden: Brill: 627–646. Steele, John (trans.) 1917. The I-li or Book of Etiquette and Ceremonial. 2 vols. London: Probsthain. (Reprinted Taibei 1966).

Gil Raz

Ritual Theory in Medieval Daoism 1. The vast majority of texts in the Daoist canon and in the extra-canonic collections possessed by Daoist priests are ritual manuals. Ritual has been the main vehicle for interaction between the Daoist priesthood and the general population. Yet, we know relatively little about the practice and meaning of Daoist ritual.1 The theoretical underpinnings of Daoist ritual remain almost unexplored.2 This is partly due to the fact that there are few explicitly theoretical treatises that deal with ritual. A more important reason for this gap, however, may be the implicit assumption that discourse on ritual in Daoist texts is simply about correct practice and not about meaning, let alone theory. I argue in this paper that Daoists did indeed construct their ritual synthesis with certain theoretical assumptions that informed their understanding of ritual, from structure and performance, to meaning and explanations of efficacy. I should stress here that I am not trying to explain claims of efficacy made in Daoist texts. Rather, I am seeking to show an underlying set of assumptions that help explain the construction or structure of Daoist ritual and the framework within which such claims make sense. These assumptions may not be framed as a coherent theory, or theories, in modern terms. Rather they should be understood as explanatory models which are often implicit and embedded in tacit cultural understanding. These assumptions are embedded in ritual instructions, commentary, or exegesis that accompany ritual manuals, or even expressed in narratives and poetry. Such texts, of course, do not express explicit theoretical statements, but careful analysis of narratological and rhetorical strategies reveals implicit understandings of ritual which should be seen as theoretical. My analysis focuses on practices and the meanings these practices have in Daoist ritual. This analysis, thus, contrasts interestingly with the Caroline Humphrey and James Laidlaw’s study of ritual which focused on the distinct quality of ritual actions. According to Humhprey and Laidlaw a ritual acts are distinct from mun1 Recent publications include Zhang Zehong 1999, Yamada Toshiaki 1999. For the most detailed and groundbreaking study of early Daoist ritual, see Lü Pengzhi 2008. For a brief survey, see Benn 2000. 2 Among the few attempts at such synthesis is Lagerwey 1987.

408

Gil Raz

dane actions due to “a particular modification of the normal intentionality of human action.” 3 These acts are performed by agents who treat these acts as “elemental”, “archetypal”, “prototypical”, which are stipulated by tradition, or other external authority, and not generated by their own intentions.4 Due to their nonintentionality ritual acts may have multiple meanings for those who perform them and are thus devoid of intrinsic meaning, rather they gain meaning only when intentionality is superimposed on them, what Humphrey and Laidlaw refer to as “meaning to mean.” 5 We will see that Daoist authors conceive of ritual acts similarly, indeed it is in the realm of ascribing meaning to ritual acts that we will find the Daoist theories of ritual. To begin with, consider the following narrative from the Scripture of the Perfected One, a text that is preserved as the final section of the Array of the Five Talismans.6 The redaction of Array of the Five Talismans is very complex and beyond the scope of this paper. Suffice it to say that the text preserved in the Daoist Canon was probably completed in the late third century. The Scripture of the Perfected One was originally an independent text and continued to circulate independently even after it was included in the Array of the Five Talismans. A version of the text is preserved as chapter 18 of Ge Hong’s 葛洪 (283–343) Inner Chapters of the Master Embracing the Unhewn 抱朴子內篇 (completed in 317). The narrative of the Scripture of the Perfected One begins with the Yellow Emperor having obtained the Scripture of the Perfected One of the Celestial Luminary. Unable to understand its teachings he set out on a journey through the realm to visit various sacred mountains seeking instruction. On these mountains, the Yellow Emperor encountered several sages who presented him with further scriptures and teachings. All these instructions, however, were lacking. The Yellow Emperor’s quest culminates on Mount E’mei where he encounters the Luminary Person 皇人, who after repeated entreaties finally agrees to instruct him. The Luminary Person begins his instruction by stating, “According to what you had said earlier what you do not understand of the words of the Perfected Scriptures is that the celestial pneumas are gathered in one’s body. If the body is clearly understood then one can obtain longevity.”7 汝向所道真經言所不解者, 蓋上天之氣歸此一身耳. 一身分明便可長生也. 3 4 5 6

Humphrey & Laidlaw 1994: 71. Ibid.: 150–151 et passim. Ibid.: 212. See also the summary in Laidlaw & Humphrey 2006: 274–281. Scripture of the Perfected One is shorthand to the Scripture of Great One and Perfected One of the Most High 太上太一真一經 that forms the final section of the Array of the Five Talismans of the Numinous Treasure of the Most High 太上靈寶五符序 DZ 388: 3.16a–23b. For details, see Raz 2004: 175–182. 7 DZ 388: 3.20a.

Ritual Theory in Medieval Daoism

409

The Luminary Person continues with a lecture on the macrocosmic correlations of the human body. Among the core ideas introduced in this passage is the notion of three Ones that dwell in the niwan (brain cavity), the Scarlet Palace (heart), and the Cinnabar Field (lower abdomen). He then provides the Yellow Thearch with precise instructions for imbibing the five sprouts 五牙, the qi of the five directions. By following these instructions the adept would integrate his bodily qi and attain the Perfect One (or Unity) 真一. Now, I suggest that in the context of Daoist scriptures this narrative, and especially the statement by the Luminary Person, represents theoretical discourse. This narrative contrasts with the narratives, practices, and understandings advocated in the main body of the Array of the Five Talismans. In the Array of the Five Talismans the five sprouts are described as the nascent vapors of the five cosmic directions as well as heavens in which dwell the five celestial thearchs. They are therefore external to the body. We should note, however, that in both texts the actual physical practice of ingesting the Five Sprouts is the same, involving controlled breathing, triggering the saliva glands through precise movements of the tongue, and swallowing the saliva while meditating. What the Luminary Person is voicing, therefore, is not an argument about orthopraxy regarding the practice of “ingesting the five sprouts”, but a position in a debate about the “meaning” of the practice that is contrary to that of the Array of the Five Talismans. Both positions are explanatory models for the same practice. Furthermore, the two texts associate the practice with distinct ritual programs. In the Scripture of the Perfected One the five sprouts are associated with the Three-Ones and the Perfect One. These terms do not appear in the ritual program advocated in the main body of the Array of the Five Talismans. Here, this practice is combined with the Five Lingbao talismans and the Five Celestial Thearchs of the imperial tradition to form the program of an “offering rite” jiao 醮, the earliest extant example of a Daoist ritual program which remains at the heart of Daoist ritual to the present. The inclusion of the Scripture of the Perfected One in the Array of the Five Talismans is in itself an example of Daoist ritual discourse, which by inclusion of rival schemata actually allows for their superscription or incorporation within a new synthesis. This strategy is augmented by the introductory narrative to the Array. This narrative, which recounts the appearance of the Five Talismans in the world, explicitly states that the Yellow Emperor did not receive them. The narrative thus undercuts the narrative of the Scripture of the Perfected One and, by extension, the ritual scheme presented therein. The placement of the rival ritual scheme in the same text is thus rhetorical strategy which we should understand as part of a debate on ritual efficacy. Another rhetorical strategy used to express ritual discourse in the Array of the Five Talismans that represents a simpler form of debate is found in the second

410

Gil Raz

chapter of the text which lists over sixty recipes dealing with various illnesses, but which are especially aimed at attaining longevity. The compiler or redactor of this chapter used simple editorial comments affixed to several of the recipes that place practices in relative hierarchies of efficacy and attainment. For example, a long passage extolling an herbal method for expelling the three worms or corpses that dwell in the body ends with the words: “The Perfected value this method, Daoist masters revere this medicine and the Worthies are joyful at its application.”8 Immediately following this celebratory pronouncement, however, we find these lines: “Those who are able to ingest pneuma do not employ these methods. The perfect pneumas are the five directional essences of the green sprouts 青芽. Daoist masters ingest them for twenty-one days and the three worms are expelled.”9 Unlike the debate about the meaning of the method of the Five Sprouts discussed above this note simply asserts the superiority of the Green Sprouts method, an alternative name of the Five Sprouts, over other methods without providing further instruction or explanation. The first point to note here is that in this case the method of the Five Sprouts is used on its own and its purpose is quite limited and clear, the expulsion of corporeal parasitic demons. There is no hint at the cosmological or ritual complexes which formed the respective matrices for the method of the Five Sprouts in the debate alluded to in the Array of the Five Talismans. This example should remind us that medieval Daoists constructed their ritual schemes from diverse traditional sources. In creating their ritual syntheses Daoists incorporated originally discrete methods for specific, practical goals, such as apotropaic, healing, divination, or exorcistic techniques. In adapting such methods into complex ritual systems with overarching cosmological frameworks Daoists provided these methods with new purposes and explanations. The same practice would thus have different meanings in the different contexts. A second point to note in here is the explicit, though incomplete, development of a hierarchy of practices and attainment. I would argue further that this hierarchy is reflected in the structure of the text itself. Following the example of the encyclopedic compilations which served as blueprints for the unified empire, the redactors of the Array of the Five Talismans signaled their totalistic knowledge by the inclusion of all methods.10 The very placement of the sixty-odd recipes in the second 8 DZ 388: 24b. 9 DZ 388: 24b. 10 Mark Lewis is particularly good on the encyclopedic compendia, see Lewis 1999: 302ff.; Robert Campany is especially cogent on the notion of “collecting” as a form of domestication, and “binding diverse data into a common and delimited field of display”, see Campany 1996: 9–17.


411

chapter of the Array of the Five Talismans marks them as inferior to the methods advocated in the first and third chapters. As one of the earliest examples of Daoist systemization, the Array of the Five Talismans is rather crude and sometimes inconsistent. It was, however, an inspiration to later Daoist authors and redactors. This inspiration is apparent not only in the fact that the ritual form advocated in the text remained among the basic ritual modules of Daoist rituals to the present, but, perhaps more importantly, in the expression of theoretical notions which informed later Daoist ritual formulations. Moreover, the different meanings of the method of the Five Sprouts advanced in the text, and the debate itself, reveal similar theoretical assumptions. These common assumptions may be summarized as: (1) The names or technical terms of practices are considered as distinct units of action. (2) Practices are transformative; that is they lead to actual psychophysical changes. (3) Particular practices are parts of larger ritual systems with cosmological implications. (4) The efficacy of practice is not inherent in the practice itself, but depends on the proper understanding of the ritual and cosmological scheme in which it is embedded. In the following pages I discuss how these assumptions continue to inform the construction of Daoist ritual. Following the lead of recent scholarship in ritual studies, rather than presenting a single, coherent theory of ritual, I discuss these different assumptions as theoretical approaches. 11 Nevertheless, I suggest that Daoist authors shared these assumptions and that we should perceive them as an implicit theory, which explain tensions between different ritual schemes, shape the strategies by which such tensions were resolved. And which finally led to the construction of complex Daoist ritual schemes, exemplified by Lu Xiujing’s 陸修靜 (406–477) synthesis in the fifth century, and which continued to develop through the following centuries.

2. The assumptions listed above may also be analyzed as themes that Daoist authors in medieval China recognized as problems that they needed to address. These thematic approaches are crucial for understanding the development of Daoist ritual. For the sake of brevity, I will simply introduce the thematic approaches or theoretical assumptions, demonstrate that they are indeed both etic analytical 11 Kreinath & Snoek & Stausberg 2006: xxi–xxv.

412

Gil Raz

categories as well as emic understandings, and then briefly show how these theoretical or thematic approaches illuminate particular instances in Daoist ritual. The approaches I examine below are: 2.1. Ritualization and Embodiment 2.2. Tension between individual / communal / universal attainment 2.3. Changes in Ritual purpose: adapting discrete practices into cosmological ritual systems 2.4. Modularity: the use of technical terminology as markers of discrete units of practice These themes, tensions, and assumptions are certainly not exclusive and often overlap. I distinguish these categories in these terms is for heuristic purposes, and ease of discussion. More importantly, these themes, tensions, and assumptions should be understood as functioning together in the construction, expansion, and systemization of ritual programs. Moreover, the construction of a systematized ritual system is understood not as creating anything new, but rather as the recovery of a primordial system. 2.1 Ritualization and Embodiment The first theme I consider consists of the related concepts of ritualization and embodiment. Due to the emphasis on the transformative effects of correct practice, I argue these analytically distinct concepts should be examined in tandem for understanding Daoist ritual. Ritualization has become a core concept in ritual studies, especially in works by Catherine Bell and by Caroline Humphrey and James Laidlaw. For Bell “ritualization” connotes the creation of ritual agents, “persons who have an instinctive knowledge of these schemes embedded in their bodies, in their sense of reality, and in the understanding of how to act in ways that both maintain and qualify the complex microrelations of power.”12 As mentioned above, Humphrey and Laidlaw view ritualization as the distinct quality of ritual actions, which are performed by agents lacking intentionality. These acts are ascribed meaning from a specific repertoire that is accepted by the performers. I argue that these two aspects of ritualization are found together in Daoist ritual, in which the ritual acts were supposed to cause a complete bodily transformation. In his explanatory treatise on Lingbao ritual, Lu Xiujing explains that the “retreats” zhai 齋 of Lingbao were revealed because the inner workings of the human body were in turmoil, “They move about recklessly in murder and robbery; thus it is appropriate to control them with ritual prostrations 禮拜. Their mouths are full of evil words, flattery, and deception; thus it is necessary to regulate them through 12 Bell 1992: 221.


413

reciting the scriptures. Their hearts are full of desire and malice; thus it is necessary to employ them in contemplating the spirits. By employing these three methods their hearts will be cleansed and their actions purified. The full attainment of cognitive and physical perfection, this is the meaning of zhai.” 用此三法洗心淨行, 心行精至, 齋之義也.13 Elaborating on the notion of ritual action as transformative that we found in the debate on the method of the Five Sprouts, Lu Xiujing now explicitly states that the purpose of ritual is to control, manage, and rectify the psychophysical entirety of the performer. The outward practice would indeed bring about complete inner transformation. In the case of Daoist ritual, than, I suggest ritualization and embodiment are equivalent. The notion that ritual is transformative was not a unique or a novel idea introduced by medieval Daoists. Rather, the notion of embodiment is at the core of both Confucian and imperial ritual systems. Among the most succinct early statements regarding ritual embodiment is Confucius’ brief autobiography in the Analects (2.4). In a few sentences Confucius traces his life-long learning and gradual mastery of ritual, culminating when at the age of seventy he “could act according to his heart’s desire without overstepping the line.” Implicit to Confucius’ claim in this passage is the assumption that ritual action is transformative. The transformation assumed is moral, emotional, and intellectual. By performing socially prescribed ritual patterns an individual internalizes and finally embodies the social mores to such an extent that he cannot conceive of an action that would contravene the social boundaries. Confucius thus claims that the purpose of ritual is, to borrow Catherine Bell’s words, to create “ritualized agents” that embody the socio-cultural matrix, and thus by their very presence legitimate that matrix. Yet, a further assumption implicit in this passage is that the ritual system is in fact a natural pattern, or “heaven’s command” in Confucius’ own terms. Thus the embodiment of the ritual system by individuals leads society to be aligned and harmonized with the cosmos itself. This implicit assumption becomes explicit in the writings of imperial ritualists during the Han. This new understanding of imperial ritual perceived the living emperor as an extra-human being, a manifestation of the ascendant phase. As I have shown elsewhere, this notion was adapted into the jiao ritual of the late third century Array of the Five Talismans, which combined the ingestion of the five sprouts with summoning the five celestial thearchs, who will then support the externalization of the adept’s bodily spirits as they travel to the celestial realms.14 The ultimate goal of these various practices is to attain a state in which one’s perfection 13 Dongxuan lingbao zhaishuo guangzhu jiefa dengzhu yuanyi 洞玄靈寶齋說光燭戒罰燈祝願儀 DZ 524.3a; I follow the translation and analysis of this passage by Bokenkamp 2001: 190. 14 Raz 2007: 83–109.

414

Gil Raz

matches the Dao 與道合真. Poul Andersen has shown that the notion of “bodily transformation” 變身 remains central to Daoist ritual.15 How Daoist notions of bodily transformation cohere with the notions of self-divinization examined by Michael Puett is a critical question in the history in Chinese religion,16 and is beyond the scope of this paper. 2.2 Individual / Communal / Universal Attainment A second theme that we can trace through the changes and developments of Daoist ritual is tension between individual and communal attainment. Among the main sources for the Daoist lineages that emerged in medieval China were the technical traditions, especially those that were concerned with bodily practices seeking health, longevity, and transcendence. The practitioners of these traditions sought individual attainment. Attainment was sought usually in reclusion, and indeed, a major motif in narratives about these practitioners is the necessity to leave their native home, change their name, and never return. 17 Nevertheless, many of these practitioners who were perceived as having attained transcendence became foci of local communal cults. The adherents of these communal cults were no longer necessarily focused on the quest for transcendence, but sought a variety of far more mundane felicities. As the territorial reach of many of these communal cults expanded beyond the traditional local boundaries, they formed new types of religious institutions. It is within the context of these communal cults that we should seek the social basis for the Daoist communities that emerged in medieval China. Many of these practitioners, and their practices, were incorporated into the narratives and ritual manuals of the Daoist lineages.18 In the process of adopting the practices of the technical traditions Daoist lineages tended to transform the goals of these practices. Rather than focusing on individual attainment, the practices became devices for communal participation, often as initiation or transmission rites. Ultimately, specific practices or devices came to be cosmologized, and the attainment they entailed was perceived as universal. Among the best examples for the shift in practice from individual attainment to communal participation is the appropriation of sexual practice in the ordination rite of early Celestial Master Daoism. The communal institution established by the Celestial Masters contrasts sharply with the tradition of individual attainment

15 Andersen 1995. 16 Puett 2002; 2004. 17 For cogent thoughts on the narrative construction of these practitioners, and whether indeed they were as reclusive as the narratives claim they were, see Campany 2009. 18 On the transition from local cults to Daoist lineages, see Raz (forthcoming).


415

within which developed the sexual practices. 19 Unlike those seeking individual attainment, all members of the Celestial Master community participated in the rituals through which they sought to attain transformation into the ranks of “seed people” 種民, the community of the elect who would repopulate the world after the coming cataclysm. This attainment was the stated goal of the initiation ritual described in the Initiation Rite of the Yellow Writ which culminated in sexual congress. While this ritual retains individual attainment as the rite culminates in the production within the practitioner of a homunculus named Peach Vigor, a perfected form of the self, the total rite is defined as initiation into full adult membership in the community. Moreover, the purpose of the rite is clearly stated in a chant that is repeated several times by the initiates: “Together, we will uphold the Way and virtue. We wish that you will release us from investigation of the three Offices; released from [celestial] net and [terrestrial] web; erase and remove us from the death records and inscribe our names for longevity on the jade registers. We will cross over and ford the nine calamities and be become the seed people of the next world.”20 A second example showing similar transition in the meaning of a practice is the use of the Lingbao talismans. The earliest references to the Lingbao talismans in Ge Hong’s Inner Chapters and describe them as one of many alternative protective devices used by masters entering the mountains. 21 While maintaining this function, in the Array of the Five Talismans these talismans are presented as unique manifestations of cosmic power, and they function as emblems of the five phases. The jiao rite, which is the core ritual of the Array of the Five Talismans, grants the performer far higher attainment then mere protection in the mountains. This rite was reformulated further in the Lingbao scriptures of the late fourth century into several different practices. In Lu Xiujing’s systemization of texts and rituals this rite continued to form the basic transmission rite of all the Lingbao scriptures.22 More significantly, however, beginning with the adaptation of the Buddhist notion of bodhisattva in the Lingbao scriptures the purpose of Daoist ritual was expanded to universal salvation. For example, in the Five Tablets of Perfect Writs in Red Script the attainment, while still individual, is extended to include protection of the family and state, and to the entire cosmos.23 We should note, however, that 19 For a translation and study of the manuscripts discovered in Mawangdui dating to the third century BCE, see Harper 1998. 20 Huangshu guodu yi 黃書過度儀 DZ 1294: 6b10; see also 3b7, 5b9, 6a7, 8a3. For a detailed study of this text, see Raz 2008. 21 Baopuzi neipian 抱朴子內篇 DZ 1185: 11.2a; Wang Ming 1983: 11.197; Ware 1981: 179. 22 Taishang dongxuan lingbao shoudu yi 太上洞玄靈寶授度儀 DZ 528. 23 Yuanshi wulao chishu yupian zhenwen tianshu jing 元始五老赤書玉篇真文天書經 DZ 22: 1.40a–42a: “You will constantly be equivalent to the spirit luminaries, protect the state and safeguard the family […] Prolong your years and attain longevity, your family will flourish

416

Gil Raz

while the universal salvation was predicated on the individual attainment of the Daoist master, bodhisattva-like vows were necessary for the individual attainment of the priest. Thus, the tension between individual, communal, and universal attainments was combined into one ritual program, and resolved by expanding the notion of embodying the Dao to encompass the cosmos. 2.3 Shifts in the Purpose of Practices and Rituals The preceding section already alluded to the fourth theme, shifts in the purpose of ritual acts and practices. These changes are similar to the developments regarding the shift from individual attainment to universal salvation. Daoists of the medieval period constructed their ritual systems by adopting and adapting particular elements of practice from several traditions. These practices tended to have limited and practical purposes, such as protective and apotropaic devices or exorcistic and divinatory techniques. When adapted into the Daoist ritual program these practices were cosmologized, that is they became symbolic of discrete cosmological factors within ever-expanding ritual programs. For example, the earliest references to the Lingbao talismans describe them as one of several apotropaic devices that were carried by alchemists and herbalists as they ventured into mountains. While maintaining this function, in the Array of the Five Talismans these talismans are presented as unique manifestations of cosmic power, and they function as emblems of the five phases. In the Five Tablets of Perfect Writs in Red Script, the talismans are said to be terrestrial replicates of the primordial Five Perfect Writs, which are described as cosmogonic factors in the emanation of the cosmos. This cosmogonic function helps explain the expanded purpose of the Lingbao ritual system as a whole. 2.4 Modularity As I noted in the discussion of the Five Sprouts, one of the primary assumptions in Daoist ritual schemes is that technical terms of practice are considered distinct units of action, and that these units are named. Humphrey and Laidlaw note that one of the primary distinctions between mundane acts and natural is that the latter are “named”.24 They go on to suggest that “ritual acts are treated as elemental, that

and the state will be blessed with peace […] Calamities will be extinguished of themselves; misfortune will be reversed and resolved spontaneously. In the four directions all will be harmonious, auspicious signs will appear daily. The kirin will arrive and the white tiger will sport in your courtyard. The state will flourish and the people rich, all the heavens will be at peace and tranquil […].” 常與神明相當, 保國寧家 […] 言年長生, 家致興隆, 國祚安寧 […] 災自滅, 凶逆自消, 四方和睦, 善瑞日生, 麒麟來歸, 白虎遊庭, 國豐民富, 普天安寧 […]. 24 Humphrey & Laidlaw 1994: 144.


417

is, that each stipulated type of ritual acts is thought as if it were a ‘separate’ thing with its own essential character.”25 This is not quite what we find in Daoist ritual. The innumerable methods and techniques for meditation, attaining longevity, pacifying the realm, healing, exorcism, and divination, which we find in Daoist ritual manuals, incorporate similar sets of technical terms. These terms, including names of stars and constellations, particular practices and movements in ritual, names of individuals and places in narratives, may appear together in almost any combination, leaving one to wonder whether they have any intrinsic meaning. I suggest that the notion of modularity is helpful in understanding how these terms were used and understood in Daoist ritual. I borrow the term modularity from Lothar Ledderose’ Ten Thousand Things in which he argues that “[...] the Chinese started working with module systems early in their history and developed them to a remarkably advanced level. They used modules in their language, literature, philosophy, and social organizations, as well as in their arts. Indeed, the devising of module systems seems to conform to a distinctly Chinese pattern of thought.”26 Referring to the system of Chinese script, Ledderose goes on to show that the modular system may be analyzed into a hierarchy of increasing levels of complexity: element, a single brush stroke; module, a building block or component; unit, a single character; series, a coherent text; and mass, all existing characters. By modularity I refer to the composition of complex ritual programs from discrete “symbolic modules” that reappear in different ritual contexts with different “meaning”. By “symbolic module” I refer to the cluster of meanings and practices associated with a specific term, such as Five Talismans or Five Sprouts, although these meanings and practices are almost never specifically spelled out. On their own, the specific terms may have no more meaning then their linguistic content, but they take on symbolic meaning in the context of specific narratives and rituals as they are placed in combination with each other. It is within these contexts that we need to be especially aware of shifts in meaning. The Five Talismans of Lingbao are a symbolic module, which as I have shown above, was given different meanings in different contexts. As a symbolic module, the meaning of the Five Talismans changes its focus in particular contexts when it is combined with other symbolic modules, each carrying a multivalent cluster of meaning. For example, in the jiao ritual in the Array of the Five Talismans the five Lingbao talismans are combined with the Five Sprouts and the Five Thearchs to form a transmission ritual which simultaneously was a reformulation of imperial notions of the extra-human body of the emperor, granting the possessor of the talismans with cosmic efficacy.27 A later text, the Five Ascendant Talismans of Lingbao, 25 Ibid.: 151. 26 Ledderose 2000: 2. 27 Raz 2007.

418

Gil Raz

one of the original Lingbao revelation texts, presents two distinct ritual schemes which include the five Lingbao talismans.28 The first ritual combines the five talismans with two other symbolic modules, the Eight Archivists and East Well, to produce a ritual scheme that replicates a specific moment in time, the establishment of the Han dynasty.29 As succinctly put in the Annals of the Grand Historian “at the rise of Han, the five planets were gathered in East Well.”30 The five talismans in this case represent the five visible planets, which were in conjunction in the constellation of East Well and thus signified an inauguration of new era. Later in the text, a new ritual system is introduced, the Twenty-four Charts.31 Each of these charts is a symbolic module, and they are here placed into a larger scheme. In the terms of Ledderose’s analysis this arrangement of modules into a coherent system is a “series”; I prefer labeling it “Ritual synthesis”. The titles of these charts are all found in earlier sources, particularly the Baopuzi. In these earlier contexts, each chart was an independent symbolic module with its own particular cluster of associated practices. Here they are arranged into a systematic order. It is unclear whether this is a hierarchy of initiation, or attainment, and, why these specific modules were selected to form this new ritual synthesis. Significantly, the initiation into this synthesis requires one first acquires the Lingbao talismans, “If you already have the Lingbao talismans, you must get the Lingbao charts. There are twenty four charts, which are heaven’s perfect pneumas, the name of my methods, and the true methods of the great Dao.”32 In the list of twenty-four charts, the Chart of East Well is first and the third is the Chart of the Eight Archivists. Between them is listed the well-known Chart of the Five Marchmounts. While the sequence of charts in the list clearly alludes to the symbolic modules that composed the ritual described above, they are listed here separately, and thus clearly carry distinct values and meaning. The system is reformulated in another text, the Scripture of the Twenty-Four Lingbao Charts, in which the charts are placed within the cosmological system of 28.Taishang wuji dadao ziran zhenyi wuchengfu shangjing 太上無極大道自然真一五稱符上經 DZ 671. A Dunhuang manuscript, P 2440 Lingbao zhenyi wuchengjing 靈寶真一五稱經, preserves an early version of the text and includes several important variants. For the text, see Ōfuchi Ninji 1978–1979: vol. 2,10–22. For a comparison of the two versions, see Raz 1996. 29 For details, see Raz 2005. 30 Shiji: 27.1348. The statement in the Hanshu is far more precise, “In the tenth month of the Inaugural year, the five planets were gathered at East Well” (Hanshu 1.22). This line places the conjunction at some time between November 14 to December 12, 207 BCE. The conjunction actually occurred in May 205. Homer Dubs considers the middle of May to be the correct date, see Dubs 1935: 311. Huang Yi-long opts for May 29 as the correct date, Huang 1990: 101. 31 Raz 2005: 64. 32 P 2440 l.421; DZ 671:2.11b6.


419

three registers, the Three Primordials (sanyuan 三元).33 I will not expand further on this ritual synthesis, but move on to a further development in which the entire scheme of twenty-four charts is absorbed and suborned within a larger scheme found in a late Six Dynasties text associated with the Shangqing ritual lineage, The Eight Effulgences of Grand Profundity. 34 After listing the twenty-four charts in order, including the verses introduced in the Scripture of the Twenty-Four Lingbao Charts, this text finally introduces the Transcendent Chart of the Seven Treasures of Jade Clarity Yuqing qibao shenxian tu 玉清七寶神僊圖 defined as a Unified Thrice- Eight Pneuma 總三八部真氣. This chart is clearly meant to supersede and suborn the twenty-four Lingbao charts. In this synthesis, the earlier modules, East Well, Eight Archivists and so on, have lost their autonomous meanings and have simply become building blocks within larger competing systems.

3. Systemization of Ritual Programs In this final section I examine Daoist ritual as consciously synthetic, by which I refer to the deliberate incorporation of various practices and discourses into a specific ritual synthesis, larger ritual complexes, and finally into totalistic systematizing ritual programs. The reasoning for such inclusion and systemization was the perception that the ritual system was perfect at the primordial moment, but the current state of performance and understanding is problematic. Thus, the authors of ritual texts explain their actions as recovering lost truths rather than constructing new forms. For example, the Jade Instructions in Red Script state: The Dao said: “the perfect writs, incantations, and explanations are the preeminent methods and measures. The ancient writs are abstruse and are not detailed. Later students found them difficult to use. Therefore, the eminent have sent down commentaries in order to explain them. I have caused the descent of the perfect essentials of the Jade Instructions in order to reveal and explicate the mystic and marvelous subtle phrases of the ancient writs.” (DZ 352: 1.4b4). The point here is that the rites to be elaborated in the Jade Instructions are no more that explanations of ancient forms. While we may take this as no more than a rhetoric flourish that resonates with Confucius’ own claim that he is a transmitter and not a creator, we should recognize that this strategy in fact assumes that a ritual acts are qualitatively different than regular actions and are not simply invented. In their analysis of ritualization, Humphrey and Laidlaw argue that ritual acts need to be considered as resembling “natural kinds”, a term they borrow from cognitive 33 Dongxuan lingbao ershisi sheng tujing 洞玄靈寶二十四生圖經 DZ 1407. 34 Taixuan bajing lu 太玄八景籙 DZ 258.

420

Gil Raz

linguistics. Natural kinds, such as gold or sand, are contrasted with artifacts, man made objects. When thinking of either kind people tend to think through prototypes. Prototypes of artifacts depend on convention, intention, use, and function. Prototypes of natural kinds do not refer to use or function, but are treated, cognitively and linguistically, as though they had an underlying essence that unifies all their manifestations, even though we may not be able to specify the precise characteristics of this essence. The idea that ritual acts are represented as natural kinds thus explains how a person can accept someone else’s definitions or theories about the real or underlying “essence” of his or her own acts. People treat different instances of the same acts as objects sharing the same essence, and the result of this is that the acts may be given different meanings and purposes.35 This is precisely what we saw in the discussion of the practice of imbibing the five sprouts with which I began. However, Humphrey and Laidlaw’s insight into the relationship between ritual acts and natural kinds may be further extended into our discussion of Daoist ritual theory. The Lingbao scriptures, for example, repeatedly define themselves as “celestial writs” 天書, and describe the ritual they advocate as “self-generated” 自然. For example, the Five Ascendant Talismans of Lingbao begins with a cosmogonic statement: “At the beginning of chaos, at the source of the subtle and wondrous, before the primordial separation, there were the self-generated perfect writs of Lingbao.”36 Later, when introducing the twenty-four charts, the text defines them as “the perfect pneumas of heaven [...] the perfect methods of the Great Dao.”37 Like the five talismans, the charts are defined as celestial in origin, and as simulacra of the primordial patterns of the cosmos. We find similar terminology in Lu Xiujing, who includes more rituals in his systematized ritual scheme. On the most superficial level, the claim for celestial origins is, of course, a legitimation strategy. This explanation, however, is too simple. The insistence on the “naturalness” of the Lingbao rites clearly identifies the Lingbao rites as natural patterns, and not as human constructs. On the one hand, this claim distinguishes the lingbao rites from other rites and practices, such as imperial ritual or those of local cults. Such a distinction implies a recognition that other ritual acts and programs belong to the same category of human behavior, albeit of corrupt form. Therefore, this claim implies a category of ritual, within which are improper forms, that are artificial, and pure forms, that are in fact natural forms. 35 Humphrey & Laidlaw 1994: 152–153. 36 DZ 671: 1.1a 37 DZ 671: 2.11b; P 2440, line 421.


421

We can push our analysis further. The Daoists’ insistence on the naturalness of their ritual may be contrasted with the Confucian philosopher Xunzi (312–230 BCE) who explicitly discusses ritual as “ornamental” 文 and “artificial” 偽: “Human nature is the root and raw material; artifice is the flourishing completion of ornament and patterns. Without human nature there would be nothing for artifice to work on; without artifice, human nature would be unable to beautify itself. Only when human nature and artifice combine can a true sage emerge, and only then can the unification of the realm be attained.”38 性者, 本始材朴也; 偽者, 文理隆盛也. 無性則偽之無所加, 無偽則性不能自美. 性偽合然後成聖人之名, 一天下之功於是就也. In his ritual theory, Xunzi dispenses with any supernatural efficacy to ritual action, claiming that ritual is no more than a method for social integration and harmonization.39 The Daoist claim is not a simple rejection of this theory, but a more subtle argument. Insisting that only the Lingbao rites are natural, the Daoists in fact assert that other rituals are not truly efficacious, and thus partially agree with a major trend in imperial or Confucian ritual theory.40 We find precisely the same attitude in the Daoist rejection of traditional medical practice, replacing moxa and acupuncture with talismanic practice and confession.41 As summary, I cite Lu Xiujing’s introduction to one of his ritual manuals, Tablets and Writs of the Numinous Treasure. Lu’s explanation for his compilation resonates with several of the approaches I discuss in this paper. First, Lu Xiujing claims that there is a primordial ritual system which has to be reassembled from the current state of confusion. Second, ritual must be systematized. Third, ritual acts have precise names and connotations, that is, they are symbolic modules: “There is a procedure to the transmission of the scrolls and oaths of Lingbao. There must be a sequence to receive the tablets and practice the ordinances. Both the Middle Oath and the Great Oath include rites of Placing the Dragon-tablets 投龍簡. Later, these are practiced on the Eight Nodal days and jiazi days. In addition, there is the rite of Placing the Jade Tablets of the Three Primordials 投三元玉簡. According to this method, one is to climb mountains on the four seasonal days and practice the rite of the perfect writs. The Contract with the Perfected and Jade Instructions both include details of these rites, 42 but later students were foolish and lazy, their will and 38 Xunzi, “On Ritual”; cf. the translation in Watson 1963: 102. 39 For an interesting analysis of Xunzi as a “Durkheimian” theorist, see Campany 1992. 40 For more on the Confucian view of ritual see the chapters by Michael Puett, Joachim Gentz, and Christian Meyer in this volume. 41 See Nickerson 1996: 352. 42 Contract with the Perfected refers to the Lingbao scripture Dongxuan lingbao changye zhi fu

422

Gil Raz character were shallow and lacking. When looking at a ritual code they did not concentrate to seek its meaning. They were lax in carrying out the rites, so that there were many failings and mistakes. Having confused heads and tails, forever they were not enlightened […] Observing this ruin, weakness, and harm to the teachings, I sincerely wrote the sequence of Placing the Tablets and writs in accord with the ancient codes, returning them to the proper order. I wrote comments and explanations to the methods in order to inspire the thoughts of the confused.”43 夫受靈寶劵盟既有梯首, 授簡修剌必由次第. 中盟, 大盟皆投龍簡. 其後八節, 甲子別投三元玉簡如此方. 得四時登山修真文之事. 明真, 玉訣並有其詳, 而晚學闇惰, 志性淺略. 遇見一科, 不加精尋, 率意施用, 遂致矯錯亡首失尾, 永不悟 […] 視其淪溺, 懼傷慈教, 謹依舊典撰投簡文次第復為甲乙, 註解法度, 以啟昧者之懷.

Lu Xiujing therefore concludes that ritual does have meaning, and the meaning resides precisely in the correct order of performance, which is the perfect primordial order. This is certainly is a theoretical claim. I suggest that in creating the meaning he sought in Daoist ritual, Lu Xiujing was aware of the tensions, problems, and issues in the development of Daoist ritual. I suggest that these emic problematic and strategies for resolution parallel the theoretical approaches I analyzed above. Like Lu Xiujing, I hope this would help clarify some confusion regarding Daoist ritual.

jiuyou yukui mingzhen ke 洞玄靈寶長夜之府九幽玉匱明真科 DZ 1411; Jade Instructions refers to the Lingbao scripture Taishang dongxuan lingbao chishu yujue miaojing 太上洞玄靈寶赤書玉訣妙經 DZ 352. 43 Taishang dongxuan lingbao zhongjian wen 太上洞玄靈寶眾簡文 DZ 410.1a.


423

References Andersen, Poul 1995. “The Transformation of the Body in Taoist Ritual.” In: Jane Marie Law (ed.). Religious Reflections on the Human Body. Bloomington: Indiana University Press: 186–208. Bell, Catherine 1992. Ritual Theory, Ritual Practice. New York, Oxford: Oxford University Press. Benn, Charles D. 2000.“Daoist Ordinations and Zhai Rituals in Medieval China.” In: Livia Kohn (ed.). Daoism Handbook. Leiden: Brill: 309–339. Bokenkamp, Stephen 2001. “Lu Xiujing, Buddhism, and the first Daoist Canon.” In: Scott Pearce & Audrey Spiro & Patricia Ebrey (eds.). Culture and Power in the Reconstitution of the Chinese Realm, 200-600. Cambridge: Harvard University Press: 181–199. Campany, Robert Ford 1992. “Xunzi and Durkheim as Theorists of Ritual Practice.” In: Frank Reynolds & Tracy David (eds.). Discourse and Practice. Albany: State University of New York Press: 197–231. — 1996. Strange Writing: Anomaly Accounts in Early Medieval China. Albany : State University of New York Press. — 2009. Making Transcendents: Ascetics and Social Memory in Early Medieval China. Honolulu: University of Hawai’i Press. Dubs, Homer 1935. “The Conjunction of May 205 B.C.” Journal of the American Oriental Society 55: 310–313. Harper, Donald 1998. Early Chinese Medical Literature: the Mawangdui Medical Manuscripts. London: Kegan Paul International. Huang, Yi-long 1990. “A Study on Five Planet Conjunctions in Chinese History.” Early China 15: 97–112. Humphrey, Caroline & James Laidlaw 1994. The Archetypal Actions of Ritual: A Theory of Ritual Illustrated by the Jain Rite of Worship. Oxford: Clarendon Press. Kreinath, Jens & Jan Snoek & Michael Stausberg 2006. “Ritual Studies, Ritual Theory, Theorizing Rituals – An Introductory Essay”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing Rituals. Vol. 1: Issues, Topics, Approaches, Concepts. Leiden, Boston: Brill: xiii–xxv. (Numen Book Series: Studies in the History of Religions 114/1). Lagerwey, John 1987. Taoist Ritual in Chinese Society and History. New York: Macmillan. Laidlaw, James & Caroline Humphrey 2006. “Action”. In: Jens Kreinath & Jan Snoek & Michael Stausberg (eds.). Theorizing rituals. Vol. 1: Issues, Topics, Approaches, Concepts. Leiden, Boston: Brill: 265–283. (Numen Book Series: Studies in the History of Religions 114/1). Ledderose, Lothar 2000. Ten Thousand Things: Module and Mass Production in Chinese Art. Princeton: Princeton University Press. Lewis, Mark Edward 1999. Writing and authority in early China. Albany: State University of New York Press.

424

Gil Raz

Lü Pengzhi 呂鵬志 2008. Tangqian daojiao yishi shigang 唐前道教儀式史綱. Beijing: Zhonghua shuju. Nickerson, Peter 1996. “Abridged Codes of Master Lu for the Daoist Community.” In: Donald S. Lopez (ed.). Religions of China in Practice. Princeton: Princeton University Press: 347–359. Ōfuchi Ninji 大淵忍爾 1978–1979. Tonkō Dōkyō 敦煌道經. 2 vols. Tōkyō: Fukutake Shoten. Puett, Michael J. 2002. To Become a God: Cosmology, Sacrifice, and Self-Divinization in Early China. Cambridge: Harvard University Asia Center. (Harvard-Yenching Institute, Monograph Series 57). — 2004. “The Ascension of the Spirit: Toward a Cultural History of Self-Divinization Movements in Early China.” In: John Lagerwey (ed.). Religion and Chinese Society. Vol. 1: Ancient and Medieval China. Hong Kong: Chinese University Press: 193–222. Raz, Gil 1996. Ritual and Cosmology: Transformations of the Ritual for the Eight Archivists. (M.A. Thesis, Indiana University). — 2004. Creation of Tradition: The Five Talismans of the Numinous Treasure and the Formation of Early Daoism. (Ph.D. Dissertation, Indiana University). Bloomington: Indiana University. — 2005. “Time Manipulation in Early Daoist Ritual: The East Well Chart and the Eight Archivists.” Asia Major 18: 27–65. — 2007. “Imperial Efficacy.” In Florian C. Reiter (ed.). Purposes, Means and Convictions in Daoism: A Berlin Symposium. Wiesbaden: Harrassowitz: 83–109. — 2008. “The Way of the Yellow and the Red: Re-Examining the Sexual Initiation Rite of Celestial Master Daoism.” Nan Nü 10: 86–120. — (forthcoming). Creation of Tradition, The Emergence of Daoism in Early Medieval China. London: Routledge. Wang Ming 王明 1983. Baopuzi neipian jiaoshi 抱朴子內篇校釋. Beijing: Zhonghua shuju. Ware, James 1981 [1966]. Alchemy, Medicine and Religion in the China of AD 230: The Nei P’ien of Ko Hung. New York: Dover Publications. Watson, Burton 1963. Hsün Tzu: Basic Writings. New York: Columbia University Press. Yamada Toshiaki 山田利明 1999. Rikuch dōkyō girei no kenkyū 六朝道教儀禮の研究. Tokyo: Tōhō-shoten. Zhang Zehong 張澤洪 1999. Daojiao zhaijiao keyi yanjiu 道教齋醮科儀研究 . Chengdu: Bashu shuju.

Julius Tsai 蔡南亭

Mutation or Permutation? A Ritual Debate in Tang-Song Daoism1 1. Introduction This essay examines a controversy over ritual fixity and flux. Its focal point, the practice known as the Nocturnal Invocation 宿啟, is a basic element within the Daoist liturgical repertoire. Historically, it was generally performed on the night prior to the commencement of a Retreat 齋 (one of the primary frameworks for Daoist communal ritual) in order to establish the altar and invoke the gods. Such was the importance of the Nocturnal Invocation that the sixth’s century imperial compendium Wushang biyao 無上祕要 (Essential Secrets of the Most High, HY 1130)2 characterized it as the first of three fundamental rituals making up the classical Retreat.3 While scholarly attention to the Nocturnal Invocation has tended to focus on the laying out of the spatio-temporal array of the Five Perfected Writs 五方真文 (a frequently-seen but by no means universal feature of the Nocturnal Invocation), recent work has begun to recognize the complexity of the Nocturnal Invocation’s historical formation out of disparate ritual components.4 I want to build on that more complex recognition of ritual formation by tracing a centurieslong debate concerning the Nocturnal Invocation, one that reveals fascinating traces of “indigenous” Daoist theorizing on the nature, function and interpretation of ritual.

1 Presented at the Collaborative Research Centre SFB 619 “Ritual Dynamics and the Science of Ritual” International Conference at the University of Heidelberg, Germany, in 2008. A Japanese version of this paper is being prepared as part of the proceedings of the 2008 Japanese-American Conference on Daoism at Toyo University, Tokyo, where portions of this material were presented. I would like to thank Professors Poul Andersen, David Mozina, Michael Puett, Gil Raz, Asano Haruji 浅野春二, Maruyama Hiroshi 丸山宏 and Nikaido Yoshihiro 二階堂善弘 for their insightful comments and suggestions. 2 Texts are numbered by the Harvard-Yenching (HY) classification system, for which see Weng 1988. 3 Lagerwey 1981: 31. 4 Davis 2001: 308.

426


2. The Rectification of Errors on the Altar of Mystery The flashpoint for this particular controversy is a set of twenty theses on Daoist ritual found in the Rectification of Errors on the Altar of Mystery 玄壇刊誤論 (DZ 1280) of Zhang Ruohai 張若海, dated 943 CE. Zhang presents new views on the “methods of ritual practice of the Three Grottoes” 三洞科教 (referring to basic divisions of the Daoist canon and its ordination ranks) within a set of dialogues with the mysterious Master Cloudlight 雲光先生, a pan-religious savant who is steeped in practical and theoretical knowledge concerning the methods of the Daoist Retreat, and who may be considered Zhang Ruohai’s literary alter-ego. Master Cloudlight’s critiques of contemporaneous Daoist practice form the basis of ritual modifications that often directly contravene then-traditional practices. Even more interesting for modern observers is the way that he rationalizes those modifications. The Rectification of Errors has been little studied in modern scholarship. However, its contemporaneous influence was surprisingly long-lived. For one thing, the Rectification of Errors was collected up into the Song Daoist Canon, and exerted an influence on subsequent rescensions of the ancient rites (DZ 1280, 20b). For another, Zhang’s proposed ritual reforms must have become pervasive enough to strike as raw a nerve as it did among the leading ritualists of the age. Indeed, ritual authorities from Du Guangting 杜光庭 (850–933) in his Liturgical Manual for the Yellow Register Retreat 太上黃籙齋儀 (DZ 507, 53.3a–4a); Wang Qizhen 王契真 (twelfth century) in the Great Rites of the Shangqing Lingbao 上清靈寶大法 (DZ 1221, 56.1a–3b); Lü Taigu 呂太古 in the Comprehensive and Requisite Manuals of Daoism 道門通教必用集 (DZ 1226, 6.3a–3b), dated to 1201; to Jiang Shuyu 蔣叔輿 (1162–1223) in the Standardized Rituals of the Supreme Yellow Register Retreat 無上黃籙大齋立成儀 (DZ 508, 16.3b–5a; 12a–13a; 18a–24b) represent nearly three centuries of attempts to refute teachings such as those represented by Zhang Ruohai.

3. Contexts and Concerns: Deconstructing the Nocturnal Invocation Tracing one particular strand of the controversy, the proper placement of the practices of the “Explanation of the Precepts” 說戒 and the “Installation of the Officiants” 署職/補職 within an overall ritual sequence, reveals an underlying reason for why the controversy over the Nocturnal Invocation has seemed so complex: the voices in this debate often talk “past” each other while appearing only superficially address the same issues. This reflects not only the shifting contents of the Nocturnal Invocation over time, but also the fact that each voice in the discussion often has its own more localized audience, opponents, and contexts of concern.

Mutation or Permutation?

427

For example, Jiang Shuyu, in notes to his rescension of Lu Xiujing’s 陸修靜 (406–477) rites for the Nocturnal Invocation, criticizes Master Cloudlight for placing the Explanation of Precepts and Installation of Officiants before and not after the Nocturnal Invocation, a move that thereby inverts the orthodox order of performance (compare DZ 1280, 3b–4b with DZ 508, 16.12a–13a, 18a–21b). However, while representing a position of orthodoxy and orthopraxy, it becomes clear that Jiang himself must strike a balance between criticizing the “fabrications” of his opponents and making allowances, which he does, for the constructed, accreted nature of the Nocturnal Invocation. Indeed, it soon becomes clear that the term “Nocturnal Invocation” is not a straightforward one, as it has both a general as well as specific usage. Differentiating between those usages does much to explain seeming discrepancies between our sources. As a general term, “Nocturnal Invocation” refers to the entire sequence of rites performed at the zi 子 hour (11pm to 1 am) on the night prior to the commencement of the main ritual. As a specific term, however, the “Nocturnal Invocation,” understood narrowly, refers to a series of practices that follow upon the Spontaneous Retreat 自然朝, a more ancient ritual core that ritualists from Du Guangting onwards consistently call the “ancestor of the methods of the Retreat” 齋法之祖 and trace back to the earliest days of Daoism. According to a passage found in both Wang Qizhen and Jiang Shuyu’s compendia, the Spontaneous Retreat is made up of the Activation of the Incense Burner 發爐; the Proclamation of Divine Names and Ranks 稱名位; the Bowing to the Ten Directions 禮十方; and the Confessing of the Three Karmic Areas 懺三業, the Triple Proclamation 三啟, and Triple Homage 三禮. For its part, the Nocturnal Invocation that follows is typically made up of the Laying Out of the Perfected Writs 敷露真文; the Reading of the Invocation 讀詞; the Explanation of the Precepts 說戒; the Installation of the Officiants 補職; and the Proclamation of Restrictions 宣禁 (DZ 508, 16.12a; DZ 1221, 56.1a). Still, Jiang’s criticism does not exactly match up with what Zhang is actually propounding. In fact, the relevant section of the Rectification of Errors is not even about the Nocturnal Invocation per se. Instead, it introduces an innovation in practice, that of Entering the Quiet Chamber 入靖 (DZ 1280, 1b–3a), a three-day period of purification prior to the commencement of the main ritual, a practice that echoes early Celestial Masters and Shangqing methods of audience with the Perfected, but also taking its cue from Confucian and imperial purification practices. On the one hand, these three days do include the performance of the Spontaneous Retreat, followed by the Explanation of Precepts and Installation of Officiants (DZ 1280, 3a–3b); this would seem to be in line with Jiang’s own position, but then render his criticism inexplicable. On the other hand, Master Cloudlight’s innovation of having a three-day rite of Entering the Quiet Chamber would indeed have

428


the Teaching of the Precepts and Installation of the Officiants taking place before the time that a traditional Nocturnal Invocation would have been performed. But it cannot have only been a question of sequence. Another way to think through this controversy would be to ask: What was bothering Jiang? What did he see as the sine qua non of the ritual that others were ostensibly tampering with? Zhang Ruohai seemed to have been championing a kind of pragmatic ritual approach, one based on an imperial model. Jiang Shuyu, on the other hand, seems to have seen the establishment of the altar through the emplacement of the Five True Writs as the most essential function of the Nocturnal Invocation, something that Zhang’s account does not emphasize.5 Could this be the reason for Jiang’s opposition to Zhang’s reformulated rites? Let us speculate a bit further: The establishment of the altar by means of the Nocturnal Invocation presupposes possession of the Five Perfected Writs of Lingbao 靈寶五符 and, perhaps we can infer, a sequentially obtained sequence of Celestial Masters registers, gained through formal rites of initiation. Could Zhang’s notable omission of the laying out of the True Writs reflect a desire to perform Daoist ritual outside of traditional frameworks of ordination? If so, this would point to an ongoing negotiation between established ritual, scriptural, and clerical frameworks and the ever-present reality of religious adaptation and innovation outside of those frameworks. To add to this emerging portrait of ritual pluralism, merely reviewing our present sampling of liturgical manuals yields still other ritual innovations and adaptations to the Nocturnal Invocation: Du Guangting criticizes some for neglecting the Spontaneous Retreat altogether (DZ 507, 53.3b); Zhang Ruohai bemoans the methods of those who perform the Explanation of Precepts and Installation of Officials after the entire audience rite 一朝法事 (DZ 1280, 4b); and Jiang Shuyu criticizes a certain Li Jingqi 李景祈 for performing an additional Spontaneous Retreat during the afternoon prior to the commencement of the rites (the rationale being to relieve a congested ritual schedule) (DZ 508, 16.13a–13b). While Zhang’s text provides the fullest example of a supposedly heterodox practice, the four examples just mentioned are all departures from a “standard” Yellow Register Retreat. This begs the question of which practice is indeed standard and which is non-standard, which is typical and which is atypical. Such cases suggest that at the level of actual practice, local adaptation rather than general systematization might in fact have been the rule rather than the exception.

5 I thank Professor Asano for his comments in this regard.


429

4. Daoist Ritual Theorizing If we may speak at the level of Zhang and Jiang’s ritual theorizing, that is, the way that they talk on a meta-level about Daoist ritual, we may make the following summary observations. Zhang’s Rectification of Errors are grounded in at least five general assumptions. First, he appeals to the charismatic, hagiographic persona of Master Cloudlight (DZ 1280, 1a–1b). Second, he displays a willingness to apply pragmatic reasoning in breaking with tradition (“If something is not working, why should one keep repeating the old ways and do things in the same way?” 事不經，何必襲舊 (DZ 1280, 3b). Third, he speaks of the fascinating principle of “eliminating complexity and inclining toward simplicity,” 去繁就簡 (DZ 1280, 6b–7b) a kind of principle of elegance governing ritual innovation. Fourth, he makes the repeated and direct analogy of Daoist ritual to imperial rites (“rites of paying court with the Perfected are equivalent to national incense offerings, sacrifices, and rites” 夫朝真之科與國家煙祀禮同, see DZ 1280, 4a), which provides a model by which he thinks through his desired ritual reforms. Fifth, Zhang makes few references to past ritual masters or lineages, or even to the divine origins of ritual. In fact, Master Cloudlight says explicitly (in a tone that calls to mind the early Confucian thinker Xunzi 荀子): “Moreover, the rites do not fall from heaven, nor do they emerge from the earth. They are simply fitted to human sentiments” 且禮非從天降，非從地出，酌人情而已 (DZ 1280, 7b). By contrast, Jiang grounds his authority in the transmissional chain that links him to his own three masters, as well as to a revered trio of past patriarchs: Lu Xiujing, Zhang Wanfu 張萬福 (fl. 700–742), and Du Guangting. Not only that, he asserts the divine origin of the rites in the revealing activity of the Perfected One of Grand Extremity 太極真人, harking back to the ancient revelations of Lingbao. For Jiang, adherence to the orthodox ways is not merely a matter of being historically scrupulous. Rather, orthopraxy is key to ritual efficacy in enabling the bureaucratic transmission of petitions from the human world up to the empyrean courts. Broadly speaking, this debate centers around three poles that are in fact not exclusive to this period, or even to Daoism, but might recur whenever human beings consider, in a self-conscious fashion, the basis of their ritual action: 1) Should one follow the charisma of the master or charisma of the lineage (teachers, texts and practices?); 2) Does ritual power come from pragmatic adaptation or traditionbound adherence? and 3) Is ritual a human construction or a divine artifact?

5. Ritual Modularity This debate also raises an issue of broader importance for the study of ritual – how we might benefit from thinking of it as “modular” in nature, for which I draw from

430


Ledderose’s comments on modularity as a mode of cultural production in traditional China.6 We might think of that modularity – characterized by such traits as the nesting, interchangeability, and recombination of ritual elements – by considering the following structural levels to Daoist ritual, all of which are brought into play in the case of the Nocturnal Invocation: 1. Ritual unit: the smallest units of ritual activity typically identified discretely, such as the Activation of the Incense Burner, Proclamation of Divine Names and Ranks, Bowing to the Ten Directions, etc.; 2. Ritual sequence: the stringing together of individual “ritual units” into a larger functional sequence, for example, those that might be concerned with establishing an audience with the divinities on the one hand (for example, this appears to be the function of the Spontaneous Retreat), or those that have more of an invocatory function on the other (the Nocturnal Invocation in its narrow sense); 3. Ritual framework: a broader liturgical framework such as the Nocturnal Invocation in its general sense, Walking the Way 行道, or Proclaiming Merit 言功, all of which were constituted by the combining together of multiple “ritual sequences”; 4. Liturgical genre: the highest-level organizational principle or structure for ritual performance comprised of a piecing together of “ritual frameworks,” for example, the Retreat or Offering 醮 liturgies. As I hope my tentative breakdown of some of the structural levels of Daoist ritual illustrates, all of the ritualists we have discussed seem to mention a “Spontaneous Retreat” (itself a retrospective construction) although not always a “Nocturnal Invocation”; and they all seem to revolve around certain sometimes overlapping understandings of what “ritual units” are at play when performing some kind of critical preparatory practice before a main ritual. For our purposes, the careful tracing out of ritual units and sequences guards against any reification or essentializing of the Nocturnal Invocation. What the various writers are calling “Nocturnal Invocation” actually fluctuates between the above levels 2 (ritual sequence) and 3 (ritual framework); its contents are variable as well as modular in arrangement. While that kind of fluidity may have been of concern for those who sought ritual fixity, this protean modularity underscores the adaptive power of Daoist ritual.

6. Conclusion Let me end by observing, first of all, that attempts at systematization are often as much attempts to silence divergent voices and eliminate the religious options that 6 Ledderose 2000. I have also profited from discussions with Gil Raz on this point.


431

those voices represent. Secondly, though the Daoist ritual tradition may appear to exhibit stability and standardization, it would be better to think of that seeming stability as that of a living organism constantly maintaining its equilibrium. From our own present-day vantage point, it is sometimes easy to forget that the ritual programs of the great ritual systematizers of the Tang and Song may not have represented the mainstream in their own day, and that their grand liturgical edifices might in fact have been urgent, countervailing efforts against the ever-present threat of divergent contemporaneous practices. We thus need to read against the grain of past and present reification to recover the undoubtedly lively discourse of the times. Master Cloudlight’s teachings reveal a path ultimately not taken within the Daoist tradition. However, in exploring the balance between ritual continuity and change, the debate over the Rectification of Errors testifies to a powerful, “homeostatic” dynamic characterizing the Daoist ritual tradition.7

7 On the notion of ritual homeostasis see Bell 1997: 201.

432


References Bell, Catherine 1997. Ritual: Perspectives and Dimensions. New York, Oxford: Oxford University Press. Davis, Edward 2001. Society and the Supernatural in Song China. Honolulu: University of Hawaii. Lagerwey, John 1981. Wu-shang pi-yao: somme taoïste du Vie siècle. Paris: Ecole Française d'Extrême-Orient. Ledderose, Lothar 2000. Ten Thousand Things: Module and Mass Production in Chinese Art. Princeton: Princeton University Press. Weng Dujian 1988. Daozang Zimu Yinde. Taipei: Xin wenfeng.

Lucia Dolce

The Contested Space of Buddhist Public Rituals The shunie of Tōdaiji Shunie 修二会, literally “liturgical assemblies of the second month”, were among the most important public rituals (hōe 法会) performed in Buddhist temples in premodern Japan. These liturgies took their name from the period in which they were originally held, the second month of the lunar calendar, and in their simplest meaning may be described as exorcistic rituals performed at the beginning of the year, which centred on repentance (keka 悔過). Repentance was understood as an action that benefitted not only the individual who carried it out, but also the community at large. Celebrant-monks, by repenting committed offences, amend their errors and, thanks to this, bring public benefits, such as peace for the entire country and wealth and well-being for the populace. Significantly, Buddhist repentance in Japan was not conceived as a mental process of individual awareness, but was enacted as a set of ritual actions, including bodily movements, melodic chanting and recitations. One may indeed speak of a ritualisation of repentance. While many of the pre-modern assemblies have disappeared from the liturgical calendar of temples and shrines, shunie are today still performed annually in temples of the Nara area. The foremost examples are those held at Yakushiji 薬師寺, popularly known as “the flower ceremony” (hanaeshiki 花会式), and those at Tōdaiji 東大寺, celebrated as the liturgy of “drawing the water” (omizutori お水取り). Whether they have been reinstated recently, as in the case of Yakushiji, or have been performed continuously for centuries, as at Tōdaiji, they are presented as ancient rituals, endorsed by mythological narratives that legitimise the temple in which they are performed, and maintaining the format and function they had in the past. In this sense, shunie constitute an extraordinary focus to explore the significance of ritual continuity in a Buddhist institutional context, and the actual changes that a representative ritual underwent historically. At the same time, their scale and complexity provide grounds to analyse the multiple agency inherent in ritual practice, and the layered religious, liturgical, and social implications of performance qua ritualised production. The interest in public assemblies has grown in recent years among Japanese historians, and a good deal of research has been published in Japanese. Surprisingly,

434

Lucia Dolce

however, these liturgies have hitherto not been the object of much attention in Western scholarship. This study focuses on the shunie conducted at Tōdaiji. In order to understand how the ritual came to be formulated in the shape we have today, I shall first sketch the origin of the second-month assembly in the context of the historical evolution of non-esoteric Buddhist ceremonies. Then I shall outline the protocol of the liturgy held at Tōdaiji, and explore the performative aspects of some ritual segments. While I refer to the discursive narratives on mythological and symbolic themes that the ritual has produced, my main concern is the enactment of the ritual in specific movements, sounds, and settings, and the distinct groups of actors and sponsors for whom the ritual yields meaning and empowerment. These elements, in fact, reveal how the ritual has been constructed as a complex but regulated space, where different dynamics are at work at the same time.

The Development of Public Repentance Liturgies Scholars have suggested that the establishment of large-scale Buddhist liturgical assemblies was affected by the interaction of three socio-religious factors: the type of prayers addressed to a deity to fulfil specific wishes; the use of distinctive media to request divine intervention, such as reading and chanting; and the status of the petitioners. Textual evidence shows that the performance of repentance rituals goes back to the second half of the seventh century (the term keka is already attested in the Nihon shoki). Thus repentance may well have been the basic form that Buddhist rituals took when they were introduced and systematised in Japan under the influence of practices witnessed in Tang China. However, the content, structure, and function of these early ceremonies greatly differed from later shunie. As a matter of fact, contemporaneous sources do not emphasise the centrality of the act of repenting in front of the buddhas. The rituals carried out in the large Nara temples sponsored by the state consisted in decorating the statues of the main deity (a buddha, a bodhisattva, or the Four Heavenly Kings) and inviting a large number of monks to read sutras. The aim of the assembly was propitiatory and apotropaic: praying for rain or to stop rain, preventing illness (of the king), and avoiding disasters and epidemics that affected the country at large. Repentance was not perceived as the direct cause of the benefit requested, but as an auxiliary element, the step that preceded the actual ritual. Reciting sutras (which exploited the magical power of the scriptures) or the names of the buddhas (which relied on the power of veneration, and of the deity venerated) was considered to be equally efficacious in delivering benefits. Yet historical records also point out a number of elements that would be developed in medieval shunie. One closely related to the act of repenting offences is the purity of body and mind of the practitioners. Interestingly, such status was extended to the people who attended the ritual and were its beneficiaries.

The Contested Space of Buddhist Public Rituals

435

Not only monks, but also lay people and patrons undertook an ascetic regime, abstaining from drink and eating meat, in order to attain a pure heart. The success of the ritual was thus understood as the outcome of the joint effort of monks and laymen. A second crucial element is the recitation of sutras and spells at night, in a loud voice and with an unusual rhythm.1 The formal structure of the repentance assemblies became more defined from the mid-eighth century, when the ceremonies started being classified according to the deity to whom they were dedicated (Yakushi keka, Amida keka, Kichijō keka). Invoking these deities by chanting their names (shōmyō keka 称名悔過) became a core part of the ritual, and a special style of Buddhist melodic chanting (shōmyō 声明), which had reached Japan around that time, was introduced into the liturgy. This move indicates the extent to which public ceremonies were susceptible to new religious fashions and apt to absorb them. At the same time, ascetic practices of purification became central to the liturgy and the means to (rather than preparation for) the fulfilment of the prayers. Sutra reading and repentance thus became a set, the two terms of a combined liturgy, which would become the model for non-esoteric Buddhist ritual practices. These two elements, however, remained differentiated in terms of their spatio-temporal performance. Sutras were read, usually in tendoku 転読, during the daytime periods. Repentance, in the form of invocative repentance (raisan 礼懴), and recitations of spells, took place during the night sections. After the ninth century such distinctions would be formalised in a two-period pattern (niji 二時) adopted by many hōe. Even assemblies, such as the Tōdaiji shunie, with a more complex six-period structure shaped by the traditional six divisions of the day (rokuji 六時), would maintain a clear differentiation between day and night segments.2 This feature was further reflected in the institution of two main celebrants in charge of different ritual protocols: the master of spells (shushi 呪師) and the master of recitations (daidōshi 大道師). At the end of the tenth century, ludic elements were added to the ceremonies, aimed at the entertainment of a lay audience. Shunie were established in this context, together with another type of public ritual held at New Year, called shushōe 修正会.3 These new types of assemblies were held at night, and included artistic performances between two nightly segments, which concluded with the appearance of demons on the temple stage. This evolution was part of a religio-political move that used the rituals as didactic occasions to teach Buddhism to the lay people who 1 Satō 2002b: 6–10. 2 On the two patterns see Satō 1994: 2–10. Sanbōe no kotoba, for instance, attests that during the misaie, assemblies that lasted for seven days and nights, lectures on the Saishōōkyō took place in the daytime, while repentances devoted to Kichijōten were staged at night. 3 These two terms appear for the first time in a document related to the Tendai abbot Ryōgen, dated 970. It is here recorded that the chase of demons and the “rite of the ox-seal” were performed at the end of a twenty-seven day liturgy. Satō 2002b.

436

Lucia Dolce

attended the performances. It was influenced by political changes, as temples were set free from the protection and control that had regulated their activities under the Ritsuryō system and gave new directions to their practices.4 Today’s shunie have maintained and exploited their ludic aspect, and indeed are mostly known for it. The hanaeshiki of Yakushiji still closes with a chasing of demons (onioishiki 鬼追い式).5 This finale caters to a “ludic catharsis” of the audience, which is at once entertainment and soteriological enactment. In the eleventh century shunie became part of the liturgical calendar (nenjū gyōji 年中行事) of Buddhist institutions, and the period of their performance was fixed. They were also exported to major provincial shrines, playing a crucial role in the acceptance of Buddhism in the countryside. They in fact replicated the aims of the original state assemblies (good harvest, prevention of disease and peace for the entire country), thus functioning to integrate local and “national” concerns.

The “Omizutori” of Tōdaiji: Ritual Protocol The second-month assembly of Tōdaiji is performed at Nigatsudō 二月堂, the hall in the upper precinct of the temple complex, which takes its name from the ceremony itself. The liturgy is dedicated to Eleven-headed Kannon 十一面観音, the honzon 本尊 of Nigatsudō, and is performed for an extended period of time, nowadays from 28 February to 15 March.6 Both the temple and the city of Nara 4 Uejima 2004: 234–236. 5 Fieldwork, April 2005. Today, the Yakushiji assembly is carried out for seven days, between the end of March and the beginning of April. On the last day of the liturgy (4/4), after the monks have concluded the rites of repentance inside the Main Hall (which can be attended only by a very limited number of people), the scene moves onto a platform built outside the hall, in front of a larger public. Performers dressed as demons appear on the stage, supplied with torches, and are chased away by the spells of the Heavenly King Bishamon, who walks around the stage in an unusual manner. An excellent recording of the ritual is the video produced by the National Museum of Japanese History (2007). I am grateful to Matsuo Kōichi for providing me with a copy. 6 A complete record of the ritual may be found in Tōkyō kokuritsu bunkazai kenkyūjo 1975. A recent paperback volume offers a comprehensive guide to the ceremonies (Satō 2009). There are several studies of the omizutori, including Horiike 1985; Gangōji bunkazai kenkyūjo 1979; Nantō bukkyō kenkyūkai 1984; and the extensive works of Satō Michiko (Satō 1994 and 2002). The medieval illustrated origin story Nigatsudō engi provides important visual representations of the pre-modern ritual. Other, more recent media products include two shōmyō recordings (Victor Entertainment 1971 & Kingu recodo 1997), and a documentary video (Shogakukan 2003). Transcriptions and studies of the chants are also available (Tōdaiji 2004). No scholarly analysis of the liturgy exists in English (Boyer & Fujiyoshi 1970 is a short introductory description), but in French there is an excellent study based on ethnographic fieldwork conducted in the nineteen-seventies (Berthier 1981). The analysis of the present paper has benefited from the fieldwork I conducted at Nigatsudō in February–March


437

advertise it as a magnificent ritual of water and fire that marks the start of spring. The spectacular nightly interludes, when huge burning torches (taimatsu 松明) are carried up to the front gallery of Nigatsudō, illuminating the clear winter skies, are pre-eminent in the public perception of the ritual (Image 1). From this perspective, the climax of the ritual is reached two nights before the conclusion of the assembly, when eleven torches weighing more than sixty kilos each are whirled around on the hall balcony, producing a rain of blessed fire over the audience. Today this spectacle attracts thousands of spectators from all over Japan, yet few of the visitors are aware of the actions that take place in the hall before and after such extravaganzas.

Image 1: Taimatsu burning in the night Photo by the author

The omizutori is in fact a very complex ritual of repentance. It is divided into two distinct sections, a nine-day preparatory retreat, and the public liturgy proper. 2009. I would like to thank Kojima Yasuko, a specialist in medieval Japanese culture, for sharing her knowledge of the ritual, and making time to accompany me, providing practical information, introductions, and occasionally a blanket for the cold nights.

438

Lucia Dolce

The first stage is an indispensable period of purification, known as bekka 別火 (lit. “different fire”), in which participating clerics observe a regime of vegetarianism and eat food prepared exclusively (i.e. on a different fire) for them. The retreat serves for the practitioners, who are called rengyōshū 練行衆, lit. “those who cultivate ascesis”, to prepare themselves and the ritual goods to be used during the liturgy. They make the paper flowers that decorate the altar, the straw mats that they will use in the inner hall, and their ritual clothing and footwear, consisting of a white papier-mâché dress and wooden clogs covered by paper.7 They practise chanting and conch-shell blowing in their own lodgings first, and in a building located within the Kaidanin 戒壇院 compounds during the second part of the retreat, when the ascetic regime becomes stricter, and they are forbidden to speak to each other or go out of the lodging. The public liturgies begin on the evening of 28 February. On this day the clerics move to a building located at the foot of Nigatsudō (sanrō shukusho 参籠宿所), which becomes their secluded lodging for the duration of the assembly. The first segments of the ritual are acts of physical and spiritual purification: the practitioners take a ritual bath (Image 2), receive a blessing in the form of a Shinto exorcism (Ōnakatomi harae 大中臣祓) performed by the master of spells (Image 7), and are administered the precepts (jūkai 授戒) in the dining hall next to their lodging. They then ascend to Nigatsudō. The inner sanctuary is unlocked and the main lay assistant (dōdōji 堂童子) makes a new fire (ittokubi 一徳火) to light the “eternal lamp” offered to Kannon, which will be kept burning until the following year (Image 12). Every day the practising clerics ascend to Nigatsudō twice, for the daytime services (Image 3) and the longer evening services. (For the evening ascent their attendants carry the huge pine torches.) The liturgy revolves around two basic ritual actions. One is the recitation of the holy name of Kannon accompanied by prostrations and other acts of veneration. The second is the recitation of sutras and other texts accompanied by fast movements inside the inner hall (walking, running). At sunset the Amida sutra is also recited in a separate room of the hall. During the liturgical period the practitioners take only one ritual meal a day, at noon, and are forbidden to eat and drink until the end of the daily observances in the middle of the night, when they can have some rice gruel. Maintaining such an ascetic regime empowers them and legitimises their practice. The ritual is supported by auxiliary practices, which are not connected with repentance and thus are not part of the daily protocol, but are staged only once during the assembly. The most important of these is perhaps the ceremony of “drawing sacred water” from the well at the bottom of Nigatsudō, which gives the second-

7 For examples see the catalogue of the special exhibition held during the shunie in Nara (Nara National Museum 2006: 59).


Image 2: Practitioners exiting from the ritual bath Photo by the author

Image 3: Ascent to Nigatsudō Photo by the author

439

440

Lucia Dolce

month assembly its popular name. It is performed during the night of 12 March, and takes the form of a kami ritual, with all its symbolic apparatus (Image 4). The scented water (kōsui 香水) drawn from the sacred well is poured into two wooden buckets, and mixed with water drawn in past rituals. It is used as an offering to Kannon, and is also given to the public that attends the liturgy. Another noteworthy auxiliary practice is the so-called “outing of the small Kannon” (ko Kannon shutsugyo 小観音出御), in which a small image of Kannon, placed in a Shinto palanquin (mikoshi 神輿), is brought into the prayer hall and is venerated in turn by each of the practitioners (Image 5). This is the only occasion on which high-ranking clerics of Tōdaiji officially attend the liturgy. Remarkable is also the presence of the head priest (gūji 宮司) of the nearby Tamukeyama Hachiman shrine 手向山八幡宮, who acts as a warden of the sacred image by remaining seated next to the altar throughout the rite, thus symbolically evoking the protective role of Hachiman.8 The Shinto flavour of such auxiliary practices serves as a reminder of the significance of associative (shinbutsu 神仏) practices, even within essentially Buddhist rituals such as that of repentance.

Performance: Sounds, Movements, Silence The protocol of the shunie thus unfolds through segments that are diversified in terms of spatial setting and type of ritual action. Dramatic effects are produced by privileging darkness over light, and the use of symbolic material such as fire and water. Such effects are amplified by the interaction of what we may consider the basic elements of performance, sound and movement. The two major liturgists mark two different performative modes, which have coexisted and fed into each other, but at the same time seem to be constructed contrastively to give rhythm to the liturgy. The master exorcist’s words and gestures are mystifying and incomprehensible. He leads what we can define as the esoteric part of the ritual. He makes mudras and mantras and reads dharani. His function is to call upon the guardian deities of the place of practice and to build the sacred boundaries (kekkai 結界) that protect the ritual space. He rings bells, makes utterances in a loud voice, and moves around “noisily”, stamping on the floor with his wooden clogs and running. The other main liturgist (daidōshi), who is in charge of the exoteric side of the ritual, leads the recitative segments while sitting or walking naturally. He pronounces the inaugural statement of purpose (kaibyaku 開白) and intones the melodic chanting that takes place at the beginning and end of the night, using his voice in a more conventional way. The performance of the two leading images thus alternates between slow and fast action, focusing on symbolic move-

8 Satō 1984.


Image 4: Procession to the sacred well Photo by the author

Image 5: Venerating the small Kannon Photo by the author

441

442

Lucia Dolce

ments of the body and of the voice with one, while focusing with the other on a gifted transmission of textual words to the audience. Perhaps because the most significant liturgical sequences take place in the dark, sound (or absence of sound) plays a remarkable role in the perception and appreciation of the ritual. Of the different sounds that emerge from inside the inner hall, the voices engaged in uttering, singing, and reciting “texts” (sutras or other “reading matter”, such as lists of names) express each liturgical element in a different mode, allowing the audience outside the hall to divide the phases of the ritual into exoteric and esoteric, and the deities venerated into buddhas and other gods.9 In some sections, words and vocal sounds are emitted in a forceful way that is not heard in temples, except on the occasion of special ascetic practices. In others, the litanies of the name of Kannon are pronounced with an accelerating rhythm, progressively abbreviating the formula namu Kanjisa bosatsu (Praise to the Bodhisattva Kannon) to namu Kanjiza, namu Kan, and then only Kan, repeated at an almost delirious speed. With such vocal segments other actions are juxtaposed, which enhance the opposite effect, that of no sound. This is the case of a circumambulatory practice where the practitioners takes off their clogs, lift up their ritual robes, and start running around the altar without producing any sound (hashiri no gyōhō 走りの行法). Legend says that it reproduces the way in which the initiator of the ritual, Jitchū 実忠, saw the repentance conducted by saintly beings in the Tuṣita heaven.10 This silent performance is interrupted only by a clean, sharp stroke on a wooden plank. This is the sound of a distinctive type of prostration called gotai tōchi 五体投地, which takes place in the outer hall. The practitioner throws himself down with all the weight of his body onto a long plank placed in the middle of the outer hall (gotai no ita 五体の板), while at the same time touching five parts of his body simultaneously (Image 6). This prostration is executed also during the daily segments of the liturgy. It is a difficult exercise that has to be acted out to perfection to avoid a broken knee. It proves that performative repentance, in so far as it consists in corporeal ascetic acts, implies readiness and virtuosity. Thus the practitioners’ movements become forms of bodily enunciation that do not necessarily refer to a scriptural text, but equally define the ritual space as sacred by virtue of their empowered gestures.

9 See Abe 2005, in particular the diagrams on p. 133. On the function of the voice in constructing the sacred see Abe 2001. 10 The origin story recounts that to re-enact the divine ritual, Jitchū had to accelerate the circumambulations, since one day and one night in the Tuṣita heaven corresponded to four hundred years in the world of mortals. This silent run takes place on the last two nights of the first seven-day period (5–6 March) and the last three nights of the second seven-day period (12–14 March).


443

Image 6: Five-touch repentance (gotai tōchi) Photo by the author

Attuning the kami: Music, Stillness, and Declamation In striking contrast with the melodic chanting that is carried out inside the inner hall, ritual segments that reiterate kami worship or are patterned after a Shinto mode (and which take place outside the inner hall) are marked out by music that is associated with shrines and court rituals, gagaku 雅楽. This music is played along the steps of Nigatsudō during the procession that goes to draw sacred water, as well as in the outer hall during the rites of veneration of the small Kannon. It clearly distinguishes these liturgical sections as devotional practices with a different set of symbolic dynamics. The liturgical segments associated with the kami are also constructed through an intersection of sound/no sound and movement. Ōnakatomi harae, the incantation that opens the shunie, is mimed by the master of spells. He kneels, the short kesa loose on his left shoulder, probably to denote that this cleansing exorcism does not appeal to the power of the buddhas, but that of the kami. Holding a small gohei 御幣, he unrolls the scroll that contains the text of the formula with almost

444

Lucia Dolce

imperceptible movements of his hands. His lips move fast but no sound can be heard (Image 7). He remains still in this position until the end of the formula, when he suddenly jumps up, extending his right leg with an abrupt movement that concludes the rite. The Nakatomi formula evokes the institutional aspects of kami worship. It was originally an important prayer (norito 祝詞) pronounced by state ritualists, although in fact its recitation was introduced in the shunie in the medieval period, when the incantation underwent compelling esoteric interpretations.11 Its exorcising power as the opening step of the second-month assembly resides both in this historical pedigree and in the peculiarity of its performance, the stillness of its recitation capturing the mystery and invisibility of the kami, and the brusque final movement awakening to their blessing.

Image 7: The master of exorcism recites the Nakatomi formula. Photo by the author

11 On the use of the Nakatomi harae during the shunie and other Buddhist exoteric rituals, see Daitō 2008.


445

Inviting guardian deities to attend Buddhist rituals was a common feature of early Buddhist rituals. During the shunie, both at Yakushiji and Tōdaiji, practitioners receive the blessings of the kami before starting the liturgy, and recite sutras and other invocations in front of guardian shrines. In the Tōdaiji assembly, however, the presence of the kami seems articulated in a more complex way, and staged in a variety of ritual actions. Another recitative segment, which is part of the daily liturgy, is the declamation of the Register of Gods (jinmyōchō 神名帳). A long list of names of deities of Japan is read every day, at the beginning of the nightly rites, by a soloist. The reading takes place in the inner hall by the light of an oil lamp, announced by the blowing of conch shells. It is done with a rough voice, slowly at first, then very quickly, then quietly again, depending on the class of gods whose name is pronounced. A rhythm is thus introduced into what would otherwise be the plain reading of a text with no content. Only clerics who have trained for three years and who have been given oral transmission can act as declaimer. This points at the secretive mode in which the matters of the gods are dealt with, a feature discernible in other segments of the liturgy as well. The reading of the Register of Gods is linked with the very origin of the secondmonth assemblies. As is well known, an early list of official shrines across Japan, whose cult was sponsored by the court, was given in Books Nine and Ten of the Engishiki. Scholars see the effort to compile a “national register” that included all deities of the “five provinces and seven highways” in the context of the expansion of Buddhism outside the capital, and in close connection with the staging of Buddhist liturgical assemblies in countryside shrines.12 In the medieval period a distinction was made between the sets of kami that represented the entire country, listed in a kokunai jinmyōchō 国内神名帳, and by then numbering 951,500, and the set of kami relevant to a specific shrine-temple complex, and which would be summoned on the occasion of its ritual practices (kanjō jinmyōchō 勧請神名帳). The ritual device of reading their names served to call upon the guardian gods of Japan to rejoice in the ritual, and remind them to protect the temple and its practices. The list produced for Tōdaiji counted around five hundred deities, divided between bodhisattvas (daibosatsu 大菩薩), great eminent kami (daimyōjin 大明神), eminent kami (myōjin 明神), heavenly deities (tenjin 天神), and restless spirits (goryō 御霊).13 This is the roster still declaimed today. In medieval literary sources the reading of the jinmyōchō was considered the starting point of the repentance rituals, and thus this liturgical segment was emphasised in the performance, as well as in the mythological narratives that legitimised the ritual. The origin story of the second-month assembly says that when Jitchū 12 Uejima 2004: 247–249. On the national register of kami see Mitsuhashi 1999. This development contributed both to instilling the concept of “the country of Japan” in the local audience, and to making it participate in the soteriological dimension of the assembly. 13 This is, by and large, the list given in Kaidanin kōyō kanjō jinmyōchō.

446

Lucia Dolce

performed the ritual for the first time, he read a Register of the Gods, summoning 13,700 gods to attend his ceremony. Onyū myōjin 遠敷明神, the god of Wakasa 若狭 province (today’s Fukui prefecture), was delayed in arriving because he was then fishing in the Onyū river. To compensate for his late arrival, he expressed his wish to offer scented water to the deity to whom the observances were devoted. At that point two cormorants, one black and the other white, took off from a rock near the hall and water began to spring from that very spot (Image 8).14 Such mythological construal of the intervention of a god from another area of Japan functioned to validate the liturgical repentance with the endorsement of the deities of the entire country, and positioned the specific ritual space of Kannon at Tōdaiji within a country-wide context. At the same time, it served to emphasise that both buddhas and kami are crucial protagonists of the liturgies.

Image 8: The sacred well (akai) of Nigatsudō with a clay cormorant on the roof Photo by the author

14 This legendary origin of the jinmyōchō ritual and of the sacred spring is narrated in the oldest origin story (engi 縁起) of Tōdaiji, Tōdaiji yōroku, dated 1160. Late Muromachi-period illustrated engi supply vivid representations of the legend. See Nara National Museum 2006.


447

Enacting the Performance: Actors and Patrons Whether one looks at the second-month assemblies as public rituals or as acts of ascetic practice, they appear to be centred on a specific group of clerics with distinct qualifications. Yet the participation of the laity is indispensable to their enactment. I would like to briefly consider these two groups of people whose interaction makes the performance of the shunie possible. a. Ritualists. Participation in the Tōdaiji second-month assembly was a selective matter from the beginning. Historically the religious affiliation of the practitioners was determined by the changing religio-political situation of Nigatsudō. Until the early Kamakura period, the celebrants were Kegon monks physically living in the upper section of the Tōdaiji precincts (Nigatsudō and the close-by Hokkedō 法華堂). This attests to the distinct links that Nigatsudō established with the Kegon school, documented in other historical material, and enhanced by the myth of the foundation of the shunie by Jitchū. (Jitchū allegedly was a disciple of Rōben 良弁, the Kegon monk who established Tōdaiji.)15 From the mid-Kamakura period, however, the second-month practices were joined by monks affiliated to other parts of Tōdaiji, in particular Sanron scholar-monks and monks of the Nakamondō 中門堂 .16 In more recent times participation has been enlarged to clerics belonging to different Tōdaiji sub-temples, and today even to affiliated temples in the provinces, as the list of this year’s participants indicates. Originally two different groups of ritualists performed in each of the seven-day periods into which the liturgy is formally divided. One sees traces of the earlier pattern in a procedure still upheld in today’s protocol, the administration of the precepts to the practitioners on the first day of the liturgy and again after seven days. The original pattern implies that there was a larger number of monks interested in undergoing the training and performing the repentance than there are today, when only eleven clerics participate. In fact diaries and other historical sources document that the number of participants in the second-month assemblies was more than twenty, until around the end of the sixteenth century, with a maximum of twenty-six for a single year. The number diminished in the seventeenth century, but it increased again at the beginning of the eighteenth century. Clearly amendment had to be made to the formal aspects of the performance to deal with this evolution effectively.17 However, even when a shortage of monks hampered the protocol, the ritual remained closed to monks who were not properly trained and had not received secret transmission of the procedures.

15 On Jitchū’s (foreign) origins, as well as his dates, various conjectures exist. He is understood to have been active in the mid-eighth century. 16 Nagamura 2001: 83–84. 17 Satō 2001: 45.

448

Lucia Dolce

Changes have also occurred in the division of ritual tasks. Until the Meiji period scholar-monks (gakutō 学頭) and ordinary monks (dōshū 堂衆) were distinguished by occupying different sections of the ritual space in the inner hall. Today this distinction is no longer made. In addition, while the four leading monks (shishiki 四職)18 originally performed the gotai prostrations along with the other practitioners, annotations in early-modern ritual manuals show that, when their physical condition was not sufficiently good, they were allowed to skip this segment. From the mid-Edo period they no longer performed it.19 This clearly affected the meaning of the liturgy as a collective repentance. Hierarchies are still maintained in today’s protocol, but they appear to be determined by ritual expertise rather than status, as demonstrated by the liturgical role played by the four leading celebrants on the one hand, and by the cleric who has most recently entered the practice (shosekai 処世界) on the other. Although they do not perform the repentance, three types of semi-ordained, semi-lay assistant ritualists (sanyaku 三役) have to be mentioned as indispensable actors of the ritual. The main one of these, the dōdōji 堂童子, still played by a male member of the Inagaki family, conducts some of the rituals that take place outside the inner hall and oversees them all, interacting continuously with the master of ceremonies (Image 9). b. Supporting the ritual. Most Buddhist assemblies in the Nara and early Heian periods were sponsored by the state and the aristocracy, who donated money and lands for their performance. Common people, too, established “karmic relations” with the temples by giving goods or offering their services on the occasion of these rituals. In the case of the Tōdaiji shunie, however, there is no evidence that aristocrats from Kyoto started attending the second-month assemblies before the midKamakura period. Scholars have pointed out that the practitioners themselves were the major sponsors of the shunie, and for this they gathered and managed finances. At the same time, in the medieval period, the assembly became the opportunity to foster links between the temple and the laity, a relation that would become stronger and expand in later periods, enclosing wider strata of the populace.20 Such performative aspects of a ritual are usually documented by historical materials such as temple records, certificates of donations, and rosters of people attending the liturgies. However, the protocol of the Tōdaiji shunie inscribes the significance of patronage and the changing relations with the laity in the liturgical script itself. Two ritual actions, in particular, attest to such concern. 18 In addition to the master of spells and master of recitations, already mentioned, there is a “head priest” (wajō 和上), who usually is the most senior cleric of the group and administers the precepts, and a “master of ceremonies” (dōtsukasa 堂司), in charge of directing the practical sides of the ritual. 19 Satō 2001: 49–50. 20 See Nagamura 2001, in particular 63–66 and 86.


449

Image 9: Master of ceremonies (dōtsukasa) and attendant (dōdōji) Photo by the author

The first is the declamation of a Register of Patrons called kakochō 過去帳. This is not the usual kakochō kept in Buddhist temples, which consists of a record of the dead cared for by each temple, and is used for their commemorative rites. Rather, it is a list of monastics and lay people who contributed to the existence of the Tōdaiji and supported the performance of the second-month liturgies in a variety of ways. It starts with emperor Shōmu, who established Tōdaiji, and unfolds through the names of famous monks and political leaders who made considerable donations to the temple, and of unknown men and women who offered smaller gifts, such as the accessories to be used during the liturgical assembly or robes for the monks. By recounting their names and achievements, the practitioners acknowledge their patronage and karmic links (Images11a–b).21 The roster of names 21 The oldest kakochō extant today is dated from the end of the sixteenth to the early seventeenth century. See Nara National Museum 2006: 46–47 and Gangōji bunkazai kenkyūjo 1979: vol. 1 (shiryōhen). Fig. 11 shows a manuscript copy of the kakochō in possession of Kamitsukasa Kaiun (1906–1975), Tōdaiji priest and later bettō, and includes the intonation marks (fushi hakase 節博士). Published in appendix to Iijima 1942.

450

Lucia Dolce

Image 11: The Register of Patrons (kakochō). Cover page (a) and section with the name “woman in green”, with notations for the declamation (b) Source: Iijima 1942

is extensive, filling a thirty-four-metre scroll. It includes 2,412 people for a span of time that goes from the mythical beginning of the liturgy to the early modern period, but today the reading is abbreviated in places. The kakochō is recited twice during the assembly, on the fifth (Jitchū’s memorial day) and on the twelfth day. It is read at the beginning of the nightly sections, after the Register of the Gods, by a cantor sitting in the inner hall. The style of declamation is similar to that of the Register of the Gods, and progresses with significant accelerations of the voice. The clerics in charge of this performance are selected from those who have taken part in the assembly five times. Thus the kakochō may be considered an extraordinary record of the devotional history of the second-month assembly and of Tōdaiji. At the same time, it serves as a device to pray for the attainment of enlightenment by all those who have established a karmic connection with the temple or with Kannon – generations of rulers, monks, and lay people. The ritual reading of the roster had its mythological underpinning in the story of the “woman in green (blue) robes” (shōe no nyonin 青衣女人). This legend first appeared in late medieval illustrated narratives on the shunie.22 It relates that, during one performance in the Jōgen years (1207–1211), a woman dressed in green clothes appeared in front of the monks, who were declaiming the kakochō, and asked why her name had been dropped from the register. Since then the name “woman in green” has been included in the roster, and it is still read with a tone and rhythm that marks the significance of this episode in the economy 22 Tōdaiji engi ekotoba (fourteenth century) and Nigatsudō engi emaki (sixteenth century). See also Nara National Museum 2006: 39 and 41.


451

of the ritual (Image 11b). Nigatsudō was constructed as the sacred space for buddhas and kami, to which only men following ascetic practices (an activity notoriously perceived in misogynistic terms in the Buddhist world) had access and could share with the deities during the ritual, but which was off limits for women. The “woman in green”, by appearing not with her actual body but as a spirit, pushed herself into the world of the sacred, demonstrating that the benefit of the ritual extended to everybody. The second ritual element that links the practitioners to the audience is the production of talismans. Such action takes place inside the inner hall on two nights of the ritual, the eighth and the ninth. Silently, between the two recitative segments performed by the master of recitations, each practitioner rubs two woodblocks to make two distinctive prints. One is a protective talisman that goes by the name of “seal of the ox-jewel” (goō hōin 牛玉宝印). It is composed of a rectangular print of three sets of logographs, the name Nigatsudō in the middle and, at its sides, two formulas praising Kannon (lit. “her buddha-head”, i.e. the highest of Kannon’s eleven heads) for averting evil and diseases. A double seal depicting two jewels in red cinnabar with Kannon’s seed-letters inscribed inside is further stamped in the middle of the print (Image 13a). The second talisman is a print of the circular sonshō darani 尊勝陀羅尼 in Sanskrit script (Image 13b). This is one of the most widespread Buddhist dhāranī in East Asia, employed to invoke good health and long life. Blessings that use an ox-seal, on the other hand, are attested in Japanese esoteric rituals since the late Heian period.23 The practitioners bestow these talismans in different ways. They are handed out to Tōdaiji clerics who do not take part in the ritual, are granted to the members of the confraternities that attend to the practical aspects of the ritual, and are available for purchase at the temple. After the assembly is concluded, on 18 March, the talismans are also sent to the imperial family – the only ones it receives officially today. The talismans are regarded as especially empowered, because of the special status of the clerics who have printed them, and the context in which they are produced. Such efficacy sanctions the benefits carried by the liturgical practice beyond the wall of the ritual hall and the restricted circle of practitioners.

23 Matsuo 2006. The term goō is probably related to the ox’s yellow (goō 牛黄 in the Japanese pronunciation), an important substance in Chinese alchemy, which was used as a magical medicinal remedy, and is often mentioned in Japanese medieval Buddhism. There is evidence that the ox-seal used during the shunie was understood as a sort of wish-fulfilling jewel or relic, and that the exorcism, which the practitioners perform during the ritual with a willow cane on which the talisman is inserted (goō tsue 牛玉杖), was seen as a ritual of empowerment of the body of the king, and, by extension, of the entire country (ibid.). Both at Tōdaiji and at Yakushiji at the end of the assembly the seal is impressed upon the practitioners’ foreheads.

452

Lucia Dolce

Image 13: Amulets: (a) ox-seal (goō); (b) sonshō darani From the author’s private collection

Today the shunie are mostly financed from the revenues of the Great Buddha Hall, visited every day by thousands of tourists who never climb up to Nigatsudō. During the liturgy the temple employs more than fifty people in various capacities to attend to the practitioners. Yet the practise of the ritual would not be possible without the labour of several lay confraternities (kōsha 講社) from different provinces of the country. Among the oldest, which scholars date back to the early Edo period, are those in charge of the wood from which the torches for the dattan rite are made (Iga ichinoi taimatsu kō 松伊賀一の井松明講, from Miura prefecture), and those in charge of the oil for the lamps that are burned in the hall during the ceremonies (Hyakunin kō 百人講, from Kyoto). Stone inscriptions near Nigatsudō attest to the existence of a Thousand People Association in 1745 and Southern and Northern Associations in 1810. The confraternities were systematised after the Meiji period.24 Today only thirty-three are left. Some are in charge of the preparatory steps for the shunie, such as selecting and taking to Tōdaiji the long bamboo poles on which the torches are installed (Yamashiro taimatsu kō 山城松明講). Others attend to the ritual steps conducted outside the hall, such as the warden of the omizutori, who, dressed as shrine attendants, escort the procession that goes to draw water from the sacred well (Kawachi Eikyūsha 河内永久社). Yet another group staffs a propitiatory rite for children, which takes place on the morning after the liturgy is concluded (Asamairi kō 朝参り講).25 The confraternities at times offer the bamboo poles for the large torches. These are also donated by individuals petitioning for the well-being of their families and harmonious relations in their homes, as the inscriptions on the poles show (Image 10).

24 Shunie to kō: 20–21. 25 Shunie to kō: 2–19.


Image 10: Bamboos with inscriptions from donors Photo by the author

453

454

Lucia Dolce

Epilogue: the Multivocal Performativity of the shunie The second-month assembly of Tōdaiji is celebrated as a “ritual practice that will never decline” (futai no gyōhō 不退の行法), a liturgy that has been performed when other ceremonies were discontinued because of lack of finances, wars, or the burning of temple buildings. At the beginning of the sixteenth century, after a fire started inside Nigatsudō, the community of monks at Tōdaiji feared that the ritual would disappear. In reaction, the compilation of ritual manuals, oral transmissions, and annotated interpretations of the protocol increased.26 A historical overview of the practice shows that such continuity was assured by the changes that the performance of the shunie underwent, in relation to movements in the broader context of Japanese religion. The medieval period, in particular, was crucial for the introduction of kami-related segments such as the Nakatomi formula, and the creation of origin stories, through which mythological interpretations of the ritual were made possible. Other additions and adaptations took place before and after, and to a certain extent still do. What, then, defines the shunie as a public ritual today, and for whom does it generate meaning? As this paper has sought to demonstrate, the second-month assembly at Tōdaiji is a complex ceremony that combines different performative elements, each playing a role in the dynamics by which the ritual is constructed, signified, and enjoyed by different agents. Yet two basic motifs seem to surface which frame the liturgy and give it its distinctive character: a ludic theme, which stems from the very origin of the rite, has turned it into a popular one, and is undergoing a continuous amplification and commercialisation; and an ascetic theme, which may be regarded as the inward and experiential dimension of the ritual, but which has affected its public presentation with notions of bodily purity, construction of sacred space, and physical expression of repentance. These motifs may be understood as two levels in the practice and perception of the second-month assembly. They coexist and are equally necessary to the ritual and its performativity. At the same time, each seems to stage a different setting for the implementation of the ritual, and to imply a different audience. On the one hand, the omizutori is acclaimed for its spectacular effects and exorcistic virtues. The fire extravaganza takes place outside Nigatsudō, and can be enjoyed by everyone. The practitioners play hardly any role in it. The size of the bamboo poles, the fire effects, and the ability of the attendants who carry the torches determine its results. On the other hand, the omizutori is also perceived as an intense and dramatic liturgy shaped by ascetic exercises and melodic chanting. The ritual segments connected to repentance are performed by trained specialists in a space almost completely inaccessible to the public. Such is the dining hall at the bottom of Nigatsudō, where the precepts are imparted and the ritual meal consumed, or the inner hall of Nigatsudō 26 Satō 2001: 44–45.


455

Image 12: The eternal lamp for Kannon Photo by the author

constructed around the altar to Kannon.27 The latter is separated from the outer hall by doors and by a white curtain that for most of the performance reveals only the shadows of the clerics circumambulating the altar and the light of the eternal lamp burned for Kannon (Image 12). These parts of the liturgy seem to be meant exclusively for the clerics. There is, however, a curious relation between inside and outside, seeing and non-seeing, which opens up the possibility of partaking in the ritual with one’s senses. As the assembly progresses, the side and back doors of the inner hall are open and people who stand or sit in the corridors around the inner hall are given visual access to the practice, albeit partially and through one or two lattice doors. Other ritual spaces also offer the opportunity to break into the liturgical secrecy. The interior of the dining hall is completely closed off, but the bestowal of the precepts can be “heard” as a whisper through a small window. Moreover, during the ritual meal, the doors behind a small shrine placed in one of the corners of the building are open, and the public can silently peep through, a few people at a time, to see what happens inside.28 Altogether it appears that the more 27 The gotai repentance, however, which may be considered a demonstration of physical ability, is always performed in the outer hall, where it can be seen through the lattice doors. 28 The ritual space in its totality, that is, the precincts of Nigatsudō, is also marked by multiple and consecutive ropes that delineate sacred enclosures (kekkai 結界). On the day on which the ritual starts there are ropes as far as the Lotus Hall and the stone lantern in front of the gate to Hachimangū, and in front of the place where the Nakatomi harae is recited. Surprisingly, these limitations are progressively eliminated.

456

Lucia Dolce

the practice enters its climax, the more it is disclosed. Some degrees of inaccessibility are maintained throughout, though. For instance, in the last days of the ritual men are allowed into the corridor surrounding the inner hall, while women never are. They may only “watch” from the outer corridors, where raised platforms are built for them to sit on. Several women I have spoken to, however, consider such seats, which are called tsubone 局 (lit. “ladies’ rooms”), as privileged places that allow a more intimate perspective on the ritual. Thus the construction of the ritual space is affected by the nature of the performance as an ascetic practice and manifests the gender discrimination determined by the proximity to the sacred. The multi-levelled connotation of the second-month assembly is reflected in the complex picture of the people to whom this is a meaningful practice. Next to the practitioners, who undergo an ascetic regime and are conscious of the soteriological role that their actions are supposed to play, there is an audience, for whom the omizutori is performed. Such an audience may be broadly divided into two groups. The first consists of occasional attendees, who attend the fire spectacle and seek its benefits by standing below Nigatsudō to receive the sparks of fire, and by collecting burned wood from the torches to take home as charms. They visit the hall briefly to pay homage to the deity and are usually not concerned with what the practitioners perform. The second group consists of those who attend the liturgy to see and to listen to the monks practising. They arrive from various parts of Japan, are very knowledgeable about each ritual step, stand outside the dining hall to listen to the recitation of the precepts, and then spend the night sitting in silence in front or in the back of the inner hall. They have as different a reason to attend the liturgy as one may find in other religious practices in Japan – acquiring merits for one’s or others’ benefit, devotion to Kannon, interest in the musical features of the ritual. Yet all seem to share a sense that the experience is uplifting and has a positive impact on their lives. For this they attend every year, if not the entire ritual then at least parts of it, making clear that it is necessary to renew such an experience. In this sense they are not simple “witnesses” to the clerical practice. By enduring the cold, the lack of sleep, and the uncomfortable sitting (although several people take pillows and blankets), and by following the sounds and movements generated in the ritual space with awe and appreciation, they partake in the performance and seal its ritual efficacy.


457

References Abe Yasurō 2001. Seisha no suisan: chūsei no koe to yoko naru mono. Nagoya: Nagoya University Press. — 2005. “Omizutori no tekusuto kagakuteki kōsatsu shiron – Tōdaiji nigatsudō shunie no tekusuto uchū”. Tōgō tekusuto kagaku kenkyū 2/2: 127–142. Berthier, Laurence 1981. Syncrétisme au Japon. Omizutori: le rituel de l’eau de jouvence. Paris: École Pratique des Hautes Études (Cahiers d’études et de documents sur les religions du Japon 3). Boyer, Martha & Fujiyoshi Jikai 1970. “Omizutori. One of Japan’s Oldest Ceremonies”. The Eastern Buddhist 3/1: 67–96. Daitō Takaaki 2008. “Jiin girei ni okeru Nakatomi harae: Tōdaiji shunie ‘Ōnakatomi harae’ no tensho”. In: Nihon ni okeru shūkyō tekusuto no shoisō tojihō. Nagoya: Nagoya daigaku bungakubu gurobaru COE puroguramu: 265–286. Gangōji bunkazai kenkyūjo (ed.) 1979. Tōdaiji Nigatsudō shunie no kenkyū. 2 vols. Tokyo: Chūōkōronsha. Horiike Shunbō 1985. Tōdaiji no omizutori. Nigatsudō shunie no kiroku to kenkyū. Tokyo: Shogakukan. (Reprint 2008) Iijima Manji 1942. Shōe no nyonin. Osaka: Zenkoku shobō. Matsuo Kōichi 2006. “Shushōe, shunie o toku: ōken to minzoku”. In: Matsuo Kōichi (ed). Girei o yomitoku. (Rekishi kenkyū no saizensen. Vol. 7) Tokyo: Yoshikawa kōbunkan: 53–110. Mitsuhashi Takeshi 1999. Kokunai jinmyōchō no kenkyū (ronkō hen). Tokyo: Ōfū. Nagamura Makoto 2001. “Jiin shakaishi no kanten kara miru chūsei no hōe”. In: Nara joshi daigaku kodaigaku gakujutsu kenkyū senta (ed.). Girei ni miru Nihon no bukkyō: Tōdaiji, Kōfukuji, Yakushiji. Kyoto: Hōzōkan: 53–89. Nantō bukkyō kenkyūkai 1984. Nigatsudō. (Special issue of Nantō bukkyō 52). Nara National Museum 2006. Omizutori. Nara: Nara National Museum. Nigatsudō engi. In: Komatsu Shigemi (ed.). Zoku zoku nihon emaki taisei: denki engi hen 6 (Tōdaiji daibutsu engi, Nigatsudō engi). Tokyo: Chūōkoronsha, 1998. Satō Michiko 1984. “Kokannon matsuri”. Nigatsudō. (Special issue of Nantō bukkyō 52): 99–126. — 1994. “Kekae: chūsei e no henjō”. In: Satō Michiko (ed.). Chūsei jiin to hōe. Kyoto: Hōzōkan: 1–50. — 2001. “Hōe no katachi: ima, shunie o chūshin ni”. In: Nara joshi daigaku kodaigaku gakujutsu kenkyū senta (ed.). Girei ni miru Nihon no bukkyō: Tōdaiji, Kōfukuji, Yakushiji. Kyoto: Hōzōkan: 25–52. — 2002a. “Nigatsudō shunie no shōmyō”. In: Satō Michiko. Kekae to geinō. Kyoto: Hōzōkan: 597–614. — 2002b. “Sange to kigan no hōe”. In: Satō Michiko. Kekae to geinō. Kyoto: Hōzōkan: 5–26. — 2009. Tōdaiji omizutori. Haru o matsu inori to sange no hōe. Tokyo: Asahi shinbun. Shunie to kō. Special issue of Yamatoji Narara 2009/2.

458

Lucia Dolce

Uejima Susumu 2004. “Chūsei kokka to jisha”. In: Rekishigaku kenkyūkai & Nihonshi kenkyūkai (eds.). Nihonshi kōza. Vol. 3: Chūsei no keisei. Tokyo: Tōkyō daigaku shuppan: 227–261. Tōkyō kokuritsu bunkazai kenkyūjo 1975. Tōdaiji shunie no kōsei to shosa. Geinō no kagaku 6 (Geinō chōsa roku 1). Tokyo: Heibonsha. Tōdaiji 2004. Tōdaiji shunie. Omizutori no shōmyō. Nara: Tōdaiji.

Audio-visual Material Kingu recodo 1997. Tōdaiji shunie. Omizutori no shōmyō. (CD) National Museum of Japanese History 2007. Yakushiji hanaeshiki: gyōhō to sasaeru hitobito. (Video) Shogakukan 2003. Tōdaiji shunie: Nigatsudō omizutori. (Video) Victor Entertainment 1971. Omizutori. (CD)

Fumi Ouchi

Buddhist Liturgical Chanting in Japan Vocalisation and the Practice of Attaining Buddhahood Introduction Intoning mystical formulas such as paritta, mantra, vidyā and dhāraṇī is one of the essential practices in the esoteric tradition of Buddhism.1 According to received understanding, whilst in the early stage of the development of the esoteric type of Buddhism in India these formulas were recited mainly with the aim of averting calamity or misfortune, in the next stage, during which the Mahāvairocana Sutra was formulated, the purpose of performing esoteric rituals shifted to attaining buddhahood. This gave the recitation of mantra or dhāraṇī another role, as one of the crucial means to identify with the Buddha, together with forming mudrā and contemplation practice.2 In Japan, when the mainstreams of the Buddhist esoteric tradition were systematised through the Heian period, mantra recitation (Jap. nenju 念誦) was emphasised in the Shingon and the Tendai schools as one of the tripartite mystic practices, or three mysteries (Jap. sanmitsu gyō 三密行),3 through which a practitioner could realise his or her original nature to be tantamount to the Buddha and attain buddhahood in the present body. This idea is a main pillar of the esoteric Buddhist traditions in Japan, the “realisation of buddha-hood in the present body” (Jap. sokushin jōbutsu 即身成仏). This means that the practice of intoning a mantra or dhāraṇī was endowed with critical power in an elaborately constructed soteriological system. Yet, in fact, as we will see below, there remained an expectation regarding the type of power possessed by the mantra and dhāraṇī that does 1 Richard K. Payne has surveyed the arguments about such mystic language by western academics (Payne 2006: 14–16). Esoteric types of Buddhist tradition are sometimes designated “tantric Buddhism”. This term, however, was originally used by European academics to refer to certain religious practices formed in the latest stage of the development of the Buddhist tradition in India. Here I will use the term “esoteric tradition” to avoid such a connotation. 2 Matsunaga 1984: 87–92. 3 As I discuss below, the tripartite mystic practice has been variously interpreted in Japanese esoteric traditions, but there is agreement that its theoretical source is the Putixin lun attributed to Nāgārjuna (T 32, No. 1665: 574b) rather than the Dazhidu lun in which the term first appeared (T 25, No. 1509: 127c).

460

Fumi Ouchi

not fit into the new frame. In this sense, mantra recitation practice is an important subject for pursuing what type of features the Buddhist tradition in Japan developed. However, no attempt has been made to explore how such recitation is performed and how it actually works in the process of attaining buddhahood. Only Robert H. Sharf has mentioned it in so far as it is involved in his reconsideration of the conventional understanding of visualisation or contemplation, one of the three mystic practices.4 This chapter explores how the practice of mantra intonation functions in esoteric ritual performance to attain buddhahood, focusing on the performance of the eighteen-method practice (Jap. jūhachidō 十八道). As this practice is paradigmatic for all other esoteric rituals and the first method an esoteric practitioner has to learn,5 it is the most promising material with which I can explore the fundamental attitude to mantra recitation in the Japanese esoteric tradition. In order to discuss the subject, I need to take a performative approach along with the doctrinal. Esoteric teaching is basically transmitted face to face from a master to a disciple, but the instructions of great masters were written down and have worked as a guiding principle for practitioners. These instructions provide me with reliable information regarding both my approaches. I will also employ ethnographical evidence, including my own experience of esoteric initiation practices,6 as supporting material, since the actual feelings of practitioners play a significant role in the process. By analysing these materials, I aim at a new understanding of the practical power of vocalisation in the Buddhist soteriological context.

The Eighteen-method Practice and the Mantra Recitation Practice I will start by examining how mantra recitation has been positioned in the eighteenmethod practice, comparing the methods of the two main streams of esoteric tradition, Tendai and Shingon.

4 Sharf 2001. 5 In the Japanese esoteric tradition, practitioners learn the basic forms of practice through a series of initiation practices. This is called the four-fold preparation ritual (Jap. shido kegyō 四度加行), consisting of the eighteen-method practice, the vajra-realm (Jap. kongōkai 金剛界) practice, the matrix-realm (Jap. taizōkai 胎蔵界) practice, and the fire offering (Jap. goma 護摩) practice. The order is slightly different between the Shingon and the Tendai schools. Whilst the eighteen-method practice, the first one, is followed by the vajra-realm in the Shingon system, in the Tendai the matrix-realm precedes the vajra-realm. In both traditions this system is considered to have started in the late Heian period at the earliest. 6 I carried out the four-fold preparation ritual in the Haguro lineage of Shugendō, a religious tradition centred on mountain ascetic practices and that is theoretically systematised by adopting esoteric Buddhism. The Haguro lineage employs the methodology of the Tendai tradition, and I used the manuals belonging to the Anō lineage 穴太流.

Buddhist Liturgical Chanting in Japan

461

When and how the eighteen-method practice was established is not clear.7 What is commonly understood is that the main body of esoteric invocation rituals translated or edited in China was structured around a narrative of inviting an honoured guest to a feast, where he is cordially entertained. This narrative is represented by a combination of different kinds of mudrā and mantra, from which eighteen sets were extracted. This is how the procedure was modelled in China and Japan. The eighteen-method practice was further developed in Japan by adding another series of segments that represent the union of the practitioner with the central divinity.8 Misaki Ryōshū, a modern scholar, has carried out comparative research into this practice in the two Japanese esoteric traditions. Through a doctrinal approach, he concluded that there were certain significant disparities between the two, derived from their different theoretical foundation: the Shingon doctrine has a dualistic system comprising the vajra realm (Jap. kongōkai 金剛界) and the matrix realm (Jap. taizōkai 胎蔵界), whilst Tendai employs a three-fold system by adding another realm, su-siddhi (Jap. soshitsuji 蘇悉地), which integrates the other two.9 However, it is commonly accepted that there is no great difference in the overall framework of the ritual in the two traditions. It is largely so, as the procedures of the eighteen-method practice from both traditions in Table 1 show. The sources on which I base my analysis are the manual compiled by Gengō 元杲 (914–995), the Shō nyoirin kanzeon bosatsu nenju shidai 聖如意輪観世音菩薩念誦次第 from the Shingon tradition,10 and the Jūhachidō shidai contained in the Asaba shō 阿娑縛抄 (1242−1259)11 from the Tendai. The former is one of the two types of manual used in the Shingon tradition. The reason for my selecting it rather than the one attributed to Kūkai 空海 (774–853), is that Gengō’s represents the procedures regarding mantra recitation that I will address below in more detail than does the other.12 The Asaba shō is one of the most important collections of commentaries on Tendai esoteric rituals and the Juhachidō shidai represents the typical form the 7 According to the Shingon tradition, this practice is attributed to the founder Kūkai 空海 (774–835) or his master Huiguo 恵果 (746–805), whereas in the Tendai, different names such as the founder Saichō 最澄 (767–822), Annen 安然 (841–889/898) who systematised the esoteric teaching of the school, his master Dōkai 道海, or even Huiguo have been mentioned as the figure who established it. Examining these accounts, a modern Tendai scholar, Nomoto Kakujō, concluded that the practice in the Tendai tradition should be ascribed to Dōkai or Kūkai (Nomoto 1982: 149–153). 8 Takai 1994: 213–233. 9 Misaki 1988: 592–596. 10 Extracted from Takai Kankai’s detailed commentary on the manual (Takai 1953: 127–216). 11 DBZ 57: 257–263. 12 According to his detailed comparative research, there are almost no other critical differences in the procedure of the two (Takai 1994). Yet, whilst the principal deity is Cintāmaṇicakra Avalokiteśvara in Gengo’s method, Vairocana of the vajra realm is designated in the manual attributed to Kūkai.

462

Fumi Ouchi

practice took in the tradition. The second row of Table 1 shows the outline of the procedure in the Shingon systems, whereas the third is of the Tendai. As these show, the main structure of both follows the same lines, although there are some minor differences in the order and nomenclature of the segments. However, from a performative dimension, there is a critical disparity between the two methods in the section centred on the ‘Regular Recitation’ (Jap. shōnenju 正念誦) rite. This section goes from the stage of identification, “[The Deity] Enters Me and I Enter [the Deity]” (Jap. nyūga ganyū 入我我入) through to the “Dispersed Recitation” (Jap. sannenju 散念誦) in the Shingon system, and from the stage of “Entering samādhi” (Jap. nyūsanmaji 入三摩地) to the recitation of the “Mudrā of Buddhalocani” (Jap. Butsugen butsumo 仏眼仏母) in the Tendai. This is the most crucial part of the practice, through which a practitioner aims to experience being one with the principal deity by the composite practice of forming mudrā, reciting mantras and contemplation. What attracted my attention is the considerable difference in the way the Regular Recitation is performed and how it is positioned. Let us compare in more detail a series of segments centred on the Regular Recitation in the two traditions. As the part boxed with double lines in Table 1 shows, in the Shingon method, this section consists of three main procedures, punctured by the rite of empowerment of the practitioner by the principal deity. First, the practitioner contemplates being one with the principal deity, forming the mudrā and intoning the mantra. After empowering his rosary, he performs the Regular Recitation rite manipulating the rosary in an elaborate fashion.13 The mantra of the principal divinity, Cintāmaṇicakra Avalokiteśvara (Jap. Nyoirin Kannon 如意輪観音), is repeated one hundred and eight times. It is followed by another contemplation on the complex movement of each syllable of the name of the principal deity or that of the five elements of all nature (Skt. pañca-bhūtāni, Jap. godai 五大).14 Additionally, mantra recitation of various deities is performed. In the Tendai method, the procedure starts with the contemplation of the emergence of the principal deity on a moon disc that symbolises the originally pure nature of human beings in the practitioner’s mind. Then, using the empowered rosary, 13 The tradition instructs as follows: the practitioner pinches the rosary with thumbs and forefingers, stretching out the other three fingers of both hands. He turns the palms outwards and holds them at a distance of around ten centimetres in front of the breast, and maintains the bodily position. After contemplating one of the seed syllables on the moon disc in his heart and that of the principal divinity, he starts mantra recitation with counting the times he recites the divinity’s name using the rosary (Takai 1953: 198). 14 According to the traditional understanding, the idea of this type of contemplation is based on the fifth chapter of the Mahavairocana Sutra (T 18, No. 848: 17b, c), while the term “contemplation of the syllable circle” (Jap. jiringan 字輪観) that designates such contemplation was first used in the manual of the veneration ritual for Akșobhya Buddha (Ch. Achurulai niansong gongyangfa, Jap. Ashukunyorai nenju kuyōhō 阿閦如来念誦供養法 T 19, No. 921: 15, 19) (Misaki 1988: 204–207).


463

the practitioner performs the Regular Recitation, reciting the mantras of the principal deity Acalanātha (Jap. Fudō Myōō 不動明王) one thousand and eighty times and the mantras of other designated deities one hundred and eight times each. After that, (s)he contemplates the meaning of the seed syllable of the principal deity. According to the Shingon tradition, each of the three main procedures represents each of the tripartite ways of realising unification with the Buddha: [The Deity] Enters Me and I Enter [the Deity] means the mystery of the body; Regular Recitation means the mystery of the mouth; and Contemplation of the Syllable Wheel represents the mystery of the mind.15 Thus, each of the three procedures is given equal significance and regarded as a series of rites as a whole.16 The Tendai interpretation is rather different. A typical explanation is given in the Asaba shō. It says that the Regular Recitation is the most important esoteric practice,17 and this interpretation remains current down to the present.18 Tendai thus puts special emphasis on Regular Recitation, over and above the other procedures. The ethnographic evidence puts into focus the points for discussion. On the one hand, the ethnographic evidence demonstrates that the Tendai practitioner experiences the ritual structure along the same lines as the doctrinal explanation. In the Tendai method, the Regular Recitation takes from sixty to ninety minutes to complete, more than half the total duration of the practice. It is the part that costs a practitioner the greatest effort, both physically and mentally, and at the same time provides him or her with a feeling of satisfaction caused by the physical stimuli of 15 This interpretation, with respect to the eighteen-method practice, can be traced back to the thirteenth century at the latest, as Raiyu’s 頼瑜 (1286–1304) instruction on the eighteenmethod practice (Jūhachidō kuketsu 十八道口決) shows (T 79, No. 2529: 70). If we do not limit ourselves to this particular ritual, we can find the earliest example of such a construction in a collection of oral transmissions on esoteric veneration rituals given by Shinzei 真済 (800–860), Betsugyō shidai 別行次第 (SZ 23: 145). He was one of the greatest disciples of Kukai. Yet, a thirteenth-century commentary on the Shingon esoteric rituals in general (Hozō shō 秘蔵抄: SZ 9: 21–133) emphasises primarily the mystery of the mouth (SZ 9: 110). This writing often quotes the interpretations of the priests involved in the movement of esoteric Pure Land thought known as Himitsu nenbutsu 秘密念仏 that I will mention below, such as Kakuban 覚鑁 (1095–1143), Jippan 実範 (n.d.–1144), Jōhen 静遍 (1166–1224), and Dōhan 道範 (1178–1252). They advocated a practice intoning Amitābha’s name or its first syllable, regarding them as an esoteric mantra (Kushida 1979: 181–232). We can see here a positive tendency towards using vocalisation as a Buddhist practice, common among them. In addition, the modern instruction on the eighteen-method practice in the Denbōin lineage 伝法院流 attaches importance to the Regular Recitation interpretation (Tobe 2006: 31). 16 Takai 1953: 192. Takai quotes here medieval commentaries such as the Usuzōketsu 薄草決 by Raiyu (T 79, No. 2535) and the Kaijinhiketsu 開心秘訣 by Kenjin 憲深 (1192–1263). 17 DBZ 57: 272. 18 For example, the beginner’s manual, the Shidogyōki shishō 四度行記私抄 edited by a Tendai master and published by Hieizan senshūin, a school for training Tendai priests, gives the explanation that this rite is called the “Regular Recitation” because it is the core of the practice (Shibuya 1933: 34).

464

Fumi Ouchi

intoning the mantra and listening to it. It is therefore the most impressive section of the procedure, where the practitioner experiences a ritual climax. In the Shingon method, though, the Regular Recitation is not felt to be such a prominent section as it is in Tendai. On the other hand, in the actual performance, the Shingon practitioner sometimes experiences the ritual climax at the Dispersed Recitation rite, not at the three segments theoretically emphasised. These differences highlight two questions to address. One is the issue of how the Regular Recitation rite of the Tendai method accomplishes the distinctive role in the ritual procedure that aims at attaining buddhahood. The other is what the significance is of the inconsistencies between the theoretical construction and the actual performance that the Shingon method suggests.

Vocal Arts for Attaining Buddhahood I will first focus on the Regular Recitation in the Tendai system to discuss how repeating the mantra works to accomplish the ritual purpose. Many of the commentaries on the eighteen-method practice instruct that the Regular Recitation should be accompanied by the contemplation of syllable movement (Jap. jidōkan 字道観). This interpretation can be traced back to the Shijūjōketsu 四十帖決 (1016–1081), which is a collection of commentaries given by one of the greatest masters who systematised Tendai esoteric practice, Kōkei 皇慶 (977–1049), and recorded by one of his disciples Chōen 長宴 (1016–1081). This is the writing on which later instructions relied for authority. Kōkei explains different types of method for performing the contemplation in proportion to a practitioner’s skill, but the underlying idea is common among them. According to it, the following contemplation should be done during the mantra recitation: the syllables of the mantra are imagined to be placed on the moon disc which is the body of the principal deity, and, emerging from his mouth and entering the head of the practitioner, are put on the moon disc in the practitioner’s body, then to emerge from his mouth and enter the feet of the principal deity. According to this source, this contemplation is crucial for Regular Recitation.19 This instruction seems to allot a supporting role to the practice of intoning mantras to accomplish the contemplation. Yet, in light of the ethnographic evidence, the function of the recitation is not so simple. In the course of the recitation practice, the practitioner has a certain distinctive experience, though what it is exactly cannot be defined. According to the traditional method, the practitioner performs the rituals alone in a hall remote from places of everyday life.20 Repeating 19 T 75, No. 2408: 835. 20 In the modern system, esoteric practitioners are mostly trained in the institutions organised by the Shingon or Tendai schools. In such a case, practitioners carry on the series of initia-


465

mantras for a long time alone in a dark and silent hall, in a special setting surrounded by Buddhist images, an altar adorned with special ritual fittings and offerings, and the smell of incense, can affect the practitioner both physiologically and psychologically. The practitioner gains a certain feeling of accomplishment simply by being able to perform this. However, at this point, what he experiences during the performance is left formless. Contemplating the syllables of the mantra that he is reciting may provide him with a framework within which he can construct a physical or mental experience and give it appropriate interpretation. In this sense, the contemplation of the syllables plays a prominently significant role, as the medieval masters explained. Nevertheless, the ethnographical evidence suggests that the contemplation does not work apart from the recitation. In conjunction with the physical activity of vocalising the syllables, the practitioner gains a vivid and actual image of what he is contemplating. We can see here that intoning mantras functions as a vocal art, or a vocal technique for the practitioner’s understanding of what he experiences during the practice on the basis of the physical senses. This indicates another line of enquiry. There is a discrepancy between the actual performance and the theoretical explanation as to the contemplation and the mantra recitation of the Regular Recitation rite. Furthermore, the fact that the great masters instructed a variety of methods for contemplation from simple to complicated may contradict the absolute characteristic of the contemplation practice the masters stated. The Shijōketsu says that a beginner can merely contemplate a seed syllable of the deity.21 We need to examine what causes the discrepancy and the relatively unstable attitude towards the contemplation. I will do this by analysing the theoretical writings and the manual for the eighteen-method practice written by a medieval Tendai monk Ninku 仁空 (1309– 1388). His commentary on the eighteen-method practice, the Jūhachidō ritsuin shō 十八道立印鈔22 has had a great authoritative influence on the practice from the fourteenth century downwards. His instructions as a whole follow the conventional explanations given by earlier masters. As to the Regular Recitation, he instructs that the mantras of the principal deity and some other mantras be repeated in conjunction with contemplating the movement of the syllables of the mantras in a similar fashion to the Asaba shō. He gives relatively simple instructions with respect to tion practice in a group. However, I did it alone at my master’s temple. I was given a room in the temple and just stayed there or in a hall where I performed the practice. I was also forbidden to speak to anyone except the master. For me, the isolation was one of the most piercing experiences of the period of practice, and it worked as a factor in causing me to experience an unusual mental state. 21 T 75, No. 2408: 835a. Kōkei also instructs the beginner to perform a simple contemplation on the principle deity, although it is not an instruction regarding the eighteen-method practice but a veneration ritual for a specific deity based on the eighteen methods (Gyōrin shō 行林抄: T 76, No. 2409: 320b). 22 TZ 21: 5–45.

466

Fumi Ouchi

the Regular Recitation, confining his reference within its practical dimension. In contrast, his strong comment on the rite of Entering samādhi implies his emphasis on the rite. He starts the section emphasising the practice of controlling the breath (Jap. chōsoku 調息), and then criticises practitioners for performing the rite of Entering samādhi without the right preparation of breath control. Then he gives instructions about the method of contemplation of identifying with the principal deity, and explains the meaning of each component of the procedure in detail. He closes the section by stressing that an honest esoteric practitioner should be mindful of what the contemplation means and its importance.23 This suggests that he intended to attach a greater importance to contemplation, even though, as we have seen, a practitioner may gain a quite different feeling during the actual performance. His construction of the practice comes from his theoretical stance as a Tendai scholar-monk. His basic position concerning the attainment of buddhahood and esoteric practice is demonstrated in his commentary on the Darijing yisi (Gishaku sōketsu shō 義釈捜決抄: TZ 10, 13). He insists on two points. First, in his view, ordinary mortals can attain buddhahood in the flesh-and-blood body only if they “realise the true nature of their mind”. When Tendai teachings were established in Japan, one of the most crucial issues was how to guarantee ordinary people that they could attain buddhahood while retaining their natural body and what the specific method for this was. Conventional Chinese Tiantai teaching did not allow that ordinary mortals could accomplish enlightenment without gaining the status of sage. This means that one cannot realise buddhahood while retaining the flesh-andblood body in this world.24 The Japanese Tendai tradition challenged this contention both in the context of the esoteric teachings and of the teachings based on the Lotus Sutra. According to the analysis of the modern scholar Ōkubo Ryōshun, the three predecessors of Ninkū in the Tendai esoteric discourse grappled with the issue of how theoretically to demonstrate that a common mortal can accomplish buddhahood with his present body in his life time. Ennin 円仁 (794–864) asserts that one can attain a body as a Buddha without discarding one’s natural body, and this view was developed by Annen 安然 (841–889/898), reaching the statement 23 TZ 21: 37–40. 24 Tiantai doctrine has two types of grading system of practitioners in the process of realising Buddhist enlightenment. One is the fifty-two or fifty-seven stages, and the other is the six levels of identity (Ch. liuji, Jap. rokusoku 六即). Practitioners who get above the sixteenth stage of the fifty-seven stages or above the eleventh stage of the fifty-two stages are regarded as sages (Ch. sengwei, Jap. shō-i 聖位), which is equivalent to the fifth of the six levels of identity, just below ultimate identity (the highest level). As Zhiyi commented in his commentary on the Lotus Sutra, reaching this stage means that the practitioner realises buddhahood (Fahua wenju 法華文句: T 34, No. 1718: 136b), but at that point he is due to throw away his natural body, according to the teaching of the Dazhidu lun 大智度論 (T 25, No. 1509: 146b, 351a, 602a).


467

that a “common mortal” (Jap. bonbu 凡夫) can attain buddhahood with his own body and in the present time. Enchin 円珍 (814–891), developing his thinking along different lines from that of the other two, insisted it was meaningless to set a boundary between common mortals and sages, and emphasised the possibility of Buddhist awakening in ordinary mortals.25 Intertwining the two different lines of interpretation, Ninkū successfully established the theoretical basis for common mortals’ attaining buddhahood while retaining the nature inherited from their parents in this world. The second point of Ninkū’s discourse is the practice for accomplishing it. The pillar of esoteric practice is the idea of the three mysteries, but the interpretation of it has varied in accordance with the master and the lineage. Some emphasised a balanced performance of each of the three practices, whereas others insisted that one could realise buddhahood by performing only one of the three. Regarding the latter, there were also different views about which one should be most emphasised.26 Ninkū’s basic view about this can be seen in the second section of the first chapter of his commentary on the Darijing yisi. He asserts that in order to attain buddhahood one should carry out the tripartite mystic practices.27 This appears merely to follow the conventional view. However, by establishing the theoretical conciliation we saw above, he reinterpreted practice as the method that leads common mortals to realise buddhahood. He next supplied more specific instruction, where, among the three mystic practices, the pivot is the mystery of mind, or contemplation. However, if a person’s faculty prevents him from realising the true nature of mind by contemplation, he is instructed to adopt the other two practices as aids.28 Apparently, Ninkū seems to attach greater importance to contemplation than to bodily movement and intoning mystic words, as Ōkubo has analysed.29 Nevertheless, in my opinion, we should not dismiss Ninkū’s statement that listening to the teaching means Buddhist awakening, wherein one can accomplish the teaching, and that there is no other way to attain buddhahood but by listening to the profound teaching of the Buddha represented by letters and language.30 This argument is based on Annen’s discourse regarding language and the truth in his Meaning of esoteric teaching and time (Shingonshū kyōjigi 真言宗教時義). Here, Annen stated 25 Ōkubo 1998: 104–105, 108; 2004: 246–253. See Ennin’s commentary on the Sushiddi Sutra (Soshitsujikyō sho 蘇悉地経疏: T 61, No. 2227: 390c, 401b, c), Enchin’s elaboration on Yixing’s commentary on the Mahavairocana Sutra (Dainichikyōsho shō 大日経疏抄: DBZ 26: 677) and Annen’s commentary on the Putixin lun and the Putixin yi (Bodaishigi shō 菩提心義抄: T 74, No. 2397: 268a). 26 See Ōkubo 2004: 64–106, Hirakawa 1992: 41–56 and Kushida 1979: 211–219. 27 TZ 10: 119. 28 TZ 10: 127. 29 Ōkubo 2004: 271–277. 30 TZ 13: 257; 10: 105.

468

Fumi Ouchi

that in the esoteric teaching one enters the Buddha world through the gate of true language (Jap. shinnyo gensetsu 真如言説), but he never mentioned the act of “listening”,31 in relation to the portion that Ninkū cited. This may suggest that Ninkū was rather well aware of the significance of listening as a means of realising enlightenment. Naturally, the action of listening cannot be identified with that of intoning. Nevertheless, as the Asaba shō typically shows, according to the esoteric tradition, in the context of the esoteric practice, intoning the mantra represents the Buddha’s preaching.32 Therefore, Ninkū’s argument implies that he endorsed the undeniable importance of the mantra recitation as the mystery of words, although he set conditions: that even though one hears the Buddha preach, this action is meaningless unless one knows that one is hearing the Buddha’s words;33 and even if one intones a mantra, this is not the same as the vajra words embodying perfect wisdom unless one understands the meaning of the mantra.34 In sum, Ninkū does not simply stress the priority of the mystery of mind, but suggests a mutually dependent relationship between the two. Consequently, despite the apparent inclination for the mystery of mind or contemplation practice, Ninkū’s view as a whole seems to be a compromise. To attain buddhahood, a practitioner, as an ordinary mortal, needs to perform esoteric practices to achieve a balance of body, speech, and mind, at the same time being aware of the supporting correlation between intoning the mantra and contemplation. Thus, Ninkū formulated a theoretical base within the Tendai esoteric framework for common mortals to attain buddhahood in the flesh-and-blood body in this lifetime, and directed the practice for it based on this theory. However, his posture towards the three mystic practices seems to be relatively inarticulate. This is similar to the attitude Ninkū displayed in his commentary on the eighteen-method practice. What caused it seems to be a double-sided conflict that confronted him. On the one hand, as a scholar-monk who tried to reconcile the Tendai system that makes cessation and contemplation (Jap. shikan 止観) the fundamental practice for realising enlightenment with the esoteric system and its three mystic practices for attaining buddhahood with the natural body, Ninkū needed to attach the greatest importance to contemplation practice. On the other hand, as an esoteric practitioner, he could experience how effectively acoustic events and the intoning activity affect a practitioner. In so far as he intended to devise a practical means for ordinary mortals to attain buddhahood, he could not ignore their flesh-and-blood nature. This is the point of origin for the discrepancy between the actual experience and the theoretical explanation of the mantra recitation practice. The inconsistent 31 32 33 34

T 75, No. 2396: 380. DBZ 57: 272. TZ 10: 114. TZ 10: 107.


469

attitude towards the Regular Recitation rite can be seen as the result of devising a specific means for attaining buddhahood accepting the nature of ordinary mortals.

Vocal Arts for Involvement In the Shingon method, mantra recitation seems to serve the ritual process in another way, as demonstrated by the Dispersed Recitation rite. This comes just after the Contemplation of the Syllable Wheel in the ritual procedure. As seen above, the sequence from the “[The Deity] Enters Me and I Enter [the Deity]” through to the “Dispersed Recitation” (Jap. sannenju 散念誦) is the centre of the procedure, but the last segment seems to be different from the three before it. While each of the other three segments is allotted to one of the tripartite mystic practices, the Dispersed Recitation is not. In fact, tradition explains the reason for performing the rite as an appeal to the deities connected with the principal divinity for help to reinforce its power (Gyōhō kan'yō ki 35 行法肝要記: SZ 23: 305).36 In this sense, this part is theoretically allotted a supporting position. Probably for this reason, the rite has not attracted academic attention. Nevertheless, we should take account of the fact that despite its secondary allocation, the Dispersed Recitation rite is indispensable to the ritual process. Almost no lineage omits it from the eighteen-method practice or the rituals based on that practice.37 Furthermore, the Shijūjōketsu, to which we have referred above, records the rite.38 This means that the rite, as a distinctive feature of the Shingon mode, attracted the attention of a master affiliated with another stream of esoteric Buddhism in Japan. Yet, where the rite derives from is not clear.39 As seen above, when the eighteen-method practice was formed in Japan, segments representing the unification of the practitioner with the central divinity were added to the main part of the esoteric invocation rituals represented by eighteen sets of mudrā and mantra. The Dispersed Recitation rite was also added into this process, taking account of the fact that the prototype of the enlarged method manuals, attributed to Kūkai, included the rite. Even though it was put in later, the Shijūjōketsu attests that the rite was performed 35 This is a commentary on the esoteric practices in general transmitted in the Ono lineage 小野流 of the Shingon tradition. This is attributed to a medieval scholar monk In'yū 印融 (1435–1519), who contributed to the development of the Shingon tradition with his large number of theoretical and practical writings (SZ 43: 166, 353–354). 36 Takai 1953: 208 and Tobe 2006: 32. 37 According to the commentary on the eighteen-method practice written by Raiyu, one of the Daigoji lineages, Rishō-in 理性院, did not use the rite (T 79, No. 2529: 70c). 38 T 75, No. 2408: 835. 39 Ōyama Kōjun explains that this rite came from the invocation ritual manual based on the Sutra of the Benevolent King (Ninnōgyō shidai hō 仁王経次第法, aka, Ninnō hannyagyō nenju shidai 仁王般若念誦次第) attributed to Kūkai (Ōyama 1993: 129). However, the authorship of the manual is disputable (KZ 4: 753–754).

470

Fumi Ouchi

in the Shingon tradition in the middle of the eleventh century at the latest. According to the Shingon construction, the newly added portion was formed on the model of the invocation ritual for Amitāyus,40 but the manual for this practice does not include the rite or an equivalent to it. This also implies that the rite is somewhat out of place in the ritual procedure. Dōhan 道範 (1178–1252), one of the most important figures for the theoretical development of the Shingon tradition, above all in the secret Pure Land discourse, advanced a clear view on the rite, which seems to be one of the sources of the present understanding that this recitation is a supporting practice. His commentary reads: “Regarding the Dispersed Recitation, [masters say that] no canonical work provides us with a clear explanation about it. [We perform it] in order to ask the buddhas, bodhisattvas, and divas related with the principal deity for their protection, with earnest intent to realise the perfect manifestation of the principal deity’s power. It is an aid to accomplishing the ritual”.41 Dōhan makes two points. First, he clearly states that the rite has no canonical foundation. This demonstrates that a tradition adopted the rite as an essential element for the ritual process while aware of its dubious origin. Hence, the rite is not set there for a doctrinal reason. The second point is linked with the reason for it being set there. According to Dōhan, the role of the rite is to ask the deities for help, with which the power of the principal deity can be strengthened. This being so, the mantra recitation performed there serves as a medium for realising the deities’ power. Thus, Dōhan’s explanation shows that tradition has allotted another important, but at the same time ticklish role to the mantra recitation practice at the stage of the Dispersed Recitation, which is not in accord with the Shingon tenet that mantra recitation serves as one of the profound mystic practices for attaining identification with the Buddha. This use of mantra recitation appears to be a relic of practices from the time before Kūkai established the new esoteric system, if following modern Shingon rhetoric that sets a distinct boundary between the “pure” and “miscellaneous” esoteric or tantric traditions.42 Prior to the new type of esoteric system that was, in brief, characterised by the idea of attaining buddhahood in the present body through the tripartite mystic practices, there had spread a type of esoteric practice centring on reciting dhāraṇī. It is a well-known fact that the oldest collection of Buddhist tales 40 T 19, No. 930; Takai 1994: 217–218. 41 Gyōhō kan'yō shō 行法肝葉鈔: SZ 23: 168. 42 Western-language academics have cast doubt on the concept of “pure” tantric (Sharf 2001: 266–267; Davidson 2002: 145; Payne 2006: 16–18). Recent scholarship in Japan does not accept the term and concept uncritically, but terms such as “pure esoteric” (Jap. junmitsu 純密) and “miscellaneous esoteric” (Jap. zōmitsu 雑密) are still generally used.


471

in Japan (Nihon ryōi ki 日本霊異記) indicates that in the earliest stage of the introduction of Buddhism, people mainly expected different types of practical merit from Buddhist teachings through practices such as reciting, copying or keeping dhāraṇī or sutras.43 Furthermore, a famous biographical anecdote relates that, in his youth, Kūkai accomplished an ascetic practice centred on the practice of repeating the mantra of Ākāśagarba (Jap. Kokuzō Bosatsu 虚空蔵菩薩). The Dispersed Recitation implies an expectation closely linked to this type of power of the intoned mantra or dhāraṇī. In this sense, the scheme of setting the “pure” esoteric system against others does not work here. In the Shingon method, the power expected of the sound of the mantra strongly supports the ritual accomplishment beyond the framework of the “three mysteries”. This apparent inconsistency suggests that the experience that a practitioner acquired through performance strongly affected ritual formulation. Ethnographic evidence supports this view. Despite the theoretical importance of the ritual sequence allotted to the three mysteries, only an extraordinarily proficient practitioner can actually accomplish the contemplation of being one with the deity and the complicated movement of the seed syllables. More than a few modern practitioners feel that the performance in fact culminates with the Dispersed Recitation rite. This practice is not accompanied by contemplation or the elaborate manipulation of the rosary, unlike the Regular Recitation rite. The practitioner simply keeps intoning a series of mantra, listening to his or her own vocalisation. For ordinary practitioners, repeating the mantras of different deities, above all the more than one thousand recitations of the mantra of the principal deity, demands their greatest efforts. Consequently the performance to some extent leads them towards a certain unusual feeling, which sometimes may be close to one of profound accomplishment.44 Naturally, the modern practitioner’s experience is not necessarily applicable to every case throughout history. However, even if a practitioner trained in a pre-modern monastery had successfully performed the ritual with skilful contemplation, he would still have experienced physical and mental stress similar to that of a modern practitioner, because he has followed the same procedures and repeated the mantras numerous times. We can assume that this type of experience can lead the practitioner to be absorbed into the ritual process. In addition, as seen above, the tradition construes that this practice is performed for the purpose of gaining the support of the other deities in order to strengthen the principal deity. When the experience combines with the theoretical construction, a practitioner can be involved in the ritual per43 A typical example can be seen in a story where a patient is cured with the power of the dhāraṇī of the Thousand-armed Avalokiteśvara (the twelfth and the thirty-fourth of the third volume, Nihon ryōi ki: 145–146, 180–181). 44 One of my informants explained that he had a feeling close to samādhi during this rite, when he successfully performed the procedures preceding it.

472

Fumi Ouchi

formance in a double sense, sensorially and intellectually. This feeling of involvement is essential to activate the ritual power. When practitioners are deeply involved in the enactment, a ritual affects them physically and psychologically and they merge into the value system or semantic system of the ritual. Here the mantra recitation functions as an effective means for involving practitioners in the ritual process. As Robert Sharf has concluded from his detailed analysis of the use of Buddhist images in Shingon invocation rituals, casting light on the performative dimension, the pillar of esoteric practice is mimicking the body, speech and mind of the Tathāgata. In a ritual hall as a sanctuary, a practitioner undertakes a rite and learns “to behave as if he were dwelling in a sacred realm, as if he were in the presence of the principal deity, as if he had merged with Mahāvairocana”. By accumulating these arduous practices, he learns the appropriate manner of behaving, responding, and thinking as an esoteric practitioner. During the performance, however, a practitioner remains fully cognisant of his physical environment, like an accomplished stage actor.45 Because his point is questioning the phenomenological model for constructing esoteric practice emphasising the inner experience, Sharf stresses here the formal or, if I might say, the mechanical dimension of the ritual. However, I think that the as if aspect is deeply connected with the psychological or mental dimension of the ritual, that is, the feeling of involvement in the enactment. Like a successfully acted play, the as if practice needs to be performed with reality in order to work effectively, and such reality cannot be brought about without a sense of being involved in the performance. This being so, I will conclude that one of the essential reasons for setting the Dispersed Recitation rite in the ritual procedure is its performative effect on the sensory nature of practitioners. This is not the only case. Certain other rituals were rendered following a similar direction in the process of seeking a practice more appropriate to the reality of human beings. A typical example is the contemplation ritual on the seed syllable “A” (Jap. ajikan 阿字観), the first syllable of anutpāda, which means the truth of primordial non-birth (Jap. honbushō 本不生) in Sanskrit. It is one of the fundamental contemplation practices in the esoteric system in Japan. If we analyse how the performative dimension of the ritual altered, we can discern a positive attitude towards using vocalisation as a soteriological aid. The first manual outlining this practice was the “Admonitions of Contemplation on A-ji” (Ajikan yōjin kuketsu 阿字観用心口決: T 77, No. 2439) attributed to Jichie 実慧 (786–847), one of the disciples of Kūkai. It laid down the traditional framework of the practice. The centre of the practice is contemplating being one with A-ji on a moon disc resting on a lotus flower, through which the truth of primordial non-birth is realised. Changes were made by later priests, but the main structure was retained. Of these changes, 45 Sharf 2001: 195–196.


473

the most critical to my argument were those that two medieval monks advanced. The first is Shōkaku 勝覚 (1057–1129), who brought prosperity to one of the leading lineages of the temple Daigo, the Sanbōin 醍醐三宝院, and the second is Kakuban 覚鑁 (1095–1143), who laid the foundations of the school known as Shingi Shingon 新義真言 (literally “new doctrine”). The new idea was the contemplation of the breath of the syllable a, which was later called asokukan 阿息観.46 Controlling the breath is one of the elements essential to every esoteric practice. Yet, whereas it is performed as a preparation for the core of the ritual in other cases, now the breath was identified with A-ji, so that a practitioner contemplated A-ji with his breathing.47 Kakuban, combining it with faith in Amitābha, later developed this idea into the secret Pure Land thought (Jap. himitsu nenbutsu 秘密念仏).48 It was quite a daring idea to insist that the breath is one with Amitābha and breathing directly leads to attaining Amitābha’s Pure Land. This line of thought culminated in Dōhan’s Commentary on the Secret Nenbutsu (Himitsu nenbutsu shō 秘密念仏抄).49 Because of its innovative characteristic of reconciling faith in Amitābha’s name and the esoteric teachings, this idea has attracted the attention of both Japanese and Western-language academics.50 Among them Kushida, Sanford, and Stone mentioned the theoretical linkage of the secret nenbutsu with the contemplation of A-ji and the importance of the notion of breathing in Kakuban’s theory.51 The development of this idea is another important subject to discuss for my argument, but here I would rather emphasise the fact that Shōkaku and Kakuban instructed that the syllable “A” should be vocalised. Stone mentioned Kakuban’s insistence on the crucial nature of the A-syllable embodied in the sound of the exhaled breath,52 but her argument is mainly contextualised in the development of the deathbed practice.53 In contrast, my point here is not confined within that particular ritual. Shōkaku states that a practitioner should be devoted to intoning “a” any time and 46 Neither Kakuban nor Shōkaku used the term “asokukan”. 47 KGZ 2: 1010. 48 Kakuban’s ideas about the secret veneration of Amitābha are discussed in two of his writings: the Secret Interpretation of the Five Wheels and Nine Syllables (Gorin kujimyō himitsu shaku 五輪九字明秘密釈) and the Secret Comments on the Essentials to the Last Moment (Ichigo taiyō himitsu shū 一期大要秘密集). Recently, scholars such as Mukai Ryūken and Tachibana Nobuo have expressed some doubt as to whether the latter was Kakuban’s authentic work (Mukai 1991: 17; Tachibana 1992: 758). 49 Kushida 1979: 201–219. 50 Kushida 1979: 181–232; Satō, Tetsuei 1995: 397–426; Satō, Mona 2000; 2001; 2002; Sanford 2006; Stone 2006; 2007. 51 Kushida 1979: 201–206; Sanford 2006: 169–174; Stone 2006: 169–170. 52 Katsumata Shunkyō also noted that contemplation on the breath of the syllable “A” was their original idea, but he did not attempt any more detailed discussion of the meaning (Katsumata 1992: 13–15). 53 Stone 2006: 169–177.

474

Fumi Ouchi

anywhere (Ajikan shidai 阿字観次第 in Kitao 1991: attached chart). Kakuban went further, saying that if a practitioner continues to devote himself to the practice, he can reach the supreme accomplishment (Ajikangi 阿字観儀 in Kōgyō daishi senjutsu shū: 228). Despite its positioning as the most basic type of contemplation, the Ajikan is not actually easy to accomplish because of its abstract nature. As Payne has argued, the symbolic value of the syllable “A” in Sanskrit is unfamiliar to native speakers of Japanese.54 Contemplating one’s breath still demands considerable concentration and skill. In contrast, anybody can vocalise “a”. The “a” intoned during the ritual, however, is not merely a syllable but the sound theoretically unified with the deepest truth of primordial non-birth. The methods that Shōkaku and Kakuban taught represent their efforts to establish a practice reconciled with the theoretical basis of the esoteric tradition, whereby a practitioner can continue retaining his nature as a flesh-and-blood being. This demonstrates the underlying stance forming the vocal art shown in the eighteen-method practice in this tradition.

Conclusion The eighteen-method practice mantra recitation plays different types of critical role. In the Tendai tradition, it provides a practitioner with the physical foundation for perceiving and constructing the distinctive experience of being one with the principal deity. The Shingon tradition devised another way with the Dispersed Recitation rite. By performing a simple repetition numerous times, a practitioner becomes deeply involved in the ritual procedure, through which he or she learns how to behave and interpret his or her experience by accumulating the as if practices. Beyond the sectarian boundaries given by these differences, however, the analysis carried out in this paper suggests two common points. One is the salvational power of vocalisation in esoteric practice. Mantra recitation practice makes efficient use of vocalisation as one of the fundamental faculties of a human being. As the discourse of the A-ji contemplation shows, the period when both the A-ji contemplation and the eighteen-method practice were being formulated and altered was a time when the Buddhist traditions in Japan were attempting to deal with the reality that the human being has a sensory nature and that this has consequences for the practice. When the esoteric traditions devised a practical method to allow for it, they started accepting vocalisation as a typical activity of a flesh-and-blood being. This suggests the somatic nature of enlightenment in the Japanese Buddhist tradition and the critical role of vocalisation for accomplishing it.

54 Payne 1998: 237.


475

The other point is a creative correlation between theory and practice. The devices of vocalisation reveal the discrepancy between the actual performance and its theoretical basis. According to my analysis, in both the Shingon and Tendai systems the actual procedure of the practice gave more importance to the physical or sensory feeling of practitioners than to theoretical consistency. However, that does not mean that the theoretical basis is useless. On the contrary, I would like to emphasise that a practitioner can rightly interpret and direct his experience using a theoretical framework. This dynamic correlation supports the accumulation of the as if experience in the practice, through which a practitioner becomes deeply involved in the esoteric system. It does not only mean that an individual is appropriately trained as an esoteric practitioner, but that the practitioner is expected to carry out the public performance of ritual with great cogency. If the performance successfully involves the audience, the salvational power can affect them as well. In this sense, it is not necessarily appropriate to criticise the esoteric tradition in Japan for emphasising the performative over contemplation and by so doing weakening its “religious” power.55 Rather, I argue that being aware of the importance of the performative dimension of rituals can lead to a dynamic reproduction of new devices for soteriological use.

55 Matsunaga 1996: 128.

476

Fumi Ouchi

Table 1.1: Procedure of the Eighteen-methods Practice (left side) Shingon method Entering the Ritual Hall

Veneration of the Buddhas Taking the Seat Preparing the Offering Veneration of the Buddhas Purification with Incense Contemplation of the Three Mysteries Purification of the Practitioner Rite of protecting the Practitioner Empowering the Holy Water Empowering the Offerings Contemplation of the Syllable “Fire” Purification of the Place Contemplation of the Buddha Awakening the Deities Veneration of the Buddhas Statement of the Ritual Purpose Veneration of the Deities Five Forms of Repentance Seeking Buddhahood Esoteric Commandments

Vow Five Vows Veneration of the Triple Powers Madrā and Mantra of the Vajra Bodhisattva

Outline of the Shingon method

Preparation Adorning the practitioner Purifying the area and offerings


477

Table 1.1: Procedure of the Eighteen-methods Practice (right side) Outline of the Tendai method

Preparation

Tendai method Rite of Protecting the Practitioner Entering the Ritual Hall Confession Purification of the Practitioner Veneration of the Buddhas Taking the Seat Purification With Incense Preparing the Offering

Empowering the Holy Water Empowering the Offerings Clapping Snapping Fingers Washing the Offerings Purification of the Offerings Polishing the Offerings Statement of the Ritual Purpose First vow as an esoteric practitioner Veneration of the Deities

Purification and Protection of the Practitioner

Verses of Veneration Veneration of the Buddhas Awakening the Deities Nine Means for Attaining the Buddha Way Mudrā & Mantra of the Vajra Bodhisattva Vow Five Vows Purification and Protection of the Practitioner

478

Fumi Ouchi

Table 1.2: Procedure of the Eighteen-methods Practice (left side) Shingon method Purification of the Place Setting up a Fence around the Place Contemplation of the Ritual Hall

Outline of the Shingon method Building the sacred realm

Adorning the sanctuary Madrās and Mantra of Ākāśagarbha Madrās and Mantra of the Vajra Bodhisattva Treasure Vehicle Inviting the Vehicle Welcoming Four Seed Syllables Madrās and Mantra of Hayagrīva

Inviting the deities

Binding and protecting the sanctuary

Spreading of the Net Setting up the Vajra Fire Showing the Invitation Offering the Water Offering the Lotus Seat Ringing the Bell Offerings Five Kinds of Offering Hymn of Four Types of Wisdom Hymn of the Principal Deity Verse of the Triple Powers Veneration of the Principal Deity The Deity Enters Me and I Enter the Deity The Principal Deity Empowers the Practitioner Regular Recitation The Principal Deity Empowers the Practitioner Contemplation of the Syllable Wheel The Principal Deity Empowers the Practitioner Dispersed Recitation

Procedure for recitation


479

Table 1.2: Procedure of the Eighteen-methods Practice (right side) Outline of the Tendai method

Tendai method Purification of the Place Setting up a Vajra Fence Contemplation of the Ritual Hall Establishment of the mandala realm Verse of the Triple Powers Fulfilment of the Offerings Ringing the Bell Sending the Vehicle Inviting the Vehicle Welcoming the Deities Defence against Hindrances Spreading of the Vajra Net Setting up the Fire Fence

Offerings

Manipulation

Offering the Water Offering the Lotus Seat Verses of Welcoming Preparing the Place for Inviting the Deity Five Kinds of Offering Comprehensive Offerings Hymn of the Principal Deity Hymn of the Four Kinds of Wisdom Entering Samādhi Mudrā of the Principal Deity Empowering the Rosary Regular Recitation The Rite of Putting Away the Rosary Entering Samādhi Mudrā of the Principal Deity Mudrā of the Buddhalocanii

480

Fumi Ouchi

Table 1.3: Procedure of the Eighteen-methods Practice (left side) Shingon method Five Kinds of Offering Offering the Water Later Ringing of the Bell Hymns Comprehensive Offerings Veneration of the Principal Deity Transfer of Merit Cordially Transferring the Merits Taking Down the Fence Sending Back the Deities Purification of the Practitioner Rite of Protecting the Practitioner Veneration of the Deity Going Out of the Hall

Outline of the Shingon method

Offerings

Buddhist Liturgical Chanting in Japan Table 1.3: Procedure of the Eighteen-methods Practice (right side) Outline of the Tendai method

Accomplishment

Tendai method Five Kinds of Offering Comprehensive Offerings Hymn of the Principal Deity Hymn of the Four Kinds of Wisdom Offering the Water Ringing the Bell Transfer of Merit Transfer Merit to Every Existence Taking Down the Fence Sending Back the Deities Purification of the Practitioner Rite of Protecting the Practitioner Veneration of the Deity Going Out of the Hall

481

482

Fumi Ouchi

References Abbreviations DBZ = Dainihon bukkyō zensho 大日本仏教全書 1970–1983. Tokyo: Suzuki gakujutsu zaidan / Kōdan Sha. KGZ = Kōgyō daishi zenshū 興教大師全集 1935. Edited by Tatsue Nakano 中野達慧. Tokyo: Sesōken. KZ = Kōbō daishi zenshū 弘法大師全集 1978 Wakayama: Mikkyō bunka kenkyūjo. SZ = Shingonshū zensho 真言宗全書 1977–1987. Edited by Zoku Shingonshū zensho kankōkai 続真言宗全書刊行会. Kōya: Zoku shingonshū zensho kankōkai. T = Taishō shinshū daizōkyō 大正新修大蔵経 1924–1928. Edited by Junjirō Takakusu et al. Tokyo: Taishō issaikyō kankōkai. TZ = Tendaishū zensho 天台宗全書 1973–1974. Edited by Tendai shūten kankōkai 天台宗典刊行会. Tokyo: Daiichi Shobō.

Canonical Works Achurulai niansong gongyangfa 阿閦如来供養法. In: T 19, No. 921. Dazhidu lun 大智度論. In: T 25, No. 1509. Fahua wenju 法華文句. In: T 34, No. 1718. Mahâvairocana Sutra 大毘盧遮那成仏神変加持経. In: T 18, No. 848. Putixin lun 菩提心論. In: T 32, No. 1665. Wuliangshourulai guanghang gongyang yigui 無量寿如来観行供養儀軌. In: T 19, No. 930.

Other Primary Sources Ajikan 阿字観. Kakuban 覚鑁. In: KGZ 2. Ajikan gi 阿字観儀. Kakuban 覚鑁. In: KGZ 2: 999–1001. Ajikan yōjin kuketsu 阿字観用心口決. Jichie 実慧. In: T 77, No. 2439. Asaba shō 阿娑縛抄. In: DBZ 57, No. 432. Betsugyō shidai 別行次第. Shinzei 真済. In: SZ 23: 141–146. Bodaishigi shō 菩提心義抄. Annen 安然. In: T 74, No. 2397. Dainichikyōsho shō 大日経疏抄. Enchin 円珍. In: DBZ 14: 2–5. Gishaku sōketsu shō 義釋捜決抄. Ninkū 仁空. In: TZ 10, 13. Gyōhō kan'yō ki 行法肝要記. In'yū 印融. In: SZ 23: 295–307. Gyōhō kan'yō shō 行法肝葉鈔. Dōhan 道範. In: SZ 23: 147–178. Gyōrin shō 行林抄. In: T 76, No. 2409. Hizō shō 秘蔵抄. In: SZ 9: 21–133. Jūhachidō kuketsu 十八道口決. Raiyu 頼瑜. In: T 79, No. 2529. Jūhachidō ritsuin shō 十八道立印鈔. Ninkū 仁空. In: TZ 21: 5–45. Nihon ryōi ki 日本霊異記. Shin Nihon kotenbungaku taikei 30, 1996. Tokyo: Iwanami Shoten.


483

Ninnōhannyagyō nenju shidai 仁王般若経念誦次第. Kūkai 空海. In: KZ 4: 720–754. Shijūjōketsu 四十帖決. In: T 75, No. 2408. Shingonshū kyōjigi 真言宗教時義. Annen 安然. In: T 75, No. 2396. Soshitsujikyō sho 蘇悉地経疏. Ennin 円仁. In: T 61, No. 2227. Usuzōketsu 薄草決. Raiyu 頼瑜. In: T 79, No. 2535.

Secondary Sources Davidson, Ronald M. 2002. Indian Esoteric Buddhism: A Study of the Tantric Movement. New York: Columbia University Press. Hirakawa, Akira 1992. “Kakuban ni okeru ichimitsu jōbutsu”. In: Kogyō daishi Kakuban kenkyūronjū henshūiinkai (ed.). Kōgyō daishi Kakuban kenkyū: Kōgyō daishi happyakugojūnen onki kinen. Tokyo: Shunjū Sha: 41–56. Katsumata, Shunkyō 1992. “Naikan no seija Kakuban shōnin”. In: Kogyō daishi Kakuban kenkyūronjū henshūiinkai (ed.). Kōgyō daishi Kakuban kenkyū: Kōgyō daishi happyakugojūnen onki kinen. Tokyo: Shunjū Sha: 5–26. Kitao, Ryūshin 1991. “Kōgyō daishi ni okeru ajikan 2”. Chizan gakuhō 40: 69–95. Kushida, Ryōkō 1979. Shingon mikkyō seiritsukatei no kenkyū. Tokyo: Sankibō busshorin. Matsunaga, Yūkei 1984. “Mikkyō no jissen”. In: Miyasaka Yūshō et al. (eds.). Kōza mikkyō 1: Mikkyō no riron to jissen, Tokyo: Shunjū Sha: 86–100. — 1996. Mikkyō. Tokyo: Iwanami Shoten. Misaki, Ryōshū 1988. Taimitsu no kenkyū. Tokyo: Sōbun Sha. Mukai, Ryūken 1991. “Ichigotaiyōhimitsushū no senja kō”. Mikkyōgaku kenkyū 23: 17. Nomoto, Kakujō 1982. “Saisho no Taimitsu jūhachidōki”. Tendai gakuhō 24: 149–153. Ōkubo, Ryōshun 1998. Tendai kyōgaku to hongaku shisō. Kyoto: Hōzōkan. — 2004. Taimitsu kyōgaku no kenkyū. Kyoto: Hōzōkan. Ōyama, Kōjun 1993. Zōho kōtei Chūin-ryū no kenkyū. Osaka: Tōhō Shuppan. Payne, Richard K. 1998. “Ajikan: Ritual and Meditation in the Shingon Tradition”. In: Richard K. Payne (ed.). Re-Visioning “Kamakura” Buddhism. Honolulu: University of Hawaii Press: 219–248. — 2006. “Introduction”. In: Richard K. Payne (ed.). Tantric Buddhism in East Asia. Boston: Wisdom Publications: 1–31. Sanford, James H. 2006. “Breath of Life: The Esoteric Nembutsu”. In: Richard K. Payne (ed.). Tantric Buddhism in East Asia. Boston: Wisdom Publications: 161– 189. Satō, Mona 2000. “Dōhan cho Himitsunenbutsu-shō in'yōbunken shutten chūki”. Bukkyō bunnka kenkyū ronjū 4: 130–141. — 2001. “Dōhan no himitsunenbutsu shisō: Myōgōkan o chūshin to shite”. Indogaku bukkyōgaku kenkyū 49/2: 614–616. — 2002. “Chūsei Shingonshū ni okeru jōdoshisō kaishaku: Dōhan Himitsunenbutsushō o megutte”. Indo tetsugaku kenkyū 9: 80–92. Satō, Tetsuei 1995. Eizanjōdokyō no kenkyū. Kyoto: Hyakkaen.

484

Fumi Ouchi

Sharf, Robert H. 2001. “Visualization and Mandala in Shingon Buddhism”. In: Robert H. Sharf & Elizabeth Horton Sharf (eds.). Living Images: Japanese Buddhist Icons in Context. California: Stanford University Press: 151–197. Shibuya, Jigai 1933. Shidogyōki shishō. Shiga: Hieizan senshūin. [Reprint Kyoto: Shibakinseido]. Stone, Jacqueline I. 2006. “Just Open Your Mouth and Say “A”: A-Syllable Practice before the Time of Death in Early Medieval Japan”. Pacific World Journal 3/8: 167–189. — 2007. “The Secret Art of Dying: Exoteric Deathbed Practices in Heian Japan”. In: Bryan J. Curvas & Jacqueline I. Stone (eds.). The Buddhist Dead: Practices, Discourses, Representations. Honolulu: University of Hawai’i Press: 134–174. Shibuya, Jigai (ed.) 1933. Shidogyōki shishō. Shiga: Hieizan senshūin shuppanbu. [Reprint Kyoto: Shibakinsei dō]. Tachibana, Nobuo 1992: “Kōgyō daishi to Heian jōdokyō”. In: Kogyō daishi Kakuban kenkyūronjū henshūiinkai (ed.). Kōgyō daishi Kakuban kenkyū: Kōgyō daishi happyakugojūnen onki kinen. Tokyo: Shunjū Sha: 754–770. Takai, Kankai 1953. Mikkyō jisō no kenkyū. Kyoto: Takai zenkeshu chosaku kankōkai. — 1994. “Jūhachidō no kenkyū”. In: Yūshō Miyasaka et al. (eds.). Mikkyō taikei 9: Mikkyō no jissen. Kyoto: Hōzōkan: 213–235. Tobe, Kinji 2006. “Denbōin ryū Jūhachidōnenjushidai ryakusetsu”. Shingonshū buzanha sōgōkenkyūin kiyō 11: 17–34.

Katja Triplett

Esoteric Buddhist Eye-healing Rituals in Japan and the Promotion of Benefits One of the fundamental ideas of esoteric Buddhist rituals is that the body, speech and mind of the participants are prepared to experience the beneficial power of the deity’s body, speech and mind, and for them to see that ultimately there is no difference between the deity and themselves. This moment of the meeting of the Buddha or bodhisattva and the ritualist is referred to in Sanskrit as adhiṣṭhāna and in Chinese jiāchí 加持 (Japanese reading: kaji). The basic meaning is “assistance” but translations such as “blessings” or the “grace” of a Buddha or bodhisattva, or “empowerment” are also found. The term, as commonly used in esoteric Buddhism, refers to the compassion of the Buddha that can be depended on. It can also mean to simply aid, or support, and therefore often implies the application of prayer and ritual to which the Buddha responds. The support from a Buddhist deity may come in the form of alleviation of suffering and protection from fire, etc. that are referred to as “this-worldly benefits”. Many rituals and ritual texts aimed at initiating these kinds of blessings or this-worldly benefits, and not aimed directly at the ultimate enlightenment as the final goal, have been classified by practitioners of Esoteric Buddhism as inferior. Nevertheless, for the lay community as well as the community of monks and nuns, the perceived benefit of having a healthy body and mind for as long as possible has been considered of utmost importance in Buddhism since the earliest times. The Buddha being portrayed as the ultimate physician who can heal all the ills of the world, is a well-known metaphor that is found in the first Buddhist texts.1 Having access to the special blessings of a deity that can protect one from calamity, from being possessed by a demon, or can heal terrible diseases, brings with it a certain amount authority over others and an association with superhuman spiritual power as well. The fact that numerous sūtras connected to Avalokiteśvara,2 a bodhisattva thought to be a most powerful saviour, have been transmitted from early on could mean that they were important objects of study for elite monastics, 1 For a standard overview of the notion of “illness” and healing in the history of Buddhism see “byō” 病 in the Buddhist encyclopedia Hōbōgirin. 2 In this article referred to in the modern Japanese rendition as Kannon 観音.

486

Katja Triplett

and therefore esoteric in the usual sense of the word. However, as we shall see, several factors lead to the conclusion that versions of an important incantation or dhāraṇī ritualistic text were indeed transmitted in East Asia beyond the control of the elite. This opens up a whole discussion on ritual and authority not only in China and Korea but also in Japan. Particular groups not belonging to the elite seem to have been even more influential in shaping Japanese Buddhist culture and religious ideas than certain priest scholars whose commentaries have been carefully guarded through the centuries and can still be studied today. But tracing the waxing and waning of ritual trends of social groups not belonging to the cleric elite often proves difficult because of the scarcity of textual records. Other sources must be found. When not only limited to the study of canonised texts and holy scriptures, texts such as popular narratives as well as iconographic aspects of statues and paintings or temple layouts become meaningful sources for the study of religious history. In the first part of this contribution I examine the textual and iconographic background of esoteric Buddhist eye-healing rituals, focusing on a medium-sized temple near Nara, Japan, called Tsubosakasan Minami hokkeji 壺阪山南法華寺 (f. 703), or Tsubosakadera 壺阪寺 for short. In the second part, I look at the activities at the temple and attempt to show that in the medieval and early modern periods eye-healing rituals may have been widely promoted, and perhaps even conducted by practitioners not officially part of a Buddhist monastic order. The healing of sight-related diseases is one particular worldly benefit of venerating the bodhisattva Avalokiteśvara offered at Tsubosakadera. The temple terms the benefit: ganbyō fūji 眼病封じ, “the stamping out of eye disease”. Needless to say, the eyes are a very sensitive and an especially valued human organ; in Buddhism eyes also have a deep symbolic meaning of seeing the truth, and tales of the miraculous healing of blindness abound in collections of Buddhist tales in all of East Asia. To propagate and promote the belief in the efficacy of the principal deity of a Buddhist temple has not only spiritual motives, of course. A constant stream of pilgrims who seek to obtain the special benefit from their visit is the foundation of a healthy economy of every temple. But, as I will show, the promotion of special benefits was not always exclusively in the hands of the temple leaders. At least since the medieval period, the fame of the bodhisattva Kannon’s particular blessings attracted groups of semi-religious entertainers who also wanted to participate in the activity of promoting the bodhisattva’s miraculous powers and in doing so earn a living. My thesis is that the wide non-elitist spread of the incantation text directed at the Thousand-armed Thousand-eyed Kannon, the Senju sengen daranikyō 千手千眼陀羅尼経, was strongly connected with the development of the culture of blind itinerant ritualists and minstrels in medieval and early modern Japan.

Esoteric Buddhist Eye-healing Rituals in Japan

487

These ritualists produced numerous well-known religious stories, had their own strictly organised hierarchies, their own networks and were only loosely connected to Buddhist temples. They performed in the garb of monks and nuns but usually remained outside the official Buddhist ordination system.3 Their performances included incantations and supplications to various deities, mainly the protective deity of their trade, Benzaiten, which is the Indian Buddhist deity of music and sound, Sarasvatī.4

The Eye-healing Miracle at Tsubosakadera According to a Ninth-century Buddhist Tale Tsubosakadera was founded in 703 – only 150 years after the official introduction of Buddhism in Japan – by a monk from one of the earliest Japanese Buddhist temples, the Moto-Gangōji 本元興寺. The special feature of Tsubosakadera is that its principal object of worship, an Eleven-faced, Thousand-armed Avalokiteśvara (Kannon), is venerated as the healer of sight-related diseases. According to a short temple chronicle written in 1211, this special power has been venerated from very early on because it cites an episode from a ninth-century collection of Buddhist tales or setsuwa 説話 called Kanrei roku 感霊録 that act as – one could say – a testimonial to a miraculous eye-healing ritual from the time when Buddhism was taking root widely in Japan. The temple chronicle was written by the Kōfukuji 興福寺 priest Kasagi Shōnin Jōkei 笠置上人貞慶 (1155–1213)5 and survives as a copy from 1496.6 Unfortunately, this chronicle was not illustrated and there are no illustrated scrolls of the foundation legend or the miraculous events contained in it; so for the imagery we have to rely on the text only. The chronicle was originally called Minami hokkeji korō den 南法華寺古老伝. The fifteenth-century copy, though, has the title Tsubosakadera korō den, Tsubosakadera being the still popular name of this temple, and is in this article simply referred to as the Korō den. 3 For an introductory treatment of medieval entertainers in Japan see Ruch 1977; for blind minstrels and ritualists see Fritsch 1996, Hirose 1994; for a somewhat dated but still standard history of Japan’s blind see Nakayama 1965; for a social history of the visually impaired in Japan see Katō 1974. 4 See the study on this goddess from its Indian origins by Ludvik 2007. 5 Jōkei may have transcribed his master’s words. Jōkei’s master was Kakken 覚憲 (1131– 1213), his Fujiwara uncle, who had been in charge of the consecration (eye-opening) ceremony of the Great Buddha at Tōdaiji 東大寺 and who had retired to Tsubosaka temple in his old age. The chronicle was compiled just after a devastating fire destroyed Tsubosakadera killing many, so it may be surmised that it had been Kakken’s wish to record its history and the miracles surrounding the site before passing away. 6 The original is kept in the temple archive and a facsimile can be found in the Nara Prefectural Library. For an edition of the chronicle see Nagai 1985b.

488

Katja Triplett

The first section of a temple chronicle usually contains the foundation story or engi 縁起 of the temple proper. However, the Korō den opens with a rather political statement that a certain Eikei 永継 forged a foundation legend in order to include his own family in the genealogy of founders. The Korō den seeks to rectify this forgery by stating repeatedly that the true founder, Benki Daitoku 弁基大徳7, was not a member of Eikei’s family. The second section of the chronicle lists the objects of worship or honzon 本尊. The first honzon mentioned is, of course, the Elevenfaced, Thousand-armed Kannon, and then follows the aforementioned episode from the ninth-century setsuwa collection Kanrei roku that can be summarised as follows: A blind novice8 named Chōnin9 who grew up as a beggar, came to Tsubosakadera and intoned an incantation (神呪 jinju) uninterruptedly for many years, with the wish to have his sight restored by Kannon. His piety to the Thousand-armed Kannon was so great that people started calling him “Senju shami”, the novice devoted to Senju or the Thousand-armed. In the end Chōnin regained his eye-sight and became a widely recognised Yin-Yang diviner. This episode of the miraculous healing of blindness by Kannon is quoted as being number 15 of the setsuwa collection, possibly of the first fascicle. The chronicle goes on to list and describe the temple buildings, affiliations forged with Kōfukuji and other temples, special ceremonies and rituals conducted and visits by important monks. The temple record also chronicles major fires and subsequent reconstruction. The final part contains the liturgical calendar, annual festivals and ceremonies. Whereas this latter part resembles more the list of a bookkeeper, the sections on the Buddhist honzon contain narratives in which miracles abound. Miracle tales – not only in Japan – often reveal very detailed aspects of the rituals conducted and provide details of the persons involved shedding light on the specific concerns of the ritualists. The Korō den contains, for example, the actual words of the incantation (神呪 jinju, or in Sanskrit dhāraṇī):

7 This individual has not been identified. However, Nagai uses his detailed knowledge of contemporary literary sources to form ideas about the temple founder Benki, especially his conclusion; Nagai 1985a: 150. 8 “Novice” is here the translation of shido no shami 私度の沙彌, denoting monks and nuns in ancient Japan who ordained “themselves” but were not part of a particular Buddhist institution and therefore not officially, fully ordained. Shido can be also written 自度, often read jido. 9 The name of this novice, Chōnin 長仁, is apparently found only in this document.


489

曩謨日精摩尼手明眼暗大悲者観自在菩薩 Nōmaku [= Namu] nisshō mani ju myō gen an Daihi[shō]ja10 Kanjizai bosatsu This can be translated as “Hail bodhisattva Kanjizai, the Great Compassionate One who makes blind eyes see with the Hand of the Jewel of the Sun Essence”. The incantation is from a group of Mahāyāna sūtras dedicated to Avalokiteśvara found in volume 20 in the Taishō edition of the Buddhist canon. There are 13 versions of such sūtras, all translated in the Tang dynasty, with the earliest translated in the seventh century.11 The Senju sengen daranikyō translated into Chinese by Bhagavaddharma around 650 A.D. can be said to be the most authoritative text among this group of texts,12 and could even be considered the “most important of all the esoteric scriptures in China”.13 The text was brought to Japan starting in the eighth century. Especially important for the transmission of the text’s teachings, imagery and ritual knowledge was the development of a repentance ritual from the Senju sengen daranikyō by the Chinese monk Zhili 知禮 (960–1028) based on earlier models. This ritual is practised even today not only in Chinese Buddhist temples in Taiwan but also in Buddhist (Zen) temples in Japan. It is often practised at memorial services and at funerals due to the specific connection of the Thousand-armed Kannon as the saving deity in the realm of the hungry ghosts, and this is deeply linked to the veneration of ancestors in East Asia. Repentance or sange 懺悔 rituals are seen as the prerequisite for the saving of suffering ancestors from hellish realms.

Canonical Ritual Practices Connected to the Thousand-armed Avalokiteśvara The dhāraṇī sūtra translated by Bhagavaddharma that is connected with the quote mentioned in the Korō den is structured like the following: the text appears as being spoken by the Buddha in the palace of the bodhisattva Avalokiteśvara located on the island of Potalaka, the bodhisattva’s realm. Suddenly there is a great illumi10 Daihi[shō]ja goes back to the Sanskrit term Mahā-kārun̩ika-muni, the monk who is of great compassion. 11 For an in-depth treatment of one sūtra in this group see Reis-Habito 1993. 12 T 1060; the full title of the sūtra is: 千手千眼觀世音菩薩廣大圓滿無礙大悲心陀羅尼經 Qianshou qianyan Guanshiyin pusa guangda yuanman wuai dabeixin tuoluoni jing. This long title can be rendered as “Dhāraṇī of the Bodhisattva With a Thousand Hands and Eyes Who Regards the World’s Sounds and Feels Vast, Complete, Unimpeded Great Compassion”. The text in the Taishō edition includes a preface by the Emperor of China, Yung-lo, dated 1411. The sūtra is fully translated into German by Reis-Habito in chapter III of her study (1993). 13 Yü 2001: 59.

490

Katja Triplett

nation and when asked the Buddha answers that it is because Avalokiteśvara is going to reveal the dhāraṇī. Avalokiteśvara then takes over centre stage and explains the origin and the efficacy of the dhāraṇī. He urges everyone to follow him in making ten vows of compassion. The keeping of the dhāraṇī will result in fifteen types of good rebirth and the avoidance of fifteen types of evil death. Those are then enumerated in the sūtra. The dhāraṇī consisting of 84 phrases is then revealed. Instructions on constructing a sacred space are also provided in detail. The text explains, as is quite usual with such sūtras, that the correct keeping of the dhāraṇī leads to spiritual and mundane benefits. The sūtra also offers numerous instructions about handling certain materials to perform “magical” deeds, and it lists several recipes to deal with various problems or to attain specific goals including gaining good physical health. Here are two examples from Senju sengen daranikyō versions, both dealing with sight-related illnesses. “If one suffers from diseases of the eye, such as having a blind pupil (blindness) or darkened eyes (cataracts), or when the white of the eye is covered or red or has a lustreless membrane, take from the three fruits hārītakī, amalā, and vibhītaka one piece each and pulverise them finely in a mortar. While doing this, one has to pay strict attention to purity and avoid the presence of women who have just given birth, dogs and pigs. One then calls out the name of the Buddha and covers the eye with a mixture of white honey and mother’s milk. The milk that one uses on the person must be the milk of a boy’s mother. Milk from a girl’s mother has no effect. When the medicine is ready, recite the dhāraṇī 1,008 times in front of a statue of the Thousandeyed. After the medicine has covered the eye for seven full days in a dark room protected from the wind, the eye comes back to life. The blind pupil and the previously covered white of the eye will shine with a marvellous lustre”14 The sūtra as translated by Bhagavaddharma ends by identifying the names of the forty mudrā (ritualistic manual signs) of Avalokiteśvara and the benefits they bestow on the worshipper: e.g. if you want to be always free from governmental oppression, then make the mudrā of holding a hatchet. If you desire an official position, then make the mudrā of holding a precious bow. In performing the “Hand of the Jewel of the Sun Essence” in unison with the deity the special benefit is indeed the healing of blindness. One version of the Senju sengen daranikyō (T 1064), translated by Amoghavajra, provides the names of all forty mudrās related to the “thousand” helping hands of Kannon. The sūtra refers to the deity’s upper left hand (no. 8):

14 T 1060.110b, translated by the author of this paper.


491

“If you want to regain your eyesight, then make the sūryamaṇi mudrā holding the sun with a crow inside in your hand.”15 The crow refers to the Chinese sun symbol of a three-legged crow. Thousandarmed Kannon’s upper right hand shows the other principal celestial body, the moon: “If you seek to remain cool in the face of fear, then make the candramaṇi mudrā of holding the moon with a tree and hare inside in your hand.”16 Zhitong’s 智通 seventh-century version and two other later versions also provide detailed rules for the making of the bodhisattva’s image and give detailed information on the iconography of the Thousand-armed Avalokiteśvara.17

Image 1: Kannon’s hand in the sūrya- Image 2: The candramaṇi mudrā of ma-ṇi mudrā holding the sun holding the moon with a tree with a three-legged crow. and hare. Taishō shinshū daizōkyō 20.1064.107a

Taishō shinshū daizōkyō 20.1064.107a

In summary, these texts acted as ritual manuals and were used as instructive material to form repentance rituals and possibly as records of healing therapies and medicines. But who applied these ritual remedies to those seeking assistance and needing medical attention? Avalokiteśvara ritual manuals and sūtra texts in addition to the commentaries were – and are still to some extent – only accessible to an elite class and were therefore esoteric in the strict sense of the word. However, and this is one of Reis-Habito’s main arguments,18 the Senju sengen daranikyō was accessible to common Buddhist practitioners right from its inception in Central Asia, a trend that has continued in Japan beginning in the eighth century. This is because of its change into a repentance ritual first developed by Zhiyi 智顗 (Chih-i) 15 16 17 18

Yü 2001: 63. Yü 2001: 63. T 1057. Reis-Habito 1993: 121.

492

Katja Triplett

(538–597). Because of this, Zhiyi’s and his successors’ rituals – among them the aforementioned Zhili – must have become exceedingly popular and widely practised in East Asia. Another interesting point is that important texts in the group, namely T 105919 and T 1064,20 containing more extensive lists of diseases and healing methods, are not part of any historical sūtra catalogue. Therefore, one can surmise that these two texts are in fact creatively elaborated copies of the Senju sengen daranikyō from the late Tang dynasty, written by individuals who were very familiar with Buddhist therapeutical practices.21 This indicates that specialists existed who actually used the dhāraṇī for the healing of a great variety of diseases. One argument in support of Reis-Habito’s thesis about the popular nature of the dhāraṇī is the passage in the Tsubosaka temple chronicle that contains a reference to the dhāraṇī mentioned above. Certainly, the chronicle functions as a memory store for future generations of monastics, but the more narrative parts, the engi and the setsuwa, provide material for sermons as well and were intended to remind the community of the special power of Tsubosaka Kannon. Relating the simple tale of the blind novice Chōnin gave the impression that Tsubosaka’s Thousand-armed Kannon was always available to pious ritualists and pilgrims from all walks of life despite the daunting political involvement of the powerful Fujiwara clan in the leadership of Tsubosaka temple. Another point is that the person of Chōnin is described as a shido no shami, a “self-ordained novice”, someone who had led a pious life but did not belong to a Buddhist order. After gaining back his eyesight, Chōnin, who remains an unidentified historical individual, gives advice to others as a diviner indicating that he himself had become a soothsayer and healer. Moreover, the ritual related in the chronicle, the uninterrupted recitation of the dhāraṇī, points to a model ritual – a ritual that is phrased as being highly successful and may be intended to serve as a template for ritual performances at this particular temple. The connection to Chōnin, who is described as a Buddhist-Daoist expert, can be seen as pointing to a wide spread of ritual practice based on the Senju sengen daranikyō. It may also be surmised that in the early centuries after the introduction of Buddhism in Japan perhaps therapeutic healing methods may have been gleaned from the teachings, either from written or from orally transmitted versions. This point requires further research, however.

19 This sūtra is often referred to in short as 治病合藥經 (Japanese reading: Jibyō gōyaku kyō), the “Sūtra on the Use of Medicinal Herbs for Healing (by Kannon)”. It was also translated by Bhagavaddharma. Partial translation into English in Unschuld 1985. 20 T 1064 was translated into Chinese by Amoghavajra. 21 Reis-Habito 1993: 211.


493

Image 3: Seated Eleven-headed, thousand-armed Kannon. Main statue of the Japanese Buddhist temple Tsubosaka Minami Hokkeji near Nara. Late Muromachi period (sixteenth century). Photo: Chr. Triplett 2008.

Since the Tsubosaka temple chronicle of 1211 quotes a fairly long narrative tale from the ninth century including numerous brief commentaries of mostly linguistic information, the section about the miracle-working Kannon makes the great importance of eye healing as the special benefit of this temple in the medieval period seem obvious. Nevertheless, the healing of blindness is an extremely common benefit, so the emphasis in the Korō den may be on the effectiveness of the temple’s Kannon, and not so much on the type of miracle. Since the temple has styled itself as being the site of the Eye Bodhisattva and has offered special memorial services – increasingly since the 1970s – for old spectacles and contact lenses as well as prayers for recovery from sight-related illnesses, an unbroken tradition of this particular service seems to have existed since the temple founding in the eighth century.

494

Katja Triplett

Blind Minstrels and the Promotion of Worldly Benefits in Medieval and Early Modern Japan Late medieval and early modern Buddhist narratives indeed reveal that Tsubosakadera and its principal deity had been believed to bestow the special benefit of healing sight-related diseases during that period as well. However, another deity, the goddess Benzaiten (Sarasvatī), is nearly always mentioned in the tales, here as the Benzaiten revered on the sacred island of Chikubushima in Lake Biwa. Since there are a great number of versions of this Chikubushima Benzaiten-Tsubosaka Kannon story extant from the late sixteenth to the eighteenth centuries it can be assumed that this was one of the more popular Edo-period stories that were transmitted in picture scrolls and books as well as through theatre performances, and like many of those Buddhist narratives it was formed in earlier centuries during the late medieval period in Japan. The title of these narratives vary but most refer to the name of the heroine in the story, Matsura Sayohime 松浦さよ姫22. Researching the plot structure in particular, we find that a twelfth-century Buddhist narrative called Hōmyō dōji 宝明童子, the Young Boy Hōmyō, must have been the basis of the story.23 The Hōmyō dōji story of a pious boy eventually evolved into a story of a young girl, Matsura Sayohime, who in the end manifests herself as an eye-healing deity, Chikubushima Benzaiten. The plot of the Matsura Sayohime legend is as follows: the girl’s antagonist, a giant man-eating serpent, bestows a wish-fulfilling jewel on her out of gratitude for being relieved of its existence as a serpent. The serpent is then revealed as being the wrathful form of a local water goddess (in Adachi). In the end, it is saved through the power of the Lotus sūtra and manifests as Tsubosaka Kannon. The oldest extant version of the story is from the sixteenth century, so there is no evidence of how the story of Benzaiten and Kannon evolved from the twelfth century to the sixteenth century. We also do not know exactly who developed the story and to what end, or who exactly was responsible for its wide transmission during the medieval period.24 However, there are clues in the Edo-period versions that point to a possible group of medieval storytellers being involved. The motif of the miraculous healing of the girl’s blind mother points to a group of “blind monks” or mōsō盲僧

22 The characters for the name Sayohime vary, e.g. 小夜姫; often it is in the manuscripts and prints only represented by phonetic syllables. 23 For a thorough study of this narrative see Schmitt-Weigand 2004. 24 Triplett 2004 lists extant versions (Matsura Sayohime legends) and provides an edition of one of the versions as well as a historical background of the narratives and their modern interpretations.


Image 4: Eight-armed Uga-Benzaiten, an esoteric form of the goddess Sarasvatī with the snake-human deity Ugajin on her head. Wooden statue at the Buddhist temple Hōgonji on Chikubushima in Lake Biwa. Photo: Chr. Triplett 2008.

495

496

Katja Triplett

who were active in the twelfth and thirteenth centuries.25 The fact that blindness and the healing of blindness plays a great role in their stories does not constitute in itself proof that visually impaired people transmitted the tale, of course. It seems, however, that since blindness was believed to result from a misdeed in a former life, the visually impaired ritualists – thus stigmatised – used pious Buddhist narratives for gaining merit in order to better their plight. At the same time, groups of blind and visually impaired musicians in religious garb carefully guarded the transmission of certain melodies and ritualistic texts so no one outside the guild could learn them properly and then compete against them, by being organised in strictly hierarchical guilds or tōdō 当道. Therefore, to have a mōsō perform was something special; their trait (the visual impairment) could be seen as a mark of quality and not of failure. Because of the mentioning of particular place names in the story26 and the fact that female singers and musicians, goze 御前 or mekura goze 盲御前, used Tenkawa Benzaiten shrine as a place for gathering and the performance of their art,27 one can also assume that both males and females were involved in the transmission of the tale. This group of female minstrels can be traced at least back to the fifteenth century. In Northern Japan, versions of the Chikubushima Benzaiten story were transmitted as plays, okujōruri 奥浄瑠璃, by goze well into the twentieth century.28

Conclusions To conclude I would like to point out that in the Japanese case miracle stories were not regarded as being only “for the masses” with the priesthood taking a more aloof view. Miracle tales have been an important means of transmitting Buddhist teachings. The Hōmyō dōji story that informed one of the foundation legends and Kannon miracle tales of Tusbosakadera, for instance, is found in a reference book with exemplary stories that priests in the Pure Land tradition used for their sermons. What is perhaps an unexpected find is that ritual performances involving semi-religious groups and individuals shaped the way in which the power of certain deities was perceived by the lay practitioners. In the medieval period, because both Tsubosaka temple priests and itinerant minstrels promoted rituals and the attractive bene25 The fact that the foundation legend (Semimaru 蝉丸) of the blind monks’ guild features in the travel episode or michiyukibun 道行き文 of the story could be one indicator that mōsō were involved in transmitting the story, but since the reference to Semimaru is a very common narrative theme it cannot count as the sole proof for the mōsō’s involvement. I thank Prof. Niels Gülberg for his insightful comments on this point. 26 Honda 1987. 27 Fritsch 1996: 31. 28 Mashimo 1986.


497

fits connected to the healing Kannon and their protective deity Benzaiten, we can discern different, but not necessarily competing networks. The ritual, narrative and performative networks also spanned different regions such as localised networks in the Kansai area and Northern Japan. In the Kansai region the connection of stations on the Kannon pilgrimage route such as Tsubasakadera and the temple Hōgonji 宝厳寺 on Chikubushima must have certainly been of importance, but even more so the activities of members of the tōdō. In Northern Japan in Fukushima-ken we find numerous written versions of ritual performances of the Benzaiten-Kannon story originally performed by goze within networks directly linking them with powerful local families, and also certain temples. For instance, in Isawa 胆沢, Buddhist priests have operated alongside mountain ascetics since early modern times and have serviced a temple, Kewaizaka Yakushidō 化粧坂薬師堂, where Sayohime was worshipped even before the erection of the hall in the seventeenth century.29 Sayohime is said to have conquered and saved the giant serpent from its terrible existence in that region. Kewai (make-up) in the place name indicates the spot where she was adorned and readied for the human sacrifice. The place of worship includes a well with water that is to this day drawn and used for its eye-healing properties.

29 Mashimo 1986: 36-37.

498

Katja Triplett

References Canonical Works T: Taishō shinshū daizōkyō 大正新修大藏經 1924–1935. Edited by Takakusu Junjirō 高楠順次郎 & Watanbe Kaigyoku 渡辺海旭. 100 vols. T 1060: 千手千眼觀世音菩薩廣大圓滿無礙大悲心陀羅尼經 Qianshou qianyan Guanshiyin pusa guangda yuanman wuai dabeixin tuoluoni jing. Translated by Bhagavaddharma. T 1057: 千眼千臂觀世音菩薩陀羅尼神呪經 Qianshou qianyan Guanshiyin pusa tuoluoni shenzhou jing. Translated by Zhitong. T 1059: 千手千眼觀世音菩薩治病合藥經 Qianshou qianyan guanshiyin pusa zhibing heyao jing. Translated by Bhagavaddharma. T 1064: 千手千眼觀世音菩薩大悲心陀羅尼 Qianshou qianyan Guanshiyin pusa dabeixin tuoluoni. Translated by Amoghavajra.

Other Works Cited Fritsch, Ingrid 1996. Japans blinde Sänger im Schutz der Gottheit Myōon-Benzaiten. Munich: Iudicium. Hirose Kōjirō 広瀬浩二郎 1994. “Chūsei mōsō to Yoshino, Kumano 中世盲僧と吉野熊野”. In: Sangaku shugen 山岳修験 13/3: 33–46. Hōbōgirin: dictionnaire encyclopédique du bouddhisme d’après les sources chinoises et japonaises 1929– [and following]. Edited by Sylvain Lévi et al. Tokyo, Paris: Maison Franco-Japonaise [et. al.]. Honda Noriko 本多典子 1987. “Sekkyō Matsura chōja no kōzu 説教「まつら長者」の構図”. Todai ronkyū 都大論究 24: 13–29. Katō Yasuaki 加藤康昭 1974. Nihon mōjin shakaishi kenkyū 日本盲人社会史研究. Tokyo: Miraisha 未来社. Ludvik, Catherine 2007. Sarasvatī: Riverine Goddess of Knowledge: from the Manuscript-carrying Vīṇā-player to the Weapon-wielding Defender of the Dharma. Leiden, Boston: Brill. Mashimo Miyako 真下美弥子 1986. “Okujōruri Chikubushima no honji ron 奥浄瑠璃「竹生島の本地」論”. Denshō bungaku kenkyū 伝承文学研究 33: 27–42. Nagai Yoshinori 永井義憲 1985a. “Minami Hokkeji no kaiki Benki to ‘Hitachi fudoki’ 南法華寺の開基弁基と『常陸風土記』”. In: Nagai Yoshinori 永井義憲 (ed.). Nihon bukkyō bungaku kenkyū 日本仏教文学研究. Tokyo: Shintensha 新典社: 137–151. — 1985b. “‘Minami Hokkeji korō den’ ni tsuite 『南法華寺古老伝』について”. In: Nagai Yoshinori 永井義憲 (ed.). Nihon bukkyō bungaku kenkyū 日本仏教文学研究. Tokyo: Shintensha 新典社: 716–728. Nakayama Tarō 中山太郎 1965 [1934]. Nihon mōjinshi 日本盲人史. Tokyo: Yagi shoten 八木書店.


499

Reis-Habito, Maria D. 1993. Das Dhāranī des grossen Erbarmens des Bodhisattva Avalokiteśvara mit tausend Händen und Augen: Übersetzung und Untersuchung ihrer textlichen Grundlage sowie Erforschung ihres Kultes in China. Nettetal: Steyler (Monumenta Serica, Monograph Series 27). Ruch, Barbara 1977. “Medieval Jongleurs and the Making of a National Literature”. In: Hall, John W. & Takeshi Toyoda (eds.). Japan in the Muromachi Age. Los Angeles: University of California Press: 279–309. Schmitt-Weigand, John 2004. Das Nara ehon “Hōmyō dōji” (17. Jh.): eine illustrierte Handschrift im Besitz des Museums für Angewandte Kunst, Frankfurt am Main. Edition, Übersetzung und entstehungsgeschichtliche Studien. Wiesbaden: Harrassowitz (Bunken 10). Triplett, Katja 2004. Menschenopfer und Selbstopfer in den japanischen Legenden: das Frankfurter Manuskript der Matsura Sayohime-Legende. Münster et al.: Lit (Religiöse Gegenwart Asiens. Studies in Modern Asian religions 2). Unschuld, Paul 1985. Medicine in China. A History of Ideas. Berkeley, Los Angeles: University of California Press. Yü, Chün-fang 2001. Kuan-yin: the Chinese Transformation of Avalokiteśvara. New York: Columbia University Press.

M.A. Butler

Ritual Specialists and Collective Agency in Song Dynasty China In 972 A.D., the first emperor of the Song dynasty (960–1279) issued a decree prohibiting private ownership of divination texts. Such proclamations had been repeatedly issued throughout Chinese history. Divination and practice of what I term “occult arts” has always held a special place in the rule of China: the famous book burning of Qin Shihuangdi in 213 B.C. set aside books of divination and prognostication for exclusively imperial use. During the Song, even though possession of books on divination and practice of occult ritual were punishable by strangulation and sometimes death by slicing, continuing prohibitions suggest that neither of these measures successfully deterred occult practice.1 Despite court crack-downs, occult ritual specialists preserved their techniques: subsequent dynasties prescribed almost identical rituals, and many are still practiced today. The Song imperial court had a contradictory attitude toward those who practiced occult rituals.2 The court restricted occult texts and practices, but could not eradicate these rituals entirely. They needed occult specialists in order to centralize power, and later, when threatened by hostile neighboring states, to hold the Chinese empire together. Military expeditions in particular called for the presence of diviners and other ritual specialists to perform complicated and newly recovered techniques, at times denying them access to necessary equipment. The contradiction in court policy arose from the diviners’ ability to access power that the court could not. Local or private networks tended to monopolize efficacious ritual performance that the Song court sought. The reason the court could not completely control these networks is related to the training and transmission of various divination techniques, and how ritual knowledge created and distributed collective agency. Much previous scholarship on agency portrays it as human, individual and intentional.3 Ritual scholar William Sax argues that in ritual, agency – “the ability to transform the world” – most often originates from non-human agents “distributed 1 McKnight & Liu 1999: 476–484. 2 I use the following terms interchangeably: occult ritual specialists, occultists, diviners and sorcerers. By occult, I mean practices performed to transcend and/or control ordinary phenomena and often violate usual social or cultural practices. See Poncé 1975. 3 Ahearn 2001: 114–116.

502

M.A. Butler

in networks…not necessarily (or even usually) the property of individual persons.”4 All ritual distributes and articulates agency, and, because “those involved submit to the superordinate authority of the group”, that agency is collective.5 Ritual performance entails both canonical and self-referential streams of communications, the latter injecting the intention and current circumstances of its participants into the ritual. Self-referential communication allows interpretive “space” in an otherwise codified, rote ritual.6 Therefore, ritual performance possesses the potential to produce new ideas, beliefs and practices; that is, new systems of thought. This is ultimately an act of collective agency, unintentional or at least not entirely intentional. Such was the case for occult ritual during middle imperial China, a defining moment in the history of Chinese thought. In this essay, I investigate the dynamics of ritual transmission, its resulting networks and their role in producing collective agency. The occult rituals of middle imperial China – numerous kinds of divination, calendrical rituals, exorcism rituals, reading qi 氣 (universal vital essence), omen reading, certain kinds of metalworking, numerology, etc. – were all transmitted through networks, some intentional, some not. Most of these rituals were prescribed for military affairs, some of them exclusively so. The dynamics of their transmission and the networks that formed around that process resulted from the interplay of court policy, traditional and new structures of governance, individual intent, and cosmological constructs of the Song philosophical renaissance. I discuss the dynamic between the court’s centralization process and the networks formed by occult ritualists. Specific rituals – cosmography rituals and battle array schema – to show how intentional and unintentional transmission distributed collective agency.

Overview Overcoming ritual regionalism was one of the missions of the early Song court, since ritual specialization competed with their own centralizing process. Traces of this process emerge in military manuals of the era: the Secret Classic of Venus, Planet of War (ca. early 760s), the Classic of the Tiger Seal (1005), and the Classic of Comprehensive Military Essentials (1044).7 Though the military book (bing shu 兵書) enjoyed an ancient history in China, the above texts constitute China’s first comprehensive manuals, documenting all aspects of warfare: armor, weapons, 4 5 6 7

Kreinath & Snoek & Stausberg 2008: 474, 477–478. Kreinath & Snoek & Stausberg 2008: 481. Rappaport 1979; Bourdieu 1990: 80–97 Respectively, Taibo yinjing written by Li Quan, Huqian jing written by Xu Dong, and Wujing zongyao written by imperial committee headed by Zeng Gongliang and Ding Du. Hereafter referred to as Classic of Venus, Tiger Seal and Comprehensive Essentials.

Ritual Specialists and Collective Agency in Song Dynasty China

503

organization, medical treatment, and so forth. Ritual prescriptions comprise over half their contents. The new encyclopedic tradition of the military manual incorporated cosmological schemes developing at the time – the Song dynasty is known as a watershed in Chinese philosophy. One of the most important of these schemes to appear in the manuals and in Song military ritual practices is that of xiang (象 constellation, image, symbolization, simulacrum, and in Buddhism, the phenomenal in contrast to ultimate reality).8 In all its definitional permutations, xiang is essentially visual, one hallmark of Song philosophy. Some occult rituals appear to have been local specialties.9 Local masters decided whom they taught, sometimes allowing themselves to be discovered by worthy apprentices. In these cases, the lineage of the transmission bestowed authority and efficacy of the ritual technique, still a theme in modern-day martial arts. In such cases, the master assumed charismatic power.10 As transmitter, the master controlled who received this knowledge, often extracting an oath from the disciple to keep the techniques “secret”.11 This was especially true for military techniques. In the text Questions and Answers Between Tang Taizong and Li Weigong, Tang emperor Taizong (r. 626–649) extracts such an oath from the general-master of the title, simultaneously commanding him to pass his secrets on to his successor.12 Similarly, Li Quan 李筌 (fl. 750s–770s?), the author of the Classic of Venus, received his arts from a mysterious Daoist adept who warned him against passing on techniques to those unworthy.13 Once the master passes on “secrets” to the apprentice, these rituals are, of course, no longer secret, as further performance and transmission lay in the hands of the apprentice. The master’s charismatic power then transfers to the technique or its lineage, rather than to the specialist. In these cases, agency originated in ritual authenticity. Therefore, agency that originated in intentional human transmission then transforms into a potentially unintentional, non-human source. Other forms of occult ritual transmission resulted from interactions of tradition, structure, and policy. I now turn to the policies, structures, and collectivities that come into play in the contest for power between the court and occult ritual networks.

8 9 10 11 12 13

Smith et al. 1990: 255–256. LDBRZ 1998. Mauss 1972. See de Weerdt 2006 on secret and restricted texts and textual transmission. Sawyer 1993: 360. Hereafter referred to Questions and Answers. Taibo: Preface; Yunqi: 38.112.5–6.

504

M.A. Butler

Song Centralization and Ritual Networks After the political fragmentation of the tenth century, the early Song emperors began centralizing their power and strengthening the central court. They initiated a number of projects to do so. The first of these was a “text project”. This included gathering up “lost” texts, re-assembling old texts, and generating new texts. Despite quadrupling the number of scrolls in the imperial library within the first sixteen years of the dynasty, the early Song emperors also ordered the compilation of four new encyclopedia collections.14 With these efforts, and their intersection with China’s transition from manuscript to print technology, the early Song emperors thus set about determining a new epistemological orthodoxy.15 At the same time, the court was eager to gather up lost occult techniques. After Song emperor Taizu’s (r. 960–975) initial prohibition in 972, Emperor Taizong (r. 976–997) ordered the “confiscation of texts used by diviners throughout the land” five years later.16 In 984, he issued an edict calling for the recovery of “lost texts”, again targeting occult texts.17 Court suspicion of occult arts motivated these efforts. At the founding of the dynasty in 960, the court offered incentives for diviners to come forward to the court. By 975, the emperor prohibited masters of esoteric arts and techniques (fangshi 方士 and fangji 方技) from assignments outside the capital. In 1017, the emperor revoked some of their official privileges; in particular, the privilege that allowed them years-long visits in their home districts. These specialists and their position within the bureaucracy were re-categorized and split up in 1110, effectively defusing them as a group.18 The court needed the talents of occult practitioners, but remained extremely wary of them. Despite proscriptions and measures to control occult practices outside the court, emperors still consulted these men. Emperor Taizong consulted with Ma Shao 馬韶 (fl. late 960s–998?), who approached the emperor despite prohibitions on private occult practices. He visited Taizong repeatedly and seems to have had an informal advisory role with the emperor. Notably, Ma initiated these consultations,

14 Dudbridge 2000: 1–3. In 960, the imperial library contained 12,000 scrolls. Sixteen years later, the library boasted 46,000 scrolls. The encyclopedias are the Taiping guangji (Extensive records of the Supreme Peace Reign Period), the Taiping Yulan (Imperial Reader of the Supreme Peace Reign Period), the Wenyuan yinghua (Finest Blooms in the Literary Garden), and the Cefu yuangui. 15 Though printing itself did not immediately bring about such orthodoxy. See Cherniak 1994; Brokaw 2005; McDermott 2006; de Weerdt 2006. 16 SS 461.13500; WJZY 16.1869; Cochini & Seidel 1968: 13. On proscriptions and punishments, see McKnight & Liu 1999. 17 Dudbridge 2000: 14. 18 WXTK 35.337; SS 461–462.13495–13534.


505

not the emperor.19 In the 1040s, Emperor Renzong summoned the recluse diviner Xu Fu 徐復 (n.d. over 70 yrs. at death) to ask about the current enemy dispositions.20 Diviner Du Shisheng 杜時升 (fl. late 1180s – ca. 1232) was repeatedly offered an official position after he successfully foretold the fall of the Jin (and later, the fall of the Song), hostile neighbors to the northeast.21 Local practices continued despite these measures to control occult ritual. The twelfth century historian Zheng Qiao (1104–1162) noted in his Comprehensive Treatise that local traditions and texts unknown to the court continued to be transmitted despite government proscription. Non-local occult practices continued also. In 1109, an official of the Bureau of Military Affairs proposed that commandersin-chief detain those suspected of having divination skills or texts, even though proscriptions were already in place. By then, the Song court saw occult and other ritual specialists as the core of the problem of large-scale desertion by military troops.22 The formation of networks of transmission of ritual knowledge made it difficult for the court to root out these ritual specialists. It appears that these networks formed in several ways. First, ritual knowledge passed from father or uncle to son or nephew or through other close family and clan relations. Occultist Wang Chune 王處訥 (915–983?) and his son, Wang Xiyuan 熙元 (960–1018) is one such case.23 Similarly, official Miao Shouxin 苗守信 (954–999) learned calendrical arts from his father, Miao Xun 苗訓 (fl. 950s–960s; over 70 yrs. at death), a military diviner and supporter in the coup of Zhao Kuangyin 趙匡胤, later Song Taizu.24 Guo Nang 郭曩 (fl. early 1100s) of Sichuan province, a recluse diviner of the Book of Change (Yijing 易經; China’s oldest and most revered divination manual) and of numerical arts (shushu 數術), trained in family arts originally learned from occultist Yan Junping 嚴君平 (86 B.C.–10 AD).25 There is also evidence to suspect that, in at least some locales, these clan networks overlapped with the system of military households (jiangmen 將門), which mandated military service through a close degree of kin relations.26 Certain occult arts were earmarked as exclusively military methods; a survey of military texts throughout Chinese history indicates these arts comprised part of the training transmitted through this system, especially in commander families. This oral tradition captured in writing and modified by court committee resulted in the encyclopedic Comprehensive Essentials noted above. Despite changing government policy 19 20 21 22 23 24 25 26

LDBRZ 782; SS 461.13500. SS 461.13500; JWDS 71.944, 131.1730; TSJC 746.50r; LDBRZ 782. JS 127. SS 196.4813. SS 461. TSJC 746.50; SS 1.3–4, 68.1498 and 461.13499; LDBRZ 883. LDBRZ 671–672; SS 459.1360 Chen Feng 2007; Fang 2003.

506

M.A. Butler

toward military commanders and organization, their degraded social position, and the institution of a professional army, this system continued to function informally during the Northern Song (960–1126) at least.27 Second, various divination techniques were transmitted geographically as local or regional specialties. Zhen Qiao argued that “survival and loss” of texts is due less to collecting texts than to their transmission. Key to the latter, Zheng asserts, is the presence of specialists who will copy and transmit texts devoted to their own field. Therefore, “traditions of medicine, of Buddhism and of Daoism have survived the turbulence and destruction of history, while the rich Han literature on the Book of Change and the Legalist school has largely been lost.”28 Other records bear this out. For instance, the famous scholar-official brothers, Cheng Yi 程颐 (1033– 1107) and Cheng Hao 颢 (1032–1085) praised the efficacy of the Sichuan method of the Book of Change. As young men traveling through Chengdu in Sichuan, they encountered an anonymous cooper who had a copy of the Book of Change at his side. Surprised that the cooper owned the book, as well as his ability to read it, the cooper explained the Change hexagram “weiji 未濟” with a clarity and depth that left the brothers with profound understanding of the art. Both went on to achieve great fame as scholar-officials, Cheng Yi especially for the “Cheng method” of Change. He directed students seeking deep understanding of Change to Sichuan. One such student, Yuan Zi (fl. 1060s?) of Fujian province, sought out the Sichuanese cooper but found instead a soy sauce peddler known only as “Old Man Xue” (details unknown). Xue, a taciturn man, did not explain much, and Yuan Zi eventually departed. Sichuan recluse Qiao Ding 譙定 (fl. ca. 1120s?), Guo Nang’s student, taught numerical arts of the Book of Change to at least five men who later achieved official rank. 29 In the above examples – those involving the Book of Change made it into official records because the text formed the foundation for so much Song philosophy – local tradition clearly effected supra-regional networks. The Cheng brothers’ transformative experience with the cooper meant they sent their own disciples to Sichuan, and Qiao Ding’s apprentices – three from Fujian province and two from Sichuan – sought him out especially. Other less official networks exceeded locale also. Occultist Wang Jie 王捷 (fl. 1008–1016) learned to transform iron to gold from his teacher, whom he met in the mountains while Wang was evading military conscription (In other accounts, he learned from a former imperial master while serving time in a remote penal colony). Through convoluted machinations, including exquisite timing, deception, and official corruption, he eventually achieved official rank and cast gold talismans for the emperor.30 Wang Jie may be one of the 27 28 29 30

Chen Feng 2007: sec. 3, jie yu. Dudbridge 2000: 10. LDBRZ 671–72, 686; SS 459.13460–13461. MXBT: 667–669.


507

occultists Hong Mai (1123–1202) referred to when he claimed Song masters of occult arts were nowhere near as good as those of former times because their practices were too clouded by their own interests at court to be effective.31 Nevertheless, Hong seemed to respect cosmograph expert Zhang Yanze 張彥澤 (fl. 1134), who managed to save Song troops (though not their commander, Duke Yang) from certain disaster by correcting an erroneous cosmograph reading. Learning of the “grave blunder” from the military attaché, Zhang immediately asked who performed the technique, indicating that these diviners knew each other.32 The third type of network involves teacher-student transmission of texts and techniques, where the lineage of the object rather than the transmission of the ritual becomes a source of agency and inspiration. Such was the case when Chen Tuan 陳摶 (styled Tunan 圖南, 872–989) passed the “Diagram of the Great Ultimate (wuji tu 無極圖)” to Chong Fang 種放 (955–1015), who transmitted it to Mu Xiu 穆修 (979–1032), who in turn, passed it to philosopher Zhou Dunyi 周敦頤 (1017– 1073).33 Similarly, the “Diagram of the Former Heaven (xiantian tu 先天圖)” was passed from Chen to Chong to Li Zhicai 李之才 (? –1045) to Shao Yong 邵雍 (1011–1077).34 The River Chart (hetu 河圖) and the Luo Writing (luoshu 洛書) – ancient divination charts, “magic squares” that add up to 15 in all directions – share similar provenance.35 All of the transmitted diagrams above exerted a major influence on Song philosophy and the development of occult arts. Indeed, many visual schema arose during the Song and lay at the origin of occult agency. Chen Tuan, who obtained his powers by suckling a Daoist deity and possessed a “mantic mirror of human relations”, was known particularly for his visual invention.36 The River Chart and the Luo Writing were first depicted in visual, not textual, form during the early Song.37 One component of the Song philosophical movement concentrated on developing visual schema as a means of transcending language to express the ineffable. As scholar Mija Milcinski points out, “The awareness of the insufficiency of words resulted in many original solutions. In the Song dynasty [...] China produced a variety of diagrams (tu 圖) by which philosophers and practitioners represented their theories. These formulations often arose on the basis of meditation techniques and could not therefore be fully transmitted by means of language alone.”38 Philoso31 32 33 34 35 36

RZ 453. YJZ 甲.2.16 Berkowitz 1996: 468. On Chen Tuan, see Kohn 1990. SS 435: 12908. Louis 2003. See Louis 2003 for the development of the taiji and other diagrams. Li Yangzheng 2008: 2; SS 194; Berkowitz 1996: 466, 468. Most of his diagrams are now lost. See Louis 2003: 152–163 and 162, n. 38; Li 1990. 37 Li Shen 1998: 231. 38 Milcinski 1999: 386.

508

M.A. Butler

pher-occultist Wang Bi 王弼 (226–249) described the “insuffi-ciency of words” and its relationship to xiang: “Images (xiang) are the means to express ideas. Words are the means to explain the images.[....] Since the words are the means to explain the images, once one gets the images, one forgets the words.”39 This notion is reflected again in Questions and Answers: “Military strategy can be transmitted as ideas but cannot be handed down as words.”40 Xiang is an ancient concept; traditionally it was an image attached to the reading of key hexagrams of the Book of Change. It also meant constellations, heavenly symbols that inspired human invention; in the Book of Change, heaven dangles the constellations (xiang) and humans model them. In their metaphysical formulations, Song philosophers dovetailed renewed conceptual roles of yin-yang (primal forces of shadow and light) and qi (universal vital essence) with a cosmology that featured the changed significance of xiang as symbolizations and simulacra.41 These elemental creative forces constitute the conceptual ground for some forms of divination, which regulated and in turn reconstituted patterns of the universe (li 理) and xiang. In one cosmological scheme, one could apply numbers (shu 數; also emblem, category) to xiang to discern universal patterns, past and future.42 In the military manuals, regulating and reconstituting the power of these forces through xiang are especially salient in the case of cosmography rituals, battle array schema, setting up camp, heavenly timing, geography (dili), fenye (sympathetic resonance between constellations and geographical locations), divining the qi, and inferring events from nature (tui). Xiang were key to occult rituals because through their re-creation, occultists could access cosmic power. Xiang provided access to yin, rather than yang, power.43 The source of yin power was located in the visual aspects of xiang. In order to access the full potential of xiang, the ritual performer created simulacrum of nature and manipulated it, thus creating access to superhuman power. Constellations, the ideal city plan, inner alchemy meditation schemes, battle array schema, and the cosmograph all exhibit visual similarities.44 Such representations acted as cultural guides, visual myths that instructed consciously and unconsciously. They modeled and mimicked the universe at the cosmic and microscopic 39 Adapted from Wang Bi 1994: 31. 40 Sawyer 1993: 341. 41 See Porter 1996; Schafer 1977; Sivin 1980 and 1986; Smith et al.1990; Hall & Ames 1995: 211–225. 42 Sivin 1986; Smith et al.1990: 100–135. 43 Yin and yang, the two fundamental universal powers that separated to create all things, originally signified the shadowed and sunny sides of the riverbank or mountain, respectively. Yin became associated with the receptive, cold, wet, female, ghost, etc.: yang with the creative, warm, dry, male, human, etc. 44 See Harley & Woodward 1994: 534–535, 140–144; Robinet 1989: 298. On these schema in motion, see Anderson 1989–90.


509

scale, and applied to all realms in between. Xiang were one of the most important sources of agency in occult rituals. Simultaneously, they acted as vehicle for the transmission and distribution of agency. As unconscious visual guiding myths, they constitute another method of ritual transmission. The court welcomed some of these networks of transmission, especially those centered on the Book of Change. Song Taizong summoned Chen Tuan to court twice in the late 970s (Chen first refused), and many of the philosophers mentioned above served as officials, or at least rubbed shoulders with those in court circles. However, the court worked hard to suppress some occult arts. While the court repeatedly prohibited local performances of occult rituals, the lack of intention in ritual transmission and of the networks that formed around them meant these networks were difficult to disrupt. The court responded to this perceived threat by appropriating these rituals and creating government bureaus and official positions for certain occult arts. They also tried to eradicate local tradition by standardizing ritual. In 1095, Emperor Zhezong (r. 1086–1100) mandated the collection and documenting of rituals “in words and charts” from all the provinces. Soon thereafter, the court issued to the provinces fixed protocols mandating the form and types of sacrifices local rituals should take.45 Similarly, the Song court re-categorized rituals within the official five-ritual canon (i.e., Auspicious, Commendation, Guest, Military, Inauspicious). Military rituals that involved bloody sacrifice were moved out of the capital and into the field. For instance, the Song court dispensed with the ritual that initiated war and inducted the commander-in-chief, moving it outside the capital city walls; the only dynasty in Chinese history to do so.46 The ritual involved the emperor handing over a ceremonial pike and axe to the commanding general, followed by sacrificing an animal, handling its dissected parts, and the emperor and the commander both smearing their mouths with blood.47 By handing over these objects, the emperor transferred imperial power outside of the court to the commander – “As state [affairs] may not be managed from without, so [the affairs of] war may not be conducted from within. The authority [of the general] should be complete”.48 Calendrical rituals for exorcism of malignant qi also involved bloody sacrifice for the field commander, who officiated: (“with the left hand, behead a gray dog and white bird, and bury them at the compass point where the enemy general’s qi was observed”). At court, the analogous ritual moved “flying chessman” around a giant altar to “exorcise miasmic qi” from the empire.49 Even though both are forms of 45 WXTK: 90.824 46 WLTK: 19.14396–14408; TZ: 592; Butler 2007: 62–69. 47 WLTK: 19.14396–14408; Taibo: 3.51–52. 48 Adapted from Tjan 1949, 1952: 516. 49.HQJ 10.99–100; SS 103.2507. All three manuals contain similar versions of the field ritual.

510

M.A. Butler

calendrical rituals, those prescribed for the field commander involved blood, inherently yin. In contrast, the court rituals used “flying chessmen”. Song court ritual emphasized non-bloody (yang) rituals that clarified boundaries between yang (living, human) and yin (dead, spirit). Court rituals did not invoke natural forces or supernatural beings in the same way field rituals did. The transfer of bloody, yinproducing rituals from the court to the field manifests especially in the Comprehensive Essentials, the culmination of the Song court’s efforts to collect and control orally transmitted rituals. This re-categorization of Song court ritual revolved around the role of yin as a source of super- and non-human agency. Rituals shifted from the court to field were messy, chaotic and bloody. They invoked yin rather than yang power, and therefore, were dangerous. Yet, the court needed powerful rituals for military victory. In designating yin rituals to specifically military use, the court wanted to make their military arm more effective by using these rituals as a powerful means against the enemy. At the same time, the court worried that too much power lay in the hands of those outside their direct control, especially hands holding thousands of troops. Cosmograph rituals profile the dynamics of xiang at work in occult rituals, and its role in collective agency.

Field Rituals and Collective Agency: Cosmograph Rituals and Battle Array Schema Cosmograph rituals are among the most highly creative of occult rituals.50 Recounted in pre-modern literature, these sorcery techniques are commonly practiced by the famous general, Zhuge Liang 諸葛亮 (181–234). In one episode, Zhuge challenges the enemy general, Sima Yi 司馬懿 (179–251), in a contest of battle array shema. Once in the array, Sima and his men are dogged by fog, wind and thunder; they become hopelessly lost inside the array, and cannot find their way out. Zhuge triumphantly sends his men in to capture Sima Yi and his lieutenants.51 The military manuals record many similar techniques, explaining the extraordinary uses of the cosmograph rituals: disappearing into thin air, changing the climate, overcoming the enemy commander’s qi, and jumping time. According to the manuals, cosmograph rituals transformed hostile territory by manipulating time and space. Ideally, cosmograph rituals are performed upon a compass-like device – the cosmograph – that simulates the heavens and constellations on a round rotating

50 For sources on the cosmograph and cosmograph rituals, see Wyatt 2008: 137, n. 6, 140, n. 56 and Anderson 1989–90. 51 Luo 1959: 100.423.


511

plate, and earth on a square fixed plate (fig. 1).52 The cosmograph also shows gates and paths that connect heaven and earth through a series of relationships determined by time. The diagonal roads open and close doors and gates – ghost, heaven, earth, and human – according to correspondences between time, space, and earthly “situations” (ju 局). Through correspondences with cyclically impossible time combinations of the sexagenary calendar system and manipulation of the cosmograph, ritual participants can access supernatural forces. Cosmograph rituals manipulate the gaps and the interstices between segments of ordinary time. The ambiguous space formed by “impossible” time make the move across or into time or other universes possible.53 Therefore, highly developed rituals for manipulating xiang as embodying time – here, constellations and symbolizations – allow supernatural feats. Cosmograph rituals originate from xiang as image in the Change, also. In the Classic of Venus, Li Quan locates the origin of these rituals in the battle between legendary Yellow Emperor and arch-nemesis Chi You. Li invokes the classic myth of the River Chart and the Luo Writing, delivered by a giant turtle and crocodile. These wondrous animals emerge from the Sheng River carrying divine talismans in their mouths, mysterious texts written on an indeterminate material. The Yellow Emperor obtains the meaning of these texts through sacrifice, thereby staging the triumph of civilization over chaos, represented by Chi You. He extracts yin and yang forces from the natural universe via these magic squares, thereby establishing the cosmograph system.54 Xiang in all its senses – as heavenly inspiration, as visual guiding myth, and as application of abstract arts to objects (shushu) – form the foundation for the cosmograph rituals. The cosmograph rituals are highly geometric and numerical. More important, cosmograph rituals transformed time into space, a space that was highly geographic. Battle array schema best exemplify the geographic quality of these rituals.

52 Specific rituals include the Hidden Period (dunjia), Six Water Cycles (liuren) and the Supreme One (taiyi). 53 Wyatt 2008: 120–122 and n.50 above. 54 Taibo 9.243–244.

512

M.A. Butler


513

The Song innovation in battle array schema synthesized the traditions of tu (圖 visual illustration), fu (符 talismans), and ji (籍 registers, especially household and conscription registers) and combined them with xiang.55 The “General Introduction to Battle Array Schema” in the Classic of Venus describes constellations as the inspiration and model for battle array schema.56 The explanation of the Wondrous Grasp (Woqi) and Venus Planet of War (Taibo) battle arrays correlates the eight gates of each array with a hexagram of the Book of Change, a star deity, a color and direction, and an open or closed designation, the latter referring, of course, to its relationship with the cosmograph “roads” (figs. 2.1 and 2.2).57 The Venus Planet of War encampment (fig. 2.2) requires only 10,000 troops to the Wondrous Grasp’s standard 12,500 and is used when the White Tiger constellation appears. In these two schema, there are four gates on the diagonals. Each gate of the array has the same attributes as those of the axis and dial of the cosmograph. Each gate of the battle array has a time aspect, also. This is inherent in its mimesis of constellations – the ultimate timekeeper – and through its correlation with the Book of Change. 58 The troops located inside the array, therefore, also have a time aspect. Li Quan elaborates this theme in his conception of battle array schema: “12,500 people make the army; 12,000 of these symbolize (xiang) the twelve months; the remaining 500 symbolize (xiang) the intercalary month.”59 By having troops stand in for time, Li identifies battle array schema with constellations. He sees battle array schema and the cosmograph as mutually constituted ritual symbolizations. Visual similarities between the cosmograph and the battle array schema clarify these overlapping correlations. Battle array schema functioned as cosmographs. The Tiger Seal prescribes particular camp arrangements for the center of the array using the Six Water Cycles cosmograph ritual, determining where and when other rituals – executions, etc. – can occur.60 According to this scheme, the entire array is ritually constituted. In several Supreme One (taiyi 太乙) cosmograph rituals, troops carry banners colored according to each of the five phases, their position within the array determined by cosmograph. Holding the banners, troops dance out

55 TPYL 609.2742 and 619.2778. Tuji (charts and registers) evokes the association of tu with geography and its corresponding rule of the “nine” regions (jiuzhou 九洲). Ji (registers) were lists of households in the various subsectors of the country: the “broad learning of the tuji”, for example, enabled the sinification (hua 華) of the “thousand gates and the ten thousand households.” (TPYL 612.2752 ). Ji were also military registers, as in the phrase, fuji. During the Tang, ji became associated with the landscape (shanshui tu, tutu). 56 Taibo 6.150–159. The Eight Arrays and Uniting to Become One Array show clear visual similarities to a tenth-century star map excavated at Dunhuang (See Needham 1959: 264, pl. XXIV). 57 Taibo 6.128–19, 132–133. 58 On the time aspect of hexagrams, see Lin 1995. 59 Taibo 3.52 60 HQJ 8.63–67.

514

M.A. Butler

Fig. 2.1: Battle array schema from the Taibo yinjing. Wondrous Grasp battle array schema for the standard 12,500 troops. All but the north and south gates are named for star deities. (The Secret Classic of Venus, Planet of War, ch. 6)


515

Fig. 2.2: Battle array schema from the Taibo yinjing. Venus, Planet of War Battle array schema is used when that planet appears, requiring only 10,000 troops. The Heaven and Earth gates shift to the ordinal, approximating the cosmograph in Figure 1. (The Secret Classic of Venus, Planet of War, ch. 6)

516

M.A. Butler

the patterns and rotations of the cosmograph. As cosmograph, troops access agency.61 Likewise, the “Duke of Thunder Method” in the Comprehensive Essentials stipulates the use of the cosmograph in concert with battle array schema. In that method, the commander stands in the center of the array, turns the dial on the cosmograph, pushes on it, then turns the cosmograph to lines up its axes with those of the array. Using ordinary objects – tallies, yarrow, stalks of grass, a knife – the army thwarts the enemy by disappearing completely. 62 In all of these rituals, xiang (of the cosmograph and the array) distribute collective agency. The performance of these rituals extend agency to troops, unwitting or not, at least for the duration of their participation in the ritual. Battle array schema performed culture; their performance “worked” by creating xiang – simulacra of nature – and manipulating them to siphon universal forces, and using them to “transform the world”. The mimetic and visual qualities were made “really real” through acting patterns that imitated the workings, and sometimes non-workings, of the cosmos. Xiang stood on the threshold of the phenomenal and the abstract. The ambiguity and conflation of subject, object and medium in the simulacrum meant that, as xiang, it could be variously interpreted and applied. Xiang offered humans the capability to change the “real” thing into the model: in the case of the cosmograph rituals, the “real things” were time and space, and the battle array-as-cosmograph modeled it. The new geographic orientation of both battle array and the array-as-cosmograph patterned landscape as distinctly Chinese. These methods exorcized residual uncontrolled yin that resulted from death, the chaos of war, and alien culture, purifying and transforming conquered landscape. So far, so good for the court, since this is what warfare should do. At the same time, though, these rituals tapped yin power to achieve their result. The ability of ritual practitioners to access so much power, combined with amassed troops, worried the court. The court took harsh measures to keep them at bay. For instance, the court removed from command Di Qing (1008–1057), a general and diviner famous for quelling the south. They did so despite his consistent record of military successes because of his popularity with the troops, at least part of which revolved around his abilities as a diviner.63 The formation of collectivities within the troops alarmed the court. Desertion in large groups plagued the military system in the twelfth century: in one case thousands deserted together. Ritual predisposed troops to form such collectives, and the court saw occult ritual practitioners as the figures around whom troops collected. If, as I assert elsewhere, the most powerful forms of occult ritual transgress various bounds – and their performance at least transgressed tabus if not prohibitions – then the act of transgression inherent in the performance of these rituals 61 WJZY 18.2005–2007. 62 WJZY 21.2219. 63 XZZTJ 183.4426–27; Fang 2003: 53; Djang & Djang 1989: 306.


517

does two things: First, it forms, if only temporarily, a community of transgressors. Second, through transgression, participation in these rituals, even if unintentional, further vilifies the social and political position of the participants, who are then reluctant or unable to return to their pre-performance socio-political standing.

Conclusions The rituals described above demonstrate collective agency because the locale and region assert their abilities against a central authority, one that insisted on the legitimacy and sanction of text and literacy over informal and ad-hoc oral and vernacular culture. Though the record of occultists is spotty – sometimes leaving mere traces – one consistent theme is that the court perceived them as a threat, yet recognized that they were crucial for the dynasty’s success. Local or private networks tended to monopolize efficacious ritual performance; their lack of formality and intention threatened the court. The court responded by continually proscribing occult ritual practices and texts, instituting severe punishments, re-categorizing official ritual, creating government bureaus to appropriate these networks, and keeping careful watch on occultists who grew too popular. Occult ritual practices construct and distribute collective agency. In the case of occultists in China, ritual knowledge, texts, xiang, and the act of transmission itself all act as agents. These are agents in the sense that they have the ability and potential for transformative action, and because all are imbued with charismatic power that makes the collective possible. Transmission occurred in networks, often locally defined, and thus escaping central control. Ritual agency was transmitted from individuals to other individuals or to groups through training, apprenticeship or ritual participation; from transmission of objects (texts, diagrams, Chen Tuan’s “mantic mirror”, the ceremonial pike and axe, and everyday objects); and from recognizing and manipulating visual conceptualizations known as xiang that lie at the root of many occult arts. Networks grew from these methods of transmission, and collectively distributed agency whose source was often unrecognized. Such networks did not always result from a specific collective intent. Occult ritual practitioners devised and performed rituals that engaged and developed xiang: constellations, images, symbolizations, and simulacra. Xiang, with its new emphasis on the visual, was ideal for the Song military project. The visual quality of xiang reinforced the oral, performative nature of military field rituals. Being visual, xiang could communicate across the social and linguistic boundaries of the military collectivity, allowing a fluid interpretation of rituals and the formation of a non-literate canon. In the field, both battle array schema and cosmography rituals produced the power to overcome immediate military obstacles with visual and performative sources of agency.

518

M.A. Butler

The means through which specialists practiced ritual expose the fault lines of the Song court’s inability to wield power effectively in a few ways. First, the court’s efforts to eradicate or appropriate the networks through which occult knowledge was transmitted were ultimately unsuccessful. Second, the type of power invoked by court rituals (yang) was inherently different than that invoked by military field rituals (yin), a difference mandated by court policy and restructuring. Third, the ability of the diviner to effectively form social networks and collectivities rivaled that of the court. In military field rituals, agency is distributed and quite literally embodied in and through ritual, especially in battle array schema and cosmograph rituals. The cosmograph and battle array schema distributed collective agency among perhaps unknowing participants, articulating and re-asserting military authority vis-à-vis the court, the enemy, and an alien landscape. This is because its efficacy lies in its performance, in the body moving through space, and in a physicality that acts as primary initiator invoking nature to do human bidding. It is “embodied” in the sense of offering bodies in violent sacrifice, again defining one group vis-à-vis the other, and embodied by re-creating collectivity in that sacrifice.


519

References Primary Sources Untranslated primary source texts have been identified by the following abbreviations: HQJ Xu Dong 許洞 1936. Huqian jing 虎鈐經. Shanghai: Shangwu yinshuguan. JS Tuo Tuo 脫脫 1977. Jinshi 金史. Beijing: Zhonghua shuju. JWDS Xue Juzheng 薛居正 1976. Jiu Wudai shi 舊五代史. Beijing: Zhonghua shuju. LDBRZ Yuan Shushan 袁樹珊 1998. Zhongguo lidai buren zhuan 中國歷代卜人傳. Taibei: Xin wenfeng chuban gongsi. MXBT Shen Kuo 沈括 & Yang Jialuo 楊家駱 (ed.) 1961. Mengxi bitan jiaozheng. 夢溪筆談校證. Shijie shuju chuban. RZ Hong Mai 洪邁. 1991. Rongzhai suibi wuji 容齋隨筆五集. Jilin wenshi chubanshe. SS Tuo Tuo 脫脫 1977. Songshi 宋史, 20 vols. Beijing: Zhonghua shuju. Taibo Li Quan 李筌 1957. Taibo yinjing 太白陰經 [Shenji zhidi taibo yinjing 神極制敵太白陰經]. Taibei: Shijie bingxueshe (=Zhongue bingxue daxi 5). TPYL Li Fang. 李昉 1960. Taiping yulan 太平御覽.Beijing: Zhonghua Shuju. TSJC Chen Menglei 陳夢雷 1934. Qinding gujin tushu jicheng 欽定古今圖書集成. Shanghai: Zhonghua shu ju. TZ Zheng Qiao 鄭樵 1935. Tongzhi 通志. Wanyou wenku, di 2 ji 1935 edn. Shanghai: Shangwu yinshuguan (Shitong di 4 zhong series). WJZY Zeng Gongliang 曾公亮 & Ding Du 丁度 1987–. Wujing zongyao 武經總要. Zhongguo bingshu jicheng. Beijing: Jiefangjun chubanshe. WLTK Qin Huitian 秦蕙田 1970. Wuli tongkao 五禮通考. 20 vols. Taibei: Xingxing shuju. WXTK Ma Duanlin 馬端臨 1936. Wenxian tongkao. 文獻通考. Shanghai: Shangwu yinshu guan. XZZTJ Li Tao 李燾 1979. Xu Zizhi Tongjian 續資治通鑑長編. Beijing: Zhonghua shuju. YJZ Hong Mai 洪邁 & He Zhuo (ed.) 1981. Yijianzhi 夷堅志. Beijing: Zhonghua shuju. Yunji Zhang Junfang 張君房 1929.Yunji qiqian, Tai xuan bu雲笈七籤太玄部. Zhengtong daozang 38 vols. 正統道藏三十八冊.

520

M.A. Butler

Secondary Sources Ahearn, Laura M. 2001. “Language and Agency”. Annual Review of Anthropology 30: 109–137. Andersen, Poul 1989-90. “The Practice of Bugang”. Cahiers d’Extreme-Asie 5: 15–53. Berkowitz, Alan J. 1996. “Record of Occultists”. In: Donald S. Lopez (ed.). Religions of China in Practice. Princeton, Chichester: Princeton University Press: 466–468. Bourdieu, Pierre. 1990 [1980]. Logic of Practice. Standford: Stanford University Press. Brokaw, Cynthia J. 2005. “On the History of the Book in China”. In: Cynthia J. Brokaw & Kai-wing Chow (eds.): Printing and Book Culture in Late Imperial China. Berkeley, London: University of California Press: 3–54. Chen Feng 陳峰 2004. “Bei Song ‘Jiangmen’ xianxiang tanxi 北宋‘將門’現象探析.” In: 宋史论稿. Xian: Shanxi renmin chubanshe: 1–26. http://songdai.com/show. aspx?id=252&cid=6 (August 2009). Cherniack, Susan 1994. “Book Culture and Textual Transmission in Sung China”. Harvard Journal of Asiatic Studies, 54/1: 5–125. Cochini, Christian & Anna Seidel 1968. Chronique de la Dynastie des Sung (9601279). München: Universität München, Ostasiatisches Seminar. de Weerdt, Hilde 2006. “What did Su Che See in the North? Publishing Regulation, State Security and Political Culture in Song China?”. T’oung Pao 92: 466–494. Djang, Chu & Jane C. Djang (trans.) & Chuanjing Ding (comp.) 1989. Compilation of Anecdotes of Sung Personalities. Taibei: St. John’s University Press. Dudbridge, Glen 2000. Lost Books of Medieval China. London: British Library (The Panizzi Lectures, 1999). Fang, Cheng-hua 2003. “Military Families and the Southern Song Court: The Lu Case”. The Journal of Song-Yuan Studies 33: 49–70. Hall, David & Roger Ames 1995. Anticipating China. Albany: State University of New York Press. Harley, J. B. & David Woodward (eds.) 1994. History of Cartography, vol. 2, book 2, Cartography in the Traditional East and Southeast Asian Societies. Chicago: University of Chicago Press. Kohn, Livia 1990. “The Life of Chen Tuan After the History of the Song”. Taoist Resources 2.1: 1–7. — 1990. “Chen Tuan in History and Legend”. Taoist Resources 2.1: 8–33. Kreinath, Jens & Jan A. M. Snoek, Michael Stausberg (eds.) 2006. Theorizing Rituals. Leiden, Boston: Brill (Numen Book Series. Studies in the History of Religions, 114). Li Shen 李申 1998. “Hetu wu Mimi 河圖無秘密”. Guoji yixue yanjiu 國際易學研究 4: 231–254. Li Yangzheng 李養正 2008. “Daojiao gaishuo-di liu zhang 道教概說第六章”. In: Pengying xianguan daojiao wenhua zhongxin. http://www.daoinfo.org/?q=node/ 120 (July 2009). (hard copy: 1989 Beijing: Zhonghua shuju).


521

Li, Yuanguo. 1990. “Chen Tuan’s Concepts of the Great Ultimate”. Taoist Resources 2/1: 32–53. Lin, Li-chen. 1995. “Concepts of Time Position in the Book of Change and their Development”. In: Chin-Chieh Huang & Eric Zürcher (eds.): Time and Space in Chinese Culture. Leiden: E. J. Brill: 89–113. Louis, François 2003. “The Genesis of an Icon: The ‘Taiji’ Diagram’s Early History”. Harvard Journal of Asiatic Studies 63/1: 145–196. Luo, Kuan-Chung & C. H. Brewitt-Taylor (trans.) 1959. Romance of the Three Kingdoms, vol 1, 2. Tokyo: Charles E. Tuttle. Mauss, Marcel. & Brain, Robert (trans.) 1972. A General Theory of Magic. London: Routledge and Kegan Paul. McDermott, Joseph P. 2006. A Social History of the Chinese Book. Hong Kong: Hong Kong University Press. McKnight, Brian E. & James T. C. Liu (trans.) 1999. The Enlightened Judgments: Ch’ing-ming chi: the Sung dynasty collection. Albany: State University of New York Press. Milcinski, Maja 1999. “Zen and the Art of Death”. Journal of the History of Ideas 60/3: 385–397. Needham, Joseph 1959. Science and Civilisation in China, vol. 3, Mathematics and the Sciences of the Heavens and the Earth. Cambridge: Cambridge University Press. Poncé, Charles 1975. The Game of Wizards: Psyche, Science and Symbol in the Occult. Harmondsworth: Penguin Books. Porter, Deborah Lynn 1996. From Deluge to Discourse: Myth, History, and the Generation of Chinese Fiction. Albany: State University of New York Press. Rappaport, Roy 1979. “Obvious Aspects of Ritual”. In: Roy Rappaport: Ecology, Meaning and Religion. Richmond (Calif.): North Atlantic Books: 173–221. Robinet, Isabelle 1989. “Original Contributions of Neidan to Taoism and Chinese Thought”. In: Livia Kohn (ed.): Taoist Meditation and Longevity Techniques. Ann Arbor: University of Michigan, Center for Chinese Studies: 297–330. Sawyer, Ralph 1993. Seven Military Classics of Ancient China. Boulder: Westview Press. Sax, William 2006. “Agency”. In: Jens Kreinath, Jan A. M. Snoek, Michael Stausberg (eds.): Theorizing Rituals. Leiden, Boston: Brill (Numen Book Series. Studies in the History of Religions, 114): 473–481. Schafer, Edward 1977. Pacing the Void: Tang Approaches to the Stars. Berkeley: University of California Press. Sivin, Nathan 1980. “Shen Kua”. In: Charles Gillespie (ed.): Dictionary of Scientific Biography, vol. 11. New York: Charles Scribner’s Sons: 369–393. — 1986. “On the Limits of Empirical Knowledge in the Traditional Chinese Sciences.” In: Julius T. Fraser & Nathaniel Lawrence & Francis C. Haber (eds.): Time, Science and Society in China and the West. Amherst: University of Massachusetts Press (The Study of Time, 5): 151–169.

522

M.A. Butler

Smith, Kidder Jr. et al. (eds.) 1990. Sung Dynasty Uses of the I-ching. Princeton: Princeton University Press. Tjan, Tjoe Som (Zeng, Zhulin) (trans.) 1949, 1952. Po Hu T’ung: The Comprehensive Discussions in the White Tiger Hall, 2 vols. Leiden: E. J. Brill. Wang Bi & Richard John Lynn (trans.) 1994. The Classic of Changes: A New Translation of the I Ching as Interpreted by Wang Bi. New York: Columbia University Press. Wyatt, Don J. 2008. Battlefronts Real and Imagined: War, Border, and Identity in the Chinese Middle Period. New York: Palgrave Macmillan.

Paul S. Atkins

The Stages of Seppuku Performing Self-Execution in Premodern Japan Seppuku originated in medieval Japan as a gesture of defiance by a defeated warrior toward his victorious enemies.1 In some cases, after disemboweling himself with a sword, the seppuku performer would remove his own intestines and cast them toward the foe. He would then end his own life by other means, such as stabbing himself in the heart, or slitting his own throat. The goal was to avoid the humiliation of capture and execution. Beginning in the fifteenth century, seppuku was adopted as a ritual element in the punishment of errant members of the military class. It was not infrequently imposed during the early modern period (1600–1868) as an alternative to execution by decapitation, which entailed the forfeiture of one’s estate (escheat) and extinction of the family line. By submitting to the issuing authority, the seppuku performer provided for the financial and social survival of his heirs and the continuation of his lineage. In turn, the issuing authority was able to carry out the sentence with minimal resistance. Instead of committing suicide by other means, the seppuku performer was decapitated by a kaishaku, or second. That is, the ritual element of seppuku was now followed by an execution. Sometimes seppuku was performed for other reasons: a vassal might wish to follow his lord in death; a retainer might wish to remonstrate with his superior; an official might seek to take responsibility for a grave error. Seppuku was outlawed as a form of official punishment in the nineteenth century, but continued as a private practice. Notable examples include the seppuku of General Nogi Maresuke in 1912, multiple cases of seppuku performed by members of the imperial Japanese military forces during and after the Second World War, and the seppuku of the right-wing novelist Mishima Yukio in 1970. From the earliest encounters between Japan and the West in the sixteenth century, seppuku has remained one of the most enigmatic of the practices that are emblematic of traditional Japanese culture. In both senses of the word it is the ultimate performance and ritual. 1 Seppuku literally means “cutting the belly” and is synonymous with the term hara-kiri, which is regarded in some circles as inauthentic or vulgar. But in his article on how the practice is depicted in sixteenth century noh plays, Ikai Takamitsu notes that “harakiri,” not seppuku, is what the characters call it. See Ikai 2001: 16–38.

524

Paul S. Atkins

Seppuku is often conceived of as a method of ritual suicide, as in Ulrich Pauly’s study Seppuku: Ritueller Selbstmord in Japan.2 It was certainly performed ritually in different ways in different times. But, as Chiba Tokuji has shown, seppuku is not really a method of suicide. There are more effective and efficient methods of ending one’s life and, if the objective is to demonstrate one’s capacity to endure physical agony, then there are also more painful ways of doing so, including more painful ways of committing seppuku than the method typically used.3 We must stress that seppuku in itself was not a form of suicide in premodern Japan; it was a ritual element that was followed by suicide or execution.4 If suicide is not the objective of seppuku, then what is? Seppuku is a symbolic act. The earliest examples of seppuku involved the removal, display, and sometimes casting away of the performer’s intestines. After this the performer might slit his own throat, stab himself in the heart, or jump into a fire. What does the display of the intestines indicate? Does it show defiance, rage, courage, or, as some argued, moral purity? In order to answer these questions I will outline the differences between medieval and early modern seppuku, focusing on the practice of entrail casting; explore what the intestines signified in premodern Japanese culture; and consider what ritual failure reveals about seppuku.

Entrail Casting in Medieval Seppuku When using the term “medieval seppuku” I have in mind an act that is performed in the heat of battle, is largely voluntary, in many cases is gratuitously bloody, and has as its goal the avoidance of capture. This is a simplification but a useful one. Medieval seppuku was often followed by the removal and casting of the performer’s entrails. As an example, let us examine the following passage from Taiheiki (Annals of the Great Pacification), a military epic that deals with the wars between the Northern and Southern Courts in the fourteenth century: “Yoshiteru mounted up to the high tower at the second gate, where he watched until the prince had fled far away. When the time was right, he cut off the board from a window in the tower to show his person, naming his name with a mighty shout. ‘I am the Prince of the Blood Son’un of first rank, minister of military affairs and second son of Go-Daigo Tenno, the ninety-fifth mikado since Jimmu Tenno, the august scion of Amaterasu Ōmikami. Beaten down by rebels, I 2 Pauly 1995. 3 Chiba 1994: 25–40. 4 This insight was inspired by Axel Michaels’ “Grammars of Rituals” (Lecture, SFB 619 Ritualdynamik conference on Ritual Dynamics and the Science of Ritual, Heidelberg, September 29–October 2, 2008).

The Stages of Seppuku

525

destroy myself to avenge my grievances in the afterlife. Mark me well, that you may know how to rip open your bellies when fortune fails the military!’ Stripping away his armor, he cast it down from the tower. Next he took off the narrow-sleeved coat of his underdress, fashioned of a woven double thickness of raw and glossed silk; and clad only in brocade trousers, he pierced his fair white skin with a dagger. He cut in a straight line from left to right, flung out his bowels onto the board of the tower, thrust his sword in his mouth, and fell forward onto his face.”5 Like other medieval martial epics, the Taiheiki cannot to be taken at face value as a historical source. Rather, it shows us how seppuku may have been regarded at a certain point in time. This example may appear inappropriate at first glance, because Yoshiteru is actually impersonating a prince and performing seppuku so that the prince may evade capture thanks to Yoshiteru’s sacrifice. But it is actually highly instructive because it portrays Yoshiteru as conscious of seppuku as a performance. In such a circumstance he must perform seppuku in what the authors of Taiheiki considered its quintessential form. This entails bravado, a lingering attachment to warfare, physical courage and daring, and defiance toward one’s enemies. Casting of the intestines is a central part of the ritual; indeed, we may say it is the entire point. One does not cut one’s belly in order to die; one cuts one’s belly in order to remove and cast the intestines; once that is done, the seppuku performer ends his life by other means. There is another episode in the Taiheiki in which a warrior commits seppuku and pulls out his entrails. It is the example of Nagasaki Takashige (d. 1333): “He stripped off the stomach armor which alone remained on his body, caused his younger brother Shin’uemon to serve him wine in a cup that had been placed before the Lord of Sagami, drank three times, and put down the cup in front of the lay monk Dōjun of Settsu, saying, ‘To you do I give the cup, with this relish!’ And he cut his body with a long cut from left to right and fell down, pulling out his inwards in front of Dōjun.”6 Like Murakami’s seppuku, in this case the act is an explicitly self-conscious performance (Takashige’s grotesque jest). It also demonstrates bravado, through humor in the face of death. Yet, curiously, there is no enmity involved in the act of removing and displaying the entrails. To the contrary, it is an act of friendship and intimacy. The two men are sharing their last sips of sake together, and Takashige offers himself as the snack to accompany Dōjun’s drink. This anomaly forces us to rethink what the removal of intestines signifies.

5 McCullough 2003: 179. For the original, see Gotō & Kamada 1960: 214–15. 6 McCullough 2003: 310. For the original, see Gotō & Kamada 1960: 358.

Paul S. Atkins

526

Another valuable source is Gikeiki, a military epic completed in the fourteenth or fifteenth century that centers on the exploits of the daring general Minamoto no Yoshitsune (1159–89). In one scene, facing capture, Satō Tadanobu (1161–86) pulls out his entrails, but again for an unusual reason: “… he made a deep incision under his left arm and ripped open his belly from the pit of his stomach almost to the navel, drawing the blade steadily up toward his right side. When he had wiped the sword he inspected his handiwork. ‘An excellent blade! Mōfusa was as good as his word when he promised to make me a fine weapon. It cuts a man’s belly with no trouble at all. If I leave it behind, I suppose it will be sent to the east with my remains. Rather than allow the youths there to pass on its merits, I shall take it with me to hell.’ He wiped it again, sheathed it, and placed it under his knees. Grasping the mouth of his wound to force it apart, he plunged his fist into his belly, tossed his bowels onto the veranda, and shoved the hilt of the sword into the pit of his stomach, with the scabbard pointing down below his hipbone. ‘That’s how I treat the sword I take to hell,’ he said.7 Tadanobu also displays bravado (or perhaps simply sheer self-honesty) in his declaration that he is headed to hell after death. His performance of seppuku shows indifference to pain, but no sign of rancor.

Early Modern Seppuku In 1439 we find seppuku being used for the first time, as far as historical records indicate, as a punishment.8 It becomes, instead of an alternative to capture, torture, and summary execution, an alternative to capital punishment. Later, the kaishaku, or second, is introduced. During the early modern period an astonishing amount of lore comes to surround, or rather to create the seppuku ritual. Every aspect of the ceremony is regulated. Who should attend? Where should the act take place? What sort of furnishings are required? How should the sword be presented? When should the kaishaku administer the coup-de-grâce? All of these questions and many more are answered by the seppuku protocol manuals. Seppuku was adapted as a punishment as a kind of compromise, to bring about the death of a victim while allowing him the outward trappings of the honorable death of a defeated warrior. In order to accomplish it, however, it needed to be modified. One aspect of medieval seppuku practice that the early modern manuals 7 McCullough 1971: 205. For an edition of the original, see Kajihara 2000: 307–8. 8 The condemned were Ashikaga Mochiuji (1398–1439), who had recently served as the shogun’s deputy for eastern Japan, and a few of his close retainers. See Ōsumi 1973: 72–73.


527

attempted to erase was the removal, cutting, and casting of intestines. When we consider that the entire point of the medieval seppuku ritual was to cast the intestines in order to express strong, negative emotions (and thereby expel them), we realize that the redesigners of the seppuku ritual had a difficult task. Somehow they succeeded transforming seppuku into an illustration and achievement of “the acme and pitch of self-control,” to use Nitobe Inazō’s phrase.9 The judicial authorities had a number of methods at their disposal to control even the behavior of a condemned man. The first was the threat of decapitation (zanshu), which was a separate category of punishment. It was regarded as a heavier punishment than seppuku, less honorable, and could entail confiscation of the deceased’s estate and the extinction of his family line. The second was the kaishaku and, with it, the possibility of a “fan seppuku.” This is described in Jijinroku (Record of Self-Disembowelment), completed by Kudō Yukihiro in 1840: “If the prisoner be an unruly, violent man, a fan, instead of a dirk, should be placed upon the tray; and should he object to this, he should be told, in answer, that the substitution of the fan is an ancient custom. This may occur sometimes. It is said that once upon a time, in one of the palaces of the Daimios, a certain brave matron murdered a man, and having been allowed to die with all the honours of the hara-kiri, a fan was placed upon the tray, and her head was cut off. This may be considered right and proper. If the condemned man appears inclined to be turbulent, the seconds, without showing any sign of alarm, should hurry to his side, and, urging him to get ready, quickly cause him to make all his preparations with speed, and to sit down in his place; the chief second, then drawing his sword, should get ready to strike, and, ordering him to proceed as fast as possible with the ceremony of receiving the tray, should perform his duty without appearing to be afraid.”10 Other examples show that even when the seppuku performer is presented with a sword, the kaishaku always has the option of striking early if he believes it necessary. Although it is not mentioned in Jijinroku, a scenario in which the seppuku performer begins to remove his entrails could be curtailed by the action of the kaishaku. Kudō specifically addresses the practice of entrail casting: “Someone said, ‘Thinking it a point of pride to die as a bold man, one might make the incision large, pulling out the intestines and throwing them about and so forth, thereby shocking the spectators; to truly face one’s death, with 9 Nitobe 1969: 110. The sense of the word “pitch” as highest point, whether literally or figuratively, is one of its lesser known meanings. 10 Redesdale 1919: 280. Redesdale identifies his source simply as “a rare Japanese ms.” but a comparison reveals it to be excerpts from the Jijinroku. For this passage in an edition of the original Japanese text, see Inobe 1943: 305.

528

Paul S. Atkins unfaltering courage and undiminished strength is no mean feat. Yet even if a strong man acts in this way, it is actually inferior. Just as there are distinctions among officers and soldiers and among those of noble and common descent, in the way of performing seppuku as well, there must naturally be an order. It is probably appropriate for someone of low birth to behave in a flamboyant manner.’ I suppose this is so. But to face death with unfaltering courage and undiminished strength is what the man of valor hopes for; therefore, to the extent such a man can, I suppose it is also proper for him to behave rather flamboyantly. As mentioned earlier, in cutting into the belly, one must not exceed three or four bu [about 1 cm] in depth. It is said that if the cut is too deep, it will be difficult to cut across. One worries whether [the kaishaku] will botch the decapitation. If one is a very strong man, one will not be troubled by this, and the best thing to do is to cut 10 bu [about 3 cm] deep. It is said that protocol dictates that the short sword used for seppuku be wrapped with paper, exposing four bu of the blade. This is because in any event the kaishaku will do his work before the blade is used. How was it in the old days? I doubt one could cut 10 bu deep with so little of the tip. In this later age, we have gotten strong at investigating protocol and other matters, but our courage has grown weak.”11

One might expect an expert in protocol to express disapproval of the gory spectacle that is entrail casting but, surprisingly, that is not the case for Kudō. As a specialist in ancient lore, he is better aware than most of the contrast between the seppuku of old and that of his day. And the typical attitude toward the past in premodern Japan, with few exceptions, was that it was superior to the present age. A deep cut was regarded not as a point of pride, but as a punishment. In another seppuku manual, the anonymous and undated Seppuku kuketsu (Orally transmitted protocol for seppuku), preparation of the short sword is described. The sword is wrapped in cotton cloth and paper, exposing 2.5 bu. “But if the condemned’s crime 11 Inobe 1943: 310–1. The translation is mine, as the passage does not appear in Redesdale’s book. Immediately following this comment, the author of Jijinroku gives several examples of the flamboyant style of seppuku that were performed in the early Edo period: in one, a man committing seppuku makes three vertical incisions and one horizontal incisions to write the character yama 山, which stands for Yamashiro province, of which he was governor. In the next, another provincial governor kills an official at the palace in a dispute; after he escapes, another official attempts to capture him by grabbing him from behind. The governor manages to draw his short sword and drives it so deep into his own belly that it pierces his back and kills his would-be capturer as well. Another governor involved in the incident hears the news and attempts seppuku. He is restrained by his servants, who think he is mad and get him medical treatment. Later he goes to the toilet, puts his finger into the wound, draws out his intestine, and wraps it around his toe. Then he stomps down hard, killing himself at last. The final example is of a man who kills himself by burning his throat with a candle.


529

is a grave one, and others detest him, one may expose 5 or 6 bu.”12 Such passages illustrate how much the practice of seppuku changed over the course of half a millennium, while the name remained the same.

The Intestines in Premodern Japanese Culture Nitobe Inazō’s book Bushido (The way of the warrior) was first published in English in Philadelphia in 1905 and played a significant role in shaping the perception of Japan in the West, which was shocked by Japan’s victory in the Russo-Japanese War. It devotes an entire chapter to seppuku, in which Nitobe states that the Japanese believed that “in the abdomen was enshrined the soul … This view of mental physiology once admitted,” he continues, “the syllogism of seppuku is easy to construct. ‘I will open the seat of my soul and show you how it fares with it. See for yourself whether it is polluted or clean.’”13 In this way, Nitobe makes two assertions: first, that premodern Japanese sited the soul in the abdomen; and second, that seppuku is a symbolic opening of the abdomen to metaphorically expose the soul and permit direct examination of the performer’s character. Nitobe’s argument, which has been widely repeated, is problematic for a number of reasons. First of all, the exposure of intestines is a dispensable part of rituals that include seppuku as an element. In punitive seppuku, for example, such an gesture would indicate dissatisfaction with the sentence, and was scrupulously avoided. 14 Nitobe’s claim ahistorically conflates medieval evasion seppuku with early modern punitive seppuku. Of course, this confusion is nothing new. It was intended by the issuing authorities who co-opted seppuku as an element in their execution rituals; by performing seppuku before decapitation, the performer mimics the actions of a brave warrior being defeated in battle, rather than that of a common criminal, and is able to save face. Chiba Tokuji, the late geographer and anthropologist whose collected work forms the most comprehensive understanding of seppuku we currently possess, agrees with Nitobe that the purpose of displaying the intestines is emotional expression. Nevertheless, it is not an expression of innocence or moral purity but rather of rage, frustration, and anger (munen).15 Therefore, the intestines are displayed in evasion seppuku and kept concealed in punitive seppuku. Moreover, we should recognize that the seppuku performer is not exposing the space of the abdomen (hara) per se in the act of displaying his intestines. He is revealing its contents, that is to say, the intestines or bowels (harawata). To under12 13 14 15

Inobe 1943: 253. Nitobe 1969: 113–14. Chiba 1991: 167. Chiba 1972: 152–54.

530

Paul S. Atkins

stand what seppuku might mean if its original purpose was to expose the intestines, we must understand what they themselves signified in premodern Japan. The earliest extant account of a seppuku under the broadest definition is a legend about the goddess Ōmi set at the Harasaki Marsh in Harima Province, as described in Harima fudoki (Gazetteer of Harima). She was searching for her husband, the god Hananami, when she came to this site, and in a fit of rage, sliced open her belly with a sword and fell into the marsh.16 The text does not state that she exposed or removed her intestines but goes on to say that the fish in that marsh lack intestines. Chiba reads this in the context of similarly patterned myths and infers that the removal of her intestines is implied.17 That is to say, the removal, if not the display, of the intestines is present at seppuku’s putative origin. Harima fudoki is believed to have been generally complete not long after it was ordered by the central government in 713. 18 This is around the time the Kojiki (Record of Ancient Matters, 712) and the Nihon shoki (Chronicle of Japan, 720) were completed, so these texts may also be of help in reconstructing how the intestines were conceived in that time. Of the two, Nihon shoki (also known as Nihongi) appears to be the more fruitful. It includes the term harawata (or chō; translated by Aston as “bowels”) in two places. The first describes a woman whose husband has been sent on an expedition to Koryō in 493: “his wife Akitame, restless and full of longing, has lost her wits and become distraught, and the sound of her lamentation is very touching, even to the rending of men’s bowels.”19 The second occurs in what is presented as an imperial edict from 562 urging war against the Korean kingdom of Silla: “Heirs in their generation to the virtues of previous governments, and themselves destined to hold high dignity in a later reign if they cannot, by making drip their gall and drawing out their bowels, join with Us in slaying the traitors, thus wiping off this bitter outrage against Heaven and Earth, and doing vengeance on the enemies of a Lord and father, even in my grave I shall be indignant that the right rule of conduct of vassal and child has not been realized.”20 In both of these instances the image of the bowels is being filtered through stock Chinese expressions. In the second case, the bowels are paired with the gall 16 17 18 19

Aoki 1997: 228–29. For the original, see Akimoto 1958: 346–49. Chiba 1972: 181–82. Akimoto 1958: 27. Aston 1972: vol. 1; 396. For the original, see Sakamoto et al. 1965–67: vol.1; 530–31. Emphasis added. 20 Aston 1972: vol. 2; 81–82. Emphasis added. For the original, see Sakamoto et al. 1965–67: vol.2; 120–21. A note (p.120, n.1) says that the edict, including the phrase in question, is based on a passage from the Chinese dynastic history Liang Shu (629).


531

bladder, which symbolized courage. In the first, the very suggestive phrase “rending of men’s bowels” brings us closer to what seppuku might have symbolized in eighth-century Japan. This suggests that, as in the West, the bowels were considered the seat of compassion and pity. (As Aston observes in a note, the “bowels of compassion” is a Biblical phrase.21) In the original, the phrase corresponding to “rending of men’s bowels” is read in Japanese as danchō (lit., to sever the intestines). It has a long, rich linguistic history in China. The locus classicus is an episode contained in Shishuo xinyu (A New Account of Tales of the World), a collection of anecdotes compiled during the fifth century. In one story, one member of a party traveling by boat catches a baby gibbon. The monkey’s mother follows the boat by running along the river bank, keening all the way. At last, after traveling 100 li, the mother monkey manages to jump into the boat and dies. When the monkey’s belly is opened, her intestines appear to have been torn to shreds.22 These anecdotes are instructive in permitting us to see that the intestines are not the seat of the soul, or of simply any old emotions; they are the seat of some of the most intense feelings we have: pity, sorrow, and anger. Defining the intestines (harawata) as kokoro (heart, mind, soul)23 is a modern way of coping with the shift of the emotions and personality from various parts of the body (the gall bladder, the liver, the intestines) to the brain or, at best, the heart. In premodern Japan and China, as in the premodern West, the emotions and personal characteristics were dispersed and experienced throughout the body.

Ritual Failure and Seppuku In attempting to understand the transformation of seppuku as a ritual element, we would do well to consider it from the perspective of ritual failure.24 What did seppuku failure mean in the medieval period, and what did it mean in the early modern period? In the medieval period, failure in committing seppuku meant death at the hands of a high-status opponent at best. At worst, it led to capture, humiliation, torture, and decapitation by a low-status opponent. The costs were social and psychic. We 21 1 John 3:17: “But whoso hath this world’s good, and seeth his brother have need, and shutteth up his bowels of compassion from him, how dwelleth the love of God in him?” (King James Version). 22 See Liu 2002: 485. For an early reference to danchō by a Japanese author, see the preface to Man’yōshū 793 by Ōtomo no Tabito (664–730): “I am constantly filled with sadness enough to rend the heart, and alone I weep gut-wrenching tears.” Levy 1981: vol.1; 343. For the original, see Satake et al. 1999: vol.1; 443–44. 23 See, for example, the third definition of harawata in Nihon kokugo daijiten dainihan henshū iinkai 2000–2002: vol. 10; 1422a. 24 This approach is inspired by Hüsken 2007.

532

Paul S. Atkins

can find examples of failure of medieval seppuku as early as the Kakuichi version of the Tale of the Heike, the thirteenth-century epic that depicts the Genpei War (1180–85). The general Kiso Yoshinaka, for example, fails to end his own life despite the entreaties of his trusted retainer Kanehira (who succeeds brilliantly himself).25 Although Kiso’s adversaries, the Taira generals, preferred other methods of suicide to seppuku, such as drowning, there are examples of them too failing to take their own lives and suffering capture, humiliation, and execution.26 In the early modern period, the “director” of the seppuku ritual was no longer the principal but rather the custodian (o-azukarinin), a person into whose custody the condemned was remanded by a higher-ranking authority. Failure for the custodian meant first of all, escape of the principal. Short of that, he feared that the condemned would grab a sword from the kaishaku or another person in hopes of dying in battle, and inflict injury or death. And even if that did not occur, the custodian was concerned about the botching of the decapitation by the kaishaku. In some cases, the custodian lacked a suitable swordsman among his retainers and would have to hire an outsider, making him “retainer for a day” in order to complete the ceremony successfully. The custodian’s social capital was at stake, as was the kaishaku’s, especially if he was an insider. In the Jijinroku, we learn that protocol dictated that the kaishaku’s cut not completely sever the head, so as to maintain the distinction, however improbably, between seppuku and simple decapitation. But this could lead to a failure to sever the head completely, so Kudō advises prospective kaishaku to “strike a fair blow.” In case the blow was not clean enough, the condemned might even attempt to stand up, in which case it was the responsibility of the assistant kaishaku to kill him. Although these are very different scenarios, they share a preoccupation with maintaining the symbolic capital of the samurai class. Indeed, seppuku itself serves to preserve a distinction between commoners and members of the samurai class through the method of execution. This is outlined clearly in the opening lines of Jijinroku: “Seppuku is the mode of suicide adopted amongst Samurai when they have no alternative but to die.”27

Conclusions Seppuku is neither a ritual nor a form of suicide. It is a ritual element that has been combined, at various times, with acts of suicide or execution.

25 See McCullough 1988: 293. 26 Shigehira prepares to commit seppuku but surrenders instead; Munemori lacks the nerve to drown himself. See McCullough 1988: 315 and 379, respectively. 27 Redesdale 1919: 263.


533

The assertion that in premodern Japan the abdomen was regarded as the seat of the soul appears to be of modern vintage and cannot be substantiated by textual evidence. Our understanding of the symbolic aspect of seppuku should instead focus on the intestines, which were often displayed by the performer in early instances of seppuku. There are long-standing associations between the intestines and specific emotions of grief, anguish, and resentment. In the medieval period, the purpose of seppuku was to express and relieve these powerful feelings. In the early modern period, its purpose was to evoke one part of the medieval version (the courage of the performer) while repressing the other (extreme emotion). Therefore entrail casting and even deep cuts were avoided. Over centuries, seppuku was gradually reduced to a pale imitation of its original, at times an empty gesture. Early modern Japanese were already aware of this transition by the mid-nineteenth century, if not earlier.

534

Paul S. Atkins

References Akimoto Kichirō (ed.) 1958. Fudoki. Tokyo: Iwanami Shoten. Aoki, Michiko Y. (transl.) 1997. Records of Wind and Earth: A Translation of “Fudoki” with Introduction and Commentaries. Ann Arbor: Association for Asian Studies. Aston, William George (transl.) 1972 [1896]. Nihongi: Chronicles of Japan from the Earliest Times to A.D. 697, repr. Rutland: C.E. Tuttle Co. Chiba Tokuji 1972. Seppuku no hanashi. Tokyo: Kōdansha. — 1991. Tatakai no genzō: minzoku toshite no bushidō. Tokyo: Heibonsha. — 1994. Nihonjin wa naze seppuku suru no ka. Tokyo: Tōkyōdō Shuppan. Gotō Tanji & Kamada Kisaburō (eds.) 1960. Taiheiki, vol.1. Tokyo: Iwanami Shoten. Hüsken, Ute (ed.) 2007. When Rituals Go Wrong: Mistakes, Failure, and the Dynamics of Ritual. Leiden: Brill. Ikai Takamitsu 2001. “‘Nishikido’ to ‘Kiyoshige’: harakiri to eiyū no nō.” Gakugekigaku 8: 16–38. Inobe Shigeo (ed.) 1943. Bushidō zensho, vol. 10. Tokyo: Jidaisha. Kajihara Masaaki (ed.) 2000. Gikeiki. Tokyo: Shōgakukan. Levy, Ian Hideo (transl.) 1981. The Ten Thousand Leaves: A Translation of the “Man’yōshū”, Japan’s Premier Anthology of Classical Poetry. Princeton: Princeton University Press. Liu I–ch’ing 2002. Shih-shuo Hsin-yü: A New Account of Tales of the World, trans. Richard B. Mather. 2nd ed. Ann Arbor: Center for Chinese Studies, University of Michigan. McCullough, Helen Craig (transl.) 1971. Yoshitsune: A Fifteenth-Century Japanese Chronicle. Stanford: Stanford University Press. — (transl.) 1988. The Tale of the Heike. Stanford: Stanford University Press. — (transl.) 2003 [1959]. The Taiheiki: A Chronicle of Medieval Japan, repr. Boston: Tuttle Publishing. Nihon kokugo daijiten dainihan henshū iinkai (ed.) 2000–02. Nihon kokugo daijiten, 2nd edition. 13 vols. Tokyo: Shōgakukan. Nitobe, Inazo 1969 [1905]. Bushido: The Soul of Japan; An Exposition of Japanese Thought. Rutland: C.E. Tuttle Co. Ōsumi Miyoshi 1973. Seppuku no rekishi. Tokyo: Yūzankaku. Pauly, Ulrich 1995. Seppuku: Ritueller Selbstmord in Japan. Tokyo: Deutsche Gesellschaft für Natur- und Völkerkunde Ostasiens. Redesdale, Algernon Bertram Freeman-Mitford 1919. Tales of Old Japan. London: Macmillan and Co. Sakamoto Tarō et al. (eds.) 1965–1967. Nihon shoki. Tokyo: Iwanami Shoten. Satake Akihiro et al. (eds.) 1999. Man’yōshū. Tokyo: Iwanami Shoten.

Poul Andersen

The Life of Images in Daoist Ritual There is a way in which all rituals may be characterized as “acts of enlivenment.”∗ A powerful example is that of the Christian Eucharist, in which the minister dispenses the central objects of bread and wine to the members of the congregation, saying (in the name of Jesus Christ): “This is my body; this is my blood.” The whole process may be viewed as a way of ensuring that these statements are true. Indeed, according to the narrative of Luke’s Gospel, the ceremony was instituted by Christ himself, who commanded his followers to continue doing this “in remembrance” of him.1 Lutherans believe that it is exactly because of these “Words of Institution,” repeated in the Eucharist, that the body and blood of Christ are truly and substantially present “in and under the bread and wine.”2 And yet, modern Christian theologians (Protestants as well as Catholics) take great pains to argue that, in fact, these statements are not to be construed as literally true, that in the end the bread and wine are not “sacred” in themselves, but rather symbols in a para-linguistic act, and that the real presence of God in the event (which many continue to believe in) is of a spiritual nature and quite independent of the actual form of the symbols. According to the Catholic liturgical theologian, Louis-Marie Chauvet, the minimalistic quality of the symbols indicate, among other things, that their specific form carries no value, that God’s “gift of grace” is gratuitous, and that he is “sovereignly free” in his response to them.3 In other words, when we consider the thematic core of Christian theology, around which debates in the church have turned throughout its history, and from which a multitude of new denominations have continued to be generated, we find ourselves in the thick of persistent ambiguity, and confronting the universal attitude about both symbols and images that has been aptly characterized by W. J. T.

∗ This article is a further development of one of the trains of thought that constituted my paper for the Heidelberg conference on “Ritual Dynamics and the Science of Ritual.” As such, it was presented at the symposium on “The Chinese Art of Enlivenment” at Harvard University a few weeks later, on October 24, 2008. 1 Luke 22: 19b-20. See also Billings 2006. 2 Martin Luther, The Large Catechism, Part 5 (1529), McCain, ed. 2007: 432. 3 Chauvet 2001: xv.

536

Poul Andersen

Mitchell as “double consciousness.” In his book from 2005, What Do Pictures Want?, he states as follows:4 “How is it […] that people are able to maintain a ‘double consciousness’ toward images, pictures, and representations in a variety of media, vacillating between magical beliefs and skeptical doubts, naïve animism and hardheaded materialism, mystical and critical attitudes? […] Let me put my cards on the table at the outset. I believe that magical attitudes toward images are just as powerful in the modern world as they were in the so-called ages of faith. I also believe that the ages of faith were a bit more skeptical than we give them credit for. My argument here is that the double consciousness about images is a deep and abiding feature of human responses to representation. It is not something that we ‘get over’ when we grow up, become modern, or acquire critical consciousness. […] Why does the link between images and living things seem so inevitable and necessary, at the same time that it almost invariably arouses a kind of disbelief: ‘Do you really believe that images want things?’ My answer is, no, I don’t believe it. But we cannot ignore that human beings (including myself) insist on talking and behaving as if they did believe it, and that is what I mean by the ‘double consciousness’ surrounding images.” What I like in particular about this account is the emphasis on the fact that the so-called double consciousness about images is far from representing a modern development, but rather constitutes “a deep and abiding feature of human responses to representation.” I shall begin this article by demonstrating that this holds true also for early Daoist reflections on images. However, it is noteworthy that in the above quotation (which comes at the end of a section in the introduction to the book), Mitchell concludes by raising a question that he doesn’t answer, and by suggesting that the proof of our belief in images as “living things” is found in the ways in which we “insist on talking and behaving” in relation to them. I agree entirely with the methodology implied in this finally sentence, which indicates that, in order to answer our questions concerning the ways in which people relate to images, we must 1) pay close attention to what they say about it (both orally and in writing), and 2) deduce what is implied about these attitudes by the practices in which they engage. It should be noted, however, that the knowledge that may be derived from these two sources speaks not only to people’s belief, but equally to their disbelief concerning the status of images as living things, i.e., that the sources at our disposal, rather than settling our doubts in this regard, typically tend to confirm the state of ambiguity and “double consciousness” about images mentioned by Mitchell. 4 Mitchell 2005: 7–11.

The Life of Images in Daoist Ritual

537

The extent to which images are treated in Daoism as “living presences” or “objects of worship” (or, indeed, as “mere decoration”) clearly varies historically, regionally, and among different categories of participants in ritual – not to mention among different types of images that these participants relate to. Moreover, we have no strong reason to believe that any given person always experiences the same images the same way. In other words, in order to approach a better understanding of these issues, we need to pay careful attention to individual cases and attempt to analyze each one of them in terms of the two methodological principles mentioned above. For the present purpose I shall give an example of this, based on my fieldwork in Taiwan in the late seventies and mid eighties, when I studied with Daoist High Priest Chen Rongsheng 陳榮盛 of Tainan – and occasionally participated in his ritual performances as an assistant priest. I shall rely, not only on direct communications from Chen Rongsheng and other priests, but also on the conclusions that may be drawn from close observation of his performances in relation to the texts that he uses. It will not be suggested that my conclusions apply universally to all forms of Daoism, at all times and in all places. Even so, in order to facilitate the debate, I shall frame the discussion by referring to some claims about the matter that have been offered by other scholars, some of whom have relied on materials from a different, though related, context. But first, let us consider the philosophy of images that may be found in Daoist texts from the early Tang dynasty (618–907), a period that clearly was formative regarding the widespread use of images within Daoism, both in the form of statues, murals, and hanging scrolls. While in the Six Dynasties the attitude toward images expressed in Daoist texts had often been rather hostile and marked by the rejection of the use of external images (as opposed to mental representations, i.e., visualizations) of the gods, the early Tang dynasty witnessed a large-scale development in this area, partly as a result of the impact of Buddhism, and along with the accommodation of many other doctrinal, and philosophical aspects of Buddhism within Daoism. A prominent (though far from unique) example of this is the encyclopedia Daojiao yishu 道教義樞, by Meng Anpai 孟安排 (fl. 699), which opens with the following classical statement, based on quotations from the Daode jing, chapters 21 and 25:5 “The Way, Dao, is perfectly empty and perfectly silent, it is very real (甚真) and very wondrous; and yet, its emptiness penetrates everywhere, its silence responds to everything. Thereupon there is the Celestial Worthy of the Primordial Beginning, Yuanshi tianzun 元始天尊, who responds to the breath and gives form to an image (應氣成象); from silence comes motion, from reality comes response. He emerges from the edge of chaos, in the middle of obscure darkness; he embraces primordial harmony and transforms yin and 5 Daozang 1129, pref.:1a.

538

Poul Andersen yang. Thus it is that the Daojing of Lord Lao (老君道經) says: ‘In the obscure darkness there is essence, in the chaotic confusion there is an image’ (窈冥中有精, 恍惚中有象). It also says: ‘There was something (a being) formed out of chaos; it was born before heaven and earth. How silent! How still! It stands alone and does not change. It moves all around and is not endangered. It may be considered the mother of the universe.’ This probably clarifies that Yuanshi tianzun in the middle of chaos responds to the breath and gives form to an image (應氣成象); thus there is a being formed out of chaos.”

The text continues with an account of the activities of Yuanshi tianzun at the beginning of each world period, through which the gods in heaven are saved, and through which the promise of salvation is transmitted to later generations. It is made clear that “the body of the Celestial Worthy lasts forever without being extinguished” (天尊之體常存不滅), but that in the present kalpa the task of saving the many living beings has been delegated to Taishang daojun 太上道君, the second of the Three Pure Ones, who embodies the Scriptures (and thus the foundation of the Lingbao tradition). Having explained this to those around him, Yuanshi tianzun continues as follows:6 “For this reason I now ascend to the mystery and enter the wondrous realm. You people with eyes of the flesh will not be able to see my true and real body (汝等肉眼不能見我真實之身); meaning that it has been utterly extinguished. Only if you cultivate the correct vision, will you be able to see me of your own, just like at this time. If someone is not yet able to clearly discern the empty marks (空相), he still can rely on images in order to attach his heart/mind to me (猶憑圖像, 係錄其心). You must melt purple gold and make a copy of my true marks (寫我真相); you must worship and make offerings to it, as if you were facing my true form (如對真形); if you apply your thoughts and pray sincerely, the merit will be of equal measure. People who are poor may make variegated images of clay, wood, or bronze according to their resources; setting up halls with curtained seats, offerings of banners and flowers, lamps and candles, according to their means, as if they were serving my true body (如事真身). Receiving the karma of these actions, they will in the end return to the supreme Way (上道).” A very similar set of ideas is found in the Dongxuan lingbao sandong fengdao kejie yingshi 洞玄靈寶三洞奉道科戒營始, another encyclopedic work of the early Tang dynasty. Chapter 2 of this work opens with a section on the production of

6 Daozang 1129, pref.:3a; quoting the now lost Tang dynasty text, Duren benji jing 度人本際經.


539

images, Zaoxiang pin 造像品, providing one of a very few (if not the only) systematic account of such practices in the Daozang. It opens as follows:7 “The great image has no form, perfect reality has no color. Deep and clear, it is empty and silent. Looking and listening do not gain access to it, and yet it transforms in response and makes its body appear (應變見身). Momentarily manifest, it returns to being hidden. That by means of which one may retain the true and real (存真) is attaching one’s thoughts to the saintly countenance (係想聖容). Thus, through painting and using gold and jade one may copy the marks of its form and make an image of the true and real countenance (像彼真容), adorning and enriching it with white lead. Whenever one seeks to attach one’s heart/mind, one always must first produce an image.” The passage continues providing a large number of details concerning the materials, sizes, and numbers of statues to be produced, as well as information concerning their arrangement and the offerings to be presented to them. It concludes:8 “Day and night one should visualize them as if one was facing the true form (如對真形). People of the past and the future all attain blessings without measure and are able to accomplish the true Way (克成真道).” It is clear from all of this that the supreme god, as well as all other gods, have a “true and real form” (真形), and that the images produced by human beings are not to be confused with this true form. Instead, they stand as gateways to the divine, and people address them “as if” or rather, “like they were facing the real form.” In a way that is parallel to the first appearance of Yuanshi tianzun at the beginning of time, as the first “image” or “figure,” xiang 象, in the universe, they represent a step down from the ultimate reality of the god. The true reality of the god is the underlying emptiness, to which he “returns to be hidden” (還隱), and which may be addressed through the images produced by mankind. And yet, history abounds with claims about actual images that directly identify them with the true form of the god in question. See, for instance, the following story about a Buddha statue in a temple in Guangxi reported in the Qixiu xugao 七修續稿 by Lang Ying 郎瑛 (1487–c. 1566):9 “The Buddha in the Temple of the Buddha of Immeasurable Longevity (無量壽佛寺) in Quanzhou 全州 in Guangxi is exactly the true body of (the Buddha of) Immeasurable Longevity (無量真身). Its eye balls and finger nails are no different from those of living human beings. Its eyes gaze at things and are able to turn. Its finger nails grow longer over the years …” 7 Daozang 1125.2.1a. 8 Daozang 1125.2.2b. 9 Lang Ying 郎瑛 1961: 841.

540

Poul Andersen

With all its reputation for being entrenched in magic, Daoism (or at least the classical liturgical kind represented by the Zhengyi/Lingbao priests with whom I have studied) actually tends to be a little more cautious with these matters. Thus, when Chen Rongsheng, in one of the major rituals of the jiao 醮, Offering, liturgy, plants the five world-creating talismans, the so-called Wufang zhenwen 五方真文, of the Lingbao tradition, in the five corners of the sacred area, he does this, not by actually drawing the talismans, but in each case by facing a basin filled with rice in which a pad of paper and the appropriate writing utensils have been stuck, chanting a prayer to the first Celestial Master, Zhang Daoling 張道陵 (the founder of his ritual tradition), and asking the latter to fill in the writs. While being present, the writs thus remain invisible, that is, zhen, true and real, in the sense this term was used in the above quotations, also in relation to the true body of Yuanshi tianzun, who was, indeed, the first to propagate these salutary writs in the universe.10 As for the images of the deities, hung by Chen Rongsheng at the boundaries of the sacred area, and representing the totality of the Daoist pantheon, it is noteworthy that none of these have been animated through the performance of an “eye-opening” ceremony, kaiguang 開光. While other figures used in his liturgy are so treated for the occasion, notably, the cruder paper figurines that are disposed of through burning at the end of the ceremony, such procedures are never applied to the treasure of ritual scrolls (some of which are very old, and certainly not intended for destruction) used by Chen Rongsheng as visual representations of the gods during his performances. It should be noted that in other parts of China, and in other Daoist traditions, the animation of the same kind of ritual scrolls through eye-opening ceremonies, is absolutely of the essence, and considered indispensable for their efficacy.11 It is, in other words, a variable, and not a practice to be taken for granted in all of Daoism. With this in mind, let us proceed with a closer examination of the status of these ritual scrolls in the Daoism of Chen Rongsheng. The first to do fieldwork on the Chen family tradition in Tainan was Kristofer Schipper, who made contact with Chen Rongsheng’s father, Chen Weng 陳聬 (1890–1975), already in 1963. In an article entitled “An Outline of Taoist Ritual”

10 See, for instance, the Daijiao yishu, quoted above, in which is stated that in the chaos of the infinitesimal primordial kalpas, in which there was no form or shadow (無形無影), Yuanshi tianzun fixated the true writs and promulgated their five “chapters” (i.e., sections) (元始安鎭, 敷落五篇); Daozang 1129, pref.:1b. 11 This in general is the case for the Daoist ritual scrolls used among the Yao of Southwest China and Southeast Asia. For an overview of Yao religion and religious artifacts, see Pourret 2002. Note also that the Daoist scrolls collected by Professor Li Yuanguo 李遠國 in the western province of Sichuan (now included in the Sichuan Museum of Primordial Daoism 四川原道博物館 in Chengdu) in many cases show traces of the chicken blood and feathers used in eye-opening ceremonies.


541

(originally presented at a conference on Asian rituals and the theory of ritual in Berlin in 1984), Schipper states as follows:12 “Overall interpretations [of ritual] have been time and again expressed in stories, apologues, theatrical rituals and ritual theater, altar decorations, et cetera, that surround the liturgical tradition as so many explanatory additions or spin-offs. Altar paintings representing gods, hells, and idyllic landscapes are quite unnecessary. They are often seen because they are very decorative. They illustrate that the liturgy is an audience at the heavenly court or a banquet of immortals.” In so doing, he clearly subordinates the role of images in relation to the actual performances of Daoist ritual. In his book from 1982, The Taoist Body, he expresses this basic stance concerning the status of images in Daoist ritual by stating as follows about the scroll representing the Golden Gate, Jinque 金闕, which is hung at the center of the sacred area behind the central Cave Table:13 “In the middle of the space, right behind the central table, a painted scroll is hung, the only one that is not merely decorative and that has a real function in the ritual.” The attitude of John Lagerwey, whose book from 1987, Taoist Ritual in Chinese Society and History, gives an overview of Chen Rongsheng’s ritual tradition, 14 clearly is quite different. It is prefigured by a suggestion offered by Schipper to the effect that “We also see the sphere of the altar as something continuing outward from the master’s own body. Moving ‘towards the outside’, he projects from his body transcendent forces which radiate and surround him in concentric circles, in an infinite number of finite spaces.”15 While Schipper carefully avoids identifying any of these projections with the concrete visual representations of the gods in the sacred area (in general, these images are, after all, to be considered as “mere decoration”), this is exactly the move made by Lagerwey. In a sustained analysis of the contents of Chen Rongsheng’s sacred area, Lagerwey identifies some of the key images found in the ritual scrolls as representations of specific parts of the body of the high priest and concludes that “the domain beyond the Metal Gate in the sacred area [i.e., the Golden Portal referred to also by Schipper] is at the top of the head of the Daoist master, 12 13 14 15

Blondeau & Schipper: 117. Schipper 1993: 87. The Golden Gate appears as the character 闕 in Figure 5. Lagerwey 1987; see also Lagerwey 1991. Schipper 1993: 99.

542

Poul Andersen and the totality of the area is the expression in images of his body: the ritual which will proceed in front of the Grotto Table [translated by me as Cave Table] will take place in his grotto body.”16

There are many problems with this interpretation of the images of the sacred area. Firstly, it has no basis in indigenous interpretations of these images found in Daoist texts or offered by the priests or any other informants, but depends entirely on the author’s understanding of the significance of the inner practices of the high priest in relation to the external proceedings of ritual. Secondly, it relies heavily on very doubtful identifications of some figures represented in the ritual scrolls with specific deities that are involved in the inner practices of the high priest, notably, the identification of Red Robe, Zhuyi 朱衣, one of the guardians of the sacred area (Image 1), with the alter ego of the high priest (likewise dressed in a red robe), who emerges out of the heart of the priest at the beginning of his inner, meditational journey to heaven during the performance of great Audience rituals. The guardian Red Robe, together with his counterpart Jinjia 金甲, Golden Armor (whose scroll is hung opposite to Red Robe in the other side of the sacred area of a jiao), both have a pre-history as assistants to the god of literature, Wenchang 文昌, who during the Tang dynasty developed out of the cult of the constellation by this name, consisting of six stars in Ursa Major.17 They are the constant companions of Wenchang, who, according to Henry Doré, in paintings generally are represented standing on each side of the god (Image 2).18 Indeed, in one of the collective scrolls discussed below, used by Chen Rongsheng to represent the various divine administrations of the universe, namely, the scroll representing the Prefecture of Heaven, Tianfu 天府, the two gods appear precisely as members of the entourage of Wenchang (Image 3).19 As such, the appearance of the two gods is closely similar to their representations as individual guardians in Chen Rongsheng’s sacred area. The official’s cap worn by Red Robe is of the same kind in all cases, and in all cases he holds a book, which appears to symbolize his connection with the god of literature. As an individual guardian of Chen Rongsheng’s sacred area, he further holds a banner decorated with a constellation (consisting of seven stars, but probably referring to the literary asterism), and he is mounted on a composite mythical animal. None of these features pertain to the Red-Robed Perfected, Zhuyi zhenren 朱衣真人, who emerges from the heart of the high priest at the beginning of his inner journey to heaven. The latter figure is identified in Chen 16 17 18 19

Lagerwey 1991: 142 (translated from the French by the author of this article). See Kleeman 1994: 46–51. Doré 1920: 67. Wenchang is the figure on horseback at the bottom of the painting turning his head toward Zhuyi and Jinjia. The scroll was painted by the famous Pan Lishui 潘麗水 (1914–95), who was from Tainan, and whose temple murals are preserved not only in many temples in southern Taiwan, but also in the Baoan gong 保安宮 in Taipei.


543

Rongsheng’s secret manual as the priest himself (自身是也),20 and in an illustration found in a manual deriving from the Zeng family of Tainan he is, indeed, portrayed in the image of a Daoist priest holding a court tablet, chaoban 朝版 (Image 4).21 In Chen Rongsheng’s secret manual, this Red-Robed Perfected and alter ego of the high priest is repeatedly identified with the Great Lord of Long Life, Changsheng dajun 長生大君, one of the supreme gods of the inner pantheon of the human body. In Tang dynasty commentaries on the Duren jing, the Great Lord of Long Life is said to be none other than the Great One, Taiyi 太一, the ancient high god of Daoism, who in the inner world of the human body is thought to reside in the heart, and who in Daoist meditation techniques throughout history has been identified with the inner, immortal persona of the Daoist adept.22 The connection between this image and the guardian Red Robe thus is limited to the color of their dress and to the fact that, on his return to the human body, the Great Lord of Long Life allegedly has been transformed into “the appearance of a venerable elder.”23 Lagerwey bases this claim on the account found in Schipper’s The Taoist Body,24 and on the description in Chen Rongsheng’s secret manual. However, this secret manual mentions no such transformation, but states quite simply that, on his return to the sacred area (via the inner stages of his own body), “the Great Lord of Long Life, who is the priest himself, stands at the center, while the marshals and generals guard the boundaries of the sacred area as before” (自己長生大君立在中央, 帥將依舊護衛壇界). It may be added that Lagerwey’s interpretation of the guardians Red Robe and Golden Armor presents them, not only as images of the inner organs of the high priest (associated with the heart and the lungs, respectively), but also as acolytes of the two main patriarchs of the ritual traditions of Daoism, Zhang Daoling and Xuantian shangdi 玄天上帝, who represent the civil and the military techniques of Daoist ritual, respectively. This interpretation depends heavily on the exact placement of the two guardians in the corners of the sacred area that sometimes are associated with the directions south and west, and next to the patriarchs. However, the placement of the two scrolls varies, and they are sometimes found adjacent to 20 Longhu shan zhengyi liuhou Zhang tianshi yujue quanji 龍虎山正乙留候張天師玉訣全集, “Complete Collection of Precious Instructions Left Behind by Celestial Master Zhang of Orthodox Unity at Mt. Longhu” (no pagination). 21 Daozang biyao 道藏秘要, Secret Essentials of the Daoist Canon (no pagination). The original manuscript in the hand of Zeng Yanjiao 曾演教 (1818–66) was acquired by Kristofer Schipper. The illustration shown in Figure 4 is from the fountain-pen copy of this manuscript that he left behind in the field, of which I obtained a copy from Zeng Chunshou 曾春壽 (1913–92), the last of the Zeng family lineage of Daoist priests in Tainan. 22 See Andersen 2005: 26–29. 23 See Lagerwey 1987: 132. 24 Schipper 1993: 99.

544

Poul Andersen

the Three Worlds Table at the entrance door to the temple, that is, in positions as “quasi-door gods.”25 It should be mentioned, finally, that generic images referred to as “red-robed ones,” Zhuyi, occur quite commonly in Daoist hagiographical literature, for instance, in the Lidai shenxian tongjian 歷代神仙通鑑 compiled by Xu Dao 徐道 (1712). They are typically portrayed there as subordinate clerks of the underworld or messengers carrying the edicts of the Emperor on High, Di 帝, and they are often said to hold a book or register in their arms.26 It seems obvious that the Red Robe who serves as a guardian of the Daoist sacred area descends from this kind of figure, rather than from the gods of the inner pantheon addressed in Daoist visualization techniques. This latter point brings up an important underlying issue concerning the determination of the appearances of the gods in this kind of religious art. I wholeheartedly agree with the following remarks by Paul Katz about the murals of the famous Yongle gong 永樂宮, Palace of Eternal Joy, in southern Shanxi, which date to the first half of the fourteenth century. Katz emphasizes that the content and appearance of the murals of this temple were determined through the interaction of at least two different groups of people, namely, “the Taoist patrons who sponsored the murals and the local artisans who painted them. The former were members of the Perfect Realization movement and had supervised the palace since its rise to prominence during the thirteenth century. They were responsible for hiring the artisans to paint the murals as well as for choosing the texts to which the artisans were supposed to adhere. However, the actual planning and execution of the murals (including perhaps even the cartouches) lay in the hands of the artisans, who worked within a centuries-old tradition of religious painting.”27 Concerning a related set of traditions, Katz states that, “[a]ccording to mainland Chinese art historians Bo Songnian and Wang Shucun, artisans who made popular prints often watched dramatic performances as a source of inspiration for their work,” and later that “the beliefs artisans held were usually closer to the local culture they lived in than the world of Taoist priests.”28 I believe that something similar may be said about the Daoist ritual scrolls used by Chen Rongsheng, and that this weighs heavily against the interpretations of the contents of the scrolls proposed by John Lagerwey. However this may be, the question of the status of images in Daoist ritual (whether they should be viewed as “mere decoration,” as symbols referring to realities that essentially are somewhere else, or as living presences) clearly is not a 25 26 27 28

See also Li Fengmao 李豐楙 1998: 146. See, for instance, Xu Dao 1975, vol. 6: 2357 and 2442. Katz 1999: 134. Ibid.: 140–141.


545

matter of objective fact, but a question of how these images are experienced by those who participate in ritual. In order to get a better understanding of this, let us have a closer look at some of the ritual practices in Chen Rongsheng’s sacred area, and in particular, at some of the inner practices carried out by the high priest, as he faces some of these images. (As you will recall, the reference to such inner practices was a crucial part of Lagerwey’s argument concerning the status of the images). Images 5 and 6 show the rite of “entering the door”, Ruhu 入戶, during the performance of the Morning Audience, Zaochao 早朝, on the third day of a five-day jiao in the Jintang dian 金唐殿 in Jiali, Taiwan, in February 1987. The Ruhu is the initial rite of entrance included in the beginning of all major rituals, during which the whole group of priests moves around the sacred area and offers three sticks of incense in each of the four directions, while singing the hymn called “Appellation of the Jade Sovereign”, Yuhuang hao 玉皇號. In Chen Rongsheng’s tradition, the groups of Officials on the eastern and the western walls normally are represented by a set of six ritual scrolls, three of which are hung on each of these walls (Image 7).29 These scrolls clearly are among the most valued in his collection, not the least because he considers them to date from the Ming dynasty. On some occasions, however, he will (as here) use a set of scrolls, which derives from another iconographical tradition of representing the groups of Officials in the sacred area. It is the one that divides them into four departments called the Four Prefectures, Sifu 四府, which include the Prefectures of Heaven, Earth, and Water together with the World of the Living, Yangjian 陽間.30 Indeed, the instructions for the inner practices carried out by the high priest during this rite of entrance, found in his personal secret manual, clearly prescribe that during this rite he should inwardly address the Four Prefectures. While facing the tables in the east and in the west, behind which the scrolls representing the officials are hung, he inwardly pronounces an incantation to the Four Prefectures, Sifu, commanding that the “golden lads and jade 29 The scroll reproduced here is normally placed at the center of the group on the eastern wall. It shows the Officials of Heaven, Tianguan 天官, who are turned toward the supreme Daoist gods placed against the northern wall, and who illustrate the theme of subordinate deities “having an audience with the Prime,” chaoyuan 朝元 (see below). 30 In Figure 6 the priests are seen facing the eastern wall and the scrolls representing the Prefecture of Heaven to the left (see Figure 3), with the Prefecture of Earth, Difu 地府, next to it. The figure to the far right is Zhang Daoling, the first Celestial Master and the main patriarch of the civil traditions of Daoist ritual. To his right, the guardians of the sacred area are represented by Marshal Zhao, Zhao yuanshuai 趙元帥. This is where the scroll representing Zhuyi will normally be hung, with two other guardians next to it, but space being limited at the event in Jiali, the group of guardians was reduced to just two scrolls. The counterpart displayed on the western wall was Marshal Kang, Kang yuanshuai 康元帥. These two marshals represent the minimal set of guardians, and their placement normally is the reverse of what was the case in Jiali.

546

Poul Andersen

lasses”, jintong yunü 金童玉女 (who attend to the incense), will convey the sincerity of his heart to the Three Heavens above (Image 8).31 Chen Rongsheng will follow these precise instructions regardless of the way in which the officals are represented on the walls, that is, no matter which of the two very different iconographic traditions is adopted for the occasion. Moreover, as is clear from the wording of the text, his attention during the performance of this inner practice is directed toward the heavens above rather than toward the images in front of him. I find it hard to reconcile these facts with the notion that at this point the high priest might be said to be engaged in the “worship of images” in the sense of an immediate address of the gods as present in their depictions in the scrolls that constitute the sacred area. Many other aspects of Chen Rongsheng’s practice count against this notion, notably, the many discrepancies between the representations of the gods in the form of ritual “statements”, zhuangwen 狀文, and their portrayals in the ritual scrolls. The statements are ritual documents that are placed in yellow envelopes behind the altar tables in the sacred area, serving as invitations to the gods. They mark the seats of the deities, shenwei 神位, invited for the event, and their structure and content clearly are more closely correlated with the actual practice prescribed in the ritual manuals than are the representations of the gods in the ritual scrolls. It is interesting to notice that the exact same rite as the one discussed above is also the one referred to by Paul Katz in his discussion of the ritual functions of the murals of the Yongle gong, as part of his argument to the effect that during performances of ritual in the temple, the images in these murals were treated as “objects of worship”. He quotes the Song dynasty ritual compendium Wushang huanglu dazhai licheng yi 無上黃籙大齋立成儀, which states about this circumambulation of the altar and presentation of incense that the priests “next make one round of the jiao altar and offer incense standing in the proper order” (次旋醮壇行香一周依班序立).32 However, in his rendering of this phrase, Katz omits the last character, li, thus arriving at the interpretation that the master of rites is instructed to “circulate around the altar, making offerings of incense in the proper order.”33 He continues by stating that this indicates that “there was a prescribed sequence according to which Taoist priests were required to make offerings to the deities 31 Figure 8 shows the two pages in the Chen family secret manual which give instructions for these inner incantations. They are copied here from a manuscript in the hand of Chen Tinghong 陳廷鋐 (1838–1908), a grandson of 陳紅 (1789–1829), who was from a family of Daoist priests in present-day Longxi 龍溪 County in Zhangzhou 漳州, Fujian, and who crossed over to Taiwan toward the end of the Qianlong 乾隆 era (i.e., around 1795). The title of the manuscript is Longhu shan laozu zhengyi tianshi Zhang zhenren yujue 龍虎山老祖正乙天師張真人玉訣, and the text is identical to that of Chen Rongsheng’s personal secret manual. 32 Daozang 508:37.19b. 33 Katz 1999: 147.


547

worshiped during ritual”, and he offers this as evidence in support of the idea that the works of art on the walls of the Yongle gong “were not merely decorative but were also the objects of offerings during Taoist rituals performed at the Palace of eternal Joy.” The problem with this is that, while the priests obviously present incense to the gods in a certain ranking order, the circumambulation of the altar as described in the Wushang huanglu dazhai licheng yi and performed still today, is carried out in a straightforward clockwise manner, which in no way corresponds to the ranking order of the gods as represented on the walls. Indeed, as appears from the reproductions of the murals in the Yongle gong, the arrangement of the gods on the walls of the temple is much more complicated than that (Image 9).34 The composition of the murals on these walls clearly confirms the dominant role of the “audience” or “procession metaphor” and contradicts the possibility that the ranking order of the deities on such walls might be traced through a simple cir35 cular movement. It may be concluded that the notion of Daoist images being treated as “objects of worship” is not borne out by the materials discussed in this article – all depending, of course, on what one means by this term. In fact, I rather symphathize with the reinterpretation offered by Craig Clunas, in which he summarizes Paul Katz’ suggestions in the following way:36 “Much early religious wall painting was expected not to be looked at, but to do something. As Paul Katz has shown with regard to one of the best-preserved and consequently best-studied of such programs, … murals in a Daoist or Buddhist temple were not there for decoration nor simply to inspire devotion on the part of the worshipper. Rather they were themselves an essential part of the rituals performed in those spaces. It was their iconography, not their style or the fame of the artist who painted them, that was from the point of view of religious practice absolutely central to their function. They were pictures that were for something, pictures placed on permanent view in public spaces and carrying shared public meanings. It was this that was to make them increasingly problematic in the eyes of the élite by at the very latest the seventeenth century, even if the piety of that élite was undimmed.”

34 Figure 9 shows the central segment of the mural on the eastern wall of the Yongle gong, which portrays two of the rulers of the universe, apparently Dong wanggong 東王公 and Xi wangwu 西王母, surrounded by a host of subordinate deities, and in the process of having an audience with the supreme gods of Daoism, that is, the Three Pure Ones, Sanqing 三清. See Jin Weinuo 金維諾 1997. 35 As indicated above, the same metaphor applies to the arrangement of ritual scrolls in Chen Rongsheng’s sacred area. Historically, temple murals precede these ritual scrolls, and there is a way in which the latter may be thought of as “movable murals.” 36 Clunas 1997: 28–29.

548

Poul Andersen

Following Clunas, we may conclude that the images were an essential part of the rituals, that they were there for something, and that they carried shared public meanings. Whether they were also treated as divine presences, however, is a different matter entirely.


Image 1: Red Robe, Zhuyi 朱衣, painted by Su Qingyuan 蘇慶元, Jiayi, Taiwan, c. 1976. Source: Collection of Chen Rongsheng, Tainan.

549

550

Poul Andersen

Image 2: Red Robe and Golden Armor, Jinjia 金甲.

Source: Henry Doré, S.J. Researches into Chinese Superstitions, vol. 6 (Shanghai: T’usewei Printing Press, 1920), p. 67.


Image 3: Prefecture of Heaven, Tianfu 天府, painted by Pan Lishui 潘麗水 (1914-95), Tainan. Source: Collection of Chen Rongsheng, Tainan.

551

Image 4: The Red-Robed Perfected, Zhuyi zhenren 朱衣真人, who emerges from the heart of the high priest at the beginning of his inner journey to heaven.

Source: Daozang biyao 道藏秘要, Secret Essentials of the Daoist Canon (no pagination). Collection of Zeng Chunshou 曾春壽 (1913-92), Tainan.

552

Poul Andersen

Images 5–6: The rite of “entering the door,” Ruhu 入戶, during the performance of the Morning Audience, Zaochao 早朝, on the third day of a five-day jiao in Jintang dian 金唐殿, Jiali, Taiwan, February 1987. Photos by the author.


Image 7: Officials of Heaven, Tianguan 天官 (Ming dynasty). Source: Collection of Chen Rongsheng, Tainan.

553

554

Poul Andersen

Image 8: Two pages from the Chen family secret manual, Longhu shan laozu zhengyi tianshi Zhang zhenren yujue 龍虎山老祖正乙天師張真人玉訣, as copied by Chen Tinghong 陳廷鋐 (1838-1908), Tainan.

Image 9: Segment of the eastern wall of the Palace of Eternal Joy, Yongle gong 永樂宮, north of the Yellow River in southern Shanxi province (first half of the fourteenth century).

Source: Jin Weinuo 金維諾, ed., Yongle gong bihua quanji 永樂宮壁畫全集 (Tianjin: Renmin meishu chubanshe, 1997).


555

References Andersen, Poul 2005. “Scriptural Traditions West and East: Foundation of Belief versus Frameworks for the Transmission of Methods”. In: Poul Andersen & Florian C. Reiter (eds.). Scriptures, Schools and Forms of Practice in Daoism: A Berlin Symposium. Wiesbaden: Harrassowitz Verlag, pp. 13–32. Billings, Bradley S. 2006. Do This in Remembrance of Me: The Disputed Words in the Lukan Institution Narrative (Luke 22:19b-20). An Historico-Exegetical, Theological, and Sociological Analysis. London: T&T Clark Publishers. Blondeau, Anne-Marie & Kristoffer Schipper (eds.) 1995. Essais sur le rituel III. Louvain-Paris: Peeters. Chauvet, Louis-Marie 2001. The Sacraments. The Word of God at the Mercy of the Body. Collegeville, Minn.: The Liturgical Press Clunas, Craig 1997. Pictures and Visuality in Early Modern China. Princeton: Princeton University Press. Doré, Henry, S.J. 1920. Researches into Chinese Superstitions. Vol. 6. Shanghai: T’usewei Printing Press. Jin Weinuo 金維諾 (ed.) 1997. Yongle gong bihua quanji 永樂宮壁畫全集. Tianjin: Renmin meishu chubanshe. Katz, Paul R. 1999. Images of the Immortal: The Cult of Lü Dongbin at the Palace of Eternal Joy. Honolulu: University of Hawaii Press. Kleeman, Terry 1994. A God’s Own Tale: The Book of Transformations of Wenchang, the Divine Lord of Zitong. Albany: SUNY Press. Lagerwey, John 1987. Taoist Ritual in Chinese Society and History. New York: Macmillan. — 1991. Le continent des esprits: La Chine dans le miroir du taoïsme. Bruxelles: La Renaissance du Livre. Lang Ying 郎瑛 1961 [1959]. Qixiu leigao 七修類稿. Shanghai: Zhonghua shuju. Li Fengmao 李豐楙 et al. 1998. Donggang Donglong gong jiaozhi – dingchou nian jiuzhao qingcheng xie’en shuihuo qi’an qingjiao 東港東隆宮醮志 – 丁丑年九朝慶成謝恩水火祈安清醮. Taibei: Taiwan xuesheng shuju. McCain, Paul Timothy (ed.) 2007. Concordia: The Lutheran Confessions. A Reader's Edition of the Book of Concord, 2nd ed. St. Louis: Concordia Publishing House. Mitchell, W. J. T. 2005. What Do Pictures Want? Chicago: The University of Chicago Press. Pourret, Jes G. 2002. The Yao: The Mien and Mun Yao in China, Vietnam, Laos and Thailand. Chicago: Art Media Resources. Schipper, Kristoffer 1993. The Taoist Body. Berkeley: University of California Press. Xu Dao 徐道 1975. Shenxian jian 神仙鑑. Taibei: Guangwen shuju.

Benedetta Lomi

The Iconography of Ritual Images, Texts and Beliefs in the Batō Kannon Fire Offering Ritual and religious practices emerging in response to the needs of worshippers or to cultural and local-specific factors are often labelled “folk” or “popular,” as opposed to “canonical”. Such a strict polarity, often supported by a division between what is based on texts and what is not, fails to address the subject of ritual practice appropriately. This calls for a reconsideration of what canonical is and what defines a canonical practice. Is canonical determined by what we find in the canon, the sacred texts and their commentaries, or is canonical, in the long run, determined also by other elements, such as beliefs, customs, historical circumstances and experience? If this was the case, religion should be considered as composed of different factors, the two most important of which would be the beliefs of a community and the “structure” or format give by a system working as an organising and interpretative principle.1 An excellent example for this investigation is provided by the goma (fire offering) for the protection of horses offered to Batō Kannon 馬頭観音2 and held in Kashōin 華正院, a Jimon Tendai temple linked to Honzan Shugenshū 本山修験宗 ,3 in the town of Miharu 三春, Fukushima prefecture. The goma ritual of Kashōin is not considered a canonical practice mainly because it is not based on any Batō Kannon ritual text,4 and is performed in response to the need of the community. Batō, originally a wrathful horse-headed deity of the esoteric Buddhist pantheon, has been associated with the protection of animals since its introduction to Japan as the form of Kannon responsible for the chikushōdō 畜生道.5 Nevertheless, besides its peculiar iconography, the connection between the worship of Batō

1 Shinno & Swanson 1993, from the concluding chapter of Shinno 1991: 269–89. 2 Batō 馬頭, in Japanese means horse-headed, and refers to the main iconographical trait of this manifestation of Kannon, bearing small horse’s head between his hair. 3 Honzan is a branch of Shugendō connected with Tendai. For further details see Miyake & Earhart 2001: 11–36 and Miyake 1999. 4 Soeda 1981: 9. 5 Within the six realms of rebirth, or rokudō 六道, this is the animal realm.

558

Benedetta Lomi

Kannon and horses is considered only a Tokugawa period popularisation of the deity’s functions. In the present analysis, however, I argue that this feature has deep roots in the Buddhist scriptures, and that its interpretation and development within the Japanese context was triggered by specific social and economic circumstances. In these terms, the interactions between local customs and “established religion”, between folk and canonical, are not understood in terms of opposition, but in terms of mutual relation, providing a fuller understanding of the religious practices connected to a single deity or to a specific area. In addition, the present analysis will acknowledge the possible coexistence of different if not contrasting understandings of a given ritual, through an investigation of the role of the practitioner and of the believers in relation to the practice and its object of veneration.

The Batō Kannon goma of Miharu Every year, on 17 April, the city of Miharu dedicates a festival to the Batō Kannon of Kashōin Temple. The statue is enshrined in the Batō Kannon Hall 馬頭觀音堂 (Image 1), which, according to the legend, was originally built in 810 C.E. (Kōnin 弘仁 1) as a treasure hall to commemorate the favourite horse (aiba 愛馬) of general Sakanoue Tamuramaro 坂上田村麻呂 (758–811), and rebuilt in 1592 as part of the Kashōin temple complex.6 Anyway, the historical records date the existing structure to 1653 (Keyan 慶安 5), and a later restoration was carried out in 1725 (Kyōhō 享保 10).7 The honzon of the hall, Batō Kannon, is not visible to the worshippers, and is a hibutsu 秘仏 (a hidden sacred image) dating to the late sixteenth century, of which no photographic reproduction exists. Nevertheless, its presence is remembered by the stele at the temple entrance (Image 2), by a cemetery of meritorious horses (Image 3), and by the numerous ema 絵馬 dedicated to the beloved horse of Tamuramaro hanging outside the hall (Image 4).8

6 This is based on information gathered at the temple. It is more likely that the hall was built at the beginning of the Tokugawa period, as suggested by the historical records, and the dating to Kōnin 1, which is frequent in many other temples of the area, was suggested in order to create continuity between the myth of Tamuramaro and the existence of Batō Kannon. 7 Miharu-machi 1975–1986: vol.1, 334. Furthermore, both the temple and the hall are present in Miharu City maps dating to the beginning of the eighteenth century; Miharu-machi 1975– 1986: vol.1, 332–333. 8 Some late Tokugawa period examples of these ema are even stored at the Miharu Folk Museum.

The Iconography of Ritual

Image 1: Batō Kannon Hall, City of Miharu, Fukushima Prefecture Photo by the author

559

560

Benedetta Lomi

Image 2: Batō Kannon Stele Photo by the author


561

Image 3: Steles bearing the inscription “Batō Kanzeon” and “Aiba” dedicated to deceased horses Photo by the author

562

Benedetta Lomi

Image 4: Ema tablets of the Batō Kannon Hall Photo by the author

The original date of the ritual, according to the historical records, was the seventeenth day of the third lunar month and was part of a larger matsuri including celebrations held at Kitano Shrine for the deity Tenjin 天神.9 The joint celebration of Batō and Tenjin reflected the role of the two deities as protectors of horses and cows respectively.10 Therefore, during the matsuri (festival), villagers would pray for the prosperity of their cattle, in a period of the year, the ides of spring, when mares and cows were blessed for a safe and prosperous pregnancy. This matsuri has been an important moment for the life of the city, throughout the late Tokugawa and Meiji period, for the importance breeding cattle, and in particular horses, had for the economy of the Tamura area.11

9 Miharu-machi 1975–1986: vol.4, 295. The Kitano matsuri is now held on the fourth Sunday of April, usually around the 25th, so that it is still quite close to the Batō Kannon celebration, but since the breeding cattle do not play such an important role any more in the everyday life of the people of Miharu, or do not have an economic impact on the city, the two festivities are now separated. 10 Miharu-machi 1975–1986: vol.4, 295. 11 Fukushima-ken 1964–1972: vol.8, 1171.


563

Originally, the preparations for the festival would begin a week in advance, reaching their apex on the night before the actual day of worship, with night-long prayers held in the Batō Kannon hall. In the morning, the service would begin with people gathering on the premises and offering mochi (sticky rice cakes) to the deity and goma sticks to burn in the fire ritual. The owner of horses that were expecting the birth of a new pony, or whose mare had just given birth, would then offer a particular bag of mochi tied with a multicoloured string, as a prayer for successful breeding of their herd. The historical sources do not go into any specific detail as far as the ritual practice is concerned but, according to the monk in charge of Kashōin temple, the ritual has been transmitted for generations and is a regular goma for the aversion of disasters, or sokusai goma 息災護摩. The goma starts at 1.00 p.m., but most of the people gather in the temple main hall much earlier, if they have not spent the whole morning, some of them even the night, helping prepare the function. At the prescribed time, three monks, all of them shugenja 修験者, (Image 5), walks the steep stairs that connects the Batō hall with the main temple, accompanied by the sound of the horagai (the conch shell blown by the Yamabushi, the Shugendō practitioners). They first chant the Heart of Wisdom sūtra (Hannya shingyō 般若心経)12 in front of the Batō Kannon stele placed just outside the hall (Image 6) and then begin the goma.13 The ritual held on this occasion is a Fudō Myōo sokusai goma 不動明王息災護 14 摩, a fire offering performed for the prevention of calamities and the exorcising of disasters, whose chief deity is Fudō Myōo. Therefore, the deity to whom the ritual is offered and the one who is invoked during the ritual are different. This is not an unusual circumstance, especially considering that the framework in which the ritual developed and is still carried out is entwined with esoteric Buddhist thought, empowerment through asceticism and local customs. Therefore, in order to understand its meaning, the kind of ritual we are dealing with, and address the issues of canonicity set forth in the introduction, we should reconstruct the historical and cultural circumstances in which the practice was created.

12 The full title being 摩訶般若波羅蜜多心経, T.8 V.251. 13 There are different types of goma that vary according to branches of Buddhism. The Kashōin sokusai goma is a uchigoma 内護摩, internal fire ritual. 14 The sokusai goma is a common esoteric ritual performed in a variety of circumstances, from esoteric initiation rituals, such as the Juhachi-dō 十八堂, or important state-protecting rituals such as the Latter Seven-day ritual, where the objective is to purify and empower the ajari, to rituals to prevent disasters performed by monks and shugenja, both independently or as part of annual ascetic practices.

564

Benedetta Lomi

Image 5: Monks preparing for the Ritual Photo by the author


Image 6: Batō Kannon stele with Offering Photo by the author

565

566

Benedetta Lomi

The Horse in Fukushima From the historical sources it is clear that the Batō Kannon hall represents the most important worship hall of the Kashōin complex and that it began to gain prominence in the late seventeenth century. The reason for the increase in popularity of the hall and of its honzon is because of two distinct but not completely unrelated factors. First, it is linked with the importance of horse breeding in present-day Fukushima; since the eighth century, and surely throughout the medieval period, Fukushima has been strategically important for the breeding of horses.15 At the time, the control of Mutsu 陸奥 and Dewa 出羽 provinces16 meant the constant flow and trade of excellent horses and horsemen to the capital, providing a good, strong, economic and military asset. Moreover, from the sixteenth century onwards, there has been a stable implementation in the already well-established horse-breeding sector, constituting a profitable source of income. This, in turn, may have led to an increase in ritual practices connected to the protection and blessing of horses.17 The Bodhisattva Batō Kannon is very popular in the Tohōku area, where steles dedicated to this deity are found outside temples and along travelling routes. The greatest amount of steles is found along those roads connecting towns and villages that benefited from the horse trade, and which relied on horses as a means of transportation and work.18 Second, Kashōin temple is station number three on the Thirty-three Kannon Pilgrimage of Tamura (田村三十三観音霊場), the township comprising Miharu and surrounding areas. The pilgrimage was instituted in 1692 by Takudō 琢道, the Chief Priest of Fukushūji, 福聚寺, station number two on the pilgrimage. The establishment of the route in the late 17th century reflects the popularity of the original Saikoku Kannon Pilgrimage, replicas of which were instituted all over Japan,19 15 Mention of the importance of the horses of Mutsu province is already found in historical and literary records of the Heian period, such as the Shin Sarugaku-ki (A New Account of Sarugaku. c.1052) composed by Fujiwara no Akihira. Further references can be found in McCullough 1964–1965: 178–211. 16 Mutsu and Dewa refer to an area that roughly covers the present-day prefectures of Miyagi, Iwate, Aomori and Fukushima. In the historical sources, this area is also referred to as Michinooku no Kuni 陸奥国 or Ōshū 奥州. 17 In spite of the long-standing horse-breeding tradition of the Mutsu region, fine horses were bred also in the Musashi region and in Kyushu. It was a shift in production policies of the former that stimulated the rapid emergence of Tohōku as a horse-breeding centre in the Tokugawa period. 18 Fukushima is not the only region where Batō Kannon is widely worshipped through dedicatory and memorial steles, for a brief and clear analysis of each Tohōku region see Hayami 2000: 126–159. For an analysis of Kiso in Nagano region, another important area for the breeding of horses where Batō Kannon was highly venerated, see Kyburz 1987. 19 MacWilliams 1997: 376.


567

and in the area of Fukushima there are several other Thirty-three Kannon routes, all inaugurated in the early Tokugawa period.20 The fact that in the original Saikoku route only one station is dedicated to Batō Kannon,21 while in the case of the Tamura pilgrimage five out of the Thirty-three Temples enshrine the horse-headed deity,22 is surely indicative of the importance ritual practices related to horses had in this area.

The Favourite Horse (aiba 愛馬) and the Precious Horse (houma 寶馬) The Batō Kannon of Kashōin was, according to a local legend, installed to commemorate the favourite horse of General Sakanoue Tamuramaro, Seii Taishōgun 征夷大将軍23 of the Mutsu and Dewa provinces and leader of the campaign against the Isawa Emishi people of northern Hōnshu. Legends concerning Tamuramaro proliferate in the area, so it comes as no surprise that numerous places are named after him (for example Miharu is part of the Tamura Township) and that several temples and shrines are dedicated to this figure and his horse. As far as Miharu is concerned, Tamuramaro is remembered not only at Kashōin temple, and in the Tamura Tengen shrine 田村天元神社, where his spirit is enshrined, but also in the folk legend concerning the origin of the Play Horses, called kosodate koma 子育駒, wooden sculptures of horses, still sold nowadays for the protection of children and good rearing of horses.24 The legend of the origins of the kosodate koma is related to that of the founding of the Batō Kannon Hall. It is told that upon leaving Kyōto for the military campaign in Hōnshu, Tamuramaro was given by a virtuous Buddhist monk a box containing a small wooden Buddha image, as a good-luck charm. The box was never to be opened by Tamuramaro, unless in time of peril and need. After several days of restless battle, the horses of Tamuramaro’s army were tired and the enemies, led by the strong Mutsu lord Ōota Kimaro, were beginning to gain ground. Tamuramaro, remembering the words of the monk, opened the box, out of which onehundred horses magically appeared. Once the battle was won, they turned back into

20 Even if the Tamura pilgrimage reflects the popularisation of the Saikoku pilgrimage it was not shaped after it, as in many other cases, there is no direct correspondence of the various forms of Kannon featured in each station. 21 Matsuo-dera 松尾寺 (station 29). 22 Besides Kashōin, there are Hayamasan Yakushiji 早馬山薬師寺, Dasihōji 大昌寺, Senkōji 東光寺, Ryukoji 龍光寺. All of these temples were built at the end of the sixteenth century, even if legends would date them back to the beginnings of the ninth century. 23 This military title was in the Nara and Heian conferred on those generals who took part in the campaign to conquer the Emishi. 24 Kyburz 1994: 1–28.

568

Benedetta Lomi

wooden figurines and Tamuramaro donated the box to the city of Miharu as a protecting gift. These horses are now remembered as the kosodate koma. At this point, the question we need to raise concerning Batō Kannon is the following: are we dealing with an autochthonous and local specific deity, whom by having an iconographical resemblance to the esoteric deity Batō Kannon, became associated with it, or is there any continuity with the Batō Kannon, described in the Buddhist canonical sources of Kamakura period?25 Enshrining the favourite horse of a famous military leader as Batō Kannon is not limited to the figure of Tamuramaro alone, but it is rather a common practice in north-eastern Japan. A similar example is provided by the records of Nikkō, describing the foundation myth of the temples and shrines of its mountain range. The initial part of the Nikkōsan engi 日光山縁起26 tells the story of General Ariu (有宇中将), who is considered the incarnation, or Suijiaku, of one of Nikkō’s mountains, Nantai-san. According to the legend, the faithful horse of the Lieutenant, who is also the incarnation of the Lieutenant’s son Uma-ō 馬王, the King of Horses, was later enshrined in the Hongū shrine 本宮, and associated with the Batō Kannon of Rinnōji Temple.27 The oldest written copy of this engi, now preserved in the storehouse of Hie Shrine of Noguchi, is dated 1532 (Kyōroku 5), but Japanese scholars believe the original version must have been of the Kamakura period (1185–1333 CE), probably transmitted orally by monks.28 The date is also corroborated by the presence of a Kamakura-period scroll, representing the deities of Nikkō, the Scroll of the three avatars of Nikkō 日光三所権現像, explaining such an association. The legend of Tamuramaro in connection with the Batō Kannon hall of Miharu and that of General Ariu in connection with the one in Rinnō-ji, are only two of the various examples of Batō Kannon being the deified horse or aiba of a military leader. From my current investigation of the dissemination of Batō Kannon images and related cults in Tōhoku, it has emerged that this is a very popular and well-known Bodhisattva, consistently represented and worshipped.29 Moreover, the diversity of legends concerning the Batō and its presence in specific locations all seem to have a common thread, which has its origin in the Buddhist scriptures: the association of 25 For example, in analysing the Batō Kannon faith of Kiso, Kyburz made clear that the popularisation of this figure was a consequence of the compassionate nature of the widely worshipped Kannon and the existence of kamis protecting animals. 26 Hagiwara 1975: 276–89. 27 General Ariu finished his life on Mount Fudaraku, and he was there enshrined in the Shingū shrine 新宮, his wife in Takinoo 瀧尾, and his horse in Hongū 本宮. The three kami are believed to embody the three mountains of Nikkō, and are associated with the three Buddhist deities of Rinnō-ji. 28 Hoshino 1960: 9. 29 For a further survey of the role of Batō Kannon in the Kiso and Fukushima area see Kyburz 1987.


569

Batō with the precious horse or aśvaratna. The origin of this association, absent in the earliest dhāraṇī text on the deity, is to be found instead in the commentary of the Dainichi-kyō (大日経疏),30 which explains that Batō Kannon is like the aśvaratna of the Chakravartin.31 The term aśvaratna means “precious horse” and is one of the Seven Jewels of the Wheel-Turning King (Chakravartin). This interpretation of Batō Kannon is subsequently found in many canonical sources on the deity, so much so that it can be considered to be its epithet. Furthermore, in the thirteenth-century ritual manual Byakuhō-shō 白寶鈔, composed by Chōen,32 it is recorded that Batō Kannon is Kantaka, the horse of Prince Siddhartha. Batō Kannon possesses all the characteristics of the Prince’s horse: a compassionate and loyal nature, the ability of moving rapidly in all directions and a neigh so powerful it removes all hindrances.33 Following the description of Batō found in this and other Japanese ritual collections, such as the Besson Zakki and the Kakuzen-shō, the interpretation of Batō as aśvaratna, originally made by Yi-xing, became one of the most prominent features of the deity. This is clearly connected, on the one hand, to its equine characteristics and, on the other, to the compassionate nature of Kannon. The association with Kantaka, the mount of Shakyamuni, is therefore straightforward in this context, and helps to strengthen the symbolic association between Batō and a brave steed in a number of tales connected to this deity. Furthermore, the idea of the precious horse represents, in this context, a likely antecedent to the definition of aiba, the favourite horse. In both cases we have the figure of a ruler or a general, and also that of a meritorious and beloved horse. Even if this practice were a Tokugawa-period adaptation of the deity’s function, its continuity with the Buddhist interpretation of Batō as aśvaratna provides extremely important evidence of consistency within the development of this deity. The association of Batō Kannon with Kantaka represented a crucial step in shaping the practice of associating Batō with a beloved horse, and created an important precedent to the Tokugawa-period custom of erecting a Batō Kannon stele along the roads, at the place where one’s horse had died, to commemorate and thank the horse for its loyalty and for its good work. Surely such customs must be further linked to the long-standing tradition of ritualising deceased animals (both wild and

30 T. XXXIX, 1796, full title Daibirushana Jōbutsukyō-so 大毘盧遮那成佛經疏 composed by a disciple of Śubhākarasiṃha, called Shan Wuwei 善無畏 in Chinese and Zenmui in Japanese, monk Yi-xing 一行 (Jap. Ichigyō) during the Tang Dynasty. 31 T. XVII, 848, 007a23: 猶如轉輪王寶馬巡履四洲, 於一切時一切處去心不息. 諸菩薩大精進力. 亦復如是. 32 Byakuhō shō 白寶鈔, TZ X 3191. 33 Byakuhō-shō 白寶鈔, TZ X, 3191, p. 827.

570

Benedetta Lomi

domestic) by building pagodas or erecting steles, or of carrying out pacification rituals for their souls.34 However, from the analysis of these legends and practices it emerges that, from the late fourteenth century onwards, there has been a clear attempt at legitimising Batō Kannon ritual practices. The establishment of continuity with the interpretation of the deity given in the existing written sources represents a perfect example of a canonising process.

The Icon, the Practitioner, the Believers In the previous section I have provided the historical and textual background for the practice of associating Batō Kannon with an “excellent horse” and, subsequently, with one’s “favourite horse”, which constitutes the framework for interpreting rituals for the protection of horses. However, this represents only part of the dynamics of the ritual that cannot be limited to the reconstruction of its origins alone. In order to analyse further the issue of canonicity and canonisation we need to shift focus to a more complex network of interpretative processes, gravitating around the ritual process. This can be done by investigating how those directly involved in the practice (the practitioner, the worshippers) by experiencing, interpreting and endorsing the ritual, ultimately make it what it is. Subsequently, by contextualising the interaction and response to different media used, to the interpretative framework discussed in the previous section. Besides dealing with a composite and stratified religious tradition, we are looking at different types or layers of experience based on role (monk/laymen) and on knowledge (of the ritual, of the legends, of the purpose, of the background and so on), which need to be further examined.

The Practitioner and the honzon The goma being an esoteric ritual, we have to start from the crucial element of its execution: the union of the practitioner with the deity, or, more specifically, the process of identification of the practitioner with the honzon. The chief deity of the 34 Even if such rituals are rarely practised today, mostly because of the change in the human/animal relationship which occurred in the past hundred years, there are still ways of memorialising and paying respect to animals, such as dōbutsu ireisai 動物慰霊祭 and petto kuyō ペット供養, which are very widely practised. These rituals, even when considered as a category in themselves, do not present any consistency besides that of being performed for animals. They tend to be local specific, and are more linked to a specific cult, in a particular temple, than to a general idea of animals being protected by a single deity. In the case of Batō Kannon rituals, we are facing a similar problem: it is hard to establish a pattern, for each temple will perform a different ritual, according to its tradition.


571

sokusai goma performed on this occasion is Fudō Myōo and the aim of the practice is, broadly speaking, for the practitioner to interact with Fudō and with a set of other deities, changing according to each section, in order to acquire specific powers. Through this power he can then grant protection, blessings and purification to the worshippers. Thus, the goma is a ritual of empowerment performed to control or influence an event. A goma can be divided up into seven sections, but in the case of Kashōin, there are only five (godan 五段):35 the katen-dan 火天段, the section focusing on the fire deity (J: katen, Sk: agni); the yōshuku-dan 曜宿段, the Celestial Bodies section; the honzon-dan 本尊段, the section of the main object of veneration, which is usually Fudō Myōo;36 the shoson-dan 諸尊段, the section on various deities; the setendan 世天段, the section of the worldly deities, which are the divided body of the honzon and are responsible for the fortune and misfortunes of human beings.37 Each section of the goma has a similar pattern, and focuses on the invitation and identification of the practitioner with the section’s deity (katen, the celestial bodies, Fudō, various deities and the worldly deities). This process is called nyūga ganyū 入我我入, meaning “The Deity enters me, I enter the Deity.” Relying on the assumption that the sacred icon is not simply a reproduction of the historical, iconographical or mythological features of a deity, but it embodies it,38 the Batō Kannon enshrined in the hall is, de facto, the chief deity. If then, following the taxonomy of the ritual manuals, we consider the identification with the chief deity as the focus of the ritual practice, this should reveal a correspondence between the images and the texts, or the images and the ritual. This is not present in the case of the Batō Kannon goma of Miharu: the ritual performed is a regular Fudō Myōo sokusai goma, where no mention is ever made of the deity Batō Kannon. Yet the presence of Fudō is not surprising: he is considered one of the most powerful deities and the Fudō Myōo goma is surely the most widespread form of goma, normally performed even when the ritual is held for other deities. This does not make the Batō Kannon functions any less valid or efficacious for the purpose: the practitioner, in this case the monk and yamabushi, aims at obtaining unity with the deity Fudō Myōo, and through the empowerment, subduing and controlling a number of other deities, thanks to which he grants protection to the community. He 35 Miyake 1999: 376. 36 There are different manuals describing the sokusai goma and dividing each goma according to the deity invoked. Examples are found, for example, in the Goma zensho 護摩全書, a collection of different besson goma rituals, which include a Batō Kannon sokusai goma 馬頭觀音息災護摩, see Soeda 1981: 9. These goma, even if focused on a different principal deity, do not differ in format, so that the Batō Kannon goma would have the same structure as the Fudō Myōo goma, and vary only as far as specific visualisations and invocations of the chief deity are concerned. 37 Payne 1991: 143–91. 38 Faure 1998: 768–813; Strickmann 1996: 203; Sharf 2001: 1–18.

572

Benedetta Lomi

is fully aware of the ritual he is performing, and of the deity he is invoking, and he is also aware that the ultimate goal of the practice is that of applying a general format to a specific circumstance, providing an efficacious means to respond to the needs of the community.39 Furthermore, according to the explanation I was given by the Kashōin monk, the interaction between the practitioner and the deities is not much different from those practices in which a deity, or a class of deities, is invited to identify with Fudō Myōo, such as rituals dedicated to the sun, moon and stars. The powers of Fudō are so great that he enables the practitioner to control other deities as well. Regardless of this, at the performance level, the texts, the implements and the invocations are those of a regular Fudō goma, and at a visual level, Batō Kannon is hidden in the altar, hardly ever shown to the participants: the deity indeed seems absent throughout the ritual. If the deity empowering the ajari is Fudō, excluding Batō Kannon from the ritual process, what then is the role of this deity?

The honzon and its Image Much has been written, in recent years, on the role of images and icons during ritual practice. Scholars have attempted to establish a pattern in the way they are canonised and ritualised, and how they are employed in a religious context. In order to bring to the fore the function of the icon, attention has shifted from issues of iconology and representation to issues of ontology and identification: the statue or the icon depicted is the deity and is perceived as such not at a popular level, but at the doctrinal one as well.40 As the identification of the practitioner with the deity is not only the focus of the ritual practice, but the conditio sine qua non of its outcome, the image should also have a pivotal relevance in relationship to the practitioner during the ritual. At this level, the role of the icon, as already suggested by Faure,41 goes beyond its symbolic and iconographic function, and has to do more with its ontological nature than its interpretation. Religious images, however, can also have a function which is purely representational, sometimes referred to as noniconic,42 meaning that the subject of the representation is religious but the image 39 In another Batō Kannon goma performed on 18 January at Myōkōin 妙光院 Temple in Kobe, the head priest used a ritual manual where the parts on Fudō Myōo have been substituted by his master with those on Batō Kannon. 40 In the esoteric tradition we find a more consistent development of this concept. According to the esoteric teaching systematised by Kūkai in the ninth century, the visual representations are the ultimate, most accurate ways of expressing and understanding the preaching of the Dhārmakaya. 41 Faure 1998: 787. 42 Foulk 2001: 18–19. Foulk stresses that images can be used as icons, in which case they are installed through a ceremony in the altar, they are worshipped as the divinity and are employed during rituals; but not all images functions as icons, they can have a symbolic mean-


573

per se is not considered charged with powers, alive. Such images provide a skilful means to convey meaning, tell a story, and explain through visual representation ideas already present in a text or in a legend.43 Distinguishing between images that have been installed in temples as honzon, which are empowered through a ritual and therefore considered alive, and those that are not, necessarily implies a differentiation in the way both categories of images are used in a ritual context. In turn, the way meaning is represented and conveyed through the visual medium will have different degrees of understanding. In the case of Kashōin, Batō is not at the core of the identification process of the practitioner with the deity. Therefore, its functions may appear simply symbolic: the statue represents the deity.

The Believers and the honzon Forcing religious images into different categories based on their function (ritual/didactic) and power might not be so straightforward if we were to contextualise them into a religious practice, which includes, but is not limited to, ritual per se. If we understand the ritual process not only as the moment in which the ritual is carried out, but also as part of a broader scope of beliefs and practices that partakes in its construction and legitimisation, then the functions of the images will necessarily be layered. The Batō Kannon of Kashōin is a hibutsu, a hidden sacred icon. Images that are considered particularly sacred and powerful are often hidden in a zushi and rarely shown directly to the worshippers. Such secrecy always implies a degree of sacredness surrounding the image which goes beyond its formal features, whatever the reasons for the statue being concealed are.44 There is no doubt, in the case of Kashōin, that the Batō Kannon statue is regarded as a lively powerful icon, which embodies more than it represents. However, issues of visuality and the impact of images, paintings, banners and objects on the ritual practice must still be taken into consideration. Approaching the temple on the day of the goma, the worshippers are constantly reminded of the presence of Batō Kannon through the colourful banners bearing the deity’s name, the numerous ema hung in and outside the halls bearing ing, and are therefore used as didactic tools; images are a way to gain merit, to reproduce a specific text, even to be used as a repository for a sacred text; images are also talismans: small maṇḍala, dhāraṇī or little statues, for example, were often worn by believers, just like a rosary, to protect from illness, misfortunes and so on. 43 This is not the outcome of recent scholarship alone, as already in Medieval China and Japan the issue of the use of images was highly problematical: ideas of images as aids for visualization, as skilful means, coexisted with the tradition of images as incarnation of a deity. Shinohara 2004: 180–224. 44 Rambelli 2002: 271–307.

574

Benedetta Lomi

the scripture “Beloved Horse of Tamuramaro”, and the steles engraved with the name of the deity and of meritorious horses. The presence of the deity through an object or in a symbolic dimension, and the existence of a specific iconography and myth, are all relevant to the ritual since they are significant to the worshippers, to those who attend the ritual, to the community. Part of the Batō Kannon functions and of its ritual paraphernalia is also that of conveying a meaning (to protect horses), recalling a story (the legend of Tamuramaro), explaining through a visual representation ideas already present, in a text or in a legend (the excellent and the beloved horse). It is therefore clear that the role the deity has been invested with, and which is deeply rooted in local history, is extremely important for the meaning of the ritual practice. I am not suggesting that the ritual is necessarily a re-enactment of a memorial ritual for the horse of Tamuramaro, or that the ritual practice is historically derived from such a circumstance. Rather, I argue that the presence of legends and beliefs connecting Batō Kannon with a favourite horse has not only helped the shaping of a particular interpretation of this deity, but is responsible for its part in its canonisation as a horse-protecting deity. This, in turn, provides an interpretative tool for the ritual for lay practitioners and believers.

Conclusions When dealing with Buddhist icons and their rituals, there is a tendency to refer to the canonical textual sources as proof of canonicity for the images and their practice. However, the process of reconstructing the background of a ritual is never limited only to the written medium, even if the textual source suggests it is the first and foremost form of authentication, contextualisation, and canonisation device, as reflected in the taxonomy of the ritual manual. The role of the icon needs to be considered not only in light of the texts, but also of the meaning that is produced and transmitted through a variety of factors and that cannot be reduced to a clearcut division between text and practice, canonical and folk. By overcoming this dualism and by uncovering the different layers of meaning shaping a belief, it emerges that the canonisation of a ritual practice or of its chief deity is neither a static nor a linear process. It involves rather a constant interaction between a received tradition, its roots in the scriptures, the beliefs external to it and the obvious changes and developments caused by socio-economic factors, which are even more relevant in the case of standard ritual, performed for a variety of benefits, such as the sokusai goma. Through analysis of the Batō Kannon goma, this issue has been addressed by tackling two different but related points. The first regards the development of Batō Kannon narratives in the Honshu area. The survey and comparative analysis of the


575

legends of Sakanoue no Tamuramaro in Miharu and that of General Airu in Nikkō, vis-à-vis the canonical sources, provides strong evidence of the association of Batō Kannon with a beloved horse within the Buddhist context, as early as the Kamakura period, through its interpretation as the royal steed, the precious horse of the Chakravartin. This surely provided a fertile background for its further development as protector of horses and for the practice of dedicating a stele of Batō Kannon to meritorious horses, emerging during the Tokugawa period. The second point regards the contextualisation of these narratives within the ritual practice, through the knowledge and experience of the monks and of the worshippers. In this light, analysis of the different roles Batō Kannon is invested with points toward a ritual pattern which is dynamic in the sense that, even maintaining a fixed and common format, it has been able to adapt its structure to a variety of circumstances, and thus can still generate meaning.

576

Benedetta Lomi

Abbreviations T.

Taishō shinshū daizōkyō 大正新修大藏經 (The Buddhist Canon Newly Compiled during the Taishō Era) 100 vols. Takakusu Junjirō 高楠順次郎 et al. (eds.) Tokyo: Taishō issaikyō kankōkai 大正一切經刊行會. 1924–1935

TZ. Taishōshinshū daizōkyō zuzō 大正新修大藏經図像 (Image section of the Buddhist Canon Newly Compiled during the Taishō Era) 10 vols. Takakusu Junjirō 高楠順次郎 et al. (eds.) Tokyo: Taishō Issaikyō Kankōkai 大正一切經刊行會. 1924–1935

References Works in Buddhist collections are cited in the text by canon and volume number Chōen 澄円, Byakuhō shō 白寶鈔 (White Jewel Notes) TZ. X, 3191. Shinkaku 心覺, Besson zakki 別尊雜記 (Miscellaneous Notes on Buddhist Images other than the Buddha) TZ. III, 3007. Shōchō 承澄, Asaba-shō 阿娑縛抄 (Anthology of A, Sa and Va), TZ. VIII, 3190. Śubhakarasiṃha (Shanwuwei, Jpn. Zenmui 善無畏) Darijing 大日経 (Jpn. Dainichikyõ; full title: Daibirushana jõbutsushinhen 大毘盧遮那成佛神變加持經, Mahāvairocana sūtra) T. XVIII, 848. Yixing (Jpn. Ichigyō 一行) Darijing-shu 大日經疏 (Jpn. Dainichikyõsho; full title: Daibirushana jõbutsu 大毘盧遮那成佛經疏 Commentary on the Mahāvairocana Sūtra) T. XXXIX, 1796.

Other References Faure, Bernard 1998. “The Buddhist Icon and the Modern Gaze”. Critical Inquiry 24/3: 768–813. Foulk, Griffith F. 2001. “Religious Functions of Buddhist Art in China”. In: Marsha S. Weidner (ed.). Cultural Intersections in Later Chinese Buddhism. Honolulu: University of Hawaii Press: 13–29. Fukushima-ken (ed.) 福島県編集 1964–1972. Fukushima-ken shi 福島県史. Fukushima-ken and Tokyo: Gannando Shoten 巌南堂書店, Vol. 8. Hagiwara Tatsuro 萩原龍夫 1975. “Nikkō-san engi” 日光山縁起. In: Sakurai Taro 桜井徳太郎 & Hagiwara Tatsuro 萩原龍夫 & Miyata Nobaro 宮田登 (eds.). Jisha engi 寺社縁起. Tokyo: Iwanami shoten 岩波書店. Hayami Tasuku 速水侑 2000. Kannon shinkō jiten 觀音信仰事典. Tokyo: Ebisukosoyo shuppan 戎光祥出版. Hoshino Riichirō 星野理一郎 1960. Nikkō no kojitsu to densetsu 日光の故実と伝説. Utsunomiya-shi 栃木県宇都宮市: Tochigi-ken rengō kyōikukai 栃木県連合教育会. Kyburz, Josef A. 1987. Cultes et croyances au Japon: Kaida, une commune dans les montagnes du Japon central. Paris: Maisonneuve et Larose.


577

— 1994. “Omocha: Things to Play (Or Not to Play) with”. Asian Folklore Studies 53/1: 1–28. MacWilliams, Mark W. 1997. “Temple Myths and the Popularization of Kannon Pilgrimage in Japan. A Case Study of Ōya-ji on the Bandō Route”. Japanese Journal of Religious Studies 24/3–4: 375–411. McCullough, Helen Craig 1964–1965. “A Tale of Mutsu”. Harvard Journal of Asiatic Studies 25: 178–211. Miharu-machi (ed.) 三春町編集 1975–1986. Miharu-machi shi 三春町史. Miharumachi (Fukushima-ken) 三春町(福島県): Miharu-machi 三春町, Vol. 1, 4 & 6. Miyake Hitoshi 1999. Shugendō girei no kenkyū 修験道儀礼の研究. Tokyo: Shunjū Publishing 春秋社. — & H. Byron Earhart (ed.) 2001. Shugendō: Essays on the Structure of Japanese Folk Religion. Ann Arbor: Center for Japanese Studies (University of Michigan). Payne, Richard K. 1991. The Tantric Ritual of Japan: Feeding the Gods – The Shingon Fire Ritual. New Delhi: Aditya Prakashan (Sata-Pitaka Series 365). Rambelli, Fabio 2002. “Secret Buddhas: The Limits of Buddhist Representation”. Monumenta Nipponica 57/3: 271–307. Sharf, Robert 2001. “Prolegomenon to the Study of Japanese Buddhist Icons.” In: Robert Sharf & Elizabeth Horton Sharf (eds.). Living Images: Japanese Buddhist Icons in Context. Stanford: Stanford University Press: 1–18. Shinno Toshikazu 真野俊和 1991. “Minkan shinkō-ron kara minzoku shūkyō-ron e – Bukkyō minzoku-ron no zentei to shite” 民間信仰論から民族宗教論へー仏教民族論の前提として. In: Shinno Toshikazu 真野俊和. Nihon yugyō shūkyōron 日本遊行宗教論. Tokyo: Yoshikawa Kōbunkan 吉川弘文館: 269–289 (Nihon rekishi monzoku shōsho 日本歴史民俗叢書 1). — & Paul L. Swanson (transl.) 1993. “Minkan-shinkō to Minzoku-shūkyō: Reflections on the Study of Folk Buddhism”. Japanese Journal of Religious Studies 20/2–3: 187–206. Shinohara, Koichi 2004. “Stories of Miraculous Images and Paying Respect to the Three Jewels: A Discourse on Image Worship in Seventh-Century China”. In: Phyllis Granoff & Koichi Shinohara (eds.). Images in Asian Religions. Text and Contexts. Vancouver-Toronto: UBC Press: 180–222. Soeda Takatoshi 添田隆俊 (ed.) 1981. Batō Kannon sokusai gomaku shidai 馬頭観音息災護摩供次第. Osaka: Tōhō Publications 東方出版 (Goma zensho 護摩全集 21). Strickmann, Michel 1996. Mantras et mandarins. Le bouddhisme tantrique in China. Paris: Editions Gallimard.

Abstracts Section I: Grammar and Morphology of Ritual Axel Michaels Professor of Cultural and Religious History of South Asia, South Asia Institute, Heidelberg University and Spokesperson of the collaborative research centre “Ritual Dynamics” The Grammar of Rituals Since long scholars have attempted to analyse the rules of rituals in linguistic terms and to de-velop “grammars” of rituals. By way of introducing into the panels’s issues, this paper discusses and summarises the problems of such approaches especially with regard of the “morphology”, “syntax”, “pragmatics” and “semantics” of rituals or ritual elements. It will also be discussed to what extent rituals can be formalised in order to understand them by computer-linguistic methods. It will be argued that the essential problem for developing grammars of rituals is to denote and demarcate the smallest ritual units or elements. This is basically achieved through categories such as authority, age, orthodoxy or extraordinariness. Jan E. Houben Directeur d’Etudes of the Ecole Pratique des Hautes Etudes, Paris Formal Structure and Self-referential Loops in Vedic Ritual The Vedic Nihnava-rite, which has been used as a major example to illustrate a supposed fundamental meaninglessness of ritual, is solidly meaningful if we properly distinguish different organizational levels in ritual – levels which are to some extent comparable to the levels of (a) phoneme-sequences, (b) morphemes, (c) words and (d) sentences in language, and to the levels of (a) genes, (b) cells, (c) organs and (d) organisms in biology. In a further analysis, the Nihnava turns out to be a case of ritual self-reference, which is the minimal meaningfulness to be accepted in a formal representation of ritual that abstracts from collateral aspects of meaning. Different cases of self-reference will be reviewed, including forms of “level-crossing self-referential loops” of which I will argue that they are at the basis both of ritual’s minimal meaningfulness and of its becoming an entity with virtual causality. In the presence of suitable agents (performers of the ritual) this virtual causality can have a wide range of real-life results that vary from transformations in identity and changes of social role, to increased consumption and political violence.

580

Abstracts

Dr. Olga Serbaeva Saraogi UFSP “Asia and Europe”, Department of Indology, University of Zurich When to Kill Means to Liberate: Structure and Meaning of Two Types of Ritual Actions in Vidyāpīṭha Texts Inspired by “Some preliminary thoughts on a grammar of ritual” sent by Axel Michaels, I decided to enlarge my subject originally limited to the ritual actions of a śaiva tantric guru provoking possession-like state in a person being initiated. I shall talk about similar “morphology” of two apparently different types of tantric ritual actions. The first one, valorised and serving as basis of transmission of the system and as a guarantee of future liberation, is already mentioned action of a guru, who makes a disciple manifest visible signs of the descent of śakti in initiation. The second, is action of terrible yoginīs, who, by the use of the very same ritual elements (mantras, mudrās, etc.), kill their victims or drain their blood. What aspect predominates in “identity” of a ritual: is it defined by its main actor, by the “prescriptive” or “descriptive” languge in a given tantric text, by the aim of ritual action, or by its morphology? I shall argue that it is similar sequence of ritual elements and thus “morphology“ and “syntax“ that allowed Kṣemarāja to proclaim the essencial similarity of the two named types of ritual actions. Anand Mishra Research Associate, Cultural and Religious History of South Asia, South Asia Institute, Heidelberg University On the possibility of a Pāṇinian Paradigm for a Rule-based Description of Rituals Formally speaking, a language (e.g. Sanskrit) could be considered as a ‘refined’ sub-set of all possible human utterances. A grammar, like the Aṣṭādhyāyī of Pāṇini, then is a rulebased characterization of this special sub-set. Similarly, rituals (Saṃskāras) can be seen as certain special and “refined” sub-set of the collection of all possible human actions. Is it possible to characterise this special sub-group and can this description be rule-based? The present paper discusses the theoretical possibilities of employing the Pāṇinian paradigm for characterizing rituals. Axel Michaels (in collaboration with Johanna Buss) Axel Michaels, see above; Johanna Buss, Assistant Professor, Department of South Asian, Tibetan and Buddhist Studies, Vienna University The Dynamics of Ritual Formality: The Morphology of Newar Death Rituals The paper deals with a specific part of death ritual called sapiṇḍīkaraṇa or latyā (“fortyfive”) in Nevārī performed by a Newar Jyāpu (farmer) family from Bhaktapur and a Newar Rājopādhyāya Brahman priest originally from Patan. Michaels/Buss would like to analyse the ritual, in which the departed person is basically transformed from a yet unpacified and helpless deceased (preta) to an ancestor (pitṛ), on three levels: the ritual performance at a certain date and place, the manual or ritual handbook used in situ by the performing priest,


581

and local sapiṇḍīkaraṇa texts. The theoretical concern is ritual formality, especially the question how and to what extent ritual sequences can be transformed, left out, added and transposed, and how the priest creates using these means his own ritual referring to a set of established ritual elements more or less known to his fellow priests and customers, thus using a kind of ritual language in both a stereotype and creative way. Cezary Galewicz Adjunct Professor of Indology, Institute of Oriental Philology, Jagiellonian University Inscribing Scripture through Ritual: On the Ritual Cycle of Trisandhā The extraordinary events of Trisandhā are rare to be seen, and outsiders generally not permitted. Due to dwindling number of eligible participants its events take place once in several years. Trisandhā is said to have a religious significance and may take place in a temple, which is not a must, however. It is seen mainly as a display of learned skills and knowledge, though it is the knowledge of the Veda and its display is by necessity ritualized. Trisandhā is an event of considerable prestige for the community of Nambudiri Rigvedis, a prestige of an independent and largely different character than that attached to the better known śrauta rituals. There are regional models for this lengthy ceremony. The one referred to here was initiated in a once famous Vedic school in central Kerala, South India. Trisandhā needs no less than 200 recitation days to complete its due cycle which may amount to two years time! Over this period recitation sessions of roughly 3–4 days long are held during which Ŗgvedasaṃhitā is chanted in different textual modalities according to a ritualized pattern nowhere else to be seen. The actual proceedings follow a complex pattern interspersed with other ritualized forms of religious and social behaviour. Only best virtuoso reciters are able to meet the challenge of long concatenated recitation. After due ritual purification the participants take their seats on a platform according to a pattern of positions reflecting specific types of recitation and levels of progress. One representing a teacher (ācārya) faces the others who for this occasion are called the students. Thus a guru-mukha transmission is ceremonially staged. The structure of the Trisandhā sequences shows how cleverly oral tradition handles the problems of text delimitation. Understood socially, it represents a sophisticated art which still gives its participants a strong feeling of belonging to an elite circle and answering their need for social identity. Martin Gaenszle Professor of Cultural and Intellectual History of Modern South Asia, University of Vienna Grammar in Ritual Speech: The Use of Binomials in Rai Invocations The ritual languages among the Kiranti groups of east Nepal are characterised by a distinct use of nouns in the form of binomials, which dominate ritual discourse while verbal forms are relatively unmarked. This can be described as a case of high referential density (a term first introduced in literary theory and further developed in linguistic typology) and appears to be a reversal of the situation in ordinary language use, where referential density is com-

582

Abstracts

paratively low (i.e. verbal phrases are relatively predominant). Thus grammatical resources are used in a contrastively different manner in ritual discourse. This and other observations on the linguistic features of Kiranti ritual speech lead me to re-consider the metaphor of ritual as based on a “grammar”. By focusing on grammar – in a literal sense – in ritual texts, it becomes clear that ritual performances are not only guided by constitutive rules which define their syntagmatically ordered subunits, but also by distinct linguistic ontologies, pragmatic presuppositions, and interpretive frames which characterize a communicative style and make up a specific ritual universe. Johannes Bronkhorst Professor of Sanskrit and Indian Studies, University of Lausanne Rituals Holophrastic Utterances, and the Symbolic Mind “Although rituals often use language, they are not a language in any strict sense.” The parallelism between language and ritual has however attracted attention, and rightly so. Of particular interest are, and have to be, the formulas (sometimes called mantras) that are used in certain rituals. Two recent developments are likely to throw more light on this phenomenon: (1) the interest for formulaic language in general, primarily associated with the name of Alison Wray; (2) the analysis of symbolic reference presented by Terrence Deacon in various publications. Neither of these two developments has paid attention to ritual formulas. My paper will do so, in the hope of elucidating some aspects of the use of these formulas and, through them, of ritual itself. J. C. Heesterman Professor Emeritus of Indian Cultural History, University of Leiden The Development and Impact of Ancient Indian Ritual The systematic construction of śrauta rituals, as arrangements of various formats of sacrifice according to increasing complexity, may facilitate an analysis of its structure. The present article, however, examines the relationship of śrauta rituals with sacrifice, underpinning the absence of contest and uncertainty in śrauta rituals in contradistinction to the uncertainty and open-endedness of sacrifice. The śrauta system can be seen as a fundamental reform of the institution of sacrifice, where the ritualists restructured the agonistic pattern of sacrifice. Annette Wilke Professor of Religious Studies, Münster University Basic Categories of a Syntactical Approach to Rituals: Arguments for a “Unitary Ritual View” and the Paraśurāma-Kalpasūtra as “Test-Case” Attempts towards a “ritual syntax,” based on linguistic theories, have been highly formalistic. The article suggests a less formalist approach which is based on the “unitary sentence view” of the Indian linguist Bhartṛhari who in the fifth century developed his own kind of


583

“universal grammar”. Bhartṛhari’s focus on syntax provides both a) basic structural categories that also apply to rituals, and b) an inherently holistic approach towards morphology, syntax, pragmatics and semantics. Bhartṛhari’s theoretical framework can account for both the “self-referentiality” of rituals (which I consider to be their primary meaning and function) and a “reflexive holism” regarding semantics (the larger sphere of cultural and social sense which rituals are embedded in, but which they do not necessarily express). To illustrate and test this new Bhartṛhari-inspired syntactical approach and its categories, the Paraśurāma-Kalpasūtra, an elaborate tantric ritual manual, is being used. Considering that tantrics are known as “superritualists” (Sanderson), the ritual script ensures a fairly comprehensive taxonomy. Timothy Lubin Professor at the Department of Religion, and Lecturer in Law and Religion, School of Law, Washington and Lee University Ritual Self-Discipline as a Response to the Human Condition: Toward a Semiotics of Ritual Indices This paper proposes to explain the first-order significance of certain widespread ascetical practices relating to ubiquitous features of bodily existence: the needs and urges of the body (for food, sleep, sex) and the phenomena of birth and death. Rule-governed restrictions on the fulfillment of such urges are features of the ritual practice of professional ascetics and lay votaries alike, across many cultures. Analyzing examples from the Vedic ritual literature, it is argued that underlying the hermeneutically grounded interpretations of Vedic exegesis are units of meaning that “suggest themselves” in the very nature of the practices. These elements are explained in terms of “natural symbols” (Mary Douglas’s phrase) or “indexical signs” (in C. S. Peirce’s semiotics) – signs associated with or affected by their objects as a matter of fact. Such signs may be “legible,” at least at a basic level, independently of any merely conventional association. The common features of bodily existence might be said to provide all necessary context. This can explain the appearance in widely disparate cultures of very similar ritual signs and even the more complex ritual gestures built upon them. The rites discussed here (Vedic ablution, fasting, vigil, sexual continence, ritual rebirth) deploy indexical signs “pointing” to renewal and effacement of decay and to a defiance of human limitations. The resulting complex ritual becomes a gestural assertion of control over what is otherwise beyond normal human power – of resistance to helplessness before the laws of nature. Rich Freeman Visiting Assistant Professor, Departments of History and Religion, Duke University Pedagogy and Practice: The Meta-pragmatics of Tantric Rituals in Kerala This chapter takes as its principal focus the multiplex layerings of discourse and practice that are entailed in the learning and performance of temple worship among Nambudiri Brahman priests in Kerala. Drawing on fieldwork in a school that trains and qualifies tem-

584

Abstracts

ple priests throughout Kerala, I consider the stratified relations between multiple levels of Sanskrit and Malayalam texts, oral and practical instruction, and their meta-discursive correlation in the processes of teaching and learning the rituals and of performing them in public contexts of worship. This provides the context for reconsidering the relation between linguistic and ritual behavior suggested by the trope of modeling a “grammar of ritual.” I argue that the evident inability of any extensional notion of “grammar” to deal even with the pragmatic features of natural language, relates directly to what I demonstrate here: that ritual forms and practices are generated and regulated only in and through the higher, metapragmatic order of culture. Frederick M. Smith Professor of Sanskrit and Classical Indian Religions, Department of Religious Studies, Department of Asian Languages and Literature, University of Iowa Historical Symmetry and Ritual Asymmetry: The Interrelations Between Vedic Ritual and Temple Construction in Modern India In February-March 2007 a Vedic Soma sacrifice (agnyādhāna, the ritual establishing of the Vedic fires that mark the inception of a ritualist’s career, followed immediately by an agniṣṭoma, the archetypal soma sacrifice) was performed in the small town of Dodamarg, in southern Maharashtra, barely one kilometre from the Goa border. The Vedic rituals were actively supported by fundraising drawn from crowds that attended a consecration (pratiṣṭhā) ceremony for the main image at a new temple dedicated to the well-known Maharashtrian saint, Shirdi Sai Baba (1838? – 1918). The Vedic rituals were performed on the grounds of the still unfinished temple. The two spheres of ritual involved, Vedic śrauta sacrifice and Hindu temple consecration, are widely divergent in their intentions, sources, and performative modes. However, on this occasion they supported each other in an interlocking manner. In the views of the temple committee, the Vedic rituals boosted the efficacy of the consecration ceremony, while the Vedic rituals could not have occurred at all if it were not for the temple consecration. The relations between the two groups of officiants were cordial, though they were guided by distinct ideologies and commanded different audiences. I will present photos that depict a number of interesting borrowings and adaptations that illustrate how subtle changes in the performances reflect a changing ritual morphology. Furthermore, I will describe how this morphology has been subtly shifting for the last fifty years. Though this might be a unique case, it is symptomatic of the manner in which different kinds of ritual have converged to create a new syntax that, in this case at least, enhanced both parties and contributed to the survival of at least one of them.


585

Frits Staal Professor Emeritus of Philosophy and South and Southeast Asian Studies at the University of California, Berkeley A Theory of Ritual: The Indo-Iranian Fire Offering The term “theory“ refers to an abstract, underlying area that explores general structures. It presupposes “facts“, disparate masses of particular happenings and occurrences that are apprehended intuitively. There is no harm in such apprehension and theories must be tested by such facts; but there are no methods or algorithms that describe or explain how theories originate from them. They seem to require an intermediary: the creative minds of human beings who are immersed in facts. One fact is important in our context: later rituals incorporate earlier ones. It applies to the particular area and period with which we are concerned: the northern half of the subcontinent of South Asia, presently Pakistan and northern India; and the period between roughly 1100 and 600 B.C.E., that is, late Vedic. The paper consists of eight sections: The Indo-Iranian Fire Offering; The Oral Transmission of the Vedas; The Four Vedas; The Verbal Root śru-, “to hear”; Ritual Versus Ordinary Activity; A Dynamic and Social, but not a Ritual Change; Rules, Metarules and Defaults options; Ritual Space: the Four Directions.

Section II: Ritual Discourse, Ritual Performance in China and Japan Michael Puett Professor of Chinese History, Department of East Asian Languages and Cultures, Harvard University Ritualization as Domestication: Ritual Theory from Classical China One of the most exciting developments in the field of ritual studies over the past few decades has been the attempt to explore the enormous body of ritual theory that has been generated in non-Western cultures throughout the world. Frits Staal has done excellent work in bringing theories of ritual from South Asia into our theoretical frameworks, and, following his example, we need to do more of the same from the many other traditions throughout the world that have developed indigenous ritual theories. This paper will be a small contribution to this larger project by discussing ritual theory from China – one of the cultures that has a lengthy, indigenous tradition of theorizing about ritual. I focus particular attention on the “Li yun” chapter of the Book of Rites (Liji), a work from early China that would ultimately become one of the most influential ritual texts in East Asia.

586

Abstracts

Joachim Gentz Reader in Chinese Studies, Asian Studies Department, University of Edinburgh Living in the Same House: Ritual Principles in Early Chinese Reflections on Mourning Garments Reflection on form, meaning and function of particular rituals and rituality in general finds its expression in early Chinese mourning literature (ca. fourth to first centuries B.C.E.) especially in the form of explanations of concrete interpretations of rituals. These reflections are based on references to social, cosmological, emotional and religious systems of order, which are taken as the ground and the aim of ritual order. Recourse is sometimes taken even to different and diverging systems of order in order to strengthen the own argument. Such orders can either just be joined together in an additional way or they are systematically interwoven. A great part of Chinese ritual literature deals with mourning garments. Since the order of mourning garments represents a highly complex and multilayered reference system of social relationships it offers an opportunity to use it for the negotiation of basic principles of the social, political, religious, cosmological, ethical, economical and emotional order. It is in this context, therefore, that we find basic reflections on and explicit formulations of abstract ritual principles, which in the talk shall be presented and related to their particular exegetical discursive contexts. Christian Meyer Assistant Professor, East Asian Institute, University of Leipzig Interpretations of Ritual (“li”) in Chinese Scholarly Discussions in the Eleventh Century The topic of Confucian rituals (li) appears prominent in Song dynasty debates of the eleventh century. Unclear were not only details of the performance of ritual at the court, but also the hermeneutical rules of interpreting the classics. In the context of public, partly political, discussions different hermeneutical approaches developed. On the one hand there were those who were open for a flexible adaptation of rituals for the present, among them the influential thinkers and politicians Ouyang Xiu (1007–1072) and Su Dongpo (1037– 1101). According to them ritual should follow the “natural human feelings” (“yin renqing”), a concept which would also allow more flexibility. On the other hand there were the so-called Neo-Confucians, among them Cheng Yi (1033–1107), who took rituals seriously as revivable institutions from the classical past and embodiments of cosmological patterns. As the most ideal and normative expression of human feelings they should not be altered. The study of the classics should rather help exercise ritual to become a “sage” (shengren). Both positions intended shaping a Confucian identity which could counter unorthodox alternatives like Buddhism and Daoism.


587

Gil Raz Assistant Professor, Religion Department, Dartmouth College Ritual Theory in Medieval Daoism While the vast majority of texts in the Daoist canon and in extra-canonic collections possessed by Daoist priests are ritual manuals and ritual has been the main vehicle for interaction between the Daoist priesthood and the general population, we know relatively little about the practice and meaning of Daoist ritual. The theoretical underpinnings of Daoist ritual remain almost unexplored. This is partly due to the fact that there are few explicitly theoretical treatises that deal with ritual. A more important reason for this gap may be the implicit assumption that discourse on ritual in Daoist texts is simply about correct practice and not about meaning, let alone theory. I argue in this paper that Daoists constructed their ritual synthesis with certain theoretical assumptions that informed their understanding of ritual, from structure and performance, to meaning and explanations of efficacy. Julius Tsai Assistant Professor, Department of Religion and Program in Asian Studies, Temple University Mutation or Permutation? A Ritual Debate in Tang-Song Daoism The tenth-century Xuantan kanwu lun 玄壇刊誤論 (Rectification of Errors on the Altar of Mystery, DZ 1280) of Zhang Ruohai 張若海 introduced innovations in Daoist liturgy that ignited centuries of passionate debate. Highlighting this controversy over ritual change, this essay draws attention to underlying tensions between charisma and lineage, pragmatism and tradition, and human and divine initiative in the construction of ritual. Analyzing the Xuantan kanwu lun not only provides a window onto the formation of the classical Daoist rite known as the suqi 宿啟 (Nocturnal Invocation), but it also offers an invitation to think more broadly about the protean, adaptive structure of Daoist ritual. Lucia Dolce Senior Lecturer in Japanese Religions, Department of the Study of Religions, School of Oriental and African Studies, University of London The Contested Space of Buddhist Public Rituals: The shunie of Tōdaiji (omizutori) The Buddhist liturgical assemblies (hōe) performed in the great temples of Nara were originally repentance rituals (keka) that the monastic communities offered for the benefit of the nation. Some of these liturgies, such as the shunie of Tōdaiji (omizutori), are still performed today as exorcistic rituals executed at the beginning of the year. Their core practice consists of ascetic exercises and repentance, the latter conceived not as a mental process of individual awareness, but as a set of ritual actions, including bodily movements, melodic chanting and recitations.

588

Abstracts

This contribution explores the performative aspects of Tōdaiji’s omizutori. While I refer to the discursive narratives on mythological and symbolic themes that the ritual has produced, my main concern is the dynamics of its enactment in a diversified space and according to different devotional protocols. The performance accommodates Buddhist deities and kami, as well as distinct groups of actors and sponsors for whom the ritual yields meaning and empowerment. The scale and complexity of omizutori thus provide grounds to analyse the multiple agency inherent in ritual practice, and the layered meaning it generates. Fumi Ouchi Professor at the Department of Music, Miyagi gakuin Women’s University Buddhist Liturgical Chanting in Japan: Vocalisation and the Practice of Attaining Buddhahood Intoning mystical formulas such as paritta, mantra, vidyā and dhāraṇī is one of the essential practices in the esoteric tradition of Buddhism. In Japan, when esoteric Buddhism was systematised in the Heian period, mantra recitation (nenju) was emphasised as one of the tripartite practice, or three mysteries (sanmitsu gyō). Through it a practitioner could realise his or her original nature to be tantamount to the Buddha and attain buddhahood in the present body. This suggests that the practice of intoning a mantra or dharānī was endowed with critical power in an elaborately constructed soteriological system. This contribution explores how mantra intonation functions in the actual ritual performance. It focuses on the eighteen-methods practice (jūhachidō), which is paradigmatic of all esoteric rituals and the first an esoteric practitioner has to learn. Taking a performative approach along with the doctrinal, I demonstrate that mantra recitation plays different roles. On the one hand, it provides a practitioner with the physical foundation for perceiving and constructing the distinctive experience of being one with the principal deity. On the other hand, by performing a simple repetition numerous times, a practitioner becomes deeply involved in the ritual procedure, through which he or she learns how to behave and interpret his or her experience. In this sense, vocalisation produces a creative correlation between theory and practice. Such an analysis allows us an understanding of the somatic nature of enlightenment in the Japanese Buddhist tradition and the ways for accomplishing it. Katja Triplett Associate Professor of the Study of Religions, Institute of Comparative Cultural Research – Cultural and Social Anthropology and Study of Religions, Marburg University Esoteric Buddhist Eye-healing Rituals in Japan and the Promotion of Benefits In the Esoteric Buddhist ritual system, having access to the benefits and special blessings of a deity that can protect one from calamity, from being possessed by a demon, or can heal terrible diseases, naturally brings with it authority and social power. I examine a specific benefit of the bodhisattva Avalokiteśvara: the healing of eyes as offered by Tsubosakadera, a medium-sized temple near Nara, Japan. This serves as an example for a situation where members of the cleric elite and non-elite groups competed in promoting a particular benefit connected to a popular sanctuary. In the first part of the chapter I look at the textual as well


589

as the iconographic background for esoteric Buddhist eye-healing rituals in East Asia, especially as to the Senju sengen darani-kyō. In the second part, I explore miracle tales and foundation stories connected to Tsubosakadera and show that in the medieval and early modern periods, eye-healing rituals may have been widely promoted by practioners not officially part of a Buddhist monastic order. Evidently, not only the temple’s priests but also itinerant – often blind – minstrels and ritualists promoted rituals and the attractive benefits connected to Tsubosakadera’s Avalokiteśvara (Kannon) as well as the deity Sarasvatī (Benzaiten) throughout a wide region. Thus, particular groups not belonging to the elite seem to have been exceedingly influential in shaping local ritual Buddhist culture and religious ideas over a period of many centuries. M.A. Butler Assitant Professor, Department of History, Case Western Reserve University Ritual Specialists and Collective Agency in Song Dynasty China I examine “magic” and occult rituals documented in comprehensive military manuals of China’s middle imperial era. Coincident with the rise of printing, officially prescribed manuals have been interpreted as the imperial court’s attempt to dictate ideology via ritual. Research shows, rather, that the court was eager to gather up “lost” rituals. Nevertheless, the court remained suspicious of occult ritual specialists and of military rituals generally. The ability of occult specialists (diviners) to create collectivities within unofficial networks threatened the court, who carefully watched them and proscribed “magical” texts. (Official court diviners were confined to the capital.) In the field, diviners were drafted on an ad-hoc basis; sometimes the military commander acted as ritual specialist. In the absence of closely-held equipment, diviners adapted field rituals by using the troops themselves to create simulacrum of supernatural emblems essential to ritual efficacy. Unexpectedly, these rituals focused on producing yin (shadow, moon, female, wet, chaos, etc.) rather than yang (light, sun, male, dry, order, etc.) power, further threatening the stability of imperial rule, at least from the court’s perspective. Paul S. Atkins Associate Professor of Japanese, Department of Asian Languages and Literature, University of Washington The Stages of Seppuku: Performing Self-Execution in Premodern Japan Seppuku (self-disembowelment by sword, also called hara-kiri) emerged in the early years of Japan’s medieval age (1100–1600) as a final gesture of defiance by defeated warriors toward their victorious enemies. In many instances, the seppuku performer tore out his own intestines and cast them toward his foe. This disturbing gesture functioned as an expression of spite, a posthumous curse, and a display of masculine bravado. From the fifteenth century, however, with the first known use of seppuku as a form of punishment for members of the military class, a new kind of performance was required. All traces of rancor were erased from the original ritual, and an elaborate system of codes was created to regulate the be-

590

Abstracts

havior of the participants. Such practices diverted attention from the fact that what was once conducted as a semi-voluntary act was now a thinly disguised beheading. This article compares the function of seppuku as a ritual element in these two scenarios. It examines the origins of seppuku, the practice of “entrail casting,” the significance of the intestines in Japanese culture, and the consequences of ritual failure in seppuku. Poul Andersen Associate Professor, Department of Religion, University of Hawaii at Manoa The Life of Images in Daoist Ritual The images of the gods that adorn the sacred areas of Daoist ritual have been characterized by some scholars as “objects of worship” (i.e., objects treated by the worshipers as imbued with the “living presence” of the gods); by others as external representations of the inner pantheon addressed by the Daoist high priest primarily within his own body; and in one case as “mere decoration” and “explanatory additions or spin-offs [that] are often seen because they are very decorative.” This article aims to demonstrate that neither of these positions are very helpful in accounting for the attitudes toward images of the participants in Daoist ritual, and that, indeed, these attitudes are subject to much variation, in different parts of China, at different times, and within different liturgical traditions. Based on the close reading of ancient Daoist texts relating to the philosophy of images, and on the analysis of specific performances of Daoist ritual in present-day southern Taiwan, it is shown that, on the contrary, these attitudes exhibit the same kind of ambivalence toward images that is found in most major religions around the world, and that has been aptly characterized by W.J.T. Mitchell as the “double consciousness surrounding images.” It follows that the question of the “life of images” in Daoist ritual is more likely to find convincing answers through careful studies of individual cases than through the deployment of sweeping generalizations of the above-mentioned kind. Benedetta Lomi PhD candidate at the Department of the Study of Religions, School of Oriental and African Studies, University of London The Iconography of Ritual: Images, Texts and Beliefs in the Batō Kannon Fire Offering Ritual and religious practices emerging in response to the needs of worshippers or to cultural and local-specific factors are often labelled “folk” or “popular”, as opposed to “canonical”. Such a strict polarity, supported by a division between what is based on texts and what is not, neglects to acknowledge the different layers of meaning that compose the ritual practice, ultimately failing to address the subject of ritual practice appropriately. In this contribution, I analyse a goma (fire offering) for the protection of horses offered to Batō Kannon, the horse-headed manifestation of the Bodhisattva of compassion, held at Kashōin, a small temple of Miharu, in north-eastern Japan. Through the investigation of the role of the practitioner and of the believers in relation to the practice and its object of veneration, I argue that different, if not contrasting understandings of a given ritual are


591

proper to “canonical“ and “folk” practices alike. The article also sheds light on the intriguing role of Batō Kannon, by looking at its development as the protector of horses and animals.

Ethno-Indology. Heidelberg Studies in South Asian Rituals Edited by Axel Michaels

4: Martin Gaenszle, Jörg Gengnagel (Eds.)

Words and Deeds

Visualizing Space in Banaras: Images, Maps, and the Practice of Representation

Hindu and Buddhist Rituals in South Asia Edited by

Jörg Gengnagel Ute Hüsken Srilata Raman

2006. 358 pages, 94 ill., hc ISBN 978-3-447-05187-3 € 48,– (D) / sFr 83,–

Harrassowitz Verlag

5: Maha-deva Veda-ntin

1: Jörg Gengnagel, Ute Hüsken, Srilata Raman (Eds.)

Words and Deeds Hindu and Buddhist Rituals in South Asia

2005. 299 Seiten, 10 Abb., br ISBN 978-3-447-05152-1 € 48,– (D) / sFr 83,–

2: Ute Hüsken, Srilata Raman-Müller (Eds.)

Ritual in South Asia Text and Context In preparation.

3: Niels Gutschow, Axel Michaels

Handling Death

The Dynamics of Death Rituals and Ancestor Rituals among the Newars in Bhaktapur, Nepal With Contributions by Johanna Buss and Nutan Sharma and a Film on DVD by Christian Bau 2005. 216 Seiten, 140 Abb., 1 DVD, gb ISBN 978-3-447-05160-6 € 68,– (D) / sFr 116,–

Mı-ma-msa-nya-yasamgraha ˙ ˙ A Compendium of the principles

of Mı-ma-msa˙ Edited and translated by James Benson 2010. 908 pages, hc ISBN 978-3-447-05722-6 Ca. € 148,− (D) / sFr 250,−

6: Niels Gutschow, Axel Michaels

Growing Up

Hindu and Buddhist Initiation Rituals among Newar Children in Bhaktapur, Nepal With a film on DVD by Christian Bau 2008. 307 pages, 138 ill., 17 maps, hc ISBN 978-3-447-05752-3 € 64,– (D) / sFr 109,–

7: Jörg Gengnagel

Visualized Texts Sacred Spaces, Spatial Texts and the Religious Cartography of Banaras 2010. Ca. 358 pages, 20 ill., hc ISBN 978-3-447-05732-5 Ca. € 54,− (D) / sFr 93,−

HARRASSOWITZ VERLAG • WIESBADEN www.harrassowitz-verlag.de • [email protected]

Orient • Slavistik • Osteuropa • Bibliothek • Buch • Kultur

Ethno-Indology. Heidelberg Studies in South Asian Rituals Edited by Axel Michaels

8: Barbara Schuler

Of Death and Birth Icakkiyamman-, a Tamil Goddess, in Ritual and Story With a Film on DVD by the Author 2009. XVI, 501 pages, 1 map, 14 ill., 1 DVD, hc ISBN 978-3-447-05844-5 € 98,– (D) / sFr 166,–

Of Death and Birth by Barbara Schuler deals with dynamics of a local (non-Brahmanical) ritual, its modular organisation and inner logic, the interaction between narrative text and ritual, and the significance of the local versus translocal nature of the text in the ritual context, while providing a broad range of issues for comparison. It demonstrates that examining texts in their context helps to understand better the complexity of religious traditions and the way in which ritual and text are programmatically employed. The author offers a vivid description of a hitherto unnoticed ritual system, along with the first translation of a text called the Icakkiyamman- Katai (IK). Composed in the Tamil language, the IK represents a substantially longer and embellished form of a core version which probably goes as far back as the seventh century C.E. Unlike the classical source, this text has been incorporated into a living tradition, and is being constantly refashioned. A range of text versions have been encapsulated in the form of a conspectus, which will shed light on the text’s variability or fixity and will add to our knowledge of bardic creativity.

9: Ute Hüsken

Vis·n·u’s Children

Prenatal life-cycle rituals in South India Translated into English by Will Sweetman With a DVD by Ute Hüsken and Manfred Krüger

2009. 322 pages, 1 DVD, hc Book + DVD: ISBN 978-3-447-05854-4 € 52,– (D) / sFr 90,– DVD: (available apart) ISBN 978-3-447-05853-7 € 24,– (D) / sFr 42,20

The history of the Vaikha-nasas, a group of Brahmanic priests in the Vis·n·u temples of south India, is characterized by the effort of claiming their status against rivaling priests. Ute Hüsken’s work discusses the controversy, ongoing for centuries, as to what makes a person eligible to perform the rituals in Vis·n·u temples. Since the 14th century CE the discussion in the relevant Sanskrit texts centers around the question of whether the Vaikha-nasas priests must undergo an initiation, or whether their particular prenatal lifecycle ritual vis·n·ubali makes them eligible to perform temple ritual. In addition to the textual perspective, three instances of local conflicts from the 19 th/20th centuries about the question of whether the Vaikha-nasas require an initiation are analysed in their contexts. Three examples of present day performances of the crucial ritual vis·n·ubali are presented and interpreted in the light of the relation between text and performance and from the perspective of the acting priests’ ritual competence.

HARRASSOWITZ VERLAG • WIESBADEN www.harrassowitz-verlag.de • [email protected]

Orient • Slavistik • Osteuropa • Bibliothek • Buch • Kultur

Ritual Dynamics and the Science of Ritual: I: Grammars and Morphologies of Ritual Practices in Asia

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual Theory, Ritual Practice

Ritual Dynamics and the Science of Ritual: V: Transfer and Spaces

Ritual Dynamics and the Science of Ritual: II: Body, Performance, Agency and Experience

Ritual Dynamics and the Science of Ritual: IV: Reflexivity, Media and Visuality

Ritual Dynamics and the Science of Ritual: III: State, Power and Violence

Ritual in the Dark

Santa Ritual

Ritual Dynamics and the Science of Ritual: I: Grammars and Morphologies of Ritual Practices in Asia

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual

Ritual Theory, Ritual Practice

Ritual Dynamics and the Science of Ritual: V: Transfer and Spaces

Ritual Dynamics and the Science of Ritual: II: Body, Performance, Agency and Experience

Ritual Dynamics and the Science of Ritual: IV: Reflexivity, Media and Visuality

Ritual Dynamics and the Science of Ritual: III: State, Power and Violence

Ritual in the Dark

Santa Ritual

Recommend Documents