Functional Approaches to Spanish Syntax: Lexical Semantics, Discourse and Transitivity

Functional Approaches to Spanish Syntax Lexical Semantics, Discourse and Transitivity Edited by J. Clancy Clements and...

Author: J. Clancy Clements | Jiyoung Yoon

184 downloads 1265 Views 1MB 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

Functional Approaches to Spanish Syntax Lexical Semantics, Discourse and Transitivity

Edited by J. Clancy Clements and Jiyoung Yoon

Functional Approaches to Spanish Syntax

This page intentionally left blank

Functional Approaches to Spanish Syntax Lexical Semantics, Discourse and Transitivity Edited by

J. Clancy Clements Indiana University

and Jiyoung Yoon University of North Texas

Editorial matter and selection and Chapter 1 © J. Clancy Clements and Jiyoung Yoon 2006 Individual chapters © contributors 2006 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2006 by PALGRAVE MACMILLAN Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N.Y. 10010 Companies and representatives throughout the world. PALGRAVE MACMILLAN is the global academic imprint of the Palgrave Macmillan division of St. Martin’s Press, LLC and of Palgrave Macmillan Ltd. Macmillan® is a registered trademark in the United States, United Kingdom and other countries. Palgrave is a registered trademark in the European Union and other countries. ISBN-13: 978–1–4039–9406–6 ISBN-10: 1–4039–9406–4 This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Clements, J. Clancy. Functional approaches to Spanish syntax : lexical semantics, discourse and transitivity / edited by J. Clancy Clements and Jiyoung Yoon. p. cm. Includes bibliographical references and index. ISBN 1–4039–9406–4 1. Spanish language – Syntax. I. Yoon, Jiyoung. II. Title. PC4361.C54 2005 465—dc22 2005051277 10 9 8 7 6 5 4 3 2 1 15 14 13 12 11 10 09 08 07 06 Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham and Eastbourne

This book is dedicated to the memory of Clements’ teacher Eugenio Coseriu

This page intentionally left blank

Contents List of Tables and Figures

ix

Acknowledgements

xii

Notes on the Contributors

xiii

1

Introduction J. Clancy Clements and Jiyoung Yoon

2

Semantic and Discourse-Pragmatic Factors in Spanish Word Order Belén López Meirama

7

Continuity and Episodic Structure in Spanish Subject Reference Llorenç Comajoan

53

3

1

4

Gustar-Type Verbs Victoria Vázquez Rozas

80

5

Primary and Secondary Object Marking in Spanish J. Clancy Clements

115

6

Null Direct Objects in Spanish J. Clancy Clements

134

7

Transitivity and the Syntax of Inalienable Possession in Spanish Richard Winters

8

Ser-estar in the Predicate Adjective Construction J. Clancy Clements

9

Spanish Adjective Position: Differences between Written and Spoken Discourse Richard J. File-Muriel

151 161

203

10 Adjective Placement and Noun Semantics in Spanish Mariche García-Bayonas

219

11 Transitivity and Spanish Non-Anaphoric se J. Clancy Clements

236

vii

viii Contents

12 ‘Juan salió contento’: Semantic Constraints on Small Clauses in Adjunct Position Jiyoung Yoon

265

13 Causative hacer and dejar Carmen Ruiz-Sánchez

278

Index

301

List of Tables and Figures

Tables 1.1 1.2 2.1

2.2 2.3 2.4 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 4.1 4.2 4.3 4.4 5.1 5.2

Component parts of Transitivity Component parts of individuation Comparison of semantic differences and syntactic patterns of prototypically transitive clauses sentences v. clauses sentences with a gustar-type verb construction (In)definiteness, (non-)referentiality and topichood The theme–rheme dichotomy The overlay between thematic and syntactic structure Percentage rates of third-person pronouns in different Spanish varieties Percentage rates of zero, pronoun, and full NPs in different Spanish varieties Nominal devices for topic continuity Subject nominal devices used in the narratives Nominal devices to introduce new characters Nominal devices to introduce same subjects from previous clauses Nominal devices to introduce different subjects from previous clauses Look-back distance for definite NPs, pronoun, and zero when they refer to a different subject from the previous clause Overall hits and misses for the episodic model Hits and misses for new referents, old referents and different subjects Nominal devices in episodes 5 and 6 Nominal devices to introduce continuous subjects in Spanish, English and Japanese Degree of topicality in indirect objects Frequencies and percentages of animate v. inanimate subjects in two-participant clauses Frequencies and percentages of clauses functioning as subject Frequencies and percentages of preposition and postposition of subject and indirect object for gustar Nominative-accusative v. ergative-absolutive marking Direct–indirect v. primary–secondary object marking ix

3 3

18 21 26 29 54 55 56 60 61 61 62 63 63 64 69 69 84 88 89 97 117 122

x

List of Tables and Figures

5.3 5.4

The etymological Spanish pronominal system Frequency of le as a function of the referents’ animacy in (1) writings of St Teresa; (2) speech of present-day Castilian professional women; (3) speech of present-day Castilian rural speakers 5.5 A probable result of leísmo in a variety of the Castilian pronominal system 5.6 General pronominal system found in the Moratín play El sí de las niñas 5.7 Probable object-marking system based on pronoun use in Moratin’s El sí de la niñas 6.1 Lazard’s definiteness scale 8.1 Characterization of verb classes 8.2 Examples of Spanish adjective types in terms of features 8.3 Comparison of commonly occurring verb forms of ser and estar 8.4 Form-by-form frequency comparison of different forms of ser and estar 8.5 Copula-adjective combination according to adjective type and reading for both animate and inanimate subject referents 9.1 Spoken and written style 9.2 Position of As to Ns for the all the data according to their respective weights 9.3 Position of As to Ns in the spoken data according to their respective weights 9.4 Position of As to Ns in the written data according to their respective weights 9.5 Preposed adjectives in spoken discourse 9.6 Adjective position 9.7 One v. two adjectives modifying same noun 9.8 Distribution of semantic categories 9.9 Chi-square test for Table 9.8 data (semantic class) 9.10 Diachronic distribution of A position in texts 9.11 Mean word length of As by semantic class 10.1 Interpretations of alto funcionario and funcionario alto in adults and children 10.2 Interpretations for azafata alta and alto piloto in adults and children 10.3 Semantic interpretations for N-pequeño and pequeño-N in adults and children 10.4 ‘Miserable’ v. ‘poor’ interpretation for preposed v. postposed pobre

124

125 125 126 127 143 169 173 182 183

198 210 210 210 210 211 212 212 212 213 214 214 225 225 226 227

xi List of Tables and Figures

10.5 10.6 10.7 10.8

11.1 12.1 13.1 13.2

‘Mere’ v. ‘simple-minded’ interpretation for preposed v. postposed simple Chi-square results for the adults Chi-square results for the children The significance of specific–general and general–specific ordering in the instrument, calculated with tokens from the specific semantic interpretation in adult and child groups Characterization of Vendlerian Aktionsart categories Licensing factors of adjunct predicates in relation to Transitivity Characterization of causative dejar Characterization of causative hacer

227 228 228

229 251 272 288 289

Figures 5.1 8.1 8.2 8.3

8.4 13.1 13.2 13.3 13.4 13.5

Possibilities for a simple lexical split of case marking: two two-way sub-systems, ‘accusative’ v. ‘ergative’ The distribution of ILPs and SLPs over states, activities and events Prototypical coding of the world in language based on time stability Time stability of adjectives with no underlying process/event v. adjectives with an underlying process v. adjectives with an underlying event Spanish adjective classification Reference frame of hacer Reference frame for causative hacer with activities Reference frame for main verb and causative dejar Implied reference frame of causative dejar in certain cases Reference to the result of an event in small clauses with an implied event or state

116 166 169

170 173 295 296 296 297 297

Acknowledgements This volume has benefited greatly from discussions with students over the years. We are grateful to those students for all their input and also to an anonymous reviewer whose comments shaped the argumentation of many of the contributions. Thanks go also to Jill Lake for presenting this project to Palgrave Macmillan, to Manolo Triano López for translating Chapters 2 and 4 and to Richa Clements for proofreading some of the typescript. J. CLANCY CLEMENTS JIYOUNG YOON The editors, contributors and publishers are grateful to Georgetown University Press for permission to use Chapter 6, Mouton de Gruyter for permission to use Chapter 5, and Elsevier Publishers for permission to use Table 6.1. Every effort has been made to contact all copyright holders, but if any have been inadvertently omitted the publishers will be pleased to make the necessary arrangement at the earliest opportunity.

xii

Notes on the Contributors J. Clancy Clements is Associate Professor of Linguistics and Spanish and Portuguese at Indiana University, Bloomington, USA. He received his MA (1979) in Spanish from the Universität Tübingen, Germany, and his PhD (1985) in Romance Linguistics from the University of Washington, Seattle. His main areas of interest are contact linguistics and functional syntax, with a focus on varieties of Iberian Romance languages. His writings include The Genesis of a Language: The Formation and Development of Korlai Portuguese (1996) and The Linguistic Legacy of Spanish and Portuguese: Colonial Expansion and Language Change (forthcoming), four co-edited volumes, as well as over 30 articles on language-contact phenomena and functional linguistics. Llorenç Comajoan is Assistant Professor of Spanish at Middlebury College, USA. His main areas of interest are second-language acquisition, language variation and language policy. Richard File-Muriel is a doctoral student in the Department of Spanish and Portuguese at Indiana University, Bloomington, USA. His main interests are phonetics, phonology and contact linguistics. Mariche García-Bayonas is Assistant Professor in the Department of Romance Languages at the University of North Carolina at Greensboro, USA. Her main areas of interest are second-language acquisition, phonetics, phonology and sociolinguistics. Belén López Meirama is Professor of Spanish in the Department of Spanish at the Universidad de Santiago de Compostela, Spain. Her main areas of interest are functional syntax and syntax of Spanish, with a focus on the syntax of the sentence. Her writings include La posición del sujeto en la cláusula monoactancial en español (1997), and various articles on Spanish syntax and pedagogically oriented writings. Victoria Vázquez Rozas is Professor in the Department of Spanish in the Universidad de Santiago de Compostela, Spain. She was a visiting professor at the Romansk Instituto of the University of Copenhagen (1989) and in the Spanish and Portuguese Department at the University of California at Santa Barbara (2001). She has published numerous studies on syntax and discourse from a functionalist perspective, including El complemento indirecto en español (1995). She is a contributor to the development of the Syntactic Data Base (http://www.bds.usc.es). Carmen Ruiz Sánchez is a doctoral student in the Department of Spanish and Portuguese at Indiana University, Bloomington, USA. Her three main areas of xiii

xiv Notes on the Contributors

specialization are Hispanic sociolinguistics, second-language acquisition and functional syntax. She received her BA (1999) in English philology from the University of Seville, Spain, and her MA from Indiana University in TESOL/Applied Linguistics (2001) and Hispanic Linguistics (2003). Richard Winters is Assistant Professor in the Department of Modern Languages at the University of Louisiana at Lafayette, USA. His main research interests are functional and formal syntax, with a focus on the Romance languages. Jiyoung Yoon is Assistant Professor of Spanish and Spanish Linguistics and a coordinator of the first- and second-year Spanish programme at the University of North Texas (Denton). She received her MA at the Universidad de Guadalajara in Mexico (1994) and her PhD at Indiana University (2002). She specializes in functional syntax and semantics as well as foreign language instruction and pedagogy. Her work includes a forthcoming monograph on Spanish small clause constructions and articles on semantic and syntactic phenomena in Spanish from both functional and construction-grammar approaches. She is currently working on a cross-linguistic study of subject and object coding.

1 Introduction J. Clancy Clements and Jiyoung Yoon

Since about 1970 the study of syntax has increasingly benefited from semantics-based approaches. In generative grammar, work by Bresnan (1982, 2001), Levin and Rappaport Hovav (1995) and Pollard and Sag (1987, 1994), among others, have acknowledged and explored the intricate links between lexical semantics and syntactic structure. In the functional/cognitive-oriented literature, work by Bybee (1985, 2005), Bybee, Perkins and Pagliuca (1994), Dean (1992), Dik (1997), Croft (1991, 2001), Foley and Van Valin (1984), Givón (1995, 2001), Goldberg (1995), Gutierrez-Ordóñez (1997), Halliday (1985), Hopper and Thompson (1980, 1984), Langacker (1987, 1991), MacWhinney (1999) Thompson and Hopper (2001), Tomasello (1998, 2003), Van Valin and LaPolla (1997), Vazquez Rozas (1997), Verhaar (1990) and many others have significantly furthered the study of the multifarious connections between syntax on the one hand and lexical semantics, discourse and cognition on the other. Levin and Rappaport Hovav (1995: 1) note that the hypothesis that the syntactic properties of lexical items (specifically verbs) are determined by their meaning is traceable as far back as P¯an.ini in the 6th century BCE. They suggest that there is a causal relation between the regularities linking arguments bearing certain semantic roles and particular syntactic expressions: [t]o the extent that the semantic role of an argument is determined by the meaning of the verb selecting it, the existence of linking regularities supports the idea that verb meaning is a factor in determining the syntactic structure of sentences. The striking similarities in the linking regularities across languages strongly suggest that they are part of the architecture of language. From the perspective of language acquisition, Bates and Goodman (1999) offer substantial empirical evidence in first-language (L1) acquisition to support this causal relation between the lexicon and grammar (comprising syntactic structure and also grammatical relation marking), arguing for a 1

2

Introduction

unified lexicalist approach to the processes involved in grammar acquisition. Numerous studies by them and their colleagues reveal the close relation between the development of the L1 lexicon on the one hand, and the development of L1 grammar on the other. That is, ‘[t]he dependence of early grammar on vocabulary size is so strong and the nonlinear shape of this function is so regular that it approaches the status of a psychological law …’ (1999: 51). The syntax of Spanish is intricately connected to the verb semantics and discourse structure. In Chapter 2, for example, it is shown for intransitive verbs that the more agentive and definite a subject is, the more likely it is to appear preverbally; and the less agentive and definite a subject is, the more likely it is to appear postverbally. Spanish syntax is also sensitive to the argument structure of verbs. In Chapter 5, Spanish is argued to be a (SV/VS-VO) subject–verb/verb–subject–verb–object type language (Dryer, 1997), with a slight predominance of verb–subject order in intransitive clauses but an overwhelming predominance of subject–verb order in transitive clauses. In Chapter 4, it is shown that two-argument verbs of the gustar type have overwhelmingly experiencer subjects, and as these are low in agency they appear overwhelmingly in postverbal positions. The syntax of Spanish is also sensitive to the semantics of the noun phrase (NP). Animacy of object arguments is shown to underlie the ergative pattern discovered in the Spanish pronoun system, discussed in Chapter 5. Overt v. null pronominalization of direct objects, examined in Chapter 6, is found to depend on the semantic nature of the direct object. Thus, the semantics of the verb and its argument NPs in Spanish has direct repercussions for its syntax. One seminal work that has informed much of the above-mentioned research, and one which addresses the intimate connection between lexical semantics, morphosyntax and discourse, is Hopper and Thompson’s (1980) Transitivity Hypothesis. Traditionally, transitive clauses are those whose verbs take a direct object (DO). Hopper and Thompson, however, cast the notion more broadly, viewing Transitivity as a set of component parts that takes into consideration all aspects of the carrying-over of an action (or the transferring of energy) from one participant to another. They argue that Transitivity is scalar (2001: 28) and that it ‘can be broken down into its component parts, each focusing on a different facet of the carrying-over in a different part of the clause’ (1980: 253). By breaking Transitivity down into its component parts, it is possible to characterize clauses, not as transitive or intransitive, but rather as more or less transitive. On this view, Transitivity becomes a question of degree, and syntactic structure and marking can be accounted for by positing degrees (high or low, more or less) of Transitivity. The factors making up the notion of Transitivity, adapted from Hopper and Thompson (1980) and shown in Tables 1.1 and 1.2, are lexical semantic properties of verbs/predicates and nouns and discourse properties of noun

J. Clancy Clements and Jiyoung Yoon 3

phrases (for example definiteness, referentiality). There are two observations to be made regarding the tables. First, the reference to objects in Table 1.1j, 1.1k and Table 1.2 refers to both direct as well as indirect objects. Hopper and Thompson (1980: 259–60) note that indirect objects should be considered Transitive objects, just as direct objects are. Secondly, the item included in Table 1.1i (in small caps and bold) is an addition to Hopper and Thompson’s original proposal, an extension originally proposed in a similar framework by Verhaar (1990: 94–7), who distinguishes between what he calls intransitive verbs of actorhood, high on the Transitivity scale such as the equivalents of ‘swim’, ‘talk’ and ‘run’, and intransitive verbs of undergoing, low on the Transitivity scale such as equivalents of ‘fall’, ‘melt’ and ‘faint’. By adopting this extension, it is possible to include intransitive clauses on the Transitivity scale as well. Thus, with the help of the values ‘high/low’ in Table 1.1 and ‘individuated/non-individuated’ in Table 1.2 one can measure the degree of Transitivity of any given clause, and use this as a basis for predicting the covariance of semantic and syntactic features of different elements that co-occur in a clause. Table 1.1 Component parts of Transitivity

a. PARTICIPANTS b. KINESIS c. ASPECT d. PUNCTUALITY e. VOLITIONALITY f. AFFIRMATION g. MODE h. AGENCY i. INDIVIDUATION of S(ubject) j. AFFECTEDNESS of O k. INDIVIDUATION of O

High

Low

Two or more participants, A(gent) and O(bject) action telic punctual volitional affirmative realis A high in potency

One participant non-action atelic nonpunctual nonvolitional negative irrealis A low in potency

S HIGHLY INDIVIDUATED O totally affected O highly individuated

S NON-INDIVIDUATED O not affected O nonindividuated

Table 1.2 Component parts of individuation (Table 1.1i and 1.1k) Individuated

Non-individuated

proper human, animate concrete singular count referential, definite

common inanimate abstract plural mass nonreferential, indefinite

4

Introduction

The Transitivity Hypothesis states that in any given clause, elements of the clause should co-vary in terms of their high- or low-Transitivity features. Two examples from Spanish will illustrate the point. In Chapter 6, overt v. null direct object pronominalization is examined. It turns out that overt pronominalization is not necessary if the direct object is equivalent to a mass noun or a bare plural. Whereas the pronominalization of el libro in (1.1a) with lo is obligatory in Spanish syntax (see 1.1b v. 1.1c), the pronominalization of a mass noun can be with a null, as shown in (1.2): (1.1)

a. ¿Compraste el libro? ‘Did you buy the book?’ b. Sí, lo compré. ‘Yes, I bought it.’ c. *Sí, compré. ‘Yes, I bought.’

(1.2)

a. ¿Compraste cafe? ‘Did you buy coffee?’ b. Sí, compré Ø;. ‘Yes, I bought (some).’

In terms of the Transitivity Hypothesis, this can be analysed as follows. Overt pronominalization co-varies with higher Transitivity, while null pronominalization co-varies with lower Transitivity. This is calculated using the component parts in Individuation listed in Table 1.2. The direct object NP el libro ‘the book’ in (1.1a) is count, referential and definite, whereas the direct object NP café ‘coffee’ in (1.2a) is a noncount (mass), nonreferential and indefinite. The NP with a higher number of high-Transitivity (Individuation) traits is overtly pronominalized in Spanish. This is discussed in more detail in Chapter 6. The other example involves the verb pair dormir ‘sleep’ and dormirse ‘fall asleep’. In Chapter 11, we argue that the presence of se co-varies with higher Transitivity. In this example, as well as in many others, it is true. In terms of the component parts of Transitivity in Table 1.1, the predicate dormir ‘sleep’ is atelic. That is, it has duration and does not contain an inherent endpoint. By contrast, dormirse ‘fall asleep’ has no duration (it is punctual) and is thus telic. In other words, the point at which the event occurs is its endpoint. Thus, sentence (1.3a) is lower in Transitivity than sentence (1.3b): (1.3)

a. Marta durmió en casa de su amiga. ‘Marta slept at her friend’s house.’ b. Marta se durmió en casa de su amiga. ‘Marta fell asleep at her friend’s house.’

J. Clancy Clements and Jiyoung Yoon 5

The present volume adopts the Transitivity Hypothesis, as well as other models in the area of lexical semantics, syntax and discourse, as the basis for examining various morphosyntactic, syntactic and discourse phenomena in Spanish. The topics include word order (Chapters 2 and 4), null arguments (Chapters 3 and 6), ergative patterns in object marking (Chapter 5), inalienable possession (Chapter 7), ser-estar (Chapter 8), non-anaphoric se (Chapter 11), small-clause phenomena (Chapter 12) and causative constructions (Chapter 13). Adjective placement is analysed in a number of ways. Chapter 9 (‘Syllable weight and adjective placement’) demonstrates statistically that processing considerations involving syllable weight of nouns and adjectives play an important role in adjective placement. In Chapter 10 (‘Spanish adjective placement and noun semantics’), it is suggested that the semantic interpretation of a noun (specific v. general) is linked to adjective pre- v. postnominal placement respectively. The application of probabilitistic methodology is shown effective in identifying the most significant variables affecting adjective placement and the rates of grammaticalization apparent in this phenomenon (Chapters 9 and 10). Moreover, the use of data from large corpora (Chapters 2, 4, 5, 6, 8, 9, 11, 12, 13) or statistical analysis of elicited/collected data (Chapters 3, 7) lends strong empirical support to the findings on overt v. null arguments, word order, case-marking systems and adjective placement. We hope that this volume will contribute to filling the void we perceive in the literature on Spanish functional syntax and in general usage-based analysis of natural language. To date, this is the first comprehensive study of Spanish syntax from the general functional perspective. We hope it will encourage other scholars to develop and extend this line of research. References Bates, E. and J. Goodman (1999) On the Emergence of Grammar from the Lexicon. The Emergence of Language, ed. B. MacWhinney. Malwah, NJ: Lawrence Erlbaum, pp. 29–79. Bresnan, J. (2001) Lexical-Functional Syntax. Malden, MA: Blackwell. —— (ed.) (1982) The Mental Representation of Grammatical Relations. Cambridge, MA: MIT Press. Bybee, J. (2005) ‘The Impact of Use on Representation: Grammar is Usage and Usage is Grammar’, Presidential address at the Linguistics Society of America annual meeting. Oakland, CA, 6–9 January. —— (1985) Morphology. Amsterdam: Benjamins. Bybee, J., R. Perkins and W. Pagliuca (1994) The Evolution of Grammar. Chicago: University of Chicago Press. Croft, W. (2001) Radical Construction Grammar. Oxford: Oxford University Press. —— (1991) Syntactive Categories and Grammatical Relations. Chicago: University of Chicago Press. Dean, P. (1992) Grammar in Mind and Brain. Exploration in Cognitive Syntax. Berlin: Mouton de Gruyter.

6

Introduction

Dik, S. (1997) The Theory of Functional Grammar. 2 vols, revd edn. Berlin: Mouton de Gruyter. Dryer, M. (1997) ‘On the Six-way Word Order Typology’, Studies in Language, vol. 21(1), pp. 69–103. Foley, W. and R. Van Valin (1984) Functional syntax and universal grammar. Cambridge: Cambridge University Press. García, E. (1975) The role of theory in linguistic analysis: the Spanish pronoun system. Amsterdam: North Holland. Givón, T. (2001) Syntax: an introduction. 2 vols. Amsterdam: Benjamins. —— (1995) Functionalism and grammar. Amsterdam: Benjamins. Goldberg, A. (1995) Constructions. A Construction Grammar Approach to Argument Structure. Chicago: University of Chicago Press. Gutiérrez Ordoñez, S. (1997) Principios de sintaxis funcional. Madrid: Arco Libros. Hopper, P. J. and S. A. Thompson (1984) ‘The Discourse Basis for Lexical Categories in Universal Grammar’, Language, vol. 60, pp. 703–52. —— (1980) ‘Transitivity in Grammar and Discourse’, Language, vol. 56, pp. 251–99. Langacker, R. (1991) Foundations of Cognitive Grammar II. Stanford: Stanford University Press. —— (1987) Foundations of Cognitive Grammar I. Theoretical Prerequisites. Stanford: Stanford University Press. Levin, B. and M. R. Hovav (1995) Unnacusativity: At the Syntax-lexical Semantics Interface. Cambridge, MA: MIT Press. MacWhinney, B. (ed.) (1999) The Emergence of Language. Malwah, NJ: Lawrence Erlbaum. Pollard, C. and I. Sag (1994) Head-driven Phrase Structure Grammar. Chicago: Chicago University Press. —— (1987) Information-Based Syntax and Semantics. Stanford, CA: Center for the Study of Language and Information. Thompson, S. A. and P. J. Hopper (2001). ‘Transitivity, Clause Structure, and Argument Structure: Evidence from Conversation’. in J. Bybee and P. J. Hopper (eds), Frequency and the Emergence of Linguistic Structure. Amsterdam: Benjamins, 27–60. Tomasello, M. (2003) The New Psychology of Language. Cognitive and Functional Approaches to Language Structure. Volume 2. Mahwah, NJ: Lawrence Erlbaum. —— (1998) The New Psychology of Language. Cognitive and Functional Approaches to Language Structure. Volume 1. Mahwah, NJ: Lawrence Erlbaum. Van Valin, R. and R. LaPolla (1997) Syntax: Structure, Meaning and Function. Cambridge: Cambridge University Press. Vázquez Rozas, V. (1995) El complemento Indirecto en español. Santiago de Compostela: Universidade de Santiago de Compostela. Vendler, Z. (1967) Linguistics in Philosophy. Ithaca, NY: Cornell University Press. Verhaar, J. (1990) ‘How Transitive is Intransitive?’, Studies in Language, vol. 14, pp. 93–168.

2 Semantic and Discourse-Pragmatic Factors in Spanish Word Order1 Belén López Meirama

Introduction Spanish is usually characterized as a free-word-order language that typologically belongs to the group of subject–verb–object (SVO) languages. In this respect, we will address two things. First, a typological description of Spanish word order is too vague given that Spanish has many intransitive clauses in which subjects are postverbal (VS order). In this chapter, we will argue that Spanish is a subject-initial, that is, an SVO/SV, language.2 Secondly, the expression ‘free order’ cannot be construed as ‘arbitrary order’ since the variations in the order of constituents are never gratuitous. Following Thompson (1978: 20), we claim that Spanish is a language that uses constituent order for both grammatical and discourse-pragmatic purposes: From a typological point of view, then, we can show that some languages utilize predicate-argument order primarily for pragmatic purposes, and some primarily for grammatical purposes. There are also languages which use predicate-argument order for both purposes without giving priority to either; Spanish may be such a language. It is not uncommon for word order to be considered a way of identifying grammatical relations between the constituents of the sentence, generally related to case marking and agreement. In typological approaches to language, it is a common practice to establish possible relationships between different markings (see, for instance, Forster and Hofling, 1987). In this sense, word order can play an important role in morphologically poorer languages such as English or French. However, in morphologically richer languages such as Spanish, in which word order does not play a crucial role in identifying grammatical relations in the sentence, word order variations are used at a discourse-pragmatic level ‘to establish a social or affective stance, to create text-level structures from sentence-level resources, and to communicate in a 7

8

Spanish Word Order

way which is optimally matched to the text receiver’s cognitive capacities’ (Downing 1995: 9). In Spanish, there are many factors that influence the order of constituents, some of which are strictly formal. For example, it is claimed that Spanish follows the quasi-universal rule of placing the heaviest constituents towards the end of the sentence (see Hernanz and Brucart, 1987, Fernández Soriano, 1993, or Gawelko, 1995). It is also argued that in certain sentence types, subjects can appear postverbally, as in the case of relative sentences, in which the relative pronoun is the element that initiates the sentence (for example, el reloj que me regaló mi padre [lit. the watch that me-gifted my father] ‘the watch my father gave me as a present’). Subjects also appear obligatorily postverbally in absolute (that is nonfinite) constructions, as in al llegar el médico … [lit. upon arriving the doctor] ‘when the doctor arrived …’, llegado el médico … [lit. arrived the doctor] ‘once the doctor had arrived …’. Moreover, some descriptive grammars of Spanish posit rules such as those in Butt and Benjamin (1988: 519) given below: ●

●

● ●

Spanish does not separate prepositions from the noun or pronoun that they modify … In general, only no should separate a preposition from its infinitive. Set phrases, particularly set verbal phrases like tener que ‘to have to’ should not be broken up by the insertion of other words. As a rule, words should not be inserted between haber and a participle. Unstressed object pronouns (me, te, se, la, lo, le, nos, os, los, las, les) are never separated from their verb.

Formal factors such as those just mentioned show minimal or no variation and thus will not be examined here. Instead, we will focus our analysis on the most important semantic and discourse-pragmatic factors that affect the word order in Spanish. Since the topic is rather complex and related to many different issues, we will focus mainly on two issues: the so-called basic sentence,3 that is, the mono-transitive sentence and the intransitive monoargumental structure, and the most relevant instances of basic word order alteration.4 In the next section we discuss the most relevant semantic factors affecting word order in the sentence, that is, agentivity and definiteness of the subject,5 and we deal in the third section with the most important discoursepragmatic factors affecting word order in Spanish. Starting with the most generalizad notion of topic (that which is spoken about, that which a sentence is about), we point out the link between the topic and other discoursepragmatic notions, such as cohesion, emphasis and the distinction between old and new information. We hasten to note, however, that the link between these notions and topic does not assume identification between the notion of topic and the others. Moreover, we shall attempt to illustrate the clearest

Belén López Meirama 9

types of a marked topic in Spanish, focusing above all on the fronting of the direct object (DO). We see our contribution to the ongoing discussion of word order in Spanish to be twofold. We present arguments in favour of considering the distinction between unergative and unaccusative to refer to construction types and not to verbs. That is, we argue that it is more reasonable to speak of verbs that prefer an unergative or unaccusative construction rather than to speak of unergative or unaccusative verbs. Second, we argue that it is possible to define the notions of theme, rheme, topic, and focus in terms of tendencies and patterns instead of hardfast rules, and illustrate our arguments with data gleaned from Arthus. Abbreviation that we have used for original texts cited are listed at the end of the chapter.

Semantic factors Linguists have approached the subject of word order by analysing certain semantic factors affecting the distribution of the elements in a sentence; factors such as polarity, modality, aspect, and so on. For example, Knauer (1989) discusses how polarity can affect word order in a sentence, arguing the Spanish double negation (NEG) construction favours a postverbal subject (for example todavía no ha venido nadie [lit. yet NEG has come no one] ‘nobody has come yet’; Hernanz and Brucart (1987) argue that subject inversion is practically obligatory in a marked modality such as interrogative and imperative types. In this chapter, we will focus on two semantic factors which have generated most interest to date in the bibliography about word order in Spanish: the agentivity and the definiteness of the subject. In the generative framework, the semantic verb type plays an important role in characterizing word order. For example, intransitive verbs, that is, monadic verbs, have been divided into two semantic classes: the unergative class, which are constructed with a preverbal agentive subject; and the unaccusative class, constructed with a postverbal, non-agentive subject. Although the Unaccusative Hypothesis is not without its drawbacks (see below), it has pointed out an indisputable fact about Spanish grammar: the relevance of the semantic role of the subject with regard to its position within a clause. In general, a semantically less prototypical subject of a transitive clause tends to be in postverbal position. In this regard, we will examine a one-argument construction with VS order, and two two-argument constructions: indirect object-verb-subject (IVS) and adverbial adjunct-verbsubject (AVS).6 In addition to agentivity and animacy, linguists have also pointed out that definiteness is a very important factor affecting the likelihood of an element occupying a topical position. As we will see later in this section, definiteness transcends the purely semantic sphere, and thus can be best characterized as semantic-pragmatic (we choose, however, not to discuss definiteness because

10 Spanish Word Order

it would break the internal coherence of the section, which deals with the most representative cases of marked order in Spanish). Agency in a subject Unaccusative constructions Some generative studies on word order in Spanish claim that unaccusativity is a factor that triggers a postverbal subject in one-argument clauses. In other words, the subjects of the so-called ‘unaccusative verbs’, when receiving the -role ‘patient’, are generated in D-Structure in the position that corresponds to objects (see de Miguel Aparicio, 1989, and Fernández Soriano, 1993). Therefore, this hypothesis postulates that unergative intransitive verbs are generated with a preverbal subject, and that the unaccusative intransitive verbs are generated with a postverbal subject, as shown in the following examples from Arthus:7 (2.1)

(2.2)

a. Mil quinientos hombres avanzan con el agua hasta la cintura, su fusil recargado en los antebrazos. ‘One thousand five hundred men advanced waist-deep in water, their rifles resting on the forearms.’ (DIEGO: 91, 18) b. Los dos equipos jugarán el sábado, a partir de las diez, en Vigo. ‘Both teams will play Saturday at ten in Vigo.’ (1VOZ: 47) a. Cuando se habla del carácter de los norteamericanos, casi siempre aparece la palabra ingenuidad. ‘When people talk about the nature of North Americans, the word ‘naivety’ surfaces almost always.’ (TIEMPO: 49, 38) b. ¡Es estúpido! Cada día, todos los días del año, muere gente. ‘That’s stupid! Every day, all year round, people die.’ (MIRADA: 119, 19) c. Pero últimamente las cosas se han torcido para ti. Ha surgido otro candidato, un rival. ‘But as of late things have been going bad for you. Another candidate has emerged, a rival.’ (HOMBRE: 34, 30)

The verbs avanzar ‘to advance’ and jugar ‘to play’ in (2.1) are unergative. That is, they are accompanied by an agent subject that thus appears in preverbal position, whereas the verbs aparecer ‘surface’, morir ‘die’, and surgir in (2.2) are considered unaccusative because they possess a patient subject that appears in postverbal position. Two additional subgroups within the group of unaccusative verbs have revealed themselves to be important: verbs of movement/direction (e.g.: arribar ‘arrive’, ascender ‘ascend’, bajar ‘go down, caer ‘fall’, descender ‘descend’, entrar ‘go in’, llegar ‘arrive’, partir ‘depart’, salir ‘leave’, subir ‘go up’, venir ‘come’, and so on) and existential verbs (e.g., abundar ‘be abundant,


teem’, acaecer ‘happen’, durar ‘last’, escasear ‘be scarce’, figurar ‘appear’, menudear ‘happen frequently, ocurrir ‘take place’, quedar ‘remain’, subsistir ‘subsist’, and so forth).8 However, Spanish still lacks a complete list of unergative verbs organized in semantic classes such as the one provided for English by Levin and Rappaport Hovav (1995). Nevertheless, research done on word order includes a sufficient number of verbs as to allow the Unaccusativity Hypothesis to be tested on a copious corpus of present-day Spanish. While such a study is beyond the scope of this chapter, a cursory examination of data provided by Arthus still provides us with highly illustrative examples. For instance, comparing the uses of the verbs avanzar ‘advance’, jugar ‘play’, aparecer ‘appear’ and morir ‘die’ in (2.3) and (2.4) with the ones discussed in (2.1) and (2.2), we note not only that clauses containing avanzar and jugar can be constructed with a postverbal subject (cf. 2.3), but that clauses with aparecer y morir are also found with preverbal subjects, as in (2.4). (2.3)

(2.4)

a. Cuando las piernas me llevaron a su vera se abrió el semáforo y avanzó el tráfico rodado. ‘When my legs led me to her, the trafficlight turned green, and the traffic moved on.’ (LABERINTO: 38, 23) b. Pero cuando después hace la interpretación de esos datos, entra a jugar su sagacidad. ‘But when he interprets those data, his cleverness begins to play a part (in the situation).’ (BAIRES: 53, 4) a. Todas las mujeres de la casa aparecieron de repente. ‘All the women of the house suddenly showed up.’ (SUR: 15, 9) b. El viudo de Xius murió dos años después. ‘Xius’s widower died two years later.’ (CRÓNICA: 41, 28)

Even more revealing than these examples are those instances regarding verbs of movement provided by the corpus, with which one can verify that the majority of such verbs can be equally constructed with an agentive subject (animate entity that volitionally controls an activity definable as movement), as in the (a) examples in (2.5)–(2.8), or with a nonagentive subject (a usually inanimate entity that is introduced in discourse by means of a presentative verb), as in the corresponding (b) examples.9 (2.5)

a. El viejo asoma con gesto contrariado que se torna jubiloso al ver al joven. ‘The old man leans out with an expression that turns jubilant when he sees the youth.’ (SONRISA: 253, 9) b. Vuelve atrás: estaban hablando, hablan, no importa qué, hablan. Asoma la sonrisa horrible.


(2.6)

a.

b.

(2.7)

a.

b.

(2.8)

a.

b.

‘Come back: they were talking, they talk, it doesn’t matter what they talk about, they talk. That horrible smile turns up.’ (MIRADA: 33, 17) El hijo sale y vuelve pronto con dos hermosas peras y un cuchillo, sobre un plato que deja en la mesilla. ‘The son leaves and comes back with two gorgeous pears and a knife on a plate that he sets on the table.’ (SONRISA: 26, 13) Abrí los grifos y salió el agua tibia y parduzca. ‘I opened the faucets and the water flowed out tepid and brownish.’ (LABERINTO: 92, 23) Tía Delia vino enseguida, y entre ella y mamá me obligaron a seguir viviendo. ‘Aunt Delia came at once, and between her and mom they forced me to continue to live.’ (SUR: 39, 13) … tenía yo treintaiún años. Entonces, pues, empezaron a venir los hijos. Tengo cinco hijos. ‘I was thirty-one. Then the children started coming. I have five children.’ (SEVILLA: 110, 3) Sin saber cómo, el hijo de David había vuelto; estaba otra vez sentado frente a él. ‘Without knowing how, David’s son had come back; he was again sitting in front of him.’ ( JÓVENES: 177, 12) El lago, la lluvia, el ángel … Volvían las imágenes que despertaban sus recuerdos. ‘The lake, the rain, the angel … The images that awakened his/her memories were coming back.’ (CARTA: 137, 14)

Moreover, it is worth noting that the majority of verbs are polysemic, having commonly several meanings, with one or more involving movement and others involving existential, presentative, or stative readings. Consider the following examples, which involve different meanings of the same verb commonly found in monolingual Spanish dictionaries. (2.9)

(2.10)

a. ¡David bailaba delante del Tabernáculo! (meaning ‘dance’) ‘David was dancing in front of the Tabernacle!’ (OCHENTA: 88, 24) b. Por las páginas del libro bailaba la luz del sur. (meaning ‘oscillate’) ‘The sunlight was dancing across the pages of the book.’ ( JÓVENES: 57, 34) a. Los chuchos corren de un extremo a otro. (meaning: ‘move fast’) ‘The mutts run from one place to another.’ (PAISAJES: 57, 12) b. Corre el año 1850 en París y está anocheciendo. (meaning ‘go by’) ‘It was the year 1850 in Paris, and it was getting dark.’ (ZORRA: 7, 6)


(2.11)

a. Tu vecina pasa. No sé si me habrá visto. (meaning: ‘go by/ through a place’) ‘Your neighbor is passing by. I don’t know if she saw me.’ (DIEGO: 87, 19) b. Por eso tienes los nervios así, destrozados. Pero bueno, ya pasó todo. (meaning ‘cease, end’) ‘That’s why your nerves are on edge. But everything is over now.’ (SUR: 88, 24)

However, the fact that many verbs may display both constructional alternatives is not the only objection to the approach we are examining. More serious is the observation that the Unaccusative Hypothesis has been applied to the study of word order in excessively reductionist terms. That is, the Unaccusative Hypothesis leads us to believe that an argument accompanying an intransitive verb can only be semantically equivalent to either the subject of the transitive clause (i.e., the verb is unergative) or the direct object (i.e., the verb is unaccusative). Although a prototypical transitive clause is a case in which agent and patient are most clearly defined and differentiated, an intransitive clause offers a whole gamut of possibilities with regard to semantic functions served by its only argument. The possibilities range from the most prototypical patient (found in the traditionally called ‘reflexive passives’) to the most prototypical agent found in active sentences. The following examples, which only represent a sampling, are organized in two series: pronominal sentences, shown in (2.12), and active sentences (2.13), in which the (b) example presents a subject whose degree of agentivity lies between that of (a) and (c) examples. (2.12)

(2.13)

a. Arruti se despidió y se fue. ‘Arruti said goodbye and left.’ (HISTORIAS: 89, 28) b. Al fondo del zaguán se abre la puerta del ascensor y … ‘At the end of the hallway, the door of the elevator opens and …’ (SONRISA: 151, 13) c. Se editarán mil copias del dibujo realizado por la niña de seis años. ‘A thousand copies of the six-year-old girl’s drawing will be published.’ (3VOZ: 32) a. La enfermera sonrió buscando mi complicidad. ‘The nurse smiled, looking for my complicity.’ (DIEGO: 145, 10) b. Pero gusta más el carácter nuestro, parece que estamos más en familia. ‘But our character is more pleasing, it seems that we are among friends.’ (SEVILLA: 291, 10) c. Comieron, y cuando entraron en la plaza ya había mesas ocupadas en las terrazas. Abundaban los hombres.


‘They ate, and when they entered the square there were already full tables on the terraces. There were many more men.’ (JÓVENES: 174, 34) This variability is related to an issue addressed in Hopper and Thompson’s (1980) seminal work, namely, that the Transitivity of a sentence involves its semantic structure, which in turn is determined by sets of different factors. It turns out that all of the factors are properties relevant to the study of word order in Spanish. In this regard, it is essential to underscore that such properties are associated with tendencies, that is, they do not set rules. Based on the discussion and the observations so far, the so-called Unaccusative Hypothesis can be refined as follows: ●

●

●

With respect to word order, it is more appropriate to talk about unaccusative and unergative constructions, rather than unaccusative and unergative verbs. In addition to agentivity, there are other semantic features that influence the pre- or postverbal placement of the arguments: a study on word order in Spanish should also include at least definiteness, which we will discuss later in the chapter. Semantic features never set categorical rules of distribution of the arguments. They do, however, mark tendencies in a perceptible and measurable way.

Two-argument constructions with postverbal subject Several researchers have proposed to extend the Unaccusative Hypothesis to some two-argument sentences, such as those with a subject-indirect object configuration with an experiencer subject (for example, gustar ‘like’), and those assuming a location whose nucleus is normally a verb of movement (for example, entrar ‘go in’). In both cases, we have a so-called internal subject in postverbal position. Examples are given in (2.14): (2.14)

a. A Miguel le gusta el cebiche. (Indirect object–Verb–Subject) ‘Miguel likes ceviche.’ b. Por esa ventana entra el viento. (Adverbial adjunct–verb–subject) ‘Through that window the wind comes in.’

With regard to Spanish grammar in particular, Hatcher’s (1956) work is noteworthy is that it has been cited as an example of semantic classification of unaccusative verbs in Spanish. In it we find the classes that allow, and even trigger, the presence of a locative element,10 as shown in the examples in (2.15), taken from Hatcher (1956). The example in (2.15a) corresponds to


a verb that belongs to the come-class. The verb in sentence (2.15b) belongs to the existence-presence-class, while the verb in (2.15c) is from the production-class. (2.15)

a. A mis ojos acudían las lágrimas como temerosas. Tears would well up in my eyes, as if they were tremulous.’ b. Muy cerca de aquí está el palacio. ‘The palace is very near here.’ c. Y en sus ojos se cristalizaba un recuerdo inefable. ‘And in his/her eyes an ineffable memory was crystallized.’

Secondly, the discussions by Morales de Walters (1982) and Delbecque (1991) address the question of verb semantics and subject position, and include mention of psych verbs. The following examples have been taken from Morales de Walters and correspond to the class labelled by this author as ‘verbs of psychological reaction’. (2.16)

a. Me fascina la cocina francesa. me fascinates the cuisine French ‘I love French cuisine.’ b. Nos gusta ver los logros de nuestro trabajo. us likes to see the fruits of our work ‘We like to see the fruits of our work.’

Examples (2.14–2.16) reveal the need to review both types of constructionlocative and psych verbs, albeit briefly, in order to account for an unmarked order that does not coincide with the basic word order in Spanish. With respect to locative verbs, the structure containing a locative verb illustrated in (2.14b), termed ‘locative inversion’, is considered by many linguists as a diagnostic of unaccusativity. They posit a series of properties in Spanish with the purpose of demonstrating the equivalence between the subject of this structure and the direct object of the transitive sentence. Such properties, however, fail a detailed analysis, and thus cannot be adduced as evidence (in this respect, see López Meirama, 1997a, and Cifuentes Honrubia, 1999). Against the position that the locative inversion construction is an example of syntactic unaccusativity, Levin and Rappaport Hovav (1995) provide an alternative account for the properties of the unaccusative type that this construction exhibits (that is, the postverbal subject and the preverbal locative or directional element). The key point, in their view, is the discursive function that the construction serves. In other words, the construction is used as a presentational type of focus used to introduce the referent of the postverbal NP. Levin and Rappaport Hovav, as well as Birner (1992, 1994), argue that the general function of all inversions, including the locative one, is that of using


sentence initial-position to link relatively unfamiliar information to already familiar information in the discourse. The information represented by postverbal NPs must always be less familiar – but not necessarily new – than that represented by preverbal NPs. That is, the locative inversion introduces less familiar information within the context of more familiar information. The construction, therefore, is presentational, but in a more general sense than the one we attribute to this term. The verb in this construction must be informationally light and, if an appropriate context is provided, it may belong, at least in principle, to any semantic type of verb (including the unergative types), although the canonical types are represented by those referring to existence (for example existir ‘exist’) and appearance (for example aparecer ‘appear [in the punctual sense]’) because these do not add information to that conveyed by the preverbal NP. The following examples illustrate the difference brought about by the positional change at the discourse level between a subject and a locative element. This difference is clear with existence and appearance verbs as in examples (2.17, 2.18) and movement verbs as in examples (2.19, 2.20): (2.17)

(2.18)

(2.19)

a. Allí donde mirase descubría señales de su existencia. Él habitaba ya de alguna manera entre nosotros. ‘Wherever I looked, I would find signs of his existence. Somehow, he was already living among us. (SUR: 100, 4) b. La residencia de Lalín no sólo acoge a los ancianos de la comarca del Deza … Últimamente en la residencia habitan también tres niños. ‘The nursing home in Lalín not only admits the aged from the Deza area. As of late, three children have lived in the nursing home, too.’ (3VOZ: 32, 2) a. Los recién casados aparecieron poco después en el automóvil descubierto, abriéndose paso a duras penas en el tumulto. ‘The newly-weds showed up afterwards in the convertible.’ (CRÓNICA: 48, 24) b. Al llegar a este punto, detrás de nosotros aparecen unas personas que permanecen de pie. Son nuestros ‘ángeles’ y van a hablar en nuestro lugar. ‘At this point, behind us, some people show up who remain standing. They are our “angels”, and are going to talk on our behalf.’ (JÓVENES: 45, 17) a. … el hecho es que Pablo Vicario entró solo en la pocilga a buscar los otros cuchillos. ‘… the fact of the matter is that Pablo Vicario walked into the pig sty by himself to look for the other knives.’ (CRÓNICA: 64, 8)


(2.20)

b. La luz no regresaba y por las ventanas mal cerradas entró el silbido del viento. ‘Light was not being restored, and through the badly-shut windows the wind blew in.’ ( JÓVENES: 51, 23) a. Pasa un rato y el niño se despierta. El viejo acude a la alcobita y consigue volverle a dormir. ‘A while goes by, and the child wakes up. The old man goes to the bedroom and manages to put him back to sleep.’ (SONRISA: 158, 8) b. Los niños celíacos de Galicia celebran el domingo su fiesta anual en Santiago . . . A la fiesta acudirán pequeños de las cuatro provincias. ‘The celiac children from Galicia are celebrating their annual party in Santiago on Sunday. Children from the four provinces will go to the party.’ (3VOZ: 28, 3)

It is not difficult to see that in the (b) examples, the locative elements convey to the interlocutor old information or, at least, more familiar or accessible information than the elements appearing as subjects. Thus, the locative elements in these examples occupy a topic position (AVS order), whereas in the (a) examples it is the subject that carries more familiar information and occurs in topic position. As expected, the subject and topic coincide (SVA order) in these examples. Regarding psychological verbs, the structure containing a psychological verb exemplified in (2.14a) has been accounted for in the literature by appealing to the notion of the unaccusative construction. Fernández Soriano (1993: 130), for instance, claims the following: Another type of verbs, closely linked to the unaccusative verbs seems to display properties similar to those of the ‘psychological’ verbs, discussed in Belletti and Rizzi (1987). Succinctly, the hypothesis of these authors is that verbs such as gustar ‘like’, asustar ‘scare’, preocupar ‘worry’, and so on, also have ‘internal’ subjects whose basic postverbal position is normally reserved for the direct objects of transitive verbs accompanied by an agentive subject. The excessive reductionism of the Unaccusative Hypothesis also manifests itself in this type of psychological verb construction. Therefore, the conclusions so for advanced are equally applicable in this case. That is, it is possible to find examples of this construction involving a preverbal subject. Although such instances do not invalidate the Unaccusative Hypothesis entirely, they nevertheless strongly suggest that there are several semantic factors at play regarding constituent order in Spanish. The Arthus corpus provides instances


of the most representative verbs of this group: (2.21)

a. ¿Tienes preferencia por algún pintor concreto? Sí, Sorolla me gusta muchísimo. ‘Do you have any preference for any painter in particular? Yes, I like Sorolla a lot.’ (SEVILLA: 279, 30) b. ¿Qué me pasa? Lo de siempre. Mi estómago vuelve a molestarme. ‘What’s the matter? The same old story. My stomach is bothering me again.’ (CARTA: 87, 20) c. Cuando ríes mucho, algo te pasa. ‘When you laugh a lot, something happens to you.’ (CAIMÁN: 21, 24)

A more promising approach, in our view, is that of Vázquez Rozas (1995, and Chapter 4 of this volume), according to which the two-argument construction (subject–indirect object) is considered a non-prototypical two-argument semantic configuration in that it reveals a low degree of Transitivity based on the parameters proposed in Hopper and Thompson (1980) (cf. Chapter 1, this volume). The semantic differences between syntactic patterns of prototypically transitive clauses and clauses containing a gustar-type-verb construction can be outlined as in Table 2.1 (adapted from García-Miguel 1995: 94). Table 2.1 Comparison of semantic differences and syntactic patterns of prototypically transitive clauses sentences v. clauses sentences with a gustar-type verb construction Subject-Predicate-Direct Obj

Subject-Predicate-Indirect Obj

Dynamic Non-involuntary subject Affected (modified) 2nd argument Animate 2nd argument Animate 1st argument

Stative Involuntary subject 2nd argument not affected physically Animate 2nd argument Animate 1st argument

In fact, some verbs have different features depending on the configuration in which they are used. As we can see in the following examples from Arthus, the intransitive meaning, which corresponds to the pattern with indirect object, shows a departure from a prototypical transitive sentence: (2.22)

a. El ridículo personaje de Action Française admira ahora su propia obra con orgullo casi paterno. (meaning: ‘to esteem, to value’) ‘The ridiculous character in Action Française now admires his own work with almost paternal pride.’ (PAISAJES: 78, 29)


(2.23)

(2.24)

b. Era Julián el que hablaba, y a David le admiró su valentía para contradecir al muchacho. (meaning: ‘surprise’) ‘It was Julián who was speaking, and David was amazed at Julian’s courage in proving the boy wrong.’ (JÓVENES: 140, 32) a. El conquistador invariablemente aprovecha las rivalidades entre los estados y las divisiones internas. (meaning: ‘take advantage of’) ‘The conqueror invariable takes advantage of the rivalries among states and internal divisions.’ (TIEMPO: 88, 34) b. ¿A quién aprovechaba aquel endemoniado desbarajuste? (meaning: ‘be advantageous’) [lit. to whom was advantageous that darn mess] ‘Who benefited from that darn mess?’ (PAISAJES: 15, 4) a. Aunque parezca extraño, en el norte pasamos un calor terrible. (meaning: ‘suffer’) ‘Although it may seem strange, we suffer high temperatures in the north.’ (MADRID: 116, 7) b. Siempre me pasan cosas malas. (meaning: ‘happen, occur’) ‘Bad things always happen to me.’ (SUR: 62, 3)

These examples display differences in the animacy of the participants of the sentences. In the transitive sentences (the [a] examples), the subject is an animate entity and the object is inanimate. The the intransitive sentences (the [b] examples), it is the opposite. Moreover, in the examples (2.22a) and (2.23a) the subject of the transitive acts volitionally, in (2.22b) and (2.23b) the respective subjects do not exhibit any volition at all. Definiteness of the subject Among the semantic properties associated with the constituents of the sentence that favour the appearance of one as a topic, the feature of definiteness stands out and it has attracted a good deal of attention among many researchers. An indication of this attention is the coining of the notion ‘Definiteness Effect’. Essentially, definiteness is said to impose on the NP subject of a sentence the restriction that it must be definite (Belletti 1987: 173). With regard to the order of constituents in a sentence, the Definiteness Effect means that definite phrases (that is, proper nouns, personal pronouns, phrases headed by definite determiners or possessives and some quantifiers) cannot form part of a presentative structure (in English with expletive there), nor can they be the only argument of a sentence whose verb is unaccusative. In other words, in Belletti’s view, unaccusative verbs are constructed with an indefinite subject generated in object (postverbal) position at D-structure, that is, at deep structure. Applying this analysis to present-day Spanish, it is easy to find counterexamples to this principle in the relatively extensive database used for this study.


The following small sample shows that so-called ‘unaccusative verbs’ also appear with definite subjects, preverbally (2.25) as well as postverbally (2.26): (2.25)

(2.26)

a. Le ofrecí mi vida a cambio de tu salvación. Yo moriría antes de cumplir los diez años. ‘I offered him/her my life in exchange for your salvation. I would die before turning ten.’ (SUR: 22, 25) b. La historia ocurrió antes de 1930. ‘The story happened before 1930.’ (HISTORIAS: 101, 10) c. Es el atardecer. El día va a decaer. Tu vecina pasa. ‘It is getting dark. The day is going to fade. Your neighbour walks by.’ (DIEGO: 87, 19) a. ANA. – Estábamos contentos y teníamos miedo. ‘We were happy and scared.’ TEO. – Claro, había muerto Franco. ‘Of course. Franco had died.’ (AYER: 50, 14) b. [Fue] una imposición militar extranjera la que estableció el comunismo en Polonia. Ocurrió lo mismo, con pequeñas variantes, en Checoslovaquia. ‘It was a foreign military imposition which established communism in Poland. The same happened, with small variations, in Czechoslovakia.’ (TIEMPO: 195, 31) c. Ha pasado la medianoche y el tiempo es tormentoso. ‘Midnight has passed, and the weather is stormy.’ (COARTADA: 23, 3)

On the other hand, it is questionable to justify the postverbal position of an indefinite subject on the grounds of verb class, since it would be difficult to account for the instances of indefinite subject NPs with verbs that are considered unergative, such as the examples in (2.27): (2.27)

a. A esa hora asoman una tras otra las estrellas y canta lejos alguien que vuelve del campo. ‘At that time, the stars come out one after the other, and, in the distance, somebody coming back from the countryside sings.’ (SONRISA: 322, 23) b. Ahora estudia mucha gente. ‘Now many people study.’ (MADRID: 151, 19) c. Ladra un perro; ladra agresivamente. ‘A dog is barking. It is barking aggressively.’ (DIEGO: 88, 22)


These examples, however, are not direct counterexamples to the Definiteness Effect. Instead, they reveal two fundamental drawbacks that weaken it. The first is that it is presented as a rule but in fact reflects a tendency: in Spanish, definite phrases do not obligatorily precede the verb nor do non-definite ones follow it, although oftentimes this is the case. The second drawback is that the restriction is linked exclusively to verb class, when actually we are dealing with a semantic feature linked to the subject argument. This is why the tendency towards postposing non-definite phrases is detected in both unaccusative and unergative structures, as evidenced in the examples we have seen so far. In fact, at the time when Milsark (1974, 1977) established the notion of the ‘Definiteness Principle’, the so-called ‘thematic hierarchies’ or ‘topicality hierarchies’ began to proliferate.11 These hierarchies make it evident that the tendency for an argument to function as Topic (that is, to occupy a sentenceinitial position) is directly related to certain semantic features of the arguments, such as animacy and definiteness. Such features cannot be reduced to a dichotomy. In other words, the influence of definiteness on the tendencies found in argument pre-or postplacement cannot be accounted by means of a purely formal interpretation based on a definite–indefinite binary opposition. In our view, it is clearly advantageous to adopt a logical tradition to grammatical studies, in which definiteness is interpreted as that which is identifiable. We can, for example, appeal to theoretical usefulness in explaining definiteness within the more general framework of reference. Thus, we can easily justify why a definite referential subject is more likely to occupy a topical position and a non-referential subject more likely to occupy a non-topical position. Givón (1984: 434) states that the reason for this is that human communication fundamentally (or prototypically) deals with real events and individual referential entities. Likewise, humans typically talk about animate and agentive entities that control some action, hence the propensity to organize sentences primarily around real and concrete entities identifiable in discourse by the interlocutor. Moreover, it is specific individuals (rather than types or classes) that carry out actions. This tendency can be visualized through the hierarchy in Table 2.2. Therefore, proper nouns occupy the highest position in the hierarchy since they are inherently definite, which accounts for the high frequency with which they appear in preverbal position. In addition to proper nouns, Table 2.2 (In)definiteness, (non-)referentiality and topichood Definite

Indefinite Topic

Non-Referential


personal and demonstrative pronouns and phrases with demonstrative determiners share the highest position in the hierarchy. As the examples in (2.28) illustrate, the NPs occupying the highest position in the hierarchy appear in preverbal position: (2.28)

a. Toniolo murió enseguida, la malaria. ‘Toniolo died very quickly, malaria.’ (SONRISA: 163, 31) b. Inf. A. – Mm. Nosotros hablamos a la mañana. ¿Te acordás? ‘Uhm. We talked in the morning, remember?’ (BAIRES: 42, 1) c. ¿Puedo lavarme las manos? Las llevo manchadas de pintura y esto delata más que nada. ‘Can I wash my hands? I have paint on them, which gives one away more than anything else.’ (HOMBRE: 21, 11) d. Estas heridas sanarán dentro de unos días. ‘These wounds will heal in a few days.’ (ZORRA: 44, 13)

At the other end of the hierarchy, we find determiner-less NPs, or NPs without determiners or modifiers (bare nouns), usually employed to convey the generic sense. These occupy a non-topic position. The Arthus data provide numerous instances of bare NPs of one-argument sentences in postverbal position and some are listed in (2.29) (see also 2.2b above): (2.29)

a. Pasaban con altoparlantes y después empezó a desfilar gente con las banderas. ‘They were passing with bullhorns, and then people carrying flags started to parade.’ (BAIRES: 74, 10) b. Apenas quedaba ya luz en la habitación. ‘There was hardly any light left in the room anymore.’ (CAIMÁN: 74, 13) c. Empieza a soplar viento. Una nueva tormenta parece acercarse. ‘The wind is beginning to blow. A new storm seems to be approaching.’ (2INFAN: 90, 15)

In López Meirama (1997a), we supply and comment in detail on statistical data with reference to one-argument sentences. Summarizing those findings, we found roughly 85 per cent preplacement of subjects that were relatively high on the hierarchy (those with a proper noun, a personal pronoun or a deictic determiner), and around 90 per cent postplacement of subjects relatively low on the hierarchy (noun phrases without determiner). Such findings underscore the argument we have proposed that the placement of the subject relative to the verb is not only a matter of verbal semantics. It is for this reason that we propose speaking of unergative and unaccusative construction types rather than of unergative and unaccusative verbs.


This approach also allows the consideration of other significant factors that affect subject order, a key one of which is variable definiteness of the subject.

Discourse-pragmatic factors Introduction Unlike the wider range of interlinguistic variability in the ordering of syntactic elements – typology, for example, has established three fundamental types of basic word orders for transitive sentences: SOV, SVO and VSO – ordering on the discourse-pragmatic level is more homogeneous. Thus, the sequence theme–rheme is the most natural order, that is the least marked order, in the majority of world languages. Some linguists have formalized this fact. For instance, Bossong’s (1980: 51) ‘principe de linéarisation thématorhématique’, or Tomlin’s (1986) ‘Theme First Principle’ would govern basic word order, along with other functional principles. It is doubtless the case that Spanish follows this principle, and that as Reyes (1985: 577–8) points out, the theme–rheme pattern tends to prevail even at the expense of the subject–predicate syntactic pattern. This means that although theme and subject frequently coincide, there are many examples in which the theme of the sentence is a functional element other than a subject. In the previous section we discussed that certain semantic features contribute to word-order flexibility. In this section we examine the extent to which word-order variation in Spanish depends on discourse-pragmatic factors. From a discourse-pragmatic viewpoint, the sentence-initial slot is unquestionably a ‘privileged’ position. We shall propose that this position is used for different discourse-pragmatic purposes, mainly for preserving discursive cohesion and for accommodating given (or presupposed) information. Moreover, the sentence-initial position is used for emphasis, since it is an intonationally prominent position. At the same time, it is also true that the discourse-pragmatic functions linked to sentence-initial position can also be rendered through other means. That is, given that new or given information relates to the structure of the text and to other factors such as ellipsis and anaphora, its status is not only expressed by its relative position within the sentence. Furthermore, intonation plays a significant role in emphasizing expression, thus implying that, theoretically, an element can be emphasized in any position in a sentence. Later in this section we provide an analysis of the most relevant cases of marked positioning of subjects due to emphasis. A good number of studies on word order analyse the distribution of elements as a manifestation of specific discourse-pragmatic functions in such a way that a given position is associated with a particular function at the discourse-pragmatic level. As it relates to sentence-initial position, such an approach is not without its problems. At its best, it renders partial analyses that leave many cases unaccounted for. In our view, it is desirable to develop


a theory that allows for the interweaving of all the relevant factors for subject order relative to the verb without blurring them. With particular regard to the notion of theme (and the theme–rheme structure), however, many definitions have been advanced and various semantic and pragmatic properties have been associated with them. To illustrate, we highlight those that we believe are the most common: ‘that which is spoken about, that which the sentence is about’, ‘known or given information’, ‘presupposed information’, ‘element around which the interest of the speaker is centred’, ‘determined or generic element’, ‘element that begins the sentence / preposed to the verb / dislocated to the right or left’. Of all these, the undoubtedly most extended definition is the first one. Hidalgo (2003: 39) notes in this respect that the meaning of aboutness ‘appears in practically all the definitions of the notion [of theme] that are found in the writings and the models that have treated the question’. Hidalgo (2003: 39) finds this important because it ‘reflects the relationship of the term to the pretheoretical and intuitive notion of the term as used in common usage, as “theme or matter which something is about”, such as a book or a movie, that is, a specific fragment of discourse’.12 This will be the concept of theme we will work with in this chapter. In adopting such a general definition as ‘what is spoken about’ or ‘what we speak about’ allows us to associate it with different discourse-pragmatic functions. In general, ‘we’ (in the generic sense) speak of that which is given, presupposed, or inferable information. Discourse will be coherent if what is spoken about allows us to link what will be communicated with what has already been communicated. However, the theme can be sometimes an element that a speaker wants to highlight, and does not have to be necessarily given or inferable information. Later in the chapter we analyse the most relevant cases of marked theme in Spanish. The theme and other discourse-pragmatic functions The presence of different discourse-pragmatic functions associated with syntactic elements has been a constant point of debate in the grammatical tradition surrounding the Spanish language, although it has generally been discussed from an impressionistic viewpoint. In particular, some manuals of grammar provide a definition of the sentence that is not strictly syntactic, whereby the subject–predicate distinction is comparable to the theme–rheme dichotomy (which further implies that subject and theme coincide in unmarked cases). That is, grammarians have long believed that speakers mold the information into fully meaningful units that in turn are divided into two segments. Functional Grammar has overcome the limitations of traditional grammar by separating the syntactic level from the discoursepragmatic (or informational) one (see, for instance, Dik, 1978, and Givón, 1984). With respect to the latter level, we espouse Halliday’s concept of thematic structure (as opposed to syntactic structure), whereby a sentence is divided into theme and rheme.


However, the distinction between the syntactic and discourse-pragmatic levels cannot per se clearly establish the concept of theme since, as we have pointed out, the element the sentence is about (which in Spanish occupies the preverbal position) is generally the initial position and associated with different informational factors.13 It would, therefore, be convenient to clearly differentiate these factors from one another. In other words, the thematic or preverbal position is discourse-pragmatically relevant, and speakers use it for several communicative purposes. Old/given information The sentence-initial position is often occupied by given (or old) information, which has led to its being identified with the theme, a view that attained great prestige especially in the 1970s, thanks to the work by the Prague School, and the so-called ‘functional sentence perspective’. Among the authors that have applied this perspective to the word order in Spanish are Contreras (1976), Silva-Corvalán (1977), Delbecque (1979), Morales de Walters (1982) and de Miguel (1989). Identifying given or old information with the theme has a serious drawback: the distinction between old and new information does not constitute a dichotomy but rather a continuum, whereas the theme–rheme structure is binary, since the placing of an element in a sentence can only be preverbal or postverbal. Silva-Corvalán (1984: 4), for instance, claims that ‘even though it is possible that there is no old/new information continuum, these values are surely relative’. That is, the information can be more or less old or new, and word order ‘responds to the relative values of newness of the information’ (1984: 4). Along similar lines, Mendieta and Medina (1997: 459) analyse the notion of ‘pragmatic function’ among the variables used in their analysis of preverbal objects. The authors argue for the existence of four different functions, which are defined in terms of their degree of newness, and which are ordered along a four-point continuum: (a) the relation function; (b) lexical cohesion by repetition; (c) lexical cohesion by association; and (d) rhematic anteposition. In the first three cases, a preverbal object introduces given information (although in [c] the information is old for been deduced from the text or the situation), whereas in the fourth case (d) (rhematic anteposition) an object introduces new information. In fact, there are many examples in which the identification between theme and given information is untenable as shown in the following examples (the first one appears in Mendieta and Medina [1997: 461]) (for this section, we have underlined the item that occupies a thematic position): (2.30)

a. … pero no les aconsejo que vayan allí, porque las chinches las sacan a peso. ‘… but I advise you not to go there, because bedbugs they take out by the truckload.’ (México, p. 100)


b. ¡Madonna, déjame vivir un mes más que el Cantanotte, por favour! ¡Un cirio te llevaré, el más gordo que encuentre! ‘Madonna, let me live one month more in the Cantanotte, please! A candle I will bring you, the fattest one I can find!’ (SONRISA: 98, 4) c. NÉSTOR – Espera. Yo querría saber, Rufina, si Charito puede seguir yendo a los ensayos. ‘Wait. I wanted to know, Rufina, if Carlos may continue going to rehearsal.’ RUFINA – Que haga lo que quiera. Remedio ya no tiene. ‘He can do whatever he wants. It’s hopeless now.’ (CAIMÁN: 78, 5) Examples (2.30b) and (2.30c) support the hypothesis put forward by Mendieta and Medina (1997: 461) that preverbal objects introducing new information tend to be associated with the feature [-definite]. This hypothesis is by no means original. The association between the feature [definite] of an argument and its capacity to express old/new information has already been addressed in studies on word order since thematic hierarchies began to spring up. In these hierarchies, however, the relationship is extended to the theme–rheme dichotomy, and can be represented as in Table 2.3. Table 2.3 The theme–rheme dichotomy Theme

Rheme

Old information [definite]

New information [definite]

Nevertheless, the possibility that a non-definite element may function as theme in Spanish has led some researchers to reject the conceptualization of theme as ‘given information’ and replace it with others. For instance, Reyes (1985) argues that the notion of given information is rather imprecise and that it would be clearer to use the notion of presupposed information,14 which, in her view, encompasses all the degrees of mutual knowledge between speaker and hearer, and allows one to integrate the representation of unknown entities or individuals into the concept of theme. Zubizarreta (1999: 4,219), in turn, defines theme as ‘what the sentence is about’, and claims the following: It should be noted that an indefinite phrase … can also function as theme … We know that the indefinite nominal phrase (unlike the definite one) introduces new information. This proves that the dichotomy of new vs. old information is not appropriate to define the theme of a sentence.


Moreover, the new/old dichotomy is not reflected only in word order, since it is related to other grammatical aspects of languages, such as ellipsis and anaphora. In light of these considerations, we conclude that, even though the element spoken about (the theme of the sentence) constitutes frequently known information, one should not make the exclusive identification between theme given/known information. As Hidalgo (2003: 60) notes, ‘the informational status of discourse referents [their nature as given or new] is a discourse notion, clearly differentiated from the function of theme, understood in the sense of “aboutness”, because of which one cannot identify one aspect with the other in the organization of discourse’. Discourse cohesion Closely related to the discussion so far is the notion of theme as a cohesive element that is necessary for the text (understood as an informational unit) to be constructed coherently, as pointed out by Silva-Corvalán’: The progression from the known to the unknown information creates cohesion in discourse. Therefore, when C (the complement) represents the known information in a sentence, whether because its referent has been evoked or inferred, it is placed in initial position, where it serves a textual-linking function (translation by the editors). (1984: 6) The claim that sentence-initial position is where cohesive elements are usually placed is bolstered by the fact that connectors, whether they are syntactic or discursive, normally occupy this position, at least in Spanish, adding a sense of connection, continuity, and, therefore, coherence to the text, as we can see in the following examples: (2.31)

a. No sé si ayer estabas en la reunión. ‘I don’t know if yesterday you were in the meeting.’ b. Escuché muchas veces aquella melodía que tanto habíamos amado. ‘Many times I listened to that melody that we had loved so much.’

With respect to the notion of theme and thematic structure, there is no identification between these concepts and that of the term ‘cohesive element’, although the initial position is frequently used for this function. This can be demonstrated by the positional tendencies of adjuncts. Consider, for instance, the sentences in (2.32), in which the theme repeats what has been mentioned immediately before: (2.32)

a. Había dos asignaturas que eran una rémora en la carrera, era el Latín, por una parte, y el Arte. El Arte no había más remedio que aprobarlo.


‘There were two courses that were an obstacle in our degree: Latin, on one hand, and Art, on the other hand. Art, you just had to pass it.’ (SEVILLA: 202, 39) b. Mi padre no era de aquí, pero vivió casi toda su vida en Sevilla. ‘My father was not from here, but lived almost all his life in Seville.’ ¿Y tu madre? ‘And your mother.’ Mi madre creo que nació aquí. ‘My mother, I think she was born here.’ (SEVILLA: 275, 13) We also find cases, given in (2.33), in which the theme is syntactically disconnected with the sentence it begins, but where the cohesiveness of the sentence containing the elements is unquestionable: (2.33)

Inf B. – ¡Yo tengo unos líos!, que ya no sé ni medir los versos, fíjate. ‘I am so confused that I don’t even know anymore how to measure the verses, imagine!’ Inf. A. – Eso te coges cualquier libro, cualquier manual. ‘[For] that, you get yourself any book, any manual.’ (MADRID: 410, 5)

Therefore, there is an undeniably close relationship between sentence-initial position and discursive coherence, but we repeat that such a link cannot lead us to an absolute link between theme and a cohesive element. Cohesion can also be achieved through means other than word order, such as intonation or the use of deictic elements, particularly anaphoric elements. Emphasis The discursive importance of the sentence-initial position also surfaces when a speaker employs it to emphasize a particular element. Jiménez Juliá (2000) claims that the theme has a potentially contrastive function, since it implies a choice of one over another, and thus it is ‘very likely to carry the intonational focus at a given moment, according to M.A.K. Halliday’s description’ (2000: 156, translated from the original). Halliday defines focus as the ‘point of prominence within the message’ (1967: 203); that is, the element that receives the most intonational emphasis as a result of heightened informational interest on the part of the speaker. Halliday claims that a speaker presents the focus as a non-recoverable element from the preceding discourse, that is, it is considered new information. In any case, it is indisputable that theme and focus are not synonymous. We merely point out that an emphasized element very frequently occupies


the rheme (that is, postverbal) position, as illustrated in the following examples (the focus appears in small capitals): (2.34)

a. Y aún pienso que podré resistir mucho más, ir más allá de donde pueda ir cualquier mujer, acercarme a las fieras, a su orgullosa autosuficiencia. Quiero abastecerme YO MISMA de libertad y de aventura. ‘And I still think that I will not be able to hang on much longer, go beyond where any woman can go, approach the beasts, their prideful self-sufficiency. I want to stock up on freedom and adventure MYSELF.’ (ZORRA: 30, 1) b. Y la acompañé, o me acompañó ELLA a mí, en la noche sin tiempo. ‘And I accompanied her, or SHE accompanied me into the timeless night.’ (CAIMÁN: 71, 15)

However, we must reiterate that the functions of theme and focus are not mutually exclusive. Since two features that allow us to identify the focus are on one hand, a marked (that is, unexpected) position within a sentence, and on the other hand, intonational emphasis, it seems clear that a marked theme can become the focus of the predication. Later in this section we will discuss specifically the marked theme. For now, we focus on the most relevant cases of a marked theme in Spanish. The thematic structure of Spanish As we have already pointed out, we take the thematic structure to be a binary mechanism of distributing information through which a theme is presented and something about it, that is, the rheme, is predicated. This informational structure is superimposed, as it were, onto the semantic and syntactic structure in the sense that the elements in the informational structure also make up these semantic and syntactic structures. However, these three structures (syntactic, semantic and informational) do not exist independently of one another, but rather are interconnected in all languages. Apart from the semantic tier, a sentence is thus the result of the superimposition of two structures: the semantic and the informational. In a nominative-accusative language such as Spanish, it is not uncommon that both theme and subject coincide, as illustrated in Table 2.4.

Table 2.4 The overlay between thematic and syntactic structure Thematic structure Syntactic structure

Theme SUBJ

PRED

Rheme DO

IO


Due to this, researchers usually regard a prototypical sentence as one in which the subject is the element spoken about. Thus, in grammars (for example in that of the Academia, 1973), we find examples such as El niño duerme en la cuna, ‘The child sleeps in the crib’, or Juan dio una limosna a nuestro vecino, ‘Juan gave some money to our neighbour’, in which El niño and Juan are both subject and theme. However, it is debatable how respresentative such sentences are in Spanish, given that it is more frequent to find other structures, especially in oral texts, such as aquí duermo ‘here I sleep’, le pegó una paliza ‘s/he beat him/her up’, or Por asuntos comerciales viajaba mucho mi hermano mayor entre España y América ‘For commercial reasons, my older brother would travel a lot between Spain and America’ (in Academia, 1973). In this regard, we make the following two observations. First, not all sentences in Spanish are thematic, nor do they respond to the theme–rheme structure. The reason for this is that the morphological system of Spanish is rich enough to be able to do without obligatory full NPs or overt subject pronouns once they have been introduced into discourse. This is frequently observed in the case of the subject, which is the first candidate to occupy a thematic position, as pointed out earlier. In fact, the majority of subjects in Spanish are ‘morphological’.15 Consider the following fragment from Crónica de una muerte anunciada [Chronicle of a Death Foretold ], by Gabriel García Márquez: Bayardo San Román, el hombre que devolvió a la esposa, había venido por primera vez en agosto del año anterior: seis meses antes de la boda. Llegó en el buque semanal con unas alforjas guarnecidas de plata que hacían juego con las hebillas de la correa y las argollas de los botines. Andaba por los treinta años, pero muy bien escondidos, pues tenía una cintura angosta de novillero, los ojos dorados, y la piel cocinada a fuego por el salitre. Llegó con una chaqueta corta y un pantalón muy estrecho, ambos de becerro natural, y unos guantes de cabritilla del mismo color. Magdalena Oliver había venido con él en el buque y no pudo quitarle la vista de encima durante el viaje. ‘Bayardo San Román, the man who had given the wife back, had come for the first time in August of the previous year: six months before the wedding. He came on the weekly boat with silver-adorned saddlebags that matched the buckles on his belt and the rings on his ankle boots. He was around 30 years old, but concealed his age very well, because he had the narrow waist of a bullfighter, golden eyes, and skin broiled by saltpeter. He came wearing a short jacket and tight-fitting pants, both made of authentic calfskin, and a pair of kidskin gloves of the same color. Magdalena Oliver had come with him on the boat, and had not been able to take her eyes off him during the voyage.’ In the text, there are seven predicates, marked in boldface, which function as nuclei of their respective sentences. Among these sentences, only the first


and the penultimate ones satisfy the characterization according to which the subject and the theme are the same argument as unmarked cases. The other five sentences illustrate the abundance of athematic structures in Spanish: note that once a topic has been introduced into discourse, it is not normally expressed again by an overt NP or pronoun for several sentences (cf. Comajoan, Chapter 3 of this volume, for details). Second, Table 2.4 represents the unmarked theme, which as expected coincides with the subject. However, as just noted, the morphological richness of Spanish may be what underlies the separation of Spanish syntactic and thematic structure: the position of an NP within the sentence is not an identifying feature of the subject, except in extreme cases, such as modernización significa abolición del pasado / abolición del pasado significa modernización ‘modernization means abolition of the past / abolition of the past means modernization’. Thus, the theme is not necessarily the subject and often may be the DO or IOs, as in (2.35): (2.35)

lo he visto en el supermercado. a. A Juan OBJ Juan him.ACC have.1s seen in the supermarket. ‘Juan I saw in the supermarket.’ b. A mi gata le pica la herida. OBJ my cat her.DAT itches the wound ‘My cat’s wound itches.’

Having made these preliminary remarks, let us now discuss the issue of themes coinciding with elements of a sentence other than the subject. Given that the distributional possibilities of constituents in Spanish are many, we will limit ourselves to discussing the most representative cases. More specifically, we will focus on what might be considered a prototypical sentence (that is, independent, active, declarative, affirmative) and examine cases of thematization of the DO. In general, when the literature refers to thematized DOs, it provides examples such as (2.35a). In fact, an examination of a large corpus of presentday Spanish reveals that this example represents the most common type of DO thematization, as illustrated in examples in (2.36), extracted from Arthus: (2.36)

lo puso Arturo, naturalmente. a. El dinero the money it.ACC put Arturo of course ‘The money Arturo put it in, of course.’ (AYER: 67, 18) b. Este muelle lo construyó mi padre antes de que yo naciera. this dock it.ACC constructed my father before that I was.born ‘This dock my father built it before I was born.’ (CINTA: 32, 12) c. Al parque lo habían invadido las grandes hierbas locas. the park it.ACC had invaded the big grasses crazy ‘The park, the big crazy grasses had invaded it.’ (DIEGO: 146, 15)


This type of construction exhibits two basic features that are normally underscored by researchers who have analysed it: first, the existence of pronominal reduplication (Suñer, 1988, and García-Miguel, 1991, among others, speak of an agreement marker); second, a postverbal subject. We might thus claim that theoretically a monotransitive sentence presents two possible distributions of subject and object: SVO (unmarked), and OVS (marked). Nevertheless, the following observations are in order. First, note that most instances of preverbal object are accompanied by non-overt subject. That is, the subject is exclusively expressed by morphological means, since it is usually considered known information. The examples (2.37) corroborate this point: (2.37)

a. ¿Te gustan los toros? A mí no. No, a mí me cansan. ‘Do you like bullfights? Not me. No, they tire me.’ (SEVILLA: 30, 30) b. Esa parte de la historia la conozco mejor. ‘That part of the story I know it better.’ (PASAJERO: 63, 2) c. El «Ya» lo leíamos en casa los domingos. ‘«Ya» we used to read it at home on Sundays.’ (MADRID: 82, 17)

Hatcher (1956) employs the abbreviation ‘O lo V’ to identify this construction. Given that overt subject deletion also occurs when the DO occupies a rhematic position, we should refine our nomenclature for accuracy: SVO/VO is an unmarked order in a transitive sentence, whereas OVS/OV (or, in Hatcher’s terms, O–lo–VS / O–lo–V) is a marked sequence, resulting from DO thematization. Secondly, with a lexically realized subject, DO thematization often takes place alongside a postverbal subject, but not in all cases, which means that the thematic position may be occupied by both arguments. Thus, in Spanish there are other possibilities of constituent distribution apart from the ones discussed earlier. In the examples in (2.38), the two elements that make up what we call the multiple theme are numbered with a subscript:16 (2.38)

(2.39)

OSV a. El artículo de Fillmore1 yo2 lo podría haber elaborado mucho mejor. ‘Fillmore1’s article2 I could have developed it better.’ (BAIRES: 36, 38) b. Pero a a Emilia1 esta idea2 no acababa de hacerla feliz. ‘But Emilia1 this idea2 didn’t make her totally happy.’ (LABERINTO: 148, 21) SOV a. ¿El árabe? Yo1 el árabe2 lo aprobé con éste, con P.


‘Arabic? Well, Arabic1 I2 passed it with this guy, with P.’ (MADRID: 418, 3) b. Una buena alfombra1, como una buena capa, todo2 lo tapa. ‘A good carpet1, like a good cloak, covers everything2.’ (HOTEL: 31, 34) There are even cases in which a speaker changes the relative order of both thematic elements, as shown in the example in (2.40), where the speaker wavers when constructing the subject: (2.40)

Pues ha cambiado todo, porque a mí estos … , esta construcción de ladrillos1 a mí2 me enferma. ‘Well, everything has changed, because me, this … , this brick construction1 makes me2 sick. (MADRID: 269, 31)

On the other hand, some studies on themes in Spanish argue that the theme can be in external position of the sentence, which has led researchers to establish different types of themes, and even to distinguish theme from another discourse-pragmatic function. The first approach is found in Zubizarreta (1999: 4,218–24), who, following Cinque (1983), claims that the ‘theme of the sentence’ becomes grammaticalized through two constructions: the Hanging Topic construction and Left-Dislocation. The second approach is epitomized by López García (1996), who distinguishes theme from topic and claims that ‘the theme merely represents a modification of the gradual order of informational succession of the elements in the sentence’, whereas ‘the topic entails a privative view based on configurationality’ (1996: 486 translated from the original). The analysis of a large corpus of real texts in modern Spanish reveals that the dissociation of the theme from a syntactic structure belongs almost exclusively to the realm of colloquial Spanish (since it is practically inexistent in narratives), and it appears to varying degrees. What follows is a synopsis of the most relevant types of dissociation found in our corpus. The theme can be separated from a sentence by means of a pause,17 although our data reveal that, unless the theme is used in a spoken language (or in texts that imitate it, such as dialogues in novels and plays), a pause is primarily used for prosodic reasons when the theme is long, as shown in (2.41): (2.41)

a. A un chico a quien se le escapara un chiste atrevido o un taco delante de una señorita, se le catalogaba inmediatamente como un grosero. ‘Any boy who would let an off-color joke or a swear word slip out before a young lady one immediately would brand him a rude person.’ (USOS: 194, 10)


b. Lo que hizo la Reforma protestante en la esfera de las creencias y los sentimientos religiosos, lo ha hecho la Unión Americana en la esfera secular. ‘What the Protestant Reform did in the realm of religious beliefs and feelings, the American Union has achieved in the secular realm.’ (TIEMPO: 37, 9) There is sometimes a discrepancy between the element in thematic position and the coreferential clitic pronoun, as in (2.42): (2.42)

a. Encuestador: ¿Pero no crees que para esto de la medicina hay que tener mucha sangre fría también? Researcher: ‘But don’t you think that in this medicine business one has to be a coolheaded person, too?’ Inf. – Sí; pero eso la adquieres a lo largo del tiempo. ‘Yes, but that you acquire over time.’ (MADRID: 133, 11) b. Creo que consigue muchas cosas por su inteligencia y otras muchas cosas lo consigue por su dinero. ‘I think that s/he gets many things because of his/her intelligence and many other things s/he gets it through his/her money.’ (MADRID: 97, 18) c. Los estudios superiores me dices que lo hiciste aquí en la Facultad. ‘Your studies in higher education you’re telling me that you did it here in the school.’ (SEVILLA: 41, 4)

That is, in all these examples, there is a lack of grammatical gender and/or number agreement between the lexical DO and the corresponding doubled pronoun: eso – la ‘that-NEUT’ – it-FEM’, otras muchas cosas – lo ‘many other things-FEM – it-MASC’, and los estudios superiores – lo ‘higher education studies-MASC – it-MASC’. We also find a certain degree of dissociation when the preverbal element is a DO without the expected DO marker a, that is, an animate DO without the prescriptively obligatory personal a. (2.43)

a. Mis amigos de la Universidad los adoro todavía. ‘My friends from college I still adore them.’ (SEVILLA: 255, 29) b. De todas formas, muchos de los actores de ahora los conocemos a través de la televisión. ‘Anyway, many of today’s actors we know them through TV.’ (MADRID: 276, 4)


c. Porque, no sé, los demás los encuentro muy distanciados. ‘Because, I don’t know, the others I find them rather distant.’ (SEVILLA: 58, 13) The dissociation is even clearer when the thematized element is not the DO of a transitive sentence, as in the examples in (2.44): (2.44)

a. Y ¿la enseñanza media? La enseñanza media fui a una academia. ‘And secondary education? [For my] secondary education I went to an academy.’ (SEVILLA: 273, 9) b. La gente universitaria está mucho más unida, de forma que Granada, por ejemplo, vas en tiempo de curso o en tiempo de vacación y notas una diferencia enorme. ‘College people are much closer to one another, so Granada, for example, you go [there] during the period of classes, or over a break, and you notice an enormous difference.’ (SEVILLA: 160, 2)

Examples in (2.43) and (2.44) suggest that in extreme cases the theme may become an element outside of a sentence, an element that is similar to conjunctive adjuncts and modals in Halliday’s (1985) sense. In our view, however, the significant difference between these types of elements must be taken into consideration: the discourse-pragmatic value (more specifically, the informational value) of adjuncts is exclusively cohesive. Although these thematized elements qualify as being ‘spoken about’ within the sentence, they cannot be identified with any of the sentence’s arguments because they are not marked as such. Nevertheless, we should bear in mind that these examples almost exclusively belong to spoken language, which is characterized by constructional immediacy. This sense of urgency sometimes generates what is called Anacoluthon. That is, since the thematic position in unmarked cases is occupied by the subject argument, and given that the subject in Spanish lacks formal marking (for example, prepositions), it follows that a speaker may associate non-marking of an element with the grammatical relation of subject. As a result, a speaker sometimes begins his/her utterance with a noun phrase, although he/she then constructs a sequence in which that phrase should be introduced by a linking particle (en/durante la enseñanza media ‘in/during secondary education’; a Granada ‘to Granada’).18 In order to support this account, we offer examples in which an element that functions as theme is introduced by a marking that does not correspond to the meaning of the construction. For example, as the speaker in (2.45) goes along, s/he changes the syntactic structure with which s/he expresses the content s/he wants to communicate. In (2.45), the speaker begins with con Jacqueline ‘with Jacqueline’ as the theme and then marks this adjunct


clause as an indirect object in se le ha creado demasiado mito ‘one has created to her to much myth’: (2.45)

Es que con Jacqueline creo que se le ha creado demasiado mito y no se la considera ya como una mujer igual que las demás. ‘It’s that with Jacqueline I think one has created around her a lot of myth, and she is not considered a woman like the others.’ (MADRID: 97, 25)

Another typical feature of spoken language is that the theme can be introduced by specific marking. In this regard, López García (1996: 485), following Metzeltin (1990), distinguishes what he calls a topicalizer, such as en lo referente a ‘with regard to’, en lo tocante a ‘with respect to’, ‘as for’, en relación con ‘in relation to’, and so on, from the lexical expression of its topicalized nature, which is realized through expressions like hablando de ‘speaking of’, la cuestión es que ‘the thing is that’, and so on. Similarly, Zubizarreta (1999: 4,220–1) points out that what she calls hanging topics can be optionally preceded by the expressions en cuanto a ‘as for’ and con respecto a ‘with respect to.’ It seems clear, then, that this way of marking an element implies a disassociation between the theme and the internal syntactic structure of the sentence. Some examples from Arthus are given in (2.46): (2.46)

a. En cuanto al grupo cuarto, cabe destacar que lo encabezan el Ignacio Gago y el San Mamed. ‘As for the fourth group, we should mention that it is led by Ignacio Gago and San Mamed.’ (1VOZ: 51, 6) b. La carrera en cuanto dificultad de, de superarla, de sacarla creo que no tiene problema mayor. ‘The degree, as far as difficulty of, of bettering it and completing it is concerned, I don’t think it’s a big problem.’ (MADRID: 100, 13) c. En cuanto a la Virgen, todos, creo, que estamos de acuerdo. La más representativa de Sevilla es La Macarena. ‘As for the Virgin [Mary], all of us, I think, agree. The most representative [Virgin] of Seville is La Macarena.’ (SEVILLA: 40, 25)

Another property that Zubizarreta considers typical of the hanging topic, and one which differentiates it from left-dislocation, is that ‘the relationship between the theme and a certain sentence-internal position is that of coreferentiality; there cannot be a relationship of grammatical dependence’ (1999: 4,224, translated from the original). This entails that within a sentence we can find not only a clitic (example 2.47a), but also a stressed pronoun as


in (2.47b), an epithet as in (2.47c), or a phrase with which the theme establishes an inalienable relationship as in (2.47d) (examples from Zubizarreta 1999): (2.47)

a. En cuanto al hermano, parece que los padres lo contemplan mucho. ‘As for the brother, it seems that his parents have him on their mind a lot.’ b. En cuanto al hermano, parece que los padres hablan de él todo el tiempo. ‘As for the brother, it seems that his parents talk about him all the time.’ c. En cuanto al hermano, parece que el desgraciado se lleva bien con todo el mundo, inclusive con los padres. ‘As for the brother, it seems that the poor soul gets along with everybody, even with his parents.’ d. [Context: an argument about Juan’s cars]: En cuanto al BMW, parece que los frenos le fallan constantemente. ‘As for the BMW, it seems the brakes are constantly failing.’

In Arthus, we find examples of coreference with an element other than the clitic, even in the absence of thematization marking (that is, even without expressions such as en cuanto a ‘as for’). Some are listed in (2.48): (2.48)

a. Y luego, claro, coches, pues se veían muy pocos. ‘And then, of course, cars, you would see very few [of them].’ (MADRID: 159, 7) b. No sé, yo, vos sabés que los tanos no tengo nada en contra de ellos, pero … ‘I don’t know, but you know that Italians I have nothing against them, but …’ (BAIRES: 413, 27) c. Bueno, pues, ese animalito, trabajo, trabajo, lo que la gente llama trabajo puede que no tenga mucho. (MADRID: 219, 22) ‘Well, that little animal, work, work, what people call work, he probably doesn’t have much [to do]’ (cf. Clements Chapter 6 this volume, who considers this type a case of null object (pronominalisation)).

Extraposition is also possible in Spanish: an element that belongs to a sentence integrated within a complex syntactic structure is placed at the beginning of the structure, that is, outside of the sentence of which it is a constituent. Rivero (1980) considers such constructions to be left-dislocations,


in which the subject has been fronted: (2.49)

a. A la jovencita no convenía apearla de su pedestal de sueños. ‘The young lady, it wasn’t a good idea to knock her off her pedestal of dreams.’ (USOS: 157, 36) b. Pero en fin, el arte me gusta apreciarlo, me gusta, me encanta, sí. ‘But anyway, art I like to appreciate it, I like it, I love it, yes.’ (SEVILLA: 92, 35) c. Éste de la Gran Vía ya no me acuerdo ahora cómo se llama. ‘This one on Gran Vía, I don’t remember what is called anymore.’ (MADRID: 249, 7)

Arthus, moreover, reveals that although it is possible to find examples in written texts as in (2.49a), extraposition is more common in spoken language, such as (2.49b) and (2.49c). In addition, many examples are found in reported speech (with matrix-sentence predicates such as creer ‘believe’, saber ‘know, learn’, and suponer ‘suppose’, generally in the first person), suggesting that this structure usually stems from the speaker’s involvement in the text, shown by the example in (2.50): (2.50)

La mentalidad creo que no me atrevo a decir que sea distinta. ‘The mentality, I think I don’t dare to say that it’s different.’ (SEVILLA: 269, 18)

As a corollary of what we have discussed so far, we believe that thematic structure in Spanish is highly productive and displays a higher degree of independence relative to syntactic structure than in languages such as French or English. The productive use of thematic structure in Spanish is, in our view, precisely what makes Spanish a so-called free-word-order language. The distributional combinations of elements are manifold, especially in the oral register. However, the possibilities are not arbitrary in that the placing of elements in a particular position depends on different communicative strategies that a speaker may adopt while producing his/her utterance. In the following section, we will show that emphasis also generates a particular word order. Focus and subject position In the previous section we claimed that the discourse-pragmatic structure is a complex network of different factors. This complexity has caused a good deal of confusion and heterogeneity regarding the notion of ‘theme’ and the traits that characterize it (cf. Casielles-Suárez, 2004). One of these factors is contrast, which can be defined as the use of various means by a speaker to draw attention to one of the elements in an utterance. In other words,


contrast is given as a result of a speaker’s choice among a limited number of candidates, that is the elements in an utterance, with the element chosen being the so-called ‘contrastive focus’ (Chafe, 1976). One of the most representative cases of contrastive focus is the equational structure (or cleft and pseudo-cleft sentence), as illustrated in the examples in (2.51) in which different distributional possibilities are shown. In Spanish, this structure has some variants with the connectives si ‘if’ and porque ‘because’, examples of which are given in (2.52): (2.51)

(2.52)

a. Lo que no acepta mi soberbia es el error, mi visión deformada de ti. ‘What my pride cannot accept is the mistake, my deformed view of you.’ ( JÓVENES: 69, 8) b. Pero horchata y zarzaparrilla era lo que se bebía … la CocaCola es zarzaparrilla. ‘But horchata [a sweet milky drink made of tiger nuts] and sarsaparilla was what one would drink. Coca Cola is sarsaparilla.’ (MADRID: 259, 28) c. Son nuestros amigos mexicanos los que me han animado a pensar que puedo ganarme la vida en México, dando lecciones. ‘It was our Mexican friends who encouraged me to think that I can earn a living in Mexico teaching.’ (DIEGO: 40, 19) a. Si algo acertado he hecho en la vida, Emi, te lo aseguro, ha sido compartirla contigo. ‘If something right I’ve done right in life, Emi, I assure you, it’s been sharing it with you.’ (CINTA: 118, 27) b. Diría que lo que sucede hoy aquí es porque no nos aguantamos las unas a las otras ni un cochino día más. ‘I would say that what is happening here today is because we can’t stand one another even one more friggin’ day.’ (HOTEL: 48, 24)

Nevertheless, a discourse-pragmatic factor whose presence is practically constant in human communication (speaking, like living, is the product of successive choices) cannot be expressed only through a few specific structures. In any sentence, one of the constituents can be highlighted as a constrastive focus. In marked modalities, a possibility turns into a rule, and thus, the sentence-initial element in exclamatory sentences and in indirect, partial questions becomes the focus of the utterance:19 (2.53)

a. ¡Ya, ya, BUENAS LABORES debía de hacerle a ella! ‘OK, OK, GOOD WORK [I/she/he] should do for her.’ (SONRISA: 163, 29)


b. ¿CUÁNTAS VECES te habrías adormecido escuchándolo desde tu habitación? ‘HOW MANY TIMES would you have dozed off listening to him from your room?’ (SUR: 40, 21) It is often said that in Spanish prosodic prominence is the mechanism that enables us to identify the focus. The pairs of examples in (2.54) and (2.55) (from Hernaz and Brucart, 1987) illustrate this point: (2.54)

(2.55)

a. Pedro se casará con María. ‘Pedro will marry María.’ b. PEDRO se casará con María (y no Luis). ‘PEDRO (not Luis) will marry María.’ a. María detesta las acelgas. ‘María hates chard.’ b. María detesta LAS ACELGAS (y no los bombones). ‘María hates CHARD (not chocolates).

Unfortunately, such examples are provided without appropriate contexts (even though the authors regard the focus as a discourse-pragmatic value that warrants analysis at the discourse level, instead of the sentence level). They are indeed invented sentences, not examples extracted from recordings of actual speech that are accompanied by the pertinent prosodic study. Arthus is not the most appropriate corpus to find examples containing this discourse-pragmatic value, since it is largely made up of written texts.20 Nevertheless, the information that we have been able to glean from it suggests that speakers usually combine intonational prominence with a modification of the word orders. On some occasions, speakers even make use of specific focalization markings (like the ones for the theme in the previous section) that we might label as focalizers: (2.56)

a. (repetition of [2.34a]) Quiero abastecerme YO MISMA de libertad y de aventura. ‘I want to stock up on freedom and adventure MYSELF.’ (ZORRA: 30, 1) b. Contra las esperanzas de Mercedes, el abuelo no ofrecía indicios de mejoría. Durante los últimos días, había telefoneado con frecuencia a médicos y hospitales o acudido ELLA MISMA a visitarlos. ‘Contrary to Mercedes’ hopes, her grandfather showed no signs of recovery. On the last days, she had often called doctors and hospitals, or SHE HERSELF had gone to visit them.’ (TERNURA: 125, 25)


(2.57)

c. – Anda, anda, a tu siestecita …. ‘There, there, take your little nap …’ – Bueno, pero te acuestas TÚ TAMBIÉN. ‘OK, but YOU go to sleep, too.’ (SONRISA: 332, 36) a. HASTA LA MUERTE la consideras como una prerrogativa exclusivamente tuya. ‘EVEN DEATH you consider a prerogative exclusively your own.’ (CINTA: 22, 20) b. Pero HASTA ESO lo hizo con tanto sigilo que … ‘But EVEN THAT he did [it] in so much secrecy that …’ (CRÓNICA: 50, 24) c. Hace rato que entró en los llanos, ya NI GUARDACRUCEROS hay. ‘It’s been a while that he got into the plains. There are absolutely no GUARDACRUCEROS left’ [i.e., people who wave a flag in order to stop traffic and thus prevent accidents]. (DIEGO: 67, 23)

Based on an examination of these examples, we deduce that the focus is placed in a postverbal position if the subject of the sentence is a focalized element and in preverbal position if the focalized element is a complement. We believe word order in focalization can be more accurately characterized as follows: the focalized element is placed in a marked position, that is non-canonical position, in relation to its syntactic and/or semantic features (cf. Gili Gaya, 1961: 81–97). With regard to subjects, we can claim that, in general, the postverbal position (in particular, the absolute final position; Fant, 1984), is emphatic, but so is the preverbal position if, for instance, the subject is a determiner-less noun. The examples in (2.58) and (2.59) for subject as focus and (2.60) for object as focus (none of which include the focalizing element) serve to illustrate this point: a. ERNESTO – [Hablando de cuatro hermanos] Seguimos igual. Cuando uno está en dificultades, allí acudimos TODOS, ¿verdad? Ernesto: (speaking of four brothers) ‘The situation is the same. Each of us is having problems. EVERYBODY goes there, right?’ (CINTA: 89, 14) b. Miguel no podía dejar de mirar con asombro a un lado y a otro, sin advertir siquiera sus propias toses. Sí las advirtió en cambio LA ABUELA. ‘Miguel could not stop looking around in astonishment, without even noticing his own coughing. THE GRANDMOTHER, however, did notice it.’ (TERNURA: 6, 11) (2.59)21 a. … consultaban en vano el mapa con la nueva nomenclatura de las calles: no entendían ni pío. AMBULANCIAS Y COCHES PATRULLA aullaban inútiles.

(2.58)


(2.60)

‘… they were vainly going over the map with the new street terminology. They did not understand a thing. AMBULANCES AND (PAISAJES: 15, 14) POLICE CARS howled pointlessly.’ b. Se ha combatido contra el extranjero, contra la antipatria: MORAL Y HONOR pueden dormir tranquilos. ‘One has combated foreigners [and] anti-homeland element. MORAL AND HONOR can sleep in peace.’ (HISTORIAS: 57, 14) a. ESTA IMPRESIÓN, si no otra peor, debimos de causarle a la recepcionista. ‘THAT IMPRESSION, if not a worse one, we probably gave the receptionist. (LABERINTO: 181, 24) b. MONTSERRAT – Yo no me mancho las manos con comida. Sólo tomaré una copa de champagne y una pata de pollo con el hueso envuelto en papel de plata. ‘I am not going to get my hands dirty with food. I will only have a glass of champagne and a chicken leg with the bone wrapped in aluminum foil.’ ROCÍO – Pues yo PAPEL DE PLATA puedo darte. De lo otro, nada. ‘Well, I can give you the ALUMINUM FOIL, but none of the other stuff.’ (HOTEL: 66, 37)

From our discussion so far, we can deduce that a focused element in initial position is also a thematic element, apart from being the topic. However, this runs against a widespread stance, especially among generativists, whereby there have to be two different types of fronting, illustrated in the following examples in (2.61) taken from Hernanz and Brucart (1987): (2.61)

a. A Sansón lo traicionó Dalila. ‘Samson, Delilah betrayed [him].’ b. A SANSÓN traicionó Dalila. ‘SAMSON Delilah betrayed.’

From this standpoint, only (2.61b) can be paraphrased with a clefted structure: Fue a Sansón (y no a otro) a quien traicionó Dalila ‘It was Samson (not another man) that Delilah betrayed’. Zubizarreta (1999: 4,240) explains these differences as follows: A preverbal-focus construction is distinguished from left-dislocation not only from a prosodic and interpretative viewpoint, but also from a syntactic one. A focalized direct object does not allow accusative clitic doubling. In contrast, a dislocated direct object obligatorily requires the presence of the accusative clitic. (Translated from the original)


Given this stance, one could conclude that thematic structure and informational structure of contrastive focus are incompatible, since they cannot cooccur in the same sentence. However, this is not always the case. The speaker can combine both structures when s/he thematizes an element belonging to a cleft structure, as shown in the examples in (2.62), taken from Arthus: (2.62)

a. El campo lo que encuentro es que es un poco aburrido. ‘The countryside what I find is that it is a bit boring.’ (SEVILLA: 136, 34) b. Ahora, ¿Esto lo que gasta es gas o petróleo? ‘Now, that, what it runs on is it natural gas or normal gas?’ (BAIRES: 103, 34) c. Y allí los coches lo que no pueden es circular porque nada más es para aparcarlo. ‘And there, cars what they cannot do is drive around, because it is for parking only.’ (SEVILLA: 166, 31)

On the other hand, the concepts of ‘theme’ and ‘focus’ as used here cannot lead us to Zubizarreta’s (1999: 4,227) conclusion that the same phrase cannot function as theme and focus at the same time. In fact, there is no reason why an element which the sentence is about (that is, the theme) cannot be at the same time the result of a choice on the speaker’s part (that is, the focus). It is clear that (2.61a) and (2.61b) are different sequences from a discoursepragmatic standpoint; the presence or absence of the coreferential clitic is not, of course, arbitrary. Although it may be that the Arthus corpus is not the most appropriate source for studying the discourse-pragmatic value of an element of a sentence that manifests itself through intonational emphasis, it is indeed a good source for real oral and written language. The instances provided by the corpus indicate that in a large portion of the cases the coreferential clitic appears in sequences in which the postverbal subject is also the contrastive focus or, to be more precise, in sequences in which there is a postverbal element that is also a focus of contrast. Examine the examples in (2.63) and (2.64), which have been contextualized in order to highlight the contrastive value of the focused elements: (2.63)

Inf. A – Mira, ya al fin y al cabo, estamos lanzados … ‘Look, finally, we have thrown ourselves into it.’ Inf. C – EL SEGUNDO AÑO hacéis, ¿no? ‘You are doing THE SECOND YEAR, right?’ Inf. B – ¡No!, ¡primero! ‘No! First!’ Inf.C – ¿Y son tres?


(2.64)

‘And are they three?’ Inf. B – ¡Cinco! ‘Five!’ (MADRID: 440, 24) Hasta cuarto, lo hice en el mismo colegio que hice la primaria, y ya, después QUINTO lo hice EN EL INSTITUTO, sexto libre y, ahora, C.O.U. otra vez en el instituto. ‘Up until fourth grade, I did at the same school that I did my elementary education, and then THE FIFTH YEAR I did IN THE INSTITUTE, sixth I tested out of, and now the Pre-University Course again in the institute.’ (SEVILLA: 55, 21)

Based on the discussion so far, we believe that the examples in (2.61) can have two interpretations (with the corresponding intonational change shown in small capitals). The sentence (2.61a) can be interpreted (with stress marked) as A SANSÓN lo traicionó DALILA, that is, fue a Sansón (y no a otro) a quien traicionó Dalila (y no otra) ‘it was Samson (not another man) that Delilah (not another woman) betrayed’. It can also be interpretated as A SANSÓN lo traicionó DALILA, that is, fue Dalila (y no otra) quien traicionó a Sansón ‘It was Delilah (not another woman) that betrayed Samson’. In this respect, we have already mentioned that sentences generally lack an overt subject in the majority of cases containing direct object thematization. Therefore, it would be erroneous to think that the postverbal occurrence of this subject is due only to the thematized object. In our view, such an occurrence is also discourse-pragmatically relevant, which may explain why (2.61b), in which the subject is not a contrastive element, is not representative of Spanish, as evidenced in the Arthus corpus where out of a total of roughly 160,000 sentences we found not one example of the type illustrated by (2.61b). A good source of examples (recurrent throughout Arthus) that seems to corroborate this hypothesis can be found in a neuter demonstrative eso used as the thematized element. For instance, the example in (2.65) can be said with an additional coreferential clitic, illustrated in (2.66); such an addition would indicate the focalization of the element in final position, shown in small capitals in (2.66): (2.65)

(2.66)

a. ESO comentábamos con Lita, ¿no? ‘THAT we were discussing with Lolita, weren’t we?’ (BAIRES: 44, 22) b. ESO hizo por mí la Salvinia. ‘THAT, Salivinia did for me.’ (SONRISA: 261, 26) a. ESO lo comentábamos CON LITA, ¿no? (posible respuesta: ‘no, con Laura’) ‘THAT we were discussing WITH LITA, weren’t we?’ (possible answer: ‘no, with Laura’)


b. ESO lo hizo por mí LA SALVINIA. (posible continuación: ‘aunque tú creas que fue Eva’). ‘THAT SALVINIA did for me (possible follow-up:’even though you think it was Eva’). In fact, there are numerous examples featuring postverbal subjects, as in (2.67), in which the context clearly shows that they carry contrastive focus: (2.67)

VÍCTOR – Pero para dejar a salvo el idealismo de tu juventud no dudarás en acusarla de delatora, aún a sabiendas de que es mentira. VÍCTOR JOVEN – ESO lo hiciste TÚ. A mí, ni siquiera me han detenido todavía. VÍCTOR – ‘But in order to salvage the idealism of your youth, you will not hesitate to accuse her of being an informer, even though you know that it’s a lie.’ VÍCTOR JOVEN – ‘THAT YOU did. I haven’t even been arrested yet.’ (HOMBRE: 35, 29)

Our interpretation of this type of structures is a tentative one that could be either confirmed or refuted in a thorough study on contrastive focus in Spanish based on a large corpus of oral texts. To our knowledge, such a study has yet to be carried out, although we do have some general works on the discourse-pragmatic context in Spanish, among which Fant (1984) stands out. The author likens LOS PLATOS los lava FERNANDO ‘The dishes Fernando washes them’ to a clefted structure such as Lo que vende son CALCULADORAS’ What he sells are calculators’. Fant (1984: 76) claims that both sequences are automatically assigned a contrastive value, which allows them to be combined, as illustrated by the example (2.68) taken from Fant (1984: 131). (2.68)

… a Carlos le voy a dar dos tortas si no pone más cuidado al lavar los platos. ‘Carlos I’m going to give [him] a couple of slaps if he doesn’t pay more attention when doing the dishes.’ Espérate, papá, Carlos es quien quita la mesa; los platos los lava Fernando. ‘Wait, Dad, Carlos is the one who clears the table, but it’s Fernando that does the dishes.’

In any case, it is indisputable that contrastive emphasis in Spanish makes use of word order as a common means of expression, and thus becomes another factor that affects the variability in the order of constituents. For this reason, contrastive emphasis should receive more attention in studies


on constituent order in Spanish, to be examined alongside thematization and topicalization.

Summary In this chapter, we have not offered an exhaustive treatment of constituent order in Spanish, but rather we have tried to point out the most relevant semantic and discourse-pragmatic aspects that affect it, using real-text examples from modern Spanish. With respect to the semantic aspects affecting Spanish constituent order, we have examined two: agentivity and the definiteness of the subject noun phrase. Here, we noted the tendency displayed by Spanish to postpose non-agentive and non-definite subjects and we suggested the possibility of recasting two hypotheses discussed in the generative grammar literature (the Unaccusative Hypothesis and the Definiteness Effect) as two parts of a complex array of factors that affect Spanish word order. Although the usefulness of these notions to account for the constituent order phenomena is evident, we consider it risky to establish taxonomic laws (for example verb is an unaccusative verb) with concrete reference to constituent order in Spanish, a language about which it is preferable to speak in terms of wordorder tendencies. That is, we prefer to speak in terms of unergative and unaccusative constructions, not unergative or unaccusative verbs. On the other hand, we have found that different features overlap considerably. That is, in the linear organization of elements in each utterance by a given speaker, various factors interact in each case, semantic as well as discourse-pragmatic factors. The reliance on unaccusativity and/or definiteness alone can become, thus, a reductionist enterprise. As regards discourse-pragmatic factors, we have centred our analysis on thematic structure in discourse, attempting to offer a basic notion of a topic that is still in need of highly detailed studies. Thus, we consider the work presented in this chapter as preliminary in various ways.22 First, it must be noted that the notion ‘theme’ has been defined at times as ‘known/given/ old information’, at others as ‘a cohesive element’, and even as a ‘potentially contrastive element’ (cf. Casielles-Suárez, 2004). In this chapter, where we have preferred the definition of a theme as ‘that which is spoken about in the sentence’, we observed that the theme can frequently be associated with the features just mentioned, but in no way can it be defined by these features. The element in sentence-initial position, about which something is predicated in the sentence, is customarily known/old/given information, but it can also be new information. Such an element is customarily a cohesive element in discourse, but cohesion can also be achieved using other means. At times, a theme may be defined as contrastive, but contrastive focus, apart from the possible trait of emphatic intonation, can be independently characterized as appearing in a non-default position, that is, a marked position. That is, no one trait defines one and only one notion.


We also briefly examined Spanish thematic structure, pointing out its characteristic features. On the one hand, we observed the frequency with which we find athematic clauses in Spanish, something rarely discussed in the literature. On the other hand, we discussed the most characteristic cases of a ‘marked theme’ in transitive clauses, whereby we noted that Spanish uses a wide variety of mechanisms to mark the theme, above all in spoken language. Again, these facts suggest that in studying Spanish constituent order it is more accurate to speak of tendencies or patterns than hard and fast rules. Notes 1

2

3

4 5

6 7 8

9

For this paper we have used data from the Base de datos sintácticos, compiled at the University of Santiago Compostela from a large corpus on present-day Spanish (Arthus). The compilation was made possible through successive research projects under the supervision of Prof. D. Guillermo Rojo. For further information, the reader is directed to http://www.bds.usc.es. Some authors, such as Delbecque (1991: 261), claim that Spanish is an SVO/VS language, whereas others, De Miguel Aparicio (1989: 758) among them, argue that it is an SVO/SV-VS language. This last characterization coincides with Dryer’s (1997) proposed typology. As Givón (1995: 32) points out, ‘the main, declarative, affirmative, active clause has been tacitly assumed, in grammatical description ever since the Greeks, to be the privileged, unmarked clause type (bold is my emphasis).’ This assumption is currently shared by different frameworks. In this chapter, we will not discuss cleft constructions or clauses containing a copula. In López Meirama (1997a and 1997b), we offer a detailed account of the reasons why Spanish is to be considered a subject-initial language. Here we remind the reader that Spanish is considered a nominative-accusative language because both the intransitive- and transitive-sentence subjects (S and A respectively) are marked in the same way, and distinctly from the object (O). As is well-known, in the languages of the world we find two opposing tendencies with regard to topicalization: the more common one consists of the preference for the topicalization of agentive items; the less common tendency, however, reveals a preference for the topicalization of patients. Bearing in mind the syntactic generalizations we are referring to, we can establish two patterns of topic selection: the nominative-accusative {S,A} and the ergative-absolutive {S,O} patterns. Spanish is generally considered a nominative-accusative language, since the subject is usually the most agentive argument of the predication (which, obviously, does not mean that all subjects are agents). This implies that the pattern of topic selection is {S,A}, and, consequently, that the basic order is SVO/SV (however, see Clements, Chapter 5 of this volume, for a discussion of ergative patterns in Spanish). ‘I’ represents ‘indirect object’, and ‘A’ stands for ‘adverbial adjunct’. In the examples, taken from Arthus, the verbal nucleus of each example has been italicized. In relation to this class some researchers also allude to the verb classes in Hatcher (1956), who subdivides them in ‘verbs of existence-presence’, ‘verbs of continuingremaining’, ‘verbs of occurrence’, and so on. See López Meirama (1997a: 158ff) for further discussion.

48 Spanish Word Order 10

Contreras (1978: 74) explicitly states that in addition to presentative verbs accompanied by only one patient, there are others that also include a locative or an expression of origin. 11 Some examples are: Hawkinson and Hyman (1974), Givón (1976), Silverstein (1976), Dixon (1979) and Lazard (1984). 12 The translations from Spanish are ours ( JCC, JY). 13 There are cases in which sentence-initial position is occupied by a cohesive element, or by an adjunct, as in the following examples from our corpus. (i)

Francamente la televisión a mí me aburre. ‘Frankly, television bores me.’ (SEVILLA: 94, 8) (ii) De repente, a Renato le alarma en el viejo cierta expresión doliente. ‘Suddenly, a certain suffering expression of the old man alarms Renato.’ (SONRISA: 142, 33) 14 15

16

17

18

19 20

21

22

That is, the speaker’s supposition about what the hearer can accept as true in each stage of the conversation (Reyes, 1985: 576). A tally of the Arthus data yielded the following figures: out of a total of 149,882 sentences with subject, 53,669 (35 per cent) contain an overt subject and 96, 213 sentences (65 per cent) have a null subject. Jiménez Juliá (2000: 164) proposes that more than one element in preverbal position can be considered having two different structures, made up of either two linked themes, Yo//, de eso// (no tengo) ni idea ‘I, about that, (I don’t have) any idea’, or a presentative theme linked to a focalized unit (in small caps), as in Eso A MÍ no me gusta nada [lit. that to me NEG me pleases nothing] ‘THAT, I don’t like at all’, or A mí ESO no me gusta nada [lit. to me that NEG me pleases nothing] ‘I don’t like THAT at all’. Following Metzeltin (1990), López García (1996: 485) considers the pause as a topicalization strategy in Spanish, in cases as the following: Dulces, no quiero ‘Candies, I don’t want’, or A estos jóvenes recelosos, un exceso de solicitud y de comprensión les parece aún más sospechoso que la rigidez ‘To these distrustful youths, an excess of solicitude and of understanding seems to them even more suspicious than rigidity.’ This could also account for instances of spurious agreement, such as ¿Qué cosas te interesan más conocer? ‘What things are you most interested in knowing?’ (SEVILLA: 75, 24), in which there is agreement between main-clause interesan [interest. PRESENT.3P.] and qué cosas, the DO of the embedded-clause verb conocer. Given that this type of focus is typically accompanied by prosody, the focus element in these examples is highlighted in small capitals. In fact, all cases of focalization are interpretations of this author based on a careful reading of the texts. Thus, the use of Arthus, especially for the examples in this chapter, should be considered a way of illustrating the most prominent tendencies in Spanish constituent order. According to Suñer (1982: 224), the coordination of determiner-less nouns is one of the strategies that Spanish uses to present a situation as contrastive: ‘They do not obey the [Naked Noun Constraint] because they are contrastive; that is, each element of the list receives heavy stress because the focus of the sentence has been attracted to this preverbal subject position.’ Here, the recent work by Hidalgo Downing (2003) could serve well as a point of departure.


References Academia, Real … Española (1973) Esbozo de una nueva Gramática de la Lengua Española. Madrid: Espasa-Calpe. Belletti, A. (1987) Los verbos inacusativos como asignadores de caso. Sintaxis de las lenguas románicas, in V. Demonte and M. Fernández Lagunilla (eds). Madrid: El Arquero, pp. 67–230. Belletti, A. and L. Rizzi (1987) Los verbos psicológicos y la teoría temática. Sintaxis de las lenguas románicas, in V. Demonte and M. Fernández Lagunilla (eds). Madrid: El Arquero, pp. 60–122. Birner, B.J. (1992) The Discourse Function of Inversion in English. doctoral dissertation. Northwestern University, Evanston, IL. —— (1994) ‘Information Status and Word Order: An Analysis of English Inversion’, Language, vol. 70, pp. 233–59. Bossong, G. (1980) ‘Variabilité positionnelle et universaux pragmatiques’, Bulletin de la Société de Linguistique de Paris, vol. 75, pp. 39–67. Butt, J. and C. Benjamin (2000) A New Reference Grammar in Modern Spanish, 3rd ed. London: Edward Arnold. Casielles-Suárez, E. (2004) The Syntax-Information Structure Interface: Evidence from Spanish and English. London: Routledge. Chafe, W.L. (1976) ‘Giveness, Contrastiveness, Definiteness, Subjects and Topics’, in C.N. Li (ed.), Subject and Topic. New York: Academic Press, pp. 25–55. Cifuentes H., José Luis (1999) ‘Inacusatividad y movimiento,’ Revista Española de Lingüística, vol. 29, pp. 35–61. Cinque, G. (1983) ‘Topic constructions in some European languages and “Connectedness”’, Connectedness in Sentence, Discourse and Text in H. van Riemsdijk and K. Ehlich (eds). Tiburg: Katholieke Hogeschool. Contreras, H. (1976) A Theory of Word Order with Special Reference to Spanish. Amsterdam: North Holland (Spanish version: El orden de palabras en español. Madrid, Cátedra, 1978). Delbecque, N. (1979) Etude quantitative d’un phénomène de variation syntaxique: la position du sujet en espagnol. Doctoral Dissertation. Universidad Católica de Lovaina. —— (1991) El orden de los sintagmas. La posición del regente. Gramática española: enseñanza e investigación, vol. II, Gramática. Ediciones Universidad de Salamanca. Dik, S. (1978) Functional Grammar. Amsterdam: North Holland. De Miguel Aparicio, E. (1989) ‘Sujetos invertidos en las construcciones ergativas del castellano. Los conceptos de tema y rema’. Lenguajes naturales y lenguajes formales, IV/2, in Carlos Martín Vide (ed.) Universidad de Barcelona, pp. 753–66. De Miguel Aparicio, E. and S. Olga Fernández (1988) ‘Proceso-acción y ergatividad: las construcciones impersonales en castellano’. Lenguajes naturales y lenguajes formales, III/2, in Carlos Martín Vide (ed.) Universidad de Barcelona, pp. 643–51. Dixon, Robert M.W. (1979) ‘Ergativity’. Language, vol. 55, pp. 59–138. —— (1987) ‘Introduction’, in R.M.W. Dixon (ed.), Studies in Ergativity. Amsterdam: Elsevier Science Publishers B.V., pp. 1–16. Downing, P. (1995) ‘Word order in discourse’, in Pamela Downing and Michael Noonan (eds), Word Order in Discourse. Amsterdam: John Benjamins, pp. 1–27. Fant, L. (1984) Estructura informativa en español. Estudio sintáctico y entonativo. Acta Univ. Upsaliensis 34. Fernández Soriano, O. (1993) ‘Sobre el orden de palabras en español’, Dicenda (Cuadernos de Filología Hispánica) vol. 11, pp. 113–52.

50 Spanish Word Order Forster, J.F. and C.A. Hofling (1987) ‘Word order case and agreement’, Linguistics, vol. 25, pp. 475–99. García-Miguel, José María (1991) ‘La duplicación de complemento directo e indirecto como concordancia’, Verba, vol. 18, pp. 375–410. —— (1995) Las relaciones gramaticales entre predicado y participantes. Lalia, Series Maior, no. 2. Santiago de Compostela: Universidad de Santiago de Compostela. Gawelko, M. (1995) ‘Sobre el orden de palabras básico en español’. Anuario de Lingüística Hispánica (Universidad de Valladolid) vol. 11, pp. 183–92. Gili Gaya, S. (1961) Curso superior de sintaxis española. Vox: Barcelona. Givón, T. (1976) ‘Topic, pronoun and grammatical agreement’, in Charles N. Li (ed.), Subject and Topic, New York, Academic Press. Givón, T. (1984) Syntax: A Functional-Typological Introduction, vol. I. Amsterdam: John Benjamins. —— (1995) Functionalism and Grammar. Amsterdam: John Benjamins. Halliday, M.A.K. (1967) ‘Notes on Transitivity and theme in English, Part 2’. Journal of Linguistics, vol. 3, pp. 199–244. —— (1985/1994)2 An Introduction to Functional Grammar. London: Edward Arnold. Hatcher, A. Granville (1956) Theme and underlying question. Two studies of Spanish word order. Word 12, supplement 3. Hawkinson, A.K. and L.M. Hyman (1974) ‘Hierarchies of natural topic in Shona’, Studies in African Linguistics, vol. 5, pp. 147–70. Hernanz, M.L. and J.M. Brucart (1987) La sintaxis. Barcelona: Ed. Crítica. Hidalgo D., Raquel (2003) La tematización en el español hablado. Madrid: Gredos. Hopper, P. J. and S.T. (1980) ‘Transitivity in grammar and discourse’, Language, vol. 56, pp. 251–99. Jiménez Juliá, T. (2000) ‘Tema en español y en inglés: dos conceptos enfrentados’, Bulletin of Hispanic Studies vol. 77, pp. 153–76. Knauer, G. (1989) ‘Verbo y orden de palabras en la estructura oracional del español actual’ Linguistische Arbeits-Beritche, vol. 68, pp. 43–52. Lazard, G. (1984) ‘Actance variation and categories of the object’, in Frans Plank (ed.), Objects. Towards a Theory of Grammatical relations. London: Academic Press in pp. 269–92. Levin, B. and M.R. Hovav (1995) Unaccusativity: at the Syntax-Lexical Semantics Interface. Linguistic Inquiry Monographs 26. Cambridge, Massachusetts: The MIT Press. López García, Á. (1996) Gramática del español II. La oración simple. Madrid: Arco Libros. López Meirama, B. (1997a) La posición del sujeto en la cláusula monoactancial en español. Lalia, Series Maior, no. 7. Santiago de Compostela: Universidad de Santiago de Compostela. López Meirama, B. (1997b) ‘Aportaciones de la tipología lingüística a una gramática particular: el concepto orden básico y su aplicación al castellano’. Verba, vol. 24, pp. 45–82. Mendieta, E. and I. Medina (1997) ‘Anteposición de objeto en el habla culta de México y Madrid’. Revista Española de Lingüística vol. 27, pp. 447–77. Metzeltin, M. (1990) Semántica, pragmática y sintaxis del español. Heidelberg: Gottfried Egert Verlag. Milsark, G. (1974) Existential Sentences in English. PhD dissertation. Massachussets Institute of Technology. Cambridge MA. —— (1977) ‘Toward an explanation of certain peculiarities in the existential construction in English’, Linguistic Analysis, vol. 3, pp. 1–30.

Belén López Meirama 51 Morales de W.A. (1982). ‘La posición del sujeto en el español de Puerto Rico a la luz de la clase semántica verbal, la oposición tema-rema y el tópico oracional’, Lingüística Española Actual, vol. 4, pp. 23–38. Reyes, G. (1985) ‘Orden de palabras y valor informativo en español’, in Julio Fernández Sevilla et al. (ed.), Philologica Hispaniensa in Honorem Manuel Alvar II. Madrid: Gredos pp. 567–88. Rivero, M. L. (1980) ‘On left-dislocation and topicalization in Spanish’, Linguistic Inquiry, vol. 11, pp. 363–93. Silva-Corvalán, C. (1977) ‘A discourse study of some aspects of word order in the Spanish spoken by Mexican-Americans in West Los Angeles’. Unpublished M.A. Thesis. —— (1984) ‘Topicalización y pragmática en español’, Revista Española de Lingüística vol. 14, pp. 1–19. Silverstein, M. (1976) ‘Hierarchy of features and ergativity’, in Robert M. W. Dixon (ed.), Grammatical Categories of Australian Languages Camberra: Australian Institute of Aboriginal Studies pp. 112–71. Suñer, M. (1988) ‘The role of agreement in clitic doubled constructions’, Natural Language and Linguistic Theory, vol. 6, pp. 391–434. Thompson, S. (1978) ‘Modern English from a typological point of view: some implications of the function of word order’, Linguistische Berichte, vol. 54, pp. 19–35. Tomlin, R.S. (1986) Basic Word Order: Functional Principles. Kent: Croom Helm. Vázquez, R.V. (1995) El complemento indirecto en español, col. Lalia, Series Maior, no. 1. Universidad de Santiago de Compostela. Zubizarreta, M.L. (1999) ‘Las funciones informativas: tema y foco’, in Ignacio Bosque and Violeta Demonte (eds), Gramática descriptiva de la lengua española (Chapter 64), (eds) 215–44. Madrid: Espasa Calpe.

Abbreviations used for the original texts cited Essays Bunge, M., Lingüística y filosofía, Ariel, Barcelona, 1983. Martín Gaite, C., Usos amorosos de la postguerra española, Anagrama, Barcelona, 19888. TIEMPO Paz, O., Tiempo nublado, Seix Barral, Barcelona, 1983. HOMILÍA Sánchez Ferlosio, R., La homilía del ratón, El País, Madrid, 1986. LING

USOS

Narrative Aldecoa, J.R., Porque éramos jóvenes, Seix Barral, Barcelona, 1986. Bioy Casares, A., Historias desaforadas, Alianza, Madrid, 1986. CARTA Colinas, A., Larga carta a Francesca, Seix Barral, Barcelona, 1986. GLENDA Cortázar, J., Queremos tanto a Glenda, Alfaguara, Madrid, 19814. CRÓNICA García Márquez, G., Crónica de una muerte anunciada, Mondadori, Madrid, 1987. SUR García Morales, A., El sur (seguido de Bene), Anagrama, Barcelona, 1985. PAISAJES Goytisolo, J., Paisajes después de la batalla, Montesinos, Barcelona, 1982. MIRADA Guelbenzu, J.M., La mirada, Alianza, Madrid, 1987. TERNURA Martínez de Pisón, I., La ternura del dragón, Anagrama, Barcelona, 19883. LABERINTO Mendoza, E., El laberinto de las aceitunas, Seix Barral, Barcelona, 1982. DIEGO Poniatowska, E., Querido Diego, te abraza Quiela y otros cuentos, Alianza/Era, Madrid, 1987. JÓVENES

HISTORIAS

52 Spanish Word Order SONRISA

Sampedro, J.L., La sonrisa etrusca, Alfaguara, Madrid, 1985.

Theatre Alonso de Santos, J.L., Bajarse al moro, Antonio Machado, Madrid, 19872. Buero Vallejo, A., Caimán, Espasa-Calpe, Madrid, 1981. AYER Díaz, J., Ayer, sin ir más lejos, Antonio Machado, Madrid, 1988. OCHENTA Diosdado, A., Los ochenta son nuestros, Antonio Machado, Madrid, 1990. COARTADA Fernán Gómez, F., La coartada, Antonio Machado, Madrid, 1987. HOTEL Gala, A., El hotelito, Antonio Machado, Madrid, 1988. ZORRA Nieva, F., Te quiero, zorra, Antonio Machado, Madrid, 1989. 1INFAN Olmo, L. y P. Enciso, Teatro infantil I, Antonio Machado, Madrid, 1987. 2INFAN Olmo, L. y P. Enciso, Teatro infantil II, Antonio Machado, Madrid, 1987. PASAJERO Reina, M.M., El pasajero de la noche, Antonio Machado, Madrid, 1988. CINTA Reina, M.M., La cinta dorada, Antonio Machado, Madrid, 1989. HOMBRE Salom, J., Un hombre en la puerta, Martín Pancorbo, 1984. MORO

CAIMÁN

Oral speech Barrenechea, A.M., El habla culta de la ciudad de Buenos Aires. Materiales para su estudio, tomo II, Instituto de Fil. y Lit. Hispánicas Dr Amado Alonso, Buenos Aires, 1987. MADRID Esgueva, M. y M. Cantarero (eds.), El habla de la ciudad de Madrid. Materiales para su estudio, CSIC (Miguel de Cervantes), Madrid, 1981. SEVILLA Pineda, M.A. de (ed.), Sociolingüística andaluza 2. Materiales de encuestas para el estudio del habla urbana culta de Sevilla, Universidad de Sevilla, 1983. BAIRES

Newspapers 1VOZ La Voz de Galicia (30 October 1991). 2VOZ La Voz de Galicia (22 November 1991). 3VOZ La Voz de Galicia (23 November 1991).

3 Continuity and Episodic Structure in Spanish Subject Reference Llorenç Comajoan

Subject reference in Spanish In Spanish, as in other Romance languages, human subject reference is encoded by three linguistic devices: lexical noun phrases (NPs), personal pronouns and null subjects (zero).1 For instance, Participant 3 in this study produced the three types of devices to refer to a human referent (el niño, ‘the boy’) in the following narrative segment: (3.1) El niño recoge una de las banastas de de peras y Ø se las lleva con él en la bicicleta. Ah mientras Ø se lleva las peras Ø se cruza con otra niña que viene en bicicleta y, al cruzarse, en un camino que es más bien estrecho, él pierde el sombrero, Ø se despista mirando hacia atrás y Ø se cae. ‘The boy takes one of the baskets of pears and Ø takes them with him on the bike. Ah while Ø is taking the pears Ø passes another girl who is coming on a bike and when passing, in a road that is narrow, he loses his hat, Ø distracts himself looking back, and Ø falls.’ Research in Spanish nominal reference has been conducted within different linguistic theories and methodologies (see reviews in Luján, 1999; SilvaCorvalán, 2001). The major research questions in Spanish functional linguistics have been related to two main topics: the variable use of pronouns in different Spanish varieties and the linguistic explanations for the variability of use. The variability in Spanish nominal devices can be examined by studying the use of devices in different varieties (for example, comparing Puerto Rican and Madrid Spanish) or the use of different devices within one variety (for example, rates of NP, pronoun, and zero in Puerto Rican Spanish). 53

54 Continuity and Episodic Structure in Spanish Subject Reference Table 3.1 Percentage rates of third-person pronouns in different Spanish varieties Pronoun 3rd sing. % 8 10 12 14 26 39 43 48

Pronoun 3rd pl. % 8 9 6 14 16 25 23 22

Variety

Study

Madrid Puente Genil (Spain) Spain Madrid Mexican (Los Angeles) Puerto Rican (San Juan) Puerto Rican (Boston) Puerto Rican (New York City)

Cameron (1992) Ranson (1991) Rosengren (1974) Enríquez (1984) Silva-Corvalán (1994) Cameron (1992) Hochberg (1986) Flores-Ferrán (2002)

Most Spanish speakers intuitively notice that Spanish subject pronouns are more frequent in certain varieties (such as Puerto Rico Spanish) than in others (for example, Madrid Spanish), and research results have shown that different varieties of Spanish indeed display different rates of pronoun/NP use. Table 3.1 displays results from studies in different varieties that have focused on third person nominal reference and shows that the percentage use of pronouns ranges between 8% and 48%. However, the data on the variability of use of pronouns need to be taken cautiously, because the types of discourse, the total number of nominal devices under study, and the syntactic positions are not always the same. Table 3.2 includes a summary of studies that have specifically studied nominal reference. Some of the discourse types for these studies include written literary texts (Rosengren, 1974), semi-guided or spontaneous personal oral conversations (Givón, 1983; Bentivoglio, 1983; Cameron, 1992; Flores-Ferrán, 2002; Ranson, 1991), and elicited spoken narratives (for example, using The Pear Film, Blackwell, 1994; Clancy, 1980). Some studies (for example, Bentivoglio, 1983, 1992; and Blackwell, 1994) analyse three linguistic devices, but others (such as Cameron, 1992; Flores-Ferrán, 2004) have included only zero and pronouns. The role of different syntactic contexts in the analysis of nominal reference is evident in the data from Blackwell (1994). When the use of nominal devices in subject and object position were analysed, full NPs and zero were the most common devices (38% and 36%, respectively), followed by pronouns (26%). However, when only anaphoric instances in subject position were examined, the proportions differed (90% were zero, 5% were pronouns, and 5% were NPs). Table 3.2 shows that the use of pronouns is variable (ranging between 20% and 45%) and that zero and NPs are used in higher proportions than pronouns.

Llorenç Comajoan 55 Table 3.2 Percentage rates of zero, pronoun, and full NPs in different Spanish varieties Zero 40 56 36 90 55 79 55

Pronoun 23 24 26 5 45 21 45 40 39 24 20 22

Full NPs 37 20 38 5

Variety

Study

Caracas Caracas Aragon (Spain) Aragon (Spain) Puerto Rican (San Juan) Madrid Puerto Rican (New York City) Puerto Rican (Boston) Puerto Rican Puente Genil (Spain) Madrid Mexican (Los Angeles)

Bentivoglio (1983) Bentivoglio (1992) Blackwell (1994) (S and O) Blackwell (1994) Cameron (1992) Cameron (1992) Flores-Ferrán (2004) Hochberg (1986) Morales (1986) Ranson (1991) Enríquez (1984) Silva-Corvalán (1994)

Functional approaches in the study of noun referentiality In order to explain the variable use of NPs, pronouns, and zero in Spanish, researchers in functional linguistic approaches have isolated a series of internal and external factors (Flores-Ferrán, 2002; Morales, 1982; Silva-Corvalán, 2001). Internal factors include switch reference, person/number of the verbal inflection, tense-mood-aspect of the verb, verbal semantics (for example, perception v. cognition verbs), ambiguity, functional compensation, and specificity. External factors include narrative style, contact with another language, social class, and the speakers’ gender and age. The most studied internal factor is switch reference, which describes two types of relationships between NPs: ‘When these two NPs have different referents, they are “switch” in reference. When these two NPs share the same referent, they are “same” in reference’ (Cameron, 1995: 4). Studies following a variationist approach have shown that switch reference is a strong factor in the use of Spanish NPs. For instance, Silva-Corvalán (1982) examined conversational data of 24 speakers in the Spanish-speaking community of West Los Angeles and studied what variables could account for the (non)expression of subjects. The results showed that the factors that favoured absence of subject were old information, unambiguous verb, and same reference; whereas the factors that least favoured absence of subject were ambiguous verb form, new information, and switch reference. In a comparison of San Juan and Madrid Spanish, Cameron (1994) found that 66% of pronouns were used to switch reference in San Juan Spanish (compared to 35% to refer to the same reference). In Madrid Spanish, the percentages were

56 Continuity and Episodic Structure in Spanish Subject Reference

38% and 14%, respectively.2 Further research has emphasized the importance of going beyond switch reference and has shown that degree of discourse connectedness, reference chains, and perseveration need to be taken into consideration as well (Bayley and Pease-Alvarez, 1997; Cameron, 1995; Cameron and Flores-Ferrán, 2004). This study takes previous research into consideration and expands on it by examining two approaches to the study of nominal reference in Spanish: the distance and episodic approaches. The distance approach Early studies of noun referentiality derived from Givón’s (1983a) theory of nominal reference and topic continuity, which argued that the function of nominal reference within discourse was to assist the speaker and the hearer in keeping track of the participants in discourse. Givón’s (1983a, b, c) methodology provided quantitative data in three measures of topic continuity: (a) referential distance (look-back) assessed the gap between a referent in a clause and its previous mention; (b) persistence or decay mirrors the previous measure and assessed for how long a referent was mentioned in discourse after first mention; and (c) potential interference or ambiguity examined how several referents interact in discourse (that is, how reference is established when more than one referent are present in the context). Results using this methodology showed that different linguistic devices could be distributed along a continuum of topic accessibility in which zero coded the most continuous/accessible topics, and referential indefinite NPs coded the most discontinuous topics (Givón 1983a: 17, see also Givón, 2001b: 463) (Table 3.3).3 Bentivoglio (1983) applied Givón’s (1983a, c) methodology to analyse Latin American Spanish (Mexico City, Caracas, and Santiago de Chile).

Table 3.3 Nominal devices for topic continuity Continuity

Nominal device

Most continuous/accessible topic zero anaphora unstressed/bound pronouns or grammatical agreement stressed/independent pronouns right-dislocated definite NPs neutral-ordered definite NPs left-dislocated definite NPs contrastive topicalization cleft/focus constructions referential indefinite NPs Most discontinuous/inaccessible topics

Llorenç Comajoan 57

She found that verbal agreement (that is, zero) and stressed pronouns were the most continuous devices, whereas indefinite existentials (‘there is X’) and definite NPs modified by a relative clause were the most discontinuous. As far as the distribution of the different nominal devices, the data from Bentivoglio (1983) showed that zero was the most common device (40%), followed by NPs (37%) and pronouns (23%). In a further study, Bentivoglio (1993) studied Caracas Spanish and found that NPs had different rates of use according to their function. The most common grammatical roles for NPs were locative, object, and subject of an intransitive verb; and the least common was subject of a transitive verb. In addition, most NPs coded inanimate referents, except for subjects of transitive verbs, which were animate in 91% of the cases. Most NPs were used to track referents already introduced into the discourse, 23% were used to code truly new referents, and only 25.5% were used to code human participants. The data from Bentivoglio (1993) included very few human subjects coded by NPs (5% of NPs). Whenever they were used and they coded new referents, they were always (100%) introduced by NPs. These data showed that full NPs can fulfill a variety of roles and that coding for new information was not their major role, except for the case of subjects in transitive verbs, which tended to be human. The episodic approach In a critique of the recency or distance approach used in the studies by Givón and Bentivoglio, Tomlin (1987) argued that this approach had two strengths – generability and simplicity – but that it could not account for two types of counterexamples: NPs that refer to an antecedent found in the previous clause, and pronouns that persist for several clauses (Tomlin, 1987; Fox, 1987). As an alternative to the recency approach, Tomlin (1987: 456) argued for an episodic/paragraph model that ‘considers the alternation between noun and pronoun to be a function of the limited capacity of working memory, which is manifested in the text artifact primarily through its paragraph, or episodic, organization.’ In discussing the strengths and weaknesses of this approach, Tomlin (1987) explained that it was superior to the distance approach, because it could account for noun and pronoun use in different types of text and created a link between language and psychology. Nevertheless, the episodic approach was challenging from a methodological perspective, because it is difficult to define constructs such as paragraph, episode, and focus in texts. In order to test the episodic model, Tomlin (1987) used a set of 21 slide pictures with a story and asked three groups to produce a story in English. Group 1 saw the slides one at a time, group 2 saw the slides in odd pairs, and group 3 saw the slides in even pairs.4 The rationale behind the experiment was that the differences in the pairing of the slides would produce boundaries of episodes that would trigger differences in the use of nouns and pronouns. Thus, if participants used nouns after a boundary (a new pair of


slides), regardless of the distance from the previous mention, it would be considered evidence for the episodic approach; but if participants used nouns and pronouns to refer to previous mention, regardless of the boundaries, it would be considered evidence for the recency approach. The results indicated that all three groups produced similar measures of discourse density (proportion of propositions to episodes) and episode boundary: NPs were used after an episode boundary and pronouns within an episode boundary. In a second experiment, 10 participants watched a videotaped cartoon and produced an online oral description. The data showed that the participants produced more propositions per clause than in the slide experiment, but the proportion of NPs after an episode boundary and pronouns within an episode was the same as for the slide experiment. In sum, Tomlin (1987) provided evidence for the episodic approach, but this study still needs to be replicated in other languages and types of narratives. The current integrates research in switch reference and topic accessibility in order to investigate (a) the use and function of Spanish NPs, pronouns, and zero in Peninsular Spanish, and (b) to what extent the data provide evidence for the distance or episodic approaches.

The study Materials Spanish narratives were elicited using The Pear Film from Chafe (1980), a seven-minute video-film used for studies in cognitive, cultural, and linguistic aspects of narrative production (see Appendix A for a summary of the plot). The data in this study come from the analysis of approximately half of the story narrated in the film: from the beginning of the story until the end of the fourth paragraph in the summary of Appendix A (when the boy riding his bike falls and the pears spill out of the basket onto the ground). Participants The 16 participants (11 female, 5 male; average age 29.5) in this study were native speakers of Peninsular Spanish varieties that do not neutralize verbal inflection (Barcelona, Valencia, Mallorca, Madrid, Bilbao, Burgos, Salamanca, Guipúzcoa, Ciudad Real, Palencia, and Valladolid). All participants were students at a United States university at the time of data collection. Their average length of stay in the United States was three years. Procedure The participants were informed that they would participate in a study about the different ways people told stories. They watched the complete film and were asked to retell the story as if they were telling it to somebody who had not seen it. It was stressed that they should tell it in as much detail as


possible. For the recording of the data, the researcher provided the participants with a tape recorder and asked them to retell the story in a separate room by themselves. Coding categories Each narrative was divided into clauses that contained one verbal form. Only clauses with human subject referents in which one of the variables (NP, pronoun, or zero) could be inserted were considered (for example, direct object relative clauses were analysed, but subject relative clauses and clauses with se le were coded in a separate category). Infinitives in modal constructions were analysed as part of the matrix clause. Constructions with the verb parecer ‘to seem’ were not considered, because it was not clear whether their subject was a human referent or an expletive subject. Verb phrases containing comments from the participants were not considered. Each clause was coded into the following categories: referent, introduction of referents, linguistic form, referential distance, and episode boundary. The human referents of the film were the following: man picking up pears, man with the goat, boy, and girl. Each subject was coded as ‘new’ if it was the first mention of the referent. If it was not the first mention, it was coded as ‘different subject’ or ‘same subject’ depending on whether the previous clause had the same or a different subject. For the coding of the linguistic form of subjects, the following categories were used: definite NP, indefinite NP, pronoun, and zero. The category ‘Definite NPs’ included definite NPs, demonstrative adjectives, and otro/a ‘another’. The category ‘Other’ included a few other devices (for example, gerunds, pronominal demonstratives, and los dos ‘both’). Relative clauses were coded as direct object relative clauses (with an NP, pronoun, or zero subject) or relative clauses (with a definite or indefinite NP subject antecedent). Finally, clauses with hay ‘there is/are’ that referred to human subjects were coded as hay-clauses, and clauses with ergative constructions (with se [le]) (for example, se le cae el sombrero ‘the hat falls [off his head]’) were coded as se (le) clauses. For the analysis of the distance approach, the data were coded for referential distance. Referential distance (or look-back) measured how many clauses intervened between one referent and its last mention. Following Givón (1983a, c), new referents were given a measure of 20, and old referents were given a number ranging from 1 (mentioned in the previous clause) up to 20 (the highest number given in this study was 13). Thus, referential distance measured switch/same reference when the look-back value was 1 and a pronoun, an NP, or zero was used. The look-back measure referred to mention of the referent regardless of whether the previous mention was in subject position or not. The procedure for coding episodic boundary and events closely followed Tomlin (1984, 1987). Tomlin defined an event as ‘the total action occurring


between event boundaries’ (1984: 121) and an event boundary as ‘the loci of abrupt visual change within the video’ (1984: 121). Two main types of abrupt visual change can occur in the video segment used for this study: a camera cut and the loss or gain of characters. The analysis identified 28 events (camera cuts) and six episodes (see Appendix B). The relationship between episode boundary and linguistic form (NP, pronoun, or zero) was coded using a system of hits and misses (following Tomlin, 1987). A hit was defined as an NP at the beginning of an episode or the use of a pronoun/zero inside an episode. A miss was defined as an NP inside an episode or a pronoun/zero at the beginning of an episode. Clauses that could not be readily classified into one of the events and episodes of the film were excluded from the percentage analysis of hits and misses and were coded as ‘out’ cases. The excluded clauses were instances of extra information that was not in the event/episode analysis of the film (Appendix B), events that were not retold in the order in which they occurred in the film (for example, descriptions of characters after first mention), clauses with se le, and subject relative clauses.

Results Nominal devices for new and old referents The most common device to code human subject referents in the narratives was zero (68.5%), followed by NPs (29%) and pronouns (2.5%) (Table 3.4).

Table 3.4 Subject nominal devices used in the narratives Participant 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Total

NP 4 7 3 9 6 9 13 5 10 7 4 8 6 12 8 6 117 (29%)

Pronoun 0 0 2 0 1 0 0 0 0 4 1 0 0 1 1 0 10 (2.5%)

Zero 15 8 11 14 16 32 23 9 18 18 19 15 23 24 14 16 275 (68.5%)

Llorenç Comajoan 61 Table 3.5 Nominal devices to introduce new characters Participant 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Total

NP indef.

NP def.

Zero

Pronoun

0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2 (3.5%)

0 (0%)

0 (0%)

3 1 2 1 4 1 2 2 2 1 3 4 1 2 3 2 34 (56.5%)

RC NP indef. 0 1 1 1 0 0 0 2 1 1 0 0 2 1 0 0 10 (16.5%)

RC NP def.

hay NP indef.

Other

0 0 1 0 0 0 1 0 1 1 1 0 0 0 0 1

0 1 0 1 0 2 0 0 0 1 0 0 0 0 0 1

1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0

6 (10%)

6 (10%)

2 (3.5%)

Table 3.6 Nominal devices to introduce same subjects from previous clauses Participant

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Total

NP indef.

NP def.

0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0

0 1 1 1 0 3 4 0 2 0 0 1 2 5 3 3

2 (1%)

26 (10%)

Zero

12 4 9 12 12 24 14 6 12 15 10 9 19 20 12 13 203 (78.5%)

Pronoun

0 0 1 0 0 0 0 0 0 3 0 0 0 0 1 0 5 (2%)

RC NP indef.

RC NP def.

(se) le

0 0 0 0 1 0 2 0 1 3 0 1 1 0 1 1

1 0 0 0 2 0 1 0 0 0 0 0 0 0 0 0

0 0 0 0 0 2 1 1 0 1 1 0 0 0 1 1

11 (4%)

4 (1.5%)

8 (3%)

For the introduction of new characters, indefinite NPs were the most common device (56.5%), followed by relative clauses (RC) with indefinite NPs (16.5%), and hay-clauses (‘there is/are’-clauses) with indefinite NPs (10%). New referents were never coded by zero or pronouns (Table 3.5). The most frequent device to code subject continuity (same subject as previous clause) was zero (78.5%). Pronouns (2%) and indefinite NPs (1%) were barely used (Table 3.6).

62 Continuity and Episodic Structure in Spanish Subject Reference Table 3.7 Nominal devices to introduce different subjects from previous clauses Participant 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Total

NP def.

NP indef.

Zero

Pronoun

RC NP indef.

RC NP def.

(se) le

Other

1 4 0 6 2 6 7 3 4 6 1 3 2 4 2 1

0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 4 2 2 4 8 9 3 6 3 9 6 4 4 2 3

0 0 1 0 1 0 0 0 0 1 1 0 0 1 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 0 0 0 1 0 0 0 1 0 0 0 0 1 0 0

1 0 0 0 0 0 1 1 2 0 0 1 0 0 0 0

0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0

52 (36.5%)

1 (1%)

72 (51%)

5 (3.5%)

0 (0%)

4 (3%)

6 (4%)

2 (1.5%)

Subject discontinuity (reference to a different subject from the previous clause) was marked with zero (51%) and definite NPs (36.5%). The percentage rate of pronouns for discontinuity was very low (3.5%) (Table 3.7). These results confirm the different functions fulfilled by nominal devices in Spanish narratives: Indefinite NPs mostly code new subjects, zero tends to code subject continuity, and definite NPs and zero code subject discontinuity. Subject pronouns were infrequent in the Spanish narratives in this study, and their function is discussed in the following section. The recency approach versus the episodic approach The coding of discontinuity in this study was along the following continuum: zero (51%) definite NP (36.5%) pronoun (3.5%) (Table 3.6). The high percentage of zero to refer to a different subject from the previous clause is contrary to what in the distance approach predicts, because zero is the most continuous device. However, the percentage rates of zero to code continuity (78.5%) and discontinuity (51%) confirm that – although zero can be used to refer to different subjects from the previous clauses – its main function is to code continuity. An analysis of the look-back measure shows the difference between the use of zero for same and different subjects from the previous clause: the average look-back measure of zero when it referred to the same subject (78.5%) was 1, whereas it was 2.5 when it referred to different subjects (4.5 for NP; 3 for pronoun) (Table 3.8). The analysis of hits and misses relating episode boundaries and the use of NP, pronoun, and zero showed that 81.5% of the data were hits and could be accounted by an episodic approach (NPs at the beginning of an episode, pronoun/zero inside an episode) (Table 3.9).

63 Table 3.8 Look-back distance for definite NPs, pronoun, and zero when they refer to a different subject from the previous clause Participant

Def. NP

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1 2, 5, 7, 1

Average

4.5

Pronoun

3 2, 2, 3, 4, 4, 5 3, 2 2, 2, 2, 11, 7 1, 2, 5, 12, 3, 4, 5 3, 4, 4 2, 4, 13, 8 2, 2, 10, 11, 3, 33 2 4, 3, 2 4, 11 2, 3, 6, 7 6, 6 5

5

3

2

3

Zero 2, 4, 4 2, 3, 3, 3 2, 3 2, 2 1, 2, 2, 3 1, 2, 2, 2, 3, 4, 7, 9, 3, 1 2, 2, 2, 2, 2, 2, 2, 3, 3 1, 2, 3 1, 2, 2, 2, 2, 3 1, 1, 2 2, 2, 2, 3, 3, 4, 10, 2, 2 2, 2, 3, 3, 3, 8 2, 2, 2, 2 1, 1, 3, 3 2, 4 1, 2, 3 2.5

Note: 1 indicates that the referent was mentioned in the previous clause, 2 indicates that it was mentioned two clauses back, and so on. The look-back measure refers to mention of the referent regardless of whether the previous mention was in subject position or not.

Table 3.9 Overall hits and misses for the episodic model Participant 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Total

Hit1

Hit2

18 11 11 19 18 32 26 11 18 19 19 14 25 28 17 16

15 10 9 17 14 32 24 9 16 18 16 10 24 25 14 14

Miss 1 3 5 4 4 9 10 0 6 3 1 2 3 7 5 5

302 (81.5%)

–

68 (18.5%) Hit 1

–

267 (80%)

68 (20%) Hit 2

64 Continuity and Episodic Structure in Spanish Subject Reference Table 3.10 Hits and misses for new referents, old referents, and different subjects Hit

Miss

Out

New referents NP indef. Zero

33 0

(100%) (0%)

0 0

(0%) (0%)

1 0

Old referents Same subject NP def. Zero

3 179

(1.5%) (82%)

22 14

(10%) (6.5%)

1 10

19 57

(19%) (57%)

19 6

(19%) (6%)

14 9

Different subject NP def. Zero

New referents were always introduced by indefinite NPs at the beginning of episodes (33 tokens, 100%), In order to avoid a possible bias in favour of the episodic approach, two analyses of hits and misses were calculated. Hit 1 included all the subject referents, whereas Hit 2 excluded new referents. Overall, the difference in the misses between the two counts was less than 2 percentage points (18.5% based on Hit 1, 20% based on Hit 2). Table 3.10 provides an analysis of hits and misses for new and old referents. Overall, the data show that the episodic model can account for most uses of NPs and zero in the Spanish narratives. New referents were always coded by indefinite NPs and were all at the beginning of episodes. Regarding old referents, zero was mostly used inside episodes. Specifically, the use of zero inside episodes represented 82% of the data for subject continuity and 57% for subject discontinuity. The largest percentage of misses in the episodic model occurred in the coding of discontinuity and old referents when they were coded by definite NPs (19%).

Discussion This section discusses the data from this study in light of previous research and focuses on three aspects: the use of NP, pronouns, and zero; the evidence for the recency and episodic approaches; and counterexamples to the episodic model. In the narrative data from this study, human referents were coded by zero or NPs (68.5% and 29%, respectively), and only in a few instances by pronouns (2.5%). Overall, the participants produced more NPs and fewer pronouns than previously reported (Tables 3.1 and 3.2). This difference may be due to the fact that the data in previous studies were conversational and included more than one participant, whereas in the current study participants produced monologic retellings of The Pear Film. The presence of an interlocutor in the retellings is very likely to influence the production of


NPs, pronouns and zero. For instance, Tannen (1980) compared Greek and English narratives of The Pear Film and found differences in the use of nominal devices in the languages in part because the Greek participants tended to assume that the researcher collecting the data had already heard the narrative several times. However, despite differences in the methodology (but using the same film), Blackwell’s (1994) Spanish data on pronoun use in subject position (5-6%) were similar to those of the current study (3.5%). In this study, the nominal devices that introduced new characters were indefinite (indefinite NPs, relative clauses with indefinite NPs, and hayclauses with indefinite NPs). These results confirm Bentivoglio’s (1983) claim that the most discontinuous devices for human subjects were indefinite existentials and presentatives. Bentivoglio’s analysis did not classify relative clauses with indefinite NPs as a separate category, but her data showed that the most continuous device after indefinite existentials and presentatives were definite NPs modified by a relative clause. The strong relationship between indefiniteness and introduction of new characters also supports Givón’s (1983a) continuity hierarchy and Levinson’s (1987) extragrammatical pragmatic principles (Levinson, 1987), whereby The speaker takes into account the hearer’s knowledge and provides statements that are informationally appropriate to the situation. The limited number of subject pronouns (10 tokens) produced by the participants in this study allows for a close examination of their use. The context in which the subject pronouns were used had two characteristics in common: the pronouns referred to the boy in the narrative (8 of the 10 instances), and they occurred in episodes 5 and 6 (see Appendix B) (8 of the 10 instances). The boy was the main character in the story, which can account for the use of pronouns to focus on this character. At first, it could be argued that the subject pronouns in episodes 5 and 6 were used to disambiguate; because in those episodes the boy and the girl are riding their bicycles, and, when they pass each other from opposite directions, the boy loses his hat and falls. In the retelling of this episode, the use of a pronoun may specify who the referent for each verb is. For instance, in (3.2a and b) the subject pronoun él ‘he’ disambiguates in two occasions the referent who loses the hat and falls: (3.2)

Participant 3 (2 tokens)5 el niño recoge una de las banastas de de peras the boy takes one of the baskets of pears y Ø se las lleva con él en la bicicleta and Ø takes them with him on the bike ah mientras Ø se lleva las peras ah while Ø is taking the pears Ø se cruza con otra niña Ø passes another girl que viene en bicicleta y al cruzarse, en un camino que es más bien estrecho,


who is coming on a bike and when passing, in a road that is narrow, (a) ➔él pierde el sombrero he loses his hat Ø se despista mirando hacia atrás Ø distracts himself looking back y Ø se cae and Ø falls (b) ➔y y como consecuencia cae él de la bici, las peras and and as a consequence he falls from the bike, the pears y su sombrero se pierde and his hat is lost y ah además Ø se hace daño en una pierna. and ah in addition Ø hurts himself in a leg. In (3.3), the subject pronoun disambiguates the referent of the person who is distracted: (3.3)

Participant 5 (1 token) el niño este va por ahí con la bicicleta con el cesto el canasto este de de peras en su bicicleta de manera que es bastante dificultosa ¿no? porque pesan y todo esto this boy goes around there on the bike with the basket this basket of pears on his bike in a way that is pretty difficult, right? because they weigh and all this pasa una niña en una en otra bicicleta a girl passes by on another bike Ø se cruza con él Ø passes him ➔él se distrae he distracts himself Ø choca contra una piedra la bicicleta Ø hits the bike against a stone y Ø se y se cae. and Ø falls.

Participant 10 produced 4 subject pronouns, but only one of the tokens had a disambiguation function. The pronoun in (3.4d) makes it clear that the boy was the referent who got off the bike, not the girl: (3.4)

Participant 10 (4 tokens) hay un señor que está subido a una escalera de mano frente a un árbol


there is a man who is at the top of a ladder in front of a tree (a) ➔él está cogiendo fruta verde, unas peras he is picking up green fruit, pears y Ø lo hace de una forma muy artesanal muy calmado así como una a una and Ø does it in a very careful way very calmly like one by one eh nuestro protagonista vuelve a subir al árbol our protagonist climbs up the tree again y Ø sigue haciendo su tarea and Ø keeps doing his task Ø recoge una a una las piezas las frutas Ø picks up the pieces of fruit one by one (b) ➔mientras él está subido while he is up hay una hay una secuencia que there is a scene that eh un niño vestido así como un poco tipo boy scout pasa con una bici ah a boy dressed like a boy scout goes by on a bike y el supuesto boy scout, por llamar de alguna forma, no sé, tendrá también unos doce años and the supposed-to-be boy scout, to give him some name, I don’t know, must also be twelve years old. (c) ➔ah el al cruzarse él se queda mirándola ah when passing he looks at her y el sombrero bueno Ø se chocan bueno no se chocan and the hat well Ø crash into each other well they do not crash into each other eh Ø gira la cabeza ah Ø turns his head around y se le escapa el sombrero and the hat flies from him y nada Ø se tropieza con una piedra and Ø hits a stone y Ø se cae and Ø falls la chica yo creo que ni se da cuenta the girl I think does not realize (d) ➔pero él tiene que bajarse y demás. but he has to get off and all that.


Finally, in (3.5), the pronoun in él mismo ‘he himself’ disambiguates the reference regarding which participant falls (the boy): (3.5)

Participant 15 (1 token) entonces en esto Ø se cruza con una chica then Ø passes a girl y al mirarla se le cae el sombrero and when looking at her his hat falls ➔y él mismo va detrás de las peras. and he goes behind the pears (falls).

However, not all pronouns have a disambiguating function. For instance, the pronouns in (3.6) and (3.7) do not disambiguate, because other contextual information makes it clear who the referent is. In (3.6), the feminine direct object pronoun la ‘her’ unambiguously refers to the girl; and in (3.7), the girl has not been introduced yet: (3.6)

Participant 11 (1 token) y en esto que en el camino ah en pues en la otra dirección viene una chica and in all this on the road ah then in the other direction a girl is coming ➔él se la queda mirando he stands there looking at her y la chica pues le da le quita al chico al niño el sombrero que Ø llevaba. and the girl then takes the hat Ø was wearing.

(3.7)

Participant 14 (1 token) el niño también lleva un sombrero porque hace mucho sol the boy is also wearing a hat because it is very hot ➔en el camino cuando él va con la bicicleta on the road when he is on his bike se cruza con una niña de largas trenzas passes a girl with long braids que también va en bicicleta. who is also on a bike.

The subject pronouns in (3.6) and (3.7) show that the participants used subject pronouns in contexts that are referentially ambiguous (two referents performing similar actions in similar situations) even though the contexts are not linguistically ambiguous. Thus, a subject pronoun may anticipate ambiguity and provide extra information to track reference of the participants.


Finally, in three instances (3.4a, 3.4b, and 3.4c), subject pronouns were used in nonambiguous contexts and did not refer to the boy in the narrative. All three tokens were in subordinate clauses and were produced by the same speaker. Only 6 of the 16 participants produced subject pronouns in their narratives to refer to the boy in episodes 5 and 6. In the same episodes, the remainder of the participants used NPs, non-subject pronouns, or zero (Table 3.11). In sum, these results show that in the Spanish data for this study the expression of nonnew human subject referents was expressed through devices along the following continuum: zero NP pronoun. The ordering of the three devices does not follow Givón’s (1983) topic-accessibility hierarchy (Table 3.3). This may be due to differences in the use of pronouns in different languages. For instance, Clancy (1980) compared English and Japanese data using The Pear Film and found that the Japanese participants produced more NPs and zero than the English participants (Table 3.12). The proportion of use of zero in Japanese and Spanish was similar (73.2% and 78.5%, respectively). Clancy (1980) argued that the difference in the use

Table 3.11 Nominal devices in episodes 5 and 6 Participant

Nominal device

1 2 4 6 7 8 9 12 13 16

se le (unstressed dative pronoun) NP NP NP NP le (unstressed dative pronoun) se le (unstressed dative pronoun) NP NP zero

Table 3.12 Nominal devices to introduce continuous subjects in Spanish, English, and Japanese English

Japanese

Spanish (this study)

NP

15.7%

26.8%

Pronoun Zero

63.8% 20.5%

– 73.2%

10% (def. NP) 1% (indef. NP) 2% 78.5%

Note: The Japanese and English data are from Clancy (1980).


of NPs between English and Japanese was probably due to the fact that Japanese only has two devices for continuity, whereas English has three. Spanish has as many devices as English, but the Spanish data on the use of NPs and zero were closer to the Japanese than the English data. These results indicate that similar devices may fulfill distinct functions in different languages. Thus, pronouns in English can be used anaphorically (in low referential distance) or to mark contrast (in higher referential distance) (Givón, 1983c); whereas in Spanish the anaphoric function is carried out by zero and to a lesser extent by pronouns. Further research needs to compare data elicited through similar tasks in order to replicate these results and study the specific functions of nominal devices within narratives. Further research needs to compare data using the same elicitation task in order to replicate these results and investigate how the same devices in different languages fulfill different functions. The data from this study can shed light on the recency and episodic approaches to noun referentiality. Research following the recency approach found that zero and pronouns were the most continuous devices, followed by NPs of different kinds (Bentivoglio, 1983, table 3.3). The general results for the current study do not provide support for the recency approach, because the data showed that NPs were more continuous than pronouns, and zero was more discontinuous than NPs. Specifically, old referents were mostly marked by zero (78.5%), followed by definite NPs (10%) and pronouns (2%) (Table 3.6); and zero marked both continuity (78.5%) and discontinuity (51%). Table 3.8 showed that the average look-back distance was 2.5 clauses for zero, 3 clauses for pronouns, and 4.5 for definite NPs. This pattern supports the results of the distance approach except that a shorter distance might be expected for zero. The fact that zero persisted for an average of 2.5 clauses suggests that distance alone cannot explain the use of nominal devices, because zero continued to be used even when discontinuity (produced by a new referent, a different human referent, a different nonhuman referent, or a comment from the speaker) occurred. Cameron (1995) referred to reference chains for cases in which same-reference devices (for example, zero) persisted over intervening switch-reference subjects, and he found a gradation in the use of subject pronouns in chains that extended two referents back (higher possibility of pronoun use when the preceding referents were also switch reference in regard to each other and the target). Cameron’s (1995) reference chains capture the limitations of the distance approach but do not explain what triggers the use of a specific nominal device at the end of the chain. A possible explanation is that the nominal devices persist or decay depending on the episodic structure of the text. The analysis of hits and misses in Table 3.10 provided evidence for an episodic model in which pronouns and zero are used inside episodes, and NPs are used to begin new episodes. The percentage values of the episodic model in the current study (81.5% hits, 18.5% misses) were similar to the values in Tomlin (1987) (84% hits and 16% misses). The remaining 16% of the


data in Tomlin (1987) were counterexamples to the episode model and were intraepisode uses of NPs or uses of pronouns at the beginning of new episodes. The intraepisode uses of NPs were found in evaluations, in the performance of one individual speaker, and in cases of ambiguity resolution that were supposed to be filtered out but were not. In this study, no exceptions to the introduction of new referents were found, because they were always introduced with indefinite NPs, relative clauses, or hay-clauses and never by pronouns or zero. Regarding the introduction of same subjects, the intraepisode uses of definite NPs (10%) can be grouped under three categories: 1 Most intraepisode uses of definite NPs were descriptions of the characters, as in (3.8): (3.8)

Participant 15 en la historia aparece un hombre in the story a man appears que está recogiendo peras de un árbol who is picking up pears ➔el señor tiene bigote the man has a moustache Ø tiene un pañuelo en el cuello rojo y un delantal Ø has a red handkerchief around the neck and an apron entonces el hombre se afana en coger muchas peras Then the man labors at picking up many pears que Ø las va bajando del árbol por una escalera. that Ø is bringing them down the tree with a ladder.

2 Three instances of intraepisode definite NPs were provided by two participants, who used them with a demonstrative (este, ‘this’), as in 3.9: (3.9)

Participant 7 ➔y ah mientras este campesino sigue recogiendo la fruta and ah while this peasant keeps picking up fruit viene un niño con un sombrero en la cabeza a boy comes with a hat on his head ➔viene este niño en bicicleta this boy comes on a bike Ø es un niño Ø is a boy que es más pequeño que la bicicleta who is smaller than the bike prácticamente Ø no puede ni ir en su bicicleta ¿no? practically Ø cannot ride on his bike, right?


pero de lo pequeño que Ø es comparado con la bici but Ø is so small compared to the bike ➔ah entonces este este niño que debe tener ¿qué sé yo? unos ah siete u ocho años rubito muy delgadito ah se se da cuenta del enorme cesto de peras. ah then this this boy who must be what do I know? about seven or eight years old blond very skinny ah notices the big basket of pears. 3 Another instance of a miss is when one of the characters is part of a collective subject, as in (3.10): (3.10)

Participant 7 entonces el niño las peras y la bicicleta se van todos al suelo. then the boy the pears and the bike go to the ground (fall).

Finally, in a few instances, intraepisode uses of definite NPs did not have a clear-cut explanation, as in (3.11): (3.11)

Participant 15 entonces el hombre se afana en coger muchas peras then the man labors at picking up many pears que Ø las va bajando del árbol por una escalera that Ø brings them down using a ladder Ø las va dejando abajo Ø is bringing them down ➔entonces cuando el hombre recoge un capazo then when the man takes a basket Ø vuelve a subir arriba Ø goes back on the tree y en esto Ø sigue recogiendo peras ¿no? and Ø keeps picking up pears, right?

Only 6% of the uses of zero that referred to the same subject as the previous clause occurred at the beginning of an episode (Table 3.10). A few of the counterexamples occurred in subordinate clauses that referred to a new episode, whose main clause was part of the previous episode (example 3.12). The others did not have a clear explanation (example 3.13): (3.12)

Participant 3 el niño recoge una de las banastas de de peras the boy picks up one of the baskets of pears y Ø se las lleva con él en la bicicleta and Ø takes them with him on the bike ah mientras Ø se lleva las peras


ah while Ø is taking the pears ➔ Ø se cruza con otra niña Ø passes another girl que viene en bicicleta who is coming on a bike. (3.13)

Participant 11 y la chica pues le da le quita al chico al niño el sombrero and the girl then hits takes from the boy the hat que Ø llevaba that Ø was wearing y entonces en esto de que quitar estar mirándole a ella pues Ø no se da cuenta bien de la carretera and then in all this looking at her then Ø does not realize about the road ➔y en una y ah Ø tropieza eh con una piedra. and Ø hits a stone.

The misses for definite NPs were instances of definite NPs inside an episode (19%). These counterexamples could be classified into two categories: 1 A definite NP to refer to a different subject within an episode usually occurred after a comment from the speaker or a description of a referent: (3.14)

Participant 3 después aparece un niño en una bicicleta también con sombrero por cierto then a boy appears on a bike also with a hat by the way y Ø lleva un pañuelo puesto ah como un cowboy a la manera de cowboy and Ø is wearing a handkerchief like a cowboy, cowboy style ➔el niño recoge una de las banastas de de peras the boy picks up one of the baskets y Ø se las lleva con él en la bicicleta and Ø takes the with him on his bike.

2 Definite NPs to refer to a different subject within an episode were also used to disambiguate (as previously discussed): (3.15)

Participant 4 cuando el niño iba en la bicicleta when the boy was on his bike Ø encontró por el camino otra niña, una niña con dos trenzas, muy guapa,


Ø found on the road another girl, a girl with two braids, very beautiful, que venía en la contrar en la dirección contraria who was coming on the opposite direction ➔y el niño llevaba un sombrero de paja en su cabeza para protegerse del sol and the boy was wearing a straw hat on his head to protect himself from the sun y la niña cuando pasó delante suyo and the girl when passed in front of him Ø se lo quitó Ø took it from him Ø se lo tiró Ø pulled it from him ➔el niño, sorprendido, miró hacia atrás the boy, surprised, looked towards his back y de repente Ø se cayó de la bicicleta con las peras en el suelo and suddenly Ø fell from his bike with the pears on the ground. The misses (6%) in the introduction of different subjects by zero were due to the methodology of the study. Once the episodes and events were identified in the film, the same division of events and episodes was applied to all narratives. Most participants provided more than one clause for each episode, but in a few cases only one clause was provided. In order to be consistant in the coding of the data, these cases were coded as individual episodes, even though from the perspective of the speaker they were likely to be part of the previous episode. To summarize, the analysis of hits and misses supports the episodic model that argues for the use of NP at the beginning of an episode and zero/pronouns inside the episode. The model cannot account for all the data, but most counterexamples can be grouped under meaningful categories (for example, NPs within an episode to describe a character).

Conclusion This study has shown that zero and NPs were the two most common nominal devices to refer to subjects in Spanish narratives. Subject pronouns were used infrequently and mostly in ambiguous contexts. Reference to same and different subjects from the previous clause was expressed by zero. The high use of zero for reference to same and different subjects can be considered


counterevidence to the distance approach (Givón, 1983 a, b, c; Bentivoglio, 1983). The results indicated that a high percentage of zero and definite NPs (80%) can be explained by a model that has the segmentation of narratives into episodes and events as the main variable (Tomlin, 1987). The remaining 20% of the data was explained by some exceptions and other more specialized variables that interacted with episodes and events, such as descriptions of characters, comments from the speaker, subordination, and disambiguation. These findings add to the growing body of research in nominal reference in Spanish. Specifically, they provide evidence for the role of episodes from a cognitive and linguistic perspective. This episodic model overcomes the main weaknesses of the recency approach, namely the use of NPs for antecedents found in the previous clause and the use of zero/pronouns that persist for several clauses. However, further research needs to investigate the episodic model in order to replicate the results found in this study and refine the model. First, future studies need to apply the episodic model in a variety of types of discourse to explore how conversations with several interlocutors differ from retellings produced as monologues. Second, further studies also need to explore the processing mechanisms and narrative strategies that make some speakers use pronouns and others use different nominal devices (only six of the sixteen participants used subject pronouns). Third, the results showed that new characters were introduced by indefinite NPs (isolated or in relative clauses) and hay-clauses, which is evidence for the role of clause structure and syntax in the introduction of new characters. Thus, further research needs to go beyond the study of NP, zero, and pronouns to examine how syntax interacts with nominal reference. Finally, a complete episodic model must incorporate other factors that have been found to correlate with the use of NPs, pronouns, and zero (such as verbal inflection and regional/social variation) in order to ascertain whether episodes and scenes can subsume other factors.

Appendix A: Summary of The Pear Film (from Chafe, 1980: xiii–xiv) The film begins with a man picking up pears on a ladder in a tree. He descends the ladder, kneels, and dumps the pears from the pocket of an apron he is wearing into one of three baskets below the tree. He removes a bandana from around his neck and wipes off one of the pears. Then he returns to the ladder and climbs back into the tree. Toward the end of this sequence we hear the sound of a goat, and when the picker is back in the tree a man approaches with a goat on a leash. As they pass by the baskets of pears, the goat strains toward them, but is pulled past by the man and the two of them disappear in the distance. We see another closeup of the picker at his work, and then we see a boy approaching on a bicycle. He coasts in toward the baskets, stops, gets off his bike, looks up at the

76 Continuity and Episodic Structure in Spanish Subject Reference picker, puts down his bike, walks toward the baskets, again looking at the picker, and lifts up a basket full of pears. He puts the basket down near his bike, lifts up the bike and straddles it, picks up the basket and places it on the rack in front of his handlebars, and rides off. We again see the man continuing to pick pears. The boy is now riding down the road, and we see a pear fall from the basket on his bike. Then we see a girl on a bicycle approaching from the other direction. As they pass, the boy turns to look at the girl, his hat flies off, and the front wheel of his bike hits a rock. The bike falls over, the basket falls off, and the pears spill out onto the ground. The boy extricates himself from under the bike, and brushes off his leg. In the meantime we hear what turns out to be the sound of a paddleball, and then we see three boys standing there, looking at the bike boy on the ground. The three pick up the scattered pears and put them back in the basket. The bike boy sets his bike upright, and two of the other boys lift the basket of pears back onto it. The bike boy begins walking his bike in the direction he was going, while the other three boys begin walking off in the other direction. As they walk by the bike boy’s hat on the road, the boy with the paddleball sees it, picks it up, turns around, and we hear a loud whistle as he signals to the bike boy. The bike boy stops, takes three pears out of the basket, and holds them out as the other boy approaches with the hat. They exchange the pears and the hat, and the bike boy keeps going while the boy with the paddleball runs back to his two companions, to each of whom he hands a pear. They continue on, eating their pears. The scene now changes back to the tree, where we see the picker again descending the ladder. He looks at the two baskets, where earlier there were three, points them, backs up against the ladder, shakes his head, and tips up his hat. The three boys are now seen approaching, eating their pears. The picker watches them pass by, and they walk off into the distance.

Appendix B: Division of the Pear Story into events and episodes

Ev. Ep. Description

Character

1

1

scene-setting, countryside

scene-setting, intro. of man picking pears

2 3 4

1 1 1

man picking up pears in the tree one pear falls to the ground man goes down the ladder, empties pears from the pocket of an apron, and cleans one pear

5

2

6

2

one man comes from the left side, the other man climbs up the ladder, the man has a goat the man and the goat go by the tree and leave

Sec. Sec./Ep. 3

50

5 2 40

intro. of man 25 with goat 6

31


7 1

man picking up pears in the tree

man picking more pears

3

8 1 9 1 10 1

closeup of a hand picking a pear man picking up pears in the tree face of man picking pears in the tree

1 1 3

11 3

a boy riding a bike can be seen from a boy distance man picking up pears in the tree the boy riding a bike is getting closer the boy bikes closer to the tree and gets off his bike the boy touches a basket full of pears and looks at the man up in the tree man picking up pears in the tree the boy looks up and picks a basket, gets on his bike and puts the basket in front of the bike, and leaves man picking up pears in the tree boy riding his bike with difficulty closeup of pears in the basket in front of the bike boy riding his bike

2

12 3 13 3 14 3 15 3 16 3 17 3

18 19 20 21

3 3 3 3

22 4 23 4 24 5 25 5

26 6 27 6 28 6

a girl riding a bike can be seen from a distance boy riding his bike

girl

boy and girl riding bikes from opposite directions, they are going to pass each other closeup of boy, girl passes him, hat flies, and boy notices his hat is off

boy and girl

closeup of a stone, the boy’s bike hits a stone pears fall to the ground boy on the ground, he feels his injured leg

accident, boy

8

69

2 15 8 3 2 25

3 4 2 3 5

7

2 3

6

3

2

14

4 8 185

185

Note: Ev. event, Ep. episode, Sec. seconds of each event, Sec./Ep. seconds of each episode.

Notes 1 2

NP refers to full lexical NPs and zero refers to null subjects. Despite the fact that the relative percentage rates seemed to differ in San Juan and Madrid Spanish, VARBRUL weights were not different. Similar rates were found in New York City Puerto Rican Spanish (Flores-Ferrán, 2004, Cameron and FloresFerrán, 2004). Other studies (Cameron, 1992, 1994; Flores-Ferrán, 2004) have confirmed the central role of switch reference in explaining pronoun use in Spanish. 3 It needs to be noted that, in further research, Givón (1992, 1995) moved beyond discourse, relating his previous studies to cognition (attention and episodic memory) and arguing for the discrete rather than scalar properties of topicality (Givón, 1992: 19).

78 Continuity and Episodic Structure in Spanish Subject Reference 4

Group 1 saw each slide separately, whereas groups 2 and 3 saw slides in pairs. That is, group 1 (singles condition) saw slide 1, followed by slide 2, followed by slide 3, and so on. Group 2 (odd condition) saw slide 1, followed by the pair of slides 2 and 3, followed by slides 4 and 5, and so on. Group 3 (even condition) saw slides 1 and 2 together, followed by slides 3 and 4, and so on. 5 Each line of data represents one clause. The nominal devices under discussion are underlined. The symbol ➔ refers to the line under discussion.

References Bayley, R. and L. Pease-Alvarez (1997) ‘Null Pronoun Variation in Mexican-descent Children’s Narrative Discourse’, Language Variation and Change, vol. 9, pp. 349–71. Bentivoglio, P. (1983) ‘Topic Continuity and Discontinuity in Discourse: A Study of Spoken Latin-American Spanish’, in T. Givón (ed.), Topic Continuity in Discourse: A Quantitative Cross-Language Study. Philadelphia: Benjamins, pp. 259–311. —— (1992) ‘Linguistic Correlations between Subjects of One-Argument Verbs and Subjects of More-Than-One-Argument Verbs in Spoken Spanish’, in P. Hirschbühler and K. Koerner (eds), Romance Languages and Modern Linguistic Theory. Philadelphia: Benjamins, pp. 11–24. —— (1993) ‘Full NPs in Spoken Spanish: A Discourse Profile’, in W. Ashby, M. Mithun, G. Perissinotto and E. Raposo (eds), Linguistic Perspectives on the Romance Languages. Philadelphia: Benjamins, pp. 211–24. Blackwell, S. (1994) ‘A Neo-Gricean Pragmatic Approach to Spanish NP Anaphora’, unpublished PhD dissertation, University of Pittsburgh, Pennsylvania. Brown, C. (1983) ‘Topic Continuity in Written English Narrative’, in T. Givón (ed.), Topic Continuity in Discourse: A Quantitative Cross-Language Study. Philadelphia: Benjamins, pp. 317–41. Cameron, R. (1992) ‘Pronominal and Null Subject Variation in Spanish: Constraints, Dialects, and Functional Compensation’, unpublished PhD dissertation, University of Pennsylvania. —— (1994) ‘Switch Reference, Verb Class, and Priming in a Variable Syntax’, Papers from the 30th regional meeting of the Chicago Linguistic Society: Vol. 2: The Parasession on Variation in Linguistic Theory, ed. by K. Beals et al., pp. 27–45. Chicago: Chicago Linguistic Society. —— (1995) ‘The Scope and Limits of Switch Reference as a Constraint on Pronominal Subject Expression’, Hispanic Linguistics, vol. 6/7, pp. 1–27. —— (1996) ‘A Community-Based Test of a Linguistic Hypothesis’, Language in Society, vol. 25, pp. 61–111. Cameron, R. and Flores-Ferrán, N. (2004) ‘Perseveration of Subject Expression across Regional Dialects of Spanish’, Spanish in Context, vol. 1, pp. 41–65. Chafe, W. (1980) The Pear Stories. Norwood, NJ: Ablex. Clancy, P. (1980) ‘Referential Choice in English and Japanese Narrative Discourse’, in W. Chafe (ed.), The Pear Stories. Norwood, NJ: Ablex, pp. 127–202. Enríquez, E. (1984) El pronombre personal sujeto en la lengua española hablada en Madrid. Madrid, Spain: Consejo Superior de Investigaciones Científicas. Flores-Ferrán, N. (2002) Subject Personal Pronouns in Spanish Narratives of Puerto Ricans in New York City: A Sociolinguistic Perspective. Munich, Germany: Lincom Europa. —— (2004) ‘Spanish Subject Personal Pronoun Use in New York City Puerto Ricans: Can We Rest the Case of English Contact?’ Language Variation and Change, vol. 16, pp. 49–73.

Llorenç Comajoan 79 Fox, B. (1987) Discourse Structure and Anaphora. New York: Cambridge University Press. Givón, T. (ed.) (1983a) Introduction. Topic Continuity in Discourse: A Quantitative CrossLanguage study. Philadelphia: Benjamins, pp. 5–41. —— (1983b) ‘Topic Continuity in Discourse: The Functional Domain of Switch Reference’, in J. Haiman and P. Munro (eds), Switch-reference and Universal Grammar. Philadelphia: Benjamins, pp. 51–82. —— (1983c) ‘Topic Continuity in Spoken English’, in T. Givón (ed.), Topic Continuity in Discourse: A Quantitative Cross-Language Study. Philadelphia: Benjamins, pp. 347–63. —— (1992) ‘The Grammar of Referential Coherence as Mental Processing Instructions’, Linguistics, vol. 30, pp. 5–55. —— (2001) Syntax. Vol. I. Philadelphia: Benjamins. Hochberg, J. (1986) ‘Functional Compensation for /s/ Deletion in Puerto Rican Spanish’, Language, vol. 62, pp. 609–621. Levinson, S. (1987) ‘Pragmatics and the Grammar of Anaphora: A Partial Pragmatic Reduction of Binding and Control Phenomena’, Journal of Linguistics, vol. 23, pp. 379–434. Luján, M. (1999) ‘Expresión y omisión del pronombre personal’, in I. Bosque and V. Demonte (eds), Gramática descriptiva de la lengua española: Volume 1. Madrid, Spain: Espasa, pp. 1277–315. Morales, A. (1982) ‘La posición de sujeto en el español de Puerto Rico a la luz de la clase semántica verbal, la oposición tema-rema y el tópico oracional’, Lingüística Española Actual, vol. 4, pp. 23–37. —— (1986) ‘La expresión de sujeto pronominal en el español de Puerto Rico’, in Gramáticas en contacto: Análisis sintácticos sobre el español de Puerto Rico. San Juan, PR: Playor, pp. 89–100. Ranson, D. (1991) ‘Person Marking in the Wake of /s/ Deletion in Andalusian Spanish’, Language Variation and Change, vol. 3, pp. 133–152. Rosengren, P. (1974) Presencia y ausencia de los pronombres personales sujeto en español moderno. Stockholm: Acta Universitatis Gothoburgensis. Silva-Corvalán, C. (1982) ‘Subject Expression and Placement in Mexican-American Spanish’, in J. Amastae and L. Elías-Olivares (eds), Spanish in the United States: Sociolinguistic Aspects. New York: Cambridge University Press, pp. 93–119. —— (1993) ‘On the Permeability of Grammars: Evidence from Spanish and English Contact’, in W. Ashby, M. Mithun, G. Perissinotto and E. Raposo (eds), Linguistic Perspectives on the Romance Languages. Philadelphia: Benjamins, pp. 19–43. —— (1994) Language contact and change: Spanish in Los Angeles. Oxford: Clarendon Press. —— (2001) Sociolingüística y pragmática del español. Washington, DC: Georgetown University Press. Tannen, D. (1980) ‘A Comparative Analysis of Oral Narrative Strategies: Athenian Greek and American English’, in W. Chafe (ed.), The Pear Stories. Norwood, NJ: Ablex, pp. 51–87. Tomlin, R. (1984) ‘The Treatment of Foreground-Background Information in the On-line Descriptive Discourse of Second Language Learners’, Studies in Second Language Acquisition, vol. 6, pp. 115–142. —— (ed.) (1987) ‘Linguistic Reflections of Cognitive Events’, in Coherence and Grounding in Discourse. Philadelphia: Benjamins, pp. 455–79.

4 Gustar-Type Verbs Victoria Vázquez Rozas

Introduction When comparing the verbal structures of English and Spanish, grammatical accounts underscore the differences between English like and Spanish gustar. Despite their closeness in meaning, these predicates exhibit a divergent syntactic behaviour: whereas like codes as subject the entity that experiences a certain feeling, and as object the stimulus responsible for that feeling, gustar expresses the experiencer through an indirect object (or dative) and the stimulus through the subject, illustrated in the examples (4.1) and (4.2). (4.1) (4.2)

I liked the book. Me gustó el libro.

Other verbs that share structural features with gustar appear in the following examples: (4.3) (4.4) (4.5) (4.6)

A Miguel ya no le apetecía jugar al parchís. (TERNURA: 50, 4) ‘Miguel did not feel like playing Parcheesi anymore.’ El dulce les repugnará. (GLENDA: 57, 11) ‘The cake will disgust them.’ Le pareció reconocer la música que sonaba. (CARTA: 185, 2) ‘S/he thought s/he recognized the music that was playing.’ Nunca una mujer le importó como Viviana. (HISTORIAS: 60, 2) ‘Never before had a woman mattered to him as Viviana did.’

In view of these examples, it could be argued that in Spanish there is a particular selection of a syntactic pattern typical of verbs of feeling or emotion. The situation, however, is more complex since the range of the so-called verba sentiendi (that is, psychological verbs) also comprises verbs 80

Victoria Vázquez Rozas 81

that adhere to a transitive syntactic pattern, with a coding similar to that for English like: (4.7) (4.8) (4.9) (4.10)

La había amado en secreto varios años. (CRÓNICA: 11, 1) ‘He had loved her in secret for several years.’ La Mujer Pirata odiaba al abuelo. (TERNURA: 99, 13) ‘The Woman Pirate hated the grandfather.’ Yo detestaba a los hombres altaneros. (CRÓNICA: 35, 6) ‘I detested haughty men.’ Tía Delia creía haber visto a abuela. (SUR: 49, 10) ‘Aunt Delia thought that she had seen Grandma.’

Therefore, the same semantic class of predicates is represented by, at least, two classes of syntactic constructions.1 This chapter presents a study of gustar-type verbs (GTVs) with the Transitivity Hypothesis model advanced by Hopper and Thompson (1980). In the following section we adduce arguments in favour of considering indirect objects (IOs) as part of the core argument structure of GTVs. We will show that constructions containing GTVs exhibit low Transitivity. We then move on to examine the syntax and semantics of the GTVs as compared to transitive constructions with emotion verbs. We argue that there is a semantic distinction between the two structures and offer typological data in support of this claim that suggest the existence of a cognitive basis for the constructional contrast. We then examine the properties of those verbs that fluctuate between the DO and IO constructions. It is found that the alternation between the two structures follows from the predictions made in the Transitivity Hypothesis. We argue, however, that it is difficult to determine at which point a change from one construction to another will take place, given that we are not dealing with categorical rules but rather with statistically significant tendencies.

The Transitivity Hypothesis In Spanish, the most frequent syntactic configuration of biactant predications corresponds to the transitive construction, in which the arguments take on the grammatical relations of subject and direct object. We have taken our examples from the Arthus corpus (Archivo de Textos Hispánicos de la Universidad de Santiago) and the quantitative data provided here have been obtained from the Base de Datos Sintácticos (BDS), based on an analysis of the corpus under Professor Guillermo Rojo’s supervision.2 According to the data from the BDS, of a total of 91,465 clauses with two participants, 68,011 (74.36%) display the subject–direct object pattern. From a semantic viewpoint, the transitive construction prototypically encodes actions initiated by

82

Gustar-Type Verbs

an agent and transferred to a patient. Thus, despite their syntactic configuration, examples (4.7–4.10) do not conform to the transitive protoytpe. With regard to GTVs, their subject–indirect object pattern deviates from the transitive prototype in that they exhibit a configuration that can be best characterized as experiencer-stimulus, which is clearly distinct from the transitive agent–patient configuration. The hypothesis defended here is that both levels – syntactic and semantic – are intimately related, and that the manifestation of a particular nontransitive configuration (that is, one with an indirect object) constitutes the syntactic manifestation of the semantically marked character (that is, marked with respect to the transitive prototype) of the clause in question. Following Hopper and Thompson’s (1980) framework, it will become evident that GTVs determine the lower Transitivity of the clauses that they configure, in opposition to the higher Transitivity of the prototypically transitive clauses. The present study will analyse the semantic characteristics of the clauses with GTVs, with special emphasis on some of the Transitivity parameters proposed by Hopper and Thompson (1980: 252). Participants of the verb The first parameter mentioned by Hopper and Thompson (1980) refers to the number of participants or arguments contained in a clause. It is understood that a highly transitive clause will have two or more participants, whereas a clause lower in Transitivity will typically exhibit one participant. Although Hopper and Thompson (1980: 252, and endnote 1) do not explicitly define the concept of participant, they cite Dixon’s (1979) A (for Agent) and O (for Object), which in a restrictive interpretation coincide with the subject and object of a prototypically transitive clause. It follows that the English clause ‘Jerry likes beer’ (1980: 254) contains two participants, whereas the Spanish clause Me gusta la cerveza ‘I like beer’ features only one, since ‘the experiencer … appears in an oblique case’ (ibid.). However, by denying an O(bject) status to the dative experiencer in Me gusta la cerveza, Hopper and Thompson (1980) weaken their assertion that ‘[w]e make no claims about the grammatical relations that the NP arguments referring to these participants [A & O] might bear to the verb’ (1980: 252, footnote 1). In order to properly tackle the study of the GTVs, it is necessary to overcome the limitations inherent in the traditional analysis of clauses, which is based on a surface interpretation of the Transitivity/Intransitivity dichotomy. Such an approach merely reduces the distinction to the presence or absence of a certain type of clausal constituent, the direct object, considered ‘the object’, with other possible complements being demoted to an ancillary or peripheral status. Valency (or dependency) grammar, which takes Tesnière’s (1959) work as its starting point, distinguishes between actants and circumstants according to their governed nature. Another, more formal distinction is superimposed


on this one; it identifies with the circumstants the presence of prepositional marking or of certain case affixes. In contrast to the circumstants, the actants lack prepositional marking and correspond to less-marked cases, such as nominative and accusative. Nevertheless, the difference between actants and circumstants (or core v. oblique) cannot be reduced to the selection of certain concrete markings, but rather is felt by many linguists to exist at a more general level, with case-marking being just one of its manifestations (Thompson, 1997: 60). Although Thompson (1997: 61) points out that ‘[l]anguages differ in the extent to which they make a morphological distinction between noun phrases serving core and oblique roles’, core participants are regarded as those that take on the functions of subject and direct object, which are precisely the arguments that lack prepositional marking in English. On the other hand, from a typological viewpoint, core functions are represented by the only argument of an intransitive verb, and the agent and patient of a transitive verb (S, A and O in Dixon’s, 1979, terminology). In this sense, we arrive at a rather restrictive interpretation of the concept of core arguments, since it mirrors the traditional distinction between transitive and intransitive. With regard to GTVs, there are reasons to argue that, apart from the subject, the indirect object (or dative) is also a core function in the clause. From a semantic viewpoint, it displays enough heterogeneity to prevent a simple association between syntactic function and semantic role. An argument with the function of indirect object can be an experiencer (Me gusta la cerveza ‘I like beer’), a recipient (Le mentí a Juan, ‘I lied to Juan’), a goal (Le entregué los documentos a tu hermano ‘I handed the documents to your brother’), a benefactive (Le preparé la cena ‘I prepared dinner for him’), or a possessor (Le lavé la cara al niño ‘I washed the child’s face’), and so on. In considering the syntactic features of core functions, we might expect that the process of grammaticalization of expressive devices would lead to less marking in morphological encoding. However, this does not mean that core functions are universally characterized by syntactic features such as a lack of prepositional marking, as in English. In Spanish, for example, the presence of the preposition a with animate direct objects, as in He visto a María ‘I have seen Mary’, does not deny the core character of the direct object, nor does its presence justify the alignment of the indirect object with oblique participants. There is a significant index of expression shared by subject, direct object and indirect object – the three core functions of the Spanish clause – which is the agreement with the verb. This agreement is marked by inflectional categories for the subject function (person and number suffixes on the verbal stem), and by unstressed pronouns for the direct and indirect object functions, (proclitics or enclitics to the verbal form). These agreement markers are illustrated in the example in (4.11), where se marks the indirect object and is coreferential with Juan, la marks the feminine singular direct object

84

Gustar-Type Verbs

and is coreferential with la bicicleta, and the suffix –mos marks 1pl and is coreferential with nosotros. (4.11)

La bicicleta se la regalaremos nosotros a Juan. The bicycle IO.3 DO.3fsg give-FUT1pl we to John ‘We will give the bicycle to John.’

The presence of the clitic totals nearly 100 per cent in the GTV examples. That is, based on the count of the tokens of gustar with the subject–predicate– indirect object structure, 1,218 of the 1,221 examples in the database (99.75%) contained the clitic. With regard to the semantico-pragmatic implications, the central functions encode the most prominent participants in the process described by the predication, which is reflected in the degree of topicality they display when compared to the degree of topicality of noncentral or peripheral participants. The indirect object undoubtedly presents a high degree of topicality, as shown in the distribution of its categorical realizations in our corpus. Within the framework of the Topic Continuity Theory (Givón, 1983), clitics would display the highest topicality; the clitic conjunction and full form would imply a lower degree of topicality; and, finally, with the full form without clitic being the least topical. The BDS data regarding of indirect objects provide the results in Table 4.1. Thus, there are semantic and syntactic reasons supporting the idea that the indirect object in Spanish is a core function, and not an oblique one. In contrast to Hopper and Thompson (1980), we claim that the clauses in (4.2–4.6) have two participants, thus matching the number of participants present in prototypically transitive clauses. However, there are obvious syntactic differences between a direct and an indirect object in Spanish. Apart from the differences in case between the accusative (direct object) and dative (indirect object) clitics (lo, los, la, las v. le, les respectively), the full NP indirect object is obligatorily marked by the preposition a, whereas the direct object is marked with a only when it possesses certain features involving animacy and/or definiteness.

Table 4.1 Degree of topicality in indirect objects Clitic only Clitic full form Full form Total

9,727 1,654 1,085 12,466

78.66% 13.38% 8.77% 100%

Clitics highly topical, clitic full form less topical, full form only least topical.


There are also differences in the possibility of passivization, which is blocked for indirect objects, as illustrated by the examples in (4.12) and (4.13): (4.12)

(4.13)

a. La nueva maestra les gusta a los niños. ‘The children like the new teacher.’ b. *Los niños son gustados por la nueva maestra. ‘The children are liked by the new teacher.’ a. Los niños quieren a la nueva maestra. ‘The children love the new teacher.’ b. La nueva maestra es querida por los niños. ‘The new teacher is loved by the children.’

An explanation for the impossibility of passivization in the case of (4.12b) involves the selection of semantic roles of GTVs. One must bear in mind that the function of passivization is to make into a theme (in the theme–rheme sense) a constituent other than the subject, that is, the direct object in Spanish. However, in GTVs the indirect object is treated as an unmarked theme, and as such there is no need to passivize it (cf. Table 4.4). Another syntactic difference between the Spanish direct and indirect objects is relativization. Whereas direct objects allow relative que without preposition, indirect objects require that the preposition a precede the relative:3 (4.14) (4.15)

La maestra que quieren los niños. ‘The teacher that the children love.’ Los niños a los que les gusta la nueva maestra. ‘The children that like the new teacher.’

Nevertheless, the analysis of a copious and diverse corpus reveals that the differences between the direct and indirect object are not as clear-cut as they seem in light of the above examples, since there are numerous verbs that show vacillation between both constructions. The uncertainty is compounded by those cases that feature a first- or second-person object clitic, whose form does not allow us to discern between accusative and dative: (4.16) (4.17)

Las leyes ya no me afectan. ‘The laws do not affect me any more.’ Te fastidia recordar el pasado. ‘It bothers you to remember the past.’

(PASAJERO: 34, 12) (AYER: 35, 4)

Even when a third-person clitic is involved, the vacillation between accusative and dative turns out to be somewhat random: (4.18) Lo que realmente lo preocupaba era una ceremonia, de fecha próxima, en la que se presentaría ante la reina (HISTORIAS: 131, 18)

86

Gustar-Type Verbs

‘What really worried him was a ceremony, coming up soon, at which he would introduce himself to the queen.’ (4.19) Dentro de cincuenta años, esos bienes que tanto le preocupan no le servirán de nada (HISTORIAS: 70, 10) ‘In fifty years from now, those possessions that worry him so much will be useless.’ Such difficulty in distinguishing direct from indirect objects represents another argument in favour of considering indirect objects as a participant and not as an oblique in Spanish. Kinesis, aspect and punctuality Three of the Transitivity parameters (kinesis, aspect and punctuality) proposed by Hopper and Thompson (1980) deal with ‘aspectuality’. The authors contend that highly transitive clauses are action clauses, that is, dynamic clauses involving telicity and punctuality. On the other hand, low Transitivity corresponds to static (that is, non-dynamic) clauses, which are atelic and non-punctual. If our approach to GTVs is correct, we would expect the predicates of GTVs take stative, atelic and nonpunctual clauses, and indeed the behaviour of GTVs seems to point in this direction. Some GTVs exhibit incompatibility with the perfective aspect; and although we find this with a reduced number of verbs (atañer ‘concern’, competer ‘be incumbent upon’, concernir ‘concern’, convenir ‘suit, be good for’, incumbir ‘be up to, concern’, which are inherently imperfective), this shows the correlation between perfectivity and high Transitivity on one hand, and imperfectivity and low Transitivity on the other.4 (4.20)

(4.21)

a. Os conviene refrescaros. ‘It is good for you guys to freshen up.’ (CAIMÁN: 37, 9) b. *? Os convino refrescaros. ‘Freshening up was good for you guys.’ a. A ti no te incumbía hablar de ellos realmente. ‘It was really not up to you to talk about them.’ (SEVILLA: 192, 17) b. *? A ti no te incumbió hablar de ellos realmente. ‘It was really not up to you to talk about them.’

A feature that generally characterizes stative clauses as opposed to dynamic ones is that the former do not carry a habitual or frequentative interpretation when used in the present (4.22), contrary to what happens with dynamic clauses, as shown by the sentence in (4.23): (4.22) A Ana le gusta Juan (*todas las semanas/*cada tarde/*con frecuencia). ‘Ana likes Juan (every weekend /every evening /frequently).’


(4.23) Ana visita a su abuela (todas las semanas / cada tarde / con frecuencia). ‘Ana visits her grandmother (every weekend / every evening / frequently).’ The incompatibility between stativity and habituality not only affects GTVs but also stative predicates with transitive argument structure, such as those in (4.7–4.10) above. In the same vein, the progressive-based test has been used as a key argument in the distinction between dynamic and stative. It is generally contended that stative predicates are anomalous in the progressive (cf. Marín Gálvez, 2000: 71). Indeed, as we have confirmed through the BDS, the verb type that combines with the estar gerund construction (analogous to English to be -ing) is the one which denotes acts and operations particularly evident to the faculties of sensation, (Fernández Ramírez, 1960: 534), such as those in (4.24) and (4.25): (4.24) (4.25)

El viejo ya está abriendo a Simonetta. (SONRISA: 147, 13) ‘The old man is already opening [the door] to Simonetta.’ Mira: yo la estoy acariciando. (TERNURA: 31, 18) ‘Look: I am caressing her.’

Nevertheless, it turns out that practically any verb, supported by the appropriate context, can be used in the estar gerund construction, although not all verbs convey the same semantic implications in this construction. Following Halliday’s (1985: 109) stance regarding English, it can be similarly claimed that, in Spanish, the estar gerund construction is the marked option with nondynamic clauses, and that its use in such instances – undoubtedly infrequent – does not add a semantic feature that can labelled as ‘progressive’, but rather other features, such as ‘ingressive’ or ‘intensive’: (4.26) (4.27)

Me está apeteciendo un café. ‘I am feeling like having a coffee.’ No nos está gustando nada tu comportamiento ‘We are not liking your behaviour at all.’

Another test that has been adduced to prove the difference between stative and dynamic situations is the possibility (or lack thereof) of combining parar de ‘stop’(de Miguel, 1999). It seems that parar de is only compatible with clearly dynamic situations: (4.28) (4.29)

*Esa película no para de gustarme ‘That movie does not stop pleasing me.’ *La propuesta no paró de interesarles. ‘The proposal did not stop interesting them.’

88

Gustar-Type Verbs

In sum, there are reasons to believe that GTVs have stative (and thus, low Transitive) clauses. Therefore, GTVs contrast with prototypically transitive predicates, which have active (and thus, highly Transitive) clauses. Nevertheless, among the clauses exhibiting the transitive syntactic pattern (subject–direct object), we also find abundant examples of stativity (with verbs such as those cited in 4.7–4.10), thus making it impossible to establish a direct relationship between syntactic pattern and the stative or dynamic nature of the situation denoted by the verb. Volition and agency Among the Transitivity features proposed by Hopper and Thompson (1980), there are two that refer to the semantic characteristics of the participant A: volition and agency. Higher Transitivity is correlated with the volitional and agentive aspect of A, whereas the absence of volition and agency in A is related to a lower Transitivity in the clause. Volition and agency refer to the degree of involvement of participant A in the situation described by a given clause. Volition depends on the voluntary and conscious nature of A’s participation, while agency is given, according to Hopper and Thompson (1980), by Silverstein’s Animacy hierarchy, since it largely coincides with the degree of topicality of the entity in question.5 Both volition and agency are interrelated, as the agentive potentiality, and in particular a high degree of animacy, is a necessary but not sufficient condition for volition. A number of researchers working on typologically unrelated languages have demonstrated that the possibilities of using a noncanonical transitive coding increase when the animacy of the subject is lowered in the Animacy Hierarchy (Dixon, 1979: 85ff; Comrie, 1981: 121; and Lazard 1994: 200–4). This tendency is also confirmed by Spanish data. The subject–direct object clauses and the subject–indirect object clauses display a highly different picture regarding the animate v. inanimate character of their subjects (Table 4.2). The low degree of animacy that characterizes the subject in the subject– indirect object pattern is also corroborated by the frequency of clausal subjects in this pattern when it is contrasted with the rather exceptional presence of a subject clause in the transitive pattern (Table 4.3). Table 4.2 Frequencies and percentages of animate v. inanimate subjects in two-participant clauses Animate Direct–object clauses Indirect–object clauses

58,318 (85.75%) 1,879 (28.35%)

Inanimate 9,693 (14.25%) 4,748 (71.65%)

Note: These data include all the subject–indirect object clauses in our database, not only the data pertaining to GTVs.

Victoria Vázquez Rozas 89 Table 4.3 Frequencies and percentages of clauses functioning as subject

Direct-object clauses Indirect-object clauses

Clausal subject

Others

589 (0.86%) 1,666 (25.14%)

67,422 (99.14%) 4,748 (74.86%)

Examples of GTVs with clausal subjects are the following: (4.30) (4.31)

Me agrada que penséis lo mismo (COARTADA: 50,33) ‘It pleases me that you guys think the same.’ Al viejo le gusta que ella no bisbisee oraciones (SONRISA: 182, 14) ‘The old man likes that she doesn’t whisper prayers.’

The data from Table 4.3 show that the clauses containing the subject–indirect object pattern clearly deviate from the transitive prototype, which includes among its components a highly agentive and, therefore, a highly animate subject, while a clausal subject such as that in (4.30–4.31) occupies the lowest level on the Animacy Hierarchy, and consequently exhibits the lowest potentiality of agency. Thus, those verbs that are combined alternatively with a direct object (accusative clitic) or an indirect object (dative clitic) reject the transitive pattern when they are combined with a clausal subject. Let us compare (4.32) and (4.33): (4.32)

(4.33)

a. Le sorprendió mucho el comienzo de la Guerra. ‘The beginning of the War surprised her/him a lot.’ b. Le sorprendió mucho que comenzase la Guerra. ‘It surprised her/him greatly that the War began.’ a. El comienzo de la guerra la sorprendió en París. ‘The beginning of the war surprised her in Paris.’ b. *Que comenzase la guerra la sorprendió en París.6 ‘That the war began surprised her in Paris.’

The tendency shown by certain verbs to adhere to the transitive pattern if the subject is animate, and to the indirect object pattern if the subject is inanimate, had already been observed by grammarians such as Cuervo (1874: fn. 121) and Fernández Ramírez (1951: 192), and, more recently, by García (1975: 307–16). Copious examples from our corpus corroborate that tendency, illustrated by the following pairs: (4.34) Una chica de Nueva York toma el barco a una isla del Caribe, donde la [DO] espera el novio para casarse. Parece una chica muy

90

Gustar-Type Verbs

buena, y llena de ilusiones, que le cuenta todo al capitán del barco, que es buen mocísimo, y él mira al agua negra del mar, porque es de noche, y después la mira a ella como diciendo ‘esta no sabe lo que le [IO] espera.’ (BMA, 163–4) ‘A girl from New York takes the ship to a Caribbean island, where her fiancé is waiting to get married. She seems like a very good girl, full of hope, who tells everything to the ship’s captain, who is staring into the black water, because it is nighttime, and then he starts looking at her as if to say “she doesn’t know what she’s in for”.’ (4.35) a. Sabía que el sólo decírselo a él iba a aliviarla, pero no lo hizo para no preocuparlo [DO]. (ATC, 484) ‘She knew that her just telling him was going to relieve her, but she did not tell him in order not to worry him.’ b. Lo que más le [IO] preocupaba de la muerte al doctor Urbino era la vida solitaria de Fermina Daza sin él. (ATC, 75) ‘What worried Dr. Urbino most about death was Fermina Daza’s lonely life without him.’ With respect to the feature of volition, which according to Hopper and Thompson’s (1980) view characterizes the participant A in canonical transitive clauses, it is absent in clauses with GTVs. Since inanimate subjects, which typically appear with GTVs (cf. Table 4.2), are inherently nonvolitional, the criterion to determine the presence of volition in participant A would be those GTV constructions containing an animate subject, as exemplified below: (4.36) Hoy precisamente, doy una comida y quiero presentarte a mis amigos, les vas a encantar. (DIEGO: 138, 11) ‘Precisely today, I am hosting a meal, and I want to introduce you to my friends. They are going to like you.’ (4.37) JAVIER – Ganas dinero y follas con quien te apetece. ADELA – Y, a veces, con quien me conviene. (CINTA: 102, 23) ‘JAVIER – You make some money and ball whoever you please.’ ‘ADELA – And, sometimes, who (also) suits me.’ (4.38) Que no seas pelmazo, que ya no me importa José. (OCHENTA: 67, 29) ‘Don’t be a bore, I am no longer interested in Jose.’ (4.39) ¿Quién de los dos le gustaba al francés? ( JÓVENES: 154, 4) ‘Which of the two did the Frenchman like?’ Among the signs that are adduced to show the volitional or controlled character of a situation, are the compatibility with the imperative, the possibility


of functioning as complement of a verb of command/obligation, and the combinability with manner adverbials (see Dik, 1989: 96–7; Di Tullio, 1996: 225). None of these tests yield positive results with clauses containing GTVs: (4.40)

*Encántales a mis amigos. ‘Make my friends like you.’ (4.41) *Te sugiero / ordeno que le convengas a Juan. ‘I suggest / order that you be right for Juan.’ (4.42) *La convencí / persuadí de que le gustase al francés. ‘I convinced / persuaded her that the Frenchman like her.’ (4.43) *José le importa deliberadamente / cuidadosamente / a propósito a Pedro. ‘Pedro matters to Jose deliberately / carefully / on purpose.’

Similarly, the impossibility of constructing an impersonal passive depends on the non-agentive character of the predication (see Levy, 1994: 357): (4.44)

*Se les encanta a los amigos. ‘One is enjoyed by the friends.’ (Rough translation)

Therefore, the semantic implications of the animate subjects accompanying GTVs include neither agentivity, nor volition, nor control of the situation described by the clause, and are interpreted in a way similar to inanimate subjects, as a cause or stimulus of the state in which the the experiencer finds her/himself. Affectedness and individuation of the O(bject) With regard to the features of the O argument, Hopper and Thompson (1980) claim that the degree of Transitivity is correlated with the affectedness and individuation of this participant. A clause will become more transitive to the extent to which O becomes affected by the process described in the clause, as is the case, for instance, with created or altered Os. With respect to the O appearing with GTVs, coded as the indirect object, it would be difficult to assert that it is affected in Hopper and Thompson’s (1980: 252–3) sense, that is, ‘[t]he degree to which an action is transferred to a patient is a function of how completely that patient is affected’, since the clauses containing these predicates do not refer to an action, nor can it be claimed that the participant O becomes affected. Rather, O is the basis of a stative situation. The fact that the affectedness feature does not apply to the O of GTVs is conditioned by the fact that such a participant is not the patient in an action clause, but rather, the experiencer of a state. For its part, the ‘individuation’ feature is interpreted as a cluster concept with the following properties: proper noun, human-animate, concrete,

92

Gustar-Type Verbs

singular, countable and referential-definite. Such properties refer to the distinctiveness of O, whose Transitivity is lowered to the extent that O is characterized as common, inanimate, abstract, plural, uncountable and nonreferential. In extreme cases, we find clauses that do not really have two participants, but rather one participant (A) with O-incorporation. In this view, the Os in sentences such as el niño tiene hambre (lit. the boy has hunger) ‘the boy is hungry’, la peonza da vueltas (lit. the top gives spins) ‘the top is spinning’, los escolares hacen novillos (lit. the student make young bulls) ‘the students play hooky’, esa cara da miedo (lit. that face gives fright) ‘that face is scary’, are not considered arguments, but rather they are part of complex predicates, with the resulting clauses being interpreted as intransitive.7 Hitherto we have seen that clauses containing GTVs are characterized as having low Transitivity. However, upon examining features of the O participant of the GTV clauses in our corpus, we find that the degree of individuation of the entity would correspond in Hopper and Thompson’s (1980) hypothesis, to clauses of high Transitivity. Indeed, almost all examples with GTVs contain a human, concrete, countable, and referential-definite indirect object. There are, however, some observations by Hopper and Thompson that seem to contradict considering a highly individuated O as an indicator of high Transitivity. That is, the authors claim that the Transitivity of the clause can be reduced ‘when there is an anomalous A–O relationship, viz. when the O is higher than the A in the [Agency] hierarchy’ (ibid.: 273). Indeed, GTVs favour a type of construction in which the O participant surpasses A on the Transitivity hierarchy, such that the high animacy of O would reduce the degree of Transitivity of the clause. In order to minimize the effect of such a contradiction, Hopper and Thompson (1980: 273) state that it ‘is the reduction of the “Agency” of the A which accounts for the anomaly in the A–O relationship, not the fact that the O is high on the hierarchy.’ This statement reestablishes the authors’ initial idea that the human-animate character of O is a sign of high Transitivity. Nevertheless, we offer the following data, which counter the above idea. First, if we assume that the canonical or deviant character of a certain syntactic configuration is correlated with its frequency, those features that characterize the object most frequently will be the ones that will allow us to qualify the object as prototypical. With regard to the animacy of the O participant in transitive constructions, the data from Spanish are convincing: of a total of 68,010 subject–direct object clauses only 12,834 (18.9%) display animate Os, from which we can conclude that the prototypical O will be inanimate. Second, according to the criterion of syntactic marking, a prototypical O will exhibit a morphologically less marked construction.8 Therefore, we must conclude that those Os marked with the preposition a (typically animate and definite) cannot be considered prototypical objects. Third, the inanimate character of O is also given by the semantic definition of Transitivity. It is assumed that a prototypical transitive clause refers to a


transfer from an agent to a patient, roles which are defined by Langacker (1991: 210) as follows: The archetypal ‘agent’ role is that of a person who volitionally carries out physical activity which results in contact with some external object and the transmission of energy to that object. The polar opposite of an agent is an inanimate ‘patient’, which absorbs the energy transmitted by externally initiated physical contact and thereby undergoes some change of state.9 Finally, the correlation between the different parameters that make up the Transitivity Hypothesis leads us to consider the inanimate O as a feature of high Transitivity. If Hopper and Thompson’s approach were correct, we would expect an animate O to be correlated with the parameters of kinesis, volition, agency, and so on that characterize highly transitive clauses. However, there are copious verbs in Spanish that show the opposite situation, that is, they present more features of high Transitivity when they have an inanimate O (coded as a direct object) than when they have an animate O (coded as an indirect object) (discussed later in this chapter).

‘Dative subjects’ In light of the relationship between clauses such as (4.45) and (4.46), the existence of a relation between the indirect object in (4.45) and the subject in (4.46) was proposed in early formalist accounts: (4.45) (4.46)

A María le gusta la música. ‘María likes music.’ María ama la música. ‘María loves music.’

The relationship was analysed transformationally as a rule of subject–object exchange called FLIP by Lakoff (1970: 126), and Psych-movement by Postal (1971).10 For its part, Case Grammar attributed the same interpretation, in terms of case frames, to both constructions, based on its supposed synonymy (see Fillmore, 1968: 30). Later, Relational Grammar interpreted constructions illustrated by (4.45) as involving inversion, which consisted of demoting the underlying subject to a surface indirect object, and promoting the underlying object to a surface subject, known as unaccusative advancement. Among the specific arguments for the existence of ‘dative-subject’ verbs in Spanish is that dealing with nominalizations. As noted by Fernández-Soriano (1999: 125), the nominalization of a three-argument verb such as entregar ‘hand in’ retains the preposition a for the indirect object, as in (4.47), whereas the indirect object of a verb like faltar ‘to lack’ – which is generally

94

Gustar-Type Verbs

thought of as a type of GTV – is obligatorily preceded by de, the preposition that appears with subjects: (4.47)

La entrega del premio a Juan ‘The presentation to Juan of the award.’ (4.48) La falta de valor de / *a Juan (nominalization of A Juan le falta valor.) ‘Juan’s lack of courage.’ (‘Juan lacks courage’.) In Fernández-Soriano’s view, the obligatory selection of de in (4.48) suggests that Juan may be an underlying (that is, a deep) subject, a so-called ‘dative subject’. As for the nature of the subject in GTVs, such as gustar ‘like’, its status as an underlying object has often been pointed out, appealing to the fact that it does not retain the preposition de, typical of subjects in nominalizations, but rather it is accompanied by the preposition por, as shown in (4.49): (4.49)

a. A Juan le gustan las cerezas. ‘Juan likes cherries.’ b. El gusto de Juan por las cerezas. ‘Juan’s liking for cherries.’

Campos (1999:1,560) mentions the coreferentiality test with temporal infinitival constructions, according to which the indirect object position of GTVs (for example gustar in 4.50) controls the subject position of the infinitive. Thus, in terms of its control properties, the indirect object in (4.50) corresponds to the subject position of verbs denoting dynamic situations (for example escribir ‘write’), where the subject controls the infinitive subject position, as indicated in (4.51): (4.50) (4.51)

A Lucii le gustaba Ronnyj antes de ei,*j conocer a Otto. ‘Lucyi liked Ronnyj before ei,*j meeting Otto.’ A Ronnyj le escribía Lucyi antes de ei,*j conocer a Otto. ‘Lucyj used to write to Ronnyi before ei,*j meeting Otto.’

However, other tests employed by generative grammarians to identify the subject do not support the presence of ‘dative subjects’ with GTVs. For example, Fernández-Soriano (1999: 123–4) attempts to show that the compatibility with small clauses is possible in constructions such as those in (4.52)–(4.56): (4.52)

Me falta café. ‘I need coffee.’


(4.53) (4.54) (4.55) (4.56)

Me pasa algo. ‘Something is wrong with me.’ Me consta que eres el mejor. ‘I know for a fact that you are the best.’ Nos dieron las dos. ‘It was two o’clock (and we were still engaged in some activity)’. A Marta le fue bien en Buenos Aires. ‘It went well for Marta in Buenos Aires.’

Nevertheless, of the examples in (4.52–4.56), only (4.55) and (4.56) are compatible with a small clause co-indexed with the indirect object, as shown in (4.57) and (4.58), respectively: (4.57)

Nosi dieron las dos borrachosi ‘We were drunk at two o’clock.’ (4.58) Mal mei fue ausentei pero peor presentei. ‘It went badly for me when I was absent, but much worse when I was present.’

However, the constructions featured in (4.52), (4.53) and (4.54) are incompatible with indirect object small clauses.11 This incompatibility is found with to all GTVs. However, GTVs do admit subject small clauses, as exemplified in (4.59): (4.59)

A Maríai Juanj le desagrada borrachoj/ *borrachai. ‘Maríai dislikes Juanj drunkj/*i.’

Both Relational Grammar, as well as Generative Transformational Grammar give precedence to the behavioural properties over the coding properties, minimizing the importance of the coding that identifies subjects (agreement with the verb, absence of preposition) from indirect complements (dative clitic showing agreement, the preposition a). For our part, we interpret such coding as crucial in the syntactic characterization of arguments. Moreover, we do not consider appropriate the distinction of several (more or less deep) levels of syntactic function. However, we acknowledge the similarities between the sentences such as (4.45) (A María le gusta la música ‘María likes music’) and (4.46) (María ama la música ‘María loves music’). Unlike the generativist formulations described above, we believe that the relationship between constructions such as (4.45) and (4.46) is to be interpreted in terms of different semantic and discourse values. Although both sequences can refer to the same reality, the different syntactic coding of the participants is not random at all, but rather it fits a different semantic and discourse configuration. With respect to the corresponding English

96

Gustar-Type Verbs

sentences ( John likes music and Music pleases John), Danes (1968: 61) proposes two different semantic patterns: the first would consist of ‘bearer of attitude – attitude – object of attitude’, whereas the second would be made up of ‘source (cause) – effecting – recipient of effection’. Before studying the features that distinguish the clauses in (4.45) and (4.46), we will discuss the similarities between both structures, starting with a diachronic reference to the issue at hand. In the earlier stages of their respective histories, both English like as well as Spanish gustar had a different argument structure than what they have now. In Old English, the element that currently takes on the form of subject was coded as the object, as in him like oysters v. he likes oysters.12 With regard to gustar, it was used as a transitive verb in Old Spanish. In the sixteenth century this transitive pattern coexists with a pattern with prepositional object, and by the eighteenth century documentation emerges showing gustar with ‘dative subjects’.13 However, the structural changes are not restricted to the verbs like and gustar. As Whitley (1998) points out, many of the so-called psych verbs have presented different syntactic patterns over time, a variability that is not observed in prototypically transitive verbs, which are more syntactically stable. Focusing on the evolution of gustar, we observe that the transition from the transitive to the ‘dative subject’ argument structure has been accompanied by a semantic change that adheres to Hopper and Thompson’s (1980) Transitivity Hypothesis. As Whitley indicates (1998: 138), ‘[i]n the semantic change of gustar from a tasting agent to a satisfied experiencer, there was a decrease in “kinesis”, “volitionality” and “agency”.’ However, regardless of the direction of the structural variation, it is unquestionable that the change reveals a close link between the subject of the transitive pattern and the indirect object of the intransitive one. From a typological perspective, this relationship is supported by data from several languages. For example, in the Indo-European context it is shown in the emergence of the verb of possession ‘have’ from a dative construction containing the copula ‘be’, as exemplified by the Latin examples in (4.60)–(4.61): (4.60)

(4.61)

Mihi est liber to me (dat) is book (nom) ‘I’ve got the book.’ Habeo librum I have book (acus) ‘I’ve got the book.’

The dative, or indirect object, represents a certain degree of activity on the part of the participant, which is linked to its prototypically animate character. This degree of activity sets the participant apart from the direct object


(characterized as non-active), while bringing it closer to the subject, which in turn corresponds to a prototypically active participant. The relationship between indirect object and subject also manifests itself in the fact that the dative can code the agent complement of the passive construction in languages such as Old Persian and Pre-Classical Latin (see Herslund, 1988: 292–3), Mongolian (Comrie, 1976: 276), or Japanese (Marantz, 1984: 140).14 Another characteristic that draws the indirect object toward the subject and moves it away from the direct object is the feature of independent existence that we attribute to the entity represented by the indirect object. This is one of the properties that Keenan (1976) assigns to the subject.15 Indeed, the independent existence characterizes the indirect object that accompanies GTVs. Finally, the pragmatic-informative properties also attest to the close relationship between the indirect object in a clause like (4.45) and the subject in clauses such as (4.46). In both instances, it is the relatively more active participant that occupies the thematic position, independently of the syntactic function it performs. The data about the position with respect to the verb that have been mentioned for gustar clearly indicate the thematization of the indirect object vis-à-viss the subject. Contrary to the most common association between subject and discourse theme, in clauses with GTVs the unmarked discourse theme is not the subject, but rather the indirect object, and as such it occupies the first position in the clause, whereas the subject follows the verb (Table 4.4).16 Table 4.4 Frequencies and percentages of preposition and postposition of subject and indirect object for gustar Subject17 Preposed Postposed Implicit, clitic, or fixed order Total

Indirect object

125 (9.98%) 678 (54.15%) 449 (35.86%)

223 (18.26%) 29 (2.37%) 969 (79.36%)

1,252

1,221

Direct construction v. inverse construction We begin this section with a terminological note. Following the terminology employed to refer to the patterns under consideration, we will consider a ‘direct construction’ the pattern found in clauses such as María ama la música ‘María loves music’ (in 4.46), and ‘inverse construction’ the pattern identifiable in a María le gusta la música ‘María likes music’ (in 4.45).18 As we have seen in the previous section, some authors claim that the differences

98

Gustar-Type Verbs

between these constructions are limited to the most superficial syntactic level, and that underlyingly we would be dealing with an identical configuration. An argument against this view is the existence of sentence pairs containing the same lexical items that exhibit divergent interpretations: (4.62)

(4.63)

(4.64)

a. María admira la rapidez con que dibujas. ‘María admires the quickness with which you draw.’ b. A María le admira la rapidez con que dibujas. ‘María is astonished by how fast you draw.’ a. Juan apetece una casa propia. ‘Juan fancies having his own house.’ b. A Juan le apetece una casa propia. ‘Juan fancies having his own house.’ a. Antonio repugna el dulce. ‘Antonio finds the sweet disgusting.’ b. A Antonio le repugna el dulce. ‘The sweet disgusts Antonio.’

If we adopt a functional approach, we need to account for the existence of two different constructions for the expression of the same content. That is, we must justify the functional output of the two constructions. Whitley (1998: 130) suggests that there is difference in terms of control between the constructions: Verbs that stayed in or moved to type 1 [the direct construction] suggest a common denominator of choice or control. One can choose to love, hate, hope, use, even take responsibility for it, while pleasure (gustar, placer), pain (doler), and sorrow (pesar) in type 3 [the inverse construction] are feelings that simply happen to the experiencer. Nevertheless, the attribution of control or responsibility to the subject participant of the direct construction is debatable. It is not clear to what extent one can love, hate or admire something or somebody as a result of a conscious, volitional effort. If we try strictly linguistic tests, the result is not any clearer. It is true that some verbs like amar ‘love’ or pensar ‘think’ are compatible with the imperative: (4.65) (4.66)

‘Amaos los unos a los otros’, dijo Jesucristo. ‘ “Love one another”, said Jesus Christ.’ Piénsalo con calma. ‘Think it over calmly.’

However, it does not seem possible to use the imperative with verbs such as apetecer ‘fancy, feel like’, repugnar ‘disgust’. Nevertheless, certain contrasts


between direct and inverse constructions lend credence to the idea that the direct construction implies a certain degree of agentivity on the part of the subject that is absent from the inverse construction. Thus, in saying he tenido una idea ‘I have had an idea’ or he pensado que podríamos ir a la fiesta ‘I thought that we could go to the party’, we assume responsibility for the mental activity represented, whereas if we say se me ha ocurrido una idea ‘an idea has occurred to me’ or se me ha ocurrido que podríamos ir a la fiesta ‘it has occurred to me that we could go to the party’, we do not admit our participation in the making of the thoughts that are reproduced. Thus, they would escape the conscious control of our mind. Proof regarding how the different degree of agentivity affects the selection of the construction is given by the verbs recordar ‘remember’ and olvidar ‘forget’. Although both refer to similar cognitive processes, the language attributes more agentivity to the action of remembering than to that of forgetting. Therefore, only the latter verb allows the inverse construction:19 (4.67)

(4.68)

a. María recordó el cumpleaños de Juan.’ ‘María remembered Juan’s birthday.’ b. *A María se le recordó el cumpleaños de Juan.20 ‘María remembered Juan’s birthday.’ a. María olvidó el cumpleaños de Juan. ‘María forgot Juan’s birthday.’ b. A María se le olvidó el cumpleaños de Juan. ‘María forgot Juan’s birthday.’

On the other hand, the differences in control between direct and inverse constructions are also reflected in the (im)possibility of functioning as complement of a verb that implies volition, such as intentar ‘try’. Thus, we can say intentó olvidar a María ‘S/he tried to forget María’, but not *intentó que se le olvidara María. From a typological point of view, there are reasons to think this. Like Spanish, other languages are equipped with more than one way of conceptualizing psychic processes (see Wierzbicka 1999: 58–9). This is found in Russian, in which on can express the fact of ‘being sad’ with three different constructions: (4.69)

a. On byl grusten. he-NOM was-MASC sad-MASC b. Emu bylo grustno he-DAT (it)was-NEUT sad(ADV)NEUT c. On grustil he-NOM sad(VERB)-PAST.MASC.

(Wierzbicka, 1999: 59)

100

Gustar-Type Verbs

As Wierzbicka (1999: 60) points out: All these sentences can be roughly glossed as ‘he was sad’, but in fact they differ in meaning. In particular sentence (b) implies that the sadness was involuntary and was, so to speak, ‘happening to the experiencer’, whereas (c) implies active involvement by the experiencer, and suggests that he is bringing about his own sadness by thinking certain thoughts. The semantic and constructional parallelism between (4.69b) and (4.69c) above with the inverse and direct construction, respectively, is evident. A similar situation is found in Polish, which presents direct constructions as well as inverse ones in order to represent the same type of processes: (4.70)

a. Ania podziwiala Piotra. Ania:NOM admired Piotr:ACC ‘Ania admired Piotr.’ b. Ani imponuje Piotr. Ania:DAT impressed Piotr:NOM ‘Piotr impressed Ania. / Ania was impressed by Piotr.’ (Dhbrowska, 1997: 70)

Thus, the experiencer is coded, according to the construction, as nominative or dative. Dhbrowska asks herself about the basis for this distinction and concludes that it is based on the existence of two ways of conceptualizing mental experiences, which revolve around two different theories. On the one hand, we have the ‘craftsman model’, according to which the mental experience is conceived as an action undertaken by the experiencer who manipulates mental objects (ideas, images, experiences). This model corresponds to the coding of experiencer as nominative. On the other hand, we have the ‘mental arena model’, which conceives the mind as a container of ideas that executes the activities of thinking, feeling, and so on, as spontaneous processes that take place in the experiencer’s personal sphere. According to this model, the sensations, feelings and beliefs are conceived by attributing to them an independent existence outside the experiencer.21 This model corresponds to the dative experiencer (see Dhbrowska, 1997: 77). The existence of the two types of constructions, direct and inverse, would corroborate the cognitive validity of both theories. According to Dhbrowska for Polish, the coding of the experiencer through the nominative is interpreted as the default option, with the dative construction requiring a special motivation (ibid.: 79). We can also explain the existence of two different codings for psychological processes through the prototypical configuration of the biactant predications, which can be summed up as the transference of an action from an


agent to a patient. The verbs that adjust more naturally to this semantic configuration have been called ‘primary transitive verbs’,22 and pose no problems with regard to the syntactic construction that they select: the agent is coded as subject and the patient as direct object. Situating ourselves within Dowty’s (1991) framework, which deals with a set of semantic implications for the Proto-Roles of Agent and Patient, we can say that for ‘primary transitive verbs’ the subject has all the Proto-Agent features, whereas the object displays all the properties of the Proto-Patient.23 Nevertheless, in the case of psychological processes, the selection of subject and object is less obvious, since the participants are not so evidently specialized as Proto-Agents or Proto-Patients. Rather, we need to acknowledge that both the experiencer and the stimulus have Proto-Agent properties: the experiencer has property b (‘sentence and/or perception’) of the Agent Protorole, whereas the stimulus has property c (‘causing an event or change of state in another participant’) of the Agent Proto-role. Thus, both the experiencer and the stimulus can be candidates to be coded as subject, as is the case in Spanish.

Indirect object v. direct object The list of verbs that combine exclusively with indirect object and show the semantic configuration stimulus-experiencer is relatively limited: (4.71) agradar ‘please’, alcanzar ‘affect’, be sufficient for’, apetecer ‘fancy’, atañer ‘concern’, competer, concernir ‘concern’, convenir ‘be good for’, desagradar ‘displease’, doler ‘hurt’, extrañar ‘find strange’, gustar ‘like’, importar ‘matter’, incumbir ‘concern’, parecer ‘seem’, pasar ‘happen’, pesar ‘regret’, placer ‘please’, repugnar ‘disgust’, suceder ‘happen’, tocar ‘be posted, win, be one’s turn, concern’ Some of these verbs also have a subject–direct object construction whose semantic configuration is no longer stimulus–experiencer, but agent–patient. If the direct object is compulsorily inanimate, both constructions correspond to two clearly distinct meanings, as shown by the examples in (4.72) and (4.73): (4.72)

(4.73)

a. Los abogados han convenido las condiciones del contrato. [DO] ‘The lawyers have agreed the conditions of the contract.’ b. Al abogado [IO] le convienen las condiciones del contrato. ‘The conditions of the contract are good for the lawyer.’ a. El gobierno importa alimentos [DO] de Australia. ‘The government imports food from Australia.’

102

Gustar-Type Verbs

b. Al gobierno [IO] no le importa tu opinion. ‘Your opinion does not matter to the government.’ In other instances, the direct object can be animate, like the indirect one, but there are also differences of meaning between the two: (4.74)

a. Voy a extrañar a Mariana [DO], por ser la última mujer de mi vida. (HISTORIAS: 56, 4) ‘I am going to miss Mariana for being the last woman of my life.’ b. A él [IO] también le extraña que, de repente, lo tan ansiado parezca recuerdo de cosa ya olvidada. (SONRISA: 232, 17) ‘He also finds strange that, suddenly, that which was so craved seems a memory of something already forgotten.’

There are some other verbs that seem to reveal differences in their semantics that go hand in hand with their use with an indirect (or direct) object, although mention of both constructions does not always appear in the dictionaries. For example, the verb encantar ‘love’ exhibits two, clearly differentiated meanings, shown in (4.75)–(4.76):24 (4.75) Someter [a alguien o algo] a una acción que sobrepasa lo natural por medio de la magia. (Seco, 1999, entry for encantar) ‘To subject [somebody or something] to an action that goes beyond the natural [world] by means of magic.’ (4.76) Gustar o complacer extraordinariamente [a alguien]. (Seco, 1999, entry for encantar) ‘To like or please [someone] very much.’ Both uses are listed as transitive, contrary to the common tendency in Spanish. Nevertheless, in the same dictionary (Seco et al., 1999), there is an example with an accusative clitic for the sense in (4.75): al escupirles en la cara logra encantarlos, y quedan como muertos por el día ‘when s/he spits in their faces, s/he manages to enchant them, and they remain as if dead for the day’; and another example with a dative clitic for the meaning in (4.76): Le encanta la música ‘S/he loves music’. The possibility of leísmo for the latter example is out of the question, since the dative clitic can have a feminine referent, as in a María le encanta la música ‘María loves music.’ According to the BDS data, of a total of 96 appearances of encantar with the meaning in (4.76) above, 26 exhibit a dative clitic (le/les, whether alone or accompanied by a phrase introduced by a) and 70 have first- and second-person clitics, that is, syncretic forms that neutralize the distinction between accusative and dative. In light of these data, we can claim that in its meaning of


predicate of affection, encantar displays the inverse construction linked to the GTVs. Significantly more complex is the behaviour observed in a large set of verbs that has the possibility of appearing with the accusative or the dative, without there being any clear semantic difference between them. Among these verbs, we cite the following: (4.77) abrumar ‘overwhelm’, aburrir ‘bore’, admirar ‘admire’, afectar ‘affect’, afligir ‘grieve’, alegrar ‘make happy’, angustiar ‘distress’, apasionar ‘fascinate’, apenar ‘sadden’, asombrar ‘amaze’, asustar ‘scare’, atemorizar ‘frighten’, aterrar ‘terrify’, aterrorizar ‘terrorize’, atormentar ‘torment’, atraer ‘attract’, avergonzar ‘shame’, cansar ‘tire’, complacer ‘please’, consolar ‘console’, convencer ‘convince’, decepcionar ‘disappoint’, deleitar ‘delight’, desanimar ‘dishearten’, descontentar ‘displease’, desconsolar ‘distress’, desesperar ‘exasperate’, disgustar ‘disgust’, distraer ‘amuse’, ‘distract’, divertir ‘amuse’, emocionar ‘move’, ‘touch’, entretener ‘amuse’, ‘entertain’, entristecer ‘sadden’, entusiasmar ‘love’, escandalizar ‘scandalize’, espantar ‘scare away’, estorbar ‘bother’, exasperar ‘exasperate’, fascinar ‘fascinate’, fastidiar ‘annoy’, favorecer ‘favour’, halagar ‘flatter’, impresionar ‘impress’, incomodar ‘inconvenience’, inquietar ‘unsettle’, interesar ‘interest’, intranquilizar ‘worry’, intrigar ‘intrigue’, irritar ‘irritate’, maravillar ‘amaze’, molestar ‘bother’, obsesionar ‘obsess’, ofender ‘offend’, pasmar ‘astonish’, perjudicar ‘be bad for’, preocupar ‘worry’, reconfortar ‘comfort’, satisfacer ‘satisfy’, seducir ‘seduce’, sorprender ‘surprise’, tranquilizar ‘calm down.’ Given that these are verbs that appear in dictionaries with the ‘transitive’ label, one would expect them to take accusative (direct) objects in nonleista dialects. However, the grammarians’ accounts about the fluctuation in casemarking of these verbs are not only recent. With respect to the verb admirar, Bello (1847: §749) states that it projects accusative case in un objeto nos admira ‘an object strikes us as odd’, while Cuervo (1886–93, s.v. admirar) claims that if we interpret the admirar ‘cause admiration’, as in Bello’s example, this entry takes the dative, since the accusative is associated with the meaning ‘consider with admiration’. Other verbs for which Cuervo (1886–93) postulates a dual case marking are agradar25 ‘please’, desagradar ‘displease’, contentar ‘please’, cansar ‘tire’, descontentar ‘displease’, and disgustar ‘annoy’. More recently, several scholars have attempted to identify the factors that determine the selection of accusative or dative case marking among verbs belonging to this class.26 The factors that have been alluded to are related to the nature of the subject and object participants, as well as to the aspectual nature of the predicate as a whole. The features that have been mentioned fit into

104

Gustar-Type Verbs

the predictions of Hopper and Thompson’s (1980) Transitivity Hypothesis (with certain modifications regarding the characteristics of the object participant; see our earlier discussion of ‘Affectedness and Individuation of O’), in the sense that the accusative object is associated with highly transitive clauses, whereas the dative object corresponds to clauses of low Transitivity. With regard to the aspectual configuration of the predicate, we argue that clauses with direct object marking tend to denote dynamic and telic events, whereas indirect object marking corresponds to stative and atelic events. The following examples with the verb atraer ‘attract, appeal’ clearly illustrate how direct object marking appears in a clause that expresses an action of a physical nature (4.78), while indirect object marking appears in a clause expressing a psychic reaction on the part of the experiencer (4.79):27 (4.78) Los rugidos del león atrajeron al cazador, quien, sacándolo del fondo del foso, lo metió en una gran jaula y se lo llevó (1INFAN: 21, 26) ‘The roaring of the lion attracted the hunter, who took it out from the bottom of the pit, put it in a big cage, and took it with him.’ (4.79) El libro estaba encuadernado en piel y tenía el canto dorado, pero a ella no le atraía. (SUR: 104, 30) ‘The book was bound in leather and had a gilded spine, but did not appeal to her.’ The aforementioned differences in telicity can be observed in the following examples with convencer (4.80) in its telic sense, and (4.81) in its atelic one: (4.80) La ha convencido y se casarán en cuanto arreglen los papeles. (SONRISA: 329, 23) ‘He has convinced her and they are going to get married as soon as they straighten the paperwork out.’ (4.81) Veo que a usted le gusta el ambiente porque veranear por ahí en un poblacho, eso no le convence. (MADRID: 31, 24) ‘I see that you like the atmosphere, because spending the summer in a dump, that does not convince you.’ For Di Tullio (1998: 256), the verbs in question ‘can denote complex events as well as states in each of their alternative forms, the accusative and the dative’. In complex events, the activity carried out by the subject triggers a change in the psychological state of the object. As Di Tullio (1998: 258) points out, the activity implies dynamism and control on the part of the subject, and the resulting event is telic since it entails culmination or outcome. On the other hand, the sentence with a dative-marking verb implies neither


dynamism nor control, nor is it delimited, since it does not entail an outcome. The impossibility of the imperative and the progressive constitutes proof of the stative nature of the predicate with dative marking. Following Croft (1986), Dowty (1991) points out that the psychological predicates that code the stimulus as subject (please, frighten, and so on) can be interpreted as inchoative or stative (unlike verbs such as like, fear, and others, whose subject is experiencer, and are interpreted only as stative). The inchoative interpretation implies a change of state in the experiencer, ‘coming to experience an emotion or a new mental state’ (Dowty, 1991: 580), such that this argument would have a Proto-Patient implication absent from the stative interpretation. Consequently, the experiencer in the inchoative interpretation would be a ‘better’ Patient (ibid.) than the experiencer in the stative interpretation. This would explain, as Ackerman and Moore (1999: 24) note, the distribution of object marking in Spanish: accusative (prototypical direct object) in the inchoative predicate, and dative (non-prototypical object) in the stative predicate. The nature of the subject argument determines the marking of the object. If the subject is animate and possesses the will to act consciously, the likelihood that the experiencer/patient is coded as an accusative object increases notably (see 4.82 below). However, if the subject is inanimate it is more likely that the experiencer will be dative (see 4.83 below): (4.82) No buscó a Diego, no quería molestarlo. (DIEGO: 59, 6) ‘S/he did not seek out Diego. S/he did not want to bother him.’ (4.83) También le molestaba mucho que Agus se metiera el dedo en la nariz. (TERNURA: 90, 26) ‘It also bothered him a lot that Agus picked his nose.’ Strictly speaking, it cannot be claimed that all the apparently (animate) subjects can be interpreted as agents with the verbs under consideration in this article. Di Tullio (1998) distinguishes those cases in which the animate subject denotes an individual with agentive potentialities, like the implicit subject of molestar in (4.82) above, from other situations in which the animate reference is understood as a summation of properties, body parts, characteristics or behaviours (Di Tullio, 1998: 257), as shown by the subject of atraer in 4.84. (4.84) Los [estudiantes] más comprometidos en la lucha seguían en la Facultad. A David le atraían. (JÓVENES: 156, 39) ‘The students that were most committed to the struggle continued on in the Department. David was attracted by them.’

106

Gustar-Type Verbs

The subject of human reference in (4.84) behaves like the subjects of GTVs, which, as we pointed out earlier in the chapter, lack volition and control over the situation, independently of their subcategorization features. With regard to the characteristics of the object, it seems that the way in which it participates in the event can also condition the functional selection. The less active the object argument is, and the more directly affected by the situation it is, the more likely it is to be coded as a direct object. In particular, when the affectedness is physical, there is a tendency to use the accusative, whereas psychic affectedness tends to correlate with the dative. In this respect we can adduce, for example, Hurst’s (1951: 76) data on the verb irritar, which favours the use of the dative in a sequence such as a ella le irritan mis atenciones ‘my affections irritated her (dative)’ vis-à-vis the physical meaning of la irritaba el roce de la cinta ‘the rubbing of the tape [on her skin] irritated her (accusative)’. By the same token, the possibilities of passivization are reduced for the psychological uses of the verbs listed in (4.77). Compare, for instance, the sentences in (4.85) to their counterparts in (4.86): (4.85)

(4.86)

a. Al ladrón lo sorprendió la policía en el interior de la vivienda. ‘The police surprised the burglar in the inner part of the house.’ b. El ladrón fue sorprendido por la policía en el interior de la vivienda. ‘The burglar was surprised by the police in the house.’ a. Al ladrón no le sorprendió la actuación de la policía. ‘The police intervention did not surprise the thief.’ b. *El ladrón no fue sorprendido por la actuación de la policía. ‘The thief was not surprised by the police intervention.’

Ackerman and Moore (1999: 9), following Treviño (1992), explain the contrast between ‘direct affectedness’ and ‘non-direct affectedness’ using the examples in (4.87)–(4.88): (4.87) (4.88)

Los perros lo molestan siempre que llega ebrio. ‘The dogs harass him (DO) every time he comes home drunk.’ Los perros le molestan (*siempre que llega ebrio). ‘The dogs bother him (IO) (every time he comes home drunk).’

In (4.87), we have an object that is more prototypically patient than the one in (4.88), which accounts for its coding as a direct object. In (4.88), however, the object participant does not undergo a change of state, and is not directly affected, but rather is the experiencer of a new stative situation. Again, one can test this: only (4.87) allows passivization. With respect to the dialectal extension of the accusative–dative alternation of the verbs in (4.77), we have to point out that there are instances of


interference due to the phenomenon of leísmo. In such circumstances, the alternation only has an effect on the feminine form, with the exception of those varieties that exhibit both leísmo and laísmo. Apart from that, it has been pointed out that in certain varieties (for example those in Argentina, Chile, Peru), accusative case marking has been generalized for the verbs listed in (4.77), and for some GTVs (see Fernández-Ordóñez, 1999: 1,325). The example with concernir ‘concern’ illustrates this solution: (4.89) Algo que lo concernía tan íntimamente quizá lo descalificara para juzgar. (HISTORIAS: 93, 2) ‘Maybe something that concerned him so intimately disqualified him from judging.’ Thus, the alternation between accusative and dative case marking for verbs of affectedness is a phenomenon that itself varies somewhat in the Spanishspeaking domain. Although the factors that we have pointed out in this section undoubtedly play a role in the syntactic coding of the object participant, it must be interpreted not as a set of rules that are applied rigidly, but rather as general tendencies that account for a large number of cases, but that do not exclude the existence of examples that go against the predictions. There is one more factor that contributes to the alternation of object marking in verbs of emotion: the relatively low frequency with which speakers are forced to choose between one or the other construction. Emotion verbs express an internal process of the experiencer in such a way that a speaker can more easily appreciate his/her own feeling when s/he is the one affected, but has only indirect evidence of the process when the experiencer is another person.28 This is why in discourse there are more cases in which objects are first-person pronouns than those in which objects are third-person pronouns. Since the syncretism of the Spanish pronominal system excludes the possibility of making a case distinction between direct and indirect object in all but the third person, speakers are freed from having to choose between the two forms when using the first (and second) person pronoun. We have provided statistical analyses in our data that seem to confirm this line of argumentation. Since it is commonly understood that conversational oral language is the most representative of the spontaneous language use, we have limited the statistical analyses to data from oral texts and the theatrical texts of the data Arthus (see note 1). We have also limited the verbs under our investigation to a small sample: alegrar ‘please’, asombrar ‘surprise’, disgustar ‘displease’, divertir ‘enjoy’, interesar ‘interest’, molestar ‘bother’, preocupar ‘worry’, sorprender ‘surprise’ and tranquilizar ‘calm’. Results indicate that there were 21 cases of third-person pronouns while there were 133 cases of first person pronouns (86.4%). Needless to say, a more extensive quantitative study needs to be done, but the data discussed in our study still provide an

108

Gustar-Type Verbs

important indication for the relative frequency of the pronominal forms in real discourse. Why, then, is it considered that a low frequency of use favours variation? The answer to this question is based on the assumption that linguistic structures result from conventionalization of frequently used discourse forms. That is, following claims by Givón (1979), Du Bois (1985) and Hopper (1998), we believe that grammar emerges from discourse and thus a frequency factor is essential to account for the emergence and evolution of the syntactic structures. As pointed out by Du Bois (1985: 363), ‘grammars code best what speakers do most’. Thus, we believe that the accusative–dative alternation in emotive verbs cannot be accounted for either way – generalizing either accusative or dative for all the cases of the same verb – since its frequency of use is not sufficient to force grammaticalization. We do note tendencies for some of these verbs to be used in a fixed construction in order to express a certain meaning (cf. example 4.76 with the verb encantar ‘like’), but the majority of verbs have not lexicalized the accusative or the dative construction, which leaves ample semantic space in which the two variants of construction coexist.

Conclusion In this chapter, we have carried out an analysis of the GTVs within the framework of the Transitivity Hypothesis proposed by Hopper and Thompson (1980). First, we argued that, together with DOs and subjects, IOs should be considered part of the core argument structure of GTVs. After examining the characteristics of the constructions containing verbs and paying special attention to the parameters proposed by Hopper and Thompson (1980), we demonstrated that constructions containing GTVs exhibit low Transitivity. We also analysed the syntax and semantics of the GTVs as compared to transitive constructions with emotion verbs. In contrast to formalist approaches that defend assigning the same semantic configuration to both types of construction, we have seen how the functionalist approach leads us to make a semantic distinction between the two structures. In support of this argument, we put forward typological data which corroborate the existence of a cognitive basis for the constructional contrast. Finally, we examined the properties of those verbs that fluctuate between the construction with IO and the construction with DO. We found that the alternation between the two structures follows perfectly from the predictions made in the Transitivity Hypothesis: sentences with accusative object tend to be dynamic and telic, and tend to have an animate, agentive subject, and their objects tend to be inactive and physically affected. By contrast, sentences with dative objects are usually stative, atelic, have no agentive subject, and their objects are affected psychologically. The choice between accusative and dative objects is, thus, an indication of the degree of Transitivity of the sentence as a whole.


In any case, we have to remember that speakers ultimately choose the most appropriate construction for each case, depending on the meaning they intend to convey. If high Transitivity is coded by means of the accusative construction, we are able to say only that as a sentence becomes increasingly prototypically less transitive, the greater the possibility is of the dative construction appearing. It is, however, difficult to determine exactly at which point such a change will systematically occur. In this area we are dealing with statistically significant tendencies, not categorical rules.

Notes 1 One should also consider as a typical construction of ‘verba sentiendi’ that one found in clauses of pronominal and prepositional object construction, exemplified in Me alegraba de tenerlo a mi lado (Historias: 156, 13) ‘Having him beside me made me happy’; Jano se olvidó de Betina (Carta: 111, 13) ‘Jano forgot about Betina’; Se encaprichó con la chiquilla (Sonrisa: 301, 22) ‘S/he took a liking to the little girl’. 2 For further information on the database, the reader is directed to http://www. bds.usc.es. 3 Nevertheless, in the oral language the use of the relative is frequent without preposition: Me han hablado muy bien de la obra; ahora, que hay gente …, que hay gente que [a la que] le gusta mucho y otra que [a la que] no le ha convencido demasiado (Madrid: 371, 27) ‘I have been told very good things about the plau; now, that there are people …, that there are people that like it very much and other that has not been convinced enough’. 4 Hopper and Thompson (1980: 271) observe that ‘if the Aspect is perfective, the interpretation – other things being equal – has properties allowing the clause to be classified as more transitive; but if the Aspect is imperfective, the clause can be shown on independent grounds to be less transitive’. 5 Dixon’s (1979: 85) representation of the hierarchy is the following:

1st p. pron. 2nd p. pron. 3rd p. pron.

Human Animate Inanimate proper nouns common nouns

Functional Approaches to Spanish Syntax: Lexical Semantics, Discourse and Transitivity

Functional Approaches to Spanish Syntax: Lexical Semantics, Discourse and Transitivity

Lexical Functional Grammar, Volume 34 (Syntax and Semantics)

Lexical-Functional Grammar (Syntax and Semantics, Volume 34)

Unaccusativity: at the syntax-lexical semantics interface

The Syntax and Semantics of Discourse Markers

Lexical Semantics

Lexical-Functional Syntax (Blackwell Textbooks in Linguistics)

Morphology and Lexical Semantics

Logical Syntax and Semantics

Contrastive lexical semantics

Approaches to Discourse Particles

Lexical Functional Grammar

Lexical functional grammar

Lexical Functional Grammar

Lexical-Functional Grammar

SYNTAX AND SEMANTICS VOLUME 32

Computational approaches to morphology and syntax

Approaches to Discourse in Dementia

Lexical Perspectives on Transitivity and Ergativity: Causative Constructions in English

Syntax in Functional Grammar

Exploring the Syntax-Semantics Interface

Formal Syntax and Semantics of Java

Latin Syntax and Semantics (Romance Linguistics)

Logical Syntax and Semantics, Their Linguistic Relevance

English Syntax: From Word to Discourse

Complex Predicates in Nonderivational Syntax, Volume 30 (Syntax and Semantics)

The Syntax of Spanish (Cambridge Syntax Guides)

The Syntax of Spanish

Syntax and semantics of programming languages

The Syntax and Semantics of Infinitary Languages

Functional Approaches to Spanish Syntax: Lexical Semantics, Discourse and Transitivity