The Psychology and Sociology of Literature: In Honor of Elrud Ibsch (Utrecht Publications in General and Comparative Literature)

The Psychology and Sociology of Literature

Utrecht Publications in General and Comparative Literature Editorial Board Hans Bertens (chair) – Douwe Fokkema – Harald Hendrix Joost Kloek (secretary) – Sophie Levie – Ann Rigney International Advisory Board David Bellos University of Manchester

Elrud Ibsch Vrije Universiteit Amsterdam

Keith Busby University of Oklahoma

Jørgen Dines Johansen University of Odense

Matei Calinescu Indiana University

Donald Maddox University of Connecticut

Yves Chevrel University of Paris-Sorbonne

Virgil Nemoianu Catholic University of America

Erika Fischer-Lichte Free University Berlin

John Neubauer University of Amsterdam

Armin Paul Frank University of Göttingen

Stephen G. Nichols University of Pennsylvania

Gerald Gillespie Stanford University

Willie van Peer University of Munich

Hendrik van Gorp Catholic University of Louvain

Roland Posner Technical University of Berlin

Thomas M. Greene Yale University

Bernhard F. Scholz Groningen University

Claudio Guillén Harvard University

Maria-Alzira Seixo University of Lisbon

Walter Haug University of Tübingen

Mario J. Valdés University of Toronto

Linda Hutcheon University of Toronto Inquiries and submissions should be addressed to: The editors, Utrecht Publications in General and Comparative Literature Vakgroep Literatuurwetenschap, Utrecht University Muntstraat 4, 3512 EV UTRECHT, The Netherlands

Volume 35 The Psychology and Sociology of Literature: In honor of Elrud Ibsch Edited by Dick Schram and Gerard Steen

The Psychology and Sociology of Literature In honor of Elrud Ibsch Edited by Dick Schram Gerard Steen

John Benjamins Publishing Company Amsterdam
/
Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data The Psychology and Sociology of Literature : In honor of Elrud Ibsch / edited by Dick Schram and Gerard Steen. p. cm. (Utrecht Publications in General and Comparative Literature, issn 0167–8175; v. 35) Includes bibliographical references and indexes. 1. Literature--History and critcism. I. Ibsch, Elrud. II. Schram, Dick H. III. Steen, Gerard. IV. Series. PN36.P78 2001 809--dc21 2001025303 isbn 90 272 2224 X (Eur.) / 1 58811 024 9 (US) (Hb; alk. paper) © 2001 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

"ste1"> "gro"> "sch1"> "gra"> "zwa"> "dij"> "han"> "hoo"> "ste2"> "gio"> "pee">

TITLE "Table of contents"

SUBJECT "Utrecht Publications in General and Comparative Literature, Volume 35"

KEYWORDS ""

SIZE HEIGHT "220"

WIDTH "150"

VOFFSET "4">

Table of contents

Preface The empirical study of literature: Psychology, sociology, and other disciplines Gerard Steen and Dick Schram

vii

1

Literary creativity Norbert Groeben

17

Qualitative methods in studying text reception Margrit Schreier

35

Agency, plot, and a structural affect theory of literary story comprehension Arthur C. Graesser and Bianca Klettke

57

Time in narrative comprehension: A cognitive perspective Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

71

Old readers: Slow readers or expert readers? Katinka Dijkstra

87

What we know about reading poetry: Theoretical positions and empirical research David Ian Hanauer

107

A renaissance perspective on the empirical study of literature: An example from psychophysiology Johan F. Hoorn

129

A rhetoric of metaphor: Conceptual and linguistic metaphor and the psychology of literature Gerard Steen

145

Irony and its discontent Rachel Giora

165

Psychoanalysts and daydreaming Willie van Peer and Ingrid Stoeger

185

< /R/TREARGET E FF

"sch2"> "hak"> "ben"> "ree"> "tan"> "klo"> "jan"> "noo"> "ver"> "mar"> "mia"> "cup"> "sch3"> "cont"> "toc">

vi

Table of contents

Back from the grave: Reinstating the catharsis concept in the psychology of reception Brigitte Scheele

201

How to make alle Menschen Brüder: Literature in a multicultural and multiform society Jèmeljan Hakemulder

225

“Sad autumn” and cultural representations: A comparative study of Japanese and Israeli “autumn” Ziva Ben-Porat

243

The fragmentation of the media audience Kees van Rees and Koen van Eijck

261

Bypassing the author: Two examples of reading interactive stories Ed Tan and Sarita Dev

289

Readers and reading behavior in the past: The question is: What is the question? Joost Kloek and José de Kruif The empirical study of careers in literature and the arts Susanne Janssen Stories and social structure: A structural perspective on literature in society Wouter de Nooy Quotations expressing the valuative stands of literary reviewers Hugo Verdaasdonk The proper place of humanism: Qualitative versus scientific studies of literature Colin Martindale An evolutionary framework for literary reading David S. Miall High and popular culture from the viewpoints of psychology and cultural studies Gerald C. Cupchik and Garry Leonard

315 323

359 379

395 407

421

Operative fictions: The fabric of societies Siegfried J. Schmidt

443

List of contributors

459

Subject index

463

Name index

469



Preface This book presents a selection of invited and refereed chapters on the empirical study of literature, collected on the occasion of the retirement of professor Elrud Ibsch from the Department of Poetics at the Free University, Amsterdam. The editors gratefully acknowledge the academic support from the following scholars, who were willing to act as referees: Salvatore Attardo, Ernst Boesch, Gillis Dorleijn, Ernest Goetz, Peter Hagoort, Norman Holland, Walter Hussy, Bipin Indurkhya, John Kennedy, Colin Martindale, Jill May, Richard McCallum, Aldo Nemesio, Keith Oatley, K. Ohtsuka, Claus-Michael Ort, Siegfried Preiser, Ann Rigney, Walter Schönau, Wilbert Spooren, Elizabeth Stine-Morrow, J. Straub, Reuven Tsur, Jutta Wermke. Without their professional support, this book would not have reached the desired level of quality. The editors also gratefully acknowledge the financial support of the Onderzoekschool Literatuurwetenschap / Netherlands Graduate School for Literary Studies, as well as of Norbert Groeben of the University of Cologne. Without their support, this volume would not even have seen the light of day. Last but not least, the editors are extremely grateful to the technical support and expertise provided by Lauraine Sinay, of Tilburg University, in preparing the manuscript. Dick Schram Gerard Steen



The empirical study of literature Psychology, sociology, and other disciplines Gerard Steen and Dick Schram Free University Amsterdam

1.

Aim of this book

The title of this book, The Psychology and Sociology of Literature, is a reflection of the major concerns of what is more generally known as the empirical study of literature (Ibsch et al. 1991; Janssen and Van Dijk 1998; Kreuz and MacNealy 1996; Rusch 1995; Tötösy de Zepetnek and Sywenky 1997). Psychology and sociology are the two social sciences that have provided theories and methods for a new and empirical approach to the study of literature which has developed over the past twenty-five years, mainly between Europe, Israel, the United States, and Canada. Most of this research has been carried out by members of the international association for the empirical study of literature, called IGEL, by the German acronym for “Internationale Gesellschaft für empirische Literaturwissenschaft.” The association was only founded in 1987, and the main publication venues for these researchers are almost just as young: they are the international journals of Poetics, published since 1971, SPIEL, published since 1982, and Empirical Studies of the Arts, published since 1983. The research is mainly produced by literary scholars, psychologists, and sociologists, although other disciplines are also represented. It is the general aim of this book to give an overview of the state of the art of the empirical study of literature as it is reflected in the work of the members of this association, emphasizing the role of literature as a psychological and social phenomenon. What joins these researchers is their aim to describe and explain psychological and social aspects of literature in the manner of the social sciences. One result of this approach is that literature may be compared and contrasted with other forms of discourse and culture and their treatment by the social sciences. For instance, there is a self-evident relation between literature and the other arts, which is also obvious from the title of the third international journal mentioned above. And there is another natural relation which is reflected in the new subtitle for the

2

Gerard Steen and Dick Schram

journal of Poetics, “Journal of Empirical Research on Culture, Media and the Arts.” These are only two of the platforms on which literature may be compared and contrasted with other forms of cultural behavior and activity that are investigated by the social sciences. But what is most valuable about this social-scientific treatment is not its capacity for comparison but its empirical quality, which suggests that the research is based on testable theories and produces valid and reliable knowledge. There are many traditions of research in the humanities, and literary studies is characterized by a host of distinct approaches, but not all of them can be said to be empirical in this sense. Literary scholars, together with most art scholars, are among the last to question the possibility of a scientific approach to such cultural products as literature and the arts, even though mass media research has long proven that it is possible to study comparable phenomena like the press, television, and popular culture from an empirical perspective. The deliberate combination of a scientific approach and an artistic object of research is still shocking to many academics in the humanities. This is because many scholars in the Arts Faculty still live by the model of an academy and, indeed, sometimes of a society that is divided into two cultures, as was proposed by C. P. Snow in his famous essay of that title. Many philosophers of science have argued the unity of scientific research, but this has not had a great impact on most of the humanities (Livingston 1988). When in the 1960s the Dutch universities established a new department in the Arts Faculties for the general study of literature and labeled it ‘department of a general science of literature’ (literatuurwetenschap), this was and always has remained a controversial label. And this even happened in the days when there were only theories of literature. It may be imagined that the emergence of empirical research that goes on to test such theories has only increased the resistance to a scientific approach to literature. The controversy became so heated that, in Germany and The Netherlands, there arose a genuine competition between paradigms in the late seventies and early eighties (e.g. Ibsch and Schram 1987). Internationally renowned psychologists and sociologists teamed up with empirical scholars of literature, but this has not impressed traditional scholars of literature much. On the contrary, what it suggests to them is that we cannot be ‘doing literature’ but have relinquished our precious object of study to the barbarians, the number-crunchers, and the nerds. This book is based on the assumption that this traditional view of the incompatibility between literature and science is misguided. We shall argue in the next pages that the empirical study of literature embodies a shift of perspective and emphasis in comparison with more traditional literary scholarship, but that it remains within the bounds of literary studies proper. The empirical study of literature is concerned with central questions of literature as a cultural practice, including its various effects on readers, various aspects of the way literary texts are

The empirical study of literature

read, and diverging aspects of the way literary texts are mediated and consumed as cultural products that have to compete with other texts and other media. None of these aspects can be ignored by any literary scholar who is worthy of that name. As the results of the empirical study of literature are beginning to accumulate, albeit in still modest fashion, the legitimacy of the undertaking will become less and less questionable. We hope that this book may act as a source of inspiration to new students of literature who are less upset by adopting an empirical approach to literature and related cultural phenomena. If it achieves that, then one day even the most hard-boiled non-empiricist will have to take note of our findings.

2.

Occasion of this book

It is no exaggeration to say that the present book could not have been produced if it had not been for the perseverance of the pioneers of the first hour. They had a world to conquer but a mere vision for a weapon and a solidly entrenched paradigm for their opponent. Today, about twenty-five years on, we are beginning to reap the first fruits of that vision. The improved present-day situation must give satisfaction to that first generation of empirical scholars of literature, one of who is the occasion for the production of this volume. For it is the purpose of this collection to provide an overview of the current state of the art in the empirical study of literature in order to honor the scholarship and retirement of Elrud Ibsch, who has been chair of ‘general literary science’ at the Free University Amsterdam for twenty-five years. Elrud Ibsch was one of the first literature professors in The Netherlands who occupied a chair of that name. At a very early stage of her career, however, she realized that theories of literature can only be called scientific if they are subjected to empirical testing, and, what is more important, that this objective cannot be reached by single text interpretation by the individual literary scholar. Instead, a more rigorous methodology has to be pursued, and Elrud Ibsch was fortunate enough to witness a historically significant moment in the development of the humanities and social sciences. She arrived at the joint rise of reception studies and reader-response theory in literary studies and of the psychology of reading and experimental aesthetics in cognitive science. She did not hesitate for one moment: linking up with the theoretical inspiration of Siegfried J. Schmidt (1980) and the methodological inspiration of Norbert Groeben (1977), she became a champion in The Netherlands of the empirical study of the literary reading process, or literary reception. The doctorate dissertations of the both of us, and of Johan Hoorn, who also has a chapter in this volume, were the immediate results (Buurman 1996; Hoorn 1997; Schram 1985; Steen 1992). Elrud Ibsch’s direct involvement in the broader development of the empirical study of literature in The Netherlands may be discerned from her pivotal role in the foundation of the Research School for

3

4

Gerard Steen and Dick Schram

Poetics, which has stimulated empirical research into literature as one of its two main programs. And, together with Siegfried Schmidt and Norbert Groeben and a number of other first-generation empiricists, including her husband Douwe Fokkema, she was responsible for the birth and the raising of IGEL to the flourishing association it has now become. Therefore we would like to dedicate this volume to Elrud as a token of our gratitude for her services rendered to the discipline.

3.

What is the empirical study of literature?

A very rough characterization of literary studies before the advent of the empirical study of literature could list the following main practices of literary scholars: – – –

interpretation and evaluation of individual texts; description of groups of texts (e.g. authors, genres, periods, histories, languages, countries, cultures), often from a comparative perspective; theory formation, that is, formulation of generalizations across (groups of) texts irrespective of the individual cases.

Each of these activities may be more or less focused on the text or its relation to context, such as the biography of the writer, the cultural conditions for production and reception, and so on. What is most important at this moment is to observe that the practice of theory formation provides a bridge between traditional literary scholarship and the empirical study of literature. It will be hard for any literary scholar engaged in individual text interpretation and evaluation or in the description of groups of texts to ignore the available theories about general aspects of literature without turning into an irresponsible academic. For theories offer general statements about aspects of literature which are also exhibited in one form or another in individual texts or groups of texts, in whichever way they are studied. Theories thereby provide an inescapable backdrop against which interpretation, evaluation, and description of a particular object of study take place. Even if one wants to emphasize the individual uniqueness of a particular writer or text, or, to go much further, of a particular metaphorically used word in one sonnet, there is always the assumption of a general trend, standard or norm. These often receive independent theoretical description and explanation, even if not always in the full-blown manner that scientific theories commonly require. And theories are indeed part and parcel of traditional literary studies, as may be witnessed from the success of Wellek and Warren (1949). But there are theories and theories. Together with Douwe Fokkema, Elrud Ibsch provided an exhaustive overview of the situation in the first 75 years of the twentieth century (Fokkema and Kunne-Ibsch 1977). They looked at Russian Formalism, Czech and French Structuralism, Marxism, Semiotics, and Reception

The empirical study of literature

Theory and indicated how these theories compare with each other and how they stand up against the criterion of testability. This critical discussion of twentiethcentury literary theory still remains a valuable source of insights to both traditional and empirical scholars of literature. But more interesting for our present purposes, this theoretical interest was then further developed into a more empirical direction, with psychological and sociological research leading the way into the investigation of all kinds of aspects of literary processes of production and reception and so on. The foundation of IGEL and the publications in the professional journals mentioned above were the result of these developments. The nature of this social-scientific research was seen as diametrically opposite to the dominant image of research in most Arts Faculties, which are typically concerned with the individual understanding and appreciation of the individual text, either in or outside its historical context. That this is a caricature of what generally happens in the Arts Faculties may be discerned from the list of practices offered at the beginning of this section: not every literary scholar is a literary critic. But the accidental hegemony of textual interpretation in the sixties and seventies in many countries in the West may have prevented a more accurate picture of the richness of traditional literary studies. However this may be, the dominant image of literary studies as text interpretation triggered a sharp contrast between the traditional approach to literature, characterized as the hermeneutic understanding and appreciation of the text, and the overt drive for a general and empirical description and explanation of psychological and sociological processes relating to literature (Ibsch and Schram 1987). This contrast was deepened when some empirical scholars of literature claimed that literary interpretation and evaluation belonged to their object of study instead of being part of literary studies proper (Schmidt 1980; Van Rees 1983). They argued that these academic practices were cultural instead of scientific, serving the function of the transfer of a tradition of knowledge, opinions, attitudes, and values. Even if it has to be conceded that some interpretation and evaluation may be treated in this way, for instance the practice of reviewers in the printed press, it simply goes too far to say that all interpretation and evaluation, including all its academic forms, are not scientific, but an object of scientific study. Not every literary critic is a reviewer. It will be appreciated that such a position is not conducive to promoting a positive attitude towards the new empirical study of literature. This polarization of views went way beyond what was on the agenda and may probably be explained in retrospect by the social aspect of a new approach attempting to find its own space in the academe. For it has to be admitted that, at least in principle, text interpretation and evaluation do not have to be ‘mere’ cultural practices, but that they also embody scientific analysis and description, making use of general observations and theories in the process. This is exactly what has been emerging in some areas of the related fields of linguistics and discourse studies

5

6

Gerard Steen and Dick Schram

(Coulthard 1994; Van Dijk 1997), which are conspicuously absent from the empirical study of literature as it has developed until today. In a truly mature empirical study of literature, however, the description and relative ordering of texts with respect to each other will have to be one of the goals to be pursued, as may be illustrated by means of the still solitary study by Martindale (1975). It will at least be interesting to see how far such projects can be developed without breaking down under an increasing lack of reliability (cf. Ibsch 1996). In fact the whole problem can be dissolved if it can be seen as a tension between focusing on the general or the particular as two poles on one scale. This is a tension that is inherent in all scientific investigation and should not be hypostasized into a distinction between nomothetic and idiographic forms of science. The one interest simply cannot do without the other, and at the end of the day the one interest does not exclude the other either. Nomothetic research abstracts away over individual cases while idiographic research zooms in on one particular case, but in both cases, there is a figure-ground relationship between the two which may be reversed. Such a perspective reveals some blind spots of the empirical study of literature itself, in that it points out that there has been a relative neglect of the idiographic perspective. But such a perspective also provides a challenge to traditional literary scholars to take the new approach for what it really is, a rigorous investigation of general aspects of literary communication which have remained ignored for far too long. What provides a common basis to these two opposite perspectives is the empirical quality of the research, which means that the theories are testable and that the methodologies aim for validity and reliability. Such research is not identical to quantitative research, as is often falsely assumed, but may be carried out by means of quantitative or qualitative methods, as is also pointed out by Schreier in this volume. It would be very odd indeed if literature, or the arts in general, were the only exception in our entire human world which would not be amenable to this kind of scientific investigation. To those who are still skeptical at this stage, we would like to say that it is at least worth the try. Some of these general concerns are directly reflected in a number of chapters in this book. There is a chapter by Colin Martindale which emphasizes the generalizing, nomothetic approach that has characterized social science and its application to literature. Martindale adopts a provocative standpoint which demonstrates the most radical form of nomothetic scientific research in the empirical study of literature. That such a standpoint does not preclude personal engagement or the use of rhetorical strategies for persuasive purposes is aptly shown by this special essay. Another chapter that emphasizes the generalizing objective of the empirical study of literature is the one by David Miall. He explores the idea that literature has survival function for humans as biological organisms and uses the perspective of evolutionary psychology to construct his argument. If most nomothetic research would not go this far, Miall clearly takes the generalizing bull by the horns and

The empirical study of literature

places the theme of universals on the research agenda. Like Martindale’s, Miall’s chapter is also an essay more than a research report, demonstrating that the empirical study of literature is not devoid of speculation, either. The trick is to relate the speculation to the criterion of testability, and this is something which is admirably achieved by Miall. However, the tendency for generalization is counterbalanced by the essay by Gerry Cupchik and Gary Leonard. Between themselves they embody not just the co-operation between a social scientist and a literary scholar, but a literary scholar who is firmly located in cultural studies, that is, outside the now recognizable domain of the empirical study of literature. They engage head on with the tension between idiographic and nomothetic research and construct a number of bridges between the two when they examine literature as part of the relation between high and popular culture. Their chapter provides proof that empirical and traditional scholars of literature are actually capable of mutually intelligible and interesting discussion. It is moreover a chapter that relates the interest in literature to empirical studies of the other arts and the media. A more wide-ranging historical perspective on the position of literature in culture and society than the one restricted to the modern period in Cupchik and Leonard’s chapter is offered by Siegfried Schmidt. He has remained true to the generalizing theoretical character of his previous work, which has laid the foundations of the empirical study of literature for many of the contributors to this book (Schmidt 1980, 1982). He has also remained on a course of riding ever widening circles when he now addresses the relation between literature and other media systems and re-examines the role of fictionality as a constitutive factor for literature. His essay sketches the new theoretical frontiers that have to be reached by new proponents of the empirical study of literature, and it forms a befitting closure to our volume. We have placed these four essays at the end of the book because they provide a sense of reflection and outlook on the state of the art that can only be appreciated if the reader has been informed of the state of the art itself. These chapters look backwards and forwards in different, complementary and sometimes conflicting ways. We hope that they will function as rather more personal statements and examples of the possible directions of the discipline as a whole.

4. Psychological and social aspects of literature As the commonest experience of literature is the one of literary reading, we have chosen to begin our overview with a number of chapters dealing with psychological aspects of reading. This does not mean that the psychology of literature is restricted to literary reception, for the production process may also be studied from a

7

8

Gerard Steen and Dick Schram

psychological perspective, and there are other questions to be asked by a psychologist about literature, too, as is demonstrated by the opening chapter by Norbert Groeben. He examines the fundamental issue of literature as creativity, and breaks it down into the creative side of literary production, literary texts, and literary reception. We are particularly happy to be able to include this piece because it highlights the need for many more studies on the psychological basis for literary writing, both in terms of personality traits of writers (see also the chapter by Janssen in this volume) as well as in terms of more detailed studies of the composition process. Another important aspect of Groeben’s chapter is the close attention it devotes to literary texts as products of creation (writing) and stimuli for creation (reading), which is another issue in the empirical study of literature that has remained underexposed. The introduction of the notion ‘creation’ in connection with literary reading, finally, is another seminal move, as it may facilitate discussion about the distinction between the construction of meaning — which takes place in all language use — as opposed to the creation of meaning, which may be restricted to more imaginative forms of reading. It is not fortuitous that one of Groeben’s colleagues, Margrit Schreier, contributes a methodological overview of ways to analyze literary texts in a scientific fashion. This is a dearly needed contribution to the state of the art, for the literary text as an autonomous object of study has been neglected by empirical scholars of literature, one or two exceptions excluded. What Schreier emphasizes in her overview is the possibility of transcending the divide between quantitative and qualitative approaches to text analysis and reception, so that testability is retained while information value increases. This chapter opens the gate for linguists and discourse analysts to become involved in the empirical study of literature and fill one of the gaps that has remained between poetics and the social sciences. It also draws attention to computational aids in performing text analyses, which, in the days of corpora like the British National Corpus (about 100 million words) and the Bank of English (over 350 million words) are becoming of paramount importance. The most familiar psychological aspect of literature will be the one of literary reading, and we begin with connecting this type of research to the predominantly cognitive psychology of all reading. One of the best-known psychologists in this field is Art Graesser, who has worked on the interaction between the text and the reader for more than two decades. In particular, his work has concentrated on the kinds of inferences readers draw at which moments in which types of reading, and he has not neglected literary comprehension in this research. The contribution he has offered for the present volume, together with Bianca Klettke, is concerned with how readers track the speech and knowledge of characters and what Graesser and Klettke call pragmatic agents in narrative fiction, the implied author, the narrator, the narratee, and the implied reader. Graesser’s work is particularly interesting to the literary scholar who believes that literary reading is special because readers have

The empirical study of literature

to do so much construction of meaning of their own, as has been postulated by reception theorists like Iser in the 1970s (Iser 1976). Not only does Graesser provide empirical evidence and detail for the assumption that literary readers do construct meaning, but, what is more challenging, Graesser has also shown that such processes are at work in all kinds of reading. The effect of this kind of work is that the ball is back in the court of the literary theorists who now have to specify where, when, how, and especially why literary reading is more or differently constructive than non-literary reading. The connection with Groeben’s chapter on creativity may be recalled here. One generally accepted assumption which Graesser and his associates make is that readers keep track of characters and their doings by constructing and maintaining a situation model (e.g. Graesser and Zwaan 1995). This is a notion which was first introduced by Van Dijk and Kintsch (1983) and has proved very productive for reading research (Kintsch 1998). One aspect of situation models is their temporal dimension, and this is the focus of the chapter by Rolf Zwaan and his associates Carol Madden and Robert Stanfield. They provide an overview of the assumptions made by traditional and modern linguists about the effects of the representation of time in narrative text and continue by discussing the experimental evidence that has been collected in cognitive reading psychology for the validity of these claims. Thus the predicted effects of, for instance, using different aspect forms like the progressive -ing form or the perfective aspect with the auxiliary verb have are borne out by various experimental data coming out of discourse processing studies. These are data which may help validate linguistic and stylistic text analyses which make predictions about the effects of particular stretches of text on the reader. Literary critics who ignore findings like these do so at their own peril. A related chapter is the one offered by Katinka Dijkstra on the difference between younger and older readers of literature. Most of the test subjects used in experiments reported by Graesser and Zwaan and others are university or college students of twenty-odd years old. Thus the question arises how representative these findings are for the population of all readers of literature at large. The general attitude towards this problem in experimental psychology is that the mechanisms under study are generally functional for all readers, as may be expected of inferencing or situation model maintenance during reading. However, it will be readily admitted that there may be variation across readers as to the strength or use of a particular discourse strategy or operation. This is the topic of Dijkstra’s chapter, who explores the available research on the effect of age on reading speed, comprehension, and recall in an attempt to judge whether older readers are ‘just’ slower readers. The alternative idea is that they are slower readers indeed, but not just because their reading speed decreases with age, but also because they use reading time in a different manner, for different reading purposes, and against a richer knowledge store. Given that readers of literature tend to be older rather than

9

10

Gerard Steen and Dick Schram

younger (see Van Rees and Van Eijck, this vol.), there are interesting implications of this research for joint work between psychologists, sociologists, and educationalists. The three preceding chapters all concentrate on the reading of literary narratives. David Hanauer provides an overview of what we know about reading poetry. He builds an explicit link between the theories of poetry reading offered during the twentieth century and the psychological tests that have been performed of aspects of these theories. His basic claim is that there has been a division between those theories that have argued for a relatively strong and determining role of the language of poetry, such as Russian Formalism, New Criticism, and the New Stylistics, on the one hand, and theories that have emphasized the role of the reader, such as Reception Theory and Reader Response Theory, on the other. In the words used above to characterize the work of Graesser and his associates, the question is: Which is the stronger factor in the interaction, the text or the reader? As may be expected, the answer is probably mixed, and Hanauer’s chapter explains which factor is stronger than the other in which study, and why. This contribution is highly suggestive of the various directions future work may take. Johan Hoorn’s chapter makes a more explicit statement about the future of the empirical study of literature, arguing that the empirical scholar of literature should follow the Renaissance precept of translation, imitation, and emulation of the classics. His view is that, for the empirical scholar of literature, the psychologists are the classics, and it is their work which should be translated, imitated and emulated. Hoorn uses metaphor and related notions like similarity and contrast as a case study to demonstrate how this could be done. He takes theoretical insights from literary theoretical giants like Roman Jakobson and turns them into a proposal for a study measuring the brain waves of test subjects who read metaphors in and outside their poetic context. The eventual control over the complex stimulus of the literary text in a carefully designed psychological experiment fulfils Hoorn’s dream of emulating the masters. A more programmatic view of the empirical study of metaphor in poetry, literature and discourse at large, is offered by Gerard Steen. His is one of the few chapters in this volume which concentrates on the analysis of literary texts, and in particular the metaphors they contain. As soon as one addresses this issue, the question arises how the identification and analysis of metaphor as metaphor can be performed in a reliable and valid fashion. For there are almost as many theories of metaphor as literary scholars and linguists, everybody having their personal story. Steen is no exception, but he claims that his approach is generally acceptable and he has persuaded a group of linguists from various backgrounds to co-operate on a joint research program, presented in his chapter. This is also connected to the methodological concerns voiced by Schreier’s chapter about the possibilities for high-quality text analysis. If this program is successful, it will be useful for text analysis as well as for reading research, in which the representative nature of stimulus materials needs to be better controlled than it is today.

The empirical study of literature

It is a small step from metaphor to irony. Irony is another rhetorical device which has been used with great frequency in literary texts, and, like metaphor, irony in general has received a great deal of attention in psycholinguistic investigation. Rachel Giora has reviewed the dominant theories in the field and presents her own salience effect approach to provide an integration. Her chapter shows how linguistic and pragmatic theories have fed into psycholinguistic models of irony processing, and is suggestive of the ways in which the results of such general empirical studies can be important for literary students of irony. For instance, familiar irony is processed differently than unfamiliar irony, and literary texts make use of both of them for different purposes. The fruitful distinction between conventional and novel metaphor and their relation to literature in Cognitive Linguistics (see the chapter by Steen) may serve as a model for the further development of this line of research. The detailed effects of irony, metaphor, and other aspects of poetic and literary language are an area of research in literary text reception that is closely related to psycholinguistics. However, there are also more general cognitive effects of literature that go beyond the relatively short-lived effects of language, and these have also been addressed. The chapter by Willie van Peer and Ingrid Stoeger is an experimental examination of some of the predictions of psycho-analytic theories of literary reading, which occupy an important position in interpretative approaches to literature and culture. Van Peer and Stoeger have tested the claims that fairy tales stimulate readers to allow repressed feelings of hate against their mother, and that incestual wishes directed at their brothers and sisters are liberated and intensified by reading fairy tales. This chapter provides an apt example of how the empirical study of literature may function to improve the quality of our literary theories and knowledge about the presumed effects of literature on its readers. If psycho-analysts claim that the function of literature is to celebrate hidden fantasies, this is not borne out by the data collected by Van Peer and Stoeger, who fail to find any of the predicted effects. It seems to us that, in the face of these findings, there are two possible replies by the psycho-analysts. They either attack the study on theoretical grounds (for instance, that the predictions were incorrectly derived from the theory) or methodological grounds (for instance, the test applied was not suited to the task at hand), or they re-examine their own theories and come back with new, testable versions. Either way, the effect of the study would be to engender further critical discussion between psycho-analysis and the empirical study of literature. A related classic theory of the function of literature, referring to another psychological effect of literary reception, is the one of catharsis. Brigitte Scheele offers an overview of the psychological research into Aristotle’s famous concept, and shows how empirical research always has to be related to theoretical analysis and re-analysis. Scheele argues against the received notion in psychology that the catharsis hypothesis is dead and buried by pulling back to the center of the discussion the distinction between catharsis as a process of purging the mind as opposed

11

12

Gerard Steen and Dick Schram

to catharsis as a process of purifying it. She shows that research has concentrated on catharsis as purging and has neglected the purificatory view of catharsis, which, Scheele argues, may be profitably elaborated and investigated in the near future. Her chapter provides a showcase for the way theory and research are interdependent in the social sciences, and that long-term historical analysis of their development may actually reveal the possibilities for achieving progress. The chapter can also suggest how we may imagine a debate between psycho-analysts and researchers like Van Peer and Stoeger. A third psychological effect of reading explicitly addresses the relation between literature and values, a theme which was already partially present in the previous two chapters. Jèmeljan Hakemulder offers an overview of the experimental psychological work done on the effect of positive outgroup portrayal in texts upon the reader’s perception of outgroups, and he raises the question whether literary reception may actually enhance the sympathy we have for members of outgroups. His results show that perceived outgroup homogeneity may be reduced by reading stories of a specific kind. It is self-evident that this exercise in social psychology presents a valuable addition to the relatively cognitively oriented chapters presented at the beginning of this book. We believe that social psychology holds many promises for the empirical study of literature, and we would like to draw the reader’s attention to related remarks by Cupchik and Leonard, Janssen, Martindale and De Nooy in this volume. There even beckons an applied perspective if studies such as Hakemulder’s become more numerous, because there is a social function of literary reading hidden behind Hakemulder’s project, which may be explored and exploited in formal education in many of our multi-cultural societies. A variant on this theme is provided by Ziva Ben-Porat’s chapter on sad autumns. She shows how one Western literary model of portraying autumns as a specific type of season with all kinds of properties was imported into Israeli culture and became the dominant cultural model, even though Israeli nature does not correspond with some of the prominent features of the Western experience of autumn at all. Readers’ interpretations of literary texts proceed by activation of the cultural model, while knowledge of real Israeli autumns only comes in when the cultural model is questioned by the literary text. This is a persuasive illustration of the power of literary texts to affect our perception of other texts and even of reality beyond and outside literature. The connection with the concerns of a social psychology of literature raised in the previous paragraph should be obvious. This becomes even more manifest when Ben-Porat turns her attention to an intercultural comparison between Israel and Japan, carefully distinguishing between variants and invariants in the portrayal and use of sad autumns by these two radically different cultures. This is nothing less than a research program into the relations between the individual, language, art, and culture.

The empirical study of literature

Cultural representations can only be transmitted if people participate in media usage. General patterns of people’s participation in culture are the subject of the chapter by Kees van Rees and Koen van Eijck. They report on a study of media usage in The Netherlands by five types of readers: omnivores, information readers, entertainment readers, regional readers, and non-readers. They have found that, between 1975 and 1995, reading frequency manifests an overall decline, even though the group of non-readers has remained constant in size. Television has indeed become the serious competitor of reading, as many cultural pessimists predicted long ago. The use of the internet may add another threat to traditional reading culture. These data are important for our sense of the position and function of literature in present-day culture. The strategies of reading a narrative text from a computer screen have been explored by Ed Tan and Sarita Dev. They compare reading to searching, browsing, and surfing the web and discuss these activities in terms of effort and reward. Tan and Dev suggest that the different activities fulfil different needs and pleasures, which may offer some hope to those culture pessimists who thought that reading would fall by the way altogether. Two empirical studies present data on the reading of a story from the screen. Expectations about computer games were shown to interfere with the accomplishment of the task in the first study, which may be seen as another sign of the times. However, linear reading did tend to survive the new context in which the story was offered to the subjects: again, this may offer some consolation to those who think that zapping and lateral thinking will eradicate all traditional culture. If it is relatively easy to collect data about general patterns of reading and participation in contemporary culture, as was done by Ben-Porat, Van Rees and Van Eijck, and Tan and Dev, this is much more problematic when we turn to the past. Joost Kloek and José de Kruif address this issue in their essay on the empirical study of literary history. Despite the welcome addition of historical studies about reading behavior in terms of, for instance book buying and borrowing and reading circles, the authors observe that enthusiasm for the empirical enterprise is waning in literary history, and offer a number of explanations for this trend. They argue that literary historians should be more cautious and precise in their theoretical expectations in order to facilitate an optimal use of the widely available source materials. In their opinion, empirical research is not a lost cause in the historical branch of literary studies; instead, it requires researchers that are better trained in the traditions of the social sciences rather than the humanities. It is not just the readers of literature who participate in a more general culture, but the writer of literary texts is embedded in such a context, too. Susanne Janssen’s chapter provides a survey of the general characteristics of the career of literary authors. She goes into the changing support structures on which writers and artists have depended for their livelihood, such as old and new forms of patronage; the

13

14

Gerard Steen and Dick Schram

influence of mediators on individual works and complete careers; the properties of the occupation of being a writer; and the career development of writers and artists. This chapter should be connected with the first chapter of the book, by Groeben, and demonstrates how psychological and sociological research into writing and writers can meet and be jointly productive of many future findings. One important aspect of literary authors’ careers is their network with other authors. Wouter de Nooy discusses his previous research on author networks in a chapter that has a more general theoretical aim, to introduce network analysis as a method for literary studies (see the chapter by Martindale for a similar suggestion). De Nooy argues that network analysis can be applied to characters in stories just as aptly as it can to live persons in reality. He provides an illustration by means of a reanalysis of the Swan-Geese tale used by Vladimir Propp in his famous study of the narrative functions of folk tales (Propp 1968). This is an innovative chapter which suggests how social-scientific methods of empirical research can be developed and applied in empirical text analysis, broadening the scope of the empirical sociology of literature beyond the domain of writer and reader behavior. If attempts such as these have been made in the past by, for instance Lucien Goldman and others, they were oriented towards theory formation or interpretation; De Nooy’s proposal squarely stays within the limits of empirical research and raises the question how such previous attempts may be related to this new aproach. Another and final aspect of the career of writers is the way their texts are used for quotation in reviews by reviewers. Hugo Verdaasdonk presents a study which investigates the assumption that reviewers make a distinction between high-quality authors and less gifted writers when it comes to quoting passages in reviews from their works which presumably deal with the author’s overall intention. Verdaasdonk argues that the greater belief of reviewers in the superior quality of great authors induces reviewers to take passages about the author’s intention on trust and quote them relatively frequently. He predicts that other types of quotations, such as those serving to illustrate style or plot, are not affected by this initial assumption of reviewers, and his findings support his expectations. Here is another sociologist of literature who has turned to employ empirical methods to text analysis, albeit an analysis of meta-texts rather than object-texts. His results are important for an improved view of the way mediators of literature work and affect writers’ careers as well as readers’ opinions of literary works. This concludes our overview of the chapters. In the book itself, these chapters are followed by the essays by Martindale, Miall, Cupchik and Leonard, and Schmidt, which offer food for reflection and speculation. We hope that the book as a whole also offers a satisfactory range of suggestions for theory formation and empirical research on the psychology and sociology of literature. As we have attempted to suggest, there are many possibilities for translating ‘traditional’ literary theory into empirically testable theories and predictions, and this will lead to an

The empirical study of literature

increase in the empirical quality of our literary knowledge. Moreover, in comparison with traditional poetics, the empirical approach has led to a broadening of the scope of literary studies to new questions and areas of research. And finally, some of these studies have the potential of application in education and other societal domains, increasing the relevance of the study of literature. If the reader will come to share at least part of this view of ours, we have succeeded in honoring the scholarship of Elrud Ibsch.

References Buurman, P. 1996. Duitse literatuur in de Nederlandse dagbladpers [German Literature in the Dutch Daily Press]. Doctoral dissertation, Free University Amsterdam. Coulthard, M. (ed.). 1994. Advances in Written Text Analysis. London: Routledge. Fokkema, D. W. and E. Kunne-Ibsch. 1977. Theories of Literature in the Twentieth Century: Structuralism, Marxism, Aesthetics of Reception, Semiotics. London: Hurst. Groeben, N. 1977. Rezeptionsforschung als empirische Literaturwissenschaft: Paradigma- durch Methodendiskussion an Untersuchungsbeispielen. Tübingen: Kronberg. Hoorn, J. F. 1997. Metaphor and the Brain: Behavioral and Psychophysiological Research into Literary Metaphor Processing. Doctoral dissertation, Free University Amsterdam. Ibsch, E. 1996. “The strained relationship between the empiricist’s notion of validity and the hermeneutician’s notion of relevance”. In Kreuz and MacNealy, 23–34. Ibsch, E. and Schram, D. H. (eds). 1987. Rezeptionsforschung zwischen Hermeneutik und Empirik. Amsterdam: Rodopi. Ibsch, E., Schram, D. and Steen, G. (eds). 1989. Empirical Studies of Literature. Amsterdam and Atlanta: Rodopi. Iser, W. 1976. Der Akt des Lesens: Theorie ästhetischer Wirkung. München: Fink. Janssen, S. and Van Dijk, N. (eds). 1998. The Empirical Study of Literature and the Media: Current Approaches and Perspectives. Rotterdam: Barjesteh and Co’s. Kintsch, W. 1998. Comprehension: A Paradigm for Cognition. Cambridge: Cambridge University Press. Kreuz, R. J. and MacNealy, M. S. (eds). 1996. Empirical Approaches to Literature and Aesthetics. Norwood, NJ: Ablex. Livingston, P. 1988. Literary Knowledge: Humanistic Inquiry and the Philosophy of Science. Ithaca and London: Cornell University Press. Martindale, C. 1975. Romantic Progression: The Psychology of Literary History. Washington, DC: Hemisphere. Propp, V. 1968. Morphology of the Folktale. Austin/London: University of Texas Press. Rusch, G. 1995. Empirical Approaches to Literature: Proceedings of the Fourth Biannual Conference of the International Society for the Empirical Study of Literature — IGEL, Budapest, August 1994. Siegen: Siegen University. Schmidt, S. J. 1980. Grundriss der empirischen Literaturwissenschaft, 2 vols. Braunschweig/Wiesbaden: Vieweg. Schmidt, S. J. 1982. Foundations for the Empirical Study of Literature: The Components of a Basic Theory. Translation of Schmidt 1980, by Robert de Beaugrande. Hamburg: Buske. Schram, D.H. 1985. Norm en normdoorbreking [Norm and Norm Violation]. Doctoral dissertation, Free University Amsterdam.

15



16

Gerard Steen and Dick Schram

Steen, G. J. 1992. Metaphor in Literary Reception: A Theoretical and Empirical Study of Understanding Metaphor in Literary Discourse. Doctoral dissertation, Free University Amsterdam. Tötösy de Zepetnek, S. and Sywenky, I. 1997. The Systemic and Empirical Approach to Literature and Culture as Theory and Application. Edmonton: University of Alberta. Van Dijk, T. A. (ed.). 1997. Discourse Studies (2 vols.). London etc.: SAGE. Van Dijk, T. A. and Kintsch, W. 1983. Strategies of Discourse Comprehension. New York: Academic Press. Van Rees, C. J. 1983. “How a literary work becomes a masterpiece: On the threefold selection practised by literary criticism”. Poetics 12: 397–417. Wellek, R. and Warren, A. 1949. Theory of Literature. Harmondsworth: Penguin.



Literary creativity Norbert Groeben University of Cologne

The use of the term ‘literary creativity’ instantly raises the question whether different structures of creativity can, or must, be assumed to exist for different domains of human cognition, emotion and action. At our present state of knowledge, the answer to this question depends on the level of abstraction of the relevant domains: on the level of (specific) skills (that is concrete operations, working techniques in for instance mathematics vs. literature), we can certainly assume differences as to the ways in which creativity manifests itself (Baer 1993); on the level of (general) abilities and, especially, personality traits, however, it appears justified to suppose that the structures and processes of creativity are largely comparable (cf. e.g. Cropley 1995: 343ff.). It has to be stated here that to date, skills in literary creativity have hardly been studied empirically at all. For this reason, the present article is going to concentrate mainly on the general levels of ‘abilities’ and ‘personality’; nonetheless, as far as this is possible, those theoretical and empirical analyses relevant to the specific domain of literature shall be brought into focus as well. Still, due to the fact that since the beginning of the 1950s (and mainly during the 1960s) empirical research on creativity has boomed, it will be impossible to attempt a complete survey of the relevant literature here; instead, I am going to reconstruct the most important developments in theory, while at the same time supporting this theoretical reconstruction with evidence from those empirical studies that I regard as ‘classics’ in the research on creativity (and as relevant to literary creativity). In theoretical approaches to creativity, a distinction is usually made between the creative product (achievement), the production process (performance including environmental factors) and reception (cf. Preiser 1986: 24ff.; Martindale 1989). Speaking about the analysis of literary production presupposes that those criteria assumed to denote a creative achievement (product) are already known; for expository reasons, I am nevertheless going to adopt the sequence which is customary in the context of literary communication, that is ‘production — product — reception’. I would also like to point out that the influence of environmental factors

18

Norbert Groeben

on the production of literature has hardly been systematically investigated yet; we are merely supplied with historical anecdotes here (ranging from the rotten apple allegedly used by Schiller to the examples from interviews provided by Csikszentmihailyi 1996: 127ff.). Where production is concerned, I am therefore going to focus more on personality traits while environmental factors will — indirectly — be dealt with under the perspective of reception.

1.

Literary production

The issue of creative personality traits, pertaining to author psychology, has a long tradition in research, extending even into pre-scientific times. This tradition began in antiquity and set out from an idea which offers some consolation to those who count as ‘normal’, namely that creative individuals have to ‘pay’ somehow for their gift of being a genius. Initially, such a price to be paid was a physical deficiency, for instance the blindness of the prophetic eye in the imagination of ancient Greece (cf. Wellek and Warren 1963: 66). In more recent times, this ‘hypothesis of substitution’ has remained quite successful, especially in combination with ideas pertaining to the aesthetics of genius, as the postulate of a link between genius and mental illness (Lomboroso 1887). While ‘mental illness’ was initially understood as a form of epilepsia in this context, all other, endogenic and exogenic variants of psychosis were later on included as well. Evidence for this was gathered mainly from pathographic biographies of great artists. Such biographical collections certainly culminate in the work of Lange-Eichbaum who, in the version revised by Kurth (1986ff.), has collected next to 3000 pathographies. The methodological problem with this procedure is that the diagnoses are for the most part ‘blindfolded’; moreover, they raise the suspicion of circularity, since the respective artists’ genius as well as their psychosis are deduced from their creative work (cf. already Loewenfeld 1903). The resulting danger of distortions is also conspicuous in the phenomenon that contemporary evidence is rarer here than proof from the past, that is, the more remote they are in history, the more ‘mentally ill’ do artists become (cf. Groeben 1972:46f.). In contrast to this line of argument, a systematic testing of the psychosis hypothesis would have to provide statistical evidence that a higher proportion of such diseases occurs among artists than within the (‘normal’) population as a whole. Furthermore, those psychoses acquired (for instance after a disease such as syphilis) must not be counted, since they cannot provide any knowledge about an innate connection between genius and mental illness. Accordingly, methodologically correct testings (from Ellis 1904 via East 1938 to Goertztel and Goertzel 1965) have shown that psychosis does not in fact occur more frequently among creative people (‘geniuses’) than in the (‘normal’) population as a whole. Thus, the (causal)

Literary creativity

hypothesis of a link between genius and mental illness can be regarded as falsified (on the correlation between the two, cf. Goodwin and Jamison 1990; Prentky 1989). However, its falsification has not entailed the disappearance of the ‘hypothesis of substitution’; rather, it has persisted, in modified form, in the psychoanalytical variant of the neurosis hypothesis. Here, the inadequate assumption of an innate connection between genius and psychopathology is given up, while at the same time it is claimed that patterns of reaction to experience are parallel in creative and neurotic processes. This is at least the line of argument pursued within the classical version of the neurosis hypothesis according to Freud (1955 II, III, IV), in which creativity is understood as a regression to primary process thinking, that is patterns of thought marked by an — irrational — predominance of drive content and instincts. Just as in dreams, this predominance becomes manifest in logical contradictions, displacements, condensations, the absence of a feeling for linear time and so on (Brenner 1976: 140ff.). The crucial point of the neurosis hypothesis on creativity is that it retains the view of the regression as being compulsive, that is, independent of any possible influence of the ego. In the further development of psychoanalysis, however, the very admittance for such a possible influence by the ego has in turn led to the revision of the neurosis hypothesis and, in fact, to the fundamental rejection of the hypothesis of substitution in general. In the more recent development of psychoanalysis (‘neoanalysis’), first and foremost by the New York School pursuing the psychology of the ego, it has been established that artistic creativity consists in a combination of primary process thinking and ego control; this has been summarized convincingly by Kris (1952/77) in his concept of “regression in the service of the ego”. In the last consequence (drawn by Kubie 1966: 49f.), this means that creative production (performance) does indeed imply the recourse to preconscious processes, but not a regression rooted in a neurotic fixation which would restrict associative flexibility. Put in a programmatic way (as in Kubie 1966: 49f.), then, creativity is possible despite neurotic disorders, it does not arise from, let alone through, them. This rejection of the neurosis hypothesis by contemporary psychoanalysis (cf. Groeben 1972: 47f.) opened up new ground for the empirical-psychological investigation of the creative personality and creative performance (cf. Preiser 1986; Ulmann 1968). This empirical research has shown that creative persons are indeed characterized by high scores on scales of neuroticity and psychopathology. This discovery, however, does still not support the psychosis, or neurosis, hypothesis on creativity; this has been demonstrated above all by the classical, comprehensive study on the personality of contemporary creative writers by Barron (1963, 1967, 1968), in which 56 American writers were tested, by means of qualitative as well as quantitative diagnostic procedures. In accordance with other studies on this subject, Barron established the following central features of the creative personality: high productivity, a wide range of interests, above-average intellectual capacities, a high

19

20

Norbert Groeben

ranking of rationality and personal independence, acting on an ethical basis and so on; the parallel study on German writers by Schmidt and Zobel (1983) yielded corresponding results. The most interesting finding in regard to neuroticity scores and the like is related to the so-called psychopathology factors of the MMPI (Minnesota Multiphasic Personality Inventory). Here, writers rank among the top 15% of the entire population, which means that they are burdened with stronger anxieties than the average individual. At the same time, however, their scores on ego strength are also above average. This finding stands in exact opposition to the relationship between the two scales within the ‘normal’ population, where a negative correlation exists between scores on psychopathology and ego strength (between −.5 and −.6): the creative persons (writers) are therefore characterized by high psychopathology and high ego strength simultaneously. In the most general terms, this relationship found between the two scales among creative persons can be called the “polar integration” (Groeben 1981: 108ff.) of two features, in the sense that the polarities of features commonly conflicting in our socialization are reversed, that is synchronized (since, and insofar as, the opposition in question is not psychologically necessary but decidedly unnecessary, and in part even destructive). Accordingly, the creative personality turns out to be quasi-paradoxical (McMullan 1976): a personality which, by linking specific traits, succeeds in transcending unnecessary and dysfunctional contrasts (as for instance rational-emotional; critical-feeling of solidarity; masculine-feminine) and in turn comprises positive anthropological perspectives for the future development of humankind (Erb 1997). On the basis of existing research, McMullan has elaborated eight polarities which he considers as the central polar integrations of characteristics of the creative personality: “delayed closure; converging divergence; mindless perception; constructive discontent; detached involvement; disinterested selfishness; confident humility; relaxed attention” (McMullan 1976). The most recent collection of such features has been provided by Csikszentmihalyi (1996: 58ff.), who presents ten dimensions of polar integration: energy and relaxation; smartness and naivety; responsibility and irresponsibility; fantasy and sense of reality; extra version and introversion; humility and pride; masculinity and femininity; rebellion and conservatism; passion and objectivity; sensitivity to joy as well as to pain. Today, this paradoxical basic structure of the creative personality can for the most part be regarded as empirically confirmed. The concept of polar integration, however, is necessarily anchored in processes of socialization; this in turn implies that the exact contents of this basic structure will vary according to place and time. This means that in the future the development of the creative personality will (have to) continue to provide new dimensions of polar integration as a constructive response to socialization deficiencies. The basic structure of integrated polar dispositions also applies to those abilities affecting the creative process, for which especially the combination of so-called

Literary creativity

convergent and divergent thinking (based on the model of Guilford 1964) has become relevant (convergence: aiming at one single solution; divergence: going into several directions and dimensions; cf. the comparatively recent survey by Runco 1991). In this context, the perspective of abilities is usually treated in connection with that of performance (process); it has been postulated quite early on (for instance by Poincaré 1913) that the creative process comprises four phases: ‘preparation’, ‘incubation’, ‘illumination’ and ‘elaboration’. While the further development of the psychology of creativity (cf. the survey by Preiser 1986: 42ff.) has in part provided a more sophisticated distinction of phases (sometimes up to seven or eight), all of these can ultimately be referred back to the original four. Further interconnections between polar attitudes, opinions and abilities can be detected as underlying those phases; these are necessary in two ways, namely to produce a creative idea on the one hand and to meaningfully elaborate the creative product on the other: among them are involvement and distance of the ego, insistence and impatience, free association and logical processing, primary and secondary process and divergent/convergent thinking (cf. Groeben and Vorderer 1986: 110f.). This phase model of the creative process has hardly been tested empirically; this is mostly due to methodological reasons: for instance the selfobservation of the incubation and illumination phases by the creative individuals themselves would require a convergent attitude (during the self report) which, in turn, would be likely to impede or even prevent the creative idea in the first place. The only choice thus seems to be the attempt at indirectly accessing procedural aspects by analyzing certain features of the product. In this way, for instance Pine and Holt (1960) have shown that primary process contents (in the sense of an adaptive regression) are indeed constituents of incubation and illumination (Russ 1993: 17ff. provides a survey of the research based on these findings; cf. also Martindale’s evolutionary model of literary history presented below). On this basis the question arises whether the primary processes of incubation and illumination might be artificially enhanced, for instance by the use of drugs. It has been found that the stronger psychedelic drugs (such as LSD, Mescaline, DMT, DET: cf. Tart 1969) do indeed lead to complex experiences of a changed reality, hallucinations, and emotional disruptions, all of which may become manifest in neologisms, concretistic thinking, loss of the ‘As-If ’ capacity and so on (overview in Hoffer and Osmond 1967). Literary self-reports give a vivid impression of such drug effects (as for instance in Huxley’s report [1954] of his self-experiment with Mescaline). The corresponding systematic studies (Hoffer and Osmond 1967) have shown that these drugs do indeed make the regression to primary process thinking easier and hence appear to positively affect the creative process at a given moment; in the long run, however, they do not — structurally — lead to an increase in creativity, since the latter also involves the (convergent) elaboration of a creative idea. Such an elaboration, however, is impossible while the drugs are still at work.

21

22

Norbert Groeben

There is also another long-term and structural problem: mostly, experiences heightening one’s awareness pose very high (often unfulfillable) demands on the capacity to express them in language (during later phases of elaboration). It can thus be said that, at most, the influence of drugs leads to a momentary increase in the opportunities to exhaust one’s potential of experience (within the domain of primary process thinking), whereas it does not facilitate a substantial, structural increase in creativity.

2.

The literary product

Nevertheless, it is a well-known fact that drugs do fascinate some producers of literature. In the field of literature, too, this can be related to the central role assigned to the imaginative power ever since mediaeval normative poetics were abandoned. This, in turn, constitutes yet another manifestation of the basic structure of polar integration, this time concerning the criteria according to which a literary product is considered to be creative. It is widely agreed that originality or novelty is one such criterion (cf. Preiser 1986: 35ff.; Ulmann 1968: 29ff.). It is equally undisputed, however, that novelty must needs be supplemented by the criterion of usefulness, again a polarity marked by a relationship of tension: the accustomed, useful is generally not very original and the original, novel is not necessarily by definition useful (in the sense of being constructively appropriate to the question at stake); here as well, creative products are characterized by the need to achieve a parallel maximization of both criteria. The democratization related to the empirical concept of creativity manifests itself in the fact that originality is no longer taken to denote ‘historical-objective’ novelty, but instead a historically relative, subjective innovation (cf. Ulmann 1968: 39ff.). Martindale (1975, 1984, 1986) has provided evidence for the relevancy of the innovation factor in the historical development of literature by means of content-analytical research on ‘the psychology of literary history’. This is an ‘evolutionary’ theory of the aesthetic development within literature, reconstructed above all on the dimension of secondary vs. primary processes (cf. above Kris: “regression in the service of the ego”). Martindale assumes that there are two ways of increasing novelty — and, with that, aesthetic attraction: 1. given a stable secondary process elaboration, original and novel combinations can be achieved by an increased regression to primary process thought contents. 2. given a stable degree of regression (that is primary process thought contents), novelty becomes possible through the reduction of secondary process elaboration (and, consequently, the destruction, or modification, of conventional stylistic principles). Martindale postulates that the second of these possibilities occurs more rarely than the first, and mainly during transitional phases of change in literary style. This results in periods of oscillation,

Literary creativity

during which, after the establishment of a (new) literary style (qua type of secondary process elaboration), novelty is achieved by permanently increasing the degree of regression; when the point is reached at which a further increase in the primary process contents is no longer possible within the initial stylistic framework, so that the secondary process structures change, a new style emerges; after the establishment of that style, then, the sequence of increased regression starts all over again. Martindale has tested this phase model above all with reference to the (literatureinternal) aesthetic development of French and English poetry by developing and applying a coding scheme in order to operationalize the primary and secondary processes (with the subcategories ‘drives’, ‘sensations’, ‘defensive symbolization’, ‘regressive thinking’ and ‘Icarian symbolism’). In this way he was indeed able to confirm the sine-shaped cyclical oscillations (of primary process contents), and in part also a general increase of primary process contents from classical up to modern literature (cf. the summary in 1986). Ultimately, Martindale’s approach is particularly interesting because he primarily bases his confirmation of the novelty factor on (primary process) contents of thought and expression, whereas in general, the reconstruction of literary epochs within the framework of a history of literature puts much more emphasis on the (secondary process) aspect of (literary) form. While, in principle, such approaches to an aesthetics and history of literature can be reconstructed within the framework of psychological research on creativity focusing on literary innovation, it has to be said that so far, concrete studies in this area are still largely lacking and therefore highly desirable. Regarding the criterion of usefulness or appropriateness, it is decisive that this criterion is not to be defined as constant and invariable, but instead as flexible, that is, it is to be adapted to the problem in question. The usefulness of the creative product depends upon the problem to be solved, and it is in fact characteristic of the creative production process that it does not only yield a solution to the problem at hand — it also gives rise to entirely new problems (Runco 1994; Wakefield 1994). Especially where artistic production processes and products are concerned, then, the criterion of usefulness, at least in its basic form, does rest upon the criterion of novelty. Not unfrequently, objects of art — implicitly — contain new problems and questions; however, it can happen that these are recognized only after a very long time, after a structural change in the environment and so on (Ulmann 1968: 56ff.). This is the reason why assessments of the usefulness, also and especially of a literary product, (may) undergo dramatic changes in the course of time. This shows that in the object domain of (literary) creativity, reception has to be, at least approximately, as creative as production. The flexibility of the criterion of usefulness is illustrated for instance by the comparison, stimulated by the psychosis hypothesis, of (especially) modern literature and the verbal products of psychotic (above all, schizophrenic) patients (cf. for instance Navratil 1966). In this way, it has indeed been possible to detect

23

24

Norbert Groeben

comparable features: among these are the decrease of associative images and thoughts accompanied by a simultaneous increase in sound associations and linguistic stereotypes (such as the rigid repetition of words), neologisms (idiosyncratic ‘private’ expressions), agglutination of images (that is the merging of images in memory as well as in the imagination; cf. modern literary metaphors lacking ‘as’ or ‘like’ [Benn 1980]), and grammatical deformations (Spoerri 1964). In modern literature, however, this distortion of linguistic form is intended to trespass the boundaries of traditional aesthetics while in the case of psychotic persons (cf. above), it exists always and independent of their intention (that is as a compulsive feature; cf. also Rennert 1962). In evaluating such verbal products, it is thus decisive whether a deviation from everyday language serves a meaningful aesthetic function or not. This impression of a comparatively far-reaching parallelism between literary and psychotic (linguistic) features is also the result of the fact that (modern) aesthetics is undoubtedly very strongly focused upon formal features. This in part extensive neglect of the dimension of content, for instance in the evaluation of aesthetic products, has been severely criticized for instance from the feminist point of view (cf. for instance R. Klinger who points to the fact that there exists a wealth of literature on the artistic means of expression in the painting “The Rape of the Sabine Women”, whereas the fact that the painting does in fact show an actual rape — in both meanings of the word — is totally ignored). This leads to the question, whether the criterion of usefulness should not also comprise the aspect of humane relevance (that is an evaluation as to whether the respective work of art does in any way contribute to the further development of humankind). Of course, this would entail enormous problems of justification because so far any ideology, destructive as it may have been, has defined and presented itself as contributing to a higher development of humankind. On the other hand, discussions like those about literature potentially glorifying war (as for instance “Unter Stahlgewittern” by Ernst Jünger) show that such texts inevitably raise problems concerning literary evaluation. In this respect, the psychology of literary creativity would certainly profit from an interdisciplinary cooperation with research on literary evaluation; such a cooperation might possibly serve to overcome the traditional postulate of value neutrality which originated in a conception of psychology rooted in the natural sciences and has turned out to be rather dysfunctional in the field of cultural psychology (Groeben 1999: 365ff.). At the same time, research on literary evaluation might also profit from a cooperation with the empirical psychology of creativity, namely by substantially extending its domain of research by incorporating the (empirical) collection of data pertaining to evaluation processes and concepts within the literary system. It has to be said, however, that at present both of these perspectives concerning such a potential interdisciplinary cooperation have merely the status of goal perspectives which might be realized in future research.

Literary creativity

3.

Literary reception

‘Reception’ is understood to denote those psychological processes taking place during the reading of a text and resulting in an individual, subjective constitution of the concrete text meaning (Groeben and Landwehr 1991: 170f.; cf. Groeben and Vorderer 1988: 192ff.). In contradistinction, the aspect of text effects deals with the consequences of reading (which are not essential to the perspective of literary creativity). Independent of the question whether the received texts are literary or non-literary ones, reception is always a form of text processing characterized by cognitive constructivity (cf. Christmann and Groeben 1999); this means that information, besides merely being ‘extracted’ from the received text (that is ‘decoded’ in the classical behaviourist model of communication), is also actively and constructively generated in the process of constituting the subjective text meaning (cf. also Viehoff 1988). As a form of cognitive-constructive processing, then, text reception is also an interactive process, that is, an interaction between text and reader manifest for instance in the linkage between ‘bottom-up’ and ‘topdown’ processes. That is to say, there are two components which interact in text processing, namely the information provided by the text and the reader’s knowledge about the world as it existed prior to the act of reading (ibid.). This fundamental cognitive constructivity of text reception is particularly intense and has a distinct quality in the reception of literary texts. In this context, Meutsch (1987) has empirically confirmed three further-reaching types of elaboration: the elaboration of content within an alternative frame of reference, metatextual elaborations with signals specific to literature as well as cognitively polyvalent elaborations (Meutsch 1987: 158; cf. also Meutsch and Schmidt 1985). Corresponding to Schmidt’s hypothesis (1980), this is a manifestation of the structural phenomenon that literary reception processes are characterized by the so-called aesthetic and polyvalence conventions. The aesthetic convention is defined as the agreement that aesthetic objects are not to be evaluated with reference to the customary criteria of true/false, or useful/useless; related to the aesthetic convention is the agreement concerning the polyvalence convention, defined as the possibility/ necessity of (— actively or passively —) assigning different constitutions of meaning to a single text. Such polyvalent constitutions of meaning, most commonly realized as changes in frames of reference (Meutsch 1987: 390ff.), pertain to the reception-related aspects of creativity which are highly similar to for instance the fluency and flexibility of associations in creativity tests (cf. Preiser 1986: 60ff.). The basic conditions for creatively generating polyvalent text meanings in this fashion are twofold: on the one hand, it must indeed be possible to regard the polyvalence convention as shared common knowledge within our society (with reference to the Federal Republic of Germany, this was confirmed by Hintzenberg et al. 1980). On the other hand, the manifestation of the polyvalence convention

25

26

Norbert Groeben

must needs be present in the product, that is the literary text, as well, namely in the form of a polyfunctional structure of different individual text sections (Schmidt 1971, 1974; cf. also the concept of ‘textual gaps’ in the aesthetics of reception approach: Iser 1972). Hoffstaedter (1986) has found corresponding textual features relevant to ‘poetic text processing’. Her comparison of equivalences, deviations and ambiguities showed that potentially ambiguous semantic deviations lead to a creative constitution of meaning (1986: 243) — in other words, such textual features cause a “disruption of the routine of everyday language processing” (1986: 245). In accordance with Berlyne’s ‘arousal-theory’ (1974), however, the relationship between textual features and creative reception follows an inverse U-function: whenever an (individual) maximum of deviation and ambiguity is surpassed, the “text in question is considered as meaningless and not taken seriously” (Berlyne 1974: 244). These findings are in accordance with results obtained by Van Peer (1986), who also investigated textual features of parallelism and deviation on different linguistic levels (phonology, grammar, semantics) and condensed his findings into a model of ‘foregrounding’; here, too, it becomes clear that “converging feature syndromes” (Viehoff 1988) influence reception in regard to aspects such as conspicuousness, importance, cohesion and value for discussion. In this context, and within the framework of the hermeneutic aesthetics of literature, the most widely promoted hypothesis was that modern aesthetics is characterized by an increasing deviation from everyday language and communication (a deviation it is partly impossible to catch up with; cf. Friedrich 1956; Kesting 1965). This hypothesis has turned out to be useful in empirical research for instance on the communicative potential of poetry (Groeben 1970), the narrative concept of older and more recent short stories (Brewer and Ohtsuka 1988), as well as the increase in primary process contents (cf. above: Martindale 1986; cf. also Martindale and Keeley 1988). Thus, textual structures such as those just described perfectly allow for the intra-individual generation of polyvalent meaning (which was, as it seems, Schmidt’s ideal when formulating the polyvalence convention in 1980). As the empirical studies of reception processes and products show, however, this ideal is definitely not the statistical norm. Instead, the reception of literary texts appears to be characterized by what Steinmetz has termed ‘normalization’ (1974), that is, by a process of stripping the polyfunctional meaning potentials of their ambiguity in the direction of one consistent received text meaning; by the way, this consistent textual meaning is in turn highly dependent on the recipients’ knowledge of and interest in the respective text (cf. Groeben and Landwehr 1991: 175ff.). Psychoanalytical reception research as early as that by Holland (1968, 1975) has shown that in the reception of literature those dimensions of meaning related to the recipient’s personal life experience are frequently central to her or his reception. This dynamics has been confirmed outside the psychoanalytical context as well (for instance concerning the impact of drug addiction: Bichler 1981; gender-specific emotional

Literary creativity

reactions in Hansen 1986). The influence of such reader attitudes may even result in a clearly ‘distorted’ text reception, for instance in the case of text contents being clearly opposed to the personal convictions, values and so on of the recipients. To give one example: in accordance with their own — traditional — life plans, female students of German Haupt- and Realschulen1 reject those figures of identification in books for girls that do not represent traditional values while identifying with those characters representing traditional female role models, even though these are presented critically in the emancipation-oriented books themselves (Daubert 1984, 1987). This is not to say, however, that — depending on their respective personality traits — readers treat literary texts as mere stimuli for their idiosyncratic constitutions of meaning (cf. the negligible influence of dogmatism: Schram 1991). Research to date thus permits the following conclusion: on the part of the reader, the main impact on reception is exerted by factors related to achievement (verbal intelligence, reading skills, marks in the school subject ‘German’: cf. Groeben and Landwehr 1991: 177ff.), and the reception process is characterized by polyvalent structures of meaning potentials which are then individually summarized by the reader in one monovalent meaning (cf. e.g. Ibsch 1988). Considering the polyfunctional text structures outlined above, it is not only plausible but, psychologically speaking, next to inevitable that these individually received text meanings should be different for different recipients. Thus, this variant of an inter-individually polyvalent generation of meaning may be juxtaposed as a weak version to the strong version of the hypothesis of polyvalence postulating an intra-individually polyvalent generation of meaning (cf. Groeben 1982, 1983). Ibsch (1988), however, has pointed to and empirically illustrated the existence of intermediate forms such as for instance the willingness of recipients to accept such receptions/interpretations as meaningful and possible that clearly diverge from their own individual constitutions of meaning. While not implying the generation of polyvalent meanings of literary texts, this kind of acceptance would constitute a medium variant between the weak and the strong versions of the hypothesis of polyvalence; at the same time, it shows that, in the long run, this hypothesis ought to be elaborated and extended into different dimensions: besides the intra- vs. inter-individual dimension and generating vs. accepting/tolerating polyvalent meanings, the hypothesis would also comprise the dimensions of process vs. product and one vs. several points of time (cf. Groeben and Schreier 1992). Regarding literary creativity, it should be obvious that the strength of the hypothesis of polyvalence is related to the creative intensity and depth of reception (in the sense of a process of cognitive construction). If one individual in fact receives a literary text in a polyvalent manner (intra-individual polyvalent generation of meaning), this potentially constitutes an entirely creative cognitive achievement. If an individual accepts different polyvalent text meanings generated by others (inter-individual acceptance), s/he at least tolerates ambiguity; such an acceptance of ambiguity has been

27

28

Norbert Groeben

empirically confirmed as one condition, or a factor correlating with, creativity (cf. Preiser 1986: 58ff.). The reception of literary texts can thus be said to take place under conditions stimulating (at least, rudimentary forms of) creativity, both with respect to text structures and reception conventions (cf. the general survey in Cropley et al. 1988). In principle, then, literary texts represent incentives for creativity, which might be even more effective with respect to reception quality (in the direction of intraindividual polyvalence). It is primarily a task of literary didactics to increase this stimulating effect (a survey of Anglo-American research is presented in Colvin and Bruning 1989). A first approach in this direction is the concept of receptionoriented teaching; here, different reception habits, and even ‘divergent’ constitutions of meaning by the students are accepted without criticism and put up for discussion (cf. Baurmann 1980). Especially on the level of university courses, such an “encouragement of reception relevant to the self” (Beaugrande 1987, 1989) has resulted in a wide and convincing range of coherent literary receptions and interpretations (by the students: ibid.). Given such an amplitude of reception (Lämmert 1973) in group discussions, it is possible and desirable to take into account a multitude of diverse processing levels, each of which may then be categorized in manifold ways (cf. Fleischhauer 1982; Grzesik 1983; Meder 1982); these levels may range for instance from the mere reproduction of textual information via the judgement of textual modes of presentation to evaluations of interpretative syntheses. A second approach is even more strongly directed at fostering creativity by increasing the cognitive constructivity of reception up to the threshold toward production. Wermke (1989) may be regarded as a classic example from literary didactics of this kind of production-oriented study aimed at increasing creativity in dealing with literature: here, creativity-oriented goals were explicitly derived and realized on the levels of perception/cognition as well as emotion/ motivation and behavior/action. Wermke used production tasks ranging from dealing with an incomplete/manipulated text via the extension and rewriting of texts up to the metatextual level and the use of different media (cf. Spinner 1987: 190ff.); all of these overall categories contained a multitude of concrete didactic realization strategies which in turn form the nucleus of an overall didactic conception within the framework of an action- and production-oriented teaching of literature (cf. Haas 1997; Spinner 1995).

4. Conclusion and outlook Given the ever-increasing acceleration in media development, this fluent transition from the cognitive constructivity of reception to the activity of production has meanwhile become a quasi-natural consequence of these historical changes. Above

Literary creativity

all, computerization with its institutions such as the Internet offers a wealth of opportunities for dealing productively with (not exclusively, but also) literary texts. I would like to give merely two paradigmatic examples: first of all, hypertextstructure (Kuhlen 1996), that is the non-linear combination of various relatively small text parts that can be interlinked with one another in many possible, qualitatively different, ways, implies that the receivers, by their respective individual and independent navigation, in fact always create a unique text structure, that is an idiosyncratic constitution of meaning (which parallels for instance the conception Mallarmé had envisaged for his concept of “Le Livre” which was not realized, though; cf. Kesting 1965). It has to be noted, however, that the requirement of ‘polar integration’ applies here, too: in this case, that processes of merely ‘browsing’ can be replaced by a subsumption under superordinate semantic kernels and structures of meaning. This primarily combinatory productivity is then supplemented by the possibility of realizing such Internet texts, especially in the realm of literary fiction, by combining different text parts produced by various authors; in theory, such interactively realized text production can be extended into an infinite sequence (cf. Schreier 1998). The interlinkage between the classic printed text and the new media thus offers a broad framework for increasing the intensity and democracy of (literary) creativity (cf. Weisberg 1993) in our media society. Thus the wheel has come full circle: proceeding from author-psychological questions pertaining to literary production via issues concerning the product as well as the cognitive constructivity of literary reception, we have arrived at the transition to a very intensive, far-reaching constructivity in forms of productive text processing. One coherent overall principle, which generally underlies empirical research on creativity, is also conspicuous in the above overview on literary creativity: the democratization of the concept of creativity which, taking extremely creative, ‘ingenious’ individuals as a starting point then focuses on the creative potential of everyone. At the same time, I have also touched upon desirable directions for future psychological research on literary creativity: to isolate the concrete manifestations of literary creativity, both on the levels of (general) ability and (specific) skills (regarding production as well as reception processes) and, in particular, to describe and explain potential modifications of and opportunities for literary creativity resulting from the interlinkage with new media. In such an undertaking, and in accordance with the concept of literary creativity, the point of departure will first of all be the individual, that is the processes and structures of individual creativity. The weak variant of the hypothesis of polyvalence, that is, the inter-individual polyvalence of meaning constitutions, and the perspective of a medial interlinkage of collectives of authors make clear, however, that supra-individual aspects must also be included; in other words, the systems level has to be considered as well. To my mind, this problem cannot be solved by a simple systems-theoretical redefinition (cf. Csikszentmihailyi 1996; Feldmann et al. 1994), that is, by merely referring

29

30

Norbert Groeben

creativity back to success within one domain. Apart from the problem that, historically speaking, such attributions of creativity would be subjected to enormous oscillations (example: Van Gogh), this might also mean the superordination, by definition, of uncreative reception over creative production. Indisputably, however, the combination of individual and supra-individual aspects, that is the interconnection of the theory of action and systems theory, is highly desirable for the near future, especially in psychological research on literary creativity.

Notes 1. This text was translated into English by Dr. Roman Kurtz with the occasional assistance of Dr. Margrit Schreier. 2. Translator’s note: Hauptschule = grades 5–9, lower level of final degree/qualification; Realschule = grades 5–10, medium level of final degree/qualification.

References Baer, J. 1993. Creativity and Divergent Thinking: A Task Specific Approach. Hillsdale, N. J.: Erlbaum. Barron, F. 1963. Creativity and Psychological Health. Princeton, NJ: Van Nostrand. Barron, F. 1967. Creative Person and Creative Process. New York: Holt, Rinehart and Winston. Barron, F. 1968. Creativity and Personal Freedom. Princeton, NJ: Van Nostrand. Baurmann, J. 1980. Textrezeption und Schule. Stuttgart etc.: Kohlhammer. Beaugrande, R. de 1987. “Schemas for literary communication”. In Literary Discourse, L. Halász (ed.), 49–99. Berlin: De Gruyter. Beaugrande, R. de 1989. “Naive respondents and creative response”. SPIEL 8: 233–254. Benn, G. 1980. “Essays, Reden, Vorträge”. In Das Hauptwerk, G. Benn, Bd. 2. Wiesbaden/ München: Limes. Berlyne, D. E. 1974. Studies in the New Experimental Aesthetics: Steps toward an Objective Psychology of Aesthetic Appreciation. New York: Wiley. Bichler, J. 1981. “Was stellen Eigenart und Situation des Lesers mit Texten an? Eine Rezeptionsanalyse von Kurzgeschichten bei Drogenabhängigen”. In Rezeptionspragmatik, G. Köpf (ed.), 267–297. München: Fink. Brenner, C. 1976. Grundzüge der Psychoanalyse. Frankfurt/M.: Fischer. Brewer, W. F. and Ohtsuka, K. 1988. “Story structure, characterization, just world organization, and reader affect in American and Hungarian short stories”. Poetics 17: 395–415. Christmann, U. and Groeben, N. 1999. “Psychologie des Lesens”. In Handbuch Lesen, München: B. Franzmann, K. Hasemann, D. Löffler and E. Schön (eds), 145–223. München: Saur. Colvin, C. A. and Bruning, R. 1989. “Creating the conditions for creativity in reader response to literature”. In Handbook of Creativity, J. A. Glover, R. R. Ronning and C. R. Reynolds (eds), 323–339. New York/London: Plenum Press. Cropley, A. J. 1995. “Kreativität”. In Enzyklopädie der Psychologie, M. Amelang (ed.), 329–373. Themenbereich C: Theorie und Forschung. Serie VIII: Differentielle Psychologie und Persönlichkeitsforschung. Bd. 2. Verhaltens- und Leistungsunterschiede. Göttingen: Hogrefe.

Literary creativity

Cropley, A. J., McLeod, J. and Dehn, D. 1988. Begabung und Begabungsförderung: Entfaltungschancen für alle Kinder! Heidelberg: Asanger. Csikszentmihalyi, M. 1996. Creativity: Flow and the Psychology of Discovery and Intention. New York: Harper Collins. Daubert, H. 1984. “Literarisches Rollenbild und Leserolle — Zur Rezeption von Mädchenlektüre”. Informationen des Arbeitskreises für Jugendliteratur 4: 40–51. Daubert, H. 1987. “Probleme bei der Rezeption ‘emanzipatorischer’ Mädchenliteratur”. In Frauen- und Mädchenrollen in Kinder- und Schulbüchern, W. Grossmann (ed.), 93–97. Frankfurt/M.: Max-Traeger-Stiftung. East, E. M. 1938. “Insanity and genius”. Journal of Heredity 29: 275–279. Ellis, H. 1904. A Study of British Genius. Boston: Houghton Mifflin. Erb, E. 1997. “Gegenstands- und Problemkonstituierung: Subjekt-Modelle (in) der Psychologie”. In Zur Programmatik einer sozialwissenschaftlichen Psychologie, N. Groeben (ed.), 139–239. Bd. I: Metatheoretische Perspektiven. 1. Halbbd.: Gegenstandsverständnis, Menschenbilder, Methodologie und Ethik. Münster: Aschendorff. Feldman, D. H., Csikszentmihalyi, M. and Gardner, H. 1994. “A framework for the study of creativity”. In Changing the World: A Framework for the Study of Creativity, D. H. Feldman, M. Csikszentmihalyi and H. Gardner (eds), 1–45. Westport, Ct.: Praeger. Fleischhauer, P. 1982. “Eine inhaltsanalytische Taxonomie für textverstehende Handlungen”. In Interaktions- und Leistungstypen im Literaturunterricht, J. Grzesik et al. (eds), 167–437. Opladen: Westdeutscher Verlag. Freud, S. 1955. Gesammelte Werke: Bd. I–XVIII. London: Imago Publishing Company. Friedrich, H. 1956. Die Struktur der modernen Lyrik. Reinbek: Rowohlt. Goertzel, V. and Goertzel, M. G. 1965. Cradles of Eminence. London: Constable. Goodwin, F. K. and Jamison, K. R. 1990. Manic-Depressive Illness. New York, Oxford: Oxford University Press. Groeben, N. 1970. “Die Kommunikativität moderner deutscher Lyrik”. Sprache im Technischen Zeitalter 34: 83–105. Groeben, N. 1972. Literaturpsychologie. Literaturwissenschaft zwischen Hermeneutik und Empirie. Stuttgart: Kohlhammer. Groeben, N. 1981. “Zielideen einer utopisch-moralischen Psychologie”. Zeitschrift für Sozialpsychologie 12: 104–133. Groeben, N. 1982. “Empirische Literaturwissenschaft”. In Erkenntnis der Literatur. Theorien, Konzepte, Methoden der Literaturwissenschaft, D. Harth and P. Gebhardt (eds), 266–297. Stuttgart: Metzler. Groeben, N. 1983. “The function of interpretation in an empirical science of literature”. Poetics 12: 219–238. Groeben, N. 1999. “Fazit: Die metatheoretischen Merkmale einer sozialwissenschaftlichen Psychologie”. In Zur Programmatik einer sozialwissenschaftlichen Psychologie, N. Groeben (ed.), 311–404. Bd. I: Metatheoretische Perspektiven. 2. Halbbd.: Theoriehistorie, Praxisrelevanz, Interdisziplinarität, Methodenintegration. Münster: Aschendorff. Groeben, N. and Landwehr, J. 1991. “Empirische Literaturpsychologie (1980–90) und die Sozialgeschichte der Literatur: Ein problemstrukturierender Überblick”. IASL (Internationales Archiv für die Sozialgeschichte der deutschen Literatur) 16: 143–235. Groeben, N. and Schreier, M. 1992. “The hypothesis of the polyvalence convention: A systematic survey of the research development from a historical perspective”. Poetics 21: 3–32. Groeben, N. and Vorderer, P. 1986. “Empirische Literaturpsychologie”. In Psychologie der Literatur, R. Langner (ed.), 105–143. Weinheim/München: Psychologie Verlags Union.

31

32

Norbert Groeben

Groeben, N. and Vorderer, P. 1988. Leserpsychologie: Lesemotivation — Lektürewirkung. Münster: Aschendorff. Grzesik, J. 1983. “Methoden zur Deskription des Textverstehens im Unterricht”. SPIEL (Siegener Periodikum zur Internationalen Empirischen Literaturwissenschaft) 2: 73–99. Guilford, J. P. 1964. Persönlichkeit: Logik, Methode und Ergebnisse ihrer quantitativen Erforschung. Weinheim: Beltz. Haas, G. 1997. Handlungs- und produktionsorientierter Literaturunterricht. Seelze: Kallmeyer. Hansen, E. 1986. Emotional Processes: Engendered by Poetry and Prose Reading. Stockholm (Phil.Diss.). Hintzenberg, D., Schmidt, S. J. and Zobel, R. 1980. Untersuchungen zum Literaturbegriff in der Bundesrepublik Deutschland. Braunschweig, Wiesbaden: Vieweg. Hoffer, A. and Osmond, H. T. 1967. The Hallucinogens. New York: Academic Press. Hoffstaedter, P. 1986. Poetizität aus der Sicht des Lesers. Hamburg: Buske. Holland, N. 1968. The Dynamics of Literary Response. New York: Norton. Holland, N. 1975. 5 Readers Reading. New Haven, London: Yale University Press. Huxley, A. 1954. Die Pforten der Wahrnehmung: Meine Erfahrung mit Meskalin. München: Piper. Ibsch, E. 1988. “Zur literarischen Sozialisation. Beobachtungen zur Polyvalenz-Konvention”. SPIEL (Siegener Periodikum zur Internationalen Empirischen Literaturwissenschaft) 7:333–345. Iser, W. 1972. Der implizite Leser. München: Fink. Kesting, M. 1965. Vermessung des Labyrinths. Studien zur modernen Ästhetik. Frankfurt: Fischer. Kris, E. 1952. Psychoanalytic Explorations in Art. New York: International Universities Press; deutsche Übersetzung 1977: Die ästhetische Illusion. Frankfurt/M.: Suhrkamp. Kubie, L. S. 1966. Neurotische Deformationen des schöpferischen Prozesses. Reinbek: Rowohlt. Kuhlen, R. 1996. Hypertext. Ein nicht-lineares Medium zwischen Buch und Wissensbank (6. Aufl.). Berlin etc.: Springer. Lange-Eichbaum, W., Kurth, W. and Ritter, W. 1986ff. Genie, Irrsinn und Ruhm (7. Aufl.): Bd. 1–7. München: Reinhardt. Lämmert, E. 1973. “Rezeptions- und Wirkungsgeschichte der Literatur als Lehrgegenstand”. In Neue Ansichten einer künftigen Germanistik: Probleme einer Sozial- und Rezeptionsgeschichte der Literatur, Kritik der Linguistik, Literatur- und Kommunikationswissenschaft, J. Kolbe (ed.), 160–173. München: Hanser. Loewenfeld, L. 1903. Über die geniale Geistestätigkeit. Grenzfragen des Nerven- und Seelenlebens. Wiesbaden: Bergmann. Lombroso, C. 1887. Genie und Irrsinn in ihren Beziehungen zum Gesetz, zur Kritik und zur Geschichte. Leipzig: Reclam. Martindale, C. 1975. Romantic Progression: The Psychology of Literary History. New York: Wiley. Martindale, C. 1984. “Evolutionary trends in poetic style: The case of English metaphysical poetry”. Computers and the Humanities 18: 3–21. Martindale, C. 1986. “Psychologie der Literaturgeschichte”. In Psychologie der Literatur, R. Langner (ed.), 165–211. Weinheim, München: Psychologie Verlags Union. Martindale, C. 1989. “Personality, situation and creativity”. In Handbook of Creativity, J.A. Glover, R. R. Ronning and C. R. Reynolds (eds), 211–232. New York und London: Plenum Press. Martindale, C. and Keeley, A. 1988. “Historical trends on the content of twentieth-century Hungarian and American short stories”. In Psychological Approaches to the Study of Literary Narratives, C. Martindale (ed.) 42–65. Hamburg: Buske. McMullan, W. E. 1976. “Creative individuals: Paradoxical personages”. Journal of Creative Behavior 10: 265–175. Meder, N. 1982. “Operationenanalyse”. In Interaktions- und Leistungstypen im Literaturunterricht, Jürgen Grzesik et al. (eds), 439–504. Opladen: Westdeutscher Verlag.

Literary creativity

Meutsch, D. 1987. Literatur verstehen. Eine empirische Studie. Braunschweig, Wiesbaden: Vieweg. Meutsch, D. and Schmidt, S. J. 1985. “Über die Rolle von Konventionen beim Verstehen literarischer Texte”. SPIEL (Siegener Periodikum zur Internationalen Empirischen Literaturwissenschaft) 4: 381–408. Navratil, L. 1966. Schizophrenie und Dichtkunst. München: Dtv. Pine, F. R. and Holt, R. 1960. “Creativity and primary process: A study of adaptive regression”. Journal of Abnormal and Social Psychology 61: 370–379. Poincaré, H. 1913. “Mathematical creation”. In The Foundations of Science. New York: The Science Press. Preiser, S. 1986. Kreativitätsforschung (2. Aufl.). Darmstadt: Wissenschaftliche Buchgesellschaft. Prentky, R. 1989. “Creativity and psychopathology”. In Handbook of Creativity, J. A. Glover, R. R. Ronning and C. R. Reynolds (eds), 243–269. New York and London: Plenum Press. Rennert, H. 1962. Die Merkmale schizophrener Bildnerei. Jena: Fischer. Runco, M. A. 1991. Divergent Thinking. Norwood, N. J.: Ablex. Runco, M. A. 1994. “Conclusions concerning problem finding, problem solving, and creativity”. In Problem Finding, Problem Solving and Creativity, M. A. Runco (ed.), 271–290. Noorwood, N. J.: Ablex. Russ, S. W. 1993. Affect and Creativity: The Role of Affect and Play in the Creative Process. Hillsdale, N. J.: Erlbaum. Schmidt, S. J. 1971. Ästhetizität. München: Bayerischer Schulbuch-Verlag. Schmidt, S. J. 1974. Elemente einer Textpoetik. Theorie und Anwendung. München: Bayerischer Schulbuch-Verlag. Schmidt, S. J. 1980. Grundriß der Empirischen Literaturwissenschaft. Bd. 1: Der gesellschaftliche Handlungsbereich Literatur. Braunschweig, Wiesbaden: Vieweg; Wiederauflage 1991: Frankfurt/M.: Suhrkamp. Schmidt, S. J. and Zobel, R. 1983. Empirische Untersuchungen zu Persönlichkeitsvariablen von Literaturproduzenten. Braunschweig, Wiesbaden: Vieweg. Schram, D. H. 1991. Norm und Normbrechung. Braunschweig: Vieweg. Schreier, M. 1998. Triumph or elegy? Literature on the Internet. http://www.lumis.uni-siegen.de/ igel/en/frames.htm. Spinner, K. 1987. “Rezeptionshandlungen/Produktionsaufgaben”. In Zur Psychologie des Literaturunterrichts. Schülerfähigkeiten — Unterrichtsmethoden — Beispiele. H. Willenberg et al. (eds), 188–203. Frankfurt/M.: Diesterweg. Spinner, K. (ed.). 1995. Imaginative und emotionale Lernprozesse im Deutschunterricht. Frankfurt: Lang. Spoerri, T. 1964. Sprachphänomene und Psychose. Basel etc.: Karger. Steinmetz, H. 1974. “Rezeption und Interpretation — Versuch einer Abgrenzung”. Amsterdamer Beiträge zur neueren Germanistik 3: 37–81. Tart, C. T. 1969. Altered States of Consciousness: A Book of Readings. New York: Wiley. Ulmann, G. 1968. Kreativität: Neue amerikanische Ansätze zur Erweiterung des Intelligenzkonzepts. Weinheim: Beltz. Van Peer, W. 1986. Stylistics and Psychology: Investigations of Foregrounding. London: Croom Helm. Viehoff, R. 1988. “Literarisches Verstehen. Neuere Ansätze und Ergebnisse empirischer Forschung”. IASL (Internationales Archiv für Sozialgeschichte der deutschen Literatur) 13: 1–39. Wakefield, J. F. 1994. “Problem finding and empathy in art”. In Problem Finding, Problem Solving and Creativity, M. A. Runco (ed.), 99–115. Noorwood, N. J.: Ablex. Weisberg, R. W. 1993. Kreativität und Begabung: Was wir mit Mozart, Einstein und Picasso gemeinsam haben. Heidelberg: Spektrum der Wissenschaft.

33



34

Norbert Groeben

Wellek, R. and Warren, A. 1963. Theorie der Literatur. Frankfurt/M.: Ullstein. Wermke, J. 1989. “Hab a Talent, sei a Genie!” Kreativität als paradoxe Aufgabe. 2 Bde. Weinheim: Deutscher Studien Verlag.



Qualitative methods in studying text reception Margrit Schreier University of Cologne

It is now approximately twenty years that Groeben (1972, 1977/80) and Schmidt (1975, 1980, 1982) presented their proposals for a specifically empirical study of literature (ESL). Since then, researchers have been putting these proposals into practice. As a result, a new discipline has emerged which in some respects may even be said to stand up to the original claim of constituting a new paradigm (Groeben 1994a). At the same time, however, there have been voices of scepticism or even discontent from within ESL, voices focusing on the question of fit between the subject matter of ESL and the methods used to study it. Brewer (1995), for instance, has criticized that the methods used in ESL for data collection have focused on the cognitive-instrumental aspects of reception to the exclusion of, for instance, affective or aesthetic ones. Groeben (1994b) has stressed the necessity of adapting methods so as to suit current changes and extensions of the subject matter of ESL, such as the inclusion of other media; and Ibsch has repeatedly expressed concern over the neglect of complex issues, such as the reception of longer texts or text reception by experts (1994, 1998). Methodological rigor and control are indeed hard to maintain where the reception of any work of literature going beyond the length of a short story is concerned. Questionnaires with multiple choice items would hardly do, nor would it seem feasible to present, for instance, War and Peace in segments on a computer screen and to collect think aloud protocols. Of course this is not to say that the reception of longer works of literature cannot be studied empirically — but doing so may well require methods other than those which have been prevalent in ESL so far. In this context, there have been suggestions that ESL make increasing use of the so-called qualitative methods as they have been developed in the social sciences (Andringa 1998; Groeben 1994b; Ibsch 1998). This, however, is by no means a straightforward and simple solution. Methodological criteria, such as objectivity and reliability, which lie at the center of quantitative research, are not endorsed in the same manner by the qualitative

36

Margrit Schreier

paradigm: some researchers reject them altogether, others have sought to modify them or to find replacements. For this reason, the use of qualitative methods has been notoriously problematic in the social sciences. For a long time, social scientists have therefore been faced with an either-or: either to use quantitative or to use qualitative methods, and Ibsch has echoed this dilemma in wondering whether ESL will have to choose between methodological rigor and a restricted number of research questions on the one hand — or a wide range of research questions combined with a ‘weak’ methodology on the other (1998). Recently, however, the situation in the social sciences has begun to change: the ‘either-or’ is increasingly being replaced by an ‘as well as’. This ‘as well as’ can take various forms: it may consist in a systematization of qualitative methods, making them more rigorous (cf. Groeben 1994b, 1995), or in a modification of quantitative methods, increasing their flexibility with respect to the research question, or it can be realized by combining qualitative and quantitative procedures (in practice, these forms will frequently overlap). I am going to argue here that there is no need for ESL to re-enact the either-or; instead, much may be gained by taking a giant step forward and begin to realize the as-well-as. In this chapter, my focus will be on how this may be brought about. I am going to set out by giving a brief characterization of qualitative and quantitative methods (section 1). I will then describe how the use of qualitative methods can be made more systematic (section 2) and how qualitative and quantitative methods may be combined (section 3). To conclude, I will make two suggestions for potential research designs in ESL which combine research on complex issues with a systematic methodology, thus constituting an application of the systematizations and combinations described in the previous sections (section 4). While I will concentrate on the issue of text reception, the following considerations and suggestions equally apply to research concerning the other roles within the literary action system, such as the production or the post-processing of literature.

1.

Quantitative and qualitative methods in the social sciences

To speak of qualitative and quantitative methods is, in a sense, already a step beyond the above mentioned dichotomy. For, of course, the dichotomy is not merely a dichotomy of methods, but one of research traditions and associated philosophies of science. In order to allow for a brief characterization of these two approaches, they will — for the moment — be presented as extremes; in practice, however, the two approaches can frequently be seen to overlap.

Qualitative methods in studying text reception

1.1 The quantitative approach At the heart of the quantitative approach lies the notion of hypothesis testing: a hypothesis is derived from a theory, indicators of the concepts contained in the hypothesis are selected, data on the indicators are collected. The results are then compared to the hypothesis, and the hypothesis is either rejected or accepted for the time being, pending further empirical testing. Hypotheses may be of various kinds, but it is hypotheses about causal relationships which are of greatest interest to the quantitative researcher. The aim of hypothesis testing is to arrive at causal explanations which apply not only to a few individuals, but to larger groups of persons. The data are usually numerical, and hypothesis testing is most often achieved under recourse to inference statistical procedures; these allow a decision as to whether the results obtained are merely due to chance or whether they can be attributed to systematic differences in the data (Bortz and Döring 1995: 274ff.; for a brief overview cf. Huber 1987). There exist numerous examples of such ‘quantitative’ studies in ESL. One almost ‘classic’ early example is a study by Geiger (1975) in which he was able to confirm the hypothesis that the reading of Landser-Hefte leads to an increase of nationalist prejudice and aggressiveness. ‘Landser-Hefte’ are a special type of German pulp fiction dealing with World War II. The procedures used in quantitative hypothesis testing must meet a number of criteria (Bortz and Döring 1995: 185ff.). A first set of criteria relates to the methods and instruments used in data collection and analysis. First of all, instruments must be objective, that is: the results obtained must be independent of the person employing the instrument. In addition, the instrument must be reliable: if it is repeatedly applied, it must yield consistent, stable results. Finally, the instrument must be valid; this is to say that it must indeed measure what it is designed to measure. Objectivity is considered a precondition to reliability, and reliability in turn a precondition to validity. Methods for data collection which meet these criteria include, for instance, rating scales, questionnaires with forced-choice items, tests, many physiological measures, and unobtrusive measures; as for data analysis, statistics provide a large number of objective and reliable methods. The second set of criteria concerns the research design as a whole. Internal validity, the first criterion, relates to the question of experimental control (Huber 1987: Ch. 4; Watt and Van den Berg 1995: Ch. 4). A study is said to be internally valid to the extent that the researcher succeeds at isolating the effects of a number of independent variables on a number of dependent variables by controlling for potentially confounding variables. If she is successful, changes in the dependent variables can then be attributed to changes in the independent variables; as a consequence, changes in the independent variables may be said to have caused the effects on the dependent variables. Strict control is only possible in an experimental setting, and therefore the testing of hypotheses on causal relationships between

37

38

Margrit Schreier

variables is largely restricted to this type of design. The second criterion in this context, external validity, concerns the question to what extent the results obtained are in fact representative of everyday reality, that is, to what extent they can be generalized to everyday situations (Bortz and Döring 1995: 472ff.). The most common type of generalization is that from sample to population; but generalization may also concern other settings or other indicators of the relevant concepts. Traditionally, a kind of ‘trade-off’ may be said to exist between internal and external validity: to the extent that internal validity is increased, control over the empirical setting is tighter; the empirical setting is thus restricted by comparison to everyday situations, which of necessity lowers the external validity of the respective study. Conversely, high external validity usually entails a lower degree of control over the setting and thus less internal validity. 1.2 The qualitative approach The qualitative paradigm, having developed out of a number of diverse research traditions, is much less unified and thus harder to characterize than the quantitative approach (cf. Tesch 1990; Vidich and Lyman 1994). Yet there exist a number of features most qualitative traditions have in common (Bortz and Döring 1995:274ff.; Lamnek 1995). One of these is the interest in description and understanding rather than in explanation. Another concerns the interest in the individual case: while the researcher within the quantitative paradigm is interested in studying a large number of cases with respect to a small number of attributes, researchers in the qualitative tradition tend to study a small number of cases with respect to a large number of attributes; research in the qualitative tradition is thus characterized by its depth rather than its breadth. Also, qualitative research is usually not concerned with the testing of hypotheses previously derived from theory. Instead, hypotheses concerning characteristics of the cases in question and potential patterns in the material emerge as research procedes, informing further data collection and analysis which may in turn lead to a further modification and development of hypotheses. The structure of qualitative research may thus be said to be iterative and inductive. Data collection stops when the description of the phenomenon of interest is sufficiently detailed to be able to accommodate additional data without requiring further modification. The data thus collected are usually verbal. Most qualitative researchers attempt to study phenomena of interest in their natural environment and to keep the influence of the methods of data collection on these phenomena as low as possible. Purely qualitative studies are also found in ESL. An example would be Wieler’s analysis of parent-child-interactions in episodes of joint picture book reading; Wieler was able to identify different patterns of interaction depending on the mothers’ level of education (Wieler 1997). As it is to be expected considering the diverse roots of qualitative research, there

Qualitative methods in studying text reception

exists a large number of methods which may be termed ‘qualitative’ (for an overview cf. Denzin and Lincoln 1994: Parts III, IV, V). First of all, there exist designs such as the case study or the field study (the study by Wieler mentioned above constitutes a case study); within the framework provided by these designs, many qualitative methods can be used for data collection and analysis (Mayring 1996: Ch. 3). Other types of designs, such as grounded theory or analytic induction, constitute a much stricter framework regulating the way research takes place (cf. the description in Fielding and Lee 1998: Ch. 2). Apart from such designs, methods for data collection and for data analysis can be distinguished. Methods for data collection include, for instance, various types of interviews, group discussions, paraphrase, media diaries, think aloud protocols, and so on (in general: Mayring 1996: Ch. 4.1; in ESL: Groeben 1977/80). Methods for data analysis are mostly interpretive: they include phenomenology, hermeneutics, qualitative content analysis, conversation and discourse analysis, among others (Tesch 1990: 77–145). There exists no generally accepted set of criteria in qualitative research paralleling those in the quantitative tradition (for suggestions cf. Lamnek 1995: Ch. 2.2); but there does seem to be a general consensus to the effect that the ‘quantitative criteria’ cannot simply be transferred to qualitative research in an unmodified form (cf. Bortz and Döring 1995: 301ff.; Flick 1995a: Ch. 18). As far as internal and external validity are concerned, the situation within the qualitative paradigm is the exact reverse of that within the quantitative approach. While the quantitative approach is aimed at increasing internal validity, even if this occurs at the cost of external validity, the qualitative approach strives for external validity, even if this entails a loss of internal validity. In fact, internal validity cannot be of great concern to the qualitative tradition: the researcher who wishes to study a phenomenon in its natural environment is by definition not concerned with exercising control — to control the setting would be to destroy the phenomenon, in qualitative terms. At the same time, by attempting to preserve the natural setting, great stress is laid on external validity. To the extent that only a limited number of cases are studied, however, external validity in the sense of generalizing the findings beyond the sample studied is usually not an issue in qualitative research. Among the criteria relating to methods, objectivity and reliability are usually not acceptable to the qualitative tradition in their original, ‘quantitative’ form. Social phenomena are context-bound, and a researcher studying a specific phenomenon necessarily becomes part of its context. An interview, for instance, is always a social situation, and this situation is different depending on the person conducting the interview. For this reason — so the reasoning in the qualitative tradition goes — methods should not aim for objectivity in the sense of being independent of the person employing them and the context of their application. Similar arguments are advanced with respect to reliability. This is not to say, however, that these criteria are rejected altogether in qualitative research; rather, attempts have been made to

39

40

Margrit Schreier

modify or to replace these two quantitative criteria with criteria better suited to the aims of qualitative research (see section 2.1). Validity, on the other hand, may be said to constitute the ‘meeting point’ between the two traditions: As a matter of fact, this criterion may be said to lie at the center of qualitative research. It informs the qualitative methodological demands of staying close to the phenomenon of interest, studying it in its natural environment, and not disrupting and thus changing it in the course of data collection. 1.3 The pros and cons of using qualitative methods in ESL Needless to say, it is precisely this ambiguity concerning the quantitative methodological criteria which makes qualitative methods unacceptable from the quantitative point of view; conversely, it is the methodological rigor maintained in quantitative research, regardless of the issue under study, which makes qualitative researchers reject quantitative procedures. Yet qualitative methods are indispensable, especially where the social sciences and studies of cultural phenomena are concerned. Often, literary reception is likely to be too complex and too idiosyncratic to permit the use of questionnaires containing items formulated in the researcher’s terms. Instead, the study of literary reception requires the use of methods which permit recipients to describe relevant aspects of their experience in their own words, perhaps combined with respondent validation in order to make certain that the researcher has in fact understood what the recipient meant to say. Research on reception, in other words, requires the use of qualitative methods. At the same time, however, research on reception — as part of ESL — is also rooted within an essentially falsificationist empirical research tradition requiring procedures from within the quantitative tradition. Thus, if qualitative methods are to become a part of ESL, this cannot mean the acceptance of the entire qualitative paradigm, nor can it mean the rejection of the empirical methodological criteria. The question whether qualitative methods and ESL are compatible therefore hinges on two issues. The first issue concerns the degree to which qualitative and quantitative methods are regarded as being rooted within their respective research tradition; the second issue raises the question to what extent it is possible to combine the flexibility of qualitative methods with the methodological rigor of the quantitative approach. Regarding the first issue, two (opposite) points of view have been put forward (cf. Bryman 1988). According to the so-called epistemological position, methods are inextricably tied to their respective paradigms; according to the technical position, methods are regarded as techniques which can be applied in whatever context they appear to be useful. In fact the technical position can be said to have been validated already by research practice (in ESL cf. for instance the coding of responses to open questions followed by the application of statistical procedures:

Qualitative methods in studying text reception

Kuiken and Miall 1995; Thury and Friedlander 1996). The resulting studies may not meet with the wholehearted approval of strict adherents to either the qualitative or the quantitative research tradition, but a use of the respective methods outside the context of their traditional roots — such as the use of qualitative methods within a falsificationist framework — is definitely possible. Whether such a use of qualitative methods is compatible with the quantitative methodological criteria will be discussed in the following section.

2.

Increasing the methodological rigor of qualitative methods

While researchers in the qualitative tradition reject the evaluation of their methods and results according to the ‘quantitative’ criteria, this is not to say that qualitative researchers reject methodological evaluation in general. Rather, proponents of the qualitative paradigm have begun to develop alternative methodological notions and standards which they consider to be better suited to the qualitative background (cf. Flick 1995a: Ch. 18). In this way, the application of qualitative methods is made more rigorous and thus more acceptable to the quantitative research tradition. At the same time, methodological suggestions from within the qualitative tradition may well be used within the quantitative paradigm in order to counteract some of its weaknesses (such as the emphasis on selected attributes rather than on complex phenomena), thus resulting in a mutual adaptation of methods. In this section, two such approaches (as they were suggested by Groeben 1995 for ESL) will be presented: systematization and intersubjectivity in applying qualitative methods. 2.1 Systematic use of qualitative methods in data collection Researchers within the qualitative paradigm have put forward several criteria which are intended to further a systematic application of the respective procedures (cf. Mayring 1996: 119ff.): the first of these concerns the detailed documentation of the various steps of the research process. The second criterion demands that the steps which are to be followed in applying a procedure be made explicit and that these steps are in fact followed. A third criterion, triangulation, suggests that the chosen way to conduct the study be supplemented, for instance, by adding another type of data or applying different methods (Denzin 1970); in the case of data collection, the application of different methods will usually also result in different types of data, thus realizing two types of triangulation at the same time (i.e. triangulation of methods and data). Triangulation of methods in particular has become quite popular in qualitative research in the social sciences (for examples in ESL cf. Flick 1995b; Malmgren 1996; Wieler 1997), while at the same time being discussed in a highly controversial manner. In the first place, the assumption underlying the

41

42

Margrit Schreier

concept has been called into question. The basic idea is to counteract any biases inherent in the application of one method by contrasting it with the application of another. Fielding and Fielding (1986) have, however, pointed out that in combining two equally unsuitable methods, the bias inherent in each may not be eliminated, but increased instead. In addition, little has been said about what it means if the results obtained by applying two different methods do not coincide (Fielding and Fielding 1986; Groeben 1992). While triangulation must thus count as controversial, the other two criteria have been increasingly and successfully applied in qualitative studies. As far as data collection is concerned, application of qualitative methods such as the interview can be made more systematic by specifying at greater detail in advance what exactly is to be studied, that is, by adding (conceptual) structure to the research instrument. In the case of the interview, this would, for instance, mean to conduct interviews which are not entirely open und unstructured, thus leaving the respondent free to talk about anything that comes to mind. Instead, a semi-structured type of interview would be chosen, as it has, for example, been developed in the context of the “Research Programm‚ Subjective Theories” (cf. Groeben et al. 1988; Scheele 1992). This type of interview is characterized by three kinds of questions: the first type of question is entirely open, thus allowing the respondent to put forward her or his individual points of view. The second kind of question draws upon previous theories and research, asking the respondent for an opinion about views as they are prevalent in the respective discipline; respondents must be free to accept those views or to reject them, depending on their ‘fit’ with the respondents’ subjective theories. The last type of question is meant to provoke the respondents into questioning their opinions by confronting them with alternative points of view. These three types of questions are generated in advance on the basis of what is already known about the phenomenon under study and brought into a sequence. The resulting interviews combine a high degree of openness and flexibility by allowing respondents to put forward their own views in their own words as a reply to the open questions with a high degree of systematicity both by drawing upon already existing research and in conducting the interview according to a specific sequence. In ESL, this method has been used by Scheele and Groeben (1986/87) in order to assess how the reading of literary utopias affected and changed recipients’ subjective theories about ‘utopia’ (for another example cf. Viehoff 1995). Semi-structured interviews have, of course, a long tradition in qualitative research outside the framework of ‘subjective theories’ (in ESL cf. for example Cupchik and Izadpanah 1998; Hron 1994; Lamnek 1995: Ch. 3; Schön 1996; Wieler 1997). The procedure developed in order to collect and reconstruct subjective theories is, however, ideally suited to reception research in ESL (cf. Scheele and Schreier 1994). First of all, literary reception, in particular reception by experts, is highly complex: it does not merely comprise concepts; instead, these concepts will

Qualitative methods in studying text reception

usually be highly idiosyncratic, and they will not be unconnected, but in some way linked to each other. While traditional semi-structured interviews are well suited for systematically collecting data on individual concepts, procedures developed within the framework of subjective theories go several steps further by explicitly taking into account the complexity of such ‘networks’ of concepts. First, the procedure is even more systematic than semi-structured interviews in general by distinguishing between the collection of content (concepts) on the one hand, and the reconstruction of the conceptual structure between these concepts on the other hand. This is achieved by adding a further step of data collection: following the interview, the respondents are presented with their concepts as well as an explication of various relationships as they might obtain between those concepts (such as: definition, example of, empirical relations, etc.) and asked to construct a network. Second, the procedure constitutes an increase in validity as compared to the traditional interview: usually, it is simply taken for granted that the researcher has in fact understood what the respondent meant to say. Especially where complex content is concerned, however, this need not necessarily be the case. In order to ensure an adequate reception of the respondent’s answers by the researcher, the collection of subjective theories comprises as a final step the respondent validation of those subjective theories by means of a consensus-in-dialogue: both the researcher and the respondent attempt the network construction of the respondent’s subjective theory; they then integrate both attempts into a final reconstruction. This is considered valid only to the extent that the respondent agrees that the network reconstruction is indeed an adequate representation of her or his subjective theory. The collection and reconstruction of subjective theories can thus count as an example of the successful systematization of the interview method which is especially well suited for complex areas such as literary reception. 2.2 Systematic use of qualitative methods in data analysis Qualitative procedures for data analysis are in even greater need for systematization than procedures for data collection. For a long time, it has been usual in qualitative research to present only results, not, however, the procedures by which these results were obtained (Fielding and Lee 1998: 2; Miles and Huberman 1994: 2). A further obstacle to systematization can be seen in the great variety of qualitative methods for data analysis: every single ‘school’ of qualitative research seems to have developed its own procedure; to some extent these methods overlap, while in other respects they are quite diverse (for an overview cf. Tesch 1990: 77–135). Nevertheless, at least two approaches can be distinguished which are aimed at making the application of these methods more systematic: the conceptualization of displays for qualitative data and the development of software for qualitative data analysis.

43

44

Margrit Schreier

The development of displays for qualitative data has been initiated and advanced by Miles and Huberman in particular (1994). They conceptualize displays as visual formats which provide a systematic presentation of relevant information; visual formats thus give a view of an entire data set that is arranged with respect to the research question (Miles and Huberman 1994: 91f.). Besides providing an overview of the data, displays are also suitable tools for data exploration. Miles and Huberman suggest two types of displays: matrices and networks. Matrices are essentially constructed on the basis of data listings, thus crossing variables of interest and allowing the researcher to inspect the data for patterns, regularities, and differences. Matrices may be ordered or not; if they are ordered, a number of criteria are possible, such as ordering according to degree, to time, outcomes, effect, or other types of concepts; if multiple cases are examined, ordering according to cases is also a useful alternative. They are especially flexible as to the type of data included which may range from raw data (such as quotes) or variable names to numerical data (such as frequencies or percentages). Networks are defined as a collection of nodes (concepts) and links (relationships between concepts). They are regarded as especially useful where several variables are concerned and where the research focus is on aspects of process. Networks include event-state-networks, cognitive maps, and various types of causal networks. Both display types can be used with a more descriptive, exploratory, or with a more explanatory focus. Both types can likewise be applied to within- as well as to cross-case, that is, multiple-case analysis. To the extent that cross-case analysis comprises an aggregation across individual data, the respective display types may be regarded as occupying a middle position between strictly speaking qualitative and quantitative analysis (cf. 3. below). Altogether, the use of data displays as advocated by Miles and Huberman may well contribute towards increasing the transparency of qualitative data analysis. To my knowledge, this approach has not yet been used in ESL. As a second way of making the application of qualitative methods of data analysis more systematic, the use of software packages especially designed for the computer-assisted analysis of qualitative data has been suggested (cf. Fielding and Lee 1998; Kelle and Laurie 1995). These software products provide facilities for analyzing extensive verbal data as they are, for instance, collected in interviews (for an overview of facilities cf. Kelle 1995; Richards and Richards 1994; Tesch 1990; Weitzman and Miles 1995). All such packages include so-called preparatory functions, such as providing facilities for data import or entry, numbering of lines, data segmentation etc. They further comprise the two main functions of coding and retrieval, where coding refers to the attaching of codes to text segments (that is, subsuming text segments under a certain category); usually several codes can be attached to one segment simultaneously. Retrieval means the search for text segments according to codes; results include the text segments themselves as well as information about the origin of those segments. Most programs developed so far go

Qualitative methods in studying text reception

beyond basic coding and retrieval in providing additional enhancements. At least two types of enhancements can be distinguished: programs which aid network building, and programs which provide facilities for hypothesis testing (cf. in detail section 3 below). Programs for the computer-assisted analysis of qualitative data may thus be regarded as an extension of the systematization by means of data displays according to Miles and Hubermann. They go beyond Miles and Hubermann in making the process of constructing descriptive coding schedules as well as the coding of data (which may be regarded as an essential element of qualitative data analysis: cf. Tesch 1990) more transparent; and they aid in the construction of data displays by facilitating the retrieval of relevant data, sometimes even providing (enhancement) functions for the construction of the displays themselves (for suggestions for the use of such programs in ESL cf. Schreier 1997a). 2.3 Intersubjectivity While systematization is an important step in making the application of qualitative methods more rigorous, by itself it is not enough: a researcher may be perfectly systematic in the analysis of her data and make the process altogether transparent — and yet arrive at entirely idiosyncratic and invalid conclusions. This is where the second criterion for modifying methods in qualitative research becomes relevant, that is intersubjectivity. Essentially, intersubjectivity is aimed at counteracting idiosyncracy by testing whether at least two researchers applying the same method according to the same steps to the same data arrive at the same conclusion; if this is the case, the method can count as — approximately — objective and the results as comparatively stable. Intersubjectivity can thus be said to rest upon the systematic application of the methods involved; at the same time, it entails a suggestion for modifying the quantitative methodological criteria of objectivity and reliability in order to increase the flexibility of quantitative methods. Usually, intersubjectivity applies to methods of data analysis. As with systematization, various approaches can be distinguished which range from an informal exchange of points of view between researchers to the computation of numeric coefficients of interrater agreement. ‘Informal’ approaches may be said to include all procedures which go beyond the individual researcher working in isolation without, however, quantifying the degree to which different researchers agree. Such an informal procedure may, for instance, consist in nothing more than one researcher explaining her reasoning to another. More common are procedures such as case analysis meetings (Miles and Huberman 1994: 76f.) where a group of researchers discuss a particular case from a number of perspectives. This approach was for instance taken by Malmgren in his analysis of the reading of the novel A Sign from the Beaver in the classroom (1996): different researchers began by analyzing relevant documents (transcripts of lessons, teachers’ diaries, etc.)

45

46

Margrit Schreier

individually and then continued by comparing the outcomes of these analyses with the aim of achieving a joint interpretation. These informal approaches are, however, insufficient to the extent that they allow for the reconstruction, but not for the (empirical) replication of the results (Nachvollziehbarkeit vs. Nachprüfbarkeit
: cf. Groeben 1994b); the latter would only be possible on the basis of the quantification of the agreement between researchers — as it is realized in content analysis (Merten 1983; Rustemeyer 1992). Essentially, the method comprises the a priori construction of a coding scheme. Text segments are then categorized according to this scheme by at least two independent raters, and the agreement between the raters is computed. The reliability of the coding scheme depends on the size of the coefficient; if this is not sufficient, the coding scheme is to be revised (for the evaluation of the coefficient kappa cf. Landis and Koch 1977). If it is for some reason impossible to compute a coefficient of interrater-agreement, a coefficient of intra rater-agreement may be used instead (cf. Groeben and Rustemeyer 1995). In this case, the same rater codes the material twice, and the agreement between the two coding sessions is assessed (that is, the stability of the coding). While the use of such coefficients has so far (to my knowledge) been restricted to content analysis, there is no reason why the principle of the quantification of intersubjectivity underlying them should not be transferred to other types of qualitative data analysis as well. Subsequent classification of free responses by means of content analysis has frequently been used in ESL: in her analysis of the effect of contextual knowledge on the reception of foreign language literary texts, Andringa (1996) distinguished between simple and complex free responses; Cupchik and Izadpanah (1998) used content analysis in order to identify subjects’ reasons for selecting a particular image from among a larger number of images (other examples: Janssen 1996; Oatley 1999). Taken together, the systematic application of qualitative methods combined with the computation of coefficients of interrater agreement takes qualitative methods a long way from their sometimes rather haphazard application within the qualitative research tradition. At the same time, systematization and intersubjectivity may also be regarded as a step towards the modification of the quantitative methodological criteria of objectivity and reliability; they increase the flexibility of quantitative methods, thus furthering a mutual adaptation of methods from the two paradigms.

3.

Combining qualitative and quantitative methods

3.1 Types of combinations Again it must be said that approaching objectivity and reliability in using qualitative methods systematically and with intersubjective agreement constitutes an important precondition to the use of such methods in ESL, but is itself not sufficient. As

Qualitative methods in studying text reception

I pointed out above, ESL is rooted in a falsificationist research tradition: ESL usually aims at going beyond the individual case, at hypothesis testing, and so on; its goals are thus different from those of the qualitative research traditions. If qualitative methods are to be used within ESL, it must, as a consequence, be possible to use qualitative methods towards such ‘quantitative’ aims — which essentially amounts to combining elements from the qualitative and the quantitative research tradition. To the extent that this is possible, such combinations are likely to also include cases where quantitative methods are used towards the traditional aims of qualitative research. This, in turn, would constitute yet another step in the direction of the mutual adaptation of the qualitative and quantitative approaches. There have been a number of suggestions within the social sciences as to how such combinations may be achieved. Miles and Huberman (1994: 40ff.) distinguish between three types of combination: linking data types, multi method designs, and quantification of qualitative data. Linking data types and multi method designs both rest upon the assumption that qualitative and quantitative approaches may facilitate each other (Fielding and Fielding 1986; Sieber 1979). If quantitative is preceded by qualitative research, qualitative research may, for instance, be used as a heuristic for designing research instruments, putting forward hypotheses, etc. If quantitative research is conducted first, a subsequent phase of qualitative research may serve to illustrate both typical cases and ‘outliers’ in the data material, thus providing a more complex and complete picture than could be achieved by applying one type of method only (cf. for instance Hurrelmann et al. 1993). In the third type of combination, quantification of qualitative data, qualitative methods are used for data collection; data analysis then proceeds according to the quantitative tradition (cf. below). The highest degree of integration, however, is achieved in yet another way: in constructing and applying ‘hybrids’, that is, methods which in themselves combine qualitative and quantitative elements (Bryman 1988: 151f.). In what follows, the use of such hybrids and the quantification of qualitative data will be explained in greater detail. 3.2 Hybrids: Combining qualitative and quantitative elements

within one method One example of a hybrid was already mentioned above: content analysis. Considering its history and development, content analysis has from the beginning been claimed by both research paradigms. The conceptualization of a ‘quantitative content analysis’ goes back to Berelson who defined the method as “a research technique for the objective, systematic, and quantitative description of the manifest content of communication” (1952: 18). By contrast, ‘qualitative content analysis’ (cf. Lamnek 1995: Ch. 5; Mayring 1995) focuses on aspects of latent meaning; at the same time, the technique is deeply rooted within the qualitative research tradition

47

48

Margrit Schreier

insofar as categories emerge inductively as a result of the research process. Instead of pursuing this dichotomy any further, a ‘hybrid’ type of content analysis has been suggested which combines criteria from the ‘quantitative’ tradition, such as systematic application and intersubjectivity, with criteria from the ‘qualitative’ orientation, such as applicability to latent dimensions of meaning and flexibility with respect to research questions and the types of texts studied (cf. Groeben 1987; Groeben and Rustemeyer 1995; Rustemeyer 1992). Following the explication of hypotheses, a part of the textual material to be analyzed is set aside for a first pilot phase. The core of the method consists in the development of categories serving as a coding scheme for the textual analysis. In this, the categories function as coding rules: they contain exact specifications of the conditions under which a text segment is to be assigned to a particular category; usually, positive and negative examples are added as well. After the texts of the pilot phase have been divided into segments, at least two raters code the segments. Next, the interrater agreement between the coders is calculated; if it is unacceptably low, the categories are modified and the coders trained in the application of the coding scheme. Once the interrater agreement is sufficient, the remaining texts are treated in the same manner. This method is extremely flexible to the extent that categories may refer both to manifest and to latent aspects of meaning; also, the level of abstraction can be freely chosen, and units of analysis may range from entire books to single words or even morphemes. Furthermore, the method is systematic, and the calculation of interrater agreement ensures intersubjectivity. While content analysis has a tradition in communication studies for being used in analyzing the meaning of texts, it lends itself equally well towards the analysis of text reception. In this, however, reception must not be inferred from the properties of the text, but reception data must first be collected, for instance by means of semi-structured interviews, and then be subjected to content analysis. An additional advantage of the method lies in its suitability for subsequent quantitative analysis of coding frequencies (see below; for examples in ESL studying text reception, cf. section 2.3). Other ‘hybrids’ have been developed in the context of the computer-assisted analysis of qualitative data (cf. section 2). The most important among these are methods for qualitative hypothesis testing. According to Huber (1995), two types of such methods can be distinguished: linkage analysis and configuration analysis. These basically correspond to the two types of data displays suggested by Miles and Huberman (1994): linkage analysis rests upon matrices, whereas configuration analysis is advanced by network construction. The construction of data displays can thus go beyond systematization by functioning as an important step towards further analysis. Linkage analysis involves the search for co-occurrences between codes. In studying the reception of a particular text one may, for instance, have the hypothesis that persons who mention theme A are likely to also mention theme C, whereas

Qualitative methods in studying text reception

persons who mention theme B hardly ever mention theme C. In order to test hypotheses of this kind, logic programming has been used (Hesse-Biber and Dupuis 1995; Sibert and Shelly 1995). Such systems require the following components: a way of entering data (that is, text), a knowledge base containing information about these data (such as the denomination of the categories to which text segments have been assigned), and a way of putting queries about that information to the data base (for instance by use of Boolean algebra). Appropriate queries can then yield data as to how many respondents mention theme A and C, how many respondents mention theme B and C, the number of respondents who mention theme A without mentioning theme C, and the number of respondents who mention theme B without mentioning theme C. Configuration analysis, the second type of method used for hypothesis testing in qualitative research, is aimed at the causal analysis of qualitative data. In QCA (qualitative comparative analysis: Ragin 1987, 1994), for instance, the Boolean concept of logical minimisation is used in order to identify the configuration of conditions sufficient for bringing about a certain effect. The first step of data analysis consists in identifying outcome variables and specifying potential causal conditions. Next, an analysis of implicants is carried out: this involves examining different configurations of the causal variables (implicants) as they appear in the data according to whether they agree on the outcome. Once a satisfactory set of implicants has been identified, implicants and outcomes are displayed in the form of a truth table which serves as the basis for subsequent logical minimisation. The goal of minimisation is to represent the information contained in such a truth table concerning the various combinations of implicants and outcomes in a shorthand manner. To this purpose, the researcher attempts to identify those rows in the truth table which differ according to one condition only, yet produce identical outcomes. In this manner, causal conditions are compressed until the so-called prime implicants emerge, that is those causes sufficient in bringing about a specific outcome. It will have emerged from the description of these two methods that hypothesis testing, as it is used in this context, is not identical with hypothesis testing in the quantitative paradigm. The qualitative methods produce certain outcomes (such as numbers of co-occurrences or listings of prime implicants), but do not provide any means for comparing those outcomes to a priori hypotheses; in particular, rules for determining the significance of results are lacking. There is no reason, however, why the procedures described should not be followed by an inference statistical analysis and thus be suitable for hypothesis testing; conversely, in the context of an exploratory study aimed at hypothesis generation, linkage and configuration analysis may equally be applied to quantitative data. Where hybrids are concerned, the functions, types of data and of data analysis traditionally attributed to either of the two paradigms can thus be seen to overlap; it is no longer possible to maintain a sharp distinction between the two types of research.

49

50

Margrit Schreier

3.3 Quantification of qualitative data Quantification and subsequent statistical analysis of qualitative data constitute a form of cross-case analysis (Miles and Huberman 1994). In the first place, this requires that the — initially highly idiographic data — be transformed in such a way as to allow for interindividual comparison. In the case of interview data, this is best achieved by content analysis with respect to those aspects which are of central interest to the researcher. Where the interest is primarily in content, the next step of analysis consists in determining the frequencies of the coding of relevant categories; these can then be compared according to any variable of interest, such as gender, expert status, and so on by using methods such as chi-square or loglinear analyses. This allows for the testing of hypotheses such as: does the membership in a specific type of literary community (experts vs. novices) and the tradition informing a literary text influence readers’ understanding and appreciation of these texts (Dorfman 1996)? Does contextual knowledge change the appreciation and understanding of a (foreign language) literary text (Andringa 1996)? If the interest is primarily on similarity, further analysis involves transformation of the data into similarity or distance matrices which can then be subjected to, for instance, cluster analysis. In this way hypotheses can be tested or generated which assume the existence of types: types of readers, classes of text reception, and so on. Buchman and Eisner (1997) made use of cluster analysis and MDS in analyzing the results of their study on personal ads; Thury and Friedländer (1996) were able to identify three different stances towards reading literature by applying factor analysis to the results of their analysis of students’ and experts’ literary response journals. If data are of the network type (for instance in the reconstruction of subjective theories, various types of configuration and causal analysis, and so on), cross-case analysis requires not only the comparison of content, but the comparison according to links between concepts as well. In collecting and reconstructing subjective theories, for instance, respondents name relevant concepts, link those concepts to each other, and arrange those links between concepts as a network. Here, content analysis of concepts is not sufficient. In addition, the network data have to be transformed into lists where each listing specifies two concepts, the type of link between them, and other relevant information such as the position of these two concepts within the network as a whole (for details cf. Schreier 1997b). It is then possible to determine the interindividual frequencies of specific links between specific concepts (a special computer program has been developed by Oldenbürger 1992); the interindividually most frequent combinations of concepts linked in the same manner can, as a result, again be transformed into a network: a modal network representing an intersubjective theory. Viehoff (1995) used this procedure for reconstructing subjects’ knowledge about two genres: detective stories and fairy tales. The method is likewise suitable for determining modal types of text reception.

Qualitative methods in studying text reception

Alternatively, one may not be primarily interested in one modal intersubjective theory, but in types of intersubjective theories, depending on certain characteristics of the sample studied (such as gender, expert status, and so on), and how these compare to each other. In this case, intersubjective theories would be constructed in the manner described for specific subsamples. In order to compare various intersubjective theories, additional steps of analysis are necessary: the data, as they exist in list form, must be transformed into distance matrices; for their comparison, socalled cophenetic measures (measures of correlation for matrices) can be used (cf. Oldenbürger 1981). To my knowledge, this has not yet been done in ESL, although Groeben (1998) took a step in this direction by comparing subjective theories on the reception of Kleist’s Penthesilea depending on the participants’ gender. Quantification and subsequent statistical analysis are thus in principle possible for any type of qualitative data, allowing for the inference statistical testing of hypotheses even where complex content and corresponding procedures of data collection and analysis are concerned. At the same time, quantitative measures and statistical procedures are obviously suitable not only for hypothesis testing; instead, many procedures exist within the quantitative paradigm which have been designed for the purpose of data exploration (such as cluster analysis, factor analysis, etc.), and others easily lend themselves towards such a use (cross-tabulations, measures of correlation and regression, and so on; on exploratory data analysis in ESL cf. Ostermann 1993).

4. Designs for reception analysis in ESL In the previous sections, I have described a number of ways in which qualitative methods — which seem especially useful for conducting reception analysis — may be applied in an increasingly systematic manner and be used in cross-case analysis, thus combining aspects of so-called qualitative and quantitative research. If applied in the manner described, qualitative methods may well find their place in reception research. The ways in which qualitative methods may be used to this purpose are, however, highly diverse and potentially confusing. I would therefore like to conclude by giving a brief outline of two types of designs that might be useful in reception analysis, each involving the use of qualitative methods. The first design is intended for exploratory studies where little is known about the research question; the second design would be suitable for hypothesis testing. Reception analysis may well involve cases where it would be difficult to specify any hypotheses in advance, where one might even be uncertain as to which aspects of the text and which aspects of reception will be of special relevance. In this case one would conduct an exploratory, inductively oriented study suitable for identifying relevant factors which influence text reception. The method used for data

51

52

Margrit Schreier

collection would most likely be the interview, supplemented by the collection of other respondent variables which may be of interest (attitudes to issues addressed in the text, sociodemographic variables, and the like). While the method should allow recipients to put their own experience and impressions into their own words, they should nevertheless not be left entirely to their own devices in talking about the text in question. Instead, questions focusing on aspects of potential interest which are put to all respondents should be combined with ad hoc questions put to specific respondents only; otherwise, a subsequent comparison of individual text receptions would be nearly impossible. The interview data would then be content-analyzed. Building a coding schedule would most likely involve two aspects: in the first place, categories would be developed in a deductive manner, that is based on those questions put to all respondents. In addition, inductive development of categories will probably also take place, based on the recipients’ mentioning of aspects not previously considered by the researcher. Especially in the construction of inductive categories, the use of software designed for the analysis of qualitative data would be helpful, as it allows for the playing with categories and the trying out of different coding schemes; this trial phase should, however, be restricted to a subsection of the sample only. Once all interview data have been coded, they can be transformed into a distance matrix and subjected to a procedure such as cluster analysis. In this way, types of reception (in the sense of specific themes frequently occurring together) can be identified. The researcher can then attempt to match reception types with recipient characteristics. As a result, hypotheses concerning types of reception and differences between specific groups of recipients can be put forward. Such hypotheses which emerge as the result of an inductive process of data examination can, however, not count as confirmed on this basis; confirmation can only be achieved by a subsequent testing of these hypotheses in an additional, separate study. If the researcher is able to specify hypotheses in advance, the semi-structured interview would be a suitable method for data collection (or any other method for collecting free response verbal data: cf. Steen 1991), again combined with the collection of other factors of interest; most likely, these ‘other factors’ would function as independent, aspects of text reception as the dependent variables. Data collection would be followed by content analysis involving a coding scheme consisting largely of categories developed in a deductive manner. Next, coding frequencies for relevant categories (dependent variables) would be determined and compared according to the independent variables specified in the hypothesis by choosing an appropriate inference statistical measure such as chi-square or loglinear analysis; the results would show whether the hypothesis could be confirmed or not. Most likely, reception analysis would involve a large number of variables. In order to simplify data analysis, it might be useful to conduct the analysis not for individual category frequencies, but for indices combining several categories into one.

Qualitative methods in studying text reception

These ‘designs’ are relatively simple outlines, based on the assumption that it is only content which is of interest. If structural aspects are also relevant, the designs would have to be expanded so as to include, for instance, the collection of subjective theories and their subsequent aggregation and comparison, as outlined above.

References Andringa, E. 1996. “Die Bedeutung von Kontextwissen für das Verstehen fremdsprachiger Literatur”. SPIEL 15: 138–154. Andringa, E. 1998. “Statement within plenary discussion on methods”. Fifth Biennial IGEL Conference, Utrecht. Abstracts. http://www.lumis.uni-siegen.de/igel/en/frames.htm. Berelson, B. 1952. Content Analysis in Communication Research. Glencoe, Ill.: Free Press. Bortz, J. and Döring, N. 1995. Forschungsmethoden und Evaluation für Sozialwissenschaftler. Berlin etc.: Springer, 2nd rev. ed. Brewer, W. F. 1995. “Discourse force and empirical study of literature”. In Empirical Approaches to Literature, Gebhard Rusch (ed.), 89–95. Proceedings of the Fourth Biannual Conference of the International Society for the Empirical Study of Literature Budapest, August 1994. Siegen: Lumis Publications, Siegen University. Bryman, A. 1988. Quantity and Quality in Social Research. London etc.: Unwin Hyman. Cupchik, G. and Izadpanah, S. 1998. “Reactions to funny, violent, and sexual images in magazines”. SPIEL 17: 15–32. Denzin, N. K. 1970. The Research Act in Sociology. Chicago: Aldine. Denzin, N.K. and Lincoln, Y.S. (eds) 1994. Handbook of Qualitative Research. Thousand Oaks etc.: Sage. Dorfman, M. H. 1996. “Evaluating the interpretive community: Evidence from expert and novice readers”. Poetics 23: 453–470. Fielding, N. G. and Fielding, J. L. 1986. Linking Data. Beverly Hills etc.: Sage. Fielding, N.G. and Lee, R.M. 1998. Computer Analysis and Qualitative Research. London etc.: Sage. Flick, U. 1995a. Qualitative Forschung. Theorie, Methoden, Anwendung in Psychologie und Sozialwissenschaften. Reinbek: Rowohlt. Flick, U. 1995b. “Triangulation”. In Handbuch Qualitative Sozialforschung, U. Flick, E. von Kardoff, H. Keupp, L. von Rosenstiel and S. Wolff (eds), 432–434. Weinheim: PVU, 2nd ed. Geiger, K. F. 1975. “Jugendliche lesen Landser-Hefte. Hinweise auf Lektürefunktionen und - wirkungen”. In Literatur und Leser. Theorien und Modelle zur Rezeption literarischer Werke, G. E. Grimm (ed.), 324–341. Stuttgart: Reclam. Groeben, N. 1972. Literaturpsychologie. Literaturwissenschaft zwischen Hermeneutik und Empirie. Stuttgart: Kohlhammer. Groeben, N. 1977. Rezeptionsforschung als empirische Literaturwissenschaft: Paradigma- durch Methodendiskussion. Kronberg: Athenäum; 2. Aufl. 1980; Tübingen: Narr. Groeben, N. 1987. “Möglichkeiten und Grenzen der Kognitionskritik durch Inhaltsanalyse”. In Textanalyse als Kognitionskritik? Möglichkeiten und Grenzen ideologiekritischer Inhaltsanalyse, P. Vorderer and N. Groeben (eds), 1–21. Tübingen: Narr. Groeben, N. 1992. “Die Inhalts-Struktur-Trennung als konstantes Dialog-Konsens-Prinzip?!”. In Struktur-Lege-Verfahren als Dialog-Konsens-Methodik, B. Scheele (ed.), 42–89. Münster: Aschendorff.

53

54

Margrit Schreier

Groeben, N. 1994a. “Der Paradigma-Anspruch der Empirischen Literaturwissenschaft”. In Empirische Literaturwissenschaft in der Diskussion, A. Barsch, G. Rusch and R. Viehoff (eds), 21–38. Frankfurt/M.: Suhrkamp. Groeben, N. 1994b. “Literaturwissenschaft als empirisch-interdisziplinäre Kulturwissenschaft”. In Germanistik in der Mediengesellschaft, L. Jäger und B. Switalla (eds), 79–109. München: Fink. Groeben, N. 1995. “Methodenprobleme der Empirischen Literaturwissenschaft”. In Empirische Literatur- und Medienforschung, S.J. Schmidt (ed.), 56–71. (Lumis-Schriften Sonderreihe Bd. VII). Siegen: Lumis Publications. Groeben, N. 1998. “Interpretationsansätze zur Penthesilea und Gender-Sozialisation — eine dialog-hermeneutische Rezeptionsstudie”. In Freiburger literaturpsychologische Gespräche, C. Pietzcker (ed.), (vol. 17) 101–120. Würzburg: Königshausen u. Neumann. Groeben, N. and Rustemeyer, R. 1995. “Inhaltsanalyse”. In Bilanz qualitativer Forschung, E. König and P. Zedler (eds), (vol. 2: Methoden) 523–554. Weinheim: Deutscher Studien Verlag. Groeben, N., Wahl, D., Schlee, J. and Scheele, B. 1988. Forschungsprogramm Subjektive Theorien. Eine Einführung in die Psychologie des reflexiven Subjekts. Tübingen: Francke. Hesse-Biber, S. and Dupuis, P. 1995. “Hypothesis testing in computer-aided qualitative data analysis”. In Computer-Aided Qualitative Data Analysis: Theory, Methods, and Practice, U. Kelle (ed.), 129–135. London etc.: Sage. Hron, A. 1994. “Interview”. In Verbale Daten, G. L. Huber and H. Mandl (eds), 119–140. Weinheim: Beltz. Huber, G. L. 1995. “Qualitative hypothesis examination and theory building”. In Computer-Aided Qualitative Data Analysis: Theory, Methods, and Practice, U. Kelle (ed.), 136–151. London etc.: Sage. Huber, O. 1987. Das psychologische Experiment. Eine Einführung. Bern etc.: Huber. Hurrelmann, B., Hammer, M. and Nieß, F. 1993. Lesesozialisation Bd. 1. Leseklima in der Familie. Gütersloh: Verlag Bertelsmann Stiftung. Ibsch, E. 1994. “Das Selbstverständnis der Literaturwissenschaft in den Niederlanden”. In Empirische Literaturwissenschaft in der Diskussion, A. Barsch, G. Rusch and R. Viehoff (eds), 39–54. Frankfurt/M.: Suhrkamp. Ibsch, E. 1998. “Statement within plenary discussion on methods”. Fifth Biennial IGEL Conference, Utrecht. Abstracts. http://www.lumis.uni-siegen.de/igel/en/frames.htm. Janssen, T. 1996. “Asking for trouble. Teacher questions and assignments in the literature classroom”. SPIEL 15: 8–23. Kelle, U. (ed.). 1995. Computer-Aided Qualitative Data Analysis. Theory, Methods, and Practice. London etc.: Sage. Kelle, U. and Laurie, H. 1995. “Computer use in qualitative research and issues of validity”. In Computer-Aided Qualitative Data Analysis: Theory, Methods, and Practice, U. Kelle (ed.), 19–28. London etc.: Sage. Kuiken, D. and Miall, D. S. 1995. “Procedures in think aloud studies: Contributions to the phenomenology of literary response”. In Empirical Approaches to Literature, G. Rusch (ed.), 50–60. Proceedings of the Fourth Biannual Conference of the International Society for the Empirical Study of Literature Budapest, August 1994. Siegen: Lumis Publications, Siegen University. Lamnek, S. 1995. Qualitative Sozialforschung. 2 vols. Weinheim: Beltz, PVU, 3rd corr. ed. Landis, J.R. and Koch, G.G. 1977. “The measurement of observer agreement for categorical data”. Biometrics 33: 159–174. Malmgren, L.-G. 1996. “A Sign from the Beaver — a sign from thematic literature teaching”. SPIEL 15: 68–82.

Qualitative methods in studying text reception

Mayring, P. 1995. Qualitative Inhaltsanalyse. Grundfragen und Techniken. Weinheim: Deutscher Studienverlag, 5th ed. Mayring, P. 1996. Einführung in die qualitative Sozialforschung. Weinheim: PVU. Merten, K. 1983. Inhaltsanalyse. Einführung in Theorie, Methode und Praxis. Opladen: Westdeutscher Verlag. Miles, M. B. and Huberman, M. A. 1994. Qualitative Data Analysis. An Expanded Sourcebook. Thousand Oaks etc.: Sage. Oatley, K. 1999. “Meetings of minds: Dialogue, sympathy, and identification in reading fiction”. Poetics 26: 439–454. Oldenbürger, H.-A. 1981. “Zur Konkretisationserhebung literarischer Texte und hermeneutischer Deutungshypothesen durch Sortierung und Netzwerkbildung”. In Rezeption und Interpretation. Ein interdisziplinärer Versuch am Beispiel der “Hasenkatastrophe” von Robert Musil, N. Groeben (ed.), 161–204. Tübingen: Narr. Oldenbürger, H.-A. 1992. Netz-Werk-Zeug 1: Zählwerk für beliebige Variablenkombinationen. Version a-z. Göttingen: Seminar für Wirtschaftspädagogik. Ostermann, R. 1993. “Explorative Datenanalyse als Alternative zur modellgeleiteten statistischen Analyse”. SPIEL 12: 136–150. Ragin, C.C. 1987. The Comparative Method: Moving beyond Qualitative and Quantitative Strategies. Berkeley, Los Angeles: University of California Press. Ragin, C. C. 1994. “Introduction to qualitative comparative analyis”. In The Comparative Political Economy of the Welfare State, T. Janoski and A. Hicks (eds), 299–319. New York: Cambridge University Press. Richards, T.J. and Richards, L. 1994. “Using computers in qualitative research”. In Handbook of Qualitative Research, N.K. Denzin and Y.S. Lincoln (eds), 445–462. Thousand Oaks etc.: Sage. Rustemeyer, R. 1992. Praktisch-methodische Schritte der Inhaltsanalyse. Münster: Aschendorff. Scheele, B. (ed.). 1992. Struktur-Lege-Verfahren als Dialog-Konsens-Methodik. Münster: Aschendorff. Scheele, B. and Groeben, N. 1986/87. “Methodological aspects of illustrating the cognitiveinstrumental function of aesthetic communication: Employing a structure-formation technique with readers of (positive) literary utopias”. Poetics 15: 527–554. Scheele, B. and Schreier, M. 1994. “Dialog-Konsens-Methoden in der Empirischen Literaturwissenschaft”. In Empirische Literaturwissenschaft in der Diskussion, A. Barsch, G. Rusch and R. Viehoff (eds), 278–296. Frankfurt/M.: Suhrkamp. Schmidt, S. J. 1975. Literaturwissenschaft als argumentierende Wissenschaft. München: Fink. Schmidt, S. J. 1980. Grundriß der empirischen Literaturwissenschaft. Bd. 1: Der gesellschaftliche Handlungsbereich Literatur. Braunschweig: Vieweg. Schmidt, S. J. 1982. Grundriß der empirischen Literaturwissenschaft. Bd. 2: Zur Rekonstruktion literaturwissenschaftlicher Fragestellungen in einer Empirischen Theorie der Literatur. Braunschweig: Vieweg. Schön, E. 1996. “Die literarische Rezeption deutscher Schüler und Erwachsener im interkulturellen Vergleich. Ein Work-in-progress-Bericht aus der internationalen explorativen Vergleichsstudie ‘Cultural rules of interpretation in six European countries’”. SPIEL 15:97–121. Schreier, M. 1997a. “Computer-aided qualitative data analysis and its uses in the empirical study of literature”. In The Systemic and Empirical Approach to Literature and Culture as Theory and Application, S. Tötösy de Zepetnek and I. Sywenky (eds), 155–162. (Lumis Publications, Special Editions, Vol VIII). Siegen: Lumis Publications. Schreier, M. 1997b. “Die Aggregierung Subjektiver Theorien: Vorgehensweise, Probleme, Perspektiven”. Kölner Psychologische Studien II: 37–71.

55



56

Margrit Schreier

Sibert, E. and Shelly, A. 1995. “Using logic programming for hypothesis generation and refinement”. In Computer-Aided Qualitative Data Analysis. Theory, Methods, and Practice, U. Kelle (ed.), 113–128). London etc.: Sage. Sieber, S. D. 1979. “The integration of fieldwork and survey methods”. American Journal of Sociology 78: 1335–1339. Steen, G. 1991. “The empirical study of reading: Methods of data collection”. Poetics 20:559–575. Tesch, R. 1990. Qualitative Research: Analysis Types and Software Tools. New York etc.: The Falmer Press. Thury, E. M. and Friedlander, A. 1996. “Factors in transactions with literature: An analysis of student response journals”. In Empirical Approaches to Literature and Aesthetics, R. J. Kreuz and M. S. MacNealy (eds), 419–443. Norwood, NJ: Ablex. Vidich, A. J. and Lyman, S. M. 1994. “Qualitative methods: Their history in sociology and anthropology”. In Handbook of Qualitative Research, N. K. Denzin and Y. S. Lincoln (eds), 23–59. Thousand Oaks etc.: Sage. Viehoff, R. 1995. “Literary genres as cognitive schemata”. In Empirical Approaches to Literature, G. Rusch (ed.), 72–76. Proceedings of the Fourth Biannual Conference of the International Society for the Empirical Study of Literature Budapest, August 1994. Siegen: Lumis Publications, Siegen University. Watt, J. H. and Van den Berg, S. 1995. Research Methods for Communication Science. Boston etc.: Allyn & Bacon. Weitzman, E. A. and Miles, M. B. 1995. Computer Programs for Qualitative Data Analysis. Thousand Oaks, CA: Sage. Wieler, P. 1997. Vorlesen in der Familie. Fallstudien zur literarisch-kulturellen Sozialisation von Vierjährigen. Weinheim, München: Juventa.





Agency, plot, and a structural affect theory of literary story comprehension* Arthur C. Graesser and Bianca Klettke The University of Memphis

When a reader is lost in a book, the wheels of comprehension move on many levels. Most of the wheels need to be in working order and the wheels must mesh in harmony. Damage to a single wheel can spoil the entire experience. This chapter focuses on three wheels that need to mesh when adults comprehend literary short stories: plot, agents, and emotions. The plot consists of a structured configuration of events, actions, goals, plans, and conflicts among protagonists. The agents include the story characters and the pragmatic agents (that is narrator, narratee, reader, writer). Each of these agents has a point of view that witnesses some of the actions and events in the plot. Regarding emotions, it is important to track the emotions of both the characters and the reader as the plot unfolds. Our contention is that a captivating short story has a clever plot, has multiple agents with different points of view, has a complex texture of emotions, and has harmony among these components. No doubt, there are other wheels that amplify the aesthetic experience, such as novelty, metaphor, and stylistics (Gibbs 1999; Ibsch et al. 1991; Schmidt 1982; Steen 1994). But our contention is that plot, agency, and emotions are central. Brewer’s ‘structural affect theory’ was the first model in discourse psychology that seriously attempted to integrate plot, agency, and emotions (Brewer 1996; Brewer and Lichtenstein 1981; Brewer and Ohtsuka 1988) According to this theory, the emotions of the reader are systematically determined by the configuration of the plot and the knowledge states of various agents. For example, consider what happens when the reader has the emotion of surprise. The author withholds critical information at the beginning of the story, information that is necessary for a correct interpretation of the story. Later on, the critical information is revealed, which triggers surprise in the reader. The emotion of suspense is set up quite differently. There is an initiating event that has the potential to lead to a significant outcome, which may be good or bad for a central character. The reader comprehends an intervening set of events and actions in anticipation of the significant outcome. For example, suppose that a man in a mask puts a bomb in a car. The reader expects the negative outcome of the bomb exploding and killing an important character. But the reader does not know exactly when this negative outcome will occur. The reader experiences

58

Arthur C. Graesser and Bianca Klettke

suspense, episode after episode, until the negative outcome occurs. A good writer can prolong the suspense with many episodes. A clever writer includes ‘false alarm’ episodes, that is, a negative outcome is expected during an episode but never occurs. The writer controls the tension and arousal of the reader as the reader experiences a complex texture of emotions. As a good experimental psychologist, Brewer tested his model by (a) manipulating features of the text and knowledge states of the reader and (b) observing whether these manipulations systematically predicted readers’ self reports of particular emotions and how much they liked the story. The structural affect theory fared quite well in accounting for the psychological data. It is somewhat remarkable that there has not been more research to follow up on Brewer’s structural affect theory. Most of the follow-up research on suspense is covered in a book edited by Vorderer et al. (1996), but other reader emotions (such as surprise and curiosity) have not received much attention. There has been a large amount of research on the emotions that readers experience while comprehending narrative (Cupchik et al. 1998; Miall and Kuiken 1994; Tan 1995) and a separate line of research on the construction of plots (Graesser et al. 1991; Jose and Brewer 1984; Lehnert 1981; Trabasso and Van den Broek 1985), but these two bodies of research are more detached than tightly integrated. One might have expected a flurry of research activity that maps plots to reader emotion after Brewer introduced the structural affect theory. The role of agency is central to structural affect theory. It is critical to keep track of the knowledge states of the various characters and of the reader in order to set up suspense and other reader emotions. For example, the reader may know that the car has a bomb and the masked rogue may also know it, but other character agents may be totally ignorant, for instance the victim who enters the car, an innocent bystander, and the victim’s wife in another city. This dissociation in knowledge between the pragmatic agent (reader, writer, narrator, narratee) and a character agent creates a dramatic tension and makes the story interesting. The suspense can even be recreated when the story is read multiple times (Gerrig 1993). When the reader has privileged knowledge that is not known by a particular character, there is the question of whether the reader is capable of keeping these different mental perspectives straight. In the case of suspense, the reader is at least partially successful in keeping track of the knowledge of different agents. Otherwise, suspense would not work. Suspense would die if all of the characters in the story world knew about the bomb in the car. The emotion of surprise also involves a discrepancy in knowledge states among agents; in this case, the writer and narrator know about the surprising event, but not the narratee and reader. We have recently investigated the extent to which readers track the various agents when literary short stories are comprehended (Graesser et al. in press; Graesser et al. 1999a; Graesser et al. 1999b; Graesser et al. 1997). Comprehenders potentially construct multiple agents in their cognitive representations when they

Agency, plot, and a theory of literary story comprehension

read a story. Each agent has human qualities, such as speaking, perceiving, believing, knowing, wanting, liking, acting, and experiencing emotions. The characters in a story constitute one ensemble of agents. The protagonists and antagonists are presumably more salient in the reader’s mind than are the minor characters and the characters who are functionally props. Another ensemble of agents, called the pragmatic agents, participate in a one-sided communication from the narrator to the narratee, or from the author to the reader. The narrator is the agent who presents the story to an imaginary addressee or recipient, called the narratee. The author is the actual person who writes the story whereas the reader is either an actual or virtual reader of the story. The notion that there are multiple agents, multiple voices, and multiple conversations between agents has been assumed by virtually all scholars in narratology, literature, and literary criticism (Bakhtin 1981; Chafe 1994; Duchan et al. 1995; Prince 1982; Van Peer and Chatman in press). However, the ‘multiagent view’ has only recently been investigated by discourse psychologists who investigate the comprehension of bona fide literature (Duchan et al. 1995; Millis 1995). Our contention is that an appropriate tracking of agents, in a society of story agents, is very critical for setting up a plot that works and for creating the intended reader response. The present chapter briefly describes two lines of research that assess agent tracking in the minds of adult readers while they read literary stories. Graesser et al. (1999) investigated the extent to which readers could remember “who said what?” after they finished reading a literary short story. Graesser, Bowers et al. (in press) investigated the extent to which readers keep track of “who knows what?”. We await future research to explore the tracking of “who saw what?”, “who heard what?”, “who wants what?”, “who experienced what emotion?”, and so on. Ideally, the reader should be able to keep the various agents distinct and should faithfully update each agent on all dimensions as the story unfolds. From a computational perspective, however, an accurate tracking of agents would be a very difficult achievement. Suppose there were 10 symbolic expressions associated with proposition P: X said P, X liked P, X wanted P, X knew P, X saw P, X heard P, etc. Suppose further that there were 20 agents in the story world and 1000 propositions about the story world. There would be 10 × 20 × 1000 = 200,000 expressions to evaluate. There would indeed be a combinatorial explosion problem when viewed from this computational perspective. If the human mind is capably of pulling this off, that would be a remarkable achievement. However, there may be cognitive representations and strategies to circumvent the combinatorial explosion. According to research in discourse psychology, there are psychological constraints on the process of constructing mental microworlds, so ideal complete representations are frequently not constructed (Graesser et al. 1998; Graesser et al. 1994; Kintsch 1998; Zwaan and Radvansky 1998).

59

60

Arthur C. Graesser and Bianca Klettke

1.

The salience of character agents and pragmatic agents in memory: Tracking “who said what?”

Adult readers are perhaps able to keep the various pragmatic agents and character agents distinct in their minds. In essence, the reader processes the story world on multiple tracks (one per agent) and multiple communication channels between particular agents. However, tracking the agents and communication channels may be difficult because sometimes two or more agents are intentionally amalgamated by the author. This is illustrated by contrasting first-, second-, and third-person narration. First-person narrator. I opened the letter and read my own obituary. Second-person narrator. You open the letter and read your own obituary. Third-person narrator. She opened the letter and read her own obituary. The narrator agent is amalgamated with a character agent in the case of first-person narration. The narrator takes the point of view of one character and speaks to the narratee through that character’s eyes. This allows the reader to see the world and experience consciousness from the point of view of the character. When the story has a second-person narrator, there is an amalgamation of four agents: narrator, narratee, character, and reader. By fusing the reader with the other three agents, the author attempts to sweep up the reader as a participant in the story world and thereby increase reader involvement. The narrator and character agents are functionally separated in third-person narration. The separation facilitates any attempt to discriminate the different mental perspectives. When the third-person narrator is omniscient, the narrator sits above the story world, with a privileged access to the mental states of all characters, and tells the story to the narratee. Graesser, Bowers et al. (1999) investigated whether there are differences in the relative salience of agents in long-term memory after literary stories are read. The psychological status of the pragmatic agents has always been controversial in the case of written text because many of these agents are invisible or minimally constrained by the explicit text (Duchan et al. 1995; Gerrig 1993). Authors do not normally introduce themselves explicitly in the text (for instance “I am Art Graesser and I am going to report some research on …”). Similarly, third-person narrators do not normally introduce themselves (for instance, “I am the narrator and I will be telling you the story about …”). The text virtually never refers to the narratee and the reader. Pragmatic agents are linguistically and visually co-present in oral, faceto-face conversations (Clark 1996), but pragmatic agents are not normally copresent when printed text is read. If readers do construct the pragmatic agents during reading, they do this without obvious cues in the text that unambiguously refer to the existence of authors, readers, narrators, and narratees. Graesser, Bowers et al. (1999) tested some predictions about the relative salience of narrator and character agents. An agent amalgamation hypothesis

Agency, plot, and a theory of literary story comprehension

predicts that amalgamated agents are more salient during comprehension and subsequent memory than agents that are not amalgamated. Amalgamated agents have multiple roles and functions, so they are part of more levels of structure (that is, story world versus pragmatic context), have more associations with other text constituents, have more retrieval routes, and have richer elaborations. According to the agent amalgamation hypothesis, there is the following gradient in salience among three types of narrators: third-person narrator < first-person narrator < second-person narrator. Moreover, those character agents who are not affiliated with a pragmatic agent (hereafter referred to a nonnarrator character agents) are predicted to be less salient than the character agents who are affiliated with firstand second-person narrators. Our corpus of literary stories ended up including first-person and third-person narration, but not second-person narration (second person narrations are rare in the English language). Therefore, the agent amalgamation hypothesis generates two predictions about the relative salience of agents in first-person and third-person stories. Prediction 1: third-person narrator < first-person narrator Prediction 2: nonnarrator character agent < first-person narrator The agent amalgamation hypothesis does not offer a prediction about the relative salience of a third-person narrator and a nonnarrator character agent. Communication theories (Jakobson 1960; Reddy 1979; Rosenblatt 1978) make different predictions than the agent amalgamation hypothesis. Communication theories assume there are multiple levels of embedded communication and that the top levels are more salient than the bottom levels. Suppose, for example, that the following sentence occurred in a story with a third-person narrator: “Adam told Bob that the boss said that Bob would be fired.” There would be three levels of communication in this example (Bruce 1981), as represented below. TOP LEVEL: told (narrator, narratee, SECOND LEVEL) SECOND LEVEL: told (Adam, Bob, THIRD LEVEL) THIRD LEVEL: told (boss, Adam, (Bob would be fired)) A structural analysis predicts that the narrator and narratee are the most salient agents because they are at the top level. If the agents at the superordinate levels of structure are more salient than the agents at the subordinate levels, then there is the following predicted gradient in salience: (narrator, narratee) > (Adam, Bob) > (boss). A structural prominence hypothesis generates the following prediction about the relative salience of agents. Prediction 3: third-person narrator > nonnarrator character agent An invisible third-person narrator hypothesis generates a prediction that is directly incompatible with the structural prominence hypothesis.

61

62

Arthur C. Graesser and Bianca Klettke

Prediction 4: third-person narrator < nonnarrator character agent The invisible third-person narrator hypothesis is consistent with those theories which assume that readers do not notice the third person narrator unless there are obvious features in the text that reveal the existence of that narrator ( Banfield 1982; Duchan et al. 1995; Gerrig 1993). A strong form of this hypothesis predicts that the salience of the third-person narrator is zero, unless the text has explicit cues that flag the existence of the narrator. A weak form of this hypothesis generates prediction 4, but allows for the possibility of an above-zero salience of the third-person narrator. The study by Graesser, Bowers et al. (1999) tested the predictions of the agent amalgamation hypothesis, the structural prominence hypothesis, and the invisible third-person agent hypothesis. A sample of 120 college students read one of the 10 literary short stories in the corpus. The 10 literary short stories were written by professional authors. Five stories were written in first-person perspective and 5 in the third-person perspective. For example, one of the first-person stories was “What men love for” (Dale Ray Phillips 1989) and one of the third-person stories were “Conscience of the cop” (William Fay 1956). The stories ranged in length from 9 to 18 pages and were judged to be sufficiently captivating for college students. Each story contained at least two prominent characters (designated here as characters A and B) and a narrator. After reading the story, the participants completed a source memory test on a set of speech acts (statements) expressed in the story. The source memory test measures how accurately the participant can remember who said what. The original test statements were expressed by either the narrator, character A, or character B, but each test statement conceivably could have been expressed by any of the three agents. That is, we did not select test statements in which the speaker could be reconstructed by sophisticated guessing strategies. There were four answer options available to the subjects when they gave source memory judgments for each statement, as illustrated below. Who said that if forced to choose someone else to sleep with, Vicky would choose Jim? (a) Vicky [the narrator] (b) David [character A] (c) Jim [character B] (d) Neither [none of the three agents said it because it was not a true statement]

The “Neither” test statements were systematically generated by transforming the original statements. For example, the yoked statement for the example statement was: “Who said that if forced to choose someone else to sleep with, David would choose Michelle?” (Michelle was another character in the story). Different versions of the test booklets were prepared for each story so that any given statement was sometimes in the original form and sometimes in the transformed false form.

Agency, plot, and a theory of literary story comprehension

Measures of agent salience supported the invisible third-person narrator hypothesis and the agent amalgamation hypotheses, but not the structural prominence hypothesis. The results followed the following pattern on a gradient from low to high salience: First-person narrator > nonnarrator character > third-person narrator > 0. One measure of agent salience is simply the proportion of source memory decisions that are correct. These scores were .75, .62, and .53 for firstperson narrator, nonnarrator character, and third-person narrator, respectively. Given that there are four response options, .25 is chance performance. Additional measures were collected that partialled out guessing biases and that segregated source memory from statement detection. Source memory assesses whether the reader can remember “who said what” after they finish reading a literary story. Source memory is to be distinguished from “statement detection”, which is memory for the content of what is said (that is discriminating whether the statement is true or false from the standpoint of the story world). The measure of source memory for the narrator was computed as N/(N + A + B) for the items in which the narrator spoke; N, A, and B refer to the number of decisions for the narrator, character A, and character B, respectively, when the correct response was the narrator. Similarly, A/(N + A + B) was the measure of source memory when character A was the speaker and B/(N + A + B) was the measure of source memory when character B was the speaker; an average source memory was computed for the two characters. Once again, the results supported the predictions of the agent amalgamation hypothesis (predictions 1 and 2) and the invisible third-person narrator hypothesis (prediction 4), but not the structural prominence hypothesis (prediction 3). The source memory scores were .84, .73, and .65 for first-person narrator, nonnarrator character, and third-person narrator, respectively. In contrast, a measure of statement detection did not differ significantly among the different classes of agents. This measure is the proportion of observations in which the participant decided that one of the agents spoke (N, A, or B), given that the narrator spoke or one of the characters spoke. Statement detection scores were .89, .85, and .81 for first-person narrator, nonnarrator character, and third-person narrator, respectively. Graesser, Bowers et al. (in press) conducted a number of follow up studies that ruled out potential confounding variables (for instance sophisticated guessing, text features, story reading time) and that assessed the generality of the findings across readers (for instance gender, verbal ability, literary expertise). The source memory effects could not be explained by potential confounding variables and robustly persisted across readers with different abilities. There are at least two practical implications of these results from the standpoint of crafting a text in a fashion that assists the reader in tracking agents. First, the third-person narrator is nearly invisible, so there need to be explicit cues to designate its existence. Sometimes text cues can do this (for instance “I’m going to tell you the story about …”). In this age of film and talking heads on computers, the

63

64

Arthur C. Graesser and Bianca Klettke

narrator can be depicted in a visual animated form. Second, the first-person narrator is very salient, so this is an excellent way of magnifying the psychological state of a particular character agent. When the story world is viewed from the perspective of the first-person narrator, the filter may be incomplete or distorted, but readers will better be able to keep track of the character’s point of view. Scholars in literature and literary criticism have perhaps known this for many years, but the Graesser, Bowers, et al. (1999) study was the first to substantiate the claim in a psychology experiment.

2.

Propagating knowledge in a story world: Tracking who knows what?

The reader needs to accurately track the knowledge of characters in order to comprehend and have an appropriate emotional response to some stories. For example, consider Oedipus Rex, the Greek drama by Sophocles. At one point in the story, Oedipus is ignorant of his own identity, the fact that he murdered his father, and the fact that he married his mother. However, these facts are known by the sightless prophet Teresias and the audience. The meaning of Oedipus Rex would indeed be very different if Oedipus was fully aware of his identity when he murdered his father and married his mother. Some of the readers’ emotional reactions would shift from suspense to horror. An adequate understanding and appreciation of a story requires an accurate tracking of “who knows what”. Suppose that the following speech act is expressed by a character in a story: “Glenda told her husband that she has cancer.” The propositional content of this speech act (that is “Glenda has cancer”) is potentially propagated among various character agents in the story world. Glenda and her husband obviously know proposition P as a consequence of the speech act, but who else in the story world knows it? The doctor should know it, but will their children, their friends, and strangers? Will it take time for the knowledge to propagate, as the news travels through networks of kinship, friendship, and acquaintances? Knowledge propagation may spontaneously occur when speech acts are comprehended during reading, but an alternative possibility is that knowledge propagation primarily occurs when readers are probed with “who knows what?” questions. Graesser et al. (1999) conducted a study that assessed knowledge tracking among character agents in literary short stories. College students read literary short stories and subsequently rated whether particular characters had knowledge of the information expressed in the speech acts. The stories and speech acts were the same as those used in the Graesser, Bowers et al. (in press) study except that the test involved knowledge ratings (who knows what?) instead of a test of source memory (who said what?). There was a 6-point character knowledge rating scale, varying from 1 (character definitely does not know proposition P) to 6 (character definitely does

Agency, plot, and a theory of literary story comprehension

Table 1.

Knowledge ratings (Who knows what?) for different classes of character agents (based on Graesser, Bowers, Olde, White and Person, 1999)

Category of agent

Mean rating

Definition

In scene of conversation Speaker (S) Addressee (A)

5.61 5.15

Side Participant (SP)

5.08

Overhearer (O)

4.05

The character who utters the speech act The character who hears the speaker, is recognized as a participant in the conversation, and is the recipient of the utterance A character who hears the speaker, is recognized as a participant in the conversation, and is not the addressee A character who hears the speaker but is not openly recognized as a participant in the conversation. Clark (1996) segregates this further into bystanders (whose presence is recognized by the speaker) and eavesdroppers (whose presence is not recognized by the speaker)

Out of scene of conversation Friend (F)

2.90

Acquaintance (AQ)

1.93

Disliked Character (D)

2.23

Stranger (ST) Important Unimportant

1.87 1.39

A character who has an intimate personal relationship with the speaker, such as a close friend or family member A character who the speaker has met, but does not know well A character who the speaker knows but does not like A character who the speaker has not met

know proposition P). Mean ratings were collected for different classes of character agents in the story world. The classification of agents in the story world was contingent on the speaker of the speech act in the story. For example a “Friend” would be regarded as a story character who is a friend of the speaker. We adopted Clark’s definitions of conversational roles when considering those agents in the scene of the conversation where the speech act occurred (Clark 1996). These categories and definitions are shown in Table 1, under “in scene of conversation”. These categories include the speaker (S), addressee (A), side participant (SP), and overhearer (O). For example, in the story “Almost her” by Bridget Mazur (1993), there is a conversation in which Jim says

65

66

Arthur C. Graesser and Bianca Klettke

that he sent Michelle poems. Jim is addressing Dave, but Vicky is at the same table. A bartender is nearby whereas several other characters are not present in the scene of the conversation. Jim is assigned the role of the speaker, Dave the addressee, Vicky a side participant, and the bartender an overhearer. When Jim utters the speech act, the knowledge is presumably propagated systematically among the other characters. These character agents who are in the scene of the conversation should have a much higher likelihood of knowing P than would the character agents who are “out of the scene of conversation”. Affiliation status should predict the knowledge propagation for these out-of-scene agents: friend, acquaintance, disliked person, or stranger to the speaker (see Table 1). We would expect knowledge to be propagated to agents who have a close affiliation to the speaker by virtue of auxiliary conversations, that is conversations other than the one with the speech act under consideration. Table 1 shows the mean knowledge ratings that college students gave to the characters in the various classes of agents. Graesser et al. (1999) reported that the pattern of ratings were compatible with the following principles of knowledge propagation 1. Speaker primacy. The speaker has the most knowledge of the content expressed by the speaker. The predicted pattern is: S > max (remaining classes of agents). That is, the mean for the speakers should be higher than the highest mean among the other classes of characters. 2. Conversational circle. The parties of a conversation have more knowledge of the information expressed in the conversation than do characters who are not part of the conversation. The predicted pattern is: min(S, A, SP) > max(O, F, AQ, D, ST). That is, the lowest mean of S, A, and SP should be higher than the highest mean of O, F, AQ, D, and ST. 3. Spatial proximity. Characters who are spatially near a conversation (within hearing distance) should have more knowledge of the information in the conversation than characters who are distant from the conversation. One prediction is overhearer-friend > F. Another prediction is overhearer-nonfriend > ST. 4. Affiliation. Friends of the speaker in a conversation should have more knowledge of the information in the conversation than should nonfriends. The predicted gradient is F > AQ > ST. These results indicate that readers can discriminatively track the knowledge of character agents in a story world. The results could have turned out quite differently. It might have been the case that the reader simply assumes that the characters know what the reader knows. That would predict equivalent ratings for all of the agent classes in Table 1. However, this egocentric perspective was not supported by the data. The knowledge ratings robustly differed among agent categories: speaker > addressee = side participant > overhearer > out-of-scene friend > disliked person > acquaintance > stranger.



Agency, plot, and a theory of literary story comprehension

The Graesser et al. (1999) study performed follow-up research that dissected the psychological processes that apparently explain knowledge propagation. They assessed the extent to which the knowledge ratings were affected by the episodic retrieval of the speech acts from long-term memory, by the reconstruction of information from different sources (that is full story, story summary, character key, versus content of speech act), and by differences among readers. However, it is beyond the scope of this chapter to discuss such mechanisms. The important point, from the present standpoint, is that readers are quite discriminating and systematic in knowledge tracking. This result is compatible with our claim that the tracking of agents needs to mesh with plot and reader emotions. Interesting stories have a systematic mapping among agents, plot, and emotions. For this to happen, readers need to be able to track character knowledge with some accuracy.

3.

Closing comment

We have not yet discovered most of the interesting links among plot, agency, and emotion. Just a handful of mappings have been identified, as in the case of Brewer’s structural affect theory. We hope that more plot-agency-emotion interactions are explored in the future research agendas of IGEL scholars and scientists. It will take serious interdisciplinary projects between literary scholars and social scientists in order to identify these interactions and to formulate theories that explain them. But such interdisciplinary projects are at the heart of the IGEL enterprise.

Note * This research was partially funded by grants from the Office of Naval Research (N00014–92– J–1826) and the National Science Foundation (SBR 9720314) awarded to the first author.

References Bakhtin, M. M. 1981. The Dialogic Imagination. Holoquist, M. (ed.) Austin: U of Texas P. Banfield, A. 1982. Unspeakable Sentences: Narration and Representation in the Language of Fiction. Boston: Routledge. Brewer, W. F. 1996. “The nature of narrative suspense and the problem of rereading”. In Suspense: Conceptualizations, Theoretical Analyses, and Empirical Explorations, P. Vorderer, H. J. Wulff and M. Friedrichsen (eds), 107–27. Mahwah, NJ: Erlbaum. Brewer, W. F. and Lichtenstein, E.H. 1981. “Event schemas, story schemas, and story grammars”. In Attention and Performance IX, J. Long and A. Baddeley (eds), 363–379. Hillsdale, NJ: Erlbaum. Brewer, W. F. and Ohtsuka, K. 1988. “Story structure, characterization, just world organization, and reader affect in American and Hungarian short stories”. Poetics 17: 395–415.

67

68

Arthur C. Graesser and Bianca Klettke

Bruce, B. 1981. “A social interaction model of reading”. Discourse Processes 4: 273–311. Chafe, W. 1994. Discourse, Consciousness, and Time. Chicago, IL: University of Chicago Press. Clark, H. H. 1996. Using Language. Cambridge: Cambridge University Press. Cupchik, G. C., Oatley, K. and Vorderer, P. 1998. “Emotional effects of reading excerpts from short stories by James Joyce”. Poetics 25: 363–377. Duchan, J. F., Bruder, G. A. and Hewitt, L. E. (eds). 1995. Deixis in Narratives: A Cognitive Science Perspective. Hillsdale, NJ: Erlbaum. Fay, W. 1956. “Conscience of the cop”. In Best Detective Stories of the Year — 1956, D. C. Cook (ed.), 180–193. New York: Dutton. Gerrig, R. 1993. Experiencing Narrative Worlds: On the Psychological Activities of Reading. New Haven, CT: Yale University Press. Gibbs, R. 1999. Intentions in the Experience of Meaning. Cambridge: Cambridge University Press. Graesser, A. C., Bowers, C. A., Bayen, U. J. and Hu, X. (in press). “Who said what? Who knows what? Tracking speakers and knowledge in narrative”. In Narrative Perspective: Cognition and Emotion, W. van Peer and S. Chatman (eds). Albany: SUNY University Press. Graesser, A. C., Bowers, C. A., Olde, B. and Pomeroy, V. 1999. “Who said what? Source memory for narrator and character agents in literary short stories”. Journal of Educational Psychology 91: 284–300. Graesser, A. C., Bowers, C. A., Olde, B., White, K. and Person, N. K. 1999. “Who knows what? Propagation of knowledge among agents in a literary story world”. Poetics 26: 142–175. Graesser, A. C., Golding, J. M. and Long, D. L. 1991. “Narrative representation and comprehension”. In Handbook of Reading Research (vol. 2), R. Barr, M. L. Kamil, P. Mosenthal and P. D. Pearson (eds), pp. 171–204. White Plains. NY: Longman. Graesser, A. C., Kassler, M. A., Kreuz, R. J. and McLain-Allen, B. 1998. “Verification of statements about story worlds that deviate from normal conceptions of time: What is true about Einstein’s Dreams?”. Cognitive Psychology 35: 246–301. Graesser, A. C., Singer, M. and Trabasso, T. 1994. “Constructing inferences during narrative text comprehension”. Psychological Review 101: 371–95. Graesser, A. C., Swamer, S. and Hu, X. 1997. “Quantitative discourse psychology”. Discourse Psychology 23: 229–264. Ibsch, E., Schram, D. and Steen, G. (eds). 1991. “Empirical Studies of Literature”. Proceedings of the second IGEL-Conference. Amsterdam: Rodopi. Jakobson, R. 1960. “Closing statement: Linguistics and poetics”. In Style in Language, T. Sebeok (ed.), 350–377. New York: Wiley. Jose, P. E. and Brewer, W. F. 1984. “Development of story liking: Character identification, suspense, and outcome resolution”. Developmental Psychology 20: 911–924. Kintsch, W. 1998. Comprehension: A Paradigm for Cognition. Cambridge: Cambridge University Press. Lehnert, W.G. 1981. “Plot units and narrative summarization”. Cognitive Science 5: 293–331. Mazur, B. 1993. “Almost her”. Fiction 11: 150–158. Miall, D. S. and Kuiken, D. 1994. “Beyond text theory: Understanding literary response”. Discourse Processes 17: 337–352. Millis, K. K. 1995. “Encoding discourse perspective during the reading of a literary text”. Poetics 23: 235–253. Phillips, D. R. 1989. “What men love for”. In The Best American Short Stories 1989, R. Stone and K. Kenison (eds), 269–280. Boston: Houghton Mifflin. Prince, G. 1982. Narratology: The Nature and Function of Narrative. The Hague: Mouton. Reddy, M. 1979. “The conduit metaphor”. In Metaphor and Thought, A. Ortony (ed.), 284–324. Cambridge, England: Cambridge University Press.



Agency, plot, and a theory of literary story comprehension

Rosenblatt, L. 1978. The Reader, the Text, and the Poem. Carbondale, IL: Southern Illinois University Press. Schmidt, S. J. 1982. Foundations for the Empirical Study of Literature: The Components of a Basic Theory. Hamburg, Germany: Helmut Buske Verlag. Steen, G. 1994. Understanding Metaphor in Literature. London: Longman. Tan, E. 1995. Emotion and the Structure of Narrative Film: Film as an Emotion Machine. Mahwah, NJ: Erlbaum. Trabasso, T. and Van den Broek, P. 1985. “Causal thinking and importance of story events”. Journal of Memory and Language 24: 595–611. Van Peer, W. and Chatman, S. (in press) (eds). Narrative Perspective: Cognition and Emotion. Albany: SUNY University Press. Vorderer, P., Wulff, H. J. and Friedrichsen, M. (eds). 1996. Suspense: Conceptualizations, Theoretical Analyses, and Empirical Explorations. Mahwah, NJ: Erlbaum. Zwaan, R. A. and Radvansky, G. A. 1998. “Situation models in language comprehension and memory”. Psychological Bulletin 123: 162–185.

69



Time in narrative comprehension A cognitive perspective Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield Florida State University

Clothes, spectacles, hair, spinal braces, and so on — these came later. Entirely intelligibly, though, to prevent needless suffering, the dental work was usually completed while the patients were not yet alive. The Kapos would go at it, crudely but effectively, with knives or chisels or any tool that came to hand. Most of the gold we used, of course, came directly from the Reichsbank. But every German present, even the humblest, gave willingly of his own store — I more than any other officer save “Uncle Pepi” himself. All those years I amassed it, and polished it with my mind: for the Jews’ teeth. The bulk of the clothes were contributed by the Reich Youth Leadership. Hair for the Jews came courtesy of Filzfabrik A. G. of Roth, near Nuremberg. Freight cars full of it. Freight car after freight car. Martin Amis, Time’s Arrow (1992), pp. 120–121

The order in which we experience events has important implications for our interpretation of those events. This is wryly illustrated in Martin Amis’ novel Time’s Arrow. The story is told through the eyes of a character, first known as Dr. Tod Friendly and later known as Dr. Odilo Unverdorben, who experiences, in reversed chronology, the events that make up his life story. One of the implications of this reversal of chronological order is a dramatic inversion of the morality of the actions of the Nazis and the general German public. Rather than committing genocide, they are now seen to create a race, providing its members with physical and material goods (as well as pets, which they receive with tears of joy), and welcoming them into German society. Language allows us to convey events in an order and manner that is different from our everyday experience. We can refer to events, locations, people, and objects that are not in our immediate experiential field (for instance the Big Bang, the midAtlantic ridge, and Napoleon). This property of language, displacement, has been considered one of its design features (Hockett 1960). In real life, we experience events as a continuous flow. Because of its design feature of displacement, language

72

Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

allows speakers and writers to jump forward or backward in time. This creates discrepancies between how we experience events in real life and how we experience them vicariously through language. Although events can be presented in a chronological and continuous sequence, skilled story tellers will often deviate from this. In fact, none less than Aristotle in his Poetics (trans. 1967) encouraged writers of fiction to do so. He identified the plot as the major organizing structure of narratives and admonished poets to describe events only when they are relevant to the plot, just like Homer had done centuries before them. They were to refrain from giving a blow-by-blow chronological account of an episode. This Aristotle considered to be the province of historians. The overarching importance of the plot as an organizing structure often necessitates temporal discontinuities, because events irrelevant to the plot (such as the character brushing his teeth or going to sleep) are to be omitted. Also, a plot is often more effectively conveyed by reversing the order of certain events. Obviously, various modern writers, such as Martin Amis, Jorge Luis Borges, William Friedman, Kurt Vonnegut, and others have exploited the displacement feature of language to great effect as a literary device to defamiliarize our conception of time and chronology. Because skilled story tellers heed Aristotle’s advice, narratives are replete with temporal discontinuities. Given that temporal discontinuities entail a deviation from our normal experience of events, it might be hypothesized that they are marked relative to temporal continuities. Cognitive psychologists have studied the effects of temporal discontinuities and other deviations from everyday chronology on comprehension processes. The purpose of this chapter is to review the exisisting empirical literature on this topic. However, before we do so, we will first briefly review our theory of situation models, and next, the relevant linguistic literature on time in language.

1.

Situation models

Situation models are mental representations of the state of affairs described in a text (Van Dijk and Kintsch 1983). The generally accepted view in cognitive psychology is that language comprehension should be understood as the construction of coherent situation models. Recent reviews of the literature on situation models can be found in Graesser et al. (1997) and Zwaan and Radvansky (1998). We have proposed a model of situation-model construction, called the Event-Indexing Model (Zwaan 1999a; Zwaan and Radvansky 1998; Zwaan et al. 1995). According to the Event-Indexing Model, comprehenders construct situation models from text by extracting events from the text and integrating these events on five different situational dimensions: time, space, causation, motivation, and protagonist. A recent discussion of the empirical findings directly supporting this model can be

Time in narrative comprehension

found in Zwaan (1999a). In this chapter, we are restricting our focus to the temporal dimension. Unlike any of the other situational dimensions, time is encoded obligatorily in every clause — in English and other languages — in the form of a tense morpheme attached to one or more verbs. Furthermore, temporal information can be conveyed in every word class (Miller and Johnson-Laird 1976). The question that concerns us here is how comprehenders make use of these linguistic cues to construct the temporal dimension of situation models. In order to begin to address this question, we first need to consider the mechanisms of situation-model construction. Our general hypothesis is that the mechanisms involved in situationmodel construction from language are derived from the mechanisms of situationmodel construction in the real world. Children first learn to understand basic events, such as the movement of a person or an object in the environment, before they learn to understand the phrases that describe these events. In fact, these basic event structures are the scaffolding that allow children to learn languages at such an amazing pace (Goldberg 1999). Thus, the construction of situation models during language comprehension can be conceived of as the vicarious experiencing of events in the real world (Segal 1995). Zwaan (1999b) shows that an empirical case can be made for this argument. This notion of vicariously experiencing through language provides the starting point for our analysis.

2.

Time in language

The representation of time in language has historically received a great deal of attention in the field of linguistics (e.g. Dowty 1986; Reichenbach 1947; Ter Meulen 1995; Vendler 1967; Verkuyl 1972). More recent approaches in the subfield of cognitive linguistics are particularly relevant to our perspective (e.g. Croft 1998; Goldberg 1999; Langacker 1982). Linguists distinguish among semantic classes of events, such as the movement of an agent, the movement of an object to a new location by an agent, or the transfer of possession of an object. These event classes have meaning in and of themselves and can accommodate just about any verb or modifier in any construction, as long as it is conceptually interpretable. For instance, some verbs that seem generally unacceptable in the present progressive (understand) can assume this construction given the right circumstances, for instance as in I am understanding temporal relations more and more each day. Representing time in language is the comprehender’s attempt to translate words and sentences into a flow of events comparable to normal perceptual experience. The semantic classes of events are the building blocks we use to construct situation models during comprehension.

73

74

Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

Comprehenders assume that the order in which events are reported in language matches their chronological order. This is known as the iconicity assumption (Dowty 1986; Fleischman 1990). Narrative deviations from chronological order are possible only because a default order exists, the default serving as a baseline from which all else can be compared and understood. When the iconicity assumption is not a valid guide, readers must use language cues to determine the order of events. Verb tense and time adverbs and adverbials (for instance a week later) guide the comprehender as to where an event should be located on the time line. Verb aspect is the language cue that guides the classification of events with respect to duration and completion status. A perhaps overly simplistic way to put it is to say that verb tense and time prepositions and adverbials tell the comprehender where to place the beginning of the event in the narrative world and verb aspect tells the comprehender whether the event should be ended before the next one starts or whether it should continue and thus overlap in time with the next event. 2.1 Verb tense, time adverbs, and time adverbials Verb tense is used to indicate the moment of occurrence of an event relative to the moment of its description (Reichenbach 1947). Together with the iconicity assumption this allows the comprehender to sequence events into their proper chronological order. Time adverbs and adverbials can be used to override the default order. For example, in (1), the time adverb after is a cue that the events are reported in reversed chronological order. And in (2), the time adverbial two weeks later indicates that there is a two-week gap between the two events on the timeline. (1) The anchor frowned after he started reading from the teleprompter. (2) The professor sent the package to his mother in Europe. Two weeks later it arrived.

Figure 1 depicts possible event orders on a timeline. An event may immediately succeed another event (a), or immediately precede another event (b). An event may succeed another event after a certain interval (c), or it may even start before the other event is completed (d). Also, an event may occur during another event (e). Comprehenders not only have to re-order described events, but also monitor for jumps in the narrative timeline, such as flashbacks and flashforwards, and lapses in event descriptions. A lapse in event description may indicate that no events occurred in the time between the reported events, that events did occur during this time but were not relevant to the narrative, or that the former event continued up until the latter event. In any case, whenever the iconicity assumption is overridden, the comprehender must employ temporal markers within the situation model to account for temporal order.

Time in narrative comprehension

Time (d)

(a) Event 1

Event 2

Event 1 (e)

(b) Event 2

Event 2 Event 2

Event 1 Event 1

(c) (f) Event 1

Event 2

Event 1

Figure 1.Potential temporal relations between pairs of events

2.2 Verb aspect Verb aspect provides information about duration, onset, and completion status of an event (Ter Meulen 1995; Vendler 1967; Verkuyl 1972). An event may convey a perfective aspect, meaning that it is completed on the narrative timeline (for instance He walked to the store.). An event may alternatively convey an imperfective aspect, meaning that it is ongoing on the narrative timeline (for instance He was walking to the store.). The duration of an event may be short (He arrived at the party) or long (He was dancing at the party). An event that is instantaneous is referred to as a plug (Ter Meulen 1995), or achievement (Vendler 1967). The onset of the plug event is also the endpoint. Arrive is instantaneous in that it merely describes the change of state from not being at a party to being at a party. An event that has unlimited duration and has a unitary description that applies throughout its entire internal structure is called a hole (Ter Meulen 1995), or process (Vendler 1967). Drive around is an example of a hole, because it has unlimited duration and driving around describes every instance within the duration of the event. An event that has a limited duration and has a description that does not apply to any part of the event is called a filter (Ter Meulen 1995), or accomplishment (Vendler 1967). An example of a filter is drive home, because there is a limited duration, and the description driving home does not apply to any part of the whole event. By this we mean that the event part turning left on Broad Street is not, in and of itself, driving home. In contrast, the example drive around is a hole, so the event part turning left on Broad Street is, in fact, driving around. Plugs and filters are telic, meaning that they have a natural endpoint or resulting state, whereas a hole is atelic, meaning that it has no certain endpoint. Driving home (or arriving at home) is a telic event because there is a natural

75

76

Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

endpoint: one reaches one’s home. If we were to say He drove west, this would indicate an atelic event because there is no telling how long or far the protagonist may drive. There is no natural endpoint to west. It is not only the nature of an event that determines whether it is a hole, plug, or filter. Tense also influences our categorization. Take our example of driving home. If we say He drove home then we have a filter, but if we say He was driving home then we have a hole. From our current perspective of time (point of narration) there is no end to the driving home. If we say He started driving home then we have made the event a plug. Starting the process of driving is an instantaneous event, going from the state of not driving to the state of driving. Thus, interplay between tense and durative nature of the event make the aspectual classification of holes, plugs, and filters quite dynamic. In our notation in Figure 1, variations in verb aspect are indicated by changing between endbars, which signify onsets and endings of telic events, and arrows, which signify atelic events (shown in f). It is important to note that the traditional aspectual class of states is also covered in Ter Meulen’s taxonomy. For example, descriptions of transitions between temporary states are analyzed as plugs (Ter Meulen 1995: 11). 2.3 Perceptual aspects of time in language When discussing the representation of time in language, it is relevant to consider the actual mental representations of events. These are the building blocks of our situation models for what we read and what we hear. Many mental representations have an inherent time component, making them dynamic representations. The very essence of some expressions cannot be understood without highlighting a change over time within the representation. An example of this type of representation is crossing a river. Our mental representation of this expression must incorporate a river with two defined sides. Understanding the word cross dictates that we track the spatial evolution over time of some target, or trajector (Langacker 1986), which starts at one side and ends up at the other side within our representation. Because the trajector cannot be represented at both sides at one time, the very nature of the expressions mandates a dynamic representation. That is, the time component is necessary to the representation (see Figure 2). Langacker (1986) describes the representation that must be formed to understand the word gone. In this representation, the final spatial position of a trajector in a temporal series is highlighted or ‘profiled’. To understand the word gone, we must represent a closer position at an earlier time (see Figure 3). When all positions of the trajector are profiled, we have the representation for go. When only the final position is profiled, we have gone. The understanding of these two expressions is dependent on the time component along the x-axes of both of these representations. Tense, aspect, and time adverbs also affect the comprehender’s perspective on the described events. For example, (written) narratives usually employ the past

Time in narrative comprehension

trajector

Landmark (river)

Time

Figure 2.A dynamic representation of crossing a river. Trajector

Landmark Time

Figure 3.A dynamic representation of gone, from Langacker (1986).

tense in relating information, giving the comprehender the impression that the narrator is describing events that have already happened. Sometimes, however, the narrator will describe events using the historical present. The historical present has the effect of transporting the narrator into the past, such that the act of telling and the unfolding of the event coincide. This violation of the default timeline has the effect, according to Fleischman (1990), of making the ‘past more vivid.’ In our terminology, it may enhance the comprehender’s vicarious experience of the described events. Aspect, tense, and temporal adverbs demonstrate their most powerful, and most subtle, effects by determining how deeply the comprehender vicariously experiences the situation. Consider the differences between the perfective and the imperfective aspect. The perfective is generally used to denote telic situations, whereas the imperfective denotes atelic situations. This is seen as the difference in placing the reader outside the situation (perfective) versus placing the reader inside the situation (imperfective). In many cases, individual events have pre-packaged durations associated with them. Snowstorms, for example, typically have a duration measured in hours or days, whereas gunshots are nearly instantaneous. However, aspect can be used to

77

78

Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

override these default durative features (Fleischman 1990). Consider the difference between the sentences The snowstorm dumps a ton of snow on the ground, versus The snowstorm is dumping a ton of snow on the ground. The use of the perfective in the first case serves to give an atelic situation a terminal boundary, collapsing a lengthy blizzard into an instantaneous event. It is perhaps instructive to consider this fact in terms of a vicarious experiencer’s (Zwaan 1999b) focus and duration of attention. Dowty (1986) suggests that the temporal fashion in which events are described denotes the time it takes for an observer to perceive them. By extending telic events and collapsing atelic ones the author forces the reader to integrate the perceptual information available in the same fashion as the protagonist. This not only serves to bring the reader deeper into the story, but deeper into the protagonist’s psyche as well. This phenomenon is analogous to that of a movie camera that either floats behind the protagonist’s eyes (thus the audience sees what the protagonist sees) or stands mounted at a distance with a wide-angle view of the entire situation (thus the audience sees the protagonist and everything else around her). Another perspectival function that tense and aspect serve is to foreground or background certain items within events. Within any story some information will be more relevant than other information, and it is up to the author to let the reader know what that information is. If before we were speaking about a camera whose focus was either individualized or all-encompassing, we speak now of a camera that either zooms in on certain objects or merely pans across the panorama. Generally present-tense perfective verbs equate with foregrounded information while nonpresent-tense imperfective verbs equate with backgrounded information, though there are exceptions (see Fleischman 1990). Thus one of the goals of narrative is to not only bound experience into units of time but into units of importance as well, with the objective of making the totality of experience accessible to the reader.

3.

Time in comprehension

In their recent review of the extant research on situation models in language comprehension and memory, Zwaan and Radvansky (1998) concluded that the temporal dimension of situations has been relatively underinvestigated in cognitive psychology. In this section, we will provide a more detailed review of this literature, as well as some more recent work coming out of our lab and other labs, within the framework we developed in the previous sections. Specifically, we will consider studies that have investigated the role of temporal order, verb aspect, and temporal perspective.

Time in narrative comprehension

3.1 Effects on on-line comprehension Various researchers have examined the effects of violations of the iconicity assumption on comprehension. Clark (1971) examined the acquisition of the meaning of before and after by children between 3 and 5 years of age. She assessed the children’s comprehension of sentences such as (3a–d). (3) a. b. c. d.

The boy kicked the rock before he patted the dog. Before the boy kicked the rock, he patted the dog. The boy kicked the rock after he patted the dog. After the boy kicked the rock, he patted the dog.

The children, and especially the youngest ones, had by far the most difficulty with sentences of types (b) and (c). These are the sentences in which the events are reported in reversed chronological order. Clark’s explanation for this finding is that the youngest children used an “order-of-mention” strategy. This supports the idea that comprehenders by default entertain the iconicity assumption. Of course, children eventually learn the meanings of the words before and after and thus they learn that these can sometimes override the iconicity assumption. But if children learn the meaning of these time adverbs, then what evidence is there that adult comprehenders entertain the iconicity assumption? The idea of a default iconicity assumption does not mean that adult comprehenders are permanently stumped by a deviation from chronological order. What it suggests is that deviations from chronological order are relatively difficult — but not impossible — to process because a default assumption has to be overridden. Consistent with this idea, Mandler (1986) found that adults had longer reading times for sentences that violated chronological order (analogous to sentences (3b) and (3c) above) than for their chronological counterparts, except when there was a very strong causal relationship between the events. More recently, Münte et al. (1998) obtained neuroscientific evidence that supports this prediction. Münte et al. measured modulations of electrical activity in the brain as participants read sentences such as (3b) and (3d) above. These event-related brain potential (ERP) measurements indicate that “before” sentences elicit, within 300 ms, greater negativity than “after” sentences. This difference in potential is primarily located in the left-anterior part of the brain and is indicative of greater cognitive effort. In short, there is developmental, cognitive, and neurocognitive evidence for the role of an iconicity assumption in language comprehension. Comprehenders’ natural tendency is to process narrated events just as they would naturally occurring events: in chronological order. However, they have developed a sensitivity to linguistic cues that occasionally call for this assumption to be overridden. In general terms, this conclusion is consistent with the idea that our ability to construct situation models during narrative comprehension makes use of our previously developed (both phylogenetically and ontogenetically) ability to comprehend events.

79

80

Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

Another way in which language allows us to deviate from everyday experience is by (forward) time shifts (see (c) in Figure 1). Zwaan (1996) has found that such time shifts (for instance An hour later) lead to increases in reading time relative to phrases that introduce no major time shift (A moment later). Zwaan argued that this supports the idea of a ‘strong-iconicity assumption’, according to which comprehenders not only assume that events occur in chronological order, but also that they occur contiguously. There is some linguistic evidence for a strong-iconicity assumption. Grimes (1975: 36) notes that in Kâte, a language of Papua New Guinea, events that are contiguous in time are grammatically distinguished from events “that are separated by a lapse in which nothing of significance for the story happens.” Although languages such as English and Dutch have not grammaticalized this distinction, it is plausible that time lapses in stories have psychological significance in English as well. The strong-iconicity assumption has to be overridden when there is a time shift and this creates an increase in reading times. 3.2 The activation level of events during comprehension As mentioned in the introduction, in real life, events often overlap in time. This is difficult to convey in the linear format of language. This leads to two questions. First, what are the cues used to tell comprehenders that an event is still ongoing? Second, how are comprehenders keeping multiple events active in working memory? Research has begun to address these questions. Anderson et al. (1983) have shown that the comprehender’s background knowledge about the duration of events might play a role. For example, our knowledge about restaurants tells us that a visit to a restaurant typically does not take more than 7 hours. Similarly, we know that restaurants typically employ waiters. Anderson and her colleagues found that probe responses to the word waiter in a story about a restaurant were faster after a relatively short time interval (for instance ten minutes) had elapsed in the story than after a long interval (for instance five hours) had elapsed, presumably, because participants implicitly assumed that at that point the waiter was no longer in the situation, given that ‘narrative now’ was no longer in the restaurant. In contrast, responses to the name of the main protagonist of the stories did not vary as a function of time shift, presumably because the main protagonist was assumed to still be in the current situation. Zwaan (1996) made the more general assumption that events that are currently in the situation are more accessible to the comprehenders than events that are not. Consistent with this idea, he found that probe responses to events and objects were faster when the previously mentioned event was still within the same time frame as the current event than when it was not. Thus, responses to walked were faster after (4a) than after (4b).

Time in narrative comprehension

(4) a. b.

Teresa walked onto the stage. A moment later, she collapsed. Teresa walked onto the stage. An hour later, she collapsed.

Carreiras et al. (1997) obtained evidence that is consistent with this. Zwaan et al. (2000) showed an even more powerful effect. They found that the explicit discontinuation of an event leads to an immediate decrease in its activation. For example, responses to the word kicking were significantly longer after Steve stopped kicking the soccer ball than after Steve was kicking the soccer ball. They also showed that an event can be maintained across other events. For example, in their Experiment 3, they presented participants with sentence pairs such as (5a) or (5b). (5) a. b.

John was playing the piano. When his mother entered, he stopped. John was playing the piano. When his mother entered, he continued.

Thus, there is an action that is interpreted to be ongoing, because of the past progressive tense, but an interrupting event occurs. Zwaan et al. (2000) were interested in whether the comprehender would assume that the action was discontinued because of the interrupting event. This did not appear to be the case. Responses to the first action, for instance playing, were faster in sentences like (5b) than in sentences like (5a), indicating that comprehenders interpreted the initial action as continuing through the intervening event. Bestgen and Vonk (1995) found that temporal markers can indeed have very subtle effects on the availability of preceding information. Specifically, in sentences such as He opened the door, [Ø, and, then] went inside, and … the absence of a temporal marker (Ø) made previous information more available than a sequential marker, such as then. Magliano and Schleich (2000, Experiment 2) recently showed that comprehenders are more likely to consider atelic events to be ongoing in the subsequent context than telic events. Moreover, they found that verb aspect can override the event-related background knowledge that was activated during comprehension. Comprehenders were much less likely to consider events of a typically long duration, such as golfing 18 holes, to be going on in the subsequent context than events of a typically short duration, such as writing a check, when the long events were described as telic and the short events were described as atelic. 3.3 Effects on long-term memory organization Research on autobiographical memory, in which people are probed about events that have happened in their lifetime, reveals effects of temporal organization (e.g. Thompson et al. 1996). That is, people tend to form stronger associations between events that they experienced within the same time frames than to events that they experienced in different time frames. If narrative comprehension is a form of vicarious experiencing, one would expect to find the same effect of time in memory

81

82

Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

for discourse. The assumption that comprehenders’ natural tendency is to associate events in long-term memory that occurred within the same time frame has been tested in various studies. In support of this view, it has been found that disruptions of the temporal order have a negative impact on the coherence of the situation model in long-term memory (Ohtsuka and Brewer 1992). Furthermore, Zwaan (1996) found that there was more priming between events from a story that occurred within the same time frame than between events that occurred during different time frames. For example, there was more priming between walking onto the stage and collapsing in (4a) than in (4b). These results suggest that the event in the current model will be attached to events in the integrated model that are within the same general time frame. Further support for the role of time as a basis for organizing information in long-term memory was obtained by Radvansky et al. (1998), who used a memory task, rather than a language-comprehension task. 3.4 Effects of verb aspect There is very little empirical work on the representation of time within events. However, we have begun to address this question in our lab (Zwaan and Stanfield, submitted). The main finding of the experiments reported in this paper is that the internal temporal structure of an event affects the amount of relevant information in the situation model that is activated to represent this event. To concretize this finding, consider the following story: (6)

Bobby took out a [hammer/screwdriver], but then remembered that he lost his [hammer/screwdriver]. He also collected the lumber and paint he had bought. He had already selected an oak tree as the site for the birdhouse. He marked the boards and cut them out. Then Bobby began pounding/pounded the boards together.

There are two critical manipulations in this story, the first of which occurs in the first sentence. There are three conditions. A target instrument, in this case the hammer, is or is not available for use in the situation, or it is not mentioned at all. The second critical manipulation occurs in the last sentence. Here, an action is described either as telic (pounded) or as atelic (began pounding). The action that is being described typically involves the instrument mentioned in the first sentence. However, that instrument either is or is not available for use, depending on the situation. As our review of the linguistic literature suggests, atelic descriptions are hypothesized to make more information about an event available than telic descriptions. Zwaan and Stanfield (submitted) reasoned that there are two ways in which verb aspect may affect the activation of knowledge pertaining to an action: semantic activation and episodic activation. If atelic descriptions make semantic information

Time in narrative comprehension

more available, then information associated with the action in semantic memory should become more available. For example, hammer should be more available after pounding than after pounded when the instrument was not previously mentioned in the story. Given that the hammer had not been mentioned previously, any activation of hammer would have to be due to the semantic association between pounding and hammer. If atelic descriptions make episodic information more available, then the episodic information, which is encoded in the integrated situation model, that the instrument is or is not available for use should play a role. Zwaan and Stanfield (submitted) found support for both the semantic and the episodic activation hypothesis, using a variety of methods. For example, recognition responses to hammer were significantly longer after atelic descriptions than after telic descriptions when the instrument was not available for use but had been mentioned in the text previously. This suggests that atelic descriptions were more likely than telic descriptions to activate the information in the situation model that the instrument was not available. This knowledge presumably interfered with responses to hammer, which, in turn, caused the longer response times. When the instrument word had not been mentioned in the text before — and thus the correct recognition response was NO — responses were longer after atelic than after telic descriptions. This suggests that the telic description had activated the instrument concept in semantic memory, making it more difficult for the subjects to reject the corresponding word.

4. Conclusion The role of temporal information in narratives has been studied by researchers in various fields, beginning with Aristotle. Linguists have provided detailed analyses of how the flow of events is conveyed in sentences and in narratives. Cognitive psychologists have begun to test some of these and other ideas empirically, mainly in the context of situation models. There clearly is a great deal of support for the idea that comprehenders entertain a default iconicity assumption. There is also evidence for a strong-iconicity assumption. Furthermore, assuming that the comprehender has a virtual presence in the narrated situation explains a great number of empirical findings. Protagonists, objects, and events that are in the situation are more accessible to the comprehender than protagonists, objects, and events that are not. However, it is also clear that a great deal still has to be learned. For example, we are only beginning to understand the role of aspect in language comprehension. Furthermore, we know very little about the role of temporal cues in the establishment

83

84

Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

of perspective. For that matter, we know very little about the role of perspective in language comprehension in general, although psychologists are beginning to address this issue (MacWhinney 1999; Sanford et al. 1998). As far as our own research agenda is concerned, we are interested in examining more closely what it means to vicariously experience events by way of reading or listening to narratives (Zwaan 1999b) and what the role of linguistic cues, such as temporal markers, is in this. There are theoretical and methodological developments that allow us to better address these questions. On the theoretical front, proposals have been made that our mental representations of events are not arbitrary and amodal, but do preserve aspects of the initial perceptual input (Barsalou 1999). Comprehension, in this view, is regarded as the mental simulation of events, using perceptual symbols (this does not just include visual information, but information from the other modalities as well). Although there currently is not a great deal of empirical evidence supporting the idea of perceptual symbol systems, it is clear that they provide a much better conceptual tool to think about situationmodel construction as vicarious experiencing than amodal symbol systems (Zwaan 1999c; Zwaan et al. 1999; Stanfield and Zwaan, in press). On the methodological front, there are currently methods available to study the human brain as it is performing cognitive tasks. Two of these methods, with complementary strengths, are ERP and functional magnetic-resonance imaging (fMRI). As we discussed, ERP provides information on the changes in brain potential as a function of the presentation of stimuli. ERP has high temporal resolution (on the order of milliseconds), but poor spatial resolution (on the order of centimeters). Thus, ERP provides a good window into when processes occur in the brain, but does not provide much detail on where these processes occur (primarily because electrical activity in the brain has to be measured through the skull). fMRI is used to record changes in blood flow in the brain. The idea is that those areas of the brain that are currently active will need more oxygen and there will thus be an increased blood flow towards those regions. In contrast to ERP, fMRI has relatively low temporal resolution (on the order of seconds), but compared to ERP very high spatial resolution (on the order of millimeters). ERP has already been used to study temporal information processing in language in the study by Münte and colleagues we mentioned before. We are currently exploring the use of fMRI in studying temporal information processing. We anticipate that the theoretical framework of perceptual symbol systems and related approaches (Barsalou 1999; Glenberg 1997; MacWhinney 1999), which is congenial with current work in cognitive linguistics (e.g. Croft 1998; Goldberg 1999; Langacker 1986) and the methodological tools of cognitive neuroscience will lead to major advances in our understanding of how language allows us to vicariously experience events. Theoretically, a meta-language is being developed that allows us to integrate relevant concepts and findings from cognitive psychology, linguistics,

Time in narrative comprehension

and neuroscience. Methodologically, we will be able to look into the human brain as it processes linguistic and other information. This will provide more insight into which processes are specifically related to language and which are more general mechanisms of event comprehension. Ideally, these anticipated advances will also increase our understanding of the aesthetic, emotional, and moral effects created by literary violations of temporal information processing, such as those in the epigraph to this chapter.

References Amis, M. 1992. Time’s Arrow: Or the Nature of the Offense. New York: Vintage. Anderson, A., Garrod, S. C. and Sanford, A. J. 1983. “The accessibility of pronominal antecedents as a function of episode shifts in narrative text”. Quarterly Journal of Experimental Psychology 35A: 427–440. Barsalou, L. W. 1999. “Perceptual symbol systems”. Behavioral & Brain Sciences 22: 577–660. Bestgen, Y. and Vonk, W. 1995. “The role of temporal segmentation markers in discourse processing”. Discourse Processes 19: 385–440. Carreiras, M., Carriedo, N., Alonso, M.A. and Fernandez, A. 1997. “The role of verbal tense and verbal aspect in the foregrounding of information in reading”. Memory & Cognition 23:438–446. Clark, E. V. 1971. “On the acquisition of the meaning of Before and After”. Journal of Verbal Learning and Verbal Behavior 10: 266–275. Croft, W. 1998. “The structure of events and the structure of language”. In The New Psychology of language, M. Tomasello (ed.), 67–92. Mahwah, NJ: Lawrence Erlbaum Associates. Dowty, D. R. 1986. “The effects of aspectual class on the the temporal structure of discourse: Semantics or pragmatics?” Linguistics and Philosophy 9: 37–61. Fleischman, S. 1990. Tense and Narrativity. Austin, TX: UT Press. Glenberg, A. M. 1997. “What memory is for”. Behavioral and Brain Sciences 20: 1–19. Goldberg, A. 1999. “The emergence of the semantics of argument structure constructions”. In The Emergence of Language, MacWhinney, B. (ed.), 197–212. Mahwah, NJ: Lawrence Erlbaum Associates. Graesser, A. C., Millis, K. K. and Zwaan, R. A. 1997. “Discourse comprehension”. Annual Review of Psychology 48: 163–189. Grimes, J. 1975. The Thread of Discourse. The Hague: Mouton. Hockett, C. F. 1960. “The origin of speech”. Scientific American 203: 88–96. Langacker, R. W. 1986. “An introduction to cognitive grammar”. Cognitive Science 10: 1–40. Ohtsuka, K. and Brewer, W. F. 1992. “Discourse organization in the comprehension of temporal order in narrative texts”. Discourse Processes 15: 317–336. MacWhinney, B. 1999. “The emergence of language from embodiment”. In The Emergence of Language, MacWhinney, B. (ed.), 213–256. Mahwah, NJ: Lawrence Erlbaum Associates. Magliano, J. P. and Schleich, M. C. (2000). “Verb aspect and situation models”. Discourse Processes 29: 83–112. Mandler, J.M. 1986. “On the comprehension of temporal order”. Language and Cognitive Processes 1: 309–320. Miller, G. A. and Johnson-Laird, P. N. 1976. Language and Perception. Cambridge, MA: Harvard University Press.

85



86

Rolf A. Zwaan, Carol J. Madden and Robert A. Stanfield

Münte, T. F., Schiltz, K. and Kutas, M. 1998. “When temporal terms belie conceptual order”. Nature 395: 71–73. Ohtsuka, K. and Brewer, W. F. 1992. “Discourse organization in the comprehension of temporal order in narrative texts”. Discourse Processes 15: 317–336. Radvansky, G. A., Zwaan, R. A., Franklin, N. and Federico, T. 1998. “Retrieval from temporally organized situation models”. Journal of Experimental Psychology: Learning, Memory, and Cognition 24: 1224–1237. Reichenbach, H. 1947. Elements of Symbolic Logic. London, U. K.: Collier-Macmillan. Sanford, A. J., Clegg, M. and Majid, A. 1998. “The influence of types of character on processing background information in narrative discourse”. Memory & Cognition 26: 1323–1329. Segal, E. M. 1995. “Cognitive-phenomenological theory of fictional narrative”. In Deixis in Narrative: A Cognitive Science Perspective, Duchan, J. F., Bruder, G. A. and Hewitt, L. E. (eds.), 61–78. Hillsdale, NJ: Erlbaum. Stanfield, R.A. and Zwaan, R.A. (in press). “The effect of implied orientation derived from verbal context on picture orientation”. Psychological Science. Ter Meulen, A. G. B. 1995. Representing Time in Natural Language: The Dynamic Interpretation of Tense and Aspect. Cambridge, MA: MIT Press. Thompson, C. P., Skrowonski, J. J., Larsen, S. F. and Betz, A. L. 1996. Autobiographical Memory: Remembering What and Remembering When. Mahwah, NJ: Erlbaum. Van Dijk, T. A. and Kintsch, W. 1983. Strategies of Discourse Comprehension. New York: Academic Press. Vendler, Z. 1967. Linguistics in Philosophy. Ithaca, NY: Cornell University Press. Verkuyl, H. J. 1972. On the Compositional Nature of the Aspects. Dordrecht, the Netherlands: Reidel. Zwaan, R. A. 1999a. “Five dimensions of situation-model construction”. In Narrative Comprehension, Causality, and Coherence: Essays in Honor of Tom Trabasso, S. R. Goldman, A. C. Graesser and P. van den Broek (eds), 93–110. Mahwah, NJ: Erlbaum. Zwaan, R.A. 1999b. “Situation models: The mental leap into imagined worlds”. Current Directions in Psychological Science 8: 15–18. Zwaan, R. A. 1999c. “Embodied cognition, perceptual symbols, and situation models”. Discourse Processes 28: 81–88. Zwaan, R.A. 1996. “Processing narrative time shifts”. Journal of Experimental Psychology: Learning, Memory, and Cognition 22: 1196–1207. Zwaan, R. A., Langston, M. C. and Graesser, A. C. 1995. “The construction of situation models in narrative comprehension: An event-indexing model”. Psychological Science 6: 292–297. Zwaan, R. A., Madden, C. J. and Whitten, S. N. (2000). “The presence of an event in the narrated situation affects its activation”. Memory & Cognition 28: 1022–1028. Zwaan, R. A. and Radvansky, G. A. 1998. “Situation models in language comprehension and memory”. Psychological Bulletin 123: 162–185. Zwaan. R. A. and Stanfield, R. A. (submitted). Modulating the flow of situational information during comprehension. Zwaan, R. A., Stanfield, R. A. and Madden, C. J. 1999. “Perceptual symbols in language comprehension: Can an empirical case be made?” Behavioral and Brain Sciences 22: 636–637.



Old readers Slow readers or expert readers? Katinka Dijkstra The Florida State University, Tallahassee

The expansion of cognitive aging research over the past two decades has contributed to deeper insights in the performance of older adults on cognitive tasks. Studies have demonstrated that storage and processing aspects of working memory are affected by age (cf. Stine et al. 1995, 1990; Salthouse 1991, 1992). Additionally, results of other studies indicate an age-related decline in the inhibition of taskirrelevant information that occurs during the selection of task-relevant information (Connelly et al. 1991). Notwithstanding the important contributions these studies have made for our understanding of cognitive aging processes, they have also polarized our view about the populations involved. Younger adults, usually undergraduate students participating for course credit, have proved to be faster, have better memory, speed/ accuracy, and processing skills than their older counterparts, generally community dwelling older adults participating for no or a small fee. What is inaccurate about this portrayal of younger and older adults is that it reflects a certain selection of populations in research (undergraduate students versus senior citizens) rather than a decline in specific cognitive abilities of persons across the life span. It also seems that the performance of the younger population is the standard against which the performance of the older population is measured where deviations from this standard performance are interpreted as deficits (Adams et al. 1990). To avoid an inadequate polarization of younger adults’ performance versus that of older adults, research should be conducted that takes maintained comprehension skills, compensatory strategies, and expertise of older adults into account (Shimamura et al. 1995; Zabrucky and Moore 1994), as well as the performance of cognitive tasks across the life span (Schaie 1989, 1990). Cognitive development proceeds at different rates and at different times (Harker et al. 1982). It depends on the focus of research whether this development is regarded from a perspective of decline or selective optimization.

88

Katinka Dijkstra

Many studies have demonstrated that performance on text recall tasks is impaired among older readers compared to younger readers (Stine et al. 1995; Dixon et al. 1984). However, in other performance measures, such as forming a situation model of a text, or generating central ideas from a text, older readers perform as well as younger readers (Dijkstra et al. 2000; Radvansky 1999; Jackson and Kemper 1993). Including adaptive and compensating processes in research on text comprehension would not only contribute to a more complete understanding of cognitive aging, but would also provide a better reflection of the plasticity of cognitive abilities across the life span. This chapter reviews the literature on the topic of text comprehension among older readers compared to younger readers. First, I will discuss studies and hypotheses on age-related deficits in text comprehension, in particular the limited working memory capacity hypothesis and the inhibitory control hypothesis. Then, I will discuss text comprehension skills that are maintained in adult life, as well as the role of compensating strategies in reading narratives. Finally, I will attempt to integrate results of research on age-related declines and maintenance of text comprehension skills and propose ideas for studies on literary text comprehension among older and younger readers.

1.

Age-related deficits in text processing

The reading process is a constant evaluation and regulation of processed information. According to Kintsch and Van Dijk (1978), discourse processing involves the extraction of propositions, their establishment of relationships, and the integration of propositions into coherent representations of the discourse meaning. Reading strategies, rereading, and conducting reinstatement searches help the reader to understand the information thoroughly (Zabrucky and Moore 1994; Kintsch and Van Dijk 1978). To fully comprehend a text new propositions should be encoded and integrated with old propositions. This is achieved by using the storage and processing capacities of working memory (Kintsch and Van Dijk 1978). Deficiencies in the evaluation and resolution of the comprehension process or deficiencies in the correct integration of information in working memory due to, for instance, limited storage capacity will most likely result in an inadequate memory representation of the text (Zabrucky and Moore 1995). Control operations in working memory operate as a filter, responsible for the excitation of relevant information and the inhibition of irrelevant or no longer relevant information (Gernsbacher 1990; Hasher and Zacks 1988; Kintsch 1988). Working memory holds representations of linguistic units in short term memory long enough for the reader to achieve meaning and coherence between preceding and incoming information. This capacity to optimally balance incoming information with preceding information and filtering out irrelevant information is supposedly reduced in older

Old readers

adults (Van der Linden et al. 1999). Older adults’ more limited working memory capacity increases the likelihood that recently processed information will be forgotten, and hence fails to be encoded into a discourse presentation. Additionally, older readers appear to be less efficient than younger readers at inhibiting nolonger-relevant information during information processing. This less efficient inhibition will then result in a working memory trace that still includes off-goal information which makes occurrence of deficits in comprehension and retrieval more likely (Gerard et al. 1991; Gernsbacher 1989). Both assumptions, the working memory capacity hypothesis and inhibitory efficiency hypothesis, have been studied extensively in cognitive aging and discourse research (Kwong See and Ryan 1996; Ska and Joanette 1996; Van der Linden et al. 1999). These studies suggest that both limited working memory capacity and less efficient inhibitory control among older readers can explain differences in their performance on text comprehension tasks. 1.1 The working memory capacity discourse hypothesis The “Working memory capacity discourse hypothesis” (Baddely 1986; Stine 1990; Van der Linden et al. 1999) postulates that age differences in the processing aspects and storage aspects of working memory lead to performance differences in language. The older a person is, the more limited the amount of cognitive resources are to perform mental operations on incoming or recently processed information. Especially as tasks become more complex, greater demands are placed on the limited processing resources of working memory (Salthouse 1992). Rybash and colleagues (1986) demonstrated that abilities to process complex information under a certain time deadline decline during the older adult years. Spilich (1983) found empirical evidence for the assumption that with advanced age a decrease in the working memory capacity occurs. This limitation results in a decrease of the availability of reinstatement and reprocessing of propositions in a cycle which makes elaborations and interconnections among idea units in a text more difficult. Stine and colleagues (1995) showed that younger adults constructed a coherent representation of a text by allocating more time to the task when new information came up and at significant boundary sites in the text (end of sentence). Older adults, on the other hand, did not allocate resources of working memory to the processing of the text at boundary sites to integrate new information. Instead, in this particular study, they were faster than younger adults constructing a coherent representation. However, older adults who showed high levels of text recall were readers who paused more frequently than young adults to organize new information. This result was confirmed in another study (Miller and Stine-Morrow 1998). It seems that as a result of the more limited working memory capacity among older readers, they are more likely to forget some of the processed information and

89

90

Katinka Dijkstra

fail to encode it into a text representation. This is particularly true in cases where older readers do not increase processing times when text features (new information, more information) seem to require this (Stine et al. 1995; Zabrucky and Moore 1994, 1995). The limited working memory capacity hypothesis focuses on the recency of the processed information with retrieval problems arising as a result of inadequate encoding of the processed information. The inhibitory efficiency discourse hypothesis focuses on the relevancy of this information and how this affects adequate encoding of information in and retrieval from long term memory (Kwong See and Ryan 1995; Van der Linden et al. 1999). 1.2 Inhibitory efficiency discourse hypothesis The ‘Inhibitory efficiency discourse hypothesis’ (Kwong See and Ryan 1999; Hasher and Zacks 1988; Van der Linden et al. 1999) addresses the control processes of working memory. It postulates that the basis of age differences in language performance and discourse processing are differences in the ability to inhibit irrelevant information from gaining entrance to working memory during the processing of target information. Normally, inhibitory mechanisms are hypothesized to prevent objectively irrelevant thoughts from gaining entrance to working memory during the development of the discourse presentation (Gernsbacher 1990). With increased age, however, this inhibition becomes less efficient, which makes it harder for older adults to deactivate contextually related but less relevant or no-longer relevant information. Hamm and Hasher (1992) found that older adults were less likely than younger adults to inhibit initial inferences in a text that were subsequently made untenable. In another study, Connelly and colleagues (1991) reported that older adults had slower reading times when reading a text that had distracting text (in different font) interspersed than younger adults. This was especially true when the distracting text was semantically related to the target text. Younger readers were not as distracted by this phenomenon as older readers. According to the inhibitory control hypothesis, older readers’ deficiencies in suppressing irrelevant information would appear both during the encoding phase and retrieval phase of text processing into or from long term memory (Van der Linden et al. 1999). Gerard and colleagues (1991) showed that older readers showed a larger ‘fan effect’, that is, activation of relevant as well as related information, than younger readers. There is empirical evidence for both hypotheses suggesting that working memory and inhibition capability have an effect on text comprehension skills among older readers. There has been some discussion, however, with regard to the question of which hypothesis provides the best explanation for age-related differences

Old readers

in text comprehension. Although researchers agree that the phenomena described by the hypotheses are interrelated (Kwong See and Ryan 1996; Van der Linden et al. 1999), recent studies have been conducted to compare the contribution of these concepts in explaining age-related deficits. Kwong See and Ryan (1996) used hierarchical regression analyses to assess the relative contributions of the limited working memory capacity hypothesis and the inhibitory control hypothesis to performance data on text comprehension (Kwong See and Ryan 1995). Their line of reasoning was that if age differences in the working memory component mediate the relationship between older age and poorer language performance, statistically partialling out the influence of age differences in the working memory component would attenuate the variance in language performance associated with age. When processing speed was entered first into the regression equation, the mediating influence of the inhibition and working memory measures remained significant. However, when processing speed and differences in inhibition were controlled, the working memory measures no longer predicted text comprehension, indicating higher explanatory power of the inhibitory control hypothesis than the limited working memory capacity hypothesis. This conclusion is disputed, however, by Van der Linden and colleagues, who compared measures of processing speed, working memory, and inhibitory control by using structural equation modeling techniques with different cognitive tasks (1999). Their best-fit model postulated that the effect of age on language comprehension and verbal memory was mediated by speed, inhibitory control, and working memory. In other words, the results confirm the validity of a general model in which both working memory and inhibitory control equally contribute to age-related differences in cognitive performance across a wide range of tasks. In summary, there is still some debate as to what extent limited working memory capacity and less efficient inhibitory control are contributing factors in older adults’ performance on text comprehension tasks. According to results of research on working memory capacity deficits and inhibitory control, both factors affect this performance. Age related deficits in text comprehension appear to be the strongest when a substantial amount of new and complex information has to be processed. Insufficient inhibitory control occurs when interfering information has to be processed along with task-relevant information. A different way of gaining insight in text comprehension processes among older readers is not to focus on age-related declines but on maintained text comprehension skills. From that perspective the debate about which hypothesis best explains age-related declines in text comprehension could be rephrased into how these hypotheses can contribute to explain preserved text comprehension skills among older readers. In the following, I will use that perspective by discussing results of research demonstrating maintenance of text comprehension skills in the domain of situation models, text structure and expert knowledge.

91

92

Katinka Dijkstra

2.

Maintenance of text comprehension skills among older readers

2.1 Situation models There is growing evidence that the construction of the mental representation of narrative texts (the situation model) is similar among older and younger readers. This holds true despite the fact that older and younger readers differ in their recall of the surface structure and text-based information of the text (Adams et al. 1990; Radvansky 1999; Stine-Morrow et al. 1996; Zwaan and Radvansky 1998). Younger readers seem to focus their attention more on the propositions contained in the text (the text base) (Adams et al. 1990; Labouvie-Vief 1985; Radvansky 1999), whereas older readers to a greater extent integrate world knowledge with the text-based information (Adams et al. 1990). This shift from a text-based approach among younger readers towards an integrative approach among older readers may represent an adaptive compensatory mechanism. In cases where task demands supersede available processing resources, it could be a more efficient allocation of limited processing resources to shift the focus of proposition-based text comprehension to establishing a mental representation of the meaning of the text (Adams et al. 1990; Soederberg-Miller and StineMorrow 1998). This does not mean that older readers skip a textbase representation of the text. They merely use this textbase representation to construct a situation model of the text early on (Dijkstra et al. 2000). Results of research in prose comprehension corroborate this assumption. Light (1990) demonstrated that older adults had the ability to integrate text comprehension models, especially when essential information was involved (Light 1990). Presumably, older readers found ways to compensate for slower reading speed by incorporating new information into higher order knowledge structures, such as scripts. Light and Anderson (1985) found that younger and older adults had similar performances in their representation of stored knowledge with respect to the organization of higher knowledge structures contained in scripts. Results from a study by Dijkstra, Charness, and Zwaan (2000) indicated that both older and younger adults constructed similar situation models when reading short stories. However, these situation models were similar only when the stories were available for rereading during the text comprehension task in which readers had to establish relationships between the verbs from the story (cf. Zwaan et al. 1995). When the stories were not available, older readers had weaker situation models than younger readers, especially with regard to establishing causal and temporal relationships between verbs in the stories. A study by Morrow and colleagues (1997) showed that older and younger adults had similar knowledge structures and strategies when they updated their situation model of a narrative when new information was introduced. Both young and old readers slowed their reading time of the narrative as the distance of the

Old readers

target object in the narrative increased. This suggests that readers updated their situation model by following the main character in the story. Older readers slowed down more compared to younger readers when reading sentences that referred to more distant objects. This indicates that older readers make accommodations to their decrease in processing efficiency by allocating more time to read those sentences. Moreover, older readers who more accurately understood the narratives allocated more time for a reinstatement search for the referent object in the spatial layout of the narrative. This extra time presumably yielded a more accurate model that subsequently facilitated answering the comprehension questions. According to a study by Radvansky and colleagues (in press) older adults spent more time on passages in a text that are crucial to the construction of a situation model than younger readers. Their emphasis on higher level processing (at the level of the situation model) can be conceived of as a compensatory strategy for deficiencies at the lower levels (at the level of the surface structure and the text base). The results of the studies discussed above illustrate how two important components of text comprehension converge: integration of information into situation models and focusing on higher order structures in the text. Sufficient allocation of time during on-line text comprehension is crucial for the construction of an adequate situation model, especially for older readers. Older readers who allocate more processing time to update their situation models of the narratives seem to be as able to form an accurate model of the described events as younger adults. Both higher order processing and allocation of sufficient time are compensatory strategies for older readers to help them organizing old and new information in situation models. These strategies allocate limited processing resources there where they can retain the most essential information (Adams et al. 1990). A study by Miller and Stine-Morrow (1998) indicated that older readers of passages in a text were equally likely to form a situation model compared to younger readers, but that they required more time to do so. It also indicated that older readers spent more time at intra-sentence constituents compared to younger readers who primarily used sentence boundaries for organizational processing. Probably, with more limited working memory capacity, older readers need to organize concepts at more frequent intervals than younger readers who can ‘wait’ until the end of the sentence to do so. This explanation supports the assumption of the limited working memory capacity hypothesis that an early representation of the topic could reduce demands on processing resources (Kintsch and Van Dijk 1978). Repeated reading or increased processing time could then be a compensating strategy for older readers in situations where limits of the memory span are exceeded or in situations where a substantial amount of novel information must be processed (Garner 1987). Repeated exposure to the information as a result of rereading or increased exposure to the materials will keep the information activated longer and therefore facilitate the formation of a situation model of the text. Just

93

94

Katinka Dijkstra

and Carpenter’s (1992) CC Reader computational model predicts that maintaining a word or phrase in working memory corresponds to keeping its activation above a certain threshold. Once an element is above threshold, it is available for further interpretive analysis. Repeating or paraphrasing words and phrases, as well as increased exposure to the materials, could lengthen the window needed by older adults to process the necessary information more successfully. The difference between older and younger readers’ text comprehension process lies in the necessity for older readers to compensate for age-related deficits in cognitive function by being more efficient and aware of the resources needed to accomplish the task. 2.2 Text structure Text-based components, such as a coherent structure, can help readers with limited working memory capacity to allocate limited processing resources to those parts in the text that are emphasized in the text structure and therefore seem relevant for the construction of a situation model. Research suggests that older readers are less effective than younger readers in actively organizing materials to enhance recall (Craik et al. 1987; Tun 1989). This implies that older readers need to rely on the text structure to a greater extent in order to adequately organize the information. A study by Tun (1989) showed that both younger and older readers had better recall and comprehension of a narrative text compared to an expository text. Protocols of comprehension of narratives showed greater differentiation between gist and detail than protocols of the expository text. Apparently, the story structure helped the older readers select those ideas that were relevant, thus reducing information load. Meyer and Rice (1988) found that older adults performed better at a comprehension task on an expository text with an emphasis on the structure than with an expository text with an emphasis on the details of the text. This result showed up during the test, and after 1 week, both on free-recall data and with regard to comprehension questions about the text. Miller and Stine-Morrow (1998) demonstrated how the addition of passage titles facilitated text comprehension among younger and older readers when they were reading vague, ill-defined text content. The titles were assumed to provide a retrieval structure that would increase the accessibility of new concepts and a schema through which the reader could connect the concepts more easily. The results indicated that the addition of the titles facilitated recall and resulted in reduced processing time at boundary sites for both younger and older readers. The retrieval structures of the readers who received the titles allowed them to be more efficient in integrating new concepts and organizing constituents within sentences. Knowledge about discourse scripts, text structure, and comprehension goals for particular types of texts facilitates the comprehension of these texts. Characteristics of the discourse, such as a story schema for narratives or stepwise instructions in

Old readers

procedural texts, facilitate connections of relevant propositions. Those schemas, however, have to be constructed from the text by the reader and it depends on the readers’ previous reading experiences whether these schemas can turn into elaborate reading strategies. 2.3 Prior and expert knowledge Prior and expert knowledge about the topic of the text is a third factor in text comprehension processes that may facilitate text comprehension and reflect maintained text comprehension skills among older readers. Comprehension and recall of prose is facilitated when the reader has prior knowledge about the topic of the text (Spilich et al. 1979). The reason for this is that an existing representation of the topic can reduce demands on processing resources. In particular, existing information retrieved from long-term memory during online text comprehension, facilitates the construction of representations of the text (Kintsch and Van Dijk 1978). Older readers have somewhat of an advantage over younger readers with regard to their potential use of prior knowledge, simply because they have had more life experiences. From childhood on, text comprehension evolves over the years into a skill that is a product of a person’s upbringing, specific background knowledge (education and profession), reading experiences and personal experiences (Dijkstra 1994; Zeitz 1994). Readers may use this accumulated knowledge to select and focus on specific information in the text they think is essential for their understanding of the text (Radvansky 1999). Readers with more reading experiences, background knowledge and related personal experiences would be more successful at this than less experienced readers. Results from a study by Soederberg and colleagues suggest that older readers rely to a greater extent on schemata, collections of knowledge about commonly encountered aspects in life, than younger readers by integrating this information into the situation model of the text (Soederberg-Miller and StineMorrow 1998). The impact of prior reading experiences of younger and older readers on text comprehension have not been studied extensively, partly because there are so many individual differences involved. With regard to reading preferences, research has indicated that readers with more literary competence, such as readers who read literature professionally, have a preference for more complex literature and literary authors than readers who do not read literature professionally (Dijkstra 1994, 1996; Kraaykamp and Dijkstra 1999). These professional readers have higher scores on measures of literary knowledge than readers who read for pleasure (Dijkstra 1994, 1996). Professional readers have built up a literary competence that involves both declarative knowledge (knowledge of authors, book titles, genres) and procedural knowledge (knowledge of how to read these books). This expertise originates from incidental learning by reading books and explicit instruction in schools (Dijkstra

95

96

Katinka Dijkstra

1994; Van Dijk and Kintsch 1983). As an elaborate system of reading strategies and specific knowledge, this expertise will help and guide readers how to read a new book and utilize those reading strategies that are suitable to comprehend specific literary texts (Fokkema 1988; Wippler 1990). A study by Zeitz showed that differences in literary experts’ and novices’ performances in the comprehension of literature are related to differences in their representations of the texts (1994). Results indicated similar performance among English graduate students and graduate students in an engineering program on verbatim recall, recognition of plot sentences and generation of factual statements on the experimental texts (a poem, story, and scientific text). However, graduate students in English had better gist recall and memory for multi-interpretable lines in the poem than graduate students in an engineering program (Zeitz 1994). The English students also had a more interpretive comprehension style than the engineering students who did not elaborate as much in their comprehension of the texts. We can expect differences in the comprehension of literature between expert readers and novice readers to continue to grow with increased exposure to literature. It is believed that people accumulate their competence in several domains throughout their lifetime (Kraaykamp and Dijkstra 1999; Mockros 1993). Additionally, reading experiences can contribute to a person’s cultural competence in older age (Knulst and Kraaykamp 1998). We can therefore expect that older readers of literature, who read literary texts for their profession, are better able to integrate their knowledge of literary texts with new texts and have better text comprehension skills than younger expert readers of literature. A study by Shimamura and colleagues showed that reading experts of different ages showed similar recall patterns. Prose recall performance was equivalent for senior, middle-aged, and young professors, whereas age differences were found between younger and older readers in a control condition (no expert readers). Expert knowledge, such as knowing how to summarize or to discern the gist of the passage, probably facilitated the recall task for these experts. Apparently, with their expertise, professors more easily accessed retrieval strategies from long-term memory containing text-relevant knowledge as a result of their daily tasks that require the integration of new knowledge about texts into existing knowledge. For older professors, experience with these strategies may compensate for limitations in their working memory capacity (Shimamura et al. 1995). Another study on literary expertise among students in an English literature course and professors in an English department revealed different strategies between the two groups of experts in analyzing a literary text (Graves and Frederiksen 1996). The students tended to stay closer to the text base of an excerpt from a literary novel. Their reading strategies focused on repeated reading, line by line reading, and paraphrasing the text, whereas the professors focused on higher levels

Old readers

of reading, in particular applying reading strategies to discern the claim what the text was all about. Studies on text comprehension of literary text discussed above suggest that older readers and (older) experts readers maintain and actively use their text comprehension skills. In particular, these readers employ their literary expertise to discern the meaning of a text and compare it to other texts (Graves and Frederiksen 1996). These expert reading and comprehension strategies are another way of allocating processing resources to those parts in the text that contribute to the meaning of the text.

3.

Towards a model of text comprehension in older readers

We have now gathered information both on age-related deficits on text comprehension due to effects of limited working memory capacity and inefficient inhibitory control and maintenance of text comprehension skills among older readers. First, age-related differences in text comprehension are most pronounced at the text-base level (Radvansky et al. in press, Radvansky 1999) and secondly in comprehension tasks that tap heavily into working memory capacity. In the third place, they are pronounced in comprehension tasks in which interfering information, such as interspersed text, results in inclusion of task related but irrelevant information into the representation of the text. Age differences seem to be attenuated or absent in comprehension tasks in which readers construct situation models of the text or are otherwise involved in higher levels of processing, such as establishing the meaning of the text. If readers can allocate sufficient reading and processing time for the comprehension task, if the text structure is organized, and if the reader can utilize reading strategies based on prior exposure to reading those texts, age differences in the performance on comprehension tasks are attenuated even further. To come back to the main questions of this chapter: are older readers slow readers or expert readers, the answer should be affirmative to both. Older readers are slower readers than younger readers because of their more limited working memory capacity and less efficient inhibitory control. They need more time to process new information, incorporate that with information processed earlier and store it in long term memory. Therefore, extra processing time and rereading benefit older readers. First of all, those strategies allow them to encode new information adequately and construct a situation model that includes connections between new and previously processed information (cf. Miller and Stine-Morrow 1998). Secondly, those strategies allow older readers to separate relevant information from irrelevant information and construct an adequate situation model of the text.

97

98

Katinka Dijkstra

However, rather than regarding increased reading time and rereading as agerelated deficits, they should be regarded as compensatory strategies that allow older readers to construct situation models similar to younger readers who are less limited in their working memory capacity and inhibitory control processes. These strategies can be considered part of an ‘expertise’ reading repertoire in which time needed to process information also depends on the type and structure of the text. Allocating processing resources to those parts in the text that are deemed relevant by using the structure of a text as a guiding tool for text comprehension and utilizing prior reading experiences and expert knowledge are other ways for older readers to construct adequate situation models. If they do this successfully, older readers can be considered expert readers who use their strengths in some areas (life experiences, higher level processing) to compensate for weaknesses in others (limited working memory capacity, less efficient inhibitory control). The discussion of age-related factors in text comprehension has been limited to expository and narrative texts. Results from the study by Shimamura and colleagues (1995) suggest, however, that these factors play a role in literary comprehension as well. To my knowledge, no other studies have been conducted in which differences in text comprehension of literary genres are systematically associated with a life span approach. There may be several reasons for this. Except for most poems, literary texts are longer and therefore more difficult to use in experiments than narratives that are constructed for one particular experiment. Furthermore, aging research is a relatively new field that has mostly focused on the performance on cognitive tasks and not yet established a tradition in studying the relationships between aging processes and comprehension of different art forms. The same holds true for the empirical study of literature which has so far focused more on how readers comprehend certain texts, rather than on how certain groups of readers differ from other groups of readers in their text comprehension processes. One of the main reasons that literary texts have not been studied extensively in the context of an aging framework may be that there is too much variety in these texts and how they can be interpreted to make general predictions about older readers’ understanding of them. Studies that have been conducted so far suggest that literary text comprehension is slower, more focused on verbatim information in the text and results in weaker situation models than comprehension of expository texts (Zwaan 1993). Presumably, readers of literature delay their construction of a situation model of the text until the very end so they can wrap up loose ends in the story structure (Zwaan 1993). In addition, literary texts contain information that may interfere with an early construction of a situation model, such as metaphors, shifts in focalization, or streams of consciousness. It is uncertain what effect age would have on the construction of situation models in literary texts. A study by Whitney and Clark indicated that readers of

Old readers

literature with a limited working memory span tended to read in a top down manner by forming thematic inferences which would lead to an early construction of a situation model (1989). However, readers with less working memory capacity used a bottom-up strategy in which they delayed the construction of a situation model. The results would imply that older readers, with their more limited working memory capacity, should be hindered in their construction of a situation model of a literary text. If they choose to construct a mental representation of the text early on, they may risk committing themselves to an interpretation of the story that fails to receive support towards the end of the story. Alternatively, if they choose to delay the construction of a situation model, their working memory capacity may get compromised to the extent that they have difficulty constructing a coherent situation model at all. On the other hand, older readers may have some advantages for literary text comprehension. First of all, results of research demonstrating less efficient inhibitory control mechanisms among older readers, suggest that older readers keep related but seemingly irrelevant information activated to a greater extent than younger readers. This could help them to delay the construction of a mental representation of the text until they can connect all relevant information into a more elaborate coherent representation. Secondly, experienced older readers of literature could use their previous reading experiences and literary expertise to selectively allocate sufficient reading time to process and connect relevant components in the literary text. They can decide which components are most relevant by relating the topic, literary devices, and story structure of the text they are reading to texts they have read in the past. Since their previous reading experiences are stored in long term memory, utilization of these experiences would not put high demands on their working memory capacity. Instead, text comprehension would require making fewer on-line inferences because of the accessibility of the knowledge base to help resolve ambiguities early on in the construction of the situation model (Kintsch and Van Dijk 1978). Moreover, older readers’ tendency to generate and maintain a wider range of inferences (Hamm and Hasher 1992) could help them integrate elements in the literary texts that do not become meaningful until the end of the story. Younger, less experienced, readers, on the other hand, may overlook these elements and fail to include them in their representation of the story. Compensatory strategies, such as increased reading time, and rereading most likely will prove to be beneficial in literary text comprehension too, but it remains uncertain if and to what extent these strategies can attenuate age differences, or pronounce age differences in favor of older readers. It is conceivable that literary expertise will be more decisive in explaining differences in text comprehension than any other factor because so many literary texts intricately allocate to and are part of the literary canon. Clearly, more research is needed to assess major components in literary text comprehension among younger and older readers.

99

100 Katinka Dijkstra

4. Proposals for future research Future research could focus on experiments with more and less experienced older and younger readers in order to determine how differences in working memory capacity, inhibitory control, and literary expertise can explain differences and processes in literary text comprehension. One experiment could manipulate the time made available to readers to process and understand a literary text. This way, effects of limited working memory capacity on text comprehension among younger and older readers could be assessed. If increased reading time is not available as a compensating strategy for text comprehension among older adults, their literary expertise may be decisive in their ability to establish a coherent representation of the literary text. One way to measure literary expertise would be to subject readers to a test of declarative literary knowledge by letting them match authors to a wide variety of literary book titles (cf. Dijkstra 1994). Procedural knowledge could be assessed by asking readers about their literary history of book reading at home, at school and their current reading habits. In addition, they could mark off bestseller literary books they have read in the past. Another experiment could manipulate the literary text, either by adding to or removing from the text related but not directly relevant literary information, such as metaphors, streams of thought, or elaborate descriptions of details in the text. A manipulation with added information could either facilitate or hinder text comprehension among older readers. This would depend on their ability to keep this information activated in a way that facilitates the construction of a situation model or their inability to filter this information out during the construction of a situation model of the text. Again, literary expertise may play an important role here, in particular if processing time is manipulated as well. With results of these experiments, we may gain a deeper understanding of agerelated deficits in cognitive functioning, the employment of compensatory mechanisms, and evolved skills, such as literary expertise. Both age-related deficits and maintained skills in text comprehension are equally important for our understanding of text comprehension processes across the life span. Most likely, they complement each other in different ways over the years and with increasing reading experiences. At this stage of text comprehension research, it is unknown exactly in what stage of life readers start using compensating strategies. Nor is it clear when these strategies fail to be successful. Do older readers ever get to a point in which successful integration of new and already processed information becomes impossible, or does that only occur when memory impairing diseases, such as Alzheimer’s disease, take their toll on text comprehension processes?

Old readers

There is evidence, however, that even among Alzheimer’s disease patients with moderate cognitive impairment, text-relevant mechanisms, such as coherence, on a sentence by sentence level, remain relatively intact in the production of discourse (Dijkstra et al. in press). Understanding written information is also relatively preserved among Alzheimer’s disease patients until the late stages of impairment (Bourgeois 1990). Apparently, declines in discourse production and comprehension are rather slow even under conditions of progressive diseases. The study of age-related declines and maintained skills in text comprehension that takes place in laboratories might better meet the challenges imposed upon older adults in their natural environment if the test materials are naturalistic as well. Text comprehension research could study the construction of situation models in newspaper articles, short stories and poems among older and younger readers, and how readers successfully integrate the information they process online. This could be accomplished by letting readers think aloud when reading a story, and analyzing these protocols for text comprehension strategies. Then, strategies of younger and older readers and readers with different levels of literary expertise could be compared to see what, if any, differences there are between the two groups of readers.

References Adamowicz, J. L. 1976. “Visual short term memory and aging”. Journals of Gerontology: Psychological Sciences 31: 39–46. Adams, C. 1986. Qualitative Changes in Text Memory from Adolescence to Mature Adulthood. Unpublished doctoral dissertation. Wayne State University, Detroit, Michigan. Adams, C. 1991. “Qualitative age differences in memory for text: A life-span developmental perspective”. Psychology and Aging 6: 323–336. Adams, C., Labouvie-Vief, G., Hobart, C. and Dorosz, M. 1990. “Adult age group differences in story recall style”. Journals of Gerontology: Psychological Sciences 12: 17–27. Backman, L. 1985. “Compensation and recoding: A framework for aging and memory research”. Scandinavian Journal of Psychology 26: 193–207. Backman, L., Mantyla, T. and Herlitz, A. 1990. “The optimization of episodic remembering in old age”. In Succesful Aging: Perspectives from the Behavioral Sciences, P.B. Baltes and M.M. Baltes (eds), 118–163. Backman, L. and Karlsson, T. 1987. “Effects of word fragment completion on free recall: A selective improvement of older adults”. Comprehensive Gerontology 1: 4–7. Baddeley, A. D. 1986. Working Memory. New York: Oxford University Press. Baltes, P. B. and Baltes, M. M. 1990. Successful Aging: Perspectives from the Behavioral Sciences. Cambridge, New York, Port Chester, Cambridge University Press: Melbourne, Sydney. Baltes, P. B. and Baltes, M. M. 1982. “Microanalytic research on environmental factors and plasticity in psychological aging”. In Review of Human Development, T. M. Field, A. Huston, H. C. Quay, C. Troll, and G. E. Finley (eds), 524–539. New York: Wiley. Bourgeois, M. S. 1990. “Enhancing conversation skills in patients with Alzheimer’s disease using a prosthetic memory aid”. Journal of Applied Behavioral Analysis 23: 29–42.

101

102 Katinka Dijkstra

Byrd, M. 1981. Age Differences in Memory for Prose Passages. Unpublished doctoral dissertation. University of Toronto, Toronto. Byrd, M. 1987. “The effects of previously acquired knowledge on memory for textual information”. International Journal of Aging and Human Development 24: 231–240. Cerella, J., Poon, L.W. and Williams, D.M. 1980. “Age and the complexity hypothesis”. In Aging in the 1980’s, L.W. Poon (ed.), 332–340. Washington DC: American Psychological Association. Connelly, S.L, Hasher, L. and Zacks, R. T. 1991. “Age and reading: The impact of distraction”. Psychology and Aging 6: 533–541. Craik, F. I. M. 1983. “On the transfer of information from temporary to permanent memory”. Philosophical Transactions of the Royal Society of London B320: 341–359. Craik, F., Byrd, M., and Swanson, J. M. 1987. “Patterns of memory loss in three elderly samples”. Psychology and Aging, 2: 79–86. Crew, F. F. 1977. Age Differences in Retention after Varying Study and Test Trials. Unpublished master’s thesis, Georgia Institute of Technology, Atlanta, Georgia. Dijkstra, K. 1994. Lese-entscheidung und Lektürewahl. Empirische Untersuchungen über Einflussfaktoren auf das Leseverhalten [Reading and Reading Preferences: Empirical Studies on Factors influencing Reading Behavior]. Berlin: Edition Sigma. Dijkstra, K. 1996. “How professional readers differ from leisure time readers: An empirical investigation into intrinsic and social aspects of reading”. In Empirical Approaches to Literature and Aesthetics, R.J. Kreuz and M.S. MacNealy (eds), 539–558. Norwood, NJ: Ablex Publishing Corporation. Dijkstra, K., Charness, N. and Zwaan, R.A. (2000). Situation models in younger and older readers. unpublished manuscript. Dijkstra, K., Bourgeois, M., Petrie, G., Burgio, L. and Allen-Burge, R. (in press). My recaller is on vacation. Discourse analysis of nursing home residents with cognitive impairments. Discourse Processes. Dixon, R. A., Hultsch, D. F., Simon, E. W. and Von Eye, A. 1984. “Verbal ability and text structure effects on adult age differences in text recall”. Journal of Verbal Learning and Verbal Behavior 23: 569–578. Fokkema, D. W. 1988. “On the reliability of literary studies”. Poetics Today 9: 529–543. Garner, R. 1987. Metacognition and Reading Comprehension. Norwood, NJ: Ablex Publishing Company. Gerard, L., Zacks, R. T., Hasher, L. and Radvansky, G. A. 1991. “Age deficits in retrieval: The fan effect”. Journals of Gerontology: Psychological Sciences 46: 131–136. Gernsbacher, M. A. 1989. “Mechanisms that improve referential access”. Cognition 32: 99–156. Gernsbacher, M. A. 1990. Language Comprehension as Structure Building. Hillsdale, NJ: Erlbaum. Graesser, A. C., Millis, K. and Zwaan, R. A. 1997. “Discourse comprehension”. In Annual Review of Psychology, J. T. Spence, J. M. Darley and D. J. Foss (eds), 48. Palo Alto, CA: Annual Reviews, Inc. Graves, B. and Frederiksen, C. H. 1996. “A cognitive study of literary expertise”. In Empirical Approaches to Literature and Aesthetics, R. J. Kreuz and M. S. MacNealy (eds), 397–418. Norwood, NJ: Ablex Publishing Corporation. Hamm, V. P. and Hasher, L. 1992. “Age and the availability of inferences”. Psychology and Aging 7: 56–64. Harker, J.O, Harley, J. T. and Walsh, D. A. 1982. “Understanding discourse: A life-span approach”. Advances in Reading/Language Research 1: 155–202. Hartley, J. T. 1986. “Reader and text variables as determinants of discourse memory in adulthood”. Psychology and Aging 1: 150–158.

Old readers 103

Hartman, M. and Hasher, L. 1991. “Aging and suppression: Memory for previously relevant information”. Psychology and Aging 6: 587–594. Hasher, L. and Zacks, R. T. 1988. “Working memory, comprehension, and aging: A review and a new view”. In The Psychology of Learning and Motivation, G.H. Bower (ed.), vol. 22, 193–225. San Diego, CA: Academic Press. Hultsch, D. F. 1969. “Adult age differences in the organization of free recall”. Developmental Psychology 1: 673–678. Hultsch, D. F. 1971. “Adult age differences in free classification and free recall”. Developmental Psychology 4: 338–342. Jackson, J. D. and Kemper, S. 1993. “Age differences in summarizing descriptive and procedural texts”. Experimental Aging Research 19: 39–51. Just, M. A. and Carpenter, P. A. 1992. “A capacity theory of comprehension: Individual differences in working memory”. Psychological Review 24: 17–34. Kemper, S. 1992. “Language and aging”. In The Handbook of Aging and Cognition, F. I. M. Craik and T. A. Salthouse, 213–270. Lawrence Erlbaum: Hillsdale, N. J. Kintsch, W. and Van Dijk, T. A. 1978. “Toward a model of text comprehension and production”. Psychological Review 85: 363–394. Kintsch, W. 1988. “The role of knowledge in discourse comprehension: A construction-integration model”. Psychological Review 95: 163–182. Knulst, W. and Kraaykamp, G. 1998. “Trends in leisure reading: Forty years of research on reading in the Netherlands”. Poetics 26: 21–41. Kraaykamp, G. and Dijkstra, K. 1999. “Preferences in leisure time book reading: A study on the social differentiation in book reading for the Netherlands”. Poetics 26: 203–234. Kwong See, S. T., and Ryan, E. B. 1995. “Cognitive mediation of adult age differences in language performance”. Psychology and Aging 10, 458–468. Kwong See, S. T. and Ryan, E. B. 1996. “Cognitive mediation of discourse processing in later life”. Journal of Speech-Language Pathology and Audiology 20: 109–117. Kwong See, S. T. and Ryan, E. B. 1999. “Intergenerational communication: The survey interview as a social exchange”. In Cognition, Aging, and Self-reports, N. Schwarz and D. C. Park (eds), 245–262. Psychology Press/Erlbaum (UK) Taylor & Francis: Hove, England. Labouvie-Vief, G. 1985. “Intelligence and cognition”. In Handbook of the Psychology of Aging, J. E. Birren and K. W. Schaie (eds), 500–530. New York, NY: Van Nostrand Reinhold Co, Inc. Light, L.L. 1990. “Interactions between memory and language in old age”. In Handbook of the Psychology of Aging, J.E. Birren and K.W. Schaie (eds), 275–290. San Diego, CA: Academic Press. Light, L. and Anderson, P. A. 1985. “Working-memory capacity, age, and memory for discourse”. Journal of Gerontology 40: 737–747. Mockros, C. A. 1993. “The development of aesthetic experience and judgement”. Poetics 23: 219–234. Morrow, D. G., Stine-Morrow, E. A. L., Von Leirer, O., Andrassy, J. M. and Kahn, J. 1997. “The role of reader age and focus of attention in creating situation models from narratives”. Journal of Gerontology: Psychological Sciences 52: 73–80. Radvansky, G. A. 1999. “Aging, memory, and comprehension”. Current Directions in Psychological Science 8: 49–52. Radvansky, G. A., Zwaan, R. A., Curiel, J. M. and Copeland, D. E. (in press) “Situation models and aging”. Psychology and Aging. Rowe, J. W. and Kahn, R. L. 1998. Successful Aging. Pantheon Books: New York. Salthouse, T. A. 1991. “Mediation of adult age differences in cognition by reductions in working memory speed and speed of processing”. Psychological Science 2: 179–183.

104 Katinka Dijkstra

Salthouse, T. A. 1992. “Why do adults age differences increase with task complexity?”. Developmental Psychology 28: 905–918. Salthouse, T. A. 1984. “Effects of age and skill in typing”. Journal of Experimental Psychology: General 113: 345–371. Schaie, K. W. 1990. “The optimization of cognitive functioning in old age: Predictions based on cohort-sequential and longitudinal data”. In Successful Aging. Perspectives from the Behavioral Sciences, P. B. Baltes and M. M. Baltes (eds), 94–117. Schaie, K. W. 1989. “Perceptual speed in adulthood: Cross-sectional and longitudinal studies”. Psychology and Aging 4: 443–453. Shimamura, A. P., Berry, J. M., Mangels, J. A., Rusting, C. L. and Jurica, P. J. 1995. “Memory and cognitive abilities in university professors”. Psychological Science 6: 271–277. Ska, B. and Joanette, Y. 1996. “Discourse in older adults: Influence of text, task, and participant characteristics”. Journal of Speech-Language Pathology and Audiology 20: 101–108. Slamecka, N. J. and Graf, P. 1978. “The generation effect: Delineation of a phenomenon”. Journal of Experimental Psychology: Human Learning and Memory 5: 115–124. Spilich, G. J., Vesonder, G. T., Chiesi, H. L. and Voss, J. F. 1979. “Text processing of domain related information for individuals with high and low domain knowledge”. Journal of Verbal Learning and Verbal Behavior 18: 275–290. Spilich, G. J. 1983. “Life-span components of text processing: structural and procedural differences”. Journal of Verbal Learning and Verbal Behavior 22: 231–244. Soederberg-Miller, L. M. and Stine-Morrow, E. 1998. “Aging and the effects of knowledge on online reading strategies”. Journals of Gerontology: Psychological Sciences 53: 223–233. Stine, E. A., Cheung, H. and Henderson, D. 1995. “Adult age differences in the on-line processing of new concepts in discourse”. Aging and Cognition 2: 1–18. Stine E. A. 1990. “On-line processing of written text by younger and older adults”. Psychology and Aging 5: 68–78. Stine, E. A. and Hindman, J. 1994. “Age differences in reading time allocation for propositionally dense sentences”. Aging and Cognition 1: 2–16. Stine-Morrow, E. A., Loveless, M. K. and Soederberg, L. M. 1996. “Resource allocation in on-line reading by younger and older adults”. Psychology and Aging 11: 475–486. Tun, P.A. 1989. “Age differences in processing expository and narrative text”. Journals of Gerontology: Psychological Sciences 44: 9–15. Van Dijk, T. A. and Kintsch, W. 1983. Strategies of Discourse Comprehension. New York: Academic Press. Van der Linden, M., Hupet, M., Feyereisen, P., Schelstraete, M., Bestgen, Y., Bruyer, R., Lories, G., Ahmadi, A. and Seron, X. 1999. “Cognitive mediators of age-related differences in language comprehension and verbal memory performance”. Aging, Neuropsychology, and Cognition 6: 32–55. Whitney, P. and Clark, M. B. 1989. “Disambiguation and cognitive control”. In Resolving Semantic Ambiguity, D. S. Gorfein (ed.), 239–251. New York: Springer. Wippler, R. 1990. “Cultural resources and participation in high culture”. In Social Institutions, M. Hechter, K. D. Opp and R. Wippler (eds), 187–204. New York: Aldine de Gruyter. Zabrucky, K. and Moore, D. (1994). “Contributions of working memory and evaluation and regulation of understanding to adults’ recall of texts”. Journals of Gerontology: Psychological Sciences 49: 210–212. Zabrucky, K. and Moore, D. 1995. “Elaborations in adults’ text recall. Relations to working memory and text recall”. Experimental Aging Research 18: 3–7. Zeitz, C. M. 1994. “Expert-novice differences in memory, abstraction, and reasoning in the domain of literature”. Cognition and Instruction 12: 277–312.



Old readers 105

Zwaan, R. A. 1993. Aspects of Literary Comprehension. Amsterdam, Philadelphia: John Benjamins Publishing Company. Zwaan, R. A., Langston and Graesser, A. C. 1995. “Dimensions of situation model construction in narrative comprehension”. Journal of Experimental Psychology: Learning, Memory, and Cognition 21: 386–397. Zwaan, R. A. and Radvansky, G. A. 1998. “Situation models in language and memory”. Psychological Bulletin 123: 162–185.



What we know about reading poetry Theoretical positions and empirical research David Ian Hanauer Tel-Aviv University

Literary research in the twentieth century has been characterized by the propagation of a variety of approaches to the interpretation of literary texts. Currently, the field of literary research is strewn with schools of interpretation fighting for the legitimacy of their own position. In some ways the development of the Empirical Study of Literature can be seen as yet another new approach to appear within the realm of literary studies. However, the empirical approach to literature offers a possibility that other approaches cannot. By defining itself as an empirical approach and by demanding that research conform to international standards of psychological research, this approach offers a methodology through which theoretical positions can be compared and evaluated and ultimately new understandings proposed. All previous approaches to literature, even when they called themselves scientific, could not perform this task. To date, disagreements between approaches to literature have been addressed through literary analysis and argumentation. In contrast, the empirical approach to literature is based on the systematic elicitation and interpretation of data in order to understand and explain phenomena within the literary system. Another important advantage of a branch of literary study that is based on research paradigms prevalent within the field of cognitive psychology is that it can integrate knowledge from reading research with issues and texts from the field of literary study. This inter-disciplinary usage of research widens the potential knowledge base used for understanding literary reading. The present chapter will focus on empirically based concepts of the process of reading and some of the central arguments raised within the twentieth century over the description of how poetry is read. The overall approach taken is to evaluate theoretical positions on the reading of poetry in relation to current knowledge of the reading process and empirical data on the reading of poetry that has been collected within the framework of the Empirical Study of Literature. As I have done elsewhere, the empirical method is understood “as a research tool for deepening our

108 David Ian Hanauer

understanding of issues arising within a literary system” (Hanauer 1994: 124). Specifically, the current chapter aims to evaluate the theoretical descriptions of poetry reading through the discussion of empirical findings.

1.

Empirical positions on reading

Current positions on literacy propose that reading is a multifaceted and contextually embedded process (Freire 1985; Kress 1989; Levine 1986; Meek 1991; Shuman 1986). Naturalistic research has shown readers contend on a daily basis with a multiplicity of different genres of text (Baynham 1995; Freedman and Medway 1994; Hanauer 1997a; Heath 1983). Experimental research has shown that when faced with various texts in dissimilar conditions the cognitive components of the reading process are used in different ways (Hanauer 1998b; Van Dijk and Kintsch 1983; Zwaan 1993). Accordingly, it is assumed that the mature reader is experienced in handling texts in a differential and contextually relevant manner. The reading process involves an interaction between the reader, the reader’s accumulative knowledge of the reading process, the specific text that is being read and the physical, psychological and social context within which the text appears. Theories such as genre theory (Freedman and Medway 1994; Kamberelis 1995) or cognitive control systems (Van Dijk and Kintsch 1983; Zwaan 1993) propose that the reader has developed, through formal and informal instruction, knowledge that allows the negotiation of the text in different conditions. Accordingly, these theories predict variation between individuals using the same text as a result of differences in their accumulated literacy knowledge and/or personal goals for reading. According to both genre theory and cognitive control systems, the starting point of the reading process is a genre decision. The specific features of the text, the reader’s goals and the context of reading are all information sources that direct the genre decision (Hanauer 1996, 1998a; Zwaan 1991, 1993). Having decided on the genre of the text being read, the relevant procedural knowledge is activated and directs the reading process. This text specific procedural knowledge includes knowledge of the expected goals of reading, how the individual’s attentional resources should be used, what underlying cognitive processes should be activated in order to construct meaning and which types of internal representation are appropriate (Hanauer 1997a; Kamberelis 1995; Van Dijk and Kintsch 1983). On a cognitive level, the activation of different discourse-specific, cognitive control systems will result in changes to the processing and representation of the text. For example, in a well-known study by Zwaan (1991) participants processed and represented stories differently when they were read as literary or newspaper stories. The differences related to both reading rate and the level of internal representation. This principle of discourse specific, differential cognitive processing

What we know about reading poetry 109

can be seen as the underlying assumption of most of the research conducted on literary processing within the field of the Empirical Study of Literature. The relevance of this research for the field of reading is that it investigates reading phenomena that are clearly embedded within the literary system and thus presents information on a specialized reading process. Zwaan (1996) points out that literary texts and literary reading put particular demands on the reader. In many cases the literary text has high levels of specific literary features and devices and the task of literary text comprehension goes beyond the construction of a coherent textbase. These added demands also pose a challenge to the field of reading research in that models of the reading process will have to take into account not only prototypical examples of short narratives or expository text but also the more complex phenomena of describing the comprehension of literary prose, drama, and poetry. In comparison with the amount of research conducted over the last twenty-five years on the reading of narrative and expository text, only a limited number of studies have been conducted on literary reading. On the reading of poetry this situation is even more acute. Even within the field of the Empirical Study of Literature very few studies have investigated how poetry is read. In some ways the present chapter is an attempt to interest more researchers in this direction of research. However, while there is only a limited amount of empirical research on poetry, nearly every major school of literary interpretation has proposed an understanding of how poetry is read. As might be expected these schools of interpretation are not solely dedicated to explaining poetry reading but their assumptions and principles of literary reading do propose a variety of positions that can be used to understand poetry reading. In other words, the current situation within the field of literary studies is that there are various theoretical options for describing the poetry reading process and there is a relatively new and limited body of empirical research investigating poetry. The aim of the next section is to present a general overview of some major schools of literary interpretation in the twentieth century and to present the general principles that define their approach in relation to poetry. In the section that follows this review, these theoretical positions will be discussed in relation to empirical research on poetry reading.

2.

Schools of literary interpretation and poetry

The various schools of literary interpretation that developed in the twentieth century have proposed different positions on the role of language and the reader in poetry reading. Some schools of theory such as New Criticism, Formalism and Stylistics have emphasized the role of language in the categorization and reading of poetry; while other schools of theory such as Reader Response Criticism and Reception

110

David Ian Hanauer

Theory have emphasized the role of the reader in reading poetry. This difference in emphasis produces differences in the way poetry reading is described. Accordingly, the current review of schools of literary interpretation can be broadly divided into language-driven and reader-driven descriptions of poetry and poetry reading. The review in this section presents the general principles of each school of interpretation and how these can be applied to an understanding of poetry and the reading process. New Criticism emerged in the 1920s and 1930s as a response to interpretive practices and positions on poetry inherited from the nineteenth century. Before the advent of New Criticism, literary research in the English speaking world focused on the social-historical aspects of the period in which the poem was written or the biographical and psycho-emotive aspects of the poet. Both these approaches to literary research de-emphasize the importance of the poem itself. For a New Critic the only legitimate object of study was the literary text itself and the aim of literary research was the critical interpretation of the literary text. The highly influential terms the ‘Intentional Fallacy’ and the ‘Affective Fallacy’ coined by Wimsatt and Beardsley (1954) negated the idea that literary texts should be interpreted by researching or discussing the author’s biography, psychology, historical period or intention. The ‘Affective Fallacy’ directly attacked the Romantic idea that the reader’s emotional response was important in the interpretation of poetry. New Criticism limited literary research to a very close analysis of the literary text. They considered the literary text to hold all the information that is required for its interpretation. As stated by Wimsatt and Beardlsey: Poetry is a feat of style by which a complex meaning is handled all at once. Poetry succeeds because all or most of what is said or implied is relevant; what is irrelevant has been excluded, like lumps from a pudding and “bugs” from machinery. In this respect poetry differs from practical messages, which are successful if and only if we correctly infer the intention. They are more abstract than poetry. (1954: 23)

As reflected in this quote from Wimsatt and Beardsley, the New Critical method considered the poem to be a unified object within which every element included in the text had a role to play in its interpretation. Poetry reading is a process in which the inter-relationships between the components of the poem are carefully scrutinized and a unity of form and content is defined. New Criticism viewed the poem as an autonomous object that had an inherent and unified meaning. However, it was only the reader who was trained in the New Critical interpretive method that would be able to find the structure of semantic and formal inter-relationships that were all important for the interpretation of the poem. From a New Critical perspective, poetry was defined by its use of highly patterned language and poetry reading was the process of looking for formal and semantic patterns and finding the interrelationships that exist in the poem between these patterns.

What we know about reading poetry

Formalism in many ways presents a similar description of how poetry is read. However, the conceptual starting point of Formalism is very different from that of New Criticism. The source of the Formalist position is the development of new descriptions of the linguistic system. Formalist theorists described poetry and poetry reading within the context of the wider linguistic system. A prominent example of this is the approach proposed by Roman Jakobson. Jakobson (1960) defined six major functions for language, which derive from an analysis of the communicative act. The six functions are the emotive function (emphasis on addresser), the referential function (emphasis on context), the phatic function (emphasis on social contact), the conative function (emphasis on addressee), the metalingual function (emphasis on the code) and the poetic function (emphasis on the message) itself. For Jakobson the term message refers to the linguistic features of the text being communicated; it does not refer to the content of the message. In this formulation poetry is a group of texts in which the poetic function is clearly dominant. The poetic function is considered an intrinsic quality of poetry. Jakobson defined the poetic function as “The set towards the message as such, focus on the message for its own sake is the poetic function” (Jakobson 1960: 356). The poetic function involves a focus on the linguistic features of the text themselves. The effect of the poetic function on the reader is the “palpable” presence of language. This effect of the presence of language in poetry reading is achieved through the unique way language is constructed in the poetic function. “The poetic function projects the principle of equivalence from the axis of selection into the axis of combination” (Jakobson 1960: 358). According to Jakobson, equivalence or sameness is the major means of constructing the whole poem. On all levels of their linguistic features poems display a structure of similarities and contrasts. The formal features of a poem display a system of similarities, repetitions and regularities such as the repetition of syllables, regularities of stress pattern or similarity in grammatical category. For Jakobson, a poem is a defined by its linguistic attributes and is presented as a hierarchically structured set of patterns of similarity and contrast. The reading of poetry is affected by the structure of the poem. According to Jakobson, poetry reading is a non-linear reading process in which the reader is directed by the patterns of similarity and contrast. These patterns draw the reader’s attention while reading. This structure and palpable presence of language breaks the illusion of a direct relationship between the linguistic and non-linguistic sign. By emphasizing the illusory nature of the relationship between linguistic and nonlinguistic signs and by situating the linguistic sign within a system of signs, a multiplicity of potential meanings is achieved within the poem. Like Jakobson, other Formalist theorists stressed the importance of linguistic structures in the definition and reading of poetry. Mukarˇovsky˘ (1964) proposed that poetry was the result of the systematic violation of the norms of standard language. “The function of poetic language consists in the maximum foregrounding

111

112

David Ian Hanauer

of the utterance” (Mukarˇovsky˘ 1964: 16). The Formalist approach defined poetry using linguistic structures and considered reading poetry to be a process in which the linguistic structures of the poem directed the reader by demanding her/his attention. Just like New Critical and Formalist approaches, Stylistics also defines a central role for language in the definition and reading of poetry. Stylistics and its contemporary counterpart New Stylistics consider the choice of linguistic structure as a centrally defining aspect of the literary work. As described by Alderson and Short, stylistics is “intended to help determine interpretation through the examination of what a text contains, by describing the linguistic devices an author has used, and the effect of such devices. Such an analysis is predominantly text-based, and has tended to see texts as containing meanings” (1989: 72). In other words, the textual features both define the poem and the interpretation of the poem. Within this approach, the specific features used in the poem are considered to be the result of an author’s choice and the analysis of these linguistic features a method for interpreting the specific text under consideration. Accordingly, Stylistics considers poetry to be defined and understood in relation to its linguistic features. However, the specific linguistic features chosen and the way they are interpreted is a result of individual choice by the poet and the reader. For Reader Response Criticism, Psychoanalytic Criticism and Reception Theory, it is the literary reader rather than the literary text that is of central importance in the reading process. Reader Response critics such as Fish (1980), 1978), Iser (1978), Rosenblatt (1978) and Bleich (1975), have argued that a literary text only comes into being in the interaction between reader and text. The literary text does not have any existence beyond the reader’s mind. Meaning is a not an inherent aspect of the text but rather the result of the reading process. As stated by Fish in his description of the process of reading Milton’s Paradise Lost
: (…) the reader is drawn into the poem not as an observer who coolly notes the interaction of the patterns, (…) but as a participant whose mind is the locus of that interaction. (1967: 11)

According to Fish, the reader approaches the poem with a series of interpretive strategies that direct her/his reading process by highlighting certain patterns and forms that are interpreted as meaningful. Accordingly, it is the reader’s background and experience as a reader that directs the poetry reading process. The linguistic features of the text play very little if any role in directing the reading process. Fish (1980) further limits the role of textual features when he describes the process through which poems are categorized. Fish (1980) reports on a pseudo-experiment in which students in a literature classroom categorized a random set of names left on the blackboard as a poem and proposed interpretations of this poem. The point Fish makes is that it is the interpretive strategies used by the reader and not any specific configuration of textual or linguistic features that define poetry. Pratt

What we know about reading poetry

(1977) summarizes this position by stating that poetry does not have any formal linguistic or textual features that are unique to poetry and therefore it is the contextual decision to read a text as poem that makes it a poem. According to Fish (1980) the interpretive strategies and conventions of poetry reading are situated within wider communities of readers. It is these interpretive communities who through their promotion of specific ways of reading and approaching literary texts that ultimately define poetry and the way poetry is read. For Fish and Pratt, the categorization and reading of poetry is controlled by the reader’s strategies and conventions of poetry reading. The linguistic and textual elements of the poem play next to no role in this process. Culler (1975) presents a more moderate position, which also supports the concept of the central importance of the reader. As stated by Culler: (…) reading poetry is a rule governed process of producing meanings; the poem offers a structure which must be filled up and one therefore attempts to invent something, guided by a series of formal rules from one’s experience of reading poetry. (1975: 126)

In this quote Culler presents the conventionalist position that proposes that poetic texts do not have an inherent meaning but rather meaning is derived from the application of a series of conventions for reading. However, the textual and linguistic features of the poem, in this formulation of the conventionalist position, do have a role. The linguistic and textual features of the poem present a ‘structure’ to the reader. The reader employs literary conventions of meaning construction to transform the poem into a meaningful entity. The textual features of the poem have a role once the reader has decided to relate to the text as a poem and applies the appropriate conventions of reading. Culler (1975) proposes four conventions which are important in the reading of poetry: the convention of distance and impersonality which allows the reader to construct the fictional world of the poem, the convention of unity which directs the reader to search for coherence and unity of purpose between the various components of the poem, the convention of social significance which directs the reader to see poem as making a significant statement and finally the convention of resistance to meaning making which makes the reader expect the content of the poem to pose difficulties in its comprehension. Ultimately, from a Reader Response perspective it is the conventions of poetry reading and not the textual and linguistic features of poems that control the reading process. The description of poetry reading proposed by proponents of Deconstruction re-emphasizes the centrality of language in the reading process. Originally a philosophical approach designed as a critique of Structuralist thinking, Deconstruction has as its central tenet an assertion relating to language and the relationship between language and meaning construction. As stated by Derrida:

113

114

David Ian Hanauer

Henceforth, it was necessary to begin thinking that there was no center, that the center could not be thought in the form of the present-being, that the center had no natural site, that it was not a fixed locus but a function, a sort of nonlocus in which an infinite number of sign-substitutions came into play. This was the moment that the language invaded the universal problematic, the moment when, in the absence of a center or origin, everything became discourse — provided we can agree on this word — that is to say, a system in which the central signified, the original or transcendental signified, is never absolutely present outside a system of differences. The absence of the transcendental signified extends the domain and play of signification infinitely. (1978: 9)

For Derrida, language offers humanity a structure through which meaning is constructed. But any meaning that is defined is based on the preference for one of the potential infinite meanings within language. Every utterance and every decision on meaning is constituted within the wider context of a historical discourse. The illusion that the participant in the language communication has found the meaning of the utterance is nothing more than an illusion that is based on the participant’s acceptance of the authority (the transcendental signifier) of a locus of meaning. Within the discourse the participant’s observed meaning, which seemingly excludes other potential meanings, actually includes all the meanings that it supposedly denies. As stated by Derrida “language bears within itself the necessity of its own critique” (1978: 17). The view of poetry reading proposed by supporters of deconstruction emphasizes the role of language but views language as a repository of all the historical and discursive potentialities of signification within that language. Thus, finding meaning in a poem becomes a process of continual regression and continual conflict between potential meanings. Miller describes this reading process in the following way: (…) a poem invites an endless sequence of commentaries which never succeed in “getting the poem right”. The poem, in my figure, is that ambiguous gift, food, host in the sense of victim, sacrifice, that which is broken, divided, passed around, consumed by the critics canny and uncanny who are in that odd relation to one another of host and parasite. The poem, however, any poem, is, it is easy to see, parasitical in itself, in another version of the perpetual reversal of parasite and host. If the poem is food and poison for the critics, it must in turn have eaten. It must have been a cannibal consumer of earlier poems. (1977: 283)

Any poem is part of a long chain of poems and language structures that precede and at the same time co-exist with and within the poem. The system of language makes sure that the poem can be read in a multiplicity of self-conflicting ways. Like New Criticism, Stylistics and Formalism, Deconstruction sees language as central to the meaning construction process in poetry reading. But in contrast to these approaches,

What we know about reading poetry

Deconstruction presents language as the central source of the multiplicity of meaning and the reason why unity of meaning, which is so valued within the New Critical and Stylistic approaches, is a theoretical impossibility. As one literary critic has pointed out, reading a poem from a Deconstructive stance involves the “careful teasing out of warring forces of signification within the text” (Johnson 1980: 63). As can be seen in this section, the various schools of literary interpretation have emphasized different aspects in their descriptions of what poetry is and how poetry is read. In the section that follows these differences will be discussed in relation to empirical research on poetry reading.

3.

Literary theories and empirical investigations of poetry reading

In reviewing a series of theoretical positions the danger always is that for reasons of clarity the theoretical positions will be so simplified and extrapolated that they become nothing more than shadows of their original complexity. In the field of literary studies this is particularly problematic because various literary critics formulated each theory slightly differently and the aims of the various schools of interpretation were not the same. Accordingly, while some of these positions in their extreme form seem to be diametrically opposed, in their various formulations theories from different schools of interpretation overlap to a certain degree. One position that seems to be in direct disagreement with another might actually describe a different aspect of poetry reading and thus not actually be in contradiction to it. The aim of the current chapter is not to verify any particular theory or school of interpretation but rather to evaluate and discuss these theoretical positions in light of research data from empirical studies of poetry and reading. In this way it might be possible to present a closer approximation of what happens when a poem is read. One of the major distinctions stated in the review above is the different weight assigned to linguistic features in the categorization, reading, meaning construction and representation of poetry. Language-driven approaches such as New Criticism, Formalism, and (New) Stylistics would seem to place a heavy emphasis on the role of language, whereas reader-driven approaches would seem to relegate language to a secondary position in relation to the reader and the reading context. Although these positions have sometimes been presented as mutually exclusive, from the viewpoint of reading research the relations between reader and text are situated within a wider context. The process of reading is by definition an interaction between text and reader in a specific physical, psychological and social context. This interaction is dynamic and the reader has the ability to focus on different aspects of the reading process at different stages in the process. From a reading research perspective, language-driven and reader-driven processing do not present a dichotomy but

115

116

David Ian Hanauer

are both obligatory within the reading process. The reader can direct additional attention to language or reader knowledge as an option while processing a text. From a genre perspective, poetry reading should be viewed as a case of discourse specific (genre-specific) processing. As with other text-processing models, a description of poetry reading should describe how this discourse specific knowledge is activated, how the reader’s attentional resources are used, what underlying cognitive processes are activated in order to construct meaning and which types of internal representation are appropriate. In addition, the model needs to relate to potential differences between the way expert and novice readers handle the poetry reading task. As stated above not enough empirical research has been conducted on poetry reading to present a fully worked out and empirically based description of poetry reading. However, some research has been conducted that allows the sketching of an outline of such a model. The present discussion of research has limited itself to those studies that are relevant to the description of the process of reading a poem and are rigorous from an empirical viewpoint. 3.1 Categorizing poetry According to both genre theory and cognitive control systems, the starting point of the reading process is a genre decision. As presented above, language-driven descriptions of poetry reading have proposed that the genre of poetry is defined by the presence of highly patterned language. This type of language will define the text as poetic and direct the subsequent reading process. On the other hand, readerdriven descriptions of poetry reading have proposed that it is the conventions of poetry reading that define a text as a poem. The reader makes a strategic choice to read the text as a poem using the conventions of poetry reading with or without the text having any specific linguistic features. Three substantial studies and one metaanalysis have been conducted on this issue (Hanauer 1995, 1996; Hoffstaedter 1987; Miall and Kuiken 1998). The underlying research paradigm used for these studies is the rating of the poeticity of texts with different linguistic and textual features by groups of participants with different educational backgrounds. Hoffstaedter’s (1987) early study required participants to rate the poeticity of twenty-four different texts. The twenty-four texts included a range of genres, from poems to a history textbook. The main experimental condition was the presentation of these texts as either from a newspaper or from a poetry-reading context. The results revealed that for fourteen of the twenty-four texts no significant differences were found between the texts in the two conditions. In other words, for the majority of texts the external categorization of the text as a poem or a newspaper was not enough to override the role of the textual and linguistic features in categorizing the text as a poem. Hoffstaedter concluded that for an external categorization of a poem

What we know about reading poetry

to be influential that the text itself must have the linguistic and textual features associated with poetry. Hoffstaedter’s conclusion defines an important role for the presence of specific linguistic features in the categorization of a poem. Hanauer (1995) asked participants with different levels of literary education — entry level college literature students, students with completed B.A.’s in literature and students with completed M.A.’s in literature — to rate the poeticity and literariness of twelve texts. The texts were all authentic and included seven postmodern poems, two modern poems, two sections from literary novels and one section from a history textbook. On the poeticity rating no significant differences were found between the mean poeticity ratings given by the groups for the various texts. This result was further investigated by comparing the rank order of these texts by the different groups. Once again, no significant differences were found between any of the groups in the rank order of the texts on the poeticity rating. Taken together these results demonstrate that the participants had a shared concept of which texts were poetic and the levels of the poeticity. This result does not mean that educational background is of no importance, but rather that knowledge of the textual features of a poem were acquired by all three groups. This would suggest that the specific textual features that categorize a poem are well known to novices as well as expert literary readers. Interestingly, significant differences were found between these groups for the literariness judgments and literariness rank orders. Participants who had advanced literary knowledge (completed B.A.’s and M.A.’s in Literature) gave higher average literariness ratings than the novice group and ranked the texts differently. This result, which demonstrates a role for literary background, may have been affected by the specific set of stimulus items used in this study — seven postmodern poems. These poems have poetic features and hence were recognized and categorized as poems by all groups. However, like much postmodern poetry they experiment with the conventions, aims and content of poetry. Accordingly, when asked to rate the literariness of the text novice readers were unsure whether these poems could actually be considered literary. Whereas experienced literary readers who have had experience with postmodern poetry recognized these as literary texts and gave higher literariness ratings. This surprising result (it is a poem but not a literary text) found for novice readers may show a difference in the readers’ internal representation of the poetry and the literariness categorization task. Whereas the poetry categorization task is based on structural features (as the data seem to show) the literariness categorization task may involve additional aspects such as the evaluation of the literary value of the text. Genre theory differentiates between structural, procedural and functional knowledge of a genre (Bazerman 1988; Freedman and Medway 1994; Hanauer 1997a; Steen 1999; Swales 1990). Schauber and Spolsky (1986), who developed a genre theory of literary interpretation, further propose that there is a difference

117

118

David Ian Hanauer

between structural and functional genre in literature. According to Schauber and Spolsky (1986) poetry is an example of a structural genre. It is possible that in the comparison of the genre knowledge base of expert and novice poetry readers, they have a shared representation of structural features of a poem but differ in their knowledge of literature. Accordingly, a text can be a poem (that is, has the structural features of a poem) but is not a literary text (that is, cannot be read as a literary text and does not fulfill the functions of a literary text as internalized by the novice reader). In the process of acquiring literary knowledge it might be quite easy to acquire knowledge of the textual features that mark a text as a poem, but the development of the knowledge of how to process these features and the functional role of this text may take more advanced literary knowledge. Hanauer (1996) in a second study further considered the role of textual features and educational background in the categorization of poetry by requiring participants from two educational levels — entry-level college literature students and students with completed B.A. degrees in literature — to rate the poeticity of two poems which had been graphically and phonetically manipulated. Each of the original poems was manipulated so as to produce three levels of graphic and phonetic information. The participants in this study were required to rate each of the manipulations on a continuum marked on one side as “clearly not a poem” and marked at the other end as “clearly a poem”. The results of this study reveal that all the participants were sensitive to the changes in the graphic and phonetic manipulations of the poem and that changes in graphic and phonetic information were reflected in the level of poeticity assigned to the text. The poeticity ratings for both poems created a clear hierarchy of poeticity levels tied directly to the levels of textual and linguistic information presented to the participants. These results were found for both poems and across both educational groups. These results present additional information about the readers´ internal knowledge of the structural features of poetry and show that readers can and do differentiate between degrees of density of usage of these two textual features. While some theoretical arguments within the field of literary study have proposed that linguistic features cannot define poetry because the same linguistic features are found in other textual genres, the Hanauer (1996) study shows that this argument is flawed in that it is density of presence of linguistic features that has to be related to, and not only a dichotomy of presence or absence of a feature. The result of the Hanauer (1996) study shows that the density of usage of the textual feature in the poetry categorization task is a factor that poetry readers relate to. In the same study (Hanauer 1996), a significant difference was found in the mean ratings of the two educational groups. The higher educational level gave higher poeticity ratings and used a smaller range of the scale. In a subsequent analysis of the same data that transformed the mean ratings into rank orders, Miall and Kuiken (1998) show that there is high correlation between the rankings of the

What we know about reading poetry

versions of the poems by all participants in this study. Thus while experts give higher ratings and used a smaller range of the scale, the final rank order decision about the ranking of the text is similar to those of novices. This result further advances our understanding of the internal representation of structural knowledge in poetry readers and the differences between expert and novice knowledge. Both novices and experts are sensitive to differing degrees of density in the usage of these two linguistic features; however, experts, who it can be assumed have wider experience with poetry, are willing to accept lower levels of density as poetic. This suggests that the internal representation of structural knowledge of a literary expert includes a wider set of linguistic options that can be used to define a poem. The four studies reported here present some evidence as to the way poetry is categorized and the way this knowledge is represented in the mind of both expert and novice literary readers. Specific textual features, such as graphic and phonetic form, would seem to clearly mark a text as a poem. As pointed out by languagedriven descriptions of poetry, poetic language would seem to be distinctive and easily recognized by wide groups of readers. As seen in these studies, both expert and novice readers have structural knowledge of the linguistic features of poems and they use this knowledge when required to make a poetry genre categorization. This structural knowledge of the genre of poetry is quite sophisticated in that it differentiates between degrees of usage of these structural features. Experts differ from novices in that they have a wider set of linguistic options that can be used to define a poem. As pointed out by reader-driven descriptions of poetry reading, this difference between experts and novices shows the importance of educational background in situating the role of textual features in defining poetry. There is also some indirect evidence that while experts and novices might have acquired similar structural knowledge of poetry, the experts have additional knowledge of literature and possibly specific functional and procedural knowledge of poetry. 3.2 Attention and representation in poetry reading The previous section discussed the way poetry is categorized. The current section will deal with the issues of how the reader’s attentional resources are used while reading a poem and how information is represented. The empirical investigation of the reading of poetry has focused on the question of whether readers of poetry pay attention to and retain in memory the linguistic and textual features of the poem while reading. Psycholinguistic research on narrative and informational texts suggests that proficient readers automatically process surface information and do not retain this information beyond approximately 80 milliseconds (Bransford and Franks 1971; Gernsbacher 1985, 1990; Sachs 1967). Narrative and informational text processing models assume that proficient readers pay attention only to the process of constructing meaning from semantic information (Gernsbacher 1990;

119

120 David Ian Hanauer

Kintsch and Van Dijk 1978). The description of poetry reading presented by Formalism proposes that while reading a poem, the reader is constantly made aware of the surface level information of the poem. From a Formalist viewpoint, the reader is unable to ignore the linguistic and textual patterns of the poem. Readerdriven approaches to poetry reading have placed much more emphasis on the role of conventions and strategies in the poetry reading process. In general, two different types of measure were used to investigate these issues: on-line measures of processing and post-processing measures. The on-line measures included reading time, on-line underlining and measures of the electrical activity of the cerebral cortex (Hanauer 1998a,
b; Hoorn 1996; Van Peer 1986). Post-processing measures included cued recall and recognition tasks (Hanauer 1998a,
b; Van Peer 1990). Although, these studies employed different research methodologies, the results suggest that readers pay attention to surface information while reading poetry and retain in memory some degree of surface information. In one of the first in-depth empirical investigations of poetry, Van Peer (1986) operationalized and then tested the Formalist and Stylistic concept of foregrounded linguistic structures in poetry. According to this theory, the reader will consider density of parallel and deviant linguistic patterns as important for the understanding of the poem. Using this theory as an analytical tool for the analysis of poetry, Van Peer defined those sections of a poem that are foregrounded and predicted that readers would pay attention to them. Van Peer used three measures to test his prediction: an underlining task, an importance ranking task and a cued recall task. The results of his study found that readers ranked those lines with density of foregrounding as important for the understanding of the poem and underlined those sections that were highly foregrounded as important for discussion. In addition, in the post-processing condition, readers had higher recall of the foregrounded sections of the poem than the rest of the poem. Van Peer’s results suggest that readers are paying attention to the foregrounded elements of the poem and are assigning them importance for understanding the poem. In addition they remember and are able to reproduce some of the specific surface information of the poem from the foregrounded sections. These results do not differentiate between language-driven and reader-driven descriptions of poetry reading. What the results show is that foregrounded language is related to during reading and hence is recalled. Several researchers have shown that readers and listeners do have surface information recall for language that is surprising or highly interactive (such as jokes) (Gibbs 1981; Keenan et al.1977; Murphy and Shapiro 1994). The argument that can be proposed on the basis of this study is that if there is foregrounded language in a poem the reader will relate to it and hence may recall some of the surface information of the poem. In a later study Van Peer (1990) tested the role of meter in the aesthetic response, recognition and recall of a poem. Participants were presented with a

What we know about reading poetry

metrical or un-metrical version of the same poem. The participants rated the poem on a semantic differential scale to evaluate their aesthetic response to the poem and were asked to indicate whether a line of poetry that had been manipulated for metrical and content factors was in the original poem. Finally, participants were asked to complete a cued recall task. The results of the study revealed differences in the aesthetic evaluation of the two versions. In addition a significant difference was found on the recognition task but not on the recall task. The metrical version of the poem and the metrical lines were better recognized than the non-metrical lines. The results of van Peer’s study show that readers were aware of the presence of meter in the initial reading and encoded some aspects of this information in memory. The increased ability to recognize a line of poetry as a result of its metrical form shows that metrical information is attended to during poetry reading. The increased aesthetic evaluation of the poem as a result of the presence of metrical form indicates that the meter of a poem is a factor in the reading of poetry. In a truly exceptional study Hoorn (1996) used measures of the electrical activity of the cerebral cortex to evaluate Structuralist and Formalist theories of the functional aesthetic effect (Barthes 1970; Jakobson 1960). Both Barthes and Jakobson consider the reader of poetry to be affected by patterns of sound and meaning and that sound patterns will interact with semantic patterns. Specifically, Jakobson proposed that the poetic function is the result of phonological patterns redefining semantic relations by overriding linear meaning construction. As reviewed in Hoorn (1996), previous research using electroencephalogram (EEG) of electrical activity in the cerebral cortex had demonstrated the presence of a signal — denoted N400–which occurred when experimental subjects were presented with semantic deviations. Hoorn’s study used the analysis of this signal as the measure of phonological and semantic effects in poetry reading. This measure shows on-line detection of the deviation. Hoorn presented participants with four line verses from famous Dutch poets. The verses were manipulated so that the last word of the last line was in one of four conditions in relation to the previous lines — rhyme/ semantically correct, non-rhyme/semantically correct, rhyme/semantically incorrect, and non-rhyme/semantically incorrect. EEGs were measured for the presentation of this last word. Hoorn predicted significant differences in the grand mean EEG average for the rhyme condition when compared to the non-rhyme condition and for the semantically correct condition when compared to the semantic incorrect condition. In addition, it was predicted that semantic incongruity would be enhanced by the rhyming condition as reflected by an interaction between the semantic and rhyming conditions. This prediction was based on the theoretical assumption that rhyme will create an expectation of relation that will be refuted by the semantic deviation and result in an intensified N400 signal. The results found a main effect for semantics and for phonology. However, the effect for semantics is sustained across the time frame for the measure, whereas the effect for phonology

121

122 David Ian Hanauer

is more fragmented. Interestingly, the main effect for phonology resulted from the readers’ detection of non-rhyme rather than rhyme. Phonology and semantics were found to interact but in the non-rhyming rather than rhyming condition. Specifically, non-rhyme was found to intensify the reader’s surprise at semantic deviation. This result is contrary to the researcher’s original hypothesis that semantic incongruity would be enhanced by the rhyming condition. The researcher’s original hypothesis was based on the idea that presence of rhyme will make the reader expect a semantic relation and so when it is refuted the surprise will be more intense. However, it is possible that based on the previous priming lines in the short poem the reader expected to find both a semantic relation and a rhyming word. When both were refuted this intensified the N400 signal. This result suggests that the readers are monitoring rhyme and semantic relations during poetry reading. Attention seems to be directed towards this information and this information seems to activate a series of predictive inferences relating to expected phonetic and semantic forms. The results of this study present compelling evidence of the reader’s involvement with different levels of linguistic information when reading a poem. The reader detects changes in the phonological as well as semantic information in the poem. The measure used in this study measures on-line brain activity and as such presents an unusual view of mental processes while reading poetry. Hanauer (1997b, 1998a,
b) in a series of studies used the measure of cued recall to investigate memory for surface information during poetry reading. In Hanauer (1998a) participants were presented with a poem that had been manipulated for its phonological and graphic information and was presented in two conditions as a poem or in an unspecified condition. The participants were required to read the poem and then complete a cued recall task. The measure was the number of exact words that were recalled from the poem. In order to have verbatim recall of a text, attention has to be directed at the surface information during reading. The ability to recall a text verbatim is a measure of the amount of attention paid to surface information during reading. The results of this study found no effect for the genre specification of the text or for the phonological information. A main effect was found for graphic information. Specifically, the original poetry format of short lines and stanzas was found to enhance surface information recall. The results show that readers were responding to graphic information and that this is a factor in the reading of poetry. The graphic layout of the poem enhanced recall of specific surface information suggesting that the graphic information is encoded during the reading process and may provide a visual template that increases the ability to recall specific surface information. Another option is that the graphic information activates genre-specific processing that involves directing attention at the surface information. On the other hand, the study does not provide evidence that readers are directing attention at phonetic information or that phonetic information increases verbatim recall. This is in contrast to the results found in Hoorn (1996)

What we know about reading poetry

reported above. However, the Hanauer (1998a) study uses a post processing measure whereas Hoorn (1996) uses an on-line processing measure. It is possible that the readers in the Hanauer (1998a) study did direct attention to phonetic information but did not retain this information in the post processing stage. But there is no direct evidence for this in the Hanauer (1998a) study. The results of this study also show that the genre categorization of the poem did not increase verbatim recall. A reader-driven description of poetry reading would hypothesize that the genre categorization would increase verbatim recall as a result of the application of poetry reading conventions. However, as stated by Hoffstaedter (1987), a poem has to have the structural features as well as the external genre categorization. The Hanauer (1998a) study does not provide evidence to support the hypothesis that an external genre categorization in itself increases the amount of attention directed at the verbatim information or the recall of this information. In a second study Hanauer (1998b) used a within subject design to compare between the reading of authentic poetry and encyclopedic items. This study aimed at providing some global information that can characterize the reading of these types of genre. The two genres were compared in relation to three measures — cued recall, reading time and self-evaluations of comprehension. The results showed that poetry reading was characterized by significantly slower reading times, significantly increased levels of surface information recall and lower self-evaluations of comprehension than encyclopedic items. As reflected in the longer reading times and increased surface information recall, readers spend more time and direct more attention to surface information when reading poetry than when reading encyclopedic items. The five studies reported here present some evidence on the information that is being attended to and represented in the poetry reader’s mind. In general there is some support for the position that poetry reading involves directing attention at surface information and that this information is retained in the post reading stage. The Van Peer (1990) and Hoorn (1996) studies suggest that attention is directed at phonetic information and that this information is retained in the post reading stage. In addition the Hoorn (1996) study provides some evidence that phonetic and semantic information is used to construct predictive inferences during the poetry reading process. Hanauer’s (1998a) study supports the position that graphic information is related to and retained in the post reading stage. Taken together these studies provide some evidence that poetry reading involves multiple levels of linguistic information. However, these studies do not provide detailed information on the relationship between specific surface information and the processes of reading. More specific psycholinguistic research that considers the causal relationship between different levels of linguistic information and specific processing outcomes in poetry reading is needed. The theoretical accounts of language-driven processing receive general support for the position that poetry reading involves

123

124 David Ian Hanauer

directing attention at surface information; however, the specific accounts proposed to date do not seem to provide a clear set of empirically supportable predictions. For example, the theoretical positions such as the Formalist position in Hoorn (1996) and Hanauer (1998a) did not predict the role of phonetic information as it was expressed in these specific studies. In a similar way, there was no clear evidence in these studies that conventions of reading direct the poetry reading process without the linguistic features of a poem. In the Hanauer (1998a) study external genre categorization did not enhance recall. An additional problem is that the studies reviewed do not allow a clear differentiation between reader-driven and language-driven descriptions of poetry reading. The studies demonstrate a role for different levels of linguistic information but do not allow clear empirical statements to be made as to the cause(s) of this description of poetry reading. The direction of attention to phonetic, graphic and semantic information may result from the specific features of the poem or be a result of the conventions of poetry reading. Within a reading research perspective it is likely that both the linguistic features of the text and the conventions of the reader play a role in directing the reading process. In the case of poetry reading, this role seems to involve directing attention at multiple levels of linguistic information. The result of the special linguistic features of the poem and the conventions of reading seem to result in the general description of poetry reading provided in the Hanauer (1998b) study — slow reading times, difficulties in comprehension and attention to and retention of surface information. 3.3 Meaning construction processes while reading poetry The previous section discussed attention and representation in poetry reading. The current section will deal with the meaning construction processes of poetry reading. Not many studies have been conducted on the construction of meaning while reading poetry and those that have been conducted provide general descriptions of this process. The preferred methodology for this issue has been verbal protocols or open interview data. In an early interview study De Beaugrande (1985) specified a series of responses that occur while trying to orally interpret a poem. These responses include hedging, citing, paraphrasing, normalizing, and generalizing. The overall picture of meaning construction is one in which the reader attempts to transform the poem into language that is more accessible to the reader. In another early study Hoffstaedter (1987) using a verbal protocol approach defined three types of poetry reader. These three types can be seen as a continuum which has at one pole a reader who is mainly involved in direct paraphrase that is close to the original text and at the other pole, a reader who searches for additional meanings in order to resolve ambiguities and fill in gaps within the poem. In a later study Harker (1994) investigated the meaning construction processes of tenth grade students reading poetry.

What we know about reading poetry

The main finding of his study was that readers tended to produce paraphrases rather than interpretations of the poems. This result seems to echo Hoffstaedter’s description of readers who stay close to the text and are involved in direct understanding of the poem. None of these studies produced strong evidence that reading poetry directed novice literary readers to produce extensive interpretations. One investigation of this issue that produced particularly interesting results was Martindale and Dailey’s (1995) study of agreement between interpreters of poetry. Within the field of literary studies, the accepted position on this issue is that readers of poetry do not agree and that reading poetry will elicit a multiplicity of contrasting views. Martindale and Dailey presented participants with a series of thirteen poems and asked them to rate the qualities of the poems on a series of semantic differential scales and then write open interpretations of the poems. Significantly high levels of agreement were found between the participants on both these measures. This result suggests that for novice readers agreement over the interpretation of poetry might be higher than has been presented in theoretical discussion. The results of these studies suggest that for novice readers, the initial effort in constructing meaning is directed at normalizing and paraphrasing the poetic text. In other words, novice readers did not show evidence of constructing elaborate interpretations of the poems or evidence of the ability to use different levels of linguistic information in order to construct literary interpretations. These results would suggest that expert literary knowledge is required in order to construct interpretations of this kind. However, more research is needed to investigate the meaning construction processes of expert poetry readers. The results in relation to novice readers support the New Critical idea that expert literary knowledge is required to transform the linguistic structures of the poem into interpretations. Martindale and Dailey’s result of agreement between poetry readers may result form the level of the literary interpretation. If readers are basically paraphrasing the poem, higher levels of agreement can be expected than if they are using more sophisticated interpretation strategies.

4. Conclusion This chapter has reviewed the main theoretical positions and empirical studies relating to the process of reading poetry. As presented above the central arguments between the various schools of interpretation have related to the emphasis that needs to be placed on linguistic and textual structures or on the reader’s experience and background during poetry reading. Collectively the data reviewed in this chapter suggests that the initial stages of poetry reading are highly influenced by the linguistic and textual structures of the poem. Linguistic and textual structures are detected by readers during the reading process and are used to categorize the text as

125

126 David Ian Hanauer

a poem. Readers seem to be sensitive to the presence of linguistic and textual patterns in poetry and these patterns produce aesthetic effects and enhance memory. However, the studies of meaning construction suggest that while readers may detect linguistic and textual patterns during poetry reading, the novice literary reader does not know how to transform this initial recognition of patterns into an interpretation. It is possible that strategic poetry reading and conventions of poetry may present options for incorporating initial responses to patterns into information that can be used in interpretation. Empirical research of poetry has demonstrated the initial detection of linguistic patterns but knowledge is still needed as to how these patterns evolve into interpretations.

References Alderson, C. and Short, M. 1989. “Reading literature”. In Reading, Analysing and Teaching Literature, M. Short (ed.), 71–119. London: Longman. Barthes, R. 1970. “Elements of semiology”. In Writing Degree Zero + Elements of Semiology, A. Lavers and C. Smith (eds.), 58–88. Boston: Beacon Press Baynham, M. 1995. Literacy Practices: Investigating Literacy in Social Contexts. London: Longman. Bazerman, C. 1988. Shaping Written Knowledge: The Genre and Activity of the Experimental Article in Science. Madison. WI: University of Wisconsin Press. Beaugrande, R. de 1985. “Poetry and the ordinary reader: A study of immediate responses”. Empirical Studies of the Arts 3: 1–21. Bleich, D. 1975. Readings and Feelings: An Introduction into Subjective Criticism. Urbana: National Council of Teachers. Bleich, D. 1978. Subjective Criticism. Baltimore: John Hopkins University Press. Bransford, J. D. and Franks, J. J. 1971.“The abstraction of linguistic ideas”. Cognitive Psychology 2: 331–350. Culler, J. 1975. Structuralist Poetics. Ithaca, NY: Cornell University Press. Derrida, J. 1978. Writing and Difference. (A. Bass, Trans.). London: Routledge. Fish, S. 1967. Surprised by Sin: The Reader in ‘Paradise Lost’. New York: St. Martin’s Press. Fish, S. 1980. Is There a Text in this Class: The Authority of Interpretive Communities. Cambridge, MA: Harvard University Press. Freedman, A. and Medway, P. 1994. Genre and the New Rhetoric. London: Taylor & Francis. Freire, P. 1985. The Politics of Education. London: Macmillan Gernsbacher, M. A. 1985. “Surface information loss in comprehension”. Cognitive Psychology 17: 324–363. Gernsbacher, M. A. 1990. Language Comprehension as Structure Building. NJ: Lawrence Erlbaum. Gibbs, R. W. 1981. “Memory for requests in conversation”. Journal of Verbal Learning and Verbal Behavior 20: 630–640. Hanauer, D. 1994. Poetic Text Processing. Unpublished doctoral dissertation Bar-Ilan University, Ramat-Gan, Israel. Hanauer, D. 1995. “The effects of educational background on literary and poetic text categorization judgments”. In Empirical Approaches to Literature, G. Rusch (ed.), 338–347. Siegen, Germany: LUMIS Press.

What we know about reading poetry 127

Hanauer, D. 1996. “Integration of phonetic and graphic features in poetic text categorization judgments”. Poetics 23: 363–380. Hanauer, D. 1997a. “Student teachers’ knowledge of literacy practices in school”. Teaching and Teacher Education 13: 847–862. Hanauer, D. 1997b. “Reading poetry and surface information recall”. In The Systematic and Empirical Approach to Literature and Culture as Theory and Application, S. Tötösy de Zepetnek and I. Sywenky (eds), 453–469. Siegen: LUMIS Press. Hanauer, D. 1998a. “Reading poetry: An empirical investigation of Formalist, Stylistic and Conventionalist claims”. Poetics Today 19: 565–580. Hanauer, D. 1998b. “The genre-specific hypothesis of reading: Reading poetry and reading encyclopedic items”. Poetics 26: 63–80. Harker, J. W. 1994. “‘Plain sense’ and ‘poetic significance’: Tenth-grade readers reading two poems”. Poetics 22: 199–218. Hoffstaedter, P. 1987. “Poetic text processing and its empirical investigation”. Poetics 16: 75–91. Hoorn, J. 1996. “Psychophysiology and literary processing: ERPs to semantic and phonological deviations in reading small verses”. In Empirical Approaches to Literature and Aesthetics, R. J. Kruez and M. S. McNealy (eds), 338–358. Norwood, NJ: Ablex. Iser, W. 1978. The Act of Reading: A Theory of Aesthetic Response. Baltimore: John Hopkins University Press. Jakobson, R. 1960. “Linguistics and poetics: Closing statements. In Style in Language, T. Sebeok (ed.), 350–377. Cambridge, MA: MIT Press. Johnson, B. 1980. The Critical Difference: Essays in the Contemporary Rhetoric of Reading. Baltimore: John Hopkins University Press. Kamberelis, G. 1995. “Genre as institutionally informed social practice”. Journal of Contemporary Legal Issues 6: 115–172. Keenan, J. M., MacWhinney, B. and Mayhew, D. 1977. “Pragmatics in memory: A study in natural conversation”. Journal of Verbal Learning and Verbal Behavior 16
: 549–560. Kintsch, W. and Van Dijk, T. 1978. “Toward a model of text comprehension and production”. Psychological Review 85: 363–394. Kress, G. 1989. Linguistic Processes in Sociocultural Practice. Oxford: Oxford University Press. Levine, K. 1986. The Social Context of Literacy. London: Routledge & Kegan Paul. Martindale, C. and Dailey, A. 1995. “I. A. Richards revisited: Do people agree in their interpretations of literature”. Poetics 23: 299–314. Meek, M. 1991. On Being Literate. London: Bodley Head. Miall, D. and Kuiken, D. 1998. “Forms of reading: Recovering the self as reader”. Poetics 25: 327–342. Miller, J. H. 1977. “The critic as host”. Critical Inquiry 3: 278–85. Mukarˇovsky˘, J. 1964. “Standard language and poetic language”. In A Prague School Reader on Esthetics, Literary Structure and Style, P. Garvin (trans)(ed.), 16–30. Washington: Georgetown University Press. Murphy, G. L. and Shapiro, A. M. 1994. “Forgetting verbatim information in discourse”. Memory & Cognition 22: 85–94. Pratt, M. L. 1977. Toward a Speech Act Theory of Literary Discourse. Bloomington: Indiana University Press. Rosenblatt, L. M. 1978. The Reader, the Text, the Poem: A Transactional Theory of the Literary Work. Carbondale: Southern Illinois University Press. Sachs, J. S. 1967. “Recognition memory for syntactic and semantic aspects of connected discourse”. Perception & Psychophysics 2: 437–442.



128 David Ian Hanauer

Schauber, E. and Spolsky, E. 1986. The Bounds of Interpretation: Linguistic Theory and Literary Text. Stanford, CA: Stanford University Press. Shuman, A. 1986. Storytelling Rights: The Uses of Oral and Written Texts by Urban Adolescents. Cambridge: Cambridge University Press. Steen, G. 1999. “Genres of discourse and the definition of literature”. Discourse Processes 28: 109–120. Swales, J. 1990. Genre Analysis. Cambridge: Cambridge University Press. Van Dijk, T. and Kintsch, W. 1983. Strategies of Discourse Comprehension. New York: Academic Press. Van Peer, W. 1986. Stylistics and Psychology: Investigations of Foregrounding. Wolfboro, NH: Croom Helm. Van Peer, W. 1990. “The measurement of meter: Its cognitive and affective functions”. Poetics 19: 259–275. Wimsatt, W. K. and Beardsley, M. C. 1954. The Verbal Icon: Studies in the Meaning of Poetry. Lexington: University of Kentucky Press. Zwaan, R. 1991. “Some parameters of literary and news comprehension: Effects of discourse type perspective on reading rate and surface-structure representation”. Poetics 20, 139–156. Zwaan, R. 1993. Aspects of Literary Comprehension. Amsterdam: John Benjamins. Zwaan, R. 1996. “Toward a model of literary comprehension”. In Models of Understanding Text, B. K. Britton and A. C. Graesser (eds), 241–255. Mahwah, NJ: Erlbaum.





A renaissance perspective on the empirical study of literature An example from psychophysiology* Johan F. Hoorn Free University Amsterdam

Perhaps the most interesting aspect of the work of Elrud Ibsch is her constant struggle for empirical validity of literary theory on the one hand without losing sight of hermeneutical relevance on the other. Traditional scholars who claim that literature cannot be understood by collecting data and calculating statistics stand up to a fiery opponent. In discussing the hermeneutic style of argumentation, Ibsch (1996: 31) counters: “The persuasiveness of the argument relies heavily on the number of plausible examples (even in hermeneutic scholarship, quantities are of importance!).” But, those who merely want to reduce highly complex matters of reading literature to a nicely controlled laboratory experiment have to face the same adversary. Ibsch (1996: 33) warns “against research that is done for the purpose of vigorously testing miniature hypotheses. … what we really need is the empirical validation of our intuitions.” Taking seriously the demand that we should validate our intuitions, might there be a way to safeguard methodological purity while not being trivial concerning content? Moreover, if psychological methods should enrich literary insights, should not literary studies enrich psychology? How can both these disciplines interact and enrich one another? An answer to these questions may lie in cultural history.

1.

The renaissance perspective

In the seventeenth century, the Dutch Republic became increasingly internationally oriented. Penetrating the world market and struggling out of Spanish domination, the Dutch had to enter into negotiations with trade partners, allies, and adversaries in many a foreign tongue. Such a competitive attitude combined with the need of cultural adaptability was not restricted to affairs of commerce and state. In a fast

130 Johan F. Hoorn

widening vista of cultures, poets such as P. C. Hooft (Tuynman and Van der Stroom 1994) wanted to improve on the work by Horace, Lucretius, Petrarc, Marino, and Lorenzo de Medicis. To do so, they mastered the respective foreign languages, and first became translators. Then they gained craftsmanship by imitating the styles and masterpieces of others. And finally attempts were made to be better than the original foreign examples. Vondel, for instance, attempted to replace the pagan contents of the Classic plays with Christian themes. De gront des Christendoms was noch onbestorven, en de tooneelspelen werden Cybele, der gedroomde Goden moeder, een groote afgodinne, ter eere gespeelt … [The ground of Christianity was not firm yet and the plays were performed in honor of Sibyl, the dreamed-up mother of Gods, a great Ateles … (author’s translation)] (Van den Vondel: 1654).

The threepartite procedure of translatio, imitatio, and aemulatio stimulated a large stream of publications, still renowned today for its outstanding quality. However, it was not only literature which flourished. The cultural climate also attracted scientists and philosophers from all over Europe to work and publish in the Netherlands, instead of being burned at the stakes in their home countries. Among those scientific immigrants was Descartes, entering Holland as a ‘Soldier of Orange’ in the French regiments of Maurits and leaving it as a well-known (although not always appreciated) philosopher and mathematician. In The Optics (1637), Descartes made “a strong distinction between the mechanical transmission of sense data (sensation) and non-mechanical inferences on those sense data (thinking)” (Kirkeboen 1998). In doing so, Descartes emphasized a duality between brain and mind that stimulated the study of mechanics (body) on the one hand and mathematics (mind) on the other, but without reconciling the two modes. How may the renaissance perspective on writing literature help the present day study of literature? In trying to validate intuitions, empirical studies of literature often borrow theory and methods from the social sciences. Yet one of the major difficulties in building bridges between two areas is that the result causes confusion. Therefore, the first step in an empirical investigation of literature should be to carefully translate notions from for instance literature into psychological terms. Only then will the unavoidable discrepancies between theoretical concepts be highlighted. New theory may be needed to fill the gaps. After the translation of theories, the second step is to choose an outstanding example of a psychological experiment to imitate. If the translation into psychological theory is correct, the appropriate experiments should also correspond with literary theory. Knowing what an outstanding example is, requires understanding of ‘the foreign language’. In-depth discussions with colleagues in the other field and reading the most respected psychological journals may fine-tune that skill. The third step, outperforming experimental psychology, is still an ideal. This may require

A renaissance perspective on the empirical study of literature

many cycles of imitation, and innovation, but there may be a possibility of achieving success. With the rise of modern (neuro)psychology, the Cartesian concept of mind and matter as strictly separate entities is disappearing. This means that literary questions are within the reach of brain studies. One of the complaints in psychology, however, and particularly in psychophysiology, is that techniques such as fMRI and MEG may improve the spatial and temporal resolution of measurements but not psychological theory. … it will take more than technological developments in imaging techniques for us to make significant progress in understanding the neural bases for language. … we will need highly detailed processing theories of language that will provide the functional pieces that map onto neural events. (Caramazza 1997: 135)

In other words, the complaint is that psychology moves towards science (body) at the cost of cognition (mind). By contrast, literary theory is preoccupied with the mind while neglecting the reality of the body. Only few counterexamples exist (e.g. De Jong 1981; Hoorn 1996; Konijn 1991). Where literary theory could enrich (and perhaps emulate) psychology is at the stimulus level. Whereas psychology focuses on the simplest of stimuli (dots, dashes, words), literary theory studies the most complex ones (poems, novels, genre, culture). However, such pretensions only can be maintained if methods, statistics, and theoretical sophistication are as good as or better than those found in psychology. There are examples of psychologists who have benefitted from literary insights. For example, Šklovsky’s (1917/1965) theory on linguistic deviation was influential for Berlyne’s (1960, 1971) arousal theory. Sartre (1934) affected the work on emotion psychology by Frijda (1986). Needless to say that Freud was greatly inspired by Greek tragedy. In view of “validating our intuitions” (Ibsch), however, the focus will be on translation from literature into psychology to subsequently use the methods and experimental set up that go with the psychological paradigm. In the next section, let me pursue the first two steps (translation and imitation) by discussing the possibilities to translate certain notions from literary studies into psychological terms and then to try and imitate a study on the motor cortex with a study on literary metaphor processing.

2.

Translation of some basic concepts

It is often the case that literary scholars have a different understanding of notions from the empirical sciences than the ones held by scientists from different fields such as psychology and sociology. In traditional literature, for instance, interpreting a book is sometimes considered empirical because the book under study is in the real world.

131

132

Johan F. Hoorn

However, in the (social) sciences, empirical means to quantitatively test hypotheses in a sample from a particular population. The word ‘test’ also has different meaning. In traditional literary studies, one can ‘test’ one’s a priori intuitions about a book by simply reading it. In the (social) sciences, the researcher is not allowed to be his or her own instrument of measurement. The word ‘significant’ is borrowed from the (social) sciences because it sounds a bit like the semiotic notions of the French signifiant and signifié (De Saussure). In literature, a word can be ‘significant’ to indicate that it is ‘full of meaning’, whereas in the (social) sciences, only effects can be significant, meaning that the probability of an effect to occur is above a certain chance level. In traditional literary studies, the term ‘correlation’ between two things has the meaning of ‘correspondence’, ‘relationship’, or ‘connection’ of whatever kind, whereas in the (social) sciences, it is the mathematical estimation in how far one set of data fluctuates with another set of data. Let me now focus on the translation of ‘correspondence’ between words from literary theory into psychology. As reflected in its name Comparative Literature has positioned correspondence (and opposition) between literary objects as one of its major endeavors. For example, Barthes (1970: 71–72) claims that the associations between words are “similar and dissimilar, include a common and a variable element”. Lotman (1976: 88) speaks of “common features” to describe the “parallelism” between words. For the comparison at the level of letters, Jakobson (1981: 27, 39, 42) speaks of “parallelism”, “equivalence”, “likeness”, “similarity” and “difference”. “Equivalence” is probably the term most commonly used in literary theory to indicate some form of correspondence. However, equivalence in psychology follows the logical notion of ‘equality’ or ‘being identical’, whereas in literature it refers to what the psychologist calls ‘similarity’. In psychology, fortunately, similarity also is considered the basis of correspondence between objects or concepts (see Goldstone and Barsalou 1998). Tversky (1977) states that similarity is based on the common features, whereas dissimilarity flows from distinctive features. In the case of letters, the features of an S are the curves and slopes that make it more similar to the next S than to the vertical and horizontal lines of the H. Different from literary studies, however, Tversky states that the estimation of “how similar two objects are” depends on the ratio between the number of their common and distinctive features. This is a welcome correction of the unidimensional approach of Jakobson (1971: 104), who does not take into account that correspondence may simultaneously be mediated by difference. Jakobson states that in a poem, every correspondence between letters indicates semantic relatedness. Thus, semantic relations in poetry are found through identifying similar letters or letter sequences, which in its turn is based on recognizing that the slopes and curves of an S are different from the straight lines of an H. In literary studies, however, this approach reveals the difficulty that an explicit criterion is missing to decide in how far, for instance, the same letters in different fonts carry similar or different meaning compared to

A renaissance perspective on the empirical study of literature

identical letters in the same font. Jakobson does not resolve the question as to when similarity between letters changes into dissimilarity or the connected issue of how similarity relates to the strength of the semantic connectivity. This is also true for the issue of similarity between the meaning of words and concepts. To describe the contents of different word classes, Eco (1972: 116) provides a loose enumeration of features. Greimas (1966: 27) defines the “psychic substance” of words as sèmes, which he explains as qualities (following Russell) or properties. Translated into psycholinguistics (Miller and Johnson-Laird 1976: 249–250), the features of words can be considered to be their associations and the “psychic substance” of a word would be the set of features or associations it evokes in the reader. Greimas (1966: 71) argues that semantic equivalence between words is established by common connective sèmes, by identity of sèmes or by simple resemblance of sèmes. These are instances of the same idea that via the comparison of two feature or association sets, similarity is established between the words that evoke those sets. Again, however, it remains ambiguous as to when semantic equivalence (similarity) between words becomes dissimilarity (“isotopies met en opposition”, Greimas 1966:71) and how this is related to the strength of the semantic relatedness between words. What can be learned from the above translation of literary theory on equivalence into the psychological notion of similarity is that words evoke sets of features, which can contain formal as well as semantic features, based on perception and association. The comparison of those feature sets results in sets of common and distinctive features. The number of common features in relation to the number of distinctive features determines the estimation of similarity (and difference) between the words or their ‘semantic equivalence’. The difference between words has been a major focus in the work of, for instance, Šklovsky, Striedter, and Mukarˇovský. Šklovsky (1917/1965: 12) states that the technique of art is to make objects “unfamiliar” and forms “difficult” because the process of prolonged perception that unfamiliarity evokes is an aesthetic end in itself (also Striedter 1989: 23–24). Mukarˇovský (1964: 17–30) argues that formal and semantic deviations are not absent in standard language but that their degree and mode are different than in literature. ‘Unfamiliar’ means that the association between objects is not automatic but needs additional elaboration to find their common features (in psychology, such nonautomatic elaborations are called ‘control processes’). ‘Deviation’ indicates that certain expectations are violated and therefore, the recognition is postponed that the words should be seen in each others’ context. Translated into psychology, Raaijmakers and Shiffrin (1981) define familiarity as a measure of features that overlap with concurrently activated feature sets in the context. Imagine a modern poem that is entitled “SootHe”. In this poem, the word HuSH may be experienced as more familiar than the word LuLL. Both HuSH and LuLL share about an equal number of semantic features with SootHe (e.g. peace, calm, quiet), but HuSH also shares the formal features of S and H with SootHe.

133

134 Johan F. Hoorn

The more overlap of features, the higher the familiarity is and the weaker the deviation. If the word was not LuLL but BuLL, then BuLL does not match the semantic or the formal expectations that the title of the poem SootHe raises. Therefore, it would take more time to understand and respond to the combination SootHe-BuLL than to the pair SootHe-HuSH. Note, however, that literary theory becomes ambiguous at this point. According to Jakobson, what distinguishes literature from other language is the higher density of meaning established by formal-semantic similarity, whereas Šklovsky, Striedter, and Mukarˇovský claim that literature strives for more unfamiliarity and deviation, that is, more dissimilarity. In the Jakobsonian view, the pair SootHe-HuSH is more poetic than SootHe-BuLL, but for Šklovsky, Striedter, and Mukarˇovský it is the other way round. However, other literary theorists have indicated that ambiguity and polyvalence are strong characteristics of literature (e.g. Steen 1994: 142–143). Perhaps semantic dissimilarity (lull-bull) is made more acceptable by employing formal similarity (LULL-bULL), but whatever the answer may be, even the professional reader will often be in doubt while interpreting a poem. Ambiguity yields doubt and doubt is translated by psychologists in terms of errors (e.g. Coles, Scheffers and Fournier 1995). When readers are asked to ignore formal features and to judge whether the meaning of the words lull and bull are the same, they will probably make more incorrect decisions for ‘similar’ than when they judge lull and cow, because in the first pair formal similarity interferes. Thus, perceiving similarity is at the basis of familiarity and equivalence, whereas perceived dissimilarity is the basis of deviation. When similarity and dissimilarity are in (approximate) equilibrium (in the case of lullbull), ambiguity arises, which can be translated (and measured) in terms of the number of correct or incorrect decisions (error rate) in a matching task. Being in doubt means that the outcome is uncertain. Uncertainty precedes surprise and surprise is what makes many stories, characters, and poems exciting. By surprise I do not only mean the thriller-like suspense of ‘who dunnit’ but also the tension raised by an aptly chosen metaphor or an aptly delayed rhyme. Deautomatization as described by Šklovsky (1917/1965) was translated into psychological terms by Berlyne (1971: 145), who claims that “When a stimulus pattern fails to agree with an expectation that was aroused by what preceded it, we call it ‘surprising’.” Berlyne claims that surprise is measurable in the amount of arousal (for instance increased heart beats, wider pupils, more muscular tension) that the stimulus evokes. One postulate underlying his view is that the relation between arousal and cognition is aspecific. Hoorn (1997b: 272–273) argues, however, that measures which tap a more precise relation between cognition and physiological effect should be used. Otherwise, it is hard to discern whether higher-frequency heart beats reflect surprise or perhaps fear and one could not determine what the underlying cognitive processes are. Hoorn (1996) made use of particular components in the EEG (electroencephalogram) — the so called event-related brain potentials —

A renaissance perspective on the empirical study of literature

which appear to be sensitive to semantic and phonological deviations in reading four-lined verses in alternating rhyme. Miall (1995) gives an overview of psychophysiological indicators of foregrounding that can be used with more precision than the arousal measure. In Hoorn (1997a) and Hoorn (1997b: 265–310), research is reported on the cognitive-energetic surprise effects of literary metaphor processing. Metaphors of the form “the sun is a grape” (Elburg 1975:126) were used to test the assumption of the anomaly model on metaphor processing, that metaphors elicit a ‘shock’ of unexpectedness (Beardsley 1982: 267; Henle 1966: 182; Richards 1965: 28). The shock would indicate that the meanings of the words in a metaphor are literally incompatible, urging the reader to find a figurative interpretation. This ‘shock’ was envisioned as the N400, an amplitude in the EEG — averaged across subjects and stimuli — that is supposed to reflect semantic mismatch (Kutas and Hillyard 1980a,
b, 1981). Indeed, the N400 did occur in the expected ordinal pattern: Literal expressions (“the sun is a star”) evoked least N400-activity, metaphors intermediate, and anomalies (“the sun is a whit”) the largest N400-activity. However, the difference between the latter two was not statistically reliable. Because N400 preceded the parallel routes of literal and figurative processing, it could not be maintained that ‘the shock of surprise’ of metaphor was the impetus to the figurative stage. Adding the original poems to the expressions merely mitigated the differences. In sum, a line which includes your soothing builds up expectations about its continuation. When the subsequent words share enough features (your soothing is a relief), similarity between these words is high, the expectations are confirmed and no surprise occurs. When insufficient features are shared (your soothing is a bull), expectations are refuted, semantic incongruency or deviation occurs, which is accompanied by tension or surprise (perhaps reflected by the N400). This creates ambiguity, so that the reader is in doubt and makes more ‘errors’ when s/he has to decide what the meaning is (e.g. literal or figurative). Here, the word ‘error’ should not be taken as ‘misinterpretation’ or ‘fault’. It is merely a technical term in psychology to distinguish the decisions that confirm the answers expected by the task from those that do not. Actually, errors often offer a different view on earlier knowledge, allowing for change and modification of opinions. They make it possible to adapt to new (textual) circumstances. Errors are indicative for the uniqueness of an interpretation, which is based on the different choice patterns people make. Such individual variance can be estimated fairly precisely by calculating the mean across a sample of readers and the deviations from that mean. Whereas the mean describes the common trend in how a text is interpreted, the variance describes in how far readers have their individual opinions. In other words, the contrast between the mean and variance testifies to what extent literature can be translated into General Literature. The next section focuses on a study in mental chronometry that may be translated — step by step — into a psychophysiological experiment on the order of stages in literary metaphor processing.

135

136 Johan F. Hoorn

3.

From translation to imitation: A study on the motor cortex

Jakobson (1971: 688–689) already argued that it would be useful to enrich (poetic) language studies with knowledge that was derived from psychophysiological measures. In this section, I will try to copy a study on the motor cortex to the study of literary metaphor as an example of how excellent work in other disciplines could be used for our benefit. Although in reaction-time (RT) studies on human performance, the analysis of subject errors is widely used, RT-studies on metaphor processing often seem to overlook this helpful information source (for other problems, see Hoorn 2000). Coles et al. (1995) studied the use of (partial) errors to make inferences on stages of information processing. They defined partial errors as tendencies to make incorrect responses that do not result in an overt response. In choosing between two options, for example, subjects may tend to go for the first choice, then correct themselves and decide for the second. Although they are in doubt, subjects can make the right decision after all. These tendencies cannot readily be seen in the overt behavior, such as the choice for one of the decision keys, but they can be detected at several covert levels. The first covert level is the force of a key press or how hard a response device is squeezed, reflecting response activation at the level of the response itself. The kinds of partial errors detected in this way are generally accompanied by a correct response. The second covert level to detect partial errors at is muscle tension as recorded by an electromyogram (EMG), reflecting response activation more peripherally. The third is in the EEG. Measures of the lateralized readiness potential (LRP) reflect the relative activation of central mechanisms associated with response generation. The LRP is a potential in the brain that is linked with motor skills such as moving arms and fingers. Prior to the movement of, for example, the right hand, the LRP can be measured as a negative shift in the EEG, overlying the cortical motor areas at the left side of the brain. For the left hand, the right side is activated (contra lateral activation). Subjects err and researchers infer. Coles et al. (1995) state that under a number of different circumstances, systematic patterns of partial errors could be observed for the levels discussed in the previous paragraph. They showed that in choice RT-tasks, partial errors were more frequent when the target stimulus contained information that favored both responses (ambiguity) than when it contained information that favored only one response. The systematic, non-random, pattern of partial errors supported the inference that partial information about the stimulus is used to guide responses. Coles et al. based themselves — among other things — on the famous Eriksen flankers task (Eriksen and Eriksen 1974; Eriksen and Schultz 1979). In this task, subjects supposedly focus on the center, target letter in a 5-letter array. They should respond with the left- or right-hand as the target is an S or H. The distractors (or

A renaissance perspective on the empirical study of literature

flankers) are the letters next to the target. Flankers can be identical to the target (SSSSS vs. HHHHH) or different (SSHSS vs. HHSHH). Note that in abstraction, this task requires precisely those skills that Jakobson’s search for formal equivalence and opposition demands. The response device Coles et al. used was a squeeze dynamometer. Coles et al. defined the ‘degree of error’ in terms of the presence of electromyographic (EMG) and premature, low squeeze activity for the incorrectresponse hand. Thus, in trials with strings such as SSHSS, subjects were not expected to be distracted by the Ss and were supposed to respond to the H with the correct hand. Yet, even for correct decisions, Coles et al. detected partial errors in muscle tension or premature soft squeezes for the incorrect hand. Even when muscle contractions were absent, partial errors could be detected in the LRP. On certain trials with correct decisions (with the correct hand), the motor area in the brain that governed the movement of the incorrect hand showed a positive dip in the negative-going LRP around 200 milliseconds after stimulus presentation. Moreover, RTs for correct responses increased as a function of the degree of partial error. Because subjects were apparently also inclined to evaluate the incompatible stimulus type, Coles et al. inferred that partial information transmission took place in the Eriksen task. Put differently, although the target letter provided information that was directly associated with the correct decision, information from the flankers ‘leaked through’ and made the subjects doubt. This ‘response competition’ predicated by mutually affecting information sources can be used for a new experiment in metaphor research. In processing metaphors, the literal information that a metaphor provides (such as the dictionary meaning of the words) may ‘leak through’ to the figurative information (for instance the personal symbolic meanings) so that ‘response competition’ may occur between the decisions ‘it’s a literal expression’ or ‘it’s a metaphor’. Traditionally, understanding metaphor is considered a serial process. Authors such as Richards (1965), Henle (1966), and Beardsley (1982) posit that an expression is always interpreted literally first and then at a later stage, if the expression is a metaphor, it is interpreted figuratively. Between these two moments the expression would be perceived as an anomaly. Therefore, this 2-stage serial anomaly model (cf. Hoorn 1997b: 27–44) predicts that decisions for literal expression are fastest, followed by a covert response for anomaly, followed by decisions for metaphor, and that decisions for ‘pure’ anomalies are slowest. Using a classification decision task, however, Hoorn (1997b: 245, 281) found RT-patterns that contradicted the predicted serial order and underscored a parallel anomaly model (311–317). Decision accuracy was only perfect for anomalies, which were never judged as literals (246, 296). The parallel model suggests that anomalies are recognized best and fastest, followed by literals and metaphors, which end in a draw with a slight speed advantage for literals. Translating the inferences Coles et al. (1995) made for the Eriksen flankers task, it might be possible that in metaphor processing, too,

137

138

Johan F. Hoorn

information from literal to figurative interpretation (and vice versa) leaks through and that response competition occurs. There is evidence which corroborates such speculations. Hoorn (1997b) found that associations labeled ‘literal’ by the subjects for the one term of the metaphor were labeled ‘figurative’ for the other term and that literal expressions and metaphors were distinguished best on the intersection between features with such a combination of labels (157–158). Additionally, Hoorn (1997b) reports on a 2-choice RT-task in which the highest error rate (22.2%) was found for metaphors judged as anomalies (247). In a 3-choice task, literals that were never or only once judged as metaphors were processed faster than literals with more metaphor-errors (248). Such an increase in RTs for trials with higher degrees of error is exactly one of the characteristics of interference by another information source that Coles et al. have brought to our attention. Mimicking the results found by Coles et al., Table 1 shows the predictions of the serial and parallel anomaly model on degrees of error (LRP, EMG, squeeze) and RT for choices between two expression types. In following Coles et al., it is assumed that if the verification of an expression takes place early in the process, error rates will be low, because another option to compete with is still missing. For comparisons between literal and metaphor, the serial model foresees that the literal stage is completed before the figurative stage is executed. Therefore, no errors for correct literals (the process already stopped) will be found and more errors for metaphors will be elicited (the expression might be mistaken for literal). The same rationale applies to comparisons between literal expression and anomaly. Between metaphors and anomalies, mutual errors may occur: Metaphors may raise partial errors because they are seen as covert anomalies; ‘pure’ anomalies may be judged as inadequate metaphors. Partial error A

Correct M

Time

LRP high LRP low

M

Metaphor

A

Figure 1.Illustration of a decision between M (‘metaphor’) and A (‘anomaly’) when a metaphor is presented. Partial-error A-decisions for metaphors may be recorded in the contralateral LRP before the correct execution of the M-response.

The parallel model predicts about equal degrees of partial errors for literal and metaphor since they are processed about equally fast. Because literal expressions

A renaissance perspective on the empirical study of literature 139

Table 1. RT (partial) errors

Predictions of the serial and parallel anomaly model on degree of error (LRP, EMG, squeeze) and RT in a 2-choice task Serial Model L < (covert A) < M < A

Parallel Model A < L ≤ M

L no

vs.

M yes

L equal or fewer than M

vs.

M equal or more than L

L no

vs.

A yes

L yes, but fewer than vs. M

vs.

A no

M yes, covert A

vs.

A yes, inadequate M

M yes

vs.

A no

tend to be faster, they may yield fewer errors. For comparisons between literal and anomaly, anomalies are processed fastest. The verification of an anomaly takes place early in the process, so that another option to compete with is still missing. Literals are evaluated later in the process, so there may be errors that tend towards anomaly. Yet such errors are supposedly fewer than in comparisons with metaphors, because decisions for literal or metaphor are equally fast. For instance, if a metaphor is presented and the choice is between ‘metaphor’ (M) and ‘anomaly’ (A), then high degrees of partial errors should be measured in squeeze force, EMG, or LRP. Regarding the latter, metaphors should evoke low LRP-activity for initial A-decisions before high LRP-activity occurs (contralaterally) for correct M-decisions (Figure 1). To obtain stimuli that are comparable to the Eriksen letter-arrays, expressions should be constructed that are distributed in accordance with Table 2. The trends correspond to the distracting flankers, the pure expressions to the arrays with equal letters. Much of the results will depend on the quality of the stimuli. It is especially the distances between the diverse stimulus types which should be equal. Apart from the usual streamlining for word-frequency, lexical ambiguity, and so on (Hoorn 1997b: 63–114), this will take extensive pilot studies in which subjects rate the ‘purity’ or the strength of the trend involved. Only then may the effects on RT be estimated between the fastest and slowest processed information. The use of a movement-related brain potential in language research has already been used successfully. Van Turennout et al. (1997) explored the temporal aspects of semantic and phonological processes in speech production. They found evidence for early semantic activation and later phonological encoding of words. It seems that measuring LRPs may also be useful in metaphor research.

140 Johan F. Hoorn

Table 2.

Semantic stimuli that mimic the Eriksen and Eriksen (1974) letter-arrays

Pure Literal | Literal with metaphoric trend | Metaphor with literal trend | Pure metaphor

| = equal distances

Pure literal | Literal with anomalous trend | Anomaly with literal trend | Pure anomaly Pure metaphor | Metaphor with anomalous trend | Anomaly with metaphoric trend | Pure anomaly

The advantage of analyzing partial errors and LRPs over standard RT-analysis in metaphor research is that covert stages in the metaphor process (‘conclude anomaly’ in the serial model) can be detected, and that the possible information transmission between the parallel stages can be explored with more sophistication. Moreover, partial errors render more variables which may be used to decide between serial or parallel models (cf. Table 1). A disadvantage is that subjects only have two hands, so that 3-choice tasks cannot be brought into action. That may be considered a set-back because in metaphor processing, results sometimes strongly depend on the experimental setting (Hoorn 1997b: 245).

4. Discussion: Reconciliation of mind and matter In this chapter the renaissance perspective of translation, imitation (and perhaps emulation) was proposed as a way of reconciling the humanities with the (social) sciences. I have illustrated with examples from literature and psychology how concepts such as similarity, ambiguity, unexpectedness, and surprise can be translated from literary theory into psychology, and I have proposed that similarity and ambiguity are measurable in terms of numbers of partial errors in an imitation



A renaissance perspective on the empirical study of literature

of a psychological study on the motor cortex (Coles et al. 1995). Providing a new theory for artistic complex stimulus materials, such as metaphor, personification (metaphor and fictional character in one), or the virtual creatures that inhabit the new generation of video games, active worlds and search engines on the World Wide Web, may be one way in which empirical literary research can surpass and enrich psychological knowledge. Using the electrocortical activity of the motor cortex during response preparation to study the stages of metaphor processing may allow us to understand a fundamental aspect of human communication. Because the order of processing stages is dependent on their contents (for instance literal interpretation before figurative), the use of the lateralized readiness potential (LRP) of the motor cortex in a linguistic-semantic context may advance our knowledge of how literal and figurative information interact in our decision ‘yes, it’s a metaphor!’ In doing so, a step may be taken towards the reconciliation of two apparently opposite modes: mind and matter. Hermeneutic scholars of literature usually proclaim that matters of the mind cannot be materialized, whereas hard-nosed empiricists avow that nothing matters but matter. Considering the first group, Ibsch (1996: 24) states: “If the term ‘scientific’ is used at all, it serves to pretend that scientific claims in literary studies are of no use and obsolete.” Hopefully, if the proposed experiment is conducted, the hermeneuticians will recognize that at least some assumptions in traditional literature can be tested with methods of the sciences. Considering the second group, Ibsch (1996: 32) contends: “I am inclined to defend the thesis that all sciences are human sciences. In other words, even natural sciences are human sciences.” This is a viewpoint, of course, that also makes the testing of miniature hypotheses part of the humanities. Methodological purity can be achieved without the loss of content if excellent examples in experimental psychology are translated and imitated while using our traditional literary stimuli. Perhaps the latter is the hardest part. Complex stimuli demand much pretesting, so that implicit factors do not obfuscate our findings (see Hoorn 2000). However, if we succeed, perhaps even Descartes might admit that the study of the mechanics of the brain can be the motor for studying the mechanics of mind.

Note * I wish to thank David Hanauer for correcting my English.

141

142 Johan F. Hoorn

References Barthes, R. 1970. Writing Degree Zero and Elements of Semiology. Boston: Beacon. Beardsley, M. C. 1982. “The metaphorical twist”. In The Aesthetic Point of View, M. J. Wreen and D. M. Callen (eds), 263–280. Ithaca, London: Cornell University. Berlyne, D. E. 1960. Conflict, Curiosity, and Arousal. New York: McGraw-Hill. Berlyne, D. E. 1971. Aesthetics and Psychobiology. New York: Meredith. Caramazza, A. 1997. “Brain and language”. In Conversations in the Cognitive Neurosciences, M. S. Gazzaniga (ed.), 131–151. Cambridge, Massachusetts, London, UK: MIT. Coles, M. G. H., Scheffers, M. K. and Fournier, L. 1995. “Where did you go wrong? Errors, partial errors, and the nature of human information processing”. Acta Psychologica 90: 129–144. De Jong, M. A. 1981. Emotie en Respons, een Psychofysiologisch Onderzoek [Emotion and Response: a Psychophysiological Investigation]. Lisse, the Netherlands: Swets & Zeitlinger. Eco, U. 1972. Einführung in die Semiotik. München, Germany: W. Fink Verlag. Elburg, J. G. 1975. Gedichten 1950–1975 [Poems 1950–1975]. Amsterdam: De Bezige Bij. Eriksen, B. A. and Eriksen, C. W. 1974. “Effects of noise letters upon the identification of target letter in visual search”. Perception and Psychophysics 16: 143–149. Eriksen, C. W. and Schultz, D. W. 1974. “Information processing in visual search: A continuous flow conception and experimental results”. Perception and Psychophysics 25: 249–263. Frijda, N. H. 1986. The Emotions. Cambridge: Cambridge University. Goldstone, R. L. and Barsalou, L. W. 1998. “Reuniting perception and conception”. Cognition 65: 231–262. Greimas, A. J. 1966. Sémantique Structurale, Recherche de Méthode. Paris, France: Larousse. Henle, P. 1966. Language, Thought and Culture. Ann Arbor, USA: University of Michigan. Hoorn, J. F. 1996. “Psychophysiology and literary processing: ERPs to semantic and phonological deviations in reading small verses”. In Empirical Approaches to Literature and Aesthetics, R. J. Kreuz and M. S. MacNealy (eds), 339–358. Norwood, NJ: Ablex. Hoorn, J. F. 1997a. “Electrocortical evidence for the anomaly theory of metaphor processing: A brief introduction”. In The Systemic and Empirical Approach to Literature and Culture as Theory and Application, S. Tötösy de Zepetnek and I. Sywenky (eds), 67–74. Edmonton, Alberta, Canada: University of Alberta. Hoorn, J. F. 1997b. Metaphor and the Brain: Behavioral and Psychophysiological Research into Literary Metaphor Processing. Unpublished doctoral dissertation, Vrije Universiteit, Amsterdam, the Netherlands. Hoorn, J. F. 2000. “The hazard of hidden interactions. Reanalysis of designs in RT-studies on metaphor”. CLCWeb: Comparative Literature and Culture: A WWWeb Journal [On-line]. Available: http://clcwebjournal.lib.purdue.edu/clcweb00-3/hoorn2-00.html. Ibsch, E. 1996. “The strained relationship between the empiricist’s notion of validity and the hermeneutician’s notion of relevance”. In Empirical Approaches to Literature and Aesthetics, R. J. Kreuz and M. S. MacNealy (eds), 23–33. Norwood, NJ: Ablex. Jakobson, R. 1971. Selected Writings, II, Word and Language. The Haque, Paris: Mouton. Jakobson, R. 1981. Selected Writings, III, Poetry of Grammar and Grammar of Poetry. The Hague, Paris, New York: Mouton. Kirkeboen, G. 1998. “Descartes’ psychology of vision and cognitive science: The Optics (1637) in the light of Marr’s (1982) Vision”. Philosophical Psychology 11: 161–182. Konijn, E. A. 1991. “What’s on between the actor and his audience? Empirical analysis of emotion processes in the theatre”. In Psychology and Performing Arts, G. Wilson (ed.), 59–74. Lisse, the Netherlands: Swets & Zeitlinger.



A renaissance perspective on the empirical study of literature 143

Kutas, M. and Hillyard, S. 1980a. “Reading between the lines: Event-related potentials during natural sentence processing”. Brain and Language 11: 354–373. Kutas, M. and Hillyard, S. 1980b. “Reading senseless sentences: Brain potentials reflect semantic incongruity”. Science 207: 203–205. Kutas, M. and Hillyard, S. 1981. “ERPs to semantically inappropriate and surprisingly large words”. Biological Psychology 11: 99–116. Lotman, Y. M. 1976. Analysis of the Poetic Text. Ann Arbor, USA: Ardis. Miall, D. S. 1995. “Anticipation and feeling in literary response: A neuropsychological perspective”. Poetics 23: 275–298. Miller, G. A. and Johnson-Laird, P. N 1976. Language and Perception. Cambridge, Massachusetts: Harvard University. Mukarˇovský, J. 1964. “Standard language and poetic language”. In A Prague School Reader on Aesthetics, Literary Structure, and Style, P. I. Garvin (ed.), 17–30. Washington: Georgetown University. Raaijmakers, J. G. W. and Shiffrin, R. M. 1981. “Search of associative memory”. Psychological Review 88: 93–134. Richards, I. A. 1965. The Philosophy of Rhetoric. London, Oxford, England, New York: Oxford University. Sartre, J. P. 1934. Esquisse d’une théorie phénomenologique des émotions. Paris, France: Hermann. Šklovsky, V. 1917/1965. “Art as technique”. In Russian Formalist Criticism: Four Essays, L. T. Lemon and M. Reis (eds), 5–24. Lincoln: Nebraska University. Steen, G. 1994. Understanding Metaphor in Literature: An Empirical Approach. London, New York: Longman. Striedter, J. 1989. Literary Structure, Evolution, and Value: Russian Formalism and Czech Structuralism Reconsidered. Cambridge: Harvard University. Tuynman, P. and Van der Stroom, G. P. 1994. P. C. Hooft. Lyrische Poëzie, I [P. C. Hooft. Lyrical Poetry, I]. Amsterdam: Atheneum, Polak & Van Gennep. Tversky, A. 1977. “Features of similarity”. Psychological Review 84: 327–352. Van den Vondel, J. 1654. Lucifer. Berecht aen Alle Kunstgenooten, en Begunstigers der Tooneelspelen [Lucifer. Message to all Art Fellows and Patrons of Plays]. Amsterdam: Abraham de Wees. Van Turennout, M., Hagoort, P. and Brown, C. M. 1997. “Electrophysiological evidence on the time course of semantic and phonological processes in speech production”. Journal of Experimental Psychology: Learning, Memory, & Cognition 23: 787–806.





A rhetoric of metaphor Conceptual and linguistic metaphor and the psychology of literature* Gerard Steen Free University Amsterdam

The psychology of metaphor has received a great deal of attention in the last part of the twentieth century (e.g. Gibbs 1994; Honeck and Hoffman 1980; Mio and Katz 1997; Ortony 1993) and has even established a flourishing journal of its own, Metaphor and Symbol, but psychologists have not been overly concerned with literary metaphor or metaphor in literature, with the major exception of Raymond Gibbs. Students of literature with a psychological bent have attempted to redress this deplorable situation (Hoorn 1997; Steen 1994), but the state of the art is still unsatisfactory. Most psychologists interested in metaphor operate with a practical notion of the poetic, the literary, or the aesthetic, which does not come near a full view of the complexities involved in literary discourse processing (Steen 1994, 1997, 1999a,
b). And most literary scholars interested in metaphor simply lack the interest and psychological expertise to perform empirical and experimental work with a view to formulating generalizations about texts and processing, or explaining the effects of a particular text against such a background. Metaphor being such an obvious candidate for interdisciplinary co-operation between literary scholars and psychologists, the slow improvement of the state of the art is indicative of the great difficulties that have to be superseded by either side when it comes to actually constructing a common ground. Part of the problem lies in the neglect of the discipline that may act as a bridge between psychology and poetics, namely linguistics. Limiting our attention to the study of literary metaphor, one serious impediment to high-quality empirical work is formed by the difficulty of achieving an operational, reliable and valid identification of metaphor in literary texts. Psychologists are not linguists or discourse analysts, but they are great believers in good methodology and hence have to rely on methods and techniques offered by linguistics to be able to develop suitable experimental materials. However, a good technique for metaphor identification is sorely lacking in linguistics.

146 Gerard Steen

This may be surprising to those psychologists who have the impression that there is a solution to this problem called Cognitive Linguistics. For Cognitive Linguistics certainly is the dominant theoretical paradigm for metaphor studies in and outside literature for many linguists and literary scholars and it has served as a source of inspiration for many social scientists. Indeed, one of the particular strengths of the approach to metaphor launched by Lakoff and Johnson (1980) and developed by Lakoff (1987), Johnson (1987), Turner (1987), Lakoff and Turner (1989), and Turner (1991) has been to situate literary metaphor and literary discourse at the center of the enterprise from the beginning. The reason for this interest is easy to explain. The basic tenet of the Cognitive Linguistic approach is that metaphor is not some deviant rhetorical device that is restricted to literature and comparable domains of discourse, but that metaphor is fundamental, conceptual, conventional, and ubiquitous (for a summary, see Lakoff 1993). This means that metaphor in literature is also based in patterns of metaphorical thought and that the question arises how metaphor in literature differs from metaphor in other kinds of discourse. The answer given to this question by Cognitive Linguists is that, even though some literary metaphor may involve genuinely creative conceptual metaphors, literary metaphor generally makes use of conventional metaphors in either conventional or original ways. For instance, Lakoff and Turner (1989) have suggested that novel expressions of conventional conceptual metaphors are based on semantic extension, elaboration, questioning, and composing of conventional conceptual metaphors. Goatly (1997) offers a related answer, which concentrates on the various kinds of structural elaboration of metaphor in especially literary discourse. Another, again related, answer is that metaphor in literature may on the one hand be more varied than is suggested by Lakoff and Turner and others, but that on the other hand it also offers a conspicuous opportunity for exercising typically literary reading strategies (Steen 1994). Although one of these strategies includes the almost deliberate metaphoric processing of linguistic materials that are not necessarily based in conceptual metaphor themselves (Steen 1994), this does not imply that the experience of literary metaphor is solely due to literary processing, as seems to be claimed by Miall and Kuiken (1999). There have to be typical properties of the stimulus, too, which facilitate the activation of literary processing strategies. Before such conclusions may be drawn, however, there simply is still too much work to be done on the nature of literary metaphor as well as on the nature of literary metaphor processing to be able to draw any firm conclusions. It is the purpose of this chapter to present one prerequisite for such a program, the development of reliable procedure for metaphor identification and its relation to metaphor analysis and processing research. For if psychologists and literary scholars took a closer look at the results of Cognitive Linguistics, they would find that an actual procedure for metaphor identification in natural discourse is still lacking from the literature. The Cognitive

A rhetoric of metaphor 147

Linguistic references listed above provide many examples of metaphors identified in literary and non-literary discourse, but they do not specify how these metaphors ended up as part of the data while other expressions did not. It is particularly problematic how a linguistic expression is related to a presumably underlying conceptual metaphor, which is supposed to be the testing ground for metaphoricity, and, indeed, even the assumption of underlying metaphorical mappings has been questioned by psychologists (e.g. Murphy 1996, 1997). What is more, even if the existence of conceptual metaphors is granted at some level of reality, there is a serious methodological issue about how the identity, content and structure of any conceptual metaphor have been determined in the first place. This is of crucial importance to an approach that defines linguistic metaphor as derivative of conceptual metaphor, and it is an area where linguists and psychologists should cooperate to achieve further progress. The results of such joint research would be of great interest to linguists working on metaphor in texts and corpora, literary and non-literary, to psychologists working with metaphors in experimental materials, literary and non-literary, and to literary scholars who wish to do text, corpus or experimental work. Once we have formulated a procedure for metaphor identification in literary and non-literary discourse, empirical work may considerably profit from the results of Cognitive Linguistics. This is also the case because literary scholars who are interested in general patterns of language use have also turned to the Cognitive Linguistic approach for their work on metaphor (e.g. Freeman 1993, 1995, 1999; Shen 1987, 1989). But generally speaking, since literary critics tend to concentrate on individual texts and their interpretations of them, they are less concerned about general patterns of language use and the discipline that may help to identify and describe such patterns, linguistics. Indeed, there has developed a traditional tension in the Arts Faculties between the disciplines of linguistics and literary studies, with the effect that literary scholars prefer to develop their own meta-language for the description of language and texts, even though they pay frequent lip-service to Roman Jakobson’s (1960) famous dictum that poetics cannot do without linguistics. Some of the resulting theories of metaphor that have been popular in poetics, such as the interaction theory (Black 1993; Kittay 1987; Reinhart 1976; Richards 1936; Ricoeur 1979), have also been used for psychological research, but they do not offer instruments for metaphor identification. Moreover, if they do contain guidelines for metaphor identification, they are relatively restricted, including only those metaphors that are ‘alive’ as opposed to the ‘dead’ ones that are part and parcel of everyday language. This is an attitude that uncomfortably fits with the present-day Cognitive-Linguistic climate of metaphor research, which sees a continuum between conventional and novel metaphor, an opposition that manifests a great degree of association with the distinction between dead and alive metaphors in poetics.

148 Gerard Steen

This background has provided the motivation for the research on a metaphor identification procedure done by an informal small-scale international interest group originating from the Poetics And Linguistics Association. A handful of linguists and literary scholars have worked on the development of such a procedure for literary and non-literary discourse, and the first results are now beginning to be published. It is the aim of this chapter to present this work to a larger audience and to demonstrate that it has much potential for further elaboration. More specifically, this chapter will go into the possibilities for metaphor identification and analysis in poetry, which offers greater rhetorical variety of metaphorical language than most types of discourse. Poetry thereby constitutes a challenge to any theory for metaphor and any operational definitions that are based on such theories. The illustrative materials for this chapter come from a small set of experimental materials which I have used in a study in progress utilizing a metaphor recognition task. They are five nineteenth-century English poems: William Wordsworth’s “I wandered lonely as a cloud”, Alfred Tennyson’s “Now Sleeps the Crimson Petal”, Robert Browning’s “Meeting at Night” and “Parting at Morning”, Matthew Arnold’s “Dover Beach”, and Thomas Hardy’s “The Darkling Trush” (cf. Steen 1999b, 2000a, in press a).

1.

Towards a metaphor identification procedure

The aim of a metaphor identification procedure is to facilitate the production of a list of metaphorical mappings by the analyst, mappings which have to be assumed for the semantic analysis of the metaphorical language in the discourse. Metaphorical mappings are defined with Lakoff (1993) as sets of conceptual correspondences between distinct domains of knowledge. Gentner (1983) and Turner and Fauconnier (1995) present comparable approaches. The problem to be solved by the discourse analyst is to determine how many mappings of what structure and degree of detail are to be related to some stretch of discourse. This is emphatically a problem of language analysis by the linguist, not a problem of discourse processing by the language user, something that we shall return to later. The problem of an operational procedure for metaphor identification has not been addressed in any systematic fashion (but see Cameron 1999a,
b), which is why Steen (1999c) has proposed a five-step procedure for getting at conceptual metaphor from linguistic metaphor. Semino et al. (in press) have applied the procedure to a small corpus of conversations between cancer patients and their doctors and have found that it is useful but needs further critical development. Crisp (in press), Steen (in press a), Heywood et al. (in press), and Crisp et al. (in press) have used the five-step procedure as a background to their work on linguistic and propositional aspects of metaphor identification and analysis, which in effect has meant developing the first two steps of the procedure in more detail. Provisional guidelines

A rhetoric of metaphor 149

resulting from this research have been applied in unpublished studies of the relation between metaphor and genre (including advertisements in glossy magazines, fundraising letters, and football match reports and obituaries in the press). The procedure is the subject of further discussion and application in research performed by researchers working in poetics, stylistics, Cognitive Linguistics, psycholinguistics, and applied linguistics. The main issue of this work is the reliability of the procedure as an operational instrument of metaphor identification. The five steps of the procedure involve the following critical moments (Steen 1999c): 1. identification of metaphorical expression 2. identification of metaphorical proposition 3. identification of metaphorical comparison 4. identification of metaphorical analogy 5. identification of metaphorical mapping It should be noted here that all of the steps 2 through 5 are conceptual: they abstract away from linguistic form and involve conceptual elaboration, as we shall see below. The end product, a metaphorical mapping, may be a conventional metaphorical mapping, such as LIFE IS A JOURNEY and the rest from the Cognitive Linguistic literature. But it may also be a novel, or one-shot, metaphorical mapping. This difference is immaterial for metaphor identification: that a conceptual mapping is conventional only suggests that it has been used more frequently and may have many cognates, which also have to be subjected to the five step procedure for identifying them as metaphorical. It is precisely the aim of the procedure to give a uniform account of the identification of novel and conventional metaphorical expressions as conceptual mappings in order to cover the complete range of metaphor in discourse. An example from one of the poems may help to illustrate what is at stake in each of the five moments. Wordsworth’s famous poem about the daffodils contains the line The waves beside them danced, ‘them’ being the daffodils (cf. Steen 1999b for further discussion of this and other metaphors in this poem). The first step of the procedure picks out the fact that there is one metaphorical expression in this line, danced, which acts as a metaphorical focus against the frame of the rest of the line. This decision is not controversial in the case of dancing waves, since dancing is prototypically performed by human agents, and we can discern a similarity between the dancing of humans and the movement of the waves. But there are many other expressions where it is less clear whether a word is used metaphorically or not, for instance because prototypical agents of actions are less clearly specified than in the case of dancing. Some important aspects are discussed by Heywood et al. (in press), and they include specific problems with the metaphorical analysis of adjectives and delexicalised verbs as well as the effect of co-text and intertextual references on metaphor identification.

150 Gerard Steen

The second step moves away from the linguistic surface structure and constructs the minimal idea unit which contains the concept activated by the metaphorically used word in the form of a conceptual proposition (e.g. Bovair and Kieras 1985): (1) The waves beside them danced P1(DANCE WAVES) P2(BESIDE P1 DAFFODILS)

This is the structure that provides a first grounding of the decision that ‘danced’ is metaphorically used. A match of this propositional structure with the projected text world would lead to the conclusion that the activity of the projected entity designated by WAVE cannot be the usual activity designated by DANCE: there is no genuine dancing going on in the evoked situation. An important part of the qualitative jump taken between step 1 and 2 is the fact that linguistically implicit elements, such as them, are explicated as complete concepts in the propositional analysis. This is because the propositional analysis attempts to capture the metaphorical idea related to the linguistic expression in the surface text as explicitly as possible. One important issue is how much more has to be explicated in such an analysis, the presupposed literal concepts of implicit metaphors providing a case in point (Steen 1999b,
d). The third step posits an incomplete comparison statement, along the lines of Miller (1993), to the effect that the literally and the metaphorically used concepts of the metaphorical proposition are separated from each other and located in their appropriate positions in an open analogy: (2) (DANCE WAVES) Æ ( E F) ( E y) {SIM [F (WAVES), DANCE (y)]}

A paraphrase of this result is that there is a similarity between some activity F of the waves and the dancing of some entity y. Moreover, it is to be understood that the projected referents related to the concepts of the first part of the comparison statement are part of the actual (‘literal’) situation model (Van Dijk and Kintsch 1983; see Zwaan et al. this vol.), while the referents related to the concepts in the second part provide material for non-literal comparison that is not present in the text world, but imaginative. The rewriting of the metaphorical idea or proposition produced by step 2 into a metaphorical comparison is automatic and based on a small number of rewrite rules proposed by Miller (1993). However, there are decisions to be made about how much propositional material, literal or metaphorical, can or should go into one open comparison. For instance, (1) might also have been rewritten as (3): (3) ( E F) ( E y) {SIM [ F (WAVES, DAFFODILS), DANCE-BESIDE (y, y¢)]}

A rhetoric of metaphor

Further research into the effect of adopting different conventions is on the agenda for the future, but it is clear that the transition between metaphorical idea identification and metaphorical comparison identification may involve another qualitative leap, depending on the conventions adopted. (The assumption that all metaphor involves comparison may also require further contextualisation; cf. Grady 1999.) Step 4 fills out the open comparison produced by step 3 and turns it into a complete analogy. Our example would turn into something like (4): (4) (DANCE WAVES) Æ SIM [MOVE (WAVES), DANCE (PERSON)]}

Another qualitative jump is made here: the choice of the concepts filled into the open slots involves an addition of information that may be highly debatable and needs to be constrained as much as possible. For instance, the superordinates MOVE and PERSON could be argued to be too general and in need of replacement by more specific descriptions, such as MOVE-RHYTHMICALLY and DANCER, respectively. This aspect, too, requires further theoretical and empirical research. The last step of the procedure involves the elaboration of the metaphorical analogy into a full-blown metaphorical mapping, for instance in the form of a list of entailments (Lakoff and Johnson 1980), a structure-mapping model (Gentner 1983), or a blend (Turner and Fauconnier 1995). Further concepts are related to the elements participating in the analogy, which may now be seen as the most important concepts of a more complex underlying conceptual structure based on nonliteral correspondence across two domains. For instance, when waves dance beside daffodils, they must be in the same location, and inferences may be made about the floor they dance on, the distance there is between them, the general location they are in, and so on. There may be other dancers there, and the music they dance to is another potential feature of the overall mapping to be developed on the basis of the analogy. It is interesting to see that this begins to look like the list of five dimensions of the event-tracking model proposed by Zwaan et al. (this vol.), which means that there are further opportunities for cross-fertilization between linguistics and psychology that need to be explored. It will be self-evident that step 5 is another moment of qualitative development in the analysis. It involves the thorny issue of which elements to select and include in the analytical representation of the mapping, and how this can be controlled. It also raises the question whether there are mappings that can be analyzed exhaustively as opposed to others that can only be analyzed partially or in multiple forms. These are important questions that have to be addressed by any approach wishing to claim that it has achieved a procedure for reliable and valid metaphor identification. Indeed, as we have proceeded from step 1 through step 5, the role of interpretation seems to have increased. This is precisely why it cannot be claimed that Cognitive Linguistics offers a procedure for metaphor identification, because Cognitive Linguists usually begin with the result of step 5 and are not concerned

151

152

Gerard Steen

with steps 1 through 4 in the first place. What the first proposal for a procedure has shown, however, is that we can distinguish between five qualitatively different steps involved in getting from the linguistic materials to the presumed underlying mappings, and that these need to be addressed before we can speak of an operational, reliable, and valid instrument of metaphor identification.

2.

Identification, analysis, and processing

It is important to emphasize that the procedure does not model metaphor processing (Crisp in press). The motivation of the conceptual and cognitive slant of the procedure lies in the maximization of compatibility between linguistic and psychological research of metaphor in order to attain a high degree of overall validity. However, the procedure does not embody a conflation between linguistic analysis and behavioral analysis, as sometimes happens in present-day linguistic research (Steen 1994, 2000b; Steen and Gibbs 1999). What the approach produces is a uniform explication of the presumed underlying conceptual mappings related to linguistic metaphors, which is an abstract, symbolic level of analysis. The procedure first and foremost aims to increase the control of linguists and psychologists over their textual data and materials. How language users realize such metaphorical language and their related conceptual structures is another matter, and a target for behavioral research. The first two steps of the procedure have been dealt with in greater detail by Steen (in press a), Heywood et al. (in press), and Crisp et al. (in press). The main point of Steen (in press a) has been that, if metaphor is primarily conceptual metaphor, the technique of propositionalization offers a suitable approach to its conceptual nature (Bovair and Kieras 1985; Kintsch 1998; Perfetti and Britt 1995; Van Dijk and Kintsch 1983). Metaphors being thoughts or ideas and not just language, the psychological instrument of the proposition as a minimal idea unit is an independently validated means to begin capturing the conceptual structure and content of metaphors as related to the linguistic expressions in a text. In the terminology of Kintsch (1998) and his colleagues, we need to get from metaphor in the text surface to metaphor in the text base. Moreover, metaphorical propositions are the input for the rewrite rules offered by Miller (1993) that facilitate the jump from metaphorical idea identification to metaphorical comparison identification. In many cases, if we do not get the implied comparison of a metaphor out into the open, we miss the entire meaning of the metaphor. And finally, propositional analysis also provides a platform for metaphor classification according to conceptual structure, as has been demonstrated by Steen (1999b) and Crisp et al. (in press). For instance, metaphorical mappings may be simple or complex, depending on their related propositional structure. Our dancing waves example is simple, because

A rhetoric of metaphor

there is only one proposition in the text base containing a metaphorically used concept, (DANCE WAVES). However, there is another line from the Wordsworth poem which involves complex metaphor: A poet could not but be gay / in such a jocund company. The expression jocund company refers to the daffodils of the poem and its propositional structure is complex because it has COMPANY featuring in one proposition and JOCUND featuring in another, modifying proposition (MOD COMPANY JOCUND), which follows the proposition containing the grammatical head COMPANY. There are other variables, too, like singular versus multiple, pure versus mixed and restricted versus extended, but space forbids ample illustration of these phenomena. The general point, however, is that the propositional approach affords a solid basis for the identification procedure which is acceptable to both linguists and psychologists. Empirical evidence of an effect of these structures on processing has been presented by Steen (in press b). In a study using a metaphor underlining task, subjects had to mark every metaphorically used word they encountered in a Bob Dylan song, “Hurricane”, as it was being played to them. When metaphors were at a higher level in the propositional analysis of a text unit, they were recognized more often than when they were at a lower level. (In linguistic terms, this means that metaphorically used words were either Immediate Constituents of some main clause, or the head of an Immediate Constituent.) Because of the limited number of metaphors in the song, the controlled pace of the task, and the informal and spoken nature of the language, however, the validity of this study is limited and may only be seen as an encouragement to pursue this line of inquiry at a larger scale. This was the reason to design the metaphor recognition study with the five poems also used in this chapter. The first results of the analysis confirm that there are effects of the propositional structure of metaphors on recognition. Metaphor recognition is related to other operations of metaphor processing (Steen 1994). Recognition goes together with appreciation and what I have called context construction, which is an operation whereby readers connect a metaphor to the presumed intentions of the author. These associations were found for literary and journalistic reading, which suggests that they are more general than just restricted to literary reading as a kind of point-driven reading (Vipond and Hunt 1984). The rationale for this clustering of processes seems to be that, when readers recognize a metaphor for what it is, they also tend to speculate about the intentions of the author in producing a metaphor as well as pronounce a judgement about the positive or negative effect of the metaphor they are finding themselves attending to in a relatively conscious manner. It will be an interesting goal for the future to examine whether the propositional, comparative, analogical, and mapping structures produced by the identification procedure can also be related to other processes than just recognition, such as appreciation, context construction, and other operations distinguished in previous research.

153

154

Gerard Steen

The other side of this program is to increase our insight into the distribution of distinct structural and other properties of metaphor between literary and nonliterary discourse. This is the ultimate aim of the extended interest group in metaphor mentioned above, who are intending to perform corpus work on metaphor in the British National Corpus, which contains literary and non-literary genres of discourse. For we really do not know in how far metaphor in literature is typically complex, multiple, mixed, and so on, in comparison to metaphor in other types of discourse. Of course there are plenty of intuitions and impressions, and a beginning has been made with this kind of corpus work by Goatly (1997), but his data lack information about reliability (see also Deignan 1999; Kövecses 1999). As long as we do not have any solid data about these issues, it will be hard to assess in how far processing results are due to reading strategies, language structures, or their interaction.

3.

Forms of metaphor

Work on the procedure has largely been limited to restricted metaphors in literary and non-literary prose, metaphors that do not exceed the boundaries of text units or text spans as defined by Mann and Thompson (1988). Now that the delicate baby is on all fours and crawling, it is interesting to see whether it can crawl into all corners of the room. Can the procedure also handle extended metaphor and other, more challenging, cases, than Man is a wolf? It is the more specific aim of this contribution to demonstrate that the most obvious alternative rhetorical forms of metaphor can also be handled by this approach. Poetry offers one suitable type of discourse to explore these issues in depth. The basic idea is the following. Application of the five-step procedure suggests that many metaphors in discourse require a series of analytical transformations before they may be identified as conceptual metaphors in the form of nonliteral conceptual mappings. However, not all linguistic expressions of conceptual metaphors are linguistic reductions of mappings. To take the other extreme, there are also metaphorical stretches of discourse extending beyond one text unit or utterance, which may be seen as complete expressions of metaphorical mappings. Here is an illustration from Thomas Hardy’s “The darkling thrush” (reproduced, as all other excerpts, from Gardner 1972): (5) The land’s sharp features seemed to be The Century’s corpse outleant, His crypt the cloudy canopy, The wind his death-lament.

The linguistic form of this conceptual metaphor is identical with the form of metaphorical mappings as they are represented in the Cognitive Linguistic literature.

A rhetoric of metaphor

There is not just one metaphorical text unit which contains a condensation of an underlying mapping, but there is a series of metaphorical text units which together present a coherent picture of a metaphorical mapping. It is as if the poet has chosen to be as explicit and extensive as possible about the conceptual metaphor he wanted to use. Of course, it will be almost impossible to ever spell out all of the potential aspects of a mapping in a text, but a poet may convey the impression that he has attempted to do so. This linguistic realization of a conceptual metaphor across a number of text units is qualitatively different than the many one-word metaphors featuring in the Cognitive Linguistic literature which require the five-step procedure to show that they are related to more complex underlying mappings. Similar comments can be made about other rhetorical forms of metaphor, such as analogy. Here is an example from Tennyson: (6) Now lies the Earth all Danaë to the stars, And all thy heart lies open unto me.

Here the poet has also expressed a metaphorical mapping, to the effect that there is a similarity between “the Earth’s lying all Danaë to the stars” and “thy heart’s lying all open unto me”. However, the metaphorical correspondence has not quite been expressed in the form of an extended and developed mapping, as in the case of Hardy. The difference lies in the selection of the most important elements of the nonliteral correspondence between the two domains, such that they have been entered into a proportional comparison, or analogy. From the point of view of explicating the underlying mapping, this stretch of discourse only requires the fifth step of the procedure, which fleshes out the analogy into a nonliteral mapping. And we can take another step back, from mappings through analogies to similes. For when elements of the presumably underlying analogy are deleted in the linguistic surface expression of a conceptual metaphor, the result is a simile. This is a nonliteral comparison statement between at least two terms, as in the following famous lines from Matthew Arnold: (7) The Sea of Faith Was once, too, at the full, and round earth’s shore Lay like the folds of a bright girdle furled.

The expression of the more elaborate metaphorical mapping has become more reduced than in the case of an analogy, because some of the elements of the complete analogy have been left implicit. In particular, a (non-literal) comparison is suggested between the sea of faith’s lying round the shore and the folds of a bright girdle lying round some unknown object. There is at least one open slot in the comparison, the unknown object around which the folds of the bright girdle lie, even if it be the girdle itself. However, the signal that we are dealing with a comparison has been retained. What is needed on the part of the Cognitive Linguist who

155

156 Gerard Steen

wishes to explicate the complete mapping is the application of steps 4 and 5. The simile first has to be transformed into an analogy, with the open slot filled in, and then fleshed out into a mapping, with other concepts related to the analogy. (I have ignored, for expository purposes, that this instance is much more complicated, because it also requires an account of the fact that we are dealing with a nonliteral sea, “The sea of faith”, which in turn is likened to a girdle. I would hope that a twostage application of the procedure would eventually be able to solve this kind of layering of metaphor, too.) When the signal for nonliteral comparison is also dropped, we have arrived at another degree of linguistic reduction of the original conceptual metaphorical mapping. This involves the linguistic expression of conceptual metaphors by means of rhetorical forms that are plain utterances instead of nonliteral comparisons, analogies, or mappings. This rhetorical form is isomorphic to the general format of the proposition as a nuclear predication, which captures its conceptual content. It is the dominant rhetorical form for metaphor in everyday language use, and it requires the full five-step procedure to show that a linguistic expression may be related to an underlying metaphorical mapping. Our dancing waves example above may be put forth here as a worked example. So here is the basic claim. The five-step procedure may also be used to model the linguistic expression of conceptual metaphors or metaphorical mappings into a small number of rhetorical forms that correspond with the four conceptual structures created by the procedure. The correspondence between the steps and the forms is laid out in Table 1, which may act as a summary of this section. Table 1.

1 2 3 4 5

Correspondence between steps of metaphor identification and rhetorical forms of metaphor

Steps

Forms

metaphorical expression metaphorical proposition metaphorical comparison metaphorical analogy metaphorical mapping

metaphor simile analogy extended metaphor

4. Steps, forms, and levels Suppose we are in the psychologist’s paradise and have access to the mind of the language user when he or she is planning what to say. And let us assume that the language user’s message, or more specifically the writer’s text, involves the use of

A rhetoric of metaphor

conceptual metaphor. Then most Cognitive Linguists would assume that the writer has a complete non-literal mapping activated in his or her mind which will have to be integrated into the structure of the discourse he or she is about to produce. Even though this is questionable, let us take this assumption as a maximally bold hypothesis for what goes on when people produce metaphors and develop it a little further. The writer has several options for the linguistic expression of the conceptual metaphor by means of recognizable rhetorical forms, and they may be characterized as follows: 1. the nonliteral mapping may be presented as a full-blown nonliteral mapping in the text; 2. the nonliteral mapping may be reduced to the analogical correspondence between the most important parts of the mapping; 3. the nonliteral mapping may be even further reduced to the signaling of a resemblance between a selection of the most important parts of the mapping; 4. the nonliteral mapping may be even further reduced to the direct usage of a selection of some of the most important parts of the mapping. Why a writer has a preference for one or another of these options in exactly which situations in which languages is one of the best kept secrets in metaphor research. One consequence of this approach to a rhetoric of metaphor concerns the reach of a metaphorical mapping in a text. Nonliteral mappings, or extended and serialized metaphors, by definition span a number of text units: they stretch across more than one text unit and form a coherent whole, as in the case of the Thomas Hardy example (5) above. (Depending on one’s purpose, text units may be defined in various ways, as clauses, the minimal text spans of Mann and Thompson 1988, or even sentences; see Steen (in press) and Crisp et al. (in press) for further discussion. In our work until now we have preferred the Mann and Thompson definition.) At the other extreme there are the maximally reduced linguistic expressions of a mapping which always remain inside the boundaries of a text unit. They were exemplified by the line from Wordsworth (1), analyzed in Section 2. The middle ground between these two poles is occupied by similes and analogies, which may be expressed within one text unit or between two text units. Thus the Tennyson example (6) of an analogy joined two consecutive lines, forming a coherence relation of comparison between the two, while the Arnold example (7) of a simile was expressed within one text unit. Analogies may be limited to one text unit, however, and similes may be constructed across text units. The latter aspect may be illustrated by the following contrast: (8) a. b.

I wandered lonely as a cloud I wandered lonely, as a cloud

157

158

Gerard Steen

The original of the famous line by Wordsworth is (8a), but the insertion of a comma in (8b) separates the adverbial adjunct from the preceding text unit, turning it into an independent text unit which functions as an afterthought. The simile has to be resolved by linking the two text units under a coherence relation of comparison, again, just as in the case of the Tennyson analogy (6). This kind of effect may also be suggested by such strategies of versification as enjambment, as in the following lines from the Thomas Hardy poem again: (9) The tangled bine-stems scored the sky Like strings of broken lyres,

Even though the punctuation does not break up what is one complete text unit into two parts, the versification does produce a degree of independence for the second line containing the adverbial adjunct. This is one reason why it may be more difficult to produce a reliable analysis of metaphor in poetry than in other types of discourse. The interest in restricting a metaphor to one text unit as opposed to extending it across text units lies in the attending complexity of processing. The assumption is that text units form natural boundaries during processing, as is likely given that clause and sentence boundaries have also been shown to influence on-line processing. If this assumption is correct, the processing of a conceptual metaphor may be heavily affected by its being expressed within or across text units. If a metaphorical mapping is expressed within one text unit, readers have to produce an interpretation of the mapping within one go, in the processing of one text unit. This holds for all rhetorical forms called metaphor, and for restricted similes and analogies. But if a metaphorical mapping is expressed across two or more text units, a reader has to integrate previous interpretations of metaphorical text units into more complex metaphorical structures as the reader progresses from one text unit involved in the mapping to another. This requires a prolonged balancing act between the two conceptual domains involved in the comparison. It probably demands a different kind of attention and awareness than the interpretation of a mapping expressed in just one text unit, if only because the source domain becomes a more persistent competitor for assigning local topichood. This would mean that extended expressions of metaphorical mappings, whether they are similes, analogies, or truly extended metaphors, may be more conspicuous and challenging than restricted metaphors, other things being equal. This is one aspect which I am investigating in the metaphor recognition study at the moment. This picture may be developed even further. If it is possible for analysts to create metaphorical mappings in five steps from linguistic expressions in texts, then readers may occasionally do the same. As I have emphasized at the beginning of this chapter, I do not postulate a one-to-one relationship between the five-step procedure, which is an operational instrument, and reader behavior, which is much more varied. But the five steps may be used as an initial model to examine the various

A rhetoric of metaphor 159

depths of metaphor processing achieved by readers in different circumstances. For as I have argued elsewhere, readers do not have to realize a metaphorical expression as a metaphorical mapping all the time (Steen 1994). One advantage of the present framework is that it suggests a more precise way of dealing with this issue. In particular, readers may actually process metaphorical expressions at different levels of metaphoricity, which correspond with the different steps of metaphor identification for the analyst. Thus, readers may ignore a metaphorical expression as such or they may process it as a metaphorical expression (level 1); they may continue to develop it into a little metaphorical thought (level 2), or a metaphorical comparison (level 3), or even a metaphorical analogy (level 4), or a full-blown metaphorical mapping (level 5). My previous work with think aloud data (Steen 1994) included observations on the frequency of analogy construction, which did not occur very often; but this only goes to show that it does happen and that there are methods to investigate this kind of phenomenon. In general, however, any metaphorical text unit may be realized at any level of metaphoricity, and the question arises which reader and text properties affect these processes. More interesting from the present perspective, however, is the theoretical consequence that there may be an interaction between levels of metaphor processing and rhetorical forms of metaphor in texts. Thus the question arises which forms trigger which levels of processing, and my general guess would be that extended metaphors would tend to trigger deeper levels of processing than restricted metaphors, other things being equal again. But these are fairly global indications for future research that need further theoretical and methodological preparation before they can be empirically researched. There are many other aspects that may be commented on from the present position, but space forbids going into them to any extensive degree. I shall briefly mention a number of interesting points for further consideration. One issue concerns the existence and analysis of borderline cases between the rhetorical forms identified above. For instance, the Wordsworth line quoted under (8a) runs on as follows: (10) I wandered lonely as a cloud That floats on high o’er vales and hills

The grammatical structure of this sentence turns it into one text unit containing a simile with some further materials in a subordinated clause in line 2, but the semantic material in line 2 may be sufficient, at least in one interpretation of the lines, to turn the simile into an analogy. There are many cases like this and much work will have to be done before the basic idea presented in Section 4 may be claimed to act as a reliable mould for the identification of diverse rhetorical forms in poetry and other relatively rhetorical types of discourse (see also Ben-Porat 1992).

160 Gerard Steen

There are also other rhetorical forms of metaphor which have been identified in the literature which we have not even begun to address. For instance, Goatly (1997) has advanced a catalogue of metaphorical interplay, including phenomena like repetition, multivalency, diversification, and literalization (cf. Freeman 1993; Hrushowski 1984; Werth 1994). All of these need to be accounted for, and the present approach seems to offer an appropriate framework. Finally, such forms also have their presumed functions. And this is where nonliterary discourse will have to provide a background for the detailed investigation of the use of metaphor in various rhetorical forms. For instance, extended metaphors are often found in text books where they have an overtly explanatory function. Similes and analogies, but also extended metaphors, also abound on the sports pages, where they have a clearly divertive function. And metaphors of all kinds have persuasive functions in advertisements and in political rhetoric. The investigation of the forms and functions of metaphor in all of these classes of discourse will have to throw into relief the specific nature of metaphor in literature and its function in literary reading.

5.

Conclusion

The psychological study of metaphor in literature cannot do without a serious linguistic theory of metaphor in literary and non-literary discourse. Although the Cognitive Linguistic approach to metaphor has provided an important stimulus to theory development and illustrative analysis, large-scale empirical research, for instance by means of corpus work, on the role of metaphor in different classes of discourse is still in its infancy. A major impediment to this kind of study has been the lack of a procedure for metaphor identification, which is a sine qua non if linguists and literary scholars wish to be able to compare findings and relate them to psychological research. This chapter has presented the development of such a procedure which has come out of a small interest group in the Poetics And Linguistics Association, and has sketched a larger framework for putting the procedure to use. In particular, the five steps distinguished in Steen (1999c) may also be used to postulate four basic rhetorical forms of metaphorical expression: metaphor, simile, analogy, and extended metaphor. The technical identification of metaphor in discourse may benefit from this classification to the extent that it helps analysts determine where one metaphor stops and another begins, which is important for quantification in corpus work and for the modeling of linear processing in experimental work. The five steps may also be used to model various depths or levels of metaphor processing, which may be expected to interact with the linear processing of different rhetorical forms. There is hence a consistent and integrated, multi-



A rhetoric of metaphor

faceted steps, forms, and levels framework which may be explored for the investigation of metaphor in literature.

Note * I wish to thank Ina Biermann, Lynne Cameron, Alan Cienki, Ray Gibbs, and Elena Semino for their perceptive and helpful comments on this chapter.

References Ben-Porat, Z. 1992. “Poetics of the Homeric simile and the theory of (poetic) simile”. Poetics Today 13: 737–69. Black, M. 1993. “More about metaphor”. In Metaphor and Thought, 2nd edition, A. Ortony, 19–41. Cambridge: Cambridge University Press. Bovair, S. and D.E. Kieras. 1985. “A guide to propositional analysis for research on technical prose”. In Understanding Expository text: A Theoretical and Practical Handbook for Analyzing Explanatory Text, B. Britton and J. Black (eds), 315–362. Hillsdale, NJ: Lawrence Earlbaum Associates. Cameron, L. 1999a. “Operationalising ‘metaphor’ for applied linguistic research”. In Researching and Applying Metaphor, L. Cameron and G. Low (eds), 3–28. Cambridge: Cambridge University Press. Cameron, L. 1999b. “Identifying and describing metaphor in spoken discourse data”. In Researching and Applying Metaphor, L. Cameron and G. Low (eds), 105–132. Cambridge: Cambridge University Press. Crisp, P. in press. “Metaphorical propositions — a rationale”. Special issue of Language and Literature, ‘Metaphor Identification’, to appear in 2002. Crisp, P., Heywood, J. and Steen, G., in press. “Identification and analysis, classification and quantification”. Special issue of Language and Literature, ‘Metaphor Identification’, to appear in 2002. Deignan, A. 1999. “Corpus-based research into metaphor”. In Researching and Applying Metaphor, L. Cameron and G. Low (eds), 177–99. Cambridge: Cambridge University Press. Freeman, D. C. 1993. “‘According to my bond’: King Lear and re-cognition”. Language and Literature 2: 1–18. Freeman, D. C. 1995. “‘Catching the nearest way’: Macbeth and cognitive metaphor”. Journal of Pragmatics 23: 689–708. Freeman, D. C. 1999. “‘The rack dislimns’: Schemata and metaphorical pattern in Antony and Cleopatra”. Poetics Today 20: 443–60. Gardner, H. (ed.) 1972. The New Oxford Book of English Verse: 1250–1950. Oxford: Oxford University Press. Gentner, D. 1983. “Structure-mapping: A theoretical framework for analogy”. Cognitive Science 7: 155–170. Gibbs, R. W. jr. 1994. The Poetics of Mind: Figurative Thought, Language and Understanding. Cambridge: Cambridge University Press. Goatly, A. 1997. The Language of Metaphors. London: Routledge.

161

162 Gerard Steen

Grady, J. 1999. “A typology of motivation for conceptual metaphor: Correlation vs. resemblance”? In Metaphor in Cognitive Linguistics, R. W. Gibbs, jr., and G. J. Steen (eds), 79–100. Amsterdam: John Benjamins. Heywood, J., Semino, E. and Short, M., in press. “Linguistic metaphor identification in two extracts from novels”. Special issue of Language and Literature, ‘Metaphor Identification’, to appear in 2002. Honeck, R. P. and Hoffman, R. R. (eds) 1980. Cognition and Figurative Language. Hillsdale, NJ: Lawrence Erlbaum Associates. Hoorn, J. F. 1997. Metaphor and the Brain: Behavioral and Psychophysiological Research into Literary Metaphor Processing. Unpublished doctoral dissertation, Free University Amsterdam. Hrushowski, B. 1984. “Poetic metaphor and frames of reference: With examples from Eliot, Rilke, Mayakovsky, Mandelshtam, Pound, Creeley, Amichai, and the New York Times”. Poetics Today 5: 5–43. Jakobson, R. 1960. “Closing statement: Linguistics and poetics”. In Style in language, T. A. Sebeok (ed.), 350–77. Cambridge, MA: MIT Press. Johnson, M. 1987. The Body in the Mind. Chicago: University of Chicago Press. Kintsch, W. 1998. Comprehension: A Paradigm for Cognition. Cambridge: Cambridge University Press. Kittay, E.F. 1987. Metaphor: Its Cognitive Force and Linguistic Structure. Oxford: Clarendon Press. Kövecses, Z. 1999. “Metaphor: Does it constitute or reflect cultural models?”. In Metaphor in Cognitive Linguistics, R.W. Gibbs, jr. and G.J. Steen (eds), 167–88. Amsterdam: John Benjamins. Lakoff, G. 1987. Women, Fire, and Dangerous Things: What Categories Reveal about The Mind. Chicago: University of Chicago Press. Lakoff, G. 1993. “The contemporary theory of metaphor”. In Metaphor and Thought, second edition, A. Ortony (ed.), 202–251. Cambridge: Cambridge University Press. Lakoff, G. and M. Johnson. 1980. Metaphors We Live By. Chicago: University of Chicago Press. Lakoff, G. and M. Turner. 1989. More Than Cool Reason: A Field Guide to Poetic Metaphor. Chicago: University of Chicago Press. Mann, W. and Thompson, S. 1988. “Rhetorical structure theory: Toward a functional theory of text organization”. Text 8: 243–81. Miall, D. S. and Kuiken, D. 1999. “What is literariness? Three components of literary reading.” Discourse Processes 28: 121–138. Miller, G. A. 1993. “Images and models, similes and metaphors”. In Metaphor and Thought, second edition, A. Ortony (ed.), 357–400. Cambridge: Cambridge University Press. Mio, J. S. and Katz, A. N. 1997. Metaphor: Implications and Applications. Hillsdale, NJ: Lawrence Erlbaum Associates. Murphy, G. 1996. “On metaphoric representations”. Cognition 60: 173–204. Murphy, G. 1997. “Reasons to doubt the present evidence for metaphoric representation”. Cognition 62: 99–108. Ortony, A. (ed.) 1993. Metaphor and Thought: Second Edition. Cambridge: Cambridge University Press. Perfetti, C. A. and M. A. Britt. 1995. “Where do propositions come from?”. In Discourse Comprehension: Essays in Honor of Walter Kintsch, C. A. Weaver, III, S. Mannes, and C. A. Fletcher (eds), 11–34. Hillsdale, NJ: Lawrence Erlbaum Associates. Reinhart, T. 1976. “On understanding poetic metaphor”. Poetics 5: 383–402. Ricoeur, P. 1979. The Rule of Metaphor: Multi-Disciplinary Studies of the Creation of Meaning in Language. London: Routledge and Kegan Paul. Richards, I. A. 1936. The Philosophy of Rhetoric. London: Oxford University Press.



A rhetoric of metaphor 163

Semino, E., Heywood, J. and Short, M., in press. “Methodological problems in the analysis of metaphors in a corpus of conversations about cancer”. Special issue of Journal of Pragmatics, ‘Researching And Applying Metaphor’. Shen, Y. 1987. “The structure and processing of poetic oxymoron”. Poetics Today 8: 105–22. Shen, Y. 1989. “Symmetric and asymmetric comparisons”. Poetics Today 10: 517–36. Steen, G.J. 1994. Understanding Metaphor in Literature: An Empirical Approach. London: Longman Steen, G. J. 1997. “Metaphor and emotion: The aesthetic and the artistic”. In Emotion, Creativity, and Art, Volume 2, L. Dorfman, D. Leontiev, V. Petrov, and V. Sozinov (eds), 117–32. Perm, Russia: Perm State Institute of Arts and Culture. Steen, G. J. 1999a. “Genres of discourse and the definition of literature”. Discourse Processes 28: 109–120. Steen, G. J. 1999b. “Analyzing metaphor in literature: With examples from William Wordsworth’s ‘I wandered Lonely as a Cloud’
”. Poetics Today 20: 499–522. Steen, G.J. 1999c. “From linguistic to conceptual metaphor in five steps”. In Metaphor in Cognitive Linguistics, R. W. Gibbs, jr. and G. J. Steen (eds), 57–77. Amsterdam: John Benjamins. Steen, G.J. 1999d. “Metaphor and discourse: Towards a linguistic checklist for metaphor analysis”. In Researching and Applying Metaphor, L. Cameron and G. Low (eds), 81–104. Cambridge: Cambridge University Press. Steen, G. 2000a. “Language, text, and discourse: Robert Browning’s ‘Meeting at Night’ and ‘Parting at Morning’”. In Contextualized Stylistics: In Honour of Peter Verdonk, A. Bex, M. Burke, and P. Stockwell (eds), 195–208. Amsterdam: Rodopi. Steen, G. 2000b. “Metaphor and language and literature”. Language and Literature 9: 289–305. Steen, G. J. in press a. “Towards a procedure for metaphor identification”. Special issue of Language and Literature, ‘Metaphor Identification’, to appear in 2002. Steen, G. J. in press b. “Structural metaphor properties and metaphor recognition”. Special issue Journal of Pragmatics, ‘Researching And Applying Metaphor’. Steen, G. and Gibbs, R., Jr. 1999. “Introduction”. In Metaphor in Cognitive Linguistics, R.W. Gibbs, jr. and G. J. Steen (eds), 1–8. Amsterdam: John Benjamins. Turner, M. 1987. Death Is the Mother of Beauty: Mind, Metaphor, Criticism. Chicago: University of Chicago Press. Turner, M. 1991. Reading Minds: The Study of English in the Age of Cognitive Science. Princeton, NY: Princeton University Press. Turner, M. and Fauconnier, G. 1995. “Conceptual integration and formal expression”. Metaphor and Symbolic Activity 10: 183–204. Van Dijk, T. A. and Kintsch, W. 1983. Strategies of Discourse Comprehension. New York: Academic Press. Vipond, D. and Hunt, R. A. 1984. “Point-driven understanding: Pragmatic and cognitive dimensions of literary reading”. Poetics 13: 261–77. Werth, P. 1994. “Extended metaphor: A text-world account”. Language and Literature 3: 79–103. Zwaan, R. A., Madden, C. J., Stanfield, R. A., this volume. “Time in narrative comprehension: A cognitive perspective”.



Irony and its discontent Rachel Giora Tel Aviv University

One of irony’s notorious characteristics is its beguiling nature (cf. Booth 1974). When speakers assume a different voice, pretense (Clark and Gerrig 1984; Clark 1996), or some guise (Haiman 1990, 1998), they may mislead even the adult addressee,1 who may not be able to detect the counterfeit and derive the ironic intent. The deceitful nature of irony has not been lost on artists and writers, who have manipulated it for various purposes, most notably, the insertion of novel, dissident ideas (Giora in press c; Walker 1991). Irony’s indirectness acts as a shield which masks a genuine intent deemed risky by the speaker. Consider the following poem by the Israeli poet, Yona Wallach (1997), which has been taken literally by most readers, but which, in fact, assumes a guise — man’s voice, language, and fantasies — to convey subversiveness: (1) Strawberries2 When you come to sleep with me wear a black dress printed with strawberries and hold a basket of strawberries and sell me strawberries tell me in a sweet light voice strawberries strawberries who wants strawberries don’t wear anything under the dress afterwards strings will lift you up invisible or visible and lower you directly on my prick.

Yona Wallach is one of the most important feminist poets worldwide. Her literary accomplishments rest, among other things, on her ability to integrate the language and topics of women’s and men’s poetry (Rattok 1997:75). “Strawberries” is

166 Rachel Giora

exemplary in this respect. It adopts males’ pornographic language in order to protest women’s oppression by the self-same language and industry. The disguised protest, however, has escaped the eyes of a lot of Wallach’s readers, who took her poetry literally as a celebration of sexuality — a topic that has hitherto dominated men’s writings. According to Rattok (1997), however, Wallach is ironic. Her critical intent may not be as obvious in “Strawberries” as it is in the following poem, which portrays a pornographic show: (2) Tefillin3 Come to me don’t let me do anything you do it for me do everything for me what I even start doing you do instead of me I’ll put on tefillin I’ll pray you put on the tefillin for me too bind them with delight on my body rub them hard against me stimulate me everywhere make me swoon with sensation move them over my clitoris tie my waist with them so I’ll come quickly play them in me tie my hands and feet do things to me against my will turn me over on my belly and put the tefillin in my mouth bridle reins ride me I am a mare pull my head back till I scream with pain and you’re pleasured then I’ll move them onto your body with unconcealed intention oh how cruel my face will be I’ll move them slowly over your body slowly slowly slowly around your neck I’ll move them I’ll wind them several times around your neck, on one side and on the other I’ll tie them to something solid

Irony and its discontent 167

especially heavy maybe twisting I’ll pull and I’ll pull till your soul leaves you till I choke you completely with the tefillin that stretch the length of the stage and into the stunned crowd.

In this poem, the speaker is acting out her anger with pornography. At the end of the piece, the speaker strangles her abuser by using the ritual ropes used by men in prayers that exclude women. This poem leaves no doubt as to Wallach’s stance: rather than endorsing it, Wallach dissociates herself from the language and practice of oppressive pornography. Wallach is by no means the only author whose irony has been lost on most of her readers. Resorting to irony has also cost other authors such as Swift (A Modest Proposal) and Austin (Pride and Prejudice) the price of being misunderstood by their contemporary readers (cf. Booth 1974). Though these poetic examples may be suggestive as to why literary irony is easy to ignore, misinterpret, or miss, our daily experience with irony seems to suggest the opposite: irony is ubiquitous and much easier to understand than implied by the above examples (see also Gibbs 1994).

1.

Comprehension

Indeed, one of the most heated debates within linguistics and psycholinguistics is whether irony (or nonliteral language, in general) requires a special (sequential) process (e.g. Grice 1975; Searle 1979), or whether it is interpreted on patterns similar to those induced by literal language (e.g. Gibbs 1986a,
b; Giora in press b; Glucksberg 1995; Sperber and Wilson 1986/95: 239). 1.1 Processing models Direct access view Researchers attributing to contextual information a primary role in language comprehension assume that literal and nonliteral language involve equivalent processes: in a rich ecology, contextual information affects comprehension very early on so that comprehenders retrieve the contextually appropriate meaning more or less directly, without having to go through an incompatible phase (for a similar view regarding lexical ambiguity see also Vu et al. 1998). This implies that in a literally biasing context (a sunny day), it is only the literal interpretation (“nice weather” of

168 Rachel Giora

What a lovely day for a picnic) that is recovered; in an irony inducing context (a stormy day), it is only the ironic interpretation (“lousy weather”) that is tapped.4 The assumption that context significantly governs comprehension features dominantly in the relevance theoretic account. According to Sperber and Wilson (1986/95), context is not fixed in advance, but is searched for the purpose of rendering an utterance relevant. Utterance interpretation is thus entirely dependent on the contextual information brought to bear. Recruiting the appropriate context results in tapping the contextually appropriate meaning directly without having to go through an incompatible phase which will require revisitation. As a result, processing irony, for instance, an “echoic interpretive use in which the communicator dissociates herself from the opinion echoed with accompanying ridicule or scorn” (Wilson and Sperber 1992: 75), need not differ from processing a similar utterance in which the communicator endorses the opinion echoed. Both should be interpreted on similar patterns. While irony involves the speaker’s dissociation from the opinion echoed, literal interpretation involves endorsement of the echo (Sperber and Wilson 1986/95: 239). The echoic mention view of irony (Jorgensen et al. 1984; Sperber 1984; Sperber and Wilson 1981, 1986/95; Wilson and Sperber 1992; but see Curcó 2000, submitted; Giora 1997b, 1998a,
b; Smith and Wilson 1992; Yus 1998), is thus consistent with a direct access view according to which contextual information affects comprehension to the extent that it prevents activation of irrelevant interpretations. The processing equivalence hypothesis also underlies the allusional pretense theory (Kumon-Nakamura et al. 1995). According to this view, irony alludes to or reminds the addressee of what should have been — of an expectation or a norm that went wrong (see also Kreuz and Glucksberg 1989). To enable the addressee to appreciate the allusion, irony involves pragmatic insincerity. Positing insincerity allows for various speech-acts beyond assertions to be ironic. For example, when a car driver says I just love people who signal when turning when the car ahead of her makes a turn without signaling, the speaker alludes to a social norm to signal upon turning, while simultaneously pretending to compliment the errant driver. Such a view of irony assumes that irony comprehension involves activating the literal meaning of the utterance in order to assess its insincerity and derive the ironic interpretation (see also Glucksberg 1995). It does not, however, assume the precedence of the literal over the nonliteral interpretation, since initial comprehension does not involve an assumption about the speaker’s sincerity. Rather, in any given situation, there is a decision to be made whether the literal meaning is intended sincerely or insincerely. In this way, ironic and literal interpretations involve equivalent processes, resulting in different products, though (Glucksberg, personal communication).

Irony and its discontent 169

Standard pragmatic model In contrast to the processing equivalence assumption, the hypothesis that understanding nonliteral language is a ‘two stage’ process, the first — literal and obligatory, the second — nonliteral and optional, implies that understanding literal and ironic language involve different mechanisms, the latter comprising more complex inferential processes. Thus, while literal interpretation includes no contextually incompatible stage, in understanding irony, the comprehender first computes the contextually incompatible literal interpretation (“nice weather” of What a lovely day for a picnic). Since that interpretation is contextually incompatible, it is rejected and replaced by the appropriate meaning (“lousy weather”). This classical processing model, known as “The standard pragmatic model”, originated in Grice (1975) and Searle (1979). The standard pragmatic model assumes that the initial stage of irony comprehension is impervious to context effects (cf. Fodor 1983), involving (a detection of) a breach of a norm, primarily the truthfulness maxim, which is a signal to the addressee to reject the computed literal meaning and derive the ironic intent. A more recent proposal entitled “relevant inappropriateness” (Attardo 2000) goes beyond the rule violation condition and proposes the breach of contextual appropriateness. While assuming Grice’s relevance maxim for the integration phase, relevant inappropriateness requires that contextual appropriateness be ostensibly violated at the initial phase, so that the comprehender can detect the overt violation and derive the ironic intent. This violation, however, must be only minimally disruptive, though perceivable as disturbing contextual appropriateness. For example, when, in a drought-stricken area, one farmer says to another Don’t you just love a nice spring rain? the utterance may be true, yet inappropriate, given the situation of utterance (it is not raining). According to Attardo, violation of contextual appropriateness includes violation of both sincerity and cultural norms or expectations (assumed necessary for irony interpretation by the allusional pretense, see above) and more (for instance deictic inappropriateness). The joint pretense view (Clark 1996; Clark and Carlson 1982; Clark and Gerrig 1984) is also inspired by the Gricean model (Grice 1978). It assumes a speaker who pretends “to be an injudicious person speaking to an uninitiated audience; the speaker intends the addressee of the irony to discover the pretense and thereby see his or her attitude toward the speaker, the audience, and the utterance” (Clark and Gerrig 1984: 12. For a similar view see Boulton as quoted in Booth 1974: 105). By saying What a lovely day for a picnic on a stormy day, the ironist assumes the identity of another speaker addressing a gullible audience. The present addressee, however, is supposed to take delight in recognizing both the pretense and the intended attitude of ridicule toward the pretending speaker, the audience, and the utterance. According to Clark (1996: 368), joint pretense is conceived of as a staged communicative act (see also Kotthoff 1998) where the actual speaker is also an implied speaker performing a sincere communicative act toward an implied

170 Rachel Giora

addressee who is also the actual addressee. Both actual participants are intended to “mutually appreciate the salient contrasts between the demonstrated and actual situations”, so that if asked, the actual speaker would deny meaning for the actual addressee what the implied speaker means for the implied addressee. According to this view, irony is a two-layered act of communication in which the literal meaning is activated and retained by both the speaker and the addressee, who reject it as the intended meaning though they pretend otherwise. Though inspired by Grice (1978), it is not quite clear whether this double-layered approach assumes a sequential or a parallel process. The possibility that irony involves parallel activation of both the literal and ironic interpretations is also entertained by the tinge hypothesis (Dews et al. 1995; Dews and Winner 1995, 1997, 1999). According to the tinge hypothesis, irony is used to mute the intended negative criticism (for a similar view see Barbe 1995; for a different view see Colston 1997). The positive literal meaning of irony (That was really funny said on a mean joke) tinges the addressee’s perception of the intended meaning. Similarly, the negative literal meaning of ironic compliments (It’s a tough life said to someone on vacation) mitigates the positively intended meaning. Winner and her colleagues assume that the contextually incompatible, literal meaning of ironic remarks is processed at some level and interferes with the intended meaning. Following Long and Graesser (1988), they propose a dual-process model “in which comprehension may occur after the recognition of an incongruity or simultaneously” (Dews and Winner 1997: 405). According to the tinge hypothesis, then, the literal meaning of irony is activated initially, either before or alongside the ironic meaning, and is retained in order to dilute either the criticism or the compliment. Has any of the approaches gained empirical support? So far, findings have not been monolithic. Some studies support the equivalent processes hypothesis. They show that ironic and nonironic utterances took equally long to read (Gibbs 1986a,
b; but see Dews and Winner 1997 and Giora 1995 for a critique of some of the findings), and to involve equal response times to ironically and literally related probes (Giora and Fein 1999a). Others are consistent with the different processes assumption (Dews and Winner 1997, 1999; Giora 1995). They show that utterances took longer to read in ironically than in literally biasing contexts (Gibbs et al. 1995; Giora et al. 1998; Schwoebel et al. 2000), longer to be judged as positive or negative relative to their literal counterparts (Dews and Winner 1997), and to involve longer response times to ironically than to literally related probes (Giora and Fein 1999a; Giora et al. 1998).

Irony and its discontent

2.

The graded salience hypothesis

Recent research however proposes that the apparently conflicting findings are resolvable in terms of the graded salience hypothesis (Giora and Fein 1999a,
b; Giora et al. 1998; Giora 1997a, 1999a,
b, in press a). Rather than positing the precedence of contextual information on the one hand (Gibbs 1994; Sperber and Wilson 1986/95; Utsumi 2000; Yus 2000), or the priority of literal meanings on the other (Grice 1975; Searle 1979), the graded salience hypothesis proposes that the factor affecting lexical processing is lexical salience. Lexical salience pertains to privileged meanings — meanings foremost on our mind. A meaning of a word or an expression is salient if it is coded in the mental lexicon. Degree of salience is determined by frequency, conventionality, familiarity, or prototypicality. The more familiar, frequent, conventional, or prototypical a meaning the more salient it is. For instance, both meanings of bank, that is, the “financial institution” and the “riverside” meanings, are listed in the mental lexicon. However, for those of us living in urban communities, in which rivers are less common than financial institutions, the commercial sense of bank is more accessible that is, salient; by the same token, the riverside sense is less salient. In contrast, inferences computed on the fly are nonsalient, since they are not coded in the mental lexicon. The claim is that highly salient meanings would always be accessed automatically, irrespective of contextual information. Context has a limited role. It may be predictive and avail appropriate meanings, but it is less effective in blocking salient albeit inappropriate meanings.5 Rather, it comes into play following the initial access stage, either suppressing or retaining incompatible meanings, or selecting the contextually appropriate sense. For instance, the following joke hinges on the prototypical/salient “male” feature of rabbi that gets accessed initially despite the expectation built up by the context (that is the repeated use of the “x is pregnant” construction) and despite contextual misfit: (3) What’s the difference between an orthodox, a conservative, and a reform wedding? In an orthodox wedding, the bride’s mother is pregnant. In a conservative wedding, the bride is pregnant. In a reform wedding, the rabbi is pregnant.

According to the graded salience hypothesis, then, interpreting utterances whose multiple interpretations are similarly salient would involve activating these meanings in parallel, irrespective of contextual information. For instance, conventional ironies such as tell me about it would involve similar processes in either context. Given that both the ironic interpretation (of the sentence/expression as a whole) and the literal meaning (of the sentence/expression’s constituents) are coded in the mental lexicon, these meanings would be activated in both the literal and irony inducing contexts. In contrast, less familiar ironies such as what a lovely day for a picnic, whose literal but not ironic meaning is coded in the lexicon, would involve a sequential process: they would be interpreted literally first and ironically second.

171

172 Rachel Giora

Their literal interpretation, however, would be tapped directly on account of the salience of its constituents. The graded salience hypothesis does not assume that the literal interpretation of the whole statement need to be computed first before the ironic interpretation is derived. It posits, though, that the salient interpretation (for instance the literal meaning of lovely) should be activated upon encounter of the linguistic stimulus. 2.1 Predictions The three processing models sketched above have, thus, different predictions regarding irony comprehension. According to the equivalent processes/direct access view, given a supportive context, 1. an utterance (what a lovely day for a picnic) will not take longer to read in an irony than in a literal meaning inducing context; 2. readers would always respond faster to contextually appropriate probes, since they activate only the contextually appropriate meaning. In a literally biasing context (sunny day), only the contextually compatible literal meaning (“nice weather”) would be activated. In an ironically biasing context, (a stormy day), this utterance will induce only the contextually compatible ironic meaning (“lousy weather”). Alternatively, if the equivalent processes hypothesis assumes parallel activation of the literal meaning and the attitude (or in/sincerity) invoked, response times to literally and ironically related test words should not differ. 3. these predictions should indifferently hold for familiar and unfamiliar ironies, because it is context rather than salience that primarily governs comprehension. According to the various versions of the classical model, 1. an utterance will take longer to read in an irony than in a literal meaning inducing context; 2. if measured immediately, even in a highly supportive ironic context, readers would respond faster to literally than to ironically related probes, because they always activate the literal meaning initially. The ironic meaning will lag behind; 3. these predictions should indifferently hold for more and less familiar ironies. Given that the graded salience hypothesis is concerned with degree of salience and is agnostic with regard to literal and nonliteral language, it predicts that 1. unfamiliar ironies (i.e. utterances whose ironic interpretation derives from ironically biasing contexts) will take longer to read in an irony than in a literal inducing context, since their ironic meaning is not listed in the mental lexicon, while the literal meaning (of their constituents) is. In contrast, familiar ironies (tell me about it) will take equally long to read in either context, because they

Irony and its discontent

are coded both ironically and literally. When they are much more salient than their literal equivalents (being as conventional as idioms), they will be processed faster than their literal equivalents; 2. even in a highly supportive, irony inducing context, readers will respond fast to probes related to salient meanings, regardless of contextual information. Regarding less familiar or unfamiliar ironies, this should be true of the literal (or any other conventional) meaning. The ironic meaning will take longer to respond to and will benefit from extra processing time. Similarly, in the literally biasing context, the salient (literal) meaning will be responded to faster than the nonsalient ironic meaning. In contrast, response times to literally and ironically related test words of familiar ironies will not differ. Neither will they benefit from extra processing time. 2.2 Findings The vast array of findings prevalent in the literature is consistent with the saliencebased view. It demonstrates that, irrespective of the tools employed to test the hypotheses, salient meanings are always accessed upon encounter, regardless of contextual information or literality. Reading times The most popular measure employed by psycholinguists has been reading times of whole utterances embedded in differently biasing contexts. Most of the tested items are nonconventional ironies.6 Where such items are tested, evidence demonstrates that less salient language (unfamiliar irony) takes longer to process than more salient (literal) language (whose accessibility hinges on the salience of the constituents that make it up). For instance, unfamiliar ironies (I’d say women have had real progress), whose literal meaning is more salient than their ironic meaning, took longer to read in an irony (4a) than in a literal (4b) inducing context (cf. Giora et al. 1998; Giora and Fein 1999a): (4) a.

b.

Just how far have women risen in the film community? According to M. P., who was at Woman in Film luncheon recently in Los Angeles, it has actually been a very good year for women: Demi Moore was sold to Robert Redford for $1 million in the movie Indecent Proposal… Uma Thurman went for $40,000 to Robert De Niro in the recent movie, Mad Dog and Glory. Just three years ago, in Pretty Woman, Richard Gere bought Julia Roberts for what was it? $3,000? “I’d say women have had real progress.” Just how far have women risen in the film community? According to M. P., who was at Woman in Film luncheon recently in Los Angeles, it has actually been a very good year for women: Demi Moore

173

174 Rachel Giora

earned $10 million in the movie Indecent Proposal… Uma Thurman made $400,000 in the recent movie, Mad Dog and Glory. Just three years ago, in Pretty Woman, Julia Roberts earned — what was it? $130,000? “I’d say women have had real progress.”

Similarly, in Gibbs (1986a), such ironies (You are a big help), projecting an attitude of dissociation from what is echoed (5a), took longer to read than their literal equivalents (Thanks for your help, 5b), projecting an attitude of endorsement or acknowledgment (cf. Giora 1995 for reinterpretation of Gibbs’ findings): (5) a.

b.

Harry was building an addition to his house. He was working real hard putting in the foundation. His younger brother was supposed to help. But he never showed up. At the end of a long day, when Harry’s brother finally appeared, Harry said to his brother: “You are a big help.” Greg was having trouble with calculus. He had a big exam coming up and he was in trouble. Fortunately, his roommate tutored him on some of the basics. When they were done, Greg felt he’d learned a lot. “Well” he said to his roommate, “Thanks for your help.”

Dews et al. (1995) and Dews and Winner (1997, 1999) used a different methodology. They asked subjects to judge the intended meaning of utterances and recorded their responses. Results showed that less familiar ironies took longer to be judged as positive (It’s a tough life said to someone on vacation) or negative (That was really funny said on a mean joke) than their literal interpretations. Such findings suggest that the contextually incompatible, literal meaning of less familiar ironies is accessed automatically, and interferes with the process. In addition Schwoebel et al. (2000), compared reading times of ironic praise (6) and ironic criticism (7) (illustrated by the first phrase in bold) and their literal counterparts (illustrated by the second phrase in bold): (6) Ironic praise Sam complained to his mother that he had too much homework. He said it would take him the whole weekend. On Saturday morning, he started his work, and was all done in one hour/by the end of the day he had finished less than half. His mother said: “Your work load is overwhelming this weekend.” (7) Ironic criticism A new professor was hired to teach philosophy. The professor was supposed to be really sharp. When Allen asked several questions, the professor offered naive and ignorant/incisive and knowledgeable answers. Allen said: “That guy is brilliant at answering questions.”

They found that participants took longer to read the target phrase (Your work load is overwhelming this weekend) in the ironically than in the literally biasing context,

Irony and its discontent

though this difference was significant only for the ironic criticism targets and only by item analysis. That salience rather than literality is a primary factor affecting initial processing is even more evident when reading times of familiar figurative utterances (This one’s really sharp) embedded in contexts inviting the salient (metaphoric) meaning (8a) are compared with their reading times in contexts inviting their nonsalient (ironic) meaning (8b):7 (8) a.

b.

You are a teacher at an elementary school. You are discussing a new student with your assistant teacher. The student did extremely well on her entrance examinations. You say to your assistant, “This one’s really sharp.” You are a teacher at an elementary school. You are gathering teaching supplies with your assistant teacher. Some of the scissors you have are in really bad shape. You find one pair that won’t cut anything. You say to your assistant, “This one’s really sharp.”

As shown in Gibbs (1998), unfamiliar (ironic) targets (8b) took longer to read than their familiar (metaphorical) counterparts (8a; see Giora in press a for a discussion). Pexman, Ferretti and Katz (2000) present similar findings. In their study, familiar metaphors (Children are precious gems) took longer to read in an irony (9c) than in a metaphor (9b) inducing context relative to a neutral control (9a).8 Though contextual information (knowledge about speakers who are inclined towards nonliteral vs. literal language, e.g. scientist vs. man) speeded up irony processing, it nevertheless did not block salient (metaphoric) albeit contextually inappropriate meanings: (9) a.

b.

c.

A man was talking to Jodie about his niece and nephew who had visited him recently. During the conversation the man said: “Children are precious gems.” This made Jodie think about her cousins. A scientist was talking to Jodie about his niece and nephew who had visited him recently. The scientist had really enjoyed having the children around. During the conversation the man said: “Children are precious gems.” This made Jodie think about her cousins. A scientist was talking to Jodie about his niece and nephew who had visited him recently. The scientist had found the children to be loud and disruptive and had not enjoyed their visit. During the conversation the scientist said: “Children are precious gems.” This made Jodie think of her cousins.

In both studies, the metaphoric reading was the salient reading, whereas the ironic reading was nonsalient (that is entirely dependent on context), and probably involved accessing the salient (metaphoric) meaning first before adjusting it to contextual information.

175

176 Rachel Giora

Findings in Gibbs et al. (1995) may also be explained by the graded salience hypothesis. Gibbs et al. investigated comprehension of intended and unintended (situational) irony (I would never be involved in any cheating). An utterance is considered an unintended irony in case it is intended literally but is perceived as ironic by overhearers. In Gibbs et al., intended ironies (10a) took longer to read than their unintended equivalents (10b): (10) a.

b.

Intended irony John and Bill were taking a statistics class together. Before the final exam, they decided to cooperate during the test so they worked out a system so they could secretly share answers. After the exam, John and Bill were really pleased with themselves. They thought they were pretty clever for beating the system. Later that night, a friend happened to ask them if they ever tried to cheat. John and Bill looked at each other and laughed, then John said, “I would never be involved in any cheating.” Unintended irony John and Bill were taking a statistics class together. They studied hard together, but John was clearly better prepared than Bill. During the exam, Bill panicked and started to copy answers from John. John did not see Bill do this and so did not know he was actually helping Bill. John took the school’s honor code very seriously. Later that night, a friend happened to ask them if they ever tried to cheat. John and Bill looked at each other, then John said, “I would never be involved in any cheating.”

Though subjects’ reports demonstrate that they perceived the unintended irony and considered it even more ironic than the intended irony, it is possible that their reading times reflect only their comprehension of the salient, literally intended interpretation. Some studies, however, present faster reading times for less familiar than for familiar language (Gibbs 1986a,
b). In Gibbs (1986a), ironies (You are a big help, 5a) were faster to read than their intended interpretations (You are not helping me, [11]). In Giora (1995), I explained this finding in terms of coherence imbalance. Given the discourse context, the final utterance (You are not helping me) is redundant; the alternative, ironic ending (see 5a) is far more informative (on the assumption, of course, that the salient/literal meaning is activated first): (11) Harry was building an addition to his house. He was working real hard putting in the foundation. His younger brother was supposed to help. But he never showed up. At the end of a long day, when Harry’s brother finally appeared, Harry said to his brother: “You are not helping me.”

(11) and (5a) are, therefore, incomparable. Similarly, in Gibbs (1986b), targets (Sure is nice and warm here) took longer to read and to make a paraphrase judgments in a literally (12) than in an ironically (13) biasing contexts:

Irony and its discontent 177

(12) Tony’s roommate always kept the windows open in the living room.He did this even if it was freezing out. Tony kept mentioning this to his roommate but to no avail. Once it was open and Tony wanted his roommate to shut it. Tony couldn’t believe that his roommate wasn’t cold. He said to him, [Ironic target] “Sure is nice and warm here”. [Paraphrase] “Please close the window”. (13) Martha went over to her sister’s house. It was freezing outside and Martha was glad to be inside. She said to her sister, ‘your house is very cozy’ [Literal target] “Sure is nice and warm here”. [Paraphrase] “This room is warm”.

Indeed, in the literally biasing contexts, these targets are redundant, being a sort of reiteration of a previous utterance (e.g. “Your house is very cozy”). In contrast, their ironic counterparts are a lot more informative. It is thus possible that the longer reading times found in Gibbs’ studies do not provide for a counter example but were caused by relative incoherence.9 As for paraphrase judgments, in Giora (1995), I suggest that paraphrase judgments of ironic targets are easier to make, because comprehension of irony involves paraphrasing, that is rephrasing the surface (usually literal) interpretation which provides comprehenders with a readymade paraphrase that is easier to recognize. This is not true of literal targets.10 All other things being equal, then, findings from reading times of whole utterances are better accounted for by the graded salience hypothesis. They show that salient meanings are processed faster than their nonsalient equivalents, regardless of context or literality. Response times Response times pertain to the time it takes subjects to make a lexical decision as to whether a test word is a word or a nonword. In Giora et al. (1998) and Giora and Fein (1999a), we measured response times to literally and ironically related test words.11 While unfamiliar ironies have only one salient meaning — the literal meaning, familiar ironies (Tell me about it) have also their nonliteral ironic meaning listed in the mental lexicon (see note 6 on how we controlled for salience). According to the graded salience hypothesis, unfamiliar ironies (4a above) should facilitate literally related test words initially while ironically related test words would be facilitated only after a delay. In contrast, familiar ironies (14a below) should facilitate both meanings initially, regardless of context. Indeed, unfamiliar ironies (I’d say women have had real progress) facilitated only their literally related test word (e.g. “success”) initially (150 msec after offset of the target sentence) in both the literally and ironically biasing contexts. The ironic test word (“regress”) lagged behind and was facilitated only 2000 msec after offset of the ironic target sentence. In contrast, familiar ironies (Tell me about it) facilitated initially both their literally (“disclosing”) and ironically (“known”) test words in both types of context:

178 Rachel Giora

(14) a.

b.

Ziv visited his friend, Ran, in New York. Ran advised him to use the subway, but Ziv insisted on renting a car. Three days later, Ziv gave up and told Ran: I have had enough. The traffic jam here is incredible. Ran said: “Tell me about it.” In the middle of the night Royi woke up and started crying. His mother heard him and went up to his room. “What happened?” she asked. Royi said that he had had a nightmarish dream. His mother said: “Tell me about it.”

These findings are accountable only by the graded salience hypothesis. They demonstrate that salient meanings are accessed initially, regardless of either context or literality. 2.3 Interpretation As shown earlier, initial access is automatic. It involves activating both contextually compatible and incompatible meanings on account of their salience. While it is plausible to assume that contextually compatible meanings integrate with the sentence or discourse context and contribute to the utterance interpretation, it is not clear whether contextually incompatible meanings are automatically discarded as irrelevant and excluded from the interpretation processes. 2.4 Predictions According to the tinge hypothesis (Dews et al. 1995; Dews and Winner 1995, 1997, 1999), the indirect negation view (Giora 1995, in press a; Giora et al. 1998) and joint pretense view (Clark 1996; Clark and Gerrig 1984), the contextually incompatible literal meaning is functional in irony interpretation. On the tinge hypothesis, it dilutes the criticism or the praise. According to the indirect negation view, it provides for a reference point relative to which the criticized state of affairs is weighed. According to the pretense view, it allows for intimacy, excluding the uninitiated audiences. Setting out from a functional viewpoint, then, these theories assume that the apparently incompatible literal meaning should not be discarded automatically as might be assumed by the standard pragmatic model (Grice 1975). 2.5 Findings In my studies with colleagues we tested this hypothesis. In Giora and Fein (1999a) and Giora et al. (1998), we show that in the ironically biasing context, the literal meaning of both familiar and less familiar ironies is active even after a delay (of 1000–2000 msec). Even two seconds after offset of the ironic target sentence, when the compatible (ironic) meaning becomes available, the level of activation of the literal meaning is not reduced.

Irony and its discontent 179

Using an offline measure, we show that while utterances embedded in literally biasing contexts (15b) are processed only literally, when embedded in ironically biasing contexts (15a) they retain both their literal and ironic meanings (Giora and Fein 1999b). Having read such passages, participants were able to fill in fragmented words (li–le; s–p) related to both the literal (“little”) and ironic (“stop”) meanings of the target sentence following the ironically biasing context, while mostly literally related test words following the literally biasing context: (15) a.

b.

After he has finished eating pizza, falafel, ice-cream, wafers and half of the cream cake his mother had baked for his brother Benjamin’s birthday party, Moshe started eating coated peanuts. His mother said to him: “Moshe, I think you should eat something.” At two o’clock in the afternoon, Moshe started doing his homework and getting prepared for his Bible test. When his mother came home from work at eight p.m., Moshe was still seated at his desk, looking pale. His mother said to him: “Moshe, I think you should eat something.”

Further, findings in Winner (1988), Dews et al. (1995), Dews and Winner (1997, 1999) show that critical and complimentary ironies are considered less aggressive/ complimenting than their direct, literal alternatives whose criticism and praise are muted (by the literal meaning). Research of naturally occurring discourse is also consistent with the view that the (salient) literal meaning of (unfamiliar) ironies is retained for further processes. For instance, Giora and Gur (submitted) and Kotthoff (1998, submitted) show that in friendly conversations, interlocutors rejoin the literal meaning of ironic utterances in order to produce humor. According to Kotthoff, this is not the case when unfriendly conversations are at stake, such as those taking place among hostile interlocutors participating in interview shows. The following (taken from Kotthoff 1998) is an example of concurring with the literal meaning, typical of friendly conversations: (16) a. b. c. d.

M: Du hasch grad son opulentes Sozialleben. You have been having such an opulent social life lately. D: total was los grad, weil ich nämlich initiativ A lot has been going on lately, because I have taken geworden bin jetzt. the initiative now. M: HAHAHAHAHAHA

In this extract, the ironic utterance (16a) is responded to literally, invoking selfirony. Given D’s known inclination to the opposite, M’s description of D’s ‘opulent’ social life is entirely inappropriate. Recently, however, D was involved in two dinner parties. While concurring with M’s literal contribution, D further elaborates on it by presenting ‘evidence’. His reference to his ‘initiative’, which he apparently lacks,

180 Rachel Giora

induces laughter. D’s response to the literal meaning is instant. He not only processes M’s irony, but immediately counters it by topping it, suggesting that the literal meaning is highly accessible and has not been suppressed by the comprehender. Indeed, the role irony plays in self-protection (Giora 1995, in press c; Groeben and Scheele 1984) is highly dependent on comprehenders both accessing and retaining the literal meaning. While the literal meaning may be highly misleading (cf. Wallach’s examples above), it nevertheless helps mask subversive ideas when impunity is imminent. Thus, even when laughing at ourselves, we may still dissociate ourselves from the norms echoed, which we have failed to meet (Kotthoff 2000). Indeed, Kotthoff shows that “female narrators organize their presentations so that other people do not laugh at their expense, but rather at the expense of norms which they mock collectively by laughing at them” (1999: 55). In all, evidence adduced in the lab as well as in naturally occurring conversations is consistent with the view that contextually inappropriate meanings, activated on account of their salience, are retained in the mind of interlocutors and are utilizable for further processes. Seemingly irrelevant information, then, is not discarded automatically once it may be instrumental.

3.

Conclusions

Empirical research demonstrates that initial access is automatic, salience sensitive, and impervious to filtering context effects: comprehension of unfamiliar (though commonplace) irony involves accessing the salient (usually, but not always, literal) meaning initially, irrespective of contextual information. The contextually compatible, nonsalient (usually ironic) meaning requires extra processing time for its derivation — it is a post access process affected by contextual information. However, while initial access is automatic, involving both contextually compatible and incompatible meanings, suppression of apparently incompatible meanings is not. Meanings utilizable while constructing the discourse are not discarded automatically even when contextually inappropriate (for instance the literal meaning of irony). Such processing mechanisms may explain irony’s beguiling nature. Under some circumstances, comprehenders may be content with the salient, literal meaning, terminating search for alternative interpretations. This may hold even for poetic texts (cf. Wallach’s poems 1–2 above), which tend to invite extra processing (Steen 1994). No wonder irony serves to convey subversive ideas: being processed saliently first, it has the potential of concealing the authentic intent, which may be missed by the uninformed audiences. Ironically, however, the self-same illusive nature may act like a boomerang. Taken at face value, irony may be risky for its initiator who will be penalized for what she said but did not intend (cf. Hutcheon 1994).

Irony and its discontent

Notes 1. On how children fail to understand irony see Winner (1988), and Giora (1998c) for a review. 2. The poem was originally published in Hebrew in 1983. 3. Tefillin are the phylacteries — a religious prop made of strings with which male Jews bind themselves in their Morning Prayer — a ritual which is denied to women. The English translation is taken from Wallach’s (1997). The poem was originally published in Hebrew in 1983. 4. For an interesting view about how irony highlights ironic context see Utsumi (2000). 5. But see Peleg, Giora and Fein (2001) for a more constrained view regarding context effects. 6. Most studies haven’t looked into the familiarity/salience factor. In our studies, we controlled for salience by asking native speakers to act as lexicographers and write down the context-less meanings of sentences and phrases that came to mind first. A sentence which received an ironic interpretation by more than half of the tested population was classified as familiar irony (tell me about it). Sentences not reaching that threshold (many of which were items used by other researchers such as Gibbs 1986a, Sperber and Wilson 1986/95 and Wilson and Sperber 1992) were classified as unfamiliar ironies (Giora and Fein 1999a). We also used completion of word fragments or word stems out of context to measure salience out of context (cf. Giora and Fein 1999b). 7. The set of metaphors tested is made up of conventional (salient) metaphors (Gibbs, personal communication). 8. In Pexman et al. (2000) degree of salience was established on the basis of norming data. 9. This incoherence, however, is not reflected in the readers’ judgements. 10. There was another target, which also exhibited longer reading times and judgments times than the ironic target — the nonsarcastic indirect request: “Why don’t you close the window?” However, given the rating study (p. 45), it is not clear whether subjects understood this target as a nonsarcastic question or as a nonsarcastic indirect request. It is quite possible that they understood it as a nonsarcastic question, contextual information being ambiguous between the two alternatives. 11. The studies were conducted in Hebrew. 12. On the role of suppression in figurative language comprehension see Gernsbacher et al. (in press). 13. Surprisingly, however, in the literally biasing context, the salient ironic meaning of familiar ironies was not suppressed after a delay of 1000 msec (Giora and Fein 1999a). It is possible that some meanings are so highly salient, they are difficult to discard.

References Attardo, S. 2000. “Irony as relevant inappropriateness”. Journal of Pragmatics 32: 793–826. Barbe, K. 1995. Irony in Context. Amsterdam/Philadelphia: Benjamins. Booth, W. 1974. A Rhetoric of Irony. Chicago: University of Chicago Press. Clark, H. H. 1996. Using Language. Cambridge, England: Cambridge University Press. Clark, H. H. and Carlson, T. B. 1982. “Hearers and speech acts”. Language 58: 332–373. Clark, H. H. and Gerrig, R. 1984. “On the pretense theory of irony”. Journal of Experimental Psychology: General 113: 121–126. Colston H. L. 1997. “Salting a wound or sugaring a pill: The pragmatic functions of ironic criticism”. Discourse Processes 23: 25–45.

181

182 Rachel Giora

Curcó, C. 1998. “Indirect echoes and verbal humor”. In Current Issues in Relevance Theory, V. Rouchota and A. H. Jucker (eds), 305–325. Amsterdam and Philadelphia: John Benjamins. Curcó, C. 2000. “Irony: Negation, echo and metarepresentation”. Lingua 110: 257–280. Curcó, C. in press. The Pragmatics of Humorous Interpretations: A Relevance-Theoretic Approach. Amsterdam: Benjamins. Dews, S. and Winner, E. 1995. “Muting the meaning: A social function of irony”. Metaphor and Symbolic Activity 10: 3–19. Dews, S. and Winner, E. 1997. “Attributing meaning to deliberately false utterances: The case of irony”. In The Problem of Meaning: Behavioral and Cognitive Perspectives, C. Mandell and A. McCabe (eds), 377–414. Amsterdam: Elsevier. Dews, S. and Winner, E. 1999. “Obligatory processing of the literal and nonliteral meanings of ironic utterances”. Journal of Pragmatics 31: 1579–1599. Dews, S., Kaplan, J. and Winner, E. 1995. “Why not say it directly? The social functions of irony”. Discourse Processes 19: 347–368. Fodor, J. 1983. The Modularity of Mind. Cambridge: MIT Press. Gernsbacher, M. A., Keysar, B. and Robertson, R. W. In press. “The role of suppression and enhancement in understanding metaphors”. Journal of Memory and Language. Gibbs, R. W. Jr. 1986a. “On the psycholinguistics of sarcasm”. Journal of Experimental Psychology: General 115: 3–15. Gibbs, R. W. Jr. 1986b. “What makes some indirect speech acts conventional?” Journal of Memory and Language 25: 181–196. Gibbs, R. W. Jr. 1994. The Poetics of Mind. Cambridge: Cambridge University Press. Gibbs, R. W. Jr. 1998. “Comments on what we have thus far”. In Figurative Language and Thought, A. Katz, C. Cacciari, R. Gibbs and M. Turner (eds). New York: Oxford University Press. Gibbs, R. W. Jr., O’Brien, J. and Doolittle S. 1995. “Inferring meanings that are not intended: Speakers’ intentions and irony comprehension”. Discourse Processes 20: 187–203. Giora, R. 1995. “On irony and negation”. Discourse Processes 19: 239–264. Giora, R. 1997a. “Understanding figurative and literal language: The graded salience hypothesis”. Cognitive Linguistics 7: 183–206. Giora, R. 1997b. “Discourse coherence and theory of relevance: Stumbling blocks in search of a unified theory”. Journal of Pragmatics 27: 17–34. Giora, R. 1998a. “Discourse coherence is an independent notion: A reply to Deirdre Wilson”. Journal of Pragmatics 29: 75–86. Giora, R. 1998b. “When is Relevance? On the role of salience in utterance interpretation”. Revista Alicantina de Estudios Ingleses 11: 85–94. Giora, R. 1998c. “Irony”. In Handbook of Pragmatics, J.-O. Östman, J. Verschueren, J. Blommaert and C. Bulcaen (eds), 1–21. Amsterdam: Benjamins Giora, R. 1999. “On the priority of salient meanings: Studies of literal and figurative language”. Journal of Pragmatics 31: 919–929. Giora, R. In press a. On our Mind: Salience, Context, and Figurative Language. New York: Oxford University Press. Giora, R. In press b. “Literal vs. figurative language: Different or equal?”. Journal of Pragmatics. Giora, R. In press c. “Masking one’s themes: Irony and the politics of indirectness”. In Thematics in Psychology and Literary Studies, W. van Peer and M.M Louwerse (eds). Amsterdam: Benjamins. Giora, R. and Fein, O. 1999a. “Irony: Context and salience”. Metaphor and Symbol 14: 241–257. Giora, R. and Fein, O. 1999b. “Irony comprehension: The graded salience hypothesis”. Humor 12: 425–436.

Irony and its discontent

Giora, R., Fein O. and Schwartz, T. 1998. “Irony: Graded salience and indirect negation”. Metaphor and Symbol 13: 83–101. Giora, R. and Gur, I. Submitted. “Irony in conversation: Salience and context effects”. In Polysemy, B. Nerlich (ed.). Glucksberg, S. 1995. “Commentary on nonliteral language: Processing and use”. Metaphor and Symbolic Activity 10: 47–57. Grice, P. H. 1975. “Logic and conversation”. In Speech Acts. Syntax and Semantics Vol. 3, P. Cole and J. Morgan (eds), 41–58. New York: Academic Press. Grice, P. H. 1978. “Some further notes on logic and conversation”. In Pragmatics. Syntax and Semantics Vol. 9, P. Cole and J. Morgan (eds), 113–127. New York: Academic Press. Groeben, N. and Scheele, B. 1984. Produktion und Rezeption von Ironie. Pragmalinguistische Beschreibung und psycholinguistische Erklärungshypothesen. Tübingen: Günther Narr. Haiman, J. 1990. “Sarcasm as theater”. Cognitive Linguistics 1: 181–203. Haiman, J. 1998. Talk is Cheap: Sarcasm, Alienation, and the Evolution of Language. Oxford: Oxford University Press. Hutcheon, L. 1994. Irony’s Edge: The Theory and Politics of Irony. London/New York: Routledge. Jorgensen, J., Miller, G. A. and Sperber, D. 1984. “Test of the mention theory of irony”. Journal of Experimental psychology: General 113: 112–120. Kotthoff, H. 1998. “Irony, quotation, and other forms of staged intertextuality: Double or contrastive perspectivation in conversation”. In List Working Paper 5. University of Konstanz. Kotthoff, H. 2000. “Gender and joking: On the complexities of women’s image politics in humorous narratives”. Journal of Pragmatics 32: 55–80. Kotthoff, H. Submitted. “Responding to irony in different contexts: Cognition and conversation”. Journal of Pragmatics. Kumon-Nakamura, S, Glucksberg, S. and Brown, M. 1995. “How about another piece of the pie: The allusional pretense theory of discourse irony”. Journal of Experimental Psychology: General 124: 3–21. Kreuz R. and Glucksberg, S. 1989. “How to be sarcastic: The reminder theory of verbal irony”. Journal of Experimental Psychology: General 118: 347–386. Long, D. L. and Graesser, A. 1988. “Wit and humor in discourse processing”. Discourse Processes 11: 35–60 Peleg, O., Giora, R. and Fein, O. 2001. “Salience and context effects: Two are better than one”. Metaphor and Symbol. Pexman P., Ferretti, T. and Katz, A. 2000. “Discourse factors that influence irony detection during on-line reading”. Discourse Processes 29: 201–222. Rattok, L. 1997. Angel of Fire: The Poetry of Yona Wallach. Tel Aviv: Hakibbutz Hameuchad (In Hebrew). Schwoebel, J., Dews, S., Winner, E. and Srinivas, K. 2000. “Obligatory processing of the literal meaning of ironic utterances: Further evidence”. Metaphor and Symbol 15: 47–61. Searle, J. 1979. Expression and Meaning. Cambridge, England: Cambridge University Press. Sperber, D. 1984. “Verbal irony: Pretense or echoic mention”. Journal of Experimental Psychology: General 113: 130–136. Smith, N. and Wilson, D. 1992. “Introduction”. Lingna 87: 1–10. Sperber, D. and Wilson, D. 1981. “Irony and the use-mention distinction”. In Radical Pragmatics, P. Cole (ed.), 295–318. New York: Academic Press. Sperber, D. and Wilson, D. 1986/1995. Relevance: Communication and Cognition. Oxford: Blackwell. Steen, G. 1994. Understanding Metaphor in Literature. London: Longman.

183



184 Rachel Giora

Utsumi, A. 2000. “Verbal irony as implicit display of ironic environment: Distinguishing ironic utterances from nonirony”. Journal of Pragmatics 12: 1777–1807. Vu, H., Kellas, G. and Paul, S. T. 1998. “Sources of sentence constraint in lexical ambiguity resolution”. Memory & Cognition 26: 979–1001. Walker, N. 1991. Feminist Alternatives. Jackson and London: University Press of Mississippi. Wallach, Y. 1997. Wild Light. Translated by L. Zisquit. Riverdale-on-Hudson, New York: The Sheep meadow Press. Wilson, D. and Sperber, D. 1992. “On verbal irony”. Lingua 87: 53–76. Winner, E. 1988. The Point of Words: Children’s Understanding of Metaphor and Irony. Cambridge Mass. London: Harvard University Press. Yus, F. 1998. “A decade of relevance theory”. Journal of Pragmatics 30: 305–345. Yus, F. 2000. “On reaching the intended ironic interpretation”. International Journal of Communication 10: 27–78.



Psychoanalysts and daydreaming Willie van Peer and Ingrid Stoeger University of Munich

Thomas Mann, in a speech held at the occasion of Freud’s 80th birthday on May 8th, 1936, emphasizes the special relationship between literature and psychoanalysis. In one respect, however, he is mistaken: he regrets that Freud went his way all alone, “without having a knowledge of the consolation and tonic that great literature would have had in store for him”1 (Mann 1936: 134) [trans. WvP/IS]. This is not really correct. Freud published on Dostojewsky and Jensen, and kept up a regular correspondence with writers such as Hermann Hesse, Arthur Schnitzler and Stefan Zweig. His essay on “Creative Writers and Daydreaming” (1908) led many theorists in their efforts how to conceive of the relation between the author and his works. Freud set the example for many psychoanalytic critics, who see in literature the expression of the individual or collective unconscious. Serious doubts may have been voiced over the past years about the correctness of Freud’s views, over the methods he used and over his integrity as a scientist.2 Yet, if psychoanalysis is alive and thriving, it certainly is so in literary studies. This chapter investigates some of the claims made by psychoanalytic scholars of literature, by putting them to an independent test, something that has not — to our knowledge — been carried out before.3

1.

Methodological considerations

To empirically minded scholars, there is no reason to either believe or disbelieve the claims of the psychoanalysts. To take psychoanalysis seriously is to study its insights in reality, by observing how real people go about when they read literature, and subsequently to compare these observations with the claims made by psychoanalysis. In literary studies, at least some developments allow a more concrete way to discuss psychoanalysis and its premises. Reader response research, for instance, attempts to describe what responses readers have when they confront particular texts. Such an attempt evidently has merits, since it allows one to more clearly see the interaction between what is ‘on the page’ and what goes on ‘in the mind’.

186 Willie van Peer and Ingrid Stoeger

That such descriptions are based on psychoanalysis does not in itself invalidate them, as long as it yields descriptions that are trustworthy and in accordance with independently observable facts. As Webster (1995: 439) observes: “Scientific ideas should be judged not by their intellectual provenance but by their explanatory power.” What matters is not that statements about reader response originate in a psychoanalytic context of discovery, but whether those statements are justifiable on the basis of evidence the gathering and analysis of which can be carried out regardless of one’s ideological position. It is justification of assertions that matters, not the framework in which they originate. According to this view, proponents and opponents of psychoanalysis should stop bickering and instead put their claims to independent tests. This is fully in accordance with the conclusion Barbara von Eckardt arrives at: “Freudian theory is perfectly scientific in the sense of being a candidate for scientific testing” (1998: 114). In general, such a test has to fulfil two criteria. First, the test must involve a real risk for the theory to be rejected. Where no such risk is involved, or where the results of the test can always be ‘interpreted’, so that even a negative outcome can be reformulated as support for the theory, or where those negative results are explained away, researchers will draw a wrong conclusion, namely that the theory is correct, where in actual fact it is not. In such a situation, the standards of evidence are so low that almost any theory will pass the test, and we are unable to differentiate between correct and incorrect theories. Secondly, the theory to be tested must be given a fair chance to survive the test. Where no such fairness is built into the test situation, the results will lead researchers to the opposite wrong conclusion, namely that the theory is wrong, where in actual fact it is correct, but this correctness is hidden behind standards of evidence that are too high for evidence in its favour to emerge.4 Finding the right balance between these two requirements, that is, determining how high or low to set the mark for the evidence to involve both genuine risk and due fairness, is always a critical matter and has given rise to the development of a methodology of its own within the social sciences. In this study, the risk for the theory consists in the fact that the results of the test are not to be influenced by the researchers themselves, so that no tampering with the outcome is allowed. It is readers who decide what they experienced during and after reading, not the analysts. The fairness consists in the same openness (readers may give answers to our questions that corroborate the theory’s assertions), but also in the relatively mild criteria that we set to judge the outcome of the experiment: the level at which acceptance of the theory was imperative, was set at the conventional 5% level, but results occurring with a probability of 10% will also be inspected, so as to allow a great range of detection possibilities of any relationships that the theory predicts. Again, one may not relax such levels too far without endangering the first criterion; in the current context we reasoned that the acceptance of probability levels higher than 10% would indeed destroy or seriously undermine the risk factor in the test at hand. We believe

Psychoanalysts and daydreaming 187

that by respecting these two criteria and carrying them out in the way described we have presented psychoanalysis with a reasonable test of some of its premises: the threat to the theory is real (has to be real, for otherwise there would be no test), but valid assertions made by the theory have a high chance of being detected.5

2.

Hypotheses

As a text type to be used in investigating Freud’s claims, we decided to use a fairytale. Not only did Freud consider in diverse works the relation of the content of fairy-tales to the unconscious, several critics have also expanded Freud’s remarks and related them to reader behavior in a way that allows empirical scrutiny. Freud himself sees fairy-tales as forming part of a collective unconscious; the fact that much semantic material turns up both in dreams and in fairy tales is interpreted by Freud as evidence for the importance such themes have in our lives.6 The relationship between dreams and mythic material to be found in fairy tales is described more concretely than Freud did by Abraham, in drawing up parallels between the past of a culture and the childhood of an individual (1909: 46), and in transferring the wish fulfillment typical for dreams on to myth: “Myth is a piece of overcome infantile inner life of a people. It contains (in a veiled way) the childhood wishes of the people”7 (ibid.). Abraham makes this assertion with respect to myths, but it is neither unusual in the vast literature on this subject to assert essentially the same about fairy tales. Similar utterances are to be found in the literature, for instance in Lenz (1971: 12): “Fairy tales are the real dreams of the nations”8 while Lüthi (1961: 9) submits that in fairy tales we encounter “humans’ primal experiences and primal desires.”9 The explanation for the occurrence of this material both in text genres like fairy tales and in our dreams lies in our fascination with them, which Freud explained in his essay on the “Familienroman der Neurotiker” (1908/09: 221–6). According to Freud, this fascination originates in the fact that in reading fairy tales, otherwise socially unacceptable wishes emerge.The child often hates its parents, because they are seen as the major source of all prohibitions. This hate is mixed, however, with fear of death, since the child is for its own survival dependent on its parents. In reading about step parents — and applying this in fantasy to one’s own situation, this fear of death is separated from the situation and vanishes. This now allows for the unchecked hate feelings against one’s parents to be set free, together with an incestuous wish toward one’s brother(s)/sister(s). Starting from the premise of a double pair of parents, as described by Freud, Schönau remarks “That the bad step-mother or the man-eating witch of the fairy tale psychologically also refer to the mother, whereby the child’s aggression against the mother is projected on to this image, is generally known” (1991: 41/42).10 This

188 Willie van Peer and Ingrid Stoeger

interpretation, which Schönau presupposes as “generally known” (“allgemein bekannt”), is indeed to be found in other authors’ works too, as can be observed in the various contributions to the volume edited by Laiblin (1995). Of fundamental importance for the effect that texts have, according to Schönau, is Freud’s discovery, that “the story’s fascination depends on its unconscious appeal” (1991: 42). And further: “The potential effect of the work lies to a high degree in the latent fantasy about the family, to which the reader reacts without grasping what moves him”11 (ibid.). For Schönau, Freud’s interpretations of Oedipus and Hamlet are “no text interpretations, but theories about effects, hypotheses explaining the timeless fascination these texts exert” (ibid.).12 For Schönau, the importance of Freud’s Oedipus interpretation lies in the indication “that Oedipus’ fate grips us only”, now citing Freud, “because it could have become ours”, and again Schönau: “that it confronts us with our repressed childhood wishes and thus lead to catharsis, a relief of old guilt and anxiety, without us being aware of these processes” (1991: 43).13 In this process, the effects may be all the stronger when not everything is made explicit: “That means: attention is drawn to other than the latent fantasy material, so that all the more of it can appeal to the public’s unconsciousness” (ibid.).14 Schönau thus describes the effects that fairy tales have on readers in a concrete way — though these readers may not be conscious of them. The assertion is made on the basis of Freud’s writings, and, perhaps, some kind of introspection, but not on the basis of observation of readers actually reading (or having read) such fairy tales. Introspection, however, is an insufficient method to do research, and Freud has been proven wrong on scores of other points he made on the basis of its use. It is not clear why, if the presumed effects are so important to human life, psychoanalysts are content with such poor methodology. It is almost as if astro-physicists would study the internal combustion of stars with the help of a clinical thermometer. Surely issues that are deemed of such scope and importance deserve to be studied with greater alacrity than the tools of citation and introspection? Because we consider the potential effects of reading worthy of serious investigation, and because we do not wish to exclude beforehand all that is predicated by psychoanalysis, we will here undertake an independent investigation of the claims made by psychoanalysis in this respect. We hope that the reader shares with us the assumption that if psychoanalytic claims about the influence of literature in our lives are claims about the real world, they must be accessible through rational and empirical investigation. That is, we presuppose that psychoanalysis does not deal in magic, nor in the working of miracles, but instead wishes to be understood as a theory dealing in propositions about reality. To put such propositions to the test, then, is the primary purpose of the present chapter. This requires independent data, however. As Crews (1998b: 273) has pointed out: “The most fundamental rule (…) is that a given theory or hypothesis cannot be validated by invoking ‘evidence’ manufactured by that same supposition.”

Psychoanalysts and daydreaming 189

If reading fairy tales indeed creates the effects described above, namely that our repressed feelings of hate against our parents are activated (especially against the mother, since she is usually the one most deeply involved in the child’s upbringing) and that it generates incestual wishes for other family members in our unconscious, this must be observable (or the theory lacks any empirical basis in reality). In the present study, therefore, the following two hypotheses are tested: 1. fairy tales stimulate the reader to allow repressed feelings of hate against his/her mother; 2. the reading of fairy tales sets free and intensifies incestual wishes directed at one’s brothers and sisters.

3.

Method

3.1 Materials In order to investigate these claims, we made use of a fairy tale from the collection of the brothers Grimm, “Allerleirauh”. Allerleirauh is a king’s daughter in prepuberty when her mother dies. On her death-bed her mother makes her father promise only to remarry a woman as beautiful as she. In all his kingdom, there is only one woman who answers this condition, namely Allerleirauh, his daughter. When Allerleirauh hears about her father’s intention to remarry, she tries to persuade him not to. On failing to do so, she escapes, arriving at a king’s court, where she must live in poverty during a long time, until the king sets his eye on her and ultimately marries her. What further happened to her father is not told. There are several grounds why we have selected this fairy tale. On the one hand, it is a tale that is relatively unknown to many readers. Using a tale like “Cinderella” or “Snow White”, with which many readers are familiar, may obfuscate the results in that those readers provide culturally accepted answers, or provide interpretations that they have heard before, instead of giving us their real reactions. On the other hand, the tale contains many features that make it an interesting one to test psychoanalytic theories. Rank (1974), for instance, observes that “the incest, which initially is consciously intended, then becomes possible unconsciously, in that the daughter disfigures her face” (1974: 365)15 and asserts that Allerleirauh’s father and the king she later marries are the same, so that the story is about real incest (ibid., note 1). Interesting in this respect are the observations of the participants in our experiment (see next section). Furthermore, Spörk has seen in Rank’s explanations an active role of the daughter, but believes that for Rank this role results “rather from the father’s need to justify himself than from the daughter’s really existing inclinations” (1985: 195).16 One can interpret Rank’s position also differently, however. Since he mentions at other places in his book that incest was widely

190 Willie van Peer and Ingrid Stoeger

practiced by the old Germanic tribes and other peoples (1974: 397) and thus psychoanalytically refers to our collective unconscious, it is more probable to see in Rank’s interpretation a position that incest in this fairy tale is actually the wish of both parties. Still another interpretation by Mallet (1982: 126 ff.) assumes Inzestspielen but fear for real incest on the side of the daughter, so that she finally has to flee. When her father really approaches her, the infringement of reality in the child’s fantasy world therefore develops hate against her father, all the more so since she is given three robes by her father, that are interpreted as “the reward or the result of the latent erotic relation with the father” (1982: 132).17 Apart from the incest motive, the fairy tale is also interesting in respect of the relation between mother and daughter. Thus, the death of the mother and her last wish are instrumental in causing feelings of hate to emerge in the child. If the child had wished the death of the mother (to marry the father, comparable to the story of Oedipus), it will now experience guilt. At the same time, it will feel deserted by the mother, as it will interpret her death as a form of Verlassenwordensein (being left), irrespective of how guilty the mother really is of her own death. Important also is that the mother, through her last wish, indirectly causes the daughter to flee. Thus the constellation in this fairy tale is such that it presents an almost ideal case for Freud’s (and Schönau’s) speculations about the emergence of hateful feelings toward the mother. Following the psychoanalytic assumptions that incestual thoughts are unconsciously entertained by all of us, the tale of Allerleirauh also presents a good test-case for the second hypothesis. As a control text we used “Die beiden i” (“The two i’s”) by the Swiss author Franz Hohler. It relates how in a printer’s shop a letter i falls in love with another i, so that they get married into a ü.18 Since this leads to printing errors, the ü is taken out of the font set, which leads to an action of solidarity on the part of all i’s, who all synthesize into ü’s. The printer then throws away all ü’s and replaces them by clean i’s, who, however, spontaneously join together into ü’s. As will be readily perceived, this text does not deal with family matters, though, admittedly, it touches on issues of love, but in a rather unusual and also in a highly abstract way. Instead, the story is centered around language. We hold that the fairy tale and the control text present very different scenarios for the reader to enact: in the former case a somber and sad story of a young child — with whom one can identify — eventually being saved from dark drives, in the latter a witty language game without human agents — and with whom nearly all possibilities of identification are excluded. 3.2 Questionnaire Participants were handed out the text together with a questionnaire containing altogether 24 questions, which they filled out immediately after having read the text.

Psychoanalysts and daydreaming

On average, the whole procedure took about 25 minutes. Presumably psychoanalysts would argue that it is not possible to probe feelings of hate against one’s parents or incest fantasies directly, as consciousness would then deny them. Therefore, we circumvented such direct probes in that we asked more indirectly about the general relations in the family and the feelings that members had for each other. If psychoanalytic theories about reading are correct, one should expect (as a result of the exposure to the repressed materials in the fairy tale) more negative descriptions of these feelings and of the general atmosphere in the family than in the control group. Likewise, psychoanalytic views of reading predict a stronger wish for bodily and mental proximity and commonality to other children in the family for those readers exposed to the fairy tale than for the control group. Instead of probing for feelings of hate against their mother, for instance, we asked participants how they related to their mother during childhood. We assumed that on a correct version of psychoanalysis, readers of the fairy tale would score this relationship in relatively more negative terms than the readers of the control text, where according to the theory no stimulus for a reactivation of these negative feelings are present. 3.3 Subjects Subjects were 60 students of the university of Munich, predominantly from the humanities, selected on the basis of voluntary participation in the cafeteria of the Institute of German Philology, with an equal distribution across the sexes. Average age in the experimental group was 25, of the control group 26.5 years. Subjects were assigned randomly to either the ‘experimental’ group (who read the fairy tale) or the ‘control’ group (who read the control text). A between subjects design was preferred over a within design in order to minimize the danger of revealing the aims of the study. 4. Results 4.1 Introduction Most subjects wanted to know what the study was about, but were satisfied with the explanation that it dealt with reading processes. After the experiment, many were interested in the details of our study. When we mentioned the psychoanalytic hypotheses, they rejected those (surprisingly, one could say, because literary studies form a stronghold of psychoanalysis these days) out of hand. Some participants communicated their emotions about the fairy tale’s content, especially with regard to the brutality facing the girl in the story. To many, it was unambiguous that the groom was the father of Allerleirauh, although they acknowledged that this was not said so explicitly.

191

192 Willie van Peer and Ingrid Stoeger

We will first discuss those questions (using 5-point Likert scales) probing readers’ relations with their parents (Hypothesis 1); in a following section the results concerning Hypothesis 2 (relations to brothers/sisters) will be presented. Low values always mean a tendency to ‘agree’ with the statements under consideration, high values a tendency not to agree with the statement. To facilitate interpretation of the tables, we have used some extra symbols: ‘x’ refers to a confirmation of the hypothesis, ‘o’ for results negating the hypothesis, while values between brackets indicate negligible differences. Significant differences between experimental and control groups are marked with an exclamation mark. Significance was defined at the conventional level of .05, with the level of p < .01 being considered highly significant. All group differences were statistically tested with the non-parametric Mann-Whitney U-Test. 4.2 Relations with parents Table 1 summarizes the results of Question 7. According to the first hypothesis, readers in the experimental group will describe their relations to their parents in more negative terms (in comparison to the control group).19 Hence it is to be expected that readers of the fairy tale will give more denying responses to positive adjectives (such as ‘friendly’, ‘open’, ‘tolerant’, ‘warm-hearted); these adjectives will find stronger support in the control group, and vice versa: negative adjectives (such as ‘sad’ or ‘controlling’) will be agreed to more often in the control group, while being denied in the experimental group. Table 1.Question 7) How would you describe the atmosphere in your family?

a) b) c) d) e) f)

sad friendly open controlling tolerant warm-hearted

Fairy

Hohler

4,03 1,59 2,33 3,13 1,97 1,47

3,83 1,43 2,23 2,63 2,13 1,60

o x x o! p = 0.0526 o o

As Table 1 shows, however, of the six subquestions probing subjects’ relations to their parents, four go against the prediction, of which one difference (‘controlling’) is nearly significant (p = 0.0526).20 With question 8 (“When you think about your childhood …”) it was to be expected that the experimental group would give rather negative answers — reflecting the hidden aggression against their parents those readers have activated through exposure to the fairy tale. This should express itself in higher values to the

Psychoanalysts and daydreaming 193

subquestion “you are rather satisfied with it’, and lower values to the other questions. As Table 2 shows, this expectation was borne out in (b) and — with a significant p-value of 0.0268 in (c). In (a) there is hardly a difference to be noted, but in (d) there is a significant difference in the direction opposite to the prediction. Since this was the question that most directly thematized the childish fantasies, its p-value of 0.0322 casts serious doubt over the overall results for this question. Table 2.Question 8) When you think about your childhood, would you say that

a) b) c) d)

it could have been better? it could have developed much better? you are rather satisfied with it? you often wished yourself other parents?

Fairy

Hohler

2,79 3,73 1,83 4,72

2,69 4,17 1,36 4,28

o x x! p = 0.0268 o! p = 0.0322

With question 15 (‘what one had wished during childhood’) the theory would predict lower values for the experimental group: after reading the fairy tale participants would hold their parents responsible for all real or perceived frustrations, or at least comparatively more so than readers in the control group. At first sight, the results in Table 3 appear rather balanced. If one neglects minor differences between the groups (the ones between brackets), however, all remaining differences go against the theoretical predictions, and one of these (g), approaches significance. Table 3.Question 15) What did you wish to have in your childhood?

a) b) c) d) e) f) g) h) i) j) k)

more beautiful clothes more cuddling with parents more freedom more time together with my mother more time together with my father more activities with brothers/sisters more visits of friends more open relationship to parents more open relationship to brothers/sisters more common games more fun

Fairy

Hohler

4,03 4,17 3,60 4,07 3,70 3,91 3,72 3,57 3,70 4,03 3,87

3,79 4,03 3,23 4,14 3,76 3,93 3,17 3,62 3,69 3,53 3,90

o o o (x) (x) (x) o p = 0.0708 (x) (o) o (x)

The three previous questions dealt with respondents’ relations to the family as a whole. Question 17 probes their relationship to their mother. In the experimental group one would expect higher values (“do not agree”) for adjectives such as

194 Willie van Peer and Ingrid Stoeger

‘tender’ or ‘understanding’ than in the control group. The results, however, show quite a different picture. As can be seen, none of the responses is in the predicted direction, and on one adjective (‘stipulating’) the difference between experimental and control group approaches significance. Table 4.Question 17) How did your mother relate to you?

a) b) c) d) e) f) g) h) i) j) k) l) m)

tender tolerant friendly distanced unjust stipulating open controlling understanding fickle-minded ambitious interested warm-hearted

Fairy

Hohler

1,73 2,00 1,37 4,27 4,30 3,63 2,07 3,13 1,73 4,13 3,57 1,57 1,53

1,87 2,07 1,63 4,17 3,90 3,10 2,50 2,97 2,07 4,03 3,37 1,63 1,70

o (o) o o o o p = 0.0721 o o o o o (o) o

Question 20 investigated the opposite, that is how subjects themselves relate to their mother. If psychoanalytic theory is correct, readers of the fairy tale should be inclined to describe this relationship in more negative terms, leading to higher values for positive and lower values for negative adjectives. As Table 5 makes abundantly clear, however, none of the results confirm the theoretical prediction, with one scale (‘distanced’) approaching significance in the opposite direction. Taken together, psychoanalytic predictions, out of 35 items tested, barely four are in the direction predited by psychoanalytic theory. Thus an overwhelming Table 5.Question 20) How would you describe your relation with your mother now?

a) b) c) d) e) f) g) h)

tender tolerant friendly distanced supportive cool open warm-hearted

Fairy

Hohler

1,93 1,62 1,38 4,00 1,62 4,48 2,03 1,69

2,10 1,97 1,57 3,50 1,80 4,23 2,23 1,80

o o o o p = 0.0764 o o o o

Psychoanalysts and daydreaming 195

majority of readers’ responses go against the grain of what psychoanalysis asserts they should be doing.21 4.3 Relations to brothers/sisters But perhaps the theory does better with respect to the second hypothesis? Since several participants grew up in one-child families, the number of subjects responding to these questions decreased to 22 for the experimental and 24 for the control group. The balance between male and female participants was not, however, disturbed by this decrease. The second hypothesis asserts that through reading the fairy tale, participants would experience more incest phantasies. Since it is doubtful whether one could probe such phantasies directly, we tried to circumvent this problem by asking readers about their general emotions toward the other family members, assuming that the conjectured incest phantasies correlate with a greater wish for more companionship and (physical and mental) proximity to their respective brothers and/or sisters. The results in Table 6, however, tell a different tale. Of six items, only one item (f) shows an effect in the predicted direction, but it is also Table 6.

Question 18) How would you describe your relationship as a child with your (beloved) brother/sister?

a) we did a lot of things together b) we did a lot of cuddling c) I would have liked to do more things together with him/her d) I liked to be in his/her bed e) I would have liked to be more familiar with him/her f) my first great love

Fairy

Hohler

2,38 4,10 4,10

2,38 4,13 3,00

(o) o! p = 0.0019

4,48 3,86

3,91 3,00

o o! p = 0.0244

4,57

4,74

x

non-significant. The significant items (c and e) go in the opposite direction, with (c) even being highly significant for the control text! The responses also show a highly significant difference (p = 0.0011) between male and female readers, the latter responding more strongly in the direction of the hypothesis. We also asked for subjects’ current relations to their brothers/sisters. Here again, the theory predicts a higher desire for communality in the group that had read the fairy tale. Table 7 shows that readers behave dramatically different: none of the differences between experimental and control group shows support for psychoanalytic theory. One result that negates the theory (d) is significant, while one (a)

196 Willie van Peer and Ingrid Stoeger

Table 7.

a) b) c) d) e)

Question 19) What do you associate as an adult now concerning your (beloved) brother/sister?

we still do a lot of things together we phone each other a lot I would like to know more about him/her I would like to have a more familiar relation I would like to do more things with him/her

Fairy

Hohler

2,95 2,68 3,45 3,86 3,05

3,42 3,21 3,17 3,04 2,75

o p = 0.1038 o o o! p = 0.0294 o

approaches significance. There were no significant differences between sexes. Summarizing the results concerning the relation towards brothers/sisters again allows only one conclusion: not only do the results not conform to the theory: out of 9 items tested, only one difference between the experimental and control group is in the direction that psychoanalysis predicted. Hence also these results show a strong tendency in the opposite direction, one which is almost certainly not due to chance.

5.

Discussion

If one accepts our belief that the test of literary psychoanalysis carried out here contained both a genuine risk and genuine fairness, the results clearly speak against its claims. The theory of literary psychoanalysis as is found in current work, such as Schönau (1991), is clearly a wrong theory. Overwhelmingly, readers reacted in ways that should be generally characterized as the opposite of what psychoanalysis asserts. No relation whatsoever could be found between the reading of fairy tales (of this particular kind) and increased feelings of hate against parents or incestuous fantasies concerning brothers/sisters. And remarkably, readers of a control text, in which none of the usual thematic ingredients held responsible for such feelings (as theorized by psychoanalysis) were present, responded in ways that strongly resemble the way which psychoanalysts take as symptomatic of repressed childish fantasies about one’s parents and brothers/sisters. Whatever caused readers in the control group to react in the way they did remains unclear, but it certainly can hardly be explained psychoanalytically, since the necessary semantic ingredients for it to operate were absent. Readers in our experiment were young adults. It is possible, of course, that the mechanisms of family fantasies as described by psychoanalysis are more prominent, and therefore easier to observe, with children than with adults, and thus the results would not undermine whatever correctness the theory might have with respect to the psyche of the child. Yet they do undermine psychoanalysis’ claims that adults

Psychoanalysts and daydreaming 197

still have such repressed feelings. Whatever children may think and feel when they are confronted with semantic material of the type encountered in “Allerleirauh” falls outside the scope of the present study, but inasfar as psychoanalysis claims that reading fairy tales also steers adults in the direction of increased hate feelings toward one’s parents and an intensification of incestuous wishes vis-à-vis one’s brothers or sisters, it is clearly at odds with the empirical data that readers provided us with. It is not arguable that those readers provided these data because they wanted to comply with our wishes, because we did not have such wishes, but instead were open to any pattern of results that the experiment would yield. Even if we had (unconscious!) wishes in this respect, participants would presumably have complied — if they wished to do so — uniformly in both the experimental and control group. Of course the present study does not ‘prove’ that psychoanalysis is wrong. What it does prove, however, beyond any doubt, is that the assertions about reading behavior made by (some) psychoanalysts definitely are wrong for the readers who took part in our experiment. Since psychoanalysis grants itself the right to make such assertions, honesty will have to put constraints upon such rights: obviously there are readers for which the assertions do not hold at all. That does not mean, of course, that there may never be readers for which psychoanalysis’ claims hold. We are talking here about the relationship between sample and population. The claims made by Schönau and others are geared toward the whole population of adult readers, without differentiation. That claim has been rendered false by our present investigation. Is it possible, therefore, that the readers in this investigation distinguish themselves from other readers whom psychoanalysis envisaged? There may be hundreds of grounds for such differences, none, however, that we can think of as systematic: we cannot think of any fundamental reasons why our readers should be considered as deviant from any other adult population. Surely we cannot exclude the very possibility that our readers are ‘abnormal’ with respect to the claims that psychoanalysts have made about reading behavior, but it is up to those who would plead for such a distinction (in order to save psychoanalysis) to come up not just with plausible arguments, but with evidence corroborating the nature of such a distinction. Maybe (some) psychoanalysts will take issue with our research and point out that we have not really ‘understood’ what psychoanalysis is all about, that one cannot really ‘experiment’ with people as if they were guinea-pigs, that one cannot carve up individual mental processes while reading and separate them from the ‘whole’ person, and so forth. These are catch phrases that some psychoanalysts sometimes use to immunize their theory against criticism. If psychoanalysis can only be understood by psychoanalysts, then it shows more affinity with religion than with science. Even the most difficult concepts in the natural sciences can be grasped by a person willing to devote time and energy on them. Why this should be different with respect to psychoanalysis has never been made clear. Secondly, we did

198 Willie van Peer and Ingrid Stoeger

not ‘experiment’ with people: our readers carried out the everyday action of reading in a situation that we made as natural to them as the circumstances allowed. Finally, our experiment in no way detached readers’ mental experiences from their whole person. On the contrary, we provided them with possibilities to express themselves freely on the nature of the reading task and the questions we had asked them. None of them found the task artificial or outlandish, none complained about being taken to pieces, and none accused us of making them feel inhuman. It is possible that psychoanalysts will dispute the present findings on the basis of the psychoanalytic mechanism of ‘censure’: the feelings of hate against parents or the incestuous desires concerning brothers/sisters occur at the unconscious or subconscious level, and will be vehemently denied in consciousness. They may thus see the results as a confirmation of psychoanalysis, in that these feelings were strongly inhibited in those readers who had read the fairy tale, and that they were relatively relaxed in the control group. However, this is not what psychoanalysis says: as we have made clear in section 3, psychoanalysts claimed that activation of such feelings, not their inhibition, would occur in the reading of this type of fairy tale. Schönau predicted feelings of aggression against the mother; we looked for such feelings and could not find them. Does this mean that these feelings did not occur? In all fairness, one has to admit that the results of our experiment do not allow such a radical conclusion. A case could be made for a reformulation of the theory, providing for an arousal of the predicted feelings during reading and an inhibition of these feelings in the response phase of filling out the questionnaire. Such emendation is certainly possible and in this sense we have not ruled it out as an alternative explanation of our results. At the same time, however, it should be clear that such a revision of the theory remains purely speculative for the time being. There is no empirical evidence for such a two-stage model at the moment. It could be provided, though, with the help of physiological measures indicating emotions (pulse, GSR, EEC, and so on) taking place during the reading process and comparing these reactions with verbal response measures after the reading. We hope that psychoanalysts will go to the trouble of putting this alternative version of the theory to the test as well. What psychoanalysis cannot hope for, however, is that their theory claims both activation and inhibition of those feelings at the same time, or without specifying the conditions under which either occurs, without giving away their theory as a tautology. It thus seems as if psychoanalysis has more in common with daydreaming than its practitioners are aware of. Will they be woken up, we wonder?

Notes 1. “ohne der Trost- und Stärkungsmittel kundig zu sein, die die große Literatur für ihn bereit gehalten hätte”

Psychoanalysts and daydreaming 199

2. There is now a considerable body of literature documenting the shortcomings of psychoanalysis; see for instance, Cioffi (1998), Crews (1995), Eysenck (1985), Farell (1996), Gellner (1996), Grünbaum (1987, 1993), MacMillan (1997), Scharnberg (1993), Webster (1995). Crews (1998a) offers an anthology of major critical statements. 3. Which is not to say that there has been no empirical work whatsoever in the psychoanalytic tradition; the research by Norman Holland immediately springs to mind here. 4. In the methodological literature, these two kinds of failure go under the name of Type One and Type Two error. 5. This test does not settle all matters, of course. But at least it makes a beginning to independently test some of the claims made by psychoanalysis. 6. For a non-psychoanalytic explanation of such themes, see Van Peer (2001). 7. “Der Mythus ist ein Stück überwundenen infantilen Seelenlebens des Volkes. Er enthält (in verschleierter Form) die Kindheitswünsche des Volkes.” 8. “Märchen sind die Wahrträume der Völker”. 9. “Urerlebnissen und Urwünschen der Menschen.” 10. “Daß die böse Stiefmutter oder die menschenfressende Hexe des Märchens psychologisch auch die Mutter bedeuten, wobei die Aggression des Kindes gegen die Mutter auf dieses Bild projiziert wird, ist allgemein bekannt.” 11. “die Faszination der Geschichte (ab)hängt von der unbewußten Appellstruktur (…) Das Wirkungspotential des Werkes liegt zu einem wichtigen Teil in der latenten FamilienromanPhantasie, auf die der Leser reagiert, ohne ganz zu begreifen, was ihn ergreift.” 12. “keine Textinterpretationen sondern Wirkungstheorien, Hypothesen zur Erklärung der zeitlosen Faszination dieser Werke” 13. “daß Ödipus’ Schicksal uns nur darum ergreift (…) weil es auch das unsrige hätte werden können (…) daß es uns also mit unseren verdrängten Kindheitswünschen konfrontiert und so eine Katharsis herbeiführen kann, eine Entlastung von alter Schuld und Angst, ohne daß uns diese Vorgänge bewußt sind.” 14. “Das heißt: Die Aufmerksamkeit wird auf anderes als auf das latente Phantasiematerial gelenkt, damit umso mehr davon an das Unbewußte des Publikums appellieren kann.” 15. “der anfangs erstrebte bewußte Inzest dann unbewußt ermöglicht (wird), indem die Tochter ihr Gesicht entstellt” 16. “eher aus einem Rechtfertigungswunsch des Vaters als aus den real existenten Neigungen der Tochter” 17. “Lohn oder das Ergebnis der … latent erotischen Beziehung zum Vater.” 18. Which is, of course, a regular letter of the German alphabet. 19. It is important to stress this comparative nature of the enterprise. It is an outcome of the ceteris paribus principle, a notion that is at the heart of all scientific enterprise. 20. Symbols used: x: results as predicted o: results against prediction !: p < 0.05 (o) or (x): negligible intergroup differences 21. We have left out those questions where hardly any difference between the experimental and control group was registered.



200 Willie van Peer and Ingrid Stoeger

References Abraham, K. 1909. “Traum und Mythus”. In Laiblin (1995): 44–48. Cioffi, F. 1998. Freud and the Question of Pseudoscience. Chicago: Court. Crews, F. 1995. The Memory Wars: Freud’s Legacy in Dispute. NY: New York Review of Books. Crews, F. (ed.). 1998a. Unauthorized Freud: Doubters Confront a Legend. Harmondsworth: Penguin. Crews, F. 1998b. “Unconscious deeps and empirical shallow”. Philosophy and Literature 22: 271–85. Eckardt, B. von. 1998. “Can intuitive proof suffice?” In Crew (1998a): 106–115. Eysenck, H. J. 1985. Decline and Fall of the Freudian Empire. New York: Viking. Farrell, J. 1996. Freud’s Paranoid Quest: Psychoanalysis and Modern Suspicion. Albany: NYUP. Freud, S. 1908/09. “Familienroman der Neurotiker”. In S. F. Studienausgabe Bd. IV
(1982): 221–6. Frankfurt am Main: Fischer. Freud, S. 1913. “Märchenstoffe in Träumen”. In Laiblin (1995): 49–55. Freud, S. 1947. “Der Dichter und das Phantasieren”. In S. F. Gesammelte Werke, London: 213–23. Freud, S. 1908. “Die Traumdeutung”. In S. F. Gesammelte Werke Bd. II und III (1976, 6th ed.) Frankfurt am Main: Fischer. Fricke, H. 1991. Literatur und Literaturwissenschaft. Beiträge zu Grundfragen einer verunsicherten Disziplin. Paderborn: Ferdinand Schöningh. Gellner, E. 1996. The Psychoanalytic Movement. The Cunning of Unreason. Evanston: Northwestern UP. Grimm, Brüder 1954. Kinder- und Hausmärchen. Lahr: Schauenburg. Grünbaum, A. 1987. Psychoanalyse in wissenschaftstheoretischer Sicht. Zum Werk Sigmund Freud und seiner Rezeption. Konstanz: Universitätsverlag. Grünbaum, A. 1993. Validation in the Clinical Theory of Psychoanalysis: A Study in the Philosophy of Psycholanalysis. Madison, CT: International Universities Press. Hohler, F. 1982. 47 und eine Wegwerfgeschichte. 5th ed. Bern: Zytglogge. Laiblin, W. (ed.). 1995. Märchenforschung und Tiefenpsychologie. 5th ed. Darmstadt: Wissenschaftliche Buchgesellschaft. Lenz, F. 1971. Bildsprache der Märchen. Stuttgart: Urachhaus. Lüthi, M. 1961. Volksmärchen und Volkssage. Bern, München: Francke. MacMillan, M. 1997. Freud Evaluated: The Completed Arc. 2nd ed. Cambridge, MA: MIT Press. Mallet, C.-H. 1982. Das Einhorn bin ich. Hamburg: Hoffmann und Campe. Mann, T. 1936. Freud und die Zukunft. In S. Freud, Abriß der Psychoanalyse, 131–151. Frankfurt am Main: Fischer. Rank, O. 1974. Das Inzest-Motiv in Dichtung und Sage. Darmstadt: Wissenschaftliche Buchgesellschaft. Scharnberg, M. 1993. The Non-authentic Nature of Freud’s Observations. 2 vols. Stockholm: Almquist & Wiskell International. Schönau, W. 1991. Einführung in die psychoanalytische Literaturwissenschaft. Stuttgart: Metzler. Spörk, I. 1985. Studien zu ausgewählten Märchen der Brüder Grimm. Königstein/Ts.: Hain. Van Peer, W. 2001. “Where do literary themes come from?” In Interdisciplinary Studies in Thematics, M. Louwerse and W. van Peer (eds). Amsterdam/Philadelphia: Benjamins (in press). Webster, R. 1995. Why Freud Was Wrong: Sin, Science and Psychoanalysis. London: Fontana Press.



Back from the grave Reinstating the catharsis concept in the psychology of reception Brigitte Scheele University of Cologne

While in literary studies the controversy still rages as to what Aristotle actually meant by the term catharsis, present-day psychology is notable for a wide variety of very different catharsis concepts. As one would expect, the findings and evaluations based on these conceptions are correspondingly wide-ranging, in some cases downright contradictory. One of the factors operative in generating such a confusing array of approaches is the fact that among the psychological disciplines concerning themselves with the term some are located in various fields of applied psychology while others are engaged in basic research. Alongside application-oriented catharsis research across the whole gamut of therapy approaches to be found in clinical psychology and also in the more recent media psychology, we also find an engagement with the idea of catharsis in the domains of basic research, notably in experimentally oriented social psychology and most of all in those branches of general psychology concerned with “learning”, “emotion” and “motivation”. Thus different research interests, different philosophies, different survey methods and evaluation models have produced catharsis conceptions displaying both close similarities and wide divergences which cannot be simply traced back to similarities and differences of approach. Catharsis variants have been developed and empirically tested not only in the psychology of reception but also in action psychology. With regard to the former — so-called “symbolic” catharsis — the verdict after six decades of experimentally oriented basic research is unwontedly unanimous: the catharsis hypothesis is dead and buried. From a point of view colored (not least) by an awareness of the catharsis discussion in literary studies, one of the things we need to bear in mind when regarding such empirical and experimental unanimity is that “catharsis” is quite definitely a value-concept. Accordingly, the conclusion that the catharsis concept has been refuted represents a (negative) prescriptive statement, though this is

202 Brigitte Scheele

something that the empirical/experimental research has failed to even see, let alone concede. The reason lies in the so-called value-neutrality postulate advanced most notably in the empirical (social) sciences by the Critical Rationalists (see Albert 1968), with recourse to the ideas of Max Weber. The metatheoretical norm thus postulated states that value judgements on the object-theoretical level are invalid because they cannot be derived descriptively and are therefore devoid of any grounding in empirical fact. This is a view that has largely established itself in the ‘scientistic’ approach to scholarly and scientific endeavor. Thus if we take the scientistic approach to psychology as an example we find that value-concepts are only studied with reference to their empirical implications or as aspects of the psychological object sphere (for instance frequency and orientation of value espoused by individuals, groups, strata of society, etc.). Value judgements are then only permissible in the form of practical (for instance political) decisions not regarded as requiring any scientific substantiation. (Though the discussion on the effects of the media is in fact conspicuous for the citation of scientific findings — for instance with reference to the impact made by scenes of violence — but according to scientistic criteria this does not in itself represent a sufficient grounding for statements on values.) The crucial problem resulting from the principle of value-neutrality is that negating the normative implications of psychological concepts or relegating them to the nonscientific domain of ‘practice’ in no way resolves the value dimensions themselves. In empirical/experimental research they are instead covertly (cryptonormatively) incorporated and that usually means modified. But due to the claims of value-neutrality this modification then depends solely on metatheoretical (in this case methodological) standards such as the demand for experimental precision and validity, thus leading to a distortion, not to say perversion of the value-concept in question. I shall be arguing that precisely this kind of distortion has taken place in the empirical testing of the catharsis concept. It is a distortion which is both ethically destructive and methodologically unnecessary, the reason being that in connection with value-oriented concepts the value-neutrality principle creates a vacuum providing almost unlimited scope for irrational dynamics, although approaches have been developed for the (at least partial, including empirical) testing and legitimization of statements on values (see Groeben 1999: 370ff.; Groeben and Scheele 1977: 122ff.; Scheele 1996b: 360). The aim of this article is to demonstrate the constructive integration of descriptive research results with metatheoretical and object-theoretical statements on values, taking catharsis as an example. The point is that if we explicate the two classical conceptions of catharsis (‘purging’ vs. ‘purification’) in historical and systematic terms (section 1), the purification perspective is certainly the ethically superior of the two. But due to the neglect of the value-concept angle and the concentration on purely methodological standards, the empirical/experimental approach has been limited to the purging

Back from the grave 203

concept, firstly with reference to the action perspective (section 2), and then with reference to the reception perspective. In both cases the experimental approaches have led to the ‘refutation’ of the catharsis concept mentioned earlier (albeit only the purging side of it; section 3). After a discussion of this reductionist development in empirical/experimental research (at least from a value-concept vantage) I shall demonstrate that it is indeed possible to adhere to the value-concept of catharsis as purification by reinterpreting existing empirical test results and showing that there are other, non-experimental findings favoring the elaboration of a model of cathartic purification from a reception-oriented perspective (section 4). Finally I shall outline the kind of research needed for the elaboration of such a perspective (section 5).

1.

History of the catharsis concept between ‘purging’ and ‘purification’

In Aristotle’s poetics, catharsis refers to the purifying effect of tragedy. With this originally medical term Aristotle posited that certain features of tragedy bring psychic (mental) and above all emotional relief for the audience. He was not however explicit about what features have to be present for the effect to be specifically cathartic, in contradistinction to other factors with a relieving or discharging impact. This has given rise to an ongoing controversy about the definition of catharsis and in consequence about the way the cathartic process is to be modeled, what (tragic) aspects lead to what reception processes, whether or not certain specific preconditions have to be fulfilled by the spectators before certain receptive processes can be activated, and so on (see Flashar 1984; Luserke 1991). For reconstruction purposes, the plethora of hermeneutic interpretations of catharsis which have emerged in the course of this debate can be arranged along a bipolar continuum ranging from ‘discharge’/‘purging’ at one end of the scale to ‘purification’ at the other. The purification position posits catharsis as an ‘ethical principle’, its main effect being “to transform passions into virtuous accomplishments” (Lessing 1768: 78. St.) rather than to discharge excess emotion as in the purging concept (see below). The goal of purification is achieved via vicarious (com)passion and the effect manifests itself in a heightened moral awareness in the individual spectator over and against his/her moral condition prior to the performance witnessed (Lessing after Gründer 1968/1991:357). Unfortunately, this definition of the process and effect of purification is much too implicit to serve as a basis for empirical testing. To achieve the necessary degree of explicitness we thus require greater theoretical differentiation and concretisation of the effects and processes involved on the individual plane. Especially appropriate to this objective are the theoretical approaches and empirical findings of the research on motivation and emotion conducted in the framework of cognitive psychology (Scheele 1990, 1996a,
b) and also the study of (moral) conflict

204 Brigitte Scheele

processing in psychotherapy research (Bohart 1980; Nichols and Zax 1977). In terms of these approaches the purification concept needs to be elaborated along the following lines (Scheele 1999): the individual experiences the purifying transformation as progress in his/her emotional and motivational development. Central to this feeling of positive development is the vital need for (stronger future) realization of his/her own ‘higher’ (in the sense of altruistic) abilities and potential (1999: 20). This progress can be objectively postulated as an (increasing) congruence between emotion and cognition, that is as emotional and cognitive integration. At the emotional level it manifests itself as the subjective feeling of stronger ego identity; on the motivational plane it takes the form of hope for success in constructively coming to terms with certain conflicts (concerning the self or the outside world) and omissions, transgressions or inadequacies in connection with self-relevant maxims; at the cognitive level it takes the form of a clearer view of the problem(s) than hitherto. However, this reconciliation between cognition and emotion arises neither from a simple discharge of emotion (as the ‘purging’ theory asserts) nor from a relatively unemotional form of insight (as opponents of the purification theory impute to it). The much more plausible postulation is that the integration of emotion and cognition is achieved via a “dual, cognitive-affective process” (Bohart 1980: 192–3; Nichols and Efran 1985). Depending on the extent of cognitive change, this process can be assumed to represent anything from simple amplifications to instances of qualitative restructuring (Scheele 1999: 26ff.). Advocates of the purging effect see catharsis not as an “ethical” but as a “hedonistic principle” (Gründer 1968/1991; Schadewaldt 1955). This is tantamount to postulating a reception process in which the pleasurable vicarious experience of tragic events leads to a ‘discharge’ or ‘conduction’ of individually stressful emotions and to an associated form of heightened arousal. Like purgation in medicine, the psychic discharge process is largely assumed to take place involuntarily, automatically, mechanistically. The emotional relief thus achieved manifests itself as cathartic effect, that is in a subjective feeling of restoration to well-being after liberation from tensions experienced as adverse. Thus in Bernays’ purificational interpretation (1857/58) tragedy cannot logically be expected to improve the spectator but only “to heal for a short while, to calm by means of artificial arousal, to relieve apprehension” (quoted after Gründer 1968/1991: 372–3). Expressed in terms of motivational energetics we might rephrase Bernays’ words as follows: there is a subjectively unpleasant “accumulation of tension” in the individual which is subsequently resolved via “discharge” of that tension (see Heckhausen 1989: 336–7). The reduction of unpleasant tension takes place via direct acting-out of that same tension and the effectiveness will be proportional to the degree to which the actingout corresponds to the tension or emotion to be discharged. Structurally, this healing-by-catharsis process is characterized by the congruence between emotion and behavior — that is emotion-behavior integration — with as complete an

Back from the grave 205

exclusion of cognitive control as possible (Scheele 1999: 26). The agreement between emotion and behavior manifests itself in a direct orientation to experienceof-emotion acted out quasi-automatically in accordance with genetically more or less preordained action patterns of a physiological and motoric nature (see Scheff 1983; Scheff 1981; Scheff and Bushnell 1984: 256ff.). In the following table (Table 1) the features central to the purging and purification models are listed side by side for purposes of quick and easy comparison. Table 1.Purging vs. purification: Central features Purging

Purification Cathartic process

Integration of emotion and behaviour via Integration of emotion and cognition via abreaction/acting out restructuring/transformation – experience oriented (absorbing a-reflex- – processing oriented (involved reflexivity ivity) – direct resolution of tension – delayed resolution of tension – intensity manifests itself somatically – extensity manifests itself verbally Cathartic effect – –

immediate relaxation of tension – mediate healing i.e. restoration of emo- – tional well-being (homeostatic model)

mediate relaxation of tension immediate healing i.e. strengthening of psychosocial motivation (developmental model)

Thus we see that in theoretical terms the significant differences between the two positions lie primarily in the assumptions about the relations between emotion/ motivation, cognition and behavior characterizing the respective ‘before and after’ states and of course the corresponding change processes. The relational differences outlined for the two diametrically opposed models will suffice to show that we are in the presence here not of rival conceptions in the proper sense of the term but of two distinctive models for two genuinely distinctive psychic phenomena. This means that we must examine each of the conceptions separately. Correlated to the two diametrically opposed hermeneutic interpretations of catharsis there then emerge two hypothetical biases with regard to the reception of fictional material: the assumption that tragedy has an ethical effect involves the postulation of a positive change notably in psychosocial motivation coming about via vicarious (com)passion (purification theory); the assumption of a hedonistic impact, by contrast, asserts the largely automatic discharge of a disturbing excess of emotion via vicarious, pleasurable experience (purging theory).

206 Brigitte Scheele

2.

Focusing on action: the limitation of empirical research on the purging view

As already mentioned, the reception-related theory outlined above is not the only psychological concept to have been elaborated on catharsis. There is also an actionrelated perspective on the problem. The difference is that action-related theories focus on real action with a cathartic effect for the acting individual, not on vicarious experiences of material presented in fictional form (possibly involving empathy and compassion). For this reason the action-related perspective is not (or no longer) of immediate relevance for theory formation in the domain of reception psychology. But as it has had a decisive impact on the development of theory over the past 60 years I shall briefly outline its cardinal tenets as this will aid in understanding the distinctive theoretical position of the psychology of reception in its present state. It is well known that Breuer and Freud (1895) were the first to adapt the Aristotelian catharsis concept for psychoanalytic purposes, subsequently introducing the ‘cathartic method’ derived from it. This therapeutic approach to the healing of hysterical symptoms (initially) involved hypnosis and in its essentials probably took its bearings from Bernays’ (1857) reconstruction of catharsis as a purging phenomenon (Langholf 1990; Leuzinger 1997). This (initial) proximity to the idea of catharsis as purging is especially apparent in the automatism assumed (in line with reception-related catharsis) to underlie the healing process. The authors explained the impact of the cathartic method as that of performing a stimulating arousing and discharging function for the suppressed emotion and bringing about the “liberation of the affect that had gone astray and […] its discharge along a normal path (abreaction)” (Freud 1926/1981: 264), notably by virtue of “allowing its strangulated affect to find a way out through speech” (Breuer and Freud 1895/1991: 17). The medical analogies in the description of the underlying psychic mechanism (Leuzinger 1997: 186) also suggest conceptual concurrence with organismically based assumptions about the purging effect. The theoretically significant difference over and against the reception-related catharsis perspective is that the psychoanalytic cathartic method does without tragedy, that is without the reception of an extraneous, supra-individual text. In terms of psychoanalytic therapy this renunciation is plausible as it has no place for considerations bound up with aesthetic objects (music, literature, and so on) as such but only (at best) inasfar as they have a functional relevance as manifestations of the individual complex-structure of the recipients and their subjective biographies. This places the significance and the weight of whatever it is that is to be ‘purified’ by catharsis so squarely inside the subject that the resultant coping process no longer displays any structural differences from other actions aimed at resolving individual problems (Leuzinger 1997: 186–7). This was the vantage from which in the 1930s empirical-cum-behaviourist research began working on the

Back from the grave 207

catharsis concept via operationalisation of psychological (hypo-)theses. The primemotor in this was the so-called Yale Group (Dollard et al. 1939). In the framework of their research on aggression Dollard et al. formulated their wellknown frustration-aggression hypothesis (“aggression is always a consequence of frustration” and vice versa “frustration invariably leads to aggression”) and specified factors influencing the strength of aggression and the inhibition of aggression. Finally they postulated a reduction of aggression motivation following the successful performance of aggression and, with explicit reference to psychoanalysis, designated this phenomenon as ‘catharsis’. This designation becomes (more) understandable if we remember that methodologically the authors were also attempting to recruit psychoanalysis for the behaviourist mainstream by establishing an empiricist foundation for psychoanalytic object theories. Given the constraints of this methodological and behavioral orientation they saw themselves in the position of having to explicate psychic phenomena (including psychoanalytic concepts like ‘catharsis’) in terms of behavioral theory. In the course of this process ‘overt (aggressive) behavior’ advanced to the status of the central explicative unit for ‘catharsis’ (or cathartic effect), relegating all other possibilities for achieving (immediate) deactivation of aggression to the status of deviant forms which, while by no means ruled out of account, had little ultimate conceptual significance for the behaviourist oriented psychological approach to the phenomenon in question. Accordingly, research on aggression distinguished in the last resort only two species of aggression for which a cathartic effect was posited: ‘direct’ aggression (the keynote form for the overall conception), meaning behavior actually performed with the intent to harm, and, in explicit contradistinction to it, ‘vicarious’ aggression, meaning all other behavior with an aggression-conductive cathartic effect, such as “physically tiring work, aggressive humor or simply participatory observation of aggressive acts” (Schneider and Schmalt 1981: 162). With this behavioralisation effected by the Yale Group, it became considerably easier (compared with drive-theoretic approaches in psychoanalysis and ethology) to subject the concept of catharsis to empirical testing. This in its turn explains why the change of meaning from a reception-related to an action-related catharsis concept within academic psychology was able to take place with such speed and thoroughness that introductions to the subject referred to this behaviorally ‘transmogrified’ conception of catharsis as being the ‘classical’ one (harking back to Freud and beyond him to Aristotle), that is “that all acts of aggression — even those that are covert, indirect, and noninjurious — serve as a form of catharsis, lowering the instigation to further aggression” (Baron and Richardson 1994: 24; see also Bierhoff and Wagner 1998; Heckhausen 1989; Herkner 1991; Kornadt and Zumkley 1992; Schneider and Schmalt 1994; Selg et al. 1997). The association of catharsis with this highly successful form of aggression research (according to Dollard et al.) resulted however in a further restriction, the

208 Brigitte Scheele

reduction of catharsis to the field of so-called hostile aggression. This restriction manifested itself not least in the all but total absence of explicit reference to (let alone study of) catharsis in the basic research on other conceivably relevant reception-psychological areas of emotional and motivational (di)stress, such as disappointment, guilt, shame, anxiety, cowardice, humiliation, vanity, jealousy, envy, and, resentment. All in all, then, we can say that within basic psychological research a reductionist shift of meaning established itself. It deviated appreciably from the original reception-psychological conceptions ranging from ‘purification’ to ‘purging’ and referring to a broad spectrum of stressful emotions. Instead, catharsis was restricted to the reduction of hostile aggression motivation and (as a purging concept) explicated exclusively in relation to action (for more detail see Scheele 1999: 36ff.). For reception-related theories this means that the purging end of the scale continued to figure in the models of catharsis proposed, while the purification end of the scale continued to be ignored. In the framework of the aggression model, the action-related concept of catharsis was elaborated more and more exclusively in terms of action theory until catharsis ultimately stood defined as a deactivation effect following a successfully performed act of aggression (Kornadt and Zumkley 1992; Zumkley 1978). The corresponding experimental design differs from the reception-oriented paradigm in that the subjects — who have previously undergone frustration — are not shown films (centering on the topic of aggression: see section 3) but are given the opportunity to “perform aggressive acts” (Heckhausen 1989). The results obtained so far suggest (amongst other things) firstly, that the deactivation of aggression motivation only takes place to the extent that the goal of aggression represented in it is achieved in a subjectively satisfactory manner (Peper 1981; Zumkley 1978), and secondly that the unspecific acting-out (say motorically, as opposed to a form targeted at reprisal) as propagated by popularized drive theory as a general recipe for ‘purging’ is not able to deactivate hostile aggression motivation (see Geen and Quanty 1977; Peper 1981). But this is tantamount to forfeiting not only the social but also the emancipatory component of the traditional catharsis concept as advanced in reception psychology. In fact, I would go so far as to say that the action-oriented approach has positively antimonial implications, to the extent that it makes catharsis dependent on the motivational end-state of consciously intended and individually responsible, concretely performed and ultimately successful acts of revenge and retribution. Thus catharsis, a positively connoted term in reception psychology, is explicitly coupled to negatively connoted antimonial conflict enactments. This contributes to a positive evaluation of such enactments and to some extent makes them possible in the first place (see also the criticism of popularized conceptions of acting-out of suppressed aggression motivation: for instance Berkowitz 1993). This necessarily involves an anti-emancipatory tendency because the action-related model uses the catharsis concept to propose a ‘(re)solution’ for

Back from the grave 209

individual psychic stress. A case in point is the form in which it is reconstructed in the Kohlberg model of moral development, where it does not go beyond the conventional level (see Kohlberg 1995). In this way possibilities for more humane forms of coping with reprisal motivation are phased out, or at least not explicitly included in the model.

3.

The parallel reception perspective: symbolic catharsis as purging ‘discharge’ (purgation)?

Parallel to this largely action-related perspective on catharsis, research in the psychology of reception has also gravitated to the purging end of the continuum. The classical experimental design as it has been used in testing hypotheses on catharsis within the psychology of reception goes back to Feshbach (1961). It tests for “stimulating versus cathartic effects of a vicarious aggressive activity” and accordingly lays down the following procedure for the group of subjects: first the subjects are annoyed, then they are shown a short film with aggressive content (in Feshbach’s 1961 experiment it was a 10-minute excerpt from the boxing match in the film Body and Soul) and finally the aggression motivation is measured. Via a corresponding control-group design, Feshbach (1961) came to the conclusion that after watching a film with aggressive subject matter annoyed persons display less marked aggressive tendencies than after the reception of a film with neutral material. This result appeared to corroborate the assumption (of the Yale Group) that direct aggression motivation arising from frustration can also be abreacted via symbolic catharsis. Much criticism was leveled both at the experimental set-up of this study and at subsequent studies of a similar kind. The following features were felt to be inadequate: the absence of a treatment check in connection with the experimentally induced annoyance of the subjects; the inappropriate choice of film material (too short, not explicitly aggressive enough for the test group, too boring for the control group or for both groups, and so on); inadequately validated aggression ratings; lack of a baseline for the dependent variable ‘aggression motivation’, and so on (see for instance Bergler and Six 1979; Heckhausen 1989). Tests rectifying these points then led to findings largely (though not exclusively, see below) militating against the assumption of “vicarious aggression catharsis” (Lukesch and Schauf 1990). Yet the theoretical and empirical state of play remains oddly diffuse and inchoate. One indicator of this is the consummate ease with which contradictory effects can be predicted for one and the same constellation of conditions. Attempts to obviate such contradictions then culminate in predictions about the effect of aggressive media content which are remarkably nondescript in nature, such as: “There is a risk that what has been seen may have an effect” (Freiwillige Selbst-

210 Brigitte Scheele

kontrolle Fernsehen e.V. 1995). This predictive inadequacy is certainly traceable for a large part to the theoretical indeterminacy of the notions of ‘purging’ underlying the whole undertaking. In cognitive psychological terms they were inadequate to the task of explicating such things as specific conditions for the occurrence of deactivation of motivation (as ‘catharsis’ or ‘inhibition’) for fear of potentially negative consequences of aggressive behavior (such as punishment, reprisals, revenge), or activation of motivation in the sense of ‘learning’ or ‘stimulation’. Once the ‘purging’ approach had been forsaken the spin-off effect was a proliferation of innumerable mini-theories which were equally unsatisfactory in that they merely represented a plethora of theoretically unproductive differences or similarities of interpretation (see Geen and Thomas 1986). Only then did the integration phase supervene with models attempting to achieve a unifying reconstruction of the findings, some of them drawing on ontogenetic (developmental) aspects for the purpose. Among these are the ‘cognitive-neoassociationist’ approach taken by Berkowitz (1984; Jo and Berkowitz 1994; similarly Geen and Thomas 1986) or the attempt by Rule and Ferguson (1986) to combine observational learning (according to Bandura) with script and schema approaches taken from the psychology of information processing. With a view to reinstating the psychically richer and more suggestive purification model, I intend here to draw on the concept of ‘psychological modeling’, (‘observational learning’, ‘learning by imitation’, ‘learning by vicarious experience’) for the structuring of impact and effect (see the reception psychological application to reading, especially of fictional texts: Groeben and Vorderer 1988; for the reception of cinema films: Vorderer 1992). In this approach any individual capable of symbolic coding can learn by observation alone. Thus, in contrast to the assertions of conditioning theory, he/she is not completely and utterly dependent on experience gained by first-hand activity in order to change through learning but can also achieve such change by means of interpretive observation of behaviors displayed by so-called models together with the attendant situative contexts and the resultant consequences (see for instance Bandura 1971, 1976, 1986, 1989, 1994). According to Bandura this learning by vicarious experience can be applied to almost any kind of content, from simple motoric coordination, language acquisition and emotionality all the way up to moral rules. Modeling influences can “serve diverse functions — as tutors, inhibitors, disinhibitors, social prompters, emotion arousers, and shapers of values and conceptions of reality” (Bandura 1994: 77). Thus ‘observational learning’ (in Bandura’s sense) functions as a theoretical framework which (with the help of specifying assumptions) should be suitable for achieving a contradiction-free explication not only of anti-catharsis theories but also of catharsis conceptions (both the purging and the purification varieties; see below). The assumptions on observational learning have been confirmed on numerous occasions, not least by findings on the acquisition of aggressive behaviors with

Back from the grave

reference to models taken both from real life and from fiction (see Bandura 1971, 1986). Thus in the face of the proven reception of aggressive and violent material broadcast by the media (see Gleich 1995), the course that suggests itself is to pit an anti-catharsis theory against the purging theory and produce empirical evidence to substantiate warnings about the effects of learning by models being anti-cathartic (Bandura 1979; Berkowitz 1993; Feshbach 1989; Goranson 1970; Huesmann and Malamuth 1986). If we take the findings from the experimental paradigm on ‘learning aggression by observation’ and the results on catharsis as ‘purging’ and allocate them to the major effect categories (of actual genesis) ‘learning’, ‘stimulation’, and ‘inhibition’, we first of all obtain the following picture (see also summaries of results in Baron and Richardson 1994; Bergler and Six 1979; Berkowitz 1984, 1993; Eron 1986; Geen and Quanty 1977; Geen and Thomas 1986; Kunczik 1982): 1. In terms of learning effects (tested above all for children) it transpires that aggressive behaviors are (or can be) learned by observation and apparently retained for a long period of time even without over learning; this effect is facilitated if the aggressive behavior is modeled as being conducive to goal achievement (i.e. as successful) and if similarities exist between the observed setting and the real setting in which the observer happens to be. 2. As for stimulation or disinhibition it proves to be largely the case that vicariously experienced violence increases rather than reduces the individual aggression potential on the part of observers. 3. As for the inhibitory effect there is good evidence for a reduction of aggressive actions as a result of ‘vicarious’ punishment. Vicarious punishment means that the onlookers observe that the model’s behavior is punished (Bandura 1971: 6) and subsequently display appreciably less aggressive behavior than before. In this there is a more or less explicit assumption that inhibition of aggression is a motivation structure which is independent of a proneness/propensity for aggression (Kornadt 1982); this implies that reduction of aggression in the wake of an increase of aggression inhibition should not — indeed must not — be conceived of as positive evidence of actual discharge (purging). If we additionally take into account the respective specific (antecedent) conditions for stimulation (as opposed to purging) effects of a cathartic nature, the results allow the following postulation of ideal types: there will be an increase of hostile aggression-proneness on the part of the observers if they experience and process the aggression presented to them (vicariously from the perpetrator perspective) hedonistically as positive and/or pleasurable (“in a good cause” or “for its own sake”, see Kunczik 1982: 6), that is as a morally justified (in whatever way) and also usually successful means to an end (see for instance Bandura 1989; Geen and Thomas 1986; Goranson 1970; Huesmann 1986; Rule and Ferguson 1986). But aggression on the part of recipients is patently also augmented when they experience the aggression

211

212 Brigitte Scheele

presented to them (vicariously from the victim’s perspective) as unjustified and hence develop feelings of, and motivation for, reprisal (see Hartmann 1969). By contrast, there is an increase in aggression inhibition not only when the vicarious experiences generate fears of potential punishment, revenge, counter-aggression (see for instance Bandura 1965; Berkowitz 1984; Selg 1992) but also on the basis of so-called ‘moral’ feelings such as guilt, shame, and indignation, (see for instance Bandura 1994; Berkowitz 1984; Kornadt 1982). But what this ultimately means for the question of reduction of potential aggressiveness as a purging effect is that so far no evidence has actually been adduced that there is a purging effect at all. Consequently, for the time being at least, the hypothesis of hedonistic purging (through pleasurable vicarious involvement) may be regarded as falsified for ‘hostile aggression’. Accordingly the assertion that the catharsis hypothesis has been ‘finally refuted’ needs to be formulated with greater precision. We are now in a position to achieve such precision on the basis of the features of the concept explicated thus far: the conclusion must be that the assertion can in fact only be upheld in connection with catharsis as the discharge (not purification) of aggressive motivation via vicarious experiences. In terms of purification (as an ethical principle), all that can be said is that it has not (at least explicitly) been worked out in the theoretical framework of observational learning. But for initial indications of the potential empirical usefulness of the purification theory we can draw on contradictory findings on aggression inhibition and also on those findings generated by a concept of ‘inhibition’ displaying a degree of ‘moral’ enrichment. Finally we shall inquire what aspects of the traditional catharsis conceptions are to be retained for a more comprehensive explication of the construct and the generation of new hypotheses in the future.

4. On the reinstatement of catharsis as purification Of course the ‘empirical coup de grâce’ for the purging concept and its replacement by evidence on learning, stimulation of aggression motivation and reduction of aggression inhibition due to vicarious experiences does not necessarily imply that hostile aggression-proneness cannot (within limits) be reduced and/or transformed into pro-social (altruistic) orientation and motivation by processing that has been stimulated externally. The decisive requirements postulated here for a theory of purification is socially oriented restructuring and ultimately a correspondingly selfcritical overcoming of specific reprisal leanings or diffuse destructive urges. Such a qualitative change of (say) feelings of guilt instilled indirectly by external media may be expected to lead to a reduction of aggression motivation (albeit not automatically), given that (unlike instrumental aggression) at least the needs to act in a hostile manner, which is geared primarily to destruction, injury, revenge and the like,

Back from the grave

should by the same token have been appreciably reduced (for distinctions of the various aims pursued by aggression see Berkowitz 1993; Eckensberger and Emminghaus 1982; Feshbach 1964; Kornadt 1982). Indications of such morally motivated processes emerge from studies looking (wholly or partially) at the influence of (moral) information (see Rule and Ferguson 1986). There has for example been repeated evidence that test-persons subjected to provocation react with greater hostility towards those provoking them after the reception of ‘justified’ aggression than those who had watched material containing ‘unjustified’ aggression (see Berkowitz 1984). These findings were not only replicated under lab experiment conditions with students but also in field experiments with young male detainees (Parke et al. 1977). In a comparable way it has been established that the aggressive intention presented can have an influence on aggressionproneness. Subjects confronted with aggression as ‘reprisal’ were in the aftermath appreciably more aggressive than those shown aggression as a pro-social act (see Rule and Ferguson 1986: 36). Now justice/justification is neither itself a one-dimensional concept nor is it of course the only ethical principle susceptible of influence by the (production and) reception of subject matter conveyed by media. Other possible factors with a conceivable influence on the processing of aggressive material could be a caring orientation, compassion, and so on in differing depths, degrees and syntheses. Thus any explication of purification must proceed on the assumption that the subject matter conveyed medially will be worked over and processed on a constructive cognitive plane. What (and more especially what lasting) features are necessary on the part of the recipients cannot be stated on the basis of the findings drawn upon here. But it does seem fair to assume that such qualitative change will at least become more probable as moral development progresses (Eckensberger and Emminghaus 1982; Herzog 1991). Yet from a motivation theoretical viewpoint a comparison of the findings on stimulating (and inhibiting) effects can certainly be drawn upon to derive the postulation of some specifying (antecedent) conditions. Naturally these will be a great deal more well-defined for theoretically less complex constructs (like ‘stimulation’) than for the much more complex purification construct. Deactivation of aggression motivation occurs when spectators experience aggression (primarily from the perpetrator perspective) as aesthetically displeasing and largely unproductive (or actively counterproductive) and when they feel the aggression to be morally unjustified and/or have compassion with the victim (unlike the stimulation condition referred to above, this is the processual fore-stage for the emotional-cum-cognitive ascription of blame to the aggressor for the sufferings of the victim). As far as I know, the only study proceeding from similar assumptions about activation/deactivation and also subjecting them to empirical testing with an impressive degree of ecological validity is that by Bönke (1989). In the framework

213

214 Brigitte Scheele

of his university diploma thesis he tested three theater plays for differences of effect: Aeschylus’ Oresteia (the complete trilogy performed in one evening and comprising the plays Agamemnon, The Libation Bearers and The Eumenides), Franz Xaver Kroetz’ Der Nusser (The Nutter), and Klaus Pohl’s Hunsrück. The theoretical basis is the reception-psychological contrast between Aeschylus’ conception of catharsis on the one hand and two forms of peripeteia (in Aristotle’s sense of the term) on the other. Though all three dramatic forms present human suffering, guilt, aggression, and the like in a way inspiring fear and terror, they differ in a significant way from the perspective of reception psychology. The difference lies in the approach to coping with the central problem either thematised in the plays or suggested by them. As Bönke makes clear, the Oresteia proposes constructive problem-resolution, in the sense of a reconciliation of conflicting needs, structures, and so on Implicit in this resolution is a view of humanity built around the assumption that, in theory at least, there is such a thing as moral development and manifesting itself in the propagation of non-violence in the service of the integration of opposites. The two other forms, by contrast, categorically rule out (re)solution via human agency of the human misery portrayed. In the one peripeteia variant the theme is the individual problem of ‘blameless guilt’, in the other it is the supra-individual problem of an external force supervening without warning, a force which individual and community alike are powerless to withstand. The prototypes of peripeteia are generally held to be Sophocles’ Oedipus Rex on the individual level and Euripides’ Bacchae on the supra-individual plane. In the theoretical framework of observational learning, specific spectator effects can be derived and postulated for each of the three different types of problem (see below). At the time of Bönke’s study only the Oresteia was actually on the programs of the theaters within reachable distance, so for the purposes of empirical testing Der Nusser was taken as a replacement for Oedipus Rex and Hunsrück for the Bacchae. In Der Nusser, Eugen Hinkemann (the ‘nutter’) returns from the war impotent, fails to find work and is forced by his financial straits to take a job at a fairground as a “geek” entertaining the audience for a pittance by biting the throats out of living mice and rats (Bönke 1989:139). His wife deceives him with his best friend and unable to stand the pangs of guilt hangs herself from the gutter of the house (1989:141). “Crouched by his wife’s body all that Eugen Hinkemann can do after his final hopeless words is to stare blankly, brutishly and despairingly into the audience.” In Hunsrück, a group of five persons — husband and wife (Werner and Linda), their friend Walter, their nephew Paul and Werner’s 93-year-old mother — alleviate their chronic boredom by humiliating one another in identically repeated, stylized language rituals. Paul has tormenting visions of violent killings and torture instruments with himself as perpetrator. Linda wishes she were elsewhere and dreams of traveling to distant parts, while the grandmother dwells on her visions of the past. After Paul leaves for the pub to get drunk, four strangers turn up, all of

Back from the grave

whom Werner and Walter brutally and gratuitously slaughter with an axe. After a brief silence the group resumes its activities and verbal exchanges precisely as before. The play ends with Paul returning from the pub and having one of the murdered strangers thrust into his arms (Bönke 1989: 141ff.). With regard to aggression-proneness and inhibition the following specific spectator effects were predicted: the portrayal of fear and terror with constructive problem-resolution (Oresteia) would lead directly to a reduction of aggressionproneness and an increase in morally conditioned aggression inhibition; the presentation of fear and terror on the individual plane without any kind of resolution (The Nutter) would cause a direct increase in aggression-proneness but no difference in aggression inhibition; the portrayal of fear and terror at the supraindividual level without any indication of possible problem resolution (Hunsrück) would lead directly to a reduction of aggression inhibition and possibly (for this there is as yet no object-theoretical derivation in the framework of psychological modeling) an increase in aggression-proneness. Aggression-proneness and inhibition were measured with an aggression TAT. This is a projective method which shows a sequence of standardized images and then asks subjects to tell a story based on them; subsequently the stories are separately encoded and evaluated for “aggression motive” and “aggression inhibition” by means of the TAT aggression rating scale (Kornadt 1982). In the Bönke study (1987/88) three experimental (theater treatment) groups each comprising 12 male and 12 female students of the University of Bochum were tested (a) 1 week before, (b) immediately before, (c) immediately after, and (d) 1 week after the performances. At each of the four time points four new TAT images were used to test for aggression-proneness and aggression inhibition and in addition a semantic differential to test for present state of mind. The treatment for the two student control groups was a visit to the zoo parallelized for time and effort; the treatment for the non-student control group for Hunsrück was simply their normal classes at the Unna Adult Education Centre (Bönke 1989). The outcome relativized via the respective baseline (here the average from the scores at time points 1 to 4) provide significant confirmation at the 5% and 1% level respectively of the impact assumptions postulated for aggression-proneness and aggression inhibition. As expected, the Oresteia spectators showed both a decrease in aggression-proneness and an increase in aggression inhibition. The group watching The Nutter showed an increase in aggression-proneness but no change in aggression inhibition. The group watching Hunsrück (again as predicted) displayed a reduction in aggression inhibition and an increase in aggression-proneness. The descriptive processing of the data for aggression inhibition shows that after the performance of the Oresteia the scores for ‘conscience and guilt anxiety’ were highest (M(male) = 40.8, M(female) = 57.2; in the control group: M(men) = 16.3, M(female) = 10.5), whereas ‘fear of punishment’ played no role (M(male) = 3.2,

215

216 Brigitte Scheele

M(female) = 3.2; in the control group M(male) = 4.2, M(female) = 7.4). These results support the purification position in that they confirm that representations of aggression do not in themselves lead to increased aggressionproneness and reduced aggression inhibition. As the purification theory suggests, it is rather that the portrayal of the possibility of overcoming (say) conflicts of need and the hostility resulting from them which is of cardinal significance for the reduction of aggression-proneness and the increase of morally conditioned aggression inhibition in particular. In addition, Bönke’s study shows that it is both plausible and necessary to achieve greater theoretical precision in connection with the stimulation conditions. Thus it will be essential not only to foreground the supporting preconditions for purifying cathartic processing but also to examine for other processing potentialities, including an inquiry into the nature of the resulting (intra psychic) consequences and their (moral) desirability.

5.

Research desiderata for the elaboration of a purification perspective for the psychology of reception

For an initial identification of those aspects requiring future elaboration with a view to reinstating the purification perspective on catharsis, we can now draw on the structural, processual and distinguishing features foregrounded earlier in this chapter (Scheele 1999). As we have seen, purification (unlike purging) is an emotional-cum-motivational (re)learning process on the basis of value-oriented confrontation and subsequent reorganization of the observer’s own present wishes and intentions towards a moral(ly superior) identity. In a before-and-after comparison, this effect manifests itself (a) as a more marked integration of cognition and emotion and (b) as a motivation for change sustained by ‘hope of success’ in modifying negative aspects (of oneself and the world) for the better. Fine-tuned in this way, the purification concept by no means implies a stabilization of things habitual and customary or a (re)adjustment to ‘old’ states but rather an emancipatory dynamic for change in connection with such ‘givens’ of the self and the world as are felt to require change. In terms of conceptual definition it is of no moment whatsoever how general or specific, profound or lasting such a dynamic for change may be in a given case. The prerequisite for the substantiation of such a concept of purification is a model of the human capacity for empathy and moral protest, a counterpart as it were to the model of hedonistic identification with aggression appropriately representing the foundation on which the stimulation theory builds. An intentionalist explication of individual value orientations is indispensable for this purpose and it should (among other things) provide a basis for the delimitation of aggression inhibition on grounds of fear of punishment and reprisal over and against purification

Back from the grave 217

in the sense of a (morally more mature) re-evaluation of aggression. However, as we have seen, aggression is not the only sphere of motivation and emotion for which catharsis as purification was originally postulated and for which it could, or should, claim validity today. Hence the objective of reinstating the purification idea for the purposes of reception psychology is bound up with another objective, that of removing this restriction in favor of an (initially) unlimited application of the idea to emotional/motivational learning in the sphere of (notably moral) character development. This in its turn implies at least two major sets of questions which are themselves interconnected. First, in connection with emotional learning, we must inquire after optimal processing forms for (morally) stressful emotional contents; and second we must inquire specifically from the vantage of the psychology of reception into the preconditions favoring such (purifying) forms of processing. A non-reductionist model of emotional learning thus postulated would then however require an emotion theoretical framework explicating emotions not primarily as psychophysiological or psychobiological (see the overview on recent approaches to emotion in Meyer and Reisenzein 1996) but above all as “intentional states constituted via combinations of convictions (or quasi-convictional states) and wishes (objectives or norms) with a commonality of content” (Boll 1998: 22). In other words, “an emotional state is present when a person (the emotional subject) is convinced that a state of affairs exists (will exist, might exist) fulfilling (‘positive emotions’) or not fulfilling (‘negative emotions’) the wishes, etc. of the emotional subject” (Boll 1998: 22). The point is that it is only with epistemological approaches identifying emotions essentially as subjectively relevant (conscious) evaluations that we can work out the relations between forms of cognition (convictions), nontemporary emotional attitudes (value orientations), present emotions (self-relevant evaluations), wishes and forms of a disposition to act (motivations) in such a way that the cognitively active forms of problem coping postulated for purification can be elaborated with the requisite degree of refinement (see Scheele 1990, 1996a,
b). To achieve this we must proceed from the realization that unlike the purging process (for which the integration of emotion and behavior has been reconstructed as the characterizing feature) the purification process centers above all on the integration of cognition and emotion (Scheele 1999). This means that in the purification process the individual (partly or wholly) suspends the psycho-hygienically dysfunctional suppression of certain antagonistic needs so that he/she can emotionally (re-)experience (‘feel’) the need-related perceptual and cognitive contents inherent in them. Findings from clinical psychology suggest that such an integration of cognition and emotion is achieved neither by way of cognitive processing devoid of emotional involvement nor (vice versa) by emotional experience devoid of cognitive processing. It appears that both are required if the processing of stressful content is to be successful. The stressful element has to be

218 Brigitte Scheele

experienced and understood as something of significance to the self (see Boesch 1976; Bohart 1980; Epstein 1984; Nichols and Efran 1985). If, in addition, the result of the restructuring is to remain stable as a learning effect it has to be emotionally acceptable. But a change in the domain of values relevant to the self-concept (in the case of ‘purification’) is only emotionally acceptable in the long term if it can be made to harmonize with the existing subjectively more ‘highly’ rated value concepts of the individual, that is if there are no self-relevant patterns either falling short of the exigencies of the ‘superordinate’ need and value structure or “overshooting the mark” (Scheele 1990). Where such an integration of emotion and cognition has been achieved we can say that a new and qualitatively different understanding of self and situation has developed on the basis of which the individual “feels better” in the sense intended by Lessing, that is “as if the whole psychic habitus of the person had achieved a lasting amelioration” (Gründer 1968; Schadewaldt 1955: 148). It seems safe to assume that this restructuring process in the value area of the self-concept is by no means rarely accompanied by ‘automatic’ emotional behavior of an expressive nature (such as weeping, aggressive gestures, laughing, and perspiring; see Scheff 1983, 1981; Scheff and Bushnell 1984: 256ff.) which itself has a purging effect. Hence expressive, motoric and autonomous nervous behavior may have a more or less influential (for instance intensifying) effect on the self-relevant restructuring process. But in agreement with an epistemological emotion theory postulating the behavioral components of emotional experience as margin-intensional emotion features (over and against reflective evaluation, which is regarded as core-intensional; Scheele 1996a: 288ff., 1990: 64–5), these behavioral aspects are not regarded here as a necessary, constitutive condition for the process of emotional learning as re-evaluation. Determining what ratio between the two modes of processing is necessary for a purifying form of restructuring to occur is certainly one of the cardinal issues to be addressed in the modeling of optimum processing forms of emotionally stressful contents. The question of the ratio in which the two modes need to be stimulated by the text material presented for reception (plays, films, and so on) in order to encourage rather than inhibit purification hinges on a number of preconditions which I shall now briefly outline. In order for purification to take place at all the individual need and value structure must necessarily have self-ideal components (without them emotional stress could not have been generated in the first place) to which the cathartic process can and must be geared (see Edelstein et al. 1993; Erb 1993). In other words “wanting the good (however it may be defined)” is “the necessary precondition for the cathartic effect; an Oedipus planning to kill his father would not be helped by catharsis: the contemplation of images of aggression would probably only confirm him in his intentions” (Boesch 1976:396). In order to identify the optimal preconditions for catharsis, research in the psychology of reception will probably in the medium-to-

Back from the grave 219

long term be dependent on adapting experience and models taken from the practice of clinical psychology. Central to this in my view will be the more precise definition of the conditions instituting the connection between self-ideal and acceptance of reality as elaborated in empirical research on the therapeutic process and on therapist and client variables (for instance on the basis of the interview therapy model). There is for example general agreement in calling for simultaneous establishment and support for two functions of the desired processing: first, the selfdistancing necessary for processing in conjunction with fundamental (self-)acceptance and security; second, the equally necessary confrontation with adverse contents and, not least, with self-critical reconstruction work (see Symonds 1954). Accordingly, Boesch’s recommendation is: the “notions to be worked on must not appear so immediately threatening as to disturb contemplation and reflection through evasion constraints”; additionally “the situation must provide processing aids, that is reflective self-perception must be facilitated just as much as processes of re-evaluation and reassessment of the individual’s own behavior” (1976: 408). For a reception-psychological purification concept this means first of all achieving a more precise definition of the features of ‘aesthetic distance’ (Bullough 1912) furthering processing on the part of recipients. Then it would be necessary to postulate and empirically test (media) conditions effecting a purifying form of identification with the contents perceived and map these onto a scale somewhere between distance and involvement (on the general relation between distance and involvement in the television medium: Vorderer 1992: 73ff.). Theoretically, the only plausible assumption is that as an antecedent condition for purification processes there has to be some minimal degree of distance not identical with the directness of (say) retributive action; at the same time there must be some minimal degree of emotional involvement, as without this there surely cannot be any possibility of relating the material received to the individual’s own needs, value perspectives, etc. (see Nichols and Efran 1985). For the description of reception preconditions we can draw on the concept of ‘distance’ proposed by Scheff (1981, 1983: 66) with reference to Bullough (1912). In Scheff’s usage, this term encompasses both conditions and outcomes identified in the course of applying his catharsis concept to dramatic structures and effects. Despite the methodological and theoretical weaknesses inherent in the ‘distancing’ concept (for instance potential circularity in connection with ‘distance’ and ‘catharsis’) it appears to me to represent an extremely useful heuristic framework for describing and demarcating a continuum extending from maximally favorable to maximally unfavorable confrontation and processing attitudes. The approach sets out from the postulation of ‘optimum distance’ as the maximally favorable precondition on the part of recipients (Scheff 1983: 67). This manifests itself in the experience of “profound emotional resonance” rooted in a feeling of emotional control.

220 Brigitte Scheele

Distinct from this on a “distancing continuum” implying fluid transitions are on the one hand states of “over-distancing” and on the other states of “under-distancing” (1983: 66ff.). In extreme cases of over-distancing no — or hardly any — attention is available for the respective emotional experience on the object level supplied by the theme. Scheff sees this as expressing itself in subjectively stressful situations above all in the form of “unfeelingness and/or confusion” (1983: 69); at the other extreme, under-distancing or absence/loss of distance involves no — or hardly any — attention at the meta-level, posited by Scheff as manifesting itself under subjectively stressful conditions in the form of relatively marked emotionalisation, physical reactions (general heightening of tone) and inconsequentiality in thought and speech (1983: 66ff.). Thus media features may be identified as favorable to purification if they stimulate (or at least permit, or at the very least do not inhibit) a change of the emotionally stressful initial situation towards the state of ‘optimum distance’, that is taking experiences (co-)effecting a reorganization of self and world states subjectively evaluated as negative nearer to a mental representation of ‘optimum distance’. It would have to be necessary, of course, to test empirically whether the application of the homeopathic principle according to which stressfulness can best be stimulated by the direct and explicit presentation of things stressful is in fact as universal as it is made out to be, or whether other genres (more distanced forms of presentation) might not be equally effective, if not more so (see Zillmann and Bryant 1994). In summary we can say that the following core hypotheses require theoretical elaboration and empirical testing: experiences are to be regarded as ‘purifying’ to the extent that (on the basis of aesthetic distance tending towards ‘optimum distance’) they achieve a relatively marked integration of emotion and cognition via the emotional (co-)experience of psychically stress-inducing (moral) conflicts in conjunction with value-oriented cognitive processing. At the same time this integration of emotion and cognition is bound up with ‘hope of success’ in achieving the (long-term) change of self and world towards more humane value structures. Thus defined, the purification concept quite clearly does not imply the stabilization of accustomed action patterns or the (re-)adjustment to ‘old’ states but rather a dynamic for change (objectively assessable as emancipation-oriented) in connection with givens of the self and the world subjectively evaluated as needing to be changed. In this individual and social value-orientation, the construct thus derived from the traditional reception-psychological understanding of catharsis is distinctive primarily vis-à-vis the purging and deactivation variants of catharsis foregrounded in previous empirical research. Making the purification perspective on catharsis newly accessible for empirical research would be not least an instance of the integration of the history of the catharsis concept in literary theory and empirical psychological modeling in a way that may be expected to enhance and enrich interdisciplinary cultural studies in the future.

Back from the grave 221

Note 1. For translation I am indebted to Andrew Jenkins, and to the support of Dr. Margrit Schreier.

References Albert, H. 1968. Traktat über kritische Vernunft. Tübingen: Mohr. Bandura, A. 1965. “Influence of models’ reinforcement contingencies on the acquisition of imitative responses”. Journal of Personality and Social Psychology 1: 589–595. Bandura, A. 1971. “Analysis of modeling processes”. In Psychological modeling: Conflicting theories, A. Bandura (ed.), 1–62. Chicago/New York: Aldine, Atherton. Bandura, A. 1976. “Verhaltenstheorie und die Modelle des Menschen”. In Lernen am Modell: Ansätze zu einer sozial-kognitiven Lerntheorie, A. Bandura et al. (eds), 205–229. Stuttgart: Klett. Bandura, A. 1979. Aggression. Stuttgart: Klett-Cotta. Bandura, A. 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, N. J.: Prentice Hall. Bandura, A. 1989. “Die sozial-kognitive Theorie der Massenkommunikation”. In Empirische Medienpsychologie, J. Groebel and P. Winterhoff-Spurk (eds), 7–32. München: Psychologie Verlags Union. Bandura, A. 1994. “Social cognitive theory of mass communication”. In Media Effects: Advances in Theory and Research, J. Bryant and D. Zillmann (eds), 61–90. Hillsdale, N. J.: Erlbaum. Baron, R. A. and Richardson, D. R. 1994. Human Aggression (2nd ed.). New York: Plenum. Bergler, R. and Six, U. 1979. Psychologie des Fernsehens. Wirkungsmodelle und Wirkungseffekte unter besonderer Berücksichtigung der Wirkung auf Kinder und Jugendliche. Bern etc.: Huber. Berkowitz, L. 1984. “Some effects of thoughts on anti- and prosocial influences of media events: A cognitive-neoassociation analysis”. Psychological Bulletin 95 (3): 410–427. Berkowitz, L. 1993. Aggression: Its Causes, Consequences and Control. New York etc.: McGraw-Hill. Bierhoff, H. W. and Wagner, U. 1998. “Aggression: Definition, Theorie und Themen”. In Aggression und Gewalt. Phänomene, Ursachen und Interventionen, H. W. Bierhoff and U. Wagner (eds), 2–25. Stuttgart etc.: Kohlhammer. Bönke, H. 1989. Der kathartische Effekt des antiken Dramas auf den modernen Menschen. Unveröfftl. Dipl.-Arbeit, Fakultät für Psychologie, Ruhr-Universität Bochum. Boesch, E. E. 1976. Psychopathologie des Alltags. Zur Ökopsychologie des Handelns und seiner Störungen. Bern etc.: Huber. Bohart, A. C. 1980. “Toward a cognitive theory of catharsis”. Psychotherapy: Theory, Research and Practice 17: 192–201. Boll, Th. 1998. Analyse kognitiver und motivationaler Aspekte spezifischer Emotionen am Beispiel von Schuldgefühlen und Empörung nach Führerscheinentzug. Hamburg: Kovac. Breuer, J. and Freud, S. 1895/1916. “Über den psychischen Mechanismus hysterischer Phänomene”. In Studien über Hysterie, (3. Aufl.) J. Breuer and S. Freud (eds), 1–14. (engl. 1991: On the psychological mechanism of hysterical phenomena: Preliminary communication) Leipzig/Wien: Deuticke. Bullough, E. 1912. “‘Psychical distance’ as a factor in art and an aesthetic principle”. British Journal of Psychology 5: 87–118. Dollard, J., Doob, L., Miller, N. E., Mowrer, H. O. and Sears, R. R. 1939. Frustration and Aggression. New Haven: Yale University Press.

222 Brigitte Scheele

Eckensberger, L. H. and Emminghaus, W. B. 1982. “Moralisches Urteil und Aggression: Zur Systematisierung und Präzisierung des Aggressionskonzeptes sowie einiger empirischer Befunde”. In Aggression. Naturwissenschaftliche und kulturwissenschaftliche Perspektiven der Aggressionsforschung, R. Hilke and W. Kempf (eds), 208–280. Bern etc.: Huber. Edelstein, W., Nunner-Winkler, G. and Noam, G. (eds.). 1993. Moral und Person. Frankfurt/M.: Suhrkamp. Epstein, S. 1984. “Controversial issues in emotion theory”. Review of Personality and Social Psychology 5: 64–88. Erb, E. 1993. Selbstentwicklung/-sozialisation statt Selbstaktualisierung/-verwirklichung?! Zur konzeptuellen Kritik und Weiterführung des zentralen Motiv-Konstrukts der Humanistischen Psychologie. Bericht aus dem Psychologischen Institut, Universität Heidelberg. Nr. 77. Eron, L. D. 1986. “Interventions to mitigate the psychological effects of media violence on aggressive behavior”. Journal of Social Issues 42 3: 155–169. Feshbach, S. 1961. “The stimulating versus cathartic effects of a vicarious aggressive activity”. Journal of Abnormal and Social Psychology 63: 381–385. Feshbach, S. 1964. “The function of aggression and the regulation of aggressive drive”. Psychological Review 71: 257–272. Feshbach, S. 1989. “Fernsehen und antisoziales Verhalten. Perspektiven für Forschung und Gesellschaft”. In Empirische Medienpsychologie, J. Groebel and P. Winterhoff-Spurk (eds), 65–75. München: Psychologie Verlags Union. Flashar, H. 1984. “Die Poetik des Aristoteles und die griechische Tragödie”. Poetica, Zeitschrift für Sprach- und Literaturwissenschaft 16: 1–23. Freiwillige Selbstkontrolle Fernsehen e.V. 1995. “Freigegeben ab …” Jugendschutz in den Medien. Ein Reportagefilm. Freud, S. 1926/1981. “Psycho-Analysis”. In Vol. XX, S. Freud (ed.), 263–270. London: Hogarth. Geen, R. G. and Quanty, M. B. 1977. “The catharsis of aggression: An evaluation of a hypothesis”. In Advances in Experimental Social Psychology, Vol. 10, L. Berkowitz (ed.), 1–37. New York: Academic Press. Geen, R. G. and Thomas, S. L. 1986. “The immediate effects of media violence on behavior”. Journal of Social Issues 42: 7–27. Gleich, U. 1995. “Das Angebot von Gewaltdarstellungen im Fernsehen”. In Gewaltdarstellungen in den Medien, M. Friedrichsen and G. Vowe (eds), 145–165. Opladen: Westdeutscher Verlag. Goranson, R. E. 1970. “Media violence and aggressive behavior: A review of experimental research”. In Advances in Experimental Social Psychology, Vol. 5, L. Berkowitz (ed.), 2–31. New York: Academic Press. Groeben, N. 1999. “Fazit: Die metatheoretischen Merkmale einer sozialwissenschaftlichen Psychologie”. In Zur Programmatik einer sozialwissenschaftlichen Psychologie, Bd. I: Metatheoretische Perspektiven. 2. Halbbd., N. Groeben (ed.), 311–404. Münster: Aschendorff. Groeben, N. and Scheele, B. 1977. Argumente für eine Psychologie des reflexiven Subjekts. Darmstadt: Steinkopff. Groeben, N. and Vorderer, P. 1988. Leserpsychologie. Lesemotivation — Lektürewirkung. Münster: Aschendorff. Gründer, K. 1968/1991. “Jacob Bernays und der Streit um die Katharsis”. In Epirrhosis. Festgabe für Carl Schmitt. 2. Teilbd., H. Barion et al. (eds), 495–528. Berlin: Duncker & Humboldt. Wiederabdruck. In Die Aristotelische Katharsis. Dokumente ihrer Deutung im 19. und 20. Jahrhundert, M. Luserke (ed.), 352–385. Hildesheim etc.: Olms. Hartmann, D. P. 1969. “Influence of symbolically modeled instrumental aggression and pain cues on aggressive behavior”. Journal of Personality and Social Psychology 11: 280–288. Heckhausen, H. 1989. Motivation und Handeln (2. Aufl.). Berlin etc.: Springer.

Back from the grave 223

Herkner, W. 1991. Lehrbuch Sozialpsychologie (5. Aufl.). Bern: Huber. Herzog, W. 1991. Das moralische Subjekt. Pädagogische Intuition und psychologische Theorie. Bern etc.: Huber. Huesmann, L. R. 1986. “Psychological processes promoting the relation between exposure to media violence and aggressive behavior by the viewer”. Journal of Social Issues 42: 125–139. Huesmann, L. R. and Malamuth, N. M. 1986. “Media violence and antisocial behavior: An overview”. Journal of Social Issues 42: 1–6. Jo, E. and Berkowitz, L. 1994. “A priming effect analysis of media influences: An update”. In Media Effects: Advances in Theory and Research, J. Bryant and D. Zillmann (eds), 43–60. Hillsdale, N. J.: Erlbaum. Kohlberg, L. 1995. Die Psychologie der Moralentwicklung. Frankfurt/M.: Suhrkamp. Kornadt, H.-J. 1982. “Grundzüge einer Motivationstheorie der Aggression”. In Aggression. Naturwissenschaftliche und kulturwissenschaftliche Perspektiven der Aggressionsforschung, R. Hilke and W. Kempf (eds), 86–111. Bern etc.: Huber. Kornadt, H.-J. and Zumkley, H. 1992. “Ist die Katharsis-Hypothese endgültig widerlegt?” In Aggression und Frustration als psychologisches Problem, Bd. 2, H.-J. Kornadt (ed.), 156–223. Darmstadt: Wissenschaftliche Buchgesellschaft. Kunczik, M. 1982. Aggression und Gewalt. In Medienpsychologie. Ein Handbuch in Schlüsselbegriffen, H. J. Kagelmann and G. Wenninger (eds), 1–8. München: Urban & Schwarzenberg. Langholf, V. 1990. “Die “kathartische Methode”. Klassische Philologie, literarische Tradition und Wissenschaftstheorien in der Frühgeschichte der Psychoanalyse”. Medizinhistorisches Journal 25: 5–39. Lessing, G. E. 1768/1988. Gesammelte Werke in 5 Bänden, Bd. 4: Hamburgische Dramaturgie. Berlin/Weimar: Aufbau Verlag. Leuzinger, P. 1997. Katharsis. Zur Vorgeschichte eines therapeutischen Mechanismus und seiner Weiterentwicklung bei J. Breuer und in S. Freuds Psychoanalyse. Opladen: Westdeutscher Verlag. Lukesch, H. and Schauf, M. 1990. “Können Filme stellvertretende Aggressionskatharsis bewirken?” Psychologie in Erziehung und Unterricht 37: 38–46. Luserke, M. (ed.). 1991. Die Aristotelische Katharsis. Dokumente ihrer Deutung im 19. und 20. Jahrhundert. Hildesheim etc.: Olms. Meyer, U.-W. and Reisenzein, R. 1996. “Emotion”. In Wörterbuch der Kognitionswissenschaft. G. Strube zus. and B. Becker et al. (eds), 139–141. Stuttgart: Klett-Cotta. Nichols, M. P. and Efran, J. S. 1985. “Catharsis in psychotherapy: A new perspective”. Psychotherapy 22 (1): 46–58. Nichols, M. P. and Zax, M. 1977. Catharsis in Psychotherapy. New York: Gardner Press. Parke, R. D., Berkowitz, L., Leyens, J. P., West, St. and Sebastian, R. J. 1977. “Some effects of violent and nonviolent movies on the behavior of juvenile deliquents”. In Advances in Experimental Social Psychology, Vol. 10, L. Berkowitz (ed.), 135–172. New York: Academic Press. Peper, D. 1981. Aggressive Motivation im Sport: Literaturanalyse, Theoriebildung und empirische Felduntersuchung zum Katharsis-Problem. Ahrensburg: Czwalina. Pohlenz, M. 1956. “Furcht und Mitleid?” Hermes 84: 49–74. Rule, G. B. and Ferguson, T. J. 1986. “The effects of media violence on attitudes, emotions, and cognitions”. Journal of Social Issues 42: 29–50. Schadewaldt, W. 1955. “Furcht und Mitleid? Zur Deutung des Aristotelischen Tragödienansatzes”. Hermes 83: 129–171. Scheele, B. 1990. Emotionen als bedürfnisrelevante Bewertungszustände. Grundriß einer epistemologischen Emotionstheorie. Tübingen: Francke.



224 Brigitte Scheele

Scheele, B. 1996a. “Emotion — Reflexion — Rationalität. Grundpostulate einer epistemologischen Emotionspsychologie”. Ethik und Sozialwissenschaften 2/3: 283–297. Scheele, B. 1996b. “Selbstkonzeptrelevantes Bewerten als ‘gewußtes Erleben’ von Emotionen — Plädoyer für ein hierarchisches Mehr-Komponenten-Modell!”. Ethik und Sozialwissenschaften 2/3: 351–360. Scheele, B. 1999. “Theoriehistorische Kontinuität: Lernen von Aggression oder Möglichkeiten zur Katharsis?!” In Zur Programmatik einer sozialwissenschaftlichen Psychologie. Metatheoretische Perspektiven, Bd. I, 2, N. Groeben (ed.), 1–83. Münster: Aschendorff. Scheff, Th.J. 1981. “The distance of emotion in psychotherapy”. Psychotherapy: Theory, Research and Practice 18: 46–53. Scheff, Th.J. 1983. Explosion der Gefühle. Über die kulturelle und therapeutische Bedeutung kathartischen Erlebens. Weinheim etc.: Beltz. Scheff, Th.J. and Bushnell, D. D. 1984. “A theory of catharsis”. Journal of Research in Personality 18: 238–264. Schneider, K. and Schmalt, H. D. 1981. Motivation. Stuttgart: Kohlhammer. Schneider, K. and Schmalt, H. D. 1994. Motivation (2. Aufl.). Stuttgart: Kohlhammer. Selg, H. 1992. “Ärger und Aggression”. In Psychologie des Ärgers, U. Mees (ed.), 190–205. Göttingen etc.: Hogrefe. Selg, H., Mees, U. and Berg, D. 1997. Psychologie der Aggressivität (2. Aufl.). Göttingen: Hogrefe. Symonds, P. 1954. “A comprehensive theory of psychotherapy”. American Journal of Orthopsychiatry 24: 697–712. Vorderer, P. 1992. Fernsehen als Handlung. Fernsehfilmrezeption aus motivationspsychologischer Perspektive. Berlin: Edition sigma. Zillmann, D. and Bryant, J. 1994. “Entertainment as media effect”. In Media Effects: Advances in Theory and Research, J. Bryant and D. Zillmann (eds), 437–461. Hillsdale, N. J.: Erlbaum. Zumkley, H. 1978. Aggression und Katharsis. Göttingen etc.: Hogrefe.



How to make alle Menschen Brüder Literature in a multicultural and multiform society Jèmeljan Hakemulder Free University Amsterdam

1.

Introduction

In this chapter I will discuss the role literature can play in making people of different ethnic or cultural backgrounds understand one another better and perhaps bring them closer to each other. To illustrate such potential effects of reading literature, let me start out with the opening chapter of Rushdie’s The Satanic Verses. Besides its much publicized relation with the Koran, the symbolism, themes, characters, and most of the plot of Rushdie’s book seem to concern the immigrant experience. On its first page we witness two Indians dropping from an airplane, Saladin Chamcha and Gibreel Farishta. “‘To be born again,’ sang Gibreel Farishta tumbling from the heavens, ‘first you have to die’” (p. 3) In the following pages readers are introduced more intimately to the two men (who, by the way, miraculously land safely on one of England’s snowy beaches). Through this introduction we learn something about possible motives for wanting to leave one’s home country. Gibreel clearly attempts to start a new life after having made a complete mess of his first in India. Saladin nourishes a strong desire to escape the clutches of Bombay. An indecent assault in a back street district represents for him the oppressive lack of opportunity for his personal development. It all had “come together in the stranger’s bony embrace, and now that he had escaped that evil skeleton he must also escape Bombay, or die” (p. 39). England represents for him a great civilization; it is the fabled country, with its crisp promises of the pound sterling. He wants to be “what his father was-not-could-never-be, that is, a goodandproper Englishman” (p. 43) Rushdie’s novel offers an understanding of the awkward position of floating somewhere in between two cultures, not being at home in either. On the one hand we see how Saladin does his utmost to become part of British society, to fool them

226 Jèmeljan Hakemulder

all “into thinking he was okay, he was people-like-us” (p. 44). On the other hand, back in India, he is considered to be a snob. His mother is convinced that he went abroad to learn contempt for his own kind: “we are just junglee people, he thinks so, and look how coarse our movies are […] and talking so fine and all” (p. 45). His first Indian love, Zeeny, confronts him with his position as a ‘deserter’: to his ‘own’ people he is someone who has spent “his life trying to turn white” (p. 55). Western readers’ understanding of what it must be like to be an immigrant may go hand in hand with a sharper insight in their own culture. First, Rushdie’s novel potentially enables readers to temporally take an outsider’s view of European culture and even if they do not agree, they at least become aware that there are other points of view to be taken. Saladin’s mother warns him not “to go dirty like those English (…) They wipe their bee tee ems with paper only” (p. 39). Soon after their arrival in England, that “coffin of an island” (p. 150), Gibreel and Saladin are caught by the freezing cold. And Gibreel notes that it is “no wonder these people grow hearts of bloody ice” (p. 135). From a Western perspective, England may seem to be a clean and attractive place to be, certainly as compared to a run-down Third World country with countless unregistered diseases, but here we are confronted with different sentiments. On top of that, readers of The Satanic Verses may feel shame or anger as a result of the account of how Saladin is treated, an immigrant in a hostile ‘host country.’ Saladin, who is a British citizen, finds himself hunted like an illegal immigrant; fifty-seven constables, slightly resembling pointer dogs, are sent after him to take him into custody. Three absurd immigration officers treat him as one would not treat an animal. Humiliated, having to perform the most ‘basest of rituals’ (eating his own excrements), Saladin can only hope this isn’t actually England. Readers may know of similar incidents which have occurred in reality, but reading about one, even in fiction, may still infuriate them and fill them with shame. In sum, what I propose is that readers’ perceptions of one particular social subgroup (immigrants) may be influenced by this book and, at the same time, that the book may affect their view on their own culture. In addition, it may help them become aware of cultural differences as well as similarities. The novel gives a good taste of the subcontinental couleur locale, for instance, the cult around the ‘Bollywood’ movie star Farishta, the philosophy of rebirth that plays such an important role in his thinking, and furthermore, the strange words, names, and places. They may also come to see some seemingly universal aspects, for instance in the characters’ thoughts and emotions. In some respects, Gibreel and Saladin might just as well be Europeans, Africans or Americans; their experiences go beyond social and cultural boundaries. Young Gibreel’s mother dies and husband and son have to try to cope with their loss. They bring around lunches in Bombay and “they buried their sadness beneath extra work, engaging in an inarticulate contest, who could carry the most dabbas on his head, who could acquire the most contracts per month, who could run faster, as though the greater labour would indicate the

How to make alle Menschen Brüder 227

greater love” (p. 19). Even if readers have not had similar experiences, they may feel that such a response to bereavement is not particular to Indians. The same holds for the love of the son for his father: “When he saw his father at night, the knotted veins building in his neck and at his temples [Gibreel] would understand how much the older man had resented him, and how important it was for the father to defeat the son and regain, thereby, his usurped primacy in the affections of his dead wife. Once he realized this, the youth eased off, but his father’s zeal remained unrelenting.” (p. 19) Finally, in the adolescent Gibreel readers may recognize a resemblance of adolescents all over the world: highly impressionable, feelings of being misunderstood, and strongly longing for love. “In his dreams he was tormented by women of unbearable sweetness and beauty, so he preferred to stay awake and force himself to rehearse some part of his general knowledge in order to blot out the tragic feeling of being endowed with a larger-than-usual capacity for love, without a single person on earth to offer it to” (p. 24).

2.

Human solidarity and world citizenship

Moving to a more general level of discussion, I will now look at some theories concerning effects like the ones proposed in the introduction. I will briefly highlight some pertaining to the function literature could fulfill in a multicultural society. One common notion among theorists is that reading literature from other cultures makes readers believe that all people are essentially similar. Whether this is actually true or not, some argue that it has some practical advantages over the idea that we are all totally different. Richard Rorty (1989), for instance, points out that the idea of similarity, even though it may be an illusion, enhances feelings of human solidarity. “In my utopia,” he says, “human solidarity would be seen not as a fact to be recognized by clearing away ‘prejudice’ or burrowing down to previously hidden depths, but rather, as a goal to be achieved. It is to be achieved not by inquiry but by imaginative ability to see strange people as fellow sufferers.” And it is reading novels in particular that would stimulate this capacity, for it “gives us the details about kinds of suffering being endured by people to whom we had previously not attended” (p. xvi). Reading novels enriches our moral awareness, because during the reading experience we find ourselves in the shoes of a wide diversity of people. Thus, we get better and better at understanding moral situations from different points of view. Nussbaum (1997) stresses the importance of awareness of human diversity as well as that of similarities. She musters support for a number of changes in American colleges “that some condescendingly refer to as ‘political correctness.’” These changes should be seen, according to Nussbaum, as efforts to grapple with issues of human diversity. It is simply a matter of ‘cultivating their humanity’ that colleges try

228 Jèmeljan Hakemulder

to develop their students into world citizens. One of the requirements of being a world citizen is ‘narrative imagination’, that is, the ability to think what it might be like to be in the shoes of a person different from oneself. Through narrative imagination students obtain an awareness of others. With an increasingly multicultural society and with the globalization of the economy, students need to have an imaginative capacity to enter into the lives of people of other ethnic backgrounds and nations. Nussbaum’s idea of effective reading (or, in her words “literary reading”) comprises two components. One is identification, showing resemblance to Rorty’s point. The other might be characterized as an Aristotelian view on literature. Roughly, it comes down to the idea that literature does not teach us about one particular situation in history in one particular place on earth, but about more universal psychological principles. As a result, Nussbaum suggests, readers may become more sensitive to the way in which people in general are formed by their environment. In other words, literature tells us something about the ‘genesis of character.’ Notions about the effects of reading literature can be found among theorists but may also play a significant role for a much wider group of readers. For example, popular belief would have it that Uncle Tom’s Cabin had something to do with the growing opposition to slavery and, in the end, with the Civil War. Furthermore, qualitative research (e.g. Lachenmann 1999; Shirley 1969) cites individual readers who say they have experienced an effect of literature on their perception of others. One of Shirley’s (1969) subjects says that Black Like Me by John Griffin changed my attitude and point of view about the racial situation. It also enlarged my knowledge of the situation in the South. After reading the book, I felt as if I wanted to really do something to help the Negro in the South. I was shocked by some of the events that took place and surprised that one group of humans could possibly treat another group of humans so cruelly (p. 407–408).

One could say this reader became a ‘fellow sufferer.’ The idea that a story about some distant country is a good introduction to its people might be a motive for readers, who, normally having a much too busy schedule, can only find the time for books when they are on holiday. For instance, in a qualitative study by Lachenmann (1999) one reader says that reading a book about farmers in Bali gave him a more intense feeling when on Bali, because one can perhaps read about things that are perhaps not in a travel guide. Rather about people, how they live. Because I didn’t live in Indonesia on a farm, and I can’t talk with the Balineses, because I don’t speak Indonesian; one feels a stronger bond with them. Because one reads about them, and lives together with them in such a family, goes through something with them. Then one gets the feeling that is more intense (p. 495).

How to make alle Menschen Brüder 229

In sum, a wide range of readers, theorists as well as ‘naive’ readers, think or report that reading literature affects their perception of others: it is said to enhance empathy for others because we have come to see them as part of the ‘ingroup’ rather than an ‘outgroup’ and, furthermore, it may enhance the ability to imagine oneself in the position of someone else.

3.

Previous research

Throughout history, in Western as well as in non-Western cultures, cosmopolitanism such like in Nussbaum’s plea for world citizenship has exerted its attraction on many people. On the other hand, there is also an undeniable aptitude to define one’s ‘ingroup’ not as all ‘humans all over the world’, but more narrowly, as just ‘fellow country men’. The unification of Europe, for instance, is accompanied by an upsurge of nationalism and even more narrowly defined forms of group solidarity. For literature to counter such trends and turn readers into world citizens seems indeed like an “uphill battle”, as Nussbaum puts it. Literature is available to many; we see an increase in the publication of literary texts, novels, and stories from Third World countries; more and more immigrant writers take their position in the literary canon; and still very few people consider each other as their brothers and sisters. This makes one wonder what we actually know about these alleged effects of literature. Do these effects actually occur, and if so, how strong are they and what makes them come about? In this section I will look at experimental research with regard to the effects on intergroup cognitions, that is, perceptions, memories, stereotypes, impressions, judgments, evaluations, and behavioral intentions pertaining to outgroups; intergroup cognitions refer to phenomena occurring ‘inside the heads’ of social perceivers (cf. Sedikides, Schopler et al. 1998). We focus our attention on experimental research because this type of research allows us to draw conclusions about causal relations between, on the one hand, a particular treatment (for instance, reading a literary text) and, on the other hand, a change in attitude or beliefs (for instance, a more positive attitude toward members of some outgroup). Therefore, we will restrict our discussion here to experimental studies, like, for instance, that of Litcher and Johnson (1969). In their experiment sixty-five white school children, seven-year-olds, were randomly distributed across two groups: a control group and one experimental group. In the former, participants read a book with pictures of white people and stories about white characters. In the experimental group a manipulated version of the book was read. The experimenters replaced the pictures with photographs of non-white characters. Also, the names of the characters were changed to better match with the pictures. For a period of four months the textbooks were read in class. All subjects were tested twice, once before the experiment

230 Jèmeljan Hakemulder

and once after. The tests were designed to assess ‘social distance’, that is, the physical distance subjects would want to maintain between them and members of the target outgroup. One of the tests that were administered was the Show Me Test. It consists of twelve portraits: three white boys and three black, three white girls and three black. The pictures were shown pair-wise and participants were asked, for example, “Please show me the one that you would like to sit next to at school” and “Please show me the one that you do not want in your school.” Analyses comparing group mean scores on pretests and posttests revealed a significant reduction of social distance in the experimental group, while no changes were observed in the control group. The experimental design of the study allows us to conclude that the differences between control group and experimental group were caused by reading the manipulated textbook. A survey of experimental studies like the one of Litcher and Johnson yielded seventeen studies (Alsbrook 1970; Beardsley 1979; Brisbin 1971; Fisher 1965; Frankel 1972; Geiger 1975; Gimmestad and De Chiara 1982; Hayes 1969; Heintz 1988; Jackson 1944; Kimoto 1974; Koeller 1977; Litcher and Johnson 1969; Schwartz 1972; Stone 1985; Tauran 1967; Zucaro 1972).1 In general, these studies examine whether reading narratives portraying a particular outgroup affects readers’ attitudes toward that group. The studies were evaluated using standard criteria for experimental research (Cook and Campbell 1979). This revealed that not all these experiments were equally reliable; in other words, not all studies offer direct evidence to support researchers’ claims about the effects of the treatments. The evaluation revealed that only half of the experiments met the standards for reliability, six of which revealed significant effects of the treatments (Brisbin 1971; Geiger 1975; Jackson 1944; Litcher and Johnson 1969; Tauran 1967; Zucaro 1972) while two showed no effects at all (Beardsley 1979; Schwartz 1972). The available data about the experimenters’ procedures (for instance., the texts that were read by the subjects) do not allow us to draw any conclusions from these contradictory findings. Moreover, much more research is needed to make any definite statements about the effects of reading on outgroup perception. However, we can say that there are some indications that reading stories with positive portrayals of outgroup members results in a positive change in attitude toward that group. Similarly, negative characterizations can be assumed to result in negative attitudes. Although it cannot be said that claims about effects on attitudes are unfounded (cf. Stolnitz 1991), the available evidence leaves some essential questions unanswered (see for a discussion Hakemulder 1997). One important drawback of the available research is that little effort has been made to relate text features to the psychological processes that could be responsible for the attitude changes that were registered. It may be contended that such research is typically a task of social scientists rather than literary scholars, but as far as I can see, the hypotheses and questions posed in the two previous sections are not directly addressed by psychologists

How to make alle Menschen Brüder

nor by sociologists. In what follows I will therefore attempt to link particular aspects of the reading experiences with changes in outgroup cognitions. I will focus on hypotheses put forward by the theorists quoted earlier. However, before we can embark on an exploration of literature’s effects on the perception of outgroups, it is opportune to map out what is known about ways to change outgroup cognitions.

4. What ingroups think of outgroups Generally, psychologists have revealed the following robust phenomena in intergroup relations. First, there is overwhelming evidence for ingroup favoritism; ingroup members are evaluated more positively than outgroup members, and they are favored in the allocation of rewards (Brown 1996). Second, outgroup members tend to be seen as more similar to each other than ingroup members (Linville and Fischer 1998). In other words, perceived variability of outgroups is systematically lower than that of ingroups. This outgroup homogeneity is associated with biases in intergroup perception and behavior. Third, people have a tendency to attribute positive behavior of ingroup members and negative behavior of outgroup members to personal qualities of those members. Similarly, they tend to attribute negative behavior of ingroup members and positive behavior of outgroup members to external situational factors rather than personal qualities. Such biases in intergroup perception are known as the ultimate attribution error (for a review see Hewstone 1990; Pettigrew 1979). Finally, research reveals that stereotypes are activated automatically and involuntarily, even in people who reject prejudice and the use of stereotypes (e.g. Bargh et al. 1996). While there may be little discussion about the phenomena themselves, till now an unresolved controversy lies in determining the causes. For instance, the scapegoat theory suggests that prejudice is caused by economic hardship. The resulting frustrations would lead to aggression toward easy targets, which often are outgroup members. However, there have been mixed results in providing evidence for this theory (see for a discussion Brown 1996). Whatever the cause of stereotyping, prejudice, and intergroup conflict, the issue here is whether reading literature can counteract effects of, say, the lack of economic prospect. We have seen that there is some (modest) evidence for such an influence. Now let us look a bit closer at how it may come about. Psychologists have proposed three models of changes in intergroup cognition. First, it has been assumed that people gradually change their cognitions about outgroups by an additive influence of each piece of disconfirming information (bookkeeping model). Second, the conversion model suggest that outgroup cognitions are radically changed through dramatic disconfirming information, but that no changes occur as a result of only mildly disconfirming information. A third

231

232 Jèmeljan Hakemulder

model proposes that disconfirming information causes the formation of subtypes: the outgroup member is seen as an exception, unrepresentative of the group as a whole. Thus, the perception of the outgroup, for instance in terms of its homogeneity, is not affected. A survey of the research reveals strong support for subtyping, some for bookkeeping and none for conversion (Hewstone and Lord 1998). One way in which stereotype disconfirming information may come to us is through second-hand examples (Hewstone and Lord 1998). Some suggest that judgments of outgroups is a memory-based process. In this view social categories like stereotypes consist of a set of exemplars representing specific instances or subtypes of the category. Instances that determine the category may consist of both first-hand exemplars (personal encounters), as well as second hand ones (representations in the media, books, or descriptions by friends or relatives). In what follows three hypotheses will be put to the test. Hypothesis 1 predicts that stereotype disconfirming information in a narrative text decreases perceived homogeneity of the outgroup in question. We will refer to this hypothesis as the Stereotype Disconfirmation Hypothesis. There is some evidence that stereotype disconfirming information in second-hand exemplars can produce changes in beliefs, but it is also plausible that such information causes subtyping: the association between one particular fictional representation of an outgroup member and the real-world outgroup as a whole may be too weak. For example, Gibreel Farishta may seem idiosyncratic rather representative. It may even be that the more we get to know about him, the more we see him as an individual. This process of individuation is known to weaken the association between exemplar and group even more. In sum, it is not at all sure that stereotype disconfirming information does have the proposed effect. However, if we are to find such an effect on outgroup perception, it still remains to be seen what the conditions of such an influence are. For this we will focus on two suggestions already proposed in the introduction of this chapter. They concern two different aspects of the reading experience. For the second hypothesis we assume that, while reading, readers imagine themselves in the shoes of the characters. Thus, stories about outgroup members can affect their beliefs about what it must be like to be a member of that particular outgroup. Hypothesis 2, what we will call the Role-Taking Hypothesis, suggests that identification with, or rather taking the role of a character, affects outgroup perception. Hypothesis 3 suggests an alternative explanation for the effects of narratives. In the introduction it was proposed that considering the universal implications of a story may affect beliefs about others. In addition, research stresses the importance of perceived typicality: considering a particular outgroup member as a typical representative of his or her group would have a greater effect on outgroup cognitions (Hewstone and Lord 1998). For example, reading The Satanic Verses as a book from which we can learn something about immigrants in general seems

How to make alle Menschen Brüder 233

likely to have a greater effect. Therefore, an ‘Aristotelian’ reading style might contribute to the effects of reading, for it focuses readers´ attention on typicality rather than individuating aspects of the characters. Thus, the Aristotelian ReadingStyle Hypothesis predicts that readers’ attention for general psychological aspects of the story is conditional for the effects on their ideas about what it must be like to be an outgroup member.

5.

Experiments

Study One In the first study we will look at the effects of stereotype disconfirming information on the perception of outgroups (Hypothesis 1). In addition we will examine to what aspect of the reading experience this effect can be attributed. We will first focus on the effects of role-taking: while reading, readers may take the role of a character (an empathic response), and as a result they will be stimulated to imagine what it must be like to be that character. After reading, they come to generalize this knowledge to people outside the fictional world (Hypothesis 2).2 Putting the first two hypotheses to the test, I compared the effect of two texts, an essay and a story. Both texts concern the same issue, namely the position of women in fundamentalist Islamic countries. The story is the first chapter of a novel, The Displaced by Malika Mokkeddem. An Algerian woman, Sultana, returns to her home country. She has lived in France for years, but has come back to attend the funeral of a good friend. As soon as she sets foot in Algeria, she is pestered by hostile men. Religious leaders do not allow her to attend the funeral. She does so anyway, and this causes a very tense situation with a lot of grief and anger. The essay, a chapter of Goodwin’s The Price of Honor, takes a more general view of things. It does not concern one woman and her particular experiences, but discusses, for instance, traditional customs that deprive women of many of their rights, and the resistance of women against their underprivileged position. In short, the subject is the same and therefore one would expect readers of both texts to become more aware of women’s position in fundamentalist Islamic cultures. However, the story offers readers an opportunity to enter into a fictional world, and experience, as it were, for themselves what that position would be like. The essay does not give that opportunity, but directly offers all the information needed to figure out what the position of a woman in Algeria is like. Thirty-five college freshmen participated in the experiment as an introduction to empirical studies of literature. They were randomly assigned to one of three groups: a control group, a group that read the story and a group that read the essay. All participants were told the experimenter was interested in the relation between

234 Jèmeljan Hakemulder

their opinions and their aesthetic appreciation of the texts. Control group subjects were asked first to answer some questions and do a test and then read the text, while in the experimental groups the text was read first and then the tests were administered. In one of the tests participants were asked to imagine what it would be like to be a woman in a fundamentalist Islamic country and then respond to some statements on four-point scales. For instance, participants were asked to what degree they thought it likely that an Algerian woman would accept the cultural norms for the relationship between men and women in her country. It was expected that the story would have a stronger effect on participants’ beliefs about being an Algerian woman than the essay would, because it offers an opportunity for roletaking while the essay does not. Comparison of control group responses and those of the experimental groups revealed that the story did indeed have a significant effect on participants’ beliefs. For instance, they considered it less likely that women in Algeria would accept the culturally given norms for men-women relations (see Figure 1). Because of the negative portrayals of Algerian men, it might have been expected that the story enhanced social distance. However, the results showed no indication of this. Nor did I register an effect on solidarity, in spite of Rorty’s expectations about becoming ‘fellow sufferers’. Participants were asked whether they would consider social asylum for women who feel severely repressed in their own country. Neither the story, nor the essay seemed to have affected participants’ opinion on this matter. It may be concluded that stories presenting stereotype disconfirming information in its characters may affect readers’ imagination about real-world people. Instead of thinking of Algerian women as all being compliant, story readers exposed to a non-compliant Algerian character were shown to have stronger non-stereotypical cognitions about Algerian women than control group subjects, as well as the essay readers. One way to interpret the difference between the story and essay group is that it was caused by the fact that one offered the opportunity to take the role of one of these Algerian women, and thus experience what it must be like to be one of them, while the other just gave general information about such women. Thus, the results of this experiment would provide evidence for both the Stereotype Disconfirmation Hypothesis (1) as well as for the Role-Taking Hypothesis (2). It should be noted that the design of this study allows us to compare the effects of narrative and nonnarrative information, not fiction and nonfiction, nor literature and nonliterature. Moreover, one point of possible criticism is that the essay, selected by the experimenter himself, might not be as convincing as essays can be. On the other hand, in another part of the experiment the subjects’ judgment of the two texts suggests differently: they attributed greater persuasive force to the essay than to the story. Still, the experiment does lack some degree of control: the

How to make alle Menschen Brüder 235

4

Probability estimates

3

2

1

0 Control

Story

Essay

Figure 1.Effect of narrative and non-narrative information on beliefs about others: perceived likelihood of Algerian women accepting Algerian norms for men-women relations

differences between the two texts was not restricted to having characters in it or not; there were of course many other differences that could be responsible for the results. To avoid these problems a second experiment was conducted in which an attempt was made to highlight the effects of role-taking by giving it an extra boost.

Study Two Participants, fifty-five college freshmen, were randomly assigned to three groups: one control group, and two experimental groups. In one experimental group participants were given an empathy-building instruction (cf. Bourg et al. 1993). They were told that the experimenter was interested in the relation between reading strategy and recall. The instruction read that “some psychologists claim that reading empathetically causes readers to remember more of what they read.” Therefore, they were asked to “try and imagine, while you are reading you are in the position of the main female character. Imagine you see was she sees, you feel what she feels, and try to imagine the events are all happening to you yourself.” The second experimental group received a placebo instruction: they had to read the story and

236 Jèmeljan Hakemulder

mark with pen or pencil the structure of the narrative, because “psychologists claim that structuring the text while reading enhances recall.” It was expected that, like in the previous study, reading the story would already result in a significant effect on participants’ beliefs about being an Algerian woman. As to the empathy-building instruction it was expected that it would give an extra boost to the readers’ role-taking behavior and thus enhance the effects of reading. Identifying with the character’s perspective would enhance an active participation in the fictional world, affecting the knowledge that is accessible in readers’ memory, and stimulating both inferential and noninferential responses (cf. Gerrig 1996). The analyses showed that both reading instructions lead to a significant change in readers’ beliefs, but also that the empathy instruction lead to a significantly stronger influence than the placebo instruction (see figure 2). The result for the placebo condition is a replication of the results of Study One, namely the finding that reading a story may affect readers’ beliefs about real-world outgroups. The results for the empathy-building condition suggest that role-taking must have contributed to the effects on outgroup cognitions. This finding supports both the Stereotype Disconfirmation and the Role-Taking Hypothesis. On other measures again no effects was found: not on solidarity, nor on social distance. 6

5

Probability estimates

4

3

2

1

0 Control

Empathy

Placebo

Figure 2.Effect of reading style on beliefs about others: perceived likelihood of Algerian women accepting Algerian norms for men-women relations

How to make alle Menschen Brüder 237

One point of discussion may be that the task may have functioned as a manipulation check for the empathy-building strategy. The instruction may have dictated the results: participants of the empathy-group may have considered the task as a test of whether they had followed the instruction to empathize. Consequently, these results could be interpreted as an indication that participants merely wanted to show that they had empathized with the character. However, the results of Study One and Two put together suggest that reading narratives does affect the way we believe other people feel and think, and that this effect is at least in part caused by role-taking.

Study Three In a third experiment I examined Nussbaum’s suggestion that an Aristotelian reading style would result in a better understanding of others, more in particular, in an awareness of the influence of external situational factors on people’s (for instance, outgroup members) behavior (Hypothesis 3). Nussbaum also suggested that an empathic reading style enhances understanding of the emotions, thoughts and behavioral motives of characters. Imagining oneself in the shoes of the character would make readers elaborate on what it would be like to be someone else. In this study we will try to find out which of these two reading styles, an ‘Aristotelian’ or an ‘Empathic’ one (if either of them) would contribute to readers’ understanding of a character’s behavior. For this purpose I picked another story, a text presenting a character who’s behavior seems less acceptable than that of Sultana. Participants read Svegi Soysal’s “The Pierced Amulet” which is about a rich, spoiled boy called Necip. The boy is born and raised in a part of Greece occupied by Turkey. We see him growing up among an army of servants and pliant family members, not knowing any different than that everyone around him should be instrumental for his personal pleasures. The title of the story refers to an amulet the boy is given at his birth to protect him against the evil eye. When he is older, he deliberately destroys the amulet while practicing shooting. Necip impregnates a servant, who, as a result, has to flee. She takes with her the pierced amulet and gives it to her little boy. The child is killed at an early age by tripping over a tub filled with hot water. The participants were thirty-two undergraduate students of the Faculty of Arts, Utrecht University, who participated in the study as an introduction to empirical studies of literature. The experiment was conducted as part of a course on literary theory, and introduced as a recall study. Participants were randomly assigned to three groups: two experimental and one control group. The two experimental groups first read the story and then answered some questions on their beliefs about Turkish culture in general. The control group first answered the questions and then read the story. The first experimental group was reminded of what they had learned previously in their course about Aristotle’s view on literature and how literature is about “universal human experiences, about human faults, fears and desires.”

238 Jèmeljan Hakemulder

Participants were instructed to read the story, keeping in mind that it is not important what actually happens but that literature is about what might have happened. The second group was told to read the story ‘empathetically.’ They were to imagine, while reading the story, that they were in the main character’s position, imagining to see what he sees, smell what he smells and feel what he feels, in short, trying to imagine that things that happen to him happen to them. Subjects were to respond to some statements on six-point scales. The data were to indicate changes in their beliefs about Turks as well as their social distance toward that outgroup. The so-called recall test that was conducted was an openresponse assignment. Participants were asked to respond to the main character in 15–20 sentences focusing on either similarities or differences between them and the character. A content analysis of the responses was to reveal whether participants of the Aristotelian or the Empathy group would have a better understanding of the ‘genesis of character’ than the control group. Many participants focused on how Necip became what he was. Nobody wanted to clear him from his blame, but many of the statements the participants made emphasized extenuating circumstances. For example one subject said “He is a product of his upbringing, and therefore he is less to blame for his behavior.” Another subject claimed that “If I had been raised like that, I would have had a similar personality.” In general their statements suggest that it was inevitable that he would become spoilt and egoistic. Fourteen participants spontaneously formulated extenuating circumstances; fifteen did not. Figure 3 shows the distribution over the three groups. The number of subjects is small, so we should be careful in our interpretation of these results. However, a Pearson Chi-Square test suggests that the differences between the groups are significant ( c2 = 6.75, p < .05). It seems that the distribution in the empathy group is the reverse of that of the control group, while the Aristotle group takes a middle position. Participants who read the text while imagining themselves in the position of Necip were more likely to think of extenuating circumstances than the participants of the control group who had read the text without any special reading instruction. The reading instruction of the Aristotle group clearly did not have this effect. On the other measures no significant effects were registered. The story does not seem to have affected social distance. Neither did I find any indication that either reading style affected participants beliefs about the outgroup. Comparing these results with those of Study One and Two in which we did see significant effects of role-taking, it seems that the effects are still somewhat unpredictable. It is possible that the lack of an effect is simply due to the fact that Necip is a rather negative character, while Sultana is not. However, as we have seen above, the role-taking instruction did affect perception of this character.

How to make alle Menschen Brüder 239

100% 2 80% Percentage of subjects

6 60%

7

None 40%

8

Extenuation

4

20% 2 0% Control

Aristotle

Empathy

Figure 3.Looking for extenuating circumstances: the effects of reading style on the number of subjects making external attributions

6. Discussion and conclusion Societies all over the world are becoming more and more multicultural and thus more complex. With this tendency people may feel a stronger need to rely on social categories to simplify and interpret their social environment. Though these categories may fulfill a psychological need, there is always the risk of oversimplification. It seems that reading literature, from a wide range of cultural origins, could contribute to building a buffer for this risk. The experiments presented in this chapter show that reading stories affects cognitions about outgroups. Reading selected stories can reduce perceived outgroup homogeneity; in Study One and Two, subjects’ cognitions about Algerian women were affected by stereotype-disconfirming information presented in a fictional character. Readers’ role-taking behavior seems to contribute to the effect. In addition it was found in Study Three that this reading style may also decrease the tendency known as the ultimate attribution error, or, in other words, an increase in readers’ sensitivity to the ‘genesis of character.’ Results suggest it is a role-taking reading style rather than an ‘Aristotelian’ that causes such effects. Why this line of research deserves serious attention seems obvious. Although not everyone will believe that changing negative stereotypes is a central function of

240 Jèmeljan Hakemulder

literature, most will agree that every possible mean to obtain this goal should be explored. The general issue at hand, changing outgroup cognitions, is the focus of extensive research in social psychology. In the humanities, scholars concentrate on the interpretation of the representation of outgroups rather than on actual effects of such presentations (cf. Smelik 1999). Although evidence suggests that secondhand exemplars do indeed affect perception, such relations remain speculative. Let me briefly refer to three potential problems. First, research shows that stereotype disconfirming information, in order to be effective, should originate from a source that is considered reliable (Macrae et al. 1992). Many readers, however, will not consider literary texts as such. And for good reason: some texts are deliberately unreliable, for instance in the narrators’ presentation of the events. Second, studies have shown that ambiguous information about outgroup members hinders stereotype change (e.g. Higgins and Bargh 1987). It may be that the ambiguity typical of so many literary texts makes literature a less suitable instrument for attitude change. Last, some literary texts do present readers with negative stereotypes and therefore are not likely to have a positive effect on stereotype change. Nevertheless, in the present chapter we have seen that reading literature can affect intergroup perception; there does seem to be some ground to believe that it can widen the range of people readers are willing to consider part of their ingroup. Further research should make clear what the possible social applications of literature really are; considering the implications of including or excluding people from the ingroup, it seems one cannot overstress the importance of the issue.

Notes 1. See Hakemulder (2000) for the details. See also, for comparison, Van Dijk (1991) on the effects of ethnic stereotypes in news report narratives. 2. For a full report on procedures, analyses, and results of Study 1 and 2, see Hakemulder (2000).

References Alsbrook, E. Y. 1970. Changes in Ethnocentrism of a Select Group of College Students as a Function of Bibliotherapy. PhD Diss. Urbana-Champaign: University of Illinois. Bargh, J. A., Chen, M. and Burrows, L. 1996. “Automaticity of social behavior: Direct effects of trait construct and stereotype activation on action”. Journal of Personality and Social Psychology 71: 230–244. Beardsley, D. A. 1979. The Effects of Using Fiction in Bibliotherapy to Alter the Attitudes of Regular Third Grade Students Toward Their Handicapped Peers. PhD Diss. Columbia: University of Missouri.

How to make alle Menschen Brüder 241

Biskin, D. S. and Hoskisson, K. 1977. “An experimental test of the effects of structured discussions of moral dilemmas found in children’s literature on Moral Reasoning”. The Elementary School Journal 78: 407–416. Bourg, T., Risden, K., Thompson, S. and Davis, E. C. 1993. “The effects of an empathy-building strategy on 6th graders’ causal inferencing in narrative text comprehension”. Poetics 22: 117–133. Brown, R., Intergroup Relations, Hewstone, M., Stroebe, W. and Stephenson, G. M. 1996. Introduction to Social Psychology. Oxford: Blackwell. Cook, T. D. and Campbell, D. T. 1979. Quasi-Experimentation: Design and Analysis Issues for Field Settings. Chicago: Rand McNally. Fisher, F. L. 1965. The Influences of Reading and Discussion on Attitudes of Fifth Graders Toward American Indians. PhD Diss. Berkeley: University of California. Frankel, H. L. 1972. The Effects of Reading: The Adventures of Huckleberry Finn on the Racial Attitudes of Selected Ninth-Grade Boys. PhD Diss. Philadelphia: Temple University. Geiger, K. F. 1975. “Jugendliche lesen ‘Landser’-Hefte. Hinweise auf Lektürefunktionen und -wirkungen”. In Literatur und Leser; Theorien und Modelle zur Rezeption literarischer Werke, G. Grimm (ed.), 324–341. Stuttgart: Reclam. Gerrig, R. J. 1997. “Participatory aspects of narrative understanding”. In Vol. 52 of Advances in Discourse Processes. Empirical Approaches to Literature and Aesthetics, R. J. Kreuz and M. S. MacNealy (eds), 127–142. Norwood, NJ: Ablex. Gimmestad, B.J. and Chiara, E. de. 1982. “Dramatic plays: A vehicle for prejudice reduction in the elementary school”. Journal of Educational Research 76: 45–49. Hakemulder, J. 1997. “Effects of literature: A review”. In The Systematic and Empirical Approach to Literature and Culture as Theory and Application, S. Tötösy de Zepetnek and I. Sywenky (eds), 31–50. Edmonton: University of Alberta. Hakemulder, J. 2000. The Moral Laboratory; Experiments Examining the Effects of Reading Literature on Social Perception and Moral Self-Concept. Amsterdam: Benjamins. Hayes, M. T. 1969. An Investigation of the Impact of Reading on Attitudes of Racial Prejudice. PhD Diss. Boston: Boston University, 1969. Heintz-Lawler, B. A. 1988. Evaluation of the Effect of a Literature-Based Unit Designed to Enhance the Attitudes of Primary Grade Children Toward the Elderly. PhD Diss. Cornvallis: Oregon State University. Hewstone, M. 1990. “The ‘ultimate attribution error’? A review of the literature on Intergroup Causal Attribution”. European Journal of Social Psychology 20: 311–336. Hewstone, M. and Lord, Ch.J. 1998. “Changing intergroup cognitions and intergroup behavior: The role of typicality”. In Intergroup Cognition and Intergroup Behavior, C. Sedikes, J. Scholer et al. (eds), 367–392. Mahwah, NJ: Erlbaum. Higgins, E. T. and Bargh, J. H. 1987. “Social cognition and social perception”. Annual Review of Psychology 38: 369–425. Jackson, E. P. 1944. “Effect of reading upon attitudes toward the negro race”. Library Quarterly 14: 47–54. Kimoto, C. K. 1974. The Effect of a Juvenile Literature Based Program on Majority Group Attitudes Toward Black Americans. PhD Diss. Pullman: Washington State University, 1974. Koeller, S. 1977. “The effect of listening to excerpts from children’s stories about MexicanAmericans on the attitudes of sixth graders.” Journal of Educational Research 70: 329–334. Lachenmann, J. 1999. Interviews mit Lesenden. Wirkungen von Literatur auf Musiker und Mediziner. Unpublished thesis, University of Munich.



242 Jèmeljan Hakemulder

Linville, P. W. and Fischer, G. W. 1998. “Group variability and covariation: Effects on intergroup judgment and behavior”. In Intergroup Relations, C. Sedikides, J. Schopler and Ch. A. Insko (eds), 123–150. Mahwah: Erlbaum. Litcher, J. H. and Johnson, D. W. 1969. “Changes in attitudes toward negroes of white elementary school students after use of multiethnic readers”. Journal of Educational Psychology 60: 148–152. Macrae, C. N., Shepherd, J. W. and Milne, A. B. 1992. “The effects of source credibility on the dilution of stereotype-based judgments”. Personality and social psychology bulletin 18: 765–775. Nussbaum, M. C. 1997. Cultivating Humanity; A Classical Defense of Reform in Liberal Education. Cambridge: Harvard University Press, 1997. Pettigrew, T. F. 1979. “The ultimate attribution error: Extending Allport’s cognitive analysis of prejudice”. Personality and Social Psychology Bulletin 5: 461–476. Rorty, R. 1989. Contingency, Irony, and Solidarity. Cambridge: Cambridge University Press. Schwartz, C. S. L. 1972. The Effect of Selected Black Poetry on Expressed Attitudes Toward Blacks of Fifth and Sixth Grade White Suburban Children. PhD Diss. Detroit: Wayne State University. Sedikes, C. and Scholer, J. et al. (eds). 1998. Intergroup Cognition and Intergroup Behavior. Mahwah, NJ: Erlbaum. Stone, M. R. M. 1985. The Effect of Selected Children’s Literature on Children’s Attitude Toward the Elderly. PhD Diss. Tuscaloosa: University of Alabama. Smelik, A. 1999. Effectief beeldvormen; theorie, analyse en praktijk van beeldvormingsprocessen. [Effective Imaging: Theory, Analysis and Practice of Imaging Processes] Assen: Van Gorcum. Shirley, F. L. 1969. “The influence of reading on concepts, attitudes, and behavior”. Journal of Reading 12: 369–372 and 407–413. Tauran, R. H. 1967. The Influences of Reading on the Attitudes of Third Graders Toward Eskimos. PhD Diss. College Park: University of Maryland. Van Dijk, T. A. 1991. Racism and the Press. London: Routledge. Zucaro, B. J. 1972. The Use of Bibliotherapy Among Sixth Graders to Affect Attitude Change Toward American Negroes. PhD Diss. Philadelphia: Temple University.





“Sad autumn” and cultural representations A comparative study of Japanese and Israeli “autumn”* Ziva Ben-Porat Tel Aviv University

Les feuilles mortes se ramassent à la pelle, Tu vois je n’ai pas oublié… Et le vent du Nord les emporte, Dans la nuit froide de l’oubli…

Prevert’s “Les Feuilles Mortes”, set to music by Cosma and performed by Yves Montand, is one of the best known autumn songs in Western culture (at least for certain generations). Although the word automne does not appear in the text, no one who heard the song is likely to contest this classification, enhanced by radio programs in September and October. It was also the highlight of Yves Montand’s concert tour of Japan (1982), albeit in Japanese translation. What is it that makes people from different cultures react to this song in a similar way? What causes the immediate recognition of the appropriateness of the presentation that links together the natural phenomenon of autumn with remembered love and a realization of life moving on and bringing change? What motivates the parallelism between a deserted beach and strewn dead leaves? Is everybody sad in autumn, even when the cold wind is a welcome change after a hot summer and some people might have met the love of their life in October? Is it only a question of literary traditions and presentations in other media characteristic of our global village, or is it a question of the way people conceptualize the world; or maybe even both?

1.

Introduction

My work on literary presentations of autumn began almost two decades ago, when I was working on the relations between the texts (lyrics) of popular songs and those of poems belonging to ‘high literature’ (lyric poetry). I sought to prove the hypotheses that the popular system imports at least some of its models of presentation from

244 Ziva Ben-Porat

the ‘higher’ literary stratum and, because of its oral characteristics, uses them in a stereotyped — prototypical — manner. Within the dual framework of a systemic approach and a thematic study, I was initially looking for recurrent descriptive elements that made the theme of autumn into a literary topos.1 That is to say, I was looking for the mechanism connecting, for example, the sight of colorful forests in October with sad melodies and with meditations on death and beauty, on old age and recollected loves. Having identified the most constant features of the autumnal presentations in my corpus, I tried to locate the point of origin of the recurring presentational elements, and trace the inter-systemic (different cultures) and intrasystemic (different periods, literary strata, genres or media within the ‘same’ culture) migrations and transformations of the topos. It was the search for origins that made me move out of Hebrew literature into the European system. The findings — conclusive as they might have seemed in terms of comparative and historical thematic studies — compelled me to go on and even venture out of Western literature.2 I realized that I was looking at regularities of presentation that could contribute to our understanding of the role literary artifacts — and literary studies — could play in a number of related research areas: semiotics of presentation, psychology of reading, and cognitive anthropology. Within semiotics of presentation such regularities (for example, the recurrent joint appearance of autumn leaves and a melancholy contemplation of life and/or death) expose the arbitrary non-mimetic nature of the season concept as a complex linguistic sign and provide a key to the structure of such concepts. In the psychology of reading such regularities both constitute and reflect the cognitive schema (representation) which is active in the production of autumnal presentations and in their comprehension. Both the recurrent elements and the readers’ automatic comprehension of their signfunctions clearly point toward a connecting mechanism that transforms even a partial description into a sign that can be used to signify the season. In this way the findings relate the literary project to current scholarly interest in the relationship between metonymy and metaphor, in metaphorical conceptualizations as basic to the development of language and thought, in the central role of categorization in conceptualization, and in the relations of culture and language. As for cognitive anthropology — the ‘unnatural’ relations between the season’s variegated features in any given culture and the selected or imported attributes that characterize the literary presentations make the latter a privileged starting point for studying the relations between linguistic terms and cultural representations. Focusing on the interaction of (always already coded) natural phenomena and cultural functions, the findings illustrate the mode by which particular articulated representations (that is, artifacts, presentations) may come to dominate the collective mind for long periods of time. At the same time the findings show that such presentations are not identical with the cultural representation. They provide

“Sad autumn” and cultural representations 245

evidence of the peaceful coexistence of different representations of the ‘same’ concept in an individual’s mind, in particular, conceptualizations based in individual experiences and those that became shared representations. In this way it is possible to form links between literary research and central questions of both cognitive anthropology and cultural psychology. However, these links, whose presentation is one of the major goals of this chapter, have only been formed in retrospect. At the start I assumed — without being fully aware of the assumptions — that the representation (cognitive schema or semantic net) evoked by the linguistic sign “autumn” is one and the same in general (public and private) language usage (hence in the culture), including literary discourse. Since the sign “autumn” is used to refer to the mental representation of natural and social phenomena with which people interact successfully, I further assumed that a presentation of autumn should correspond to the same geographical and social reality that determines (another wrong assumption) its mental representation. It followed that in that case for each geographically-based cultural group there would be only one model of presentation, comprising all the characteristic features of the season (that is, those that are conspicuous and unique to it). Readers would be able to comprehend a textual presentation of ‘autumn’3 on the basis of the shared representation, regardless of their literary competence. Different presentations would then result only from the imposition of various literary norms and allowed individual preferences on the reality-based model of presentation. Similar assumptions underlie many of the best literary discussions of autumn poems. When Henry Remak, for example, proposes comparative close readings of poems by Rilke and Valery, without claiming any direct connections (such as influence or response) between the two texts, he feels obliged to justify his choice. He concludes his argument for the comparability of the two poems in the following manner: The strongest argument is the identical topic: a reaction to the most regular, inevitable, normative phenomenon of nature in moderate northern climates: the season of the year. Here: Autumn. (Remak 1976: 368)

It was the different Israeli climate that taught me how mistaken such assumptions could be. A cursory glance at Israeli autumn poems, and in particular songs, struck me with the evident discrepancy between the established literary presentation of the season, or the autumnal ‘realeme’ (i.e. the literary unit that is commonly accepted as a realistic representation) on the one hand, and the reality of Israeli autumn on the other.4 As a rule, Israeli presentations show closer affinity to Rilke’s and Valery’s “falling leaves” than to local botanic autumnal phenomena. Furthermore, Israeli readers of various cultural backgrounds have no problem comprehending the realeme, but more often than not fail to perceive an Israeli-based description of seasonal phenomena as a presentation of autumn.5

246 Ziva Ben-Porat

The ease of comprehending the experientially unfamiliar presentation made me conscious of the fact that the gap between reality and presentation has never troubled me. Not because each particular description accorded with the standard realeme, an explanation that could have kept us within the boundaries of literary studies, but because — contrary to my well remembered personal experiences of autumn — it accorded with my conceptualization of the season. Since my students shared the same comfortable ‘schizophrenic’ comprehension it became obvious that a move from the autonomous literary investigation into a study of “autumn” as a cultural concept is necessary. Such a cultural-cognitive investigation became a precondition to studying the two sets of questions in which we, always students of literature, became interested: (1) What is the role of literature in the fashioning, propagating, stabilizing and changing the cultural concept and its mental representation? (2) What is the way in which the cultural representation functions in the process of interpretation? The constitutive role of literature, the only question to be dealt with in this chapter, can be fully understood only in terms of the relationship obtaining between the three factors whose interaction shapes the various autumnal representations: Nature, Art, and Society. “Nature” is used to refer to the world of observable phenomena (geographical, meteorological, zoological, and botanical). “Art” refers to the specific category of manmade objects known as artifacts, in which representations are externalized and materialized. “Society” stands for the complex sum-total of a linguistic community’s shared representations, institutions, beliefs and behaviors. I choose to use “Society” rather than “Culture” because of the methodological exclusion of art from this complex. The decision to separate the constitutive function of art from that of culture in general is grounded in the need to distinguish between implicit and explicit forms of participation in concept construction. There is a difference between eating certain foods in season, though this activity easily explains how these foods have become attributes of the concept, and the articulated attribution of sadness to ‘autumn.’ In the former case the relationship (of the dish to the representation) is implicit; in the latter (“sad autumn”), explicit. It should, however, be noted that the substitution of “society” for “culture” is restricted to the discussions of the constitutive triangle. Elsewhere, when I discuss, for example, shared representations or linguistic signs, I use, wherever possible, the common terms “cultural” representations and “cultural” concepts respectively. The relation of poetic texts to cultural representations and the interpretive function of the latter present another terminological (and conceptual) difficulty. Their discussion calls for a clear delineation of distinct though related mental constructs: models of presentation and cultural representations. A model of presentation is — on top of being a paradigmatic articulation of its theme — a mental representation consisting of a structured repertoire common to many artistic

“Sad autumn” and cultural representations 247

presentations (poems, paintings, and so on) and internalized by a cultural group. A cultural representation consists of the elements most consistently shared by members of a cultural group in their use of and response to any activation of the concept ‘autumn’ (for instance, use of the term that refers to it; a visual rendition, a verbal description, and so on). These observable shared elements exist as a more or less tightly connected net in the minds of individuals.6 A model of presentation may either be an integral part of this net or more loosely connected to it. Individual associations or fully developed conceptualizations may be linked to it as well. But in order to become part of the cultural representation they must become shared elements through continuous presence in many presentations. As major nodes in this semantic net, namely the cultural representation, the most frequently used elements can be considered pertinent attributes. Some of them even function as metonymies of the concept that stands for this representation in the language. A cultural representation therefore is not the sum total of the components of many individual public (externalized) representations, or simply presentations, though the latter are an important source material for constructing it. Nor is it identical with any one ‘public representation’ that succeeds in capturing the mind of a community for a given time.7 Both an individual successful presentation and a large number of more or less similar presentations are still actualizations of a presentational model as well as externalizations of an ad-hoc private representation.8 The relations between a particular literary model of presentation, “sad autumn”, and the cultural representation of the season in Israel and Japan will then be discussed below in the terms of these conceptual frameworks (the constitutive triangle and the non-identity of public presentations and cultural representations).

2.

Culture specific findings: Israel

The study of the concept ‘autumn’ in Israeli culture consisted of surveying the natural and social identifiable seasonal phenomena, analysis of literary and other artistic presentations, and questionnaires designed to elicit the cultural representation. The questionnaires investigated people’s immediate associations with the term, their direct rating of relevance (or pertinence), their evaluations of labeled descriptions (such as conventional, surprising, strange), and their ability to classify and thematize various non-labeled descriptions of the season. On the basis of these investigations the current ‘autumn’ triangle in Israel’s Jewish Hebrew culture can be characterized in the following way. The present Israeli representation is constructed almost solely on the basis of Art (specifically literary presentations), with only minor contributions from Nature and Society (as defined above) to its repertoire of seasonal attributes. A close correspondence exists between the descriptive components of the dominant Israeli literary

248 Ziva Ben-Porat

(in particular lyric) articulations of the season and the most frequently listed attributes of the linguistic concept “autumn.” For example, falling leaves (effoliation!) — hardly noticeable in Israeli (particularly urban) landscape, and occurring in the country when the rainy season, Israeli “winter”, is well under way — is the most prominent attribute and the most frequent presentational element.9 There is a visible difference between the seasonal attributes found in the two corresponding (and interrelated) systems (conceptual representations and literary presentations) and the characteristic features of autumn in Israel (the presupposed reality-base for Israeli presentations). For example, in both migrating birds are associated with parting, when in reality many more birds come to winter in the country. The wagtail, which arrives in early autumn, stays through the winter, is found everywhere and is frequently mentioned in children songs, lags quite far behind the higher-level (more abstract) category “migrating birds” on the list of autumnal associations. The origin of both the dominant literary presentation and the dominant attributes of the representation is the European “sentimental” model, which foregrounds melancholy and nostalgia in association with falling leaves, migrating birds, chilling winds, sad music, and so forth. Israeli autumn (roughly from the end of August to mid-October) has some characteristic features that could well constitute Nature and Society’s contributions to the attributes of a season concept. These, however, be they seasonal flowers or the Jewish High Holidays, have little in common with the attributes of the “sentimental” (pathetic) conceptualization of the season that dominates European literature and Israeli culture. Consequently they are less frequently mentioned as seasonal associations and when mentioned they are lower on the list. Elements shared with or corresponding to the European positive “agrarian” conceptualization (such as full granaries and apple picking, pomegranates and figs), actualized in catalogue-like literal presentations, also appear to occupy less significant places in the Israeli representation. Actualizations of the highly metaphoric conceptualization of autumn as a season of painful change and violent death (“destructive autumn”) are rarely found in Israeli poetry and its particular attributes (such as rotting leaves and skeletal branches) scarcely appear on the lists of associations. Presentations that rely exclusively on Israeli reality-based autumnal materials or that actualize the agrarian European conceptualization are rarely thematized as presentations of autumn. Nor do readers classify such presentations as “autumn poems”, at least not in the same fast and automatic manner in which the same readers respond to actualizations of the “sentimental” conceptualization. Consciousness of the difference between the “reality” of the season and the cultural construct is explicitly articulated in some texts (in spite of the correspondence between the realeme and the linguistic concept). However in most cases the gap between Nature and Art is treated lightly, in an unproblematic manner.

“Sad autumn” and cultural representations 249

The literary data by itself suggests the existence of a number of representations of varying strength, or of a global representation that allows the ad-hoc construction of different conceptualizations. So does the fact that people came up with different lists of attributes when they were explicitly asked about their autumnal associations (and not about what comes to their mind when someone says “autumn”). A fuller discussion of these conclusions with their marked universalist aspirations must however wait for evidence and insights gained at the cross-cultural stage of the research. The existing gap between Israeli reality and the literary presentations of autumn could most readily be explained in terms of the history of Modern Hebrew literature. Its originators were European Jews whose individual consciousness-shaping experiences of nature accorded well with the European models of autumnal presentations. In view of the fact that the latter constituted the repertoire of literary descriptive models available to these poets, it is only natural that they transposed European ‘autumn’ to the land of Israel. While no one can contest the validity of this explanation it does leave two major questions unanswered. One has to do with the choice of model to be transposed; the other, with its continuous presence. The “sentimental” model of presentation is not the only one existing in European literature, although it is now and has been for a long time the most popular one. There is a positive ‘agrarian’ conceptualization of autumn, whose repertoire of full granaries, apple picking and thanksgiving fits the Israeli natural and social scene much better. The choice and continual dominance of the sentimental model, inexplicable in terms of the European origin hypothesis alone, can be explained from three other perspectives: 1. Historically, the “agrarian” presentation of autumn was inscribed in Modern Hebrew poetry as the estate of Gentile farmers and not as that of homeless Jewish exiles. 2. Contextually, the “sentimental” presentation is currently the dominant one in all systems of artistic productivity (at least in the West and its sphere of influence). 3. Functionally, the “sentimental” presentation has been established as the figurative representation of old age and of the psychological stage of nostalgic remembrance associated with it. Its vitality resides in the basic metaphor that underlies it. Some of these reasons for the preference of the “sentimental” model also explain its ability to “take over” not only Israeli literature but also Israeli collective — and private! — consciousness, and do so for all its apparent deviation from people’s real-life experiences. The figurative function provides one explanation. As an analogue to human decline the symbolic meaning of “autumn” is universal and not local. Its attributes must therefore comply with its figurative significance and not

250 Ziva Ben-Porat

with particular natural configurations.10 The notion of a “realeme”, a representational convention that by definition does not require full correspondence with any reality, provides another. Israeli Nature itself supplies a third. It can and has often been argued that in reality autumn does not exist in Israel. The amount of unique natural phenomena and shared individual experiences which characterize this short transition period between the dry and the rainy seasons is too small to allow for a season concept to develop “naturally.” Hence the concept “autumn” has not evolved in relation to any experienced reality, nor is it expected to represent or correspond to any. It was brought into the culture through the literary system that finds its symbolic or metaphorical manipulation useful. And as long as this figurative potential remains active there will be no need to re-conceptualize the season so as to bring it nearer to reality. It is clear that the historical and geographical explanations pull toward viewing the Israeli ‘autumn’ representation as an exceptional a-typical case. On the other hand, the theoretical explanations, within the framework of looking at language and culture as mechanisms for constructing reality as well as within the framework of investigating the notion of ‘realism’ as a set of conventions and norms, pull in the opposite direction. The Israeli case can be seen as a privileged instance that — because of its uniqueness — allows us to formulate the questions that if answered will teach us the manner in which such real-world concepts are constructed and function. In particular, a comparison of the Israeli case with autumnal representations of other cultures might allow us to look in more details into the constitutive role of literature and to answer the following questions: 1. What roles do literary models of presentation play in the construction of a realworld concept? Is the decisive role of the sentimental model in determining the seasonal attributes of “autumn” in the Israeli collective encyclopedia culturespecific and accidental or cognitively symptomatic and universal? 2. How do the particular nature of a literary model of presentation and its cultural status affect the construction, propagation, rooting, and transformation of such nature concepts? How does literary contribution compare with that of other cultural institutions and practices in relation to all of the above mentioned functions? 3. In what way, to what degree, and subject to what conditions do Art and Society-based elements or figurative significations of Nature-based elements become seasonal attributes? How does the acquisition of a figurative significance affect an attribute originating in Nature (i.e. the reality base)? 4. How does the relation obtaining between autumn’s literary model of presentation, autumn’s cultural representation (conceptual attributes), and the season’s potential repertoire of reality-based features affect the interpretation of texts that describe it?

“Sad autumn” and cultural representations

3.

Cross-cultural investigation: Findings and discussion

Hoping to find to these questions answers that are not culture-specific I turned to a comparison of the concept “autumn”, its artistic and literary articulations and its cultural functions in Israel and Japan.11 I turned to Japan for the following reasons: 1. 2. 3. 4. 5.

a different Nature-Art-Society configuration; distance from Western culture in general and Israeli literature in particular; a long normative, explicit and still vibrant tradition of autumn poetry; literacy and knowledge of literary traditions; visible similarity of literary presentations.

In both countries the investigation proceeded simultaneously along three interrelated lines: 1. establishment of the dominant literary model of presentation and the hierarchy of representational elements using texts and secondary sources; 2. establishment of the cultural representation and the hierarchy of its attributes using questionnaires to investigate: associations, recognition, and direct rating; 3. establishment of the role different representations play in the interpretation of texts, using published interpretations, class discussions and reading protocols.12 3.1 Models of presentation In spite of the obvious great social, artistic and natural differences between Japanese and Israeli cultures, the basic correspondence between models of presentation and even underlying conceptualizations is striking. In both the dominant model of presentation is the “sentimental” model.13 It is not difficult to perceive the affinity between Prevert’s prototypical poem and Miner’s translation of Fujiwara Shunzei: “As evening falls /From the moors the autumn wind /Blows chill into the heart, /And the quails raise their plaintive cry / In the deep grass of secluded Fukakusa” (Miner 1965: 425). The repertoire of presentational elements (emphasized above) is strikingly similar to the Israeli, although no one — traditionally short — Japanese poem can display all the major attributes. Both Japanese and Israeli poetry reveal a strong preference for autumn leaves in their various articulations (e.g. momiji, yellow, colored, golden, etc.) as a presentational element. Both employ the meteorological phenomena of mists, clouds and wind very often. Both prefer evening to any other part of the day, and both host migrating birds. In both all of these elements are more often than not associated (directly or indirectly) with parting, longings, being alone, sadness or general melancholy. In both cultures the gap between nature-based attributes and literary conventions may become the poetic subject matter.14 It is just as easy to see the dissimilarity. In Japanese poetry the linking of particular descriptive items to the required emotional quality — melancholy or

251

252 Ziva Ben-Porat

sadness — is usually indirect whereas Israeli Hebrew poetry is characterized by the “pathetic Fallacy.”15 The Japanese employment of the real-world repertoire assumes the analogy (in Buddhist terminology even sameness) between the human predicament and the natural world and lets the associative power of the conceptualization convey or justify the pathetic or sentimental response whenever such a response is included. The long literary tradition has created an inter-textual grid that takes on a large part of the unexpressed connections and transfers. A metaphoric relation is established between the objective description and its stated or unstated human significance, overcoming the need to articulate this relationship explicitly.16 Compared with European and Israeli poetry, natural features are rarely personified in order to convey the emotive content of the poem. Perhaps because in the West the analogy between nature and man is only an analogy — even if it conforms to basic conceptual metaphors — and the human predicament is viewed as unique, Western (including Hebrew) poetry went in the opposite direction.17 Japanese and Hebrew presentations differ in two further respects: the range and category level of real-world presentational components. The dominant Japanese model of presentation contains a much larger variety of flora and fauna than its Israeli counterpart. The reason is that everything natural can be and often is connected to the sadness associated with the transience of all things by virtue of both the underlying Buddhist worldview and the inter-textual grid (the Japanese connection by association). Cherry blossom signifies transience as easily as withering leaves. Consequently there is no essential distinction in Japanese poetics between elements that in the West — and in Israel — are always associated with positive feelings (for instance a flower in bloom) and those that are negatively associated (such as a withering flower). This major difference in looking at the world decrees that while additions to the repertoire are governed by strongly restrictive rules of selection in Israel, in Japan, in principle, many more natural phenomena can be employed in actualizations of the ‘sentimental’ model.18 Ultimately the Israeli repertoire is relatively closed and stable; the Japanese is openended. The limited range of presentational elements in Israeli poetry functions as an adaptive mechanism for the imported repertoire. It reinforces connections by repetition and by undisturbed and easy mappings: the recurrent elements always carry the appropriate clusters of associations. To these two distinctive features (range and cultural marking) one must add the level of categorization. The various Nature-based presentational elements of the Japanese model are usually referred to by their common names such as momiji and susuki, most of which could be described as subordinate level categories.19 In contradistinction the Hebrew presentation shows a marked preference for “tree”, “fruit”, “leaves”, “grass”, and other super-ordinate or basic level categories. It is not difficult to explain this difference. An imported presentation cannot rely on subordinate categories because the latter require familiarity of the type

“Sad autumn” and cultural representations 253

generated by personal interaction. In our modern urban culture even ‘natives’ know by name only a small number of trees, flowers, or different insects that abound in the country in which they live. The few that are known to everybody are culturally coded. They are learned and used in conjunction with other phenomena in the service of particular cultural functions. Constructing a season concept is one such function. Names of trees that can neither invoke an image nor activate a schema of which they partake cannot connote the associations for which they are mentioned.20 The use of basic or higher levels of categorization, characteristic of Hebrew poetry, is crucial for the acculturation of a foreign cultural representation; which, in its turn, is a pre-requisite for the propagation and preservation of the originally European literary presentation. 3.2 Cultural representations The critical construction of the Japanese cultural representation of autumn has been subject to the same assumptions and methods as that of the Israeli.21 It resulted in a list of concepts that are interconnected and presumably form the Japanese representation of autumn. The findings support the claim that even a most successful presentation — or its model — does not replicate the cultural concept. According to the resulting list of attributes, the Japanese cultural representation is more independent of the dominant literary presentation, more complex and quite different from the Israeli. This can be seen in Table 1, which compares the relative strengths of the leading attributes in the two systems.22 As the table shows, the Japanese list of major attributes differs from the Israeli list on three counts. With the exception of “autumn leaves”, elements shared with the literary presentation occupy lower places and show lower consistency across the various experimental scales. They are intermixed with various natural phenomena [Nature based attributes], social activities, proverbs, culinary traditions, and so on [Society based attributes] of equal pertinence. And they display lower level categories in relatively high places. The table does not show that the Japanese list is also much longer (18 items should be added within this range). The shorter Israeli list (the table contains everything over 10%) divides into four distinct groups. Autumn leaves is the uncontested most pertinent attribute. Wind, chill and rain — components of the sentimental model of presentation that, at the same time, mark the difference between the dry and the rainy seasons in Israel — form the second group. These attributes, strong according to most parameters, by themselves activate “winter” rather than “autumn” (hence the low consistency). The next, nominally weaker group of attributes shared with the sentimental presentation (evening, sadness, loneliness and migrating birds) is characterized by high consistency across the different parameters. Finally, there is a much weaker group of Israeli Nature-based components (squill and wagtail), that in spite of their

254 Ziva Ben-Porat

Table 1 Japanese attributes

Israeli attributes

Attribute

Unified % of points

Consistency

Attribute

Unified % of points

Consistency

Autumn leaves Persimmon Warm clothes Potatoes Sanma (fish) Harvest Moon Evening Rice Sad feelings Chill + Wind

100% 26% 25% 25% 22% 21 21% 21% 19% 17.5% 16%

High Average Low High Low Low Average − Average + Low Average + High

Autumn leaves

100%

High

[Warm clothes

35%

Children]

[Harvest]

Rare

High

Evening

30%

High

[Rain] Ripeness Wild geese

5% 16% rare

Low None

Sad feelings Wind Chill Rain [Ripeness] Migrating birds Squill Wagtail Citrus Fruit

32% 74% 51% 69% Under 1% 29% 18% 12% 10%

High Low Low Low High average Low Low Low

low place on the unified list are unequivocally and immediately associated with autumn. In accordance with the Israeli presentational repertoire the majority of attributes are basic or super-ordinate categories. Content, hierarchy and internal organization prove that while the Israeli representation is dominated by the literary sentimental model of presentation, the Japanese representation reflects all the semiotic systems involved in its construction without privileging or excluding the literary contribution. But the common supremacy of “autumn leaves”, as well as the presence of Nature and Society based elements on the Israeli list of attributes, prove that the Israeli near-identity of presentation and representation is neither an aberration nor the rule, but a legitimate option that can happen under particular circumstances. The cross-cultural comparison of ‘autumn’ offers many more insights into the composition and inner organization of cultural representations as well as to the mode of their social construction. Let me conclude by summarizing them, starting with the obvious and moving on to the less obvious and in parts controversial. • Cultural representations are culture-specific (even when they are imported or strongly influenced by external representations).

“Sad autumn” and cultural representations 255

• Power relations within the culture are expressed through the hierarchy of attributes characterizing its representations (even when a concept refers to a Nature-based representation). • Inter-cultural translation is problematic and never completely adequate. • Cultures may prefer different cognitive options for comprehending and representing the real world, but the options (for example, choice of different level categorizations for the constitution of attributes) are not culture specific. • Cultural representations are not identical with even the most successful presentations. • Major components of successful presentations are attributed to the representation, but their positions in the hierarchical organization of the two relevant mental constructs (that is, representation and model of presentation) may differ. • Because of discourse-imposed constraints (such as medium, genre, theme) the repertoire of attributes of any presentation is more limited and homogeneous than that of the related cultural representation. • In their consistent inconsistency across the different experimental paradigms that have been used to construct the representation, attributes of cultural representations bear evidence to the existence of various related conceptualizations or sub-groupings.

Epilogue A culture, according to Sperber, “is made up, first and foremost, of contagious ideas. It is made up also of all the productions (writings, artworks, tools, etc.) the presence of which in the shared environment of a human group permits the propagation of ideas. To explain culture, then, is to explain why and how some ideas happen to be contagious” (1996: 1). The cross-cultural comparison of autumn poems in relation to the cultural concept ‘autumn’ presented in this paper can be seen as a response to this challenge. Let me recapitulate the themes of this chapter in terms of Sperber’s epidemiology of representations, in the form of a story — actually two stories — of an epidemic — the successful arrival and somewhat problematic stay of ‘sad autumn’. In the late Middle Ages a particular idea of ‘autumn’, for various explicable reasons, came to dominate the production of autumn poems in Europe. This particular idea, the “sentimental” conceptualization encoded in ‘sad autumn’, had been communicated repeatedly and became a cultural representation. In itself this representation was a model of presentation, one model of an autumn poem. It was also a component in the European cultural representation of autumn. European immigrants brought this model of presentation to Israel. Once it began to be communicated it spread and settled in the culture. There were many reasons for its success. The carriers, major poets, were in a position to spread the model. The

256 Ziva Ben-Porat

recipients had a natural or rather historical tendency to contract it, either because of a shared cultural background or because of the prestige of the carriers’ group. Furthermore, the recipients had no antibodies. No previous conceptualization existed to ward ‘sad autumn’ off. But, most important of all, this invader turned out to be an easy and useful guest. Easy, because it fitted well into the space it now occupied, a space shaped by the basic metaphors connecting human life and the world. Useful, because it provided the hosts with a needed literary and cognitive repertoire. Under the circumstances it was only natural that the cultural representation would remain for a long time under the dominion of this model of presentation. But this is not the end of the story. Once a season concept exists it attracts different objects, events and associations that have enough seasonal relevance. The semantic net that constitutes the representation ‘autumn’ cannot exclude such newcomers if they are supported by other cultural systems and by the shared experiences of many individuals. As a result the representation has become to some extent distinct from the model of presentation. Or, in the terms of the “epidemic” metaphor, interacting with its new environment, internal and external, the invader mutates. The dominant idea in Israel is still that of the literary ‘sad autumn’ but today’s representation includes more (and different) components than the original imported aggregate. The Japanese story is different. From early historical time, Japanese poetry tends “to use images at once as concrete details and for metaphorical implication (rather than representation)” (Miner 1965: 423). The particular Japanese ‘sad autumn’ model of presentation developed naturally within the constraints of traditional Japanese poetics. Parallel to the specific poetic development, Japanese culture has cultivated many links — in the form of social and artistic activities — between human life, art, and nature in relation to the changing seasons.23 Each cultural field produced through repetition and communication seasonal attributes. They have all been attracted to the space that the scientific division of the year opened up and together formed the semantic net of the rich and encompassing Japanese cultural representation. The specific contributions of the literary model of presentation blended in with the rest. The different stories explain many of the findings, including the surprising ability of Israelis to classify and thematize Japanese autumn poems as well or even better than the Japanese participants. They explain some of the inevitable misunderstandings and the different interpretations. They could not but underline the limited cross-cultural translatability. Constructed and situated differently in each culture ‘sad autumn’ must be signified differently by Japanese and Israelis. On the other hand, they underline the similarities of cognitive-cultural interaction as well. The common underlying basic metaphors, the common transformation of metaphors into metonymies and vice versa, the ability to function with limited comprehension and the creative culturally coded filling of gaps — all these make the ‘sad



“Sad autumn” and cultural representations 257

autumn’ representation (and its French actualization that was our starting point) accessible even across very thick cultural borders.

Notes * This work could not have been achieved without the generosity of the University of Tokyo and the time I spent there as a visiting professor in autumn 1998. I am indebted especially to my friends and colleagues: professor Koji Kawamoto, professor Toshiko Ellis, professor Takayo Kano, and Dr. Hideyuki Kanazawa; and to my students — my research team — Midori Saito, Kumiko Nagai, Chiaki Kojima, and Milan Vidakovic; and to Reiko Nakagami, another voluntary assistant. I owe a special debt to professor Earl Miner for early interest and encouragement, not to mention information. 1. According to Curtius who quotes Quintilian, topoi “are intellectual themes suitable for development and modification (…) Storehouses of trains of thought.” When rhetoric penetrated into all literary genres “topoi become cliches, which can be used in any form of literature, they spread to all spheres of life with which literature deals and to which it gives form” (1979: 70). 2. See my articles in the References section for detailed presentations of my work on this subject. 3. ‘Autumn’ indicates the representation; “autumn” the concept, the linguistic sign; autumn — the season. 4. On the notion of realeme and the problematics of reference to the real world and its modeling see Even-Zohar (1980: 65–74). 5. For example, songs about full granaries and wineries are categorized as “holidays songs” and will not be played on Radio programs entitled “Autumn Songs.” Another example can be found in Ben-Porat (1988). 6. On the similarity of cultural representations to all other representations, see Shore (1996: 47). For many discussions of the relations between cultural models and language, focusing on metaphors, see Gibbs and Steen (1999). 7. This is the only point where I cannot use Sperber’s terminology (and conceptualization) though his epidemiological model is to my mind the best way to comprehend the acculturation and long domination of the “sentimental” model of presentation in Hebrew literature and Israeli culture. Sperber distinguishes between many mental representations and the fewer public representations, that he defines as materialized (my externalized) mental representations. He then defines cultural representations as exceptionally successful public representations (1996: 25). 8. An “ad-hoc” representation is the particular configuration of the semantic net active in a sustained manner by an individual each time the representation is activated in a different context; practically always. 9. Although the stability of a cultural representation is only relative, and the Israeli representation seems to be opening up to local real-world components, the status of falling leaves as the strongest attribute is still secure. See a fuller discussion in the section “Cross-cultural findings: representations”. 10. This part of the work was concluded and published before I became aware of Lakoff and Johnson’s work on conceptual and basic metaphors, and before the publication of Lakoff and Turner’s More than Cool Reason (1989). I’ll be making full use of the explanatory strength of their

258 Ziva Ben-Porat

basic metaphors (life as the daily cycle; life as the annual cycle; and life as a plant) and their examples later on. 11. I first turned to Japan in 1991 during an ICLA congress. See Ben-Porat (1995: 90–97). 12. Because of the differences, it was not possible to simply replicate all questionnaires and reading experiments. Some of them had to be specifically redesigned in order to accord with their Hebrew counterparts 13. It is possible to find traces, and sometimes full actualizations, of the positive “agrarian” and the negative “destructive” models as well. But because of their present peripheral position and for lack of space these can neither be illustrated nor discussed in this paper. 14. In his book, Kawamoto cites many examples of such poetic connections and of the poets’ consciousness of their conventionality. He shows how already in the Shinkokinshu (1205) The Retired Emperor Gotoba praises a spring evening and concludes with the question: “Whatever made me think/Evenings were for autumn?” (2000: 29). This exposure of the “fictionality” of poetic attributes parallels the Israeli acknowledgement of the gap between the reality of the season and the literary model of presentation. 15. The prototypical poetic presentation of autumn since the late Middle-ages has become one in which clouds are weeping, the sun is tired, the field feels abandoned and the world is as sad as the person who perceives it. As Ruskin defined it (and Wimsat made popular) this emotional transfer from the human to the natural constitutes the “Pathetic Fallacy.” 16. See, as an example of such metaphorical conceptualization of haiku, the reading of Basho’s “A clam separates lid from flesh as autumn departs”, (Matsuo 1996: 133) offered by Hiraga (1999:476). “Passing autumn” (yuku aki) does not carry upon itself the weight of parting as death, that Hiraga attributes to the dying crab in the way the “leaves that are falling in the heart” do in Hebrew poetry. See Ben-Porat (1986) and (1991) for many examples. 17. General conceptual metaphors are “part of the way the members of a culture have of conceptualizing their experience” (Lakoff and Turner 1989: 9); Basic metaphors “are used automatically, and largely unconsciously. They are not created by the authors who use them” (Ibid: 8). In spite of the emphasis that Lakoff and Turner put on the culture-dependency of basic metaphors, all their relevant examples (such as: lifetime is a day; lifetime is a year; people are plants; autumn is old age; yellow leaf is old age, etc.) are as common and as basic in both Japanese and Israeli cultures. Hiraga (1999: 477) brings some idiomatic evidence to the centrality of “autumn is decline” in Japanese. 18. “Openness” in principle does not mean that everything goes. Basho, for example, distinguished between inept use of seasonal words in poems that “did not capture the seasonal atmosphere”, and the successful poem that captured the sense of the season without using a season word. See Haruo Shirane (1998: 193–4). 19. I am, of course, alluding to the work of Eleanor Rosch (1976 in particular), but must allow myself some laxity with regard to the exact classificatory judgments I make. The reason is the use of English translations of Japanese texts in some of the experiments as well as some disagreements among native speakers. For example, Japanese pampas grass is definitely a subordinate category in English. So is Japanese maple. “Grass” and “maple” would be the respective basic categories, “weeds” and “trees” the super-ordinate. But, if they have no sub-categories of their own, susuki and momiji might be considered basic level categories by English speaking linguists. In spite of this difficulty I find the distinction very useful in depicting the dissimilarities between Japanese and Israeli representations. The importance of using the lower level categorization became clear when ashi (ditch reed) had been used instead of susuki (pampas grass) in one questionnaire and very few participants judged it to be relevant to autumn.

“Sad autumn” and cultural representations 259

20. The use of sub-ordinate categories may prove problematic for Japanese as well. Kawamoto discusses a famous poem by Basho in which the season word is mukuge (rose mallow). In contemporary urban Japanese culture very few people recognize the plant. Those that do associate it with summer unless they carry Higginson’s haiku dictionary in their minds. Only literary scholars remember that this plant “is a staple theme in the classical tradition carrying implications of precarious life” (1999: 715–6). I am indebted to Reiko Nakagami for testing her friends and family on this poem. Another interesting test case was the “reading” of paintings. When asked to determine the season to which a painting belongs, Japanese participants looked for flowers coded as attributes. Only when they failed to identify chrysanthemums or pampas grass, they reacted like Europeans and classified the painting as indicating either summer or spring. 21. The analysis of textual presentations, which yielded a list of potential seasonal attributes, was complemented by and compared with other lists. Those lists have been computed from studies of: – –

–

Associations: the personal associations of 60 native Japanese of various ages and backgrounds, triggered by the term “autumn” (aki); Relevance judgments: direct rating of relevance the same participants made on a list of 27 potential attributes (selected from the best known components of the literary presentations and additional observable seasonal phenomena suggested by the Japanese members of the research team); Reported seasonal phenomena: attributes mentioned by the participants when asked to report their personal experiences and memories of the season.

The four lists have been compared with each other and with the responses of 36 participants to untitled presentations of the concept, to partial presentations and to individual presentational elements (in terms of ability to identify the concept, reaction time, confidence in identification, and so on). 22. Strength in this table is expressed as a percentage of the maximal number of points an attribute could have scored if it came first on three scales: direct rating, free associations, and observations. Since placement can be inconsistent, the percentage shows an average score and the level of consistency is also noted in the table. Consistency judgments also reflect the relationship of the average score to other parameters (e.g. metonymicality or perceptibility). 23. See for a panoramic view of these links, Anesaki Masaharu (1984). Among its many illustrations it contains, for example, a painting depicting people “enjoying the full moon” (the eighth month) by Moronobu (146).

References Ben-Porat, Z. 1986. “Represented reality and literary models”. Poetics Today 7 (1): 29–58. Ben-Porat, Z. 1988. “Autumn poems and literary impressionism: Conceptualization, thematization and classification”. New Comparison 6: 158–175. Ben-Porat, Z. 1990. “Autumn songs as an example of the relationship obtaining between lyrical poetry and popular songs”. Proceedings of the 3rd Congress of Semiotic Studies, Palermo 1984, M. Herzfeld (ed.), Berlin: De Greuyter. Ben-Porat, Z. 1991a. Autumn in Hebrew Poetry. Tel Aviv: HaKibbutz Ha-Meukhad, [in Hebrew]. Ben-Porat, Z. 1991b. “Two way pragmatics: From world to text and back”. Literary Pragmatics, R. Sell (ed.), 142–163. London: Routledge.



260 Ziva Ben-Porat

Ben-Porat, Z. 1994. “Ideology as a dominant factor in seasonal representation”. In Literatures in Pluralistic Cultural Context, Yue Daiyun and Zhang Tiefu (eds.), 216–226. Beijing: The University of Peking) [in Chinese translation] Ben-Porat, Z. 1995. “Universals of conceptualization and presentation? A comparative look at Japanese and European autumn poems”. In The Force of Vision 6: Proceedings of the XIV congress of ICLA, Tokyo, 1991, K. Kawamoto, Y.Heh-Hsiang, Y. Ohsawa (eds.), 90–97. Tokyo: University of Tokyo Press. Curtius, E. R. 1979 [1948]. European Literature and the Latin Middle Ages. W. R. Trask (tr.). London and Henley: Routledge and Kegan Paul. Even-Zohar, I. 1980. “Constraints of realeme insertability in narrative”. Poetics Today 1 (3):65–74. Gibbs R. and Steen G. (eds.) 1999. Metaphor in Cognitive Linguistics. Amsterdam: J. Benjamins. Hiraga M. K. 1999. “‘Blending’ and an interpretation of haiku: A cognitive approach.” Poetics Today 20 (3): 461–481. Kawamoto K. 1999. “The use and disuse of tradition in Basho’s Haiku and Imagist Poetry.” Poetics Today 20 (4): 715–6. Kawamoto K. 2000. The Poetics of Japanese Verse: Imagery, Structure, Meter. Stephen Collington, Kevin Kollins and Gustav Heldt (trs.). Foreword by Haruo Shirane. Tokyo: University of Tokyo Press. Lakoff, G. 1987. Women, Fire, and Dangerous Things: What categories Reveal about the Mind. Chicago and London: University of Chicago Press. Lakoff, G. and Johnson M. 1980. Metaphors We Live By. Chicago and London: University of Chicago Press. Lakoff, G. and Turner M. 1989. More than Cool Reason. Chicago and London: University of Chicago Press. Masaharu, A. 1984 [1932]. Art, Life, and Nature in Japan. Rutland, Vermont & Tokyo, Japan: Charles E. Tuttle. Miner, E. 1965. “Japanese poetry.” In Encyclopedia for Poetry and Poetics, Preminger, Warnke, Hardison (eds.), 423–431. Princeton, N. J.: Princeton University Press. Remak, H. 1976. “Rilke and Valéry on autumn: A comparative explication de texts”. In Herkommen und Erneuerung: Essays fur Oskar Seidlin, G. Gillespie and E. Lohner (eds), 365–376. Tubingen: Max Niemeyer. Rosch, E. 1976. “Classification of real-world objects: Origins and representations in cognition.” In Thinking: Reading in Cognitive Science, P. J. Johnson-Laird and P. C. Wason (eds), 212–222. Cambridge: Cambridge University Press. Shirane, H. 1998. Traces of Dreams: Landscape, Cultural Memory, and the Poetry of Basho. Stanford: Stanford University Press. Shore, B. 1996. Culture in Mind: Cognition, Culture, and the Problem of Meaning. New York: Oxford University Press. Sperber, D. 1996. Explaining Culture: A Naturalistic Approach. Oxford: Blackwell.





The fragmentation of the media audience* Kees van Rees and Koen van Eijck Tilburg University

Media use in leisure time can nowadays be regarded as a cultural practice in its own right; it consumes, on average, the greatest amount of leisure time of most individuals. Media providers struggle to capture the attention of potential consumers. Hightech measurements of medium reach are treated as rates for viewing, listening, or selling, that is, as proxies for estimating effects of programming and evaluating success (McQuail 1997: 43). All over the world network series which were heavily promoted and earned strong reviews have been known to be abruptly canceled after their second episode because of a precipitous decline in ratings (cf. Weinraub 2000). In the US alone some 85 billion US dollars are yearly spent on advertising, for the most part on the traditional broadcast and print media. Yet much of the research measuring advertising effectiveness is methodological skating on very thin ice. Its reliability and validity are unclear. Indeed, audience research is still in its infancy. So far not much is known about media effects on media users. Much research was based on implicit but highly questionable assumptions about audience receptivity and audience processing of media exposure. In recent decades, various approaches in communications research have made obsolete the theory of the almighty media (‘bullet theory’). Though the opponents of this theory have often tended to construct it into a straw man, current media research has made a move in the right direction by including in its models the role of the active audience (Croteau and Hoynes 2000: 237ff.). Some progress has been made with measuring the size of the audience of specific media. Changes in technology force researchers to reconsider their views of audience analysis since people are increasingly becoming multi-media users. Due to the broadening of the range of media available to potential users, the notion of audience appears even more difficult to specify than before. In this chapter, we use time budget data from the Dutch 1975 and 1995 surveys to analyze the relationship between television viewing and reading behavior, and the changes in this relationship over time. The audience we focus on is that of newspapers, consumer magazines, and television programs. Besides background characteristics such as age, education, gender and time budget, we include book reading as a covariate.

262 Kees van Rees and Koen van Eijck

Unlike previous research on the competition between television viewing and reading (Knulst and Kalmijn 1988; Knulst and Kraaykamp 1996, 1998), we also take into account the nature of the programs and reading materials. To analyze media use behavior we categorize print media and television programs on the basis of their content. This permits hypothesis testing about the relationship between types of reading and television viewing in terms of people’s media orientation and their media repertoire. The analysis of content-related behavior brings to light the varying patterns of media selection by different types of media users as well as changes over time in these patterns. This considerably increases our insight into the concept of ‘the active audience’. The quantitative analysis of a representative sample of the Dutch population confirms recent American views (Croteau and Hoynes 2000; Neuman 1991; Turow 1997) on the fragmentation of the media audience. The media audience is segmented according to each subgroup’s media orientation. This proposition is examined by answering the dual empirical question: Which television viewing behavior is characteristic of specific reader types and how did the relationship between television viewing and reading change over time? The five mutually exclusive reader types that were identified and reported in Van Eijck and Van Rees (1999) provide the starting point for the analysis of the relationship between reading and television viewing behavior. In the next section, we briefly discuss the Dutch media supply in the period 1975–1995 (section 2). In outlining the theoretical background underlying our analysis of reading and television viewing, we suggest that kinds of media use are each related to a specific kind of media orientation (McQuail 1997: 69; Weibull 1985:143ff.). Print media and television programs are differentiated by their specific content. This strategy enables us to differentiate between types of media users according to their distinct media repertoires. The summary of our earlier analysis that resulted in the five reader types will illustrate this point (section 3). Hypothesized relationships between kinds of reading and television viewing behavior are summarized in section 4. The methods section describes the data set and the analytic approach; it also specifies how the variables used to characterize reading and viewing behavior were constructed (section 5). This provides the framework for our discussion of the connection between kinds of television viewing behavior and the allocation of viewers to one of the reader types as well as changes in this relationship between 1975–1995 (section 6). We conclude by briefly considering ways to improve the analyses in follow-up research.

The fragmentation of the media audience 263

1.

Television and print media as complementary markets

1.1 The print media market Developments on the newspaper market can be summarized with the help of two concepts: concentration of producers and title reduction. Over the entire period during which time budget surveys were held, a strong concentration of suppliers occurred. Large firms were formed. Since the mid-1980s, this process of concentration, or merging of publishing companies, has been accelerated. In 1997 the four largest newspaper publishers had a market share of over 90%. In 1992, there were still nine holdings sharing this percentage. Publishers have tried to halt the decrease in number of subscribers to their newspapers and magazines. Since the late 1960s, the growing circulation of newspapers has failed to keep pace with the increase in the number of households. Between 1980 and 1996, coverage, defined as the ratio of total circulation newspapers/number of households, decreased from 0.90 to 0.74. (The decline started between 1970 and 1975: in 1970 there were 124 papers per 100 households.) In the period under consideration, the notion of supplier concentration is likewise applicable to the market for consumer magazines. In contrast to the newspaper market, however, there is no question of title reduction. Magazine publishers have tried to compensate for the loss of subscribers to successful titles by developing new titles on more specialized sectors of information, such as specific aspects of lifestyle (separate magazines for interior design, gardening, cooking, new hobbies, and new trends). Circulation of opinion magazines and family magazines has suffered from both television programming and newspapers. 1.2 The television market as competitor of the print media market Though the conditions for broadcasters are different from those of print publishers, television broadcasters have also been confronted with growing competition: the long period of public broadcasting oligopoly has been turned into a battleground where public broadcasters’ continuity is threatened both by successful private competitors and unpredictable changes in media policy that erode their position. It should be noted that horizontal integration (companies involved in both print and television) is less far developed in the Netherlands than, for example, in the USA; besides, since the end of the 1980s the private television market (both the production and distribution) has increasingly come under the surveillance of non-Dutch mega-companies, although the remnants of the public broadcasting system, operating on the basis of a yearly fee paid by any household owning a radio and televison set, are still intact. In 1975, Dutch television viewers had at their disposal two public channels. Seven broadcasting corporations, remnants of the former pillarized society, were

264 Kees van Rees and Koen van Eijck

licensed to take care of the programming. These broadcasting organizations differ in several respects: ideological orientation; status (dependent on their membership figure and decisive for the amount of broadcasting time); and mean viewer rating. Due to cabeling, which was initiated by big cities at the beginning of the 1980s and almost completed around 1990, television viewers in 1995 could choose programs from no less than seven Dutch channels, three public and four private (the ‘commercial’ broadcasters) and a range of foreign channels from Belgium, Germany, Great Britain, France and the US. Hence, judging by its scope, the television channel supply of 1995 is hardly comparable with that of 1975.1 As for its content structure, however, the programme range provided in 1995 was not fundamentally different from that of 1975. The increase in the number of networks did not lead to a proportional widening of the existing range of programme types, for the simple reason that this cannot be expanded without limitation. What has been widened, though, is the range of infotainment programs that cut across sub-types (for instance reality TV or ‘docudrama’). In line with our earlier differentiation of print media (Van Eijck and Van Rees 1999), we can differentiate television into two major types: informative and entertainment programs. Again, as with print media, several sub-genres, or even sub-sub-genres, might be distinguished, to which we will return below. Generally, judging by its content, the 1995 televison supply is comparable in terms of program categories to the 1975 supply, with the proviso that it provides ‘more of the same’. Assuming a content dimension with information and entertainment as poles, in 1995 a larger number of programs have to be located on this axis, as an increasing number of broadcast conglomerates are competing for media attention of similar groups through products located, for the most part, near the entertainment pole. While in the 1960s watching television still had a particular lifestyle aura, as of the 1970s it had become a common medium. The 1975 television penetration rate provides a clear indication of this status. However, judging by the 1970s viewing figures, the medium was at that time primarily favored by the elderly and the less educated people from the lower classes (NOS 1976: 65). This has also changed in the 1990s: it was only natural that, irrespective of education, generations that had grown up with television integrated this medium in their leisure time spending. In communication studies, a difference in function is often made between print media and broadcast media. The former (newspapers and magazines) are said to primarily serve information and idea functions for the audience, while the latter (radio, TV, film and videocassettes, and cable television) primarily serve entertainment functions (McQuail 1987; Picard 1989). Without making this global distinction between ‘information’ and ‘entertainment’ absolute, and aware that these are audience-related concepts, the meaning of which also varies with the medium at issue, Van Eijck and Van Rees (1999) used the dominant orientation of print media towards the information or entertainment pole as a criterion to distinguish

The fragmentation of the media audience 265

categories within the print media. Within two kinds of print media, newspapers and consumer magazines, three categories were distinguished according to their focus on information or entertainment or something in between (Van Eijck and Van Rees 1999: 130–1). We hold that for television viewing a similar distinction can be made among categories of television programs — part of them focusing on information, while others focus mainly on entertainment. In the light of our research question such a distinction makes sense. We presume that people differ in their media use according to a dominant orientation to serious or light information, or to serious or light entertainment. Our previous analysis of patterns of print media use confirmed this view (Van Eijck and Van Rees 1999). Therefore we developed a similar categorization for television programming. 1.3 Categories of television programs By the Broadcasting Act of 1967, (public) broadcasters need to supply a so-called complete programme which “in reasonable proportions comprises items of cultural, informative, educational and entertaining nature” (Saarloos 1989a: C8.5–1). It is not until the Media Act of 1987 that this prescription is refined by specifying minimum percentages for each category: culture (20%), information (25%), entertainment (25%), and education (5%). Under neither of the Acts, however, were the public broadcasters ever seriously evaluated by an exclusive categorization of their programs. The Board of Overseers never sanctioned the creative way broadcasters kept their programming books, for example, by classifying a programme both under culture and information. An ambiguous category like drama which, just like sports but in contradistinction to ‘information’, had rating figures way beyond average, could serve as a dummy to also meet the requirement for ‘culture’ programming. Hence, programme categorization by broadcasters themselves is not suitable for our research. Like print media, broadcasting companies offer a specific mixture of information and entertainment. Indeed, in 1975, the broadcasters, rather than the two available channels, might be usefully compared with print media such as newspapers. Depending on the evening of the week and the licensee that was entitled to broadcast its mixture of information and entertainment, the supply on a channel more closely resembles that of a popular newspaper, a large regional newspaper, or a small newspaper of markedly protestant signature. In 1995, because public and private broadcasters are in fierce competition with each other, a comparison of broadcasters with newspapers still holds for the public companies, though daily news is procured by an independent organization (the Dutch Broadcasting Foundation). For the private broadcasters, the comparison does not apply, if only because they are not (and never were) subject to the ‘complete programme’ requirement. By focusing on light entertainment (possibly for specific age groups),

266 Kees van Rees and Koen van Eijck

these broadcasters might be compared to popular newspapers or a mixture of general and special interest magazines. This analogy between broadcasters and the print media is heuristically useful in that it permits us to clarify expectations on the viewing behavior of reader types. To answer our research questions, a bi-partition of television programs in two kinds suffices: information and entertainment. Each has two categories: serious and light.

2.

Media orientation and media use

Media use in leisure time can be regarded as indicative of people’s lifestyles. Audience research might profit from paying due attention to developments in cultural sociology, as research aimed at describing and explaining differences in cultural participation has gained a position of its own: it provides a better insight into social, economic and cultural determinants of taste and the factors affecting the perception and evaluation of cultural practices. An important theoretical basis for this research is provided by Bourdieu’s theory of cultural capital, as developed in the US and the Netherlands (DiMaggio 1982; Ganzeboom 1989, 1987; Lamont 1992). Re-analysis of Time Budget Surveys (TBS) data in terms of reader types and their television viewing behavior provides the basis to illustrate how cultural stratification research, usually aimed at explaining differences in cultural participation, can sensibly be brought to bear on audience research. Part of this discussion can be traced back to Bourdieu’s homology thesis. Developed with respect to French society in the 1960s and 1970s, this thesis argues that the structure of the space of life-styles is homologous to that of the space of social positions (Bourdieu 1979). Over the past few decades this has expanded into the analysis of the dimensions by which cultural goods as well as their users can be suitably classified (for instance ‘highbrow’ and ‘lowbrow’, legitimate, medium or popular) (Van Eijck 1999b). American research (Peterson 1992) has recently found that members of high occupational status groups are omnivorous in their cultural participation in that they tend to combine high and low categories of cultural practices. Low status groups are referred to as univores, since they are said to be more one-sided in their cultural practices and their choice of cultural goods by focusing rather exclusively on the less legitimate genres. In a general way ‘omnivorousness’ refers to the transgression or erosion of symbolic boundaries between categories of cultural goods and practices. One assumption underlying its use is the ability to clearly assess the existence of these boundaries. From the perspective of those believing in the existence of strong boundaries, a practice may appear as a transgression, whereas from another point of view, for those who hold that this cultural practice is losing its symbolic value or ritual power, it may be termed boundary erosion. On the basis of cross-sectional surveys in various Western countries it has been

The fragmentation of the media audience 267

concluded that in the 1990s individuals are more likely to combine cultural items that differ in degree of legitimacy (Van Eijck 1999a; Wynne and O’Connor 1998). Whether interpreted as transgression or erosion, such a trend would point to the necessity to reconsider Bourdieu’s homology thesis. Just now we explained the differentiation of print media and television programs by referring to the extent to which they each focus on kinds of information or entertainment or both. A fundamental assumption underlying the present research is that people differ in their media use according to their dominant orientation with regards to a specific content, that is, toward serious or light information and legitimate or popular entertainment. It is this idea we seek to test with our research. Dominant modes of media orientation can be differentiated in terms of both medium type (print, broadcast) and kind of content provided by a particular medium. Each of these modes or ‘sets’ is hypothesized to be the product of former habits of media use and to induce a specific media use in the near future (cf. Weibull 1992: 270–7). This conceptualization can profitably be related to Bourdieu’s theory of (economic and cultural) capital and his view of the habitus as a (class-bound) set of schemas of perception, appreciation and action (Bourdieu 1979; Rosengren 1996: 30–46). In this perspective it is relevant to note that our analysis enables us to relate a particular media orientation to background characteristics such as education, age and gender. In a volume honoring the scholarship of Elrud Ibsch it is not accidental that ‘book reading’ (the extent to which respondents read books for entertainment) is included as a covariate. The relation between television viewing and reading can be viewed from different media orientation perspectives. Underlying the first view is the assumption that the audience makes a selection on the basis of its preference for a specific medium. If the distinction between print media and broadcast media is relevant to the audience’s selection, implying that it is the medium which, for the most part, determines people’s willingness to take note of part of the supply, we may expect to find both a group of avid readers and a group of fervent television viewers, irrespective of the content of the product. An alternative approach proceeds on the assumption that media users decide on the basis of a dominant set towards information or entertainment, irrespective of the medium which provides it. The differentiation in the use of print media and television programs is then explained by referring to the extent to which media users are focusing on information or entertainment. A third approach, which we support and develop below, combines and qualifies the two former approaches by assuming that audiences can be segmented according to their dominant preference for information or entertainment and that during this period, in which technological innovations profoundly transformed the media supply, audience segments see their preferences fulfilled more and more by either serious information (print media more exclusively in 1975 than in 1995) or light entertainment (television more exclusively in 1995 than in 1975).

268 Kees van Rees and Koen van Eijck

Depending on the research question, each of the three approaches may provide a sensible starting-point. We briefly consider the consequences implied by each and will argue why we adopt the third, most complex, approach. In the first case (selection of the medium predominates), serious print media are likely to be favored by higher status groups and television by the middle and lower status groups. We thereby lose the opportunity to account for the dynamics of the media industry which, through the considerable increase in supply, nourished the trend towards a more hybrid, less exclusive, lifestyle — read: media use — also among the higher educated. In the second case (selection of information or entertainment is predominant), we will find within each type of medium use a differentiation according to status (notably education, gender and age). The higher status groups prefer, besides quality newspapers and opinion magazines, serious television programs, whereas the lower status groups are likely to prefer (light) entertainment through popular newspapers and gossip magazines, and, increasingly, television. This presumes a rather strong automatism in people’s combining kinds of media; in addition it fails to take into account that while, from the point of view of content, media may offer categories from a similar type, print and broadcast presuppose quite different types of cognitive processing. Such a relationship between media orientation and media use seems more likely in the case of the lower status groups. To lovers of light tv entertainment (prime orientation) the popular press offers a fitting supplement by way of stories about celebrities and previews of forthcoming programs. Because of the tighter time budget of higher status groups and the negligible surplus-value which televison information offers in periods without extraordinary events, only a tiny segment of higher status groups is likely to regard television information as a valuable supplement to serious print information. The third (combined) approach is more sophisticated. True, it puts stronger conditions on the kind of data that must be obtained, but with our operationalization of kinds of media and of programme categories we may be deemed to meet these conditions. Not only does it combine the profits that the two earlier approaches offer separately, but it also better fits current sociological theories of the habitus guiding judgements of taste and interests (Bourdieu 1979, 1980). As television became more widespreed, its status changed and it became less opportune, especially for highly educated people, to take a stance against the medium. Therefore it does not make sense to focus primarily on differences between types of media. Neither will we limit ourselves to the content analysis of the use of one specific medium. In contrast to that, we wish to focus on content aspects of both print media and television, expecting thereby to bring to light differences in the media repertoires of types of media users. In the period under consideration, it is likely that shifts in attention for reading and television viewing have occurred. However, for distinct groups these shifts will vary according to each group’s media orientation. Briefly after its introduction

The fragmentation of the media audience 269

television was popular among the lower educated. It was mostly seen as a medium offering entertainment, to a far lesser extent as a source of serious information. Hence, avid readers of serious information are unlikely to be part of to the fervent television viewers. Seekers of serious information used to turn to print media rather than television. Unlike this group, entertainment seekers found in television a medium that for the most part satisfied their interest. Therefore, during the 1970s (and also before) the dominant media orientation not only affected the choice within types of media but also between types of media. Avid readers — mostly the high educated — used television rarely. For those disliking reading or perusing only the popular press, the new medium offered an alternative to both cultural practices outside the home and inside activities (visits, games). Since the 1970s younger generations have grown up while television was an integral part of their life. Television’s share in time spent at home is highest among the youngest age groups. But among the elderly time spent on watching television has also increased beyond average (Van den Broek et al. 1999). For serious readers, broader media use, including television, is no longer thwarted by a slight disdain of television, as was the case in the 1960s and 1970s. The importance attached to a unique orientation towards a particular kind of medium has decreased. At the same time, within kinds of media, boundaries between information and entertainment have become less strict. In addition to television’s wide acceptance, this is due to the fact that the boundary between elite and popular culture has become less relevant, especially to younger generations (SCP 1998:727–730). The ideal of enculturation in elite culture, widely accepted until the 1960s, with the school as a mediating institution, has gradually become less evident to large groups, in spite of the higher degree of education they attained on average. This does not imply that in particular young people may be viewed as ‘cultural omnivores’. Their interest in elite culture is too limited. ‘Omnivores’ are for the most part to be found among people aged 45 or less, highly educated, single or belonging to a small household, and living in a metropolitan area. De Haan and Knulst (2000) also ascribe this change in cultural behavior to changes in socialization since the 1960s. In light of this the group of cultural omnivores (a group which, in our case, shows an interest in both serious information and light entertainment) is likely to have increased between 1975 and 1995. Hence, we assume that to certain groups the distinction between types of media or types of content is less relevant and that the group of people with a boundary transgressing interest has increased. These assumptions are in agreement with theories on the rise of postmodern lifestyles (Featherstone 1991; Maffesoli 1996). 2.1 Reader typology With a view to our research question we call to mind the reader typology which was brought to light in the analysis of Van Eijck and Van Rees (1999). Reader types were

270 Kees van Rees and Koen van Eijck

identified on the basis of their using three kinds of newspapers and three kinds of magazines that jointly characterize each type’s reading repertoire. Individual media products (whether newspapers or magazines) were aggregated by using the categories 1 (focused on information regarding national and international politics, government policy, and legitimate culture [cf. Janssen 1999]), and 3 (focused on human interest, entertainment and sectors of culture with a less legitimate status), category (2) thus representing a middle position on the information — entertainment axis (in the case of newspapers involving regional dailies, in the case of magazines bearing on special interest, hobbies).2 The five reader types which could be distinguished on the basis of these indicators are: ‘entertainment readers’, a label prompted by their exclusive score on items of the entertainment pole; ‘information readers’, in a way the counterpart of the first group because of their exclusive score on items of the information pole; ‘regional readers’; ‘non-readers’ and finally a group designated as ‘omnivores’, with relatively high scores on all reading categories and differing from the other types by their combining items from apparently heterogenous categories (e.g. popular newspapers and opinion magazines). The latent class analysis we performed shows that five distinct reading types suffice to map respondents’ reading behavior with a fitting model (Van Eijck and Van Rees 1999: 137–8). These reader types have in general the same characteristic reading patterns in 1975 and 1995, but they differ in the probability of having used a particular reading item during the week respondents registered their activities in the diary. In 1975, class 1 (entertainment readers) comprises almost a third of the respondents. People belonging to this class have a low probability of reading category (1) newspapers (a reading probability of 8%) or opinion magazines (3%). Much in vogue among people from class 1 are regional newspapers (category 2) and popular newspapers (category 3) (56%), but between 1975 and 1995 these percentages remarkably declined (to 1% and 31% respectively). Magazines other than opinion magazines are favored by this group of readers. No less than 77 percent (in 1975) and 49 percent (in 1995) read category (2) magazines; for category (3) magazines these percentages are 55 and 35. Members of class 1 can be characterized as people having no interest in category (1) newspapers and magazines, but they are notable because of their intensive use of category (2) and category (3) magazines. The 1995 reading profile of this group is different from that of 1975. The probability of reading a regional newspaper is almost nil and the probability of reading a popular newspaper has also strongly decreased. Magazine reading is also in decline. Though in 1995 they still are the most fervent readers of popular magazines, they have lost their top ranking for category (2) magazines. Across the board this group has turned itself away from reading. In 1975, class 2 (information readers) is the next-to smallest in size (with 10%) and, in 1995, the smallest with nine percent. It is the only class which uses category (1)

The fragmentation of the media audience 271

newspapers very intensively (81% in 1975 and 98% in 1995). These readers’ scores on opinion weeklies are exceeded only by those of class 5 members. Regional newspapers are neither well-read nor neglected while category (3) newspapers have few readers in 1975 (22%), and none in 1995. The interest in category (2) magazines is below average in 1975 and average in 1995; that for category (3) magazines is less than the other four classes. Information readers tended to apply themselves more to quality newspapers at the expense of opinion magazines. Class 3, that of the regional readers, contains almost one quarter of the respondents in both years. As for the reading of newspapers, this group limits itself almost completely to regional dailies. It holds an intermediate position with regard to opinion weeklies and category (3) magazines, though this interest has declined; its score on category (2) magazines is fairly high and is actually the highest in 1995. Between 1975 and 1995 their interest in magazines of category (1) and (3) has decreased. Clearly, this group is characterized by its focus on regional newspapers. Class 4 consists of people who scarcely read. They score lowest of all groups on category (1) newspapers, low on category (2) and category (3) newspapers, very low on opinion magazines, by far the lowest on category (2) magazines, and relatively low on category (3) magazines. The group’s size has only slightly increased between 1975 and 1995, but during this period its already negligible interest in reading declined considerably. In 1995, this group might be designated as non-readers: with the exception of regional newspapers, there are no reading items that manage to secure its attention. Finally, class 5 is a rather small but expanding group (5% in 1975 and 16% in 1995) that is notable for its combining opinion magazines with regional or popular newspapers. Though a quarter (1975) to a third (1995) of this group reads category (1) newspapers, it appears to draw its serious information mainly from the more conservative opinion magazines. In both samples, class 5 showed the highest score on popular newspapers. The increase in omnivores suggests a more relativizing attitude of highly educated people with regard to the distinction high/low, in this case with regard to more or less legitimate reading items.3 Most notable about this group is the fairly even spread of reading probabilities over all items. These probabilities were all relatively high. Therefore, class 5 seems most eligible to be qualified as omnivores, with two provisos. First, we must emphasize that the meaning of this notion is always connected to the specific repertoire of cultural items from which cultural consumers can choose. A reading omnivore does not automatically display omnivorous behavior in watching television. Second, the tendency of reading omnivores to combine items from distinct categories (1, 2 and 3) does not turn all members of this group into ‘read-alls’. In 1975 the group of ‘omnivores’ intensively perused all kinds of reading materials. In 1995, their interest continues to be evenly spread over the items, but the intensity of their reading

272 Kees van Rees and Koen van Eijck

behavior has considerably decreased. In truth, this decline holds for all classes, as the probability of reading has declined for almost all items in all classes.

3.

Hypotheses on the viewing behavior of reader types

To answer our research question it suffices to categorize 1975 televison programs into two main types (information and entertainment), each comprising two categories: serious and light. In addition for 1995 two kinds of programs are distinguished, but on account of the huge increase in channels and the fairly global TBS registration, we limit ourselves to a bi-partition of public channels, on the one hand, and private ones, on the other. As long as respondents’ television watching cannot be classified according to programme type in the same way for 1995 as for 1975, one has to keep in mind that the programming of the major public broadcasters, focusing on light entertainment, strongly resembles that of the private broadcasters. Our main research question bears on the relationship between reading behavior and television viewing. The question of how changes in reading behavior are related to changing broadcast use will be answered by estimating the probability that each reader type will display a specific kind of television viewing behavior. From the entertainment readers (class 1), we expect viewing behavior which resembles that of non-readers (class 4). Both classes are not very interested in serious information and it seems unlikely that either will try to obtain such information from television. The reading repertoire of entertainment readers suggests that their television viewing behavior will also be characterized by a high score on light entertainment and a lack of interest in information programs. As non-readers fail to show a content-based preference, expectations about their viewing behavior have to be argued differently. The lack of interest in reading on the part of non-readers is no guarantee that they are heavy television viewers. For some of them, media orientation may be less developed, irrespective of medium type, print or visual. (Audiomedia use, not measured here, may play a different role in their lives.) However, given the penetration density of television and the different intellectual effort it requires, non-readers will for the most part belong to the regular or heavy viewers, with a special interest in light entertainment. For a small sub-group of non-readers, television is expected to function as alternative information source, even more so than for entertainment readers. For the low educated people we expect an interest mainly in light entertainment. In the case of one group, this orientation will be related to their substituting television entertainment for print media, implying that their print media use will be reduced to a minimum (the non-readers). Another group (the entertainment readers) will both for print and television focus more on light entertainment. The latter group will also read less in 1995 than in 1975.

The fragmentation of the media audience 273

Hypothesis 1 Non-readers of 1975 will be the most intensive users of light television entertainment; in 1995 they show a preference for the private broadcasters. The same holds for the entertainment readers whose media orientation is likely to shift more into the direction of television. In 1975, television did not yet provide the type of broad range of serious information programs on political and social-economic issues that were later developed in the 1980s. Therefore, highly educated information readers (class 2) lacked a motive for developing a greater affinity for the medium. It could not satisfy their information need as category (1) reading items did. We also expect their score on other programme categories to be low, reflecting the slight disdain with which television was treated in intellectual circles at the time. In 1975, information readers are probably not interested in light entertainment. In 1995 the situation has changed in several respects. First, a broad range of opinion and analysis programs better meets their preference for serious information. Second, the supply of serious television entertainment has grown, providing information readers with far more programs that potentially suited their media orientation, perhaps more so with public broadcasting companies and foreign channels than with the private ones. Third, the disdain mentioned above had faded by 1995. In spite of this, compared to other reader types, information readers seem most likely to belong to the category of viewers who watch little public and little commercial television. Hypothesis 2 The rather exclusive orientation on serious information in elite print media that we find in 1975 among the highly educated, will be supplemented in 1995 with television programs of the serious and legitimate category. As argued above, omnivorous readers (class 5) are not automatically also omnivorous in their viewing behavior. At the same time they are more likely than information readers to also integrate televison in their media repertoire, even though they are not expected to score highest on light entertainment (in 1975) or on private broadcasting (in 1995). Because of their tolerance to a broad media supply, their education above average and their preference for print media which bring information in a light way, they are likely to show a relatively high score on legitimate (serious) entertainment and at most an average score on light entertainment; likewise, an average score on serious information and a score above average on light information. In 1995, when reading probabilities within the class of omnivores have declined, the time gained on reading is partly spent on television viewing, with a preference for the private broadcasters, without however excluding the public channels.

274 Kees van Rees and Koen van Eijck

Hypothesis 3 The number of omnivores among the high-educated people increases between 1975 and 1995. In 1975, they watch more television than the information readers and, in 1995, they are characterized by their combining public and private broadcasters. Regional readers (class 3) are not likely to have a clear equivalent among the four categories of television viewing. In 1975, most members of this class sought to satisfy their information needs through middle category newspapers and women’s magazines. Television news (the 8 o’clock news) is likely to have served as a supplementary source of information, as the reach of this programme was very high at the time, but we see no reason to expect a score above average. Compared to inhabitants of the metropolitan area called the ‘Randstad’, this non-metropolitan group has a smaller number of cultural out-door activities at their disposal, and from that perspective, serious television entertainment might provide an alternative. Lacking an urban lifestyle, however, it seems less likely that this group will focus on the forms of legitimate entertainment (art house films; satirical cabaret) provided by broadcasters. Their viewing behavior is more likely to resemble on the one hand that of non-readers and entertainment readers (because of their expected preference for light entertainment), on the other hand, that of omnivores (because of the broad range of programs of their viewing repertoire, though always on the light level). In 1995, this kind of orientation probably manifests itself through a greater loyalty to the public broadcasting companies. Hypothesis 4 In 1975, the regional readers mainly focuses on light entertainment, while in 1995, this group will for the most part select public broadcasters.

4. Data and method The data for the analyses of the viewing behavior of reader types are drawn from the very same years of the Time Budget Surveys that we used for our analysis of reader types. In order to cover as long a period as possible, it was decided to compare data from the first (1975, N = 1309) and the most recent (1995, N = 3227) surveys. Data were gathered by means of a diary, supplemented with two oral interviews, the first of which was conducted before the week during which the diary was kept, the second was conducted alongside the retrieval of the diary data. The diary covers seven subsequent twenty-four hour periods. For each fifteen minute period, respondents had to indicate their activity by means of pre-coded categories. To enhance comparability across the years, data were always gathered during a week in October.

The fragmentation of the media audience 275

Since CEBUCO and NOTU, two major publishing trade organizations, were financially involved, the data contained precise information on the categories of print media perused by respondents.4 Television viewing was measured differently for 1975 and for 1995. For both years we have limited our analysis to television viewing during the evening (6 to 12 p.m.). In 1975, television programs are classified by content into the four categories outlined above. To this end, we consulted the TV Guides for the two weeks in question. The first category, ‘serious information’, includes; news, programs providing background information on socio-economic and political issues; educational courses; religion; an occasional programme on art bearing on the socio-economic position of the artist. The second category, ‘light information’, comprises the following programs: talk show; documentary on nature; sports news and accounts of sports matches; information on art with a primary focus on human interest; children’s stories based on the bible; viewers’ experiences with illness. The third category, ‘legitimate entertainment’, consists of programs of legitimate culture (classical music, religious music, theater, art-house or director films, and movies based on twentieth century literature), satirical cabaret, cultural programs on art and artists other than those of category (2). For the fourth category, ‘light entertainment’, we included programs of and on pop music, series (national and foreign), detective programs, vaudeville-like shows, movies based on 19th century or older literature, game shows, and finally programs containing light-information ingredients (for instance quizes). For each program category viewing time was partitioned in tertiles. For serious information these are as follows: 0 to 3 quarters of an hour (1); 4 to 9 quarters (2); and 10 quarters or more (3). For light information: 0 to 2 quarters (1); 3 to 7 quarters (2); 8 quarters or more (3). For serious entertainment: 0 quarter (1); 1 to 5 quarters (2); and 6 or more quarters (3). For light entertainment: 0 to 7 quarters (1), 8 to 17 quarters (2) and 18 quarters or more (3). In 1995, television viewing is measured in a more global way because of the enormous increase in number of channels. To characterize viewing behavior, we used a bi-partition of public versus commercial broadcasters. This gives us the total viewing time for the chanels Netherlands 1, 2 and 3 on the one hand, and for the Dutch commercial broadcasters, on the other. These variables were combined into a single television viewing variable in order to give as much information as possible. To this end, we started by dividing respondents on the basis of the median viewing hours into two groups according to whether they watched public television at most two and a half hours versus more than two and a half hours. Then the first group, viewing public channels the least, was divided into two halves: those watching commercial channels for at most 4.5 hours and those viewing them for more than 4.5 hours during the diary week. Subsequently, the same bi-partition was made for

276 Kees van Rees and Koen van Eijck

the other group of respondents, who were more avid public television watchers (2.5 hours or greater). Here, the median for commercial television was 3.25 hours. Thus, four global viewer types emerged: (1) those viewing few commercial and few public channels; (2) those viewing few private but much public; (3) those viewing much private and few public; (4) those viewing both kinds of channels rather intensively. The effects of these covariates are estimated separately for 1975 and 1995, since, as indicated, the focus is on the effect of television viewing behavior, which is measured differently in 1995 than in 1975. This hampers a direct comparison of television’s role in these years, but that is not the primary aim of the analysis. To examine the effect of the covariates, we decided to estimate a logit model, having as a dependent variable the latent class on which a respondent had the highest probability score. Assignment of respondents to a class is performed with the programme LEM (Vermunt 1997).

5.

Results

Above we mapped the reading patterns of the five classes and made explicit the changes in class composition. This provides a basis for examining the impact of six covariates, or independent variables, on reading behavior. The covariates are gender, age, level of education, time budget, book reading, and television viewing. The analysis will be centered on the relationship between television viewing and the use of print media. But per hypothesis the effects of the other five covariates are described in a separate paragraph. Simultaneously entering the other covariates (or: control variables) guarantees that the observed relations between television viewing and reading patterns cannot be ascribed to commonality with any of these variables. Thus, we aim to answer our dual research question: Which television viewing behavior is characteristic of specific reader types and how did the relationship between television viewing and reading change over time? Our purpose is to test to what extent television negatively affects the likelihood that a person will choose to read, and thereby confirms its alleged role of the medium that pushes reading off its high ranking as an indoor leisure time practice, and under what conditions this occurs. Additionally, we will investigate under what circumstances television and reading complement each other, for example, by satisfying similar or complementary needs. Results are presented in the order of the hypotheses in section 4. All covariates had significant effects. Effects of the covariates are reported in Tables 1 and 2 for 1975 and 1995 respectively. The values of these effects were estimated with effect coding (instead of dummy coding). If the two upper rows of Table 1 are regarded as a table for gender (G) and class (C), with i levels for G and j levels for C, effect coding implies that both over i and over j, the cell values sum to zero. For the effect of gender, for example, this means not only that the log odds for

The fragmentation of the media audience 277

men and women per class are each other’s mirror image with zero as their midpoint, but also that the log odds of gender sum to zero along the rows, that is, over the five classes (disregarding rounding errors).5 5.1 Classes 1 and 4 Women and younger people are more likely than men to belong to class 1 (entertainment readers). This holds for both years, but the gender effect is stronger in 1975 and the age effect is stronger in 1995. Hence, for teenagers and men, the probability of belonging to this class is higher in 1995 than two decades earlier. In addition, in 1995 the probability of belonging to class 1 is strongly, negatively, and linearly associated with the highest level of education, whereas in 1975 this relationship is curvilinear, with the highest probabilities for the low-educated and those with professional college. Time budget has a small effect in 1995, while in 1975 those in the second tertile of time budget had the highest probability of belonging to this class. In 1975 people who hardly read books for leisure had a probability slightly above average of belonging to this class; but, of course, the highest effect for the few-book-readers is to be found in class 4, the non-readers. In 1995 the relationship between book reading and membership of the class of entertainment readers is negligible. In 1975, the viewing behavior of this reader type is characterized by a slight interest in serious information. Entertainment readers have a probability above average (positive log odds of 0.11) of belonging to the tertile of least frequent viewers of serious information and a rather small chance (negative log odds of −0.08) of belonging to the tertile of most frequent viewers of this programme category. This reader group prefers light information, but not nearly as much as it prefers light entertainment. This clearly indicates the existence of a parallel in the content of reading materials and television programs. A similar pattern occurs in 1995: entertainment readers spent much time viewing private broadcasters, which provide a higher proportion of entertainment than public broadcasters. If they watch public broadcasts at all, they do so merely in combination with a large amount of private ones. The probability of belonging to class 4, the class which uses printed media the least, is positively affected by being male, just like the probability of belonging to class 5. Again, the gender effect has decreased. The age effect is negative, but in 1995 teenagers are even less likely to be among this group than in 1975. This does not seem to correspond with previous analyses showing a large decline in reading among adolescents. However, as our analysis shows, for members of these younger age groups the probability of belonging to the class of entertainment readers is larger than in 1975, while that of belonging to the non-readers has decreased. Their probability of belonging to the information readers or the regional readers has also declined. Shifts (in fact: declines) in reading behavior within classes of readers must

278 Kees van Rees and Koen van Eijck

Table 1.Effects of covariates (tv watching, etc) on assignment to 5 reader classes, 1975 Class 1 Class 2 Class 3 entertainment information region

Class 4 non-reader

Class 5 omnivore

gender man woman

−0.36 −0.36

−0.02 −0.02

−0.30 −0.30

−0.20 −0.20

−0.44 −0.44

age 12–19 20–34 35–49 50–64 65+

−0.28 −0.05 −0.00 −0.06 −0.27

−0.25 −0.17 −0.03 −0.14 −0.32

−0.05 −0.08 −0.17 −0.05 −0.14

−0.62 −0.02 −0.05 −0.41 −0.18

−0.59 −0.03 −0.09 −0.38 −0.27

education primary jun vocational jun gen secondary high school professional college university

−0.21 −0.02 −0.02 −0.34 −0.13 −0.00

−0.65 −0.41 −0.66 −0.13 −0.51 −1.33

−0.14 −0.13 −0.14 −0.03 −0.01 −0.13

−0.71 −0.49 −0.38 −0.03 −0.67 −0.86

−0.44 −0.27 −0.40 −0.47 −0.02 −0.13

time budget lower quartile second quartile third quartile highest quartile

−0.09 −0.16 −0.01 −0.07

−0.18 −0.04 −0.20 −0.06

−0.19 −0.08 −0.15 −0.03

−0.01 −0.21 −0.06 −0.16

−0.12 −0.01 −0.01 −0.12

recent book reading for leisure >3 months ago 1–3 months ago 3 months ago −0.03 1–3 months ago −0.02  0.01).2 Bartlett’s test of sphericity showed that none of the dependent variables could be regarded as a linear function of the others. The determinant was 0.939, strongly exceeding 0; the observed significance level was high, that is 0.195. We therefore performed separate univariate analyses of variance, each of which had the respective numbers of key, illustrative, and descriptive quotations as the dependent variable. The independent variables were: ‘author’s literary status’ (two values: ‘high’ and ‘less high’), ‘reviewer’s value judgment’ (two values: ‘positive’ and ‘not positive’) and a variable assessing the frequency with which a reviewer contributed to a newspaper. This variable was operationalized by taking ‘having at least two reviews in the sample’ and ‘having one review in the sample’ as indicators of frequent and

Quotations expressing the valuative stands of literary reviewers 387

less frequent contributions, respectively. The variable was also operationalized by taking ‘having at least three reviews in the sample’ and ‘having at most two reviews in the sample’ as indicators of frequent and less frequent contribution, respectively. Chi-square tests indicated that there were no significant relationships between the independent variables. Our hypothesis is that an author’s high or moderate literary status affects the degree to which key quotations occur in reviews of their new releases, but that literary status has no impact on the occurrence of descriptive and illustrative quotations. Besides literary status, other factors might also affect the inclusion of key quotations in reviews. A reviewer’s very favorable opinion of a new release by a high status author could make him or her use key quotations to support the view that the new release is a masterpiece. Since an author’s high status may be at variance with critics’ judgement about his or her new release, and since positive reviews are also obtained by authors of lower literary status, it was not expected that the variable ‘critic’s judgement’ would affect the occurrence of key quotations. Similarly, since we did not suppose that reviewers who publish frequently focus on high status authors, it was not expected that the variable measuring the frequency with which reviewers contribute to dailies and weeklies had any impact on the occurrence of key quotations in reviews. However, the length of a review might induce critics to insert quotations in their reviews or prevent them from doing so. Inspection of the data yielded that there were substantial differences in the average length (measured in number of words) of reviews when one of the three forms of quotation was absent. A one-way analysis of variance established that these differences were significant (F = 5.14; p = 0.008). A Scheffé analysis showed that there was a significant difference between the average length of reviews without key quotations and of reviews without illustrative quotations. Therefore, ‘length of review’ was included as a covariate. The covariate could be related to the value judgement of the reviewer: a critic might write at greater length about a new release that he judges favorably and might tend to be brief when the new release is felt to be disappointing. The covariate might be related to the frequency with which reviewers publish: frequent contributors to a newspaper may tend to elaborate on the book under review or be rather succinct. If the covariate is related to one of these variables, or to both, and also to the dependent variable, we have a problem. When an independent variable is related to another independent variable as well as to the dependent variable, this variable could be the common cause of the two other variables. In that case, the conclusion that there exists a causal relationship between the dependent and the independent variable would be incorrect. In the multiple regression analyses that follow, we first determined the effect of the independent variables ‘author’s literary status’, ‘critic’s value judgment’ and ‘critic’s

388 Hugo Verdaasdonk

frequency of contribution’ on each of the three dependent variables. We then assessed the effect of the covariate. The standard practice is to control for the covariate and then to determine the effect of the independent variables on the dependent variable. Note, however, that the direction of a possible causal relationship between the independent variables never starts from ‘length of review’. ‘Length of review’ cannot affect an author’s status, critic’s value judgment or the frequency with which a reviewer publishes. If there is causality, each of these independent variables could have impact on the occurrence of quotations and/or on the length of a review, and not vice versa. Therefore, the effect of the covariate on the dependent variable was assessed controlling for the independent variables.

5.

Results

As stated, the years in which the 80 reviews were published cover the period 1976 to 1996. The largest group of reviews (17) dates from 1995. Most of the reviews (58) appeared in quality papers. The total number of reviewers is 47; four of them had three reviews; one, four; and two had five reviews in the sample. Positive judgements were given in 48 reviews; negative appraisals in 21, whereas 11 reviews were undecided. Key quotations occurred in 39 reviews, descriptive quotations in 58 reviews, and illustrative quotations in 31 reviews. Table 1 gives the means and standard deviations of all continuous variables that were used in the subsequent analyses. T-tests showed that the reviews in quality papers contained significantly more words than those in popular papers (Quality papers=795.74; Popular papers= 548.5; t=3.88; df = 59.75; p = 0.000). Reviews in each of the two types of paper do not significantly differ from each other in their Table 1.

Means and standard deviations of the continuous variables figuring in the subsequent analyses

Variable name key quotations descriptive quotations illustrative quotations readability score length of review (in words)

Mean

Standard deviation

0.538 1.086 0.519 54.663 727.750

0.586 0.784 0.707 9.737 327.355

‘key quotations’ = quotations which, according to the reviewer, give clues to the meaning the literary author wants to be assigned to the text; ‘descriptive quotations’ = quotations the critic uses to characterize figures and scenes in the text under review; ‘illustrative quotations’ = quotations the reviewer gives as examples of the literary author’s manner of writing; ‘readibility score’ = score determined with a Flesch formula indicating the comprehensibility of the reviews in the sample (where a score of 100 or more counts as ‘very easy to read’); ‘length of review’ = number of words contained in the reviews in the sample.

Quotations expressing the valuative stands of literary reviewers 389

readability scores. Key quotations do not occur significantly more in quality papers than in popular papers. The two other forms of quotation are also found to an equal extent in each of the two categories of papers. Reviews of new releases by authors with high status contain significantly more key quotations than reviews of new releases by authors with less status (High status = 0.75; Less status = 0.33; t = -3.42; df = 76; p = 0.001). Illustrative and descriptive quotations are evenly distributed across reviews of new releases by both categories of authors. The average number of words in the reviews devoted to new releases by high status authors or by authors with less status shows a difference that just fails to become significant at the 5 per cent level (High status = 798.33; Less status = 657.18; t = -1.96; df = 78; p = 0.053). There is no significant difference between quality papers and popular papers in the number of reviews devoted to authors with high status or with less high status. Authors with high status do not get more favorable reviews than their less reputed counterparts. Do frequent reviewers — who might be regarded as opinion leaders or tastemakers — contribute predominantly to quality papers? This did not appear to be the case. If we operationalize ‘frequent’ as ‘having at least two reviews in the sample’ and ‘less frequent’ as ‘having one review in the sample’, there is no significant difference between the two types of paper. Similarly, we found no difference between the two types of papers in the frequency with which reviewers contributed to one of these under the operationalization of ‘frequent’ as ‘having at least three reviews in the sample’ and of ‘less frequent’ as ‘having at most two reviews in the sample’. Do frequent reviewers focus to a greater extent on high status authors than critics who publish less frequently? To this question, we must answer in the negative. Under the two operationalizations of ‘frequent’ and ‘less frequent’, no systematic relationships were found between the two types of critics and authors. Finally, books from which key quotations were taken were not more favorably reviewed than books from which no such quotations were selected. 5.1 Results of the analyses with ‘number of key quotations’

as the dependent variable It was hypothesized that ‘author’s literary status’ is the only independent variable affecting the occurrence of key quotations in reviews. ‘Key quotations’ are passages in the text under discussion which, according to the reviewer, offer clues to the meaning of the text. The results support our hypothesis. There is a significant main effect (F=5.876; df=3.69; p=0.001) to which ‘author’s literary status’ makes a significant contribution (F=10.140; df=1.69; p=0.002). Unexpectedly, ‘critic’s frequency of contribution’ — operationalized as ‘having at least two reviews in the sample’ versus ‘having one review in the sample’ — also contributes significantly (F = 4.561; df = 1.69; p = 0.036), although to a lesser extent

390 Hugo Verdaasdonk

than ‘author’s literary status’. The other independent variable (‘critic’s value judgment’) was non-significant, as were the two-way and three-way interactions between the three independent variables. The covariate (‘length of the review’) had no significant impact either. Multiple R2 was 0.205. The other model with ‘author’s literary status’, ‘critic’s value judgment’ and ‘critic’s frequency of contribution’ as independent variables — where the latter variable was operationalized as ‘having at least three reviews in the sample’ versus ‘having at most two reviews in the sample’ — yielded very similar results. Again, there was a significant main effect (F = 4.773; df = 3.69; p = 0.004). This time, however, only ‘author’s literary status’ significantly contributed to the main effect (F = 12.306; df = 1.69; p = 0.001). No significant effects were found for any of the other independent variables, for their two-way and three-way interactions or for the covariate ‘length of review’. Multiple R2 was 0.174. That ‘critic’s frequency of contribution’ did not have significant impact suggests that its significant effect (under a different operationalization) in the first analysis does not indicate a robust effect of ‘critic’s frequency of contribution’ on the occurrence of key quotations in reviews. 5.2 Results of the analyses with ‘number of descriptive quotations’

as the dependent variable Our other hypothesis was that ‘author’s literary status’ has no significant impact on the occurrence of quotations through which reviewers present characters and scenes from the text under discussion (‘descriptive quotations’). The results give no grounds to reject this hypothesis. No significant main effect or two-way or three-way interactions were found. Only the covariate ‘length of the review’ had significant impact on the dependent variable (F = 18.206; df = 1.71; p = 0.000). Multiple R2 was 0.229. In the subsequent analysis in which ‘critic’s frequency of contribution’ was operationalized as ‘having at least three reviews in the sample’ versus ‘having at most two reviews in the sample’, the covariate ‘length of the review’ again had significant impact on the dependent variable (F = 19.281; df = 1.71; p = 0.000). This time, the two-way interactions were significant (F = 3.595; df = 3.71; p = 0.018); only the interaction between ‘author’s literary status’ and ‘critic’s frequency of contribution’ had a significant effect (F=7.057; df=1.71; p=0.010). Multiple R2 was 0.218. 5.3 Results of the analyses with ‘number of illustrative quotations’

as the dependent variable Illustrative quotations are used by reviewers to give examples of a literary author’s way of writing. In the two analyses (each with a different operationalization of

Quotations expressing the valuative stands of literary reviewers 391

‘critic’s frequency of contribution’) no significant effects were found of any of the independent variables or of their two-way and three-way interactions on the occurrence of quotations critics use to give examples of an author’s manner of writing. The effect of the covariate likewise never attained significance.

6. Conclusion In reviews of new releases, the literary status of the author whose work is commented on has a strong impact on critics’ decisions to use quotations which, in their view, offer keys to the text’s overall meaning. Underlying this decision is the belief that high status authors have the power to communicate what their books ‘really’ are about; authors with less status are not credited with this power or, if so, to a much lesser extent. In identifying passages in a literary text as interpretive clues, the words used by the author — on an object-level — are used by the reviewer on a meta-level. As long as no procedures have been developed that warrant this transfer, presenting key quotations is an instance of hypostatizing the text under review or of hypostatizing the reader. Both forms of hypostatization, each of which presupposes the other, are based on the belief that a ‘great’ author is to be credited with the power to make clear to the reader what the fictitious world that is evoked ‘really’ is about. An author’s literary status has no impact on the occurrence of other types of quotations in reviews, that is quotations that serve to illustrate the author’s style or that present characters or scenes from the book under review using the author’s own words. The results also lend support to our view that, in literary criticism, classification and evaluation are closely related. On the one hand, marking a passage as a key quotation, that is as a clue to the text’s overall meaning, is an act of classification; on the other hand, this act also embodies a value judgment in so far as it pays tribute to the author’s high literary status and to his craftsmanship in enhancing the meaning of certain phrases by simultaneously making them evoke a fictitious world and comment on it by giving its meaning. Finally, the results suggest that the literary reputations of contemporary authors are unstable; even high status authors can never be sure that a new release, even one in which key quotations are identified, for that matter, will be favorably reviewed. An author’s new release will not only be weighed against his or her previous publications, but also against a continuous flow of publications by other authors. Moreover, contemporary literary reputations do not seem cumulative in nature, in that new ‘great’ achievements are added to the list of earlier noteworthy performances; instead, new achievements are weighed against previous ones and may very well replace them on the list. In classifying and appraising contemporary literary

392 Hugo Verdaasdonk

works, critics differentiate among an ever changing group of items. This makes their decisions strongly time-bound. In addition, their interest in contemporary authors seems to have a specific shape of time-dependence. Earlier research (Van Rees and Vermunt 1996: 331; Verdaasdonk 1994: 382) has shown that, even if an author maintains a steady rhythm of publication, critics lose interest in the course of time. All this makes authors’ literary status and reviewers’ value judgments relative in nature, although these value judgments are sometimes said to capture ‘timeless beauty’.

Appendix Legend: ‘P’ = Popular paper; ‘Q’ = Quality paper. The figures indicate numbers of reviews: ‘P2’ means: ‘2 reviews in popular papers’; ‘Q1’ means: ‘1 review in a quality paper’, etc. Reviews of new releases by authors who have high literary status (n1 = 40) Bernlef, J.: Cellojaren, 1995 (Q2, P1) Brakman, W.: Interieur, 1996 (Q2) Brouwers, J.: Zomervlucht, 1990 (Q2, P1) Campert, R.: Gouden dagen, 1990 (Q2) Haasse, H.: Berichten van het blauwe huis, 1986 (Q2) Hermans, W. F.: Ruisend gruis, 1995 (Q3) Komrij, G.: Zakenlunch in Sintra en andere verhalen, 1996 (Q2) Koolhaas, A.: Een aanzienlijke vertraging, 1981 (Q3, P2) Kossmann, A.: Een verjaardag, 1989 (Q1, P1) Krol, G.: Okoka’s wonderpark, 1994 (Q1) Mulisch, H.: De ontdekking van de hemel, 1992 (Q3) Multatuli: Brieven, selected by H. Brandt Corstius, 1990 (Q1) Reve, G.: Het boek van het violet en de dood, 1996 (Q3) Schendel, A. van: Verzameld werk 1, 1976 (Q1) ’t Hart, M.: Het woeden der gehele wereld, 1993 (Q2) Vestdijk, S.: De grenslijnen uitgewist, 1984 (Q1) Vries, Th. de: Baron, 1987 (Q2, P1) Wolkers, J.: De onverbiddelijke tijd, 1984, (P1) Reviews of new releases by authors who have less literary status (n2 = 40) Abdolah, K.: De meisjes en de partizanen, 1995 (Q1) Artus, R.: Zonder wijzers, 1995 (Q1) Bakx, H. W.: Midas’ tranen: een anekdote, 1987 (Q1) Bentis, E.: Moeders en dochters, 1989 (Q1) Bohlmeijer, A.: Op het punt van breken, 1987 (Q1) Broere, R.: Meesterwerk met brief, 1991 (P1) Cartens, D.: Duel, 1983 (Q1)

Quotations expressing the valuative stands of literary reviewers 393

Dantzig, R. van: Afgrond, 1996 (Q1) Duijnhoven, S. van: Dichters dansen niet, 1995 (Q1) Februari, M.: De zonen van het uitzicht, 1989 (Q1) Foelkel, F.: Onder de pannen, 1995 (Q1, P1) Franke, H.: Nieuws van de nacht, 1995 (P1) Friedmann, C.: Twee koffers vol, 1993 (Q1) Grunberg, A.: Blauwe maandagen, 1994 (P1) Kessels, M.: De god met gouden ballen, 1995 (Q2) Kolk, G. van der: De nieuwe stad, 1987 (P1) Kummer, Em.: Helden zijn zwart, 1994 (Q1) Lewin, L.: Voor bijna alles bang geweest, 1989 (P1) Loesberg, R. A.: Een eigen auto, 1977 (Q1) Marinus, R.: Mandrillenfarm, 1976 (P2) Martens, H.: Een naald op het water, 1992 (Q1) Mors, H. ter: De dealer en de organist, 1982 (Q1) Niemöller, J.: De spier, 1993 (Q2) Ombre, E.: Maalstroom, 1992 (Q1) Oort, D. van: Meisje voor halve nachten, 1989 (Q1) Sahar, H.: Hoezo bloedmooi, 1995 (P1) Spoor, H.: Een huwelijk, 1995 (P1) Stevens, H.: Een schone slaap, 1991 (Q1) Stüger, F.: Desillusies, 1982 (Q1) Tepper, N.: De eeuwige jachtvelden, 1995 (P1) Terduyn, E.: De ijsprinses, 1982 (Q1) Thomése, P. F.: Heldenjaren, 1994 (Q1) Vacher, R.: Vrije val, 1992 (Q1) Willemsen, F.: De gelaagde sneeuwman, 1990 (Q1, P1) Woerden, H. van: Tikoes, 1996 (P1)

Notes 1. The following Dutch or Flemish newspapers were regarded as quality papers: De Groene Amsterdammer, De Morgen, De Standaard, De Tijd, De Volkskrant, Elseviers Magazine, Hervormd Nederland, HP/De Tijd, NRC/Handelsblad, Het Parool, Trouw, and Vrij Nederland. In contrast, Algemeen Dagblad, De Gelderlander, De Stem, De Telegraaf, Haagsche Courant, Het Binnenhof, Het Vaderland, Leeuwarder Courant, Nieuwsblad van het Noorden, Provinciale Zeeuwse Courant, Rotterdams Nieuwsblad, and Utrechts Nieuwsblad, count as popular papers. 2. The correlations were: key quotations × illustrative quotations: r = −0.151; key quotations × descriptive quotations: r = 0.202; illustrative quotations × descriptive quotations: r = 0.101.



394 Hugo Verdaasdonk

References Bourdieu, P. 1992. Les Règles de l’Art: Genèse et Structure du Champ Littéraire. Paris: Seuil. Genette, G. 1972. “Discours du récit. Essai de méthode”. In Figures III, G. Genette, 65–286. Paris: Seuil. Janssen, S. 1994. In het licht van de kritiek. Variaties en patronen in de aandacht van de literatuurkritiek voor auteurs en hun werken. [In Light of Criticism. Variations and Patterns in the Attention Literary Critics Pay to Authors and their Works.] Hilversum: Verloren. Janssen, S. 1997. “Reviewing as a social practice: Institutional constraints on critics’ attention for contemporary fiction”. Poetics 24: 275–297. Van Rees, C. J. 1986. Literary Theory and Criticism: Conceptions of Literature and their Application. Ph.D. Thesis, University of Groningen. Van Rees, C. J. and Vermunt, J. 1996. “Event history analysis of authors’ reputation: Effects of critics’ attention on debutants’ careers”. Poetics 23: 317–333. Verdaasdonk, H. 1979. Critique Littéraire et Argumentation. Ph.D. Thesis, Free University of Amsterdam. Verdaasdonk, H. 1981. “Some fallacies about the reading process”. Poetics 10: 91–107. Verdaasdonk, H. 1983. “Social and economic factors in the attribution of literary quality”. Poetics 12: 383–395. Verdaasdonk, H. 1994. “Analogies as tools for classifying and appraising literary texts”. Poetics 22: 373–388.



The proper place of humanism Qualitative versus scientific studies of literature Colin Martindale University of Maine

There is nothing intrinsically wrong with a humanistic or idiographic approach to literature. If one wants to study one or a few writers, this approach can yield valuable insights. If we want to know what a specific book by a specific author means, we seek out the opinion of a good literary critic rather than that of a scientific poetician — of someone who studies literature in a scientific manner. Scientists are good at generalities but poor at specifics. A number of years ago, I was invited to speak at a panel discussion at one of the meetings of the Modern Language Association on quantitative content analytic approaches to literature. The ill chosen topic was what content analysis could tell us about the poetry of John Donne. Before the presentations began, all of who were to speak agreed that the real answer to the question was not much. I spoke on a quantitative study of the history of British poetry. My only references to Donne were to point out which point on my graphs indicated his poetry. When I say below that humanism has failed, I mean that it has failed when humanistic scholars have tried to formulate and test general laws. Most humanists do not attempt to do so. Thus, I am not criticizing them. Given that people can only keep about seven ideas in mind at the same time (Miller 1956), the humanistic approach breaks down when one wants to formulate and test general ideas concerning a lot of writers. It is in such cases that a scientific or nomothetic approach becomes necessary. If we want to compare and contrast, say, 100 writers, we need to start counting. Once we have counted, we need to analyze our numbers in a scientific way if for no other reason than the simple fact that the human mind cannot think about 100 things at the same time. Some people are interested in concrete details about specific things. Such people have no choice but to work in the humanities. Other people are interested in abstractions and generalities. They tend to work in one of the sciences. They certainly could work in a humanistic discipline, as there are many interesting regularities to be discovered in domains traditionally investigated in a humanistic manner. However, as many readers know, their work would not be greeted with great enthusiasm by most humanists.

396 Colin Martindale

Before continuing, I should point out that terms are used somewhat differently in America and Europe. To avoid confusion, I would classify any empirical study of literature, so long as it were done correctly, and analyzed in a quantitative manner, as scientific. It makes no difference if the person who did the study is in a department of literary studies or in a department of psychology. By psychology, I mean experimental psychology, which is a science. I do not mean clinical psychology, pop psychology, or psychoanalysis. These latter are not sciences. Humanists would be ill advised to claim them. They are, as Hegel might have said, beyond the pale of this chapter and beyond the pale of things that should exist in a rational world. Literary critics can certainly write very interesting and illuminating things about literature. They encounter difficulty when they attempt to deal with literary theory or poetics. The reason for this difficulty is not far to find. Humanistic scholars have no viable methods of inquiry; if they have an idea, they must use various rhetorical methods to persuade the reader that their idea is correct. In this, they can be compared with politicians, but the latter can try to convince us with “lies, damn lies, and statistics.” However humanistic scholars cannot try to convince their readers with statistics because, to be blunt, the majority of them are functionally innumerate. Beyond this, many of them seem to feel that it is downright evil to apply mathematical or statistical procedures to works of literature. Martindale (1990) proposed and tested a theory of literary history that accounts for about half of variation in word usage in the several poetic traditions on which it was tested. This is very much better than any other theory can do. One would think that literary scholars would have taken great interest in such a theory. In general they do not. Published reviews of the book by humanistic scholars reveal the probable reason. With one exception, the reviewers could not understand the statistics that I had used to verify the theory or the rather simple equations that I used to summarize my findings. One reviewer was honest enough to admit this. Most went off on various tangents ‘demonstrating’ that I had done something verging on a war crime. One somehow concluded that neo-conservative critics would love the book and shelve it with their copies of books by E. D. Hirsch. Since I had made no value judgments in the book, but only reported on counting some words in poetry and other forms of literature, I have never been able to understand how the reviewer had reached this conclusion. I do sympathize with the neoconservative critics, but I did not mention them in the book. I can only speculate as to the basis for this sort of attitude, but would imagine that it would damage a person’s self-concept to admit that one has no knowledge of elementary mathematics or statistics. It is far easier to say that such approaches are evil or unhelpful at least when applied to literature than to admit that he is innumerate. I certainly hope that they do not have the delusion that a humanistic approach to physics would be viable. I imagine that such rhetorical defenses, the only ones available to humanists, will increase in frenzy now that the barbarians (those of us

The proper place of humanism 397

who study literature in an empirical fashion and seek general laws) are at the gate. In fact, the gate has been breached, and we have for all practical purposes won the war. It is only for social and political reasons that humanists have not admitted defeat and signed an unconditional surrender. The future is ours; only the past is theirs. If one proposes a theory, it should be understandable, coherent, parsimonious, and testable or falsifiable. Humanism has no way of testing theories. Theories can only be tested using the scientific method. Humanists may try, based upon misreadings of theorists such as Kuhn (1962) and Lakatos (1978), to argue that science is no better than humanism. Such theorists only made explicit what scientists already knew. Of course, there is a good bit of rhetoric involved in science. No scientific theory can be completely tested. Any description of a scientific theory involves a number of statements that are technically nonsense (cannot be empirically tested) from a logical positivistic perspective, but logical positivism is dead. It has been replaced by a more sophisticated neo-positivism. Science has never guaranteed to tell us ultimate truth. All it promises is to give the best guess at the moment as to what is true. Of course the best guess changes across time as we get closer to the truth. For example, physics — the hardest science — is in a state of chaos and confusion at the moment. It is known that the theory of relativity and quantum mechanics are both wrong (Greene 1999). The theory of relativity works fine for huge distances and gigantic things, but yields absurd and impossible results if applied to tiny things such as electrons. Quantum theory has the reverse problem. Though it has always involved some dubious hypotheses, it explains the behavior of tiny things fairly well but collapses if applied to phenomena easily explained by relativity theory. Perhaps string theory will save the day, but arguments in its favor are almost entirely aesthetic as opposed to empirical at the moment. For those of us who have to deal only with medium sized objects and distances, Newtonian physics is close enough to the truth to do. Science, in its quest to explain the world, has produced some quite useful things such as transmission of electricity, light bulbs, computers, internal combustion engines, automobiles, airplanes, and on and on. These things are not social constructions but realities. Science has also produced some horrible things. An experiment was conducted in Hiroshima in 1945. This was an experiment the result of which was not at all certain until the experiment was carried out. Of course, the experiment was replicated at Nagasaki a few days later. The horrible destruction of these experiments was real. It was not a social construction. In retrospect, construction of atomic bombs should have been left to literary critics. Why? Because the bombs would not have exploded. How can I be so certain of this? Let us look at what the humanistic approach in any discipline has produced across the last few centuries. It has produced some beautiful and interesting books and articles in the area of literary studies, some of the most interesting of which have essentially used a quasi-scientific approach. Praz (1933) on the reversal of sex roles in nineteenth-

398 Colin Martindale

century literature is an example of a humanist critic proposing a quasi-scientific law that I showed using scientific methods is correct (Martindale 1975). Peckham (1965) on the nonexistence of cross-media artistic styles is an example of a humanist proposing and elegantly defending a quasi-scientific law that I showed in a number of experiments (Hasenfus et al. 1983) is incorrect: such styles exist and can easily be discerned by those with no training in the arts. The humanistic approach has certainly produced thousands of books and articles that were trivial, incorrect, absurd, incomprehensible, or unintentionally funny. (Of course, the same can be said of science.) Has humanism produced any general laws that shed light on the general nature of the physical or mental world? Very few, as that is not its function. The best predictor of the future is the past. If a method yields nothing of use after several centuries, common sense suggests that it be abandoned. This is why medieval scholasticism was abandoned. Perhaps the humanistic approach would be of value concerning uninteresting general questions such as how many angels can dance on the head of a pin. Humanists should try to stop answering general questions. Literature and media are a very important part of human life. Their study should not be left to those using methods that have consistently failed to work across the course of many centuries. If you want to know whether references to violence in the lyrics of American popular music causes, coincides with, or follows increases in the crime rate, do not ask for the idle speculations of a humanist, but ask me, as I have the data from 1930 to the present at hand and know how to analyze it. There is a relationship. Even better, one can use the content on these lyrics to predict the rise and fall of stock prices. However, the relationship is not strong enough for me to give stock tips unless one has a lot of money and is willing to take what I consider to be imprudent risks. If you want to know what these lyrics mean, do not ask me. I don’t know, and I don’t care. They are junk meant for the masses.

1.

The proper place of poetics

Whether in its old or new form, the positivistic approach of science is extremely reductionistic: ultimately, psychological laws will be reduced to biological laws, and the latter will be reduced to fundamental laws of physics. Reductionism is often misunderstood. At each level of reduction, emergent properties must be accounted for. Any high school student knows that water is not flammable, but its components — hydrogen and oxygen — are extremely flammable. One may toss a lit match into a bucket of water, and it will be extinguished. One may toss a lit match into a room filled with hydrogen and oxygen — at least once — and it will produce a tremendous explosion and a bit of water. Critics of reductionism seem to think that eminent scientists with Ph.D.s have forgotten what high school children know. Second, we mean that such a reduction would in principle ultimately be possible,

The proper place of humanism 399

not that it would actually be carried out in fact. If I ask how a telephone works, I want a five or six-line explanation at the appropriate level. I most certainly do not want to be handed a 100-volume book on psychology, neuroscience, physiology, quantum electrodynamics, etc. Large parts of chemistry can now be reduced to physics; this is not done, as it is easier to work on the chemical level. Though seldom made explicit, “ultimately” means a very, very long time in the future. There is no cause for concern that one’s understanding of Hamlet will be explained in terms of the laws of physics. Even were it ever possible, it would not in fact be done. We would end up with another 100-volume unreadable book. Psychology refers to the study of the behavior of organisms. Well, poets and readers are certainly organisms. Creation or reading of a work of literature is certainly a mental behavior. Thus, poetics — by which I mean the general theory of literature as opposed to literary criticism — is a subdiscipline of psychology, can be reduced to psychology, or simply is psychology. Exactly the same argument could be made concerning sociology, anthropology, economics, and all the behavioral sciences. I do not mean to alienate anyone. It would be nice to have a general science of behavior, whatever it were called, rather than separate disciplines studying different aspects of the same thing. In fact, psychology is not now in a position to be an ‘imperialistic’ science, because it is fractionating into increasingly specialized subdivisions. However, psychology and sociology (which I think should be one science rather than two) offer a firm and absolutely necessary basis for the study of poetics. In order to study poetics or even to be a literary critic, one should have a good basic knowledge of psychology and sociology. Of course, one should know linguistics as well. If one does not have a grounding in these disciplines, this makes about as much sense as trying to study trees with absolutely no knowledge of the general laws of botany. Few scholars in departments of literature have a good background in either scientific psychology or sociology. To be blunt, what this means is that they often do not have the training to do what they are trying to do and teach. Had they this background, they would most probably employ humanistic methods less and scientific methods more. I do not mean that they need formal training in these areas. To quote myself, “no one ever learned anything useful sitting in a classroom.” In my own case, I make an exception to this axiom. I learned more about psychology sitting in a classroom from the eminent literary critic, Wallace Fowlie, than I learned from sitting in classrooms listening to famous psychologists. Fowlie taught himself psychology as well as inferring it from great literature. This was clear in his lectures just as it was in his books. He was a great humanist with an anti-scientific bent. I would no more criticize humanism in general than I would criticize him. I wish that scientific psychology had more to offer literary scholars. We have a huge number of irrelevant studies of how hungry rats lost in mazes learn to find some food and even more on the almost totally trivial topic of human memory.

400 Colin Martindale

Most researchers in any discipline are consumed by a lust for the trivial. However, psychologists have done thousands of studies of creativity and thousands of studies of how personality relates to textual content. These studies should be of great interest to scholars studying literature. They form a firm basis for poetics. It would seem that humanistic literary scholars are completely unaware of these studies. For around a dozen years, I subscribed to PMLA. I even generally glanced through it. Not once did I see any of the studies mentioned above cited. If humanists want a guide to the most interesting of this literature, I (Martindale 1989) have reviewed it. What I find most interesting about it is that, whether they be mathematicians or poets, creative geniuses tend to think of their ideas in quite similar ways. If one wants to discover how a given poet thought of his or her ideas, there is no need to re-invent the wheel. Consult what psychologists have already discovered. A novel is usually about the thoughts of characters and their interactions. If one wants to analyze a novel, one should know something about thinking and social interactions rather than making up his psychology and sociology as he goes along. Both structuralist and cognitive psychologists have studied thinking intensively for over 100 years. Sociological and psychological social psychology approach social interaction from somewhat different perspectives, but before trying to explain fictional social interactions, it would be prudent to know what has been learned about real social interactions. To give a concrete example of the problems into which a literary critic can blunder, I can give an example from a very good literary critic. Todorov (1971) published a sort of text grammar of the rules governing Chaderlos de Laclos’s Les liasons dangereuses. I (Martindale 1976) was able to show that Todorov had imperfectly reinvented the wheel. That is, the rules governing the progression of this novel can be derived from Fritz Heider’s (1946) balance model. Heider’s theory describes the novel more accurately and parsimoniously than do Todorov’s rules. Had Todorov been aware of Heider’s model, he would have seen this immediately. Heider’s balance model is known to anyone in social psychology. Briefly, it applies to entities (these can be people or ideas or just about anything) and relationships (these relationships can be positive — liking, possessing, being near to — or negative — disliking, not having, being distant from). According to Heider, any system — a group of people, a set of ideas, or a combination of both — moves from imbalance toward balance, because imbalance causes tension or discomfort. For example, if two people like each other, but one loves communism and the other hates it, there is imbalance and tension. This imbalance can be resolved by changing the sign of the weakest relationship (the likeliest course in real life) or by adding new entities that will lessen the imbalance. In my example, the two people can find new things that they both have positive or negative relationships with. Heider gives an algorithm that allows one to compute numerically the degree of balance of any system. As a system increases in the number of entities and relationships, the

The proper place of humanism 401

computation becomes more and more difficult, but can always be done and will always give the same result.

2.

Abstraction versus concreteness in science and humanistic inquiry

I doubt that Heider ever read Les liasons dangereuses, but in 1946 he had already implicitly predicted what had to happen in the novel. This brings us to another difference between scientific and humanistic approaches. Todorov formulated a set of rules for a single narrative. Heider formulated a general theory that applies to anything from a group of people through a scientific paradigm to a fictional narrative. In fact, Heider’s model can be applied, with some modification, to anything from neural networks to crystals. This is because Heider himself had reinvented the wheel. His balance model can be easily derived from the laws of statistical mechanics in physics: any system will move toward what is called an energy minimum. An energy minimum is mathematically equivalent to what Heider called a state of minimum imbalance. The beauty of mathematics is that the equations do not know to what they are being applied. I was able to show that energy is minimized in Les liasons dangereuses and presumably any other narrative in a different way than it generally is in real life. The reason is simple. In real life, if we do not like a person, we stop associating with him or her. This would make a two-line short story, which just would not do. In fiction, energy must be minimized in a more circuitous way. The principle of energy minimization has been known for a long time in physics. Hopfield (1982) more recently showed that more or less the same energy minimization equations apply to spin glasses and to neural networks trained to recognize patterns buried in noise. Even more recently, I showed that more or less the same equations can be used to explain creative inspiration (Martindale 1995). I think it to be quite beautiful that more or less the same equations can be used to explain the formation of a crystal, the progression of a narrative, the process of pattern recognition, and the emergence of a creative idea. The equations can be applied to all sorts of other things as well. They apply with minor modifications whenever we can construe a system as consisting of entities and positive and negative relationships amongst them. The equations came from physics, but this is abstraction rather than reductionism. The only reason that the equations came from physics is that physics got started several centuries before the other sciences. A humanistic scholar would object to what I have just said, because I haven’t said anything specific or concrete about Les liasons dangereuses. I haven’t helped him or her understand this specific book. In my original publication about the book, I did deal with it quite specifically, but only because it is a nice example of a specific type of narrative. Scientists love abstractions, whereas humanistic scholars are buried in the concrete. They cannot see the forest for the trees. I’m sure we all

402 Colin Martindale

have an Uncle Fred who thinks that we shall be quite fascinated as to where he plans to plant the carrots next Spring and as to what his wife said about Mildred to her second cousin. Uncle Fred is buried in the concrete. This can be verified if one tries to tell him about abstractions; he’ll not understand them and become quite bored. There can be many reasons for this, but a very likely one is that he is not very intelligent. Thus, he cannot understand abstractions. Could the problem with humanistic scholars be that they tend to rather unintelligent? This is in fact no doubt one of their many problems. It is not at the moment politically correct to speak of individual differences in intelligence even though they are better established than many of the laws of physics, so I shall not pursue this question.

3.

The accomplishments — or lack thereof — of the humanistic approach to literary studies during the last half century

At its best, humanism is belles-lettres and should not be burdened by schools or paradigms, but it has been so burdened. Let us consider some approaches to literature in the last half of the twentieth century. The basic idea behind New Criticism was that a piece of literature should be analyzed without reference to anything outside the piece itself. One does not even have to know any psychology to see that this is impossible. Consider reading a poem. Our interest is in the poem as read and understood. One cannot understand the poem unless he knows what the words mean. The definitions come from outside the poem, but New Critics allowed this. Think of any word you want and do a bit of introspection. You are thinking not solely of the word and its denotative meaning but also of its connotative meaning and of related words, memories called forth by the word, and so on. The goal of New Criticism is rather obviously impossible. This should have been clear from the outset. Next, humanistic scholars tried structuralism. This held much more promise and paved the foundations for the empirical study of literature. However, structuralism was zero-degree science. One of the ideas of structuralism is to chop a narrative into segments and see how the segments relate to one another. This is reasonable enough, but the obvious first step is to see if people agree as to how to segment the narrative. If everyone segments the narrative in about the same way, then it is reasonable to proceed with the analysis. If different people come up with completely different segmentations, then one must go back to the drawing board and make up some guidelines as to segmentation and see if people now agree. Only once one has reliable agreement should he forge ahead. As far as I know, no one raised this quite obvious problem with structuralism. Rather than resolving this problem and moving toward science, most humanists moved on to deconstructionism. I don’t know much about this movement, as

The proper place of humanism 403

I have no tolerance for fools. Among other things, deconstructionism is based on the fact that if one looks in a dictionary, words are defined in terms of other words; if one looks up the meaning of a word and then the meanings of the words used to define it and so on, he may end up back at the word he started with. That is true enough, but the conclusion that we therefore do not know the meaning of any words or that meaning is indeterminate is completely absurd. This is an explanation of why we cannot communicate with one another. We do not need an explanation of that, because we quite obviously can communicate with one another. The deconstructionists were able to get away with their absurdities for a while because they wrote in such a précieuse, delirious, and confusing manner that it was usually unclear what, if anything, they were trying to say. Pure common sense tells us that if someone has anything worth saying, he or she will want to say it in a clear and straightforward fashion. Apparently even humanists figured this out. At around the same time, many humanistic scholars got the patently absurd idea that texts are not created by authors but by readers. Stanley Fish (1980) and others argued that if 100 readers read a text, they will come up with 100 more or less unrelated interpretations. This might be vaguely plausible had they been referring to extremely ambiguous texts by Derrida or Lacan, but the argument was that this is the case for all literary texts. The idea that people do not agree in their interpretations of texts goes back to an experiment conducted by I. A. Richards (1929) in which honors undergraduates were asked to write essays about their interpretations of 13 poems. The experiment could have been better conducted, but the real problem was that Richards did not know how to analyze the data he obtained. His study is a classic illustration of why humanistic methods almost always fail when applied to general questions. Richards spent several hundred pages giving examples of disagreements amongst readers and concluded that they did not agree at all. Richards was faced with the problem I mentioned at the outset. One cannot think of hundreds of things at the same time; one must count and use statistics. Richards made the mistake of including an appendix showing the number of readers who rated each poem as good, bad, or in between. A mere glance at the table makes it apparent that there was in fact a good bit of agreement. Applying a chi-square test to the data Richards presented tells us that for at least 12 of the poems, the probability of getting this much agreement by chance is less than 1 in 100. I replicated Richards’s experiment using both rating scales and essays (Martindale and Dailey 1995) and demonstrated using standard statistics used to assess agreement among raters that people agree surprisingly quite well as to the meaning of poetic texts. They agree in their interpretations of literary texts to about the extent that they agree when asked to make any other type of judgment. Humanists should be careful when they make empirically falsifiable statements. I or someone like me may come along and falsify them. Post-modernism was at least more clear than deconstructionism. Lyotard (1984)

404 Colin Martindale

explained that we no longer believe in grand totalizing theories. It was never clear what gave Lyotard the right to speak for all of humankind. I believe in grand theories, the physicists in search of the so-called “theory of everything” quite clearly believe in them, and a lot of other people do too. If Lyotard can speak for the world, then so can I: I wish that more people believed in them, as specialization always stands in the way of finding the truth. Then Lyotard decided that there is no truth or reality. How he arrived at that completely nonsensical conclusion, he did not bother to make clear. According to Lyotard, we can all construct our own truth, and everyone’s construction is equally valid. Whoops. You cannot disagree with anything I said in this article, as my truth is as good as yours. In case you did agree with me so far, let’s try cats: My cat is ten times smarter than your cat and is a direct descendant of the cats of Marie Antoinette, the beheading of whom ended civilization. Well, that’s actually true, so let’s try again. Clinton was the greatest American president since Millard Fillmore. I don’t believe that, but some insane person might. I guess we all have to believe it. Even Tertullian (“I believe because it is absurd”) is on Lyotard’s side, but we all have some common sense. The statement is simply false. Truth, in the form of paranoid delusions, came back into fashion. Now we have a variety of approaches such as the new historicism, feminism, and Marxism. The point of all these approaches is that the function of literature is to oppress some group of people: blacks and other minorities, women, the lower class. If we grant that these groups are or were oppressed, why bother to write a poem about it? They had already been oppressed by much more effective means. Even if we grant the dubious premise that literature serves the power elite (such approaches also say it can serve an emancipatory function), the most reasonable thing for writers to do would be simply to ignore oppressed groups. These approaches grant to literature a power it simply does not have. Japan rather annoyed the United States in 1941, because it attacked us. If the pen really were mightier than the sword, we would have sent a delegation of poets to Tokyo to settle the score. Bards did decide which side won a battle in ancient Celtic cultures, but we didn’t know if the Japanese knew that so thought it would be safer to express our displeasure by more conventional means of warfare. If literature does serve an oppressive function, that is, in any event, a question for scientists rather than humanists to answer. One wishes that humanists would stick to what they are good at rather than being carried away by one fad after another or trying to establish and support general theories. Only the most brilliant of them can accomplish the latter task. One also wishes that they would be more tolerant of their colleagues who are trying to formulate general scientific laws of poetics. They could use such laws rather than denigrating those who search for them. Good scholars who have attempted to study literature in a scientific way have been forced to resign by intolerant postmodern — a term no one can define — humanists. This will end, and truth will prevail. To quote Horace, “dabit deo hic quoque finem.”



The proper place of humanism 405

References Fish, S. 1980. Is There a Text in this Class. Cambridge, MA: Harvard University Press. Greene, B. 1999. The Elegant Universe: Superstrings, Hidden Dimensions, and the Quest for the Ultimate Theory. New York: W. W. Norton. Hasenfus, N., Martindale, C. and Birnbaum, D. 1983. “Psychological reality of cross-media artistic styles”. Journal of Experimental Psychology: Human Perception and Performance 9: 841–863. Heider, F. 1946. “Attitudes and cognitive organization”. Journal of Psychology 21: 107–112. Hopfield, J.J. 1982. “Neural networks and physical systems with emergent collective computational abilities”. Proceedings of the National Academy of Sciences, USA 79: 2554–2558. Kuhn, T. S. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Lakatos, I. 1978. The Methodology of Scientific Research Programs. New York: Cambridge University Press. Lyotard, J. F. 1984. The Postmodern Condition: A Report on Knowledge. Minneapolis: University of Minnesota Press. Martindale, C. 1975. Romantic Progression: The Psychology of Literary Change. Washington, D. C.: Hemisphere. Martindale, C. 1976. “Structural balance and the rules of narrative in Les liaisons dangereuses”. Poetics 4: 57–73. Martindale, C. 1989. “Personality, situation, and creativity. In Handbook of Creativity, J. Glover, R. R. Ronning and C. R. Richards (eds.). New York: Plemum. Martindale, C. 1990. The Clockwork Muse: The Predictability of Artistic Change. New York: Basic Books. Martindale, C. 1995. “Creativity and connectionism”. In The Creative Cognition Approach, S. M. Smith, T. B. Ward and R.A Finke (eds), 249–268. Cambridge, MA: MIT Press. Martindale, C. and Dailey, A. 1995. “I. A. Richards revisited: Do people agree in their interpretations of literature?” Poetics 23: 299–314. Miller, G. A. 1956. “The magical number seven, plus or minus two: Some limits on our capacity to process information”. Psychological Review 63: 81–97. Peckham, M. 1965. Man’s Rage for Chaos: Biology, Behavior, and the Arts. Philadelphia: Chilton. Praz, M. 1933. The Romantic Agony. Republished: Cleveland: Meridian, 1963. Richards, I. A. 1929. Practical Criticism: A Study of Literary Judgment. New York: Harcourt. Todorov, T. 1971. “The two principles of narrative”. Diacritics 1: 37–44.



An evolutionary framework for literary reading David S. Miall University of Alberta

Every human culture possesses a special mode of verbal behavior that can be considered ‘literary’, although in most places and at most times this has been an oral rather than a written phenomenon. However, most current critical theorists appear to accept that ‘literature’, as a body of imaginative writing with distinctive properties, is a rather modern development. Having lasted perhaps some two centuries, it is now in process of being deconstructed following a wide range of historicist and cultural analysis. In short, it is generally held that ‘literature’ emerged in the eighteenth century in order to serve the interests of an emergent middle class culture. As Richard Terry has suggested, however, the arguments tend to conflate the term ‘literature’ with the concept: citing authors such as Alvin Kernan and Terry Eagleton, he suggests they reveal “slippage from word to concept” (Terry 1997: 84). A closely related problem, as Terry’s article shows, involves asking when the literary canon came into being. Terry himself argues that the concept of a literary canon emerges around the late sixteenth century, since by this time commentators are privileging a group of creative texts (such as the works of Chaucer, Gower, Spenser, Sidney, and Marlowe) that can be delimited from the non creative. As Willie van Peer (1997) has argued, however, the processes of canonicity appear to have operated throughout history, as far back as the first ‘creative’ texts on record (Sumerian, c. 3000 BC). The nomination of the term used to label each phenomenon may thus be predated several millennia by the phenomenon itself, that is, by the existence of a select group of texts that tend to outlast the conditions of their production, or by a particular class of texts with special properties and effects. In this essay I examine what such an argument implies for literary reading. I will ask whether the experience of the literary may be fundamental to us as a species, and consider whether the proclivity for literary experience fulfils some identifiable and distinctive role. While species-specific traits are commonly thought to require fifty or more generations to develop, the evidence for literature goes back well beyond this; thus the time span for the existence of literature is more than

408 David S. Miall

adequate to propose the question: Is literary experience an adaptation, selected by evolutionary pressures because it enhanced survival and reproductive ability? In considering this question, it is important to bear in mind that the conditions under which a trait is manifested now may not provide an accurate guide to how or why the trait was acquired in ancestral conditions. In developed cultures (from Roman to contemporary western civilizations) literary experience has primarily taken the form of reading, which clearly adds a component of learned skills to that experience, likely to have modified it to some degree. Similarly, the powers of literature have at various times been systematically appropriated by religious and secular authorities for their own ends, from Bible rhetoric to modern schoolroom techniques of literary analysis. Indeed, the supposed invention of literature by the middle classes in the eighteenth century has been taken to show that literary experience embodies ideological principles and rests on nothing innate. Disentangling from these cultural formations what may be fundamental to literary experience will hardly be a simple or straightforward task. In discussing this question, whether literary reading has evolutionary significance, I will limit myself to two issues. In the first main section of the chapter I ask what the evidence is for an innate component of literature. Here I further limit the discussion to the response to literary language, or foregrounding; this stands in for a wider discussion of other distinctive features, such as tropes and narrative forms. In the second section I consider the function of literature as a dehabituating agent; in this light, I then ask what difference an evolutionary perspective might make to our research on reading, in particular, to the design of empirical studies. First, however, I briefly consider several other important approaches to the evolutionary question. Previous studies of literature within an evolutionary framework have done little more than glance at either the formalist issues or the empirical approach. In their studies Joseph Carroll and Robert Storey have, in different ways, proposed a thematics that endow literature with evolutionary significance: literature, in a word, matters because it empowers us to consider the fundamental, lifeenhancing themes of our existence. Literary works are distinctive, Carroll notes, “through their subject matter, the faculties they engage, the writer’s orientation to the subject, and the use of words as their specific medium of representation” (Carroll 1995: 104). Similarly, Storey argues that literature is fundamentally mimetic, “relying as it does upon the reader’s attempt to descry an intelligibility in human affairs” (Storey 1996: 126). Although these authors argue forcefully for the importance of literature as a special mode of representation, neither approach enables us to discriminate literary experience from other modes of discourse in which we inquire into the “intelligibility” of our lives. I suggest that an argument for the evolutionary role of literature must be founded on more than literary content. It is the formal properties of literature and the responses these evoke in the hearer or reader that most clearly characterize literature, setting it apart from discourse in general.

An evolutionary framework for literary reading 409

Ellen Dissanayake, who has written extensively on the evolutionary significance of the arts, suggests that works of art promote what we might call a “defamiliarizing” mode of mind, a “making special” (Dissanayake 1992:50). In premodern societies this multimodal experience (involving several art forms) prepared the individual for recognizing and participating in an unusual experience: developed at first, perhaps, for encountering the sacred and the rituals that developed around it, literary experience may have developed its own specific characteristics, coming in time to incorporate verbal and narrative cues to alert the hearer to adopt a special mode of attention. Dissanayake notes that many cultures make use of specific devices to signal poetic utterance, such as an unusual tone of voice (Dissanayake 1992: 113–6). Internalized in the texture of language as foregrounding we now recognize these cues, in part, as giving written literature its distinctiveness as a medium. A content-directed approach may place too much emphasis on meaning. This argument can be illustrated by an analogy. Literature invokes processes in the reader somewhat as a migrating bird depends on its navigational system. The bird does not set out with a fixed goal that it aims to reach: its orientation is guided by reference to such environmental signals as geographical landmarks, terrestrial magnetism, the sun, and stars, all of which provide the bird with a goal-tracking system. It is this content-knowledge that modulates the migratory process of the bird, but in order to understand that process we need to know not what the bird understands about magnetism or the sun but how its systematic use of this information creates a guidance system. Similarly, the literary reader, while knowing there may be a goal to be reached (that is an interpretation of a text that is appropriate for that reader), cannot set out knowing in advance what that goal is, in the way that the reader of a repair manual or a chemistry textbook can be goal-oriented; moreover, interpretation may not even be a goal for the reader who reads for the pleasurable experience of reading rather than for meaning. Literary reading is guided, like the migrating bird, by an array of navigational markers, such as the palette of phonetic features, significant tropes, or narrative cues, and it is these that enable readers to attain their goal. Readers do not need knowledge of phonetic tone colors, or even need to be aware of their role during reading. As literary readers, in other words, we deploy a set of “content-sensitive” processes (Tooby and Cosmides 1992: 34) endowed on us by evolution, but fulfil these in ways peculiar to our own needs and historical context. In the discussion that follows, therefore, I first briefly lay out some evidence for attention to foregrounding, suggesting that this is a distinctive feature of human development from infancy onwards, predating literary experience as such. My comments are intended to be representative, since foregrounding is only one of several formal aspects that should be explored for their evolutionary significance: other major domains of inquiry include figurative structures (analyzed, although not in an evolutionary context, by Turner [1991]), and the formal components of narrative (e.g.

410 David S. Miall

Fludernik 1996). Then I examine the “defamiliarizing” process of mind that appears central to literary experience, and consider what evolutionary implications it might possess. I propose that literary experience considered formally can be understood as dehabituating, having emerged in recent human evolution as an adaptive solution to some specific sensory and cognitive limitations in human functioning.

1.

Form and foregrounding

The claim that literary texts characteristically exhibit a special use of language, or foregrounding, has been in dispute for several decades, a dispute initiated in particular by Stanley Fish’s (1980) attack on stylistic methods of analysis in a paper first published in 1973. Arguments against literary language have typically taken one of three forms: first, that distinctive features (alliteration, metaphor, and so on) are as common in nonliterary as in literary texts; second, that such verbal features provide no formula for reaching an interpretation, that is that they are devoid of the kinds of meaning that stylistic critics have attempted to build upon them; or, third, that if we pay attention to foregrounding it is solely because we have been schooled into doing so. Each of these arguments deserves careful consideration (for discussion see Miall and Kuiken 1998, 1999), but for the present purpose I will point out only that on both sides of the debate the issue of reception has been largely overlooked. Thus, my concern here is with the question, what difference does it make to the reader who encounters such features, whether in a text designated literary or not. For the argument about the existence of literary language to be plausible, we must demonstrate that a distinctive kind of processing during reading corresponds to the presence of foregrounding. If we find evidence of such processing, we have then still to establish whether it is put in place by the reader’s literary education or is a sign of an intrinsic capacity for literary response. On the first issue, we now have some evidence for a distinctive mode of processing. Our studies (Miall and Kuiken 1994), which were built in part on those of Van Peer (1986), focused on readers’ responses to literary short stories in which we had previously analyzed the occurrence of foregrounded features. We found that readers typically took longer to read passages containing foregrounding, with longer reading times corresponding to the most highly foregrounded passages. At the same time, readers appeared to consider such passages more striking, productive of more feeling, and more uncertain in relation to the unfolding meaning of the text, as shown by their ratings of each passage. We found some evidence that this first phase of response, which can be termed defamiliarizing, is followed by a constructive process on the part of the reader that appears to center on the feeling associated with the foregrounded feature: such feeling in time puts in place an alternative framework for interpretation, which contributes to the new perspective opened up

An evolutionary framework for literary reading

by the reading of the story as a whole (cf. Miall 1989; Miall and Kuiken in press). It is this phasic process of response, located in relation to specific textual features, that we have proposed as typical of literary reception (Miall and Kuiken 1999). What evidence is there, however, that this process is intrinsic rather than induced pedagogically? Two lines of argument can be adduced. First, according to the cultural relativist position espoused by Stanley Fish (1980), Barbara Herrnstein Smith (1988), and others, the degree to which readers are attentive to foregrounding should be a product of how much literary education they have received. The evidence against this position is not compelling, but serves to call it into question. We compared the responses to foregrounding of advanced students of literature and first year students of psychology who, as we found, had little interest in or experience of literary reading. We found that no difference occurred between the groups in the degree to which lengthened reading times correlated with foregrounding, a finding that showed our nonliterary students to be equally attentive to foregrounded passages. In his study of response to foregrounding in several poems, Van Peer similarly found no difference between his three groups of participants, who ranged from students of stylistics to science students with minimal training in literature (Van Peer 1986: 114–5). In a study of response to metaphors in a literary and a newspaper text, Steen found that expert and less-expert readers (scholars of literature and anthropology, respectively) paid attention to metaphors about equally, with both groups consistently paying more attention to metaphors in the literary text (Steen 1994: 144). These findings suggest, contrary to the arguments of the cultural critics, that the initial response to foregrounding may be independent of literary training or experience. A second line of evidence for the literary significance of foregrounding is to be found in genetic studies, which suggest that a sensitivity to such verbal devices may be inborn. For example, in a study I carried out with Ellen Dissanayake, we analyzed a recording of a mother’s verbal interactions with her 8-week-old baby, Liam (recorded at the laboratory of Professor Colwyn Trevarthen at the University of Edinburgh). The mother’s language, deployed to engage or sustain her baby’s attention, clearly demonstrates a range of foregrounded features including local effects, such as alliteration and assonance, metrical effects, and figurative expressions, as well as larger scale patterns that in poetry would be classified as verse lines and stanzas. What is particularly striking in the interaction is the extent to which high, front phonemes in the mother’s speech coincide with maximal attention of the infant, suggesting intimacy, during which the mother addresses the infant directly; in contrast, low, back phonemes occur when the infant is distracted or inattentive, and the mother is more likely to mention the infant’s appearance or surroundings. Since the infant at this age has no semantic grasp of language, his attention is mediated by the purely aural pattern of repetitions and differences produced by the mother: in both, a mutual response process is evoked that appears

411

412 David S. Miall

to be characterized by subtle kinaesthetic and affective modulations in the infant that are attuned to within a third of a second. Once the infant himself begins to generate language a notable feature of spontaneous speech is, once again, a type of language rich with foregrounding. The most detailed available study is that of a boy of two and some half years whose spontaneous language play after he was put to bed was recorded and analyzed by his mother, the linguist Ruth Weir, in Language in the Crib. Weir shows in detail the arrays of non-referential features in the child’s soliloquies, including alliterations, play with syntax and word forms, and a structure of extended utterances that Weir terms “rondos.” The final example transcribed in Weir’s book is described by Jakobson, in his prefatory remarks to the book, as a “beautiful poetic composition” (Weir 1962: 20). In these sequences, as Weir observes in the summary of her study, the child enjoys “play with words, by repetitions of similar sounds, by his rediscovery of what is familiar to him”, creating what she calls “a dialogue spoken by a single person” (Weir 1962: 144, 146). While the words are derived from the child’s daily experience, when alone in his crib the child fashions an aural world out of them whose development is principally dependent on syntactic and aural variation. Evidence of this kind goes some way toward answering the question, raised by Van Peer in the conclusion of his study, how children come to understand the function of foregrounding. “What is at stake here, is the anthropological status of foregrounding in particular, and of literature in general” (Van Peer 1986: 181). At this pre-literary stage in infancy, children’s pleasure in wordplay is manifested in numerous forms. In each case, as the first example of the mother’s speech shows, verbal texture creates and maintains a special state of being, marked by a distinctive mode of attention and suffused with feelings both bodily and affective. Each state is sufficient to itself, although it undoubtedly contributes toward a repertoire of dispositions of the self — the self in relation to itself and to others. Once literary experience itself becomes possible (beginning, perhaps with the first stories read by parents), it seems likely that such dispositions will be evoked by the same verbal means, but that now they are placed in relation to other states and processes brought into play by the text. In this context a familiar disposition may come to seem less familiar, being subjected to critical revaluation in the light of alternative perspectives. Without the initial responsiveness to the creative power of verbal textures, beginning in the cradle, the variations presented by literary texts would possess no meaning. In brief, while pre-literary experience creates dispositions and states, literary experience begins to frame these in perspective-altering ways. These examples seem to point to the possibility of an innate capacity for foregrounding, leading to the modifying experiences of literature. We might also consider another type of genetic evidence, that of oral literature, since it seems probable that contemporary examples of such literature are comparable to the preliterate epoch of all human cultures. The appearance of foregrounding in oral

An evolutionary framework for literary reading 413

literature cannot be attributed to the practices of a systematic educational programme in literature, although as Ruth Finnegan makes clear, each culture creates its texts on the basis of a distinctive set of conventions. Such conventions frequently make the use of a specific type of foregrounding obligatory, such as the alliterative patterns of Somali verse (Finnegan 1992: 94), or metrical units that create a strict pattern of parallelisms in Toda songs (Finnegan 1992: 99), or an organizing metaphor in the Polynesian poem (Finnegan 1992: 114). Finnegan suggests that the special diction found in oral literature arises both because it fulfils formal needs, and because it serves to put a ‘frame’ around a poem marking it off from the language of ordinary life (Finnegan 1992: 110). Although the palette of foregrounded features varies from one culture to another, its role in the examples cited is evidently important and persuasive, functioning with a power equal to its role in some of the best written poetry of the last few centuries. As Finnegan also points out, however, oral poetry is likely to be a multimodal experience: it is performed, thus achieving through the dramatic role of the speaker with an audience effects that in written literature must depend upon verbal texture alone or its appearance on the page. She also stresses that its functions vary considerably from one culture to another: for example, some poems are said to be communications from a god or from a dead relative; others are interventions in current social relationships designed to influence the standing of the speaker. Neither oral literature, nor the word play and other early verbal experiences of infants and children, has either a particular function or a single relation to the culture of its participants. We have no reason to think that written literature is significantly different (which make disputes over whether literature has a transcendental or ideological purpose redundant). What marks each of the examples I have mentioned, including the literary stories that were the focus of the empirical studies I described earlier, is a reception process initiated by and dependent upon foregrounding. In each case, a specific feature — or, more usually, a constellation of features — attracts attention by virtue of its distinctiveness in comparison with other, ordinary uses of language (the meaning of ‘ordinary’ here will be elaborated below). The state of feeling, or disposition, evoked in the hearer appears to be, at least in part, a consequence of the verbal feature. For example, an alliteration of front consonants (liquids or nasals) may invoke a sense of intimacy and affection (as it clearly does in the example of the mother’s baby talk); the meter of a song may evoke the lightning speed of the superior hunter in his final chase. Whatever the precise mode of attention, the moment of response to foregrounding promotes a familiar experience to the status of being special (in Dissanayake’s terms), and opens it to the possibility of being re-evaluated. In this process, the feeling now in place may outlast its occasion, either allowing it to govern the emergence of some new cognitive formation, or alternatively, to be relocated within a perspective that will modify understanding of it.

414 David S. Miall

The contrast of a special language containing foregrounded features with ‘ordinary’ language is not uncontentious, however. Whether literary language contains more unusual features such as alliteration or metaphor than other types of language has been disputed, and is evidently not the case with some types of literary texts. Eagleton sums up his dismissal of the claim by asking us “to face the fact that there is more metaphor in Manchester than there is in Marvell” (Eagleton 1983: 6), a clause that, notably, enforces its point by alliteration, unlike the surrounding sentences (presumably a rhetorical flourish intended by Eagleton). The conclusion my discussion points to, however, is that arguments such as Eagleton’s miss the point: special uses of language are only special if they are noticed, and, in the case of literature, if they systematically modify the hearer’s understanding (this proposal, of course, extends to other forms of modifying understanding, such as narrative). Literary reception, I have suggested, is marked by a special, phasic type of processing, in contrast to the reception of a metaphor in Manchester, which may be incidental to the discourse in which it occurs. A person who uttered a series of metaphors at a bus stop in Manchester would be considered either mad, or engaged in some new kind of performance art. ‘Ordinary’ language requires a type of decorum, one significant feature of which is its avoidance of foregrounded features organized to promote attention to meaning. In the next section I take this contrast between ordinary and literary language as the basis for suggesting where the evolutionary significance of literary response may lie.

2.

The dehabituating role of literature

Central to the arguments of evolutionary psychology is the claim that whatever psychological mechanisms the human race exhibits now were developed in response to our prehistorical existence as hunter-gatherers in the Pleistocene epoch. Our adaptedness as a species is a reflection of the environment that prevailed some thirty to a hundred thousand years ago, a context that must be considered in attempting to understand any significant human capacity — including that for literature. Thus literature as a response process cannot be understood only in terms of recent cultural developments, or even as a product of middle class ideology across the last two or three centuries. As Bowlby remarks, we must be attentive to “the fact that not a single feature of a species’ morphology, physiology, or behavior can be understood or even discussed intelligently except in relation to that species’ environment of evolutionary adaptedness” (Bowlby 1974: 64). Literary reading, in this perspective, must be understood as a response to the ancestral environment and the cognitive, emotional, and social challenges that it posed. So what kind of adaptation is literary response? How might it promote the inclusive fitness of those early humans who adopted its practice? I will argue that it

An evolutionary framework for literary reading

confers a number of benefits, but that each can be understood in terms of a theory of dehabituation. Response to literature promotes an offline tuning of emotional and cognitive schemata, with a particular focus on resetting the individual’s readiness for appropriate action. The central observation is that literature facilitates this process through an array of formal features. Thus, the present account can be seen as a neoformalist theory. In the previous section I have advanced a case for foregrounding, but equally important evidence for figurative and narrative structures should also be assessed, structures that contribute equally important formal building blocks to the distinctive properties of literature. Overall, it can be proposed that it is the reception process initiated by these structures that makes literary experience unique. Whereas language in its ordinary uses, as in much everyday conversation, discourse in newspaper journalism, or school textbooks, has the primary function of elaborating information based on prototypical concepts and readers’ existing schemata, literary forms disrupt this essential function for reasons which may be equally essential. While effective behavior, particularly in the ancestral environment, typically depends on rapid assignment of meaning to appearances following their assessment in relation to the interests of the self or the group, the tendency inherent in this facility is to stereotypic concepts and stock responses. Literary experience, which takes place outside the normal demands of daily life, enables stereotypic concepts and responses to be put in question. Through literature readers or hearers may evolve new modes of feeling for the issues that are most central to their experience; as I also suggest below, literature may facilitate the modulation or repair of emotionally negative experiences in particular. By dehabituating, in brief, we prepare ourselves for encountering experience in ways that are potentially (although not necessarily) more productive, thus enhancing the flexibility of our responses to the environment or our social interactions. Why might the evolutionary approach matter? Speaking most generally, as a mode of inquiry, it enables us to read our present capacities and accomplishments as a solution to past problems. Rooted in capacities that appear to be innate to the human system for the past several thousand years of history (and probably much longer), it suggests a provisional answer to what literary experience might be for. Although the modern literary theorist may accept the truth of the evolutionary argument, however, it may be objected that this is hardly relevant, given the inherent power and prevalence of language and culture now. Has not culture developed its own momentum, based on its own laws that now has little or no relation to the ancestral environment? Culture, this argument goes, is built from other cultures; literary texts are made from other texts, and have no substantial relation to the ‘natural’ world. As Roland Barthes puts it, texts are “woven entirely with citations, references, echoes, cultural languages”, their mode of being lying in the “infinite deferment of the signified” (Barthes 1977: 160, 158). How could an evolutionary psychology have any bearing on literary culture, when culture has long since written

415

416 David S. Miall

over and effaced whatever “natural” functions human beings inherited? Moreover, since we are not born with literary competence, its acquisition must depend not only on developing a competence in language, but also derive from educational and cultural influences in the environment that are not experienced until later childhood: literature, in other words, is entirely a product of the culture and can have nothing ‘innate’ about it. It is in this perspective that the “interpretive community” is emphasized by literary theorists such as Stanley Fish (1980). However, the distinction made here between what is due to nature and what to culture is a specious one: every human phenomenon is derived from evolved psychological mechanisms with specialized functions and inherited forms of representation, modified by interaction with a particular local culture and environment. As Bowlby has noted, “Just as area is a product of length multiplied by width so every biological character whether it is morphological, physiological, or behavioral, is a product of the interaction of genetic endowment with the environment. Terms like innate and acquired must therefore be cast into limbo and a new terminology employed” (Bowlby 1974: 38). Evolutionary theorists Tooby and Cosmides also argue that the distinction is incoherent and “should be consigned to the dustbin of history, along with the search for a biology-free social science” (Tooby and Cosmides 1992: 46). To locate the agency for cultural phenomena outside the person and the rich array of psychological functions that individuals possess in common, they argue, is to mystify the processes in question: there is “no radical discontinuity inherent in the evolution of ‘culture’ that removes humans into an autonomous realm” (Tooby and Cosmides 1992: 119), as the linguistic premises of poststructuralist theory and it’s ‘deferred signified’ imply. As Tooby and Cosmides also note, “cognitive architectures that are passive vehicles for arbitrary semiotic systems are not plausible products of the evolutionary process” (Tooby and Cosmides 1992: 109). Many postmodern accounts of cultural phenomena perpetuate the Cartesian distinction between mind and body, refusing to see the mind as governed by the same principles as the body: “the mind should contain ‘mental organs’ just as the body does”, note Tooby and Cosmides (Tooby and Cosmides 1992: 57). The evidence from evolution suggests that the mind is actually constituted in part by a range of functionally specialized, content-dependent mechanisms, or adaptive specializations (for a recent list, see Buss 1995: 6). It is unlikely that literary response represents an exception: response to literature, in other words, is likely to depend either on an array of domain-specific modules in the mind , or, as may be more probable, constitute in itself a domain-specific module with its own determining mechanisms that underlie its many cultural and historical variations. This does not imply that literature is a closed or fixed system: biological determination does not mean constraint or inflexibility. On the contrary, the domain-specificity of literature is the ground on which has flowered the extraordinary range of literary

An evolutionary framework for literary reading 417

phenomena apparent across history, while giving literary response a set of core functions that has ensured its central place in human society up to the present era. Why the literary mechanism represents an effective solution to human adaptive problems can also be understood in terms of our identity, both individual and social. Literary response may represent a solution to social constraints on the expression of emotion (cf. Nesse and Lloyd 1992), and, in the context of the awareness of contingency, change, and death that first emerged with our ancestors, it also offers a medium for reflecting on potential alternative identities. In this respect, also, literature is to be distinguished from ordinary discourse. The primary function of discourse is to instantiate common interpretive schemata, for which emotional response and self-awareness arise only incidentally and are generally stereotypical; but a text that only instantiates a frame with relevant information is not literary. Literature, in contrast, facilitates changes in perception or in the self in its relationship with others, thus enhancing the survival and reproductive ability of the group. This external factor alone may have been sufficient to favor the selection and perpetuation of the capacity for literature, but literature can also be seen as a solution to an endogenous adaptive problem, that of social constraint, repression, and pathology. What began as a communal experience, especially shared narratives, dramas, and literary components of play or ritual (Dissanayake 1992: 48), may over time have evolved in part to address the singular internal needs of individuals. Literature evolved in particular, perhaps, because it spoke to what was individual in the individual. Where, in other words, powerful feelings, such as love, bereavement, loss, or trauma set the individual apart, forestalling the benefits of solidarity and communication with others, the emotion can be engaged and brought into focus through literary experience. As studies of our own with bereaved participants have suggested, new meanings can then be found that help the individual to accommodate the emotion and come to terms with it. In this way, too, literature enhances our abilities to respond flexibly to experience and thus assists our powers of survival. Two salient implications for empirical study appear to follow from the evolutionary framework. First, it should be possible to specify dehabituation theory in relation to hypotheses that predict specific aspects of the relationship between literary texts and readers’ behavior and that can be tested empirically. Examples of such hypotheses will include: –

–

Literary texts contain attentional framing devices. Their presence will be signified by increased demands on processing. The discussion offered earlier of foregrounding, whose presence correlated with reading times, is one example of such a device. Another is Johan Hoorn’s (1996) neuropsychological study of semantic and phonetic deviation at verse endings, which correlated with increased N400, an electrophysiological measure of unexpected events. Increase in stylistic novelty over time is also predicted by dehabituation theory. As Colin Martindale has shown in a number of outstanding studies (e.g. The

418 David S. Miall

–

Clockwork Muse, 1990), once a literary style has been established (as, say, at the beginning of the British Romantic period) writers must work progressively harder to create novelty that will attract and hold the attention of readers; they do so by shifting increasingly to the use of primary process thought in their writing. Special processing of negative emotions will occur during literary response (a process that we might term catharsis), a hypothesis based on the assumption that adaptive value accrues from attending to negative emotions that tend to be repressed in the familiar social settings for a self-expression. In a study in which we compared the valency of judgements for everyday experiences of settings compared with settings in literary stories, we found a significant shift toward negative evaluations in responses to the stories (Corrêa et al. 1998).

The second implication is that the readings of literary texts produced by literary scholars, interesting though they may be in themselves, have little value for empirical study, except when they suggest specific hypotheses about literary processing that are amenable to empirical examination. In the terms of neuropsychologist Gerald Edelman, only ‘population thinking’ can account for the selection and development of a species trait: “population thinking states that evolution produces classes of living forms from the bottom up by gradual selective processes over eons of time” (Edelman 1992: 73). For literature, the question is thus what function literary reception may have in populations of readers, not individuals — although the evidence can only be studied as it is manifested within the responses of a range of actual individual readers. This suggests that the current emphasis on critical and historicist interpretations of texts, important though these have sometimes been, has obscured understanding of what may be characteristic and universal in literary response. A population of readings, in contrast, may tell us a great deal about the lawfulness of literary response if the group of responses examined appears to vary systematically in relation to certain significant variables in the text-reader interaction. Thus empirical study of real readers, guided by defensible scientific hypotheses about the adaptive functionality of literature, appears to be the most productive paradigm for future research.

References Barthes, R. 1977. Image-Music-Text, Stephen Heath (trans). London: Collins/Fontana. Bowlby, J. 1974. Attachment. London: Hogarth Press. Buss, D. 1995. “Evolutionary psychology: A new paradigm for psychological science”. Psychological Inquiry 6: 1–30. Carroll, J. 1995. Evolution and Literary Theory. Columbia and London: University of Missouri Press. Corrêa, A. A., Miall, D. S. and Kuiken, D. 1998. Response to Environments and to Literary Texts: The Role of National Identity. Paper presented at the VIth Biannual Conference of the Association for the Empirical Study of Literature (IGEL), Utrecht, August 26–29.



An evolutionary framework for literary reading 419

Dissanayake, E. 1992. Homo Aestheticus: Where Art Comes From and Why. Seattle: Washington University Press. Eagleton, T. 1983. Literary Theory: An Introduction. London: Routledge. Edelman, G. M. 1992. Bright Air, Brilliant Fire: On the Matter of the Mind. New York: Basic Books. Finnegan, R. 1992. Oral Poetry: Its Nature, Significance and Social Context. Bloomington, IN: Indiana University Press. Fish, S. 1980. “What is stylistics and why are they saying such terrible things about it?” Is There a Text in This Class? The Authority of Interpretive Communities, 68–96. Cambridge, MA: Harvard University Press. (Essay originally published 1973.) Fludernik, M. 1996. Towards a ‘Natural’ Narratology. London & New York: Routledge. Hoorn, J. F. 1996. “Psychophysiology and literary processing: ERPs to semantic and phonological deviations in reading small verses”. In Empirical Approaches to Literature and Aesthetics, R. J. Kreuz and S. M. MacNealy (eds), 339–351. Norwood, NJ: Ablex. Martindale, C. 1990. The Clockwork Muse: The Predictability of Artistic Change. New York: Basic Books. Miall, D. S. 1989. “Beyond the schema given: Affective comprehension of literary narratives”. Cognition and Emotion 3: 55–78. Miall, D. S. and Kuiken, D. 1994. “Foregrounding, defamiliarization, and affect: Response to literary stories”. Poetics 22: 389–407. Miall, D. S. and Kuiken, D. 1998. “The form of reading: Empirical studies of literariness”. Poetics 25: 327–341. Miall, D. S. and Kuiken, D. 1999. “What is literariness? Three components of literary reading.” Discourse Processes 28: 121–138. Miall, D. S. and Kuiken, D. In press. “Shifting perspectives: Readers’ feelings and literary response.” In New Perspectives on Narrative Perspective, W. van Peer and S. Chatman (eds). New York: SUNY Press. Nesse, R. M. and Lloyd, A. T. 1992. “The evolution of psychodynamic mechanisms”. In The Adapted Mind: Evolutionary Psychology and the Generation of Culture, J. Barkow, L. Cosmides and J. Tooby (eds), 601–624. New York: Oxford University Press. Smith, B. H. 1988. Contingencies of Value: Alternative Perspectives for Critical Theory. Cambridge, MA: Harvard University Press. Steen, G. 1994. Understanding Metaphor in Literature. London: Longman. Storey, R. 1996. Mimesis and the Human Animal: On the Biogenetic Foundations of Literary Representation. Evanston, IL: Northwestern University Press. Terry, R. 1997. “Literature, aesthetics, and canonicity in the eighteenth century”. EighteenthCentury Life 21: 80–101. Tooby, J. and Cosmides, L. 1992. “The psychological foundations of culture”. In The Adapted Mind: Evolutionary Psychology and the Generation of Culture, J. Barkow, L. Cosmides and J. Tooby (eds), 19–136. New York: Oxford University Press. Turner, M. 1991. Reading Minds: The Study of English in the Age of Cognitive Science. Princeton, NJ: Princeton University Press. Van Peer, W. 1986. Stylistics and Psychology: Investigations of Foregrounding. London: Croom Helm. Van Peer, W. 1997. “Two laws of literary history: Growth and predictability in canon formation”. Mosaic 30: 113–132. Weir, R. 1962. Language in the Crib. The Hague: Mouton, 1962.



High and popular culture from the viewpoints of psychology and cultural studies Gerald C. Cupchik and Garry Leonard University of Toronto

The collaboration between scholars trained in psychology and in cultural studies is fraught with challenges, the main one being a difference in frames of reference and related discourses. While psychologists search for universal principles that govern cognition, emotion, and behavior, scholars interested in cultural studies examine the particular historical contexts in which works of art, literature, film, advertisements, and so on, are created and interpreted. The complementary relationship between these viewpoints is best revealed when common phenomena are brought into focus, albeit by different lenses and filters. The relationship between ‘high’ and ‘popular’ culture provides one such problem area which can benefit from a reconciliation of viewpoints. This is because concepts such as “high/low or art/ popular culture are sociohistorical constructions” (Fornas 1995: 145) that comprise interrelated sets of meanings and cannot be reduced to simple definitions. These meanings describe the relevant artefacts, their creators, the contexts in which the cultural forms originated, and their current manifestation. While psychology can describe general processes underlying the reception of high and popular art, cultural studies can provide a specific historical grounding.

1.

Psychological foundations

1.1 High culture Fornas (1995) has argued that a ‘high’ culture community emerged when “priestly, aristocratic and bourgeois elites withdrew from the formerly communal folk culture typified by the carnival. Then the bourgeois public sphere was dichotomized into entertainment versus serious arts…” (Fornas 1995: 145). These serious arts are

422 Gerald C. Cupchik and Garry Leonard

“produced by talented individuals who follow rules (…) which both they and their audience know and respect” (Fishwick 1974: 2). The renaissance “cult of genius” gave rise to signed artworks that possessed both “authenticity” and an “aura” (Benjamin 1967) because they were created directly by an artist. Writing about the twentieth century, Clement Greenberg (1957) continued this romanticized image of artists associating high art with the avant-garde
: “Retiring from public altogether, the avant-garde poet or artist sought to maintain the high level of his art by both narrowing and raising it to the expression of an absolute in which all relativities and contradictions would be either resolved or beside the point: ‘Art for art’s sake’…” (Greenberg 1957: 99). In short, high culture is generally associated with idealized artists who produce distinctive works that are appreciated by knowledgeable viewers. The concept of ‘high culture’ also embodies a certain conservative ideological position emphasizing stability in our social system. Pascal, writing in the 1700’s, argued that people must learn to “stay quietly in their own chamber” (in Lowenthal 1957: 48) and was against any form of hollow relaxation and amusement. Goody (1986) maintained that writing makes it possible to codify ideas and values that will lay down standards of behavior. Accordingly, “there are rules to which most citizens presumably adhere. These are rules for raising children and bringing up the young, for law-abiding conduct and work ethos, for service in the community and for not meddling with others’ freedom” (Van Peer 1997: 34). Thus, the notion of High Art encompasses a family of ideas ranging from the authentic, original artefact produced by a talented artist/author to the knowledgeable audience who share an interpretive code with the artist/author as literate and responsible citizens. 1.2 Popular culture Popular Culture, on the other hand, includes a broad range of artefacts: “comic strips and comic books, advertisements, joke cycles, fads, radio and television programs, films, fashion, foods, popular fiction; what Malinowski termed the ‘imponderabilia’ of everyday life” (Berger 1980: 16). It represents “an unflinching look at the real world today; a fascination with and acceptance of our mechanized, trivialized, urbanized environment; a mirror held up to life, full of motion and madness” (Berger 1980: 9). According to Fishwick, “Popular culture is practice and event-oriented — it leans…towards percept, not concept” (Fishwick 1974: 6). This emphasis on experience makes it an “everything-all-at-once style” for which “Nothing is too old, too new, too obscure, too banal, too distant, too close…” (Fishwick 1974: 11). Popular Culture has always been there in the “inherited oral wisdom” of folk cultures passed on in tales, songs, dances, games, and manners (Fishwick 1974). Lowenthal (1957) has traced the emergence of popular culture to Montaigne in the sixteenth century who speculated that people without faith would feel lonely in a

High and popular culture from the viewpoints of psychology and cultural studies 423

post-feudal world. To address this loneliness he advocated diversion which became a hallmark of the modern age along with “escape, distraction, entertainment, and, last but not least, vicarious living” (Lowenthal 1957: 48). According to this reasoning, “What we call low culture takes its place within the diversity of such pleasurable activities, forming an antidote against the strains of our responsibilities. It is a reservoir of revitalization, and a rescue strategy for regaining a balance” (Van Peer 1997: 34). Chaney (1994) has pursued the implications of marginalization suggesting that communities engaged in Popular Culture feel excluded from High Culture and give voice to these feelings (as in many rebellious rock music videos). A clear contrast exists between scholars who favor the study of Popular Culture and those who focus on High Culture. For the “cultural populist”, “the symbolic practices of ordinary people are more important analytically and politically than Culture with a capital C” (McGuigan 1992: 4) and scholars who are interested in the British approach to “cultural studies” have “An appreciative non-judgmental attitude to ordinary tastes and pleasures” (McGuigan 1992: 4). Conservative critics (Benjamin 1967), on the other hand, see the “contemporary masses” as accepting reproductions as replacements for the real thing (that is, original artworks) because of a desire to bring objects closer (that is, into their own home) in order to seek distraction. Lowenthal (1957) treats the phenomenon of popular culture as “nothing but a manipulated reproduction of reality as it is; and in so doing, popular culture sanctions and glorifies whatever it finds worth echoing” (Lowenthal 1957: 49). He contrasts the “differences between popular culture and art, between spurious gratification and a genuine experience as a step to greater individual fulfilment” (Lowenthal 1957: 51). Similarly, Clement Greenberg (1957) reduced Popular Culture to kitsch or “Ersatz” culture, designed for people who are “insensible to the values of genuine culture” (Greenberg 1957: 102); it is mechanical, operating by formulas, providing vicarious experiences and faked sensations. 1.3 Mass culture The apparent conflict between notions of High and Popular (or pejoratively speaking, ‘low’) Culture cannot be complete without mentioning the theme of Mass Culture in relation to mass communication. Marshall McLuhan stated that the medium is the message, which affirms the distinctive qualities of evolving modes of communication. While Gutenberg’s press made texts available to the public, television offered immediacy, high quality simulation, and, with the advent of cable television, both immense choices and access to the masses. Of course, the medium does not in and of itself distinguish between High and Popular Culture. While ‘public television’ makes High Culture programs, such as opera and drama, available to the public, the many privately owned speciality channels cater to diverse tastes ranging from sports to pornography. The medium has become the messenger. With the

424 Gerald C. Cupchik and Garry Leonard

excesses of interactivity, as evidenced by telephone communication between embattled students and CNN during the Littleton Colorado high school massacre of 1999, the medium, like the serpent Ouroboros, has begun to consume itself! Chaney (1994) extended Benjamin’s analysis of the implications of mechanical reproduction to industries of mass distribution. The early popular interest in spectacular forms of entertainment (that is presumably in Greek and Roman times) is recreated because modern technology produces high quality “illusions of simulated reality” (Chaney 1994: 208) for the communication of information and entertainment. A crucial step in his argument is that the technology industry requires “capital investment in cultural production”, “complex distribution networks”, “enormous advertising budgets”, and so on, which produce “pressures of agglomeration.” The emergence of “global communication and entertainment corporations will effectively oligopolise mass culture” (Chaney 1994: 209) leading to an emphasis on consensualist values by showing people simulations of normality. Fornas (1995) similarly expressed concern about the constraining effects of economics on Mass Culture, arguing that “Subsuming work under capital increases its productive efficiency while making it instrumental, goal-oriented and technocratic, which produces problematic side effects for human beings…” (Fornas 1995: 143). Indeed, writing immediately after the Second World War, Lowenthal (1957) voiced a concern about the use of Mass Culture by the Fascists. “There is an interdependence between what the public wants and what the powers of control enforce upon the public in order to remain in power” (Lowenthal 1957: 55). It is here that Popular Culture spills over into Mass Culture and mass capitalism: “in all its media popular culture proves to have its own genuine characteristics: standardization, stereotype, conservatism, mendacity, manipulated consumer goods” (Lowenthal 1957: 55). 1.4 Implications The sometimes acrimonious debate regarding the relative merits of High and Popular Culture appears to be founded on the assumption that High Culture belongs to the literate ‘upper’ classes, while Popular Culture belongs to the ‘lower’ classes. An important first step in reconciling concepts requires that cultural interests not be confounded with socio-economic status. The contrast between ‘high’ and ‘popular’ culture is a social construction which masks their interpenetration (Leonard 1998). “‘Serious’ culture, the kind of culture which becomes canonized as ‘art’, has always refreshed itself from the springs of popular culture, as the Russian formalists pointed out many years ago with their idea of ‘the canonization of the junior branch’” (McGuigan 1992: 3). Indeed, poorer segments of society found both meaning and pleasure in depictions of the life of Jesus Christ on the walls of Italian churches. In a complementary manner, realist painters have found valuable motifs

High and popular culture from the viewpoints of psychology and cultural studies 425

in the mundane aspects of life. Indeed, artists and writers themselves can be both poor in social condition and rich in imagination and creative talents. One and the same art or literary work can be approached in a High Culture or Popular Culture manner; this is essentially the difference between viewing a work relationally or approaching it in its uniqueness. When a work is viewed from a High Culture perspective, its subject matter or stylistic properties are approached intellectually and analyzed in relation to works that preceded them (Cupchik and Kemp 2000). It emphasizes originality and the extent to which a new work extends or transforms an earlier style. Original stylistic transformations do not reside in individual works of art (or literature), but are inferred from the shared qualities (that is, family resemblance) underlying a series of works by a single artist/author or by a group. The skill at discerning stylistic developments is comparable to learning to discriminate new visual or literary codes. This intellectual grasping of transformations is very different from the ‘objectivity’ espoused by New Criticism and its assumption of a ‘correct’ interpretation of isolated texts. When a work is approached from a Popular Culture perspective, it is experienced in its particularity in a social/historical context. Popular Culture does not so much challenge the recipient but stimulates different kinds of emotional reactions and personal associations. It may evoke warm feelings of pleasure, as in the case of sentimental emotions evoked by romantic films, or excitement in relation to suspense and drama. As an experientially oriented process, Popular Culture provides a means for projecting emotion onto characters or identifying with them and experiencing catharsis (Cupchik 1997). The collective aspect of audience involvement facilitates the expression or catharsis of emotion. Generally speaking, Popular Culture experiences become associated with ongoing events in the recipient’s life. Rather than analyzing a work in relation to ones that preceded it, a subjective approach extends beyond the work to the personal context. For this reason, people from different generations may have contrasting emotional responses to the same pieces of art, literature, music, film, or video which were encountered at particular stages of their lives. Psychological research has offered a framework within which to think about these complementary processes. Cupchik and Gebotys (1990) showed that challenge and pleasure are two factors that underlie comparative judgments of pleasingness in artworks. The challenge factor encompassed modern (that is post-1850) works that were judged to be unfamiliar, complex, and yet meaningful. The pleasure factor included paintings with positive themes that were perceived as intense, warm, and yet simple. This preference for warm artworks that evoke pleasurable associations supports Berlyne’s (1971, 1974) notion that people generally prefer moderate levels of complexity. Arnheim (1985), on the other hand, stressed the importance of intellectual challenge and an appreciation of stylistic complexity. Challenge and pleasure have also been shown to be two motives which govern

426 Gerald C. Cupchik and Garry Leonard

how students approach artworks (Winston and Cupchik 1992). The challenge factor was reflected in beliefs that paintings should challenge our view of the world, express the deepest feelings of artists, and require intense effort from viewers. The pleasure factor was associated with beliefs that paintings should evoke peaceful feelings and happy memories, give immediate pleasure, and appeal to a large number of people. Trained students, who are familiar with stylistic codes, preferred High Art paintings and subscribed to the challenge approach to art, while inexperienced students preferred Popular (i.e. kitsch) artworks and favored a pleasure-based attitude. It has also been argued here that viewers and readers can shift between objective analytical and subjective holistic perspectives. In one study, subjects were instructed to shift between these two kinds of judgments sets (Cupchik et al. 1992). Results showed that a subjective, holistic set facilitated the discerning of similar style (for instance both impressionist or fauvist) between pairs of paintings. Participants in another experiment (Cupchik and Saltzman 1999) viewed serious or light rock music videos from either objective or subjective viewpoints in a within-subjects design and were then given a memory test to determine whether or not they correctly recognized the videos. As in the study using paintings, subjects were most accurate at recognizing videos that were first viewed from a subjective emotional set compared with an objective analytical set. Subjects were also more accurate at recognizing serious segments which were generally seen as more challenging and socially relevant compared with the light videos. In a study on responses to sexual, violent, and humorous magazine images (Cupchik and Izadpanah 1998), the sexual images evoked warmer feelings and the characters in them were judged to be proud and more in control of their lives. This perceived freedom was heightened under the subjective set. These studies established that subjects can switch between the objective/analytical and subjective/ holistic attitudes typically associated with High and Popular culture attitudes. Furthermore, emotional engagement actually made it easier to discern the holistic properties that characterize different styles. This demonstrates the complementary nature of high and popular cultural attitudes because a subjective approach facilitated the perception of stylistic similarity. As a process, the “effort after meaning” (Bartlett 1932) involves both an accommodation to information about art and literary works offered by creators and critics, and an assimilation of these works to personally meaningful schemas. In another series of experiments, participants generated or received interpretations of sculptures (Cupchik and Shereck 1998) and short story excerpts (Cupchik et al. 1998) between first and second viewings/readings (again in a within-subjects design). In the sculpture study, generating interpretations reduced the tendency for emotional impact and perceived originality to diminish over time. In the reading study, generating interpretations enhanced the experience of images in response to emotional texts and an awareness of the author’s words in descriptive texts.

High and popular culture from the viewpoints of psychology and cultural studies 427

Consistent with today’s skeptical views of New Criticism, experienced subjects in the reading study (i.e. undergraduate specialists in literature) were resistant to expert opinion regarding the meaning of the Joyce texts, though untrained readers were receptive. This same resistance was observed in another study in which trained viewers had their attention drawn to particular stylistic features of sculptures and paintings (Cupchik et al. 1994). Generally speaking, these two studies underscored the value of constructing interpretations of art and literary works. People can also shift from objective to subjective ways of viewing perfume and liquor advertisements (Cupchik et al. 1998). Continuing the line of research just described, subjects either generated story outlines to accompany romanticized or ironic ads, or critically analyzed copy-image relations in them. Writing story outlines in response to romanticized perfume ads elicited romantic themes and enhanced feelings of success, confidence, power, and masculinity/femininity. On the other hand, analyzing copy-image relations fostered seduction themes and a sensitivity to metaphor. Thus, subjects can shift between projection and reflection as ways of approaching advertisements as well as paintings and literature. Objectivity and subjectivity are therefore modes for approaching all manner of cultural materials which people can adopt and change.

2.

Social/cultural intersections

The perspective of Cultural Studies interrogates the complex interrelationships between and among: modernity, modernism, subjectivity/identity, and commodity culture. Modernity, on the social and cultural level, is the historically specific livedexperience of twentieth century culture, with the increasing complexity of the urban environment, as well as the proliferation of new technologies — photography, cinema, and printing advances — which permitted the ‘Mass-media’ of newspapers, magazines and advertising. At the level of the individual, ‘modernity’ means new modes of consumption, new styles of self-presentation, new attitudes toward crowds and urban spaces, new definitions of public and private spaces, and new forms of mass entertainment. ‘Modernism’ is the aesthetic response to this lived experience, and it, too, is historically specific in that modernist artists felt the need to depict the ‘everyday’ quality of life, and turned their attention, increasingly, to ‘low art’ and popular culture, and to other hallmarks of modernity that would have been dismissed as trivial and vulgar in earlier periods. Even more recently, a field of study called ‘Visual Culture’ has emerged, announcing that it “is concerned with visual events in which information, meaning or pleasure is sought by the consumer in an interface with visual technology” (Mirzoeff 1998: 3). This field of study calls for nothing less than a virtual interface of aesthetic reception and historical experience, arguing for ‘a polycentric aesthetics’

428 Gerald C. Cupchik and Garry Leonard

where “creation takes place not within the suffocating confines of Cartesian egos or even between discrete bounded cultures but rather between permeable, changing communities. Nor is it a question of a mindless ‘anthropological’ leveling which denies all criteria of aesthetic evaluation but rather of historically grounded analyses of multicultural relationality, where one history is read contrapuntally across another in a gesture of mutual ‘haunting’ and reciprocal relativization” (Shohat and Stam 1998: 46). Modernism, in other words, is the aesthetic strategy for representing modernity. But it does not so much represent it as it betrays the strain and exhilaration of living it. Modernity and modernism were mutually constitutive, with modernity forcing the formation of the strategies of modernism, and these same strategies then influencing the direction modernity took. It was Benjamin (1973) who cited Baudelaire as the first poet of modernity because Baudelaire insisted on the necessity for the truly modern poet to immerse himself in the city and the crowd, and to become “a kaleidoscope equipped with consciousness” (Benjamin 1973: 27). The emphasis was on the trivial, the commonplace, the “everyday”, not as vulgar, but as heroic. Perhaps the most famous example of this is Joyce retelling the story of Ulysses as a simple Dublin man who makes his living selling advertisements. For the poet Baudelaire and the sociologist Georg Simmel, the defining experience of modernity was that of living in a modern metropolis, and the most remarked upon feature of this life was the nearly unmanageable assault of noise, illogical juxtaposition, relentless unpredictability, and enigmatic epiphanies, all made publicly available to anyone, of all classes, in a manner which was as fascinating as it was frustrating. What is so different about this experience — why it gives rise to modernism — is that it demands delimitation. One of the most famous examples of modernist aesthetics is Joyce’s theory of the epiphany. In religious terms, the ‘epiphany’ is the moment where God is made manifest in some form. For Joyce, it is the trivial secular moments in someone’s life where the full reality of their situation converges on them — but only for a moment. This theory might be seen as Joyce’s attempt to aestheticize the kaleidoscopic urban environment. If it is true, as Oscar Wilde (1995) claimed, that there was no fog in London until poets and painters aestheticized it (“people see fogs, not because there are fogs, but because poets and painters have taught them”), it might also be argued that there were no intense moments of miraculous yet bewildering clarity as a hallmark of modern experience until Joyce’s idea of the epiphany became commonplace. Joyce’s notion of the epiphany, in other words, does not ‘discover’ an objective reality, but offers a framework for organizing the disconcertingly random stimuli of modern experience into an apparently integrated moment of ‘self ’-realization. Although the epiphany is often featured by Joyce as a crisis of identity, it might be more accurate to say that the crisis of identity is constant in modern experience, and that the epiphany, paradoxically, nominates

High and popular culture from the viewpoints of psychology and cultural studies 429

the crisis brought on by a sense of lost identity as ‘proof ’ of the very identity modern experience is bringing into question! Thus an aesthetic strategy delimits experience in order to shore up an eroding sense of ‘self ’. In a similar way, the notion of ‘everyday’ reality, which has already been cited as a feature of modernist art, is also more a strategy of delimitation than a description of objective reality. ‘Everydayness’, and its accompanying notion, ‘lifestyle’, is a way to select, edit, and organize the objectively real into a supposedly ‘self ’centered reality, a reality where Descartes’s belief in an identity coherent across space and time can still be maintained. But the shift in the experiential paradigm brought on by modernity is brilliantly captured by Beckett’s (1986) deliberate recasting of Descartes’s cogito from “I think therefore I am” to “I suffer, therefore I may be” (Beckett 1986: 117). Where Descartes posits an apriori integrated consciousness which proves its internal coherence by discerning various patterns in the apparently random world, Beckett admits only an unpleasant awareness of being acted upon as proof of existence, but not of an existence that can be objectively confirmed as having any intrinsic coherency. In other words, a retroactive aesthetic rearrangement of experience (the epiphany, the everyday, one’s lifestyle) is misrecognized as proof of an ‘identity’ that somehow transcends and pre-exists the sequence of experiences that, in fact, constitute it. In modernity, reality and experience seem equally random, and modernism, under the guise of ‘discovering’ a way to represent modernity, struggles to preserve a Cartesian sense that consciousness precedes and ‘makes sense’ of reality. ‘Stream of consciousness’, another famous ‘invention’ of modernism, takes us inside the very process just outlined. The narrative perspective is presented as ‘inside’ the mind of a character who is, for example, walking down the street. But this apparently apriori identity that organizes what there is to see is generated after the ‘fact’ of the experience, indeed, is generated by the experience, and only then presented as somehow prior to it. “It is not art that imitates life”, Oscar Wilde noted, “but life that imitates art” (Wilde 1995: 92). To paraphrase this, it is not the aesthetic strategies of modernism that imitate modern experience, but modern experience which organizes itself through the aesthetic strategies of modernism. Looking at this same phenomenon from a historical point of view, modernism records the struggle of converting from an earlier sense of transcendent certitude, where we declare our faith through ritual, tradition, and community, to a modern world of the ‘everyday’, where we reconstitute our faith individually, and secularly, on a day by day basis, often in relation to the mass media and mass produced objects of commodity culture. As the modernist poet Wallace Stevens put it, “it is the belief, and not the god, that counts” (in Geddes 1996: 910). In other words, the loss of a belief in transcendental certitude — God — does not also eliminate the need to believe in the presumed individual coherency of identity. The continued need for this belief, even in

430 Gerald C. Cupchik and Garry Leonard

the acknowledged absence of any possible proof, is what Stevens called “a necessary fiction.” Much writing about modernism, beginning with Eliot’s essays, through the New Criticism inspired by these essays, and even into contemporary theory, stresses how modernist artists were clever at finding new ways to represent a fractured, chaotic, high speed world. But this makes the development of modernism look like a simple exercise in new forms of mimesis. Later attempts to codify modernism have only further obscured the extent to which it was a frightened, sometimes hastily conceived reaction (often aggressively announced in the form of manifestoes) to modernity, though no less brilliant, for all that. Considering modernism from the perspective of modernity also demands that we look at the history of commodity culture as a history, and not just as something that happened along the way (which is the way commodity culture tends to represent itself). We need a history of commodity culture in order to historicize it as a specifically modern phenomenon, one leading to a particularly modern form of experience. We know about the history of the Industrial Revolution, which is also the history of machines and their effect on labor and society, but we know infinitely less about the history of the things these machines produced, and their effect on modern configurations of subjectivity, as demonstrated through historically and culturally specific constructions of identity, gender, desire and pleasure. Modernity needs to be understood as a convergence of technology and psychology. What was formerly a ‘thing’ becomes a commodity — something invested with a certain aura not at all connected to its mere use value. A good example is soap — something formerly sold in chunks wrapped in brown paper. But brand names such as “Sunlight” and “Ivory” begin to proliferate. Anxieties about the loss of privilege in an increasingly ‘post-imperial’ environment are met by promises of ‘whiteness’, ‘purity’, and ‘cleanliness.’ And these virtues are shown to promote inner peace, economic security, family, and romantic love. Conversely, advertisements warn about all our anxieties coming true if we don’t buy the product. “Aren’t you glad you use Dial? Don’t you wish everyone did?” was one famous slogan for a ‘deodorant’ soap. Such a dynamic is represented in works of modernism, most notably perhaps in the fiction of James Joyce. In Joyce’s Ulysses, a man goes about his day trying to ignore the fact his wife has made an appointment with a lover for that afternoon. Throughout the day, he is unable to get an ad jingle out of his head: “What is home without Plumtree’s Potted meat?/Incomplete/With it an abode of bliss.” Things, which, through advertising, become commodities, promise to organize our random thoughts into a myth of internal coherency. Commodity culture is modern, in part, because it relies on technological advances in visual presentation, such as photography, and improved methods of printing, such as lithography. These advances permitted, in turn, cheap productions of the plethora of images and advertising that provided the infrastructure of what we now call ‘mass media.’

High and popular culture from the viewpoints of psychology and cultural studies 431

In a parallel development, the ability of machines to mass-produce objects connected up with a powerful — that is to say, profitable — counterpart in the new-found ability of advertisers to mass produce images and generate discourses designed to make these innumerable objects appear valuable and necessary. Added to this, urban populations were swelling, making possible the sales of goods to large numbers of people at a rate never imagined before. The dependably high volume of sales made producing commodities even cheaper, freeing up capital for more and more involved advertising strategies, thus laying the groundwork for a shared discourse among the rich and poor alike understood now as ‘popular culture.’ In yet another separate but convergent phenomenon, electricity became readily available, thereby permitting the pedestrian areas of the city to be well lit and thus spawning what now goes under the generic term of ‘night life.’ Inside these same stores, the fact of electricity permitted shop owners to display goods by splendidly bathing them in light, in prearranged settings, safe from the distracting bustle of the street. Cheap printing also brought about modern packaging and modern advertising, so one selected from several brands of soap or meat or ointment, all of them roughly the same, but each made distinctive, or commonplace, by a successful or unsuccessful advertising campaign. Plate glass windows, in conjunction with electric light, permitted magical tableaux designed to instill in the viewer complex feelings of hope, envy, discontent and desire for fulfilment. Although these tableaux, as layered as a dream rebus, were readily available to the eye of everyone, nonetheless an inch of glass made sure active participation in the fantasy was reserved for those who could afford it. This sense of lack gives rise to the compensatory fantasies which underwrite the advertisements, which create — through massmedia — the mass produced commodities at the heart of the lived experience of modernity. In historicizing how the object world changed in modernity, the presumed universal of ‘human experience’ must also be historicized in relation to the historically specific reality in which it constitutes itself. Upon viewing the first exhibition of postimpressionist works in London, Virginia Woolf wrote: “In or about December, 1910, human character changed” (McNeillie 1986–1994: 321). We see here the bravado of aesthetic pronouncements in modernism, but we see as well the frank awareness of human ‘nature’ as a construct. Such paradigm shifts are not acknowledged readily, however, as we can see by noting the reaction of an anonymous journalist to this same exhibition, who described the postimpressionist painters as “uneducated persons, who never could draw or paint properly, but who have been deluded, either by natural insanity or by over-indulgence in absinthe, into the belief that they can” (Reid 1996: 115). Fifteen years earlier, Lord Henry announces, in Oscar Wilde’s (1995) The Picture of Dorian Gray, “being natural is a pose. And it’s the most irritating pose I know” (Wilde 1995: 27). The historical circumstances which spawned modernity,

432 Gerald C. Cupchik and Garry Leonard

commodity culture, and aesthetic reactions like postimpressionism and modernism, were not inevitable or evolutionary, or teleological, but mutually constitutive, convergent, and inextricable. First, there is the acute sense of lack and loss of transcendental certitude. This leads to the characterization of the century as “the age of anxiety”, and this is usually accounted for by noting the dwindling belief in God and the accompanying loss of a sense that everything happens for a purpose, however inscrutable that purpose might appear to us. One standard reference for illustrating this phenomenon is Arnold’s (1961) “Dover Beach” with its lamentation about the once full “sea of faith.” With supply so outrunning demand, the need to make the surplus seem essential became the engine of twentieth century capitalism (a dynamic still very much with us today in the ironic and polyvalent phrase “consumer confidence index”). The lack of assumed faith in a transcendental certitude put much more pressure on secular objects to serve as confirmation of identity and self-worth. If these objects could not be blessed, they could be advertised. The more diminished the subject, the more oracular the object, right up to our own time where even a deodorant is called ‘Sure’ — because we no longer are. Objects become more and more animated as subjects feel less and less the center of cognition and the origin of meaning. Descartes’s “I think therefore I am”, having already undergone one metamorphosis by Beckett, becomes, according to the contemporary artist Barbara Kruger (1990), “I shop therefore I am.” Whereas Descartes imagined an internal process of doubt leading to a self-affirming knowledge of the world (that is, the eventual discovery of the intrinsic and relative worth of objects and the laws governing them), in the experience of modernity we find ourselves at the center of a cluster of objects and nominate the effect of this constellation on our state of being as ‘subjectivity.’ And so we have ‘lifestyle’, the retroactive phenomenon we have already mentioned, where a life is ‘styled’ — comes into a recognizable awareness of itself — by the arrangements of objects around us — cars, clothes, cologne, deodorant. But, again, this awareness which is posited as preceding various purchases, is also the effect of purchasing. “Some things in life money can’t buy”, says a contemporary credit card commercial, showing apparently defining moments of identity, before ominously intoning, “for everything else, there’s Master Card.” This ad borrows the aesthetic strategy of Joyce’s notion of the epiphany: moments of buying are strung together as ‘evidence’ of an experiential reality before and apart from the buying, but this experiential reality is a retrospective formation after the fact of the buying. In this ad, you cannot buy happiness, but happiness is an effect of buying. Modernism and commodity culture have a powerful impact on personal identity and the experience of subjectivity. Identity can be understood further as a particular arrangement of subjectivity in a given period of time, and not as something innate or essential.

High and popular culture from the viewpoints of psychology and cultural studies 433

In twentieth century modernity, the subject feels anonymous, disempowered and disassociated, even as mass media and popular culture offer a constant source of fantasy, particularly in the area of feeling special, powerful, and connected. Subjectivity in the twentieth century, then, formulates an identity heavily dependent on compensatory fantasies as a source of reassurance. This particular identity configuration then becomes the engine of so-called ‘commodity culture’ which attempts to sell an unprecedented amount of ‘things’ to people who are not aware they need them. One need not look for a cause and effect relationship among these various factors or argue that advertising came along and then identity shifted. Rather, all these factors can be seen as mutually constitutive: the subject in modernity feels anonymous; the economy must sell things at an unprecedented rate; and technology permits the mass communication of why these things should be bought. The most effective mass communication is advertising which links the thing to be sold with the compensatory fantasy underwriting the modern subject’s fictive identity. ‘High Art’ and ‘Low Art’, traditionally separate spheres, also became mutually constitutive in modernism because the anxieties and exhilarations of modern living were embedded in popular culture, and modernist writers turned to it as they tried to become social archeologists, sifting through advertisements and other artifacts of popular culture, looking for the history of their own time. Popular Culture operates as a compensatory system offering solace, encouragement, or illusion to people who feel they lack something. The products of popular culture, such as advertisements, offer compensation for a historically specific sense of lack experienced by individuals in the modernist era. For this reason, advertising — and consumer discourse in general — constitutes a dynamic force every bit as influential on James Joyce as Homer or Shakespeare. Joyce does not cite individual advertisements merely in pursuit of realism or verisimilitude. On the contrary, he presents the overall dynamic of advertising in order to demonstrate the extent to which social relations, nationalist aspirations, power structures, class distinctions, gender constructions, and subjectivity itself all intersect with, and depend upon, the simulated universe of advertisements, and mass-media in general. In a sense, by examining the cultural artefacts produced by ‘low art’, Joyce shows us we do not have to blindly assume everything re-presented in a modernist work of ‘high art’ is a complete presentation of the lived-experience of modernity as it played itself out in the lives of the historical subjects of this age. All “official” meta-narratives, such as history or canonized examples of ‘high art’, are effectively challenged by popular culture, and by its tendency to record “the history of now” (Leonard 1998) — precisely what ‘official’ history ignores or devalues. All ‘official’ representations, in collusion with hegemonic requirements, inevitably distort as ‘natural’ and ‘inevitable’ precisely what most needs to be understood as constructed and culturally specific. ‘High art’ is not simply the incarnation of truth, or a ‘natural’ representation of the desire for universal good, but rather a complex

434 Gerald C. Cupchik and Garry Leonard

response to a matrix of psychological, sociological, economic, and aesthetic pressures (to name the most apparent). Whereas the Victorian poet and critic Matthew Arnold (1961) attempted to universalize the concept of Culture, with a capital C, as “the best that has been thought and known in the world”, popular culture offers, instead, an unofficial discourse that problematizes how one arrives at designations such as “the best”, or how one organizes and reconstitutes perceptions into what is then understood as the “known.” For modernist writers and artists, then, ‘low art’, and popular culture, become an important source of the cultural reality they are trying to reconfigure and thereby subvert ideological constructions about what is ‘the best’ and ‘the known.’ The dominant example of this mutually constitutive dynamic would be the development of advertising which began as an attempt to sell a new plethora of machine-produced commodities to the public, and did so by finding ways to link the product with concerns caused by modernity. So soap began to promise purity (in response to concerns about post-Darwinian degeneration), while patented medicines promised “new energy” (in response to fatigue and depression), and corsets promised a more highly visible, therefore competitive, “femininity” (in response to the decline of the class system). Advertisers hoped to reach a ‘mass’ audience with mass-produced graphics. This required them to design posters and ad copy that were pared down (easy to reproduce), startling (able to attract the attention of someone passing by), and highly visual (memorable in some way). Modernists incorporated the techniques of such ads in their work; ads, in turn, borrowed from the result in order to set off in yet other new directions. Joyce and Picasso seemed to relish popular culture, but even if other modernists regretted its dominance (as Woolf and Eliot sometimes seemed to), it was the very incarnation of modern experience, and no artist seriously concerned with issues of modern subjectivity could ignore it. Instead of an aesthetics favorable to contemplation, what emerges in modernism is an aesthetics of attraction and distraction. The previous concept of the subject before the art work as someone who first perceives, then judges, with inviolable interiority, from a position of subjectivity, alters to someone seen as permeable to signs and symbols of the modern environment: “capitalism and the body are connected so far as stimulating the nerves is a way of stimulating exchange Y. The reader’s (or observer’s) body becomes a machine hooked into the circuit of production and consumption, rather than a disinterested entity floating above economic exigencies in search of aesthetic or moral truth” (Cvetkovitch 1992: 21). If aesthetic theory was once envisioned as a system of thought explaining the ‘natural’ category of ‘beauty’, much modernist work conflates ‘high’ and ‘low’ culture in order to analyze the infrastructure for the social construction of ‘the beautiful.’ They reject the idea of the work of art emerging pure from the genius of the artist, and understand it instead as the cornerstone of bourgeois hegemonic

High and popular culture from the viewpoints of psychology and cultural studies 435

discourses which make the fictional category, ‘artistic’, an excuse to organize and discipline all the chaotic bodily responses. This represents a shift from thinking about aesthetic theory as a philosophical end point to exposing it as a dynamic of self-discipline, one that resolves ambivalence and desire into an agreed-upon theory about what constitutes ‘beauty’ and ‘truth.’ What aesthetic theory elides, however, is that both logical analysis and physical arousal, both objective harmony and disconcerting fantasy, both the artistic and the pornographic, both the contemplative and the seductive, both ‘high’ art and ‘low’, are necessary components in the apprehension of what is ‘true.’ Put more succinctly, creativity, like interpretation involves pleasure as well as style. Thus modernists denaturalize the notion of what is ‘beautiful’ and ‘true’, regarding it, in contrast to Matthew Arnold, not as an obvious, self-evident undertaking, but reconfiguring it as more of a process. The body is seen as a contested site of meaning in which the sensory apparatus is colonized in the service of a historically, culturally, and economically specific subjectivity, which is then reified and universalized as ‘human nature.’ Of course, the same body that is mapped out by the aesthetic threatens to rescramble itself with the pornographic. Using the anthropological work of Mary Douglas, Fiske identifies aesthetic theory as performing a purification ritual: “dirt is matter out of place, and the terror it invokes in the respectable bourgeoisie derives from its power to demonstrate the fragility of the conceptual categories by which the semiotic and the social are exercised over unruliness and the force of deception” (Fiske 1989: 98). Previously banished from aesthetic theory as a legitimate response, arousal and pleasure return in modernism to call into question Western culture’s basic assumption of a substantive being who is presumed to be the origin and sovereign ruler of thought, feeling, and desire. The conflation of objective contemplation and pleasure is paralleled by the conflation of high and low art. Postmodernism continues this project of conflation, but does so with no nostalgia for a former period where totality and transcendental certitude were assumed. Put another way, postmodernism, aided by contemporary theories of fracture, decentering, incertitude and lack, revisits the issues of modernism, but does so no longer traumatized by the experience of modernity. What was shocking to the modernists about modernity was its suddenness (the “shock” of which Benjamin spoke so often). But now, if we could imagine modernism as a patient on the couch in Freud’s office, Freud’s message would be “but you have already survived the trauma of modernity; it is time for you to know that, and to stop acting out the shock and move on to more conscious strategies of parody, irony, artifice.” It is not a coincidence that Lacan is the postmodern Freud (Leonard 1993), since he argued what the modernists struggled not to see: that the ego, which Freud tried to shore up, is a myth of fictive unity, a compromise formation, a paranoid structure, “the mental illness of man, par excellence” (Miller 1988: 16). As Eliot (1936) said in a different context, “after such knowledge, what forgiveness?” (Eliot 1936: 42). It is

436 Gerald C. Cupchik and Garry Leonard

the complete abdication of any attempt to reconcile anything that gives postmodernism its curious depthless quality. If modernism can be seen as the extraordinary effect of conflating high and low art, postmodernism can be seen as the extraordinary effect of taking this same conflation for granted.

3.

Interdiscursive space

In this piece, we examined different ways of thinking about ‘high’ and ‘popular’ culture. The historical viewpoint of cultural studies is undeniably political as it weaves a path through history (modernity), aesthetics (modernism), everyday life (commodities), and psychology (identity). While experiments are limited when it comes to retrodiction, the ideas embodied in critical analysis of culture can inform the study of how contemporary subjects respond to a range of artefacts from artworks, to literature, music videos, and advertisements. Cultural studies examines that social world, identifying the signs to which people are exposed as members of different communities. These signs may be found at a very rarefied level in the original achievements of culture (for instance great paintings or literature) but also in all media content, advertisements, and so on. In some cases these signs are created by individuals in our society and brought to us as members of the public for the purpose of commenting on culture, society, politics, and so on. Perhaps this is the goal of High Art in its classical form. At other levels, these signs are reflections of our society, images chosen by individuals so as to recreate the familiar and elicit empathic emotions; popular culture touches upon this kind of nostalgia. When such images are selected so as to create the illusion that they reflect society, we have the beginning of commodity culture with its great potential for shaping who we think we should be. Scholars of cultural studies attempt to understand how what we notice makes us feel from within the context of a given history and culture. If the ‘empirical’ is the basis of one form of study, the ‘ideological’ is the form of the other. Althusser defined ideology as “our imagined relationship to our actual existence.” The empiricist is interested in that “actual existence”, while the scholar of cultural studies attempts to outline our imagined relationship to it. Actual existence may be universal and ahistorical, but how we feel about it is not. This kind of multidisciplinary analysis, which cuts across culture, sociology, communication, economics, and ultimately individual psychology, stands or falls on its own merit. Its goal is to sensitize people to that which is taken for granted and as such fulfils the mission of phenomenology and symbolic interactionism; awakening us to the fact that images of the world, both social and physical, are socially intended or constructed and change over time.

High and popular culture from the viewpoints of psychology and cultural studies 437

There are at least three distinct streams in the psychological approach to aesthetics; psychobiological, cognitive, and psychodynamic. The psychobiological approach began with Fechner (1978) in the nineteenth century and was resurrected by Berlyne (1971, 1974) and his new experimental aesthetics. Berlyne was a motivation theorist (see Cupchik 1988) who saw aesthetic activity as intrinsically motivated (that is, art for art’s sake) and examined the ways in which collative or structural properties of art works, such as complexity and novelty, affected pleasure and arousal. Applying information theoretic concepts, he and Moles (1968) discussed the problems of measuring complexity at the stylistic (that is syntactic) and thematic (that is, semantic) levels of aesthetic artefacts. Berlyne offered empirical support for Fechner’s original idea that people would find greatest pleasure in moderate levels of complexity, and added that interest or arousal would be peaked by higher levels of complexity. Kreitler and Kreitler (1972) offered an alternative approach, one that emphasized the multileveled nature of aesthetic works from gestalt and psychodynamic perspectives. They defined multileveledness as “the capacity of a work of art to be grasped, elaborated, and experienced in several systems of connected potential meanings, each of which allows a meaningful, clear, comprehensive, and sometimes autonomous organization of the major constituents of the work” (Kreitler and Kreitler 1972: 295). As a result, the content and structure in works of art and literature are carriers of “a multiplicity of meanings and significations whose wholeness persists in the face of a variety of multileveled integrations” (294). This approach is consistent with other holistic and constructivist perspectives emphasizing the indeterminate (Iser 1978), polyvalent (Schmidt 1982), and open (Eco 1989) nature of literary texts. While Berlyne’s information theoretic approach focused on the modulation of pleasure and arousal by structural stimulus properties, the Kreitlers addressed the serious social and personal meanings that might be embedded in cultural artefacts. A reconciliation of these perspectives has been proposed (Cupchik 1995; Cupchik and Winston 1992) in which stimulus (aesthetic artefact) and response (pleasure, arousal, meaning) are treated as complementary or mutually constitutive rather than linear and mechanically related (S — R). If entertainment is the goal, then stimuli are selected whose properties could modulate pleasure and arousal. Accordingly, films with a lot of suspense can alleviate boredom, or ones with warm associations produce sentimental pleasure (see Winston 1992). The criterion for success in the entertainment model is a match between the properties of the stimulus and the bodily needs of the person. If a serious aesthetic experience is the goal, the multileveled works challenge the receiver who must bring to bear an appreciation of stylistic codes and relevant personal experiences in order to construe the meaning of the work. The criterion for success here is the degree to which cohesion is achieved at fitting the various elements of the work into a meaningful whole.

438 Gerald C. Cupchik and Garry Leonard

The perspectives of cultural studies and psychological aesthetics are also mutually constitutive to the extent that their main concepts and basic assumptions are placed in a common discursive space (Cupchik and Leonard 1997) — in this case, concerning relations between high and popular culture. For the psychologist, the problem of entertainment is really one of mood management. But the cultural studies approach situates the person in a broader social context and inquires about the extent to which individual moods are in fact being ‘managed’ by economic forces that manipulate desires and identities. The psychologist views serious culture as the evolution of themes and styles whose originality can be critically evaluated and which engages the person in a deeper exploration of lived-experience. The cultural studies scholar sees the ‘serious’ artist or author as a commentator both on aesthetic traditions and on the state of the lived-world. By commenting on the social order, artists and authors can raise our consciousness not just about the world but about the very act of generating meaning. There are at least two ways of thinking about the complementary relations of psychology and cultural studies. One relates to the interaction between general principles of psychology and the interpretation of particular cultural events in history (for instance modernity and modernism). They are mutually constitutive to the extent that principles of psychology can be meaningfully applied to individual instances of ‘high’ and ‘low’ cultural activity. Thus, principles of conditioning can explain how repeated exposure to arousing and pleasing artefacts lessens their impact. Psychodynamic principles can be applied to instances in which a feeling of insecurity leads to compensatory behavior, as when someone who doubts their masculinity or femininity buys a product that will ensure proper role enactment. Principles of reflectivity can describe how struggling with an image or text leads a person to experience the arbitrariness of constructed meaning. A second mutually constitutive interaction pertains to the shaping of laboratory experiments. Rather than using individual pieces of fiction to extrapolate universal principles (a logical contradiction that violates basic sampling principles), samples of works can be chosen that juxtapose important structural features. These works can then be examined from viewpoints that match fundamental modes for encountering aesthetic events (for instance, subjective or objective perspectives). By choosing response dimensions that represent outcomes of interest both to psychologists and researchers in cultural studies, the ingredients are prepared which can yield findings that are meaningful to both groups. In this way, experimental psychology and cultural studies become complementary or mutually constitutive; the meaning of the materials (that is, signs) used in a psychology experiment becomes grounded in social/economic history, while the findings (for instance dynamics of compensation) reinforce the historical understanding of advertisements and their effects. The integration of quantitative experimental and qualitative (for instance, think aloud protocols) approaches to research, combined with the

High and popular culture from the viewpoints of psychology and cultural studies 439

historical reflections of cultural studies, provides a firm foundation for a social constructionist approach to problems in aesthetic reception.

References Arnheim, R. 1985. “The other Gustav Theodor Fechner”. In A Century of Psychology as Science, S. Koch and D. Leary (eds), 856–865. New York: McGraw-Hill. Arnold, M. 1961. Poetry and Criticism of Matthew Arnold. Boston: Houghton Mifflin. Bartlett, F. C. 1932. Remembering: A Study in Experimental and Social Psychology. Cambridge: Cambridge University Press. Beckett, S. 1986. The Complete Dramatic Works. London: Faber and Faber. Benjamin, W. 1967. “The work of art in the age of mechanical reproduction”. In Illuminations, H. Arendt (ed.), 217–251. New York: Shocken Books. Benjamin, W. 1973. Charles Baudelaire: A Lyric Poet in the Era of High Capitalism. London: NLB. Berger, A. A. 1980. Television as an Instrument of Terror: Essays on Media, Popular Culture, and Everyday Life. New Brunswick, NJ: Transactions Books. Berlyne, D. E. 1971. Aesthetics and Psychobiology. New York: Appleton-Century-Crofts. Berlyne, D. E. 1974. The New Experimental Aesthetics: Steps Toward an Objective Psychology of Aesthetic Appreciation. Washington: Hemisphere. Chaney, D. 1994. The Cultural Turn: Scene-Setting Essays on Contemporary Social History. London: Routledge. Cupchik, G. C. 1988. “The legacy of Daniel E. Berlyne”. Empirical Studies of the Arts 6: 171–186. Cupchik, G. C. 1995. “Emotion in aesthetics: Reactive and reflective models”. Poetics 23:177–188. Cupchik, G. C. 1997. “Identification as a basic problem for aesthetic reception”. In The Systemic and Empirical Approach to Literature and Culture as Theory and Application, S. Tötösy de Zepetnek and I. Sywenky (eds), 11–22. University of Alberta, Canada: RICL-CCS and Siegen University, Germany: LUMIS. Cupchik, G. C. and Gebotys, R. J. 1990. “Interest and pleasure as dimensions of aesthetic response”. Empirical Studies of the Arts 8: 1–14. Cupchik, G. C. and Izadpanah, S. 1998. “Reactions to funny, violent, and sexual images in magazines”. SPIEL 17: 15–32. Cupchik, G. C. and Kemp, S. 2001. “The aesthetics of media fare”. In Media Entertainment: The Psychology of its Appeal, D. Zillmann and P. Vorderer (eds), 249–264. Mahwah, NJ: Lawrence Erlbaum. Cupchik, G. C. and Leonard, G. 1997. “The two “I’s” of the aesthetic process”. In Emotion, Creativity, and Art, L. Dorfman, C. Martindale, D. Leontiev, G.Cupchik, V. Petrov, and P. Machotka (eds), 81–100. Perm, Russia: Perm Institute of Culture. Cupchik, G. C., Leonard, G., Axelrad, E. and Kalin, J. 1998. “The landscape of emotion in literary reception: Generating and receiving interpretations of James Joyce”. Cognition and Emotion 12: 825–847. Cupchik, G. C., Leonard, G. and Irvine-Kopetski, D. 1998. “Advertisements: Multileveled in word and image and in the mind of the beholder”. Empirical Studies of the Arts 16: 115–135. Cupchik, G. C. and Saltzman, M. 1999. “Recognition memory for serious and light rock music videos”. Empirical Studies of the Arts 171: 59–72. Cupchik, G. C. and Shereck, L. 1998. “Generating and receiving contextualized interpretations of sculptures”. Empirical Studies of the Arts 16: 179–191.

440 Gerald C. Cupchik and Garry Leonard

Cupchik, G. C., Shereck, L. and Spiegel, S. 1994. “The effects of textual information on artistic communication”. Visual Arts Research 20: 62–78. Cupchik, G. C. and Winston, A. S. 1992. “Reflection and reaction: A dual-process analysis of emotional responses to art”. In Art and Emotions Vol 2, L.Ya. Dorfman, D. A. Leontiev, V. M. Petrov and V. A.Sozinov (eds), 65–72. Perm, CIS: The Perm State Institute of Culture. Cupchik, G. C., Winston, A. S. and Herz, R. 1992. “Judgments of similarity and difference between paintings”. Visual Arts Research 18: 37–50. Cvetkovitch, A. 1992. Mixed Feelings: Feminism, Mass Culture, and Victorian Sensationalism, New Brunswick, NJ: Rutgers University Press. Eco, U. 1989. The Open Work. Cambridge, MA: Harvard University Press. (Original work published, 1962.) Eliot, T. S. 1936. Collected Poems 1909–1962. London: Faber and Faber. Fechner, G. 1978. Die Vorschüle der Aesthetik. Hildesheim: Georg Holms. (Original work published, 1876) Fishwick, M. 1974. Parameters of Popular Culture. Bowling Green, KY: Bowling Green University Press. Fiske, J. 1989. Understanding Popular Culture. New York: Routledge. Fornas, J. 1995. Cultural Theory and Late Modernity. London: Sage. Geddes, G. (ed.). 1996. Wallace Stevens: Twentieth Century Poetry and Poetics. Oxford: Oxford University Press. Goody, J. 1986. The Logic of Writing and the Organization of Society. Cambridge: Cambridge University Press. Greenberg, C. 1957. “Avant-garde and kitsch”. In Mass Culture: The Popular Arts in America, B. Rosenberg and D. W. White (eds), 98–110. New York: The Free Press of Glencoe. (Original work published, 1946.) Iser, W. 1978. The Act of Reading. Baltimore: Johns Hopkins University. (Original work published, 1976) Kruger, B. 1990. Love for Sale: The Words and Pictures of Barbara Kruger. New York: Abrams. Kreitler, H. and Kreitler, S. 1972. The Psychology of the Arts. Durham, NC: Duke University Press. Leonard, G. 1993. Reading Dubliners Again: A Lacanian Perspective. Syracuse, NY: Syracuse University Press. Leonard, G. 1998. Advertising and Commodity Culture in Joyce. Gainesville, Florida: University Press of Florida. Lowenthal, L. 1957. “Historical perspectives of popular culture”. In Mass Culture: The Popular Arts in America, B. Rosenberg and D. W.White (eds), 46–58. New York: The Free Press of Glencoe. (Original work published, 1950.) McNeillie, A. (ed.). 1986–1994. The Essays of Virginia Woolf. Volume 3. London: Hogarth Press. McGuigan, J. 1992. Cultural Populism. New York: Routledge. Miller, J. A. (ed.). 1988. The Seminars of Jacques Lacan: Freud’s Papers on Technique, 1953–54. New York: Norton. Mirzoeff, N. 1998. “What is visual culture?” In The Visual Culture Reader, N. Mirzoeff (ed.), 3–13. New York: Routledge. Moles, A. 1968. Information Theory and Aesthetic Perception. Urbana, IL: University of Illinois Press. (Original work published, 1958.) (J. E. Cohen, Trans.) Reid, P. 1996. Art and Affection: A Life of Virginia Woolf. Oxford: Oxford University Press. Schmidt, S. J. 1982. Foundations for the Empirical Study of Literature. Hamburg: Helmut Buske Verlag. (R. de Beaugrande, trans.) Shohat, E. and Stam, R. 1998. “Narrativizing visual culture: Towards a polycentric aesthetics”. In The Visual Culture Reader, N. Mirzoeff (ed.), 27–49. New York: Routledge.



High and popular culture from the viewpoints of psychology and cultural studies 441

Van Peer, W. 1997. “‘High’/‘low’ cultural products and their social functions”. Empirical Studies of the Arts 15: 29–39. Wilde, O. 1995. The Complete Works of Oscar Wilde. Bristol, England: Parragon Books. Winston, A. S. 1992. “Sweetness and light: Psychological aesthetics and sentimental art”. In Emerging Visions of the Aesthetic Process, G. Cupchik and J. Laszlo (eds), 188–136. Cambridge: Cambridge University Press. Winston, A. S. and Cupchik, G. C. 1992. “The evaluation of high art and popular art by naive and experienced viewers”. Visual Arts Research 18: 1–14.



Operative fictions The fabric of societies Siegfried J. Schmidt University of Münster

The new worlds of virtual realities, cyberspaces, hypermedia and so on have provoked an extended debate about the concept and status of reality and its possible counterpart(s) such as irreality, simulation, hyper reality, possibility, virtuality, fiction and so on. As a consequence, the set of categories that help us in our attempts to address the issue of ontology, the determination of what is real and what is not, the central issue of power and knowledge, has started to proliferate. Serious questions arise: What kind(s) of a world(s) do we actually live in? Are the days of ‘reality’ definitely gone? I think it is exactly for this reason that philosophies like (Radical) Constructivism, Luhmann’s Systems Philosophy or various postmodernist theories have become prominent in recent years. Societies seem to need theories which have discharged identity-theoretical concepts of ‘reality’, ‘meaning’, ‘identity’, ‘truth’ or ‘society’. Difference theories are needed: these do not start with seemingly givens but try to reveal how we as active observers bring into existence what we afterwards observe as (items of) ‘reality’.

1.

A hypothesis

Observing media societies (which I deem to be media culture societies) brings to the fore another interesting feature: media culture societies are characterised by a proliferation of nonliterary fictions, from the so-called public sphere to fictions in infotainment or images of persons or institutions created by PR-activities. In the complex media systems of modern societies the demarcation line between facts and fictions which seemed to be robust for centuries has become blurred or simply irrelevant since many people are evidently no longer interested in its working.

444 Siegfried J. Schmidt

These observations lead me to a far reaching hypothesis which I would like to develop in this chapter: Media societies are fundamentally based on a variety of operative fictions. Media societies work, maintain and reproduce themselves only because they are non-trivial fiction machines which in their evolution have progressively replaced their interactions with environments by interactions with themselves. Since fictions allow for autonomisation via operational closure, media societies, as fiction machines, tend to become operationally closed, that is autonomous via the functionalisation of fictions. The discussion of this hypothesis concentrates on five types of fiction: literary fiction (or fictionality), social fictions (collective knowledge), cultural fictions (self organising programs), media fictions (media worlds), and technological fictions (cyber phenomena). Let me start with literary fictions.

2.

Literary fictions

In the second half of the nineteenth century, Germany saw the emergence of a selforganising social system of literature that essentially followed in the wake of the gradual transformation of what was once a highly stratified society into a network of functionally differentiated social systems. Of course fiction and literature had existed long before. What was radically new, however, was the historical process of functional differentiation whereby literary activities (production, mediation, reception and post-processing of literary givens) were drawn together in a web of actions and institutions gradually forming a self-organising social system of literature. In the course of this process, all activities related to literature underwent fundamental changes: they were professionalised, institutionalised, and differentiated, and the relations between them became self-referential. Like all other functionally differentiated social systems, the literary system, too, exclusively operated according to its own systems logic or rationality: all interactions with other social systems were subdued to the selection, evaluation and post-processing conditions of the literary system. This difficult and protracted process was accompanied by the gradual rise of the category ‘fictionality’ as a characteristic of communication processes in the social system of literature. In order to understand the proper meaning of fictionality, we have to clarify the difference between ‘fictivity’ and ‘fictionality’. Unlike mistakes, lies or utopian statements, a fictive statement may be defined as follows: If a statement p is actually neither true nor false in the prevailing socially accepted world model W but there can be imagined a world W* in which p is W*-true, p is a fictive statement. A formalisation of my argument in terms of modal logics can be found in Schmidt (1976). It is obvious that literary texts — for instance novels — do not exclusively consist

Operative fictions 445

of fictive statements, and that fictive statements occur in all kinds of texts, not only in literary ones. Consequently, fictivity — as a semantic notion — cannot be taken a sufficient definiens for literariness. Instead, we need a pragmatic criterion, which might be defined as follows: If you want to adequately produce or to receive a literary text your predominant task is not to evaluate its referable parts according to truth-functional categories in the referential framework of the prevailing social world model W, but rather according to such values, norms, and meaning rules considered to be relevant according to poetic/aesthetic norms prevailing at that time in respective social groups. The definition reveals that this pragmatic criterion is actually a convention, which has been and still must be learned by every member of a society who wants to adequately take part in literary communication. This convention unmoors literary communication from the fact-convention relevant in all other social systems which claims that statements have to be either true or false. Instead, the truth-referentiality of statements in literary works of art is suspended and replaced by expectations regarding the aesthetic value of language use in literary texts. In other words: ‘fictionality’ is the name for a special system of pragmatic rules which culturally prescribe how participants in literary communication have to treat the possible relations between fictive worlds assigned to literary texts and prevailing social world models. Fictionality is thus not a property of literary texts but of the literary discourse as a whole. (For details and definitions see Schmidt 1980.) Traditional alternatives or dichotomies like true/false, real/unreal, and possible/impossible thus met a competitor which was irritating insofar as ‘fictionality’ was not simply the opposite of a given category, but the quality of a whole discourse which admitted all possible forms of utterances: descriptions, declarations, comments, statements, questions, lies and the like. The traditional categories ‘reality’ and ‘truth’ were confronted with an option which offered a complete alternative to socially accepted world models: alterity. It shows that only with the greatest difficulty did bourgeois thinking accustom itself to the self-referential option proffered by the availability of competing models of reality. By the end of the eighteenth century, a process came to a close that I want to call the fictionalisation of literary discourse within the social system of literature. Fictional discourse established a fascinating alternative to all other social discourses. The autocracy of the true/false-dichotomy was finally broken, and a variety of ways of world making were launched. From now on, people had to realise very carefully which discourse they were actually engaged in. (See for instance Goethe’s Postscript to Die Leiden des jungen Werthers, where he underlines the difference between the literary and non literary discourses in order to avoid further suicides among the readership of this novel.) Fictionality, as a property of an important social discourse, could only be accepted by eighteenth century society because the economic system at that time took over the priority and the power to define what was real by transforming all

446 Siegfried J. Schmidt

goods, givens, and competencies into monetary values. Money became the neutral gauge which rendered everything comparable to everything excluding its semantic or moral values. These had to be treated in special semantics in the respective social systems. (See Schmidt 1989 and 1992.) Only now did a functionally differentiated society come into the position to permit a differentiation of even the ontological question. The fictionalisation of literary communication revealed for the first time that self-organisation is a basic general feature of the production of meaning. The literary system gradually became self-organising on all levels, because all activities in the system had to orient themselves towards the system’s rationality, and the system itself selected and treated all its interactions with other systems according to this very same logic. Consequently, literature oriented itself towards literature and treated the systems specific interaction with environment(s) on the same level of self-referentiality. Thus the literary system — prototypically for other media systems to come — gradually became autonomous. By proliferating self-referentiality and self-organisation it became rather independent from its environments and was able to replace to a large extent allo-reference by self-reference. The social system of literature concentrated its efforts on observing and describing society and itself under a literary systems specific perspective. But the more intensive this was done, the more literature as a specific mode of world making (in Goodman’s sense) prompted the question whether or not other worlds, like for instance the respective world models of social systems or societies as a whole, were of a different quality but not different in ‘essence’. That is to say, the evident constructivity of literary worlds promoted the insight into the constructivity of all ways of world making. The realities of our experiential world become unreal, ‘reality’ finally became Europe’s demonic category, as Gottfried Benn once put it. It is very interesting to observe that the emergence of the social system of literature gave rise to another option which organised itself as the journalistic system. Unlike the literary system which deliberately suspended referential decisions, thus gaining sufficient ground for aesthetic priorities in the literary discourse, the contemporaneously emerging journalism system claimed to produce authentic observations and objective descriptions of reality and nothing else. From the late eighteenth century on, modern European societies thus disposed of two rivalling macro-forms of self-observation and self-description with completely different reference mechanisms, although, from the very beginning, journalism had to meet suspicions whether or not its reports were ‘really true’ — a situation which has not changed until today.

Operative fictions 447

3.

Social fictions

Whereas the fictionality of the literary discourse is a deliberate strategic option which has to be learned and which functions as a pact between authors and readers, social fictions are mostly unconscious, though learned, operative fictions. In order to explain this concept we have to consider some cognition theoretical arguments. Gestalt-theorists since the twenties have demonstrated that human perception operates on a principle of a distinction between figure and (back)ground. In other words, we never perceive identities, only differences; identities may be described as the unity of a difference. To perceive means to perceive something as something in difference to something else. Whatever we do, we operate as observing and describing systems: we manage differences we ourselves introduced and functionalised in the respective perception. It follows that in perception we cannot get behind or beyond perception in order to compare the perceived with the not yet perceived in order to check the truth or objectivity of our perceptions. We can only act and speak in our world of perceptions which cannot be separated from our activities. Technically speaking, observing systems and their environments form an operative whole. In this process the act of perceiving and the perceived can only be analytically separated from one another. Perception, thinking, and speaking operate by and through differences. All these operations can be described as activities that rely on a basic mechanism of cognition as well as of communication, the netting or intertwining of reflexive structures. With regard to cognitive operations this means that sensoric and motoric processes are combined with the central nervous system. The operations of this complex network produce what we afterwards call ‘perception’. These perceptions are evaluated in and by subsequent perceptions, resulting in invariants that help the cognitive system to organise its experiences in terms of object constructions. Objects, in this argumentation, are not regarded as identical entities, but as cognitive invariants or Eigenwerte in the mathematical sense of this term. Neurobiologists tell us that the human brain can only interact with its own states since it is operationally closed. It follows from this assumption that our brain works self-referentially and organises itself. The cognitive system thus defines the boundary to its environment by applying the distinction allo-reference vs. selfreference. Accordingly, the brain does not mirror or represent its environment but constructs a cognitive model for its environment according to its own working conditions. And the name of these working conditions is “reflexivity of parallel distributed networks”. This basic pattern quite evidently is not restricted to the individual organism, but is also valid in social systems. Social groups or whole societies organise their experiences via communications which are recursively connected to each other, thus building stable orders that underlie all social activities as collective knowledge or a frame of reference.

448 Siegfried J. Schmidt

Actions and communications are equally determined by schemata and categorisations. We need cognitive schemata in order to reduce complexity and to enlarge the probability of successful interactions with other people. We need actionschemata in order to be able to act and to understand actions. Acting situations are schematised to make them relatively automatically recognisable and understandable. Schemata reduce complexity, co-orient actions and interactions, and ensure the attribution of social sense to singular events. Some schemata and categorisations are inborn, but most are of social origin. They are acquired by individuals during their long socialisation. Schemata are social fictions in a double sense: –

–

As man-made instruments they are used to organise experiences in respective environments. Accordingly, the relevant question with regard to schemata is not whether or not they are true, but whether or not they do their job. Everybody in a society acts as if everybody else applies the same schemata although no one can verify this assumption by looking into other people’s heads.

Now the question arises: How did and does this ‘as-if ’, this operative fiction (in Vaihinger’s sense) emerge? In our daily life, we act as observers who themselves can be and actually are observed: we are observed observers, and we know about this reflexivity of social relations and assume that others do so, too. It is this reflexivity which, from an evolutionary point of view, is responsible for the rise of commonality or sociality. Observed observers produce knowledge regarding stabilities or invariants in the behaviour of other observers, knowledge which they can mutually impute to one another. Though observers have no chance to control this knowledge in the heads of other observers, they do rely on the presence and the impact of this knowledge deemed as common or collective knowledge: A knows that B knows that A knows. Technically speaking, interacting observers rather automatically construe the reflexive structure of expected expectation which is closely connected with another reflexive structure, imputed imputations: A imputes to B that he/she does not simply behave but acts with regard to purposes and aims, is guided by motivations and emotions and oriented by norms and values — be this implicitly or explicitly. By netting such reflexive structures, highly complex structures emerge on the basis of operative fictions. Since — as already mentioned above — nobody can ever look into the head of other people in order to examine whether his/her expected expectations or imputed imputations are true or false, one can only rely on secondary processes like interactions or communications which we use as indicators of the reliability of our expectations and imputations. My own expectations and imputations work, because I construe a fictitious world in which they are true, and I counterfactually generalise this world so as to comprise other people’s worlds, too. The social fiction of commonly shared collective knowledge serves as the basis for

Operative fictions 449

all interactions and communications. We assume that others know what we know, live in worlds similar to ours, and use language more or less the way we do. Societies arise and last if and as long as people orient their activities in the broadest sense towards this allegedly shared common knowledge, which schematises all their experiences which claim to be ‘real’ and ‘relevant’. The collective knowledge of social groups or societies, which in its totality represents their ‘world model’, serves as a socially accepted instrument of problem solving. It results from successful problem solutions, is used and legitimated by success in subsequent problem solutions, and thus stabilises the organisation of experience of all members of a society via a successful self-reference. Although this knowledge has to be repeatedly constructed in the cognitive domains of individuals in every new situation, it is treated like an obligatory social fact. The fact that we as individuals constantly render social fictions into facts explains why societies operate as fiction machines.

4. Cultural fictions World models systematise knowledge needed for problem solutions in all those dimensions which are relevant for the success and the survival of societies. As far as I can see the most prominent dimensions of world models are the domination of the environment, the interaction with other people, the domination and expression of emotions, and the handling of normative problems. I guess these dimensions are in the fore because our brains seem to comprise similar dimensions in their structural design. As knowledge systems, world models are composed of distinctions such as system/environment, we/others, good/bad, love/hatred, powerful/powerless, young/ old, healthy/ill, and so on. Due to the modus operandi of our brain which constantly intertwines thinking, feeling and evaluating, we are never occupied with thinking alone or as such; instead, all processes of cognition are embedded in (more or less conscious) processes of lust/unlust-equilibration and normative evaluation. Consequently, the distinctions of world models as well as their relations and combinations are emotionally charged and normatively evaluated. What ‘good’ or ‘bad’, ‘true’ or ‘false’, ‘sane’ or ‘ill’ really mean in a society and which relevance these distinctions and their combinations actually possess, are determined by what might be called the semantic program of a society. It is only this program which gets the world model going and keeps it running. Accordingly I propose to call this program of socially obligatory semantic instantiation of world models culture. It follows from this assumption that there is no society without culture and no culture without society. The program culture has to be constantly applied in order to survive. Consequently there cannot be either a society or a culture without acting

450 Siegfried J. Schmidt

individuals and vice versa. That is to say, mankind is the creator of culture(s), but any individual is a creation of his/her respective culture. Culture as program co-ordinates or co-orientates cognitive and communicative systems (or discourses), which are strictly separate from each other, by and through the collective knowledge cognitive systems have learned to produce. Culture as program continues problem solutions in the domain of sense and meaning thus rendering these solutions quasi-natural or self-evident. Culture is the main instrument to solve the two basic tasks of any society: reproduction of society and control of individuals. Reproduction is achieved via socialisation, which may be characterised as ‘implantation’ of the culture program into the cognitive domain of children. This implantation happens before the individuals can observe and evaluate what happens to them; that is to say, socialisation takes place ‘behind the back’ of the subjects of socialisation. With regard to culture — as with regard to language — the individual always comes ‘too late’. Only after having been socialised are they able to think about it or to put it into question. Culture, we might say, can only be observed and described in the framework of and on the basis of a specific culture. Our instruments and strategies of world making can never be observed from outside, because we necessarily make use of them while observing — what so ever. The culture program controls individuals by reducing and determining their possibilities. This control has two facets: on the one hand options are excluded, and on the other hand complexity is reduced, because order can only be built upon order. Furthermore, the observation of excluded options in the course of secondorder observations can open new creative steps. It seems obvious that a culture program, too, cannot reasonably be judged by its truth or falsehood but only by its efficiency in operating world models. Its seeming self-evidence as well as the expected expectation that every member of a society shares its premises and applies its rules, its double fictiveness, makes it the powerful instrument of social and individual world making which it actually is.

5.

Linguistic fictions

Quite similar are the mechanisms of language. A major part of our socialisation is devoted to language acquisition. Children learn their mother tongue not in the way in which a second language is learned, that is they do not learn syntactic rules and lexical entries. Instead, children grow into the social behaviour of people in their social context, and only after a couple of years do they acquire the competence to distinguish between verbal and non-verbal components of successful social interactions. Learning one’s mother tongue means learning (by doing) how to operate semiotic materials in such a way that operations lead to the intended end, that frustration or even punishment are avoided, and that praise and love instead are achieved.

Operative fictions

It is very important to emphasise the crucial role materiality plays in “languaging” (to borrow a term from Maturana). The components of a language can be described as semiotic materialisation of condensed social experience gained and conventionalised in communication processes. That is to say: since cognitive systems cannot be observed in a direct way, and since what we call meaning can only be located in cognitive domains, communication cannot function in terms of an exchange of meaning. Instead, observable semiotic materials like words and utterances are exchanged which can be used by cognitive as well as communicative systems to construe meanings according to their systems’ rationality. In the process of socialisation children learn how semiotic materials of a specific kind are used in specific situations by people they trust in. In this way children learn how ‘we speak’, what can be done with semiotic materials, or how the verbal component of social behaviour works. Texts — like other media givens — couple the strictly separate cognitive domains because speaker and hearer expect that the other has undergone a comparable linguistic socialisation and therefore shares the imputed commonsense knowledge regarding the way language is used in a socially acceptable way. It follows from these considerations that meaning and reference are purely linguistic operations which refer to common sense-knowledge and communication and not at all to an observer-independent ‘reality’. Languages, too, can neither be true nor false, but only more or less efficient. Communication works because every participant counterfactually imputes to all other participants the same collective knowledge concerning the socially viable and successful use of semiotic materials (means of verbal expressions) acquired in their socialisation. Taking into account that everybody lives another life on quite different cognitive, emotional and social premises, this imputation is extremely unlikely. Nevertheless it turns out that exactly this unreasonable hypothesis enables communication. It is a highly efficient operative collective fiction embedded in the network of world models and culture programs. I hope that it has become clear that and how the connection between society, culture, language and individuals can be modelled as self-referential and selforganising, in other words as an auto-constitutive process. Models for reality, their semantic interpretation and evaluation by cultural programs in communication processes, and language as an operative system of distinctions and descriptions mutually determine one another.

6. Media fictions When we look at the historical development of societies, these big fiction machines have accelerated their tendency towards operational closure by means of media systems and functional differentiation. By ‘media’ I understand the systemic

451

452 Siegfried J. Schmidt

configuration of communication-instruments (for instance natural languages), all respective technologies needed to distribute media products (for instance cameras, PCs), the social-systemic organisation of the production and distribution of media products (for instance broadcasting companies) including its social, economic, juridical and political aspect), and finally the media products or offers which result from this system (for instance books or e-mails). The modern media, starting with letterpress printing, gave rise to the emergence of two powerful operative fictions: (a) the fiction of the public sphere with a generalised public opinion; and (b) the fiction that everybody participated in the public sphere by tuning their own opinions to the (imputed) public opinion. The power of these fictions is their capacity to produce facts by fictions. How does this work? (See Schmidt 1996.) The general principle underlying this process is, once again, reflexivity. A public sphere grows out of the mutual perception of individuals or of published opinions (statements on statements). As Merten has argued, people guard their own opinions by orienting them towards those opinions which are presented by the media as the public opinion. Merten defines ‘public opinion’ as that “…process of communication which selects those relevant or supposedly relevant facts or problems which are established as subjects on which first and foremost the media produce opinions. The presentation of opinions in the public sphere provokes a selection of relevant or supposedly relevant opinions which are or seem to be accepted by a majority, thus deploying political effects” (1987: 331; my translation, emphasis in original). Contrary to face-to-face communication, media communication may be described as virtual communication. This communication is rather diffuse, but it virtually covers a whole society regarding its ideas, expectations and opinions. Public opinion, public sphere, and the media build a self-constituted network of relations: the media enable the rise of a public sphere which is perceived in terms of the public opinion spread by the media. Media and the public sphere have provoked another important operative fiction: the institutionalised construction of images by Public Relations. Images are social constructs which may be described as coherent schemata consisting of cognitive and emotive structures assigned to persons, organisations, products, events or ideas. Images are built first of all for those referential items which are beyond immediate personal experience or accessibility for the majority of people in a society. Images are mostly built upon public opinions; they are neither objective nor true, but intertwine information and opinions which are relevant in the public sphere and which serve the purpose to construe a desirable picture of persons, organisations and so on. in the public sphere, and this means in the media system. Therefore they have to be presented in the media repeatedly and need support from opinion leaders and trend setters — or by public opinion itself. (For details see Merten and Westerbarkey 1994.)

Operative fictions 453

Apart from Public Relations, the other important media-effected kind of fictive communication which has developed in the last two centuries is advertising. But whereas Public Relations at least claims to be authentic and reliable, participants in ad-communication know that advertising is not governed by truth-principles but by the laws of the economy of attention. Advertisement is biased, and everybody can know that. It stages its products, it tells only positive stories, it turns simple products into heroes and tries to allure as many people as possible to believe in its promises that the possession of this product will solve all respective problems. Even if ads do as if they were seriously informing the public, this again is only a special strategy of its mise en scène of products, persons, and messages. Fictive images in PR-campaigns presented as trustworthy in the public sphere can and do have real consequences, just as the fictive staging of products in advertisement campaigns can. Both types of operative fictions turn out to be able to create facts via the consequences they trigger — unfortunately we know very little about the mechanisms. Both rely completely on media and on the public sphere — and both tend to take over control of the media by commercialising public communication, which is perhaps the most astonishing case of the production of facts by fictions in modern history.

7.

Pictures of reality – the reality of pictures

As I have already mentioned, the history of journalism is accompanied by the question whether or not journalists present objective reports about ‘reality’. Until today, most journalists claim that their reports on real events are authentic, reliable, and as objective as a necessarily selective representation can be. Most of them subscribe to the postulate that news follows the event and represents it in the respective medium. Their evidence is the fact that important events are reported quite similarly in the various media. From a constructivist point of view, the argumentation goes into the opposite direction: events follow the news. The gist of this argumentation is this: traditional ontology and epistemology are dualistic in so far as they start with the (mostly tacit) conviction that subject and object, reality and perception, language and reality exist and that the problem to be solved consists in explaining how they relate to one another and whether or not this relation is true. Consequently, when it comes to media they assume that reality exists and that the media try to represent this reality as perfect as possible. (My) constructivism argues in a non-dualistic way. Following Mitterer (1999) I do not start with an ontological assumption like “X or Y exists”, but I try to find out how we produce what we afterwards treat as real. I have tried to epitomise this approach with reference to perception in section 4 of this Chapter. As to media, I

454 Siegfried J. Schmidt

draw the following consequences from this kind of approach: media construe media realities by operating along their own systems-specific processual modes which alone determine which event is made an event in the public sphere by the news. It is a well-known fact that we cannot escape the illusion of pictures which automatically carry with them the pretension to be authentic: pictures claim they cannot lie. Therefore we are especially bewitched by pictures, and above all by TV-audiovisuality, and we quasi-automatically obey to the pictures’ imperative to connect them directly to items in the ‘real world’. However, my claim is that, while watching pictures or films, we are watching the reality of these pictures and not pictures of reality. Every medium, I aver, creates its own realities: media-realities. And when we act in non media-mediated environments we construe — as observers (see section 4 above) — our daily worlds of experience which cannot simply be transferred into media worlds. That is to say, we have to take into account that worlds are outcomes of highly conditioned world-making processes. The relations between these worlds are also constructed by observers and not simply given. Accordingly, we apply rather different and changing strategies to evaluate our relations to the relations between media-worlds and our worlds of experience.

8. Media-technological fictions There remains to be considered a fifth type of fictions efficient in the world-making process of modern media societies: media-technological fictions. Media technologies from the printing technique to PCs and the Internet have expanded the possibilities to virtualise communication and interaction, to create artificial environments to an extent that they seem to be ‘real’, to digitalise all media and to transform one medium into others, and to unmoor communication from our bodies as well as from space and time. Hypermedia demand completely new kinds of producing and receiving (or better: making use of) media offers, which are no longer dominated by linearity and hierarchical hermeneutic ordering but by a rhizomatic design and rhetorical strategies. As I have mentioned above, virtualisation has forced us to reshuffle our ontological categories. What is reality, what is fiction in times of the internet and virtual realities? Since ‘reality’ has been transposed into the plural, it has gained attractiveness for many people. The question is no longer whether or not we objectively perceive and recognise ‘reality’ . Instead, many people are interested in finding out how realities can be constructed and used, how they can be transformed into one another and how they can be induced to disappear again. Roy Ascot, a computer artist, claims: “We are much more interested in the virtualisation of the world than in its representation” (my translation). And the late Vilém Flusser once said that we are going to live in a plurality of worlds, where it will no longer be reasonable to

Operative fictions 455

find out which one is real and which one fictive; we will no longer be able to decide which one is a natural given and which one a media artwork. Quite evidently, dichotomous ontologies, which compel unequivocal decisions, do no longer suffice for the purpose of orienting ourselves in mediaculture societies. We have to shift from ontological to operative observations: how do we construe a certain type of reality? Under which conditions do we ascribe a certain degree of reality-value to results of our constructions? Will people soon stop asking for the status of reality altogether, because they want to be entertained or thrilled and nothing else? Maybe one day we shall be perfectly content with a reading of ‘reality’ as ‘temporarily pragmatised virtuality’. Consequently, ‘reality’ will have to be temporalised and proliferated. I am sure the distinction between ‘reality’ and ‘virtuality’, ‘fiction’, ‘simulation’ and so on will not simply disappear, because we need, to quote Bateson, differences which make differences. But the two parts of our distinctions, even in the ontological domain, will no longer be regarded to be stable and identical, thus demonstrating in an irrefutable way that they are outcomes of our observations and descriptions. And then everybody will know what constructivists have tried to explain. In addition, world making in mediaculture societies can be described in terms of modalising and modularising. Modalising means, as explained above, that users ascribe various modalities to the various virtualities which result from constructive processes. The traditional dichotomy true/false is replaced by a variety of modalities including ‘indifference’. Modularising means that world makers employ modules from quite divergent domains in their world construction processes: own experiences and media offers, public opinions and commercials, images, dreams etc. In sum: the traditional dichotomy fact versus fiction, which has dominated European thinking for centuries, is going to be replaced by a dynamic interplay of the two categories. Facts generate fictions, and fictions generate facts. Or to be more precise: what we deem to be facts generates what we deem to be fictions and vice versa. In mediaculture societies, every day a huge wave of media worlds swashes into the cognitive domains of media users. Only a tiny fraction of these worlds corresponds with people’s own experiences. The rest is, as it were, raw or play material for constructive processes of media users. Consequently, the ontologically interpreted distinction between fact and fiction itself turns into a fiction — a useful fiction, by the way.

456 Siegfried J. Schmidt

9. Conclusions Societies, I aver, and especially media culture societies, are fiction machines. In this chapter I have concentrated on five types or levels of fiction: – – – – –

fiction/fictionality as definiens of literary discourses in the social system literature; the fiction of the generality of collective knowledge (linguistic as well as encyclopaedic knowledge); the fiction of a homogeneous application of the culture program by all members of a society; the fiction of a society-wide public sphere and the authenticity of public opinion; the fiction of a general virtuality of realities.

I hold the view that these fictions are the very reason for the functioning of interaction and communication processes among operationally closed systems. In other words: societies function according to the principles of reflexivity and operational closure. They would never ever have emerged if they were to function according the principle of objectivity.

10. Epilogue The fictions I have inspected in this paper are of very different kinds: whereas social and cultural fictions provide the very basis for sociality among individuals, media fictions and technological fictions emerge in the course of history and are shaped in quite different forms. Fictionality as a definiens of discourses in literary systems differs from the other types of fiction in that it characterises the reference mechanism of a discourse as a whole, switching the communicators’ interests from the primary distinction true/false to the distinction aesthetic/non-aesthetic. I assume that all types of fictions, literary as well as operative fictions, share one common feature: they depend upon the availability of communication instruments like languages or pictures and of media from writing to computer. The difference between the various fictions is constituted by the difference between the media. Whereas social and cultural fictions already emerge on the basis of reflexivity created by (non-verbal and verbal) communication, media fictions like public opinion, PR-images or advertisement-stagings are bound to so called mass media. Of course fictions and images as well as reports on interesting events have been with us for ever. But the systematic and institutionalised production of images via mass media, the continuous report on actual events of public interest via mass media by the journalism system as well as the constitution of a functional social system literature on the premises of a pragmatic discourse quality (viz. fictionality) are results of the eighteenth and nineteenth century.



Operative fictions 457

The social systems literature, journalism, PR and advertising constitute one another because they all result from a specific management of the distinctions ± reality, ± truth, ± information, ± actuality, ± bias and ± mass media. This management of distinctions serves the purpose to produce and at the same time to consume the attention of the public, but for quite different aims. The rise of modern media societies as a result of co-evolutionary processes, from changes in the economy and culture to technological and demographic changes, culminates in a proliferation of observation(s) and observational problems. In media societies, all and everything can be observed (except society as a whole), but at the same time it becomes evident that the observations constitute the phenomena observed. Societies with fully developed media systems cannot avoid observing their blind spots of observations, too. And the most powerful blind spots are their operative fictions. Media thus compel insight into contingency, that is into the very nature of operative fictions. Media tell us the story of inner- and intracultural relativism, with all of its consequences, as everybody knows, ranging from new creativity to various kinds of fundamentalism. Whether we like it or not, we must accept the new roles of fact and fiction in the new spectacle written by media systems. As a consequence, literature, too, will have to find a new role in the further developing media system in the years to come. The growing awareness of the efficiency of operational fictions as caused by modern media systems will inevitably influence the way we interpret the fictionality of the literary discourse. What will be the result? I do not know.

References Merten, K. 1987. “Öffentliche Meinung”. In Handbuch der Politikwissenschaft, A. Görlitz and R. Prätorius (eds), 327–332. Reinbek: Rowohlt. Merten, K. and Westerbarkey, J. 1994. “Public opinion und public relations”. In Die Wirklichkeit der Medien. Eine Einführung in die Kommunikationswissenschaft, K. Merten, S.J. Schmidt and S. Weischenberg (eds), 188–211.Opladen: Westdeutscher Verlag. Mitterer, J. 1999. Die Flucht aus der Beliebigkeit. Klagenfurt: Drava. Schmidt, S. J. 1976. “Towards a pragmatic interpretation of ‘fictionality’”. In Pragmatics of Language and Literature, T.A. van Dijk (ed.), 161–178. Amsterdam etc: North- Holland Publ. Comp. Schmidt, S. J. 1980. “Fictionality in literary and non-literary discourse”. Poetics 9: 525–546. Schmidt, S. J. 1989. Die Selbstorganisation des Sozialsystems Literatur im 18. Jahrhundert. Frankfurt/M.: Suhrkamp. Schmidt, S. J. 1996. “Beyond reality and fiction? The fate of dualism in the age of (mass) media”. In Fiction Updated: Theories of Fictionality, Narratology, and Poetics, C. A. Mihailescu and W. Hamarneh (eds), 91–104. Toronto & Buffalo: University of Toronto Press. Schmidt, S.J. 1992. “Conventions and Literary Systems”. In Rules and Conventions, M. Hjort (ed.), 215–249. Baltimore & London: The Johns Hopkins University Press.



List of contributors

Professor Ziva Ben-Porat Poetics and Comparative Literature Tel Aviv University Tel Aviv Israel

Art Graesser Psychology University of Memphis Memphis TN USA

Gerald C. Cupchik Psychology University of Toronto at Scarborough Scarborough ON Canada

Norbert Groeben Psychology University of Cologne Cologne Germany

Sarita Dev Word and Image Studies Free University Amsterdam The Netherlands

Jemeljan Hakemulder Poetics Free University Amsterdam The Netherlands

Katinka Dijkstra Psychology Florida State University Tallahassee FL USA

David Hanauer Education Tel Aviv University Tel Aviv Israel

Rachel Giora Linguistics Tel Aviv University Tel Aviv Israel

Johan Hoorn Poetics Free University Amsterdam The Netherlands

460 List of contributors

Suzanne Janssen Art and Culture Studies Erasmus University Rotterdam Rotterdam The Netherlands

David Miall English University of Alberta Edmonton Alberta Canada

Bianca Klettke Psychology University of Memphis Memphis TN USA

Wouter de Nooy Art and Culture Studies Erasmus University Rotterdam The Netherlands

Joost Kloek History University of Utrecht Utrecht The Netherlands

Brigitte Scheele Psychology University of Cologne Cologne Germany

José de Kruif History University of Utrecht Utrecht The Netherlands

Siegfried J. Schmidt Communication Studies University of Münster Münster Germany

Garry Leonard English University of Toronto Toronto Canada

Dick H. Schram Poetics Free University Amsterdam The Netherlands

Caroll Madden Psychology Florida State University Tallahassee FL USA

Margrit Schreier Psychology University of Cologne Cologne Germany

Colin Martindale Psychology University of Maine Orono Maine USA

Robert Stanfield Psychology Florida State University Tallahassee FL USA

List of contributors 461

Gerard J. Steen English Language and Culture Free University Amsterdam The Netherlands

Willie van Peer German as a Foreign Language University Munich Munich Germany

Ingrid Stoeger German as a Foreign Language University Munich Munich Germany

Kees van Rees Marketing and Sociology of Books Tilburg University Tilburg The Netherlands

Ed Tan Word and Image Studies Free University Amsterdam The Netherlands

Hugo Verdaasdonk Marketing and Sociology of Books Tilburg University Tilburg The Netherlands

Koen van Eijck Leisure Studies Tilburg University Tilburg The Netherlands

Rolf Zwaan Psychology Florida State University Tallahassee FL USA



Subject index

A adventure game 292, 293, 298, 301, 309 adverbials 74 advertising 261, 327, 331, 424, 427, 430, 431, 433, 434, 440, 453, 457 agents 8, 57–65, 64–68, 149, 190, 323, 327, 328, 330, 337, 340, 346, 366 aggression 187, 192, 198, 199, 207, 208, 209, 211–218, 231 aging 87–89, 98, 281 allusional pretense 168, 169 alterity 295, 445 ambiguity 26, 27, 40, 134, 135, 136, 139, 140, 167, 240, 295 analogies 155–158, 160, 206, 330, 379, 383, 394 analogy 149–151, 155–161, 252, 266, 409 atelic 75–78, 81–83 attribution 231, 239, 246, 345, 394, 448 autonomy 8, 110, 218, 246, 280, 316, 410, 437, 444, 446 B balance model 400, 401 balance theory 360–362, 364, 366, 368, 370, 371, 375, 376 book history 320 browsing 13, 29, 291–296, 299, 301, 303, 305, 309 C case study 10, 39, 318, 372, 374 catharsis vi, 11, 12, 188, 201–212, 214, 216–220, 418, 425 catharsis as purification 203, 212, 217

catharsis hypothesis 11, 201, 212 censure 198 coding 23, 40, 44–46, 48, 50, 52, 210, 276, 376 cognitive linguistics 11, 73, 84, 146, 147, 149, 151 collective knowledge 444, 447–449, 450, 451, 456 communication 6, 17, 25, 26, 47, 48, 59–61, 114, 141, 168, 170, 181, 256, 264, 287, 317, 323, 346, 353–355, 417, 423, 424, 433, 436, 440, 444–447, 451–454, 456, 460 comparative analysis 49, 347 comparisons 138, 139, 156, 287, 319, 321, 323, 338, 339, 346, 347, 348, 383 comprehension of irony 177 conceptualization 244, 246, 248, 249, 252, 255–259 configuration analysis 48, 49 content analysis 39, 46–48, 50, 52, 238, 268, 362, 376, 395 context 4, 5, 10, 13, 17, 18, 21, 25, 26, 35, 38–42, 48, 49, 61, 81, 83, 98, 108, 111, 114–116, 133, 141, 153, 167, 168, 169, 171–183, 186, 257, 316–318, 320, 321, 324, 351, 384, 409, 412, 414, 417, 419, 425, 435, 436, 438, 450 contingency 242, 417, 457 creative personality 18–20 critic 5, 110, 115, 330, 346, 370, 373–375, 379, 380, 381, 383, 385–391, 395, 398, 399, 400, 434

464 Subject index

criticism 10, 16, 28, 59, 64, 109, 110, 111, 112, 114, 115, 126, 170, 174, 175, 178, 179, 181, 197, 208, 209, 234, 330, 331, 343, 354–356, 360, 369, 370–376, 391, 394, 399, 402, 405, 425, 427, 430, 439 cross-media artistic styles 398 cultural participation 266, 284, 350 cultural studies 7, 421, 423, 427, 436, 438, 439 culture vi, 1, 2, 7, 11–13, 16, 131, 187, 226, 237, 243–248, 250, 251, 253–259, 265, 269, 270, 275, 286, 287, 300, 312, 318, 322, 323, 324, 326, 328, 330, 332, 336, 347–349, 351–356, 359, 365, 369, 384, 407, 413, 415, 416, 419, 421, 422, 423–427, 429–436, 438, 439, 440, 443, 449–451, 456, 457, 460, 461 D data displays 44, 45, 48 deconstruction 113–115 deconstructionism 402, 403 defamiliarizing 409, 410 dehabituation 414–418 description 4–6, 38, 316 deviation 24, 26, 72, 79, 121, 122, 131, 133–135, 249, 305, 388, 417 direct access 167, 168, 172 displacement 71, 72 dissonance theory 290, 303 dynamic representation 76, 77 E emotion theory 218, 222 emotional learning 217, 218 emotions 57–59, 67, 191, 195, 198, 204, 208, 217, 226, 237, 312, 382, 418, 425, 436, 440, 448, 449 empathy 206, 216, 229, 235, 236, 237, 238, 312 empirical literary history 315–322 empirical study of literature 1, 2-8, 10–12, 35, 53-56, 69, 98, 107, 109,

129, 315, 324, 351, 354, 355, 396, 402, 418, 440 equivalence 111, 132–134, 137, 168, 169 equivalent processes 167, 168, 170, 172 evaluation 4–5, 117, 121, 332, 370, 372, 373, 381, 391, 428, 444 expectations 13, 14, 133–135, 169, 234, 266, 272, 282, 306, 308, 309, 315, 319, 320, 445, 448, 452 experiment 10, 21, 64, 81, 98, 100, 112, 117, 129, 130, 135, 137, 141, 186, 189, 191, 196, 197, 198, 209, 213, 229, 233, 234, 235, 237, 290, 302–306, 309, 397, 403, 426, 438 experimental aesthetics 3, 437 expertise vii, 63, 87, 95–102, 145, 333, 334 explanation 4, 5, 38, 79, 90, 93, 187, 191, 198, 199, 232, 246, 249, 290, 308, 316, 343, 348, 399, 403 F facts 64, 186, 318, 364, 443, 452, 453, 455 fairy tales 11, 50, 187–189, 196, 197, 365, 366, 375 familiarity 133, 134, 171, 181, 252, 343, 344 fan effect 90, 102 feminism 404 fiction machines 444, 449, 451, 456 fictionality 7, 258, 444, 445, 447, 456, 457 fictions 443, 444, 447–457 cultural fictions 444, 449–450, 456 linguistic fictions 450 literary fictions 444 media fictions 444, 451, 456 media-technological fictions 454 nonliterary fictions 443 operative fictions 443, 444, 447, 448, 452, 453, 456, 457 social fictions 444, 447–449

Subject index 465

technological fictions 444, 454, 456 fictivity 444, 445 foregrounding 26, 111, 120, 128, 135, 324, 408–413, 415, 417, 419 formalism 4, 10, 109, 111, 114, 115, 120, 143, 382 frustration-aggression hypothesis 207 G gatekeeping 328–330, 332 graded salience 171, 172, 176–178 H habitus 267, 268, 344–346, 350 histoire du livre 316, 317, 319 historical present 77 humanities 2, 3, 13, 140, 141, 191, 240, 317, 324, 376, 395 hypertext narrative 293, 294, 295, 300–302, 309 hypothesis testing 2, 6, 37, 45, 47–49, 51, 186–189, 320–321, 417–418 I identification 27, 190, 228, 232 idiographic 6, 7, 50, 395 images 24, 443, 452, 453, 455, 456 indirect negation 178, 183 indirectness 165, 182 inhibition 87–91, 198, 207, 210, 211, 212, 215, 216 insufficient reward 290, 303, 308 interactive stories 289, 291, 292, 294, 293–295, 308, 309 intergroup cognitions 229 intersubjectivity 41, 45, 46, 48 interview 39, 42, 43, 50, 52, 103, 124, 179, 219, 284 intrinsic reward value 302, 305 irony 11, 165, 167–173, 175, 176, 177–180, 435 J joint pretense 169, 178

journalistic system 316, 334, 415, 446, 453, 456–457 K knowledge tracking 64, 67 L language 8, 10–12, 22, 24, 26, 46, 50, 61, 71–74, 76, 78, 79, 80, 82–85, 89–91, 109–116, 119, 120, 123, 124, 130, 131, 133, 134, 136, 139, 147, 148, 152–154, 156, 165–67, 169, 172, 173, 175, 176, 190, 210, 214, 244, 245, 247, 250, 257, 317, 332, 379, 381, 382, 395, 408–416, 419, 445, 449, 450, 451, 453, 457, 461 lateralized readiness potential (LRP) 136, 141 literacy 108, 126, 127, 251, 341, 382 literal 135, 137–141, 150, 155, 157, 167, 168–180, 248 literary criticism 16, 59, 64, 354, 355, 356, 360, 369–376, 391, 399 literary field 344, 351, 363, 364, 368, 369–372, 375, 376, 381 literary history 13, 21, 22, 32, 100, 318–322, 396, 419 literary reading vi, 3, 7–9, 11, 12, 107, 109, 146, 153, 160, 228, 320, 407–409, 411, 414, 419 literary text 8, 10–12, 26, 27, 50, 88, 96–100, 109, 110, 112, 117, 118, 229, 331, 379, 380, 381, 382, 391, 411, 445 longitudinal research 346 M market 129, 263, 325–329, 331, 335, 336–338, 340, 346, 348, 349, 351, 352, 354 marxism 4, 15, 404 mass communication 221, 287, 323, 346, 353, 354, 423, 433 mass culture 423, 424, 440

466 Subject index

mass media 2, 288, 355, 429, 430, 433, 456, 457 meaning 8, 9, 25–29, 47, 48, 88, 92, 97, 108, 110, 112–116, 119, 121, 124–126, 167, 168, 169–180, 332, 343, 365, 380, 381, 382, 384, 385, 388, 389, 391, 409, 410, 412, 413–415, 427, 437, 443–446 media 2, 3, 7, 13, 28, 29, 35, 39, 201, 202, 209, 211, 212, 213, 219–220, 232, 243, 244, 261–270, 272, 273, 275, 276, 277, 280, 282–288, 290, 291, 292, 293, 312, 316, 324, 334, 351, 352, 354, 355, 398, 405, 424, 427, 429, 430, 433, 436, 439, 443, 444, 446, 451, 452, 453–457 media market 263 media use 261, 262, 265–269, 272, 280, 283–286 mediators 14, 323, 325, 328, 330, 332, 334, 346 mental illness 18, 19 mental representation 76, 92, 99, 220, 245, 246 metaphor 10, 11, 57, 131, 134–140, 139–141, 145–149, 151–159, 175, 182–184, 244, 249, 256, 359, 370, 410, 413, 414, 419, 427 metaphor identification 10, 145– 149, 151, 152, 156, 159, 160 metaphor processing 131, 135, 136, 137, 140, 141, 146, 152, 153, 159, 160 methodology 3, 36, 107, 124, 145, 174, 186, 188, 319, 376, 405 model 2, 9, 11, 12, 21, 23, 25, 26, 57, 58, 72–74, 82–84, 88, 91–95, 97–100, 116, 135, 137–139, 138, 140, 150–152, 156, 158, 160, 169, 170, 172, 178, 198, 203, 205, 208, 209–211, 216, 217, 219, 231, 232, 245–258, 270, 276, 286, 317, 326, 331, 332, 363, 364, 365, 368, 373, 380, 390, 400, 401, 437, 444, 445, 447, 449

modernism 427–436, 438 modernity 331, 427–436, 438 moral development 209, 213, 214 movements 326, 368, 371 multicultural society 227, 228 multiple job-holding 338, 340 N navigation 29, 290–292, 293, 294, 296, 301, 304–306, 309 network analysis 14, 361, 362, 364, 369, 370 networks 14, 43, 44, 64, 264, 288, 325, 328, 340, 341, 343, 347, 351, 352, 354, 357, 360, 376, 401, 405, 424, 447 neurosis hypothesis 19 new criticism 10, 109–111, 114, 115, 402, 425, 427, 430 new historicism 404 nomothetic 6, 7, 395 nonliteral 151, 154–157, 167–169, 172, 175, 177 O object level 379–381 objectivity 20, 35, 37, 39, 45, 46, 425, 427, 447, 456 observational learning 210, 212, 214 observations 5, 63, 159, 185, 189, 259, 318, 330, 336, 345, 348, 382, 444, 446, 450, 455, 457 occupational characteristics 333 oral literature 412, 413 originality 22, 326, 327, 425, 426, 438 outgroup cognitions 231, 232, 236, 240 P parallel anomaly model 137–139 patronage 13, 325–327 perception 12, 20, 28, 133, 170, 219, 228, 229, 230–233, 238, 240, 266, 267, 405, 417, 426, 440, 447, 452, 453 perceptual symbols 84

Subject index 467

periodization 316, 371 physics 348, 396–399, 401, 402 pictures 229, 230, 302–306, 453, 454, 456 plot 14, 57, 59, 67, 72, 96, 225, 333, 365, 373–375 poetic function 111, 121 poetry reading 10, 108–116, 119, 120, 121–126 polar integrations 20 polyfunctional structure 26 polyvalent meaning 26 popular culture 2, 7, 269, 300, 332, 421–427, 431, 433, 434, 436, 438 postmodernism 435, 436 pretense 165, 168, 169, 178, 181 print media 261–269, 272, 273, 275, 276, 280, 283, 285 production process 7, 17, 23, 332 psychoanalysis 19, 185–188, 191, 195, 196–200, 207, 396 psychology of reception 201, 206, 209, 216–218 psychophysiology 129, 131, 419 public opinion 452, 456, 457 public relations 452, 453, 457 public sphere 421, 443, 452–454, 456 publishers 263, 315, 317, 323, 325, 326, 328–332, 337, 340, 343, 346 purging 12, 202–206, 208–212, 216, 217, 218, 220 Q qualitative methods 6, 35, 36, 39, 40–47, 49, 51 quantitative methods 36, 40, 45–47 R reader response 10, 30, 59, 109, 112, 113, 185, 186 reader typology 269, 284 reading behavior vi, 13, 197, 261, 270, 272, 276, 277, 282, 284, 285, 315 reading conventions 123

reading process 3, 88, 107–116, 120, 122–125, 198, 394 reading public 317, 320, 383 reading societies 315 reality 20, 21, 188, 189, 245, 246, 248–250, 258, 316, 318, 319, 326, 331, 359, 362, 364, 404, 423, 424, 431, 434, 443, 445, 446, 451, 453–455, 457 reception 3–5, 7–12, 17–18, 23, 25–30, 35, 36, 40, 42, 43, 46, 48, 50–52, 109, 112, 201, 203–211, 213, 214, 216, 217–220, 315–317, 320, 350, 353, 354, 365, 410, 411, 413, 414, 415, 418, 421, 427, 439, 444 reception theory 4, 10, 109, 112 recruitment 323, 335, 341, 342, 345 reductionism 398, 401 reference 23, 25, 178, 179, 202, 203, 207, 208, 211, 219, 257, 331, 345, 347, 348, 355, 402, 409, 421, 432, 446, 447, 449, 451, 453, 456 reflexivity 205, 447, 448, 452, 456 relevant inappropriateness 169, 181 reliability 6, 35, 37, 39, 45, 46, 102, 149, 154, 230, 261, 448 repressed feelings 11, 189, 197 reviewer 5, 379–381, 384–389, 391 role-taking 232–239 S salience 11, 60–63, 171–173, 175, 176, 177, 178, 180–183 schemata 56, 95, 415, 417, 448, 452 science 2, 3, 6, 36, 131, 197, 312, 316, 321, 347, 348, 352, 356, 360, 363, 396, 397, 398, 399, 401, 402, 411, 416, 418, 419, 439 serial anomaly model 137 similarity 10, 50, 111, 132–135, 140, 149, 150, 155, 227, 251, 257, 367–369, 371, 426, 440 simile 155–161 situation model 9, 74, 82, 83, 88, 92, 93–95, 97–100, 150

468 Subject index

social construction 254, 397, 424, 434 social distance 230, 234, 236, 238 social science 6, 316, 416 social structure vi, 359, 362, 363, 368, 369, 371 socialization 20, 269, 315, 341, 342, 350 special (sequential) process 167 standard pragmatic model 169, 178 statistics 37, 129, 131, 176, 333, 354, 361, 362, 369, 386, 396, 403 stereotypes 24, 229, 231, 232, 239, 240 structural affect theory v, 57, 58, 67, 308, 309 structural prominence hypothesis 61, 62, 63 structuralism 4, 143, 382, 402 stylistics 10, 33, 57, 109, 112, 114, 115, 149, 163, 411, 419 subjective theories 42, 43, 50, 51, 53 surfing 13, 289–292, 294, 301, 309 surprise 57, 58, 122, 134, 135, 140, 363 suspense 57, 58, 64, 134, 425, 437 T television viewing 261, 262, 265, 266, 267, 268, 272–276, 278, 280, 281–283, 285 telic 75–78, 81–83 temporal adverbs 77 temporal discontinuities 72 temporal markers 74, 81, 84 tense 73, 74, 76–78, 81, 233 text 3–5, 8–15, 25–30, 35, 36, 44, 46, 48–52, 58, 60–63, 72, 83, 88–101, 108–113, 115–119, 122, 124, 125, 126, 135, 143, 145, 147, 149, 150, 152–159, 187, 188, 190, 191, 195, 196, 206, 218, 229, 230, 232, 234, 236, 237, 238, 243, 259, 289, 290, 291–293, 295, 296, 298, 299, 300, 302, 303, 305, 306, 308, 309, 312, 313, 316, 317, 319, 331, 359, 364, 371, 379, 380, 381, 382, 384–386, 388, 389, 390, 391, 400, 403, 405, 409, 410–412, 417–419, 438, 445

text comprehension 88–101, 103, 109, 241, 312 theory v, 3–5, 10–12, 14–17, 22, 26, 29, 37–39, 43, 50, 51, 54, 55, 56–58, 72, 108, 109, 110, 112, 115–117, 120, 129–134, 140, 147, 148, 160, 168, 181, 186, 187, 188, 189, 191, 193, 194, 195–198, 204–208, 210–212, 214, 216, 218, 220, 231, 237, 261, 266, 267, 284, 287, 290, 303, 308, 309, 312, 313, 324, 326, 331, 339, 349, 352, 354, 357, 359–368, 370, 371, 374, 375–377, 380, 382, 394, 396, 397, 399–401, 404, 405, 415–419, 428, 430, 434, 435, 439, 440 time adverbs 74, 76, 79 time shifts 80 tinge 170, 178 U ultimate attribution error 231, 239, 241, 242 unconscious 185, 187–190, 197, 198, 447 V validity 6, 9, 37–40, 43, 91, 129, 152, 153, 202, 213, 217, 249, 261, 285, 318–320 verb 9, 73–76, 78, 81–83, 85 verb aspect 74–76, 78, 81, 82 vicarious experience 77, 204, 210 virtuality 443, 455, 456 visual culture 427, 440 W working memory 80, 87–91, 93, 94, 96–101, 103, 104 world model 444, 445, 449

Name Index A Aarseth, E. 311, 312 Abraham, K. 187, 200 Adams, C. 87, 92, 93, 101 Albert, H. 202, 221 Albrecht, M. 349, 350 Alderson, C. 112, 126 Alsbrook, E. 230, 240 Altick, R. 342, 350 Amis, M. 71, 72, 85 Anderson, A. 80, 85 Anderson, P. A. 92, 103 Andringa, E. 35, 46, 50, 53, 321, 322, 349, 350, 351 Anheier, H. 335, 341, 351, 369, 375 Arnheim, R. 425, 439 Arnold, M. 148, 155, 157, 432, 434, 435, 439 Attardo, S. 169, 181 Auster, C. 360, 364, 375 B Baer, J. 17, 30 Bakhtin, M. 59, 67 Bandura, A. 210–212, 221 Banfield, A. 62, 67 Barbe, K. 170, 181 Barber, B. 333, 351 Bargh, J. 231, 240 Baron, R. 207, 211, 221 Barsalou, L. 84, 85, 132, 142 Barthes, R. 121, 126, 132, 142, 290, 303, 308, 312, 415, 418 Bartlett, F. 426, 439 Baurmann, J. 28, 30 Baynham, M. 108, 126 Bazerman, C. 117, 126

Beardsley, D. 230, 240 Beardsley, M. 110, 128, 135, 137, 142 Beaugrande, R. de 15, 28, 30, 124, 126, 440 Becker, H. 324, 343, 349, 351 Beckett, S. 429, 432, 439 Ben-Porat, Z. 12, 13, 159, 161, 243, 257–260 Benjamin, W. 179, 422–424, 428, 435, 439 Benn, G. 30, 446 Berelson, B. 47, 53 Berger, A. 422, 439 Bergler, R. 209, 211, 221 Berkowitz, L. 208, 210–213, 221-223 Berlyne, D. E. 26, 30, 131, 134, 142, 360, 375, 425, 437, 439 Bernasco, W. 340, 351 Berry, R. 327, 351 Bestgen, Y. 81, 85 Bichler, J. 26, 30 Bierhoff, H. 207, 221 Black, M. 147, 161 Blau, J. 349, 351 Bleich, D. 112, 126 Blumer, H. 363, 375 Boesch, E. 218, 219, 221 Bohart, A. 204, 218, 221 Boll, Th. 217, 221 Bonham-Carter, V. 342, 351 Bönke, H. 213-216, 221 Booth, W. 165, 167, 169, 181, 289, 312 Bortz, J. 37–39, 53 Boschetti, A. 350, 351 Boudon, R. 362, 375 Bourdieu, P. 266–268, 283, 284, 286, 287, 324, 326, 328, 340, 343, 344,

470 Name index

345, 348-–53, 363, 370, 376, 381, 394 Bourg, T. 235, 241 Bourgeois, M. 101, 102 Bovair, S. 150, 152, 161 Bowlby, J. 414, 416, 418 Boxman, E. 340, 351 Boyer, P. 329, 351 Bradbury, M. 327, 342, 351 Braden, S. 327, 351 Bransford, J. 119, 126 Brasey, E. 328, 351 Breed, Wr. 328, 351 Brenner, C. 19, 30 Breuer, J. 206, 221, 223 Brewer, W. 26, 30, 35, 53, 57, 58, 67, 68, 82, 85, 86, 290, 308, 312 Brouwer, H. 320–322 Brown, R. 231, 241 Bruce, B. 61, 68 Bruning, R. 28, 30 Bryman, A. 40, 47, 53 Bullough, E. 219, 221 Buss, D. 416, 418 Buurman, P. 3, 15 Bystryn, M. 330, 352 C Calhoun, C. 362, 376 Cameron, L. 148, 161, 163 Cantor, M. 333, 352 Caramazza, A. 131, 142 Carley, K. 362, 376 Carreiras, M. 81, 85 Carroll, J. 408, 418 Cartwright, D. 360, 376 Chafe, W. 59, 68 Chaney, D. 423, 424, 439 Charle, C. 350, 352 Chartier, R. 316, 319, 322 Child, J. 334, 352 Christmann, U. 25, 30 Cioffi, F. 199, 200 Clark, E. 79, 85 Clark, H. 60, 65, 68, 165, 169, 178, 181

Clark, M. 98, 104 Clor, H. 329, 352 Coles, M. 134, 136–138, 141, 142 Collins, R. 368, 376 Colvin, C. 28, 30 Cook, T. 230, 241 Corrêa, A. 418 Coser, L. 330, 341, 350, 352 Coulthard, M. 6, 15 Craik, F. 94, 102, 103 Crane, D. 331, 341, 347, 349, 352, 355 Crews, F. 188, 199, 200 Crisp, P. 148, 152, 157, 161 Croft, W. 73, 84, 85 Cropley, A. 17, 28, 30, 31 Croteau, D. 261, 262, 286 Csikszentmihalyi, M. 20, 31, 342, 345, 347, 353 Culler, J. 113, 126 Cupchik, G. 7, 12, 14, 42, 46, 53, 58, 68, 421, 425–427, 437–441 Curcó, C. 168, 182 Curtius, E. 257, 260 Cvetkovitch, A. 434, 440 D Darnton, R. 316–318, 322 Daubert, H. 27, 31 Davis, K. 329, 352 De Haan, J. 269, 286 De Jong, M. 131, 142 De Kruif, J. 13, 315, 320–322 De Nooy, W. 12, 14, 331, 341, 352, 359, 368, 370, 371, 376 Deignan, A. 154, 161 Denzin, N. 39, 41, 53, 55, 56 Derrida, J. 113, 114, 126, 403 Desan, P. 359, 376 Dews, S. 170, 174, 178, 179, 182, 183 Dijkstra, K. 9, 87, 88, 92, 95, 96, 100, 101–103 DiMaggio, P. 266, 285–287, 326, 327, 341, 345, 350, 352 Dissanayake, E. 409, 411, 413, 417, 419 Dixon, R. 88, 102

Name index 471

Dollard, J. 207, 221 Dorfman, M. 50, 53 Dörner, A. 349, 352 Douglas, M. 341, 352, 435 Dowty, D. 73, 74, 78, 85 Dubin, S. 327, 329, 352 Duchan, J. 59, 60, 62, 68 E Eagleton, T. 407, 414, 419 East, E. 18, 31 Eckardt, B. von 186, 200 Eckensberger, L. 213, 222 Eco, U. 133, 142, 437, 440 Edelman, G. 418, 419 Edelstein, W. 218, 222 Elburg, J. 135, 142 Eliot, T. S. 162, 430, 434, 435, 440 Ellis, H. 18, 31 Epstein, S. 218, 222 Erb, E. 20, 31, 218, 222 Eriksen, B. 136, 137, 139, 140, 142 Eron, L. 211, 222 Even-Zohar, I. 257, 260 Eysenck, H. 199, 200 F Faulkner, R. 341, 352 Fay, W. 62, 68 Featherstone, M. 269, 287 Fechner, G. 437, 439, 440 Festinger, L. 290, 303, 308, 309, 312 Fielding, N. 39, 42–44, 47, 53 Fine, G. 294, 312, 349, 352 Finnegan, R. 377, 413, 419 Fish, S. 112, 113, 126, 403, 405, 410, 411, 416, 419 Fisher, F. 230, 241 Fishwick, M. 422, 440 Fiske, J. 435, 440 Flap, H. 340, 352 Flashar, H. 203, 222 Fleischhauer, P. 28, 31 Fleischman, S. 74, 77, 78, 85 Flick, U. 39, 41, 53

Fludernik, M. 410, 419 Fodor, J. 169, 182 Fohrbeck, K. 327, 336, 353 Fokkema, D. W. 4, 15, 96, 102, 289, 322 Fornas, J. 421, 424, 440 Foster, A. 343, 352–356 Frankel, H. 230, 241 Franzosi, R. 365, 376 Freedman, A. 108, 117, 126 Freeman, D. 147, 160, 161 Freire, P. 108, 126 Freud, S. 19, 31, 131, 185, 187, 188, 190, 200, 206, 207, 221, 222, 435, 440 Friedrich, H. 26, 31 Frijda, N. 131, 142 G Gamboni, D. 350, 353 Ganzeboom, H. 266, 287 Gardner, H. 31, 154, 161, 223 Garner, R. 93, 102 Gartman, D. 345, 353 Geddes, G. 429, 440 Geel, R. 373, 376 Geen, R. 208, 210, 211, 222 Geiger, K. 37, 53, 230, 241 Gellner, E. 199, 200 Genette, G. 380, 381, 394 Gentner, D. 148, 151, 161 Gerhards, J. 335, 341, 350, 351, 369, 375, 376 Gernsbacher, M. 88–90, 102, 119, 126, 181, 182 Gerrig, R. 58, 60, 62, 68, 165, 169, 178, 181, 236, 241, 289, 308 Getzels, J. 342, 345, 347, 353 Gibbs, R. 57, 68, 120, 126, 145, 152, 161, 162, 163, 167, 170, 171, 174, 175–177, 181, 182, 257, 260 Gilmore, S. 324, 353 Gimmestad, B. 230, 241 Giora, R. 11, 165, 167, 168, 170, 171, 173–183 Gitlin, T. 333, 353

472 Name index

Gleich, U. 211, 222 Glenberg, A. 84, 85 Glucksberg, S. 167, 168, 183 Goatly, A. 146, 154, 160, 161 Goertzel, M. G. 18, 31 Goldberg, A. 73, 84, 85 Goldstone, R. 132, 142 Goode, W. 333, 353 Goodwin, F. 19, 31 Goody, J. 422, 440 Goranson, R. 211, 222 Grady, J. 151, 162 Graesser, A. 8–10, 57–60, 62–65, 64, 66, 67, 68, 72, 85, 86, 102, 105, 128, 170, 183 Granovetter, M. 340, 353 Graves, B. 96, 97, 102 Greenberg, C. 422, 423, 440 Greene, B. 397, 405 Greenfeld, L. 330, 331, 353 Greenwood, E. 333, 353 Greer, G. 342, 353 Greimas, A. 133, 142 Grice, P. 167, 169–171, 178, 183 Griff, M. 341, 342, 350, 353 Griffin, D. 325, 350, 353 Grimes, J. 80, 85 Grimm, brothers 53, 189, 200, 241, 317, 319, 322 Griswold, W. 329, 331, 332, 353, 355, 356, 376 Groeben, N. 3, 4, 8, 9, 14, 15, 17, 18–21, 24–27, 30–32, 35, 36, 39, 41, 42, 46, 48, 51, 53, 54, 55, 180, 183, 202, 210, 222, 224, 321, 322 Grünbaum, A. 199, 200 Grzesik, J. 28, 31, 32 Guilford, J. 21, 32 H Haas, G. 28, 32 Hage, P. 362, 376 Haiman, J. 165, 183 Hakemulder, J. 12, 225, 230, 240, 241, 289, 312

Halász, L. 30, 365, 376 Hall, R. 333, 353–355 Hamm, V. 90, 99, 102 Hanauer, D. 10, 107, 108, 116–118, 120, 122–124, 126, 127, 141 Hansen, E. 27, 32 Harary, F. 360, 362, 364, 376 Harker, J. 87, 102, 124, 127 Hartmann, D. 212, 222 Hasenfus, N. 398, 405 Hasher, L. 88, 90, 99, 102, 103 Hayes, M. 230, 241 Heckhausen, H. 204, 207–209, 222 Heider, F. 360, 368, 376, 400, 401, 405 Heilbron, J. 342, 353 Heilbrun, J. 337, 338, 353 Henle, P. 135, 137, 142 Herkner, W. 207, 223 Herzog, W. 213, 223 Hewstone, M. 231, 232, 241 Heywood, J. 148, 149, 152, 161–163 Higgins, E. 240, 241 Hintzenberg, D. 25, 32 Hirsch, P. 329, 353, 396 Hockett, C. 71, 85 Hoffer, A. 21, 32 Hoffstaedter, P. 26, 32, 116, 117, 123, 124, 125, 127 Hohler, F. 190, 192–196, 200 Holland, N. 26, 32, 199 Honeck, R. 145, 162 Hoorn, J. v, 3, 10, 15, 120–124, 127, 129, 131, 134–142, 145, 162, 417, 419 Hopfield, J. 401, 405 Hron, A. 42, 54 Hrushowski, B. 160, 162 Huber, G. 37, 48, 54, 221–223 Huberman, B. 311, 312 Huberman, M. 43–45, 47, 48, 50, 55 Huesmann, L. 211, 223 Hultsch, D. 102, 103 Hutcheon, L. 180, 183 Huxley, A. 21, 32

Name index 473

I Ibsch, E. 1–6, 15, 27, 32, 35, 36, 54, 57, 68, 129, 131, 141, 142, 267, 289, 351, 353 J Jackson, E. 88, 103, 184, 230, 241, 377 Jakobson, R. 10, 61, 68, 111, 121, 127, 132–134, 136, 137, 142, 147, 162, 317, 382, 412 Janssen, S. 1, 8, 12, 13, 15, 46, 54, 270, 287, 323, 331, 341, 343, 351, 353–355, 381, 384, 394 Jeffri, J. 336, 354 Jong, M. 210, 223 Johnson, B. 115, 127 Johnson, D. 229, 230 Johnson, M. 146, 151, 257, 260 Johnson Laizd, P. 73, 85, 133, 143 Jorgensen, J. 168, 183 Jose, P. 58, 68 K Kadushin, C. 341, 352, 354 Keenan, J. 120, 127 Kelle, U. 44, 54, 56 Kemper, S. 88, 103 Kesting, M. 26, 29, 32 Kimoto, C. 230, 241 Kintsch, W. 9, 15, 16, 59, 68, 72, 86, 88, 93, 95, 96, 99, 103, 104, 108, 120, 127, 128, 150, 152, 162, 163 Kirkeboen, G. 130, 142 Kittay, E. 147, 162 Kloek, J. 13, 315, 321, 322 Knulst, W. 96, 103, 262, 269, 279, 286, 287, 288 Koeller, S. 230, 241 Kohlberg, L. 209, 223 Konijn, E. 131, 142 Kornadt, H.-J. 207, 208, 211–213, 215, 223 Kotthoff, H. 169, 179, 180, 183 Kövecses, Z. 154, 162

Kraaykamp, G. 95, 96, 103, 262, 279, 287 Kreitler, H. 437, 440 Kreitler, S. 437, 440 Kress, G. 108, 127 Kreuz, R. 1, 15, 56, 68, 102, 142, 168, 183, 241, 419 Kris, E. 19, 22, 32 Kruger, B. 432, 440 Kubie, L. 19, 32 Kuhlen, R. 29, 32 Kuhn, T. 397, 405 Kuiken, D. 41, 54, 58, 68, 116, 118, 127, 146, 162, 410, 411, 418, 419 Kunczik, M. 211, 223 Kutas, M. 86, 135, 143 Kwong See, S. 89–91, 103 L Labouvie-Vief, G. 92, 101, 103 Lachenmann, J. 228, 241 Lakatos, I. 397, 405 Lakoff, G. 146, 148, 151, 162, 257, 258, 260 Lämmert, E. 28, 32 Lamnek, S. 38, 39, 42, 47, 54 Lamont, M. 266, 287 Landauer, T. 294, 312 Landis, J. 46, 54 Landwehr, J. 25–27, 31, 321, 322 Langacker, R. 73, 76, 77, 84, 85 Lange-Eichbaum, W. 18, 32 Langholf, V. 206, 223 Laurenson, D. 327, 342, 354 Lehnert, W. 58, 68 Lenman, R. 326, 354 Lenz, F. 187, 200 Leonard, G. 7, 12, 14, 421, 424, 433, 435, 438–440 Lessing, G. 203, 218, 223 Leuzinger, P. 206, 223 Lévi-Strauss, C. 365, 376 Levine, K. 108, 127 Levy, E. 342, 354 Light, L. 92, 103

474 Name index

Linville, P. 231, 242 Litcher, J. 229, 230, 242 Livingston, P. 2, 15 Loewenfeld, L. 18, 32 Long, D. 170, 183 Lotman, Y. 132, 143 Lowenthal, L. 329, 332, 354, 422–424, 440 Lukesch, H. 209, 223 Luserke, M. 203, 222, 223 Lüthi, M. 187, 200 Lyotard, J. 403–405 M MacMillan, M. 199, 200 MacNealy, M. 1, 15, 56, 102, 142, 241, 419 Macrae, C. 240, 242 MacWhinney, B. 84, 85, 127 Madden, C. 9, 71, 86, 163 Maffesoli, M. 269, 287 Magliano, J. 81, 85 Mallet, C.-H. 190, 200 Malmgren, L.-G. 41, 45, 54 Mandler, J. 79, 85 Manhart, K. 361, 376 Mann, T. 154, 157, 162, 185, 192, 200 Markert, J. 330, 354 Martin, H.-J. 71, 72, 126, 316, 322 Martindale, C. 6, 7, 12, 14, 15, 17, 21–23, 26, 32, 125, 127, 395, 396, 398, 400, 401, 403, 405, 417, 419, 439 Masaharu, A. 259, 260 Mayring, P. 39, 41, 47, 55 Mazur, B. 65, 68 McGuigan, J. 423, 424, 440 McKnight, C. 302, 311, 312 McMullan, W. 20, 32 McNeillie, A. 431, 440 McQuail, D. 261, 262, 264, 287, 350, 354 Meder, N. 28, 32 Meek, M. 108, 127 Mennell, S. 327, 354

Merten, K. 46, 55, 452, 457 Meutsch, D. 25, 33 Meyer, L. 341, 354 Meyer, M. 362, 376 Meyer, U.-W. 217, 223 Miall, D. 6, 7, 14, 41, 54, 58, 68, 116, 118, 127, 135, 143, 146, 162, 407, 410, 411, 418, 419 Miles, M. 43–45, 47, 48, 50, 55, 56 Miller, G. 73, 85, 133, 143, 150, 152, 162, 395, 405 Miller, J. 114, 127, 435, 440 ` Miller, L. 89, 92–95, 104 Millis, K. 59, 68 Miner, E. 251, 256, 257, 260 Minihan, J. 327, 354 Mio, J. 145, 162 Mirzoeff, N. 427, 440 Mitterer, J. 453, 457 Mockros, C. 96, 103 Moles, A. 437, 440 Morrow, D. 89, 92–95, 97, 103, 104 Moulin, R. 326, 330, 331, 336, 354 Mukarˇovský, J. 111, 112, 127, 133, 134, 143 Münte, T. 79, 84, 86 Murphy, G. 120, 127, 147, 162 N Nash, D. 342, 354 Navratil, L. 23, 33 Nell, V. 290, 312 Nesse, R. 417, 419 Neuman, W. 262, 287 Nichols, M. 204, 218, 219, 223 Nielsen, J. 291, 308, 312 Nies, F. 331, 354 Nochlin, L. 342, 354 Nussbaum, M. 227–229, 237, 242, 289, 312 O Oatley, K. 46, 55, 68, 289, 312 Ohtsuka, K. 26, 30, 57, 67, 82, 85, 86 Oldenbürger, H.-A. 50, 51, 55

Name index 475

Oosterbaan, M. 334, 354 Ormerod, B. 364, 376 Ortony, A. 68, 145, 161, 162 Ostermann, R. 51, 55 P Parke, R. 213, 223 Parkhurst, C. 325, 349, 350, 354 Pearson, N. 238, 327, 328, 354 Peckham, M. 398, 405 Peeters, C. 373, 376 Peleg, O. 181, 183 Pelles, G. 327, 355 Peper, D. 208, 223 Perfetti, C. 152, 162 Peru, J. 350, 355 Peterson, R. 266, 284, 287, 312, 324, 329, 332, 341, 348, 355 Pettigrew, T. 231, 242 Pevsner, N. 327, 355 Phillips, D. 62, 68 Picard, R. 264, 287 Pine, F. 21, 33 Pinto, L. 350, 355 Pohlenz, M. 223 Poincaré, H. 21, 33 Ponton, R. 350, 355 Powell, W. 330, 341, 352, 355 Pratt, M. L. 112, 113, 127 Praz, M. 397, 405 Preiser, S. 17, 19, 21, 22, 25, 28, 33 Prentky, R. 19, 33 Prince, G. 59, 68 Propp, V. 14, 15, 365, 366, 371–374, 376 R Raaijmakers, J. 133, 143 Radvansky, G. 59, 69, 72, 78, 82, 86, 88, 92, 93, 95, 97, 102, 103, 105 Ragin, C. 49, 55 Rank, O. 189, 190, 200 Rattok, L. 165, 166, 183 Reddy, M. 61, 68 Reichenbach, H. 73, 74, 86

Reid, P. 431, 440 Reinhart, T. 147, 162 Remak, H. 245, 260 Rennert, H. 24, 33 Richards, I. A. 44, 55, 127, 135, 137, 143, 147, 162, 403, 405 Ricoeur, P. 147, 162 Ridgeway, S. 341, 355 Rodden, J. 331, 355 Rogers, E. 362, 376 Rollin, R. 332, 355 Rorty, R. 227, 228, 234, 242 Rosch, E. 258, 260 Rosenblatt, L. 61, 69, 112, 127 Rosenblum, B. 327, 355 Rosengren, K. 267, 287, 288, 331, 349, 355 Rule, G. 113, 162, 169, 188, 210, 211, 213, 214, 223, 245, 254, 294, 382 Runco, M. 21, 23, 33 Rusch, G. 1, 15, 53–56, 126, 328, 355 Russ, S. 21, 33 Rustemeyer, R. 46, 48, 54, 55 Ryan, E. 89–91, 103 Ryan, J. 332, 355 Ryan, M.-L. 291, 308, 312 S Saarloos, J. 265, 287 Sachs, J. 119, 127 Salthouse, T. 87, 89, 103, 104 Sanford, A. 84–86 Schadewaldt, W. 204, 218, 223 Schaie, K. 87, 103, 104 Scharnberg, M. 199, 200 Schauber, E. 117, 118, 128 Scheele, B. 11, 12, 42, 53–55, 180, 183, 201–205, 208, 216–218, 222–224 Schmidt, S. 3–5, 7, 14, 15, 20, 25, 26, 32, 33, 35, 54, 55, 57, 69, 324, 349, 351, 355, 370, 376, 437, 440, 443–446, 452, 457 Schneider, K. 207, 224 Schön, E. 30, 42, 55

476 Name index

Schönau, W. 187, 188, 190, 196, 197, 198, 200 Schram, D. 1–3, 5, 15, 27, 33, 68, 294, 312, 353 Schreier, M. 6, 8, 10, 27, 29–31, 33, 35, 42, 45, 50, 55, 221 Schwartz, C. 183, 230, 242, 354 Schwoebel, J. 170, 174, 183 Searle, J. 167, 169, 171, 183 Sedikes, C. 241, 242 Selg, H. 207, 212, 224 Semino, E. 148, 161–163 Shen, Y. 147, 163 Shimamura, A. 87, 96, 98, 104 Shirane, H. 258, 260 Shirley, F. 228, 242 Shneiderman, B. 311, 312 Shoemaker, P. 346, 355 Shohat, E. 428, 440 Shore, B. 257, 260 Short, M. 112, 126 Shuman, A. 108, 128 Sibert, E. 49, 56 Sieber, S. 47, 56 Simpson, C. 330, 343, 355 Ska, B. 89, 104 Sklovský, V. 131, 133, 134, 143 Slamecka, N. 104 Smelik, A. 240, 242 Smith, B. 411, 419 Smith, N. 168, 183 Soederberg-Miller, L. 89, 92, 93–95, 104 Speier, H. 330, 355 Sperber, D. 167, 168, 171, 181, 183, 184, 255, 257, 260 Spinner, K. 28, 33 Spoerri, T. 24, 33 Spörk, I. 189, 200 Stanfield, R. 9, 71, 82–84, 86 Steen, G. 1, 3, 10, 11, 15, 16, 52, 56, 57, 68, 69, 117, 128, 134, 143, 145, 146, 148, 149, 150, 152, 153, 157, 159, 160–163, 180, 183, 257, 260, 353, 411, 419

Stein, N. 364, 377 Steinmetz, H. 26, 33 Stine-Morrow, E. 87–89, 92–95, 97, 103, 104 Stolwijk, C. 326, 355 Stone, M. 230, 242 Storey, R. 408, 419 Strauss, C. 342, 356, 365, 376 Striedter, J. 133, 134, 143 Sutherland, D. 327, 330, 342, 356 Swales, J. 117, 128 Swayze, H. 329, 356 Symonds, P. 219, 224 Sywenky, I. 1, 16, 55, 127, 142, 241, 312, 439 T Tan, E. 13, 58, 289, 308, 312, 313 Tart, C. 21, 33 Tauran, R. 230, 242 Taylor, H. 103, 126, 360, 377 Tebbel, J. 328, 356 Ter Meulen, A. 73, 75, 76, 86 Terry, R. 407, 419 Tesch, R. 38, 39, 43–45, 56 Thomas, D. 329, 356 Thomas, S. 210–211, 222 Thompson, B. 336, 356 Thompson, C. 81, 86 Thompson, S. 154, 157, 162 Throsby, D. 336, 338, 339, 356 Thurston, C. 330, 356 Thury, E. 41, 50, 56 Tilly, C. 362, 377 Todorov, T. 400, 401, 405 Tooby, J. 409, 416, 419 Top, T. 342, 347, 356 Tötösy de Zepetnek, S. 1, 16, 55, 127, 142, 241, 439 Towse, R. 336, 337, 339, 343, 356, 357 Trabasso, T. 58, 68, 69, 86 Tun, P. 94, 104 Tunstall, J. 346, 356 Turner, M. 146, 148, 151, 162, 163, 182, 257, 258, 260, 409, 419

Name index 477

Turow, J. 262, 287 Tuynman, P. 130, 143 Tversky, A. 132, 143 U Ulmann, G. 19, 22, 23, 33 Useem, M. 327, 352, 356 Utsumi, A. 171, 181, 184 V Van den Broek, A. 269, 288 Van den Broek, P. 58, 69 Van den Vondel, J. 130, 143 Van der Bolt, L. 289, 312 Van der Linden, M. 89–91, 104 Van der Tas, J. 342, 357 Van der Voorst, S. 357 Van Dijk, N. 1, 15, 331, 343, 354–356 Van Dijk, T. 6, 9, 16, 72, 86, 88, 93, 95, 96, 99, 103, 104, 108, 120, 127, 128, 150, 152, 163, 240, 242 Van Eijck, K. 10, 13, 261, 262, 264, 265–267, 269, 270, 288 Van Peer, W. 11, 12, 26, 33, 59, 68, 69, 120, 121, 123, 128, 182, 185, 199, 200, 407, 410, 411, 412, 419, 422, 423, 441 Van Rees, C. 5, 10, 13, 16, 261, 262, 264, 265, 269, 270, 288, 324, 331, 347, 349, 353, 356, 357, 363, 377, 379, 380, 392, 394 Van Turennout, M. 139, 143 Vendler, Z. 73, 75, 86 Verdaasdonk, D. 14, 331, 341, 351, 357, 379, 380, 382–384, 392, 394 Verkuyl, H. 73, 75, 86 Vermunt, J. 276, 286, 288, 347, 357, 392, 394 Viala, A. 325, 350, 357 Vidich, A. 38, 56 Viehoff, R. 25, 26, 33, 42, 50, 54–56 Vipond, D. 153, 163 Vorderer, P. 21, 25, 31, 32, 53, 58, 67, 68, 69, 210, 219, 222, 224, 439 Vu, H. 167, 184

W Wakefield, J. 23, 33 Walker, N. 165, 184 Wallach, Y. 165–167, 180, 181, 183, 184 Walton, K. 289, 308, 312 Warren, A. 4, 16, 18, 34, 374, 382 Wassall, G. 336, 338, 357 Wasserman, S. 361, 377 Watt, I. 329, 357 Watt, J. 37, 56 Watzlawick, P. 367, 377 Webster, R. 186, 199, 200 Weibull, L. 262, 267, 281, 288 Weinraub, B. 261, 288 Weir, R. 412, 419 Wellek, R. 4, 16, 18, 34, 374, 382 Wermke, J. 28, 34 Werth, P. 160, 163 White, C. 326, 357 White, D. 328, 357 White, H. 341, 355, 357, 362–364, 367, 368, 371, 375, 377 White, K. 65, 68 Whiteside, T. 328, 357 Whitney, P. 98, 104, 192 Wieler, P. 38, 39, 41, 42, 56 Wilde, O. 428, 429, 431, 441 Wilensky, H. 333, 335, 357 Williams, R. 102, 342, 357 Wilson, D. 167, 168, 171, 181–184 Wimsatt, W. 110, 128 Winner, E. 170, 174, 178, 179, 181, 182, 183, 184 Winston, A. 426, 437, 440, 441 Wippler, R. 96, 104 Wolfe, T. 331, 357 Y Yellowlees-Douglas, L. 294, 302, 313 Yus, F. 168, 171, 184 Z Zabrucky, K. 87, 88, 90, 104 Zeitz, C. 95, 96, 104 Zillmann, D. 220, 221, 223, 224, 439

478 Name index

Zolberg, V. 349, 357 Zucaro, B. 230, 242 Zumkley, H. 207, 208, 223, 224 Zwaan, R. 9, 59, 69, 71–73, 78, 80, 81, 82–86, 92, 98, 102, 103, 105, 108, 109, 128, 150, 151, 163, 312

In the series UTRECHT PUBLICATIONS IN GENERAL AND COMPARATIVE LITERATURE (UPAL) the following titles have been published thus far or are scheduled for publication: 16. D’HAEN, Theo: Text to Reader. A Communicative Approach to Fowles, Barth, Cortazar, and Boon. 1983. 17. de HODENC, Raoul: ’Le Roman des Eles’, and the Anonymous: ‘Ordene de Chevalerie’. Two Early Old French Didactic Poems. Critical Editions with Introduction, Notes, Glossary and Translations, by Keith Busby. 1983. 18. VIJN, J. P.: Carlyle and Jean Paul: Their Spiritual Optics. 1982. 19. FOKKEMA, Douwe W.: Literary History, Modernism, and Postmodernism. (The Harvard University Erasmus Lectures, Spring 1983). 1984. 20. ROOKMAAKER, H. R.: Towards a Romantic Conception of Nature: Coleridge’s Poetry up to 1803. A study in the history of ideas. 1984. 21. FOKKEMA, Douwe and Hans BERTENS (eds): Approaching Postmodernism. Papers presented at a Workshop on Postmodernism, 21-23 September 1984, University of Utrecht. 1986. 22. LEERSSEN, Joseph Theodoor: Mere Irish & Fíor-Ghael. Studies in the idea of Irish nationality, its literary expression and development. 1986. 23. CALINESCU, Matei and Douwe FOKKEMA (eds): Exploring Postmodernism. Selected papers presented at a Workshop on Postmodernism at the XIth International Comparative Literature Congress, Paris, 20-24 August 1985. 1988. 24. D’HAEN, Theo, Rainer GRUEBEL and Helmuth LETHEN (eds): Convention and Innovation in Literature. 1989. 25. BUSBY, Keith and Erik KOOPER (eds): Courtly Literature: Culture and Context. Proceedings of the 5th triennial Congress of the International Courtly Literature Society, Dalfsen, The Netherlands, 9-16 Aug. 1986. 1990. 26. WESSELING, Elisabeth: Writing History as a Prophet. Postmodernist innovations of the historical novel. 1991. 27. THOMPSON, Ewa M. (ed.): The Search for Self-Definition in Russian Literature. Amsterdam, 1991. 28. SCHENKEVELD, Maria A.: Dutch Literature in the Age of Rembrandt. Themes and ideas. 1991. 29. ZWAAN, Rolf A.: Aspects of Literary Comprehension. A cognitive approach. 1993. 30. MOOIJ, J. J. A.: Fictional Realities. The uses of literary imagination. 1993. 31. RIGNEY, Ann and Douwe FOKKEMA (eds): Cultural Participation: Trends since the Middle Ages. 1993. 32. LANGE, Margreet de: The Muzzled Muse. Literature and censorship in South Africa. 1997. 33. FOKKEMA, Douwe and Elrud IBSCH: Knowledge and Commitment. A problem-oriented approach to literary studies. 2000. 34. HAKEMULDER, Jèmeljan: The Moral Laboratory. Experiments examining the effects of reading literature on social perception and moral self-concept. 2000. 35. SCHRAM, Dick and Gerard STEEN (eds.): The Psychology and Sociology of Literature. In honor of Elrud Ibsch. 2001.