[Harke a. Bosma, E. Saskia Kunnen] Identity and Em(Bookos.org)

Identity and EmotionDevelopment through Self-Organization

Recent ideas concerning the development of self and identity have stressedthe importance of moving away from an approach which is mainly con-cerned with outcomes, to one which focuses instead on processes ofdevelopment and, more specifically, on a relational perspective on theseprocesses. Identity and Emotion focuses on the individual development ofidentity and the processes involved. By working from emotions and adynamic systems perspective the book offers a new and exciting approach tohuman identity and its development across the lifespan. The contributors tothe book are specialists in this approach, and offer challenging ideas on thedevelopment of identity as a self-organizing process. The book offers awealth of new ideas and insights, but also concentrates on the ways theseinsights can be translated into research.

Harke A. Bosma is Associate Professor of Developmental Psychology at theUniversity of Groningen where he lectures in adolescent psychology withresponsibility for research and teaching in this area. He has publishedin both Dutch and English and is co-editor of Coping and Self-Concept inAdolescence (1990) with A. E. Jackson and Identity and Development (1994)with T. L. G. Graafsma, H. D. Grotevant, and D. J. de Levita.

E. Saskia Kunnen is Associate Professor of Developmental Psychology atthe University of Utrecht where she researches and lectures in the areas ofpersonal development in adolescence and adulthood, and application ofdynamic systems theory and models in this field. She has published in bothDutch and English including articles in the International Journal of BehaviouralDevelopment and New Ideas in Psychology.

STUDIES IN EMOTION AND SOCIAL INTERACTIONSecond series

Series editors

Keith OatleyUniversity of Toronto

AntonyMansteadUniversity of Amsterdam

This series is jointly published by the Cambridge University Press andthe Editions de la Maison des Sciences de l’Homme, as part of the jointpublishing agreement established in 1977 between the Fondation de laMaison des Sciences de l’Homme and the Syndics of the CambridgeUniversity Press.

Cette publication est publiee co-edition par Cambridge University Presset les Editions de la Maison des Sciences de l’Homme. Elle s’integre dansle programme de co-edition etabli en 1977 par la Fondation de la Maisondes Sciences de l’Homme et les Syndics de Cambridge University Press.

Titles published in the second series

The psychology of facial expression0 521 49667 5 hardback and 0 521 58796 4 paperbackEdited by James A. Russell and Jose Miguel Fernandez-Dois

Emotions, the social bond, and human reality: part/whole analysis0 521 58491 4 hardback and 0 521 58454 7 paperbackThomas J. Scheff

Intersubjective communication and emotion in early ontogeny0 521 62257 3 hardback and 2 7351 07728 hardback (France only)Edited by Stein Braten

Emotion across languages and cultures: diversity and universals0 521 59042 6 hardback and 0 521 59971 7 paperbackAnna Wierzbicka

Communicating emotion: social, moral and cultural processes0 521 55315 6 hardback and 0 521 55741 0 paperbackSally Planalp

The social context of nonverbal behavior0 521 58371 3 hardback and 0 521 58666 6 paperbackEdited by Pierre Philippot, Robert S. Feldman, and Erik J. Coats

Feeling and thinking: the role of affect in social cognition0 521 64223 X hardbackEdited by Joseph P. Forgas

Gender and emotion: social psychological perspectives0 521 63015 0 hardback and 0 521 63986 7 paperbackEdited by Agneta H. Fischer

Causes and consequences of feelings0 521 63325 7 hardback and 0 521 63363 X paperbackLeonard Berkowitz

Emotions and beliefs: how feelings influence thoughts0 521 77138 2 hardback and 0 521 78734 3 paperbackEdited by Nico H. Frijda, Antony S. R. Manstead, and Sacha Bem

For a list of titles in the First Series in Studies in Emotion and Social Interaction,see end of book.

Identity and EmotionDevelopment through Self-Organization

edited by

Harke A. Bosma and E. Saskia Kunnen

Editions de la Maison des Sciences de l’HommeParis

CAMBRIDGE UNIVERSITY PRESSCambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University PressThe Edinburgh Building, Cambridge CB2 2RU, UK

Editions de la Maison des Sciences de l’homme54 Boulevard Raspail, 75270 Paris Cedex 06, France

Published in the United States of America by Cambridge University Press, New York

www.cambridge.orgInformation on this title: www.cambridge.org/9780521661850

© Maison des Sciences de l’Homme and Cambridge University Press 2001

This publication is in copyright. Subject to statutory exceptionand to the provisions of relevant collective licensing agreements,no reproduction of any part may take place withoutthe written permission of Cambridge University Press.

First published 2001This digitally printed first paperback version 2005

A catalogue record for this publication is available from the British Library

Library of Congress Cataloguing in Publication data

Identity and emotion: development through self-organization / edited byHarke A. Bosma & E. Saskia Kunnen. p. cm. – (Studies in emotion and social interaction)Includes bibliographical references and index.ISBN 0 521 66185 41. Identity (Psychology) 2. Personality and emotions. I. Bosma, Harke,1945– II. Kunnen, E. Saskia. III. Series.BF697.I349 2001155.2–dc21 00-046743

ISBN-13 978-0-521-66185-0 hardbackISBN-10 0-521-66185-4 hardback

ISBN-13 978-0-521-02156-2 paperbackISBN-10 0-521-02156-1 paperback

Contents

List of contributors page ixPreface xiii

1 Introduction 1E. Saskia Kunnen, Harke A. Bosma, Cor P. M. Van Halen, andMatty Van der Meulen

2 Developments in self-concept theory and research:affect, context, and variability 10Matty Van der Meulen

Commentary: the self-concept is dead, long live . . .which construct or process? Differentiation andorganization of self-related theories 33Annerieke Oosterwegel

3 The self and emotions 39Nico H. Frijda

Commentary: the self and emotions 58Seymour Epstein

4 Fish, foxes, and talking in the classroom: introducingdynamic systems concepts and approaches 64Paul L. C. Van Geert

Commentary: fish, foxes, identity, and emotion 89Linda A. Camras and George F. Michel

5 A relational perspective on the development of selfand emotion 93Alan Fogel

Commentary: the personal experience of coherence 115Jeroen Jansz

vii

6 Affective processes in a multivoiced self 120Hubert J. M. Hermans and Els Hermans-Jansen

Commentary: affective processes in a multivoicedself in action 141Leni Verhofstadt-Deneve

7 Old–new answers and new–old questions forpersonality and emotion: a matter of complexity 151Jeannette Haviland-Jones, David Boulifard, and Carol Magai

Commentary: emotions as sources of informationabout the self 172Peter G. Heymans

8 Cognitive–emotional self-organization in personalitydevelopment and personal identity 177Marc D. Lewis and Michel Ferrari

Commentary: two faces of identity 199Carol Magai

9 A self-organizational approach to identity andemotions: an overview and implications 202E. Saskia Kunnen, Harke A. Bosma, Cor P. M. Van Halen,and Matty Van der Meulen

References 231Author index 259Subject index 265

viii Contents

Contributors

Harke A. BosmaDepartment of Developmental Psychology, University ofGroningen, Grote Kruisstraat 2/1, 9712 TS Groningen,The Netherlands

David BoulifardDepartment of Psychology, State University of New Jersey,New Brunswick, NJ 08903, USA

Linda A. CamrasDepartment of Psychology, DePaul University,2219 N. Kenmore Ave., Chicago, IL 60614, USA

Seymour EpsteinDepartment of Psychology, University of Massachusetts atAmhurst, Amhurst, MA 01003, USA

Michel FerrariDepartment of Human Development and Applied Psychology,Ontario Institute for Studies in Education, University of Toronto,252 Bloor Street West, Toronto, Ontario M5S 1V6, Canada

Alan FogelDepartment of Psychology, University of Utah, 390 S. 1530 E.,Salt Lake City, UT 84112-0251, USA

Nico H. FrijdaDepartment of Psychonomics, University of Amsterdam,Roeterstraat 15, 1018 WB Amsterdam, The Netherlands

Jeanette Haviland-JonesDepartment of Psychology, State University of New Jersey,New Brunswick, NJ 08903, USA

ix

Hubert J. M. HermansDepartment of Clinical and Personality Psychology, University ofNijmegen, PO Box 9104, 6500 HE Nijmegen, The Netherlands

Els Hermans-JansenCenter for Personality-Research, Bosweg 18, 6571 CD Berg en Dal,The Netherlands

Peter G. HeymansDepartment of Developmental Psychology, University of Utrecht,Heidelberglaan 2, 3584 CS Utrecht, The Netherlands

Jeroen JanszDepartment of Political and Social-Cultural Sciences, University ofAmsterdam, Oude Hoogstraat 24, 1012 CE Amsterdam,The Netherlands

E. Saskia KunnenDepartment of Developmental Psychology, University of Utrecht,Heidelberglaan 2, 3584 CS Utrecht, The Netherlands

Marc D. LewisDepartment of Human Development and Applied Psychology,Ontario Institute for Studies in Education, University of Toronto,252 Bloor Street West, Toronto, Ontario M5S 1V6, Canada

Carol MagaiDepartment of Psychology, Long Island University, Brooklyn,NY 11201, USA

George F. MichelDepartment of Psychology, DePaul University, 2219 N. KenmoreAve., Chicago, IL 60614, USA

Annerieke OosterwegelDepartment of Developmental Psychology, University of Utrecht,PO Box 80.140, Heidelberglaan 2, 3584 CS Utrecht, The Netherlands

Paul L. C. Van GeertDepartment of Psychology, University of GroningenGrote Kruisstraat 2/1, 9712 TS Groningen, The Netherlands

Cor P. M. Van HalenDepartment of Developmental Psychology, University ofGroningen, Grote Kruisstraat 2/1, 9712 TS Groningen,The Netherlands

x Contributors

Matty Van der MeulenDepartment of Psychology, University of GroningenGrote Kruisstraat 2/1, 9712 TS Groningen, The Netherlands

Leni Verhofstadt-DeneveDepartment of Developmental and Personality Psychology,University of Ghent, Henri Dunantlaan 2, B-9000 Ghent, Belgium

xiContributors

Preface

Recent ideas concerning the development of self and identity havestressed the importance of moving away from an approach which ismainly concerned with outcomes, to one which focuses instead onprocessesof development and, more specifically, on a relational perspec-tive on these processes. This change has also led to increased attentionto the role of emotions in the development of self and identity. Thesedevelopments offer new possibilities and challenges for theory andresearch. However, they also lead to new concerns and questions at atheoretical, as well as a methodological level.In 1996, a workshop on the development of self and identity was

organized with the explicit intention of focusing on these new trends.The main topics of the workshop were the conceptualization of thedevelopment of the person as an emotional, relational, and self-organizing process and the way in which such a dynamic con-ceptualization can be translated into empirical research employingmethodological approaches which are adapted to the study of dynamicprocesses in self-stability and change.During the intense and lively workshop discussions, new ideas were

developed, and serious attemptsweremade to clarify and elaborate thedevelopment of self and identity as an inherently emotional processembedded within a relational context. This book can be seen as a nextstep in this discussion. Most of the contributors to this volume wereparticipants in the workshop. Using the workshop discussions as astarting point, they were asked to elaborate their perspective boththeoretically and methodologically. Their ideas and the comments pro-vided by others reflect and extend the nature of the workshop dis-cussions and provide an illustration of the self-organizing, dialogical,and open approach which is advocated in this volume.

In the organization of the originalworkshop and the preparation of thisvolume we have received considerable support from individuals and

xiii

organizations. Here we wish to express our gratitude for their help.The workshopwas supported financially by the Dutch Science Foun-

dation (NWO), the Association for Developmental PsychologicalResearch, and the Departments of Developmental Psychology of theUniversity of Utrecht and of the University of Groningen. The latterdepartment provided additional help by paying for the transcription ofthe audio-taped workshop discussions. This was greatly appreciatedsince the resulting transcripts not only brought the workshop dis-cussions back to mind, but also helped the authors to prepare theirchapters for the book. The Departments of Developmental Psychologyof Utrecht and Groningen Universities also gave financial support forthe preparation of the book itself.A variety of individuals provided us with essential help andwewish

to express our deep thanks to each of them. Nel Wiersma transcribedthe workshop discussions and also helped to abbreviate some of thecontributions. Fiona Buiter corrected the English of the non-nativeauthors. Leen Van Geert provided editorial assistance in the prepara-tion of the manuscript in its final form. Their help was truly indispens-able. We are also very grateful to our publishers, Catherine Max andSarah Caro, for their trust in the project and the series editor TonyManstead for his very careful and supportive review of themanuscript.This book would never have been prepared without the enthusiastic

and creative efforts of the presenters/authors. They all kept on scheduleand conscientiously revised their texts on the basis of the feedback weprovided. We are thankful (and proud) that they were all willing toshare their scholarship in their contributions to the book. Finally, wewant to give our special thanks to Matty Van der Meulen and Cor VanHalen, who helped us organize the workshop, who greatly contributedto the introductory and concluding chapters of the book, and who gaveus the essential comradely support during times when the completionof the book seemed a distant prospect.

Harke A. Bosma and E. Saskia Kunnen

xiv Preface

CHAPTER 1

IntroductionE. Saskia Kunnen, Harke A. Bosma, Cor P. M. Van Halen, andMatty Van der Meulen

Over the years, the topics of self and identity have received a great dealof attention in the field of psychology. The literature is replete withinvestigations into self-concept, people’s perceptions, ideas, and feel-ings about themselves, and into identity, people’s perceptions of theirown sameness and continuity (Oosterwegel and Wicklund, 1995). Al-though researchers in the field choose to focus on different facets of selfand identity, broad theoretical trends can be identified.Traditionally, theorists have conceptualized self and identity as

cognitive structures (Hattie, 1992; Marcia, Waterman, Matteson,Archer, and Orlofsky, 1993). These structures have mostly been re-garded as stable mental representations that – once they have becomecrystallized through the repeated processing of personal information –control our further behavior (e.g. Markus and Wurf, 1987). As a result,the phenomena that are seen as indicative of self and identity areimplicitly reduced to a self-concept: a set of beliefs about oneself. Thisset of beliefs, moreover, is considered to have dynamic implications forthe regulation of our actions. The self-concept is thought to serve as aninterpretative framework that integrates our personal experiences andas a regulative basis to guide further behavior. However, since themental representations that constitute the self-concept are seen as stablecarriers of personal information, deeply engraved in our memory, thetraditional approaches are more suitable for accounting for aspects likestability and continuity than for the dynamics that emerge in self andidentity. For instance, how do aspects of self and identity develop overtime, why do they change with different situations and circumstances,how are they affected by emotional states or emotions as such? Cogni-tivists have tried to solve this dilemma by postulating increasinglycomplex representational structures underlying the self-concept, oper-ating through intra-individual mechanisms of information processing.At this point, however, a purely cognitive approach to self and identityruns into serious conceptual problems. Neither the supposed represen-tational complexity, nor the supposed information processing mechan-isms can be inferred directly from what people tell about themselves.

1

Thus, the self-concept is no longer what people feel, think, or say aboutthemselves, but has become a hypothetical, almost homunculary con-struct with a life of its own. This comes close to a kind of ‘‘agencysmuggling,’’ as Gergen once put it (1984). In fact, a cognitivistic viewturns us into rule-instructed operatorswithout clarifying howwe adaptto the ever-changing contexts of everyday life or even initiate changes(e.g. Bruner, 1990). Concepts like self and identity have thus becomeabstracted from the behavior that we exhibit in relation to the naturalcircumstances in which we have learned to function.Empirical research also suggests that the common conceptualiz-

ations have strong limitations. Given the assumed functional import-ance and personal relevance of self and identity, the empirically foundrelations between cognitive representations about oneself and con-structs such as well-being, adaptation, motivation, and performancebehavior, turn out to be less straightforward than theoreticallyexpected. One reason may be that the prevailing approach is too staticand cognitivistic. If we consider how opinions about oneself may playa role in psychological functioning, it is quite clear that cognitiveopinions alone do not indeed tell the whole story. For example,negative self-evaluations are important only if they concern topicsthat matter to us (Harter, 1999; James, 1890; Tesser, 1988). We couldhappily admit that we are very bad at playing tennis, as long as we donot care about our tennis performance. For someone else, the evalu-ation that he is not as good as the top-ten players in the world may bedevastating.Thus, self-evaluations appear especially important if they are about

our basic concerns, and, thus, are connectedwith emotions. The type ofemotions involved may differ greatly. Take, for example, an ambitiousplayer who is uncertain about his capacities. Baumeister, Smart, andBoden (1996) describe how people with a high but brittle self-esteemmay act aggressively in situations that are perceived as threatening totheir self-esteem. A certain player may present himself as a great playerand become very angry if he loses a game, attributing the loss toanything but himself. Another may feel sad and see himself as a lousyplayer, after losing an important match. But even for the most ambi-tious tennis player, his evaluation of his tennis performance or hisidentity as a tennis player does not matter always and everywhere. Inhis relationwith his intellectualmother, for instance, hemight even be abit embarrassed about his tennis reputation. Whether it matters, de-pends on the situation, has to dowith the specific relationship involved.Thus, opinions about oneself are intertwined with emotions, arise inrelations, and may be more or less stable. They emerge, become rel-evant, motivate, and change in the ever-changing relations. Focusing

2 E. Saskia Kunnen, Harke A. Bosma et al.

solely on non-contextual cognitive statements about oneself may be toonarrow an approach.From the history of thinking about self and identity it becomes

evident that the cognitive approach has not always predominated.Founding fathers of self and identity psychology, such as WilliamJames (1890) and Erik Erikson (1950, 1968), described much broaderconcepts. They perceived self and identity as dynamic phenomena,which include cognitions, but also emotions and perceptions, andwhich are always embedded in the person’s relationship with thecontext. The more restricted, cognitive focus became dominant after-wards, in all probability as a consequence of the demands of formaliz-ation and standardization since the time of behaviorism. Since then,self-related research has focused mainly on self-representations, oftenin written form. In identity research the structuralistic status approachof Marcia (e.g. Marcia, 1980) has dominated for more than thirty years.The recent chapter of Harter in the Handbook of Child Psychology (1998)clearly shows the dilemma. Harter begins by stating that emotions doplay an important role in the self-concept, and that research should payattention to them. However, the current knowledge in this field, asHarter notices, mainly concerns cognitive aspects. That is why heroverview of the research and her further elaboration are cognitive too.But, things appear to be changing. Newer approaches to the study of

self and identity have challenged aspects or implications of this cog-nitive approach and now focus on factors that have been hithertoneglected. Demo (1992), for example, criticized the notion that theself-concept is always stable. Identity researchers such as Bosma,Graafsma,Grotevant, andDe Levita (1994) have begun to conceptualizeidentity as a process of ongoing adaptation and havemade suggestionsfor a relational and interdisciplinary approach to identity development.Recent research efforts (e.g. Baumeister, 1998) started to investigate therole of emotions and situational influences on the development of selfand identity. In a similar way many other authors have begun toaddress the limitations of a merely cognitive approach to self andidentity (see chapter 2 for a comprehensive overview).Although these recent developments have resulted in new insights

and have provided new perspectives, fundamental questions remain.For example, the role of emotions in self and identity requires furtherexplanation. How do emotions and cognitions influence each other?What direction or directions does this influence take? In terms ofcontextual factors, what kinds of events tend to be relevant to anindividual’s sense of self and why do the same events often impacton different individuals in different ways? Furthermore, we lack in-sight into the processes that can account for stability and change in

3Introduction

self-perception and self-evaluation over different time spans. Whatprocesses are responsible for long-term developmental change frominfancy into adulthood? Is fundamental and enduring change possiblein adulthood? If so, what mechanisms are responsible for such change?These fundamental questions remain largely unanswered. Onemajor

impediment for addressing them is the lack of a framework in whichthe role of emotions, the role of the context, and issues of stability andchange in self and identity are considered in combination. The morerecent approaches offer new perspectives, but it is unclear how theyrelate to each other. Moreover, a majority of these approaches are stillfirmly rooted in the conventional view of self and identity as a cogni-tive, internal structure. Emotions and context are primarily regarded ascorrelates of this structure or influences upon its expression. And forthese structures increasingly complicated conceptualizations have beensuggested (compare Higgins, 1987; Markus and Wurf, 1987).Recently, conceptualizations of self and identity as an internal, entity-

like structure have been criticizedmore radically. Several theorists havebegun to tackle self and identity from a completely different perspec-tive. Instead of working with an established construct of ‘‘self’’ or‘‘identity,’’ and trying to relate emotions and context to it, they startfrom the opposite direction: emotions and context are seen as formativeconditions fromwhich self and identity emerge in a self-organizationalprocess. Fogel (1993), for example, considers relationships fundamentalto all development and discusses how the self evolves in a social andemotional context. Hermans (1996) also takes relationships as a startingpoint and has argued for a ‘‘dialogical interchange’’ instead of an‘‘information processing’’ perspective on the self. Lewis’ (1995) theoryof personality development is based on principles of self-organization,and it describes how stable characteristics emerge from feedback andcoupling in cognition–emotion interactions. Haviland (Haviland,Davidson, Ruetsch, Gebelt, and Lancelot, 1994) and Magai (Magai andMcFadden, 1995) turn to emotional processes to explain identity andpersonality development. Most of these authors apply principles ofnon-linear dynamic systems theory, e.g. the notion of self-organization,to the study of developmental processes. Groundbreaking work in thisarea has recently been done by Van Geert (1991, 1994).These new andmore radical perspectives provide promising insights

that may serve to begin to answer many of the questions listed earlier.They also offer intriguing opportunities for the integration into a muchbroader theoretical framework based on self-organizational perspec-tives. The aim of this volume is the discussion and further elaborationof these new ideas. In fact, according to many of the authors in thisvolume, emotions – central in one’s self-experience – emerge, change,


anddissipate in relationships as a dynamic, self-organizing system. Theauthors suggest conceptualizing self and identity as rooted in emotion,emerging in relationships, and developing as a dynamic, self-organiz-ing system.Most of the authors of the chapters in this book share an active

interest in a dynamic perspective on self and identity. In such a per-spective the various elements – cognitions, emotions, and context –need no longer be considered separately, but can be viewed as acomplex, interacting network. Processes within such a network maygive rise to ‘‘self’’ phenomena: conscious and non-conscious, reflectiveand non-reflective self-experiences. If these experiences become stabil-ized in this network, and relevant for the person, they can be seen aspart of the identity. Such a process approach has the advantage that theembeddedness in the context and the changeability are inherentlygiven. This approach has yet to explain stability. Several chapters in thisvolumewill explicitly focus on this topic. Herewewish to warn againsta source of conceptual confusion. ‘‘Self-organization’’ in psychologycan have two completely different meanings: it can refer (a) to thestructure/organization of the self, and (b) to a very general process inwhich higher-order phenomena emerge from interacting lower-ordercomponents (Lewis and Granic, 1999a). The first meaning is mostcommon in psychology, the second stems from Dynamic SystemsTheory and could also be referred to as ‘‘auto-organization.’’ These twomeanings have as much in common as ‘‘wine table’’ and ‘‘table wine’’(Marc Lewis, personal communication, April 2000). Although mostauthors in this volume use ‘‘self-organization’’ in the sense of ‘‘auto-organization,’’ it is sometimes used in the sense of ‘‘the organization ofthe self .’’A conceptualization of self and identity as rooted in emotion, emerg-

ing in relationships, and developing as a dynamic, self-organizingsystemhas implications for themethods thatareused forassessing them.Most current operationalizations of self and identity are based on theideaofa stablecognitivestructure,mostclearlyexemplifiedbyquestion-naires and standardized interviews with questions concerning the per-son’s ideas and opinions about him- or herself (see Byrne, 1996, for anoverview).Moreopenmethods, suchas openessayquestions and ‘‘Whoare you?’’ interviews, are also common. In contrast to the standardizedquestionnaires and interviews, the more open methods allow the sub-jects togivetheirownself-description. Identitymeasuresmostlyconcernsemi-open or fully standardized interviews and questionnaires also(e.g. Marcia et al., 1993). The assumptions behind most of these instru-ments are that the construct to be measured is context-independent,stable, and mainly cognitive. Such instruments can hardly be of use

5Introduction

when self and identity are seen as dynamic, self-organizing systems.Measurementwill thereforebea recurrent issue throughoutthisbook. Insome chapters alternativemeasurementmethodswill be presented thatare more in line with the new conceptualizations. In the concludingchapter we will come back to this topic and discuss suggestions formethodology and methods that are more appropriate.Another recurrent issue concerns the terminology that is used with

regard to self and identity. This terminology has always been a sourceof inconsistency and confusion (e.g. Bosma, 1995). As the reader of thedifferent chapters will note, there is a huge diversity of concepts that allhave to do with how people perceive themselves within their context.All these concepts can be classified in two main groups with verydifferent theoretical backgrounds. The first group consists of a broadcollection of ‘‘self’’ terms: self-concept, self-system, self-schema, theself, the Self, self-evaluation, self-esteem. James, Cooley, and Mead areseen as the founding fathers of this theoretical stream. Different ‘‘self’’terms connote different theoretical backgrounds. The term ‘‘self-concept’’ refers to a cognitive construct, while self-system definitionssuggest broader and more dynamic connotations. The other groupconsists of ‘‘identity’’ terms, like ‘‘ego’’ or ‘‘self ,’’ but themost frequent-ly used term in this group is ‘‘identity’’ itself. Erikson (influenced byFreud as well as by James) is the founding father of this theoreticalstream. In the ways these concepts are used (also in this volume) it isoften not very clear what the differences are between self and identity.Regardless of the different and often completely separate theoreticalbackgrounds, however, there are good reasons to assume that, in fact,both groups of theories and concepts concern more or less the sameempirical phenomena. This is quite evident when the focus is on self-conception problems and identity problems (Van der Werff, 1990; VanHalen, in preparation). For this reason ‘‘self’’ and ‘‘identity’’ will beused interchangeably in this book.The aim of this book is to discuss and elaborate a dynamic systems

perspective of self, identity, and emotion.We think that the applicationof such a perspective can help to overcome certain limitations of thecognitive approach. It can help to integrate emotions and the context inthe study of self and identity and it offers promising possibilities for theconceptualization of stability and change. To achieve this aim, the bookis organized according to a certain ground plan.After this introduction the book continues with a contribution (chap-

ter 2) about three problematic issues in recent self-concept theory andresearch. In this chapter, Van der Meulen will show that the perspec-tives from which the authors in this book approach ‘‘the self’’ are notmerely applications of modern paradigms in psychology to a particular


domain, ‘‘the self.’’ They can (also) be understood as a continuation andextension of recurring arguments within self-concept theory and re-search over the years in different branches of psychology. Van derMeulen discusses how the former conception of a stable cognitiveself-concept has been challenged, and how the role of emotions and ofcontext, and the issue of stability versus variability have become threemajor issues in the theoretical discussion. The overview makes clearthat all these new ideas have resulted in a shattered picture with manyloose ends at this moment. She concludes that till now the emotional,dynamic, and relational aspects of self-concept and identity have beenaddressed in isolation.We see the integration of the three issues in theory and research as

one of the major challenges in the psychology of self and identity. Thediscussion of such an integration is the main aim of this volume.Chapter 2, thus, sets the agenda for the rest of the book. It is followed bychapters 3 and 4 focusing on emotions and dynamic systems theory in ageneral way, and by chapters 5, 6, 7, and 8 offering theory and research,in which the issues are integrated in a particular way.In chapter 3, Frijda discusses emotions. Emotions, according to

Frijda’s componential emotion theory (1986), are always about some-thing: they emerge in the person’s relationship with the world. Inaddition, emotions signal that one’s own person is at stake. Moreover,emotions can be conceptualized as fluid processes rather than struc-tures or entities. Frijda addresses the nature of emotions and discusseshow emotions are related to the self. He argues that emotions do notrequire a representation of self, because they include responses toperceived events in which the self is not explicitly appraised. However,from an early age emotions imply what William James called a notionof ‘‘I’’ – the center of experience and action.Van Geert (chapter 4) presents a general and non-technical introduc-

tion to some of the major concepts of dynamic systems theory and itsapplication to developmental psychology. Various contributions in thisbook rely heavily on the theory of dynamic systems. Readers unfamiliarwith either dynamic systems theory or its application to developmentalpsychology will probably find the proliferation of technical terms con-fusing rather than revealing. Because dynamic systems theory fits innaturally with the basic concepts and models of developmental psy-chology, it may serve as an adequate theoretical and methodologicaltool in the study of developmental phenomena. VanGeert’smain aim isto demonstrate that these concepts and models, exotic as they mayseem, function as convenient and handy tools for conceptualizing andstudying processes of development and of the development of self andidentity in particular.

7Introduction

Chapter 5 is about the intrinsic relatedness of person and context, inother words, about a relational perspective of self and emotion. Thepremises in Fogel’s chapter are that psychological experience alwaysimplies a relation, and that this experience is always dynamic andchanging. Still, people have a sense of uniqueness and permanencethrough time, because they experience their changing relationshipsaccording to different types of emotions. These emotions provide themwith information about the self. The experience and the type of emo-tional process varies at different time scales. This chapter forms abridge to the following chapters. Its main focus is on social-relationalprocesses, but ideas about emotions, self, and the dynamics of changeare incorporated as well. This chapter, therefore, also offers a recentapproach which integrates the emotional, dynamic, and relational as-pects of self and identity.The following three chapters each describe a specific approach to the

dynamics of self, identity, and emotion and provide illustrations of andempirical support for the model by means of examples and case stu-dies. Hermans and Hermans-Jansen (chapter 6) consider the relation-ship between self – viewed as multivoiced and dialogical – andemotion. They argue that during human development a person’s affec-tive responses are increasingly influenced by the relationship the per-son has with him- or herself. In chapter 7, Haviland-Jones, Boulifard,andMagai discuss the function of emotions in identity. Both emotionalexperiences and identity show discontinuous and non-linear develop-mental changes. The authors assert that the application of dynamicsystems modeling will give a better insight into these processes. Inchapter 8, Lewis and Ferrari address the problem of the continuity ofidentity despite ongoing change in the person and the world. On thebasis of principles of self-organizationwhich can be observed in naturalsystems, they discuss personality and identity self-organization. Cen-tral in this process is the developmental consolidation of recursivelyinteracting cognition–emotion elements.Each of the chapters 2 through 8 will be followed by a commentary,

written by a distinguished scholar with regard to the topic addressed inthe chapter. These scholars were asked, first, to comment on the ideasexpressed in the chapter in the light of the aims of the volume, and,second, to give their own ideas about these aims. As a consequence ofthis twofold request some commentaries have the form of real dis-cussions, while others mainly present perspectives in addition or com-plementary to the topic of the chapter and the book. Some authors havereplied to the commentaries on their chapters. These replies are notincluded in the book, but their main points are discussed in the finalchapter. They are referred to in the text and included in the list of


references. As the combination of chapters and commentaries willshow, the attempts to combine emotion theory, dynamic systems think-ing, and recent developments in theory and research on self and ident-ity are very new and still leave much room for debate. We do not wantto give the impression that with this volume all of the problems in thisdomain have been solved. On the contrary, the critical commentaries ofoutstanding scholars on the various contributions to the book aremeantto broaden the discussion beyond the work and ideas of the authors ofthe different chapters. By asking the commentators to discuss the ideasin the chapter, we hope to get an overview of discussion points, contra-dictory viewpoints, unsolved problems, and unanswered questions.In this sense, the commentators can be seen as a forum of highlyspecialized colleagues. Because the topic is so new, it is important to getan overview of the diversity of the many viewpoints, before boilingdown our thinking about the topic into a comprehensive body oftheoretical statements.In chapter 9, the final chapter of the book, we will discuss how

emotion, dynamic systems thinking, and a relational perspective arerelated to each other in the different theoretical and empirical ap-proaches to the study of self and identity development. The variousauthors all have their own focus and the chapter will elaborate howtheir approaches, in combination, offer a broad and lively picture ofhow self and identity can be seen as rooted in emotion, emerging inrelationships, and developing as a dynamic, self-organizing system.The chapter will also discuss to what extent they complement eachother, and to what extent contradictions exist between the differentviewpoints. The resulting overviewwill present the current knowledgein this new field.

9Introduction

CHAPTER 2

Developments in self-concepttheory and research: affect,context, and variabilityMatty Van der Meulen

Theory and research on constructs such as the self-concept, which havea considerable history in psychology, are inevitably influenced by pre-vailing scientific opinions and developments at a particular point intime (Baumeister, 1987; Logan, 1987). In the sixties and seventies theself-concept was a rather unproblematic construct, predominantlyhandled as a trait: a relatively stable, generalized, cognitive set orsystem of descriptive features, characteristic of a particular individual.The majority of methods still used to investigate the self-concept,mainly variations of self-esteem questionnaires, underline this view(Byrne, 1996; Wylie, 1989).This solid picture has, however, been questioned during the last two

decades from different angles.Markus andWurf’s (1987) proposal for adynamic self-concept, in which situational influences are taken intoaccount, has been very influential in this process. Not surprisingly, thefocusing on situational aspects of the self-concept undermines its sup-posed stability. Furthermore, the strictly cognitivistic interpretation ofthe self-concept has been differentiated (Byrne, 1996; Damon and Hart,1988; Greenwald and Pratkanis, 1984), most explicitly by Epstein(1993a).In this chapter the attention is focused on the foregoing three issues.

The traditional conceptualization of the self-concept will be examinedin the light of recent developments: (a) the cognitive view of the self-concept will be confronted with the role of emotion, (b) the idea of ageneralized construct faces the problem of how to take the context intoaccount, and (c) the assumed stability of the self-concept has to standup against observed situational and temporal variability of the self-concept. These issues were originally highlighted by James (1890),Cooley (1902), and Mead (1934), the nestors of self-concept literature.The renewed interest in self-concept psychology on these issues canfurther be placed within the broader context of current psychologicaltheorizing and empirical research. Emotion, context, and variability aretopics of general interest in developmental, social, and personalitypsychology, as will be indicated below.

10

These three issues are not the only ones in self-concept theory andresearch. Other recurring issues, beyond the scope of this chapter, arefor instance: the self-as-subject versus the self-as-object, the globalself-concept versus hierarchical multidimensional models, real versusideal self-concept, true versus false selves, and the motivational orexecutive function of the ‘‘self’’ (Baumeister, 1998; Byrne, 1996; Harter,1996, 1998; Marsh, Byrne, and Shavelson, 1992).At this point a description of the phenomenonwe are dealing with in

this chapter is called for. To avoid terminological complications andlengthy explanations, I will use the term ‘‘self-concept’’ to refer to ‘‘theset of beliefs a person has about him- or herself’’ (Harre, 1987). Con-trary to common practice, I will not use the terms self-concept andself-esteem interchangeably. The term ‘‘self-esteem’’ is confined to theevaluation of one’s entire person, that is global self-esteem, or of speci-fic aspects or components of oneself. These terms are used here ascollective terms, referring to constructs the characteristics of which willbe dealt with in the rest of this chapter. (For definitions and self-conceptmodels see Bracken, 1996; Brinthaupt and Lipka, 1992; Byrne, 1996;Damon and Hart, 1988; Greenwald and Pratkanis, 1984; Hattie, 1992;Oosterwegel, 1992; and chapter 1 of this volume.)In the following sections the question of affect, context, and variabil-

ity in relation to the self-concept will be discussed separately. In thefinal section these three issues will be considered in combination andthe implications for the conceptualization, measurement, and future ofthe self-concept will be examined.

The place and role of emotions in the self-concept

Although the cognitive foundation of the self-concept has not been amajor subject of debate over the years, it has been present from the start.In the early years of self-psychology James (1890) and also Cooley(1902) included self-feelings as an important part of the ‘‘self.’’ Accord-ing to James, self-feelings are aroused when one’s success is not inaccordance with one’s pretensions. For Cooley self-feelings are theaccompanying emotionswhen individuals imagine how they appear toothers, and how otherswould judge that appearance. Self-feelings ariseout of a desire for appreciation or an attack on one’s ‘‘me.’’ Thirty yearslater, however, Mead (1934) emphasized that ‘‘self’’ has to do withthinking in the first place and is therefore a cognitive construct, ratherthan an emotional phenomenon.Nowadays one finds a renewed interest in emotions in several do-

mains of psychology, e.g. intelligence (Goleman, 1995). This interest isreflected for instance in Eisenberg’s (1998) introduction to the latest

11Developments in self-concept theory and research

edition of the Handbook of Child Psychology. She writes that emotion isone of the emerging themes in social and personality development, andtherefore ‘‘an integral aspect of conceptions of the self’’ (p. 3). This is,however, not yet generally accepted. On the one hand, there is still a fairamount of agreement about the interpretation of the self-concept as acognitive construct (Hattie, 1992; Marsh, Byrne, and Shavelson, 1992;Oosterwegel, 1992). On the other hand, several authors consider emo-tions and affect important factors in relation to the self-concept (e.g.Epstein, 1993a; Greenwald and Pratkanis, 1984). How emotions havebeen given a place and role in the self-concept will be worked outhereafter. The terms emotion and affect include different features ofemotions: emotional states, emotional experiences, or emotional ex-pressions (Lewis, 1993). Furthermore, emotions indicate an individ-ual’s relationship with the environment and a readiness for action, inparticular when matters of importance are at stake (Saarni, Mumme,and Campos, 1998; Fischer and Tangney, 1995).

Affective responses to self-referent stimuli

James and Cooley observed that one’s own person – one’s appearance,behavior, or possessions – is an object of special significance: one’sreflections in the mirror give rise to a peculiar interest, one’s childrenare unquestionably the prettiest and brightest. A number of studiesprovide empirical evidence for these everyday observations.When children have acquired objective self-awareness, from the sec-

ond half of the second year onwards (Lewis and Brooks-Gunn, 1979),they show self-conscious emotional reactions: to their reflections in themirror (Lewis, Sullivan, Stanger, andWeiss, 1989; Schneider-Rosen andCicchetti, 1991), to failure (Kagan, 1981) and success (Bullock andLutkenhaus, 1990), and to overpraise and exposure in interaction situ-ations (Lewis et al., 1989). When in the third year children are able toevaluate their own behavior in relation to rules and standards, theymay experience self-conscious evaluative emotions like pride andshame (Lewis, 1993; Stipek, Recchia, and McClintic, 1992).That one’s own person is not a neutral object is also illustrated by

people’s judgment of attributes which are not consciously or immedi-ately recognized as referring to themselves. People, for example, evalu-ate distorted pictures of themselves, their handwriting, or recordings oftheir own voice less neutrally than comparable information of others(Huntley, 1940; Van der Werff, 1967; Wolff, 1943). Though these phe-nomena are intriguing, one of the problems in these studies is to presentstimuli that are self-referent and at the same time sufficiently unrecog-nizable. This is less of a problem in a study by Hoorens and Nuttin

12 Matty Van der Meulen

(1993), who found that letters contained in one’s own name are likedbetter than other letters.Thus the point is not whether emotions have to do with one’s own

person or one’s self-concept, but rather what precisely their relation-ship to the self-concept is.

Types of relationships between the self-concept and emotions

The relationship between the self-concept and emotions has been givenshape in different ways and degrees in the self-concept literature,running from relations of emotions with early self-awareness to a closeassociation or intertwinement of cognition and emotion within theself-concept.

(a) Self-awareness as a prerequisite for emotional experiencePeople may not perceive their own emotions or may be unaware oftheir emotional states. Emotional experience, that is the meaning indi-viduals attach to their emotional state and expressions, requires aparticular cognitive ability, i.e. self-awareness, the capability of attend-ing to oneself (Lewis, 1993). This capacity usually emerges in thesecond half of the second year (Kagan, 1981; Lewis and Brooks-Gunn,1979) and is inferred from particular self-referential behavior, likepointing to self-reflections in a mirror. Lewis et al. (1989) found thatembarrassment, a self-conscious emotion, occurred more in youngchildren who showed signs of self-recognition. The capacity to reflectupon oneself is necessary to recognize one’s own emotions and also tocommunicate about them (Saarni, 1997).

(b) Emotions as indicators of self-awarenessSelf-awareness is thus considered to be a prerequisite for emotionalexperience. Conversely, emotional expressions have been used asindications for the existence of self-awareness. In young childrenemotional reactions such as ‘‘outcome reactions’’ after completing abuilding task (Bullock and Lutkenhaus, 1990) or signs of distress inreaction to the behavior of a model (Kagan, 1981) have been interpretedas such.

(c) Emotions as a dimension of the self-conceptSome recent multidimensional, hierarchical models of the self-conceptcontain dimensions such as Emotional Self-Concept, Emotional Stabil-ity, Self-Confidence, or Affect (Bracken, 1996; Hattie, 1992; Marsh andHattie, 1996; Marsh et al., 1992). Suchmodels have been developed and


extensively investigated through factor-analyzing self-descriptionquestionnaires. Moreover, classification systems of open, experientialmethods, like the ‘‘Who are you?’’ interview, may have a contentcategory Feelings and Emotions (l’Ecuyer, 1992).

(d) Self-esteem as the affective component of the self-conceptIn the literature self-esteem is usually seen as representing the affectivedimension of the self-concept (Damon and Hart, 1988; Keith andBracken, 1996). In contrast to Emotional Self-Concept as one of thedimensions of the self-concept, self-esteem may concern variousdomains of one’s own person or one’s own person in general. This isoften operationalized through separate subscales in self-esteem ques-tionnaires: e.g. Scholastic Competence, Social Acceptance, AthleticCompetence, Physical Appearance, Behavioral Conduct, and GlobalSelf-Worth in Harter’s Self-Perception Profile for Children (Harter,1985).

(e) Causal relationships between self-concept and emotionsThough not everyone considers emotions an integral part of the self-concept, it is clear that certain evaluations or appraisals of one’s ownperson can trigger emotions. A first example is the kind of emotionalreactions experienced after success or failure, e.g. in young children asdescribed above (Stipek et al., 1992). Also adult self-conscious emotionslike shame, guilt, embarrassment, and pride have recently becomesubjects of investigation (Tangney and Fischer, 1995). Guilt and pridearise, for instance, when one holds oneself responsible for wrong-doings or for socially approved outcomes respectively. When high butfragile self-esteem is threatened (Baumeister, Smart, and Boden, 1996),or in adult narcissism (Rhodewalt andMorf, 1998), emotional outburstsand behaviors may become quite violent and aggressive.A second type of relationship can be found in the literature on

self-esteem, depression, and suicide. Depression and suicide are asso-ciated with chronically low self-esteem (Baumeister, 1993). Thoughglobal self-worth and depressed affect correlate highly, these are twodistinct constructs that influence each other reciprocally. This occursthrough different pathways, in which depressed affect may be blendedwith anger (Harter and Marold, 1989). Harter and Marold suggest thatlow self-worth precedes depressed affect when anger is directed to-wards oneself. Conversely, in cases where depressed affect precedeslow self-worth anger is predominantly directed towards others.A third type of relationship concerns emotional problems that result

from discrepancies between divergent perspectives on one’s own per-


son. In Higgins’ work specific patterns of discrepancies between ‘‘ac-tual,’’ ‘‘ideal,’’ and ‘‘ought’’ self give rise to specific emotional reactionsand also health problems (Higgins, 1987; Higgins, Vookles, andTykocinski, 1992).

(f) Emotions and cognitions intertwined within the self-conceptA further step is to see the cognitive foundation of the self-conceptclosely associated or inextricably intertwined with emotions or affect.Epstein is the most explicit advocate of the association between the

self-concept and affect. This has been worked out in his cognitive–experiential self-theory (CEST; e.g. 1973, 1991a, 1993a). In this theorytwo distinct systems of information processing are assumed, a rationalsystem and an experiential system. The rational system operates ac-cording to analytic, logical rules, mediated by conscious appraisals ofevents, and therefore works more slowly. The experiential systemprocesses on a more holistic, context-specific level, in an implicit way,therefore works more rapidly, and is particularly related to emotionalexperiences. With regard to the ‘‘self’’ as an object, both systems oper-ate simultaneously, but independently. The rational system concernscognitions about one’s own person that can be retrieved and describedby regular self-report methods. The experiential system regards self-referent cognitions derived from emotionally significant experiences,which may not always be explicit, but which nevertheless may have asignificant influence on behavior. Self-referent information obtainedfrom these two distinct systems may be discrepant, e.g. high reportedself-esteem versus low self-esteem behavior. It is only in the experien-tial system that cognition and emotionwithin a person’s self-theory arestrongly associated, in cognitive–affective units or modules.The cognitive–experiential self-theory has acted as an explanatory

framework for several phenomena, for instance the acquisition andmaintaining of negative self-schemata, and depressive realism (Ep-stein, 1992a; Pacini, Muir, and Epstein, 1998).In Greenwald and Pratkanis’ (1984) conception the association be-

tween cognition and affect goes further: both are part of the self-concept. They consider ‘‘self’’ as the object of an attitude, more precise-ly, as an attitudinal schema. Such a schema consists of (a) a cognitivestructure, (b) behavioral tendencies, and (c) affect, the central compo-nent of an attitude. Parallels between the findings in attitude researchand self-concept research support this view.Summing up, several authors agree that emotions are important for

the self-concept, but the divergent approaches above illustrate that sofar theory and research provide no clear-cut standpoint about the na-ture of the relationship. Self-concept and emotions are predominantly


viewed and investigated as two constructs which may be related toeach other, influence each other reciprocally, sometimes correlate high-ly, but which remain distinct. In Epstein’s self-theory emotions areassociated with but not part of the experiential self-system, though theidea of cognitive–affectivemodules suggests a strong interweaving.Anexample of emotions as part of the self-concept is Greenwald andPratkanis’ approach. In their attitudinal schema both cognition andaffect are represented. Another candidate for this is the widespreadview that self-esteem – global self-evaluation or evaluation of specificcomponents – represents the affective dimension of the self-concept.However, this last case is disputable. It is a questionwhether or to whatdegree affect or emotions are involved in self-evaluation (Baumeister,1998). James (1890) was the first to make a distinction between coldintellectual self-estimation and the passionate warmth of self-feeling.Evaluation refers, strictly speaking, to the cognitive component of anemotion, the appraisal (Frijda, 1986; Hattie, 1992; Oosterwegel, 1992),which may arouse emotions but is not emotional itself.Epstein and also Greenwald and Pratkanis base their theoretical

expositions partly on information-processing models and research.This domain may provide further insight into the relationship betweenthe self-concept and emotions.

Processing of self-relevant material

The question here is whether or how affect influences the processing ofinformation about one’s own person. Studies by Kuiper and Rogers(Kuiper and Rogers, 1979; Rogers, Kuiper, and Kirker, 1977) and Mar-kus and co-workers (Markus, 1977; Markus and Sentis, 1982) show thatself-relevant information is selected, processed, and retrieved moreeasily than other information. Processes like that have been called‘‘self-reference effects’’ (Greenwald and Pratkanis, 1984). Such effectscan be attributed to four specific features of self-schemata (Markus andSentis, 1982): compared to other cognitive structures, self-schemata aremore extensive and complex, better connected, more frequently ac-tivated, and more associated with affect. The role of affect is thought tobe themost important of these features, but precisely why and how thisfunctions remains unclear. Epstein’s experiential system, in which af-fect has a prominent role, provides a possible explanation for thespecific processing of self-relevant stimuli. One of his assumptions isthat the experiential system processes information relatively rapidlyand in an uncontrolled way, at a preconscious level of awareness.Epstein stresses that there are different levels of processing. Otherspoint out that different channels of processing should be considered:


propositional/verbal, affective/emotional, behavior/motor (Ashmoreand Ogilvie, 1992).In considering the processing of affect-loaded stimuli, the controver-

sial issue about the primacy of affect or cognition comes up again. Indiscussing multiprocess models for emotional experience, Robinson(1998) concludes that what has primacy depends on the type of infor-mation. For information that is valenced and at the same time urgent,preattentive mechanisms lead to a rapid response, without consciouscontrol, as in fear and anxiety. In controlled processing the role ofvalence is to direct the attention. Self-relevant stimuli are examples ofvalenced stimuli, which may give rise to both types of processing,depending on the urgency of the situation. The problem for empiricalresearch, according to Robinson, is how to disentangle different typesof processing, in particular when they seem to interact so highly.

Basic concerns, values, and needs as instigators

A first observation was that one’s own person is an object of specialsignificance and affective responses. After that several ways and de-grees of the relationship between the self-concept and emotions weredescribed. Research and models of processing information about one’sown person were also briefly mentioned. A last point now is to find anexplanationwhy people react affectively to aspects of their own person,to self-referent stimuli.The explanations vary in the terms that are used, but they all boil

down to the assumption of one or more underlying basic concerns,values or needs, which have the function to protect or enhance self-regard or self-esteem (Allport, 1955; Epstein, 1991a; Frijda, 1986;Greenwald and Pratkanis, 1984; Kaplan, 1986; Sedikides and Strube,1997). Positive emotions are associated with the approximation of suchconcerns, negative emotionswith the threatening or thwarting of them.Such emotions signal a self-regarding attitude, which probably has anadaptive function (Baumeister, 1998; Greenwald and Pratkanis, 1984;Sedikides and Strube, 1997).

Conclusion: the function of emotions for the self-concept

Early self-psychologists such as James and Cooley explicitly stated thatemotions, or self-feelings, were an important part of the ‘‘self.’’ In thefollowing decades, however, emotions have not held a prominent placein mainstream self-psychology theory and research. Admittedly, sev-eral authors write that emotions are of importance and are involved inthe self-concept, but theoretical elaborations of this idea or empirical


research into it have been scarce. Eisenberg (1998) announced thatemotions are considered to be an integral aspect of the conceptions ofself. Though agreeing with this, Harter (1996) admits in a historicaloverview that she still predominantly handles the self-concept as acognitive construct. Despite the unclear status of emotions, people’sreactions towards self-relevant stimuli indicate that one’s ownperson isnot a neutral object, but ‘‘hot’’ information accompanied by affect. Thishas its effects in the attention for and the processing of self-relevantmaterial. How exactly this works needs further explication and investi-gation. The most elaborate ideas on this subject can be found in thework of Epstein (e.g. 1991a, 1993a) and the literature on self-consciousemotions (Tangney and Fischer, 1995).Several functions are assigned to the self-concept: e.g. the processing

of information about oneself, the motivation and guidance of behavior(Harter, 1998; Higgins, 1996), the integration of developmentalprocesses, such as physiological and cognitive functioning (Baltes,Lindenberger, and Staudinger, 1998). Clarifying the place and role ofemotions may possibly specify and extend these functions of the self-concept. Emotions, for instance, may indicate what aspects of one’sownperson andwhat situational elements are important for an individ-ual (Saarni et al., 1998; Fischer and Tangney, 1995). What is desirable inthe discussions around this issue is amore extensive treatment of recentemotion theory and research than has been the case so far in self-concept literature.

The significance of context for the self-concept

The significance of context in the emergence and development of theself-concept has been recognized from the earliest psychological in-volvement with this subject by the classic authors James, Cooley, andMead. Notwithstanding the importance assigned to context, the self-concept was for a long time considered to be a fairly context-freephenomenon, consisting of a set of descriptive features which appliedto one’s own person in general, across situations.This view of the self-concept has been challenged lately under the

influence of a reassessment of the relationship between context anddevelopment in general, among others things.

The nature of context

During the last two decades theory and research on the self-concepthave been directed more to the influence of the situation, environment,or context. Emphasis on the role of context in psychological functioning


is not limited to the self-concept literature. In developmental psychol-ogy, for instance, the environment within which individuals grow upand live, are ‘‘embedded in’’ so to speak, has always been consideredan important factor (seeMagnusson and Stattin, 1998;Moen, Elder, andLuscher, 1995 for overviews of the role of context in psychologicalresearch and developmental psychology).Context is a global construct, which needs further specification. In the

first place, different levels of context can be distinguished. Bronfen-brenner and Crouter (1983) provide a taxonomy containing four hier-archical context systems: microsystems (single settings), mesosystems(systems of microsystems), exosystems (larger social structures), andmacrosystems (society at large).Second, context can be viewed either as a relatively stable variable,

exerting a constant influence, like socio-economic status, or as a chang-ing variable during the course of development. In the latter case,Kindermann and Skinner (1992) distinguish three different models:(1) one-way influences: contexts which themselves change, but whichare not influenced by the subjects under study, e.g. birth of a sibling;(2) reciprocal influences: context and subjects induce change in eachother through feedback mechanisms, but the change in context is notparticularly tailored to the subjects, e.g. mutual socialization in peergroups; and (3) adjusted contextual change: changes in the context areattuned to developmental changes in the subjects, e.g. parental child-rearing goals and strategies.Third, for a clear understanding of the role of the context, the pro-

cesses through which the context influences the behavior under study(and vice versa) need to be explained. What elements of the context arerelevant to a particular person at a particular time? What seems to bemost important is not the objectively observable context, i.e. the actualcontext, but the meaning context has for the individual, i.e. the per-ceived context. In that case, context becomes a psychological variable,part of the characteristics of the individual and less a separatelymeasurable variable (Magnusson and Stattin, 1998; Sansone and Berg,1993). In such interactional subject–context models subjects are nopassive receivers of contextual influence, but active agents participat-ing in their own development (e.g. Brandtstadter, 1998).

Types of relationship between self-concept and context

In the self-concept literature context has been assigned a twofold role:(1) context plays an important part in the emergence and developmentof the self-concept, and (2) context determines or influences the currentself-concept, the self-concept at a particular time and place.


There exists considerable agreement that context, at the level ofinteractionswith the primary carers, is a prerequisite for the emergenceof self-awareness in infants (Lewis and Brooks-Gunn, 1979; Pipp, 1993;Trevarthen, 1993).In the further development of the self-concept, significant others –

parents, peers, teachers, partners, or colleagues – play a role throughtheir feedback in direct interactions or, more indirectly, through socialcomparison (e.g. Demo and Savin-Williams, 1992; Feiring and Taska,1996; Harter, 1996, 1998). In social comparison the social groups towhich persons belong or compare themselves with, the referencegroups, are particularly important (Juhasz, 1992). Over the life spanso-called normative contextual transitions, such as entry into school,the transition from school to work, military service, marriage, or a firstchild (Gecas and Mortimer, 1987; Leifer, 1980; Mummendey, 1988),may affect individuals’ self-concept. In addition to that, stressful lifeevents or otherwise important turning points may exert their influence(Clausen, 1995; Prout and Prout, 1996).Throughout these contextual influences – in infant–carer interac-

tions, socialization practices, peer interactions, normative life transi-tions, or stressful life events – culture resonates. Culture may beclassified as a macrosystem (Bronfenbrenner and Crouter, 1983), but itmanifests itself, gets its meaning, in concrete practices and tools, and ininteractions between people (Goodnow, 1996), especially in the rela-tionships between infants and carers (Fogel, 1993).Emerging through interactions, the self-concept incorporates the

meaning a cultural community attaches to it (Schweder, Goodnow,Hatano, LeVine, Markus, and Miller, 1998). Cultures may differ in theway the self-concept is perceived, but a preliminary question iswhether the self-concept is a universal phenomenon. Two aspects seemto exist across cultures: a sense of self as physically distinct from othersand a sense of continuity over time (Hart and Edelstein, 1992). Otheraspects may differ considerably between cultures. A salient distinctionis one between cultures in which the self-concept is individualistic orindependent versus cultures in which the self-concept is relational orinterdependent (Markus and Kitayama, 1991; Schweder et al. 1998;Triandis, 1989). This distinction has important consequences for otheraspects of the self-concept, like the perception, processing, and retrievalof information about one’s own person, the nature of self-consciousemotions, and motives of self-preservation or self-enhancement.In social psychology the interest in context concerns the specific set of

elements of the self-concept, which becomes actualized in a particularsituation at a particular moment. Markus and Wurf (1987) have intro-duced the term ‘‘working self-concept’’ to refer to this. This notion of a


working self-concept can explain why persons may have differentself-concepts in different situations. An example of empirical evidencefor this can be found in the work of Ashmore and Ogilvie (1992). Theystudied variations in self-concept, in what they call self-with-otherrepresentations. With a specific clustering method (HICLAS; see DeBoeck and Rosenberg, 1988), clusters of self-ratings on a set of featuresin combination with certain target persons can be obtained.To get away from the idea that one’s self-concept is in constant flux,

as Gergen represents it (1984), Markus and Wurf (1987) and others(Damon and Hart, 1988; Hattie, 1992) distinguish ‘‘core’’ aspects of theself-concept, which are relatively stable over situations, and more pe-ripheral aspects, which are tied to specific contexts. To the core aspectscould belong factual information, such as ascribed characteristics, ma-jor roles and memberships, and important or salient self-conceptions(Markus and Nurius, 1986). Authors such as Fogel (1993) and Hermansand Kempen (1993) also search for some kind of unity or synthesis. ForFogel ‘‘self ’’ is something that is constantly being created in dialogue.He tries to find a kind of unity or cohesion, not in core aspects, but incommon informational themes that arise across separate dialogicalrelationships.Hermans andKempen assign to the ‘‘self’’ the function ofsynthesizing the self as a whole, to counterbalance for parts of the selfwhich become dominant in particular situations. The term workingself-concept has found its equivalence applied to the context in ‘‘work-ing context.’’ Sansone and Berg (1993) use this term to make clear thataspects of the context influence a person’s behavior or activity in so faras they have a specific meaning for that particular person. In Sansoneand Berg’s view there exists a reciprocal influencing: elements of thecontextwhich are relevant to the individual define theworking context,but the working context in turn defines which elements become partof the working self-concept. Together they form what Sansone andBerg call ‘‘the activated life space.’’ Which elements of the context arerelevant at a certain moment may be inferred sometimes from a per-son’s emotional reactions (cf. Epstein, 1991a; Fischer and Tangney,1995; Robinson, 1998; Saarni, 1997).The notion of a working self-concept has consequences for the level

of self-esteem. The working self-concept consists of a variable, context-specific set of elements, each of which can be valued differently by aperson. Therefore the evaluation of such sets may change from onesituation to another. Research with the Experience Sampling Method,developed by Csikszentmihalyi, nicely illustrates this. With thismethod subjects carry a beeper and have to fill in self-report question-naires at randomly signaled moments several times a day for someweeks. Using this method, Wells (1992) found that young professional


women’s current on-going self-esteem was higher when they were atwork or engaged in leisure activities than when they were at hometaking care of their children. A similar type of study with this methodwas carried out with adolescents to investigate how their self-esteemwas influenced by physical setting, activities they were engaged inor persons they interacted with (Jaquish and Savin-Williams, 1981;Savin-Williams and Demo, 1983).

Conclusion: the dynamic interaction between the self-concept and context

Interactional or transactionalmodels of individual and context relation-ship (Magnusson and Stattin, 1998; Sameroff and Fiese, 1990) are wide-ly accepted, also for a phenomenon like the self-concept (Stratton,1988). In line with this, the self-concept is often viewed as a dynamicsystem (e.g. Baltes et al., 1998;Hermans andKempen, 1993;Markus andWurf, 1987; Oosterwegel, 1992). Exactly how such a system functions,and what kind of interchanges between individual and context takeplace, can be worked out very differently.This depends, first, on one’s general conceptualization of the self-

concept, or the aspects of the self-concept under study. For the interper-sonal self-concept (Fogel, 1993; Neisser, 1993), the interchange occursvery directly, through mother-and-child attunement. In the workingself-concept the core aspects are more or less stable over situations. Theinterchange between context and self-concept in the first place concernsthemore peripheral aspects. Formultidimensional, hierarchicalmodels(Hattie 1992; Marsh and Hattie, 1996), context exerts the most influenceon concrete, situational self-concept aspects at the bottom level of themodel. Here the system is most flexible or, so to speak, ‘‘dynamic.’’Second, the level of context and whether context itself is considered

to be fixed or changeable, both characterize this interchange. At amicrolevel, the interchange between context and an individual’s self-concept seems to occur directly in child–carer interactions. In the rela-tionship between working self-concept and working context both areopen systems which reciprocally constitute each other. At the macro-level, culture seems to be a variable to be reckoned with. Culture mustnot be viewed here as a kind of abstract variable exerting one-wayinfluence on the self-concept, but as something which is expressed ininteractions at every contextual level and, in turn, is defined by these.Third, the relationship between context and individual is also typi-

fied by what is precisely meant by the term ‘‘dynamic.’’ The term‘‘dynamic’’ may be used in a loose way, meaning malleable or flexibleas opposed to stable or fixed. It refers in particular to processes bringingabout change. ForMarkus andWurf (1987), the self-concept is dynamic


in the sense that different sets of self-representations become active indifferent situations. Moreover, dynamic also points to the agentivefeatures of these self-representations or self-structures. In that casedynamic means the function of self-structures in processing input andoutput of information and behavior. For Oosterwegel (1992), the termdynamic refers to the processes of the organization of the self-conceptor, in her terminology, the self-system. With this she means the way inwhich various aspects of the self-concept, and divergent perspectivesone can take towards it, become differentiated or integrated in order tomaintain a kind of balance.Dynamic is a term that is also used more specifically, in relation to

dynamic systems theory. In order to understand the complex function-ing of the person–environment system in general, Magnusson andStattin (1998) advocate a holistic viewpoint and a dynamic systemsapproach. Hermans andKempen (1993) use dynamic systems theory toclarify the function of the synthesizing ‘‘self.’’ An advantage of thistheory is that dynamic systems are not closed, but open to their envi-ronment. In addition, the principle of self-organization, the iterativefeedback of the system’s output into itself, might explain how syn-thesizing processes between the multiple positions one can take to-wards oneself could work.The relationship between the self-concept and context has been given

shape very differently in the course of time. Instead of two separatevariables that somehow correlate, this relationship is nowadays viewedas a dynamic interaction between two interdependent systems. Thisnevertheless leaves several questions to be answered. What aspects ofthe self-concept relate towhat levels of the context, howdo processes ofinteraction precisely work theoretically, and what empirical evidencecan be found for that? Principles of dynamic systems theory may beapplied heuristically, to form ideas about this, or empirically, to testspecific models. It is obvious that suitable methods and measurementsare needed to investigate this.

Stability and variability within the self-concept

Ashas been shown by the previous section, an interactional individual–context view has certain implications for the stability and variability ofthe self-concept. Stability and variability of personality characteristics,the difference between traits and states, is an old question, which is atopical subject again these days (e.g. Baltes et al., 1998). James made adistinction between stable self-feeling, the average overtone indepen-dent of objective reason, and self-feelings as a result of one’s actualsuccess or failure. Several authors agree now that both stability and


variability characterize the self-concept (Brinthaupt and Lipka, 1992;Demo, 1992;Markus andWurf, 1987). Others restrict the use of the termself-concept to enduring, relatively stable dimensions (Marsh et al.,1992). The prevailing viewpoint on the construct thus sets limits to thevery possibility of stability and variability of the self-concept before-hand. In speaking about stability or variability, it is important to makea distinction between the self-concept and self-esteem. A second pointto take into account is the time-perspective. It makes a differencewhether short-term fluctuation or variability is meant or long-termdevelopmental change.

Life-span developmental change of the self-concept

The development of the self-concept has been described with somefrequency for the period up to and including adolescence or for theentire life span (e.g. Baltes et al., 1998; Damon and Hart, 1988; Demo,1992; Harter, 1983, 1998; Hattie, 1992). Inherent to a life-span perspec-tive is a focus on changes which occur over relatively long periods oftime and which concern traitlike characteristics (Nesselroade, 1991).Self-concept studies covering several age periods usually concern thechanging content or dimensions of the self-concept, but also morestructural aspects, like the organization of self-concept dimensions.Recentlymore attention is requested for processes of change. A proces-sual view on change does not consider the self-concept to unfold itselfin isolation, but in person–environment interaction, as discussed in theprevious section.The number and nature of self-concept dimensions change over

the years. Young children’s self-conceptions concern a few concreteobservable dimensions, like abilities, activities, and possessions, whileadolescents’ self-conceptions refer to more abstract dimensions, likepersonality characteristics and attributes related to various roles(Harter, 1998). Within certain age periods the main dimensions areconsidered to remain relatively stable.In the conceptualization of the working self-concept (Markus and

Wurf, 1987) more enduring change concerns the entire set or universeof self-conceptions, which form the basis for the working self-concept.The meaning of existent self-conceptions may change or new elementsmay be added to the set. Within the working self-concept certain‘‘core’’ elements may possess a certain cross-situational and temporalstability.In addition to change in content or dimensions, change in structural

aspects of the self-concept has been investigated. Structuremay refer tocharacteristics of separate dimensions, like abstractness and centrality,


as well as the relationship between dimensions, like consistency anddifferentiation (Bailey, 1970). Studies of structural aspects are scarce.Damon and Hart (1988) studied the level of self-understanding inseveral age groups. Oosterwegel (1992) investigated differentiation anddiscrepancies between several perspectives on the self-concept – ownreal and ideal perspective and that of others – in children of six tosixteen years old. Structural aspects may change in age periods wheredevelopmental transitions occur, but also remain stable for some time.Strauman (1996), for instance, found that the content of self-conceptionsof students changed over the period of three years, but that discrepan-cies between their self-conceptions remained stable.Changes in content and structure of the self-concept are related to

cognitive development and life transitions in the social environment ofthe individual. It is a question which of the two factors exerts moreinfluence on self-concept change. Hattie and Marsh (1996) find thatevents people are confronted with in the course of their lives are moreinfluential than cognitive development. Demo (1992), on the contrary,thinks that developmental changes outdo variability due to situationaldifferences. Baltes et al. (1998) present a model of interacting sources,related to age, history, and non-normative occurrences, providingopportunities and constraints influencing the development of ‘‘selfand personality.’’ Such a model makes clear that the answer to thequestion of which factors influence the development of the self-con-cept requires ingenious research designs and at least a study of com-plex processes.

Short-term situational variability of the self-concept

In the section on context, the variability of the self-concept related todifferent roles and situations has already been discussed. What isvariable in the idea of the working self-concept is the set of self-conceptions that is activated in particular situations. Within this setsome elements recur quite often, i.e. the stable core aspects, others aremore tied to the specific situation or context.Marsh and others (Hattie, 1992; Hattie and Marsh, 1996; Marsh et al.,

1992) in the first place stress the stability of the self-concept. Thehierarchical models they propose, however, leave open the possibilityof variability. This variability occurs at the bottom of the hierarchy,where concrete aspects of the self-concept are placed. Themore abstractdimensions at the top are thought to be relatively stable.L’Ecuyer (1992) also presents a hierarchical model, based on theoreti-

cal considerations and the coding of self-descriptive interviews withsubjects over the entire life span. In thismodel, hierarchy is represented


as increasing abstraction, starting with daily experiences that aredirectly felt, then perceived, and finally conceptualized. The dailyexperiencesmay bemanifold, but in l’Ecuyer’s view all experiences canultimately be classified in five principal dimensions – Material Self,Personal Self, Adaptive Self, Social Self, and Self/Nonself – whichremain the same for all age groups.In the above views variability is one of the characteristics of the

self-concept. Variability concerns peripheral, concrete, directly experi-enced aspects, which are contextually grounded. Core aspects andabstract dimensions refer to stability and continuity over time and thusare the aspects that are usually studied over longer time periods.Variability, however, also offers possibilities to be studied in a life-spanperspective. Certain age trajectories, like adolescence, may show morevariability than others.

Self-esteem: fluctuations along a baseline

Self-esteem has been defined in this chapter as the evaluation of one’sentire person or of specific components of oneself. It is usually meas-ured by self-report questionnaires, several of which now possess satis-factorily theoretical and psychometric qualities (see for overviewsByrne, 1996; Hattie, 1992; Keith and Bracken, 1996; Wylie, 1989). Theseinstruments ask for an opinion about one’s ownperson in general, one’stypical self-esteem, and focus by their very nature on stability over timeand place.Though stability is characteristic of self-esteem investigated with

these types of instruments, slight changes can be found over the lifespan. Overviews of research in this area (Baltes et al., 1998; Damon andHart, 1988; Demo, 1992; Harter, 1998) report a decline and a subsequentrecovery in global self-worth in middle childhood. The decline prob-ably is due to the emergence of a more realistic view of children afterthe tendency in early childhood to overrate themselves: ‘‘I’m the big-gest and strongest boy of my class’’ (Van der Meulen, 1987). In earlyadolescence again a slight decline can be noticed, possibly as a result ofchanges in school settings and other social settings at that age. Fromthen on a steady increase occurs until late adulthood. In the elderlyself-esteem may drop somewhat again.Evidence from several studies in the last few years confirms what

most people experience in their daily life, namely that one’s currentself-esteem is not always the same, but has its ups and downs over theday or week. An explanation for this has been given above. The work-ing self-concept is constituted of a changing array of self-conceptionsthat are evaluated differently and this has its effect on current, global


self-esteem. Current self-esteem is often measured with well-knownself-esteem questionnaires, like Harter’s Self-Perception Profile orRosenberg’s Self-Esteem Scale, with the instruction to answer how oneevaluates oneself right now (Kernis and Waschull, 1995; Wells, 1992).The findings of relative stability over several years and short-time

situational variability have been integrated into a picture of a baselinelevel of self-esteem with moderate fluctuations around it (Demo, 1992;Kernis and Waschull, 1995; Rosenberg, 1986; Wells, 1992). This picture,however, cannot be generalized right away, but is probably true foronly part of the people. Demo and Savin-Williams (1992) found highstability over a period of three years in adolescents for three dimen-sions: self-reported self-esteem, self-esteem reported by peers, andself-feelings. Over a period of a week, about one-half of the adolescentscould be classified as stable, the other half showing moderate to highinstability. This variability may be partly a characteristic of adoles-cence. It is possible that in older age groups, when relations andsituations have become more stabilized, large fluctuations in self-esteem decrease.

The interaction between stability and level of self-esteem

For a long time people with low self-esteem seemed to form the prob-lematic group in self-esteem literature (Baumeister, 1993). There exists alarge body of research about the relationship between psychologicaland clinical problems or membership of a problematic group and levelof self-esteem. The results of this type of research are not unequivocal inthe direction of expected low self-esteem. Why some people’s self-esteem seems not to be influenced by adverse circumstances andothers’ by seemingly trivial occurrences cannot be satisfactorily ex-plained by only considering level of esteem.Kernis and associates found that psychological processes, serving

self-preservation or self-enhancement, appear to be different for peoplewho vary not only in level but also in stability of self-esteem. Level andstability of self-esteem hardly correlate and can be treated as twodistinct dimensions. In this, level of self-esteem refers to people’s rela-tive stable baseline self-esteem and (in)stability to the short-term, con-textual fluctuations around this baseline (Kernis, 1993; Kernis andWaschull, 1995; Kernis, Grannemann, and Barclay, 1989).The attention of Kernis and associates in the first place focuses on

reactions of people with unstable self-esteem. People with high butunstable self-esteem seem to depend on positive evaluations by them-selves and others. Furthermore, they may react defensively and emo-tionally, with anger, hostility, and aggression, when their fragile


self-esteem is threatened (see also Baumeister et al., 1996; Rhodewaltand Morf, 1998). In contrast, people with stable high self-esteem needno validation for this by contingent, external feedback. They possessauthentic or ‘‘true’’ positive self-regard, which is not easily altered bypositive or negative events.The role of instability for low self-esteem is less clear. People with

low self-esteem may incidentally experience positive self-esteem, butprobably lack the skills to refute negative information about them-selves. Low and stable self-esteem seems to be associatedwith a chronicdislike of oneself and a resistance to use self-protective or self-enhance-ment strategies.

Conclusion: stability and variability as dynamically interrelated states

In contemporary viewpoints, both the self-concept and its evaluativeequivalent self-esteem can be characterized by stability as well asvariability.For the self-concept, stability refers to certain main dimensions or to

the entire collection of possible self-conceptions. These are assumed toremain relatively stable during a particular period in life, but may alsoshow changes along the life course. Variability within the self-conceptconcerns concrete, contextual elements or different sets of elements,resulting from changing roles and contexts.These two characteristics, stability and variability, at first sight seem

to exist relatively independently of each other. Yet, the distinction isone of degree rather than of type. What seem to be relatively stable coreelements of one’s self-concept at a certain time and place may beirrelevant in another and vice versa. New experiences and insights intooneself may add new elements to the open set of beliefs about oneself,while others disappear. At a particular turning point this may regroupexisting dimensions and also influence the structure between dimen-sions. Instead of being fixed, the self-concept can be imagined as acontinuously shrinking, expanding, and reorganizing collection of self-conceptions.With regard to self-esteem, stability concerns the more or less stable

baseline of general or typical self-esteem. Overall, this baseline self-esteemmay show slight changes – decreases or increases – along the lifespan. Variability in self-esteem is expressed in fluctuations along thebaseline, the current or barometric self-esteem, due to differences inself-evaluation in different relations, contexts, or through particularpositive or negative events.Though correlations between baseline and current self-esteem are

usuallymoderate to low, fluctuationsmay induce change in a relatively


stable pattern. This may happen through frequently occurring largefluctuations, but originally small changes may also get a process ofchange going. On the other hand, a pattern of more or less stableself-esteem may set limits to variability. People with a stable highself-esteem, for instance, have a ceiling for upward peaks and, similar-ly, people with stable low self-esteem have a bottom for peaks goingdownward.In sum, stability and variability in the self-concept and self-esteem

should be conceptualized rather as dynamically interrelated states ofongoing processes of self-concepting and self-evaluating than as twoindependent states with their own internal laws. What processes exact-ly take place, how stability may evolve out of variability, and at whatpoint variability moves towards stability needs closer inspection.

Conclusions and implications

The traitlike, cognitivistic self-concept of the sixties and seventies hasknown several revisions and reconceptualizations over the years (e.g.Brinthaupt and Lipka, 1992; Damon and Hart, 1988; Epstein, 1973;Gorden and Gergen, 1968; Greenwald and Pratkanis, 1984; Markus andWurf, 1987), but so far this has not resulted in a generally acceptedcircumscription. Regarding the issues discussed in this chapter, thefollowing general conclusions can be drawn. The set of beliefs a personhas about him- or herself (a) is not an exclusively cognitive matter, butsomething that is in several ways closely associatedwith affect, (b) doesnot (solely) concern general, decontextualized beliefs, but beliefs thatemerge in close interaction with the physical, social, and cultural con-text, and (c) consists of bothmore or less enduring, stable beliefs as wellas more short-term, variable ones.Several problems and questions, however, remain inconclusive and

need further investigation. First, the precise place and role of emotionsin the self-concept are still unclear. Are emotions excluded or to beincluded in the self-concept? Which role do emotions play in the pro-cessing of self-relevant information?Which functions do emotions havefor the self-concept? Second, the relationship between context and theself-concept should be worked out more thoroughly. What is preciselymeant by dynamic interaction between context and the self-conceptand how can this be investigated empirically? Third, how exactlyshould the relationship between stability and variability of the self-concept and self-esteembe represented?Howdoes stability emerge outof variability and when does variability tip over to stability?So far affect, context, and variability have been treated separately.

These issues are, however, related to each other in several ways. An


integrated approach could be beneficial in dealing with the abovequestions.

The interplay between affect, context, and variability

After considering stability and variability of the self-concept and self-esteem, these three issues turn out to be interrelated. In the variability ofthe working self-concept and in fluctuations in current self-esteem,context plays a prominent role. What elements of the context becomeimportant at a particular moment depends on the interaction betweencontext and individual in the ‘‘activated life space’’ (Sansone and Berg,1993). One of the functions of emotions is to provide insight into whatsituations are relevant and what is valenced information (Epstein,1991a; Fischer and Tangney, 1995; Robinson, 1998; Saarni et al., 1998). Inparticular, those elements and events come to the fore that potentiallyfacilitate or threaten individuals’ need to preserve or enhance theirself-esteem. A further integration of these three issues in theoryand research may contribute to the development of self-conceptpsychology.

A shift from content and structural aspects to dynamic processes

To understand the interaction of emotion, context, and variability forthe self-concept the focus should be on processes rather than on theoutcomes of these processes. Processes can be characterized as ‘‘acontinuous flow of interrelated, interdependent events’’ (Magnussonand Stattin, 1998: 699). This means that time is a fundamental elementin the study of processes. To grasp the complex interactional processesof emotions, context, and behavior (Saarni et al. 1998) or those of personand environment (Magnusson and Stattin, 1998), the key word is‘‘dynamic,’’ in particular a dynamic systems approach. Several authorsargue that this approach should also be applied to the study of theself-concept (e.g. Baltes et al., 1998; Hattie and Marsh, 1996; Hermansand Kempen, 1993).

A need for adequate measuring methods

In accounting for affect, context, and variability and focusing on pro-cesses rather than content and structure, other measuring methods arerequired than the general self-report questionnaires usually used inself-concept research. Several authors realize this and plead for the useof more idiographic and qualitative methods, like autobiographies andcase studies (Demo and Savin-Williams, 1992; Hattie andMarsh, 1996).


Epstein (1991a) suggests that we should observe situations in whichpersons react emotionally. Robinson (1998) and Izard (1992) suggestthat facial expressions and physiological indicators should supplementself-report instruments. Markus and Kunda (1986) conclude that thevariability of the self-concept should be investigated with methodsthat clarify the entire behavior repertoire that is involved in self-categorization, self-defining etc. Reaction times of behavior in responseto self-relevant stimuli may be one of these methods. The ExperienceSampling Method, devised by Csikszentmihalyi and used in severalstudies (Jaquish and Savin-Williams, 1981; Kernis and Waschull, 1995;Savin-Williams and Demo, 1983; Wells, 1992) to investigate variabilityin level of current self-esteem in different contexts and in relation toaffect, is an example of a fruitful procedure in this respect.

Prospects for the self-concept

Judging from recent overviews (e.g. Baltes et al., 1998; Baumeister, 1998;Harter, 1998) and the abundance of articles on empirical research, theself-concept (or ‘‘self’’ as noun or prefix) is alive and well. Two thingsare important for the prospects of the self-concept: terminological clar-ity and keeping on track with developments within psychology ingeneral.Self-psychology is notorious for its terminological confusion and

diversity. The term ‘‘self-concept’’ has been used throughout this chap-ter as a comparatively general and neutral term to cover most of thefield. New conceptualizations of a construct often lead to the introduc-tion of new terms. Some authors have substituted the term ‘‘self-concept’’ for terms which take into account the hierarchical complexstructure (‘‘self-system’’; e.g. Oosterwegel, 1992) or active, motivationalaspects and flexible adaptation to altering information (‘‘self-schema’’and ‘‘self-theory’’; e.g. Epstein, 1973; Markus and Sentis, 1982). Theterm ‘‘self’’ is also used quite often to include both objective andsubjective aspects of selfhood, product as well as process (Baumeister,1998; Harre, 1998). This may lead to confusion about what exactly ismeant and to reification. Since it appears difficult to arrive at some sortof agreement in this area, the best solution for the present is, first, todescribe carefully the phenomena under study: specific aspects ofpeople’s beliefs about themselves, their current or typical evaluations ofthemselves, their self-preserving or self-enhancing behavior, or theselection and processing of self-relevant information etc. Second, inreferring to such phenomena a relatively neutral term can be applied,such as the ‘‘self-concept’’ or ‘‘self-concepting,’’ or ‘‘self’’ can be used asa prefix, as in ‘‘self-relevant information.’’


The connection of self-concept theory and research with currentissues in psychology at largemay contribute to preserving its role in theunderstanding of human behavior. A focus on the functions of theself-concept, like the processing of information, the motivation andguidance of behavior, or the integration of developmental processes(Baltes et al., 1998; Harter, 1998; Higgins, 1996), could further thatunderstanding. A dynamic systems approach, which incorporates af-fect, context, and variability, may provide heuristic models for this aswell as theoretical models which should be tested empirically.


COMMENTARY

The self-concept is dead, longlive . . . which construct orprocess? Differentiation andorganization of self-relatedtheoriesAnnerieke Oosterwegel

In her chapter Van der Meulen organizes both theory and research onthe ‘‘self-concept’’ around three key themes: cognition versus affect,interactions between self and context, and something termed ‘‘variabil-ity,’’ which can perhaps be characterized as a function of the formertwo plus motivation. This structure offers an excellent framework forthe chapters that follow,where the specifics of these issues are taken upin some detail. At the same time, it appears that both the structure andthe global nature of Van der Meulen’s account illustrate the majortension in contemporary ‘‘self-concept’’ research: important aspects arestudied in isolation, and yet few specifics are known. Given thesethemes, and all the relevant issues which surround them, how do wecombine them into an integrated whole, whilst also seeking to specifythe precise nature of the mechanisms involved? This is the task whichthe current volume invites us to embark upon and for which its authors‘‘set the stage.’’ Van derMeulen provides us with a useful listing of keyissues which prevents us from losing sight of the broader picture as weimmerse ourselves in detailed discussions of specific aspects. Here, Iwish to go one step further and tentatively suggest an initial frame-work, based on the argument that we need (a) clear definitions thatdistinguish between the ‘‘self-as-object’’ and the ‘‘self-as-subject’’; (b)scope for several types of interactions between cognition and affect; and(c) clarity concerning what we understand as constituting ‘‘context’’and where we position its boundaries.

The not-so-omnipotent self-concept

Van der Meulen makes a plea for clearer definitions of terms. Theneed for such clarity cannot be stressed enough. I would like to be evenmore rigorous, and, for the moment, dismiss any pretension of an

33

overarching label such as ‘‘The Self,’’ or Van der Meulen’s ‘‘the self-concept.’’ It would be clearer to refer to each termwith the prefix ‘‘self-’’as a separate construct until we know how they are related.With respect to definition, the term ‘‘Self’’ refers to the person as a

whole, taken from that person’s own perspective. In other words, theSelf refers to the ‘‘I’’ of the individual, the acting entity of whichaffective, biological, neurological, and cognitive processes are part.Seen in these terms, however, studying the Self comes close to studyingpsychology; a statement that may be true, but does not afford muchexplanatory power. The self-concept, on the other hand, is nothingmore than a sample of cognitive self-representations, a cross-section ina process of personal construction, one frame in a movie. Affect andcontexts interact with self-evaluation and self-regulation, but only be-come a part of a self-concept once the individual has reflected on thataffective process or interaction with the context, and has found it to beself-descriptive. Between the self and a self-concept lies the wholerange of constructs we tend to study. Some of them, like self-regulationor self-evaluation, are conceivedmore actively and dynamically, others(fortunately fewer and fewer) are understood more statically. Theseconstructs appear to be inextricably intertwined with context and af-fect, but again (as Frijda, this volume, argues) do not necessarily co-incide. Too often and too early, in my opinion, we try to encompassdifferent aspects and mechanisms into one omnipotent construct. Onemight say that self–other differentiation is the first step in healthyself-research development. It is not only important to be clear aboutwhat we study, but also to be aware of the aspects we do not cover inany given piece of research.Differentiation among self-related theoretical constructs becomes es-

pecially important if one takes a dynamic systems approach to research– as this volume advocates. More specifically, in order to do justice tothis approach a distinction between the self-as-subject and the self-as-object is in order; a distinction between activities themselves and prod-ucts of self-reflection. Dynamic systems theory aims to study processes.This implies that the focus of dynamic systems theory is on the self-as-subject. Ignoring the distinction between the self-as-subject and theself-as-object, between processes and products, would seriously handi-cap a dynamic systems approach because it would confound its objec-tives. And so too would sticking to the old ways of talking in terms ofproducts. In other words, if we intend to study dynamics, we shouldalso think in terms of processes and mechanisms instead of products.Moreover, a proper study of dynamic systems implies that the self-as-subject cannot simply be taken as one ‘‘Self’’ or some sort of homun-culus, but rather should be understood as being differentiated, and

34 Annerieke Oosterwegel

encompassing a number of simultaneous and interactive executiveprocesses and functions.

The multiple roles of affect

Differentiation is also warranted with regard to the relationshipbetween self-related mechanisms such as self-verification and self-enhancement, on one hand, and affect, on the other. If the relationshipbetween these two fields is not clear, this may be because there aremultiple interactions instead of just one causal relationship.Emotional intensity, for instance, functions as an instigator of self-

reflection (e.g. Salovey, 1992), or can be used as an indicator for Me’sversus Not-Me’s. One might, then, infer from Frijda’s theory (e.g. thisvolume) that focusing on the self is the action tendency of intenseglobal arousal. At the same time, this emotional intensity signals whatindividuals consciously or unconsciously consider to be part of theirinterests. Hermans’ Self-Confrontation Method (e.g. Hermans andHermans-Jansen, this volume) uses the latter principle. Positive emo-tions instigate approaching or repeating, whilst negative emotionsmotivate avoidant self-regulating processes (cf. Van der Meulen, thisvolume). Because of these approaching or avoiding tendencies andtheir interpretative framework, specific emotions may – by means ofpositive feedback loops – function as attractors and organizers of self-related events (cf. Haviland-Jones, Boulifard, and Magai, this volume;Lewis and Ferrari, this volume). Moods (to be distinguished fromemotions) are known to color one’s interpretative framework, butemotions also follow from such interpretations.Van der Meulen has already reviewed the vast majority of available

options. The examples described above simply serve to underscore thatthe interactions and their functions can be many – depending on theself-related construct under study. The fact that diverse ones appearthroughout the different chapters in this book suggests that we are stilllearning about each of them and that it may well be too early tointegrate them.

Contexts: systems of relationships

As becomes clear from Van der Meulen’s review, contexts are con-strued in many ways. Personally, I have difficulty in defining theconstruct. One may focus on the dyad at the microlevel (cf. Fogel, thisvolume) and collect a very rich set of data. But the dyad as the unit ofstudy and analysis is comparable to observing one contingency at aneo-Piagetian first stage (e.g. Fischer, 1980). At some point, we will

35Commentary on Van der Meulen

want to study triads, tetrads, and complex social systems, which willthen extend beyond direct interaction. Or we may start from the otherside and study the role of the social environment (past and present)through the perception of the individual, as more or less internalized‘‘voices.’’ I speculate that we will soon find (in so far as we have not yetdone so) that these voices depend on the . . . context. In sum, is it not thecase that context is all that remains as Not-Self? How parsimonious issuch a construct, and, if one takes the self-context interaction seriously,where do we set the boundaries between them?Much research concerned with the self in context would be more

accurately described as the study of (the development of) subjectivetheories about the world and one’s stance in it (e.g. Kelly, 1955). Thesetheories, then, have spin-offs for self-related mechanisms. These spin-offs, in turn, will interact with the context. The result is a theoreticalperpetuum mobile of interactions between the various aspects of theself-as-subject and the self-as-object and various aspects of the context.In sum, with contexts as well, it is important to define the actual unitand focus of study.

Everyone in his or her place

All that I have said so far should not be interpreted as implying that anyof the issues mentioned here are not worthy of study. On the contrary, Ithink that they represent vitally important attempts to understandhuman behavior. Perhaps so important (emotional intensity!) that wecannot be self-critical and precise enough in defining where we standand which (part of which) mechanism or product we are trying toexplain. Such specificity requires an awareness of the general field,where we stand, and, by implication, what we leave out or fail toaddress. A list such as the one offered by Van der Meulen helps andrepresents an important initial conceptualization of the relevant issues,but some sort of coherent framework would be even better. Here, Isuggest the need for a three-dimensional space, created by shootingthree theoretical vectors through the issues in Van derMeulen’s list andthe research presented in the remainder of this volume.Perhaps recent research on the individual’s stance in theworld could

be positioned in a three-dimensional space of self-awareness, contex-tual complexity, and internalization. The self-awareness axis refers tointeractions between cognition and affect in complementary grada-tions. On one side of the self-awareness axis we find experientialbehavior (Epstein, e.g. this volume, commentary on chapter 3). Here,affect seems to predominate. The process may remain in this loop ofsensation (global affect) and appraisal without the self becoming the


focus of attention. That is, the individual may remain in the organisticstate (Rogers, 1951) or flow (Csikszentmihalyi, 1993) without objectiveself-awareness (Duval andWicklund, 1972). The focus of study is solelyon the self-as-subject. The individual may, however, also be drawn outof this state by the intensity of affect, the completion of a task, orphysical (mirrors, cameras) or social triggers (an audience, feedback)into self-awareness (of self-discrepancies) and the inference of emotion.In other words, the flow of action is interrupted and attention shiftsfrom an external orientation to a focus on the self. On this side of theaxis we find self-narrating, commitment (identity), and rational, goal-directed behavior. Here, cognition predominates. In short, the axis ofself-awareness represents recurrent loops of affect and attention withan increasing amount of cognitive self-reflection (cf. Oosterwegel,1996). Again, in such a conceptualization, the question is not whetherself-processes are cognitive (see Van der Meulen, this volume), butwhich particular part of a potential process one studies. The focus ofthis volume is to a large extent on this dimension.The process described above may take place whilst the individual is

alone, in a dyad, triad, or complex social environment (the second axis).Itmay also occur in concrete social interactionwhere the other is largelyextrinsically present or in a situation in which others are increasinglyinternalized into discrepant, conflicting or integrated internal ‘‘voices’’(the third axis). The second axis reflects the number of actors in theactual physical social context. This is a truly social dimension, rangingfrom carer–infant dyads to group processes such as peer interactions.With the exception of work by Fogel and colleagues, this dimension ispoorly represented in research on self-related concepts and mechan-isms. In most research, the dimension mainly appears to provide asetting for individual functioning, and interactions with the contextsand within the context itself are taken to be relatively static.The third axis adds the personal history and experience each individ-

ual brings into the actual context. Albeit not lacking a social component,this dimension is more motivational than social. It refers to the degreeto which experiences are internalized. On one side of the axis we findappraisals, beliefs, and standards (self-as-object), tendencies and reac-tions (self-as-subject) that appear to have their origins in the self. Here,we find drives, intrinsic motivation, and autonomous behavior. Furtheralong this axis we find reflected appraisals and introjected motivation.Here, we find the internal voices of significant others and issues ofself–other differentiation. Towards the end of the axis the input comesfrom outside the individual and motivation becomes extrinsic.Taken together, the three dimensions appear to provide a framework

within which cognitive, affective, social, and motivational aspects of


the self-as-object can be located. These aspects of the functioning of theself-as-subject can be defined, and perhaps in due course everythingcan be understood in relation to each other. It would gowell beyond thescope of this reply, andwell beyondmy potential, to map all theories ofself-related constructs and mechanisms along these axes. I have, so tospeak, merely offered an initial attempt to fill in the corners and edgesof the theoretical jigsaw puzzle.My impression is that much research could be located at particular

points or along one narrow line in this three-dimensional space. Thereader may also find that many of the conceptualizations presented inthis volume tend to move through this space more freely, each assum-ing a slightly different track. And what about the people we try tostudy? They seem to jump around like fleas. A dynamic systems per-spective forces us to be specific about what makes them jump andwhere, and provides a tool for organizing some of this jumping into acomplex integrated process.


COMMENTARY

The self-concept is dead, longlive . . . which construct orprocess? Differentiation andorganization of self-relatedtheoriesAnnerieke Oosterwegel

In her chapter Van der Meulen organizes both theory and research onthe ‘‘self-concept’’ around three key themes: cognition versus affect,interactions between self and context, and something termed ‘‘variabil-ity,’’ which can perhaps be characterized as a function of the formertwo plus motivation. This structure offers an excellent framework forthe chapters that follow,where the specifics of these issues are taken upin some detail. At the same time, it appears that both the structure andthe global nature of Van der Meulen’s account illustrate the majortension in contemporary ‘‘self-concept’’ research: important aspects arestudied in isolation, and yet few specifics are known. Given thesethemes, and all the relevant issues which surround them, how do wecombine them into an integrated whole, whilst also seeking to specifythe precise nature of the mechanisms involved? This is the task whichthe current volume invites us to embark upon and for which its authors‘‘set the stage.’’ Van derMeulen provides us with a useful listing of keyissues which prevents us from losing sight of the broader picture as weimmerse ourselves in detailed discussions of specific aspects. Here, Iwish to go one step further and tentatively suggest an initial frame-work, based on the argument that we need (a) clear definitions thatdistinguish between the ‘‘self-as-object’’ and the ‘‘self-as-subject’’; (b)scope for several types of interactions between cognition and affect; and(c) clarity concerning what we understand as constituting ‘‘context’’and where we position its boundaries.

The not-so-omnipotent self-concept

Van der Meulen makes a plea for clearer definitions of terms. Theneed for such clarity cannot be stressed enough. I would like to be evenmore rigorous, and, for the moment, dismiss any pretension of an

33

overarching label such as ‘‘The Self,’’ or Van der Meulen’s ‘‘the self-concept.’’ It would be clearer to refer to each termwith the prefix ‘‘self-’’as a separate construct until we know how they are related.With respect to definition, the term ‘‘Self’’ refers to the person as a

whole, taken from that person’s own perspective. In other words, theSelf refers to the ‘‘I’’ of the individual, the acting entity of whichaffective, biological, neurological, and cognitive processes are part.Seen in these terms, however, studying the Self comes close to studyingpsychology; a statement that may be true, but does not afford muchexplanatory power. The self-concept, on the other hand, is nothingmore than a sample of cognitive self-representations, a cross-section ina process of personal construction, one frame in a movie. Affect andcontexts interact with self-evaluation and self-regulation, but only be-come a part of a self-concept once the individual has reflected on thataffective process or interaction with the context, and has found it to beself-descriptive. Between the self and a self-concept lies the wholerange of constructs we tend to study. Some of them, like self-regulationor self-evaluation, are conceivedmore actively and dynamically, others(fortunately fewer and fewer) are understood more statically. Theseconstructs appear to be inextricably intertwined with context and af-fect, but again (as Frijda, this volume, argues) do not necessarily co-incide. Too often and too early, in my opinion, we try to encompassdifferent aspects and mechanisms into one omnipotent construct. Onemight say that self–other differentiation is the first step in healthyself-research development. It is not only important to be clear aboutwhat we study, but also to be aware of the aspects we do not cover inany given piece of research.Differentiation among self-related theoretical constructs becomes es-

pecially important if one takes a dynamic systems approach to research– as this volume advocates. More specifically, in order to do justice tothis approach a distinction between the self-as-subject and the self-as-object is in order; a distinction between activities themselves and prod-ucts of self-reflection. Dynamic systems theory aims to study processes.This implies that the focus of dynamic systems theory is on the self-as-subject. Ignoring the distinction between the self-as-subject and theself-as-object, between processes and products, would seriously handi-cap a dynamic systems approach because it would confound its objec-tives. And so too would sticking to the old ways of talking in terms ofproducts. In other words, if we intend to study dynamics, we shouldalso think in terms of processes and mechanisms instead of products.Moreover, a proper study of dynamic systems implies that the self-as-subject cannot simply be taken as one ‘‘Self’’ or some sort of homun-culus, but rather should be understood as being differentiated, and


encompassing a number of simultaneous and interactive executiveprocesses and functions.

The multiple roles of affect

Differentiation is also warranted with regard to the relationshipbetween self-related mechanisms such as self-verification and self-enhancement, on one hand, and affect, on the other. If the relationshipbetween these two fields is not clear, this may be because there aremultiple interactions instead of just one causal relationship.Emotional intensity, for instance, functions as an instigator of self-

reflection (e.g. Salovey, 1992), or can be used as an indicator for Me’sversus Not-Me’s. One might, then, infer from Frijda’s theory (e.g. thisvolume) that focusing on the self is the action tendency of intenseglobal arousal. At the same time, this emotional intensity signals whatindividuals consciously or unconsciously consider to be part of theirinterests. Hermans’ Self-Confrontation Method (e.g. Hermans andHermans-Jansen, this volume) uses the latter principle. Positive emo-tions instigate approaching or repeating, whilst negative emotionsmotivate avoidant self-regulating processes (cf. Van der Meulen, thisvolume). Because of these approaching or avoiding tendencies andtheir interpretative framework, specific emotions may – by means ofpositive feedback loops – function as attractors and organizers of self-related events (cf. Haviland-Jones, Boulifard, and Magai, this volume;Lewis and Ferrari, this volume). Moods (to be distinguished fromemotions) are known to color one’s interpretative framework, butemotions also follow from such interpretations.Van der Meulen has already reviewed the vast majority of available

options. The examples described above simply serve to underscore thatthe interactions and their functions can be many – depending on theself-related construct under study. The fact that diverse ones appearthroughout the different chapters in this book suggests that we are stilllearning about each of them and that it may well be too early tointegrate them.

Contexts: systems of relationships

As becomes clear from Van der Meulen’s review, contexts are con-strued in many ways. Personally, I have difficulty in defining theconstruct. One may focus on the dyad at the microlevel (cf. Fogel, thisvolume) and collect a very rich set of data. But the dyad as the unit ofstudy and analysis is comparable to observing one contingency at aneo-Piagetian first stage (e.g. Fischer, 1980). At some point, we will


want to study triads, tetrads, and complex social systems, which willthen extend beyond direct interaction. Or we may start from the otherside and study the role of the social environment (past and present)through the perception of the individual, as more or less internalized‘‘voices.’’ I speculate that we will soon find (in so far as we have not yetdone so) that these voices depend on the . . . context. In sum, is it not thecase that context is all that remains as Not-Self? How parsimonious issuch a construct, and, if one takes the self-context interaction seriously,where do we set the boundaries between them?Much research concerned with the self in context would be more

accurately described as the study of (the development of) subjectivetheories about the world and one’s stance in it (e.g. Kelly, 1955). Thesetheories, then, have spin-offs for self-related mechanisms. These spin-offs, in turn, will interact with the context. The result is a theoreticalperpetuum mobile of interactions between the various aspects of theself-as-subject and the self-as-object and various aspects of the context.In sum, with contexts as well, it is important to define the actual unitand focus of study.

Everyone in his or her place

All that I have said so far should not be interpreted as implying that anyof the issues mentioned here are not worthy of study. On the contrary, Ithink that they represent vitally important attempts to understandhuman behavior. Perhaps so important (emotional intensity!) that wecannot be self-critical and precise enough in defining where we standand which (part of which) mechanism or product we are trying toexplain. Such specificity requires an awareness of the general field,where we stand, and, by implication, what we leave out or fail toaddress. A list such as the one offered by Van der Meulen helps andrepresents an important initial conceptualization of the relevant issues,but some sort of coherent framework would be even better. Here, Isuggest the need for a three-dimensional space, created by shootingthree theoretical vectors through the issues in Van derMeulen’s list andthe research presented in the remainder of this volume.Perhaps recent research on the individual’s stance in theworld could

be positioned in a three-dimensional space of self-awareness, contex-tual complexity, and internalization. The self-awareness axis refers tointeractions between cognition and affect in complementary grada-tions. On one side of the self-awareness axis we find experientialbehavior (Epstein, e.g. this volume, commentary on chapter 3). Here,affect seems to predominate. The process may remain in this loop ofsensation (global affect) and appraisal without the self becoming the


focus of attention. That is, the individual may remain in the organisticstate (Rogers, 1951) or flow (Csikszentmihalyi, 1993) without objectiveself-awareness (Duval andWicklund, 1972). The focus of study is solelyon the self-as-subject. The individual may, however, also be drawn outof this state by the intensity of affect, the completion of a task, orphysical (mirrors, cameras) or social triggers (an audience, feedback)into self-awareness (of self-discrepancies) and the inference of emotion.In other words, the flow of action is interrupted and attention shiftsfrom an external orientation to a focus on the self. On this side of theaxis we find self-narrating, commitment (identity), and rational, goal-directed behavior. Here, cognition predominates. In short, the axis ofself-awareness represents recurrent loops of affect and attention withan increasing amount of cognitive self-reflection (cf. Oosterwegel,1996). Again, in such a conceptualization, the question is not whetherself-processes are cognitive (see Van der Meulen, this volume), butwhich particular part of a potential process one studies. The focus ofthis volume is to a large extent on this dimension.The process described above may take place whilst the individual is

alone, in a dyad, triad, or complex social environment (the second axis).Itmay also occur in concrete social interactionwhere the other is largelyextrinsically present or in a situation in which others are increasinglyinternalized into discrepant, conflicting or integrated internal ‘‘voices’’(the third axis). The second axis reflects the number of actors in theactual physical social context. This is a truly social dimension, rangingfrom carer–infant dyads to group processes such as peer interactions.With the exception of work by Fogel and colleagues, this dimension ispoorly represented in research on self-related concepts and mechan-isms. In most research, the dimension mainly appears to provide asetting for individual functioning, and interactions with the contextsand within the context itself are taken to be relatively static.The third axis adds the personal history and experience each individ-

ual brings into the actual context. Albeit not lacking a social component,this dimension is more motivational than social. It refers to the degreeto which experiences are internalized. On one side of the axis we findappraisals, beliefs, and standards (self-as-object), tendencies and reac-tions (self-as-subject) that appear to have their origins in the self. Here,we find drives, intrinsic motivation, and autonomous behavior. Furtheralong this axis we find reflected appraisals and introjected motivation.Here, we find the internal voices of significant others and issues ofself–other differentiation. Towards the end of the axis the input comesfrom outside the individual and motivation becomes extrinsic.Taken together, the three dimensions appear to provide a framework

within which cognitive, affective, social, and motivational aspects of


the self-as-object can be located. These aspects of the functioning of theself-as-subject can be defined, and perhaps in due course everythingcan be understood in relation to each other. It would gowell beyond thescope of this reply, andwell beyondmy potential, to map all theories ofself-related constructs and mechanisms along these axes. I have, so tospeak, merely offered an initial attempt to fill in the corners and edgesof the theoretical jigsaw puzzle.My impression is that much research could be located at particular

points or along one narrow line in this three-dimensional space. Thereader may also find that many of the conceptualizations presented inthis volume tend to move through this space more freely, each assum-ing a slightly different track. And what about the people we try tostudy? They seem to jump around like fleas. A dynamic systems per-spective forces us to be specific about what makes them jump andwhere, and provides a tool for organizing some of this jumping into acomplex integrated process.


CHAPTER 3

The self and emotionsNico H. Frijda

What is the role of the self in emotions? Is it essential for emotions?Emotions are subjective; therefore, what place has subjectivity in emo-tions, and what place has the subject’s representation of this subjectiv-ity? It appears to be legimate to question the extent to which a notion of‘‘self’’ is indispensable to understanding emotions.I find it difficult to discuss this issue in a coherent manner. One

reason is my lack of familiarity with discussions of ‘‘the self.’’ Anotheris that the word ‘‘self’’ seems to possess many different meanings thatmay ormay not havemuch to dowith one another.Most certainly, eachindividual is in some sense a ‘‘self,’’ by definition. But does he or shenecessarily ‘‘have’’ a self? For myself, I am not sure. The best way toapproach these issues, therefore, seems for me to briefly review myanalysis of emotions, and to examine whether and where a notion like‘‘self’’ appears needed, and in what function.Such an approach may be useful because by concentrating upon the

emotional phenomena one may perhaps steer clear of certain concep-tual confusions. The literature on emotional development suffers fromthemhere and there. For instance, it sometimesworries about the age atwhich a given emotion, say shame (Lewis, 1992), appears. This givenemotion is thereby identified by a name that itself is undefined. Atwhatage shame appears is, however, a meaningless question when it is notindependently made clear what ‘‘shame’’ refers to, and which of thedefining features appear at what age; they usually will not all maketheir appearance at the same time. Involvement of some notion of theself may be one, but some other role for that notion may be another,independent one. When, for instance, a certain development of self-awareness is considered a necessary defining condition for shame itcannot be an empirical finding that appearance of shame depends upona certain development of self-awareness.The perspective of my analysis will, therefore, be to examine to what

extent a reference to self is needed for describing or understanding thephenomena, and what that reference then is needed for.

39

Emotions and subjectivity

Emotions, admittedly, are subjective. They depend upon the subject’sviewpoints, appraisals, and personal attributions of meaning. They aresubjective also in that each subject may have different emotions inresponse to the same event. The subject (the ‘‘transcendental subject’’)is a logical condition for explaining emotions. Is then some awarenessof the self not a precondition for emotions? I think this conclusion is amisconception. It confuses the standpoints of the experiencer of theemotion with those of an outside observer or explainer. It confuses thephenomenological and the causal-analytical perspectives. When tryingto understand how emotional phenomena come about, an observer or atheoretician observes that two entities are involved, the subject and anobject. He knows that the emotional phenomena depend upon both,and that these often allow more inferences regarding the state of thesubject than the nature of the object. A dog is not by its very naturefearsome or endearing; the subject sees it as such.Phenomenologically, however, this subjectivity need not be present.

To the subject, the object’s emotional nature may be the truth. The dogis fearsome, period, just as the Kosovars simply are evil to the Serbs,and the Serbs to the Kosovars. Emotions are perceptions in the firstplace. Phenomenologically, emotions are out there, just as the red andgreen of an apple are out there. One perceives an object or a scene thatreveals its meaning. And it is not subjective that the object may makeone shrink. Neither is it subjective that a tall tree makes one bend one’sneck backwards.In this perceptual experience the subject is not present. There is an

object and no self. The subject him- or herself may not be part of theexperience. He or she may be – and I will come back to this – but this isnot necessary in order for an emotion to occur. A representation by thesubject of him- or herself may or may not be a condition for thatexperience; we cannot know in advance. But it is not so by necessity.Appearance of an emotional reaction, be it an experience or some overtbehavior or behavioral impulse, does not allow a compelling inferenceregarding the representations that play a role. It is necessary to separateexperiences of ‘‘the self,’’ or those of the subject of him- or herself, fromthe subject as a logical condition for the explanation of emotions. Andthus: what the role of the self is in the constitution of emotions is anempirical matter.

40 Nico H. Frijda

Analysis of emotions

Determining empirically where ‘‘the self’’ plays a role in emotions, andhow, necessitates an analysis of what emotions are, that is, of thephenomena that the term ‘‘emotion’’ refers to, and how these are bestunderstood.‘‘Emotion’’ does not necessarily refer to subjective experience. Emo-

tions, in fact, aremulticomponential (Frijda, 1986; Scherer, 1984); by thiswe mean that the human and animal reactions that are designated‘‘emotions’’ or ‘‘emotional’’ consist of a number of components. Themajor components are appraisal (that is, the way an event is appraised),state of action readiness, physiological response, and one’s feelingsstemming from these components. Several of these components tend tooccur together, and there rarely is only one of them. Emotions can thusbe described in terms other than only those of experience, and emo-tional experience is usually part of a larger whole. More importantly,the reactions that are called ‘‘emotions’’ can be described in functionalterms that apply to each of the components as well as to the reaction as awhole.Such functional terms characterize emotions as involuntary changes

in an individual’s relationship with an object – most often an object inthe environment, often the subject him- or herself, sometimes an objectin thought or a conceptual object such as an ideal. From a functionalpoint of view, emotions (the reactions or states that we call emotions)are interactive states rather than inner, felt states. They operate betweenan individual and his or her environment. They are ways for the personto position him- or herself towards the object and ways of being in-clined to deal with it correspondingly. This is, I think, the most profit-able perspective to take on the phenomena concerned, such as experi-ence, motivation, behavior, and bodily changes. Emotions areinteractive events. They have to do with the relationship between twoentities, a subject and an object. That of course is the case with anypsychological event, even a simple perception, in that it somehowaffirms the independent existence of some entity out there.Whatmakesan emotion different from a perception is that a particular, specificrelationship between subject and object is affirmed or constructed.Emotional behavior and emotional feeling fit this description. The

behavior regarded by bystanders or by the subject as emotional by andlarge represents activities that change the relationship with an object orthe environment as a whole. Emotional behavior can be interpretedas behavior that manifests changes in action readiness (Frijda, 1986).It is relational behavior, implemented to establish, maintain, ordisrupt such a relationship, or it may be behavior that shows that no

41The self and emotions

relationship is being sought after. Manifesting a particular form ofaction readiness is the common denominator of the various behaviorsshownunder particular emotional conditions; states of action readinessare in fact inferences from what those behaviors have in common.Behaviors called ‘‘angry’’ are largely behaviors meant to block theprogress of an antagonist or to hurt or damage an antagonist; behaviorscalled ‘‘joyful’’ are largely exuberant behaviors, that is, interactivebehaviors in excess of what might have been needed under the circum-stances, and so forth.A systematic description can be given of the modes of action readi-

ness that can meaningfully be distinguished in humans and higheranimals. It is indeed likely that they form a limited collection, eachelement of which represents the activation of some distinct neuraldisposition. Modes of action readiness can be divided into modes ofactivation (activation increase and decrease, and inhibition, includingtenseness) and action tendencies. The action tendencies each seek toimplement a particular type of relationship or relationship change.Major modes that are usually distinguished are desire or seeking toobtain or handle, attending, proximity-seeking, exuberance or playing,avoidance, antagonism, rejection, dominance, and submission. Each ofthese modes of action readiness can be identified in self-reports, andtends to appear from factor analyses based upon them (Davitz, 1969;Frijda, Kuipers, and Terschure, 1989) as well as in behavior (e.g. VanHooff, 1972, for chimpanzees). They thus represent data of subjectiveexperience, of motivational states, and of behavior. Each of the distinctmodes of action readiness tends to correspond to one or several of themajor common emotion categories, as illustrated in table 3.1.All modes of action readiness have to do with dealing with the

environment in a particular fashion, or with not dealing with it. ‘‘Deac-tivation’’ means not simply doing nothing, but also denotes lack ofstriving and lack of interest, as in sad apathy. The behaviors suggestingthese states of action readiness tend to be involuntary. The suggestedstates share the characteristic of ‘‘control precedence,’’ that is, they tendto interfere with voluntary behaviors or other ongoing goals, and regu-lation processes appear needed to prevent this.Behaviors resulting from states of action readiness may take a form

determined by cultural habit and individual preference. However, thebehavioral repertoire contains core features that appear to belong to theendowment of the species. These are called expressive behaviors. Thebehaviors are ‘‘expressive’’ in the sense that onlookers can and domakeinferences concerning the states of readiness involved (and thus con-cerning the emotions). However, they are not expressive in the sensethat they aremeant to communicate emotions to others. In part, they are

42 Nico H. Frijda

Table 3.1.Modes of action readiness and corresponding illustrativeemotions

Modes of action readiness Emotion names

desire, seeking desireattending interestproximity seeking affectionexuberance or playing joyavoidance fearantagonism angerrejection disgustdominance pridesubmission shameinterrupting surprisedepending sadness

activationdeactivationinhibition

elements of relational behavior repertoires. The standard surprise ex-pression, for instance, is part of an orienting response, and the standardfear expression (Ekman, 1982) a self-protective one. In part they areindeed communicative signals to the social environment, not so muchto make one’s emotions known, but to influence the behavior of on-lookers so as to favor the prevailing action readiness; for example,crying incites help-giving and exists for that purpose (Fridlund, 1994;Frijda and Tcherkassof, 1997).The point of the preceding analysis was to show that emotions

include more than experiences or cognitive representations of one’semotion. The self, a representation of the person by the person, does notneed to be part of them. Events bring an individual into a state ofself-protection, or of aimless agitation, or of defensive action impulse,or perhaps they only diffusely increase the state of activation, with noparticular object at all. Physiological changes usually accompany this,as the ‘‘logistic support’’ of action, in case action is produced. Maybethis description is adequate only for primitive or elementary emotions,such as reflex-like instances of fear and anger; but these are among theemotions nonetheless.Even emotional experience may on occasion be entirely non-

representational. There exist object-less moods (Morris, 1989). Further-more, a major aspect of emotional experience is experience of affect,that is, of feelings of pleasure and pain. Pleasure and pain are


non-cognitive states that can be felt, and that can also be experiencedwithout, or regardless of, feeling (Zajonc, 1980). They are best under-stood as states of acceptance or non-acceptance of stimuli or of theindividual’s own condition (Frijda, 2000). The individual may justsuffer them, or, in the case of negative affect, be driven to action toescape from them. The latter, for instance, holds good for battery-farmanimals – chickens, pigs, cows – that increase their endorphin-releasevia self-mutilation. Cognitive representations will help to focus orinstigate actions, but the urges to action may be there in their absence.All this is of course a description of a crying newborn baby in which theaction urge avails itself of the innate crying repertoire. Furthermore,there exist experiences of anger, fear, panic, depression, or joy reflectingthe awareness of states of action readiness thatmay have no representa-tional content.

Emotional experience

Changes in action readiness do not usually come about spontaneously,by physiological process (they may do so, in exceptional and notablypathological cases; Izard, 1993). They are instigated by stimuli orevents, as appraised by the subject. Emotions, in the typical case, are theproducts of appraisal. Let us define emotions by including this restric-tion: emotions are response structures centering upon changes in actionreadiness, called forth by events as appraised. ‘‘Events’’ include recol-lections and thoughts, as well as perceived bodily changes.‘‘Appraisal’’ means that some event or thought obtains affective

value, and usually further aspects of meaning, by virtue of certainprocesses in the individual. These processes may be simple. This is thecase when certain sensory stimuli and bodily states impinge uponelementary affective sensitivities like pain sensitivity and those produc-ing bodily discomfort, or pleasure given by sweet taste (Blass and Shah,1995). Almost equally elementary is elementary matching with storedinformation, as with the effects of mere exposure (Zajonc, 1980). Moreoften it is more cognitive, calling upon learned expectancies and cogni-tive schemas (Lazarus, 1991a), and sometimes conscious inferences.Appraisal includes two rather distinct sets of processes. One of these

includes the processes by which a stimulus event obtains affectivevalue; it is sometimes referred to as ‘‘primary appraisal.’’ The secondconsists in picking up or generating information that gives clues for thekind of action readiness that might suit the context at hand; it is called‘‘secondary appraisal’’ (Lazarus, 1991a). Secondary appraisal may beabsent when no response-relevant cues are there. A crying newbornbaby is simply distressed, and its helplessness response is the only

44 Nico H. Frijda

available one. A little later it may become aware of the presence andrelevance of the mother, and modify its emotional urge accordingly.Clearly, the nature of emotions and of the emotion repertoire changeswhen the repertoire of cognitive cues changes or is extended.Different emotions differ in appraisals, and theymay also differ as to

state of action readiness. Appraisals can be described at two levels.They can be described globally as ‘‘themes’’ (Lazarus, 1991a) such as‘‘loss,’’ ‘‘success,’’ ‘‘threat,’’ or ‘‘frustration’’; or more analytically interms of the variables of which the themes are composed, such aspresent or expected, favorable or harmful, increase or decrease, control-lable or uncontrollable, or being due to some other person’s agency orone’s own agency (as these are often prominent in anger and guiltfeeling, respectively). Different proposals for appropriate sets of vari-ables have beenmade (see Scherer, 1999, for a review of current apprai-sal theory).The point of all this in the present context is again that the subject is

not a necessary element of the cognitive domain that constitutes apprai-sal. This is consistent with what was said earlier about the nature ofemotional feelings.Emotional experience or feeling consists in the first place in experi-

encing the emotional event as appraised. It is, as I said, a form ofperceptual experience; it is out there. One sees an object or scene that isdisgusting, fearful, attractive, or desirable. The epithets ‘‘disgusting,’’‘‘fearful’’ or ‘‘desirable’’ translate a perceived quality of what is outthere. They indicate what the object or event is in the process of doing,withholding or offering, or the potential contained in them for furthereffects that might ensue with further contact or longer confrontation.‘‘Disgusting’’ is what is felt to be unbearable upon close contact, topossibly soil indelibly, to contaminate (Rozin and Fallon, 1987; Miller,1997); ‘‘fearful’’ is something or someone that spells danger or spellssomething unknown; and so on.Emotional experience is not in the first place ‘‘in here.’’ It is not felt as

my state; it does not necessarily have the component ‘‘I feel.’’ This, ofcourse, is not always so. We do come to realize that our feelings are inour minds, just as we come to realize that our perceptions are in ourheads, and that the origin of their meaning is in here. Of course, peoplecan, when reflecting, realize that emotional experience is a subjectiveexperience. They know they can have such an experience withoutanybody else having an inkling of it. They can be vividly aware of thefact that they have a certain feeling. But it does not need to be the case,and experiencing an event-as-appraised can exist without it perfectlywell. Emotional experience does not presuppose a sense of self. Itpresupposes a sense of the world.


This applies to all facets of emotional experience. Emotional experi-ence is a combination of experience of the world with a particularsituational meaning structure (appraisal awareness as discussed), ex-perience of affect, experience of a state of action readiness, and experi-ence of arousal. Experience of affect is most often a form of experienceof the world. Affect is primarily experienced not as a subjective state,but rather as an encounter with a repulsive or attractive object. Theattractiveness, aversiveness, acceptability, or unacceptability are all outthere, if not in the lightness and ease or heaviness and difficulty of one’sactions. Action readiness, to the extent that it is experienced and notjust ‘‘had,’’ is experienced as a reachable, to-be-escaped-from, to-be-examined, to-be-approached event in the world or goal to be achieved.In joy the world is felt as an accessible world, in desire the world is feltas a world that has to be accessed. And the experience of one’s body? Isit not one’s own? Well, in some sense it is (in some sense ‘‘own’’ hasexperiential content), and I will come back to this. But primarily, inemotional experience, engagement of the body is experienced as theglow of reality of the emotional event, the fact that its meaning is to betaken seriously, imminent, there (this formulation is taken from Sartre,1939).All this implies that the subject does not necessarily figure in emo-

tional experience. The subject is not necessarily part of the emotionalobject. It – the subject – is of course a condition of potential emotionalexperience, as it is for any other experience. We have to suppose a‘‘transcendental subject,’’ that reflects at Dennett’s (1978) ‘‘intentionallevel of analysis’’ the sum-total of the information-processingprocessesat Dennett’s ‘‘functional’’ level. We do not, however, have to suppose,nor dowe find, a ‘‘self’’ in the sense of a representation of an entitywithproperties on a level with the perceived objects.This analysis, as already indicated, does not apply to all emotions.

Not all emotional experiences conform to it, nor do some emotioncategories as defined or as used in labeling emotion instances. ‘‘Con-ceit’’ implies explicit reference to the self, and so does ‘‘remorse’’ or‘‘guilt feeling.’’ But there is a large range of emotions for which it doesnot apply. These emotions, by consequence, are within reach of bothanimals and small children at an age at which no representation of theself can be assumed to be established. The emotional experience doesnot imply more than the presence of a valenced perception, plus thecalls for action that may or may not be part of the state of awareness.The calls for action presumably require no awareness to push forsimple actions like crying or defensive movements; nor does the activa-tion of what are best considered support mechanisms for coping ac-tions, the changes in autonomic functioning. Phenomenally, one may

46 Nico H. Frijda

be drowned in one’s perceived environment. This is the case for ananimal that, presumably, has no awareness of self, or at leastmost of thetime does not have one. It is the case for a baby. It is the case for adults inrapture or rage, where they ‘‘forget themselves.’’There is no self of which awareness has to have grown before emo-

tions are possible. There need not be a sense of the subjectivity ofemotions for emotions and emotional experiences to be possible, be-cause no subjectivity whatsoever is required in emotional experience.There need not be any awareness of ‘‘I feel sad,’’ andmost emphaticallynot of ‘‘the sadness is mine.’’ No. The sadness is there, in the event, inthe world.

The self as subject

Analytically, of course, there is a transcendental self, the logical refer-ence point for any experience and striving, for any experience of‘‘being-in-the-world.’’ The transcendental self alone ‘‘has’’ no self. Ithas no properties. It is just there as the condition for the possibility ofexperience. It is not usually an object of awareness, but simply a logicalcondition for such awareness.Yet, I think there is something special in awareness and in the

conditions of behavior that would seem an empirical content of thetranscendental self. I think there is something in emotional exper-ience that, from a particular moment onward, allows the use of thefirst-person pronoun in verbal expressions, but that was there inexperience before such use and irrespective of such use. Probably,the awareness has little or no cognitive content. It is a form ofknowledge-by-acquaintance rather than knowledge-by-description.The knowledge-by-acquaintance, it would seem, is built upon experi-ences of motor action, motor preparation, and motor planning, that is,experiences of familiarity and of action initiation or spontaneity.Familiarity. One of the sources for this supposition is the experience of

familiarity of objects, events, and one’s body and actions.Objects, surroundings, and oneself are experienced as familiar or, at

least, as real; I will use ‘‘familiarity’’ and ‘‘reality’’ interchangeablybecause both are meant here in an emotional sense, referring to impres-sions rather than to justifiable knowledge. Usually, even unfamiliarobjects do not entirely lack this characteristic because they do notnecessarily appear ‘‘strange.’’ The ‘‘awareness’’ of familiarity, strangeas it may seem, is not something one is usually aware of. It is a bit likethe awareness of water for a fish. It becomes articulate only after itdisappears, in feelings of strangeness and unreality. Such feelings arecommon during fever, under the influence of drugs or brain damage


(sensory neglect), in response to sudden threatening events such aswhen one’s car skids or upon being notified that a close person has diedor desires to break up one’s relationship. They are also common undertorture and other traumata, and after traumata, where they have beencalled states of numbness or ‘‘denial states’’ (Horowitz, 1992). Onemayroam the ruins of one’s bombed city in a daze, as if none of it is true andas if none of it has happened to oneself. These experiences are thusalmost ubiquitous during and after threatening events. By inference,under normal conditions the opposite feelings of familiarity, reality, orof being in contact with the events are always there, as a background.How the sense of familiarity develops I do not know, nor do I knowat

what age it develops, or what processes are involved. Mere previousexperience is clearly neither sufficient nor necessary. The ubiquitous-ness of feelings of estrangement and their short latency after the onsetof the eliciting event (to judge from descriptions of experiences duringskidding and under torture) suggest that endorphin mechanisms areinvolved. The feelings of estrangement and numbness themselves sug-gest that they stem from a disturbance in the automatic involuntarymotor responses to sensory stimuli which represent their ‘‘knowledgeby acquaintance,’’ such as implicit orienting, readying for grasping,reaching, or using; or perhaps the disturbance is in the awareness ofthose responses. Responses of this nature emerge from the beginning oflife onward, when the coordination of eye movements and perceivedstability of the environment develops. It is well known that the visualworld is perceived as stationary under eye movement as long as move-ments across the retina are compensated for by self-initiated eyemuscleinnervations, that is, by movement intentions. Involvement of incipientor actualmotor actions has of course been emphasized andwas studiedat the time in the so-called ‘‘sensory-tonic field theory’’ by Werner andWapner (1953), which obtained a solid extension in the work on adjust-ment to distorted visual fields (Held, 1965).Spontaneity. There is not only coordination between planned and

executed movements (as evident from the stationary visual field witheye movement), but also between the planning of actions and theirperceived outcomes. As Piaget (1935) argued at the time, the begin-nings of intentional action are seen in the secondary circular reaction,that is the child noticing, for instance, the coincidence of its movementswith the sound of a rattle or the rocking of its crib, and seeking toreproduce the latter. Anyway, the cycle of intent-operate-test-exit is oneof the most fundamental building blocks of human functioning. So, itwould seem, is some awareness of it by the individual. Monitoring thecompletion of those cycles is fundamental because the purposive na-ture of planned movement depends upon it, whether that of crawling

48 Nico H. Frijda

towards a goal or anything more complicated. Planned movement canonly be executed when a movement outcome can be observed to co-incide with what the movement set out for.Self-initiated movement, spontaneity, intentional action, and initiat-

ive: all four are expressions to denote this process structure. It shouldnot surprise us that spontaneity appears to have a special status inexperience, to judge from the emotions connected with its vicissitudes.To begin with, movement interference is one of the earliest frustra-

tions to be observed. Watson regarded it as the unconditional stimulusfor anger. It leads to protest by animals and babies alike. Movementrestraint (rolling a rat in a towel) was the first experimental stressorexamined by Selye (1956). Interruption of voluntary action quite gen-erally generates distress or aggression. This would seem to be a directconsequence of the fact that monitoring task completion is at the heartof voluntary behavior. Indeed, a strong case can be made for thehypothesis that monitoring the effective functioning of an individual’sprocesses is the major function of pleasure and displeasure experience(Frijda, 2000); and executing intentions is one of the very basic functionsof those processes.Non-intentional processes, by contrast but probably by virtue of

these same facts, have a special status in experience. This presumably isbecause they impose themselves upon intended ones. The special statusI mean is that of emotional experience. Ancient philosophy labeledwhat we now call ‘‘emotions’’ as ‘‘passions’’ – that which the soulreceives passively – or ‘‘affections’’ – affects, that which affects the soul– as distinct from the soul’s actions. Emotional experience, as I notedabove, has control precedence. It is felt to overrule ongoing actions, ortake over control of action. The self does not figure in emotional experi-ence, but infringement upon the self by emotional events certainly does.Spontaneity and self-initiation are emotionally valued, from an early

age onward. They are desired and appraised positively. Even before theuse of verbal self-reference, whether in the form of ‘‘I’’ or by the child’suse of its own name, a child makes efforts at independent actions. Itmay violently resent when a helpful parent takes over. I once observeda one-year-oldwhowanted his toy to be put backwhere it was when hebegan to reach for it, but his father picked it up for him, so that his ownreaching could be completed. Evidently, it was the game that countedrather than the toy.My understanding of all this is that, at this stage, there is no self that

obtains satisfaction. Rather, there are planned processes that are recog-nized and monitored as such, actions in the philosophical sense, therealization of which gives satisfaction. It may well be the recognition ofthese very processes that verbal and conceptual development takes


hold of in acquiring the proper name and the personal pronoun, whichthen is used as the object of self-reference. ‘‘I’’ is a way to talk with apronoun about a way of existing or functioning that contrasts with allother observable ways of functioning. The contrast is based upon one’sknowing what is going to happen the next instant and upon havingemotions when the next instant yields what one is prepared for. James(1890) indeed designated ‘‘I as actor’’ as one of the modes of self-awareness. ‘‘I as actor’’ thus exists, not only transcendentally but alsofunctionally, as a process well before it becomes a conceptual entity.It eventually becomes a conceptual entity, of course, but when it

does, this is still in a more or less formal role. ‘‘I’’ as actor has no otherproperties than that of being a causal agent. It refers to the causal agentwho topples the tower, and has a continuous existence beyond thatevent. I can cause harm and take the blame for it, in the sense of beingaware of having caused it. The constellation may give rise to proto-guilt. That is, it may motivate actions directed at atonement and theregaining of love, to my mind long before there is any notion ofresponsibility, of obligation, or of morality.Social interactions. There are more emotional constellations in which

‘‘I’’ is a formal background or condition, rather than an object in thefield of awareness. That is, when the person enters the field as an actualor possible target of actions by others.To be a target of others as such may cause emotions, without any

realization that they are caused by being a target.Maltreatment hurts orleaves the individual in an emotional cold, in a very elementaryfashion. Things change when comparisons with others or with oneselfat other times start to play a role. Counterfactual comparisons have, Ithink, to be assumed in understanding why the good fortune of othersmay cause distress, as it does in envy and jealousy. The fact that I getupset about what another gets must be because I do not get it. Cogni-tively, it must be primitive because even small infants may becomeexcited when other infants receive food or attention; yet a frustration isfelt that is not directly taking place, because nothing is actually beingwithheld. But it is not that primitive, after all, since at least an under-standing is involved that eating or receiving in others can be recognizedas similar to actions of one’s own; there is an elementary empathyinvolved, perhaps related to the possibilities of imitation identified byMeltzoff and Moore (1979).No counterfactual comparisons are necessary to recognize that one

can be a target of others’ actions. The glow of anticipating approval forwhat one did also centers around oneself as an actor. It may even be afirst step towards becoming an object of one’s own emotions. Expectingpraise contingent upon what I do or will do would seem to hover

50 Nico H. Frijda

midway between expecting a pleasurable event with oneself as actorand having a representation of oneself as both the actor and the recipi-ent of the praise.Perhaps negative emotions show this more clearly. There of course

exists distress felt and shown by being rejected from participating in agroup. In a variant of such distress the rejection is attributed to whatone did orwho one is (‘‘We don’t want you! Red hair!’’). Itmay give riseto non-moral shame (Ausubel, 1955), that is, to distress linked to ident-ifying oneself as the cause of the rejection, and leading to withdrawaland desire to hide, or change one’s hair-color. No self-schema (Markus,1977) would seem to be involved, let alone a falling short of someself-ideal; but something akin to oneself-as-actor would appear to playa role. One is the causal agent of rejection, with the shame reaction(distress, submission to the judgment of others as well as submission ina posture with bent head) the appropriate one.Shame is close to being humiliated, which can also emerge, I think,

with an elementary cognitive structure. Humiliation is here defined asbeing pushed to an inferior role in a relationship. It is the structure offrustrating expectations and the construction of a role around them thatcarries the emotion beyond mere distress. Yet it does not necessarilyinclude a conceptualization of one’s role. One may live a life of humili-ationwithout conceptualizing oneself accordingly. Conceptually pride,shame, and humiliation all have to do with the self. They all concep-tually imply counterfactuals, and they may also do so psychologically(Niedenthal, Tangney, and Gavanski, 1994). But psychologically theymay exist without such a cognitive structure and I think, therefore, thatthey do not need the corresponding cognitive development. The con-cepts should not lead the analysis of the emotions themselves astray.

The self as object

There are of course many emotions in which the person him- or herselffigures as an object. There are two forms of this.First, there is consciousness of oneself as an experiencing subject.

Traditionally, emotional feelings have been regarded as subjective ex-periences in their phenomenology as well as their genesis. They are myexperiences, my inner experiences. Introspective reflectiveness makesthem so, and attention to body feelings adds to this. There is in fact adoubling of awareness: there is awareness of somemeaningful object orevent, and then this awareness is the object of reflective awareness.‘‘There is a frightening event’’ changes into ‘‘I feel fear.’’ And there isawareness of the relationship between the two, in that there is an ‘‘I’’who has, feels, or perhaps thinks he feels, the emotion. Adult


emotionality is hardly ever free of reflective self-observation. We al-most always know what we feel and that, in fact, we feel that we feel.Only when self-forgetful, in ecstasy, under extreme suffering, or whenblinded by anger or desire, is this different.Second, however, one can oneself be an object of appraisal and thus

of an emotion: one of one’s actions, thoughts or emotions, or oneself as aperson. Anger towards oneself, self-hatred, pride, shame, and guilt areobvious examples. At least, they are in their more prototypical forms,which are close to the definitions of the concepts and to the legalmeaning of the word ‘‘guilt.’’There are intricacies here. I just mentioned shame and guilt, these

words being used to denote states giving rise to certain responsepatterns. Both words are indeed used in this way in daily conversation.In those usages they likewise fit the eliciting circumstances as appraisedby the subject reasonablywell. ‘‘Shame’’ mostly follows devaluation byothers, which the subject accepts, and ‘‘guilt’’ follows having causedharm to others, even if inadvertently. Neither devaluation nor havingcaused harm need imply any moral overtone. One can be ashamed ofhaving red hair, and feel guilty about having hurt someone who be-haved unpredictably and imprudently (Frijda, 1993b). By consequence,emotions given those names, and involving the action tendencies ofdesiring to hide from view and to atone, respectively, can occur withoutthe cognitive complexities of comparisons with self-ideals or moralstandards.But that does not do away with the fact that the more prototypical

shame and guilt emotions also exist, and that in them a representationof the self and a self-schema are elements. In most situations of shame,as defined by appraised evaluative rejection by others and the actiontendency to withdraw or hide, there also is an image of one’s person asbeing of lesser value. In most situations of guilt feeling, as defined byfeeling responsible for harm to others and an impulse to atone, there isan explicit notion of oneself as a blameful agent. Pride often is not justjoy at achievement or position, but also a high evaluation of one’sperson and a high view of one’s capacities for achievement. And inmany emotions that have the appearance of simple joy, fear, or sadness,the value of oneself as a person or as an actor plays a role in thebackground. Often, indeed, it plays a role in the foreground. Theemotions may include a representation of oneself as valuable or worth-less, leading to actions that communicate that evaluation in socialinteraction (in timidity or conceit), and in expectations to be treatedaccordingly by others or to fill a particular social-status slot.Still, one has to be careful in assessing precisely what that role is. Is

the self-reference or self-evaluation an elicitor of the emotion, or merely

52 Nico H. Frijda

something that gave form to an emotion that would have been thereanyway? Or is it merely something that served social integration in thewake of a more elementary emotional response? These possibilitiesbecame apparent in self-report studies of shame and guilt (reported inFrijda, 1993b). The self-evaluations in shame and guilt, at least some-times, appeared to have come after an emotion had been aroused,rather than being one of its antecedents. Sometimes they appeared togive structure to the diffuse distresses of rejection and empathic dis-tress, respectively. The negative self-evaluation in shame also functionsto bridge the gap between the view of cherished others and one’s ownview (coming to dislike one’s red hair at least removes one difference);taking the blame in guilt feeling at least goes some way in equalizingsuffering (Baumeister, Stillwell, and Heatherton, 1994). To repeat: theconceptual analysis of the emotion words, and even the subject’s self-reports, provide no convincing evidence of what was going on, psycho-logically.

Emotion significance

Not only one’s acts, properties, and person can be the object of emo-tional appraisal. So, too, can one’s emotions.Emotions are appraised; many and perhaps most emotions are. This

applies to individual emotion instances as well as to emotion categories(‘‘anger,’’ ‘‘fear,’’ ‘‘fear of spiders’’). One may feel anger to be a blame-worthy passion, one may dislike having fallen out with someone on agiven day, one may be proud of having stood up for oneself whenscolded, one may despise oneself for being afraid to speak in public.Emotions are appraised because of the implications connected to

having, showing,or feeling them. Iwill refer to knowledge (conscious ornon-conscious) of these implications as the emotion’s ‘‘significance’’(Frijda, 1986). Emotion significance is what motivates emotion regula-tion; one regulates one’s emotions, or one’s emotions are regulated,because of their possible consequences. Emotions may show variousforms of significance. Theirmost immediate consequences are practical.Other peoplemay retaliate to one’s burst of anger; nervousness tends todecrease motor efficiency; examination fear blocks the availability ofknowledge; some desires bring one into trouble. Emotions may alsoconflict with how one wants to be or to appear to others. They may, onthe contrary, also improveone’s image in the eyes of others or of oneself.A careless and carefree attitude in an amorous encounter may bethought to work better and, anyway, one might wish to possess daring.There also exist moral rules and social norms to which one wouldwantto conform and which, if one does not, might generate shame.


The various forms of significance may affect the actual occurrence ofemotion or whether or not one feels it, or these formsmaymerely affectone or some of its manifestations. In the latter case significance tends togo by the name of display rules (Ekman, 1972). Not all manifestations ofsignificance involve representations of the self, as these distinctionsindicate. Some of them are outright and basically emotional; the fearevoked by sexual temptationwhen the consequences are unknown (theseducer may wish to take advantage) is an example. But some formsmost distinctly do involve representations of the self, because theyinvolve who one knows one is, who one wants to be, and how onewants to be and to be seen.

Concerns

From a dynamic, that is, motivational point of view the transcendentalself or, rather, what and how the subject ‘‘is,’’ plays a much moreimportant role in emotions than what individuals think they are or do.Many of the properties that characterize an individual as a particularindividual are decisive in emotions. They underlie emotions whetherone knows it or not; and whether one knows it or not also makes littledifference.The most relevant properties are the individual’s concerns. ‘‘Con-

cerns’’ is the concept used for an individual’s motives, major goals,interests, attachments, values, ideals, sensitivities, and aversions andlikings. It covers the dispositions that give events their emotionalmean-ing. In the analysis of emotions, concerns are pivotal. Emotions resultfrom the interplay of an individual’s concerns and events confrontinghim or her. Emotions result from events appraised as relevant – asharmful or beneficial – to one’s concerns (Frijda, 1986; Oatley, 1992;Stein and Trabasso, 1992). No concern, no emotion.When someone dieswhom one does not care about, the reaction is not very emotional. Griefcomes when someone dies whom one loved.Properties of the concerns to which an event is relevant (that is, is

appraised as relevant) determine properties of the emotion. Thestrength or centrality of the concern determines emotional intensity. Sodoes the number of concerns to which a given event is appraised asrelevant. Both variables have been demonstrated to be correlated torated emotional intensity (Sonnemans and Frijda, 1995). Note that mostevents are relevant to more than one concern; most events havemultiple meanings. Loss of a love partner is emotionally so seriousbecause it represents loss of company, loss of a place for confidence,loss of sex, and loss of familiar social surroundings. The point is ofimportance because so many events have relevance for concerns less

54 Nico H. Frijda

obvious than the central one. An event conducive to jealousy not onlymeans a threat to one’s relationship, but, in addition, to self-esteem andto a desire to be in control. Often those secondary meanings are whatgives the event its sting.An individual’s concerns change and grow in number over a life-

time. Their range increases, and so does the number of objects thatappeal to a given concern (people whom one likes, ideals that onestrives for, the ‘‘surface concerns’’). The set of concerns can be said toconstitute the individual’s ‘‘self’’; the term ‘‘self’’ is indeed often usedin that sense. The word ‘‘self’’ in this sense means something quitedifferent from one’s self-schema or self-conception; the former canexist without the latter. One is this self; one does not have such a self.The fact that the same word is used for both is confusing. Emotionsalways reflect the self in the former sense (except perhaps when theconcerns are only focused at their most impersonal, general, andelementary level). They need not involve ‘‘self’’ in the second sense.Nor do emotions necessarily presuppose the cognitive complexitiesthat are involved in self-awareness, let alone the self-awareness thatgoes beyond the formal one that underlies recognizing another humanindividual as similar. One can of course take stock of one’s concerns,and start to act accordingly. The self as the sum-total of one’s concerns,as one knows them, can become an object of thought, and it can be usedto guide behavior.In line with this, people also have concerns that have one’s own

person or one or more of its properties as their target. Themost obviousexamples are strivings for self-esteem and for social regard, and formaintaining and enhancing identity. Identity striving is one of thoseconcerns that operate at the background of encounters with differentsuperficial meanings, such as the defense of a power position, reachinga given goal, or undergoing humiliating treatment (Frijda, 1994).Still, the role of the self in self-regard in a strict psychological sense is,

however, not easy to assess. Self-enhancing self-presentation in eroticcontexts need not have anything to dowith it. Even themale blue lizardself-presents to females in a manner that is not entirely unlike self-presentation in the humanmale under similar circumstances (Maclean,1990). Social-status strivings are prominent among male baboonsand among chimpanzees, but only the latter show evidence of self-awareness in mirror experiments and self-decoration (De Waal, 1996).Efforts at status achievement and maintenance, whether in lizards,monkeys, or humans maywell be drivenmerely bymomentary desiresfor social-dominance achievements. On the other hand, the distress atbeing dethroned as an alpha male chimpanzee is so poignant thatdamage to self-conception is hard to deny (DeWaal, 1982). The lines are


difficult to draw, and research on children as well as animals wouldappear to be needed.In any case, concerns for one’s personal value and valuation do

emerge as results of development. They are of obvious importance asguiding powers in individual as well as in social life (Carver andScheier, 1990), where they give rise to the whole gamut of the morecomplete forms of the emotions of pride, contempt, shame, and guilt.

Competence

All this is relevant to yet another aspect of the self as person and his orher properties that is an essential factor in shaping emotions, that is,coping competence and trust in one’s coping resources, or self-efficacy(Bandura, 1986).Emotions are a function of three classes of variables: an eliciting event

as appraised, the individual’s concerns that give that event its emo-tional impact, and the individual’s repertoire of action for dealing withthe event. This is one of the most general statements about emotionsthat can be made. Imagine an event such as losing one’s physicalbalance. It threatens one’s concern for absence of pain, aswell as that forknowing what is going to happen next and for being able to deal withthat. Being able to deal with what is going to happen next partlydepends on being able to deal with uncertainty. When the latter abilityis weak, loss of balance is a much more upsetting event. When it isstrong, it may be a suspenseful event, a challenge. You straighten upwith a laugh.When in addition one knowsperfectlywell that it will endinnocently, it may just be felt as a loss of time. This, of course, describesthe development of how infants between three and sevenmonths of agerespond to being lifted, thrown up, and caught again. First they cry orlook anxiously; then they laugh; then they just want to be put downagain (Sroufe and Waters, 1976). This threefold relationship holds forall emotions. Satisfactions that come as amatter of course usually fail toevoke joy or happiness. Satisfactions give positive emotions when theycome after uncertainty about whether they will come or whether onecan indeed avail oneself of them, and their intensity is proportional tothe previous uncertainty (Ortony, Clore, and Collins, 1988).Trust in one’s competence is in principle entirely implicit in expecta-

tions concerning the outcome of an encounter. There need be no ‘‘self-image,’’ no comparison between what one would want to do and whatone thinks one can accomplish. The object one confronts simply looksweak or powerful, dangerous or challenging, within reach or out ofreach, in the projective way described earlier. Self-confidence in thissense is open to animals and human infants as well as to human adults.

56 Nico H. Frijda

Mice that have been victorious in previous antagonistic encounters aremore aggressive than those who have lost a fight. There also are innatedifferences between daring and fearful mice and rats. Assertive behav-ior in chimpanzees goes up and down with position in the socialhierarchy, which itself is dependent upon earlier success in powercontests (De Waal, 1982).

Conclusions

In this chapter I examined the role of the self in emotions. We encoun-tered reference to the person who experiences emotions in differentroles. It appeared useful or necessary to distinguish: (a) the transcen-dental self, as a logical premise of the causal as well as the phenom-enological analysis of emotions; (b) the self as an equivalent of theperson with given properties, notably competences and concerns interms ofwhich events can obtain emotionalmeanings; (c) ‘‘I as actor,’’ afunctional formulation for an individual’s spontaneity and intentionalactivity, and the basis for the emotions that emerge when events inter-fere with that spontaneity; and (d) a notion of ‘‘self’’ experienced asexperiencer and as an object of certain emotions.A transcendental self and oneself as a personwith concerns and other

properties are logical prerequisites of emotions but need not be el-ements of experience. They even need not have informational content,because emotions may be caused by strivings and sensitivities that aresimply there, as the mechanisms that cause goal-directed behavior. ‘‘Ias actor’’ is at first implicit in experience, mainly through the pleasuresof spontaneous action and the pains of interferences. It becomes acontent of awareness only gradually, presumably when language pro-vides a reference in the form of a name or a pronoun. Oneself as aperson, and thus a ‘‘self’’ with properties, would, from the analysis ofemotions, only appear much later, and much later than certain emotionconcepts like shame and pride would seem to suggest. It would behelpful if separate concepts were developed for the various notions thatthe word ‘‘self’’ points to, at some stage.


COMMENTARY

The self and emotionsSeymour Epstein

In order to properly examine the relation between emotions and theself-concept, Frijda believes it is necessary to distinguish between theself as object of knowledge and the transcendental self. He makes acompelling case for the importance of the self as object of knowledge,but I have trouble understanding exactly what he means by the tran-scendental self. At one point he defines it as ‘‘the sum total of theinformationprocesses at a functional level’’ and at another as a person’s‘‘set of concerns.’’ These are both vague constructs that require clarifica-tion, and, unless I am missing something, they are not even the same.Others, as well as I, have proposed more clearly articulated distinc-

tions between two somewhat similar constructs to the ones proposedby Frijda. For example, William James (1910) initially, and GordonAllport (1955) later, expressed the difference between the two selves interms of the self as an object of knowledge and the self as an executiveself that is the source of behavior. Both, however, eventually disownedthe executive self as scientifically indefensible, likening it to a homun-culus residing inside a person’s head and directing the person’s behav-ior. The problem with the executive self and its homunculus analogy isthat understanding their behavior is no simpler than understanding thebehavior of the whole person and therefore does nothing to further ourunderstanding. I will discuss a more meaningful representation of theexecutive self later, when I present my own views.Although I amnot enamoredwith Frijda’s transcendental self, I am in

agreement with almost everything he says about emotions, includingthe important role he assigns to appraisal in the instigation of almost allemotions, the possibility of the occurrence of emotions in the absence ofrepresentations of the self, and the view that the most central aspect ofall emotions consists of an action tendency, such as flight for fear,withdrawal for depression, and attack for anger. However, I do notagree with him and many others, including Lazarus (1991a) and Beck(1976), that the specific instigator of an emotion is the appraisal of astimulus situation. Rather, I believe it is the appraisal of a desirableaction. For example, the appraisal of a situation as threatening does not

58

necessarily produce fear. If the perceived threat is viewed as unjustifiedand the perpetrator as vulnerable to attack, and this results in anappraisal of attack as the most desirable response option, the emotionwill be anger rather than fear. Moreover, to be logically consistent, onceit is assumed that the essence of an emotion is an action tendency, thenthe implicit appraisal of an action tendency should be the criticalinstigator of a specific emotion. Of course, certain stimulus situationsare so closely associated with specific action tendencies that the apprai-sals of these situations are likely to be predictive of the appraisals of thepreferred action. This near correspondence between the two kinds ofappraisal, no doubt, is the source of the conceptual confusion that isprevalent on this issue.Frijda emphasizes the importance of ‘‘concerns’’ in producing emo-

tions. I agree with him that, in the absence of concerns, there is noemotion. However, unlike Frijda, I distinguish between concerns(which I refer to as motives), in two different information-processingsystems that I refer to as the experiential and rational systems. It is onlythe motives in the experiential system that are associated with emo-tions. Intellectual concerns that reside in the rational system, unlessthey are also associated with experiential concerns, do not elicit emo-tions. In fact, one way of distinguishing between motives in the twosystems is by observing whether or not they are accompanied byemotions. I will have more to say about this shortly.It is time now to turn to a consideration of my own views on the

executive self. In response to the editors’ request for commentators tocompare their views to those expressed in the chapters they reviewed, Iwill brieflydescribe the thoughts that I had about the self-conceptmanyyears ago that led to the introduction of a new global theory of personal-ity (Epstein, 1973) that I have since labeled cognitive–experiential self-theory (CEST). At the time, I gave the matter of an executive selfconsiderable thought because I intuitively felt there had to be some-thing like it but that it needed to be better articulated. I posed thefollowing riddle for myself: ‘‘What is it that has neither material sub-stance nor fixed content, yet can interpret and organize experience,direct behavior, interactwith the environment, and grow in the process,all properties that have been attributed to an executive self?’’ When theanswer came to me, it not only solved the problem of what the execu-tive self is, but it became the nucleus of a highly integrative,psychodynamic theory of personality, CEST, which, unlike other self-theories that emphasized a single cognitive system and ignored emo-tions, proposed two cognitive systems, one that operates consciouslyand analytically, and is relatively affect-free, and the other that operatespreconsciously and intuitively, and is intimately associatedwith affect.

59Commentary on Frijda

The solution to the riddle of the executive self lies in the recognitionthat it corresponds to an implicit personal theory of reality that, like ascientific theory, organizes experience, directs behavior, and growsthrough its interaction with the data of experience. An implicit theoryof reality, according to CEST, resides in the experiential system andincludes a self-theory, a world-theory, and their interaction. As animplicit theory that operates automatically and preconsciously, theself-theory in a person’s experiential system must be distinguishedfrom a person’s conscious explicit self-theory in the rational system.Thus, according to CEST, a person has two self-theories, and thereforetwo self-concepts, one in the experiential system and one in the rationalsystem. The two may correspond or diverge to different degrees, withthe degree of divergence being an important source of stress andmaladjustment. I believe that this conceptualization of the self-conceptalong with other aspects of CEST go a long way toward resolving theproblems concerning the self that the editors listed in the introductionto this volume requiring a solution. Following are the questions pres-ented by the editors and the solutions proposed by CEST:1. What is the role of emotions in the self-concept? According to

CEST, emotions are intimately associated with the operation of theexperiential system (in which the experiential self-concept resides).First they are considered to act as a barometer of the significance ofevents in a person’s self-concept. Accordingly, by noting the eventsthat trigger emotions, the significant schemas in a person’s self-concept can be inferred. For example, if I react with greater emotionto an assault on my appearance than to an assault on my intelligence,it can be inferred that my appearance is more important to me (in myexperiential system) than my intelligence, no matter what I may con-sciously believe (in my rational system). Second, since particular emo-tions are instigated by corresponding appraisals, it is possible to inferthe schemas in a person’s self-system from a person’s characteristicemotions. Thus, knowing that a person is characteristically angry(over a representative sample of life events) suggests that the person(in his or her experiential system) tends to view himself or herself asgood and just and other people as bad and unjust. Third, positiveand negative affect, which are most often determined by automaticappraisals in the experiential system, bias conscious thinking in aperson’s rational system. Through this mechanism the experientialself-system routinely influences the rational self-concept in the ab-sence of awareness. The implications of this for the widespreadpresence of human irrationality and for the meaning of ‘‘knowingonself’’ are evident. Fourth, emotions reinforce schemas in both theexperiential and rational self-concepts.

60 Seymour Epstein

2. How do emotions and cognitions influence each other? This ques-tion has been answered in my response to the first question.3. What direction does this influence take? As indicated in my re-

sponse to the first question, the influence is reciprocal. That is, emotionsdetermine which cognitions are reinforced, and preconscious cogni-tions determine which emotions are experienced. Also, conscious, re-flective cognitions are often used to correct preconscious, automaticcognitions that are the source of maladaptive emotions and behavioraltendencies. Further, the relatively primitive emotions available in earlydevelopment serve as nuclei aroundwhich broader and more differen-tiated cognitive-emotional networks develop.4. What kinds of events are relevant to an individual’s sense of self

andwhydo the same events influence different individuals differently?I already indicated that the degree of relevance of an event for aperson’s experiential self-concept is indicated by the degree of emotionit arouses. In turn, the emotion-eliciting properties of events are deter-mined by certain inherent properties (e.g. a universal fear of heights)and by their acquired significance as a result of previous experiences.As experiences differ for different individuals, people acquire differentemotional sensitivities to the same events.Of additional relevance to this question is the assumption in CEST

that there are four basic beliefs that develop as a result of their associ-ation with four basic needs. The four basic needs include the need tomaximize pleasure and minimize pain, the need to assimilate the dataof reality (by maintaining and extending one’s implicit theory of real-ity), the need for relatedness, and the need to enhance one’s self-esteem.The four related beliefs, respectively, include the belief in the degree towhich the world is benign relative to malevolent; the belief in thedegree to which the world is organized, predictable, and controllablerelative to chaotic, unpredictable, and uncontrollable; the belief in thedegree towhich people are supportive, trustworthy, and loving relativeto dangerous, untrustworthy, and rejecting, and the belief in the degreeto which one is worthy, lovable, and competent relative to unworthy,unlovable, and incompetent. Any event that is appraised as relevant tothese needs or beliefs is experienced as significant to the self andtherefore produces an emotional response.5. What processes can account for stability and change in self-

perception and self-evaluation over different time spans? The self-system is hierarchically organized. Central concepts are highly stable,whereas peripheral ones readily change. Holding centrality constant,although momentary changes occur frequently, they tend to occuraround a relatively stable mean. Thus, greater change is observed inperceptual, evaluative, behavioral, and any other kind of response


when measures are aggregated (intuitively or objectively) over shorterthan over longer time spans.6. What processes are responsible for long-term developmental

changes from infancy to adulthood? Change occurs as a result of theinteraction of maturation and emotionally significant experiences.There are critical periods in life in infancy, childhood, adolescence,young adulthood, middle age, and old age, each with its characteristicadaptive demands, as well as unique, individual experiences (e.g.traumas; transforming love relationships), that, depending on theirstrength, duration, and repetition, have the capacity to produce long-term changes in a person’s conception of self and world, and thereforein a person’s personality.7. Is fundamental and enduring change possible in adulthood, and, if

so, what mechanisms are responsible for such change? Such changescan occur when there are emotionally significant experiences of suffi-cient intensity, duration, or repetition to change fundamental schemasin a person’s experiential self-concept and/or world-concept.The editors concluded that these questions have not yet been ad-

equately addressed because of a lack of an integrative framework forconsidering them in combination. I submit that CESTprovides just sucha framework. Moreover, not only have these issues been addressed inCEST at a theoretical level, but many of the assumptions in CEST havebeen supported by an extensive body of research conducted over thepast twenty-six years (see review in Epstein and Pacini, 1999).In addition to providing answers to the above questions, CEST

identifies and provides a solution to a fundamental issue that theeditors have overlooked, and which, until it is resolved, makes a resol-ution of the other problems impossible. The issue concerns the differ-ence between implicit versus explicit cognitions, or, if you prefer,conscious versus preconscious or unconscious processing. Accordingto CEST, until it is recognized that there are two self-systems, precon-scious experiential and conscious rational, that operate by differentrules and contain different schemas (sometimes coinciding and some-times discrepant), it is impossible to answer questions about the selfbecause the answers differ for the two selves. It is beyond the scope ofthis brief comment to discuss these systems or the responses to theabove questions in greater detail. The interested reader can obtainfurther information in a number of publications that are readily avail-able in the literature and can also be obtained by writing to the author(e.g. Epstein, 1983, 1984, 1985, 1989, 1990, 1991a, 1991b, 1992b, 1993a,1993b, 1994, 1998a, 1998b, 1998c; Epstein and Pacini, 1999).In summary, Frijda has provided several important insights about

the relation of the self-concept and emotions, which is all he was

62 Seymour Epstein

requested to do. In contrast, my efforts have been devoted to construct-ing a global theory of personality in which the self-concept, emotions,motivation, levels of processing, and their interactions are centralconstructs.


COMMENTARY

The self and emotionsSeymour Epstein

In order to properly examine the relation between emotions and theself-concept, Frijda believes it is necessary to distinguish between theself as object of knowledge and the transcendental self. He makes acompelling case for the importance of the self as object of knowledge,but I have trouble understanding exactly what he means by the tran-scendental self. At one point he defines it as ‘‘the sum total of theinformationprocesses at a functional level’’ and at another as a person’s‘‘set of concerns.’’ These are both vague constructs that require clarifica-tion, and, unless I am missing something, they are not even the same.Others, as well as I, have proposed more clearly articulated distinc-

tions between two somewhat similar constructs to the ones proposedby Frijda. For example, William James (1910) initially, and GordonAllport (1955) later, expressed the difference between the two selves interms of the self as an object of knowledge and the self as an executiveself that is the source of behavior. Both, however, eventually disownedthe executive self as scientifically indefensible, likening it to a homun-culus residing inside a person’s head and directing the person’s behav-ior. The problem with the executive self and its homunculus analogy isthat understanding their behavior is no simpler than understanding thebehavior of the whole person and therefore does nothing to further ourunderstanding. I will discuss a more meaningful representation of theexecutive self later, when I present my own views.Although I amnot enamoredwith Frijda’s transcendental self, I am in

agreement with almost everything he says about emotions, includingthe important role he assigns to appraisal in the instigation of almost allemotions, the possibility of the occurrence of emotions in the absence ofrepresentations of the self, and the view that the most central aspect ofall emotions consists of an action tendency, such as flight for fear,withdrawal for depression, and attack for anger. However, I do notagree with him and many others, including Lazarus (1991a) and Beck(1976), that the specific instigator of an emotion is the appraisal of astimulus situation. Rather, I believe it is the appraisal of a desirableaction. For example, the appraisal of a situation as threatening does not

58

necessarily produce fear. If the perceived threat is viewed as unjustifiedand the perpetrator as vulnerable to attack, and this results in anappraisal of attack as the most desirable response option, the emotionwill be anger rather than fear. Moreover, to be logically consistent, onceit is assumed that the essence of an emotion is an action tendency, thenthe implicit appraisal of an action tendency should be the criticalinstigator of a specific emotion. Of course, certain stimulus situationsare so closely associated with specific action tendencies that the apprai-sals of these situations are likely to be predictive of the appraisals of thepreferred action. This near correspondence between the two kinds ofappraisal, no doubt, is the source of the conceptual confusion that isprevalent on this issue.Frijda emphasizes the importance of ‘‘concerns’’ in producing emo-

tions. I agree with him that, in the absence of concerns, there is noemotion. However, unlike Frijda, I distinguish between concerns(which I refer to as motives), in two different information-processingsystems that I refer to as the experiential and rational systems. It is onlythe motives in the experiential system that are associated with emo-tions. Intellectual concerns that reside in the rational system, unlessthey are also associated with experiential concerns, do not elicit emo-tions. In fact, one way of distinguishing between motives in the twosystems is by observing whether or not they are accompanied byemotions. I will have more to say about this shortly.It is time now to turn to a consideration of my own views on the

executive self. In response to the editors’ request for commentators tocompare their views to those expressed in the chapters they reviewed, Iwill brieflydescribe the thoughts that I had about the self-conceptmanyyears ago that led to the introduction of a new global theory of personal-ity (Epstein, 1973) that I have since labeled cognitive–experiential self-theory (CEST). At the time, I gave the matter of an executive selfconsiderable thought because I intuitively felt there had to be some-thing like it but that it needed to be better articulated. I posed thefollowing riddle for myself: ‘‘What is it that has neither material sub-stance nor fixed content, yet can interpret and organize experience,direct behavior, interactwith the environment, and grow in the process,all properties that have been attributed to an executive self?’’ When theanswer came to me, it not only solved the problem of what the execu-tive self is, but it became the nucleus of a highly integrative,psychodynamic theory of personality, CEST, which, unlike other self-theories that emphasized a single cognitive system and ignored emo-tions, proposed two cognitive systems, one that operates consciouslyand analytically, and is relatively affect-free, and the other that operatespreconsciously and intuitively, and is intimately associatedwith affect.


The solution to the riddle of the executive self lies in the recognitionthat it corresponds to an implicit personal theory of reality that, like ascientific theory, organizes experience, directs behavior, and growsthrough its interaction with the data of experience. An implicit theoryof reality, according to CEST, resides in the experiential system andincludes a self-theory, a world-theory, and their interaction. As animplicit theory that operates automatically and preconsciously, theself-theory in a person’s experiential system must be distinguishedfrom a person’s conscious explicit self-theory in the rational system.Thus, according to CEST, a person has two self-theories, and thereforetwo self-concepts, one in the experiential system and one in the rationalsystem. The two may correspond or diverge to different degrees, withthe degree of divergence being an important source of stress andmaladjustment. I believe that this conceptualization of the self-conceptalong with other aspects of CEST go a long way toward resolving theproblems concerning the self that the editors listed in the introductionto this volume requiring a solution. Following are the questions pres-ented by the editors and the solutions proposed by CEST:1. What is the role of emotions in the self-concept? According to

CEST, emotions are intimately associated with the operation of theexperiential system (in which the experiential self-concept resides).First they are considered to act as a barometer of the significance ofevents in a person’s self-concept. Accordingly, by noting the eventsthat trigger emotions, the significant schemas in a person’s self-concept can be inferred. For example, if I react with greater emotionto an assault on my appearance than to an assault on my intelligence,it can be inferred that my appearance is more important to me (in myexperiential system) than my intelligence, no matter what I may con-sciously believe (in my rational system). Second, since particular emo-tions are instigated by corresponding appraisals, it is possible to inferthe schemas in a person’s self-system from a person’s characteristicemotions. Thus, knowing that a person is characteristically angry(over a representative sample of life events) suggests that the person(in his or her experiential system) tends to view himself or herself asgood and just and other people as bad and unjust. Third, positiveand negative affect, which are most often determined by automaticappraisals in the experiential system, bias conscious thinking in aperson’s rational system. Through this mechanism the experientialself-system routinely influences the rational self-concept in the ab-sence of awareness. The implications of this for the widespreadpresence of human irrationality and for the meaning of ‘‘knowingonself’’ are evident. Fourth, emotions reinforce schemas in both theexperiential and rational self-concepts.

60 Seymour Epstein

2. How do emotions and cognitions influence each other? This ques-tion has been answered in my response to the first question.3. What direction does this influence take? As indicated in my re-

sponse to the first question, the influence is reciprocal. That is, emotionsdetermine which cognitions are reinforced, and preconscious cogni-tions determine which emotions are experienced. Also, conscious, re-flective cognitions are often used to correct preconscious, automaticcognitions that are the source of maladaptive emotions and behavioraltendencies. Further, the relatively primitive emotions available in earlydevelopment serve as nuclei aroundwhich broader and more differen-tiated cognitive-emotional networks develop.4. What kinds of events are relevant to an individual’s sense of self

andwhydo the same events influence different individuals differently?I already indicated that the degree of relevance of an event for aperson’s experiential self-concept is indicated by the degree of emotionit arouses. In turn, the emotion-eliciting properties of events are deter-mined by certain inherent properties (e.g. a universal fear of heights)and by their acquired significance as a result of previous experiences.As experiences differ for different individuals, people acquire differentemotional sensitivities to the same events.Of additional relevance to this question is the assumption in CEST

that there are four basic beliefs that develop as a result of their associ-ation with four basic needs. The four basic needs include the need tomaximize pleasure and minimize pain, the need to assimilate the dataof reality (by maintaining and extending one’s implicit theory of real-ity), the need for relatedness, and the need to enhance one’s self-esteem.The four related beliefs, respectively, include the belief in the degree towhich the world is benign relative to malevolent; the belief in thedegree to which the world is organized, predictable, and controllablerelative to chaotic, unpredictable, and uncontrollable; the belief in thedegree towhich people are supportive, trustworthy, and loving relativeto dangerous, untrustworthy, and rejecting, and the belief in the degreeto which one is worthy, lovable, and competent relative to unworthy,unlovable, and incompetent. Any event that is appraised as relevant tothese needs or beliefs is experienced as significant to the self andtherefore produces an emotional response.5. What processes can account for stability and change in self-

perception and self-evaluation over different time spans? The self-system is hierarchically organized. Central concepts are highly stable,whereas peripheral ones readily change. Holding centrality constant,although momentary changes occur frequently, they tend to occuraround a relatively stable mean. Thus, greater change is observed inperceptual, evaluative, behavioral, and any other kind of response


when measures are aggregated (intuitively or objectively) over shorterthan over longer time spans.6. What processes are responsible for long-term developmental

changes from infancy to adulthood? Change occurs as a result of theinteraction of maturation and emotionally significant experiences.There are critical periods in life in infancy, childhood, adolescence,young adulthood, middle age, and old age, each with its characteristicadaptive demands, as well as unique, individual experiences (e.g.traumas; transforming love relationships), that, depending on theirstrength, duration, and repetition, have the capacity to produce long-term changes in a person’s conception of self and world, and thereforein a person’s personality.7. Is fundamental and enduring change possible in adulthood, and, if

so, what mechanisms are responsible for such change? Such changescan occur when there are emotionally significant experiences of suffi-cient intensity, duration, or repetition to change fundamental schemasin a person’s experiential self-concept and/or world-concept.The editors concluded that these questions have not yet been ad-

equately addressed because of a lack of an integrative framework forconsidering them in combination. I submit that CESTprovides just sucha framework. Moreover, not only have these issues been addressed inCEST at a theoretical level, but many of the assumptions in CEST havebeen supported by an extensive body of research conducted over thepast twenty-six years (see review in Epstein and Pacini, 1999).In addition to providing answers to the above questions, CEST

identifies and provides a solution to a fundamental issue that theeditors have overlooked, and which, until it is resolved, makes a resol-ution of the other problems impossible. The issue concerns the differ-ence between implicit versus explicit cognitions, or, if you prefer,conscious versus preconscious or unconscious processing. Accordingto CEST, until it is recognized that there are two self-systems, precon-scious experiential and conscious rational, that operate by differentrules and contain different schemas (sometimes coinciding and some-times discrepant), it is impossible to answer questions about the selfbecause the answers differ for the two selves. It is beyond the scope ofthis brief comment to discuss these systems or the responses to theabove questions in greater detail. The interested reader can obtainfurther information in a number of publications that are readily avail-able in the literature and can also be obtained by writing to the author(e.g. Epstein, 1983, 1984, 1985, 1989, 1990, 1991a, 1991b, 1992b, 1993a,1993b, 1994, 1998a, 1998b, 1998c; Epstein and Pacini, 1999).In summary, Frijda has provided several important insights about

the relation of the self-concept and emotions, which is all he was

62 Seymour Epstein

requested to do. In contrast, my efforts have been devoted to construct-ing a global theory of personality in which the self-concept, emotions,motivation, levels of processing, and their interactions are centralconstructs.


CHAPTER 4

Fish, foxes, and talking in theclassroom: introducing dynamicsystems concepts and approachesPaul L. C. Van Geert

Dynamic system: a general definition

The phrase dynamic system contains thewords dynamic – referring to theGreek ‘‘dynamikos,’’ which means ‘‘powerful’’ – and system – a wordthat stems from a Greek verb that means ‘‘to combine.’’ Thus, a dy-namic system is a combination of things to which certain powers orforces apply. If a force is applied to something, it moves or changes(unless the force is counteracted). Defined in this way, the term dy-namic systemhas an extremely broadmeaning. For instance, a bunch ofdry leaves blown by the autumn wind can already be considered adynamic system. It consists of a collection of leaves all subject to thesame external force, the wind.In order to avoid such trivial applications, we should confine the

term to something that is more conceptually appealing. Let us begin byconfining the notion of system to collections of things that are related toone another in a way that corresponds with the notion of dynamic, thatis force- or power-related. We shall call something a dynamic system ifit consists of elements that exert specific influences or forces upon oneanother and, by doing so, change each other’s and their own properties(for general, technical introductions to dynamic systems, see, amongothers, Beltrami, 1987; Jackson, 1991a, b).In this chapter I shall adopt examples from different fields: physics,

biology, and, of course, psychology. My point is to show that dynamicprinciples apply to systems, irrespective of those systems’ actual formor nature. What matters are the relationships, not the content matter.Dynamic systems’ thinking is basically a way of thinking about sys-tems, not about psychology per se. Its importance for the field ofpsychology is that psychology, on the one hand, abounds with interest-ing opportunities for applying systems principles, while, on the otherhand, systems thinking has not yet fully permeated this field of inquiry(for applications of dynamic systems thinking to psychology, see,among others, Abraham, 1990; Levine and Fitzgerald, 1992a, b;Molenaar and Newell, 1998; Port and Van Gelder, 1995; Smith andThelen, 1993; Thelen and Smith, 1994; Van Geert, 1999).

64

Because we shall wander through different universes, let me beginwith a little Star Trek.Assume we join Captain Picard on one of his dramatic missions. One

day we arrive at that part of Deep Space known as the NewtonianUniverse. There we spot two freely moving bodies, exerting gravi-tational forces upon one another andmoving in an endless complicatedswirl. Since this is the Newtonian Universe – says Captain Picard – thetwo bodies are completely isolated from anything else in the Cosmos.The two bodies surely constitute a dynamic system. They affect oneanother through gravitational forces and by doing so change theirposition in space. The result, it turns out, is a simple, cyclical motion.But then our Captain breaks in and introduces a third celestial body andimmediately closes the Newtonian Universe again, leaving the threebodies on their own. But now we observe something quite differentfrom the simple cycle: the three bodies move around each other in acomplicated swirl that never seems to repeat itself.The three-body problem has been notoriously difficult to under-

stand, in spite of the fact that the two-body problem –which is only oneless – was the basic Newtonian idealization of a gravitational system. Itwas the French mathematician Poincare who presented a formal ap-proach to this and similar problems and by doing so paved the way fordynamic systems theory (Jackson, 1991a; Stewart, 1989).

Dynamic systems: general properties

The three-body problem illustrates one of the most appealing featuresof dynamic systems theory: you don’t need complicated systems toarrive at complicated behavior. It illustrates several other propertiesthat I shall explain further. It shows, among other things, that compli-cated patterns such as the motion of the three celestial bodies can beproduced by a few interacting elements. An external guidance plan,leading the paths in their various directions, is not necessary. Bywatch-ing other simple systems evolve over timewe observe that the nature ofthe systems’ behavior varies widely. The variations are caused by theway the elements interact, they are not caused by varying externalforces (this is not to deny the possibility, of course, that external forcescan eventually cause similar effects). However much the patterns maydiffer, they are not just chaos.Most of the patterns produced by interac-ting elements that exchange forces are not justmumbo-jumbovariation.They display both complexity and regularity that intuitively appeals tothe human eye.The ideal self-containedNewtonianUniverse visited by our Star Trek

crew lies of course outside the realm of the real. There is no part of the

65Introducing dynamic systems concepts and approaches

existing world that is completely isolated from all the rest. Even thethree celestial bodies deep in outer space will be subject to gravitationaland other forces from outside. The issue is, however, whether or not itmakes sense to treat the three objects as a semi-isolated system wherethe most important forces are those that the objects exert on one an-other. It makes sense to do so if the introduction of the dynamic systemsnotion allows one to come up with a better explanation of the behaviorof the system than some other approach would allow. The question is,however, what kind of explanation is this?Before introducing applications to psychology, let us proceed to

some biological examples (see, among others, Burghess and Wood,1985; Hofbauer and Sigmund, 1988; Kingsland, 1995; Murray, 1989). In1926, an Italian biologist by the name of Humberto d’Ancona made astudy of certain fish populations in the Adriatic Sea and observed apattern of increases and decreases in populations of fish. How couldthis pattern be explained? Did populations of fish have some inbuilt,genetic tendency to rise and fall in an endless cycle? Was there someunknowndisease that killed fish in a cyclical pattern? D’Ancona turnedto his father-in-law, Vito Volterra, who was a famous mathematician.Volterra transformed the problem into a dynamic systems issue bytreating the rise and fall of the populations as the result of an interactionbetween predator and prey fish (which his son-in-law had alreadyobserved from the biological point of view).

Dynamic systems: an example from biology

One can conceptually isolate the system by assuming that prey fishthrive on some unspecified food resource that, for the sake of simplic-ity, is viewed as something that comes in more or less constant quanti-ties. Prey fish die because they are eaten by predator fish (mackerel forinstance). Predator fish, in turn, thrive on the prey fish and the moreprey there is to catch themore predatorswill live. Predators die becauseof some external cause, the nature of which need not be taken intoaccount for the moment. For the sake of simplicity, we can treat thecauses of the predators’ deaths as a constant factor. Thus Volterra andd’Ancona’s fish system is a semi-closed system. It has a constant inflowof food for the prey fish and a constant outflow of dead predator fish.How could one ever expect to find a pattern of oscillations if both theinflow and outflow are a simple linear, constant flux (see figure 4.1)?Here’s where the dynamic nature of the interaction between predator

and prey fish comes in. Volterra showed that the interaction that Ispecified above in verbal terms (prey eating food, dying because they

66 Paul L. C. Van Geert

Figure 4.1. Volterra and d’Ancona’s model of population changes inpredator and prey fish; in order to make the populations graphicallycomparable, relative population sizes have been used

were eaten by predators, predators thriving on prey and dying fromsome external cause) could be described in simple mathematical terms.The solution took the form of two coupled differential equations. Oneequation described how predators were affected by prey, the otherspecified how prey was affected by predators.

Dynamic systems and geometric representation

A system like this one – and any other dynamic system for that matter –can easily be represented in a geometric form. Consider the number ofpredators – in a specific region of the Adriatic Sea – as one dimensionand the number of prey fish as another dimension. What we get then isa simple plane with two coordinates, one for the prey and one for thepredators. Each point in the plane represents a specific population sizeof prey and a population size of predator fish. The differential equa-tions mentioned above are used to calculate how the two populationswill evolve over a specific amount of time. The series of changingpopulations sizes – which are points in the plane – specify trajectories.The characteristic trajectory of a predator–prey dynamics is a some-what deformed cycle (see figure 4.2). This example demonstrates vari-ous aspects of what a dynamic systems approach to a phenomenonusually entails. Let us now take a closer look at those aspects.


Figure 4.2. Geometric representation of the predator and preypopulations over time

Dynamic systems as simplifying, conceptually semi-closed structures

First, the example makes clear that the system is in fact a conceptualconstruction. More precisely, we conceptually isolate some part ofreality (say the predator and the prey fish). By doing so, we create asystem (the relationship between the predator and prey fish) and itsenvironment. The environment is basically anything else in the uni-verse, except the system. The relationship between the system and itsenvironment is then simplified as much as possible. The easiest simpli-fication is to assume that the exchange between the system and itsenvironment is a constant flow. For instance, we assumed that the preyfish feed on a constant supply of food that comes from a source outsidethe predator–prey system (let us assume the prey fish feed on algae).We also assumed that the predator fish die at a constant rate (e.g. onefish out of ten dies per week).It is important to note that the assumptions about the system–

environment exchange need not be empirically correct, in the sense thatthey do not need to refer to an actual empirical system. It suffices thatthe assumed exchange is empirically possible. Once we understand thesystem, we can apply actual empirical conditions to the system and seewhether the system helps us understand the actual state of affairs (e.g.the food supply is variable, the death rate varies stochastically as aconsequence of diseases or the presence of predators that feed onpredators). To put it differently, the system is in fact amodel that shows


some deep similarities with the reality it models but that otherwisesimplifies and reduces that reality as much as possible.Being a conceptual construction, a dynamic system is basically a

semi-closed (or semi-open, depending on one’s perspective) system. Bytreating the predator–prey relationship as an isolated event, we try toprotect it as much as possible from interference from external sources,whatever their nature. However, given that the system refers to somepart of the natural universe, it must have a connection with that uni-verse. We have seen that this connection is conceptually reduced to thesimplest form possible.

Dynamic systems are characterized by a structure of interactions

Once the system has been conceptually isolated, it must be specified bycharacterizing the time-dependent interaction of its components. Thefirst step is to provide a conceptual description of the interaction. Thatdescription must confine itself to the most elementary and essentialaspect of the interaction. For instance, we observe that, first, the preyfish population grows at some constant rate that directly reflects theconstant external food supply; second, that the prey fish die at someconstant rate that directly reflects the species’ natural longevity; third,that the prey fish die at some variable rate that reflects the number ofpredators that feeds on them; fourth, that the predator fish populationgrows as a function of the food supply, i.e. the prey fish; and, fifth, thatthe predator fish die at a rate that reflects the species’ longevity. Thiselementary conceptual description is of course based on empirical factsor at least on empirically justifiable assumptions. However, there are alot of empirically established facts that can and in fact should be left outof the conceptual specification because they are not essential. For in-stance, prey fish not only die of old age or because a predator catchesthem, they also die from diseases. Diseases, however, are not explicitlymentioned in the model and are in fact treated as a constant.The art of building a dynamic systemsmodel is to reduce the concep-

tual description to its bare minimum. What this bare minimum isdepends on whether the minimal description of the relationships be-tween the components suffices to make the system do what it is sup-posed to do. In the case of the predator and prey fish example, thesystem should produce the empirically observed cyclic, directed pat-tern of oscillations in the predator and prey population. As far as theminimum is concerned, it turns out that we can actually leave out thesecond assumption (the constant death rate of prey assumption): theremaining four suffice to produce the required pattern.Fish – whether or not swimming in the warm waters of the Adriatic


Sea are not exactly the psychologist’s staple. However, principles thatgovern dynamic systems do not depend on the actual content matter towhich they apply. System properties apply to specific types of interac-tion and it does not really matter whether such interactions occur withbiological organisms or psychological variables. Let me try to explainhow the dynamics that characterized the nature of the interactionbetween prey and predator fish also apply to human behaviors andpsychological variables.Some actions that people undertake lead to negative consequences.

For instance, if a child talks toomuch or too loudly in the classroom, it islikely that the teacher will show her discontent, for instance by punish-ing the talking by a negative remark spoken in a loud voice. Assumingthat such remarks are so-called aversive stimuli for the child in questionand that the child experiences them as negative, for instance becausethey cause negative emotions in the child, we know that the punish-ment will contribute to the child’s avoidance of the punished behavior.However, the unwantedbehavior – the talking in the classroom– has itsown rewards and reinforcements, for instance because exchangingfeelings and ideaswith other children brings about positive emotions inthe child and thus increases the frequency with which such behaviorwill occur in the future. All this amounts to the principles of simple,general learning theory (see, for instance, Novak, 1996). We also knowthat the so-called aversive stimuli, the teacher’s reprimands and angrylooks, are related to the cause of the stimulus, which is the child’stalking.Moreover, it is reasonable to assume that the gravity of the aversive

stimulus, let us say the seriousness of the punishment, is related to thefrequency with which the unwanted behaviors actually occur. Thisrelationship could be a simple linear one – the gravity of the punish-ment is proportional to the frequency of the unwanted talking – but it isunlikely that this would be so. Teachers have an image of their pupilsand that image changes as a consequence of their experiences with thepupils (and their eventual biases or prejudices with regard to thosepupils). The relationship between a pupil’s unwanted-talking fre-quency and the gravity of the punishment – let us say the intensity ofnegative emotions that accompany the teacher’s reprimands – is moreprobably one of change or growth over time. As the pupil’s frequencyof unwanted behavior increases, the average seriousness of theteacher’s reprimands also grows bigger. On the other hand, we mayalso assume that the reprimands are subject to some formof habituationor, in learning-theoretical terms, extinction. The reprimands of theteacher do not always have the desired effect, in that the pupil stillshows the unwanted behavior (its frequency is less than the frequency


it would have without punishment, but there is no possibility for theteacher to know what this frequency would have been and to reckonwith it; that is, the unknown frequency cannot play a reinforcing orrewarding role with regard to increasing the frequency of the punish-ment behaviors). Thus we may assume that some amount of extinctionwill apply to the frequency of the punishment or to the gravity of eachpunishment event.What is the relationship between the fish, the talking in the class-

room, and the reprimands? It is easy to see that the unwanted talking inthe classroom is formally equivalent to the prey fish. The talking in-creases proportionally as a consequence of being rewarded and rein-forced through social processes of the classmates, just as the preypopulation grows proportionally as a consequence of the availability ofa food resource. On the other hand, the talking decreases as a conse-quence of the punishment (the reprimands) in line with the principle ofavoidance learning. It decreases, first, as a consequence of the fre-quency of the punishment, which is linearly related to the frequency ofthe unwanted talking, and, second, as a consequence of the gravity ofthe punishment (more serious punishment causes more avoidancelearning effect). The prey population, in turn, decreases because indi-vidual fish are being eaten by predators. The decrease depends, first, onhowmany fish are available as food for the predators (the prey popula-tion) and, second, on how many predators there are (the predatorpopulation).The variable seriousness-of-the-reprimands is formally similar to the

variable size-of-the-predator-population. The seriousness or gravity ofthe reprimands increases proportionally as a consequence of the fre-quency of the unwanted talking in the classroom, similar to the pred-ator population that increases proportionally to the available foodsupply, the prey fishes. The gravity, however, also decreases as aconsequence of extinction or habituation, which directly acts upon thefrequency of the punishments, the gravity, or both. This system ofrelationships can be compared to that applying to the predator fish,Their population increases proportionally as a consequence of theavailable food supply (the prey) but meanwhile decreases as a conse-quence of natural deaths of predator fish, which is of course alsoproportional to the predator population.In summary, a dynamic systemmodel applies to a set of relationships

between components of a system, not to the actual physical instanti-ation of those components. From that point of view, the relationshipsbetweenpredator and prey fish, on the one hand, and unwanted talkingand reprimands, on the other hand, can in principle be described by asimilar dynamic model.


From a structure of interactions to a mathematical model

Now that we have obtained some minimal conceptual specification ofthe time-dependent interactions in the system under scrutiny, we areready to specify those interactions in a mathematical form. I take theterm ‘‘mathematical’’ in its broadest possible form, i.e. as any algorith-mic procedure that can be used to generate certain results or outcomes.In general, the mathematical expression of the conceptual relationshipsis written in the language of differential calculus. I shall take theunwanted-talking–reprimands situation as the domain of applicationof the mathematical model, although it should be noted that the result-ing model applies to any situation that is compliant with the system ofinteraction described in this particular situation and in the predator–prey case

dBdt= �B− �BP

equation 1dPdt=�PB−�P

In spite of its – probably – somewhat unusual appearance (most readerswill not be used to reading equations with Greek characters), theequation is in fact astonishingly simple. It says that any change in thefrequency of unwanted talking over time (represented by B for ‘‘behav-ior’’; dB/dt means ‘‘the difference in the frequency of the behavior oversome change in time’’) is equal to the increase in the behavioral fre-quency thanks to some factor alpha (which depends on the rewardingor reinforcing social effects of the talking to other children in theclassroom and which can be represented by any reasonable constant)minus the decrease in the frequency of talking (a decrease that dependson the actual frequency of the reprimands, which is a linear function ofthe frequency of the unwanted talking B itself, the gravity of thepunishment, P, and some constant that specifies the magnitude of theeffect of punishment on the decrease in talking frequency; note that thisconstant, lambda, is the product of an effect size constant and addi-tional constants, such as the punishment ratio schedule). The secondpart of the equation does a similar thing with the gravity of the punish-ment. It says that the change in the gravity of the punishment depends,first, on the automatic decrease in the gravity of the punishment due toextinction or habituation effects (specified by a constant beta; note thatbeta may affect both the gravity of the punishment itself and thefrequency of the punishment: if the frequency decreases, the resultingaverage gravity of the punishment per unit time also diminishes); and,second, on the proportional increase in the gravity P of the punishment


due to the level of unwanted behaviors B on which it ‘‘feeds,’’ so tospeak, and on a constant mu that moderates the effect size of P and B.The remarkable thing is that the complicated interaction between un-wanted behavior and punishment or other aversive stimuli – or thenatural interaction between predators and their prey, for that matter –can in fact be simplified to this extremely frugalmathematical specifica-tion. In spite of its incredible simplicity, it does capture the essence ofthe real dynamics between behaviors or predators and prey.

From the mathematical model to the geometric relationship

Given the equations that formally specify ourmodel, we can proceed tothe next step in exploring the dynamic model. It is easy to see that themodel and the equation are related to a simple geometric structure,namely a two-dimensional plane. One dimension of the plane is thepossible frequency of the unwanted talking behavior, B, the other is thepossible gravity of the punishment, P. The beauty of combining themathematical equation with the representation in the two-dimensionalplane is that we can literally see how the talking and the punishment, orthe predator and prey populations, for that matter, evolve across time.Since frequencies of behavior and level of gravity of punishments arenot easy to imagine, I shall revert to the original biological predator–prey example (recall that the dynamics as such remain the same,irrespective of the domain of application; what applies to this domainare specific assumptions regarding characteristic levels and time units;such assumptions may differ from domain to domain). For the sake ofsimplicity, I shall use a classical predator–prey example, which consistsof rabbits, represented by r and foxes, represented by f.Assume that, given the time interval is a year, we can make some

reasonable estimations of the values of the alpha, beta, lambda, andmuparameters. Assuming that we know at least something about rabbitand fox populations, or about predator and prey fish in the AdriaticSea, it shouldn’t be too difficult to arrive at an initial guess of the valuesof those parameters. The initial guess need not be empirically correct. Itsuffices that it is empirically possible, given our knowledge of theanimals involved. Armedwith these values, we can take any number offoxes and rabbits that we want and calculate how many rabbits andfoxes – or predator and prey fish – we shall find after a year (dt). Again,we use our common-sense knowledge of those animals in order to set areasonable initial guess. For instance, it is reasonable to assume that thenumber of rabbits is considerably bigger than the number of foxes (andnot the otherway round). For the sake of illustration, I shall assume thatthe parameter values are as follows: alpha .11, beta .2, lambda .01 and


Figure 4.3. Numbers of predators and prey in the predator–preyspace over three years

mu .0002. The starting values are 1000 for the rabbit population and 13for the fox population. With these values, it is easy to calculate therabbit population next year (which is the actual population – 1000animals – plus the change, which is given by .11 * 1000 – .01 * 13 * 1000),which is 980. The fox population next year will again be 13 (the actualpopulation, 13, plus the change, which is –.2 * 13+ .0002 * 1000 * 13). Ourequations tell us that two years later, the rabbit populationwill be 960.4and the fox population will be 12.948 . . .Confronted with these numbers, I take the opportunity to make two

important remarks. The first remark addresses the issue of geometricrepresentation. If we temporarily forbear the ghostly .4 rabbit and .948fox and take the numbers as they are, we can easily see why thedynamic model produces a geometric representation of the change inpredator and prey populations. Since both the number of predators(foxes) and the number of prey (rabbits) are considered as dimensionsof a two-dimensional space, the combinations of a rabbit and a foxpopulation (1000 and 13, 980 and 13, 960.4 and 12.984) correspond withthree separate points in the predator–prey plane. This is a simpleapplication of the principle of the Cartesian plane (see figure 4.3).Since the points are connected by the predator–prey dynamic prin-

ciple (the first pair of values produces the second, the second the third,etc.) we may connect the consecutive points by a line. By doing so, weproduce a line in the predator–prey space. If we go on calculating the


Figure 4.4. Numbers of predators and prey in the predator–preyspace over 100 years

predator and prey populations for, say, 100 consecutive years, weobtain a cyclical pattern (which, in the case of a model that operateson fixed steps of one year looks more or less like a menhir fallen on itsside; see figure 4.4). This cyclical pattern corresponds with sinusoidalpatterns over time of the population sizes of predators and preyrespectively (see figure 4.1). The cyclical pattern in the predator–prey space is the typical geometric format of a predator–preydynamics.If we experiment with various parameter values, we find in facta structure of nested cycles, reminiscent of the structure of an onion(see figure 4.5).Note that a qualitatively similar pattern will apply to the unwanted-

talking–punishment example, because the behavioral situation is basedon a similar dynamics. Quantitative differences may apply to the sizeand form of the cycles, but not to the principle of cyclical evolution,which is a direct consequence of the current dynamics and of nothingelse.

Mathematical analysis versus modeling

The second remark Iwant tomake concerns the strange .984 fox and the.4 rabbit. Presumably .4 rabbit is what we may find on the plate of acarnivorous gourmet, but it is certainly not something that runs aroundin the wild. At this point we hit the distinction between mathematical


Figure 4.5. Geometric representation of the predator and preypopulations over time; different cycles result from applying differentstarting values

analysis and modeling. From the point of view of mathematical analysisof a dynamic model, the dimensions are just numbers, not representa-tions of rabbit and fox populations.Moreover, since the dynamicmodelis based on differential equations, they should be solved as differentialequations. More precisely, the dr/dt specification is a description ofcontinuous change, which can be found if dt (the time interval) ap-proaches the zero limit. Approaching the predator–prey dynamicsfrom the mathematical side, the American mathematician Lotka,together with Volterra, whom we have already mentioned, succeededin solving the differential equations that form the backbone of anypredator–prey relationship. The typical solution of a Lotka–Volterrasystem looks a bit like the pattern of layers of an onion, or moreprecisely, like those of a clove of garlic (see figure 4.5).From a biological point of view, however, a population consists of

whole numbers of animals. This means that if we want to reckon withthe fact that the Lotka–Volterra equations refer to numbers of animalsinstead of mere numbers, we must change the equations in order toaccount for this whole-number restriction. Put differently, if wewant tomodel the fate of two populations we must take into account all factorsthatwe consider relevant from the viewpoint of describing the aspect ofreality in question. It goes without saying that a similar adaptation hasto be made with models that specify discrete behavioral interactions,for instance, discrete behavioral events or discrete units of knowledge.


Accounting for the fact that populations – or discrete behavioralevents, for that matter – consist of whole numbers of animals is just afirst step. We might also wish to consider the fact that real dynamicsinvolve a host of additional factors.For instance, with regard to the relationship between whiting (prey

fish) and mackerel (predator), biologists discovered that there is a thirdparty that complicates the predator–prey cycle. They found that thesquid Illex feeds upon young mackerel, whereas the adult mackerelfeed upon young squid. In addition, squid also are cannibalistic, whichimplies that they feed on themselves (speaking population-wise). Putdifferently, the predator–prey cycle consists of a complicated set ofcycles and subcycles, each of which complies with the basic predator–prey structure described earlier.The behavioral case poses similar difficulties. For instance, neither

the frequency nor the gravity of the punishments is really fixed. Al-though the model itself specifies only a single value for both variables,the actual values will fluctuate considerably. If the model provides areasonable image of the process, the fluctuations should center on thecomputed values. One way to bridge the gap between a mathematicalmodel – for instance the predator–prey model – and a process model –that also accounts for additional factors – is to add random factors to themathematical model, in an attempt to delineate the potential effects ofrandom or accidental variation. By randomizing the parameters andvariables, one can carry out a sensitivity analysis of the model, todetermine how sensitively the model behaves with regard to empiri-cally acceptable variations. Models that are too sensitive, i.e. that pro-duce strongly varying outcomes as a result of minor randomvariations,should in general not be considered empirically adequate. Note,though, that this rule does not hold if the variations occur in the vicinityof bifurcation points, i.e. places in the space of parameters that lie at theboundary of different attractor regions (different equilibria).In spite of these complications, however, two conclusions remain

valid. The first is that complicated dynamics may be modeled on thebasis of the relationships exemplified in the simple dynamics. Thesecond is that the simplifications, for instance of the most elementarypredator–prey dynamics, nevertheless cover a considerable part of theactual population phenomena.In summary, we have now seen that a dynamic system is in fact a

conceptual structure that specifies a semi-closed system. It isolates aparticular interaction that takes place between particular parts or as-pects of the world and specifies the relationship with the rest of theworld in terms of the simplest possible flow. In the case of the un-wanted behavior-aversive stimuli model (or the predator–prey model


for that matter), the isolated property concerns the interaction betweenan unwanted behavior and the gravity of an aversive stimulus (punish-ment) and describes how both the unwanted behavior and the aversivestimulus change as a consequence of this interaction. The relationshipwith the rest of the world concerns everything in the model that is notspecified by the interaction term. In the behavioral case of the predator–prey equations described earlier, this amounts, for instance, to allconditions that make the unwanted behavior change in frequency (e.g.the social rewards the child gets by talking to his or her classmates).

Dynamic systems and self-organization

Given the distinction between the internal interaction componentand the external–internal flow component (food, deaths, . . .), we cannow proceed to an important, if not central property of dynamicsystems, namely that of emergent properties and self-organization.Self-organization is sometimes treated as a somewhat mystical prop-erty. In this overview I shall try to explain it in its simplest possibleform, although it remains true that many forms of self-organization areamazing in their complexity (Holland, 1998; Kauffman, 1993; Prigogineand Stengers, 1984; van Geert, 1989).I shall begin with the first and second laws of thermodynamics

(Atkins, 1984). Whereas the first law concerns the overall conservationof energy, the second law states that physical systems move spon-taneously and irreversibly toward a state of disorder. For instance,when I get myself a cup of freshly made, hot tea and leave the cup onthe table, physically speaking I start with a state of increased order.There exists a concentration of high temperature in the cup, whichdiffers from the ambient temperature inmy study by about 60°C.WhenI come back after an hour I shall discover, probably to my regret, thatmy tea has cooled off and that the heat from the tea has been transferredfrom the cup to the room, resulting in a – barely noticeable – increase inthe room’s temperature. We refer to this state as a state of increaseddisorder, in that the concentrated heat has now been distributed moreor less evenly across the room. This process of increasing disorder isalso known as entropy.The Russian-born, Belgian physical chemist Prigogine showed that

complex systems that receive and transmit a flow of energy and matterfrom and to their environment eventually show a process that runscounter to the second law. In these systems order increases rather thandecreases. Another way to put this is to say that the system self-organizes into a state of increased order.There are several concepts that need further explanation here. Let me


begin with the term ‘‘self-organization.’’ Self-organization means thatthe order comes from the system itself, not from the environment (in thelatter case it would be organization, not self-organization). The secondimportant concept, order, implies a degree of patterning or similarity.The predator–prey model nicely illustrates the principles implicit inboth self-organization and order (the behavioral example does that too,of course, since it is based on the same dynamics; the explanation isprobably a bit easier to follow if I use the concrete example of popula-tions of animals as from the predator–prey model).Let us first assume that the foxes and the rabbits are not coupled in a

predator–prey relationship, the rabbit–fox system (which would just bea collection of rabbits and foxes in some bounded region). In that case,the system would be characterized by a proportional inflow of energy,which is the food consumed by the ever-increasing population of rab-bits (the net result of rabbit births minus rabbit deaths). There wouldalso be a constant outflow of energy, which consists of dead foxes (inthis simplifiedworld, the foxeswould have nothing to eat – they are notcoupled to the rabbits – and hence only the population decrease factor �would apply).Formally speaking, the ratio between any value of r (the rabbit

population) and any preceding value of r is a constant (namely 1 + r; thesame principle holds for the fox population f ). Thus the flow exhibits asimple order, expressed by the constants 1 + r and 1+ f. However, oncewe couple the rabbits and foxes in the form of a predator–prey dynam-ics an entirely different pattern results which consists of a regular andcoherent repetition of rises and falls in the populations. This patternshows a considerably higher level of complexity than the simple con-stant from the original system–environment flow. In order to visualizethis pattern I shall use a simple trick. Recall that the ratio between anylevel of f (or r for thatmatter) and any preceding level equals a constant,namely 1 + f. I can plot the value of that ratio (in fact the first derivativeof the growth function) in a two-dimensional space. One dimension isthe value of the ratio at time t, the other dimension is the value of thatratio at some later time, t + n (n could be 10, for instance). Since the ratiois a constant in the case of the flow, the values at time t and t+nare identical. They specify a point in the two-dimensional spacewith coordinates (1 + f, 1 + f ). I now compute the ratio between anyfox population level and any preceding level in the predator–preycase. When I now plot the ratio at time t against that of time t+ n,the resulting image takes the form of a rounded triangular shape(figure 4.6).The simple point – the simplest possible form of order in the

two-dimensional plane – has now been replaced by a regular but


Figure 4.6. The ratio of the present population level (time t) and thepreceding level (time t−10) for 1000 consecutive years.

considerablymore complicated form of the rounded triangle. This newform must be the product of the predator–prey relationship and ofnothing else but the predator–prey relationship. It cannot come fromthe environment, since the interaction with the environment has thegeometric form of a simple point, which refers to a considerably lowerorder of organization. Likewise, it has also not been imported or smug-gled into the form of the dynamics, since the only thingswe have addedto the flow part of the equations are the components that specify thecoupling (which is �BP and �BP in the behavioral example and �rf and�rf in the rabbit–fox predator–prey system; see equation 1). Conse-quently, the higher order is the sole product of the interaction dynami-cs. In that sense, we say that the system shows self-organization,mean-ing that it produces a higher level of order (organization) by itself. Sincethe form of the interaction, represented by the rounded triangularshape, is in no way prefigured or pre-represented in the form orparameters of the dynamics, we can also say this form is a property thatemerges from the dynamics.Assume that, instead of running into the oscillatory pattern, the

predator–prey dynamics produced a completely random pattern ofpopulation sizes. Thismeans that the ratio between any current popula-tion level and a preceding level would be random as well (again, theexample directly generalizes to the unwanted behavior-aversive stimu-lus case). Plotting those ratios in our t and t +n space would finally


result in covering all points in the t and t +n space. This result (if itapplied to the predator–prey dynamics, but recall that it does not)would actually amount to a complete loss of order. Instead of the highlyconstricted order of the flow (it is specified by one and only one point inthe space), we would now have a completely unlimited range of possi-bilities. This case would be similar to maximal entropy on the ther-modynamics level. However, the order produced by the predator–preydynamics is of a completely different kind. In contrast with the com-pletely random case, it is highly specific: only a very limited and highlyspecific set of points are allowed (namely the points that form therounded triangular shape). On the other hand, in contrast with thesingle-point case, it contains much more information, that is, a higherdegree of order.In summary, we can define a dynamic system as a semi-isolated

system, characterized by a specific internal structure and a specificflow from the environment. If it is a non-linear dynamic system, itwill increase the amount of order in comparison with the amount oforder present in the environment flow. In principle, it will increasethe amount of order up to a level that is characteristic of the system(which, in the case of the predator–prey or of the unwanted-behavior–punishment relationship is the oscillatory pattern of the populationsand of the frequency–punishment relationship, respectively, specifiedby the rounded triangular and egg-shaped forms).

Dynamic systems and time

An important, though almost tautological, aspect of dynamic systems isthe aspect of time. The components or variables out of which the systemis composed affect one another across time. The patterns that emergeare temporal patterns, characteristic changes in the values of the vari-ables involved or characteristic replacements and transformations ofthe system’s physical elements. The time-governed aspect of dynamicsystems is what distinguishes them from the structures of relationshipsbetween variables described in terms of regression models or compar-able models of association. An association model describes the way avariable is linked with another variable, for instance how level ofschooling is associated with a person’s income, to mention a standardexample from introductory courses in statistics. Association does notimply that one variable causes the other to change (or the other wayround), as would be implied in a dynamic system. Most often, associ-ation measures such as regression describe the coherence of variablesacross a sample at a specific moment in time. For instance, in casesof relationships between unwanted behavior, such as talking in the


classroom, and the punishment that follows such behavior, the associ-ation could easily be summarized in the form of a simple linear relation-ship, a correlationmeasure. However, because of the oscillatory patterncharacteristic of this particular dynamics, the Pearson product momentcorrelation between the values of the coupled variables is about zero.This finding from the regression model implies that – in the model –unwanted behavior and punishment are not related at all, whereas weknow that there exists a strong, deterministic, and mutual relationshipbetween the two.Dynamic systems describe changes in variables acrosstime for a specific case (e.g. a single individual). The dynamic relation-ship does not always show in the form of a linear association betweenthe variables (which is not to say, of course, that all cases where thecorrelation coefficient is small or almost zero hide some complicateddynamics; the relationship between linear and dynamic associations isclearly not symmetrical).

Dynamic systems and attractors

Although the properties described so far constitute the basic features ofdynamic systems – the semi-closedness, the geometric nature of therepresentation, the creation of order – there is of course far more to sayabout them. There is a host of concepts and notions that are regularlyused in the context of dynamic systems modeling and application. Animportant concept concerns the notion of attractor. An attractor can bedefined as the region of state space that captures the long-termbehaviorof the system. By state space I mean the geometric space formed by thecomponents or variables that constitute the system. In the unwanted-behavior–punishment case, the state space consisted of the variableunwanted-behavior-frequency and the variable gravity-of-the-punish-ment, respectively. We have seen that any combination of a behaviorfrequency and a specific punishment level defines one point in thebehavioral state space. The dynamics explains how, for any givenstarting level of unwanted behavior and of punishment and any givenset of parameter values, the variables will evolve either towards somefixed set of values or towards a repeated pattern of changes. The changein the variables is described by the geometric pattern from figures 4.5and 4.6 (note that similar patterns apply to the fate of predator–preypopulations in the biological example). This pattern occupies a speci-fied region of the state space, hence the definition of attractor. Wedistinguish between four types of attractors. The first is a single point inthe state space and corresponds with a so-called steady state (forinstance, a fixed level of unwanted behavior and a fixed level of punish-ment for which the system is in equilibrium). The second is a closed


loop or periodic cycle, as in the current behavioral or predator–preyexample. The third is a torus, which is in fact a combination of cycles.The fourth is a relatively new discovery, dating back to the 1960s, and iscalled a strange attractor. A strange attractor is like a bounded region ofthe state space (for instance the space circumscribed by a cycle in thestate space). In that region, the changes in the system are chaotic. Chaosis another highly appealing concept from dynamic systems theory. Inclassical mechanics, randomness and unpredictability of processeswere considered apparent and were attributed to the fact that theconditions under which the processes occurred, including their causes,were not sufficiently known.We now know that there exist very simpleand completely deterministic systems that nevertheless display behav-ior that is – seemingly – completely random and unpredictable. A goodexample of such a system is the logistic map. The logistic map is thediscrete form of the logistic growth equation (by discrete is meant thatthe time steps occur in discrete units and not continuously). The equa-tion of the logistic map with an equilibrium level equal to 1 is

Lt+1=Lt + Lt � (1 − Lt) equation 2

If the parameter � is between 2.54 and 3 the consecutive values of Lt+1show a chaotic oscillation. An important feature of deterministic chaosis its extreme sensitivity to initial conditions. Any infinitesimally smalldifference between the starting values of an initial Lt and another initialLt will after a while lead to an arbitrarily large deviation between theirchaotic series. The practical importance of chaos is that it provides apotentially internal source of arbitrarily large or small variation todynamic processes. In the standard view, variation or more preciselyfluctuation over time is viewed as the result of error terms, that is, termsresulting from causal processes that are not being controlled for. Itfollows then that the more variables are being controlled by a processoperator (e.g. an experimental researcher), the smaller the contributionof random factors to the process (because the number of random factorsis reduced by the extensive control). The existence of deterministicchaos, however, opens the possibility for randomization based on theprocess itself. Although the actual fluctuations in the processwill be theresult of both the intrinsic fluctuations based on the internal chaoticprocess and the extrinsic fluctuations based on uncontrolled externalfactors, reduction of the latter will not considerably reduce the degreeof randomness produced. The reason is that chaotic processes canfill up any arbitrarily large region of the state space. As a result,externally added randomness will not change the nature of the processfluctuations.


Dynamic systems and bifurcations

The concept of bifurcation refers to the occurrence of sudden shifts inthe behavior of a system as a result of a gradual upgrading of the valueof one of the process parameters. In a linear process we expect that theincrease in the value of a parameter leads to an increase in a dependentvariable. In principle, it does not matter whether the increase is fromsome very low value to a somewhat higher value or from some highvalue to an even higher value. The effects will be quantitatively differ-ent, but qualitatively similar (an increase, say). In non-linear systems,that is, systems that are capable of producing or increasing order,quantitative changes in parameter values sometimes lead to suddenchanges in the quality of the system’s output. Some systems – includingpredator–prey systems – home in onto a point attractor for any valueof a parameter under a critical value pc. If the parameter crosses thecritical value, the system no longer settles into a point attractor butmoves towards a cyclical attractor. This transition is known as a Hopfbifurcation.Another, very simple example of bifurcation is provided bythe logistic map. The growth rate parameter, � in the equation, leads toa point attractor for values smaller than 2; for values between 2 and 2.54it leads to cyclical attractors with a doubling period (from 2 points to 4,from 4 to 8, etc.); above 2.54 the attractor becomes chaotic. Bifurcationsare interesting phenomena because they demonstrate that discontinu-ity in the system’s output (the sudden change in the qualitative behav-ior) can be based on continuity in the system’s input (for example, aflow that causes the parameter at issue to increase in a continuousway).Or, to put it differently, discontinuity can be the intrinsic product of thesystem dynamics and does not necessarily depend on a discontinuousexternal input (Van der Maas and Molenaar, 1992).

Using dynamic systems thinking in understanding psychologicalphenomena

After giving this overview of general properties and principles ofdynamic systems,wemay now ask ourselves howall this could eventu-ally be put to use in psychological theory building and research. Ouraim is to understand psychological phenomena, not the fate of rabbitsand foxes or mackerel and whiting. One of the major advantages ofdynamic systems models is their extreme generality. They are almostliterally stripped of anything that reminds us of the complexity andspecificity of real natural phenomena. In the preceding sections, wehave seen that the dynamics that applied to fish, or to rabbits and foxes,also applied to unwanted behaviors, such as talking in the classroom


and the resulting punishment. Dependent on the nature of the phenom-ena under scrutiny or themodels available for such phenomena,we canapply either a quantitative or a qualitative approach to dynamic sys-tems modeling.

A quantitative approach

We accepted that the equations of the predator–prey model apply tobiological populations because the components of the equations wereseen as highly generalized descriptions of biological phenomena suchas ‘‘animals die, hence the population decreases’’ and ‘‘animals feed onprey, hence the population increases relative to the available prey.’’ Theequations themselves, however, referred to additions and subtractions,to multiplication and comparable operations. We have seen that thelogic of predator–prey interactions generalizes to interactions betweenbehaviors or actions in different persons. The meaning of this – prob-able – similarity is not that the relationship between correctional andantisocial activities is a form of predator–prey interaction (although it isnot wrong to state the relationship in this way). It is probably morecorrect to say that there exists a general asymmetrical model, describ-able by equation 1, that describes a wide variety of phenomena, includ-ing biological predator–prey relationships and relationships betweencorrectional and antisocial or unwanted behaviors. According toThatcher (1994) the model also specifies the changes in post-natalintra-cortical connections, with different regions of the brain corre-sponding to the ‘‘predator’’ and the ‘‘prey’’ parts.Comparable models can be built for many other phenomena in

psychological change, growth, and development. For instance, it can beargued that the so-called elementary or primitive emotions are not likepre-wired emotional schemes but rather self-organize rapidly as aresult of simple competitive and supportive relationships betweenconstituents of various kinds. Examples of such constituents are motorpatterns of facial muscles, internal hormonal secretions, perceptualpatterns, and so forth (see De Weerth and Van Geert, 2000; Camras,Lambrecht, and Michel, 1996).At this point in the discussion, we should mention two kinds of

difficulties associated with this quantitative approach to dynamicmodel building in psychology. The first difficulty concerns the possibil-ity of representing the phenomena at issue by a simple mathematicalmodel. It is often assumed that psychological phenomena are farmore complex than biological interactions between foxes and rabbits,for instance, and that, consequently, simple models that may be applic-able to such biological interactions are not suited for the complex


psychological realm. There is no reason, however, to believe that bio-logical phenomena should be simpler than psychological phenomena.The point of building dynamic systems is to reduce the complexityinherent in the phenomena under scrutiny and to distill the simplestpossible dynamic relationship that nevertheless specifies the essentialaspect of the ongoing process. Such reduction is difficult to accomplish,but it is no more difficult with psychological than with biologicalprocesses. However, if quantitative models are just too far out ofreach, it is always possible to confine oneself to applying the generalprinciples of dynamic systems thinking in a more qualitative way. Ishall discuss the qualitative approach in the next section.The second difficulty concerns the relationship between the math-

ematical model and the empirical data it attempts to model. If onemodels population sizes, the variable and its empirical correlate have atransparent relationship. A population is a countable set and 200 rab-bits is twice as much as 100 rabbits, for instance. Things become differ-ent with psychological variables where the numerical representation ofan underlying variable – for example the ‘‘level’’ of antisocial behavioror the coherence between specific facial motor patterns – amounts to ahighly idealized representation that bears only a relatively vague rela-tionship to its observable counterpart. If that is indeed the case, thereremains only an indirect way of fitting the model to the data. Forinstance, if the model specifies a regression or a cycle, we can check ifthe observable data, whatever their exact nature, show qualitative signsof a regression or cycle respectively. A comparable problemwith quan-titative models is that they can easily be viewed as mere metaphors,because there is no direct mapping between the components of themodel and certain aspects of the modeled reality. For instance, al-though a growth model applies transparently to the growth of a popu-lation or the growth of crops, it allegedly applies onlymetaphorically tothe growth of syntactic understanding or moral reasoning, to name justtwo examples (Van Geert, 1991). The point is that a problem like this –metaphor or model – cannot be answered in general. By their verynature, dynamic models are very abstract and general specifications ofrelationships. The question of whether or not they adequately apply toa specific empirical phenomenon should be answered by taking allrelevant aspects into account and not by drawing a line between adomain where such models apply and where they do not.

A qualitative approach

As said before, it often makes no sense to try to represent interestingaspects of reality by one-dimensional numerical variables that form the


centerpieces of dynamic mathematical models. The reason could bethat such parts of reality are simply too complex or involve relation-ships that are too numerous. Another reason could be that we have noidea whatsoever as to what kind of dynamic relationships hold be-tween the aspects or constituents of interest. If there are theoretical andempirical reasons to believe that the phenomenon under study canfruitfully be conceptualized as a dynamic system, it becomes interest-ing to try to check whether that phenomenon displays the propertiesthat are often identified with non-linear dynamic systems in general.Such properties are, among other things, self-organization in the firstplace, but also the existence of attractors, threshold phenomena, bifur-cations, chaos, discontinuities, self-similarity, and so forth (see forinstance Fogel, Nwokah, Dedo, and Messinger, 1992; Fogel, 1993;Lewis, 1995, 1996; Lewis and Granic, 1999b). They are properties thatmake particular sense in the context of processes of change and devel-opment (they are less interesting from a purely differential psychologi-cal point of view, for instance).The concept of self-organization provides an interesting illustration.

In principle, processes of whatever kind obey the general entropicprinciple. That is, the transmission of information or the creation oforder always leads to a decrease of order or to a loss of information.Manymodels in psychology are implicitly or explicitly built upon this –generally correct – assumption. Examples are the general theory ofinformation, but also theories like the Chomskyan view of languagedevelopment. The basic argument of the latter runs as follows. Thegrammatical structure of language is – obviously – something that achild acquires by being confrontedwith a specific language, the mothertongue. That is, the structure of language comes from the ‘‘outside’’ andis transferred to the ‘‘inside.’’ It can be shown, however, that theobservable language input is technically insufficient to get the structureof grammar across the outside/inside border, so to speak. That is,language input under-determines language structure. Thus, if gram-matical structure cannot be transferred from the outside to the inside,where does it come from? There is, according to the Chomskyan view,only one logical possibility left, namely that it is already present inside.That is, it must be an innate property. The input is needed to set certaingeneral innate parameters to the values adopted by the child’s mothertongue. From a dynamic systems point of view, the problem lies withthe phrase ‘‘only one logical possibility left.’’ It is true only underthe entropic assumption, which states that a process of informationtransmission can at most retain the order or information given and can,in principle, by no means increase the order or information containedin the signal (the input, say). However, non-linear dynamic systems


are characterized by the fact that they are order- and information-increasing systems. Hence, the statement that there is only one logicalpossibility left must be rejected. More precisely, there are two logicalpossibilities left, namely, first, that grammatical structure is alreadyinternally present and, second, that grammatical structure arises byself-organization, given the properties of the learner and the propertiesof the ambient language. Both are mere statements-of-possibility. Atpresent, we have no idea of how a structure such as grammar couldbecome posted inside the brain preceding any actual confrontationwith language. Similarly, we have no idea of how grammatical struc-ture can emerge by self-organization, given the properties of the learnerand the ambient language. From the viewpoint of conceptual likeli-hood, however, the self-organizational option constitutes the safestbet. The reason is that considerably more complicated forms of self-organization than the genesis of a grammar – for instance morphogen-esis – are known to occur in living systems. Another reason is that theinnateness assumption transfers the problem of the genesis of languagestructure from the ontogenetic to the phylogenetic domain. It remainsto be explained how the structure of grammar came to be established inthe human genome by some process that, by definition, must have beena process of self-organization, since there was no grammar aroundbefore it got materialized in the human brain along the path of humanevolution.In summary, the grammar example shows how the application

of a general property of dynamic systems – the likelihood of self-organization – throws a new and interesting light on old problems. Itgoes without saying that dynamic systems thinking has considerablymore to offer than that. The present book provides a number ofexamples of how this particular approach may be fruitfully appliedto a host of problems concerning the self, emotions, and humanrelationships.


COMMENTARY

Fish, foxes, identity, and emotionLinda A. Camras and George F. Michel

Paul Van Geert’s chapter provides a lucid explanation of the dynamicsystems approach that will be invaluable in orienting the reader to theother contributions in this volume. Dynamic systems theorywas devel-oped to explain the organization and maintenance of physical andbiological systems that appear to defy the second law of ther-modynamics (i.e. the law of entropy). The editors of this volume hopeto demonstrate that dynamic systems theory can provide a frameworkfor understanding one’s ‘‘identity’’ or ‘‘self’’ as a stable system thatmay yet change over time in a lawful manner. They propose that thissystem emerges from the interaction of component elements, includingcognition, emotion, and contextual experience.If the dynamic systems approach is to be a guiding framework for

studying the development of self and identity, then it should offernovel solutions to old problems or redefine what constitutes a problem.Only the experts who study these phenomena can determinewhether itdoes so.We would like to highlight certain issues raised by Van Geert’sdescription of the characteristics of dynamic systems that should bekept in mind during such evaluation.Van Geert begins by emphasizing that dynamic systems involve a set

of component elements. This seemingly obvious fact is actually quiteimportant because several features of dynamic systems might alsoseem applicable to phenomena conceived as unanalyzable wholes.Thus, one who views identity as a unitary phenomenon might observethat it develops in a discontinuous fashion and assumes a limited rangeof values within a particular time period. For example, categories ofidentity status (e.g. Marcia, 1980) might appear to be sets of ‘‘attractorstates’’ through which one moves as the result of a ‘‘bifurcation’’process. Van Geert believes that making such analogies has value andcan lead us to a fruitful reconceptualization of a phenomenon. In fact,such analogies are the cornerstone of the qualitative approach to usingdynamic systems theory described at the end of his chapter. However,by preceding this discussion with a clear explanation that the character-istic features of dynamic systems are the result of lawful interactions

among component elements, Van Geert implicitly reminds us thateventually the set of component elements in any dynamic systemmustbe accurately determined. Defining identity clearly (as urged by theeditors of this volume) and determining whether it is an emergentphenomenon whose components are emotion, cognition, and experi-ence or whether identity is itself a component of some broader systemof individual functioning that includes also cognition, emotion, andexperience is an important challenge. If we view the identity as emerg-ing from the pattern of interaction among emotion, cognition, andexperience, then this emergent phenomenon cannot itself affect thestates of its constituent components. Rather it is the interactions amongcomponents (i.e. emotion, cognition, and experience) that enhance andconstrain specific aspects of each other and, in doing so, give rise to theemergent phenomenon (i.e. ‘‘identity’’). At the same time, most apprai-sal theories of emotionwould consider one’s identity to be an importantinfluence on one’s emotional response to a stimulus event. One possiblesolution to this apparent theoretical incompatibilitymight be to make aconceptual distinction between identity and self-concept. That is, self-concept (i.e. one’s cognitive self-representation) might be consideredone component of a dynamical system of identity that also includesemotion, experience, and other aspects of cognition as additional com-ponents.One feature of dynamic systems that Van Geert strongly emphasizes

is self-organization. As illustrated by the differential equations describ-ing predator–prey interactions, the non-linear relationship amongelements is produced by forces within the system, i.e. the coupledinteraction of components. Thus a dynamic system is not just anysystem that changes over time. Rather it is an internally coherentsystem whose changes are primarily determined by internal (ratherthan external) forces. This feature might make the dynamic systemsapproach particularly appropriate to the study of identity since identityis indeed considered bymost psychologists to be an internally coherentphenomenon. However, if identity-dependent behavior varies with thesituational context, then selected aspects of environment and experi-ence must now be incorporated into the system. This may requireextending the concept of identity beyond its usual boundaries. As aresult, the complex phenomenon of identity-functioning might be suc-cessfully modeled as a self-organized system requiring no externalguidance plan controlling identity-related behavior.Currently, psychology is not experiencing a headlong rush toward

the adoption of dynamic systems approaches. One reason may be thatmany researchers are simply unacquainted with the concept of dy-namic systems. Thus, the current volume can play a seminal role in

90 Linda A. Camras and George F. Michel

encouraging more widespread adoption of this perspective. Van Geertexplicitly addresses two other reasons why dynamic systems perspec-tives have notmademuch headwaywithin the psychology community.First, many researchers feel that certain psychological phenomena aretoo complex to be adequately represented by a simple mathematicalmodel. Van Geert rightly asserts that biological phenomena are ascomplex as psychological phenomena. Nevertheless, with appropriatesimplifying assumptions, complex biological (and physical) phenom-ena can be productively modeled as dynamic systems. This is the ‘‘art’’of model building. Second, and less easy to dismiss, is the issue ofmeasurement. As Van Geert points out, many psychological phenom-ena (e.g. self-esteem) are measured ‘‘very indirectly’’ with only a vaguerelation to an observable counterpart. Therefore, it may be difficult toassign numerical values to states of emotion, cognition, and contextualexperience that are robust enough to permit the empirical testingneeded for the development and refinement of dynamic models of selfand identity. For example, there is no universal consensus regardingthe measurement of emotion. Although systems for identifying andquantifying the production of facial expression are well developed, therelationship between facial expression and emotion itself is subject todebate. Until such thorny problems of numerical representation can beresolved, Van Geert suggests taking a qualitative approach rather thanseeking to establish precise parameters for one’s model. However,this solution may be unsatisfactory to those for whom the very appealof the dynamic systems approach is its potential for mathematicalrepresentation.Possibly more common than the dissatisfaction engendered by

under-mathematizing is the dissatisfaction engendered by over-mathematizing. Many psychological theories are primarily verbal nar-ratives rather than formal mathematical models. In terms of Aristotle’sdoctrine of causation, such narratives traditionally have presented effi-cient (mechanistic) causes or final (teleological) causes for a phenom-enon rather than a formal cause (Bambrough, 1963). Formal equationsmay describe a phenomenon but, for many psychologists, they will notbe able to explain it. Using Van Geert’s classroom-talking example,some constant value (alpha) may be found to accurately represent themagnitude of increase in classroom talking, butmost psychologistswillbe more concerned with confirming or disconfirming the proposal thatsocial reinforcement is the (mechanistic) cause of increased talking.Thus itmay be important to emphasize that dynamic systemsmodelingdoes not exclude the possibility of explaining a phenomenon on a‘‘different level of analysis.’’ Within the field of biology, researchersmay describe phenomena such as morphogenesis as a self-organizing

91Commentary on Van Geert

dynamic system but they further seek to understand the mechanisticforces that produce the mutual influences among the elements of thesystem. Perhaps for psychologists, one important virtue of the dynamicsystems approach will be the guidance that such formal (or informal)modeling can provide in our search for non-obvious mechanistic ordevelopmental causes.In our own work (Camras, 1992, 2000; Camras, Lambrecht, and

Michel, 1996; Michel, Camras, and Sullivan, 1992), we are attempting touse the dynamic systems perspective to achieve a greater understand-ing of the relationship between facial expression and emotion.We cameto this perspective out of a sense of dissatisfaction with the approachesthat have dominated the field in recent years. The story of our workbears a resemblance to Van Geert’s example of the potential value thatthe dynamic systems perspective provides for the study of languagedevelopment.Many emotion theorists have viewed their target phenomenon as the

manifestation of an internal controller (i.e. a central emotion ‘‘pro-gram,’’ analogous to Chomsky’s genetically provided LanguageAcqui-sition Device). Such theories predict a strong linear coherence amongthe components of emotion (e.g. feelings, facial expressions, neuro-physiological substrates). In contrast, we were struck by the fact thatstereotypic emotional facial expressions do not always seem to occurduring episodes of emotion and, in fact, they may sometimes occurwhen the predicted corresponding emotion is unlikely to take place.For example, focusing narrowly on the phenomenon of infant facialexpression, we have empirically demonstrated that non-emotion fac-tors (i.e. head and gaze movements) may systematically recruit specificfacialmovements resulting in the production of an ‘‘emotional’’ expres-sion. Yet such ‘‘inappropriate’’ responses do not produce confusion inthe emotion communication process and, in fact, go largely unnoticedduring social interactions. We have sought to account for this by recon-ceptualizing emotion as a ‘‘softly assembled’’ system in which relation-ships among components are orderly but non-linear.Currently, we are attempting to delineate more specifically those

circumstances under which the various components of emotion aremanifested. This step must precede any attempt to produce a precisemathematicalmodel of emotion responding. Because emotion is indeeda complex psychobiological phenomenon, we are far from achievingour ultimate goals. Nonetheless, we feel that adopting a dynamicsystems perspective has offered us a means of escape from earlierparadigms that do not adequately capture the real phenomenon ofemotion.


COMMENTARY

Fish, foxes, identity, and emotionLinda A. Camras and George F. Michel

Paul Van Geert’s chapter provides a lucid explanation of the dynamicsystems approach that will be invaluable in orienting the reader to theother contributions in this volume. Dynamic systems theorywas devel-oped to explain the organization and maintenance of physical andbiological systems that appear to defy the second law of ther-modynamics (i.e. the law of entropy). The editors of this volume hopeto demonstrate that dynamic systems theory can provide a frameworkfor understanding one’s ‘‘identity’’ or ‘‘self’’ as a stable system thatmay yet change over time in a lawful manner. They propose that thissystem emerges from the interaction of component elements, includingcognition, emotion, and contextual experience.If the dynamic systems approach is to be a guiding framework for

studying the development of self and identity, then it should offernovel solutions to old problems or redefine what constitutes a problem.Only the experts who study these phenomena can determinewhether itdoes so.We would like to highlight certain issues raised by Van Geert’sdescription of the characteristics of dynamic systems that should bekept in mind during such evaluation.Van Geert begins by emphasizing that dynamic systems involve a set

of component elements. This seemingly obvious fact is actually quiteimportant because several features of dynamic systems might alsoseem applicable to phenomena conceived as unanalyzable wholes.Thus, one who views identity as a unitary phenomenon might observethat it develops in a discontinuous fashion and assumes a limited rangeof values within a particular time period. For example, categories ofidentity status (e.g. Marcia, 1980) might appear to be sets of ‘‘attractorstates’’ through which one moves as the result of a ‘‘bifurcation’’process. Van Geert believes that making such analogies has value andcan lead us to a fruitful reconceptualization of a phenomenon. In fact,such analogies are the cornerstone of the qualitative approach to usingdynamic systems theory described at the end of his chapter. However,by preceding this discussion with a clear explanation that the character-istic features of dynamic systems are the result of lawful interactions

among component elements, Van Geert implicitly reminds us thateventually the set of component elements in any dynamic systemmustbe accurately determined. Defining identity clearly (as urged by theeditors of this volume) and determining whether it is an emergentphenomenon whose components are emotion, cognition, and experi-ence or whether identity is itself a component of some broader systemof individual functioning that includes also cognition, emotion, andexperience is an important challenge. If we view the identity as emerg-ing from the pattern of interaction among emotion, cognition, andexperience, then this emergent phenomenon cannot itself affect thestates of its constituent components. Rather it is the interactions amongcomponents (i.e. emotion, cognition, and experience) that enhance andconstrain specific aspects of each other and, in doing so, give rise to theemergent phenomenon (i.e. ‘‘identity’’). At the same time, most apprai-sal theories of emotionwould consider one’s identity to be an importantinfluence on one’s emotional response to a stimulus event. One possiblesolution to this apparent theoretical incompatibilitymight be to make aconceptual distinction between identity and self-concept. That is, self-concept (i.e. one’s cognitive self-representation) might be consideredone component of a dynamical system of identity that also includesemotion, experience, and other aspects of cognition as additional com-ponents.One feature of dynamic systems that Van Geert strongly emphasizes

is self-organization. As illustrated by the differential equations describ-ing predator–prey interactions, the non-linear relationship amongelements is produced by forces within the system, i.e. the coupledinteraction of components. Thus a dynamic system is not just anysystem that changes over time. Rather it is an internally coherentsystem whose changes are primarily determined by internal (ratherthan external) forces. This feature might make the dynamic systemsapproach particularly appropriate to the study of identity since identityis indeed considered bymost psychologists to be an internally coherentphenomenon. However, if identity-dependent behavior varies with thesituational context, then selected aspects of environment and experi-ence must now be incorporated into the system. This may requireextending the concept of identity beyond its usual boundaries. As aresult, the complex phenomenon of identity-functioning might be suc-cessfully modeled as a self-organized system requiring no externalguidance plan controlling identity-related behavior.Currently, psychology is not experiencing a headlong rush toward

the adoption of dynamic systems approaches. One reason may be thatmany researchers are simply unacquainted with the concept of dy-namic systems. Thus, the current volume can play a seminal role in


encouraging more widespread adoption of this perspective. Van Geertexplicitly addresses two other reasons why dynamic systems perspec-tives have notmademuch headwaywithin the psychology community.First, many researchers feel that certain psychological phenomena aretoo complex to be adequately represented by a simple mathematicalmodel. Van Geert rightly asserts that biological phenomena are ascomplex as psychological phenomena. Nevertheless, with appropriatesimplifying assumptions, complex biological (and physical) phenom-ena can be productively modeled as dynamic systems. This is the ‘‘art’’of model building. Second, and less easy to dismiss, is the issue ofmeasurement. As Van Geert points out, many psychological phenom-ena (e.g. self-esteem) are measured ‘‘very indirectly’’ with only a vaguerelation to an observable counterpart. Therefore, it may be difficult toassign numerical values to states of emotion, cognition, and contextualexperience that are robust enough to permit the empirical testingneeded for the development and refinement of dynamic models of selfand identity. For example, there is no universal consensus regardingthe measurement of emotion. Although systems for identifying andquantifying the production of facial expression are well developed, therelationship between facial expression and emotion itself is subject todebate. Until such thorny problems of numerical representation can beresolved, Van Geert suggests taking a qualitative approach rather thanseeking to establish precise parameters for one’s model. However,this solution may be unsatisfactory to those for whom the very appealof the dynamic systems approach is its potential for mathematicalrepresentation.Possibly more common than the dissatisfaction engendered by

under-mathematizing is the dissatisfaction engendered by over-mathematizing. Many psychological theories are primarily verbal nar-ratives rather than formal mathematical models. In terms of Aristotle’sdoctrine of causation, such narratives traditionally have presented effi-cient (mechanistic) causes or final (teleological) causes for a phenom-enon rather than a formal cause (Bambrough, 1963). Formal equationsmay describe a phenomenon but, for many psychologists, they will notbe able to explain it. Using Van Geert’s classroom-talking example,some constant value (alpha) may be found to accurately represent themagnitude of increase in classroom talking, butmost psychologistswillbe more concerned with confirming or disconfirming the proposal thatsocial reinforcement is the (mechanistic) cause of increased talking.Thus itmay be important to emphasize that dynamic systemsmodelingdoes not exclude the possibility of explaining a phenomenon on a‘‘different level of analysis.’’ Within the field of biology, researchersmay describe phenomena such as morphogenesis as a self-organizing

91Commentary on Van Geert

dynamic system but they further seek to understand the mechanisticforces that produce the mutual influences among the elements of thesystem. Perhaps for psychologists, one important virtue of the dynamicsystems approach will be the guidance that such formal (or informal)modeling can provide in our search for non-obvious mechanistic ordevelopmental causes.In our own work (Camras, 1992, 2000; Camras, Lambrecht, and

Michel, 1996; Michel, Camras, and Sullivan, 1992), we are attempting touse the dynamic systems perspective to achieve a greater understand-ing of the relationship between facial expression and emotion.We cameto this perspective out of a sense of dissatisfaction with the approachesthat have dominated the field in recent years. The story of our workbears a resemblance to Van Geert’s example of the potential value thatthe dynamic systems perspective provides for the study of languagedevelopment.Many emotion theorists have viewed their target phenomenon as the

manifestation of an internal controller (i.e. a central emotion ‘‘pro-gram,’’ analogous to Chomsky’s genetically provided LanguageAcqui-sition Device). Such theories predict a strong linear coherence amongthe components of emotion (e.g. feelings, facial expressions, neuro-physiological substrates). In contrast, we were struck by the fact thatstereotypic emotional facial expressions do not always seem to occurduring episodes of emotion and, in fact, they may sometimes occurwhen the predicted corresponding emotion is unlikely to take place.For example, focusing narrowly on the phenomenon of infant facialexpression, we have empirically demonstrated that non-emotion fac-tors (i.e. head and gaze movements) may systematically recruit specificfacialmovements resulting in the production of an ‘‘emotional’’ expres-sion. Yet such ‘‘inappropriate’’ responses do not produce confusion inthe emotion communication process and, in fact, go largely unnoticedduring social interactions. We have sought to account for this by recon-ceptualizing emotion as a ‘‘softly assembled’’ system in which relation-ships among components are orderly but non-linear.Currently, we are attempting to delineate more specifically those

circumstances under which the various components of emotion aremanifested. This step must precede any attempt to produce a precisemathematicalmodel of emotion responding. Because emotion is indeeda complex psychobiological phenomenon, we are far from achievingour ultimate goals. Nonetheless, we feel that adopting a dynamicsystems perspective has offered us a means of escape from earlierparadigms that do not adequately capture the real phenomenon ofemotion.


CHAPTER 5

A relational perspective on thedevelopment of self and emotionAlan Fogel1

Begin with two premises.First, psychological experience always implies a connection, a rela-

tionship: with another person, with cultural tools or language, or withthe natural environment. Life is a network of relationships.Second, psychological experience is always dynamic and changing.

The simplest visual perception requires a change, either in a movementof the object or a movement of the eyes, head, or body. Thoughts andfeelings fluctuate in a continuous pattern of change. These patterns ofchange themselves change as people develop. Life is a series of changes.On the other hand, part of psychological experience is a sense of one’s

uniqueness (the self) and a sense of one’s permanence through time(identity). How can this occur? How can people have a sense of them-selves and their stability over time if psychological experience is funda-mentally relational and dynamic?The answer proposed in this chapter is that people experience the

changes in their relationships according to different types of emotionand that emotions provide information about the self. Consistent withthe two premises, emotions are conceptualized as dynamic experiencesof harm or benefit, perceived as personally meaningful with respect tothe individual’s changing relationship with the environment (Barrett,1993; Barrett and Campos, 1987; Fischer, Shaver, and Carnochan, 1990;Frijda, 1986; Lazarus, 1991b). According to Frijda, emotional experience‘‘is glued, as it were, to its object, coinciding entirelywith apprehendingthat object’s nature and significance . . . [negative] emotional experienceis perception of horrible objects, insupportable people, oppressiveevents’’ (Frijda, 1986: 188). And DeRivera suggests that ‘‘emotionsmaybe conceived as existing between people, as various sorts of attractionsand repulsions . . . which transform their bodies and perceptions’’ (DeRivera, 1992: 200). As we shall see, this perspective suggests that thereare an unlimited variety of emotional experiences corresponding to thelimitless variety of relationship dynamics.From the perspective of this chapter, emotion is one way of

93

discovering the meaning of a relationship for the self and, hence, theunique position of the self in the relationship. Over long periods oftime, the perception of consistency in one’s emotional experience be-gins to yield a sense of permanence of the self through time. In thissense, perception and emotion are two aspects of the same process ofdiscovering the way in which the self is related to the world (Dewey,1934; Stern, 1985).

The relational perspective on self and emotion

Before developing this perspective, a fewwordsmay be said aboutwhyit is different from traditional models of self and emotion. If one beginswith a different premise, that each person is a unique individual fromthe outset, then the existence of a sense of self does not need to beexplained. The self-contained individual, existing as a totality indepen-dent of the surround, is one of the main features of Western cultureinherited from Greek philosophy (Levinas, 1969). In contrast to theview presented here, of emotion as the perceived meaning of relation-ships, emotions in the individual perspective are conceived as basic‘‘internal’’ states or motivations, generated from the neurophysiologyin response to an ‘‘external’’ cause. In this perspective, individuals‘‘have’’ an emotion that requires expression, via action, with respect tothe environment. The individual perspective predicts that there are alimited number of discretely different emotions that are geneticallyprogrammed into the neural system and become available to defineharms and benefits for the individual.The relational perspective, by contrast, sees individuals as open and

changing components in systems of relationships. Cooperative com-munication, for example, brings people together to create an outcomethat no single person could achieve. Consider choral singing. Individ-ual singers have different parts and vocal ranges that, when mixed,form a coherent aural aesthetic. The concept of coregulation describesthis type of social process that is jointly created (Fogel, 1993). Whencoregulating, individuals’ behavior in the group is not theirs alone, butthe manifestation of the group’s dynamics in each of their bodies. Inthe same way, emotion is experienced as harms and benefits withinindividual bodies but it reflects the body’s experience of a relationaldynamic. According to Gergen (1991: 157), for example,

If it is not individual ‘‘I’’s who create relationships, but relationshipsthat create the sense of ‘‘I,’’ then ‘‘I’’ ceases to be the center of successor failure, the one who is evaluated well or poorly, and so on. Rather,‘‘I’’ am just an I by virtue of playing a particular part in a relationship.

The relational perspective does not deny that people often perceive

94 Alan Fogel

Table 5.1. Time scales and self experience. It takes only seconds to becomeaware of some aspects of the self and years to become aware of other aspects ofthe self.

Topic Event Frame Development

Time scale Seconds Minutes, hours YearsSense of self Orientation Authorship/agency IdentityType of experience Orientational

emotionsNarrative emotions Reference

emotions

their part in a communicative process as ‘‘their own’’ contribution or‘‘their own’’ failure to make an effective contribution. On the contrary,these attributions are one of the undeniable manifestations of the self,the sense of one’s own uniqueness. Each singer in the chorus manifeststhe group dynamics in a unique way, experiencing his or her own partin the larger musical creation. The relational perspective described inthis chapter does, however, seek a principled explanation of the senseof self. How, in fact, does the self arise in spite of the obvious fact ofecology: that everyone and everything is merely an incomplete locationin a network of relationships comprising the cultural and physicalworld?There are different ways in which the sense of uniqueness can be

experienced and they differ according to the time scale in which theindividual is a participant in a relationship. At each time scale, adifferent aspect of self arises with respect to a different type of emo-tional process. In this chapter, three time scales are discussed. They aresummarized in table 5.1 and introduced briefly here, followed by amore detailed description in the remainder of the chapter.The first time scale is microseconds and seconds. Even during these

very brief periods of time, individuals can have a sense of their orienta-tion with respect to others corresponding to emotions related to ap-proach or avoidance. The second time scale is minutes, hours, or longer.This is the time it takes for orientations to form into a sequential patternof communication. During this time scale, individuals can have a senseof their unique role in the authorship of the pattern. Authorship is asense of one’s agency and occurs with regard to emotions such assecurity versus insecurity or togetherness versus loneliness. The thirdtime scale occurs over years. This is the period of time it takes forindividuals to have a sense of the uniquely enduring aspects of them-selves, their identity. These experiences typically occur with respect toemotions such as harmony or conflict, satisfaction or dissatisfactionwith oneself over time. At any given moment in time during the lifecourse, the individual may experience the self and its related emotions

95A relational perspective on self and emotion

at all three time scales. These ideas are developed in the remainder ofthis chapter.

Orientation

This section discusses the emergence of the most basic form of self, thesense of a unique orientation with respect to a relationship. In order tocoregulate, people orient their actions toward the other within a periodof microseconds and seconds. Orientation is not a fully formed actionsystem with a predefined goal but orientation may become more goal-like during the process of coregulation.Orientation is similar to the concept of intention defined as ‘‘behav-

ioral object directedness’’ (Merleau-Ponty, 1962; Vedeler, 1991). In thiscase, an object can be a physical object or a person. Intention implies anorientation movement of the body relative to someone or somethingrather than toward some fully formed goal. Because there is always anorientation, it means that there is always an intention. Trevarthen’s(1993: 123) concept of ‘‘motive . . . a readiness for perceiving informa-tion needed for acting’’ is similar. Orientation is also similar to theconcept of ‘‘dialogical position’’ of actual or imagined selves (Hermansand Kempen, 1993), one’s stance vis-a-vis another. Another relatedconcept is ‘‘action readiness,’’ a propensity to move toward or awayaccompanied by an emotional experience of approach or avoidancetoward a particular relationship (Frijda, 1986). Frijda’s position is theonly one that makes clear that orientations are connected to particularforms of emotional experience. Since there is always an orientation,there is always some kind of emotional experience.Orientations can be relatively open or relatively closed. Postural

orientations, for example, may differ with respect to whether stance,gaze direction, and limb position is open or closed vis-a-vis a partner.Mental orientations differ with respect to the willingness to embrace orreject the consideration of concepts or ideas. More open orientationsallow a freedom in how the process unfolds and permit creativitywithin the system. More closed orientations reflect degrees of controlover how the process unfolds. The dimension of open–closedhighlightsthe emotional aspect of orientation. Open orientations correspond toemotional attraction or a feeling of wanting to approach the other.Closed orientations correspond to feelings of repulsion or withdrawal.The sense of uniqueness, the self, arises in part because orientations

imply the location of the self with respect to the environment. Orienta-tions may be directed from the individual toward the environment orthey may be receptive, from the environment toward the individual(Vedeler, 1994). When orientations are more directed, individuals have

96 Alan Fogel

the experience of acting on the environment with varying degrees ofcontrol. When orientations are more receptive, one has the experienceof being influenced by the environment or of ‘‘abandoning ourselves tothe world’’ (Vedeler, 1994: 346). There is typically an alternation be-tween being directed or receptive.

Hardness and softness, roughness and smoothness, moonlight andsunlight, present themselves in our recollection not preeminently assensory contexts but as certain kinds of symbioses, certain ways theoutsideworld has of invading us and certainwayswe have ofmeetingthat invasion. (Merleau-Ponty, 1968: 194)

Thus, although orientations are inherently relational, implying a rela-tionship between self and other, individuals can perceive their positionvis-a-vis others. Since the individual occupies a location in physical andpsychological space, the body is perceived as fundamentally in anorientation because all movement is from one location and towardanother, either directed away from the self or toward the self. Allpsychological processes, including thought and narration, have anorientation because of their fundamental embodiment (Hermans andKempen, 1993; Lakoff and Johnson, 1999; Yasuo, 1987).The self’s uniqueness, then, is experienced from its particular per-

spective at one pole of a relational orientation (Gibson, 1966; 1979;Michaels andCarello, 1981; Reed, 1987). Recent discoveries suggest thatthis type of self-awareness can be directly perceived at an early age.Human infants, for example, can identify the movements of their ownbodies with respect to the environment and other persons, the so-called‘‘ecological self’’ that differs from the mirror recognition of self at 18months (Butterworth, 1995; Fogel, 1995; Legerstee, Anderson, andSchaffer, 1998; Rochat, 1995; Rochat and Hespos, 1997; Stern, 1985).The dynamics of coregulation and the continual shifts of orientation

also play a role in establishing a particular emotional quality to thesense of a unique self. Because to coregulate one must continuallyreadjust one’s actions based on the continuously changing actions ofthe partner, individuals can experience their relative degree of creativityas a participant in a relationship (Fogel, 1993; Ganguly, 1976; Pickering,1999; Whitehead, 1978). Because the self is inherently creative, theexperience of uniqueness is not an experience of being but an experi-ence of becoming, a process of improvisation during communication(Bakhtin, 1981; Barclay, 1994; Boesch, 1991; Bosma, 1995; Hermans andKempen, 1993; Jansz, 1995; Josephs, 1998; Shotter, 1981). It is thoughtthat the earliest awareness of self in human infancy, called the emergentself, is just this experience of being a unique participant in the creativeprocess (Stern, 1985).


A similar perspective is offered by philosophies of intersubjectivity(Buber, 1958; Jopling, 1993; Levinas, 1969; De Quincey, 1998) in whichself and other arise as a result of communication and dialogue. Duringthe give and take of ordinary discourse, each person’s statements arecountered by the other so that mutual reformulation (coregulation) ofmeaning is required. As speakers make clearer their point of view tolisteners, that point of view becomes clearer to the speakers themselves(Hermans and Kempen, 1993).The sense of self that arises from these creative processes, however, is

most likely to be experienced just following the creative act or creativemoment. It is necessary for the experience to achieve a resolution. Afterhaving peak experiences of creative flow, as in mountain climbing ormusical performance, adults report feeling a sense of uniqueness thatdid not occur during the experience itself (Csikszentmihalyi, 1990;Csikszentmihalyi and Rathunde, 1998). Dewey (1934) has pointed outthat there is a crucial difference between experiencing (being in the flowof activity) and having an experience. The latter implies a sense of bothparticipation and completeness from which a characteristic emotionmay be felt as well as a sense of self. This leads us to consider the issueof the completion of orientations into events.

Events

At particular times during the process of communication, the creativeflow becomes punctuated. There is a change from experiencing as flowto having an experience. An experience is a sense that the flow hascoalesced, a noticeable pattern has formed, a pattern that will be calledan event. The sense of one’s uniqueness comes into being as a concreteevent that emerges from the orientational flow (Fogel and Branco,1997; Kegan, 1982; Pedrosa, Carvalho, and Imperio-Hamburger, 1997;Sampson, 1989; Whitehead, 1978).Events are the psychological experiences of real or actual entities such as

objects, actions, feelings, or thoughts. When we experience something asreal, we are participating in a process by which orientations havecoalesced into an awareness of events, a process that may take onlymicroseconds (Whitehead, 1978). The emergence of events is the per-ception of an island of stability against the background of the dynamicflux of change. From a dynamic systems theory perspective, events areself-organizing processes that owe their stability to the mutual interac-tions among the elements in the relationship and not to the existence ofstored, fully formed representations inside the individual (Fogel andThelen, 1987; Thelen and Smith, 1994).As orientations coalesce into events, there is the simultaneous emergence

98 Alan Fogel

of a stable emotional experience of value vis-a-vis the relationship betweenself and environment (Hermans and Kempen, 1993). All events areexperienced as inherently emotional (Dewey, 1934). These orientationalemotions yield the experience of values such as goodor bad, approach oravoid, accept or reject, tasteful or distasteful, pleasure or pain. Value isthe most basic form of awareness, occurring on a time scale of secondsand microseconds. Value is also the earliest form of awareness inhuman development. Newborn infants have sophisticated approachand withdrawal reflexes by which they can evaluate a wide range ofsensory experiences. Newborns suck sweet liquids and spit out bitterones; they turn toward voices with particular pitch and intensity levelsand not others; they follow the movements of high-contrast visualobjects and not others (Fogel, 2001). These experiences of value arecenters of psychological consistency, called the emergent self (Stern,1985), bywhich the self’s unique relationship to the environment can beestablished.To the extent that orientations are open, rather than closed, they are

perceived with some sense of uncertainty. When engaged in an openconversation, for example, there are flow periods in which the out-comes are unknown and participants are willing to remain in a relative-ly uncertain cocreative process. Events, although they appear whenuncertainty in orientations is reduced over time, always have someorientational aspect that leads them from present to future. Although itis theoretically useful to distinguish orientations from events, in prac-tice there are neither pure orientations nor pure events: orientationscoalesce into events but events-as-orientations set the stage for the nextevents (James, 1976; Merleau-Ponty, 1962; Whitehead, 1978). Events,therefore, can be more or less orientational, referring to the possibilityfor the event to become transformed in the process of coregulation. Inscience, for example, the ideal situation is that concepts are workingmodels. The scientist is supposed to maintain an orientational stancetoward the data using those concepts but the concepts are always opento modification. The values that are perceived as part of an event arerarely finalized. They are always open to a process of revaluation.

Becoming and being: balancing orientation and event

Orientations and events are two related aspects of awareness of self.The former is connected with the awareness of becoming and the latterwith the awareness of being, the formerwith change and the latter withstability. Within individual experience, there are moments when beingand becoming are present in a fruitful and self-sustaining balance. Oneremains open to change while at the same time one has a sense of


Table 5.2. The spectrum of self-awareness, from being to becoming

Primarily being Balance of being and becoming Primarily becoming

Events are stable Events are orientational Events do not coalesceOrientations are minimal Orientations flow into events Orientations are salientThe self is isolated The self is co-regulated The self is lost or merged

stability and uniqueness. In an open conversation, for example, theparticipants begin with a frame of mind, a loosely defined set oforientations that partake of their individuality. Via coregulation, eachperson opens to change but that change is integrated into their priororientations, thus preserving their individuality and at the same timechanging it. In this case, there is an intimate and directly perceivedconnection between self and other that has the special quality ofbeing cocreative. One is aware of the self, one is aware that the other is aself, and one is aware that the emergence of those selves (as events)depends upon the cocreative process (as mutual orientations) and theself-events are always orientational. Cobeing arises from and flows intocobecoming.Buddhist thinking refers to the concept of dependent coarising, which

refers to the awareness of the creative process by which cobecoming isbalanced with cobeing. In this process, the individuals become awarethat their attempts to ground themselves in the permanence of beingare set against the background of the continuous living flow of becom-ing (Pickering, 1999; Varela, Thompson, and Rosch, 1991;Wilber, 1979).When individuals become fully open to each other during interper-sonal encounters, the experience of orientation toward each other isboth indeterministic and endless, the experience of ‘‘infinity’’ (Levinas,1969). In these meetings, the other person becomes the self and vice-versa, there is no sense of being directed or receptive, only a feeling ofselfhood dissolving into union. One sees the same humanity in theother and meets the other without reserve and in the fullness of theother’s vulnerability to be changed. These have been called‘‘I–Thou’’ as opposed to ‘‘I–It’’ relationships (Buber, 1958). This balancebetween being and becoming is only one part of the spectrum ofself-awareness. One can shift the balance either more toward becomingor more toward being (see table 5.2).One pole of self-awareness is that of pure becoming,when experience is

completely orientational. One is oriented toward another person or thingwithout an end or resolution. There is a merger with that thing,the experience ‘‘of eternity, the complete absorption in being’’(Loewald, 1980: 141). Because the sense of self arises only as orienta-

100 Alan Fogel

tions coalesce into events, purely orientational awareness cannotinvolve a sense of self. In Frijda’s (1985) view, for example, emotionalexperience has this character. When you attack someone or somethingin rage, for example, there is a complete absorption in that thing orperson (as when a person says, ‘‘I completely lost myself,’’ or ‘‘I didn’tknowwhat I was doing’’). At the time ofmaking such statements, one isreflecting on having had an experience of anger. During the experience,however, there is no sense of I or you. In relationships, partners fail toprovide each other with closure and confirmation for their actions.When infant self-initiated acts are not facilitated by the family system tohelp the infant appreciate what acts lead to what consequences, forexample, then the inner experience will not be felt as the infant’s own(Sander, 1962).The other pole of self-awareness is pure being, when events lose their

orientational character. There is no change because no matter what theorientation is, the result is the same event. These are ‘‘experiences offragmentation’’ such that ‘‘each instant loses its relation to any otherinstant and stands by itself’’ (Loewald, 1980: 143). One feels estrange-ment, depression, helplessness, depersonalization, and alienationwhen faced with unchanging circumstances, stubborn people, irrecon-cilable differences, chronic tensions or anxieties, or persistent negativeemotions. In these cases, the self fills one’s awareness with a kind ofheaviness. Thoughts and emotions are self-focused and attempts toorient away from the self typically flounder in a sense of self-negationor self-inflation. Being lost or losing all of one’s freedoms may lead tothis experience of an absence of self.Periods of being or becoming may last only seconds or minutes, as

when the resolution of a communication process takes longer thanexpected, ormuch longer periods as in chronic depression. The orienta-tional experience of self, therefore, is not a static, stored, or structuralthing. It is an appearance and disappearance, a coming and going. Inconversation, for example, one goes between a selfless merger with theother, an awareness of a cocreated self experienced in the effort todialogue, and a heightened focus on one’s own needs at the expense ofthe other. In athletics and othermovement activities one alternates fromconcentration to flow, from an awareness of the self–other relationshipand its implications for action and emotion, to selfless immersion in themoment, or to excessive pride or shame at winning or losing. Ideally,then, a sense of a unique self occurs when dynamically unfoldingorientations coalesce into stable events, each having a particular emo-tional quality and a particular connection between self and other. In thesame way that orientations self-organize dynamically into events dur-ing microseconds and seconds, observation suggests that events are


perceived to self-organize into sequential patterns of events overminutes, hours, or days. These sequential patterns of events in com-munication systems are called frames. The sequential coherence ofevents and its connection to the self is the topic of the next section.

Authorship

The psychological coherence of patterns of event sequences is capturedby the concept of frame. Frames are sequences of events that have a coherenttheme, that take place in a specific location, and that involve particular forms ofmutual coorientation between participants. Frames have a temporal organ-ization – a beginning, middle, and end – in which the events in theframe cohere (Bateson, 1955; Fogel, 1993; Goffman, 1974; Jones, 1990;Kendon, 1985). Examples of frames are greetings, topics of conversa-tion, conflicts, or children’s social games. Frames are coherent patternsthat result from self-organizing processes.Because events are partly orientational; the events in a sequence can

change in order to establish a psychological relationship with eachother. In this way, all the events in the frame are coconstructed with allthe others, an example of self-organization. In the sameway that eventscoalesce and become stable from the dynamics of orientations, framescoalesce as coherent patterns out of the dynamics of event sequences.Because of these psychological stabilities in the flowof action over time,individuals have the opportunity to perceive a stable sense of self inrelation to the frame.Frames are similar to narratives, the latter referring primarily to

coherent sequences of verbal communication. Narratives have a stabletheme that emerges from the self-organization of the events; they havean orientational direction of flow over time that motivates the move-ments of the actors and events (Ginsburg, 1985; Hermans and Kempen,1993; Haviland and Kahlbaugh, 1993; Jones, 1990; Ricoeur, 1983). Singleevents are created in a period of microseconds and seconds, enoughtime for one to experience an approach or withdrawal orientation.Narratives, on the other hand, are assembled over longer periods oftime, over minutes or hours or even longer periods. At this time scalethe individual can perceive new stable aspects of the self, beyondsimple orientation and emotional value.As the individual becomes aware of the narrative/frame as a whole

unit (e.g. a particular form of play or conflict in a dyad or group), thenarrative is perceived as an event in itself. The same considerations thatapply to the prior discussion of orientations and events can apply tonarratives. The perception of the narrative as whole event can occur inmicroseconds or seconds, while at the same time the individual can also

102 Alan Fogel

experience the actual unfolding of the narrative over minutes, hours ordays. In this way, the entire history and future of frames that extend forlong periods of time can come into existence as an event in the present(Fogel, 1993; Merleau-Ponty, 1962).Some theorists (Stern, 1985; Tomkins, 1962) suggest that a new type

of emotional experience, different from that of orientational emotions,arises over the time scale of narratives and that these narrative emotionspartly account for the perceived coherence of narratives. In the creationof a narrative emotion, the specific values of orientational emotionsfrom each event in the frame do not simply sum together. Rather, theemerging coherence of the events in the frame is partly accounted for bythe emerging coherence of a narrative emotional process that is per-ceived to link events meaningfully together through time.Observational research suggests that each narrative emotion forms

into smoothly unfolding temporal contours, patterns of feeling experi-ence (Tomkins, 1962) that have been called ‘‘protonarrative envelopes’’(Stern, 1985; 1993). These contours create ‘‘feeling qualities best cap-tured by such kinetic terms as ‘crescendo,’ ‘decrescendo,’ ‘fading,’‘exploding,’ ‘bursting,’ ‘elongated,’ ‘fleeting,’ ‘pulsing,’ ‘wavering,’‘effortful,’ ‘easy,’ and so on’’ (Stern, 1993: 206). The temporal contour ofa crescendo, for example, is a relatively slow build-up of emotion,reaching a peak and then declining rapidly.Each narrative emotion also has a meaning for the self, the personal

harms or benefits of the relational connections in the frame (Frijda,1986; Stern, 1993; Tomkins, 1962). These are experiences that connectevents over time in some meaningful way for the self. Narrative emo-tions include, therefore, experiences that take some time to unfold andthat establish a personal relational meaning such as love versus hate,respect versus disrespect, security versus insecurity, togetherness ver-sus loneliness, safety versus threat, pride versus shame. Security canonly be felt in reference to the consistency of interpersonal contact overthe course of a frame. Security has a different temporal contour thanother narrative emotions, a steadiness that accompanies feelings ofenjoyment or comfort.Narrative emotions self-organize over time through a process that magnifies

or minifies the psychological meaning of particular events by focusing ordiffusing the experience of initial relational orientations over a sequenceof events (Demos, 1982; Fogel et al., 1997; Lewis, 1995; Lewis andFerrari, this volume; Sarbin, 1986a; Tomkins, 1962; 1978). Consider thefollowing example. Walking down a lonely street at night one hears, orthinks one hears, a sound coming from behind. One’s orientation (at-tention and posture) toward the direction of movement becomeschanged to having more urgency and one’s orientation toward what is


behind comes into existence for the first time as wariness. What hap-pens next depends upon the specific way in which emotion magnifiesor minifies this orientation. Perhaps one turns around and notices a cat,which minifies the potential fear and magnifies the orientation to con-tinue at the same pace, lowering the potential to perceive threat. Orperhaps one turns to see a person. The subsequent process dependsupon whether that person is a man or woman, whether he or she isfollowing or merely walking in the same direction, whether there arepeople ahead, or street lights.The experience of participation in frames, therefore, inevitably high-

lights the self in a particular manner. Is there anything the personcan do to protect him- or herself from harm? To calm him- or herself?To take charge of the situation? The time scale of narratives allows the selfto be aware of a sense of its own agency. I will follow Day and Tappan’s(1996) use of Bakhtin’s (1981) concept of authorship to describe thatsense of agency. Authorship refers specifically to the sense of self thatcan only be experienced as a coagent in the construction of a socialnarrative, a story about the self meant to be communicated to anotherperson. Authorship, however, can vary from being to becoming, asshown in table 5.2. There are other similar terms for agency, such asself-efficacy and perceived control, but they fail to specify explicitlytheir links to the relational dynamics. The purpose of this chapter is toshow how particular features of self-experience can be linked to theemerging stabilities present in communication processes at differenttime scales.Authorship can also be experienced in non-verbal frames in parent–

infant communication (Fogel, 1993; Stern, 1985). Stern (1985) refers tothe sense of self that involves an awareness of agency and narrativeemotions as the core self, which begins around the thirdmonth of life. Asan example, my colleagues and I made weekly videotapes of mother–infant communication for the first year of life as mother and infantplayed with toy objects. The examples below came from one weeklysession of one of the dyads at three months. In this frame, the infant islying on her back on the floor and her mother is sitting beside herholding a toy rattle. The infant is capable of putting her hand in hermouth yet does not have the ability to reach for objects. In the sequencedescribed below, however, there is a transfer of the object from themother’s to the infant’s hand. How does this occur?

The infant is looking at the rattle with her hand in her mouth, asmother moves the rattle in the infant’s line of vision but out of theinfant’s reach. The infant’s hand comes out of her mouth and movestoward the mother. At the same time, the mother begins to move theobject toward the infant’s hand. A period of ten seconds ensues in

104 Alan Fogel

which the mother adjusts the orientation of the rattle while the infantopens her palm and watches. The mother then places the rattle in theinfant’s hand. The infant grasps it and then brings the hand, contain-ing the rattle, back to her mouth.

In this frame, the infant is oriented toward the self (hand-in-mouth) andat the same time toward the object held by the mother. The infant’sacquisition of the object, with themother’s assistance, is then integratedinto the same pattern of self-directed activity as the handwith the objectis returned to the infant’s mouth. The infant’s sense of authorship ingetting the object to her mouth arises as her orientations form intoevents and events form into a frame during a communication with themother. It would be difficult to imagine any sense of authorship for theinfant outside of a frame or narrative construction. One must have atleast aminimal cause–effect sequence for the experience of self as agent.Another instance of the same frame occurs a few moments later, afterthe infant drops the rattle.

The mother moves the rattle in the infant’s line of vision and out ofreach as the infant looks at the rattle. This time, however, her hand isnot in her mouth but is held out in front of her. The mother continuestomove the rattle and the infant looks off to the side and puts her handin her mouth. The mother calls her name and she turns to look at themother as her hand comes out of her mouth. At this point, the mothertries to insert the rattle into the infant’s palm but the infant looks awayonce again. The infant turns to look at the rattle and the mother triesagain but the infant closes her palm, displays a distressed expression,and turns to the side, fussing.

By contrasting these two examples, we can see the variability in theemergence of authorship between instances of the frame. In the firstexample, the emotion – attentive interest in the object – is orientedsimultaneously toward self and toward the object. The continuation ofthat emotion depends, first of all, upon the changing communicativeevents within the frame, specifically, upon the mother and infant’scoregulated mutual adjustment of their hands. The continuation of theemotion of attentive interest also appears to depend upon the connec-tion of that coregulated communication with the infant’s orientationtoward the self. The object transfer event depends, in otherwords, uponthe incorporation of the object into the infant’s self-orientation (movingthe hand from themouth, to get the object, and back to themouth). Notealso that the temporal contour of the emotion during the successfulobject transfer appears to be a steady focus of attention on the coordina-tion of action. The infant is more likely to have a sense of self as authorof the narrative when the communication is coregulated, the orienta-tions become events, and the events address the emotional concerns ofthe self over the course of the frame (see table 5.2).


In the second example, the infant does not appear to have a coherentsense of authorship. This lack of a sense of authorship/agency is con-nected to the distinction, made earlier, about whether orientations aredirected or receptive,whether one perceives a sense of control by actingtoward the environment or whether the environment is perceived asacting on the individual. When orientations are open, the result is acoregulated balance between directed and receptive stances, a dialoguebetween self and other.When receptiveness occurs in this context, it canbe perceived as a part of the sense of authorship. In the second example,however, the experience of receptiveness is not pleasant for the infantandmost likely is accompanied by a loss of a sense of perceived controlover the sequence of events. The unsuccessful transfer also has a differ-ent temporal contour of emotional experience, beginning with fluctu-ation rather than steady attention (wavering) and ending with a rapidbuild-up (crescendo) of distress.In the following example, also frommy research, there is a balance of

being directed and receptive and a corresponding sense of authorship.Once infants are able to reach and grasp objects (about five months),they are reluctant to release them. It takes another three or four monthsfor infants to learn how to give as well as to receive objects. Thefollowing example comes from the same dataset of weekly videotaperecordings. It reports on an observation of a different mother–infantdyadwhen the infantwas ninemonths (see Fogel, 1993, formore detailsabout this example). The example is the first weekly session in whichwe observed the infant voluntarily release the object into the mother’shand.

Mother and infant are seated across from each other at a child-sizedtable. The infant has a toy fork in his hand and holds it out toward hismother. As he moves the fork toward his mother, she begins to moveher hand toward his hand. She opens her hand in a palm up gesturejust underneath the infant’s hand. The infant orients his hand as if toplace the fork into the mother’s but does not release the fork. He looksintently at his hand and at the fork and gradually begins to open hishand. The tines of the fork catch on the mother’s open palm as sheslowly begins to pull her hand back toward herself, but withoutclosing her palm. At the same time and at the same rate, the infant’shand continues to open. At some point, the fork falls into the mother’sopen hand and she grasps it and brings it up to her chest, smiling. Asshe does this, the infant watches the movement of the fork away fromhim while his hand remains extended outward and open. Finally, theinfant gazes at his mother’s face and begins to smile.

In this third example, like the first example, mother and infant com-municative actions were coregulated into a gently increasing temporalcontour. The mother’s movements of her hand toward her body, for

106 Alan Fogel

example, were coordinated with the infant’s opening of his palm. Theinfant’s orientation toward concerns of the self was seen over the courseof the frame in the intensity of the gaze at his own actions with theobject and the care with which he released the object into the mother’shand. The infant’s emotional experience of joy at the accomplishment ofthe object transfer, reflecting a sense of authorship within the frame,emerged from this intensity of interest in combination with thecoregulated communication with the mother.This third example also shows the first instance of a new develop-

mental achievement: the infant’s voluntary transfer of an object to themother. In everyday communication, orientations serve the partici-pants by providing an initial stance from which communication be-comes possible and a relative openness to emergent and creative pro-cesses that allow communication to be spontaneous and to address theconcerns of the participating selves. We can see from this example thatthese same properties of orientation additionally allow for the emerg-ence of a developmental change, the emergence of novel events andframes that have never before appeared in the communication system.In the weeks following the novel object transfer event in the lastexample, a new frame of give-and-take appeared, allowing the emerg-ence of new orientations that did not previously exist in this dyad andnewways of experiencing the authorship of the self-in-relation (the selfwho can give and receive with another). These new forms of mutualactivity also brought new emotional and self experiences with them.Development is conceptualized as a change in the process by which a system’sconstituents change each other to create a newly emergent frame, a change inthe process of change. As frames change developmentally, so do thecorresponding experiences of self and emotion.

Identity

It is possible for the individual to develop a sense of oneself overdevelopmental time. Because any particular instance of framing in-volves an experience of self, and because framing involves the reenact-ment of similar patterns that have occurred in the past, there is thepossibility during frames for the individual to perceive the similaritiesbetween the present experience of self and the past experiences of self.Identity is the experience that the self endures through time and that this selfthrough time has preferential orientations that enter into particular forms ofcreativity, particular propensities for change and stability, and predispositionsfor particular patterns of framing (Bosma, 1995; Erikson, 1968; Hermansand Kempen, 1993; Lewis, 1995).Previous authors used the term identity to mean the particular


manifestation of self through time that begins during adolescence andinvolves a growing commitment to career, gender role, or religion.Without negating that use of the term, one can see more general instan-ces of between-frame comparisons that begin as early as the end of thefirst year of life. During Piaget’s third substage of sensorimotor devel-opment, infants begin to compare and to integrate past and presentschemes. During the fourth sensorimotor substage, infants begin to useone frame as a means for another. It is at this age, for example, thatinfants begin to tease their parents, showing that they can take a pastframe and transform it into a means to create something emotionallydifferent in the present (Reddy, 1991). These developments indicatethat infants develop a sense of their own agency and an understandingof the causal links between current actions and past actions. This is alsothe age at which secondary intersubjectivity emerges, the ability to referto a different frame outside the context of the current frame, as inpointing to an object in order to bring it into the discourse betweenmother and infant (Trevarthen and Hubley, 1978). According to Stern(1985), this is the period of emergence of the shared self.The coherence of the past and present in relation to the self as a

whole, however, does not develop until the age of four years. After thisage, children can consistently remember past experiences of themselvesand weave them into a narrative account about themselves. ‘‘Thechild’s representational system can begin to ‘temporalize’ what werepreviously successive and unrelated states of the self into an organized,coherent autobiographical self-concept’’ (Provinelli and Simon, 1998:189). This has been called the ‘‘proper self’’ (Provinelli and Simon,1998), the ‘‘extended self’’ (Neisser, 1991), the ‘‘narrative self’’ (Stern,1985), and the ‘‘autobiographical self’’ (Nelson, 1992).What is common across all these forms of self-awareness is the sense

of an identity between frames. It could be the current frame in relationto the past (e.g. I am the self who typically is shy in social situations).Alternatively, it may be the current frame in relation to cultural frames(e.g. I am the self who is similar to a shy character in my family orcommunity, or in literature, TV, or film). Identity is also about one’srelationship to cultural definitions of identity (I am the self who canmore readily remain active in old age because there are more culturalmodels for successful aging).Identity is experienced with respect to reference emotions involving the

relative success of making these comparisons or of meeting the standardsimplied by the comparisons. These emotions include satisfaction versusdissatisfaction, harmony versus conflict, achievement versus failure,frustration versus elation, and approval versus disapproval. One canaccept one’s self as shy in social situations, for example, or feel dissatis-

108 Alan Fogel

fied with one’s self because of it. One can feel in harmony with culturalstandards or in conflict. The self-comparison emotions can be observedin infants beginning at the end of the first year and the comparisonswith others or with cultural standards (e.g. gender roles) emerge by thethird year, both inside and outside the family (Emde and Oppenheim,1995; Kitayama et al., 1995; Reddy, 1991).Although the topic of identity as typically construed brings up

the issue of a cognitive representation of the self through time, I amsuggesting that identity is an emotional experience, not cognition orrepresentation. Following along the theme of this chapter, identity isthe emotional perception of the coherence of a relationship over adevelopmental time scale. Identity, like orientation and authorship,arises in one’s experience of being a participant in a communicativerelationship.Ideally, identity is created and recreated dynamically, always partly

orientational, a balance between being and becoming. On the otherhand, identity can become lost at the pole of becoming or centered onthe self in the form of a personality disorder (cf. table 5.2 and Lewisand Ferrari, this volume). On the side of becoming, identity can losecoherence, especially during periods of developmental change (Dunne,1996). Individuals may experience periods during which they have losttheir identity, as during transitions in career, marital status, or genderrole (Erikson, 1968).Ricoeur (1996), for example, defines ‘‘narrative identity’’ as the per-

sonal stories that endure against the background of life changes, adefinition that could apply at any age. Identity emerges in the retellingof something about the self in the past. This is because, as shown earlier,the self is defined in the process of communication. In autobiographicalwriting, authors do not merely recount the past but are ‘‘deciding whatto make of the past narratively at the moment of telling’’ (Bruner, 1990:122). Identity also arises when one selects existing cultural narratives asframes in which to tell one’s own story. An individual may frame his orher identity in terms of cultural narratives for gender, speaking asfemales from the perspective of stories of communion or as males fromthe perspective of stories of independence (Gergen and Gergen, 1995;Miller et al., 1992; Mistry, 1993).The sense of identity can be perceived as an event in real time, as one

reflects upon oneself across developmental time. Identity then, as anevent, is partially orientational over a developmental time scale, be-coming stabilized into an event over a period of years. In thesecoregulated dynamics, self perspectives become magnified andminified in relation to others’ narratives about the self and theresulting emotions (see Bosma, 1995; Haviland and Kahlbaugh, 1993;


and Lewis, 1995; for more detailed descriptions of identity/personalitydevelopment).One can move this discussion to the cultural level. The types of

identity narrative in a culture also develop over historical time, pre-sumably by the same types of change dynamics discussed earlier.During the nineteenth century in Western Europe and North America,gender identity was defined with respect to the differences betweenmen and women. In the mid-twentieth century, androgyny becamemore salient as a theme in gender narratives. In the latter decades ofthe twentieth century, gender narratives involve the self-constructionof one’s own gender identity and sexual orientation (Gergen and Ger-gen, 1995). One of the features of the end of the twentieth century israpid cultural change in identity narratives.When cultural change is thebackdrop for individual and relationship changes, the indeterminism of self-development is increased and places increasing demands on individuals tomaintain an open orientation to change at all levels in order to allow identity todevelop freely (Cole, 1996; Erikson, 1968; Gergen, 1991).

Time and the self

The experience of self, then, can occur with respect to different timescales: events, frames, and development, as shown in table 5.1. One canexpand the ideas in table 5.1 by including longer time scales. Thedevelopmental time scale, mentioned above, refers to the changes thatoccur every few years. The individual can become aware of cumulativechanges, leading to increasingly complex senses of self and emotion.The particular form of identity that begins in adolescence seems torequire a time scale of fifteen years or more before the individualbecomes aware of the frame comparisons related to the self’s rolecommitments. This may be due in part to the fact that one needs thelengthy perspective on self and also in order to integrate the complexarray of identity frames of the culture into a view of the self. Accordingto Erikson’s (1950) theory of psychosocial development, each stage ofthe life course brings with it a newly emergent sense of self, involvingdifferent types of self experiences and different kinds of emotions.Preschool children, for example, take on an identity in the eyes ofothers, feeling pride and shame (emotions related to their view of howothers perceive them) for the first time. As young adults make choicesabout marriage or career, they acquire sociocultural identities and feelemotions of commitment or alienation. The older adult becomes awareof an identity of a person who dies, of the uniqueness of his or her lifecourse, and experiences emotions such as wisdom and detachment. Aspeople near endings, especially those at the end of life, emotional

110 Alan Fogel

Table 5.3. Summary of different types of emotion according to type of changeprocess

Type of change process Emotional experience

Orientation to event Orientational emotions: approach vs.withdrawal, directed vs. receptive, open vs.closed

Event to frame Narrative emotions: love vs. hate, respect vs.disrespect, security vs. insecurity,togetherness vs. loneliness, safety vs. threat,pride vs. shame

Frame to frame comparison Reference emotions: harmony vs. conflict,satisfaction vs. dissatisfaction, approval vs.disapproval, achievement vs. failure, elationvs. frustration

All changes Change emotions: certainty vs. uncertainty,stability vs. instability, determinism vs.indeterminism, order vs. chaos

awareness becomes heightened aswell as a corresponding appreciationfor one’s unique life course (Carstensen, Isaacowitz, and Charles, 1999).In addition to the time scale of life-course duration, as described in

the last paragraph, change itself has an emotional meaning for the self,at all time scales (Fogel et al., 1997).

Personal histories are processes of change in time, and the change itselfis one of the things immediately experienced. ( James, 1976: 25)

A critical dimension in defining and describing emotional experience,therefore, focuses on the concept of changing states. (Stein, Trabasso,and Liwag, 1993: 281)

The emotions of change are certainty versus uncertainty, stability versusinstability, determinism versus indeterminism, order versus chaos.Note thatthese emotions reflect the types of experience along the pole of beingand becoming (see table 5.2). Order emerges on the being side whilechaos may be felt on the becoming side. These emotions of change, aswell as experiences that appear at different time scales of emergence,are summarized in table 5.3. The emotions of change refer to the relativechangeability of the other emotions. During transitions in romanticrelationships, for example, the relational emotionsmay fluctuate rapid-ly, between security and insecurity or between joy and despair. Thesefluctuations create the change emotions of uncertainty and chaos. It isby these emotions of change during transitions in togetherness orintimacy that one recognizes the importance of the relationship for theself.


The maintenance of creativity and orientational flexibility in the faceof these uncertainties reflects a balance of being and becoming and isthought to facilitate optimal developmental change (Antonofsky andSagy, 1990; Block and Block, 1980; Csikszentmihalyi, 1990; Csikszent-mihalyi and Rathunde, 1998; Isen, 1990; Malatesta and Wilson, 1988).Change, therefore, can be experienced either as highly creative or asdistressing and disorganizing, depending upon how individuals aresupported in their important relationships while in the process ofchanging. When most creatively balanced, individuals may seek change,magnifying their emotions of fear or success, uncertainty or certainty, in orderto continue acting near the edge of chaos. At other times, people may findchangemore aversive, keeping their emotionswithin prescribed boundaries andmaintaining the stability of previously ‘‘safe’’ frames for communication andemotion when faced with perturbations that are likely to lead to change(Antonofsky and Sagy, 1990; Butz, Duran, and Tong, 1995; Fogel andLyra, 1997; Lewis, 1995; Rogers, 1961; Stern, 1985).Both positive and negative emotions, then, may be important for

navigating periods of change because all emotions provide informa-tion about the self’s concerns in the relationship. The idea thatemotions can be usefully magnified in the service of change runscounter to traditional emotion narratives from Western culture. Inthese narratives, emotions are irrational impulses, originating frominside the individual, driving us to act in involuntary ways. Tradi-tional Western narratives for mature identity, on the contrary, reflectthemes of rationality, self-determination, and responsibility. ‘‘Emo-tions, with their alleged irrationality, would seem to seriously under-mine this cultural ideal of personhood’’ (Fischer and Jansz, 1995: 61).In adopting these traditional narratives, we devalue the opportunitiesof change, times when events become more orientational and emo-tions more intense and unpredictable (Butz et al., 1995).There is a diverse variety of therapeutic methods – ranging from

psychotherapy, to body and movement awareness, to meditation andspiritual practices – involving the enhancement of flexibility and cre-ativity in the face of pain or changing, unpredictable circumstances(Antonovsky, 1993; Rogers, 1961; Wilber, 1979). The goal is finding abalance, which in the perspective of this chapter is optimal creativityand optimal stability, a balance between being and becoming, and aheightened awareness of being a coparticipant in a relational process.These approaches can help individuals develop an alternative culturalidentity with narratives for connection between persons in caring rela-tionships with others and with the environment and for trusting theinherent value of emotional experience to guide one through the cur-rents of chaos and change.

112 Alan Fogel

Conclusions

In this chapter, I have presented a view of communication, self, andemotion based on the primacy of relationships and their change dynamics.Each time scale allows the individual to perceive different aspects of theself. It is as if these different avenues of self-awareness are the naturalconsequence of the time it takes for events on different time scales toemerge in relationships and for the psyche to detect them. In addition,single events, authorship events, and identity events can all be per-ceived in the present and at the same time. Single events can change aframe more readily than they can change a culture or an identity but,theoretically, every single event is in some way part of the events thatchange at all levels. One of the issues left unexplored in this chapter is adetailed accounting of the coregulation that occurs across these differ-ent levels.Also unexplored in this chapter is the relationship of the authorship/

identity system to events that emerge over longer time scales. Theseinclude historical, phylogenetic, geological, and cosmological timescales. The way in which events are constituted from orientationsdepends in part upon the sensorimotor, emotional, and cognitive sys-tems that evolved on this planet. Themain narratives of culture dependas much on interpersonal relationships and their changes as on rela-tionships with the earth and its changes.Another unexplored topic is the historical pathways by which orien-

tations and events enter into frames and by which frames enter intoidentities. Each specific single event or frame in a relationship has ahistory. In some cases, events and frames form and later disappear,with no perceptible effects on the system. In other cases, particularframes and events become amplified into seeds for developmental andhistorical change (Fogel and Lyra, 1997; Lyra andWinegar, 1997). Theserelational–historical dynamics are part of a more complete understand-ing of communication systems and the self.More can be said about the role of the body in self-awareness. The

sense of self and identity emerges from relational dynamics because thebody is specialized to create emotional experiences for the temporalcontours and relationalmeanings of eventswithin frames. And recipro-cally, the particular features of orientations, events, frames, and cul-tures owe their identity to the possibilities of emotion as experienced bythe human body.Research methods were not discussed in detail. In general, research

that follows from the theoretical principles in this chapter must befocused on the process of change over time by which orientationscoalesce into events. This requires a sequential analytic approach as


illustrated in the narratives taken from my research. It is essential totrack how a psychological process unfolds over time – at all time scales– in the context of coregulated communication between the person andthe environment. It is important to understandwhat kinds of sequencespromote open compared to closed orientations and howdevelopmentalcreativity can be facilitated in relationships.Writing this chapter was an opportunity to author my own theoreti-

cal identity on the problem of how the self is created in the midst ofrelational changes, to become aware of my position in the scholarlycommunity, and to realize where I am incomplete. I was comforted bythe work of many others who have authored narratives of relationshipand the indeterministic creativity of emotion and self. The individualis-tic narrative forms of the scientific culture are changing as they coregu-late with narratives of dynamic systems and relationships. In this fluidculture of change, the challenge is to create a relatively stable narrativeidentity granting autonomy within connection, accepting self-lossagainst the backgroundof self-emergence, preserving individuality as apart of the human relational ecology.

Note

1 This workwas funded in part by a grant to the author from the United StatesNational Institute of Mental Health (MH48680 and MH57669). I am gratefulto the following individuals for their comments on this chapter: Kari Apple-gate, Trevor Burnsed, Jacqueline Fogel, J’lene George, Ilse de Koeijer, TomMalloy, Andrea Pantoja, Cory Secrist, and Dankert Vedeler.

114 Alan Fogel

COMMENTARY

The personal experience ofcoherenceJeroen Jansz

Introduction

As soon as we conceptualize ‘‘the self’’ as a dynamic structure that isrooted in communicative relations, we bump into the question of howto account for the continuity and consistency most people experienceacross time and situations. Fogel addresses this perennial question atthe very beginning of his chapter when he asks how people can have asense of ‘‘stability over time if psychological experience is fundamen-tally relational and dynamic?’’ (p. 93). His answer focuses on the role ofemotions: people perceive consistency in their emotional experiencesand this contributes to a sense of stability of the self through time. Inthis commentary I will first criticize Fogel’s proposal with respect toemotions, and then propose two alternative candidates for sustainingthe sense of stability. The first is individual embodiment. Fogel alreadytouches upon the importance of embodiment, but I will attribute amorefundamental role to embodied being than he does. My second candi-date for sustaining stability is autobiographical memory. The recollec-tions of people’s personal past are organized in such a way that theygenerate consistency and continuity in normal individuals.

Emotional experiences are context bound and subject to evaluation

In this section I will take issue with Fogel’s argument of the perceivedconsistency in emotional experiences. According to me, it is difficult toexperience consistency in emotions because the emotion process is asmuch subject to the dynamisms of communicative interaction as selvesare. Emotions unfold in a particular communicative context, and arealways evaluated with respect to their appropriateness. This generallyleads to adjustments in the emotional experience. I will first brieflydiscuss the inextricable link between emotions and evaluation, andthen sketch the standards that are used in evaluation. Emotions are inmost cases judged in terms of the social context (or the culture at large),

115

but individuals may also employ a more personal standard across timeand situations.The importance of evaluation with respect to emotion stands out at

the beginning of the emotion process: an emotion unfolds when anevent or a situation is appraised as significant for the (personal) inter-ests of the individual (Frijda, 1986). This kind of evaluation is anintuitive and swift cognitive operation, which proceeds almost auto-matically. Once the emotion is experienced, it is under constant scru-tiny by the person who has the feeling. It generally results in decisionsabout what to do about the emotion and the event that caused it(Lazarus, 1991b). In the same time span, the appropriateness of theemotion is evaluated. Someone who experiences an emotion will judgewhether the emotion conforms with or transgresses the feeling rules inthat context (Hochschild, 1983). These rules specify the normativerelationship between the situation that evokes the emotion (and inwhich it is experienced and expressed) and the emotion itself. Anemotion is only seen as appropriate if it emerges in the right context, atthe right time, and in response to the right antecedent. As a matter ofconsequence, emotions are strongly regulated by the possibilities andconstraints in particular social situations.Other standards of evaluation are employed in addition to the feeling

rules. In this volume about identity and emotions the standards that arederived from the individual’s identity are of particular interest. Thecultural conception of the person is the source model of self andidentity (Harre, 1983).Western personhood can be characterized brieflyby an emphasis on rationality, autonomy, and responsibility (Fischerand Jansz, 1995). These building-blocks of identity contrast sharplywith the characteristics thatWestern culture attributes to emotions. It iswidely assumed that emotions are impulsive bodily forces that over-whelm individuals all of a sudden and result in irrational and irrespon-sible behavior (Fischer and Jansz, 1995;White, 1993). As a consequence,individuals who experience an emotion will be well aware of thepotential threat it implies for their status of being. In order to be able toestablish the actual damage emotionality has done (or will do) to theirrespectability, people evaluate the emotion in that particular context.Their evaluation is based on the relevant feeling rules, but it is enrichedwith evaluative standards that are derived from their actual or aspiredidentity. In our own research we focused on situations in which indi-viduals judge a particular emotion as being at odds with their identity.Teachers at Dutch high schools were interviewed about their (profes-sional) identity and about the emotions they experienced in class. Theprofessional identity of a teacher requires, among other things, copingwith the particularities of pupils’ behavior. What pupils do (or fail to

116 Jeroen Jansz

do)may, however, incite emotions that conflict with the aspired profes-sional identity. This is what happens, for example, when the poorperformance of a ‘‘dull’’ child causes irritation in a teacher who isstronglymotivated to support disadvantagedpupils. For him, irritationis not the kind of reaction a supportive teacher should have. In ourinterviews with the teachers we found that this kind of a collisionbetween a particular emotion and the teacher’s identity produces a newfeeling, that is, a negative affect of a special kind. We have coined theterm emotional dissonance to describe this particular feeling of uneasethat motivates people to reduce the unpleasant dissonance (Jansz andTimmers, in press).The experiential effects of the conflict between identity and a particu-

lar emotion underline the point that both are dynamic processes thatare subject to change. On this Fogel and I agree. But ourways partwhenFogel says that ‘‘identity is an emotional experience, not cognition orrepresentation . . . identity is the emotional perception of the coherenceof a relationship over a developmental timescale’’ (p. 109). I would saythat identity is not a subclass of emotion because identity and emotionunfold at different levels: in judging the appropriateness of emotions,identity is used as a standard. In addition, I underline the representa-tional nature of knowledge about oneself. The ‘‘facts and figures’’ aboutoneself as a person are by all means represented in that part of thecognitive system that we tend to call memory (see below). If we mustlook for consistency in the sense of self we can therefore better focus onthe stability implied in the construction of identity rather than onemotional experiences. Consistency and continuity are, after all, im-portant aspects of the notion of identity (Bosma, 1995). I will now linkthe sense of stability to our embodied being as singular entities.

Embodied being

The fact that people experience themselves as (rather) stable and per-manent over time has a solid physical basis: their bodies and embodiedexperiences do not change very quickly over time. It is important tonote that the body always functions as a point of location. In Fogel’swords:

Since the individual occupies a location in physical and psychologicalspace, the body is perceived as fundamentally in an orientation be-cause all movement is from one location and toward another. (p. 97)

In social interaction the physical location of the body comes to func-tion as a point of reference. A newborn baby is addressed as if he or sheis a person long before the child can make any self-reference in its

117Commentary on Fogel

words and actions. Parents, carers, family-members and others choosethe physical location of the child’s body to address the newborn in thesecond person singular. In due course, all kinds of characteristics, traits,and capacities are attributed to the embodied child. It takes years beforechildren have the capacity to refer to themselves and their bodies interms of ‘‘I.’’ In this sense the ‘‘you’’ is older than the ‘‘I’’ (Shotter, 1989).The phenomenological continuity of the body that is present from an

early age onwards is enriched with the continuity that is taken forgranted in social interaction (Harre, 1998). A particular identity isascribed to a person in social interaction, and this identity is physicallylocated in the individual’s body. This underlines the point that personsconstruct their identities in social interaction, and both actors andaudiences locate the identity in the physical body. Furthermore, theidentity is kept ‘‘in’’ mind once it has been constructed. These personalmemories are the subject matter of the next section.

Autobiographical consistency

In living their lives, people accumulate a huge amount of information intheir memory. The capacity of people to recollect their lives is called‘‘autobiographical memory’’ (Baddeley, 1992). The actual autobio-graphical memories are of a varied nature. Most are episodic, becausethey are concerned with events that gained significance in the individ-ual’s past. Other autobiographical memories are semantic. They areconcerned with factual knowledge of personal significance. Next toepisodes and semantics, people keep other kinds of memories in storelike, for example, feelings, thoughts, movements, and, of course, odorsand tastes. The present argument about personal stability draws atten-tion to three important aspects of autobiographical memory. First, itsnarrative nature; second, the role of imagery; and third, the accuracy ofautobiographical remembering (cf. Rubin, 1996).The narrative nature of autobiographicalmemory stands out inmany

research studies as well as in ordinary experience (Barclay, 1993). Assoon as people start recalling their personal past, they generate story-like sequences with a beginning, a middle, and an end. The narrativestructure of autobiographical memories brings a sequential order intoinformation that might have been endlessly varied and chaotic at themoment of storage. Schank and Abelson (1995) claim that recall isnarrative as a result of the narrative representation of autobiographicalmemories in the cognitive system. Other researchers have shown thatthis claim is too radical: personal recollections are also representedin non-narrative ways (Brewer, 1996). This brings us to the secondcomponent.

118 Jeroen Jansz

Reports of autobiographical memories frequently involve vivid im-ages of the personal past. Imagery is in particular related to the sense of‘‘reliving’’ an episode that is often experienced during autobiographi-cal recall (Brewer, 1996). In recall, the imagery becomes part of anarrative structure as soon as people try to communicate the visualrepresentation of the earlier episode.When they tell a story about a pastexperience they translate the image into a narrative language that willdominate their recollections. The fact that images are traded for narra-tive accounts is probably due to the fact that people are accustomed tomaking sense of their experiences by employing stories (Baumeisterand Newman, 1994).The third aspect of autobiographical memories is their alleged accu-

racy. Reports of personal recollections show that people have moreconfidence in the accuracy of their autobiographical memories than inmemories that are not related to the self (Rubin, 1996). The high per-ceived accuracy of autobiographicalmemories is generally explained interms of the active reconstruction that takes place in the memoryprocess. When people recollect their personal past they establish anarrative order in their lives whatever the actual course of their liveshas been. Thus, ‘‘whereas our lives may not be coherent, our storiesare’’ (Schank and Abelson, 1995: 34). The final result is that autobio-graphical remembering preserves ‘‘a sense of being a coherent personover time’’ (Barclay, 1996: 99).In conclusion, the argument I submit here holds that coherence in the

sense of self is not preserved by consistency in emotional experiencesbut rather by three other constants in individual experiences. The firstone is the immediate and constant experience of being embodied singu-larly. This is linked to the second constant: the body is the point atwhich to locate the individual in social reference. This underlines thepoint that consistency and continuity are to a large extent structured bypractical patterns of interaction. The narrative nature of autobiographi-cal recollections is the third, and final, contribution to the coherencepeople say they experience in their sense of self. It is worthwhile toinvestigate the links between identity and autobiographical memory indetail. There are strong indications that a focus onmemorywill help usto gain a more precise understanding of the relation between self,identity, and emotion (Christianson and Safer, 1996).


COMMENTARY

The personal experience ofcoherenceJeroen Jansz

Introduction

As soon as we conceptualize ‘‘the self’’ as a dynamic structure that isrooted in communicative relations, we bump into the question of howto account for the continuity and consistency most people experienceacross time and situations. Fogel addresses this perennial question atthe very beginning of his chapter when he asks how people can have asense of ‘‘stability over time if psychological experience is fundamen-tally relational and dynamic?’’ (p. 93). His answer focuses on the role ofemotions: people perceive consistency in their emotional experiencesand this contributes to a sense of stability of the self through time. Inthis commentary I will first criticize Fogel’s proposal with respect toemotions, and then propose two alternative candidates for sustainingthe sense of stability. The first is individual embodiment. Fogel alreadytouches upon the importance of embodiment, but I will attribute amorefundamental role to embodied being than he does. My second candi-date for sustaining stability is autobiographical memory. The recollec-tions of people’s personal past are organized in such a way that theygenerate consistency and continuity in normal individuals.

Emotional experiences are context bound and subject to evaluation

In this section I will take issue with Fogel’s argument of the perceivedconsistency in emotional experiences. According to me, it is difficult toexperience consistency in emotions because the emotion process is asmuch subject to the dynamisms of communicative interaction as selvesare. Emotions unfold in a particular communicative context, and arealways evaluated with respect to their appropriateness. This generallyleads to adjustments in the emotional experience. I will first brieflydiscuss the inextricable link between emotions and evaluation, andthen sketch the standards that are used in evaluation. Emotions are inmost cases judged in terms of the social context (or the culture at large),

115

but individuals may also employ a more personal standard across timeand situations.The importance of evaluation with respect to emotion stands out at

the beginning of the emotion process: an emotion unfolds when anevent or a situation is appraised as significant for the (personal) inter-ests of the individual (Frijda, 1986). This kind of evaluation is anintuitive and swift cognitive operation, which proceeds almost auto-matically. Once the emotion is experienced, it is under constant scru-tiny by the person who has the feeling. It generally results in decisionsabout what to do about the emotion and the event that caused it(Lazarus, 1991b). In the same time span, the appropriateness of theemotion is evaluated. Someone who experiences an emotion will judgewhether the emotion conforms with or transgresses the feeling rules inthat context (Hochschild, 1983). These rules specify the normativerelationship between the situation that evokes the emotion (and inwhich it is experienced and expressed) and the emotion itself. Anemotion is only seen as appropriate if it emerges in the right context, atthe right time, and in response to the right antecedent. As a matter ofconsequence, emotions are strongly regulated by the possibilities andconstraints in particular social situations.Other standards of evaluation are employed in addition to the feeling

rules. In this volume about identity and emotions the standards that arederived from the individual’s identity are of particular interest. Thecultural conception of the person is the source model of self andidentity (Harre, 1983).Western personhood can be characterized brieflyby an emphasis on rationality, autonomy, and responsibility (Fischerand Jansz, 1995). These building-blocks of identity contrast sharplywith the characteristics thatWestern culture attributes to emotions. It iswidely assumed that emotions are impulsive bodily forces that over-whelm individuals all of a sudden and result in irrational and irrespon-sible behavior (Fischer and Jansz, 1995;White, 1993). As a consequence,individuals who experience an emotion will be well aware of thepotential threat it implies for their status of being. In order to be able toestablish the actual damage emotionality has done (or will do) to theirrespectability, people evaluate the emotion in that particular context.Their evaluation is based on the relevant feeling rules, but it is enrichedwith evaluative standards that are derived from their actual or aspiredidentity. In our own research we focused on situations in which indi-viduals judge a particular emotion as being at odds with their identity.Teachers at Dutch high schools were interviewed about their (profes-sional) identity and about the emotions they experienced in class. Theprofessional identity of a teacher requires, among other things, copingwith the particularities of pupils’ behavior. What pupils do (or fail to

116 Jeroen Jansz

do)may, however, incite emotions that conflict with the aspired profes-sional identity. This is what happens, for example, when the poorperformance of a ‘‘dull’’ child causes irritation in a teacher who isstronglymotivated to support disadvantagedpupils. For him, irritationis not the kind of reaction a supportive teacher should have. In ourinterviews with the teachers we found that this kind of a collisionbetween a particular emotion and the teacher’s identity produces a newfeeling, that is, a negative affect of a special kind. We have coined theterm emotional dissonance to describe this particular feeling of uneasethat motivates people to reduce the unpleasant dissonance (Jansz andTimmers, in press).The experiential effects of the conflict between identity and a particu-

lar emotion underline the point that both are dynamic processes thatare subject to change. On this Fogel and I agree. But ourways partwhenFogel says that ‘‘identity is an emotional experience, not cognition orrepresentation . . . identity is the emotional perception of the coherenceof a relationship over a developmental timescale’’ (p. 109). I would saythat identity is not a subclass of emotion because identity and emotionunfold at different levels: in judging the appropriateness of emotions,identity is used as a standard. In addition, I underline the representa-tional nature of knowledge about oneself. The ‘‘facts and figures’’ aboutoneself as a person are by all means represented in that part of thecognitive system that we tend to call memory (see below). If we mustlook for consistency in the sense of self we can therefore better focus onthe stability implied in the construction of identity rather than onemotional experiences. Consistency and continuity are, after all, im-portant aspects of the notion of identity (Bosma, 1995). I will now linkthe sense of stability to our embodied being as singular entities.

Embodied being

The fact that people experience themselves as (rather) stable and per-manent over time has a solid physical basis: their bodies and embodiedexperiences do not change very quickly over time. It is important tonote that the body always functions as a point of location. In Fogel’swords:

Since the individual occupies a location in physical and psychologicalspace, the body is perceived as fundamentally in an orientation be-cause all movement is from one location and toward another. (p. 97)

In social interaction the physical location of the body comes to func-tion as a point of reference. A newborn baby is addressed as if he or sheis a person long before the child can make any self-reference in its


words and actions. Parents, carers, family-members and others choosethe physical location of the child’s body to address the newborn in thesecond person singular. In due course, all kinds of characteristics, traits,and capacities are attributed to the embodied child. It takes years beforechildren have the capacity to refer to themselves and their bodies interms of ‘‘I.’’ In this sense the ‘‘you’’ is older than the ‘‘I’’ (Shotter, 1989).The phenomenological continuity of the body that is present from an

early age onwards is enriched with the continuity that is taken forgranted in social interaction (Harre, 1998). A particular identity isascribed to a person in social interaction, and this identity is physicallylocated in the individual’s body. This underlines the point that personsconstruct their identities in social interaction, and both actors andaudiences locate the identity in the physical body. Furthermore, theidentity is kept ‘‘in’’ mind once it has been constructed. These personalmemories are the subject matter of the next section.

Autobiographical consistency

In living their lives, people accumulate a huge amount of information intheir memory. The capacity of people to recollect their lives is called‘‘autobiographical memory’’ (Baddeley, 1992). The actual autobio-graphical memories are of a varied nature. Most are episodic, becausethey are concerned with events that gained significance in the individ-ual’s past. Other autobiographical memories are semantic. They areconcerned with factual knowledge of personal significance. Next toepisodes and semantics, people keep other kinds of memories in storelike, for example, feelings, thoughts, movements, and, of course, odorsand tastes. The present argument about personal stability draws atten-tion to three important aspects of autobiographical memory. First, itsnarrative nature; second, the role of imagery; and third, the accuracy ofautobiographical remembering (cf. Rubin, 1996).The narrative nature of autobiographicalmemory stands out inmany

research studies as well as in ordinary experience (Barclay, 1993). Assoon as people start recalling their personal past, they generate story-like sequences with a beginning, a middle, and an end. The narrativestructure of autobiographical memories brings a sequential order intoinformation that might have been endlessly varied and chaotic at themoment of storage. Schank and Abelson (1995) claim that recall isnarrative as a result of the narrative representation of autobiographicalmemories in the cognitive system. Other researchers have shown thatthis claim is too radical: personal recollections are also representedin non-narrative ways (Brewer, 1996). This brings us to the secondcomponent.

118 Jeroen Jansz

Reports of autobiographical memories frequently involve vivid im-ages of the personal past. Imagery is in particular related to the sense of‘‘reliving’’ an episode that is often experienced during autobiographi-cal recall (Brewer, 1996). In recall, the imagery becomes part of anarrative structure as soon as people try to communicate the visualrepresentation of the earlier episode.When they tell a story about a pastexperience they translate the image into a narrative language that willdominate their recollections. The fact that images are traded for narra-tive accounts is probably due to the fact that people are accustomed tomaking sense of their experiences by employing stories (Baumeisterand Newman, 1994).The third aspect of autobiographical memories is their alleged accu-

racy. Reports of personal recollections show that people have moreconfidence in the accuracy of their autobiographical memories than inmemories that are not related to the self (Rubin, 1996). The high per-ceived accuracy of autobiographicalmemories is generally explained interms of the active reconstruction that takes place in the memoryprocess. When people recollect their personal past they establish anarrative order in their lives whatever the actual course of their liveshas been. Thus, ‘‘whereas our lives may not be coherent, our storiesare’’ (Schank and Abelson, 1995: 34). The final result is that autobio-graphical remembering preserves ‘‘a sense of being a coherent personover time’’ (Barclay, 1996: 99).In conclusion, the argument I submit here holds that coherence in the

sense of self is not preserved by consistency in emotional experiencesbut rather by three other constants in individual experiences. The firstone is the immediate and constant experience of being embodied singu-larly. This is linked to the second constant: the body is the point atwhich to locate the individual in social reference. This underlines thepoint that consistency and continuity are to a large extent structured bypractical patterns of interaction. The narrative nature of autobiographi-cal recollections is the third, and final, contribution to the coherencepeople say they experience in their sense of self. It is worthwhile toinvestigate the links between identity and autobiographical memory indetail. There are strong indications that a focus onmemorywill help usto gain a more precise understanding of the relation between self,identity, and emotion (Christianson and Safer, 1996).


CHAPTER 6

Affective processes in amultivoiced selfHubert J. M. Hermans and Els Hermans-Jansen1

In studying affective processes, one can start from the assumption thatduring human development, affective responses are increasingly in-fluenced by the relation individuals havewith themselves. To illustratethis, let’s look at a crawling infant’s fascination for a bouncing ball.Initially, the ball is not experienced as part of the infant’s self, but theinfluence of self-reflectivity becomes explicit when the child is approxi-mately two years old. When playing with another child, it may shout:‘‘Ball mine!’’ The child has developed a special relationship with theball, considering himself its ‘‘possessor.’’ The ball falls within an invis-ible boundary between Mine and not-Mine. The child’s definition ofhimself as the possessor not only indicates that the self is explicitlyinvolved, it also has immediate repercussions for the affective pro-cesses both between the child and the ball and between the child andothers. Still later, when the child has become amember of a sports club,he may consider himself a ‘‘good soccer player’’ and feel very proud ofthis. This self-evaluation indicates that the child, being able to adopt thecomplex language-games of a community, applies to himself a particu-lar standard which is, as an organizing principle, accepted by a groupor community. In Mead’s (1934) terms, the child is organizing his selfby ‘‘taking the role of the other’’ and his feelings are also organized bythis role-taking.As the above examples suggest, the existence of primary affect (e.g.

joy in playing with a ball) is influenced by the development of the self(‘‘ball is mine’’), which strongly colors and organizes the child’s, andlater the adult’s, secondary affective responses to a great variety ofsituations.

Self and affect: dynamics and complexity

Any concept of the self is unthinkablewithout the distinction between IandMe, originallymade by James (1890) and later adopted as one of theclassic distinctions in the psychology of the self (Rosenberg, 1979). In

120

James’ view, the I is the ‘‘self-as-knower,’’ whereas the Me is the‘‘self-as-known.’’ The I has three features: continuity, distinctness, andvolition. The continuity of the self-as-knower manifests itself in a‘‘sense of sameness’’ through time (James, 1890: 332). A feeling ofdistinctness indicates that the I exists separately from others. Finally,the sense of personal volition shows the I as an active processor ofexperience, expressing itself through continuous appropriation andrejection of thoughts. These three features imply the self-reflectivityessential for the self-as-knower (Damon and Hart, 1982).James (1890), identifying theMe as the ‘‘self-as-known’’ or the ‘‘em-

pirical self,’’ observed a gradual transition between ‘‘Me’’ and ‘‘Mine.’’He expressed this in a frequently cited quotation saying that the empiri-cal self is composed of all that the person can call his or her own, ‘‘notonly his body and his psychic powers, but his clothes and his house, hiswife and children, his ancestors and friends, his reputation and works,his lands and horses, and yacht and bank-account’’ (p. 291). As thisquotation indicates, the self is not simply ‘‘within the skin,’’ but ratherextends to the outside world. When someone makes a remark about‘‘my mother’’ or ‘‘my child,’’ the feelings aroused by this remark aresimilar to when the remark is directed at ‘‘myself.’’ James (1890) wasexplicit about the relationship between self and affect: ‘‘The words ‘me’. . . and ‘self,’ so far as they arouse feelings and connote emotionalworth, are objective designations, meaning all the thingswhich have thepower to produce in a stream of consciousness excitement of a certainpeculiar sort’’ (p. 319, emphases in the original).In summary, any theory of the self in the Jamesian tradition implies

two fundamental principles. First, that the self is reflective, assumingan Iwhich becomes aware of itself as aMe. Second, that the self is not a‘‘thing in itself,’’ not a self-contained entity, but a relationship, oftenvery intense, between a person and those parts of the environmentwhich have personal value and affective significance. A combination ofthe two principles leads to the proposition of a reflective self which isaffectively extended towards the outside world.The concept of a self as a self-reflecting agency, continuously in-

volved in I-world connections, as Boesch (1991) calls it, has profoundimplications for understanding the relationship between self and affect.A personwho reacts fearfullymay feel ashamed of his fear, being awareof other peoplewatching him.When someone blusheswith shame, thatperson may be ashamed of his shame so that the initial affective re-sponse is intensified. People may even correct or suppress an initialfeeling by fighting against it with anger and thus ‘‘give themselves akick.’’ In other words, as a result of self-reflectivity, an initial feelingmay become intensified, combined with, or even suppressed by other

121Affective processes in a multivoiced self

feelings. Feelings evoke feelings as a result of one’s own judgment orthe supposed judgment of other people. Looking at feelings this way,they become phases or subphases of dynamic processes which take theform of highly complex affective structures.A self-reflecting person not only responds to her own affective

impulses, she is also able to remember, foresee, and compare experien-ces which are seen as similar, different, or contrasting. Humans are, asBecker (1973) has argued, ‘‘time-binding animals,’’ being capable ofrelating their present to past or future. This active and continuousprocess of relating deeply influences present experiences. When some-one says ‘‘I am already sixty-nine years old,’’ the word ‘‘already’’ mayindicate that this person feels his life has run fast and little time is left.Such a statement may imply the ambivalent feeling of satisfactionabout things achieved and dissatisfaction about running out of time.This condensing of experiences dispersed across time and space re-sults in affective structures which can only be understood by virtue ofthe specifically human capacity of transcending the immediate situ-ation. This complex affective structuring was recently revealed by aformer Dutch prime minister who was asked how he felt at the end ofhis career. He answered that he felt ‘‘dishonestly happy’’ and ex-plained that, although he felt happy for the rich life he had lived, heworried at the same time about the ‘‘unequal distribution of luck in theworld.’’In summary, the use of terms like ‘‘already,’’ ‘‘again,’’ ‘‘still,’’ and

‘‘the first time,’’ suggest that great realms of human affect are part of thedynamics of the self. With its reflectivity and I-world connection, theself evokes complexities of similar, different, and contrasting feelingswhich together form highly dynamic structures. Self-feelings shouldnot simply be understood as anonymous or automatic processes, but,rather, as affective processes in which human agency and intention(James’ ‘‘volition’’) play an organizing role.

The multivoicedness of the self

Whereas the preceding considerations correspond closely with James’founding work, we will make two additional theoretical steps in orderto explain the multivoicedness of the self. The first step concerns thebasic character of time and space in the narrative construction of themind; the second concerns themetaphorical importance of the notion ofvoice.

122 Hubert J. M. Hermans and Els Hermans-Jansen

The narrative turn in the psychology of the self

Social psychologist Sarbin (1986b) made an important theoretical stepby translating the I–Me distinction into a narrative framework. Heproposed that the pronoun I stands for the author andMe for the actor.The self as author, the I, imaginatively constructs a story in which theMe functions as the protagonist, and other figures that are part ofMineas antagonists (e.g. ‘‘my wife,’’ ‘‘my father,’’ ‘‘my friend,’’ ‘‘my en-emy’’) (Verhofstadt-Deneve, 1995, 1999).Sarbin (1986b) views the narrative as a way of organizing episodes,

actions, and accounts of actions in time and space, as an achievementthat brings together fact and fiction in coherent patterns. Our fantasiesand daydreams are organized as stories, and our plannings, our re-memberings, even our loving and hating, are guided by narrative plots.Bruner’s (1986, 1990) writings on narrative have been particularly

influential in drawing on the distinction between the imaginative qual-ity of the narrative mode and the deductive quality of the logico-scientific mode. The narrative mode leads to ‘‘good’’ stories, grippingdrama, and believable historical accounts; the logico-scientific mode,also called ‘‘argumentation,’’ refers to the ability to see formal connec-tions between statements as part of a reasoning process. Whereas thenarrativemode emphasizes human intention as an organizingprinciplebehind specific events and actions, the logico-scientific mode seeks totranscend the particular by reaching for increasingly higher levels ofabstraction. Whereas the logico-scientific mode aims to make abstrac-tion from the specific location of people in action, the narrative modeputs the generalities of the human condition into the particularsof experience, and attempts to locate experience in time and space(Bruner, 1986).The particulars of time and space as central ingredients of storytell-

ing have also been discussed by Jaynes (1976), who also used thedistinction between I and Me in describing the self as ‘‘mind-space.’’The I constructs an analog space and metaphorically observes the Memoving in this space. When we plan to visit someone, we imagine thehouse we are going to and we see ourselves talking to a friend. In thefunctioning of our consciousness, the I sees theMe as themain figure inthe story of our life. Seated where I am, Jaynes explains, I am writing abook and this fact is imbedded in the story of my life: ‘‘Time beingspatialized into a journey of my days and years’’ (p. 63). Like Sarbin(1986b), Jaynes points to the capacity of stories to include not only thecourse of actions but also the motives behind these actions. Thievesmay narratively explain their act as being due to poverty, poets tobeauty, scientists to truth. The narrating I observes the Me as an actor


located in time and space and relates a story about the movements oftheMe, including its motivation, as parts of a narrative construction.

Voices in Bakhtin’s polyphonic novel

In his book Problems of Dostoyevky’s Poetics (1929/1973), originally pub-lished in Russian, Bakhtin develops the thesis thatDostoyevsky– one ofthe most brilliant innovators in the history of fictional writing – createda new form of artistic thought, the polyphonic novel. The principalfeature of the novel is the retreat of the author as the central organizingfigure who is ‘‘above’’ his writings and has an overview of all hischaracters. In the polyphonic novel there is not one single author, butseveral authors or thinkers (Raskolnikov, Myshkin, Stavogin, IvanKaramazov, the Grand Inquisitor) each having his own voice andtelling his own story. The hero is not simply the object of Dostoyevsky’sfinalizing vision, but comes across as the author of his own ideology.The novel is composed of a number of independent and mutuallyopposing points of view embodied by characters involved in dialogicalrelationships. Dostoyevsky himself is only one of the many characters.Each character is ‘‘ideologically authoritative and independent,’’ whichmeans that each character is perceived as the author of his own legit-imate ideological position instead of being an object of Dostoyevsky’sall-encompassing vision. This ‘‘retreat of the omniscient author,’’ socharacteristic of modern novels (Spencer, 1971), implies that the charac-ters are not ‘‘obedient slaves,’’ in the service of Dostoyevsky’s inten-tions, but are capable of standing beside their creator, disagreeing oreven rebelling against him. As in a polyphonic composition, the severalvoices or instruments have different spatial positions and accompanyand oppose each other in a dialogical fashion.

Logical versus dialogical relationships

In order to understand Dostoyevsky’s polyphonic novel, it is necessaryto establish the difference between logical and dialogical relationships.Bakhtin gives the following example (see also Vasil’eva, 1988). Con-sider two phrases that are completely identical, ‘‘life is good’’ and,again, ‘‘life is good’’. In terms of Aristotelian logic, these two phrasesare related in terms of identity; they are, in fact, one and the samestatement. From a dialogical point of view, however, they are differentbecause they may be seen as two remarks voiced by two spatiallyseparated people, who in this case entertain a relationship of agreement.The two phrases are identical from a logical point of view, but differentas utterances: the first is a statement, the second a confirmation. Similar-


ly the phrases ‘‘life is good’’ and ‘‘life is not good’’ can be analyzed. Inlogical terms, one is a negation of the other. As utterances from twodifferent speakers, however, a dialogical relation of disagreement exists.According to Bakhtin, agreement and disagreement are, like questionand answer, dialogical forms. To better understand Bakhtin’s positionit should be added that he certainly does not reject the rules of logic:‘‘Dialogical relationships are totally impossible without logical andconcrete semantic relationships, but they are not reducible to them;they have their own specificity’’ (Bakhtin, 1929/1973: 152).The notion of dialogue opens for Bakhtin the possibility of differenti-

ating the inner world of one individual in the form of an interpersonalrelationship. By transforming an ‘‘inner’’ thought of a particular char-acter into an utterance, dialogical relations spontaneously occurbetween this utterance and the utterance of imaginal others.Dostoyevsky’s novel The Double may serve as an example. The secondhero (the double) was introduced as a personification of the interiorthought of the first hero (Golyadkin). This externalization of a thoughtin a spatially separated opponent instigates a fully developed dialoguebetween two relatively independent parties. Such a dialogical narrativenot only includes space and time, but temporal relations are eventranslated into spatial relations. The result is that temporally dispersedevents are contracted into spatial oppositionswhich are simultaneouslypresent. In Bakhtin’s terms: ‘‘This persistent urge to see all things asbeing coexistent and to perceive and depict all things side by side andsimultaneously, as if in space rather than time, leads him [Dostoyevsky] todramatize in space even the inner contradictions and stages of develop-ment of a single person . . .’ (p. 23, emphasis added).

Imaginal dialogues in daily lives

In her book Invisible guests, Watkins (1986) describes the powerfulinfluence of imaginal dialogues on people’s daily life. We may findourselves speaking to the photograph of someonewemiss, to our cat ordog, or to our reflection in the mirror. Even when we are outwardlysilent, we may be talking with our mothers or fathers, opposing ourcritics, or questioning some personification of our conscience.After a critical discussion of the theories of Piaget, Vygotsky, and

Mead, Watkins (1986) concludes that most psychological theories viewimaginal phenomena only through the eyes of the ‘‘real.’’ The imagin-ation is often seen as a tricky opponent of the real, as little more than amimic or a helpmate, always ready to respond to the real. Most theoriesgive a clear ontological priority to actual or ‘‘real’’ others, usuallyconsidering imaginal others as derivative from or subordinate to them.


Caughey (1984), another student of imaginal phenomena, is alsocritical of the narrow assumptions that many Western social scientistsmake about imaginal phenomena. As a cultural anthropologist heperformed field work in Micronesia and Pakistan, and compared hisobservations with the daily life of Americans. He concluded that it iserroneous to think that only people in cultures with thunder gods andfertilizing spirits live in an imaginal world. The daily world of theaverage American, Caughey observed, is also populated by a throng ofbeings with whom no physical interaction exists. He identified at leastthree categories of imaginal figures: (a) media figures with whom theindividual engages in imaginal interactions (e.g. pop stars and othercelebrities); (b) purely imaginary figures produced in dreams or fanta-sies; and (c) imaginal replicas of parents, friends, or lovers who aretreated as if they are really present. For Caughey, the interactionalnature of these relationships was reason to speak of an (imaginal)‘‘social world’’ rather than a purely ‘‘inner world.’’ Given the personalimportance of these imaginal interactions, Caughey was critical of theidentification of ‘‘social relationships’’ with only ‘‘actual social relation-ships.’’ He saw this conception as incomplete because ‘‘such a suppos-edly objective approach to social organization actually representsan ethnocentric projection of certain narrow assumptions in Westernsocial science’’ (p. 17).Discussing themultivoicedness of the self, it is important to note that

the individual is not only able to talk about the variety of imaginalothers, but also to talk with them as relatively independent parts of anextended self. Cassirer (1955) described the role of a guardian spirit inmythical conciousness, emphasizing that the spirit might be closelyassociated with someone, and perhaps even – like thoughts – inhabithis body or govern his behavior. Even in this case, Cassirer continued,the spirit should not be conceived ‘‘as the man’s I, as the ‘subject’ of hisinner life, but as something objective, which dwells in man, which isspatially connected with him and hence can also be spatially separatedfrom him’’ (p. 168, quoted byWatkins, 1986: 93). All these observations,frombothWestern and non-Western cultures, suggest that the imaginalis not simply ‘‘outside the I.’’ Rather, together with a multiplicity ofimaginal figures the I constructs a multivoiced world. In other words,imaginal figures are voiced positions in a self-space, and sometimesthese positions entertain dialogical relationships.The preceding observations referring to a process of intersubjective

exchangewith an imaginal figure suggest that the frequently used term‘‘projection’’ is an unsatisfactory label. It reduces the process to onlyone voice and, consequently, to only one author. Moreover, it ignoresthe fact that as well as the person creating an imaginal figure, the


imaginal figure is creating the person (see, for example, the manyaccounts by artists of an inspirational figure).

I-positions in a multivoiced self

The metaphor of the polyphonic novel expands on the narrative con-ception of the I as author and the Me as observed actor. Whereas inSarbin’s (1986b) conception of the self-narrative, a single author (the I) isassumed to tell a story about him- or herself as an actor (the Me), thepolyphonic novel as metaphor for the self goes one step further. Itpermits one individual to live in a multi-authorial world, in a multiplic-ity of worlds with each world having its own author telling a storyrelatively independent of the authors of the other worlds. Moreover,the several authors have the possibility to enter into dialogue.Elaborating on the polyphonic metaphor and its implication of

spatialized dialogue, Hermans, Kempen, and Van Loon (1992) con-ceptualized the self in terms of a dynamic multiplicity of relativelyautonomous I-positions in an imaginal landscape; the I has the ability tomove from one position to the other in accordance with changes insituationand time. The I fluctuates betweendifferent and even opposedpositions, and has the capacity to imaginatively endow each positionwith a voice so that dialogical relations between positions can beestablished. The voices function like characters in a story, involved in aprocess of question and answer, agreement and disagreement. Each ofthem has a story to tell about their own experiences from their ownstandpoint. As different voices, they exchange information about theirrespectiveMe’s, resulting in a complex, narratively structured self.In the following section we will translate the preceding consider-

ations into a more specific psychological theory (valuation theory) andconcrete procedure (self-confrontation method) allowing us to studythe dialogical self in an empirical way.

Valuation theory: personal meanings in a multivoiced narrative

Valuation theory (Hermans, 1987a, b; Hermans and Hermans-Jansen,1995) was originally developed to study the ordering and developmentof experiences into a narrative structure. The underlying view of theperson was inspired by philosophical–phenomenological thinking(James, 1890; Merleau-Ponty, 1945) and, later, by Bakhtin’s work. Invaluation theory, the self is conceived as an ‘‘organized process ofvaluation.’’ The process aspect refers to the historical nature of humanexperience and implies a spatio-temporal orientation: the person livesin the present and is, from a specific point in space and time, oriented


towards both past and future. The organizational aspect emphasizes thatpeople not only orient to different aspects of their spatio-temporalsituation but also tell different stories about their past, present, andfuture depending on the actual or imaginal positions from which thesestories are told. Through the act of self-reflection and telling, the personis able to bring different experiences and views together into a compos-ite whole. In this composite whole some parts become more influentialthan others.The central concept of valuation is very open to the everydayworld of

the individual. It refers to anything people find important when talkingabout their life-situation. It is any unit of meaning that has a positive(pleasant), negative (unpleasant), or ambivalent (both pleasant andunpleasant) value to them. Valuations can include a broad range ofphenomena: a dear memory, a difficult problem, an unreachable ideal,an intriguing dream, and so forth. The person is differentially orientedtowards present, past, and future, resulting in different valuations. Thedifferent valuations are combined into a valuation system which ex-presses an individual’s story from the perspective of one position.When someone is invited to tell his or her story from the perspective ofdifferent positions (e.g. I as a teacher, I as the person I am in mydreams), they are supposed to be able to formulate different valuationsystems. In other words, an individual may tell a different personalhistory and value this history differently, depending on the positionfrom which it is told.The theory assumes that each valuation has an affective connotation

represented by a particular pattern of affect. When we know whichtypes of affect are characteristic of a particular valuation, we knowsomething about the valuation itself, which also implies that the affec-tive meaning of a valuation cannot be separated from it. Moreover, theaffective component reveals the organization of a specific valuationsystem, and, in the case of several valuation systems, the organizationof the different systems as a differentiated whole.As part of the affective component of valuation theory, the latent–

manifest distinction was introduced. At the manifest level, a largenumber of personal valuations come and go. This phenomenologicalrichness varies not only across individuals but also within a singleindividual across time and space. At the latent level, a limited number ofbasic motives are assumed to be in operation. Active in everyone in allperiods of life, these motives are reflected (latently) in the affectivecomponent of the valuation system. Study of the affective componentcan therefore reveal which particular motive is active in a particularvaluation or valuation system. The basic motives reflect the person as amotivated storyteller (Hermans and Hermans-Jansen, 1995).


Two motives have been elaborated: the striving for self-enhance-ment, or S motive (i.e. self-maintenance and self-expansion), and thelonging for contact and unionwith the other, or Omotive (i.e. the searchfor alliance with somebody or something else). This distinction was theresult of a literature review inwhich the perspectives of various authorson the basic duality of human experience were considered. Forexample, agency and communion for Bakan (1966); autonomy (or self-determination) and homonomy (or self-surrender) for Angyal (1965);Bindung (solidification) and Losung (dissolution) for Klages (1948);power and intimacy forMcAdams (1985). Note that in valuation theorythe basic motives are considered not as causal explanations but ratheras basic themes in people’s self-narratives. Moreover, the suppositionof two basic motives has no exhaustive pretensions. It is rather a choiceon the basis of theoretical parsimony.When a valuation represents a realization of the S motive (e.g. ‘‘I

passed a difficult test’’), the person experiences a feeling of strengthand pride in connection with the valuation. Similarly, when a valua-tion represents a realization of the O motive (e.g. ‘‘I enjoy the momentswith my friend’’), feelings of intimacy and tenderness are associatedwith the valuation. In other words, the latent motivational base be-comes manifest in the affective pattern associated with a particularvaluation.

Generalization and idealization

Two concepts, generalization and idealization, particularly representthe organizational aspect of valuation theory and play a central role inthe methodology presented in the next section. The more a particularvaluation generalizes as part of a system, the more it determines the‘‘general feeling.’’ When people tell how they generally feel, it is highlyprobable that particular experiences color this general feeling morethan others. Not all valuations are equally influential in the system. Themore generalizing power a valuation has, the more influential theaffective component of this valuation is in determining how someonegenerally feels during a particular period of life.Similarly, valuations differ in the extent of idealization. This concept

refers to the fact that certain valuations better fit how someone wouldlike to feel (the ‘‘ideal feeling’’) than others. Valuations that color theideal feeling are often different from those that influence the generalfeeling. When people are faced with personal problems which affecttheir selves to a significant degree, the ideal feeling typically has anaffective modality that is in contrast to the affective modality of thegeneral feeling.


Self-confrontationmethod: organization and reorganization ofvaluations

The self-confrontation method is an idiographic procedure based onvaluation theory. It is designed to study relations between valuationsand types of affect and the way these variables become organized intoone or more systems (Hermans and Hermans-Jansen, 1995). The pro-cedure involves elicitation of a set of valuations and then association ofthese valuations with a standardized set of affect terms. The result is amatrix in which each cell represents the extent to which a specific affectis characteristic of a specific valuation. As we will see, the self-confron-tation method will enable us to study valuational and affective struc-tures belonging to I-positions in detail.The valuations (rows in the matrix) are elicited using a series of

open-ended questions. The main questions are intended to bring outsome important units of meaning for the past, present, and future (fordetails see Hermans and Hermans-Jansen, 1995). The questions inviteindividuals to reflect on their life and tomention those concerns that aremost relevant from the perspective of the present situation. Also, sub-jects are encouraged to phrase the valuations their own way so that theformulations reflect the intended meaning. The typical form of expres-sion is the sentence as the basic unit of text.In the second part of the investigation, a standard list of affect terms

(columns in the matrix) is presented. Concentrating on the first valu-ation, subjects indicate on a 0–5 scale the extent to which they experi-ence each affect in relation to the valuation. The subject, working alonenow, rates each valuation with the same list of affect terms. The differ-ent valuations can then be compared according to their affective pro-files. Once the affective ratings for the different valuations have beenobtained, a number of indices that represent the motivational structureof the valuation system are calculated.

1. Index S is the sum of the scores for four affect terms expressingself-enhancement.

2. Index O is the sum of the scores for four affect terms expressingcontact and union with the other.

3. Index P is the sum of the scores for four positive affect terms.4. Index N is the sum of the scores for four negative affect terms. (Notethat the scores for each of the four indices S, O, P, andN range from 0to 20 for each valuation.)

In summary, the degree of S and O affect indicates the extent ofrealization of the basic motives, whereas the degree of positive andnegative affect reflects the extent in which these motives meet super-able or insuperable obstacles.


5. For the extent of generalization (G) of a valuation, the followingquestion is asked at the end of the valuation construction phase:‘‘How do I generally feel these days?’’ This question does not ask fora specific valuation but is intended to assess the ‘‘general feeling.’’The person answers directly with the list of affect terms that wasused for the characterization of the valuations. The product-momentcorrelation between the pattern of affect that belongs to a specificvaluation and the pattern of affect that belongs to the general feelingreflects the extent of generalization of this valuation. The morepositive the correlation, the more this valuation is supposed togeneralize.

6. For the assessment of the idealization (I) of a valuation, the questionis: ‘‘Howwould I like to feel?’’ The correlation between the affectiveprofile belonging to a specific valuation and the affective profilebelonging to the ideal feeling indicates the ideal quality of thisvaluation. The strength of the correlation indicates the extent ofidealization. The more positive the correlation, the higher the ideal-ization. When a valuation has an affective profile that contrasts withthe ideal feeling, this is expressed in a negative (minus) correlation.(G and I represent the last rows in the matrix, see table 6.1). (Forreliability and validity data of all the indices presented so far, seeHermans, 1987b.)

The results are discussedwith the person, usually oneweek later so thatthe psychologist can analyze and study the data and the person candigest the many impressions, associations, and thoughts evoked by theself-investigation. The discussion has the quality of an intensive self-reflection and profound dialogue with the interviewer. The person andthe interviewer base their discussion on the overall picture provided bythe system of elicited valuations.A second (and sometimes a third and a fourth) self-investigation

usually follows, typically after some months. In this case, however, thesubjects do not start ‘‘from scratch.’’ Instead, the interviewer reads theoriginal questions and, after each question, produces the statementsthat they provided in the preceding self-investigation. The subjects areinstructed to consider, for each statement separately, whether theywould still respond the same.Where this is not the case, the intervieweroffers various options: an old valuationmay be reformulated (modifica-tion); replaced (substitution); discarded altogether (elimination); or anew response may be added (supplementation). This way, subjectshave considerable freedom and indirectly point to the constant andchanging elements in their own valuation systems.


Sharon’s I-positions: telling and retelling a self-narrative

In one of our projects on the multivoiced self, we considered oppositepersonality sides as I-positions, approaching them as voices. The cen-tral question we posed was: ‘‘Can you consider two sides of yourpersonality, which you see as opposite, with one of the sides moredominant than the other side?’’ One subject may say: ‘‘I have an openside, but also,more hidden and less familiar, a closed side.’’ Othersmaystate different opposites, for example, ‘‘serious versus playful,’’ ‘‘activeversus passive,’’ ‘‘hard versus soft,’’ etc. After describing the oppositesin their own terms, we invite them to focus on the first, dominant sideof their personality and to think and feel about themselves, say, as anopen person, about their past, present, and future and tell about im-portant experiences and circumstances from this particular perspective(I-position). This results in a valuation system of this subject as an openperson. Next, they are invited to reflect about themselves as a closedperson, and tell about their life from this particular perspective. In otherwords, the person constructs two valuation systems, each with theirown content and affective organization, which then can be studied andfollowed across time.Sharon, a 34-year-old married mother of two children and part-time

teacher, contacted a psychotherapist (the second author) in a period ofdistress. The immediate reason was a nightmare, in which she saw herhusband and two children at an unnatural distance, experiencing her-self as if she were dead: ‘‘In this dream I thought: ‘perhaps I’m deadnow . . . theywill manage, it makes no difference tome’; I felt no sorrow,no anger, I felt nothing.’’ She became convinced that this dream sym-bolized the way she often experienced herself in daily life: being con-tinuously busy, feeling increasingly alienated from her own feelings.Because the dream frightened her, she told her mother, but her motherdidn’t respond. Sharon increasingly became aware of her dream asrevealing an important but neglected part of her life, both present andpast, and therefore she decided to contact a psychotherapist.After some sessions the psychotherapist proposed that she investi-

gate herself using the self-confrontationmethod. As part of this investi-gation, Sharon was asked the question about the existence of twoopposite personality sides (see above). She phrased her dominant sidethis way: ‘‘My I which puts my respect outside of myself’’ and ex-plained that she always tried to conform to the high expectations thatothers, in particular her parents, had of her achievements and visiblepresentation. Her opposite side, which she considered less dominant,even suppressed, was labeled ‘‘My own world of feelings.’’ This siderepresented her deeper, affective life from which she felt closed-off, as


indicated by her dream. The valuations and affective indices whichSharon formulated from the perspective of her ‘‘outside’’ (Position A)and her ‘‘inside’’ (Position B) are presented in the left part of table 6.1.A direct insight into the difference between Sharon’s two opposite

positions (A and B) can be found by inspecting the affective propertiesof her ideal feeling.Whereas for Position B this is high for both S and O,for Position A it was high in S and low in O, suggesting that from theperspective of this position, she is oriented to an ideal characterized byhigh self-enhancement and low contact and union. Looking at thespecific valuations of Position A, we see that valuation 1, concerning‘‘being important,’’ is associated with much self-enhancing affect incombination with a high level of positive affect. Also, this valuationcorrelates highly (.89) with the ideal feeling, suggesting that ‘‘beingimportant’’ has an ideal quality from the perspective of this position.Strikingly, as part of valuation 3, the same theme has a high level ofnegative affect in combinationwith low levels of both self-enhancementand contact. Moreover, valuation 3 strongly generalizes within thisposition (A), which suggests that this valuation is most influentialwithin Position A. Apparently, Sharon finds herself in an internalconflict: what she remembers as positive in the past (valuation 1), isperceived as negative in referring to her future (valuation 3). In otherwords, from the perspective of Position A, Sharon is oriented to self-enhancement but does not see much prospect for pursuing this‘‘outside goal’’ in the future (see also the contrast between general andideal feeling).Whereas Position A is governed by the self-enhancement motive,

Position B focuses primarily on contact and union (see the high levelsfor O affect for valuations 6, 7, 8, and 9). This strong orientation is alsoindicated by the high level for O affect in the ideal feeling. Despite thisstrong orientation towards contact and union, the valuations high in Odo not generalize (see the low levels of G for valuations 6, 7, 8, and 9).Instead, valuations 4 and 5, referring to contact with her parents, showlow levels for both S and O, and strongly generalize within the systemof Position B (.93 and .86, respectively). Apparently, Sharon is longingfor contact and union, but cannot fulfill her desire because she feels herawareness of her parents’ inability to express their emotions is blockingher. This lack of fulfillment is also represented by the affective contrastbetween general and ideal feeling.In summary, Sharon’s self-investigation shows that whereas Position

A is oriented to self-enhancement, Position B is primarily oriented tocontact and union, both orientations being unrealized. The two posi-tions represent disagreeing voices in Sharon’s self and are associated bycontrasting affect.


Table 6.1. Sharon’s valuations and corresponding affective indices for positions A, B, and C

Position A/Time 1 Position C/Time 2My I which puts my respect outside of myself My own world of feelings as basis for a well-functioning outside

S O P N G I S O P N G I

(past)1. My work, my educationaltraining, that is, my visibleappearance, that was mostimportant for my parents

16 1 16 7 −.23 .89 My work, that’s what I cando, that’s what I am, and,moreover, I work with agroup in which people arevery involved with each other

15 7 13 0 .75 .75

(present)2. I’ve a desire for work andfor a position in society

15 1 9 11 .22 .66 I’ve a desire for work and fora position in society

12 0 11 9 .06 −.05

(future)3. Self-confidence, which I getfrom my work, knowledge,and being important, is myoutside goal

2 1 1 18 .92 −.32 I’m satisfied with what I havenow; in my private life and inmy work, I no longer putmyself under pressure toachieve more

18 7 18 1 .70 .78

General feeling 4 3 2 19 — −.38Ideal feeling 12 2 11 3 −.38 —

Position B/Time 1My own world of feelings

S O P N G I S O P N G I

(past)4. I’ve missed having amother who had feelings andemotions and showed them

0 0 0 17 .93 −.94 [valuation eliminated]

(past)5. My father also didn’t showhis emotions; I’ve thesuspicion that he has thembut can’t cope with them

2 3 0 13 .86 −.83 [valuation eliminated]

(present)6. When I talk with mychildren, when I’m angry,powerless or distressed, Iwant them to see somethingof this

12 20 14 4 −.62 .80 When I talk with my children,when I’m angry, powerless ordistressed, I want them to seesomething of this

9 13 11 6 .38 .50

(present)7. When I was sick, andSheila [daughter] was lookingfor something, I loved hervery much and felt sad that Icouldn’t care for her

8 18 13 17 .29 −.16 I’m feeling intensely in touchwith my children and peoplearound me

10 14 17 3 .75 .83

(present)8. After the nightmare I feltagain that I loved mychildren; I also felt in touchwith the people around me

9 17 15 13 .12 .10 [valuation combined withvaluation 7]

(future)9. I wish that my feelingswould leave these extremities,that I could become aware ofthem in time, and get more ofa grip on this

14 15 14 2 −.80 .84 My feelings and emotionscome sooner to the surface; itbecomes clear to me morequickly where they comefrom and to which aspects ofmy life they are related

13 8 12 4 .73 .65

General feeling 3 3 4 20 — −.90 General feeling 15 14 15 8 — .82Ideal feeling 17 20 16 0 −.90 — Ideal feeling 18 18 20 1 .82 —

Note: S= affect referring to self-enhancement; O=affect referring to contact with the other; P =positive affect; N=negative affect;G=generalization; I = idealization

After her self-investigation, Sharon continued seeing the therapist ona two-weekly basis. After six months they decided to perform a secondself-investigation to examine whether the positions and valuationsfrom Investigation 1 had been changed. When asked to consider thetwo personality sides as distinguished in Investigation 1, Sharon nowproposed to combine Positions A and B into a third Position C,which she described as ‘‘My own world of feelings as basis for awell-functioning outside.’’ She explained that her original Positions Aand B, which she had experienced as strongly opposed and inconsist-ent, had been losing much of their conflicting character so now theycould be combined. The valuations belonging to Position C are pres-ented at the right side of table 6.1. As can be seen from table 6.1, Sharoneliminated some of the valuations and modified others, in accordancewith the procedure described earlier.Comparing the first self-investigation (Time 1) with the second (Time

2), we see a conspicuous change in the general feeling. Being stronglynegative at Time 1, it is positive with high levels for both S and O atTime 2. The valuations with the highest degree of generalization arenos. 1, 3, 7, and 9. Valuation 9 is quite significant from a psycho-therapeutic point of view, telling that Sharon has learned to bring herfeelings ‘‘sooner to the surface’’ and has developed a capacity to relateher feelings to specific experiences. As a result of her increased psycho-logical skills, Sharon has succeeded in bringing her ‘‘outside’’ and her‘‘inside,’’ represented by two strongly contrasting Positions A and B atTime 1, together into a more integrativeGestalt, represented by PositionC at Time 2. This change is also manifest in valuation 1: at Time 1 thisvaluation referred to her ‘‘visible appearance,’’ the modified valuationat Time 2 combined an achievement element (‘‘. . . what I can do . . .’’)with a social element (‘‘I work with a group in which people are veryinvolved with each other’’). The overall change is reflected by the highcorrespondence between the affective modalities of the general andideal feelings at Time 2.Notably, Sharon had deleted all references to her parents which

played such an important role at Time 1 (see valuations 1, 4, and 5).Despite the valid reasons which may exist for concentrating on herpositive social relationships, Sharon’s parents are of direct significanceto her well-being because they are not simply ‘‘external’’ persons, butpart of her self. Sharon recognizes the emotional problems of herparents in herself, as exemplified by her nightmare in which she foundherself empty of feelings, a problem which is similar to the problemperceived in her parents. On the assumption that the parents are rel-evant parts of Sharon’s self, the question that remains is whetherSharon is capable of developing her self-system further without


working through her relationship with her parents. On the other hand,it should be emphasized that Sharon has made progress in coping withher feelings (see valuation 9 at Time 2), so that the possibility exists forher to develop further the coherence of her self without directly focus-ing on the relationship with her parents.From a theoretical point of view, Sharon’s case illustrates some of the

ideas and concepts discussed at the beginning of this chapter. She did aself-investigation not from ‘‘one and the same I.’’ She did her self-investigation not from one centralized position, as is usual in interviewsor test administration. Rather, we approached Sharon as a multivoicedindividual, able to reflect on her self from two opposite positions andtell about significant meaning units from the perspective of these posi-tions. She answered the questions in the self-confrontation methodtwice, coming up with different answers and producing different affec-tive profiles. The two initial positions can be considered as twodisagreeing voices or oppositional characters in a multifaceted self-narrative. To a large extent, Sharon’s emphasis on solidarity andintimacy with significant others (Position B) can be understood as ananswer to the threatening shallowness and emptiness of her ‘‘outside’’(Position A). In the second investigation, she made a significant step inthe direction of agreement between the two voices as combined in athird, more integrative system (Position C). As Josselson (1995) wouldsay, from the latter position she tells a self-narrative superseding andencompassing the preceding ones.

Different voices, different affect

Aswe have seen in our case study, voices, or, using amore spatial term,I-positions, are not simply ‘‘internal affairs’’ representing an intra-psychic world of thoughts and feelings. As Voloshinov (1929/1986) hasargued, a word is a two-sided act. It is not only determined by thespeaker but also by the imaginal or real person for whom the word ismeant. The word is the product of the reciprocal relationship betweenspeaker and listener, between addresser and addressee. In otherwords,a word is a bridge.As Sharon’s case study shows, I-positions are embedded in social

relationships. Her positions can be understood as localized speakersentertaining active, reciprocal relationships with the social environ-ment and with each other. Not only her first position (her outside), butalso her second position (her feeling world rooted in her inside) wereembedded in reciprocal relationships with the environment, as thevaluations formulated from the perspective of this position show. Her‘‘own feeling world’’ explicitly referred to her relationship with the


people around her. These reciprocal relationships are even at the heartof so-called ‘‘inner feelings’’ and are a challenge to our inside–outsidedichotomies. As Sampson (1993) has argued, the boundaries betweeninside and outside or between intrapersonal and interpersonal are bynomeans as neat as we had believed: ‘‘What is inside is simultaneouslyoutside, and vice versa’’ (p. 134). People who are ‘‘outside’’ are inter-nally reconstructed as imaginal others and are also present when wedefine ourselves as being ‘‘alone.’’ In fact, we are never alone, as ourminds are continuously populated by the voices of ‘‘external’’ people.As we have seen in this chapter, imaginal others can be considered as

examples of I-positions. This concept, however, can be applied in moreways. Personality traits or Kelly’s (1955) personal constructs can also beconceived of as I-positions (Hermans, 1996a, b). In Sharon’s case study,we have examined idiosyncratically phrased personality sides as I-positions. All these approaches suppose that people can tell a variety ofstories about themselves, from the perspective of voiced positionslocalized in an imaginal space. Given the blurring boundaries betweeninside and outside, this imaginal space is closely intertwined withpeople’s actual space. By implication, interpersonal and intrapersonalrelationships are always side-by-side and even interwoven.The term I-position can be best understood as a reaction to the

traditional Cartesian split between subject and object. As Overton(1998) explains, Descartes’ Cogito ergo sum was a historically definingmoment because it set the subject (I) against the (external) object, and,hence, created the overarchingdichotomy thatwas to become knownastheCartesian split. This became the origin of ontologicaldualism: subjectversus object, mind versus body, thinking versus space were eachpresented as independent, isolated realities. As a consequence of thesplit, psychologists conceived the I as a soul, supposed to have anexistence by itself, unified in itself, and separate from externally locatedspace.As discussed at the beginning of this chapter, James (1890) corrected

the traditional notion of the separate soul by conceiving of a dynamicself which was explicitly extended to the environment. His emphasis,however, remained on the I as unified in itself and consistent over time.It should immediately be added that James was well aware of themultiplicity of the self as demonstrated by his discussion of the ‘‘rivalryand conflict of the different selves’’ (p. 309). Despite James’ insight intothe multiplicity of the self, Bakhtin’s (1929/1973) dialogical approachcan be seen as a decisive step beyond James’ and even Mead’s theor-izing, because he accepted the full consequence of the linguistic natureof the human being.The I-position concept can be seen as a correction to the Cartesian


split by bringing space back into the self. The same idea can be found inJohnson’s (1985) book The body in the mind. The word ‘‘in’’ reflects aconceptual attempt to transcend the boundaries of the traditional Car-tesian self. Similarly, the term ‘‘I-position’’ is devised as a spatial termsuggesting that space is in the mind and not only outside. Moreover, aposition is always in relation or opposition to something else. A posi-tion never exists separately from an opposite position. Just as a particu-lar meaning can only be defined by its opposite, as Rychlak (1988) hasso eloquently argued, a position can only be properly understood bytaking its counterposition(s) into account. When I am in contact withanother person, the other person is an opposite position reconstructedas part of my extended self. Both my own place (here) and the place ofthe other (there) function as positions in a real or imaginal space.Moreover, as Harre and Van Langenhove (1991) have argued, the verbspositioning and repositioning have, in contrast to the more static termrole, the theoretical advantage that they are sensitive to the negoti-ations, contradictions, agreements, and disagreements among andwithin people which suppose an iterative moving to and fro betweendivergent or opposite stances in an extended field of positions.The preceding considerations have immediate consequences for the

way we conceive of the relationship between self and emotion. Thereare two ways of considering this relationship. One is to start from aparticular conception of emotions and then examine its implication forthe domain of the self. Another is to start from a particular conceptionof the self and explore its implications for the functioning of emotions.The present contribution is based on the second option. At the begin-ning of this chapter we argued that the capacity of self-reflection,together with the I–world relationship, deeply influences our (second-ary) affect. The succeeding discussions of the multivoiced, dialogicalself as elaborated in valuation theory again emphasize the implicationsthat a conception of the self has for affective processes. As illustratedby Sharon’s case, different voices carry different affective structures(or profiles). Instead of a single-voiced investigation, Sharon did amultivoiced investigation with the result that not only different andcontrasting valuations were found between positions, but alsodifferent and contrasting affective profiles. One and the same personcan experience opposite feelings depending on the position evoked in aself-investigation. A multivoiced view on the self involves someonewho is able to shift to and fro between positions and to experiencedifferent affects accordingly. This view implies that there is notone center from which feelings are experienced, but more centers orpositions from which feelings are structured and contrasted and fromwhich they receive their specific meaning.


As part of his plea for a narrative conception of the mind, Bruner(1986, 1990) holds that the self is ‘‘distributed.’’ Thinking of the self thisway means we are forced to dissociate ourselves from a monolytical,centralized, and unified conception of the self, which has for so longdominated the psychology of the self, both before and after James. Themultivoiced approach implies that attributing an isolated emotion (e.g.anxiety) as a trait or state to a particular person is inadequate forunderstanding affective processes. Such attributions should be re-phrased as part of the dynamic interplay between andwithin positions.An emotion or emotional process receives its specific meaning in thecontext of other emotions within a particular position, and, moreover,in the context of the emotions of other positions (see the dynamic andmeaningful contrasts between the feelings within and between Posi-tions A and B for Sharon). Emotions receive their subjectivemeaning aspart of a dynamic process of positioning and repositioning and thisprocess arouses different, contrasting, and contradicting feelings, notonly at different moments but even simultaneously. It is our thesis thatthis requires a complex and dynamic view of the self and, as a conse-quence, of the process of emotion.

Note

1 We thank Sue Houston for her editorial remarks. In this chapter HubertHermans as a personality psychologist cooperated with Els Hermans-Jansenas a psychotherapist. The therapeutic assistance to the client of the case studywas provided by the second author.


COMMENTARY

Affective processes in amultivoiced self in actionLeni Verhofstadt-Deneve

Introduction

The theoretical framework ofHermans andHermans-Jansen, includingthe so-called valuation theory, is grounded in James’ well-known I–Meconcept of the self with the I as the reflecting ‘‘self-as-knower’’ and theMe as the reflected ‘‘self-as-known.’’ An important aspect of this view isthat the Me is much wider than the mere result of reflection on the self(in its narrow sense); it also contains the whole social and materialworld which has a certain personal value and affective significance forthe person. This view implies an I–world connection with a strong andcomplex relational and affective component in the self-reflecting per-son in present, past, and future situations. Hermans and Hermans-Jansen: ‘‘With its reflectivity and I–world connection, the self evokescomplexities of similar, different, and contrasting feelings, which to-gether form highly dynamic structures’’ (p. 122).An original turn (inspired by Sarbin) is the interpretation of the self

as a sort of narrative psychodramatic system in which the I and the Mefunction as ‘‘author’’ and ‘‘observed actor’’ respectively. The self asauthor (the I) constructs an imaginary story in which the Me functionsas (1) the protagonist, and (2) other figures that are part of Mine asantagonists (or imaginal others e.g. father, sister, partner). I presumethat Hermans and Hermans-Jansen do not restrict these antagonists tosignificant others but include the whole social and material world towhich the I attributes a personal affective value. From this point of viewthe I constructs a multivoicedworld in which the individual is not onlyable to talk about the variety of imaginal others (antagonists) but also totalk with them as relatively independent parts of an extended self. ‘‘Inother words, imaginal figures function as voiced positions in a self-space, and, at times, these positions entertain dialogical relationshipswith each other’’ (p. 126).It is interesting to note that this valuation theory functions as a basis

141

for the Self-Confrontation Method (SCM), an idiographic procedurewhich is constructed ‘‘to study the relation between valuations andtypes of affect and the way in which these variables become organizedinto one ormore systems’’ (p. 130). The procedure involves elicitation ofa set of valuations and subsequent association of these valuations witha standardized set of affect items.The valuation theory is highly attractive, as it is based on the organ-

ized but complex subjective phenomenological cognitive–affective nar-ratives of the person. In the elicitation of a set of valuations the personhas the opportunity to take diverse (also contradictory) I-positions,which over time can develop ‘‘into a more integrative Gestalt.’’However, one can question whether or not the SCM remains on a

merely cognitive, rational self-reflective level. Itmainly stresses a reflec-tion on cognitions and emotions from self and others and the dialoguesbetween those constructions. Even when in the SCM the focus is ex-plicitly on affects, what remains is mainly a reflection on affects and notnecessarily an intense involvement with those affects. The ‘‘psycho-drama’’ in the SCM exclusivelyworks on an internal level with a strongcognitive–rational accentuation.A similar limitation concerns the important aspect of conflict experi-

ence. For example: ‘‘Sharon did her self-investigation not from onecentralized position . . . Rather we approached Sharon as a multivoicedindividual, able to reflect on her self from two opposite positions andtell about significant meaning units from the perspective of these posi-tions’’ (p. 137) Real conflict experience is not at stake here, but thereflectionon and dialoguewith different I-positions seems again crucialfor personality integration.This highly cognitive accent (which no doubt has value of its own)

could effectively be enriched with cognitive–affective components bymeans of the introduction of real action and drama techniques, inwhich the protagonist not only reflects on, or engages in dialogue with,but also becomes diverse antagonists or imaginal others in a specifictime and space.In the following comments I will try to defend this thesis by first

explaining the importance of conflict for personality development andsecondly by illustrating how, in association with the SCM, differentsorts of conflicts could be worked with through the introduction ofaction and drama techniques.

The role of conflict in personality and personality development

Although Hermans and Hermans-Jansen do pay attention to the possi-bility of conflicts between different parts of the person, between differ-

142 Leni Verhofstadt-Deneve

ent I–Me positions, they mainly treat them from a clinical perspective,with a strong rational–cognitive accent. In this view conflict is not seenas the driving force in psychological development. In developmentaltheories such as that of Piaget (1949), conflict is seen as an essentialcharacteristic in the developmental process. Conflicts within the‘‘multivoiced self’’ can drive the individual’s development. Opposi-tions and conflict, from this perspective, can be considered a necessary(but not sufficient) condition for development. Such an interpretation isunderpinned by my own follow-up research, therapeutic practice, andby theoretical analyses on the basis of dialectical developmental psy-chology (Brown, Werner, and Altman, 1998; Conville, 1998; Verhof-stadt-Deneve, 1985; Verhofstadt-Deneve and Schittekatte, 1996).In this view development is seen as a self-organizing process, con-

sisting in dialectical developmental movements, which are inherent inthe qualitative development of each living organism (see also Piaget,1968; Prigogine and Nicolis, 1977). Or in the words of Piaget, referringto the dialectical course of cognitive development: ‘‘c’est une demarcheinevitable de la pensee’’ (p. 104). A comparable view is to be found inthe rather rigid epigenetic model of development as described byErikson (1968): every stage of life is built on a crisis, the content ofwhich is characterized by a positive and a negative pole. There is astrong resemblance here to Kelly’s personality theory and its bipolarconstructs as the cornerstones of an active and constantly evolvingsystem of constructs. The experiencing of oppositions causes the entiresystem of constructs to shift, so that change, activity and personalitydevelopment can take place (Kelly, 1955). An explicitly dialectical biasis also inherent in Vygotsky’s approach (1979; Zender and Zender,1974), since he looks upon crises and conflicts as criteria marking offthe stages of a dialectical development. It is worth noting that inVygotsky’s view the crisis (or negative phase) in the child’s psycho-logical development acquires a pronounced positive meaning. Thebasic concept of contradiction and conflict as stimuli for developmentalprocesses is related to Kohlberg’s (1976) notion of cognitive conflictas the stimulus of moral development. His way of thinking wassupported by Lawrence Walker’s (1983) experiments. Walker foundthat a range of conditions which trigger cognitive conflicts stimulatethe transition from one stage to the next in the moral developmentof children. In my theoretical framework dialectical movements andself-organization determine respectively the process and direction ofqualitative developments.

143Commentary on Hermans and Hermans-Jansen

Conflicts between what?

The idea that conflicts drive development has been further elab-orated in my phenomenological–dialectical model of personality(Verhofstadt-Deneve, 1988, 1995, 1997, 1999, 2000). In this model thephenomenological aspect refers to the unique subjective content whichevery human being attributes to him- or herself and to the surroundingworld and is also based on the William James I–Me paradigm. Thedialectical aspect refers to the underlying process of interaction andconflict between these contents, and their development.Among the multitude of I–Me personality constructions six main

domains or images (comparable with fundamental I-positions inHermans and Hermans-Jansen) are focused on, each correspondingwith a specific question:

The Self-Image: ‘‘Who am I, with my potential and my short-comings?’’

The Alter-Image: ‘‘What are the others like?’’The Meta-Self: ‘‘How is one’s own person viewed by the others?’’

These three basic images are counterbalanced by three ideal-images:

The Ideal-Self: ‘‘Who would I like to be or become?’’The Ideal-Alter: ‘‘What should the others be or become like?’’The Ideal-Meta-Self: ‘‘How should the others perceive me?’’

For each image a distinction can be made between an external aspect(the way we behave externally) and an internal aspect (what we thinkand feel at the same time). In this model the person can also be seen as a‘‘multivoiced self’’ with six main parts which can interact and be inconflict.On the basis of themodel,many types of conflict can be distinguished

according to the nature of the opposing poles. I shall just give someexamples:

– There may be inter-dimensional oppositions, for example betweenthe Self-Image and the Meta-Self.

– Similarly, theremay exist intra-dimensional oppositions, for examplebetween external and internal contents of one dimension.

– Oppositions may be experienced between the subjective phenom-enological constructions of oneself and of theworld, on the one hand,and possible alternative interpretations, on the other.

– Vivid oppositions can also be felt between the three time dimensions.

Within the scope of these comments I will, in the presentation of actionand drama techniques, confine myself to the first and the second kinds


of oppositions. For more information I refer to Verhofstadt-Deneve(1995, 1997, 1999, 2000).

How to stimulate conflicts and development?

As Hermans and Hermans-Jansen have shown, the use of the SCM cantrigger self-organizing processes in the ‘‘multivoiced self’’ by means ofthe stimulation of cognitions, emotions, and language. This happens onan internal level, but from my clinical experience and didactic groupworkwithmy students I have learned that direct spatial, external actionin the context of an intense and supportive group situation can lendvital support to such internal self-organizing processes. To me it seemsthat the SCM is too exclusively aimed at an internal, mental action level:the ‘‘protagonist’’ has to construct his or her personal life story purelyin the imagination. Such internal action can be effectively supportedand will receive more impetus by the application of action and drama-techniques, within a supportive audience. Psychodrama can enable the‘‘protagonist’’ to literally step into his or her own personal universe.Following Bruner’s line of thought on the distinction between the

narrative and logico-scientific mode (in Hermans and Hermans-Jansen), one could say that both in the SCM and in psychodrama theemphasis is mainly on the ‘‘narrative mode,’’ although the ‘‘logico-scientific mode’’ is not absent. Still, psychodrama (by literally staging aspecific and concrete situation) probablymeets Bruner’s requirement ofthe narrative approach more explicitly: the narrative mode puts thegeneralities of the human condition into the particulars of experience,and attempts to locate experience in time and space (Hermans andHermans-Jansen: 123).In order to facilitate internal, dialectical self-organizing processes as

well as the individual’s personality development, therefore, a struc-tured application of action and drama techniques is suggested. Actiontechniques, however, are much more than simple action tricks. Theymust be grounded in an appropriate theory of personality and person-ality development. It is for this reason that the phenomenological–dialectical personality model (see above) was developed.

Conflicts and development: the application of action and drama techniques

There are various theoretical frameworks and methods for activatingthe I–Me reflection. The SCM on the basis of the Valuation Theory, onthe one hand, and action and drama techniques based on the develop-mental psychotherapeutic view outlined above, on the other, are bothvaluable methods. The latter, due to its action-orientated character and


by its direct affective-situational impact, seems best suited to triggerintensive affective processes in a multivoiced self. A short practicaldescription based on a variant of the psychodramamethod as describedby Moreno (Moreno, 1946, 1959; Moreno and Moreno, 1969; Verhof-stadt-Deneve, 1999) and Elefthery (Moreno and Elefthery, 1982) isgiven below.

Case presentationKevin (K) is a seventeen-year-old boy, the eldest of five children, withpoor results at school; he finds it difficult to make friends and is ratherhot-tempered. His father is an alcoholic and unemployed; he behavesaggressively towards his children and particularly towards his wife,and there are many conflicts between father and K. Mother cannotcope with her family and relies on K to support her.Specific event: father comes home drunk and wants his wife to give

himmoney. He bullies her and K gives his father a push so that he hasa nasty fall. The father turns out to be permanently paralyzed andwillnever be able to walk again. After this event, K becomes very with-drawn; he leaves home from time to time, plays truant and wandersaround aimlessly; he even makes an attempt at suicide. His father isrebellious but has stopped drinking. The mother takes care of herhusband to the best of her abilities. The two parents suffer from K’sbehavior.

One major task of the director of a psychodrama (Dir.) is to create anatmosphere providing the greatest possible feeling of security andunconditional respect for all group members. The idea is that throughan in-depth I–Me reflection participants themselves should (as in theSCM) find solutions to their own problems. This will be illustrated bymeans of some examples from sessions where K was the protagonist ina group of eight adolescents. The case study shows how, during thepsychodrama session, conflicting I–Me constructions appear, and howthis conflict results in psychological change.

Example 1 (inter-dimensional): Self-Image versus Alter-Image/Meta-SelfK shows clear signs of depression. As mentioned above, he has anextremely negative self-evaluation and strong feelings of guilt ( = Self-Image). Before action techniques are applied, the protagonist has tocreate for himself a concrete familiar situation.

Dir.: Kevin, you’re going to see your father now. Where do you wantto meet him? What time is it? What are you going to do?

K: He’s in the kitchen in his wheelchair, it’s 7 p.m. . . .Dir.: OK, tell us what the kitchen looks like . . . Where’s the door,where’s the worktop?

(K describes the kitchen, and a few chairs and a table are brought in.These simple objects considerably enhance Kevin’s affective in-volvement in this specific situation.)


Dir.: I know that it’s difficult for you to meet your father, but here youcan safely give it a try, you will not be alone, we will all help you. Ifyou like you can choose somebody from the group to represent yourfather (‘‘Antagonist’’).

K: OK, I take Bob, because he is the tallest of the group . . . (Bob entersthe inner group-space).

Dir.: OK, Kevin, come here and stand behind Bob. Youwill now try tobecome your father and tell who you are, using the I-form (= role-taking, cf. Mead: ‘‘taking the role of the other’’). Take your time,Kevin . . . Bob, youwill listen very carefully so that you knowhow torepresent Kevin’s father.

K: I am Kevin’s father. I am thirty-six. I am paralyzed, tired, I’ve lostall my strength and vitality. I watch TV all day long, I feel angry andsad. The only reason I go on living is my youngest son, Jerry, he’sthe only one left who loves me . . . ( = a part of Kevin’s Alter-Image:his father and brother).

This Alter-Image is comparable to Hermans and Hermans-Jansen’sI-position in a multivoiced self within an imaginal landscape. We nowproceed to Kevin’s Meta-Self (cf. Hermans and Hermans-Jansen: an-other ‘‘I-position’’):

Dir. (addressing Kevin in the role of his father): Well, father of Kevin,what do you think about your eldest son?

K (still playing the role of his own father): He is the cause of all thismisery . . . he beat me . . . I might have been dead . . . I wish I was . . .As a child he was such a lovely kid . . . But how much he haschanged! . . . Yet, I think I still love him . . . ( =Kevin’s Meta-Self).

(Kevin is finding it hard to cope and starts crying softly . . . )Dir.: OK, Kevin, take your time . . . Come here, you can now beyourself again, you are no longer your father . . . Who are you now?

K: I’m Kevin, I had a difficult moment just now. I don’t know why, Idon’t usually cry, but it came so suddenly! ( = again Kevin’s Self-Image).

The former actions are an example of a dialectical process, namely adouble negation movement following a three-stage course. The firststage (or thesis) hadKevin’s Self-Image as its startingpoint. Through therole taking in the second stage (or anti-thesis) the Self-Image is disre-garded and the Alter-Image and the Meta-Self are focused on. This isthemovement of the first negation (or the negation of the Self-Image). Kbecomes his own Alter-Image and Meta-Self, though historically theSelf-Image is still present. This vividly incites the cognitive–affectiveopposition between Self-Image and Alter-Image/Meta-Self. This strongexperience of opposition often goes together with a crisis or a catharsis.This was indeed the case when K, in the role of the father, felt hisfeelings for the father and the pain and the love of the father at the sametime. In the third stage (or synthesis), through the second negation (or


the negation of the Alter-Image and Meta-Self), the protagonist returnsto his own Self-Image.In psychodramatical practice this process is not at all a rigid three-

stage movement. Mostly there is a repeated toing and froing betweenthe different poles, which can lead to successive partial synthesisstages.For Kevin, after this dialectical movement, his Self-Image, his Alter-

Image, and his Meta-Self had changed mainly through the experienceof conflict in connection with the first negation. In the next session hesaid to the group members: ‘‘The guilt feelings are lessened and theimage of my father is not so negative anymore; I can now think abouthim quietly without being completely upset.’’What is important in this process is that the experiencing of a conflict

incited by the first negation can stimulate change and integration of thetwo opposite poles in the synthesis stage, or at least open them up to amore intense I–Me reflection.The above example describes the experience of oppositions between

images (or I-positions) of the person. The following example illustratesoppositions within one image. It still concerns Kevin’s relationshipwith his father. This illustrates how the opposition between internaland external Self-Images can be used.

Example 2: External versus internal Self-Image

(The following scene is a dialogue in the kitchen.)K: Pa, I’m off for a walk.Bob (as Kevin’s father): All right son, but be sure to be back beforemidnight.

Dir. (to K): Could your father say that?K: No, my father would never say that!Dir.: OK, reverse roles! Kevin, you’ll be your own father now. Bob,you’ll be Kevin, and you repeat the last few words he has just said.

Role reversal is necessary at this stage because we are proceeding onK’s subjective phenomenological constructions. The dialogue thereforemust fit perfectly with K’s usual environment, for if it does not wewould have a kind of role play with a much weaker affective andcognitive impact. The I constructs a ‘‘multivoiced world’’ but withinpsychodrama this construction is stimulated by action within a tem-poral and spatial, affectively involved situation.

Bob (as K): Pa, I’m off for a walk.K (as his father): Yes, do, and get out of sight for the rest of theweek aswell!

Dir.: Role reversal once again!K (as himself, addressing his father): Oh shut up! Look at yourself!


What did you do with your life . . . Misfit!Dir. (to K): What are you thinking of now? What do you feel? Go onestep to the left and try to say what is going on in your mind, yourfather will not hear you now.

(K nowhas the possibility of expressing his internal part acting from anew I-position, as if another person was speaking from a differentspace.)

K (thinking aloud): I feel bad, guilty . . . Why am I saying all this? Iknow he feels miserable . . . and strangely enough I feel pity for him. . . Sometimes I think I still love him . . . Is this the little boy in me?

For adolescents (and not only for adolescents) it can be much easier toexpress harsh and insolent behavior to significant others than affectiveapproach attitudes (of course, the reverse can also be true). By freezingthe external action (cf. ‘‘exterior dialogue’’) and stimulating the protag-onist to reflect on the simultaneous internal affects and thoughts (cf.‘‘internal dialogue’’) the protagonist can vividly experience the diver-gent poles. Comparablewith an intra-psychic role reversal he canmovefrom the exterior to the interior following the same three-stage processback and forth. Through this dialectical movement he can discover thatboth have sense and can enrich each other. K, for example, discoveredthat his harsh behavior towards his father was grounded in pain, guilt,anxiety, and love. Action and drama sessions are meant to offer anopportunity to express such hidden internal contents and to activelywork with them.This way of working is perfectly in tune with Hermans and

Hermans-Jansen’s reference to Watkins’ ‘‘Invisible Guests’’ andBakhtin’s ‘‘interior and exterior’’ dialogues referring to Dostoyevsky’snovel The Double. ‘‘The second hero (the double) was introduced as apersonification of the interior thought of the first hero (Golyadkin). Thisexternalization of an interior thought in a spatially separated opponentinstigates a fully developed dialogue between two relatively indepen-dent parties’’(ibid., p. 125). This was exactly what was put into practicewith K.

Conclusion

The notion that conflicts drive development is inherent in thephenomological–dialectical personality model. Action and drama tech-niques as applied within this framework are certainly not a panacea.Practice does show, however, that thanks to the combination of speak-ing, thinking, feeling, and acting, this method offers a very strongstimuluswithin the complex process of self-actualization and personal-ity development.Both the self-confrontation method and psychodrama techniques


can produce a picture of an individual’s personality structure and bothcan also engender personality development. The accent of the twomethods seems different, however. The self-confrontation methodprobably encourages a more systematic survey of the personality struc-ture at different developmental moments, whereas psychodrama,thanks to its strong and specific affective-relational, emotional, andcognitive involvement, probably has a more direct impact on personal-ity development.Through the construction of a concrete space and time dimension

within psychodrama, an extra relational accent can be created. Theprotagonist has the opportunity to objectify and exteriorize his or herimage of the antagonist. This encounter intensifies and surpasses theimaginary self-reflective dimension. Thanks to this intense physicaland mental action the protagonist is able to experience the self andsignificant others as part of amultivoiced self not onlymentally but alsoto meet and be those significant others in vivid cognitive–emotionalsituations in different times and spaces. Through the process of becom-ing the other, the protagonist feels the differences and disharmoniesintensively but also becomes aware of similarities and harmony be-tween Self and Other. Through action, personal identity is given anexplicit relational component.All this means that the internal psychodrama activated in the protag-

onist by the SCM, with the I as author and the Me as actor, can bestrongly affectively supported and intensified if the person as a multi-voiced self could effectively move from one I-position to the other, inorder to really meet and become the antagonists in a concrete time andspace experience. A combination of the Self-ConfrontationMethod andpsychodrama therefore offers considerable promise for the future.


COMMENTARY

Affective processes in amultivoiced self in actionLeni Verhofstadt-Deneve

Introduction

The theoretical framework ofHermans andHermans-Jansen, includingthe so-called valuation theory, is grounded in James’ well-known I–Meconcept of the self with the I as the reflecting ‘‘self-as-knower’’ and theMe as the reflected ‘‘self-as-known.’’ An important aspect of this view isthat the Me is much wider than the mere result of reflection on the self(in its narrow sense); it also contains the whole social and materialworld which has a certain personal value and affective significance forthe person. This view implies an I–world connection with a strong andcomplex relational and affective component in the self-reflecting per-son in present, past, and future situations. Hermans and Hermans-Jansen: ‘‘With its reflectivity and I–world connection, the self evokescomplexities of similar, different, and contrasting feelings, which to-gether form highly dynamic structures’’ (p. 122).An original turn (inspired by Sarbin) is the interpretation of the self

as a sort of narrative psychodramatic system in which the I and the Mefunction as ‘‘author’’ and ‘‘observed actor’’ respectively. The self asauthor (the I) constructs an imaginary story in which the Me functionsas (1) the protagonist, and (2) other figures that are part of Mine asantagonists (or imaginal others e.g. father, sister, partner). I presumethat Hermans and Hermans-Jansen do not restrict these antagonists tosignificant others but include the whole social and material world towhich the I attributes a personal affective value. From this point of viewthe I constructs a multivoicedworld in which the individual is not onlyable to talk about the variety of imaginal others (antagonists) but also totalk with them as relatively independent parts of an extended self. ‘‘Inother words, imaginal figures function as voiced positions in a self-space, and, at times, these positions entertain dialogical relationshipswith each other’’ (p. 126).It is interesting to note that this valuation theory functions as a basis

141

for the Self-Confrontation Method (SCM), an idiographic procedurewhich is constructed ‘‘to study the relation between valuations andtypes of affect and the way in which these variables become organizedinto one ormore systems’’ (p. 130). The procedure involves elicitation ofa set of valuations and subsequent association of these valuations witha standardized set of affect items.The valuation theory is highly attractive, as it is based on the organ-

ized but complex subjective phenomenological cognitive–affective nar-ratives of the person. In the elicitation of a set of valuations the personhas the opportunity to take diverse (also contradictory) I-positions,which over time can develop ‘‘into a more integrative Gestalt.’’However, one can question whether or not the SCM remains on a

merely cognitive, rational self-reflective level. Itmainly stresses a reflec-tion on cognitions and emotions from self and others and the dialoguesbetween those constructions. Even when in the SCM the focus is ex-plicitly on affects, what remains is mainly a reflection on affects and notnecessarily an intense involvement with those affects. The ‘‘psycho-drama’’ in the SCM exclusivelyworks on an internal level with a strongcognitive–rational accentuation.A similar limitation concerns the important aspect of conflict experi-

ence. For example: ‘‘Sharon did her self-investigation not from onecentralized position . . . Rather we approached Sharon as a multivoicedindividual, able to reflect on her self from two opposite positions andtell about significant meaning units from the perspective of these posi-tions’’ (p. 137) Real conflict experience is not at stake here, but thereflectionon and dialoguewith different I-positions seems again crucialfor personality integration.This highly cognitive accent (which no doubt has value of its own)

could effectively be enriched with cognitive–affective components bymeans of the introduction of real action and drama techniques, inwhich the protagonist not only reflects on, or engages in dialogue with,but also becomes diverse antagonists or imaginal others in a specifictime and space.In the following comments I will try to defend this thesis by first

explaining the importance of conflict for personality development andsecondly by illustrating how, in association with the SCM, differentsorts of conflicts could be worked with through the introduction ofaction and drama techniques.

The role of conflict in personality and personality development

Although Hermans and Hermans-Jansen do pay attention to the possi-bility of conflicts between different parts of the person, between differ-


ent I–Me positions, they mainly treat them from a clinical perspective,with a strong rational–cognitive accent. In this view conflict is not seenas the driving force in psychological development. In developmentaltheories such as that of Piaget (1949), conflict is seen as an essentialcharacteristic in the developmental process. Conflicts within the‘‘multivoiced self’’ can drive the individual’s development. Opposi-tions and conflict, from this perspective, can be considered a necessary(but not sufficient) condition for development. Such an interpretation isunderpinned by my own follow-up research, therapeutic practice, andby theoretical analyses on the basis of dialectical developmental psy-chology (Brown, Werner, and Altman, 1998; Conville, 1998; Verhof-stadt-Deneve, 1985; Verhofstadt-Deneve and Schittekatte, 1996).In this view development is seen as a self-organizing process, con-

sisting in dialectical developmental movements, which are inherent inthe qualitative development of each living organism (see also Piaget,1968; Prigogine and Nicolis, 1977). Or in the words of Piaget, referringto the dialectical course of cognitive development: ‘‘c’est une demarcheinevitable de la pensee’’ (p. 104). A comparable view is to be found inthe rather rigid epigenetic model of development as described byErikson (1968): every stage of life is built on a crisis, the content ofwhich is characterized by a positive and a negative pole. There is astrong resemblance here to Kelly’s personality theory and its bipolarconstructs as the cornerstones of an active and constantly evolvingsystem of constructs. The experiencing of oppositions causes the entiresystem of constructs to shift, so that change, activity and personalitydevelopment can take place (Kelly, 1955). An explicitly dialectical biasis also inherent in Vygotsky’s approach (1979; Zender and Zender,1974), since he looks upon crises and conflicts as criteria marking offthe stages of a dialectical development. It is worth noting that inVygotsky’s view the crisis (or negative phase) in the child’s psycho-logical development acquires a pronounced positive meaning. Thebasic concept of contradiction and conflict as stimuli for developmentalprocesses is related to Kohlberg’s (1976) notion of cognitive conflictas the stimulus of moral development. His way of thinking wassupported by Lawrence Walker’s (1983) experiments. Walker foundthat a range of conditions which trigger cognitive conflicts stimulatethe transition from one stage to the next in the moral developmentof children. In my theoretical framework dialectical movements andself-organization determine respectively the process and direction ofqualitative developments.


Conflicts between what?

The idea that conflicts drive development has been further elab-orated in my phenomenological–dialectical model of personality(Verhofstadt-Deneve, 1988, 1995, 1997, 1999, 2000). In this model thephenomenological aspect refers to the unique subjective content whichevery human being attributes to him- or herself and to the surroundingworld and is also based on the William James I–Me paradigm. Thedialectical aspect refers to the underlying process of interaction andconflict between these contents, and their development.Among the multitude of I–Me personality constructions six main

domains or images (comparable with fundamental I-positions inHermans and Hermans-Jansen) are focused on, each correspondingwith a specific question:

The Self-Image: ‘‘Who am I, with my potential and my short-comings?’’

The Alter-Image: ‘‘What are the others like?’’The Meta-Self: ‘‘How is one’s own person viewed by the others?’’

These three basic images are counterbalanced by three ideal-images:

The Ideal-Self: ‘‘Who would I like to be or become?’’The Ideal-Alter: ‘‘What should the others be or become like?’’The Ideal-Meta-Self: ‘‘How should the others perceive me?’’

For each image a distinction can be made between an external aspect(the way we behave externally) and an internal aspect (what we thinkand feel at the same time). In this model the person can also be seen as a‘‘multivoiced self’’ with six main parts which can interact and be inconflict.On the basis of themodel,many types of conflict can be distinguished

according to the nature of the opposing poles. I shall just give someexamples:

– There may be inter-dimensional oppositions, for example betweenthe Self-Image and the Meta-Self.

– Similarly, theremay exist intra-dimensional oppositions, for examplebetween external and internal contents of one dimension.

– Oppositions may be experienced between the subjective phenom-enological constructions of oneself and of theworld, on the one hand,and possible alternative interpretations, on the other.

– Vivid oppositions can also be felt between the three time dimensions.

Within the scope of these comments I will, in the presentation of actionand drama techniques, confine myself to the first and the second kinds


of oppositions. For more information I refer to Verhofstadt-Deneve(1995, 1997, 1999, 2000).

How to stimulate conflicts and development?

As Hermans and Hermans-Jansen have shown, the use of the SCM cantrigger self-organizing processes in the ‘‘multivoiced self’’ by means ofthe stimulation of cognitions, emotions, and language. This happens onan internal level, but from my clinical experience and didactic groupworkwithmy students I have learned that direct spatial, external actionin the context of an intense and supportive group situation can lendvital support to such internal self-organizing processes. To me it seemsthat the SCM is too exclusively aimed at an internal, mental action level:the ‘‘protagonist’’ has to construct his or her personal life story purelyin the imagination. Such internal action can be effectively supportedand will receive more impetus by the application of action and drama-techniques, within a supportive audience. Psychodrama can enable the‘‘protagonist’’ to literally step into his or her own personal universe.Following Bruner’s line of thought on the distinction between the

narrative and logico-scientific mode (in Hermans and Hermans-Jansen), one could say that both in the SCM and in psychodrama theemphasis is mainly on the ‘‘narrative mode,’’ although the ‘‘logico-scientific mode’’ is not absent. Still, psychodrama (by literally staging aspecific and concrete situation) probablymeets Bruner’s requirement ofthe narrative approach more explicitly: the narrative mode puts thegeneralities of the human condition into the particulars of experience,and attempts to locate experience in time and space (Hermans andHermans-Jansen: 123).In order to facilitate internal, dialectical self-organizing processes as

well as the individual’s personality development, therefore, a struc-tured application of action and drama techniques is suggested. Actiontechniques, however, are much more than simple action tricks. Theymust be grounded in an appropriate theory of personality and person-ality development. It is for this reason that the phenomenological–dialectical personality model (see above) was developed.

Conflicts and development: the application of action and drama techniques

There are various theoretical frameworks and methods for activatingthe I–Me reflection. The SCM on the basis of the Valuation Theory, onthe one hand, and action and drama techniques based on the develop-mental psychotherapeutic view outlined above, on the other, are bothvaluable methods. The latter, due to its action-orientated character and


by its direct affective-situational impact, seems best suited to triggerintensive affective processes in a multivoiced self. A short practicaldescription based on a variant of the psychodramamethod as describedby Moreno (Moreno, 1946, 1959; Moreno and Moreno, 1969; Verhof-stadt-Deneve, 1999) and Elefthery (Moreno and Elefthery, 1982) isgiven below.

Case presentationKevin (K) is a seventeen-year-old boy, the eldest of five children, withpoor results at school; he finds it difficult to make friends and is ratherhot-tempered. His father is an alcoholic and unemployed; he behavesaggressively towards his children and particularly towards his wife,and there are many conflicts between father and K. Mother cannotcope with her family and relies on K to support her.Specific event: father comes home drunk and wants his wife to give

himmoney. He bullies her and K gives his father a push so that he hasa nasty fall. The father turns out to be permanently paralyzed andwillnever be able to walk again. After this event, K becomes very with-drawn; he leaves home from time to time, plays truant and wandersaround aimlessly; he even makes an attempt at suicide. His father isrebellious but has stopped drinking. The mother takes care of herhusband to the best of her abilities. The two parents suffer from K’sbehavior.

One major task of the director of a psychodrama (Dir.) is to create anatmosphere providing the greatest possible feeling of security andunconditional respect for all group members. The idea is that throughan in-depth I–Me reflection participants themselves should (as in theSCM) find solutions to their own problems. This will be illustrated bymeans of some examples from sessions where K was the protagonist ina group of eight adolescents. The case study shows how, during thepsychodrama session, conflicting I–Me constructions appear, and howthis conflict results in psychological change.

Example 1 (inter-dimensional): Self-Image versus Alter-Image/Meta-SelfK shows clear signs of depression. As mentioned above, he has anextremely negative self-evaluation and strong feelings of guilt ( = Self-Image). Before action techniques are applied, the protagonist has tocreate for himself a concrete familiar situation.

Dir.: Kevin, you’re going to see your father now. Where do you wantto meet him? What time is it? What are you going to do?

K: He’s in the kitchen in his wheelchair, it’s 7 p.m. . . .Dir.: OK, tell us what the kitchen looks like . . . Where’s the door,where’s the worktop?

(K describes the kitchen, and a few chairs and a table are brought in.These simple objects considerably enhance Kevin’s affective in-volvement in this specific situation.)


Dir.: I know that it’s difficult for you to meet your father, but here youcan safely give it a try, you will not be alone, we will all help you. Ifyou like you can choose somebody from the group to represent yourfather (‘‘Antagonist’’).

K: OK, I take Bob, because he is the tallest of the group . . . (Bob entersthe inner group-space).

Dir.: OK, Kevin, come here and stand behind Bob. Youwill now try tobecome your father and tell who you are, using the I-form (= role-taking, cf. Mead: ‘‘taking the role of the other’’). Take your time,Kevin . . . Bob, youwill listen very carefully so that you knowhow torepresent Kevin’s father.

K: I am Kevin’s father. I am thirty-six. I am paralyzed, tired, I’ve lostall my strength and vitality. I watch TV all day long, I feel angry andsad. The only reason I go on living is my youngest son, Jerry, he’sthe only one left who loves me . . . ( = a part of Kevin’s Alter-Image:his father and brother).

This Alter-Image is comparable to Hermans and Hermans-Jansen’sI-position in a multivoiced self within an imaginal landscape. We nowproceed to Kevin’s Meta-Self (cf. Hermans and Hermans-Jansen: an-other ‘‘I-position’’):

Dir. (addressing Kevin in the role of his father): Well, father of Kevin,what do you think about your eldest son?

K (still playing the role of his own father): He is the cause of all thismisery . . . he beat me . . . I might have been dead . . . I wish I was . . .As a child he was such a lovely kid . . . But how much he haschanged! . . . Yet, I think I still love him . . . ( =Kevin’s Meta-Self).

(Kevin is finding it hard to cope and starts crying softly . . . )Dir.: OK, Kevin, take your time . . . Come here, you can now beyourself again, you are no longer your father . . . Who are you now?

K: I’m Kevin, I had a difficult moment just now. I don’t know why, Idon’t usually cry, but it came so suddenly! ( = again Kevin’s Self-Image).

The former actions are an example of a dialectical process, namely adouble negation movement following a three-stage course. The firststage (or thesis) hadKevin’s Self-Image as its startingpoint. Through therole taking in the second stage (or anti-thesis) the Self-Image is disre-garded and the Alter-Image and the Meta-Self are focused on. This isthemovement of the first negation (or the negation of the Self-Image). Kbecomes his own Alter-Image and Meta-Self, though historically theSelf-Image is still present. This vividly incites the cognitive–affectiveopposition between Self-Image and Alter-Image/Meta-Self. This strongexperience of opposition often goes together with a crisis or a catharsis.This was indeed the case when K, in the role of the father, felt hisfeelings for the father and the pain and the love of the father at the sametime. In the third stage (or synthesis), through the second negation (or


the negation of the Alter-Image and Meta-Self), the protagonist returnsto his own Self-Image.In psychodramatical practice this process is not at all a rigid three-

stage movement. Mostly there is a repeated toing and froing betweenthe different poles, which can lead to successive partial synthesisstages.For Kevin, after this dialectical movement, his Self-Image, his Alter-

Image, and his Meta-Self had changed mainly through the experienceof conflict in connection with the first negation. In the next session hesaid to the group members: ‘‘The guilt feelings are lessened and theimage of my father is not so negative anymore; I can now think abouthim quietly without being completely upset.’’What is important in this process is that the experiencing of a conflict

incited by the first negation can stimulate change and integration of thetwo opposite poles in the synthesis stage, or at least open them up to amore intense I–Me reflection.The above example describes the experience of oppositions between

images (or I-positions) of the person. The following example illustratesoppositions within one image. It still concerns Kevin’s relationshipwith his father. This illustrates how the opposition between internaland external Self-Images can be used.

Example 2: External versus internal Self-Image

(The following scene is a dialogue in the kitchen.)K: Pa, I’m off for a walk.Bob (as Kevin’s father): All right son, but be sure to be back beforemidnight.

Dir. (to K): Could your father say that?K: No, my father would never say that!Dir.: OK, reverse roles! Kevin, you’ll be your own father now. Bob,you’ll be Kevin, and you repeat the last few words he has just said.

Role reversal is necessary at this stage because we are proceeding onK’s subjective phenomenological constructions. The dialogue thereforemust fit perfectly with K’s usual environment, for if it does not wewould have a kind of role play with a much weaker affective andcognitive impact. The I constructs a ‘‘multivoiced world’’ but withinpsychodrama this construction is stimulated by action within a tem-poral and spatial, affectively involved situation.

Bob (as K): Pa, I’m off for a walk.K (as his father): Yes, do, and get out of sight for the rest of theweek aswell!

Dir.: Role reversal once again!K (as himself, addressing his father): Oh shut up! Look at yourself!


What did you do with your life . . . Misfit!Dir. (to K): What are you thinking of now? What do you feel? Go onestep to the left and try to say what is going on in your mind, yourfather will not hear you now.

(K nowhas the possibility of expressing his internal part acting from anew I-position, as if another person was speaking from a differentspace.)

K (thinking aloud): I feel bad, guilty . . . Why am I saying all this? Iknow he feels miserable . . . and strangely enough I feel pity for him. . . Sometimes I think I still love him . . . Is this the little boy in me?

For adolescents (and not only for adolescents) it can be much easier toexpress harsh and insolent behavior to significant others than affectiveapproach attitudes (of course, the reverse can also be true). By freezingthe external action (cf. ‘‘exterior dialogue’’) and stimulating the protag-onist to reflect on the simultaneous internal affects and thoughts (cf.‘‘internal dialogue’’) the protagonist can vividly experience the diver-gent poles. Comparablewith an intra-psychic role reversal he canmovefrom the exterior to the interior following the same three-stage processback and forth. Through this dialectical movement he can discover thatboth have sense and can enrich each other. K, for example, discoveredthat his harsh behavior towards his father was grounded in pain, guilt,anxiety, and love. Action and drama sessions are meant to offer anopportunity to express such hidden internal contents and to activelywork with them.This way of working is perfectly in tune with Hermans and

Hermans-Jansen’s reference to Watkins’ ‘‘Invisible Guests’’ andBakhtin’s ‘‘interior and exterior’’ dialogues referring to Dostoyevsky’snovel The Double. ‘‘The second hero (the double) was introduced as apersonification of the interior thought of the first hero (Golyadkin). Thisexternalization of an interior thought in a spatially separated opponentinstigates a fully developed dialogue between two relatively indepen-dent parties’’(ibid., p. 125). This was exactly what was put into practicewith K.

Conclusion

The notion that conflicts drive development is inherent in thephenomological–dialectical personality model. Action and drama tech-niques as applied within this framework are certainly not a panacea.Practice does show, however, that thanks to the combination of speak-ing, thinking, feeling, and acting, this method offers a very strongstimuluswithin the complex process of self-actualization and personal-ity development.Both the self-confrontation method and psychodrama techniques


can produce a picture of an individual’s personality structure and bothcan also engender personality development. The accent of the twomethods seems different, however. The self-confrontation methodprobably encourages a more systematic survey of the personality struc-ture at different developmental moments, whereas psychodrama,thanks to its strong and specific affective-relational, emotional, andcognitive involvement, probably has a more direct impact on personal-ity development.Through the construction of a concrete space and time dimension

within psychodrama, an extra relational accent can be created. Theprotagonist has the opportunity to objectify and exteriorize his or herimage of the antagonist. This encounter intensifies and surpasses theimaginary self-reflective dimension. Thanks to this intense physicaland mental action the protagonist is able to experience the self andsignificant others as part of amultivoiced self not onlymentally but alsoto meet and be those significant others in vivid cognitive–emotionalsituations in different times and spaces. Through the process of becom-ing the other, the protagonist feels the differences and disharmoniesintensively but also becomes aware of similarities and harmony be-tween Self and Other. Through action, personal identity is given anexplicit relational component.All this means that the internal psychodrama activated in the protag-

onist by the SCM, with the I as author and the Me as actor, can bestrongly affectively supported and intensified if the person as a multi-voiced self could effectively move from one I-position to the other, inorder to really meet and become the antagonists in a concrete time andspace experience. A combination of the Self-ConfrontationMethod andpsychodrama therefore offers considerable promise for the future.


CHAPTER 7

Old–new answers and new–oldquestions for personality andemotion: a matter of complexityJeannette Haviland-Jones, David Boulifard,and Carol Magai1

We are very fortunate to see the beginning of an exciting time inpsychology, an era in which we may bring together complex intuitiveideas with appropriate technical tools for expanding them. This may beparticularly interesting in the area of emotion and personality develop-ment. The present authors have described changes in theories of ident-ity and emotion from dichotomous sets to systems of sets across thetwentieth century (Haviland and Kahlbaugh, 1993) and have alsoexamined the impact of emotion on personality development (Magaiand McFadden, 1995). Here we intend to take a further step in thisdirection with some analyses of the development of identity, or self-concept, in two individuals as these are interwovenwith emotion.Witha dynamic mathematical model of emotion we will also show howemotional processes may be variable without the necessity for variablecognitive content in the analysis of the stimulation that led to anemotional change. We will use our examples to show how differentcontexts at the origin of a fearful stimulus might lead either to fear or toanger without any necessity for awareness by the individual.In this chapter we will first briefly review a few instances of our past

research in which we employed some of the older tools of categor-ization and linear analysis. We will use these examples to demonstratethe questions that arose that we could not answer with these tools. Forillustrativepurposeswewill then present two cases usingAnne Frank’sdiary and Albert Ellis’ autobiography and we will focus on the impactof one emotion on self-concept or identity development, that of fear andits neighbors such as anxiety and terror. These cases will offer cluesabout the functions of emotion in identity and lead us to consider asimple dynamic model for emotion, again focusing on fear, to bring usto new questions.

151

Old answers

At the very beginning of our research, there was some debate aboutwhether emotions could be studied at all as various writers referred toemotions as epiphenomena. To refute these points, we and other re-searchers began with a very concrete system that involved classifyingfacial movements into categories according to the systems of Ekman(1972), Izard (1971), and our own early coding system (Haviland, 1976).Using these categories ofmovementwe could track the development ofemotion expressions in the young infant (Haviland and Lelwica, 1987;Malatesta, Culver, Tesman, and Shepard, 1989) and into old age(Malatesta-Magai, Jonas, Shepard, and Culver, 1992).The categorical and linear approaches provided interesting answers

to some old questions. The categories provided us with insight intomother–child interaction and the development of individual and famil-ial emotional expression. For example, we found that infants matchtheir mothers using different muscles in a familial way. If the mother’sface was primarily active in the brow region, her baby was likely todevelop that pattern. We also found that very young infants respondwith different facial expressions to their mothers’ repeated expressionof a particular emotion such as anger or sadness or happiness but donot necessarily match their mothers. For example, many babies ‘‘froze’’when their mothers acted angrily. So we learned that emotional pat-terns and developing responses were sometimes responsive to simpleassociative connections and sometimes not. How this might be relatedto later appearing identity was quite obscure and why they sometimesled to one response and sometimes to another was also unclear.Partly because we did not have particularly good tools for following

continuity in development, we shifted to an older age group, workingmore directly on emotion and identity in early adolescence. We knewby the time we designed this project that adolescents’ self-perceptionof their emotional lives changed developmentally (Stapley andHaviland, 1989). Younger children are oriented to how concrete eventsinfluence their feelings – toys breaking, presents arriving, sportsgames won, and so forth. In early adolescence there is a dramaticchange. Older children orient toward their relationships with peers toexplain their most salient emotional experiences. Later in adolescence,young people orient toward their personal identity issues as instiga-tors of salient emotion. Memorable emotional experiences may comefrom self-discovery and self-exploration, self-expectation and self-monitoring. But the diversity of emotional experiences among individ-uals begins to emerge here as well. To what this diversity might beattributed was still not clear.

152 Jeannette Haviland-Jones, David Boulifard, and Carol Magai

These changes in self-report again suggest that there are qualitativelydifferent meanings to emotional experience for a variety of reasons andthat acquiring a particular set of meanings for emotional experiencemight not be a simple linear process. The investigator cannot alwayscount on a dark and stormy night to inspire fear or a broken doll toinspire anger or a family birthday party to inspire happiness. This is notto claim that there are no fairly universal events, only that they arelikely to be limited to extremes – death, mutilation, new life, and soforth. We still do not know quite how to approach this, so the data thatwe have tell a story that we know is too simple. Consider that it isnot only the self-perception of emotion that changes during this time,but also the prominence of the type of emotion that is expressed(Kahlbaugh andHaviland, 1994). Genuine signals of emotional distanc-ing emerge. The youngest adolescents hide their faces with their hair orhands or turn away or examine their fingers during conversation.Under the very same circumstances, older adolescents add more con-fronting expressions of distance such as contempt – eye rolling, block-ing with arms and legs, and even snorting and head tossing. Thesechanges appear to parallel in some way the changes in the adolescents’sense of identity but they are not conscious, controlled changes. Thisresearch alerts us to levels of subconscious change. It also alerts us tothe possible importance of repetitive, brief expressions in establishingthe styles or tone of personality and of dialogue, even when people areunaware of the changes at the level that we are noting them. Perhapssmall iterations of emotion expression across changing contexts lead tolarge changes in the organization of identity and interpersonal interac-tion. Perhaps such iterations even lead to categorical shift rather thanmere enlargement of the same category. This suggests that it is notnecessarily a gross and acknowledged change in the adolescent’sknown social status in society, but could easily be an accumulation ofsmall, subconscious changes in the adolescents’ emotions or moti-vations that is responsible for major changes in identity as well as socialstatus.In our next studies we tried to get a picture of the changes in relations

among self-defined roles and emotional traits. We asked children andyoung adolescents to tell us about their personal and social roles andtheir personality traits, particularly their emotional ‘‘traits.’’ To get asense of how they related these aspects of their self-concept to eachother we asked them to consider themselves in each role and to tell ushow likely each trait was to emerge in that role. For example, onlyshyness might emerge as the emotion for a new role such as ‘‘enteringhigh school’’ whereas many emotions might emerge in a familiar rolesuch as ‘‘daughter’’ – anything from anger to shyness to happiness and

153Old–new answers and new–old questions

sorrow. Did adolescents just assimilate new roles to old emotionalstructures – adding the high school role to the daughter role – or is therea different sort of shift?Theway that one analyzes such networks of data is to run a hierarchi-

cal clustering program (hiclas; see De Boeck and Rosenberg, 1988) thatorders the roles so that roles with many defining traits are consideredinclusive of most of the self’s possibilities and thus ‘‘higher’’ in theself-concept hierarchy. Roles with very few traits are subsumed by theroles that are associated with more traits. It is also possible to have setsof roles that do not fit together but form separate compartments or evento have roles that do not fit into the hierarchy but become residuals. Inany case, the hierarchical clustering technique presents one with apicture of the relations that different roles have to each other in terms oftraits.We found that ten- and eleven-year-old children had elaborate hier-

archies of roles and traits that were quite predictable. The family rolestended to be the most inclusive, with school and friend roles secondaryand very few unique roles such as ‘‘Auntie’s favorite.’’ In adolescence,however, the familiar roles such as daughter actually lost many of theirpositive emotions and traits while retaining their more negative ones.Because they lost traits, such ‘‘familiar’’ roles were unlikely to remainthe most salient or at the top of the hierarchy, as they had been inchildhood. New roles such as ‘‘apprentice beautician’’ emerged thatdid not fit well in the hierarchy and that did not seem as complex as theold ones in terms of their inclusiveness of possible traits. Roles droppedout and were not replaced by complex ones or inclusive ones. This maybe a vortex during which unpredictable changes occur and the place ofemotional traits seems to play a pivotal role. Clearly we could not use asimple linear additivemodel. At best, a rise and fall model might work,but what emerged after the fall was not the same as what existedbeforehand.Our previous research seemed to tell us that we would not be able to

detect how emotional processes entered into identity when takingaccount of data collapsed across many people. Trying to sum acrossvaried time lines for emotions, roles, interactions, and changing self-perceptions for many people would obscure the developmental pro-cesses. It began to seem that we need to take one case (the x case) andthen the next case (x+ 1) (see Haviland and Kramer, 1991) and so forth,looking for the rules of process, not summation. We were being led todynamic processes and network or matrix analyses by the lack oflinearity in the data themselves.


Questions about identity and emotion: two special cases

Anne FrankAt the beginning of our analyses of Anne Frank’s diary we predictedthat we would see how the emotion of fear became a major attractor forevents and thoughts, a significant organizer of her self-concept oridentity.We assumed that the war circumstances of her life terrified heroften and perhaps continuously. We predicted that a search for thewords that indicated fear would show that they became more frequentand more related to diverse corners of Anne’s life. Eventually fearwould become an attractor in her self-concept. Using a simple modelfor the development of an anxious clinical syndrome we predicted thatwhen fearful things happen, the person is sensitized towatch for fearfulevents, the slightly fearful aspects of unrelated events are exaggerated,the person develops an interpretation of life that revolves entirelyaround fear and anxiety, and so forth. Does Anne Frank’s diary showevidence of such an associative process? Is this a reasonable approachto conceptualizing fear – to think of it as a stable, global category, asingular thing?We know quite a bit about changes in Anne Frank’s emotional

expression and also about her styles of cognitively solving everydayproblems from an earlier analysis (Haviland and Kramer, 1991). Bytracing change across time we had already discovered that cognitivedevelopment was preceded by changes in emotional expression. How-ever, across the diary as a whole, different types of emotional expres-sion tended to be more or less related to different cognitive forms,suggesting two processes. On the one hand, the rise in emotionalitymight predict a kind of disorganization or loosening of previous con-nections, just as we had seen in the cross-sectional studies. On the otherhand, singular emotions, experienced repetitively, seemed to have thepotential to become organizers that could affect mental processes. Forexample, when Anne was expressing her fears she had a tendency touse relativistic logic or thinking.We learned, contrary to our prediction,that the number of references to fear did not rise and fear did notbecome more variously associated with diverse events. Did it notchange its place at all? What is faulty about our predictions?We divided Anne’s diary into four sections so we could examine the

self-concept structure at each time and see whether fear was a signifi-cant organizer of the self-concept structure at any point. In each sectionwe coded traits, especially emotional traits, and people or events thatwere described by or associated with the traits. We created data sets towhich we could apply the hierarchical clustering techniques as we had


done in the previous cross-sectional studies on identity and the place ofemotion. In the first section, before the Franks go into hiding, we founda matrix that resembled those collected from other children. Anne’sroles were daughter and student, sister and friend. The events werefamily- and school-related with just a subordinate pocket for war wor-ries. Most of her roles were related to both negative and positiveemotional traits and the most frequent traits were more positive thannegative. Anne was a fairly typical child in a happy family at thebeginning of the diary. Her thoughts about bombs, the dark of night,being Jewish, death, and war were highly negative and fearful; how-ever, these thoughts made a separate small cluster or section in herself-concept, a cluster that was not integrated with her day-to-day life,nor an integral part of her self-concept.In the next section of the diary inwhichAnne describes the beginning

of their life in hiding, her childhood roles are emotionally disconnected.Her daughter role is largely negative with an emphasis on anger; herschool and friend roles only exist as memories or as she learns orfantasizes what the grim fate of her school friends has been. These rolesemphasize depressive emotions. Happy moments form no repeatingcluster. There is no longer any well-elaborated, complex set of roles forAnne that attracts both positive and negative emotions. She has movedinto a period of segmented or disintegrated role structure, aware ofbeing pulled by different sides of her emotional self in different scenes.The cognitive shift into a more relativistic style in which there was nocertainty and no possibility of establishing certainty along with resol-ution of that relativism into absolute statements became common. Thecluster of fearful war thoughts also is missing. The elements of it areseparate but visible in the mix of fearfulness, anger, and sadness – thebeginning of a complex depression syndrome – as if the fear hadrecruited other emotions, or perhaps they recruited fear. The war is notentirely outside in the night to be afraid of and separate from herself,but has begun to be a small part of hermemories of better times. It is notan attractor for anything, however.In the third slice that we took from Anne Frank’s diary there is the

beginning of a new organization of self around her relationship withher friend in the annex, Peter, and her sister, Margot. Both of theserelationships have happy episodes and contents mixed with anger inMargot’s case and with some fearfulness as regards Peter and her newsexual feelings. A large, diverse cluster now surrounds sadness (notfear) – dead friends, memories of flirting with boys, outsiders, justice,her mother’s feelings of sorrow. Her self-concept is still segmented –one fairly broad and diverse cluster surrounds sadness with lowerlevels of other negative emotions and smaller clusters are related to


different issues. Fear is once again associated directly with bombings,terror of discovery. It was still not a major organizer.In the last section of Anne Frank’s diary a more interrelated self-

concept is developing, one that looks in its organization more like theearliest structure. But this new self-concept is far removed from thehappy child we knew at the beginning of the diary. Anne’s roles withPeter and her annex family are complex and high in the hierarchy ofclusters. What is surprising is that the most positive parts of herself arenow intertwinedwith thewar cluster, to hope for its ending, and beliefsabout the ideal life that will follow the end of the war. The least positiveor most dreaded aspects of Anne’s self-concept are in another clusterthat contains her memories of her old self and her old life, her judg-ments of her own mood regulation and her personal criticism of hertwo-sided identity.The beginning self-concept in the diary was set around external

objects and people and events in Anne’s young life. The ending self-concept was set around internal and idealized distant objects. Therewere intervening glimpses of separated clusters of feelings and roles.None of this supported a linear process of change. None of this sup-ported our original prediction about fear, either. When we come to theendof thediary,Anne’smostpositive feelingsand thoughtsarecenteredin a wished-for and idealized future that is intertwinedwith her fearedand mourned war experiences. Her most elaborate dreaded self is notassociated with actual frightening war experiences, but with her ownability or lack of abilities to control hermoods and express her true self.Is fear an organizer of either Anne’s idealized or self-evaluative

self-concept clusters? Or is fear a process that is open to changing intoother emotional states, highly variable, non-linear in its process? Fear iscertainly contained in the idealized cluster and its absence may helpdefine the second, but one cannot argue easily that fears are simplymultiplying in Anne’s identity – in spite of her terror through manydays and nights. There is an accumulation of depressing events, and anaccumulation of Anne’s attempts to counter and control her depressivereactions. Perhaps because she cannot escape to a bigger world, beingfrustrated by the restraints of living in the annex, she turns to an inwardexamination of controlling depression and anxiety. Perhaps the con-straint, independent of her thoughts about it, changes her long-termfeelings.

Albert EllisIn the case of Albert Ellis we have another opportunity to ask howfearful events might emerge to command a significant place in identity.Thiswork on a particular emotion and its place in identity comes from a


larger study by Magai and Haviland-Jones (in press). We have exten-sively analyzed the ideology, attachment patterns, and emotions ofAlbert Ellis over several decades of his life. Ellis is exemplary as aperson with a very stable adult self-concept that is associated with fearand the management of fear. Over many years, Ellis assimilates newinformation to a successful identity structure. He does not go throughdetectable episodes of disorganization and reorganization in his adultyears.Fear (fright, terror, horror, panic, anxiety – the whole panoply of

words that Ellis repeats over decades of writing) is a particularlycompelling and toxic emotion for him. It is usually an emergencyemotion that, when chronically activated, is physiologically enervatingand can be ultimately deadly (Selye, 1956). It may have played animportant role even in Ellis’ childhood illnesses and his chronic adultsomatic problems, but it obviously did not prevent him from living along and productive life, as he is still working and now is in his ninthdecade. This gives rise to the possibility that fearfulness has many facesor functions.Ellis’ (1972) autobiography reveals that he believes he developed

coping strategies when very young that were the forerunners of histheory of Rational Emotive Therapy (RET). Around the age of four, hebegan drilling into his head rules such as: ‘‘Life is full of hassles youcan’t control or eliminate’’ (i.e. Things are terribly, frighteningly, un-controllable) or ‘‘Wait before you panic’’ (i.e. Panic is imminent). Ellis’childhood rules are obviously tools for managing forms of fearfulness.They are remarkable because there is no expectation that the sources offright might be located and eliminated; in fact, the first rule is that theycannot be eliminated. The problem is therefore identified as the emo-tional feeling, as part of oneself, not as the external stimulus for theemotion. This is the sort of identity–emotion interaction that we sawbeginning in Anne Frank. For Anne, the terror and sadness of the warbecame issues of personal mood control as she began to judge herselfbased on her emotional control. Even more clearly for Ellis, fears andfearlessness are personal traits and need to be controlled internally butthis seems to carry even more weight than we had anticipated.In his autobiography, Ellis describes many roles and describes these

roles with many traits, including many emotional traits. From this weextracted a self-concept model, as before. We took some liberties toreduce the number of traits by coding together ones that seemed to beused to indicate the same thing. For example, enjoyment and happywere grouped together; stupid, idiotic, and so forth were groupedtogether.Nevertheless, we ended upwithmore roles and traits thanwehad for Anne, probably indicative of a long and rich life. Again, using


hierarchical clustering methods, we developed a picture of how rolesand traits were associated. For example, Ellis’ clients and himself-in-a-crisis are clustered together because they share traits of anger (includ-ing hostility) and psychoticism (including emotionally disturbed).There are no roles that are superordinate to this cluster, meaning thatEllis’ descriptions of clients are not a piece of his description of someother role. Fear is the trait that is superordinate to this cluster, aspredicted. As it turns out, fear is equally well associated with all thedefined roles except for himself-as-a-therapist and himself-as-a-child.These last two roles are distinguished by their associationwith fearless-ness, confrontation, logic, shyness, and happiness, and are not relatedto fearfulness. In other words, all of Ellis’ roles in the hierarchy arecharacterized either by fear or by the management of fear. The possibleexceptions are some relationship roles that do not fit in any cluster inthe hierarchy. This could indicate that the matrix is not as good a fit aswewould like – perhaps that we need more data, or possibly that somemore complex aspects of Ellis’ identity do not fit this type of matrix. Orit could indicate that many of Ellis’ roles are not systematically relatedto other roles. In any case, the earlier prediction that fearful experiencescould organize self-concept does appear to be upheld in this caseleading us to wonder why it is so different.Fear for Ellis, because it is a superordinate emotional trait, does not

differentiate roles well; it is omnipresent in the identity construction.Ordinarily we think of fear as restricted to certain types of situations,such as fear of strangers, of illness, of heights, and so forth, as itoriginally was stimulated by the threat of bombs in Anne Frank’saccount. But if fear can associate with any situation, there might bemore variation in adult development than we previously suspected (onemotional biases in personality, see Magai and McFadden, 1995).Ellis’ autobiography is a history of stoicism and cleverness triumph-

ing over what anyone else might construe as wretched early circum-stances. He almost single-handedly battled forces of neglect, crosseddangerous streets, went to school at an extraordinarily young age,superintended his two younger siblings, and later in young adulthoodcured himself of painful shyness, social inhibition, and sexual inepti-tude. In addition to managing this collection of hardships, Ellis hadmajor physical trauma. When he was five he was hospitalized fortonsillitis, developed a severe streptococcal infection that requiredemergency surgery to save his life and culminated in nephritis. He hadrecurrences of the nephritis, developed pneumonia, and was hospital-ized repeatedly. He remembers being in the hospital for almost anentire year during which he was left largely alone for days at a time.Rather than recounting the negative feelings he might have had about


this time, he related how he took a scientific interest in the proceedingsof one operation, which allowed him to watch pus coming out of hisabdomen.From the account of Ellis’ early years one might predict depression

and anxiety but Ellis emphasizes his successful management of thesefeelings. The organization of roles and emotional traits with fear at itscenter tells us that in order to enable this management, Ellis becamevigilant for fearful events, finding evidence of manageable fear in allaspects of his life and continuing to find fear in new roles that wereadded as he reached maturity. These instances of fearfulness are separ-ated from each other in his self-concept, however. As a homely meta-phor one can imagine reciting all the meals one could have with cheeseas an ingredient, but onewould not imaginemixing them all together tohave the ultimate in cheese dishes. Cheese is in every dish, but it has adifferent array of tastes and textures in each one. Similarly, Ellis haskept his experiences of fear, ubiquitous though they appear to be,separate and manageable.One of the striking things about Ellis’ autobiography is clear evidence

of disconnection as an organization strategy, operating at many levels.In fact, he opens the autobiographywith an assertion of disconnection:‘‘I do not believe that the events of my early childhood . . . oriented meto becoming the kind of individual and the type of therapist that I nowam’’ (p. 103). And yet elsewhere he ascribes the origins of RET tochildhood ideas. For another example, there are a number of inconsis-tences in the narrative that leave a disconnected picture of his family.Ellis’ engagement and disengagement in love affairs in his early adult-hood also mirror the segmentation or separation he values. Even in hiscognitive strategies, as analyzed from his theoretical contributions,there is a strong preference for absolute, separable analyses rather thanintegrated systems of analysis. Segmentation and separation are apattern across the various modes of looking at Ellis and fear appears tobe the key. Fear is not just a simple reaction, here and gone, nor is itsimply additive. Something is happening to the experience of fear thatis changing its nature and its associations, so it is the backbone of aparticular style of information processing.Ellis claimed that his manageable experience of fear happened at the

level of gross intention – that he decided at the age of four to think offear in this manageableway. Shouldwe be persuaded by this? He is notlikely to be aware of the continuing ubiquitous, if transformed, functionof fear in his self-concept, nor is he likely to be aware of its associationwith social-cognitive strategies of separation. The belief that one has (ordoes not have) an emotional trait without necessarily knowing whatprocess might have led one to that conclusion, or how its very existence


is controlling one’s thoughts about it, is not restricted to our two cases.As Ohman (1993) pointed out, the source of fearful reactions may beperceived only subconsciously. When the person reacts fearfully, buthas no conscious grasp of the connections, he believes that the source ishimself and further believes that he is irrational – that there is no realsource. The paradox is that such a person may in fact become anunusually fearful person in the sense that he becomes hypersensitive topartial cues of potentially fearful things.Thus far, in our use of old questions and modes of analysis we have

stayed at a gross level of categorization in our analyses. We haveconsidered emotions, traits, thoughts, intentions, goals, beliefs, and soon as things that we could identify, separate, and hold constant. Therehas been some usefulness to this level of analysis but there are puzzlinggaps in our predictions at this level.What couldwe learn about emotionand identity that would take us beyond such often faulty global predic-tions? To focus on only one problem, why might an Albert Ellis headtowards fearfulness and an Anne Frank head towards depression? Ifwe were to narrow our focus and expand our view of one aspect of thisprocess, if we could use a microprocess to analyze just ‘‘fear’’ and itspossible transformations,whatmight change in our understanding andcould we then reintegrate with what we find at the grosser level?

Microprocesses of fear: one special case

When we think of emotions in dynamic systems terms, we necessarilybegin to move from gross categories of emotions, thoughts, beliefs, andso on, to microprocesses. What kinds of process would we like to positfor fear? Fear most often seems to be the result of exposure to informa-tion that is increasingly intense and which overloads the perceptualand cognitive system. As Tomkins (1962) might have explained it, it is aglobal response to increasingly rapid and intense change from theperson’s point of view, change that cannot be assimilated. ‘‘Our theoryposits three discrete classes of activators of affect, each of which furtheramplifies the sources which activate them. These are stimulation in-crease [startle, fear, interest], stimulation level [sadness, anger], andstimulation decrease [joy] . . . such a set of mechanisms guaranteessensitivity to whatever is new, to whatever continues for any extendedperiod of time and to whatever is ceasing to happen, in that or-der’’(Tomkins, 1962: 252).Tomkins suggested that all emotions, not just fear, were activated

by changing patterns of neural firing, by which he meant changingreceptivity, reactivity, and behavioral activity. Later on we willrefer to emotions that involve stimulation increase or decrease as


‘‘change-based’’ emotions and those that involve stimulation level as‘‘level-based’’ emotions. While most theories of emotion have eithercategories or dimensionality, but not both, they usually do not allowcontinuous transition from one emotion to another. Tomkins suggestedthat one emotional state could dissolve or morph into another. Hisproposal is analogous to color perception. Continuous change in thewavelength of light moves in and out of categories of color becausereceptors have special sensitivity curves. The stimulation is continuous,but the categories occur because of receptor qualities. Our emotionmodel could operate similarly with either receptor specificity or bio-chemical rate of change specificity and not violate modern theories ofemotion. Fear occurs because the match in stimulation (itself notnecessarily a categorical event) matches the receptors for fearfulness.Neural processes are linked to emotion (see Feyereisen, 1989; LeDoux,1989; Panksepp, 2000a) and the ‘‘amplification’’ of sensory informationresults in behavioral changes which in turn produce more informationbased on the latest summary of stimulation (including the internalprocess as a source of stimulation) so that emotion is a continuousprocessing variable as well as a potential state.The ‘‘color’’ model of emotion does move us a bit away from naive

psychology’s intuitive understanding of emotions as separate catego-ries of information, but we can come back to the categories just as wecan come back to colors. In this model, qualities of the information suchas its intensity and rate of intrusion on the processor are part of themessage, potentially leading to particular processing routes. It wouldnot just be the content of what is assimilated, but qualities of theassimilationprocess that influence emotional responses. In our case, toomuch information intruding too quicklywould be a source of anxiety orfear no matter what the content of the information might be. Panksepp(2000a) also proposes that there are different neurological systems orcircuits for different emotions such as the ‘‘Rage,’’ ‘‘Fear,’’ and ‘‘Lust’’circuits; interestingly, he notes that the effects of stimulation and activa-tion of one circuit can spread to another.In the physical sciences we have learned that to study physical

process one looks at rates of change and what or how different ‘‘forces’’induce different rates. For example, to understand the acceleration ofbodies in classical physics, time derivatives are used to model howforces are related to the motion of bodies. Influences (for us, this mightbe sensory input) on rates of change (e.g. patterns of sensory process-ing) show up ultimately in the state of systems, and the states then maybe described in categories (e.g. discrete emotions). It is always possiblethat human systems do not work in ways analogous to such physicalsystems, but testing the models will reveal this in the long run.


An advantage of amicromodel for continuous emotional processes isthat it allows us to demonstrate complex interactions that the computercan follow farther with iterations than we can predict analytically withgrosser categories. Perhaps there are multiple outcomes to situationsthat beginwith the same conditions,manymore thanwe can anticipate.Partly, this may occur because randomeffects aremagnifiedwithmanyrepetitions, as they tend to be in dynamic systems, especially morechaotic systems. (While our present model does not take all thesepossibilities into account, it lends itself to including them.) The systemswe have been using historically to study emotions are very simple,categorical, and linear, as we have shown. This probably severely limitsour understanding of individual differences.We have attempted to model emotion processes based on the above

framework and assumptions. The basic idea in our model is that eachemotion is activated by a particular pattern of stimulation and that suchpatterns feed back upon themselves, producing dynamic systems.Either the level of information or stimulation from low to high or therate of change in the amount or density of information defines theemotion pattern.To make a demonstration model of this we chose a time series, the

familiar sine wave, because it goes through the full range of theoreticalpossibilities for the model – a range of accelerating rates, deceleratingrates, and some relatively stable levels. We then defined boundaries foreach emotion along the time series, based largely on our interpretationof Tomkins’ original formulations, as they have been refined by oursubsequent years of research. For example, fear was defined as occur-ring along a band of rapidly accelerating stimulation. Anger was de-fined such that it would occur when the wave stabilizes at a certainintensity and happiness was defined as being stimulated by a band ofdeceleration.In this model, any change in sensation occurring internally or exter-

nally could change the emotion, which raises the question of timeboundaries in a general sense. Frijda and colleagues (e.g. Frijda,Mesquita, Sonnemans, and Van Goozen, 1991) propose that people areaware of moods, if not emotions, as lasting days and weeks. On theother hand, discrete emotions theorists argue that ‘‘real’’emotions lastat most a few minutes and that most expressions can be measured inseconds or microseconds (e.g. Ekman, 1972). We suggest two answersto the problem of duration that can occur in our model. First, wesuggest that durations are related to the type of emotion (Malatesta-Magai and Culver, 1991). Emotions of level – that is, emotions that tellus that nothing is changing (e.g. depression) may be of any duration,but emotions of change, such as surprise, would necessarily be of


shorter duration. So part of the model would have to take the type ofemotion into account. On the other hand, people’s ways of reportingemotions also must be taken into account. A frequently occurringemotionmay be perceived as a constant emotion. Different people maydiffer in their awareness and in the sophistication of their reporting ofdifferent categories of emotion, as Ellis is highly sensitive to cues ofanxiety. Some of the differences may be largely a matter of experience,culture, and training. It might be noted that awareness and ability torecognize emotions is another amplifier of information and hence couldbe construed as an additional stimulus for emotion.Next, we intend to give the reader an example so that some benefits

of modeling emotions as microprocessors might emerge and so that wecan reflect on the questions that remain in our cases of Anne andAlbert.In our example, we suggest that a storm (the stimulus), with lightningfollowed later by a rumble of thunder, is in the vicinity of our subject,called Anne–Albert. We will show how our model predicts the emo-tional experience. This is a Gedanken experiment – neither Anne norAlbert has actually reported this experience. We somewhat arbitrarilymodeled the stimulus change of lightning and thunder events bytracing the general shape of the lightning flash and that of the ensuingthunder clapwith a normal probability curve and a chi square probabil-ity density curve. Superimposed on that curve are deviations obtainedfrom random draws from a normal distribution. These manipulationsmake the stimulus irregular and convey some of the chaotic nature ofreal-world stimuli rather than working with the smoothing over ofinformation into large categories that we usually prefer.Figure 7.1 shows the patterns of rate change and level in the amount

of information in the storm at the top. Then it shows the effect of thatstimulus on the processing of the storm as change-based emotions andas level-based emotions. In particular, when the stimulus intensity rises(e.g. lightning flashes) the level of neural response (density of neuralfiring) will rise. The rate of change will be directly proportional to thedifference between the stimulus intensity and current level of neuralresponse, but inversely proportional to the difference between themaximum possible level and the current level. We adopted this func-tional form because we thought the response to sensory inputs shouldreflect not only the magnitude of these inputs but also the finite re-sources of the response system. (For a fuller description: Haviland-Jones, Magai, and Boulifard, unpublished.) The level of processedinformation always lags slightly behind the sensory input to producefear or surprise and the iterativemodel requires that prior state influen-ces the processing of the present stimulus, so the curves are not exactlythe same.


0

20

40

60

80

100

0 1 2 3 4 5 6 7 8

ST

IMU

LU

S

0

20

40

60

80

100

0 1 2 3 4 5 6 7 8

DN

F L

EV

EL

0

50

100

150

0 1 2 3 4 5 6 7 8

LE

VE

L-B

AS

ED

EM

OT

ION

DISTRESS

ANGER

PAIN

0

50

100

150

0 1 2 3 4 5 6 7 8

TIME, secs

CH

AN

GE

-BA

SE

D E

MO

TIO

N

JOY

INTEREST

FEAR

STARTLE

Figure 7.1. Emotion model time series for storm stimulus

The power of this simple little model lies in the interesting things ittells us about the complex stimulus and the emotions evoked. Forexample, we can see in the graphs that the lightning evokes roughlyequal amounts of startle, fear, and interest but relatively little distress


and essentially no anger or pain. There is even some momentary joy asthe lightning quickly dissipates. On the other hand, the thunder evokesmostly distress and anger.Why does the model work as it does? What you see in the lightning

event is a rapid onset and release of the sensory input. It occurs tooquickly for the level of information to rise to the required thresholds forrelative stability in level-based negative emotions. What you see in thethunder event is a greater duration of high sensory input levels whichallows the information level to reach these thresholds for distress andthen anger and to sustain those values. To generalize, the model sug-gests that brief but dramatic changes result primarily in change-basedemotions. On the other hand, slower and more pervasive stimuli resultin level-based emotions, if the persistent level of stimulation is highenough.Thismodel provokesmany interesting speculations. Obviously, if we

change the parameters of the model even a little (as one might whencomputing individual differences) there could be shifts in the catego-ries of emotions produced. A change in parameter that affected the rateof processing could move fear towards more interest, for example, justas a change in threshold for constant information could change oneperson’s frequent anger to another person’s more frequent distress.This model also demonstrates that information processing can be re-lated to emotional categories without any need for prior categoricallearning or self-evaluation. One does not need to have been thrashedduring a thunder storm in order to be angry. It also shows how verysmall differences in original state or in the stimulus itself could makelarge categorical differences that would have implications both forimmediate behavior and long-term identity scripts.If the model we propose is even close to veridical it would help to

make sense of one of the central problems in emotional processing –namely that similar stimuli evoke different emotions in differentpeople. These different responses have generally been attributed tocognitive processes, as Ellis attributes his management of fear to hispowerful decision-making abilities.While this is one source of variationin the process, in ourmodel there could bemany others. These thoughtsled us to a simple variation in the emotion example.Not only may two people respond differently to the same stimulus,

but even the same person will not be likely to respond the same waytwice to repeated exposure. Our model handles this problem by mak-ing the previous state influence the processing of new information.Suppose, for example, youngAnne–Albert was already angrywhen thenext thunder and lightning storm unexpectedly hit. Instead of havingthe storm sequence impinge upon a tabula rasa as we had proposed in


0

20

40

60

80

100

0 1 2 3 4 5 6 7 8

DN

F L

EV

EL

0

20

40

60

80

100

0 1 2 3 4 5 6 7 8

ST

IMU

LU

S

0

50

100

150

0 1 2 3 4 5 6 7 8

LE

VE

L-B

AS

ED

EM

OT

ION

DISTRESS

ANGER

PAIN

0

50

100

150

0 1 2 3 4 5 6 7 8

TIME, secs

CH

AN

GE

-BA

SE

D E

MO

TIO

N

JOY

INTEREST

FEAR

STARTLE

Figure 7.2. Emotion model time series for storm while upset

the first instance, we now have the subject previously engaged in apersonal internal ‘‘storm’’ at the point of origin for ourGedanken experi-ment. When we look at the figure 7.2 we see that this makes interestingdifferences. This time there is very little change-based emotion because


there is less difference between the new and background stimulus. Fearand surprise, as well as joy after the storm, are reduced because of thechange in the original state of the system. On the other hand, distressand anger are pervasive and pain is increased during the lightning aswell as the thunder. This little dynamic system for describing fearshows that when the state of origin is rather angry and the externalstimulus is fearful, the resultantmost frequent response can be distress.Anne, with her anger at having to adapt to the restrictions of the annex,may have beenmore in this situation than Albert who was free to roamand experiment, with the exception of his hospital stays. When theoriginal conditions for processing are different, it is possible to showthat fearful stimuli exacerbate distress, not fear and not necessarilyanger, though both fear and anger are possible with a slight change inconditions. This very simple dynamicmodel of emotional process leadsto testable hypotheses that do not emerge obviously from simple ana-lytical approaches of gross categories.We couldplaywith other parameters in this simple dynamicmodel to

reflect individual differences, ad infinitum.We could change the level atwhich particular emotions shift categories. We could change the rangeover which each emotion is defined. We could get more complicatedanduse different parameters for each categoryof emotion. It is also easyto conceive of different slopes for transitions from one emotion toanother. One of the remarkable things about the model is how shiftingthese parameters – manipulating small, continuous variables – changesthe categories of emotions, often in ways that we had not predicted.Already, shifting our conceptualization of emotion to a dynamic systemand focusing on a microlevel process reveals whole new types ofquestions that we were unlikely to ask about developmental emotionalprocess just using grosser categorical approaches. These new questionseither bypass or precede the old ones because they do not require thatprocesses include awareness and synthesis of contextual information.

Concluding thoughts

In our earlier work we asked the old questions and tried to understandidentity processes and their relation to emotion by using categoricalmethods of analysis and linear assumptions of additivity. In the presentwork,we abandoned these assumptions and approached the issue froma dynamic systems framework and now we find new questions. Wemodeled an emotion process mathematically with a microprocessor.We found that the distinction between classes of emotion and dimen-sions of emotion is perhaps an artificial one. We saw that informationparameters havingmainly to do with rates of change and levels of state


are, in themselves, capable of eliciting changing emotion categories.Wesaw that conscious or goal-oriented knowledge was not even funda-mental to the shift. This may tell us something aboutwhy our semantic/categorical sorting approaches would be bound to miss certain occur-rences and reasons for surface appearance changes.There is often acknowledgment that emotions, considered as large

summaries of contextual interaction, influence a wide variety of devel-opmental processes that are collapsed under the term ‘‘identity.’’Coming from one end of the spectrum and embedding this idea in theirown dynamic model, we have Kunnen and Bosma (2000) putting‘‘conflict’’ in their dynamic model of identity as a meaning-makingdevelopmental phenomenon for identity. ‘‘Conflict’’ is one way ofconceptualizing where emotional process enters into the identity sys-tem, but it is as yet highly non-specific and does not include our conceptof emotion as an ongoing process. Could positive emotional processesalso have developmental potential for change, asMagai andMcFadden(1995) have suggested? Perhaps they enter the Kunnen and Bosmamodel as the ‘‘support’’ functions. Such positive processes could bevery significant forAnne Frank andAlbert Ellis, for both of whom thereare many happymoments and events, as well as clear conflicts. But thismodel does not explain why Anne with many ‘‘conflictful’’ eventsbecomes somewhat depressed, whereras Albert with many ‘‘conflict-ful’’ events becomes sensitized to fear. This difference in emotional biashas a large effect on identity.Coming from a very different level of analysis, we see the work of

researchers like Marc Lewis (1995) also suggesting that emotions aid inthe association of concepts, causing them to form semantically mean-ingful systems leading to self-concept. Lewis argues that some aspect ofemotional experience enters into a larger, self-organized form not un-like the self-concept structureswe presented for Anne Frank andAlbertEllis. Once again, at even this more limited or perhaps more process-oriented level where we consider the interaction of two proposedindependent systems – the emotional and the cognitive – we arestill missing the possibility that different emotions might enter theprocesses in different ways to produce not only different outcomes butalso different associative processes. It is possible that the emotional andcognitive systems are not independent. Such a possibility is closer to theconclusions from themodel of emotional experience we have proposedinwhich brief emotional changes, occurring beyond a conscious level ofawareness, are the dynamic and continuous part of more general infor-mational processing activities. The difference between a human thinkerand a machine thinker (at this point in time) would be that the machinethinkerwould always use the sameways of combining bits and solving


problems, whereas the human processor would shift the mode ofcombining and solving problems because it is dependent upon the rateand intensity of the information. This shift would be called the emo-tional phase of the process.The concept of ‘‘phase shift,’’ useful in dynamic systems approaches,

suggests that an existing system or state can shift in the sense thatevents or concepts that were stable and related to each other in particu-lar patterns in one phase now change. When Anne Frank’s self-conceptspace shows evidence of change it seems that it could have been due torepeated emotional experiences and that this would lead to a realiz-ation of changing opportunities and threats. Those realizations wouldthen lead to further change, but are not required to initiate change. Itwas just this sort of concept that we tried to access in our microproces-sor model of emotions, admittedly at a very low level. These phaseshifts offer enormous flexibility if the phase shifting lends itself tolong-term adaptation under diverse circumstances, even though itmight also lead to shorter periods of disorganization in self-concept.In our little model, when we showed emotional responses to the

stormmoving from one emotion to another – in brief and not necessar-ily conscious processes – we meant to illustrate that emotional processat this level instantiates phase shifts in attending to the stimulus. Thesevery brief emotional phases are pieces of a response pattern in whichdifferent processes might be signaled and brought into play in Anne’sor Albert’s reaction to, and possibly even understanding of, stormyevents.Beliefs and knowledge about emotions can operate to change emo-

tion aswell, in that such beliefs are themselves a part of the information,not just a way of managing emotion. Telling oneself not to jump at thecrash of thunder because it is not dangerous is probably minimallyeffective in preventing fear. If one wanted to diminish fear and knewabout the process, then singing loudly before the thunder is probablyeffective, but then one might respond with increased distress, as ourmodel showed. Having a meta-affective knowledge of oneself and ofemotional process in general may enable one to change many situ-ations, thereby changing the emotional process but not always in theway that one predicts. The rise in distress is not intuitively obvious, forexample.When we look back on earlier analyses of Anne Frank and Albert

Ellis we might have been persuaded that we could understand theprocesses by asking about them at the semantic level. For example, Ellishas an explanation for developing his particular self-concept and forplacing emotions within a very controlled cognitive system. AnneFrank also has explanations for the two sides to her self-concept – an


inner one with true feelings and an outer one with managed feelings.As reasonable as these explanations seem to them, our microprocessexample encourages us to be more cautious in interpreting emotional-ity as a global construct or as one that is understood by the personherself or himself. Emotional processes may be about as accessible toconsciousness as blood pressure fluctuation, even though the conse-quences of emotional or of blood pressure change may be perceivedand explanations for such change concocted. Our secondary analysis ofhowwe come to be as we are maymake a good linear story, but it is notnecessarily true especially at the level of understanding emotionalprocess and how it is placed in the identity process of change andstability.As far as we know, we are the first researchers to propose that

emotions might be modeled as continuous processes in dynamic sys-tems. An examination of the many possible brief reactive as well asactive emotional processes reveals much about emotional process thatwould not be predictable from more categorical or linear theories.Dynamic systems have the potential to trace enormous quantities ofcomplex data and are made barely possible by the latest technology.This chapter opens up possibilities for us all in how we approach ourquestions about self-concept and emotion. Shall we ask ourselves or thesubject what we believe or shall we study microscopic changes? Obvi-ously it makes a difference. The difference has led us to propose thatdifferent forms of emotion are responses to different phases of informa-tion process both internally and externally, and that this constant phaseshifting occurs on a microlevel as well as on the semantic meaning-making level. The iteration of these processes leads to the changingformations of self-concept.

Note

1 This work was partly supported by a grant to Carol Magai from theMinorityBiomedical Research Support Program and the National Institute of Aging(1 SO6 GM54650-01).


COMMENTARY

Emotions as sources ofinformation about the selfPeter G. Heymans

Haviland, Boulifard, and Magai propose interesting points of view onthe study of the self and emotions, and illustrate their positions withadmirable research. I especially admire their efforts to present parts ofthe life trajectories of real people and illuminate their plights by ana-lyzing underlying processes. The main plea, as I see it, is an appeal toleave the traditional methods of analysis in the field and switch tonon-linear model building accompanied by computer/spreadsheet-based simulations of the model. Such a model-based simulation of theprocessing of fearful stimuli has been presented by Haviland et al. todemonstrate the power of non-linear models over classical approaches.The Haviland et al. model predicts the course of the intensities of sevenemotions over a time period of eight seconds in the life of a single(hypothetical) individual.Below I will try tomitigate the enthusiasm in the Haviland et al.-plea

by pointing out four warnings, extending beyond the Haviland et al.model. They concern: (1) anchoring the model-time unit in real devel-opmental time; (2) the Haviland et al. (and my own) plea that theanalysis of individual developmental trajectories should be accom-panied by a compatible type of (statistical) treatment of the simulatedtrajectories; (3) the turn toward non-conscious processing of self-relevant information at microlevel, which is implied in many non-linear dynamical models, should not divert attention from solvingunanswered questions about the cultural construction of the self and itsdevelopment; (4) classical methods of analysis have not (yet) beenexhausted in advancing our understanding of the development of theself, they were just not applied to the right type of data.

Time in the modeling and simulation of developmental processes

To construct a model one needs to combine in a suitable way assump-tions about the processes involved. In order to compare predictionsfrom the model with observed data parameters have to be estimatedquantitatively. When the derivation of predictions from the model is

172

mathematically cumbersome, simulations of the model-reality bringrelief. One of the more important decisions for the model builder is thechoice of a timescale in which model-reality unfolds. Is time measuredin (milli-)seconds, hours, months or years? Of course we know that aday has 24× 60× 60 seconds. But the interesting ripples in simulateddevelopmental curves which are compared to parts of an observeddevelopmental curve might occur ten thousand seconds after startingthe process, while in real life a similar ripple might be observable afterten thousand days. Moreover, at which point in the lifecourse does thesimulated process start? Anchoring the time-unit and the starting pointmore firmly in real-life phenomena would greatly increase the value of(non-linear) model simulations in developmental psychology. In theHaviland et al. simulations presented time runs in seconds, startingfrom a well-defined event (lightning followed by thunder). But ingeneral, we do not know (yet) which life events can serve as triggers fordevelopmental processes.

The estimation of the simulated developmental trajectories

Usually several (ten-)thousands of process-unfoldings are run. Theresults of all these runs are averaged to obtain an estimate of thesimulated developmental curve(s). These multiple runs can show quitevariable results, due to the random components which are assumed tobe part of the model’s assumptions. In fact the unexpectedness ofprocess progression, given certain combinations of parameter values, iswhat seems to attract many researchers. The outcome of any singlesimulation-run is to be considered as one possible realization of adevelopmental trajectory. As such, a single run is comparable to a (partof a) life trajectory of a single individual. The simulated developmentalcurves presented in many textbook applications are thus averaged overmany individual fates. (Note that this averaging is not the case in theHaviland et al. example.) Indices for variation around these meanvalues should reflect (a) inherent variability in the phenomenon understudy, and/or (b) systematic inter-individual differences. Usually onlythe mean simulated values are given, variances not. In the case ofsimulating habitualways of responding of single individuals, variancesaround the mean developmental curve can only refer to inherent un-predictablity of the process.Haviland et al. seem to downplay this unpredictabilitywhen they are

showing how the previous emotional state influences an individual’semotional reactions (comparison of figures 7.1 and 7.2). Systematicinter-individual differences in emotional responding between a hypo-thetical Anne and Albert are supposed to be explainable by assuming

173Commentary on Haviland-Jones, Boulifard, and Magai

that (1) both individuals react emotionally according to the Haviland etal.model, and (2) both individuals differ in the habitual emotional state,which serves as partial input to the Haviland et al. model, and (3) thishabitual emotional state is the resultant of previous life history. Itremains however to be explained how this habitual emotional state hasbeen built up out of the thousands of residues of emotional encounterswith life events. There is certainly a possibility for extending the pres-entedHaviland et al. model for present-day concurrent emotionswith amechanism for cumulating residues of emotional encounters into ahabitual mood-state. In this last part I see definitely a place for non-linearmodeling, as thesemodels allow for catastrophe-like bifurcationsin the course of the unfolding of a developmental process. Indeed, is itnot the singular, unique encounters with events that individuals lateron report as having been so influential in their lives?

The cultural construction of the self and its development

‘‘Shifting our conceptualization of emotion to a dynamic system andfocusing on a microlevel process reveals whole new types of questionsthat we were unlikely to ask about developmental emotional processesjustusinggrossercategoricalapproaches’’ (Havilandetal., p. 167). I tendto agree, but warn – in my self-chosen role of skeptic – also of thepossibility that results obtained at that microlevel might be difficult torelate to the grosser level of observable behavior. Ties between levels ofanalysis have to be demonstrated as well. Haviland et al. proceed bystating that ‘‘These new questions . . . bypass . . . the old ones becausethey do not require that processes include awareness and synthesis ofcontextual information’’ (emphasis added). Indeed,muchof the success ofpsychology consists in having shown that the causal sequence of eventsleading up to a human perception, judgment, decision or act is notwhatit appears to be to the individual concerned, but follows a sequencewhich is (partly) hidden from awareness.However, it is also an undeni-able fact that people are aware (erroneously or not) of some connectionspredisposing to behavior, and that people tend to act on the basis of thatawareness, especially in social interaction.An individual’s identitymayserve as an example. Let us focus upon identities connected to havingsuccesfully completed so-called developmental tasks, e.g. as describedby Erikson’s lifespan theory which can be found in every introductorytextbook in (developmental) psychology. Not many people are awarethat Erikson’s theory embodies a combination of two powerful culturalnarratives. The first is the narrative of man on his earthly way toredemption and heaven being distracted by Cardinal Sins (with anorigin in fourth-century contemplation experiences of monks in the

174 Peter G. Heymans

Syriandeserts). The secondnarrativedepicts life as a journey in searchofthe realization of an ideal inwhich the hero is hindered by obstacles putby representatives of Evil; by fighting and overcoming the obstacles thehero receives honors from the audience (i.e. the narrative of Romance,Murray, 1985; Heymans, 1992). Many people map their life experienceson such a (combined) narrative, or are being mapped when a psychol-ogist analyses their life from an Eriksonian perspective; the benefit ofsuch a mapping is order in one’s set of experiences. Personhood andmeaning spring from this order-generating mapping (Heymans, 1994).Let us have a closer look at the developmental task of adolescence:overcoming the ‘‘identity-crisis,’’ made famous by Erikson and hisfollowers. Erikson assigned the Cardinal Sins to a specific phase in life,one for each phase, and used an 8-Sin instead of the standard 7-SinDoctrine (Capps, 1989). Cardinal Sinswere thought bymedievalman tobe conditions which hindered spiritual progress; these sins could beavoided or overcome by acquiring sin-specific virtues. Erikson postu-lated that the Cardinal Sin typical of adolescence was Pride, sometimesthought of by Church Fathers to be the basis of all other Cardinal Sins.Thevirtue to be acquired as antidote againstPride (a tendency to requirecontinuous adulation from others) was Fidelitas (fidelity) conceived aspersonalpredictabilityover situations.Thisfidelitas comesquite close tothe demonstrationof personality traits, which psychologists conceive astrans-situational behavioral consistencies. For Erikson, these situationsare mainly diverse social roles which the adolescent is offered forenactment. Solution of the identity-crisis is nothing else than demon-strating relative cross-situational stability of behaviors. When the adultworld arranges for a fifteen- to seventeen-year-old to enter into a varietyof new situations (roles), none of the persons involved is aware that theyare setting up a test of fidelitas for the adolescent. The cultural modelstructuring the interactions and the judgments is beyond awarenesswhen adults and the adolescent are ‘‘conspiring’’ to make part of theCardinal Sin model come true again. This example points – I hope – tothe need to invest also in searching formesolevel processes occurring atthe nexus of the individual and his social context. Microlevel modelingshould not distract researchers from what happens at the interface ofculture and the developing person. Incantation-theorywas proposed toclarify such processes (Heymans, 1994, 2000).

Classical methods of analysis have not been exhausted

The emphasis on non-linear dynamical models in developmental psy-chology has directed attention to the need for (multiple) timeseries datafrom single individuals (or other units of analysis, e.g. child–mother


dyad). There is no substitute for such data as a basis for statementsabout intra-individual, patterned changes. The near reflex-like applica-tion of psychometric methods of analysis (developed to serve a cleardescription of inter-individual differences) has led to a depreciation ofindividual case data, which until now were not frequently collected.Classical methods can help the developmental psychologist to uncoveraspects of the developing self, given that multiple timeseries from asingle individual are available. Space restrictions permit only someindications of what is possible.Emotions indeed are a powerful source of information about the self.

Therefore, part of my favorite research design is to collect timeseries ofthe intensities of multiple emotions (between 40 and 140) from singleindividuals, over long stretches of time (e.g. between 60 and 260 con-secutive days). This is the same type of data that can serve in tests ofnon-linear dynamical models. However, I use classical methods, allwithin the SPSS/Timeseries Statistical Software package. Calculation of(partial) autocorrelations within an emotion timeseries can tell aboutperiodicity in this emotion, just as does spectral analysis. Cross-laggedcorrelations between timeseries for different emotions tell whether oneemotion leads to another emotion, and after how many days. Mostimportantly, by clustering the patterns of activation-over-time ofmultiple emotions from a single individual the resultant emotion clus-ters can be interpreted as ‘‘concerns’’ in the sense of Frijda’s (1986)emotion theory, and the activation of these concerns over time can bedescribed. Internal dynamics of these concern-activations over time canbe described by ARIMA-models (see the SPSS/Timeseries package).When information about life events is also available it is possible todetermine their effects on concern-activations, e.g. by adding to therelevant ARIMA-model a regression model representing this effect.Cross-lagged correlation analysis of the ARIMA-whitened concern-activation scores will give information about the intra-individual inter-nal dynamical relations between concern-activations. For an example ofthese analyses see Heymans (2000).In conclusion: the best short-term consequence of the increased atten-

tion to non-linear dynamic models in developmental psychology is theneed to revalue and collect (multiple) timeseries data from single cases.Inmy opinion, these cases should be individuals-in-their-sociocultural-context, and the data should comprise also indicators for processessupposed to occur at the nexus between developmental context andindividual.


COMMENTARY

Emotions as sources ofinformation about the selfPeter G. Heymans

Haviland, Boulifard, and Magai propose interesting points of view onthe study of the self and emotions, and illustrate their positions withadmirable research. I especially admire their efforts to present parts ofthe life trajectories of real people and illuminate their plights by ana-lyzing underlying processes. The main plea, as I see it, is an appeal toleave the traditional methods of analysis in the field and switch tonon-linear model building accompanied by computer/spreadsheet-based simulations of the model. Such a model-based simulation of theprocessing of fearful stimuli has been presented by Haviland et al. todemonstrate the power of non-linear models over classical approaches.The Haviland et al. model predicts the course of the intensities of sevenemotions over a time period of eight seconds in the life of a single(hypothetical) individual.Below I will try tomitigate the enthusiasm in the Haviland et al.-plea

by pointing out four warnings, extending beyond the Haviland et al.model. They concern: (1) anchoring the model-time unit in real devel-opmental time; (2) the Haviland et al. (and my own) plea that theanalysis of individual developmental trajectories should be accom-panied by a compatible type of (statistical) treatment of the simulatedtrajectories; (3) the turn toward non-conscious processing of self-relevant information at microlevel, which is implied in many non-linear dynamical models, should not divert attention from solvingunanswered questions about the cultural construction of the self and itsdevelopment; (4) classical methods of analysis have not (yet) beenexhausted in advancing our understanding of the development of theself, they were just not applied to the right type of data.

Time in the modeling and simulation of developmental processes

To construct a model one needs to combine in a suitable way assump-tions about the processes involved. In order to compare predictionsfrom the model with observed data parameters have to be estimatedquantitatively. When the derivation of predictions from the model is

172

mathematically cumbersome, simulations of the model-reality bringrelief. One of the more important decisions for the model builder is thechoice of a timescale in which model-reality unfolds. Is time measuredin (milli-)seconds, hours, months or years? Of course we know that aday has 24× 60× 60 seconds. But the interesting ripples in simulateddevelopmental curves which are compared to parts of an observeddevelopmental curve might occur ten thousand seconds after startingthe process, while in real life a similar ripple might be observable afterten thousand days. Moreover, at which point in the lifecourse does thesimulated process start? Anchoring the time-unit and the starting pointmore firmly in real-life phenomena would greatly increase the value of(non-linear) model simulations in developmental psychology. In theHaviland et al. simulations presented time runs in seconds, startingfrom a well-defined event (lightning followed by thunder). But ingeneral, we do not know (yet) which life events can serve as triggers fordevelopmental processes.

The estimation of the simulated developmental trajectories

Usually several (ten-)thousands of process-unfoldings are run. Theresults of all these runs are averaged to obtain an estimate of thesimulated developmental curve(s). These multiple runs can show quitevariable results, due to the random components which are assumed tobe part of the model’s assumptions. In fact the unexpectedness ofprocess progression, given certain combinations of parameter values, iswhat seems to attract many researchers. The outcome of any singlesimulation-run is to be considered as one possible realization of adevelopmental trajectory. As such, a single run is comparable to a (partof a) life trajectory of a single individual. The simulated developmentalcurves presented in many textbook applications are thus averaged overmany individual fates. (Note that this averaging is not the case in theHaviland et al. example.) Indices for variation around these meanvalues should reflect (a) inherent variability in the phenomenon understudy, and/or (b) systematic inter-individual differences. Usually onlythe mean simulated values are given, variances not. In the case ofsimulating habitualways of responding of single individuals, variancesaround the mean developmental curve can only refer to inherent un-predictablity of the process.Haviland et al. seem to downplay this unpredictabilitywhen they are

showing how the previous emotional state influences an individual’semotional reactions (comparison of figures 7.1 and 7.2). Systematicinter-individual differences in emotional responding between a hypo-thetical Anne and Albert are supposed to be explainable by assuming


that (1) both individuals react emotionally according to the Haviland etal.model, and (2) both individuals differ in the habitual emotional state,which serves as partial input to the Haviland et al. model, and (3) thishabitual emotional state is the resultant of previous life history. Itremains however to be explained how this habitual emotional state hasbeen built up out of the thousands of residues of emotional encounterswith life events. There is certainly a possibility for extending the pres-entedHaviland et al. model for present-day concurrent emotionswith amechanism for cumulating residues of emotional encounters into ahabitual mood-state. In this last part I see definitely a place for non-linearmodeling, as thesemodels allow for catastrophe-like bifurcationsin the course of the unfolding of a developmental process. Indeed, is itnot the singular, unique encounters with events that individuals lateron report as having been so influential in their lives?

The cultural construction of the self and its development

‘‘Shifting our conceptualization of emotion to a dynamic system andfocusing on a microlevel process reveals whole new types of questionsthat we were unlikely to ask about developmental emotional processesjustusinggrossercategoricalapproaches’’ (Havilandetal., p. 167). I tendto agree, but warn – in my self-chosen role of skeptic – also of thepossibility that results obtained at that microlevel might be difficult torelate to the grosser level of observable behavior. Ties between levels ofanalysis have to be demonstrated as well. Haviland et al. proceed bystating that ‘‘These new questions . . . bypass . . . the old ones becausethey do not require that processes include awareness and synthesis ofcontextual information’’ (emphasis added). Indeed,muchof the success ofpsychology consists in having shown that the causal sequence of eventsleading up to a human perception, judgment, decision or act is notwhatit appears to be to the individual concerned, but follows a sequencewhich is (partly) hidden from awareness.However, it is also an undeni-able fact that people are aware (erroneously or not) of some connectionspredisposing to behavior, and that people tend to act on the basis of thatawareness, especially in social interaction.An individual’s identitymayserve as an example. Let us focus upon identities connected to havingsuccesfully completed so-called developmental tasks, e.g. as describedby Erikson’s lifespan theory which can be found in every introductorytextbook in (developmental) psychology. Not many people are awarethat Erikson’s theory embodies a combination of two powerful culturalnarratives. The first is the narrative of man on his earthly way toredemption and heaven being distracted by Cardinal Sins (with anorigin in fourth-century contemplation experiences of monks in the


Syriandeserts). The secondnarrativedepicts life as a journey in searchofthe realization of an ideal inwhich the hero is hindered by obstacles putby representatives of Evil; by fighting and overcoming the obstacles thehero receives honors from the audience (i.e. the narrative of Romance,Murray, 1985; Heymans, 1992). Many people map their life experienceson such a (combined) narrative, or are being mapped when a psychol-ogist analyses their life from an Eriksonian perspective; the benefit ofsuch a mapping is order in one’s set of experiences. Personhood andmeaning spring from this order-generating mapping (Heymans, 1994).Let us have a closer look at the developmental task of adolescence:overcoming the ‘‘identity-crisis,’’ made famous by Erikson and hisfollowers. Erikson assigned the Cardinal Sins to a specific phase in life,one for each phase, and used an 8-Sin instead of the standard 7-SinDoctrine (Capps, 1989). Cardinal Sinswere thought bymedievalman tobe conditions which hindered spiritual progress; these sins could beavoided or overcome by acquiring sin-specific virtues. Erikson postu-lated that the Cardinal Sin typical of adolescence was Pride, sometimesthought of by Church Fathers to be the basis of all other Cardinal Sins.Thevirtue to be acquired as antidote againstPride (a tendency to requirecontinuous adulation from others) was Fidelitas (fidelity) conceived aspersonalpredictabilityover situations.Thisfidelitas comesquite close tothe demonstrationof personality traits, which psychologists conceive astrans-situational behavioral consistencies. For Erikson, these situationsare mainly diverse social roles which the adolescent is offered forenactment. Solution of the identity-crisis is nothing else than demon-strating relative cross-situational stability of behaviors. When the adultworld arranges for a fifteen- to seventeen-year-old to enter into a varietyof new situations (roles), none of the persons involved is aware that theyare setting up a test of fidelitas for the adolescent. The cultural modelstructuring the interactions and the judgments is beyond awarenesswhen adults and the adolescent are ‘‘conspiring’’ to make part of theCardinal Sin model come true again. This example points – I hope – tothe need to invest also in searching formesolevel processes occurring atthe nexus of the individual and his social context. Microlevel modelingshould not distract researchers from what happens at the interface ofculture and the developing person. Incantation-theorywas proposed toclarify such processes (Heymans, 1994, 2000).

Classical methods of analysis have not been exhausted

The emphasis on non-linear dynamical models in developmental psy-chology has directed attention to the need for (multiple) timeseries datafrom single individuals (or other units of analysis, e.g. child–mother


dyad). There is no substitute for such data as a basis for statementsabout intra-individual, patterned changes. The near reflex-like applica-tion of psychometric methods of analysis (developed to serve a cleardescription of inter-individual differences) has led to a depreciation ofindividual case data, which until now were not frequently collected.Classical methods can help the developmental psychologist to uncoveraspects of the developing self, given that multiple timeseries from asingle individual are available. Space restrictions permit only someindications of what is possible.Emotions indeed are a powerful source of information about the self.

Therefore, part of my favorite research design is to collect timeseries ofthe intensities of multiple emotions (between 40 and 140) from singleindividuals, over long stretches of time (e.g. between 60 and 260 con-secutive days). This is the same type of data that can serve in tests ofnon-linear dynamical models. However, I use classical methods, allwithin the SPSS/Timeseries Statistical Software package. Calculation of(partial) autocorrelations within an emotion timeseries can tell aboutperiodicity in this emotion, just as does spectral analysis. Cross-laggedcorrelations between timeseries for different emotions tell whether oneemotion leads to another emotion, and after how many days. Mostimportantly, by clustering the patterns of activation-over-time ofmultiple emotions from a single individual the resultant emotion clus-ters can be interpreted as ‘‘concerns’’ in the sense of Frijda’s (1986)emotion theory, and the activation of these concerns over time can bedescribed. Internal dynamics of these concern-activations over time canbe described by ARIMA-models (see the SPSS/Timeseries package).When information about life events is also available it is possible todetermine their effects on concern-activations, e.g. by adding to therelevant ARIMA-model a regression model representing this effect.Cross-lagged correlation analysis of the ARIMA-whitened concern-activation scores will give information about the intra-individual inter-nal dynamical relations between concern-activations. For an example ofthese analyses see Heymans (2000).In conclusion: the best short-term consequence of the increased atten-

tion to non-linear dynamic models in developmental psychology is theneed to revalue and collect (multiple) timeseries data from single cases.Inmy opinion, these cases should be individuals-in-their-sociocultural-context, and the data should comprise also indicators for processessupposed to occur at the nexus between developmental context andindividual.


CHAPTER 8

Cognitive–emotionalself-organization in personalitydevelopment and personalidentityMarc D. Lewis and Michel Ferrari

The continuity of identity, despite ongoing change in the person andthe world, has challenged thinkers since ancient times. Identity has itsroots in the Latin word for same. According to the Oxford EnglishDictionary (on-line), identity involves ‘‘the sameness of a person orthing at all times or in all circumstances’’ (OED 2.a) and, more specifi-cally, personal identity involves ‘‘the condition or fact of remaining thesame person throughout the various phases of existence.’’ Locke (1690),perhaps the first to propose a modern sense of personal identity, wrotethat ‘‘The Identity of the same Man consists . . . in nothing but a partici-pation of the same continued Life, by constantly fleeting Particles ofMatter, in succession vitally united to the same organized Body’’ (ii.xxvii. Sect. 6). Yet the roots of Locke’s statement go at least as far back asPlato’s symposium (Plato, ~ 390, 207.d).Psychologists in our era have attempted to solve the riddle of identity

by proposing the construction of a self structure (Marcia, 1980), con-cept, or theory (Schlenker andWeigold, 1989), built out of cognitive andsocial constituents. Whether viewed as a schema of the self, a theory, aset of traits or dispositions, or a hierarchy of defenses and goals,identity is understood by conventional theories as a stable structurebuilt up over development. For such an identity, continuity over time isnot difficult to explain. Building-block structures maintain their same-ness by virtue of an invariant relation among their parts and an invari-ant set of functions or transactions with the world.The present volume advances the premise that traditional models of

identity, with their constructivist flavor, do not offer adequate explana-tions. Rather, as Bosma (1995) and others have suggested, the interplayof continuity and change is central to identity, just as it is to personalitymore generally (e.g. Block and Robins, 1993; Mischel and Shoda, 1995).For example, conventional models have a difficult time explainingidentity change, because their explanation of continuity relies on prin-ciples of organization that are foreign to natural systems. Change and

177

stability are antithetical in these models, but they are different facets ofthe same phenomena in nature. Indeed, as proposed by Locke over300 years ago, personal identity can only be understood in relation tothe changing, or ‘‘constantly fleeting,’’ substrate out of which itemerges. In keeping with this insight, principles of self-organizingdynamic systems have been proposed as the basis for a new generationofmodels of identity and identity development (Bosma, 1995;Havilandand Kahlbaugh, 1993; Kunnen and Bosma, 1994; Lightfoot, 1997). Thepresent chapter is one attempt to further this agenda.Our chapter addresses the development and continuity of personal-

ity and identity. We begin with principles of self-organization andself-perpetuation in natural systems. We then go on to discuss person-ality self-organization, modeled as a developmental consolidation ofcognition–emotion interactions, and describe its relations with real-time appraisals and mood states. Finally, we move into the realm ofreflective, autobiographical, and dialogical activities through whichidentity continuously emerges as a specialized product of the personal-ity system.The relation between personality and what we usually refer to as

‘‘personal identity’’ can be schematized as implicit identity versusexplicit identity. Personality involves interpretations, emotions, goals,and intentions that persist over time (McAdams et al., 1997; Mischeland Shoda, 1995). This sort of identity (i.e. sameness) remains largelyimplicit, unknown, and unconscious, and it describes young infantsand non-human animals. Perhaps more complex and certainly morespecifically human is personal identity. This is experienced by individ-uals as statements, stances, policies, or stories about who they are andhow they live their lives. This sort of identity is at least partly explicitand conscious, it reflects the sociocultural matrix in which individualsare raised, and it involves unifying narratives that make sense of – orexplicate – disparate actions in one’s past. In order to arrive at a modelof explicit identity, we first attempt to establish an understanding ofpersonality, the implicit identity that serves as a foundation for psycho-logical sameness over the lifespan.

Continuity despite change in self-organizing systems

Continuity of form in the natural world is evident everywhere, despitethe massive fluctuations which surround and penetrate it. It is presentin the demarcations of species which can last for tens of millions ofyears despite changes in weather, ecology, and even the genome itself.It is present in climatic conditions that persist for eons, despite extremeseasonal changes and gradual upheavals in the topography of land and

178 Marc D. Lewis and Michel Ferrari

sea. It is present in ecosystems, with their uncanny capacity tomaintainor renew a balance among predators, prey, waste products, and vegeta-tion. But the constancy of these forms is not that of a predeterminedstructure or rule system. For example, the genome of each species is nota program that dictates adult morphology. Nor is it a product of someexternal source by way of instruction or copying. The ecosystem in alake may depart radically from that of the river flowing into it or thelake previous to it in a chain of lakes. Natural forms are notwholes builtpiecemeal out of parts, and there is nobodymonitoring their constancy,to repair it when it fails.Principles of prespecification, programming, instruction, or con-

struction, so central to psychological explanations, have no place in thenatural world and in the continuity of its forms. Rather, principles ofself-organization and self-perpetuation are seen as the basis for theorderliness of nature (Kauffman, 1993). Models of self-organizationfocus on the emergence of order from disorder, and in particular theemergence of complex, higher-order forms from simpler, lower-ordercomponents. Examples are the emergence of living microbes throughthe interaction of complex molecules, the emergence of herding behav-ior out of the interactions of individual animals, and even the emerg-ence of embryonic structure through the interaction of proliferatingcells (e.g. Goodwin, 1987). These processes exemplify the convergenceor crystallization of global organization from recurring interactionsamong simpler components, providing structure and pattern in all ofnature (Prigogine and Stengers, 1984).There are at least three ways in which principles of self-organization

explain growth and change in natural systems. First, and most import-ant, is the emergence of novel, higher-order forms from recursiveinteractions among lower-order elements. When systems are unstable,these interactions give rise to positive feedback loops that amplifynovel coordinations into macroscopic patterns which replace the previ-ous organizational regime (Prigogine and Stengers, 1984). These pat-terns are actually arrangements of coupled elements that persistbecause they move energy through the system efficiently. Second,recurring patterns of coupling or coordination change the elements andthe connections that give rise to them. These changes facilitate similarcouplings on subsequent occasions, so that active habits grow instrength and replace competing organizations (Thelen and Smith,1994). Third, self-organizing systems shift between alternative patternsof coupling in abrupt changes called phase transitions (Kelso, 1984).Often, small changes in a particular environmental parameter (e.g. heat,concentration) trigger transitions between stable states. But they do notdetermine them, as causation resides within the system itself.

179Cognitive–emotional self-organization

While principles of self-organization are usually applied to under-standing growth and change, they must also account for stability,continuity, and themaintenance of order. Particularly in living systems,it is obvious that the maintenance of organization is necessary, not onlyfor survival, but for systems to grow and evolve (Maturana and Varela,1987). Our discussion of personality and identity will be grounded indevelopmental considerations, but its greatest challenge is to modelcontinuity. Personal identity exemplifies continuity of form as much asanything in the natural world. We therefore begin by asking howcontinuity is accomplished in nature.Biological models of self-maintenance are complex and difficult, but

a number of general principles can be extracted. Self-continuity resultsfrom recurrent interactions – change rather than stasis (e.g. Maturanaand Varela, 1987). This is as true of a peaceful nap after a large meal asthe ordering of birds in flight or the steadfast continuity of a runner ormountain-climber. Continuity emerges out of enduring or recurringpatterns of coupling among constituents that are in flux. This is becauseself-organized systems are constantly taking in and dissipating energyand information through their transactions with the environment, andthe coupling of their constituents funnels this energy along lines of leastresistance. Higher-order continuity through lower-order coupling canbe found at many hierarchical levels in complex living forms – fromcells to cell assemblies, to modular structures such as organs or leaves,to systems of such structures, and finally to the organism as a whole.Thus, the continuity of (self-organized) biological systems resides in anongoing, macroscopic ordering of microscopic interactions across sev-eral scales.How does this ordering come about? The microscopic interactions at

any level of a living hierarchy fall into preferred or familiar patterns ofcoupling, giving rise to an emergent, higher-order form. These patternsof coupling guide the flow of energy through the system (or subsystem)despite decay, perturbation, and even structural change to the systemitself. Meanwhile, the higher-order forms that emerge in this processcan be described as attractors on the state space of the system, demon-strating its tendency to occupy only a few states despite the multitudeof potential states allowed by the physical properties of its constituents.These preferred configurations are constrained by the nature of theelements, their modes of interaction with each other (e.g. damping,activating, competing, catalyzing), and the pattern of connections be-tween them. But also, most importantly, they are constrained by theirfunctional relationship to the whole system – the levels of the hierarchyabove and below – and the environment beyond. Such interactingconstraints can be thought of as a matrix of complementarities that


supports or favors particular patterns and permits spontaneous changefrom one preferred pattern to another.Complementarities favor the emergence and continuity of orderli-

ness, but they do not act like a set of rules dictating a particularoutcome. For example, the population level of a species of birds stabil-izes through complementarities across mating patterns, nestingsupplies, food supplies, and predation, coupled with overall reproduc-tive success. But there is no set of rules that stipulates what thatpopulation level should be. Thus, continuity arises from interactionsguided by complementarities among constituents and functions, andthese reflect the essential character of the system through its patternedbehavior in real time. Concepts such as ‘‘decentered structure’’ (Good-win, 1987), ‘‘relative stability’’ (Landauer, 1987), and ‘‘dynamic stabil-ity’’ (Thelen and Ulrich, 1991) capture continuity of this sort.Self-perpetuation in biological systems is of course much more com-

plex and difficult than implied by this brief discussion. For one thing,mechanisms of self-regulation emerge in self-organizing systems ofsufficient complexity and protect them from deviation and change. Foranother, self-replication is a particular type of continuity, adopted byall biological systems, through which identity is transmitted over timeand space. We will not deal specifically with these ideas in this chapter,but they will be important considerations for further theorizing in thisarea. For now, we can begin to model the emergence of continuity inpersonality and identity armed with the basic principles of recursion,coupling, and complementarities.

A dynamic systems perspective on personality development

Before going on to model identity from a self-organizational perspec-tive, it is necessary to establish a more general picture of personalitydevelopment and emerging personality consistency. Our approachfollows the general strategy of dynamic systems (DS) approaches indevelopmental psychology. These approaches assume that developingsystems consolidate through recurrent interactions among psychologi-cal, social, or perception-action components, passing through phases ofreorganization at transition points (Fogel, 1993; Lewis, 2000b; Thelenand Ulrich, 1991; Van Geert, 1991). Developing systems are thus self-organizing systems, accruing order through increasingly complex coor-dinations. This orderliness can be represented by attractors on a statespace (Lewis, Lamey, and Douglas, 1999; Thelen and Smith, 1994) or bythe stabilization of a growthprofile or time series (VanGeert, 1994). It isimportant to note that the term self-organization can describe emergentprocesses at many time scales. It can refer to the movement of behavior


to attractors in real time (e.g. over seconds) and to the emergence andconsolidation of attractors in developmental time (e.g. over the courseof months or years) (Thelen and Ulrich, 1991). Thus, DS approaches arecapable of tracking convergence, reorganization, and crystallization atseveral scales at once, and the interaction of time scales has been ofparticular interest to psychologists with a dynamic systems view (e.g.Van Gelder and Port, 1995; Thelen and Smith, 1994).To date, motor development (Thelen and Ulrich, 1991), cognitive

development (Smith, 1995; Van der Maas and Molenaar, 1992; VanGeert, 1991), and communicative development (Fogel, 1990, 1993) havebeenmost thoroughly studied using DSmethods. However, DSmodelsof personality development are now beginning to appear (Derryberryand Rothbart, 1997; Lewis, 1995, 1997; Magai and Nusbaum, 1996;Schore, 1997, 2000; Cloninger, Svrakic, and Svrakic, 1997).Most of theseapproaches see personality as being knit together from psychological,psychosocial, and/or biological constituents over developmental time,with early patterns setting the course for the subsequent channeling oflearning and experience. Many of these models also describe relationsamong emotion, cognition, and behavior as central to the developmentof individual styles. The promise of these approaches is to explain – notjust describe – personality as an emerging, consolidating, and continu-ous organization (Lewis and Granic, 1999a). From this perspective,personality is not specified by genes or environment or constructed outof parts; it is a self-organizing system that converges to its own uniqueform and perpetuates that form as its implicit identity.

Cognition–emotion interactions and personality self-organization

According to the present approach, personality development can bedescribed as change, stabilization, and refinement of recurrent patternsof cognition–emotion interactions, represented by attractors on thestate space of the psychological system (Lewis, 1995, 1997). Like othertheoreticalmodels of personality development, this account has little incommon with trait or factor-analytic approaches to adult personality,the latter beingunconcernedwith the fundamental processes of person-ality formation and change. More specifically, the idea that personalityarises from cognition–emotion interactions derives from theories ofemotion and emotional development, though it also resonates withsome personality theories (e.g. Mischel and Shoda, 1995). Tomkins(1978) and Izard (1984) viewed personality as the construction of affec-tive-cognitive or ideoaffective structures, arising through the child’smost frequent or important experiences. These structures are defined aslinkages between frequently felt emotions and accompanying cognitive


interpretations. Affective-cognitive structures are assumed to generatecharacteristic behaviors that indicate trait-like qualities in the eyes ofothers (Malatesta and Wilson, 1988). As to the mechanisms underlyingthese structures, Tomkins and Izard emphasize the cueing of cognitiveprocesses by emotions, whereas cognitive-developmental approacheshighlight the opposite causal direction – the elicitation of emotions bycognitive appraisals (e.g. Sroufe, 1979). These two causal directions arebrought together by Malatesta and Wilson (1988), who propose a ‘‘re-cursive and sustaining relationship’’ (p. 103) between cognitive inter-pretations and emotions underpinning personality development.The premise of cognition–emotion recursion has been crucial for

modeling emotional and personality self-organization in real time anddevelopment (Lewis, 1995, 1996, 1997). Such modeling efforts havemuch in common with dynamic systems approaches of Camras(1992), Fogel (1993), Magai and Nusbaum (1996), and recently Izard,Ackerman, Schoff, and Fine (2000) and Mascolo, Harkins, and Harakal(2000). A comprehensive theory of cognition–emotion interactions iscurrently under construction (Lewis, 2000a). This account specifiesthree time scales of emotional self-organization: appraisal–emotionepisodes in real time, moods that last from minutes to days, andpersonality patterns that persist over a lifetime. At each of these scales,cognition–emotion configurations are proposed to converge frominitial indeterminacy and variability to coherent and lasting states.Thus, appraisals, moods, and personality each develop, giving rise tonested scales of emotional self-organization in microdevelopment,mesodevelopment, and macrodevelopment.According to this account, a cognitive appraisal or interpretation

converges in real time from the coordination of (lower-order) elementssuch as concepts, associations, and perceptions, but it does so only ininteractionwith a consolidating emotional state. The convergence of anappraisal augments and constrains emotional activation while emo-tional activation simultaneously augments and constrains attentionalprocesses atwork in appraisal (Lewis, 1995). In contrast to conventionalappraisal theory, appraisals are not presumed to be independent ofemotion, nor to precede emotion in time. Rather, as argued by Frijda(1993b, in press), it is not until the end of an appraisal process thatcognitive interpretations are comprehensive and complete. Thus,rather than view cognitive appraisals as eliciting emotions, or emotionsas cueing cognitive activities, appraisal and emotion are proposed toarise in tandem (cf. Buck, 1985) and to stabilize through ongoing feed-back. This is consistent with the notions of reciprocal causation men-tioned by previous theorists (e.g. Malatesta and Wilson, 1988).The synchronization of appraisal and emotion can be described as a


macroscopic coupling between the cognitive and emotional systems,giving rise to a coherent emotional interpretation (EI) in real time.Emotional interpretations are similar to Izard’s affective–cognitivestructures, but they are construed as temporary organizations ratherthan lasting structures. In an EI of blame, for example, anger rapidlycoupleswith the sense that someone is at fault. This coupling subsumesmore intricate couplings amongmicroscopic cognitive constituents. Forexample, the sense that someone is at fault results from couplingsamong an image of a person, an inference of harm, and the attributionof power or intention. In a more complex example, the cognitive coup-lings that contribute to an EI of inadequacy might include links betwenconcepts of ‘‘small,’’ ‘‘helpless,’’ and ‘‘baby,’’ expectancies of humili-ation, associations between adulthood and power, and images of par-ental disdain. However, these cognitive couplings would not cohereunless nested in a global coupling with emotion, such as anger in thefirst example and shame in the second. In a recent look at emotionalself-organization, Izard et al. (2000) suggests that coupling among basicemotion systems creates comprehensive emotion patterns, just as cog-nitive couplings underpin comprehensive interpretations in the presentaccount. EIs may thus involve the coupling of constituents in bothsystems. (At a neural level, this is inevitable, but psychological termi-nology does not make clear distinctions between networks and singleentities.)The idea that coherent psychological states arise from coupling with-

in and between cognitive and emotional systems rests partly on neuro-biological evidence (Damasio, 1994; Freeman, 1995; Schore, 1997;Harkness and Tucker, 2000). According to this research, cognition–emotion interactions are centered in the relation between the prefrontalcortex (e.g. ventromedial prefrontal or orbitofrontal cortex), which actsas a convergence zone for all cortical regions, and the emotional cir-cuitry of the limbic system, which includes various connected nuclei.Appraisals in the prefrontal convergence zone arise from the entrain-ment ofmany other cortical regions, including all perceptual andmotorregions, and they mediate perceptual expectancies or preafference,motor rehearsal, and the articulation of specific behaviors (e.g. Free-man, 1995; Schore, 1997). These appraisals also resonatewith the limbicsystem and brain stem, fashioning a macroscopic feedback loop be-tween emotional and ideational activity (e.g. Tucker, 1992).Because EIs are higher-order forms, arising from preferred patterns

of coordination among lower-order cognitive and emotional constitu-ents, they can be represented as attractors on a state space of ‘‘possible’’psychological states. Perceptual categories (Thelen and Smith, 1994),linguistic categories (Smith, 1995), motor coordinations (Hopkins and


Butterworth, 1997), cognitive skills (Van Geert, 1994), memories (Ab-raham, 1995), belief systems (Goertzel, 1995), and communicativeframes (Fogel, 1993) have all been proposed as attractors on a psycho-logical or biological state space. More recently, modal emotions(Scherer, 2000) have been added to the list. We suggest that recurrentemotional interpretations (EIs) are psychological attractors whoseunique configuration represents an individual personality. The pres-ence of several attractors, representing several EIs, indicates a range ofstates (cf. Schore, 1994) to which interpretations can converge for thesame individual. Thus, personality is depicted as a cluster of probablecognitive–emotional states, indeterminate and variable in real time butstable or continuous over development (Lewis, 1997). Such a depictionis consistent with Mischel and Shoda’s (1995) definition of personalityas an individual cognitive-affective system that produces predictablebehaviors in each of several classes of situations.Self-organizing EIs are not merely psychological events; they have

specific implications for action. The function of emotion in (human andnon-human) animals is to induce, urge, and constrain actions that aredesigned to alter the physical world – and very often the social world –in ways that are conducive to the achievement of goals (Oatley andJohnson-Laird, 1987; Stein and Trabasso, 1992). Thus, blame urges andsupports some form of attack in order to get rid of an obstacle, and theshamed individual is propelled to hide some part of him- or herself inorder to avoid rejection.According to the present theory, it is the type ofrelation between emotion and action that gives rise to three distinctscales of emotional self-organization.At the first scale, microdevelopment, the organization that builds

over seconds in an emergent EI rapidly dissipates when goals areachieved through actions. We notice a part of a headline in the news-paper held by the passenger beside us on the subway. We feel interestand excitement, peek over her shoulder, notice that the headline reportsnothing out of the ordinary, and go back to our daydreaming. How-ever, when actions cannot achieve goals, then the goal state can perse-vere, and the organization of interpretation and emotion may persistfor minutes, hours, days, or weeks (cf. Scheff, 1987; Teasdale andBarnard, 1993). The development of such prolonged states of cognitive-affective patterning constitutes the second scale of emotional self-or-ganization – what we have termed the mesodevelopment of moods.We are loosely defining moods as lasting states of cognitive–

emotional organization. That moods are characterized by reducedcognitive variance is supported by copious data on mood-congruentperception, learning, and interpretation. Moods also imply a limitedrange of long-lasting emotions or affects, though not necessarily a


single, persistent emotion (Frijda, 1993a). For example, in an angrymood, resentment is recurrent and enduring whereas interest, sadness,and happiness are less probable and shorter lasting. Thus, mood can bedepicted as a temporary constriction of the state space, representing acognitive–emotional bias: attractors are strengthened for some EIs andweakened or absent for others, and trajectories between states are morelimited (happy mood may be an exception; see Lewis, 2000a).In moods, not only does action (anticipated or actual) fall short of

achieving goals, but action tendencies continually arise as long as anemotional constellation remains in place. Thus, action is both urged andobstructed in moods. This hypothesis of a prolonged state of psycho-logical organization, resulting from backed-up action, is congenial withtheorists going back at least to Freud and continues to be verifiedempirically (Horowitz, 1998; Polivy, 1998). It is also consistent withrecent neurobiological accounts. According to Freeman (1995, 2000) thecorticolimbic feedback or resonance described earlier gives rise to aglobal intentional state, beginning in the hippocampus and rapidlyself-organizing across the entire brain. Action is initiated by this inten-tionality, fueled by emotion in the limbic system, and articulated andelaborated in the frontal and prefrontal cortex (Freeman, 1995; Tucker,1992). Thus, intentions keep the brain organized in the pursuit of goals,and emotions serve as their handmaidens, supporting a perpetualreadiness for intended actions. This picture is consistent with Frijda’s(1986) emphasis on states of action readiness as key attributes ofemotions.If this picture of brain functioning is accurate, then emotional inter-

pretations (EIs) represent global patterns of corticolimbic coherencemanifesting recurrent intentional states. Yet these patterns should beshortlived when EIs give rise to successful actions. Freeman’s researchinto the perception–action cycle demonstrates electroencephalogram(EEG) coherence during the perception phase, dissipating when actionis initiated. This dissipation of coherence leaves the brain in a chaotic‘‘background’’ state fromwhich new coherences can emerge in the nextcycle (Freeman and Baird, 1987; Skarda and Freeman, 1987). But whathappens if action is ineffective or impossible? If the goal cannot beabandoned (as is typical of humans, with their advanced symbolic andmemory capacities), then the orderliness of intention and emotionwould persist. In this state, goal-related behaviors might be rehearsed,expected responses monitored, and contingencies checked and re-checked. Rumination is characterized by these mental activities, andresearch has demonstrated a reliable correlation between the tendencyto ruminate and the endurance of depressive mood states (Nolen-Hoeksema and Morrow, 1991; Teasdale and Barnard, 1993). In neural


terms, a lasting state of goal-specific corticolimbic organization wouldresist switching to novel pursuits (Schore, 1997). Moreover, enduringaffective states cause the release of neurohormoneswhose function it isto commit the cortex to mammalian goals such as aggression, mating,caring, and so forth (Panksepp, 1998). Thus, mood states may harnessbrain activity to an agenda that cannot be readily satisfied. Instead ofpeeking at somebody’s newspaper, one may feel outraged by a rudepassenger or sexually attracted to another. The intentions that corre-spondwith these feelings – aggression and courting – cannot readily beput into action on the subway. They therefore have the potential toresonate for long periods, oftenwithout one’s awareness of their originsor their object.The third and longest scale of emotional self-organization is person-

ality development. In previous work, personality development hasbeen described as the evolution and refinement of a cognitive–emotional state space, whereon attractors for EIs emerge, crystallize, orvanish over the course of months and years (e.g. Lewis and Douglas,1998). As mentioned earlier, EIs are the real-time manifestation of thisdevelopmental process, and the means by which personality is dis-played by states of shyness, grumpiness, friendliness, and so forth.Moods constitute temporary modifications to the personality statespace, strengthening some attractors and weakening others. But howdoes the state space consolidate in the first place? In other words, howdoes personality develop?As mentioned at the beginning of this chapter, a universal conse-

quence of recursive self-organizing processes is their tendency to leavesome trace behind them. Wherever water pools in the garden during arain storm it is more likely to pool in the future. Connection strengths ina neural network are adjusted according to patterns of activation,enhancing the probability that those same patterns will recur. In thebrain, Hebbian learning describes the increased likelihood of coactiva-tion of neurons that have been coactivated on a previous occasion. Andeveryday experience tells us that interpreting a novel event in a certainway increases the chance of interpreting similar events in a similar way.These examples demonstrate a principle of change by which the coup-ling of elements in a self-organizing system alters the elements, theirconnections with each other, and their relations with other levels oforganization – in short, their complementarities – such that similarpatterns of coupling are facilitated in the future.This general principle of accrual and crystallization has several im-

plications for personality development. First, if interpretive organiz-ation is always embedded within cognitive–affective patterns, thencognition-emotion couplings are the source of personality structure


(Izard, 1984; Malatesta and Wilson, 1988). Second, any converging EIwill increase the likelihood of its own recurrence, contributing to per-sonality development through the repetitive incidence of similar inter-pretations and the ongoing accretion of emotional patterning. Third,EIs that recur frequently in mood states will have a particularly strongimpact on personality development. The longer cognitive–emotionalcouplings resonate and themore often they recur, the more deeply theywill become entrenched in underlying complementarities. Moreover,neurohormonal activation in moods extends intentional states, main-tains cortical configurations for longer periods, and promotes ‘‘long-lasting changes in the strengths and durations of synaptic actions’’(Freeman, 2000: 226). Moods thus have the greatest potential to guidesynaptic learning as well as emotional conditioning over long periods,fashioning permanent modifications to global attractors for develop-ment (Panksepp, 2000b).It is often assumed that personality starts to crystallize in early

childhood. How might this early consolidation be influenced by recur-rent mood states? Young children’s moods have not been systemati-cally studied, but no parent would doubt their ubiquity and power.Schore’s work is particularly relevant here. According to Schore (1997,2000), neuromodulator action during prolonged affective states main-tains enduring cortical configurations. These configurations fashionentrenched habits that gradually supersede the brain’s early plasticity.Moreover, it is the affective context of caring that constrains corticalpatterning. Habitual appraisals emerge during caring interactions witha characteristic affective tone, consolidate in the first two years, andassume a hierachical dominance over subsequent cortical organizations(Schore, 1994, 2000). According to Harkness and Tucker (2000), moreintense emotions andmoods, reexperienced overmany occasions, havethe strongest effects, as when the recurrence of abuse and neglect earlyin life ‘‘kindle’’ depression or anxiety with increasing predictabilityover development. Thus, the synaptic architecture of the brain is lit-erally sculpted during states of prolonged affect in infancy and earlychildhood.Most important from the present perspective, prolonged affect in

mood states is very often negative – a consequence of blocked actionand unresolved intentions. These states reflect tensions, losses, andfrustrations arising from recurrent patterns of familial interaction butthey resonate long after interpersonal exchanges are over. Examplesinclude anxiety about parental rejection, jealousy of siblings, shameand anger following punishment, guilt about harm done to others, andsadness resulting from parental distance or withdrawal. These states


reflect goals and needs that are difficult to satisfy (yet difficult torelinquish) given the real contingencies of the interpersonal world. Thecontinuities they hold in mind and brain thus tap continuities in theworld. That they also fashion continuities at a longer time scale – that ofpersonality development – expresses the psychoanalytic axiom thatpersonality derives from unresolved wishes, but does so in a scientifi-cally plausible manner.Before going on to a discussion of identity, let us summarize how

this theoretical model explains personal continuity in psychologicalfunctioning. Specifically, how do processes at longer time scales con-strain those at shorter ones? We have already defined moods as modi-fications or refinements to a previously established (personality) statespace. What this means is that the connections laid down in personal-ity (macro)development constrain the possibilities for moods (inmesodevelopment). Hostile people tend to have angry moods and shypeople tend to have anxious-inhibited moods. Personality also con-strains feeling and thinking in microdevelopment, as indicated by themovement of emotional interpretations (EIs) to personality-specificattractors in real time. However, this effect is necessarily mediated bymoods (broadly defined) which constrain microdevelopment moreimmediately. Thus, it is the nesting of mood in personality that con-strains EIs in real time, and their joint effects curtail the varianceavailable for making sense of and feeling about the world. One couldsay that real-time EIs grow out of system orderliness maintained inmoods by an inability to achieve goals in the present, and moods growout of the orderliness maintained in personality by an inability toachieve goals in the past. Thus, continuity is to be found in real-timeinterpretations that echo the incomplete resolution of longstandingneeds and wishes.This account of continuity does not rely on static structures, rules, or

programming. As with all natural systems, it relies on hierarchicallynested self-organizing processes, by which the emergent products ateach level assemble themselves according to complementarities amongtheir elements and constraints from above and below. A psychologicalstate in real time is always emergent, a product of synaptic activationand corticolimbic resonance. However, the pattern that arises is one ofonly a few that can arise given complementarities that have evolvedover longer time scales. The more constraining the complementarities,the more frequent the recurrence of the pattern, as epitomized bypsychopathological states such as paranoia and depression. Yet thesecomplementarities do not comprise a structure or rule: they comprise acontoured state space, a matrix of tendencies. According to the present


theory, these tendencies express unfinished business, by which theindividual remains preoccupied with an idealized version of an imper-fect world.The conclusion is that continuity builds on itself wherever goals are

blocked, tension is unresolved, and conflict interrupts the smooth flowof action from intention. This may seem like a new idea, but it mirrorsthe classic psychoanalytic insight that personality is a byproduct of thestance we take toward a world that habitually obstructs our goals.Personal continuity, viewed in this way, is a compensation for the lossof continuity between intentions and actions, and it expresses theuniquely human capacity to hold onto wishes rather than realities. Weare hypothesizing that this principle of continuity underlies the crystal-lization of personality and its manifestation in moods, emotions, andappraisals, creating an implicit identity or sameness over time. Thequestion that remains is how explicit identity can be illuminated by thesame principle.

Explicit identity: recurrent appraisals of an idealized self

Explicit identity, or personal identity as we define it, relies on thecontinuity of personality but goes beyond it. Personal identity could notbe fashioned and maintained without the interpretive and behavioralcontinuity resulting from a stable personality system. Yet personalitystability is not sufficient to explain identity. Animals and infants havestable personalities but do not demonstrate identities as we understandthem. We suggest that personal identity is a specialized product ofpersonality that relies on semantic, reflective, and often conscious activ-ities.We also suggest that this identity is not a static cognitive structure,but rather a lineage of recurring self-appraisals intrinsically linkedwithemotion (Haviland and Kahlbaugh, 1993). Like other appraisals, theyembody cognition–emotion resonances embedded in intentional states.Like other appraisals, they should recur more often in the presence ofunsatisfied goals. Unlike other appraisals, they are necessarily aboutthe self and are both autobiographical and dialogical in nature. Thus,we construe identity appraisals as the recounting or enacting of a storyof oneself to a real or imagined listener. This story is an explicit accountof who one is – an explication of self. Yet, epitomizing appraisals thatpersist and recur, it suggests a wish or a need to be someone that cannotquite be accomplished, rather than a readout of who one really is. Itsstability is situated at a juncture between who one is and who onewould like to be, and this tension, this non-resolution of personal goals,is the grounds for its remarkable persistence.


Identity as autobiography

A number of contemporary models of identity stipulate its autobio-graphical nature, reflecting a growing emphasis on narrative in devel-opmental, social, and personality psychology (Bruner, 1990; Freemanand Brockmeier, in press; Labouvie-Vief, Orwell, and Manion, 1995;McAdams, 1993). Narratives are key towhatmost people consider theirinnermost selves and are integral to how they construe their participa-tion in culture (Bruner, 1996; Lucariello, 1995; McAdams, 1993). Narra-tives are also considered instrumental in orienting individuals’interpretations of life events, including choices theymake as they striveto construct their own personal identity or to assume the identities thatculture constructs for them (Chandler, Lalonde, and Sokol, in press;Ferrari andMahalingam, 1998; Labouvie-Vief et al., 1995). McAdams etal. (1997) suggest that ‘‘contemporary American adults make sense oftheir sometimes scattered lives by fashioning and internalizing storiesthat integrate their reconstructed past, perceived present, and antici-pated future’’ (p. 678). From adolescence on, people construct suchnarratives in order to give their lives a sense of unity and purpose.‘‘Identity, therefore, may be viewed as an internalized and evolving lifestory, a way of telling the self to the self and others, through a story orset of stories complete with settings, scenes, characters, plots, andthemes’’ (McAdams et al., 1997: 678).Good narratives are about action, and identity narratives concern the

actions that aremost integral towhowe thinkwe are (or wishwewere).Connecting these actions into a coherent account establishes identity asan autobiography. Narratives are also powerful goads to action – one’sown action and that of others. Indeed, McAdams et al. (1997) suggestthat narratives are highly effective in promoting and supporting gener-ative action. Most people actually develop their stories in the service ofaction, and not at arm’s length from lived events (Bourdieu, 1997). Thefuture is lived as the ‘‘anticipated present’’ (Bourdieu, 1997; Cole, 1996)as one acts to make one’s narratives actual. These actions, in turn,become the object of further narratives in a recursive cycle of action andinterpretation.The ability to anticipate events stems from personal narratives but is

also acquired tacitly and historically. It thus constitutes what Bourdieu(1997) refers to as socially inherited cultural capital. This capital ismanifested in an enactive narrative (an ongoing interpretation ofevents) that is experienced as the power and possibility to act in aparticular sociocultural setting. For example, narratives associatedwithgender, ethnic identity, or family serve to frame and orient action.


Action in particular cultural contexts, in turn, transforms narratives byenriching and validating them, or by posing problems that must beovercome in order to sustain them. Such encultured action often tailorscognitive development to the narratives in which one participates (Fer-rari, 1998; Ferrari and Mahalingam, 1998).Thus, identity can be viewed in terms of autobiographical narratives

which concern actions in the past, express cultural ideals, and help toguide and organize actions in the present and future. To rephrase thisidea in the language of appraisals and emotions, the actions andevents that come to mind in self-appraisals are named, referred to, andwoven into autobiographical stories in subsequent appraisals in anongoing lineage. These autobiographical appraisals, like all appraisals,are coupled with emotions in recurring emotional interpretations(EIs). For example, excitement or hope couple with interpretations ofaccomplishments, and jealousy or guilt couple with interpretations offailed relationships. Emotions make identity narratives compellingand absorbing and are central to their creation (Haviland andKahlbaugh, 1993). Moreover, these emotions are instrumental ingenerating and framing further actions, extending our auto-biographies into belief systems and intellectual achievements (Magaiand Hunziker, 1993).But this is only half the story. Autobiographical accounts, or recur-

rent appraisals of oneself and one’s actions, change and grow in detailand nuance but remain highly consistent in theme. This consistencyconveys the essence of one’s identity.Moreover, personal consistency isdemanded by all cultures and it is a psychological necessity for devel-oping individuals (Chandler et al., in press). How do we explain thisconsistency in explicit identity? Our account of implicit identity stipu-lates that only unsuccessful or partially successful actions producelasting states of psychological organizationand only unresolvedwishesand goals create lineages of recurrent appraisals. In order to explain theconsistency of explicit identity, we should therefore examine not onlythe actions recalled and promoted by autobiographical accounts, butthe very immediate actions of telling these accounts, and we should askwhether these actions successfully resolve the goals which initiatedthem. To do this, we turn now to the dialogical nature of identityappraisals. The act of telling (or enacting) one’s story is inherentlydialogical because it presupposes a listener who will be interested andresponsive. Thus, the telling of one’s story, whether to real or imaginedlisteners, must be evaluated in terms of the interpersonal goals inherentin dialogue.


The dialogical nature of identity

The idea that personal identity is essentially dialogical is a naturalextension of the notion of a narrative self, and it follows from theproposition that self-narratives are collaborative efforts involving oneor more listeners (Sarbin, 1993). The dialogues in which self-narrativesare embedded may take place with anyone, but they often end uptaking place with imagined listeners who have a particular voice andperspective (Bakhtin, 1929/1973). Hermans (1996a) argues that the selftakes up various positions in such internal dialogues. Voiced positions,each functioning as a separate person, engage in disagreements, con-flicts, and confrontations. Thus, each position is a center of initiative orintention, laying the groundwork for the other voices’ reactions (Her-mans, 1996a; Hermans and Kempen, 1993).Not surprisingly, the idea of self-dialogue has a long history. It goes

back at least as far as Plato’s claim that thinking involves, ‘‘a discoursethat the mind carries on with itself about any subject it is considering’’(Plato, ~ 390). In terms of personal identity, C. S. Lewis (1967) eloquent-ly stated that ‘‘A person cannot help thinking of himself as, and evenfeeling himself to be (for certain purposes), two people, one of whomcan act upon and observe the other. Thus he pities, loves, admires,hates, despises, rebukes, comforts, examines, masters or is mastered by,‘himself’’’ (p. 187).To return to the language of appraisal, emotion, and action, all

emotional interpretations (EIs) include action readiness, or an emergentplan for action, as well as anticipated feedback from the world – anexpectation of what will happen when one actually does take action.Identity appraisals include an emergent plan for telling others aboutoneself and an expectation of how theywill respond.Whenwe actuallytalk about ourselves to others, these anticipations resolve into realdialogues. But most of the time, the telling remains at the level ofrehearsal and imagination, and much of what is ‘‘heard’’ internally isthe expected response of the other.As we groom ourselves in the mirror or prepare to teach a class, the

following voices might be heard: you look terrific today . . . thanks, I’mactually in great shape for someonemy age; you look tired . . . well, I’vebeen working too hard lately, not surprisingly considering the de-mands placed on me . . . he carries so much responsibility, we couldn’tdo without him; that was a great point you made in class . . . yes, I’mactually an excellent teacher . . . you’re/he’s much brighter than peoplerealize. These internal dialogues permit a constant editing and elabor-ating of one’s autobiography. They also fall naturally into the intersec-tion set of one’s own and others’ expectations (Bakhtin, 1929/1973).


They are quasi-realistic exchanges contextualized in the interpersonalworld (even though we would rarely speak them aloud). For Fogel(1993), such internal dialogues are examples of consensual frames, andthey express the coregulation of two (or more) participants who adjustto each other in expectable and familiar ways. Consensual framesself-organize in real time, maintaining their configuration within occa-sions, but, according to Fogel, they also recur across occasions, estab-lishing a developmental pattern or lineage. This is certainly the casewith personal identity.In our view, each voiced position in an internal dialogue or frame

corresponds with a global intention that entrains cognition–emotion.This means that others’ voices are indeed one’s own, at least for themoment it takes to formulate them. But why are these voices, and theintentions they serve, so consistent across occasions? We can begin toanswer this question by examining the position of the ‘‘I’’ and the‘‘you.’’ The ‘‘I’’ consistently intends to put forth a reasonable story ofsomeone who will be accepted, esteemed, or liked. This is no easy job,because the ‘‘you’’ is often critical rather than sympathetic, or simplynon-committal. Thus, wishes to be accepted, admired, and liked, to beclose, to avoid rejection, and even to hide unattractive parts of the selfcannot be fully or permanently realized, because other people in theworld and in imagination are not always accommodating. Yet thesewishes are what give rise to internal (and external) dialogues repeated-ly, because they propagate the goal of ‘‘explicating’’ oneself in the firstplace. It may be the very difficulty in achieving these intentions thatentrenches them and extends their orderliness over time. Thus, the goalof narrative in dialogue is never quite fulfilled, either internally orexternally, because the wish for closeness, acceptance, and admirationis never completely satisfied.There may be another crucial obstruction to the dialogical goals for

self-narrative: it is nearly impossible to fashion an account of the selfthat is coherent and unified given the disparity of positions and voices(Bruner and Kalmar, 1998; McAdams et al., 1997). As a result, theintention to unify what is so obviously disparate cannot be fully realiz-ed, and the justification of discrepancies among our various selves iselaborated and extended over many recurring efforts. Narratives arecommunicative acts with a particular listener in mind. The same storyrequires major adaptation to appeal to different audiences, who valuedifferent things and judge one’s stories (including one’s explanations,justifications, and plans) by different criteria (Ferrari and Mahalin-gham, 1998). It is commonplace to tell conflicting stories in differentcontexts (Cooper et al., 1998; Hermans, 1996a), and we do not alwaysvoice a particular narrative out loud (Bruner and Kalmar, 1998). Thus,


the demand for narrative consistency, highlighted by McAdams et al.(1997), clashes with the need to please different audiences with differ-ent accounts of the same actions.The different voices in Hermans’ (1996a) dialogical melee cannot all

be satisfied with the same narrative, yet the actions being explainedcannot be completely rewritten. Aggressive actions and intentions con-cerning one’s younger brother must be hidden, or at least disguised,when recounting them to one’s mother, but amplifiedwhen recountingthem to the other boys in the playground. We suggest that the need toreconcile these accounts constitutes an unrealizable agenda: one thatcannot be satisfied through action and which therefore maintains theorderliness of cognition–emotion coupling and behavioral rehearsal.This claim is not far removed from the psychoanalytic idea that tensionbetween wishes and justifications is the crucible of psychological life.The backing up of orderliness resulting from this impasse would thengive rise to a richly articulated, historical, and continuous account of anidealized, justified self.However, the explication of self can succeed, as whenwe admit to the

contradictions in our stories and the listener accepts the many facets ofwho we are. If our model is correct, then the achievement of thisdialogical goal ought to dispel explicit identity – at least for a while. Asevidence for this prediction, consider the experience of identity whenfalling in love. It is no accident that the feeling of being completelyaccepted by another person corresponds with the sense that the self isno longer finite, unitary, describable, or even known. A similar experi-ence may derive from psychotherapy. These ‘‘self-actualizing’’ experi-ences free us from carrying around a story of ourselves in mind andarguing its validity to a real or imagined respondent. However, the lossof identity may come about much more drastically when meaningfuldialogue with real or imagined listeners is truncated by developmentalor cultural upheaval. When the dialogue of self-definition comes to ahalt, loss of identity can lead to suicide, presumably because the futurehas no relevance and the present is too painful (Chandler and Ball,1989).

Identity change

Given our explanation for continuity of personal identity, howdoes thisidentity change? Early in this chapter we reviewed three kinds ofchange in self-organizing systems: (1) the emergence of novel, higher-order forms that replace their precursors, (2) the entrenchment ofparticular habits to the exclusion of others, and (3) the abrupt shiftingbetween stable patterns denoted by phase transitions. Identity change


in development can be modeled in all three ways, but first we brieflyconsider identity change in real time.Identity change in real time consists of switching from one story to

another, or shifting dialogical frames (Fogel and Lyra, 1997), as impliedby the previous discussion. We will not discuss this type of change indetail, but Fogel and his colleagues have thoroughly researched thequalitative distinctions between frames and the discontinuity seen inbehavior when dyads switch from one to another (Fogel and Lyra,1997). The DS notion of phase transitions seems to capture the abruptand global nature of such change, but research on this topic is stillpending (Fogel, 1993: 116). For the present, we note that each of thesestories or frames would anchor identity only if it had a historicallineage. Thus, each identity story (as a robust and recurrent EI) wouldrequire a cognitive–emotional complementarity that had arisen overmany occasions, expressing concerns and wishes intrinsic to one’spersonality. This is one obviousway inwhich real-time identity apprai-sals makes use of longstanding personality patterns.Change in identity across developmental time would involve change

to one or more of these narrative lineages or, importantly, to theintentions with which one recounts and integrates them in dialogue.The three kinds of change each suggests different directions formodeling.First, new identity constellations (either stories or meta-stories) can

be viewed as novel, higher-order assemblies that cohere from emotion-driven coordinations among the cognitive constituents of emotionalinterpretations (EIs) in general and previous stories and dialogues inparticular. This general mechanism of self-organization would be re-sponsible for the consolidation of identity in the first place, probablyin the second to fourth year of life. It may also be a good way toconceptualize subsequent identity change, especially when new ident-ity patterns start off inchoate and unstable and consolidate over manyemotionally significant occasions. For example, in adolescent transi-tions, needs for autonomy and intimacy can be brought to bear on one’sautobiographical stories, forcing individual narratives to change inorder to maintain authenticity (Harter, 1998; Mascolo and Fischer,1998). Yet, it takes many repeated dialogical appraisals – justifyingactions and mending discrepancies – before those narratives becomecoherent and consistent. Marriage also brings about identity change,sometimes driven by the intention to have children and to place thembefore one’s own interests. Again, narratives of oneself as a parent taketime to self-organize, and they do so by assembling emotionally mean-ingful constellations of narrative constituents already present in one’spersonality or one’s culture.


In the second kind of systemic change, recursion strengthens newhabits such that they supersede competing tendencies. The retelling ofnewly fashioned autobiographical accounts to real and imagined lis-teners is therefore a likely mechanism for identity change. These ac-countsmay initally be imported frompeers or themedia but then retoldrepeatedly in one’s imaginary dialogues. Elkind (1974) has emphasizedthe impact of the ‘‘imaginary audience’’ on the development of adoles-cent identity. Moreover, actions would seem to be particularly power-ful instruments of identity change. We have already suggested thatidentity narratives bothmake sense of actions and promote new actionsin an ongoing cycle of recursive interpretations. Such recursion wouldrapidly strengthen complementarities that support one’s favouredstory-lines. Finally, new narratives are likely to be problematic, emo-tional, and unfinished because of their inconsistency with previousaccounts. They are thus well suited to capture psychological organiz-ation for hours and days in mesodevelopment and to modify personal-ity itself over macrodevelopment. In this way, identity change mayprovide a leading edge to personality development while at the sametime being shaped and constrained by underlying personality comple-mentarities.Finally, the idea of phase transitions, now at the scale of develop-

ment, may be useful for modeling identity change. Life stories areconstantly under revision (Freeman and Brockmeier, in press), butsome actions or events act as pivot points in people’s life stories.Dramatic life events can serve as junctures in one’s autobiography,specifically when one story is abandoned and another is begun (Brunerand Kalmar, 1998). These pivotal changes can be modeled as phasetransitions, and the events that elicit them as triggers. As discussedearlier, the nature of such changes is determined by the attributes of thesystem, not the attributes of the triggering events, a point also empha-sized byMagai andNusbaum (1996). Changes in personal identitymayalso be revealed by changed emotions (Haviland and Kalbaugh, 1993)and emotionally powerful life events have been shown to precipitateidentity change (Magai and Nusbaum, 1996). A temporary period offluctuation at a point of rapid developmental change is the necessarycriterion for phase transitions (Van Geert, 1994). Indeed, Magai andNusbaum (1996) integrate a number of lines of evidence indicating thatpersonal identity and self-concept change abruptly, even explosively,while underlying personality dispositions endure; but personality andidentity are not clearly distinguished in their account. It is possible thatidentity change is more precipitous than personality change, but fur-ther research is needed to decide this issue.


Conclusion

Early in this chapter we reviewed a newly formulated theory of emo-tional self-organization at three time scales. According to this theory,real-time interpretations, mood-like states, and crystallizing personal-ity dispositions were linked through principles of psychological en-trenchment related to the non-resolution of longstanding goals andwishes. Personal identity was then modeled as a specialized product ofpersonality that relies on reflective and often conscious activities thatare inherently dialogical and autobiographical. We described identityas a lineage of real-time self-appraisals that generally retains continuityyet is subject to periodic transitions. These appraisals revise and extendnarrative accounts of who we are, or who we would like to be, ininterpersonal, dialogical contexts. As a final word on the relation be-tween personality and identity, we suggest that identity not only de-pends on personality as grist for its mill; it also guides the direction ofpersonality change and consolidation, at least in the years followinginfancy. It thus exists in a mutually causal relation to both personalitycontinuity and personality change.These ideas are consistent with a new family of models that apply

principles of self-organizing dynamic systems to psychological pro-cesses in general and socioemotional processes in particular. However,much of the present framework is still untested and may need substan-tial revision before it is authenticated.We present it here, speculatively,in order to demonstrate how self-organization can explain not onlydevelopmental change, but also continuity or sameness in cognitive–emotional processes. Some theorists suggest that all dynamical systemshave an implicit memory – that is, an accretion of identity – resultingfrom subtle changes to their elements and their connections whenengaged in self-organizing processes (Schwartz and Russek, 1997). Thisprovocative hypothesis will take years to explore fully. However, iftrue, it would suggest that both implicit and explicit identity expresstendencies that extend well beyond human psychology to all naturalphenomena.


COMMENTARY

Two faces of identityCarol Magai

Marc Lewis, who was one of the first to apply dynamic systems con-structs to an explication of personality development, and who did so ina very exciting way, provides us in the present chapter with morestimulating ideas and further food for thought. Here Lewis and Ferraridistinguish between implicit identity, which refers to personality char-acteristics, and explicit identity, which is a ‘‘specialized product’’ of thepersonality system and is self-reflective and dialogic. Both of these‘‘identities’’ are subject to change as modeled in a dynamic systemsframework. This is an interesting and potentially useful distinction; Iwould like to explore a comparison between these two aspects a bitfurther.It seems tome that implicit and explicit identitymay be differentially

stable and differentially permeable to change. An examination of theliteratures on ‘‘personality’’ and ‘‘self systems’’ suggests that implicitidentity or personality may be more stable than explicit identity or selfidentity. According to a number of accounts, personality is in largemeasure grounded in emotional dispositions that have roots in tem-perament (McCrae and Costa, 1996; Goldsmith, 1994), though regulari-ties and repetitive emotional experiences triggered by environmentalconditions also shape dispositions (Malatesta, 1990). In any event,whether or not emotional dispositions of particular individuals accruelargely from temperamental substrates or from repetitive emotionalevents during early development, it is clear that mood states, which areproducts of personality, are fairly stable by the time individuals reachadulthood. The literature indicates that stability coefficients for positiveand negative affect are quite high. For example, Deiner and Larsen(1994) report that the correlation between positive affect scores over aone-year period was .74; similarly, Ormel and Schaufeli (1991) foundthe correlation for negative affect over seven years was .57.It seems more likely that explicit identity, given its dialogic nature,

might be more susceptible to change, though one suspects that the coreemotional dispositions would retain their overriding influence. Anumber of individuals have noted that even people who appear toundergo a striking change in personality retain important components

199

of their personality. For example, DavidMcClelland (1984) recounted achance encounter with Richard Alpert, some twenty-five years aftertheir short acquaintance when Alpert was a graduate student studyingpsychology under him. McClelland describes what at first blush ap-peared to be a major personality transformation – one that transformeda young, highly intelligent, and charasmatic Harvard assistant profes-sor into Ram Das, a servant of the Lord, under the influence of theIndian guru, Maharaji-ji. Alpert’s whole appearance had changed, ashad his attitudes, career, and purpose in life; his intellectual and spiri-tual inspiration shifted from neo-Freudianism to the eastern psycholo-gies of Buddhism and Hinduism. One rarely sees such striking change,and yet, as McClelland pointed out, fundamental aspects of Alpert’spersonality could still be perceived. He still had the same expressivemovements, he still had a strong interest in internal psychic states, hewas still charming, he was still very much involved in power games,and still felt guilty about being interested in power. This suggests thatalthoughAlpert’s explicit identity had changed substantially, there hadbeen far less change in implicit identity.It is these kinds of observations as well as a growing research litera-

ture on self-concept that makes us question how the two ‘‘identities’’relate to one another. As the editors of this volume have noted in theirintroductory chapter, the construct of self-concept (Lewis and Ferrari’sexplicit identity) has undergone significant change within the socialand personality literature in recent times. Indeed, within this literaturethere has been a movement away from viewing the self-concept as aunified, stable, integrated entity, to one that involves seeing the self as arepresentational system that is large, complex, dynamic, and subject tochange. The self is now seen as a cognitively complex entity constitutedof self-schemas, which are active, functional structures shaping a per-son’s emotional and behavioral responses to events (Aron, Aron, Tu-dor, and Nelson, 1991; Markus and Kitayama, 1991).In a recently completed study (Magai, 1999), we attempted to exam-

ine the stability of these two aspects of personality in the same peopleover an eight-year period. In effect, we were studying both implicit andexplicit identity, thoughwe referred to the two aspects of personality asdispositional tendencies (emotion traits) and characteristic adaptations(views of self). Stability coefficients for the emotion traits of anxiety,depression, interest, anger, anger-in, anger-out, and aggression rangedfrom .47 to .75, with an average of .62; only anger-out showed signifi-cant change over the eight years. These findings are in accord withother research, noted earlier, on the robust nature of individual dif-ferences in positive and negative affect. On the other hand, our respon-dents reported moderate changes in perspectives, goals, personality,

200 Carol Magai

feelings, and ways of relating (an average score of 2.9 on a scale of1=not at all, to 5 = extremely). Moreover, ratings by family and friendsof the informants were significantly correlated with self-reports ofchange for all variables except goals. Another interesting finding wasthat this kind of personality change was associated with positive andnegative interpersonal life events of an intimate nature, such as mar-riage, divorce, and death of loved ones that took place during the eightyears;moreover, it was not associatedwith other high and lowpoints inpeoples’ lives involving careers, changes in residence, andmore distantsocial relationships. This suggests that interactions with others have asubstantial impact on aspects of the self that relate to self-concept andmotives, which would seem to support Lewis and Ferrari’s idea thatexplicit identity is dialogic in nature, and perhaps may even representan interiorization of dialogueswith others. In fact, elsewherewe (Magaiand McFadden, 1995; Magai and Nusbaum, 1996) have argued thatalthough intense emotion of a positive or negative nature is a necessarycondition of personality change, change is unlikely to be sustained inthe absence of interpersonal support and mental elaboration.Lewis and Ferrari suggest that the same three dynamic system prin-

ciples may govern change in explicit and implicit identity. Given theirapparent differential permeability to change it is possible that theunderlying dynamic principles may be less similar than would at firstappear to be the case. It might help to adopt an ideographic approach tothe issue, say contrasting cases that exemplify dramatic, structuralchange involving change in emotion traits, as in the case of post-traumatic stress disorder, and cases of explicit identity change (as in thecase of Richard Alpert), to begin to model these different kinds ofidentity change (Magai and Haviland-Jones, in press).

201Commentary on Lewis and Ferrari

COMMENTARY

Two faces of identityCarol Magai

Marc Lewis, who was one of the first to apply dynamic systems con-structs to an explication of personality development, and who did so ina very exciting way, provides us in the present chapter with morestimulating ideas and further food for thought. Here Lewis and Ferraridistinguish between implicit identity, which refers to personality char-acteristics, and explicit identity, which is a ‘‘specialized product’’ of thepersonality system and is self-reflective and dialogic. Both of these‘‘identities’’ are subject to change as modeled in a dynamic systemsframework. This is an interesting and potentially useful distinction; Iwould like to explore a comparison between these two aspects a bitfurther.It seems tome that implicit and explicit identitymay be differentially

stable and differentially permeable to change. An examination of theliteratures on ‘‘personality’’ and ‘‘self systems’’ suggests that implicitidentity or personality may be more stable than explicit identity or selfidentity. According to a number of accounts, personality is in largemeasure grounded in emotional dispositions that have roots in tem-perament (McCrae and Costa, 1996; Goldsmith, 1994), though regulari-ties and repetitive emotional experiences triggered by environmentalconditions also shape dispositions (Malatesta, 1990). In any event,whether or not emotional dispositions of particular individuals accruelargely from temperamental substrates or from repetitive emotionalevents during early development, it is clear that mood states, which areproducts of personality, are fairly stable by the time individuals reachadulthood. The literature indicates that stability coefficients for positiveand negative affect are quite high. For example, Deiner and Larsen(1994) report that the correlation between positive affect scores over aone-year period was .74; similarly, Ormel and Schaufeli (1991) foundthe correlation for negative affect over seven years was .57.It seems more likely that explicit identity, given its dialogic nature,

might be more susceptible to change, though one suspects that the coreemotional dispositions would retain their overriding influence. Anumber of individuals have noted that even people who appear toundergo a striking change in personality retain important components

199

of their personality. For example, DavidMcClelland (1984) recounted achance encounter with Richard Alpert, some twenty-five years aftertheir short acquaintance when Alpert was a graduate student studyingpsychology under him. McClelland describes what at first blush ap-peared to be a major personality transformation – one that transformeda young, highly intelligent, and charasmatic Harvard assistant profes-sor into Ram Das, a servant of the Lord, under the influence of theIndian guru, Maharaji-ji. Alpert’s whole appearance had changed, ashad his attitudes, career, and purpose in life; his intellectual and spiri-tual inspiration shifted from neo-Freudianism to the eastern psycholo-gies of Buddhism and Hinduism. One rarely sees such striking change,and yet, as McClelland pointed out, fundamental aspects of Alpert’spersonality could still be perceived. He still had the same expressivemovements, he still had a strong interest in internal psychic states, hewas still charming, he was still very much involved in power games,and still felt guilty about being interested in power. This suggests thatalthoughAlpert’s explicit identity had changed substantially, there hadbeen far less change in implicit identity.It is these kinds of observations as well as a growing research litera-

ture on self-concept that makes us question how the two ‘‘identities’’relate to one another. As the editors of this volume have noted in theirintroductory chapter, the construct of self-concept (Lewis and Ferrari’sexplicit identity) has undergone significant change within the socialand personality literature in recent times. Indeed, within this literaturethere has been a movement away from viewing the self-concept as aunified, stable, integrated entity, to one that involves seeing the self as arepresentational system that is large, complex, dynamic, and subject tochange. The self is now seen as a cognitively complex entity constitutedof self-schemas, which are active, functional structures shaping a per-son’s emotional and behavioral responses to events (Aron, Aron, Tu-dor, and Nelson, 1991; Markus and Kitayama, 1991).In a recently completed study (Magai, 1999), we attempted to exam-

ine the stability of these two aspects of personality in the same peopleover an eight-year period. In effect, we were studying both implicit andexplicit identity, thoughwe referred to the two aspects of personality asdispositional tendencies (emotion traits) and characteristic adaptations(views of self). Stability coefficients for the emotion traits of anxiety,depression, interest, anger, anger-in, anger-out, and aggression rangedfrom .47 to .75, with an average of .62; only anger-out showed signifi-cant change over the eight years. These findings are in accord withother research, noted earlier, on the robust nature of individual dif-ferences in positive and negative affect. On the other hand, our respon-dents reported moderate changes in perspectives, goals, personality,

200 Carol Magai

feelings, and ways of relating (an average score of 2.9 on a scale of1=not at all, to 5 = extremely). Moreover, ratings by family and friendsof the informants were significantly correlated with self-reports ofchange for all variables except goals. Another interesting finding wasthat this kind of personality change was associated with positive andnegative interpersonal life events of an intimate nature, such as mar-riage, divorce, and death of loved ones that took place during the eightyears;moreover, it was not associatedwith other high and lowpoints inpeoples’ lives involving careers, changes in residence, andmore distantsocial relationships. This suggests that interactions with others have asubstantial impact on aspects of the self that relate to self-concept andmotives, which would seem to support Lewis and Ferrari’s idea thatexplicit identity is dialogic in nature, and perhaps may even representan interiorization of dialogueswith others. In fact, elsewherewe (Magaiand McFadden, 1995; Magai and Nusbaum, 1996) have argued thatalthough intense emotion of a positive or negative nature is a necessarycondition of personality change, change is unlikely to be sustained inthe absence of interpersonal support and mental elaboration.Lewis and Ferrari suggest that the same three dynamic system prin-

ciples may govern change in explicit and implicit identity. Given theirapparent differential permeability to change it is possible that theunderlying dynamic principles may be less similar than would at firstappear to be the case. It might help to adopt an ideographic approach tothe issue, say contrasting cases that exemplify dramatic, structuralchange involving change in emotion traits, as in the case of post-traumatic stress disorder, and cases of explicit identity change (as in thecase of Richard Alpert), to begin to model these different kinds ofidentity change (Magai and Haviland-Jones, in press).

201Commentary on Lewis and Ferrari

CHAPTER 9

A self-organizational approach toidentity and emotions: anoverview and implicationsE. Saskia Kunnen, Harke A. Bosma, Cor P. M. Van Halen, andMatty Van der Meulen

Introduction

The perceptions, ideas, and feelings we hold of ourselves, whetherconsciously or not, are a natural, intrinsic part of our daily functioning.They seem a necessary ingredient of our attempts to sustain a basicsense of identity, or as Erikson (1950) described it, to protect a sense ofpersonal continuity, unity, and social recognition. As such, they makeus an individual in the eyes of others as well as ourselves. But they arealso the orientating instruments by which we try to bring some coher-ence to our own life or to set out new directions. For psychologists,self-referential processes retain an intriguing, though somewhat elu-sive character, which makes it difficult to capture them in a single, neatconceptual and empirical framework. In a sense, the content of suchprocesses seems to typify the person one has grown to be. Yet, inanother sense, they constitute the foreshadowing of the person onecould become. Moreover, in both cases, as James explained, these pro-cesses introduce an element of recursivity in the way we experienceourselves, without totally merging with ongoing mental, behavioral,and social processes.In chapter 1, we stated that in past decades psychologists have tried

to deal with the indefinite status of such generic concepts as self andidentity by exclusively focusing on the more tangible aspects. Thisgenerally meant that self and identity were equated with the self-concept, which was seen as a relatively stable, generalized set of self-representations that people have formed during their lives. Thisopened the way to the still predominant empirical practice of usingself-descriptions and self-ratings as reliable indications of an underly-ing self-concept. Thus, the more static, decontextualized, and cognitivefeatures were stressed at the expense of the more dynamic, contextual,and emotional qualities.A general dissatisfaction with the far from spectacular results of

traditional self-concept research (e.g. Wylie, 1979) eventually forced

202

self-psychologists to venture in new directions. Van der Meulen’s re-view of recent developments in self-concept theory and research inchapter 2 provides an extensive discussion of the different attempts tocompensate for the lack of attention to affect, context, and variability.She described how the various reconceptualizations and ensuing em-pirical research to date have shown that the beliefs persons have ofthemselves are closely associated with affect, and are a mixture of moregeneralized, enduring self-attributions and more situational and tem-poral ones. Unfortunately, when taken together, the empirical findingsconstitute a rather fragmentary picture of partial aspects, without thepromise of an adequately integrated framework. Moreover, these newapproaches remain rooted in a view of self and identity as a cognitive,internal structure that analytically and empirically stands apart fromthe accompanying emotional and contextual processes.The authors represented in this volume propose a radically new

perspective on self and identity, theoretically as well as empirically.The central notion in this approach is the dynamics of self-organization,emerging from the reciprocity of cognitive, emotional, and contextualprocesses. More specifically, the following key issues are addressed.First, in what way(s) do emotional and cognitive processes implicateeach other within the dynamics of self and identity? Second, how dothese processes in interaction take shape within the contexts in whichone is functioning? Third, how do seemingly contradictory characteris-tics such as stability and variability combine in self-organizing pro-cesses? While such issues can be posed in simple and straightforwardterms, answering them is a totally different matter. This is reflected inthe contributions to this volume. In dealing with the complexity of thedynamics involved in self and identity, each author has developed heror his own vantage point and explanatory devices. Hence, we canimagine that a first reading may have resulted in some confusion aboutthe underlying divergences and similarities between the different con-tributions. In this final chapter, we will, therefore, start with a briefoverview of the chapters in order to identity both the underlyingassumptions and important future issues.

A concise overview

Setting the stage for a dynamic perspective

In disputing the exclusively cognitive connotation of self and identity atthe outset of this volume, it seemed desirable to approach questionsabout the role of emotions from the standpoint of emotion theory. Thiswas the task Frijda fulfilled in chapter 3. Basic principles in Frijda’s

203A self-organizational approach to identity and emotions

(1986) emotion theory are that emotions indicate what matters to aperson, and that they motivate human behavior. One only has toreverse this line of reasoning to come to the conclusion that if self andidentity do matter and motivate human behavior they should logicallyinclude emotions. This subsequently raises the question in what ways,according to this theory, self and emotions may be related. In none ofthe emotion components that Frijda distinguishes is the self a necessaryingredient, at least not in its most common meaning as explicit self-awareness or self-representation. The only prerequisite is what Frijdacalls the ‘‘transcendental self’’: the logical reference point of any experi-ence (including emotions). The transcendental self has no mentalcontent beyond being the locus of a vague sense of familiarity andinitiative. In the remainder of his analysis, Frijda actually turns thingsaround. Though aspects of self and identity may not be prerequisites ofemotion, it can be shown that, if properly specified, different aspects ofself and identity are implicated by emotions.First, there are many emotions in which the self figures as an object,

either in becoming aware of undergoing a particular emotion (e.g. ‘‘Ifeel fear’’) or as the referent of an emotional appraisal (e.g. ‘‘I hatemyself’’ or ‘‘I feel proud of myself’’). Such self-representations andself-evaluations are often involved in the more adult forms of shame,guilt, and pride. Besides, one’s own emotions can become objects ofemotional appraisal, because of the personal and social implications ofhaving or showing these emotions. This is the ‘‘significance’’ of anemotion, which motivates the further regulation of emotions. The sig-nificance often involves self-representations such as ‘‘a Friesian doesn’tcry,’’ or ‘‘I’m an anxious person.’’Second, a sense of self may also be considered a possible factor in

estimating one’s own coping resources. The perceived competence indealing with the emotion-eliciting event is one of the determinants inthe coming about of an emotion, in addition to the appraised event andthe individual’s concerns. However, such a sense of competence isentirely implicit in the appraisal of the outcome of an emotional event,and no self-representation or conscious self-evaluation needs to beassumed here.The most pervasive sense in which the self is implicated in emotions

is in the form of concerns, of ‘‘what and how the subject is’’ (Frijda,chapter 3, p. 54). ‘‘Concern’’ is a broad term for the individual’s mo-tives, goals, attachments, aversions, etc., in other words all that mattersto the person. Sometimes, the term ‘‘self’’ is used to denote thesepersonal concerns. Persons can have concerns about themselves, forexample the wish to achieve and maintain a certain identity. Accordingto Erikson (1950), the need for a sense of identity is one of the most


fundamental human concerns. Frijda (1986) himself mentions self-esteem as one of the most basic concerns. These concerns, in turn, canlead, in the course of the individual’s development, to the morecomplete forms of self-conscious emotions (e.g. pride, shame).Summarizing, Frijda’s chapter provides several pertinent contribu-

tions to the central issues in this volume. Though self-representationsmay not be a necessary ingredient of emotions, emotions do play amajor part in self and identity. Most importantly, they indicate what isessential to the person: his or her concerns and competencies. Theneed for a sense of identity and self-esteem are among the most funda-mental human concerns. Moreover, by integrating emotions into thestudy of self and identity, a direct link between the individual and thecontext is guaranteed. Because emotions are fundamentally relational,fusing the subject and the object of an emotion, this integration canresolve the problems inherent in the decontextualized nature of self-representations research. The conceptualization of emotion as a systemof loosely coupled affective, cognitive, and behavioral components(see also Frijda, 1993) offers new opportunities for an understanding ofhow emotions and self may be intertwined. It also preludes a dynamicsystems perspective, because the mutual interactions between thedifferent emotion components and aspects of self are the definingelements in the eventual course of the emotion processes.In his critical commentary, Epstein elaborates on the relation between

emotions and self-aspects in a way that to an important extent goesalong with Frijda’s componential and dynamic view. However, herelegates much of Frijda’s emotion theory to the so-called ‘‘experientialsystem.’’ He describes the experiential system as the executive part ofthe self, and as something that has to be clearly separated from the selfas the object of rational considerations. The systemoperates on the basisof an intuitive theory about oneself, about the world, and about theirinterrelationship. In this system, emotions indicate what is of signifi-cance to a person. Frijda and Epstein also agree that the dynamicinteractions between implicit appraisals and emotions are largely non-conscious, but both describe the dynamics involved only in globalterms. In this connection, the application of dynamic systems theorymight be very useful.Chapter 4 by Van Geert is intended to be a general introduction to

dynamic systems thinking. Since many contributors to this volume useor advocate a dynamic systems perspective on issues of self, identity,emotions, and development, this chapter offers the necessary back-ground information for a preliminary evaluation of the usefulness ofsuch a perspective. A dynamic systems perspective offers insight intoprocesses, and thereby into questions concerning stability and change.


Van Geert describes a dynamic system as something that ‘‘consists ofelements that exert specific influences or forces upon one another andby doing so, change each other’s and their own properties’’ (chapter 4,p. 64). In their commentary, Camras and Michel correctly point to oneof the critical implications of this definition. In applying a dynamicsystems perspective, one should define constructs in terms of systems,that is, in terms of a conglomerate of elements (e.g. prey and predatorfish). In this volume, therefore, the question has to be posed as to whatconglomerate of elements the generic terms ‘‘identity’’ or ‘‘self’’ referto. It should also become clear what is part of the system and what ispart of the system’s environment. Indeed, most authors in this volumehave in one way or another, implicitly or explicitly, tried to deal withthese implications.An important view of dynamic systems theory is that in order to

explain variation over time, that is patterns of stability and change, it isnot necessary to resort to external causal forces, or an inbuilt, determin-istic ground plan. Instead, the most conspicuous characteristic of dy-namic systems is self-organization. Complex, non-linear systems tendto self-organize into a state of increased order/complexity. Higher-order phenomena, for instance self-constructs or identity, can be seen asemerging from the dynamics of the system rather than being alreadypresent as internal structures. In dynamic systems terms, changes resultfrom the time-dependent interactions between the constituent compo-nents of the system. These interactions – the set of relationships be-tween the components of the system – form the core of a dynamicsystems model, and not the physical instantiation of the components(e.g. molecules, fish, classroom reprimands, self-representations, ident-ity choices). This means that the general process characteristics ofdynamic systems, such as different types of attractors and bifurcations,can be applied to describe the dynamics in a specific domain such asself and identity. At a qualitative level, this can have an importantheuristic value, but, as Van Geert has demonstrated, the relationshipsbetween the components can also be specified in mathematical func-tions which are surprisingly simple but which nevertheless capture theessence of the system’s behavior over time.

Four elaborations on a theme

The chapters by Frijda and by Van Geert have informed us about therole of self-phenomena in emotions, and about the application of adynamic systems perspective in order to approach questions concern-ing stability and change. Chapters 5 through 8 provide in-depth dis-cussions of emotions and context as formative conditions from which


self and identity emerge in a self-organizational process. However,each of these chapters has a different focus. Nevertheless, in certainways, they also complement each other and build on the foundationslaid by Frijda and Van Geert.Frijdamakes it clear that one’s self can become the object of emotions.

In order to experience oneself as an object of an emotion, some notion ofoneself as enduring and stable over time must exist. But, when andhow, in ontogeny, do these self-related phenomena emerge? These arethe kinds of questions Fogel explored in chapter 5.Fogel based his approach on two premises: psychological experience

(a) always implies a relationship (with the object-world, others, oroneself) and (b) is always dynamic and changing.Within this inherent-ly relational and dynamic world, the crucial question then becomeshow people can ever develop a sense of uniqueness and continuity.1Fogel considers emotions as playing a central role in this process,because they are a way of discovering the meaning of a relationship forthe self and, hence, the unique position of the self in a relationship. Theperception of consistency in one’s emotional experience begins to yielda sense of permanence of the self over time.With these premises and hisalmost programmatic statements, Fogel occupies a distinctive positionin the context of this volume: all experience is dynamic, self is rooted inrelationships, and emotions (which in themselves are inherently con-textual) play a constitutive role.Central to Fogel’s analysis of the dynamics of self and identity is his

distinction between three time scales, and the observation that, depend-ing on the time scale, different types of emotion go together withdifferent experiences of uniqueness. A first and basic sense of selfalready comes into existence when the emotional experience that ac-companies a currently ongoing relationship stabilizes into a primaryorientation towards the relationship. This takes place on the time scaleof (micro)seconds. In turn, orientations coalesce into identifiable events.On the time scale of minutes to days, sequences of events may self-organize into coherent and stable frames. Frames are similar to narra-tive scripts and enable individuals to experience a more explicit, stablesense of self. The third time scale, the scale of ‘‘development,’’ extendsover years. The corresponding sense of self is identity: the experience ofan enduring self. It is rooted in the relationwith one’s past or future, butalso in the surrounding cultural frames. Although identity may devel-op into a narrative account of oneself, Fogel considers it to be anemotional experience rather than a representation.Regarding the main questions considered in this volume, Frijda and

Fogel have much in common. Both argue for a differentiated concep-tualization of self and identity, and both agree that emotions are


relational experiences. In a certain sense, they also complement eachother. Frijda makes clear that self-related phenomena are not a necess-ary part of the emotional experience. Fogel gives a similar primacy toemotions but seeks to understand how aspects of the sense of selfemerge from emotional experience in relationships. In our opinion,Fogel’s description of the emergence of a sense of self on three differentbut interrelated time scales offers a significant and elegant contributionto our understanding of the constitutive role of emotions in the devel-opment of aspects of self. It is interesting to note that Fogel hardly refersto explicit representational content. Apparently, he, like Frijda, con-siders self-representations and self-awareness to be unnecessary inunderstanding the way in which the self becomes organized.In his commentary on Fogel, Jansz suggests an alternative explana-

tion for the emergence of a sense of identity. In contrast to Fogel, hesees identity and its constituent feelings of continuity and consistencymainly as a social construction that people have learned to ascribe tothemselves. In this they are guided by the prevailing notion of person-hood, which in Western societies is characterized by an emphasis onpersonal consistency, uniqueness, autonomy, and rationality. In hisreply to Jansz (not published) Fogel states that the difference betweenhis and Jansz’s view stems from a difference in starting point. InFogel’s view structures have to be explained, whereas relations needno explanation. Jansz starts from structures and has to explain rela-tions. Thus, Jansz emphasizes the role of cultural scripts in formattingidentity through dialogical exchange, and the cultural fact that emo-tions count as a violation of the aspired image of equanimity. Fogel, onthe other hand, concentrates on the stabilizing role of emotions in theemergence of a clear sense of identity within dialogical relationships.Both ways of relating emotions to the dialogical nature of identityreturn in chapter 6. Hermans and Hermans-Jansen combine Fogel’srelational perspective with culturally determined patterns of meaning-making by introducing a new element: a strong recognition of thedevelopmental possibilities that are brought about by the human fac-ulty for self-reflection. They start from the assumption that in thecourse of human development emotional responses are increasinglyinfluenced by the relationship individuals have with themselves. Interms of the differentiation Frijda makes, Hermans and Hermans-Jansen focus on experiences in which the subject becomes the object ofthe emotion, and in particular on the significance of the emotion. Thesignificance plays an important role in emotion regulation, in the wayindividuals interpret and express, or do not express, their emotions.This is also the domain in which the culturally determined emotionscripts figure.


In their analysis of the dynamics and complexity of the relationbetween self and affect, Hermans and Hermans-Jansen elaborate onJames’ distinction between ‘‘I,’’ i.e. the self-as-knower, and ‘‘Me,’’ i.e.the self-as-known. They see self-awareness as an essential feature of theI. Individuals can reflect on their feelings, appraise them, and developfeelings about these feelings. Thus, the self authors dynamic and com-plex compositions of personally significant, affect-laden meanings (i.e.valuations). Such narratives typically take the shape of an ongoingself-dialogue. Instead of being the omnipotent author of oneself, theself-reflecting individual continuously changes perspectives. Each per-spective in fact implies a different I-position, entailing its own inter-nalized voice and telling its own story. The different narrative positionsusually revolve around recurring conflicts and obstacles one encoun-ters in becoming the person one wants to be or should be. Thus,self-narratives have developmental implications. To serve as a fruitfulpoint of departure for further identity development, the inner voicesthat speak for the own person should be brought into open dialoguewith each other.The notion of the ‘‘dialogical self’’ returns in the valuation theory and

the self-confrontation method, which can be employed to study thedialogical self empirically. Both in its conceptual connotations and inits methodology, the approach of the dialogical self has much incommon with the phenomenological–dialectical personality model ofVerhofstadt-Deneve, that is used as starting point in the application ofpsychodrama techniques.The authors of chapter 7, Haviland-Jones, Boulifard, and Magai,

present a similar componential conceptualization of the self. By the selfthey mean self-concept, identity, or personality, terms which they usemore or less synonymously. They take emotions as a starting point andstudy how emotions organize personality, in contrast to Hermans andHermans-Jansen, who argue that I-positions organize patterns of emo-tions. Haviland-Jones and her colleagues give a lively account of theirquest for tools that enable them to arrive at a better understanding ofthe dynamic interplay of cognitive and affective components of the self.They suggest that dynamic systems thinking and modeling offer excit-ing new possibilities. Their earlier studies failed to provide adequateinsight into the role of emotions in processes of developmental changein self-concept and identity. It was their idea, therefore, that analyzingone aspect of these developmental processes might provide furtherinsight. This brought Haviland-Jones and colleagues to the study ofmicroprocesses of fear by means of a dynamic systems model. Theybase their model of emotion on the theory of Tomkins. In the modelthey combine the idea of continuous change in the system’s input with


various discrete emotions as its categorical output. Such a model alsoallows for continuous transitions from one emotion to another. Thedynamic systems methodology developed by Van Geert makes itpossible to run computer simulations of these continuous emotionalprocesses and to study the complex interactions involved.The computer simulations of emotion processes carried out by Havi-

land-Jones, Boulifard, and Magai suggest that differences in the orig-inal state of the person, or in the stimulus itself, may make largedifferences in the emotional experience, with implications for immedi-ate behavior as well as long-term identity scripts. Haviland-Jones andcolleagues have demonstrated that information parameters are inthemselves capable of producing shifts in emotion categories. Further-more, conscious, goal-oriented knowledgedoes not seem to be essentialfor these shifts.Themodel and simulations also give room to interesting speculations

about the interdependencies of emotional and cognitive processes: briefand unconscious emotional processes could evidence ‘‘phase shifts,’’which could even lead to changes in stable patterns in the person’sself-concept. Beliefs and knowledge about emotions also play a role inthe processing of self-relevant information. They may enable individ-uals to change themselves, but due to the intricacies of the mostlyunconscious dynamics of microemotional processes, such changes cantake unpredictable courses. These may or may not correspond withhow individuals understand and describe themselves: A ‘‘good linearstory’’ about how one came to be – call it one’s narrative or one’smeaning-making – does not necessarily reflect the workings of themicroemotional processes in the change and stability of the self. Havi-land-Jones and colleagues conclude by suggesting that the iteration ofthe phase shifts on the microlevel and the meaning-making level maylead to changes in one’s self-view.In his commentary on chapter 7 Heymans does not contest this

suggestion, but he warns us that one should not forget the role ofcultural constructions of the self in the individual’s developmentalprocess, or the individual’s own active involvement. In line with this,he also observes a conceptual gap between the description of self-organizational processes occurring at themicrolevel of ongoing,mostlyunconscious emotion processes, and at the macrolevel of active, con-scious identity construction.Heymans broaches an important and thorny issue that is still left

unanswered. Throughout this volume, there seems to be a dichotomyin the kind of explanations that are used to describe the dynamicorganization of self and identity. Self and identity are seen either asemergent qualities of emotion-driven self-organizational processes


(Fogel, Haviland-Jones et al.) or as a self-reflective construction ofmeanings (Hermans and Hermans-Jansen, Jansz, Heymans).In chapter 8, Lewis and Ferrari attempt to unite these two types of

explanation under the header of ‘‘implicit’’ versus ‘‘explicit identity.’’Unlike Epstein’s CEST, Lewis and Ferrari do not try to solve this issueby postulating two totally independent self-systems. Instead, theystrictly adhere to the organizing principles of dynamic systems, prin-ciples that help to explain how complex, higher-order organization andstable states emerge from the interactions between lower-order compo-nents. As such, Lewis and Ferrari present an encompassing applicationof dynamic systems thinking in the domain of self and identity.According to Lewis and Ferrari, implicit identity – the sameness and

uniqueness of one’s personality – is a naturally evolving characteristicthat we have in common with most animals and that already exists inyoung infants. It emerges from recurrent, self-stabilizing patterning ofappraisal–emotion interactions. The self-organizational processes in-volved build on each other along three different time scales. In realtime, the appraisals and emotional states that accompany our actionsare coupled into temporary ‘‘emotional interpretations’’ (EIs). TheseEIs dissipate as soon as the initial goals are achieved. When we areunable to achieve the goals, the goal state perseveres. The emotions thatthis situation generates have a consolidating effect on the current EIs,turning them into more or less lasting moods that may persist forminutes to weeks. Such moods have a facilitating effect on the interpre-tations and action-tendencies that are congruent with them. They alsoconstrain the likelihood that alternative EIs develop into a mood.Through this habitual shaping of themind, the configuration of EIs thatfit in with a recurrent mood may start to function as a psychologicalattractor state, and become a more permanent ingredient of one’spersonality. This is especially the case when long-lasting needs andwishes are involved that are difficult to accomplish, given the personalcircumstances one lives in. In a nutshell, then, psychological continuitycrystallizes from themind’s preoccupationwith ‘‘unfinished business,’’rather than being the intrinsic property of an underlying, invariantself-structure.A fully developed personal identity relies on the sameness of the

implicit identity, but also implies a coherent composition of self-appraisals. This is what Lewis and Ferrari call the ‘‘explicit identity.’’ Itcontains people’s ‘‘statements, stances, policies, or stories about whothey are and how they live their lives’’ (p. 178), and stems from seman-tic, reflective, and often conscious activities of a more autobiographicaland dialogical nature. Again, such self-reflective activities occur es-pecially when people have to deal with unresolved needs and goals or


the conflicts that this arouses. The main function of self-narratives andself-dialogues is to account for the fact that one has not yet succeeded inbecoming the person onewants to be. This is done in anticipation of realor imagined listeners. However, since the potential audience can differwidely in their appreciation of the narrator’s intentions, and the narrat-ing individual can adopt a multitude of I-positions, there can be noconclusive understanding of oneself.Lewis and Ferrari also stress that from a dynamic systems perspec-

tive, identity change can be modeled through different kinds of changein self-organizing systems. Identity change across developmental timemay stem from the emergence of new patterns, the strengthening ofnew habits, and dramatic events. However, in her commentary, Magaitries to qualify Lewis and Ferrari’s assertion that the same dynamicsystem principles govern change and stability in implicit and explicitidentity. She refers to the empirical evidence that, precisely because ofits dialogical nature, explicit identity appears to be much more suscep-tible to change than the more deeply entrenched implicit identity. Shethereforewonders whether at a more specific level each type of identitymay imply rather different types of organizing mechanisms and pro-cesses. This is one of the fundamental questions that will be discussedin the following section. In their reaction, Lewis and Ferrari (unpub-lished) agree that there are important differences between identity andpersonality: not only does identity change more easily, but probably itcan also changemore partially. They suggest however that the differen-ces between both systems do not reside in the general dynamic systemsprinciples, but in themore specificmechanisms and processes onwhichthese general principles operate.In the course of Lewis and Ferrari’s comprehensive discussion,

many of the conclusions of previous chapters reappear: the emphasisof Frijda, Fogel, and Haviland-Jones and her colleagues on the or-ganizing role of emotions in self and identity; Van Geert’s outline ofdynamic system principles; Fogel’s distinction between three timescales and levels of organization; and Hermans and Hermans-Jansen’ssketch of the dialogical self. However, as Magai has already observed,this comprehensiveness also carries the cost of addressing thedynamics of self and identity in a general and abstract way. In theirstudy of these dynamics, researchers need to be aware that there aresome decisive analytical implications that have to be taken into ac-count when the general principles of self-organization are translatedto the domain of self and identity. We will, therefore, conclude thisclosing chapter by discussing some of the most compelling of theseimplications.


Implications of a dynamic systems approach to self and identity

In the first two chapters of this book three broad problem areas wereidentified in recent developments in the study of self and identity: therole of emotions in a predominantly cognitivistic approach, the ques-tion of how to take the context into account, and the issue of stabilityversus variability and change. Most of the authors in this book share aperspective that differs fundamentally from the approaches that haveto contendwith these three problems. They see emotions and context asformative conditions from which self and identity emerge in a self-organizational process. In such a dynamic systems perspective thevarious elements ‘‘cognitions, emotions, and context’’ are viewed as acomplex, interacting network in which the embeddedness in the con-text and the changeability are inherently given. The limitations of thetraditional cognitivistic approach can be overcome via this new per-spective. The use of a dynamic systems perspective has brought us awealth of new ideas, but it also raises new questions; the ensuingimplications primarily have to do with how the concept under study isapproached and defined.In his chapter Van Geert distinguishes four steps necessary for ap-

plying such a perspective. The first step, conceptual construction, con-cerns the definition of the components of the system. In addition,Camras and Michel stress the point that a dynamic system alwaysconsists of multiple layers, and that the phenomena emerging fromthese components should also be defined clearly. Further, it must alsobe clear what exactly is meant by the different concepts used in the selfand identity literature (e.g. self-concept versus identity). A first implica-tion therefore concerns an elaboration of self and identity in terms ofmultiple layers of components and emerging phenomena. As a furtherimplication we will discuss the distinction between the various con-cepts that are central to this volume.The second step in Van Geert’s description concerns the formu-

lation of the interactions between the components. Applying a self-organizational perspective to self and identity implies that theprocesses involved in the interactions between the components, withinand between levels, have to be described. This is another implication tobe elaborated here. Furthermore, as Camras and Michel point out, therelation between identity development and its context has also to bedefined. Yet another implication is that there are alternative ways ofconceptualizing this relation.The third step mentioned by Van Geert concerns a quantitative

approach: the translation of the conceptual model into a mathematical


one. Elaboration of this step goes beyond the scope of the presentchapter, although we will, by way of an example, briefly describe howdifferent conceptualizations of the context could be translated into amathematical model.Van Geert’s final step concerns the exploration of the model. This, of

course, is needed both in a qualitative and in a quantitative approach:researchers have to formulate hypotheses and test these against empiri-cal data. A final implication of a dynamic systems approach has to dowith the kind of data and methods of analysis that are needed fortesting such hypotheses.

Components and emerging characteristics

A dynamic system is – by definition – multilayered. This means thatcomponents and emerging characteristics should be defined separatelyfor each level of aggregation: a phenomenon that is an emerging charac-teristic on one level can be a component on a higher level. Although thehierarchies and components discussed by the various authors are notalways comparable, we will elaborate on how different levels of organ-ization can be distinguished and ordered with regard to self and ident-ity; wewill then try to relate the concepts discussed in this book to thesedifferent levels.The lowest level (see table 9.1) discussed in this volume consists of

components such as feelings, bodily sensations, cognitions, orienta-tions, perceptions, appraisals, action tendencies, etc. This is a somewhatarbitrary choice of ‘‘lowest level,’’ because these components, forexample bodily sensations, can also be seen as higher-order phenom-ena that emerge from components of an even lower order. Couplingthese components results at the second level in the emergence of (high-er-order) temporary emotional interpretations, which are probablycomparable – in terms of level of aggregation – with what Fogel callsevents. On the third level these events self-organize over time intoenduring phenomena: recurrent emotional interpretations or EIs(Lewis and Ferrari), frames (Fogel), roles (Haviland-Jones et al.), andvaluations (Hermans and Hermans-Jansen). Self-evaluations and self-representations can also be situated on this level.On the fourth level, self-organizing EIs give rise to implicit and

explicit identity (Lewis and Ferrari). ‘‘The self one is’’ (Frijda), sense ofself, sense of identity (Fogel), and self-theory (Epstein) can also besituated on this fourth level of organization. (For reasons that will beexplained below, self-concept is not included in the table.)Table 9.1 provides an overview of how – from a self-organizational

perspective – different concepts can be related to each other. The table is


Table 9.1. Aggregation levels of self-organization

1. feelings, bodily sensations, cognitions, perceptions, appraisals,orientations, etc.

2. temporary emotional interpretations, events3. Emotional Interpretations, roles, valuations, self-representations,self-evaluations, frames

4. self, self-theory, implicit identity, explicit identity, sense of identity, senseof self

not intended to be the basis of an exhaustive definition of the one andonly systemof self and identity. The definition and the defining compo-nents that are used depend on the concept on which a particular studyfocuses. Hermans and Hermans-Jansen, for example, focus on thenarrative organization of valuations in their case study. Each valuationcan be seen as a system, consisting of events, temporary emotionalinterpretations, emotions, and cognitions to begin with, but they be-come indicative of self and identity because they cohere into ameaning-ful whole. Haviland-Jones and colleagues, on the other hand, describe asystem that consists of lower-order components and emerging patternsof emotions.The various levels are related to different time scales. Lowest-level

components self-organize in events, that is, momentary emotional ex-periences, on a microscopic time scale; experiences subsequentlyorganize in EIs on a medium time scale; and identity, finally, emergeson the highest time scale. The role of emotions and the relation withthe context differs across the different time scales. It is essential, there-fore, to be clear about the time scale one is talking about. From asystems perspective it is also important to realize that the levels ofself-organization presented here are in turn part of a much more exten-sive system of systems. Fogel, for example, sees phenomena like historyand culture as constructs of an even higher order.Conceptualizing self and identity as emerging from a multilayered

system of interacting components has some major implications. First,the question ‘‘Which comes first, emotions or cognitions?’’ becomes ameaningless one. There is a continuous interaction between all compo-nents, EIs themselves being a conglomeration of emotions, perceptions,cognitions, etc. Second, this conceptualization implies that constructslike self-concept, self-esteem, self-evaluations, and all kinds of ideal,real, feared and other selves, are entities that do not pre-exist in aperson’s mind. Stability does not mean that the construct is alwaysthere (in this sense the word construct is inadequate, because it sug-gests something that is always there), but that comparable phenomena


emerge in different situations over time. For example, a similar self-evaluation may emerge during many years in different situations inwhich an individual’s attention to his or her own person is drawn.What situations they are depends on the person, the moment, and thesituation. Since self-related phenomena self-organize at specific mo-ments, they will be comparable, but they will probably also vary sys-tematically in different situations. A self-organizational perspectivedoes not necessarily require the defining of different kinds of selves(‘‘self at school,’’ ‘‘self as I would like to be,’’ etc.) to account for thesesystematic differences. It is therefore much more economical than es-sentialistic theories, which do need such a priori selves. Moreover, sinceemotions and context are fully present in the system at the most basiclevel of self-organization, there is no need to add them as separateconstructs later on.

Sense of identity, implicit and explicit identity

We have so far argued that in dealing with self and identity it isessential to make a distinction between different levels of aggregation.Processes on all levels play an important role but studies of self andidentity usually focus on long-term developmental processes, that is,concepts on the third and fourth levels. In general, the concepts men-tioned on the third level are clearly defined throughout the book.Concepts and definitions on the fourth level, however, cause consider-able confusion. The concepts on this level are all global ones, theyconcern the person as a whole, and emerge from the same underlyinglayers. What they have in common, and what distinguishes them fromother constructs in psychology, for example ‘‘the person’’ and ‘‘person-ality,’’ is that they refer to the person’s own perspective upon andexperience of him- or herself.‘‘Self’’ and ‘‘identity’’ are the most encompassing constructs in table

9.1. As stated in chapter 1, we treat them as interchangeable. We preferto continue with the term ‘‘identity’’ only because of its more circum-scribed meaning. The differentiation made by Lewis and Ferrari be-tween ‘‘implicit’’ and ‘‘explicit’’ identity is especially helpful. Apartfrom ‘‘a sense of identity,’’ all fourth-order constructs in table 9.1 arecovered by these two constructs. Thus, sense of identity, implicit ident-ity and explicit identity will be the three central constructs used in theremainder of this chapter.Sense of identity is a general feeling, a sense of ‘‘sameness and continu-

ity’’ (Erikson, 1968; see also Blasi and Glodis, 1995) that runs throughthe whole system. Because of its coordinating character, it should besituated on the highest level of abstraction (in table 9.1). The emergenceof a sense of identity has nothing to do with the content of any of


the components, but everything to do with the way the differentcomponents fit together over time. Continuity of emotions, of frames,and of emotional interpretations gives rise to a feeling of being thesame over time. Thus, a sense of identity may emerge from all threelower levels of aggregation: the underlying system of components,self-organized events, and EIs.Paradox is inherent in a sense of identity: if people have a sense of

identity, they are (generally) not aware of it (Erikson, 1968). It manifestsitself mainly as a sense of well-being; from a self-organizational per-spective this means having a total absence of any positive feedbackloops in the system. Positive feedback loops fixate, stabilize. In theirabsence, the system can be seen as a completely free-floating continu-ous relationship inwhich all kinds of change are possible, as in the statethat Fogel describes as purely orientational. This conceptualizationunderlines the fact that a sense of identity refers to a fit and is funda-mentally relational. The absence of a sense of identity, on the otherhand, may be felt keenly and consciously, and may give rise toheightened and preoccupied attention to identity issues (Erikson, 1968).Implicit identity is the organization of especially non-conscious EIs a

person has of him- or herself. It resembles Epstein’s concept of ‘‘implicitself-theory,’’ which – togetherwith a ‘‘world theory’’ and notions aboutthe relations between these two – makes up the experiential system.This experiential system is comparable to what Lewis and Ferrari callpersonality. As Epstein points out, there is no sharp boundary betweena person’s self-theory andworld theory. In fact, they are two sides of thesame coin. For example, the notion that ‘‘the world is a dangerousplace’’ implies that the person sees him- or herself as vulnerable.Nevertheless, the chance that mutual couplings will emerge is higherwhere both EIs concern oneself. This will result in the formation of asubsystem, the internal connections of which are much stronger thanthe connections with components that do not belong to the subsystem.Implicit identity can be distinguished, therefore, as a self-organizedsystem within personality.This implies that specific EIs can be more or less categorized in terms

of whether they are part of identity. Most components on the lowerlevels cannot in themselves be categorized as being or not being part ofimplicit identity. A feeling may be part of an EI concerning oneself, butalso of some EI concerning the world. Frijda distinguishes severalaspects of emotions that have to dowith self: emotions that have oneselfas object, the significance of the emotion, implicit estimates about one’sown competence, and the twobasic concerns of self-esteemand sense ofidentity.We assume that at least one such aspectmust be present beforean EI concerning oneself can possibly emerge. However, all elementsthat are part of personality may become part of the implicit identity.


Explicit identity is an organization of conscious EIs about oneself. It isautobiographical and dialogical. It resembles the concept of narrative(Hermans and Hermans-Jansen), and explicit self-theory (Epstein). Ex-plicit identity and related concepts have been clearly described else-where in this book, and need no further explanation here. As comparedwith implicit identity, explicit identity is organized in a more compart-mentalized way, which means that parts may change, while the restremains the same.Other concepts in the field differ from self and identity mainly with

regard to their inclusiveness. Camras and Michel, for example, suggestthat self-concept could be conceived as a component of identity viewedas a dynamic system.We prefer to view self-representations as compo-nents of identity. Self-concept could then be defined as the collection ofdescriptions of ‘‘who one is,’’ ‘‘who one likes to be,’’ ‘‘how one thinksothers see oneself,’’ etc., and therefore as conscious self-representationsthat are part of one’s identity. In this view self-concept is not anemerging higher-order construct: as a collection it is no more than thesum of its components. Consequently, self-evaluations are defined asthose self-representations that include emotions, values, and judg-ments about oneself.Apart from the differences in content described above and elsewhere

in the book, differentiating the concept of identity in implicit identity,explicit identity, and sense of identity is also necessary because, as willbe argued below, the processes involved and the assessment pro-cedures differ according to which construct is entailed.

Processes

Now that identity has been described as a self-organizing system, wecan examine the processes involved. Both implicit and explicit identityemerge from underlying levels. Lewis and Ferrari suggest that themechanisms involved might be comparable. However, as Magai sug-gests, the processes involved in the emergence of implicit identitymightwell differ from those playing a role in conscious phenomena likeexplicit identity. She points out that explicit and implicit identity differin susceptibility to external influences and in probability of change.This would mean that there are different processes involved in the twosystems, as is also argued by Epstein and Pacini (1999).Camras and Michel’s distinction between verbal explanation versus

formal mathematical models is relevant in this controversy. It may bethat the differences are restricted to the verbalmodel. The formal causesand types of mechanism can be the same in systems that differ stronglywith regard to the verbal causes and variables involved.


We will first consider the processes within the two identities on averbal-psychological level. Various theories suggest that these are quitedifferent. Epstein states that the processes in the experiential system arebased on non-conscious, rapid association, and those in the rationalsystem on deduction and reasoning. In a similar way, Dannefer andPerlmutter (1990) distinguish between two different processes thatcoordinate between inner-biological, individual-psychological, cul-tural-sociological, and outer-physical processes; such a distinction canalso be related to the two types of identity. The first process, habitu-ation, is a more-or-less automatic, self-regulatory process that is charac-terized by routine, non-reflective interaction with the environment.This process contributes to the non-consciousmaintenance of a sense ofidentity. It works on the level of emotional interpretations and includesall kinds of biased types of information processing that prevent peoplefrom becoming aware of discrepancies with the desired perception ofthemselves. This is a discrepancy-reducing, equilibrium-seeking pro-cess. The second process, cognitive generativity, is an intentional, re-flective form of coordination: the ability to take oneself as object ofthought, to reflect upon the past and the present, to anticipate thefuture, and to speculate about the unknown. This resembles internaldialogues as described by Hermans and Hermans-Jansen, and explicitidentity or narratives as described by Lewis and Ferrari. Cognitivegenerativity is a conscious attempt to modify and mobilize habitualprocesses that have become maladaptive. The process works on thelevel of the deliberate change of one’s narrative and is based on therules of communication and narratives. One of the consequences of thismay be that explicit identity can change in a more compartmentalizedway (Lewis and Ferrari, unpublished). One of the commitments orvoices in the dialoguemayweaken, disappear, and be replacedwithoutnecessarily turning the whole system upside down.It is not clear how theprocesses in the twosystemsare related.Do they

affect each other? Are they active alternately, or do they run in parallel?Epstein describes how the rational system may ‘‘correct’’ the experien-tial system if reflection and conscious reasoning ‘‘detect’’ maladaptivethoughts or actions. The experiential system, in turn, may affect or biasthe rational system in a non-conscious way. People do not generallyreflect upon themselves most of the time, however. Cognitive generat-ivity might be activated following the inability of the habituation pro-cess todissipate emotions (compare Jansz). Thismaydisrupt one’s senseof identity and result in the emergence of strong emotions. Such strongemotions may trigger self-reflection, as Oosterwegel suggests.On a macroscopic time scale, we see that the developmental courses

of the two types of identity differ strongly. According to Erikson (1968),


early characteristics of implicit identity are present from birth. One’sexplicit identity becomes an imperative fromadolescence onwards. Theperson is then confronted with a combination of biological, social, andcognitive changes that makes his or her implicit identity insufficient forsafeguarding a sense of identity, resulting in feelings of identity con-fusion. People will then try to retrieve their sense of identity, and theydo so by constructing a personal identity: a tentative, reflectively con-strued configuration of self-descriptions that allows the person to findhis or her own place in the world. A new qualitative dimension isthereby added to the system. This new dimension does not differ fromthe sense of identity in terms of level of aggregation, but in terms oflevel of consciousness. Its emergence is the beginning of an ongoingdynamic process of identity maintenance: a more deliberate process ofstabilization and change. Every time that – due to developmentalgrowth, changes in situations or in underlying goals and needs –feelings of identity confusion occur, active and conscious changes inexplicit identity may follow. Such changes in explicit identity probablyhave consequences for implicit identity. Identity development can thusbe seen as an ongoing interaction between implicit and explicit identity,between non-conscious and conscious processes.The case studies discussed byHermans andHermans-Jansen, and by

Verhofstadt-Deneve, and the intervention methods used in these casesprovide some additional suggestions about the howandwhen of devel-opmental switches and of the dynamics involved in such changes. Bothof the clients involved sought help in a period of distress. Initially, thereason for this distress was not part of their conscious, explicit identity.During the therapy, different parts of the non-conscious system werebrought into awareness. Especially in the case of Sharon, the client’stroubles seem to stem from the existence of two different organizations,or attractors, and from a lack of communication between them. The aimof the therapy is to bring the two attractors together, by making theperson aware of their existence, and by stimulating the person tocommunicate between the two perspectives. From a self-organizationalperspective, one can describe both therapies as attempts to lower thedivisions between different ‘‘basins’’ of attraction, by coupling compo-nents of the two attractors in new and different ways, as attempts tobuild a new self-organizing attractor, new EIs. In Verhofstadt-Deneve’spsychodrama, one of these attractors or voices is as it were ‘‘moved’’ toanother person. Distress, in general, may result in a focus on what isgoing on, in an increasing awareness of previously non-conscious EIs.This awareness enables a deliberate change of the couplings and thusthe EIs. The non-conscious system is then changed via conscious inter-vention. This is not to say that all couplings can be changed at will.


Haviland-Jones and colleagues have argued that the construction andchange of one’s own identity may also result from unconscious micro-level processes and have consequences other than those expected.Although the existence of different attractors in the self-system can

be problematic, people in general have several EIs. This raises thequestion of when the existence of different attractors becomes trouble-some. One could imagine that conflicts between EIs arise if they areboth triggered in similar sorts of situations, inwhich they fulfil differentneeds, or if they threaten each other’s needs. For example, an interpre-tation of oneself as being strong and independent may threaten theneed for belonging. Epstein (1998b) has described how different con-flicts and dis-balances between needs can result in distress.From the foregoing, it may be obvious that on a verbal level of

understanding the processes involved in implicit and explicit identityclearly differ. This does not tell us anything, however, about the dif-ferences from a formal, dynamic systems perspective. From this per-spective, self and identity comprise multiple levels of aggregation,withtwo types of self-organizational process: within one level of aggrega-tion (horizontal) and between different levels (vertical: bottom-up andtop-down). These processes take place on and between all levels ofaggregation.Lewis and Ferrari introduce two basic principles of self-organization,

which together explain how lower-order components give rise to stablehigher-order phenomena (bottom-up processes). First, componentscouple, and this coupling results in the emergence of higher-orderforms; and second, subsequent recurrent coupling may lead to thestabilization and strengthening of these forms. Crucial in their ap-proach is the idea that recurrence takes place when experiences cannotbe brought to a completely satisfying completion, which leads to un-resolved intentions and blocked-off actions. Depending on the strengthof the accompanying emotions, such experiences may form the basis ofrecurrent and salient attractors. Mechanisms like thesemake it possibleto explain why some experiences develop into recurrent, salient, char-acteristic patterns and others do not. Another elaboration of bottom-upprocesses can be found in Nowak, Vallacher, Tesser, and Borkowski(2000). They demonstrate, by means of a cellular automaton, howinteracting elements give rise to higher-order phenomena such as theglobal evaluation of the self. The driving force in this model is the presstowards integration, i.e. the striving for coherence between differentcomponents in the self-concept.In contrast to the emergence of higher-order phenomena from the

interaction of lower-order components, other processes work top-down: higher-order phenomenamay affect lower-order phenomena. In


his reply, Van Geert (unpublished) has discussed two possible forms ofthese top-down influences. First, newly emerging phenomena mayaffect streams of energy or information in the lower-order system.One’s narrative may affect the way one appraises new situations. Forexample, the appraisal of computer-related problemswill change if onestarts to see oneself as a computer specialist. People may in a consciousand deliberate way change their behavior and environment and try toadjust their interpretations, feelings, and thoughts to their own narra-tive. Second, a new phenomenon may be of the same order as thecomponents from which it emerges, and become part of the system. Anewly emerging role, for example becoming a mother, will lead to newcouplings and may destroy old ones. In this way, the new role affectsother roles.A process that occurs within one level (horizontal) is the so-called

phase transition: a change in the organization of the higher-orderphenomenon, which can be triggered by small changes in underlying,control variables. Sudden changes may depend on gradual changes incontrol parameters in a very non-linear way, and their occurrenceoften depends on all kinds of system characteristics. This means thata transition never has just one cause, but takes place as a consequenceof the coincidence of many interacting variables: the state of thewhole system and its context. Moreover, major and sudden changesdo not need major external events for their occurrence, but can becaused by gradual, and possibly invisible, processes within the systemitself.In physical systems, self-organization is assumed to be activated by a

striving for the lowest level of energy. Lewis and Ferrari assume thatpsychological systems may be activated (or motivated) by the strivingfor a smooth stream of information or energy – flow in Csikszent-mihalyi’s (1990) terms. In flow, the stream of information or energy isnot hampered, does not get stuck in recurrent patterns, but is freefloating. Several authors in this volume assume the existence of so-called basic needs that activate development. Frustration of these needsmight block the stream of energy and disrupt the sense of identity. On anon-conscious level, this may result in the formation of new EIs. On aconscious level, in their narrative identity people aim at stability, conti-nuity, and coherence. A stable, enduring and coherent narrative en-ables rapid processing and integration of information. A disruption ofone’s narrative means that there is new information that cannot beprocessed and integrated. This calls for reflective forms of coordinationto change the narrative in such a way that the new information can beintegrated.


At this moment there is little to say about the difference betweenimplicit and explicit identity as represented in formal (mathematical)models. Little is known about the formal characteristics of the mechan-isms we described above in verbal terms. In their reaction to Magai,Lewis and Ferrari (unpublished) refer to Allan Schore who, in a per-sonal communication (August, 1999), suggested that implicit and ex-plicit identity may rely on altogether different memory systems, yetboth develop in parallel and interact continually. For one thing, wehave no idea about the developmental trajectories of either system,which leaves us with several questions. Do changes in the two systemsalways have the form of sudden transitions from one attractor toanother, as Lewis and Ferrari describe? Probably, gradual changes arealso possible. Are there comparable process parameters for the twosystems? What does it mean on a formal level if underlying memorysystems differ? Implicit identity aims at equilibrium, but is equilibriumalso important for explicit identity? Our guess is that equilibrium is notespecially important in the case of explicit identity. A narrative shouldbe convincing, but does not need to be consistent. A good narrativeeven requires some degree of inconsistency and dramatic tension(Bruner, 1990). In Lewis and Ferrari’s opinion, retelling a narrative overand over again strengthens it in the sameway as happenswith repeatedcouplings on the implicit level. Are change processes within the twosystems comparable? This is hard to determine. To be able to answerthe question of differences on the formal level, much more needs to beknown about the formal characteristics of the emergence and develop-ment of self and identity.

Systems and their context

Van Geert states that in a dynamic systems approach the context mayrefer to the whole universe, with the exception of the system itself.However, it is far from economical to include the universe in one’smodel. Instead, it is the task of the researcher to define what aspects ofthe context are expected to be relevant for the system under study. Inthis regard, the authors in the present volume are not of one opinion.Fogel’s approach is fundamentally relational: the focus of his study ison the characteristics of relations rather than on those of the person orthe context. Haviland-Jones and colleagues have a different approach.In their view, the context is important in another way. First, it offersspecific situations, such as a prolonged, frightening war-time situation.Second, the influence of context is manifested in the different roles asperceived by the person him- or herself.


The way in which the context is brought into the system has nothingto do with the real characteristics of that context, but rather reflectsthose relations with the context that are considered essential in thisspecific model. An important insight from dynamic systems theory isthat living dynamic systems have no clear and fixed boundaries. Thecontext cannot simply be defined as ‘‘outside the skin’’ and the systemas ‘‘inside the skin.’’ It is at least partly a characteristic of the perspec-tive of the observer: an individual may perceive unwanted ‘‘innervoices’’ as being outside herself, outside her system, as ‘‘not me.’’Thoughts and emotions that are felt as unacceptable by the person mayalso be seen as ‘‘context,’’ or at least as ‘‘not me.’’ A researcher, on theother hand, studying this person’s conflicts,may regard these oppositesas – conflicting – parts of a single system (e.g. Kunnen and KleinWassink, 1999). In a dynamic systems approach one should, therefore,definewhich aspects belong to the system,which contextual factors canbe considered important, and how these contextual factors may affectthe system under study.For the researcher who seeks to specify how the system is affected by

the context, the three distinct conceptualizations of Dannefer andPerlmutter (1990) may be helpful. The influence of the context can bedefined as (1) a random influence, (2) a factor having ordered effects, or(3) a system in itself. The differences between these conceptualizationsbecome most explicit when the researcher wants to translate the rela-tionwith the context into a quantitative dynamic systemsmodel. There-fore we will illustrate each conceptualization by describing such a‘‘translation.’’Context as random influence means that the self-organization of the

system is affected in an unpredictable and non-systematicway from theoutside. There is no systematic relation between subsequent disturban-ces or between the disturbances and the system. In a mathematicaldynamic systemsmodel, this could be expressed by adding or subtract-ing a small random number for each iteration. If context is seen as afactor that has ordered effects, there is a systematic relation between thecontextual influences at different moments, and the context is more orless predictable. We already mentioned the example of living in a warsituation providing a long sequence of potentially frightening experien-ces. Context as an ordered effect can be represented in a model as aparameter. A lower or higher value of this parameter, then, expresseswhether the context is more or less frightening. Such a parameter maydiffer between individuals and situations, and change over time. Ifcontext is seen as a system in itself, it has its own internal dynamics, apartfrom the system under study. For example, the context could have itsown stabilizing forces, which can counteract changes in the ‘‘identity’’


system. In this case, a mathematical model could consist of two interac-ting systems or represent the context in the form of a separate logisticequation that is related to the other equations. Note that the system andthe environment could equally well be seen as two interacting systems,or as one system. This ambiguity illustrates the point that systems haveno intrinsically defined borders and environments.Conceptualizing the context as a system has several advantages.

First, it makes it clear that people are intrinsically related to theirenvironment, while at the same time the environment is seen as beingorganized on its own terms, having its own dynamics and its own self-perpetuating tendencies and transitions. Second, the context can bemeaningfully conceptualized at different levels of generality, in thesameway that there are different levels of identity and self as shown intable 9.1. Society at a macrolevel concerns nations and institutions; onan intermediate level, schools, corporations, families etc.; and on amicrolevel a face-to-face interaction between individuals (compare thediscussion of Bronfenbrenner’s work in chapter 2). The different levelsof generality offer possibilities for a much more detailed and specificfocus on the relation between identity and context. The conception ofcontext as a system in itself represents themost thorough account of therelation between system and context. This characterizes the work ofFogel. He integrates the infant-system and the mother-system into onesystem that consists of the interaction between mother and infant,thereby including the dynamics of both the infant and the mother.However, this mother-plus-infant system also has a context. Whichcomponents are part of the system and which belong to the contexttherefore depends on the researcher’s objectives.A consequence of a self-organizational perspective is that, regardless

of whether the context is conceptualized as random factor, orderedfactor, or system, the effects of the contextual factor on the developmentof the system depend on its state. Random influences, for example,often have no effect at all. In an unstable system, a small randomdisturbance may cause a transition, however. In the same way, anordered factor may have a positive effect, no effect, or a negative effect,again depending on the state of the system. If the context is seen as asystem in itself, the effects of the context not only depend on the targetsystem, but the target system in turn affects the context. For example,whether an individual’s self-esteem is affected by a critical remarkstrongly depends on that individual’s appraisal. This appraisal resultsfrom the interaction between the identity-system and the context. Theindividual reaction, in turn, will affect the behavior of the person whomade the remark, and thereby change the contextual system. Whichmay change the appraisals, and so on.


Assessing self and identity as self-organizing systems

The new and promising approach towards identity presented here willonly be successful if it generates and is supported by empirical re-search. A final implication we therefore want to address is the oper-ationalization, measurement, and analysis of key constructs.The assessment of identity as a self-organizing process poses specific

demands on the way information is measured and analyzed. Twonotions of dynamic systems are especially relevant here. First, selfand identity are not structures that are always present, but ratherphenomena that emerge in a self-organizing system. Second, essentialcharacteristics of self and identity as a self-organizing system manifestthemselves in the way they change over time.A fundamental difference from previous conceptualizations is that

stability of self and identity does not refer to structures that are alwayspresent. Stability means that comparable phenomena (EIs) emerge indifferent situations over time. We therefore have to find ways to‘‘grasp’’ these phenomena at the moment at which they are present.These ways may differ in accordance with the various types and as-pects of self and identity that have been distinguished in this chapter.Implicit identity is the emerging system of self-referent EIs. The

phenomena we want to observe, for example the emotional reactionswhen something concerning the person is at stake, only emerge inspecific situations. This implies that the system should be studiedwhenit is ‘‘at work’’: in situations that matter with regard to self and identity.It cannot be assessed at any time and in any place, for example byadministering a questionnaire. Several authors in this volume havedeveloped different methods to assess the system ‘‘at work.’’ Fogel andVerhofstadt-Deneve observe their subjects directly in situations thatmatter to them. Fogel observes mothers and infants during their dailyinteractions, and Verhofstadt-Deneve creates situations in psycho-drama in which self-referent EIs are expected to emerge. Haviland-Jones et al. andHermans andHermans-Jansen assess theworking of thesystem more indirectly, by asking subjects about personally relevantroles and situations. By analyzing the relation between assessments,Haviland-Jones and colleagues and Hermans and Hermans-Jansen tryto understand the structure of the higher-order system: the self oridentity. Another possibility is to gather personal reports, such asdiaries (Haviland-Jones et al.) and frequent open interviews (Kunnenand Klein Wassink, 1999), which yield relevant and emotional experi-ences in situations when the system was operating. In these methods,the researchers try to reveal non-conscious aspects of EIs by analyzingcommonalities across situations.


Acquiring information about explicit identity is less difficult. Explicitidentity is, by definition, the conscious and reflected narrative or ident-ity. Although this identity need not always be present either, its emerg-ence can be controlled by the person concerned. It can be assessed byasking people for their narratives, commitments, or simply ‘‘who theyare.’’ This is closest to the open type of traditional assessment methodsof self and identity. This openness is a prerequisite here: one’s storiesand commitments are individual characteristics, and differ frompersonto person.A sense of identity has been described as an all-encompassing feeling

of sameness, continuity, and uniqueness. This can be deducted fromobservations or assessed by direct questions concerning these feelings.Such questions have been developed, for example, by Blasi and Glodis(1995).The second relevant notion is that a dynamic system is a system of

related components that changes over time. Amajor difference betweena dynamic systems, self-organizational approach and more traditionalapproaches is that the former approach studies phenomena from adiachronous perspective instead of a synchronous perspective. That is,it focuses on relations over time, instead of relations between variables(in subjects) at the same time. In the preceding chapters identity hasbeen described as a time-related process that can only be studied from adiachronous perspective. Fogel, therefore, stresses that research shouldfocus on sequences, in order to track how a psychological processunfolds over time in the context of coregulated interactions. This im-plies that longitudinal observations are required for the study of selfand identity. In addition, criteria should be formulated for the kinds ofpatterns of stability and change that are relevant, and methods areneeded to determine whether or not these patterns are present.Most methods of analysis in the social sciences are directed towards

assessing differences, rather than towards assessing change. Dynamicsystems theorymay help to identify the phenomena that are relevant inthe search for change. The notion, for example, that a system oftenstabilizes in one of only a limited number of stable attractor states canbe of help in the search for stable states. In a comparable way, dynamicsystems theory describes specific types of behavioral characteristicsthat precede, accompany, and follow transitions in different types ofsystems. For example, an increased vulnerability to disruptions, and alonger return time (that means, a prolonged period of turbulence fol-lowingdisruptions) often precede a transition. These characteristics canbe operationalized in observable behavior. Knowledge of bifurcationpoints – critical points at which development may follow either onepathway or the other – can also be helpful when observing the behavior


of a system over time. Important questions are then: what triggers abifurcation, and in what situations is such a transition more likely tooccur? Because sudden changes may depend on gradual changes inunderlying process parameters, such parameters should be found, andconditions should be defined under which bifurcations are more or lessplausible.Techniques that assess and analyze relevant dynamic systems behav-

ior are still scarce. Lewis, Lamey, and Douglas (1999) show how attrac-tors and state space can be operationalized and demonstrated bysimply registering well-defined behavior of the system over a certainperiod. Fogel focuses on interactions that have to do with mutualhandling of toys and he demonstrates qualitative changes in this inter-action. More traditional methods for analyzing change can also be veryuseful, however. Hermans and Hermans-Jansen compare assessmentsof an individual’s valuations at different times, and Haviland-Jonesand colleagues assess change by hierarchical cluster analysis to com-pare an adolescent’s old roles with his or her new ones. Heymansstresses that some ‘‘classical’’ methods of time-series analysis also offerpossibilities to analyze change in systems. Often, these methods(ARIMA for example) appear to be more common in fields other thanthe social sciences. In addition, we point to a form of factor analysis thatcan be used to assess change in individual time-series, developed byMolenaar (1985).Another way to assess characteristics of a system is to build a quanti-

tative model of the system, and to simulate its behavior. Such a modelreflects the theoretical assumptions concerning the dynamics of thesystem. These assumptions can be tested by comparing the system’sbehavior with empirical phenomena or with other theoretical assump-tions. To build such a model requires defining the components in thesystem (this step is needed in a qualitative approach as well), choosingthe time scale, and turning the process of development into a fewessential principles; moreover, a translation of these components needsto be made into numbers, and of the process principles into equations.Haviland-Jones et al. give a clear example of this approach. Theydemonstrate how a few basic principles can be translated into a dy-namic systems model of the emergence of different emotions in areal-time situation. Differences in personal and situational characteris-tics can be expressed in terms of the values of a few parameters.Relations between these characteristics and emerging outcomes can beused as a basis for the validation of the model. Other examples of theapplication of quantitative dynamic systems models in the field of selfand identity are given in Kunnen, Bosma, and Van Geert (in press), andin Kunnen and Bosma (2000).


Mathematical models can also be used to model the relations be-tween microlevel systems. Emerging behavior on a lower level can beincluded as a parameter or variable in a higher-level system.Microlevelprocesses, then, may result in changes in parameters on a higher levelof aggregation. Repeated frightening experiences – to follow theexample of Haviland-Jones and colleagues – could result in an in-creased sensitivity to such experiences, and an increased chance thatnew situations will be experienced in a comparable way.Most changes in identity will concern a macroscopic time scale, and

can be assessed by comparing assessments over longer time periods.The duration of the therapies in the case studies of Hermans andHermans-Jansen and Verhofstadt-Deneve and the time covered by thediaries in Haviland-Jones et al., and the interview assessments of Kun-nen and Klein Wassink (1999) range from months to years. Fogelobserves mother–infant dyads over a shorter time span but here, too,the changes that take place within ‘‘frames’’ cover a period of weeks oreven months.Studying and assessing self and identity as self-organizing systems

require a thorough thinking and planning of assessment procedures. Atthis moment suitable techniques and methods are not commonly avail-able. However, more and more applications of dynamic systems viewsand techniques are appearing in many different fields. Wemention justa few examples: social psychology (Nowak and Vallacher, 1999), lan-guage development (Van Geert, 1991; Ruhland, 1998), cognitive psy-chology (Van Geert, 1991; Eckstein, 1999; Van der Maas and Molenaar,1992), infant research ( Fogel, Nwokah, and Karns, 1993), motor devel-opment (Thelen and Ulrich, 1991), psychopathology (Msihara, 1996),and personality (Svrakic, Svrakic, and Cloninger, 1996). Since (on ageneral level) the dynamic systems in these fields have much incommon, the publications offer a wealth of inspiring and illustrativeapplications of dynamic systems.The authors in the present volume offer elaborate and specific views

on self and identity as emotional, dynamic systems, rooted in context.Despite clear difference of views, there is sufficient agreement to be ableto formulate general as well as more specific characteristics of this newapproach. The componential approach and the attention to time scalesand aggregation levels provide new opportunities for the analysis ofstability and change in the development of self and identity. As men-tioned in our preface, we see this book as a next step in a discussion andexploration of a new perspective in the field of self and identity. Not allthe questions we posed at the beginning have been answered. Severalstill remain, some have disappeared, others have been reformulated.The question of how self and identity develop in adulthood,mentioned


in chapter 1, has, for instance, not been discussed at all. The conceptual-ization of self and identity as a self-organizing system, as worked out inthis volume, however, does offer a framework, we hope, for a furtherexploration of these questions.

Note

1 Fogel’s question is a direct consequence of his relational and dynamic per-spective. Authors working from an individualistic perspective, on the otherhand, tend to posit stability and uniqueness and have to go to great lengths toconceptualize change and connectedness (e.g. Grotevant and Cooper, 1986;Josselson, 1994).


References

Abraham, F. D. (1990). A visual introduction to dynamical systems theory forpsychology. Santa Cruz: Aerial Press.

Abraham, F. D. (1995). Dynamics, bifurcation, self-organization, chaos, mind,conflict, insensitivity to initial conditions, time, unification, diversity,free will, and social responsibility. In R. Robertson and A. Combs (eds.),Chaos theory in psychology and the life sciences, pp. 155–174. Mahwah, NJ:Erlbaum.

Abram, D. (1997). The spell of the sensuous: perception and language in a more-than-human world.New York: Vintage Books.

Allport, G. W. (1955). Becoming.NewHaven: Yale University Press.Angyal, A. (1965). Neurosis and treatment: a holistic theory.New York: Wiley.Antonovsky, A. (1993). Complexity, conflict, chaos, coherence, coercion, and

civility. Social Science and Medicine, 37, 969–981.Antonovsky, A., and Sagy, S. (1990). Confronting developmental tasks in the

retirement transition. The Gerontologist, 30, 362–368.Aron, A., Aron, E., Tudor, M., and Nelson, G. (1991). Close relationships as

including the other in the self. Journal of Personality and Social Psychology, 60,241–253.

Ashmore, R. D., and Ogilvie, D. M. (1992). He’s such a nice boy . . . when he’swith grandma: gender and evaluation in self-with-other representations.In T. M. Brinthaupt and R. P. Lipka (eds.), The self: definitional and method-ological issues, pp. 236–290. Albany, NY: State University of New YorkPress.

Atkins, P. W. (1984). The second law.New York/Oxford: Freeman.Ausubel, D. (1955). Relationships between shame and guilt in the socializing

process. Psychological Review, 62, 378–390.Baddeley,A. D. (1992).What is autobiographicalmemory? InM.A. Conway,D.

C. Rubin, H. Spinnler, andW.A.Wagenaar (eds.), Theoretical perspectives onautobiographical memory, pp. 13–29. Dordrecht: Kluwer.

Baerveldt, C., and Voestermans, P. (1996). The body as a selfing device: the caseof anorexia nervosa. Theory and Psychology, 6, 693–713.

Bailey, S. T. (1970). Independence and factor structure of self-concept meta-dimensions. Journal of Counseling Psychology, 17, 425–430.

Bakan, D. (1966). The duality of human existence. Chicago: Rand-McNally.Bakhtin, M. M. (1973). Problems of Dostoyevsky’s poetics, 2nd. edn. Trans. R. W.

Rotsel. USA: Ardis. First published in 1929 under the title Problemy tvor-chestva Dostoevskogo [Problems of Dostoyevsky’s Art].

Bakhtin, M. M. (1981). The dialogical imagination: four essays by M. M. Bakhtin.Austin: University of Texas Press.

231

Baltes, P. B., Lindenberger, U., and Staudinger,U. M. (1998). Life-span theory indevelopmental psychology. In R. M. Lerner (vol. ed.; W. Damon, generaled.),Handbook of child psychology, 5th edn., vol. 1, Theoretical models of humandevelopment, pp. 1029–1143. New York: Wiley.

Bambrough, R. (1963). The philosophy of Aristotle. Trans. J. L. Creed and A. E.Wardman. New York: New American Library.

Bandura, A. (1986). Social foundations of thought and action: a social cognitivetheory. Englewood Cliffs: Prentice Hall.

Barclay, C. R. (1993). Remembering ourselves. In G. M. Davies and R. H. Logie(eds.),Memory in everday life, pp. 285–310. Amsterdam: Elsevier.

Barclay, C. R. (1994). Composing protoselves through improvisation. In U.Neisser and R. Fivush (eds.), The remembering self: construction and accuracyin the self-narrative, pp. 55–77. Cambridge: Cambridge University Press.

Barclay, C. R. (1996). Autobiographical remembering: narrative constraints. InD. C. Rubin (ed.), Remembering our past: studies in autobiographical memory,pp. 94–126. Cambridge: Cambridge University Press.

Barrett, K. C. (1993). The development of non-verbal communication andemotion: a functionalist perspective. Journal of Nonverbal Behavior, 17, 145–169.

Barrett, K. C. and Campos, J. J. (1987). Perspectives on emotional developmentII: functionalist approach to emotion. In J. Osofsky (ed.),Handbook of infantdevelopment, 2nd. edn. New York: Wiley.

Bateson, G. (1955). The message: ‘‘This is play.’’ In B. Schaffner (ed.), Groupprocesses, vol. 2. Madison, NJ: Madison Printing Co.

Baumeister, R. F. (1987). How the self became a problem: a psychologicalreview of historical research. Journal of Personality and Social Psychology, 52,163–176.

Baumeister, R. F. (ed.) (1993). Self-esteem: the puzzle of low self-regard.New York:Plenum.

Baumeister, R. F. (1998). The self. In D. T. Gilbert, S. T. Fiske, and G. Lindzey(eds.), The handbook of social psychology, vol. 1, pp. 680–740. New York:McGraw-Hill.

Baumeister, R. F. andNewman, L. S. (1994).How storiesmake sense of personalexperiences: motives that shape autobiographical narratives. Personalityand Social Psychology Bulletin, 20, 676–690.

Baumeister, R. F., Smart, L., and Boden, J. M. (1996). Relation of threatenedegotism to violence and aggression: the dark side of high self-esteem.Psychological Review, 103, 5–33.

Baumeister, R. F., Stillwell, A. M., and Heatherton, T. F. (1994). Guilt: aninterpersonal approach. Psychological Bulletin, 115, 243–267.

Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New York:International Universities Press.

Becker, E. (1973). The denial of death.New York: Free Press.Beltrami, E. (1987).Mathematics for dynamic modeling. Boston: Academic Press.Butz, M. R., Duran, E., and Tong, B. R. (1995). Cross-cultural chaos. In R.

Robertson and A. Combs (eds.), Chaos theory in psychology and the lifesciences, pp. 319–329. Mahwah, NJ: Erlbaum.

Blasi, A., and Glodis, K. (1995). The development of identity: a critical analysisfromtheperspectiveof the self as subject.DevelopmentalReview,15, 409–433.

Blass, E. M., and Shah, A. (1995). Pain reducing properties of sucrose innewborns. Chemical Senses, 20, 29–35.

232 References

Block, J. H., and Block, J. (1980). The role of ego-control and ego-resiliency in theorganization of behavior. InW. Andrew Collins (ed.),Development of cogni-tion, affect, and social relations, Minnesota Symposia on Child Psychology,vol. 12, pp. 39–110. Hillsdale, NJ: Erlbaum.

Block, J., and Robins, R. W. (1993). A longitudinal study of consistency andchange in self-esteem from early adolescence to early adulthood. ChildDevelopment, 64, 909–923.

Boesch,E.E. (1991).Symbolic action theoryandculturalpsychology.Berlin:Springer.Bosma, H. A. (1995). Identity and identity-processes: what are we talking

about? In A. Oosterwegel and R. A. Wicklund (eds.), The self in Europeanand North American culture: development and processes, pp. 5–17. Dordrecht:Kluwer.

Bosma, H. A., Graafsma, T. L. G., Grotevant, H. D., and de Levita, D. J. (eds.)(1994). Identity and development: an interdisciplinary approach. London: Sage.

Bourdieu,P. (1997).Meditations pascaliennes (Pascalianmeditations). Paris: Seuil.Bracken, B. A. (ed.) (1996). Handbook of self-concept: developmental, social, and

clinical considerations.New York: Wiley.Brandtstadter, J. (1998). Action perspectives on human development. In R. M.

Lerner (vol. ed.; W. Damon, general ed.), Handbook of child psychology, 5thedn., vol. 1, Theoretical models of human development, pp. 807–863. NewYork:Wiley.

Brewer, W. F. (1996). What is recollective memory? In D. C. Rubin (ed.),Remembering our past: studies in autobiographical memory, pp. 19–67. Cam-bridge: Cambridge University Press.

Brinthaupt, T. M., and Lipka, R. P. (eds.) (1992). The self: definitional and method-ological issues.Albany, NY: State University of New York Press.

Bronfenbrenner, U., and Crouter, A. C. (1983). The evolution of environmentalmodels in developmental research. In W. Kessen (ed.), Handbook of childpsychology, vol. 1, pp. 357–414. New York: Wiley.

Brown, B., Werner, C. M. and Altman, I. (1998). A dialectical-transactionalperspective on close relationships. In B. M. Montgomery and L. A. Baxter(eds.), Dialectical approaches to studying personal relationships, pp. 137–183.Mahwah, NJ, London: Erlbaum.

Bruner, J. S. (1986). Actual minds, possible worlds. Cambridge, MA: HarvardUniversity Press.

Bruner, J. S. (1990).Acts of meaning.Cambridge,MA: Harvard University Press.Bruner, J. S. (1996). The culture of education.Cambridge,MA: HarvardUniversity

Press.Bruner, J. S., and Kalmar D. A. (1998). Narrative and meta narrative in the

construction of self. In M. D. Ferrari and R. J. Sternberg (eds.), Self-awareness: its nature and development, pp. 308–331.NewYork: GuilfordPress.

Buber, M. (1958). I and thou. Trans. R. G. Smith. New York: Scribners.Buck, R. (1985). Prime theory: an integrated view of motivation and emotion.

Psychological Review, 92, 389–413.Bullock, M., and Lutkenhaus, P. (1990). Who am I: selfunderstanding in tod-

dlers.Merrill-Palmer Quarterly, 36, 217–238.Burghess, D. N. and Wood, A. D. (1985). Mathematical models in the social,

management and life sciences. Chichester: Ellis Horwood.Butterworth, G. (1995). An ecological perspective on the origins of self. In J. L.

Bermudez, A. Marcel, and N. Eilan (eds.), The body and the self, pp. 87–105.Cambridge, MA: Bradford.

233References

Byrne, B. M. (1996). Measuring self-concept across the life span: issues and instru-mentation.Washington, DC: American Psychological Association.

Camras, L. A. (1992). Expressive development and basic emotions. Cognitionand Emotion, 6, 269–283.

Camras, L. A. (2000). Surprise!: facial expressions can be coordinative motorstructures. In M. Lewis and I. Granic (eds.), Emotions, development, andself-organization, pp. 100–124. New York: Cambridge University Press.

Camras, L. A., Lambrecht, L., and Michel, G. F. (1996). Infant ‘‘surprise’’expressions as coordinative motor structures. Journal of Nonverbal Behavior,20, 183–195.

Cantor, N., and Harlow, R. (1994). Social intelligence and personality: flexiblelife-task pursuit. In R. J. Sternberg and P. Ruzgis (eds.), Personality andintelligence, pp. 137–168. New York: Cambridge University Press.

Capps, D. (1989) The deadly sins and saving virtues: how they are viewed bylaity. Pastoral Psychology, 37, 229–253.

Carstensen, L. L., Isaacowitz, D. M., and Charles, S. T. Taking time seriously: atheory of emotional selectivity. American Psychologist, 54, 165–181.

Carver, C. S., and Scheier, M. F. (1990). Origins and functions of positive andnegative affect: a control-process view. Psychological Bulletin, 97, 19–35.

Cassirer, E. (1955). The philosophy of symbolic forms, vol. 2,Mythical thought. NewHaven, CT: Yale University Press.

Caughey, J. L. (1984). Imaginary social worlds: a cultural approach. Lincoln: Uni-versity of Nebraska Press.

Chandler, M. J., and Ball, L. (1989). Continuity and commitment: a develop-mental analysis of identity formation process in suicidal and non-suicidalyouth. In H. A. Bosma and S. Jackson (eds.), Coping and self-concept inadolescence, pp. 149–166. Berlin: Springer.

Chandler, M. J., Lalonde, C., and Sokol, B. (in press). Continuities of selfhood inthe face of radical developmental and cultural change. In L. Nucci, G. Saxe,and E. Turiel (eds.), Culture, thought, and development. Mahwah, NJ:Erlbaum.

Christianson, S., and Safer, M. A. (1996). Emotional events and emotions inautobiographical memories. In D. C. Rubin (ed.), Remembering our past:studies in autobiographical memory, pp. 218–244. Cambridge: CambridgeUniversity Press.

Clausen, J. A. (1995). Gender, contexts, and turning points in adults’ lives. In P.Moen, G. H. Elder, and K. Luscher (eds.), Examining lives in context: perspec-tives on the ecology of human development, pp. 365–389. Washington, DC:American Psychological Association.

Cloninger, C. R., Svrakic, N. M., and Svrakic, D. M. (1997). Role of personalityself-organization in development of mental order and disorder. Develop-ment and Psychopathology, 9, 881–906.

Cole, M. (1996). Cultural psychology: a once and future discipline.Cambridge, MA:The Belknap Press.

Conville, R. (1998). Telling stories: dialectics of relational transition. In B. M.Montgomery and L. A. Baxter (eds.), Dialectical approaches to studyingpersonal relationships, pp. 17–40. Mahwah, NJ: Erlbaum.

Cooley, C. H. (1902). Human nature and the social order.New York: Scribner.Cooper, C. R., Jackson, J. F., Azmitia, M., and Lopez, E. M. (1998). Multiple

selves, multiple worlds: three useful strategies for research with ethnicminority youth on identity, relationships, and opportunity structures. In

234 References

V.C. McLoyd and L. Steinberg (eds.), Studying minority adolescents:conceptual, methodological and theoretical issues, pp. 111–125. Hillsdale, NJ:Erlbaum.

Costa, P. T., Jr., and McCrae, R. R. (1996). Mood and personality in adulthood.In C. Magai and S. McFadden (eds.),Handbook of emotion, adult development,and aging, pp. 369–383. San Diego: Academic Press.

Csikszentmihalyi,M. (1990). Flow: the psychology of optimal experience.NewYork:Harper and Row.

Csikszentmihalyi, M. (1993). The evolving self.New York: Harper Collins.Csikszentmihalyi,M., and Rathunde, K. (1998). The development of the person:

an experimental perspective on the ontogenesis of psychological complex-ity. In R. M. Lerner (vol. ed.; W. Damon, general ed.), Handbook of childpsychology, 5th edn., vol. 1, Theoretical models of human development, pp. 635–684. New York: Wiley.

Damasio, A. R. (1994). Descartes’ error.New York: Avon.Damon, W., and Hart, D. (1982). The development of self-understanding from

infancy through adolescence. Child Development, 4, 841–864.Damon, W., and Hart, D. (1988). Self-understanding in childhood and adolescence.

Cambridge: Cambridge University Press.Dannefer, D., and Perlmutter, M. (1990). Development as a multidimensional

process: individual and social constraints.Human Develoment, 33, 108–137.Davitz, J. R. (1969). The language of emotion.New York, Academic Press.Day, J. M., and Tappan, M. B. (1996). The narrative approach to moral develop-

ment: from the epistemic subject to dialogical selves. Human Development,39, 67–82.

De Boeck, P., and Rosenberg, S. (1988). Hierarchical classes: model and dataanalysis. Psychometrika, 53, 361–381.

De Quincey, C. (1998). Engaging presence. Noetic Sciences Review, 45, 24–27.De Rivera, J. (1992). Emotional climate. Social structure and emotional dynam-

ics. In K. T. Strongman (ed.), International review of studies on emotion, vol. 2.New York: Wiley.

De Waal, F. B. M. (1982). Chimpanzee politics: power and sex among apes. London:Jonathan Cape.

De Waal, F. B. M. (1996). Good natured. Cambridge, MA: Harvard UniversityPress.

DeWeerth, C., and Van Geert, P. L. C. (2000). The dynamics of emotion-relatedbehaviors in infancy. In M. D. Lewis and I. Granic (eds.), Emotion, develop-ment, and self-organization, pp. 324–348. Cambridge: Cambridge UniversityPress.

Deiner, E., and Larsen, R. J. (1984). Temporal stability and cross-situationalconsistency of affective, behavioral, and cognitive responses. Journal ofPersonality and Social Psychology, 47, 871–883.

Demo, D. H. (1992). The self-concept over time: research issues and directions.Annual Review of Sociology, 18, 303–326.

Demo,D.H., and Savin-Williams,R. C. (1992). Self-concept stability and changeduring adolescence. In R. P. Lipka and T. M. Brinthaupt (eds.), Self-perspec-tives across the life span, pp. 116–148. Albany, NY: State University of NewYork Press.

Demos, E. (1982). Facial expressions of infants and toddlers: a descriptiveanalysis. In T. Field and A. Fogel (eds.), Emotion and early interaction,pp. 127–160. Hillsdale, NJ: Erlbaum.

235References

Dennett, D. C. (1978). Brainstorms: philosophical essays on mind and psychology.Montgomery: Bradford Books.

Derryberry,D., andRothbart,M.K. (1997).Reactiveandeffortfulprocesses in theorganization of temperament.Development and Psychopathology, 9, 633–652.

Dewey, J. (1934). Art as experience. New York: Perigee.Dewey, J. (1963). Experience and education. New York: Macmillan. (Originally

published in 1938).Dunne, J. (1996). Beyond sovereignty and deconstruction: the storied self. In P.

Ricoeur, The hermeneutics of action, pp. 137–157. London: Sage.Duval, S., and Wicklund, R. A. (1972). A theory of objective self-awareness. New

York: Academic Press.Eckstein, S. G. (1999). A dynamic model of cognitive growth in a population:

spatial tasks and conservation. Journal ofMathematical Psychology, 43, 34–70.Eisenberg,N. (1998). Introduction. InN. Eisenberg (vol. ed.;W. Damon, general

ed.), Handbook of child psychology, 5th edn., vol. 3, Social, emotional, andpersonality development, pp. 1–24. New York: Wiley.

Ekman, P. (1972). Universals and cultural differences in facial expressions ofemotions. Nebraska symposium on motivation, vol. 19, pp. 207–283. Lincoln:University of Nebraska Press.

Ekman, P. (ed). (1982). Emotion in the human face, 2nd edn. New York: Cam-bridge University Press.

Elkind, D. (1974). Children and adolescents: interpretive essays on Jean Piaget. NewYork: Oxford.

Ellis, A. (1972) Psychotherapy without tears. In A. Burton (ed.), Twelve thera-pists, pp. 103–126. San Francisco: Jossey-Bass Inc.

Emde, R. N., and Oppenheim, D. (1995). Shame, guilt, and the Oedipal drama:developmental considerations concerning morality and referencing ofcritical others. In J. P. Tangney and K. W. Fischer (eds.), Self-consciousemotions: the psychology of shame, guilt, embarrassment, and pride, pp. 413–438.New York: Guilford Press.

Epstein, S. (1973). The self-concept revisited or a theory of a theory. AmericanPsychologist, 28, 404–416.

Epstein, S. (1983). The unconscious, the preconscious, and the self-concept. In J.Suls and A. Greenwald (eds.), Psychological perspectives on the self, vol. 2,pp. 219–247. Hillsdale, NJ: Erlbaum.

Epstein, S. (1984). Controversial issues in emotion theory. In P. Shaver (ed.),Annual review of research in personality and social psychology, pp. 64–87. Bev-erly Hills, CA: Sage.

Epstein, S. (1985). The implications of cognitive-experiential self-theory forresearch in social psychology and personality. Journal for the Theory of SocialBehavior, 15, 283–310.

Epstein, S. (1989). Values from the perspective of cognitive-experiential self-theory. In N. Eisenberg, J. Reykowski, and E. Staub (eds.), Social and moralvalues, pp. 3–22. Hillsdale, NJ: Erlbaum.

Epstein, S. (1990). Cognitive-experiential self-theory. In L. Pervin (ed.), Hand-book of personality theory and research: theory and research, pp. 165–192. NewYork: Guilford Press.

Epstein, S. (1991a). Cognitive-experiential self-theory: implications for develop-mental psychology. In M. Gunnar and L. A. Sroufe (eds.), Self-processes anddevelopment,Minnesota Symposia on Child Psychology, vol. 23, pp. 79–123.Hillsdale, NJ: Erlbaum.

236 References

Epstein, S. (1991b). The self-concept, the traumatic neurosis, and the structureof personality. In D. Ozer, J. M. Healy, Jr., and A. J. Stewart (eds.), Perspec-tives in personality, vol. 3A, pp. 63–98. London: Jessica Kingsley Publishers,Ltd.

Epstein, S. (1992a). Coping ability, negative self-evaluation, and overgeneraliz-ation: experiment and theory. Journal of Personality and Social Psychology, 62,826–836.

Epstein, S. (1992b). The cognitive self, the psychoanalytic self, and the forgottenselves. Comment on DrewWesten’s The cognitive self and the psychoanalyticself: can we put our selves together? Psychological Inquiry, 3, 34–37.

Epstein, S. (1993a). Implications of cognitive-experiential self-theory forpersonality and developmental psychology. In D. Funder. R. Parke, C.Tomlinson-Keasey, and K. Widaman (eds.), Studying lives through time:personality and development, pp. 399–438. Washington, DC: American Psy-chological Association.

Epstein, S. (1993b). Bereavement from the perspective of cognitive-experientialself-theory. In M. S. Stroebe, W. Stroebe, and R. O. Hansson (eds.), Hand-book of bereavement: theory, research, and intervention, pp. 112–125. NewYork:Cambridge University Press.

Epstein, S. (1994). Integration of the cognitive and the psychodynamic uncon-scious. American Psychologist, 49, 709–724.

Epstein, S. (1998a). Cognitive-experiential self-theory: a dual process personal-ity theory with implications for diagnosis and psychotherapy. In R. F.Bornstein and J. M. Masling (eds.), Empirical perspectives on the psycho-analytic unconscious, vol. 7, pp. 99–140.Washington, DC: American Psycho-logical Association.

Epstein, S. (1998b). Emotions and psychopathology from the perspective ofcognitive-experiential self-theory. In W. F. Flack, Jr., and J. D. Laird (eds.),Emotions in psychopathology: theory and research, pp. 57–69. New York: Ox-ford University Press.

Epstein, S. (1998c). Cognitive-experiential self-theory. In D. Barone, M. Hersen,and V. B. Van Hasselt (eds.), Advanced personality, pp. 211–238. New York:Plenum.

Epstein, S., and Pacini, R. (1999). Some basic issues regarding dual-processtheories from the perspective of cognitive-experiential self-theory. In S.Chaiken, and Y. Trope (eds.), Dual-process theories in social psychology,pp. 462–482. New York: Guilford Press.

Erikson, E. H. (1950). Childhood and society.New York: Norton.Erikson, E. H. (1968). Identity: youth, and crisis.New York: Norton.Feiring, C., and Taska, L. S. (1996). Family self-concept: ideas on its meaning. In

B. A. Bracken (ed.),Handbook of self-concept: developmental, social, and clinicalconsiderations, pp. 317–373. New York: Wiley.

Ferrari, M. (1998). Being and becoming self-aware. In M. Ferrari and R. J.Sternberg (eds.), Self-awareness: its nature and development, pp. 387–422.New York: Guilford Press.

Ferrari, M., and Mahalingam, R. (1998). Personal cognitive development andits implications for teaching and learning. Educational Psychologist, 33,35–44.

Feyereisen, P. (1989). Theories of emotions and neuropsychological research. InF. Boller and J. Grafman (ed.),Handbook of neuropsychology, vol. 3, section 6,pp. 271–281. Amsterdam: Elsevier.

237References

Fischer, K. W. (1980). A theory of cognitive development: the control andconstruction of hierarchies of skills. Psychological Review, 87, 477–531.

Fischer, A. H., and Jansz, J. (1995). Reconciling emotions with Western person-hood. Journal for the Theory of Social Behaviour, 25, 59–81.

Fischer, K. W., Shaver, P. R., and Carnochan, P. (1990). How emotions developand how they organise development. Cognition and Emotion, 4, 81–127.

Fischer, K. W., and Tangney, J. P. (1995). Self-conscious emotions and the affectrevolution: framework and overview. In J. P. Tangney and K. W. Fischer(eds.), Self-conscious emotions: the psychology of shame, guilt, embarrassment,and pride, pp. 3–22. New York: Guilford Press.

Fogel, A. (1990). The process of developmental change in infant communicativeaction: using dynamic systems theory to study individual ontogenies. In J.Colombo and J. W. Fagen (eds.), Individual differences in infancy: reliability,stability, prediction, pp. 341–358. Hillsdale, NJ: Erlbaum.

Fogel, A. (1993). Developing through relationships: origins of communication, self,and culture.New York: Harvester Wheatsheaf.

Fogel, A. (1995). Relational narratives of the pre-linguistic self. In P. Rochat(ed.), The self in early infancy: theory and research, pp. 117–139. Amsterdam:Elsevier North Holland.

Fogel, A. (2001). Infancy: Infant, family, and society, 4th edn. Belmont, CA:Wadsworth.

Fogel, A. (unpublished). Reply to Jansz. Salt Lake City: University of Utah,Department of Psychology.

Fogel, A., and Branco, A. U. (1997). Meta-communication as a source of indeter-minism in relationship development. In A. Fogel, M. Lyra, and J. Valsiner(eds.), Dynamics and indeterminism in developmental and social processes,pp. 65–92. Hillsdale, NJ: Erlbaum.

Fogel, A., Dickson, K. L., Hsu, H., Messinger, D., Nelson-Goens, G. C., andNwokah, E. (1997). Communicative dynamics of emotion. In K. C. Barrett(ed.), New directions in child development: the communication of emotion, Cur-rent research from diverse perspectives, vol. 77, pp. 5–24. San Francisco:Jossey-Bass.

Fogel, A., and Lyra, M. (1997). Dynamics of development in relationships. In F.Masterpasqua and P. Perna (eds.), The psychological meaning of chaos: self-organization in human development and psychotherapy, pp. 75–94. Washing-ton, DC: American Psychological Association.

Fogel,A.,Nwokah,E.,Dedo, J.Y., andMessinger,D. (1992).Socialprocess theoryof emotion: a dynamic systems approach. Social Development, 1, 122–142.

Fogel, A., Nwokah, E., and Karns, J. (1993). Parent–infant games as dynamicsocial systems. In K. B. MacDonald (ed.), Parent–children play, pp. 43–70.Albany, NY: SUNY Press.

Fogel, A., and Thelen. E. (1987). Development of early expressive and com-municative action: reinterpreting the evidence from a dynamic systemsperspective.Developmental Psychology, 23, 747–761.

Freeman,W. J. (1995). Societies of brains: a study in the neuroscience of love and hate.Hillsdale, NJ: Erlbaum.

Freeman,W. J. (2000). Emotion is essential to all intentional behaviors. In M. D.Lewis and I. Granic (eds.), Emotion, development, and self-organization,pp. 209–235. New York: Cambridge University Press.

Freeman, W. J., and Baird, B. (1987). Relation of olfactory EEG to behavior:spatial analysis. Behavioral Neuroscience, 101, 393–408.

238 References

Freeman,M. and Brockmeier, J. (in press). Narrative integrity: autobiographicalidentity and the meaning of the ‘‘good life.’’ In J. Brockmeier and D.Carbaugh (eds.), Narrative and identity: studies in autobiography, self, andculture.

Fridlund, A. J. (1994). Human facial expression: an evolutionary view. New York:Academic Press.

Frijda, N. H. (1986). The emotions. Cambridge: Cambridge University Press.Frijda, N. H. (1993a). Moods, emotion episodes, and emotions. In M. Lewis and

J. M. Haviland (eds.), Handbook of emotions, pp. 381–403. New York:Guilford Press.

Frijda, N. H. (1993b). The place of appraisal in emotion. Cognition and Emotion,7, 357–387.

Frijda,N. H. (1994). The Lex Talionis: on vengeance. In S. H.M. VanGoozen, N.E. Van de Poll, and J. A. Sergeant (eds.), Emotions: essays on emotion theory,pp. 263–290. Hillsdale, NJ: Erlbaum.

Frijda, N. H. (2000). The nature of pleasure. In J. Bargh (ed.), Unravelling thecomplexities of social life: a festschrift in honor of R. B. Zajonc, pp. 71–94.Washington, DC: American Psychological Association.

Frijda, N. H. (unpublished). Reply to Epstein. Amsterdam: University of Am-sterdam, Department of Psychology.

Frijda, N. H., Kuipers, P., and Terschure, E. L. (1989). Relations betweenemotion, appraisal, and emotional action readiness. Journal of Personalityand Social Psychology, 57, 212–228.

Frijda, N. H., Mesquita, B., Sonnemans, J., and Van Goozen, J. (1991). Theduration of affective phenomena or emotions, sentiments, and passions. InK. T. Strongman (ed.), International review of studies on emotion, vol. 1,pp. 187–225. Chichester: Wiley.

Frijda, N. H., and Tcherkassof, A. (1997). Facial expression and modes ofaction readiness. In Russell, J. A. and Fernandez-Dols, J. M. (eds.), Thepsychology of facial expression, pp. 78–102. Cambridge: Cambridge Univer-sity Press.

Ganguly, S. N. (1976). Communication, identity, and human development.Communication, 2, 221–244.

Gecas, V., and Mortimer, J. T. (1987). Stability and change in the self-conceptfrom adolescence to adulthood. In T. Honess and K. Yardley (eds.), Self andidentity: perspectives across the lifespan, pp. 265–286. London: Routledge andKegan Paul.

Gergen, K. (1984). Theory of the self: impasse and evolution. In L. Berkowitz(ed.), Advances in experimental social psychology, vol. 17, pp. 49–115. NewYork: Academic Press.

Gergen, M. M. (1991). The emergence of the relational self. New York: BasicBooks.

Gergen, M. M., and Gergen, K. J. (1995). What is this thing called love? Emo-tional scenarios in historical perspective. Journal of Narrative and Life His-tory, 5, 221–237.

Gibson, J. J. (1966). The senses considered as perceptual systems. Boston, MA:HoughtonMifflin Company.

Gibson, J. J. (1979). The ecological approach to visual perception systems. Boston,MA: Houghton Mifflin Company.

Ginsburg, G. P. (1985). The analysis of human action: current status and futurepotential. In G. P. Ginsburg, M. L. Brenner and M. von Cranach (eds.),

239References

Discovery strategies in the psychology of action, pp. 255–279. London: Aca-demic Press.

Goertzel, B. (1995). Belief systems as attractors. In R. Robertson and A. Combs(eds.), Chaos theory in psychology and the life sciences, pp. 123–134. Mahwah,NJ: Erlbaum.

Goffman, E. (1974). Frame analysis: an essay on the organization of experience.Cambridge, MA: Harvard University Press.

Goldsmith, H. H. (1994). Parsing the emotional domain from a developmentalperspective. In P. Ekman and R. J. Davidson (eds.), The nature of emotion,pp. 68–73. New York: Oxford University Press.

Goleman, D. (1995). Emotional intelligence.New York: Bantam Books.Goodnow, J. J. (1996). From household practices to parents’ ideas about work

and interpersonal relationships. In S. Harkness and C. Super (eds.), Par-ents’ cultural belief systems: their origins, expressions, and consequences,pp. 313–344. New York: Guilford Press.

Goodwin, B. C. (1987). Developing organisms as self-organizing fields. In F. E.Yates (ed.), Self-organizing systems: the emergence of order, pp. 167–180. NewYork: Plenum.

Gordon, C., and Gergen, K. J. (eds.) (1968). The self in social interaction. NewYork: Wiley.

Greenberg, J., Solomon, S., Pyszczynski, T., Rosenblatt, A., Burling, J., Lyon, D.,Simon, L., and Pinel, E. (1992). Why do people need self-esteem? Converg-ing evidence that self-esteem serves an anxiety-buffering function. Journalof Personality and Social Psychology, 6, 913–922.

Greenwald, A. G., and Pratkanis, A. R. (1984). The self. In R. S. Wyer and T. K.Scrull (eds.), Handbook of social cognition, vol. 3, pp. 129–178. Hillsdale, NJ:Erlbaum.

Grotevant, H. D., and Cooper, C. R. (1986). Individuation in family relation-ships. Human Development, 29, 82–100.

Harkness, K. L., and Tucker, D. M. (2000). Motivation of neural plasticity:neural mechanisms in the self-organization of depression. In M. D. Lewisand I. Granic (eds.), Emotion, development, and self-organization, pp. 186–208.New York: Cambridge University Press.

Harre, R. (1983). Personal being: a theory for individual psychology. Oxford: Black-well.

Harre, R. (1987). The social construction of selves. In K. Yardley and T. Honess(eds.), Self and identity: psychosocial perspectives, pp. 41–52. New York:Wiley.

Harre, R. (1998). The singular self: an introduction to the psychology of personhood.London: Sage.

Harre, R., and Van Langenhove, L. (1991). Varieties of positioning. Journal forthe Theory of Social Behaviour, 21, 393–407.

Hart, D., and Edelstein, W. (1992). Self-understanding development in cross-cultural perspective. In T. M. Brinthaupt, and R. P. Lipka (eds.), The self:definitional and methodological issues, pp. 291–322. Albany, NY: State Univer-sity of New York Press.

Harter, S. (1983). Developmental perspectives on the self-system. In P. H.Mussen (ed.), Handbook of child psychology, vol. 4, pp. 275–385. New York:Wiley.

Harter, S. (1985). Manual for the self-perception profile for children. Denver, CO:University of Denver.

240 References

Harter, S. (1996). Historical roots of contemporary issues involving self-concept. In B. A. Bracken (ed.),Handbook of self-concept: developmental, social,and clinical considerations, pp. 38–90. New York: Wiley.

Harter, S. (1998). The development of self-representations. In N. Eisenberg (vol.ed.; W. Damon, general ed.). Handbook of child psychology, 5th edn.,vol. 3, Social, emotional, and personality development, pp. 553–617. New York:Wiley.

Harter, S. (1999). The construction of the self: a developmental perspective. NewYork: Guilford Press.

Harter, S., and Marold, D. B. (1989). A model of the determinants and medi-ational role of self-worth: implications for adolescent depression and sui-cidal ideation. In J. Strauss and G. R. Goethals (eds.), The self: interdisciplin-ary approaches, pp. 66–92. New York: Springer.

Hattie, J. (1992). Self-concept. Hillsdale, NJ: Erlbaum.Hattie, J., andMarsh, H.W. (1996). Future directions in self-concept research. In

B. A. Bracken (ed.),Handbook of self-concept: developmental, social, and clinicalconsiderations, pp. 421–462. New York: Wiley.

Haviland, J. M. (1976). Looking Smart. In M. Lewis (ed.), Origins of Intelligence,pp. 353–378. New York: Plenum Press.

Haviland, J. M., Davidson, R. B., Ruetsch, C., Gebelt, J. L., and Lancelot, C.(1994). The place of emotion in identity. Journal of research in adolescence, 4,503–518.

Haviland, J. M., and Kahlbaugh, P. (1993). Emotion and identity. In M. Lewisand J. M. Haviland (eds.), Handbook of emotions, pp. 327–339. New York:Guilford Press.

Haviland, J. M., and Kramer, D. A. (1991). Affect-cognition relations in anadolescent diary: the case of Anne Frank.Human Development, 34, 143–159.

Haviland, J. M., and Lelwica, M. (1987). The induced affect response: 10-week-old infants’ responses to three emotion expressions.Developmental Psychol-ogy, 23, 97–104.

Haviland-Jones, J. M., Boulifard, D., and Magai, C. (unpublished manuscript).New models of emotion and identity development: towards dynamicsystems.

Held, R. (1965). Plasticity in sensory-motor systems. Scientific American, 213,84–94.

Hermans, H. J. M. (1987a). Self as an organized system of valuations: toward adialogue with the person. Journal of Counseling Psychology, 34, 10–19.

Hermans, H. J. M. (1987b). The dream in the process of valuation: a method ofinterpretation. Journal of Personality and Social Psychology, 53, 163–175.

Hermans, H. J. M. (1996a). Voicing the self: from information processing todialogical interchange. Psychological Bulletin, 119, 31–50.

Hermans, H. J. M. (1996b). Opposites in a dialogical self: constructs as charac-ters. The Journal of Constructivist Psychology, 9, 1–26.

Hermans, H. J. M., and Hermans-Jansen, E. (1995). Self-narratives: the construc-tion of meaning in psychotherapy.New York: Guilford Press.

Hermans, H. J. M., and Kempen, H. J. G. (1993). The dialogical self: meaning asmovement. San Diego, CA: Academic Press.

Hermans, H. J. M., Kempen, H. J. G., and Van Loon, R. J. P. (1992). Thedialogical self: beyond individualism and rationalism. American Psychol-ogist, 47, 23–33.

Heymans, P. (1992). Lifespan-learning: developmental tasks and their manage-

241References

ment. In A. Tuynman and M. Van der Kamp (eds.), Learning across thelifespan: theories, research and policies, pp. 33–51. Oxford: Pergamon Press.

Heymans, P. (1994). Developmental tasks: a cultural analysis of human devel-opment. In J. ter Laak, A. Podolski, and P. Heymans (eds.), Developmentaltasks: towards a cultural analysis of human development, pp. 3–33. Dordrecht:Kluwer.

Heymans,P. (2000). Talent fordevelopment: responding to contextualpromises.In C. F. M. Van Lieshout and P. G. Heymans (eds.), Developing talent acrossthe lifespan. Hove: Psychology Press.

Higgins, E. T. (1987). Self-discepancy: a theory relating self and affect. Psycho-logical Review, 94, 319–340.

Higgins, E. T. (1996). The ‘‘self-digest’’: self-knowledge serving self-regulatoryfunctions. Journal of Personality and Social Psychology, 71, 1062–1083.

Higgins, E. T., Vookles, J., and Tykocinski, O. (1992). Self and health: how‘‘patterns’’ of self-beliefs predict types of emotional and physical prob-lems. Social Cognition, 10, 125–150.

Hochschild, A. (1983). The managed heart. Berkeley: University of CaliforniaPress.

Hofbauer, J., and Sigmund, K. (1988). The theory of evolution and dynamicalsystems. Cambridge: Cambridge University Press.

Holland, J. H. (1998). Emergence: from chaos to order. Reading, MA: Helix Books/Addison-Wesley.

Hoorens, V., and Nuttin, J. M. (1993). Overvaluation of own attributes: mereownership or subjective frequency? Social Cognition, 11, 177–200.

Hopkins, B., and Butterworth, G. (1997). Dynamical systems approaches to thedevelopment of action. In G. Bremner, A. Slater, and G. Butterworth (eds.),Infant development: recent advances, pp. 75–100. Hove: Psychology Press.

Horowitz, M. J. (1992). Stress response syndromes, 3rd edn. Northvale, NJ: JasonAronson.

Horowitz, M. J. (1998). Cognitive psychodynamics: from conflict to character. NewYork: Wiley.

Huntley, C. W. (1940). Judgments of self based upon records of expressivebehavior. Journal of Abnormal Social Psychology, 35, 398–427.

Isen, A. M. (1990). The influence of positive and negative affect on cognitiveorganization: some implications for development. In N. Stein, B. Leven-thal, and T. Trabasso (eds.), Psychological and biological approaches to emotion,pp. 75–94. Hillsdale, NJ: Erlbaum.

Izard, C. E. (1971). The face of emotion.New York: Appleton-Century-Crofts.Izard, C. E. (1984). Emotion-cognition relationships and human development.

In C. E. Izard, J. Kagan, and R. B. Zajonc (eds.), Emotions, cognition andbehavior, pp. 17–37. Cambridge: Cambridge University Press.

Izard, C. E. (1992). Basic emotions, relations among emotions, and emotion-cognition relations. Psychological Review, 99, 561–565.

Izard, C. E. (1993). Four systems of emotion activation.Psychological Review, 100,68–90.

Izard, C. E., Ackerman, B., Schoff, K., and Fine, S. (2000). Self-organization ofdiscrete emotions, emotion patterns, and emotion-cognition relations. InM. D. Lewis and I. Granic (eds.), Emotion, development, and self-organization,pp. 15–36. New York: Cambridge University Press.

Jackson, E. A. (1991a). Perspectives of nonlinear dynamics, vol. 1. Cambridge:Cambridge University Press.

242 References

Jackson, E. A. (1991b). Perspectives of nonlinear dynamics, vol. 2. Cambridge:Cambridge University Press.

James, W. (1976). Essays in radical empiricism. Cambridge, MA, and London:Harvard University Press.

James, W. (1890). The principles of psychology.New York: Holt.James, W. (1910). Psychology: the briefer course.New York: Holt.Jansz, J. (1995). Self-narratives as personal structures of meaning. In A. Ooster-

wegel and R. A. Wicklund (eds.), The self in European and North Americanculture: development and processes, pp. 65–72. Dordrecht: Kluwer.

Jansz, J., and Timmers, M. (in press). Emotional dissonance: when emotionsjeopardize identity. Theory and Psychology.

Jaquish, G. A., and Savin-Williams, R. C. (1981). Biological and ecologicalfactors in the expression of adolescent self-esteem. Journal of Youth andAdolescence, 10, 473–485.

Jaynes, J. (1976). The origin of consciousness in the breakdown of the bicameral mind.Boston: HoughtonMifflin.

Johnson, F. (1985). TheWestern concept of self. In A. J. Marsella, G. de Vos, andF. L. K. Hsu (eds.), Culture and self: Asian and Western perspectives, pp. 91–138. New York: Tavistock Publications.

Johnson, M. (1987). The body in the mind: the bodily basis of meaning, imagination,and reason. Chicago: University of Chicago Press.

Jones, D. (1990). The matrix of narrative: family systems and the semiotics of story.New York: Mouton de Gruyter.

Jopling, D. (1993). Cognitive science, other minds and the philosophy of dia-logue. In U. Neisser (ed.), The perceived self, pp. 290–309. Cambridge: Cam-bridge University Press.

Josephs, I. E. (1998). Constructing one’s self in the city of the silent: dialogue,symbols, and the role of ‘‘as-if’’ in self-development. Human Development,41, 180–195.

Josselson, R. (1994). Identity and relatedness in the life cycle. In H. A. Bosma, T.L. G. Graafsma, H. D. Grotevant, and D. J. de Levita (eds.), Identity anddevelopment: an interdisciplinary approach, pp. 81–102. Thousand Oaks, CA:Sage.

Josselson, R. (1995). Imagining the real: empathy, narrative, and the dialogicself. The Narrative Study of Lives, 3, 27–44.

Juhasz, A. (1992). Significant others in self-esteem development: methods andproblems in measurement. In T. M. Brinthaupt and R. P. Lipka (eds.), Theself: definitional and methodological issues, pp. 204–235. Albany, NY: StateUniversity of New York Press.

Kagan, J. (1981). The second year: the emergence of self-awareness.Cambridge, MA:Harvard University Press.

Kahlbaugh, P. E. and Haviland, J. M. (1994) Nonverbal communication be-tween parents and adolescents: a study of approach and avoidance behav-iors. Journal of Nonverbal Behavior, 18, 91–113.

Kaplan, H. B. (1986). Social psychology of self-referent behavior. New York: PlenumPress.

Kauffman, S. A. (1993). The origins of order: self-organization and selection inevolution.NewYork: Oxford University Press.

Kaye, K. (1982).The mental and social life of babies.Chicago:University of ChicagoPress.

Kegan, R. (1982). The evolving self: problem and process in human development.

243References

Cambridge, MA: Harvard University Press.Keith, L. K., and Bracken, B. A. (1996). Self-concept instrumentation: a historical

and evaluative review. In B. A. Bracken (ed.), Handbook of self-concept:developmental, social, and clinical considerations, pp. 91–170.NewYork:Wiley.

Kelly, G. A. (1955/1963). A theory of personality: the psychology of personal con-structs.New York: Norton.

Kelso, J. A. S. (1984). Phase transitions and critical behavior in human bimanualcoordination. American Journal of Physiology, 15, A1000–A1004.

Kendon, A. (1985). Behavioral foundations for the process of frame attunementin face-to-face interaction. In G. P. Ginsburg, M. Brenner, and M. vonCranach (eds.), Discovery strategies in the psychology of action, pp. 229–253.Orlando: Academic Press.

Kernis, M. H. (1993). The roles of stability and level of self-esteem in psycho-logical functioning. In R. F. Baumeister (ed.), Self-esteem: the puzzle of lowself-regard, pp. 167–182. New York: Plenum Press.

Kernis, M. H., Grannemann, B. D., and Barclay, L. C. (1989). Stability and levelof self-esteem as predictors of anger arousal and hostility. Journal of Person-ality and Social Psychology, 56, 1013–1022.

Kernis, M. H., and Waschull, S. B. (1995). The interactive roles of stability andlevel of self-esteem: research and theory. In M. P. Zanna (ed.), Advances inexperimental social psychology, vol. 27, pp. 93–141. London: Academic Press.

Kindermann, T. A., and Skinner, E. A. (1992). Modeling environmental devel-opment: Individual and contextual trajectories. In J. B. Asendorpf and J.Valsiner (eds.), Stability and change in development: a study of methodologicalreasoning, pp. 155–197. Newbury Park, CA: Sage.

Kingsland, S. E. (1995). Modeling nature: episodes in the history of populationecology. Chicago: University of Chicago Press.

Kitayama, S., Markus, H. R., and Matsumoto, H. (1995). Culture, self, andemotion: a cultural perspective on ‘‘self-conscious’’ emotions. In J. P.Tangney and K.W. Fischer (eds.), Self-conscious emotions, pp. 439–464. NewYork: Guilford Press.

Klages, L. (1948). Charakterkunde (Characterology). Zurich: Hirzel.Kohlberg,L. (1976).Moralstagesandmoralization: thecognitive-developmental

approach. In T. Lickona (ed.),Moral development and behavior: theory, researchand social issues, pp. 31–53. New York: Holt, Rinehart and Winston.

Kuiper, N. A., and Rogers, T. B. (1979). Encoding of personal information:self-other differences. Journal of Personality and Social Psychology, 37, 499–514.

Kunnen, E. S., and Bosma, H. A. (1994, June). A growth model of commitmentsbased on a dynamic systems perspective. Paper presented at the Meetingsof the Society for the Study of Behavioral Development, Amsterdam, TheNetherlands.

Kunnen, E. S., and Bosma, H. A. (2000). Development of meaning making: adynamic systems conceptualization.New Ideas in Psychology, 18, 57–82.

Kunnen, E. S., Bosma, H. A., and Van Geert, P. L. C. (in press). A dynamicsystems approach to identity formation: theoretical background andmethodological possibilities. In J.-E. Nurmi (ed.),Navigating through adoles-cence: European perspectives. Michigan State University Series on Children,Youth and Families. New York: Garland Publishing

Kunnen, E. S., and KleinWassink,M. E. (1999). The transitional process in adultmeaning making. Presentation at the XIth European Conference on Devel-

244 References

opmental Psychology, September 1–5. Spetses, Greece.l’Ecuyer, R. (1992). An experiential-developmental framework and methodol-

ogy to study the transformationsof the self-concept from infancy toold age.In T. M. Brinthaupt and R. P. Lipka (eds.), The self: definitional and method-ological issues, pp. 96–134. Albany, NY: State University of New York Press.

Labouvie-Vief, G., Orwell, L., and Manion, M. (1995). Narratives of mind,gender, and the life-course. Human Development, 38, 239–257.

Lagerspetz, K. (1961). Genetic and social causes of aggressive behavior in mice.Scandinavian Journal of Psychology, 2, 167–173.

Lakoff, G. and Johnson, M. (1999). Philosophy in the flesh. New York: BasicBooks.

Landauer, R. (1987). Role of relative stability in self-repair and self-mainte-nance. In F. E. Yates (ed.), Self-organizing systems: the emergence of order,pp. 435–443. New York: Plenum.

Lazarus, R. S. (1991a). Emotion and adaptation. New York: Oxford UniversityPress.

Lazarus, R. S. (1991b). Cognition and motivation in emotion. American Psychol-ogist, 46, 352–367.

LeDoux, J. E. (1989). Cognitive-emotional interactions in the brain. Cognitionand Emotion, 3, 267–289.

Legerstee,M., Anderson, D., and Schaffer, A. (1998). Five- and eight-month-oldinfants recognize their faces and voices as familiar and social stimuli. ChildDevelopment, 69, 37–50.

Leifer,M. (1980).Psychological effects of motherhood: a study of first pregnancy.NewYork: Praeger.

Leont’ev, A. N. (1978). Activity, consciousness, and personality. Trans. M. J. Hall.Englewood Cliffs, NJ: Prentice-Hall.

Levinas, E. (1969). Totality and infinity. Trans. A. Lingus. Pittsburgh: DuquesneUniversity Press.

Levine, R. L., and Fitzgerald, H. E. (1992a). Analysis of dynamic psychologicalsystems: basic approaches to general systems, dynamic systems and cybernetics.New York: Plenum Press.

Levine, R. L., and Fitzgerald, H. E. (1992b). Analysis of dynamic psychologicalsystems: methods and applications.New York: Plenum Press.

Lewis, C. S. (1967). Studies in words, 2nd edn. Cambridge: CambridgeUniversityPress.

Lewis, M. (1992). Shame, the exposed self.New York: Free Press.Lewis, M. (1993). The emergence of human emotions. In M. Lewis and J. M.

Haviland (eds.), Handbook of emotions, pp. 223–235. New York: GuilfordPress.

Lewis, M., and Brooks-Gunn, J. (1979). Social cognition and the acquisition of self.New York: Plenum Press.

Lewis, M., Sullivan, M. W., Stanger, C., and Weiss, M. (1989). Self developmentand self-conscious emotion. Child Development, 60, 146–156.

Lewis, M. D. (1995). Cognition-emotion feedback and the self-organization ofdevelopmental paths. Human Development, 38, 71–102.

Lewis,M. D. (1996). Self-organising cognitive appraisals.Cognition and Emotion,10, 1–25.

Lewis, M. D. (1997). Personality self-organization: cascading constraints oncognition-emotion interaction. In A. Fogel, M. C. Lyra, and J. Valsiner(eds.), Dynamics and indeterminism in developmental and social processes,

245References

pp. 193–216. Mahwah, NJ: Erlbaum.Lewis, M. D. (2000a). Emotional self-organization at three time scales. In M. D.

Lewis and I. Granic (eds.), Emotion, development, and self-organization,pp. 23–69. New York: Cambridge University Press.

Lewis, M. D. (2000b). The promise of dynamic systems approaches foran integrated account of human development. Child Development, 71,36–43.

Lewis, M. D., and Douglas, L. (1998). A dynamic systems approach to cogni-tion-emotion interactions in development. In M. F. Mascolo and S. Griffin(eds.), What develops in emotional development?, pp. 159–188. New York:Plenum.

Lewis, M. D., and Ferrari, M. D. (unpublished). Reply to Carol Magai. Toronto:OISE/University of Toronto, Department of Human Development andApplied Psychology.

Lewis, M. D., and Granic, I. (1999a). Who put the self in self-organization? Aclarification of terms and concepts for developmental psychopathology.Development and Psychopathology, 11, 365–374.

Lewis, M. D., and Granic, I. (1999b). Self-organization of cognition-emotioninteractions. In T. Dalgleish and M. J. Power (eds.), Handbook of cognitionand emotion, pp. 683–701. Chichester, UK: John Wiley and Sons Ltd.

Lewis, M. D., Lamey, A. V., and Douglas, L. (1999). A new dynamic systemsmethod for analyzing early socio emotional development. DevelopmentalScience, 2, 457–475.

Lightfoot, C. (1997). Transforming the canonical cowboy: notes on the deter-minacy and indeterminacy of children’s play and cultural development. InA. Fogel, M. C. Lyra, and J. Valsiner (eds.), Dynamics and indeterminism indevelopmental and social processes, pp. 165–174. Mahwah, NJ: Erlbaum.

Locke, J. (1690). An essay concerning human understanding.New York: Penguin,1998.

Loewald, H. W. (1980). Papers on psychoanalysis.New Haven and London: YaleUniversity Press.

Logan, R. D. (1987). Historical change in prevailing sense of self. In K. Yardleyand T. Honess (eds.), Self and identity: psychosocial perspectives, pp. 13–26.New York: Wiley.

Lucariello, J. (1995). Mind, culture, person: elements in a cultural psychology.Human Development, 38, 2–18.

Lyra, M., and Winegar, L. T. (1997). Processual dynamics of interation throughtime: adult–child interactions and process of development. In A. Fogel, M.Lyra, and J. Valsiner (eds.),Dynamics and indeterminism in developmental andsocial processes, pp. 93–109. Mahwah, NJ: Erlbaum.

Maclean, P. (1990). The triune brain in evolution: role in palaeocerebral functions.New York: Plenum Press.

Magai, C. (1999). Personality change in adulthood: loci of change and the role ofinterpersonal process. International Journal of Aging and Human Develop-ment, 49, 339–352.

Magai, C., and Haviland-Jones, J. M. (in press). Emotion and personality: dynamictrajectories of lives.New York: Cambridge University Press.

Magai, C., and Hunziker, J. (1993). Tolstoy and the riddle of developmentaltransformation: a lifespan analysis of the role of emotions in personalitydevelopment. In M. Lewis and J. M. Haviland (eds.), Handbook of emotions,pp. 247–259. New York: Guilford Press.

246 References

Magai, C., and McFadden, S. H. (1995). The role of emotions in social and personal-ity development: history, theory, and research.New York: Plenum Press.

Magai, C., and Nusbaum, B. (1996). Personality change in adulthood: dynamicsystems, emotions, and the transformed self. In C. Magai and S. H. McFad-den (eds.), Handbook of emotion, adult development, and aging, pp. 403–420.San Diego: Academic Press.

Magnusson, D., and Stattin, H. (1998). Person-context interaction theories. InR. M. Lerner (vol. ed.; general ed. W. Damon),Handbook of child psychology,5th edn., vol. 1, Theoretical models of human development, pp. 685–759. NewYork: Wiley.

Malatesta, C. (1990). The role of emotion in the development and organizationof personality. In R. Dienstbier and R. Thompson (eds.), Nebraska Sympo-sium on Motivation, vol. 36, pp. 1–56. Lincoln: University of NebraskaPress.

Malatesta, C., Culver, C., Tesman, J., and Shepard, B. (1989). The developmentof emotion expression during the first two years of life. Monographs of theSociety for Research on Child Development, 54, nos. 1 and 2, pp. 1–104.Chicago: University of Chicago Press.

Malatesta, C., and Haviland, J. M. (1982). Learning display rules: the socializ-ation of emotion expression in infancy. Child Development, 53, 991–1003.

MalatestaC. Z., andWilson, A. (1988). Emotion cognition interaction in person-ality development: a discrete emotions, functionalist analysis. British Jour-nal of Social Psychology, 27, 91–112.

Malatesta-Magai, C., and Culver, C. (1991). Identification of emotions, moods,and personality via diary data. In K. T. Strongman (ed.), InternationalReview of Studies on Emotion, pp. 163–186. New York: JohnWiley and Sons.

Malatesta-Magai, C., Jonas, R., Shepard, B., and Culver, C. (1992). Type Apersonality and emotional expressivity in younger and older adults. Psy-chology and Aging, 7, 551–561.

Marcia, J. E., (1980). Identity in adolescence. In J. Adelson (ed.), Handbook ofadolescent psychology, pp. 159–187. New York: Wiley.

Marcia, J. E., Waterman, A. S., Matteson, D. R., Archer, S. L., and Orlofsky, J. L.(1993).Ego identity: a handbook for psychosocial research.NewYork: Springer.

Markus, H. R. (1977). Self-schemata and processing information about the self.Journal of Personality and Social Psychology, 35, 63–78.

Markus, H. R., and Kitayama, S. (1991). Culture and the self: implications forcognition, emotion, and motivation. Psychological Review, 98, 224–253.

Markus, H. R., and Kunda, Z. (1986). Stability and malleability of the self-concept. Journal of Personality and Social Psychology, 51, 858–866.

Markus, H. R., and Nurius, P. (1986). Possible selves. American Psychologist, 41,954–969.

Markus,H. R., and Sentis, K. (1982). The self in social information processing. InJ. Suls (ed.), Psychological perspectives on the self, vol. 1, pp. 41–70. Hillsdale,NJ: Erlbaum.

Markus, H. R., and Wurf, E. (1987). The dynamic self-concept: a social psycho-logical perspective. Annual Review of Psychology, 38, 299–337.

Marsh, H. W., Byrne, B. M., and Shavelson, R. J. (1992). A multidimensional,hierarchical self-concept. In T.M. Brinthaupt and R. P. Lipka (eds.), The self:definitional and methodological issues, pp. 44–95. Albany, NY: State Univer-sity of New York Press.

Marsh, H. W., and Hattie, J. (1996). Theoretical perspectives on the structure of

247References

self-concept. In B. A. Bracken (ed.), Handbook of self-concept: developmental,social, and clinical considerations, pp. 38–90. New York: Wiley.

Mascolo, M. F., and Fischer, K. W. (1998). The development of self through thecoordination of component systems. In M. D. Ferrari and R. J. Sternberg(eds.), Self-awareness: its nature and development, pp. 332–384. New York:Guilford Press.

Mascolo, M. F., Harkins, D., and Harakal, T. (2000). The dynamic constructionof emotion: varieties in anger. In M. D. Lewis and I. Granic (eds.), Emotion,development, and self-organization, pp. 125–152. NewYork: Cambridge Uni-versity Press.

Maturana, H. R., and Varela, F. J. (1987). The tree of knowledge: the biological rootsof human understanding. Boston: New Science Library.

McAdams, D. P. (1985). Power, intimacy, and the life story: personological inquiriesinto identity. Chicago: Dorsey Press (reprinted by Guilford Press).

McAdams, D. P. (1983). The stories we live by: personal myths and the making of theself. New York: Morrow.

McAdams, D. P., Diamond, A., de St. Aubin, E., and Mansfield, E. ( 1997).Stories of commitment: the psychosocial construction of generative lives.Journal of Personality and Social Psychology, 72, 678–694.

McClelland, D. C. (1984).Motives, personality, and society.New York: Praeger.Mead, G.H. (1934).Mind, self, and society.Chicago: University of Chicago Press.Meltzoff, A. N., and Moore, M.K. (1979). Interpreting ‘‘imitiative’’ responses in

early infancy. Science, 205, 217–219.Merleau-Ponty, M. (1945). Phenomenologie de la perception [Phenomenology of

perception]. Paris: Gallimard; trans. Colin Smith (1962), Phenomenology ofperception. London: Routledge and Kegan Paul.

Merleau-Ponty, M. (1968). The visible and the invisible. Trans. A. Lingis. Evans-ton, IL: Northwestern University Press.

Michaels, C. F., and Carello, C. (1981). Direct perception. Englewood Cliffs, NJ:Prentice-Hall.

Michel, G. F., Camras, L. A., and Sullivan, J. (1992). Infant interest expressions ascoordinativemotor structures. Infant Behavior andDevelopment, 15, 347–358.

Miller, P. J., Mintz, J., Hoogstra, L., Fung, H., and Potts, R. (1992). The narratedself: young children’s construction of self in relation to others in conversa-tional stories of personal experience.Merrill-Palmer Quarterly, 38, 45–67.

Miller,W. I. (1997). The anatomy of disgust.Cambridge,MA: Harvard UniversityPress.

Mischel, W., and Shoda, Y. (1995). A cognitive-affective system theory ofpersonality: reconceptualizing situations, dispositions, dynamics, and in-variance in personality structure. Psychological Review, 102, 246–268.

Mistry, J. (1993). Cultural context in the development of children’s narratives.In J. Altarriba (ed.), Cognition and culture: a cross-cultural approach to psychol-ogy, pp. 207–228. Amsterdam: Elsevier Science Publishers.

Moen, P., Elder, G. H., and Luscher, K. (eds.) (1995). Examining lives in context:perspectives on the ecology of human development.Washington, DC: AmericanPsychological Association.

Molenaar, P. C.M. (1985). A dynamic factormodel for the analysis of multivari-ate time series. Psychometrika, 50, 181–202.

Molenaar, P. C. M., and Newell, K. M. (1998).Applications of non-linear dynamicsto developmental process modeling.Mahwah, NJ: Erlbaum.

Moreno, J. L. (1946). Psychodrama, vol. 1. New York: Beacon House.

248 References

Moreno, J. L. (1959).Psychodrama, vol. 2, Foundations of psychotherapy. NewYork:Beacon House.

Moreno, J. L., and Elefthery, D. G. (1982). An introduction to group psycho-drama. In G. M. Gazda (ed.), Basic approaches to group psychotherapy andgroup counseling, 3rd edn., pp. 101–131. Springfield, IL: Charles C. Thomas.

Moreno, J. L., and Moreno, Z. T. (1969). Psychodrama, vol. 3, Action therapy andprinciples of practice. New York: Beacon House.

Morris, W. N. (1989).Mood: the frame of mind.New York: Springer.Msihara, B. L. (1996). A dynamic developmental model of suicide. Human

Development, 39, 181–194.Mummendey, H. D. (ed.) (1988). Verhalten und Einstellung (Behavior and atti-

tude). Berlin: Springer.Murray, J. D. (1989).Mathematical biology. Berlin: Springer.Murray, K. (1985). Life as fiction. Journal for the Theory of Social Behaviour, 15,

173–188.Neisser, U. (1991). Five kinds of self-knowledge. In D. Kolak and R. Martin

(eds.), Self and identity: contemporary philosophical issues, pp. 386–406.Toronto: Cassell and Collier Macmillan.

Neisser, U. (ed.) (1993). The perceived self: ecological and interpersonal sources ofself-knowledge. Cambridge: Cambridge University Press.

Nelson, K. (1992). Emergence of autobiographical memory at age 4. HumanDevelopment, 35, 172–177.

Nesselroade, J. R. (1991). Interindividual differences in intraindividual change.In L. M. Collins and J. L. Horn (eds.), Best methods for the analysis of change,pp. 92–105. Washington DC: American Psychological Association.

Niedenthal, P. M., Tangney, J. P., and Gavanski, I. (1994). ‘‘If only I weren’t’’versus ‘‘if only I hadn’t’’: distinguishing shame and guilt in counterfactualthinking. Journal of Personality and Social Psychology, 67, 585–595.

Nolen-Hoeksema, S., and Morrow, J. (1991). A prospective study of depressionand distress following a natural disaster: the 1989 Loma Prieta earthquake.Journal of Personality and Social Psychology, 61, 115–121.

Novak, G. (1996). Developmental psychology: dynamic systems and behavior analy-sis. Reno, NV: Context Press.

Nowak, A., and Vallacher, R. R. (1999). Dynamical social psychology. London:Guilford Press.

Nowak, A., Vallacher, R. R., Tesser, A., and Borkowski, W. (2000). Society ofself: the emergence of collective properties in self-structure. PsychologicalReview, 107, 39–61.

Oatley, K. (1992). Best laid schemes: the psychology of emotions. Cambridge: Cam-bridge University Press.

Oatley, K., and Johnson-Laird, P. N. (1987). Towards a cognitive theory ofemotions. Cognition and Emotion, 1, 29–50.

Ohman,A. (1993). Fear and anxiety as emotional phenomenon. InM. Lewis andJ. M. Haviland (eds.), Handbook of emotions, pp. 511–536. New York:Guilford.

Oosterwegel, A. (1992). The organization of the self-system: developmentalchanges in childhood and adolescence. University of Amsterdam, unpub-lished doctoral dissertation.

Oosterwegel, A. (September 1996). Away from a rational self in psychology:how re-inventing the wheel can provide clarity. Paper presented at aMeeting of the Frankfurt/Southampton Link, Frankfurt.

249References

Oosterwegel, A., and Wicklund, R. A. (eds.) (1995). The self in European andNorth American culture: development and processes.Dordrecht: Kluwer.

Ormel, J., and Schaufeli, W. B. (1991). Stability and change in psychologicaldistress and their relationship with self-esteem and locus of control: adynamic equilibriummodel. Journal of Personality and Social Psychology, 60,288–299.

Ortony, A., Clore, G., and Collins, A. (1988). The cognitive structure of emotions.Cambridge: Cambridge University Press.

Overton, W. F. (1998). Developmental psychology: philosophy, concepts, andmethodology. In R. M. Lerner (vol. ed.; W. Damon, general ed.), Handbookof child psychology, 5th edn., vol. 1, Theoretical models of human development,pp. 107–188. New York: Wiley.

Pacini, R., Muir, F., and Epstein, S. (1998). Depressive realism from the perspec-tive of cognitive-experiential self-theory. Journal of Personality and SocialPsychology, 74, 1056–1068.

Panksepp, J. (1998). Affective neuroscience: the foundations of human and animalemotions. New York: Oxford University Press.

Panksepp, J. (2000a) Emotions as natural kinds within the mammalian brain. InM. Lewis and J. M. Haviland-Jones (eds.), The Handbook of Emotion, 2ndedn., pp. 137–156. New York: Guilford Press.

Panksepp, J. (2000b). The neuro dynamics of emotions: an evolutionary-neurodevelopmental view. In M. D. Lewis and I. Granic (eds.), Emotion, develop-ment, and self-organization, pp. 236–264. New York: Cambridge UniversityPress.

Pedrosa, M. I., Carvalho, A. M. A., and Imperio-Hamburger, A. (1997). Fromdisordered to ordered movement: attractor configuration and develop-ment. In A. Fogel, M. C. D. P. Lyra, and J. Valsiner (eds.), Dynamics andindeterminism in developmental and social processes, pp. 135–152. Hillsdale,NJ: Erlbaum.

Piaget, J. (1935). La naissance de l’intelligence chez l’enfant.Neuchatel: Delacheauxet Niestl.

Piaget, J . (1949). La psychologie de l’intelligence. Paris: Colin.Piaget, J. (1968). Le structuralisme. Paris: Presses Universitaires de France.Pickering, J. (1999). The self as a semiotic process. Journal of Consciousness

Studies, 6, 31–47.Pipp, S. (1993). Infants’ knowledge of self, other, and relationship. In U. Neisser

(ed.), The perceived self: ecological and interpersonal sources of self-knowledge,pp. 185–204. Cambridge: Cambridge University Press.

Plato (1961). The collected dialogues, ed. E. Hamilton and H. Cairns. Princeton,NJ: Princeton University Press. (Original work written 390 BC)

Polivy, J. (1998). The effects of behavioral inhibition: integrating internal cues,cognition, behavior, and affect. Psychological Inquiry, 9, 181–204.

Port, R. F., and Van Gelder, T. (1995).Mind as motion: explorations in the dynamicsof cognition. Cambridge, MA: Bradford Books/MIT Press.

Prigogine, I., and Nicolis, G. (1977). Self-organization in non-equilibrium systems.New York: Wiley-Interscience.

Prigogine, I., and Stengers, E. (1984). Order out of chaos: man’s new dialogue withnature. Toronto: Bantam Books.

Prout, H. T., and Prout, S. M. (1996). Global self-concept and its relationship tostressful life conditions. In B. A. Bracken (ed.), Handbook of self-concept:developmental, social, and clinical considerations, pp. 259–286. New York:

250 References

Wiley.Provinelli, D. J., and Simon, B. B. (1998). Young children’s understanding of

briefly versus extremely delayed images of self: emergences of the autobio-graphical stance. Developmental Psychology, 34, 188–194.

Reddy, V. (1991). Playing with others’ expectations: teasing andmucking aboutin the first year. In A. Whiten (ed.), Natural theories of mind: evolution,development and simulation of everyday mindreading, pp. 143–158. Oxford:Basil Blackwell.

Reed, E. S. (1987). Why do things look as they do? The implications of J. J.Gibson’s:The ecological approach to visual perception. In A. Costall andA. Still(eds.), Cognitive psychology in question, pp. 90–114. New York: St. Martin’sPress.

Rhodewalt, F., and Morf, C. C. (1998). On self-aggrandizement and anger: atemporal analysis of narcissism and affective reactions to success andfailure. Journal of Personality and Social Psychology, 74, 672–685.

Ricoeur, P. (1983). Temps et recit, vol. 1, L’ordre philosophique. Paris: Seuil. Trans.K. McLaughlin and D. Pellauer, Time and narrative, vol. 1. Chicago: Univer-sity of Chicago Press, 1984.

Ricoeur, P. (1996). The hermeneutics of action. Thousand Oaks, CA: Sage.Robinson, M. D. (1998). Running from William James’ bear: a review of preat-

tentive mechanisms and their contributions to emotional experience. Cog-nition and Emotion, 12, 667–696.

Rochat, P. (1995). Early objectification of the self. In P. Rochat (ed.), Theself in early infancy: theory and research, pp. 53–72. Amsterdam: ElsevierScience.

Rochat, P., and Hespos, S. J. (1997). Differential rooting response by neonates:evidence for an early sense of self. Early Development and Parenting, 6,105–112.

Rogers, C. R. (1951). Client-centered therapy: its current practice, implications, andtheory. Boston: HoughtonMifflin.

Rogers, C. R. (1961). On becoming a person: a therapist’s view of psychotherapy.Boston: HoughtonMifflin.

Rogers, T. B., Kuiper, N. A., and Kirker, W. S. (1977). Self-reference and theencoding of personal information. Journal of Personality and Social Psychol-ogy, 35, 677–688.

Rosenberg, M. (1979). Conceiving the self.New York: Basic Books.Rosenberg, M. (1986). Self-concept from middle childhood through adoles-

cence. In J. Suls and A. G. Greenwald (eds.), Psychological perspective on theself, vol. 3, pp. 107–135. Hillsdale, NJ: Erlbaum.

Rozin, P., and Fallon, A. E. (1987). A perspective on disgust. PsychologicalReview, 94, 23–41.

Rubin, D. C. (ed.) (1996). Remembering our past: studies in autobiographical mem-ory. Cambridge: Cambridge University Press.

Ruhland, R. (1998). Going the distance: a non-linear approach to change inlanguage development. University of Groningen: unpublished doctoraldissertation.

Rychlak, J. F. (1988). The psychology of rigorous humanism, 2nd edn. New York:New York University Press.

Saarni, C. (1997). Emotional competence and self-regulation in childhood. In P.Salovey and D. J. Sluyter (eds.), Emotional development and emotional intelli-gence: educational implications, pp. 35–66. New York: Basic Books.

251References

Saarni, C., Mumme, D. L., and Campos, J. J. (1998). Emotional development:action, communication, and understanding. In N. Eisenberg (vol. ed.;W. Damon, general ed.),Handbook of child psychology, (5th edn. vol. 3, Social,emotional and personality development, pp. 237–309. New York: Wiley.

Salovey, P. (1992). Mood-induced self-focused attention. Journal of Personalityand Social Psychology, 62, 699–707.

Sameroff, A. J., and Fiese, B. H. (1990). Transactional regulation and earlyintervention. In S. J. Meissels and J. P. Shonkoff (eds.), Handbook of earlychildhood intervention, pp. 119–149. Cambridge: Cambridge UniversityPress.

Sampson, E. E. (1989). The deconstruction of the self. In J. Shotter and K. J.Gergen (eds.), Texts of identity, pp. 1–19. London: Sage.

Sampson, E. E. (1993).Celebrating the other: a dialogic account of human nature. SanFrancisco: Westview Press.

Sander, L. W. (1962). Issues in early mother–child interaction. Journal of theAmerican Academy of Child Psychiatry, 1, 141–166.

Sansone, C., and Berg, C. A. (1993). Adapting to the environment across the lifespan: different process or different inputs? International Journal of Behav-ioural Development, 16, 215–241.

Sarbin, T. R. (1986a). Narrative psychology: the storied nature of human conduct.New York: Praeger.

Sarbin, T. R. (1986b). The narrative as a root metaphor for psychology. In T. R.Sarbin (ed.), Narrative psychology: the storied nature of human conduct, pp. 3–21. New York: Praeger.

Sarbin, T. R. (1993). The narrative as the root metaphor for contextualism. In S.C.Hayes and L. J. Hayes (eds.),Varieties of scientific contextualism, pp. 51–65.Reno, NV: Context Press.

Sartre, J. P. (1939). Esquisse d’une theorie des emotions. Paris: Hermann. Trans.B. Frechtman, The emotions: sketch of a theory. New York: PhilosophicalLibrary, 1984.

Savin-Williams, R. C., and Demo, D. H. (1983). Situational and transituationaldeterminants of adolescent self-feelings. Journal of Personality and SocialPsychology, 44, 824–833.

Schank, R. C., andAbelson, R. P. (1995). Knowledge andmemory: the real story.In R. S. Wyer (ed.), Advances in social cognition, vol. 8, pp. 1–87. Hillsdale,NJ: Erlbaum.

Scheff, T. J. (1987). Creativity and repetition: a theory of the coarse emotions. InJ. Rabow, G. Platt, and M. Goldman (eds.), Advances in psychoanalyticsociology, pp. 70–100. Malabor, FL: Robert E. Krieger.

Scherer, K. R. (1984). Emotion as a multicomponent process: a model and somecultural data. In P. Shaver (ed.), Review of personality and social psychology,vol. 5, pp. 37–63. Berverly Hills: Sage.

Scherer, K. R. (1999). Appraisal theory. In T. Dalgleish and M. Power (eds.),Handbook of cognition and emotion, pp. 637–663. London: Wiley.

Scherer, K. R. (2000). Emotion as episodes of subsystem synchronization drivenby nonlinear appraisal processes. In M. D. Lewis and I. Granic (eds.),Emotion, development, and self-organization, pp. 70–99. New York: Cam-bridge University Press.

Schlenker, B. R., and Weigold, M. F. (1989). Goals and the self-identificationprocess: constructing desired identities. In L. A. Pervin (ed.), Goal conceptsin personality and social psychology, pp. 243–290. Hillsdale, NJ: Erlbaum.

252 References

Schneider-Rosen, K., and Cicchetti, D. (1991). Early self-knowledge and emo-tional development: visual self-recognition and affective reactions to mir-ror self-images in maltreated and non-maltreated toddlers. DevelopmentalPsychology, 27, 471–478.

Schore, A. N. (1994). Affect regulation and the origin of the self: the neurobiology ofemotional development.Hillsdale, NJ: Erlbaum.

Schore, A. N. (1997). Early organization of the nonlinear right brain and devel-opment of a predisposition to psychiatric disorders. Development and Psy-chopathology, 9, 595–631.

Schore, A. N. (2000). Contributions from the neurobiology of emotional devel-opment to models of self-organization. In M. D. Lewis and I. Granic (eds.),Emotion, development, and self-organization, pp. 155–185. New York: Cam-bridge University Press.

Schwartz, G. E., and Russek, L. G. (1997). Do all dynamical systems havememory? Implications of the systemic memory hypothesis for science andsociety. In K. H. Pribram and J. S. King (eds.), Brain and values: behavioralneurodynamics, pp. 249–273. Hillsdale, NJ: Erlbaum.

Schweder, R. A., Goodnow, J., Hatano, G., LeVine, R. A., Markus, H., andMiller, P. (1998). The cultural psychology of development: onemind,manymentalities. In R. M. Lerner (vol. ed.; W. Damon, general ed.), Handbook ofchild psychology, 5th edn. vol. 1, Theoretical models of human development,pp. 865–937. New York: Wiley.

Sedikides, C., and Strube, M. J. (1997). Self-evaluation: to thine own self begood, to thine own self be sure, to thine own self be true, and to thine ownself be better. InM. P. Zanna (ed.),Advances in experimental social psychology,vol. 29, pp. 209–269. New York: Academic Press.

Selye, H. (1955). Stress and disease. Science, 122, 625–631.Selye, H. (1956). The stress of life. New York: McGraw-Hill.Shotter, J. (1981). Vico, moral worlds, accountability, and personhood. In P.

Heelas and A. Lock (eds.), Indigenous psychologies: the anthropology of the self,pp. 266–284. New York: Academic Press.

Shotter, J. (1989). Social accountability and the social construction of ‘‘you.’’ In J.Shotter and K. Gergen (eds.), Texts of identity, pp. 133–151. London: Sage.

Skarda, C. A., and Freeman, W. J. (1987). How brains make chaos in order tomake sense of the world. Behavioral and Brain Sciences, 10, 161–195.

Smirnov, S. D. (1994). Intelligence and personality in the psychological theoryof activity. In R. J. Sternberg and P. Ruzgis (eds.), Personality and intelligence,pp. 221–247. Cambridge University Press.

Smith, L. B. (1995). Self-organizing processes in learning to learn words: devel-opment is not induction. In C. A. Nelson (eds.),New perspectives on learningand development, Minnesota Symposia on Child Psychology, vol. 28, pp. 1–32. New York: Academic Press.

Smith, L. B., and Thelen, E. (1993). A dynamic systems approach to development:applications. Cambridge, MA: Bradford Books/MIT Press.

Sonnemans, J., and Frijda, N. H. (1995). The determinants of subjective emo-tional intensity. Cognition and Emotion, 9, 483–507.

Spencer, S. (1971). Space, time, and structure in the modern novel.New York: NewYork University Press.

Sroufe, L. A. (1979). Socio emotional development. In J. Osofsky (eds.), Hand-book of infant development, pp. 462–518. New York: Wiley.

Sroufe, L. A., and Waters, E. (1976). The ontogenesis of smiling and laughter: a

253References

perspective on the organization of development in infancy. PsychologicalReview, 83, 173–189.

Stapley, J., andHaviland, J. M. (1989). Beyond depression: gender differences innormal adolescents’ emotional experiences. Sex Roles, 20, 295–308.

Stein, N. L. and Trabasso, T. (1992). The organization of emotional experience:creating links between emotion, thinking, and intentional action. Cognitionand emotion, 6, 225–244.

Stein, N. L., Trabasso, T., and Liwag,M. (1993). The representation and organiz-ation of emotional experience: unfolding the emotion episode. InM. Lewisand J. M. Haviland (eds.), Handbook of emotions, pp. 279–300. New York:Guilford Press.

Stern, D. (1985). The interpersonal world of the infant.New York: Basic Books.Stern, D. (1993). The role of feelings for an interpersonal self. In U. Neisser

(ed.), The perceived self, pp. 205–215. Cambridge: Cambridge UniversityPress.

Stewart, I. (1989). Does God play dice?Oxford: Basil Blackwell.Stipek, D. J., Recchia, S., and McClintic, S. (1992). Self-evaluation in young

children.Monographs of the Society for Research in Child Development, 57(serial no. 226). Chicago: University of Chicago Press.

Stratton, P. (1988). Parents’ conceptualization of children as the organizer ofculturally structured environments. In J. Valsiner (ed.), Child developmentwithin culturally structured enviroments, vol. 1, Parental cognition and adult-child interaction, pp. 5–29. Norwood, NJ: Ablex.

Strauman, T. J. (1996). Stability within the self: a longitudinal study of thestructural implications of self-discrepancy theory. Journal of Personality andSocial Psychology, 71, 1142–1153.

Svrakic, N. M., Svrakic, D. M., and Cloninger, C. R. (1996). A general quantitat-ive theory of personality development: fundamentals of a self-organizingcomplex. Development and Psychopathology, 8, 247–272.

Tangney, J. P., and Fischer, K. W. (eds.) (1995). Self-conscious emotions: thepsychology of shame, guilt, embarrassment, and pride. New York: GuilfordPress.

Teasdale, J. D., and Barnard, P. J. (1993).Affect, cognition, and change: re-modellingdepressive thought.Hillsdale, NJ: Erlbaum.

Tesser, A. (1988). Toward a self-evaluationmaintenance model of social behav-ior. In L. Berkowitz (ed.), Advances in experimental social psychology, Vol. 21,pp. 181–227. San Diego, CA: Academic Press.

Thatcher, R. W. (1994). Psychopathology of early frontal lobe damage: depend-ence on cycles of development. Development and Psychopathology, 6, 565–596.

Thelen, E., and Smith, L. B. (1994). A dynamic systems approach to the developmentof cognition and action. Cambridge, MA: Bradford/MIT Press.

Thelen, E., and Ulrich, B. D. (1991). Hidden skills: a dynamic systems analysis oftreadmill stepping during the first year. Monographs of the Society for Re-search in Child Development, 56, 1 (serial no. 223). Chicago: University ofChicago Press.

Tomkins, S. S. (1962). Affect, imagery, consciousness, vol. 1: The positive affects.New York: Springer.

Tomkins, S. S. (1978). Script theory: differential magnification of affects. In H. E.Howe, Jr., and R. A. Dunstbier (eds.), Nebraska symposium on motivation,pp. 201–236. Lincoln: University of Nebraska Press.

254 References

Trevarthen, C. (1993). The self born in intersubjectivity: the psychology of aninfant communicating. In U. Neisser (ed.), The perceived self: ecological andinterpersonal sources of self-knowledge, pp. 121–173. Cambridge: CambridgeUniversity Press.

Trevarthen, C., and Hubley, P. (1978). Secondary intersubjectivity: confidence,confiding and acts of meaning in the first year. In A. Lock (ed.), Action,gesture and symbol: the emergence of language, pp. 183–227. New York: Aca-demic Press.

Triandis, H. C. (1989). The self and social behavior in differing cultural contexts.Psychological Review, 96, 506–520.

Tucker, D. M. (1992). Developing emotions and cortical networks. In M. R.Gunnar and C. Nelson (eds.), Developmental behavioral neuroscience, Min-nesota symposia on child psychology, vol. 24, pp. 75–128. Hillsdale, NJ:Erlbaum.

Van der Maas, H. L. J., and Molenaar, P. C. M. (1992). Stagewise cognitivedevelopment: an application of catastrophe theory. Psychological Review,99, 395–417.

Van der Meulen, M. (1987). Self-references in young children: content, metadimen-sions, and puzzlement.Groningen: Kinderstudies.

Van Geert, P. L. C. (1989). Cognitive development: a process of organizedself-organization. In G. J. Dalenoort (ed.), The paradigm of self, pp. 146–166.London: Gordon and Breach Science Publishers.

VanGeert, P. L. C. (1991). A dynamic systemsmodel of cognitive and languagegrowth. Psychological Review, 98, 3–53.

Van Geert, P. L. C. (1994). Dynamic systems of development: change betweencomplexity and chaos. New York: Prentice Hall/Harvester Wheatsheaf.

Van Geert, P. L. C. (in press). Time, change and intentional action in a dynamicframework: on the nature of psychological causes and measurement. In N.Granott and J. Parziale (eds.),Microdevelopmental clues: transition processes indevelopment and learning, pp. 00–00. Mahwah, NJ: Erlbaum.

Van Geert, P. L. C. (unpublished). Reply to Camras and Michel. Groningen:University of Groningen, Department of Psychology.

Van Gelder, T., and Port, R. F. (1995). It’s about time: an overview of thedynamical approach to cognition. In R. F. Port and T. van Gelder (eds.),Mind as motion: explorations in the dynamics of cognition, pp. 1–43. Cam-bridge, MA: MIT Press.

Van Halen, C. P. M. (in preparation). Experiencing self-definition problemsover the life span. Groningen: University of Groningen, doctoral disserta-tion in preparation.

Van Hooff, J. A. R. A. M. (1972). A structural analysis of the social behavior of asemi-captive group of chimpanzees. In M. Van Cranach and J. Vine (eds.),Social communication andmovement, pp. 75–162. NewYork: Academic Press.

Van der Werff, J. J. (1967). ‘‘Onbewuste’’ zelfbeoordeling: een experiment in destijl vanWernerWolff (‘‘Unconscious’’ self-evaluation: an experiment afterWerner Wolff). Nederlands Tijdschrift voor de Psychologie, 22, 150–161.

Van der Werff, J. J. (1990). The problem of self-conceiving. In H. Bosma and S.Jackson (eds.), Coping and self-concept in adolescence, pp. 14–33. Berlin:Springer.

Varela, F. J., Thompson, E., and Rosch, E. (1991). The embodied mind: cognitivescience and human experience.Cambridge, MA: MIT Press.

Vasil’eva, I. I. (1988). The importance of M. M. Bakhtin’s idea of dialogue and

255References

dialogic relations for the psychology of communication. Soviet Psychology,26, 17–31.

Vedeler, D. (1991). Infant intentionality as object directedness: an alternative torepresentation. Journal for the Theory of Social Behavior, 21, 431–448.

Vedeler, D. (1994). Infant intentionality as object directedness: a method forobservation. Scandinavian Journal of Psychology, 35, 343–366.

Verhofstadt-Deneve, L. (1985). Crises in adolescence and psycho-social devel-opment in young adulthood: a seven-year follow-up study from adialectical viewpoint. In C. J. Brainerd and V. F. Reyna (eds.), Develop-mental psychology, pp. 509–522. Amsterdam, New York, Oxford: NorthHolland.

Verhofstadt-Deneve, L. (1988). The phenomenal-dialectic personality model: aframe of reference for the psychodramatist. Journal of Group Psychotherapy,Psychodrama and Sociometry, 41, 3–20.

Verhofstadt-Deneve, L. (1995). How to work with dreams in psychodrama:developmental therapy from an existential-dialectic viewpoint. Interna-tional Journal of Group Psychotherapy, 45, 405–435.

Verhofstadt-Deneve, L. (1997). Using conflict in a developmental therapeuticmodel. International Journal of Adolescent Medicine and Health, 9, 151–164.

Verhofstadt-Deneve, L. (1999). Action- and drama-techniques with adolescentvictims of violence: a developmental therapeutic model. International Jour-nal of Adolescent Medicine and Health, 11, 351–367.

Verhofstadt-Deneve, L. (2000). Theory and practice of action and drama techniques:developmental psychotherapy from an existential-dialectical viewpoint. London,Philadelphia, Sydney: Jessica Kingsley Publishers.

Verhofstadt-Deneve, L. and Schittekatte, M. (1996). Adolescents have becomeadults: a 15-year follow-up. In Verhofstadt-Deneve, L., Kienhorst, I., andBraet, C. (eds.), Conflict and development in adolescence, pp. 47–60. Leiden:DSWO Press, University of Leiden.

Voloshinov, V. N. (1929/1986). Marxism and the philosophy of language. Cam-bridge, MA: Harvard University Press.

Vygotsky, L. S. (1979). Mind in society: the development of higher psychologicalprocesses, 2nd edn. Cambridge, MA: Harvard University Press.

Walker, L. J. (1983). Sources of cognitive conflict for stage transition in moraldevelopment.Developmental Psychology, 19, 103–110.

Watkins, M. (1986). Invisible guests: the development of imaginal dialogues. Hill-sdale, NJ: Erlbaum.

Wells, A. J. (1992). Variations in self-esteem in daily life: methodological anddevelopmental issues. In R. P. Lipka and T. M. Brinthaupt (eds.), Self-perspectives across the life span, pp. 151–185. Albany, NY: State University ofNew York Press.

Werner,H., andWapner, S. (1953), Sensory-tonic field theoryof perception. In J.S. Bruner and D. Krech (eds.), Perception and personality, pp. 83–107. Dur-ham, NC: Duke University Press.

White, G. M. (1993). Emotions inside out: the anthropology of affect. In M.Lewis and J.M.Haviland (eds.),Handbook of emotions, pp. 29–41.NewYork:Guilford Press.

Whitehead,A. N. (1978).Process and reality, an essay in cosmology.NewYork: TheFree Press.

Wilber, K. (1979). No boundary: eastern and western approaches to personal growth.Boston: Shambala.

256 References

Wolff, W. (1943). The expression of personality.New York: Harper.Wylie, R. (1979). The self concept, vol. 2. Lincoln: University of Nebraska Press.Wylie, R. C. (1989). Measures of self-concept. Lincoln: University of Nebraska

Press.Yasuo, Y. (1987). The body: toward an eastern mind-body theory. Albany: State

University of New York Press.Zajonc, R. B. (1980). Thinking and feeling: preferences need no inferences.

American Psychologist, 35, 151–175.Zender, M. A., and Zender, B. F. (1974). Vygotsky’s view about age periodiz-

ation of child development.Human Development, 17, 24–40.

257References

Author index

Abelson, R. P., 118, 119Abraham, F. D., 64, 185Ackerman, B., 183Allport, G. W., 17, 58Alpert, R., 200, 201Altman, I., 143d’Ancona, H., 66Anderson, D., 97Angyal, A., 129Anne–Albert, 164, 166Antonovsky, A., 112Archer, S. L., 1Aron, A., 200Aron, E., 200Ashmore, R. D., 17, 21Atkins, P. W., 78Ausubel, D., 51

Baddeley, A. D., 118Bailey, S. T., 25Baird, B., 186Bakan, D., 129Bakhtin, M. M., 97, 104, 124, 125, 127, 138,149, 193

Ball, L., 195Baltes, P. B., 18, 23–26, 30–32Bambrough, R., 91Bandura, A., 56Barclay, C. R., 97, 118, 119Barclay, L. C., 27Barnard, P. J., 185, 186Barrett, K. C., 93Bateson, G., 102Baumeister, R. F., 2, 3, 10, 11, 14, 16, 17, 27,28, 31, 53, 119

Beck, A. T., 58Becker, E., 122Beltrami, E., 64Berg, C. A., 19, 21, 30Blasi, A., 216, 227Blass, E. M., 44Block, J., 112, 177

Block, J. H., 112Boden, J. M., 2, 14Boesch, E. E., 97, 121Borkowski, W., 221Bosma, H. A., 3, 6, 97, 107, 109, 117, 169,177, 178, 228

Boulifard, D., 8, 35, 164, 172, 209, 210Bourdieu, P., 191Bracken, B. A., 11, 13, 14, 26Branco, A. U., 98Brandtstadter, J., 19Brewer, W. F., 118, 119Brinthaupt, T. M., 11, 24, 29Brockmeier, J., 191, 197Bronfenbrenner, U., 19, 20, 225Brooks-Gunn, J., 12, 13, 20Brown, B., 143Bruner, J. S., 2, 109, 123, 140, 145, 190, 194,197, 223

Buber, M., 98, 100Buck, R., 183Bullock, M., 12, 13Burghess, D. N., 66Butterworth, G., 97, 185Butz, M. R., 112Byrne, B. M., 5, 10–12, 26

Campos, J. J., 12, 93Camras, L. A., 85, 92, 206, 213, 218Capps, D., 175Captain Picard, 65Carello, C., 97Carnochan, P., 93Carstensen, L. L., 111Carvalho, A. M. A., 98Carver, C. S., 56Cassirer, E., 126Caughey, J. L., 126Chandler, M. J., 191, 192, 195Charles, S. T., 111Chomsky, N., 92Christianson, S., 119

259

Cicchetti, D., 12Clausen, J. A., 20Cloninger, C. R., 182, 229Clore, G., 56Cole, M., 110, 191Collins, A., 56Conville, R., 143Cooley, C. H., 6, 10–12, 17, 18Cooper, C. R., 194, 230Costa, P. T. Jr., 199Crouter, A. C., 19, 20Csikszentmihalyi, M., 21, 31, 37, 98, 112,222

Culver, C., 152, 163

Damasio, A. R., 184Damon, W., 10, 14, 21, 24–26, 29, 121Dannefer, D., 219, 224Davidson, R. B., 4Davitz, J. R., 42Day, J. M., 104De Boeck, P., 21, 154De Levita, D. J., 3De Quincey, C., 98De Rivera, J., 93De Waal, F. B. M., 55, 57De Weerth, C., 85Dedo, J. Y., 87Deiner, E., 199Demo, D. H., 3, 20, 22, 24, 26, 27, 30, 31Demos, E., 103Dennett, D. C., 46Derryberry, D., 182Descartes, R., 138Dewey, J., 94, 98, 99Dostoyevski, F. M., 124, 125, 149Douglas, L., 181, 187, 228Dunne, J., 109Duran, E., 112Duval, S., 37

Eckstein, S. G., 229Edelstein, W., 20Eisenberg, N., 11, 18Ekman, P., 43, 54, 152, 163Elder, G. H., 19Elefthery, D. G., 146Elkind, D., 197Ellis, A., 158, 160Ellis, Albert, 151, 157–161, 166, 169, 170Emde, R. N., 109Epstein, S., 10, 12, 15–18, 21, 29–31, 36, 59,62, 205, 214, 217, 218, 221

Erikson, E. H., 3, 6, 107, 109, 110, 143, 174,175, 202, 204, 216, 217, 219

Fallon, A. E., 45

Feiring, C., 20Ferrari, M. D., 8, 35, 109, 191, 192, 194, 199,200, 201, 211, 212, 214, 216, 218, 219,221–223

Feyereisen, P., 162Fiese, B. H., 22Fine, S., 183Fischer, A. H., 112, 116Fischer, K. W., 14, 18, 21, 30, 35, 93, 196Fitzgerald, H. E., 64Fogel, A., 4, 8, 20–22, 35, 87, 94, 97–99,102–104, 106, 111–113, 115, 117, 181–183,185, 194, 196, 207, 208, 211, 212, 214, 215,225–230

Frank, Anne, 151, 155–157, 159, 161, 167,169, 170

Freeman, M., 191, 197Freeman, W. J., 184, 186, 188Freud, S., 6, 186Fridlund, A. J., 43Frijda, N. H., 7, 16, 17, 34, 35, 41–44, 49,52–55, 58, 59, 62, 93, 96, 101, 103, 116,163, 176, 183, 186, 203–208, 212, 214,217

Ganguly, S. N., 97Gavanski, I., 51Gebelt, J. L., 4Gecas, V., 20Gergen, K. J., 2, 21, 29, 94, 109, 110Gergen, M. M., 109, 110Gibson, J. J., 97Ginsburg, G. P., 102Glodis, K., 216, 227Goertzel, B., 185Goffman, E., 102Goldsmith, H. H., 199Goleman, D., 11Goodnow, J. J., 20Goodwin, B. C., 179, 181Gordon, C., 29Graafsma, T. L.G., 3Granic, I., 5, 87, 182Grannemann, B. D., 27Greenwald, A. G., 10–12, 15–17, 29Grotevant, H. D., 3, 230

Harakal, T., 183Harkins, D., 183Harkness, K. L., 184, 188Harre, R., 11, 31, 116, 118, 139Hart, D., 10, 11, 14, 20, 21, 24–26, 29,121

Harter, S., 2, 3, 11, 14, 18, 20, 24, 26, 27,31, 32, 196

Hatano, G., 20,Hattie, J., 1, 11–13, 16, 21, 22, 24–26, 30

260 Author index

Haviland, J. M., 4, 8, 102, 109, 151–155,172–174, 178, 190, 192, 197

Haviland-Jones, J.M., 8, 35, 158, 164, 201,209–212, 214, 221, 226, 228, 229

Heatherton, T.F., 53Held, R., 48Hermans, H. J. M., 4, 8, 21–23, 30, 35,96–99, 102, 107, 127, 128, 130, 131, 138,141, 142, 145, 147, 149, 193–195, 208,209, 211, 212, 214, 215, 218–220, 226,228, 229

Hermans-Jansen, E., 8, 35, 127, 128, 130,141, 142, 145, 147, 149, 208, 209, 211,212, 214, 215, 218–220, 226, 228, 229

Hespos, S. J., 97Heymans, P., 175, 176, 210, 228Higgins, E. T., 4, 15, 18, 32Hochschild, A., 116Hofbauer, J., 66Holland, J. H., 78Hoorens, V., 12Hopkins, B., 184Horowitz, M. J., 48, 186Hubley, P., 108Huntley, C. W., 12Hunziker, J., 192

Imperio-Hamburger, A., 98Isaacowitz, D. M., 111Isen, A. M., 112Izard, C. E., 31, 44, 152, 182–184, 188

Jackson, E. A., 64, 65James, W., 2, 3, 6, 7, 10–12, 16–18, 23, 50,58, 99, 111, 120–122, 127, 138, 140, 141,202

Jansz, J., 97, 112, 116, 208, 211, 219Jaquish, G. A., 22, 31Jaynes, J., 123Johnson, F., 139Johnson, M., 97Johnson-Laird, P. N., 185Jonas, R., 152Jones, D., 102Jopling, D., 98Josephs, I. E., 97Josselson, R., 137– 230Juhasz, A., 20

Kagan, J., 12, 13Kahlbaugh, P., 102, 109, 151, 153, 178, 190,192, 197

Kalmar, D. A., 194, 197Kaplan, H. B., 17Karns, J., 229Kauffman, S. A., 78, 179Kegan, R., 98

Keith, L. K., 14, 26Kelly, G. A., 36, 138, 143Kelso, J. A. S., 179Kempen, H. J. G., 21–23, 30, 96–99, 102,107, 127, 193

Kendon, A., 102Kernis, M. H., 27, 31Kindermann, T. A., 19Kingsland, S. E., 66Kirker, W. S., 16Kitayama, S., 20, 109, 200Klages, L., 129Klein Wassink, M. E., 224, 226, 229Kohlberg, L., 143Kramer, D. A., 154, 155Kuiper, N. A., 16Kuipers, P., 42Kunda, Z., 31Kunnen, E. S., 169, 178, 224, 226, 228,229

Labouvie-Vief, G., 191Lakoff, G., 97Lalonde, C., 191Lambrecht, L., 85, 92Lamey, A. V., 181, 228Lancelot, C., 4Landauer, R., 181Larsen, R. J., 199Lazarus, R. S., 44, 45, 48, 93, 116l’Ecuyer, R., 14, 25, 26LeDoux, J. E., 162Legerstee, M., 97Leifer, M., 20Lelwica, M., 152Levinas, E., 94, 98LeVine, R. A., 20Levine, R. L., 64Lewis, C. S., 193Lewis, M., 12, 13, 20, 39Lewis, M. D., 4, 5, 8, 35, 87, 103, 107, 109,110, 112, 169, 181–183, 185, 187, 199–212,214, 216, 218, 219, 221–223, 228

Lightfoot, C., 178Lindenberger, U., 18Lipka, R. P., 11, 24, 29Liwag, M., 111Locke, J., 177Loewald, H. W., 100, 101Logan, R. D., 10Lotka, 76Lucariello, J., 191Luscher, K., 19Lutkenhaus, P., 12, 13Lyra, M., 112, 113, 196

Maclean, P., 55

261Author index

Magai, C., 4, 8, 35, 151, 158, 159, 164, 169,172, 183, 192, 197, 200, 201, 209, 210, 212,223

Magnusson, D., 19, 22, 23, 30Mahalingam, R., 191, 192, 194Malatesta, C. Z., 112, 152, 183, 188, 199Malatesta-Magai, C. 152, 163Manion, M., 191Marcia, J. E., 1, 3, 5, 89, 177Markus, H. R., 1, 4, 10, 16, 20–22, 24, 29,31, 51, 200

Marold, D. B., 14Marsh, H. W., 11–13, 22, 24, 25, 30Mascolo, M. F., 183, 196Matteson, D. R., 1Maturana, H. R., 180McAdams, D. P., 129, 178, 191, 194, 195McClelland, D. C., 200McClintic, S., 12McCrae, R. R., 199McFadden, S. H., 4, 151, 159, 169, 201Mead, G. H., 6, 10, 11, 18, 120, 125, 138,147

Meltzoff, A. N., 50Merleau-Ponty, M., 96, 97, 99, 103, 127Mesquita, B., 163Messinger, D., 87Michaels, C. F., 97Michel, G. F., 85, 92, 206, 213, 218Miller, P., 20Miller, P. J., 109Miller, W. I., 45Mischel, W., 178, 182, 185Mistry, J., 109Moen, P., 19Molenaar, P. C. M., 64, 84, 182, 228, 229Moore, M. K., 50Moreno, J. L., 146Moreno, Z. T., 146Morf, C. C., 14, 28Morris, W. N., 43Morrow, J., 186Mortimer, J. T., 20Msihara, B. L., 229Muir, F., 15Mumme, D. L., 12Mummendey, H. D., 20Murray, J. D., 66Murray, K., 175

Neisser, U., 22, 108Nelson, G., 200Nelson, K., 108Nesselroade, J. R., 24Newell, K. M., 64Newman, L. S., 119Nicolis, G., 143

Niedenthal, P. M., 51Nolen-Hoeksema, S., 186Novak, G., 70Nowak, A., 221, 229Nurius, P., 21Nusbaum, B., 182, 183, 197, 201Nuttin, J. M., 12Nwokah, E., 87, 229

Oatley, K., 54, 185Ogilvie, D. M., 17, 21Ohman, A., 161Oosterwegel, A., 1, 11, 12, 16, 22, 23, 25,31, 37, 219

Oppenheim, D., 109Orlofsky, J. L., 1Ormel, J., 199Ortony, A., 56Orwell, L., 191Overton, W. F., 138

Pacini, R., 15, 62, 218Panksepp, J., 162, 187, 188Pedrosa, M. I., 98Perlmutter, M., 219, 224Piaget, J., 48, 108, 125, 143Pickering, J., 97, 100Pipp, S., 20Plato, 177, 193Poincare, 65Polivy, J., 186Port, R. F. 64, 182Pratkanis, A. R., 10–12, 15–17, 29Prigogine, I., 78, 143, 179Prout, H. T., 20Prout, S. M., 20Provinelli, D. J., 108

Rathunde, K., 98, 112Recchia, S., 12Reddy, V., 108, 109Reed, E. S., 97Rhodewalt, F., 14, 28Ricoeur, P., 102, 109Robins, R. W., 177Robinson, M. D., 17, 21 30, 31Rochat, P., 97Rogers, C. R., 37, 112Rogers, T. B., 16Rosch, E., 100Rosenberg, M., 27, 120Rosenberg, S., 21, 154Rothbart, M. K., 182Rozin, P., 45Rubin, D. C., 118, 119Ruetsch, C., 4Ruhland, R., 229

262 Author index

Russek, L. G., 198Rychlak, J. F., 139

Saarni, C., 12, 13, 18, 21, 30Safer, M. A., 119Sagy, S., 112Salovey, P., 35Sameroff, A. J., 22Sampson, E. E., 98, 138Sander, L. W., 101Sansone, C., 19, 21, 30Sarbin, Th. R., 103, 123, 141, 193Sartre, J.-P., 46Savin-Williams, R. C., 20, 22, 27, 30, 31Schaffer, A., 97Schank, R. C., 118, 119Schaufeli, W. B., 199Scheff, T. J., 185Scheier, M. F., 56Scherer, K. R., 41, 45, 185Schittekatte, M., 143Schlenker, B. R., 177Schneider-Rosen, K., 12Schoff, K., 183Schore, A. N., 182, 184, 185, 187, 188,213

Schwartz, G. E., 198Schweder, R. A., 20Sedikides, C., 17Selye, H., 49, 158Sentis, K., 16, 31Shah, A., 44Shavelson, R. J., 11, 12Shaver, P. R., 93Shepard, B., 152Shoda, Y., 177, 178, 182, 185Shotter, J., 97, 118Sigmund, K., 66Simon, B. B., 108Skarda, C. A., 186Skinner, E. A., 19Smart, L., 2, 14Smith, L. B., 64, 98, 179, 181, 182, 184Sokol, B., 191Sonnemans, J., 54, 163Spencer, S., 124Sroufe, L. A., 56, 183Stanger, C., 12Stapley, J., 152Stattin, H., 19, 22, 23, 30Staudinger, U. M., 18Stein, N. L., 54, 111, 185Stengers, E., 78, 179Stern, D., 94, 97, 99, 103, 104, 108, 112Stewart, I., 65Stillwell, A. M., 53Stipek, D. J., 12, 14

Stratton, P., 22Strauman, T. J., 25Strube, M. J., 17Sullivan, J., 92Sullivan, M. W., 12Svrakic, D. M., 182, 229Svrakic, N. M., 182, 229

Tangney, J. P. 12, 14, 18, 21, 30, 51Tappan, M. B., 104Taska, L. S., 20Tcherkassof, A., 43Teasdale, J. D., 185, 186Terschure, E., 42Tesman, J., 152Tesser, A., 2, 221Thatcher, R. W., 85Thelen, E., 64, 98, 179, 181, 182, 184,229

Thompson, E., 100Timmers, M., 117Tomkins, S. S., 103, 161, 182, 183Tong, B. R., 112Trabasso, T., 54, 111, 185Trevarthen, C., 20, 96, 108Triandis, H. C., 20Tucker, D. M., 184, 186, 188Tudor, M., 200Tykocinski, O., 15

Ulrich, B. D., 181, 182, 229

Vallacher, R. R. 221, 229Van der Maas, H. L. J., 84, 182, 229Van der Meulen, M., 6, 7, 26, 33–37, 203Van der Werff, J. J., 6, 12Van Geert, P. L. C., 4, 7, 64, 78, 85, 86, 89,90–92, 181, 182, 185, 197, 205–207,210, 212–214, 222, 223, 228, 229

Van Gelder, T., 64, 182Van Goozen, J., 163Van Halen, C. P. M., 6Van Hooff, J. A. R. A. M, 42Van Langenhove, L., 139Van Loon, R. J. P., 127Varela, F. J., 100, 180Vasil’eva, I. I., 124Vedeler, D., 96, 97Verhofstadt-Deneve, L., 123, 143–146, 209,220, 226, 229

Voloshinov, V. N., 137Volterra, V., 66Vookles, J., 15Vygotsky, L. S., 125, 143

Walker, L. J., 143Wapner, S., 48

263Author index

Waschull, S. B., 27, 31Waterman, A. S., 1Waters, E., 56Watkins, M., 125, 126, 149Watson, J. B., 49Weigold, M. F., 177Weiss, M., 12Wells, A. J., 21, 27, 31Werner, C. M., 143Werner, H., 48White, G. M., 116Whitehead, A. N., 97–99Wicklund, R. A., 1, 37

Wilber, K., 100, 112Wilson, A., 112, 183, 188Winegar, L. T., 113Wolff, W., 12Wood, A. D., 66Wurf, E., 1, 4, 10, 20–22, 24, 29Wylie, R., 10, 26, 202

Yasuo, Y., 97

Zajonc, R. B., 44Zender, B. F., 143Zender, M. A., 143

264 Author index

Subject index

action readiness, 41–43affect, discrete classes of activators,

161–162Albert Ellisanalysis of autobiography, 157–160development of, 158–160

Anne Frankanalysis of diary, 155–157development of, 155–157

appraisal, 44, 58levels of, 45primary and secondary, 44self as object of, 52–53

assessmentof change over time, 227–228of self and identity as self-organizingsystem, 226–230

of the system ‘‘at work,’’ 226–227attractors, 82–83in identity development, 220–221

authorship, 102–107a definition of, 95in parent–infant communication, 104

autobiographical memories, 118–119

becoming and being, 99–102bifurcations, 84

Cardinal Sins, 174–175Cartesian split, 138case study research versus averaging

across cases, 154, 173–174change-based emotions, 162, 164change emotions, 111–112chaos, 83cognition–emotion interactions, 182–190cognition–emotion recursions, 182–190cognitions, implicit versus explicit, 62Cognitive Experiential Self-Theory

(CEST), 58–63experiential and rational system, 59,60

cognitive generativity, 219color model of emotion, 162competence, 56concerns, 54–56, 176and self, 55

conflictbetween I-positions, 144in development, 169in personality development, 142–143

conscious versus non-consciousprocessing, 62

consistency and blocked goals, 188–189,194–195

contextas system, 224–225as ordered effect, 224as random influence, 224in a dynamic systems approach,223–224

coupling of elements, 179

development, a dynamic systemsperspective, 181–182

developmental processeslevels of analysis, 169–170time scales, 173

diachronous perspective, 227dialectical developmental psychology,

143dialogical relationships, 124–125dialogues, 125imaginal, 125

dynamic systemsa definition of, 64a mathematical non-linear model, 72a qualitative approach, 86–88a quantitative approach, 85–86and environment, 68and geometric representations, 67–68and self-organization, 78–81as structure of interactions, 69change across time, 81–82

265

dynamic systems (cont.)mathematical analysis versus modeling,75–78

dynamic systems model of emotionalprocesses, 163–168

emergence of continuity, 179–181emotional dissonance, 117emotional experience, 45–47, 93and sense of self, 45–47as non-linear process, 153as iterative process, 153in adolescence, 152–153

emotional expression in mother–childinteractions, 152

emotional interpretation (EI), 220–221neurobiology of, 186–187real time, 184–186recurrent, 184–186

emotional processescomputer models of, 163–168microlevel versus semanticmeaning-making level, 171

emotional traits, 153–154emotionsa definition of, 41a functional perspective, 41–43and awareness of self, 40, 45–47and cognitions, 60and competence, 56and concerns, 54–56and facial expressions, 92and self-representations, 43–44and subjectivity, 40and the self as subject, 47–51appraisal of (significance), 53as microprocesses, 161–162as organizer of identity, 155, 160–161as subject–object relationship, 41conceptual confusions, 39duration of, 163–164modeling of individual differences,166–168

self as object of, 51–53standards of evaluation, 116the evaluation of, 116

emotions narratives in western culture,112

events, a definition of, 98explicit identity, 190a definition of, 218consistency of, 192–195

feedback loops, 179framesdefinition of, 102

nonverbal, 104

Gedanken experiment, 164

habituation, 219

identitya definition of, 177and autobiographical consistency,118–119

as autobiography, 191as beginning in infancy, 108as cognitive structure, 177as dialogue, 193as embodied being, 117implicit and explicit, 178stability and change, 177–178

identity change, 195–197, 220self-organizing processes, 196–197time scales, 196–197

identity crisis, 175identity development, as dialectical

process, 145–149implicit identity, a definition of, 217I-positions, 127, 132–133, 137–139

level-based emotions, 162, 164logistic map, 83Lotka–Volterra equations, 76

methods of analysis, 175–176moods, 188–189multivoicedness of the self, 126

narrative, 102–103, 123, 191–192narrative emotions, 103–111narrative identity, a definition of, 109

O-motive (contact and union with others),129, 132–133

orientation, as most basic form of self,96

orientational emotions, 111

personal identity, a definition of, 178personality, a definition of, 178phase shift, 170phase transitions, 179, 220Phenomenological–Dialectical Personality

Model, 144–145polyphonic novel, 124principles of self-organizationrecursion, coupling andcomplementarities, 181

processes, 218–223bottom-up and top-down, 221–222

266 Subject index

in implicit and explicit identity,218–223

verbal psychological models, 218–223formal mathematical models, 218–223

psychodrama, 145psychological theoriesas verbal narratives, 91as formal mathematical models, 91

reference emotions, 108–109, 111relational pespective, 94–95roles, development of, 154

selfa definition of, 34and emotions, 60–61, 139–140as object in emotions, 51–53as self-reflecting agency, 121as set of concerns, 55as subject in emotions, 47, 51cultural construction of, 174–175development in infancy and childhood,108

executive, 58–60meta-affective knowledge, 170multivoiced, 127stability and change, 61–62terminological issues, 31, 34

self and emotion, a relational perspective,94–95

self and identityand self-organization, 4–5as cognitive structures, 1–2as multilayered system, 214–216components and emerging phenomena,90, 214–216

definition as a system, 89–90from a dynamic systems perspective,206

fundamental questions, 3, 59–63history of thinking about, 3implications of a dynamic systemsapproach, 213

new perspectives, 4terminological issues, 6the issue of quantification, 91

self-as-knower, 121self-as-known, 121self-awareness, the spectrum of, 100–101self-concepta definition of, 11history of the concept, 10–11measurement methods, 30–31

short-term variability, 25–26life-span developmental change, 24–25stability and variability, 23–29terminological issues, 31the dynamic interaction with context,22–23

the role of context, 18–23, 35–36the role of emotions, 11–18, 35

self-confrontationmethod, 130–131self-esteema definition of, 11stability and fluctuations, 26–28

self-experience, and time scales, 95,110–112

self-evaluation and emotions, 2self-investigation, 131, 136–137self-organization, 5, 78–81, 178–181a definition of, 79and continuity, 178–181

self-organization of personalitythe principle of accrual andcrystallization, 187–189

self-perpetuation, 178–181self-related concepts, a framework, 36–38sense of identity, 107a definition of, 216–217

S-motive (self-enhancement), 129, 132–133stability, differences between implicit and

explicit identity, 199–201systems, lower-order elements and

higher-order forms, 179

time scales, 215, 219–220and levels of organization, 215in emotional self-organization,184–187

in the modeling of developmentalprocesses, 173

time series analysis, 175–176transcendental self, 47empirical contents, 47–48

uniqueness of self, 97

valuation theory, 127–129generalization (G), 129idealization (I), 129

valuations, 128organization and reorganization,130–131

working context, 21working self-concept, 20–21

267Subject index

Studies in Emotion and Social Interaction

First SeriesEditors: Paul Ekman and Klaus R. SchererHandbook of methods in noverbal behavioral researchEdited by Klaus R. Scherer and Paul EkmanStructures of social action: studies in conversational analysisEdited by Max Atkinson and John HeritageInteraction structure and strategyStarkey Duncan, Jr., and Donald W. FiskeBody movement and speech in medical interactionChristian HeathThe EmotionsNico FrijdaConversations of friends: speculations on affective developmentEdited by John M. Gottman and Jeffrey G. ParkerJudgment studies: design, analysis, and meta-analysisRobert RosenthalThe individual, communication, and society: essays in memory of Gregory BatesonEdited by Robert W. RieberLanguage and the politics of emotionEdited by Catherine Lutz and Lila Abu-LughodFundamentals in nonverbal behaviorEdited by Robert Feldman and Bernard RimeGestures and speechPierre J. M. Feyereisen and Jacques-Dominique de LannoyLandscapes of emotion: mapping three cultures of emotion in IndonesiaKarl G. HeiderContexts of accommodation: developments in applied sociolinguisticsHoward Giles, Justine Coupland, and Nikolas CouplandBest laid schemes: the psychology of emotionsKeith OatleyInterpersonal expectations: theory, research, and applicationsEdited by Peter David BlanckEmotional contagionElaine Hatfield, John T. Cacioppo, and Richard L. Rapson

Exploring affect: the selected writings of Silvan S. TomkinsEdited by E. Virginia Demos

[Harke a. Bosma, E. Saskia Kunnen] Identity and Em(Bookos.org)

Documents

Transcript of [Harke a. Bosma, E. Saskia Kunnen] Identity and Em(Bookos.org)