Advances in Child Development and Behavior Volume 28

ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR

Volume 28

Contributors to This Volume

Shelli Avenevoli

Steven C. Hayes

Dermot Barnes-Holmes

V. E Reyna

Yvonne Barnes-Holmes

Bryan Roche

C. J. B rainerd

Karl S. Rosengren

Gregory S. Braswell

Laurence Steinberg

Michael J. Chandler

ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR edited by Hayne W. Reese Department of Psychology West Virginia University Morgantown, West Virginia

Robert Kail Department of Psychological Studies Purdue University West Lafayette, Indiana

Volume 28

ACADEMIC PRESS A Harcourt Scienceand TechnologyCompany

,San Diego ,San Francisco New York

Boston London Sydney Tokyo

This book is printed on acid-free paper.

Copyright © 2001 by ACADEMIC PRESS All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher's consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-2001 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0065-2407/2001 $35.00 Explicit permission from Academic Press is not required to reproduce a maximum of two figures or tables from an Academic Press article in another scientific or research publication provided that the material has not been credited to another source and that full credit to the Academic Press article is given.

Academic Press A Harcourt Science and Technology Company 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.academicpress.com

Academic Press Harcourt Place, 32 Jamestown Road, London NW1 7BY, UK http://www.academicpress.com International Standard Book Number: 0-12-009728-1 PRINTED IN THE UNITED STATES OF AMERICA 01 02 03 04 05 06 SB 9 8 7 6 5 4

3

2

1

Contents Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Variability in Children's Reasoning K A R L S. R O S E N G R E N A N D G R E G O R Y S. B R A S W E L L I. II. III. IV. V. VI.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Assessing the Importance of Variability in Children's Thinking . . . . . . . . . . . . . . . . . . . . . . Types of Variability in Children's Reasoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theories and the Meaning of Variability in Children's Reasoning . . . . . . . . . . . . . . . . . . . .

2 3 8 16

Methodological Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35

Fuzzy-Trace Theory: Dual Processes in Memory, Reasoning, and Cognitive Neuroscience C. J. B R A I N E R D A N D V. F. R E Y N A I. II. III. IV. V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M e m o r y - R e a s o n i n g Relations: Necessity, Constructivism, and Dissociation . . . . . . . . Dual-Process Conceptions of Reasoning, Memory, and M e m o r y - R e a s o n i n g Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Four Research Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

42 43 47 65 89 91

Relational Frame Theory: A Post-Skinnerian Account of Human Language and Cognition Y V O N N E B A R N E S - H O L M E S , S T E V E N C. H A Y E S , DERMOT BARNES-HOLMES, AND BRYAN ROCHE I. II. III.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

101

Relational Frame Theory as an Approach to Language and Cognition . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

103 135

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

136

Contents

vi

The Continuity of Depression across the Adolescent Transition SHELLI AVENEVOLI AND L A U R E N C E STEINBERG I. II. III. IV. V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Disorder, Syndrome, and Mood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stability and Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Genetic, Biological, Cognitive, and Social Influences on Continuity . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

140 142 148 158 163 165

The 77me of Our Lives: Self-Continuity in Native and Non-native Youth M I C H A E L J. C H A N D L E R I.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

176

Narrativity and Essentialism Writ Large . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

177

III.

The Essentialist Tradition: Strict Essentialism, Nominal Essentialism, Neo-essentialism, and Any Other Essentialism You Can Think Of . . . . . . . . . . . . . . . . . . .

182

IV.

The Narrative Tradition as Another Singular Possibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

190

A Theory in Search of Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Real Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of Conclusions: Five Easy Pieces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

197 205 215 215

II.

V. VI. VII.

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

223

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

233

Contents of Previous Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

239

Contributors Numbers in parentheses indicate the pages on which the authors' contributions begin. SHELLI AVENEVOLI

Department of Epidemiology and Public Health, Yale University School oj Medicine, New Haven, Connecticut 06510 (139) DERMOT BARNES-HOLMES

Department of Psychology, National University of Ireland, Maynooth, Kildare, Ireland (101) YVONNE BARNES-HOLMES

Department of Psychology, National University of Ireland, Maynooth, Kildare, Ireland (101) C. J. B R A I N E R D

Department of Special Education, Rehabilitation, and School Psychology, University of Arizona, Tucson, Arizona 85721 (41) G R E G O R Y S. B R A S W E L L

Department of Psychology, University of California-Santa Cruz, Santa Cruz, California 95064 (1) M I C H A E L J. C H A N D L E R

Department of Psychology, The University of British Columbia, Vancouver, British Columbia V6T 1ZA, Canada (175) S T E V E N C. H A Y E S

University of Nevada at Reno, Reno, Nevada 89557 (101) V. E R E Y N A

Departments of Surgery and Medicine, Arizona Health Sciences Center, University of Arizona, Tucson, Arizona 85721 (41) BRYAN R O C H E

Department of Psychology, National University of Ireland, Maynooth, Kildare, Ireland (101) K A R L S. R O S E N G R E N

Department of Psychology, University of Illinois at Urbana-Champaign, Champaign, Illinois 61820 (1) LAURENCE STEINBERG

Department of Psychology, Temple University, Philadelphia, Pennsylvania 19122 (139)

vii

This Page Intentionally Left Blank

Preface The amount of research and theoretical discussion in the field of child development and behavior is so vast that researchers, instructors, and students are confronted with a formidable task in keeping abreast of new developments within their areas of specialization through the use of primary sources, as well as being knowledgeable in areas peripheral to their primary focus of interest. Moreover, journal space is often simply too limited to permit publication of more speculative kinds of analyses that might spark expanded interest in a problem area or stimulate new modes of attack on a problem. The serial publication Advances in Child Development and Behavior is intended to ease the burden by providing scholarly technical articles serving as reference material and by providing a place for publication of scholarly speculation. In these documented critical reviews, recent advances in the field are summarized and integrated, complexities are exposed, and fresh viewpoints are offered. These reviews should be useful not only to the expert in the area but also to the general reader. The series is not intended to reflect the development of new fads, and no attempt is made to organize each volume around a particular theme or topic. Manuscripts are solicited from investigators conducting programmatic work on problems of current and significant interest. The editors often encourage the preparation of critical syntheses dealing intensively with topics of relatively narrow scope but of considerable potential interest to the scientific community. Contributors are encouraged to criticize, integrate, and stimulate, but always within a framework of high scholarship. Although appearance in the volumes is ordinarily by invitation, unsolicited manuscripts will be accepted for review. All papersmwhether invited or submitted--receive careful editorial scrutiny. Invited papers are automatically accepted for publication in principle, but usually require revision before final acceptance. Submitted papers receive the same treatment except that they are not automatically accepted for publication and may be rejected. The Advances series is generally not a suitable place of publication for reports of a single study or a short series of studies, even if the report is necessarily long because of the nature of the research. The use of sexist language, such as "he" or "she" as the general singular pronoun, is not acceptable in contributions to the Advances series; the use of "they"as a singular pronoun is incorrect. Use of the phrase "he or she" (or the like) is acceptable. We acknowledge with gratitude the aid of our home institutions, West Virginia University and Purdue University, which generously provided time and facilities

x

Preface

for the preparation of this volume. We also thank Judy Garber and Karen Rudolph for their editorial assistance and Ann Davis at West Virginia University for her excellent secretarial services. Hayne W. Reese Robert Kail

VARIABILITY IN CHILDREN'S REASONING

Karl S. Rosengren DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN CHAMPAIGN, ILLINOIS 61820

Gregory S. Braswell DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF CALIFORNIA-SANTA CRUZ SANTA CRUZ, CALIFORNIA 95064

I. INTRODUCTION II. ASSESSING THE IMPORTANCE OF VARIABILITY IN CHILDREN'S THINKING A. THE STABILITY OF COGNITIVE STRUCTURES B. INTER- AND INTRAINDIVIDUAL VARIABILITY C. D E V E L O P M E N T VERSUS L E A R N I N G D. NEGATIVE ASPECTS OF VARIABILITY E. POSITIVE ASPECTS OF VARIABILITY III. TYPES OF VARIABILITY IN CHILDREN'S REASONING A. VARIABILITY ACROSS AGE B. VARIABILITY ACROSS CHILDREN OF THE SAME AGE C. VARIABILITY IN A CHILD'S P E R F O R M A N C E ACROSS RELATED TASKS D. VARIABILITY WITHIN A CHILD OVER REPEATED TRIALS E. VARIABILITY WITHIN A CHILD WITHIN A SINGLE TRIAL IV. THEORIES AND THE M E A N I N G OF VARIABILITY IN CHILDREN'S R E A S O N I N G A. PIAGETIAN VIEW OF VARIABILITY B. S O C I O C U L T U R A L VIEW OF VARIABILITY C. D O M A I N SPECIFICITY VIEW OF VARIABILITY D. I N F O R M A T I O N PROCESSING VIEWS OF VARIABILITY E. DYNAMIC SYSTEMS VIEWS OF VARIABILITY V. M E T H O D O L O G I C A L ISSUES A. ASSESSING R A N D O M N E S S IN C H I L D R E N ' S RESPONSES B. ISSUES OBSCURING VARIABILITY IN C H I L D R E N ' S REASONING

ADVANCESIN CHILDDEVELOPMENT AND BEHAVIOR,VOL.28

Copyright 9 2001 by AcademicPress. All rights of reproduction in any form reserved. 0065-2407/01 $35.00

2

Karl S. Rosengren and Gregory S. Braswell

C. HOW TO EXAMINE VARIABILITY IN C H I L D R E N ' S THINKING VI. CONCLUSIONS REFERENCES

I. Introduction How should we best characterize the development of children's thinking? Is children's thinking highly stable within any period of development? Or does children's thinking have a more fleeting or variable quality, changing from one moment to the next, one task to another? Traditionally, children's thinking has been described in terms of global properties, stages, or modes of reasoning with Piaget's theory of cognitive development serving as the prototype for this kind of model of children's reasoning. From this view, the development of children's thinking is characterized by periods of relative stability followed by periods of rapid, dramatic change. For the most part, this view of children's thinking was characterized by the uniformity of cognitive structures at a given point in the developmental process. That is, a child's thinking was assumed to be relatively stable across time (except for periods of developmental change) and across similar tasks. This view of the stability of reasoning has been maintained even though much of Piaget's theory has come under attack. Indeed, Siegler (1994, 1996) has noted that much of the research in cognitive development has characterized children's thinking at any particular age in terms of relatively static statesmregardless of whether these static states are used to refer to children's mental models (Vosniadou & Brewer, 1992, 1994), naive theories (Carey, 1985; Wellman, 1990), symbolic representations (DeLoache, 1995), or strategy use (Siegler, 1996). One reason that acceptance of Piaget's theory has generally declined is the finding that children's thinking is more highly variable than his theory predicts. Specifically, research in the 1980s and 1990s revealed an exceptional amount of variability across different domains of reasoning. For example, a child's performance on some sort of animal categorization task (which taps into the domain of biology) is unlikely to have much relation with his or her performance on various other cognitive tasks, such as a false-belief task (which taps into the domain of psychology) or a balance-scale task (which taps into the domain of physics). This variability in thinking across tasks, often characterized as "domain specificity," is merely one way that children's thinking has been characterized as variable (Wellman & Gelman, 1998). However, despite seemingly widespread agreement that the nature and quality of children's thinking do in fact vary by domain, a dominant view continues to hold that children's thinking is relatively coherent and

Variability in Children's Reasoning

3

consistent within a particular domain. For example, children's reasoning about beliefs and desires is believed neither to vary from one day to the next nor to vary across related questions. Rather, the belief is that at a particular age, children have a relatively coherent theory-of-mind that guides their behavior and enables them to interpret the behavior of others in a relatively consistent manner (Wellman, 1990). In this chapter we go beyond issues of domain specificity to explore more broadly the issue of variability in children's thinking. Specifically, we take a new look at variability in children's thinking, drawing in part from the work of Siegler (1994, 1996), who has written extensively on variability in children's use of counting strategies, and of Thelen and her colleagues (Smith & Thelen, 1993; Thelen & Smith, 1994), who have examined issues of variability from a dynamic systems perspective. In this chapter we focus on children's cognitive processes generally rather than examine a particular aspect of cognitive development. We start by examining different broad views of variability in psychology. We then examine various ways in which children's thinking might be found to varymdrawing to a large extent on the work of Siegler (1994, 1996) for the general framework of this section. We expand on his framework by explicitly examining how different theoretical viewpoints construe variability in children's thinking. In particular we examine how variability is seen from a traditional Piagetian view, a sociocultural view, a domain specific view, an information processing view, and a dynamic systems view. The point of this exercise is to evaluate these views with respect to how they address the issue of variability in children's reasoning. We suggest that in most of these theoretical views the issue of variability in children's thinking has often been seen as a side issue, obfuscating the true nature of development. We argue, however, that the time has come to take a more serious look at variability in children's thinking, focusing on both the meaning and the causes of variability. Part of this task involves determining when variability is meaningful and when it is merely random noise. To this end, in the section following our theoretical review we examine the issues of assessing the nature of variability. Within this section we discuss ways of determining whether the variability we observe is meaningful or not. A central theme of this chapter is the idea that researchers in the field of cognitive development should embrace variability as an important and integral aspect of children's thinking. Thus, a primary goal of this chapter is to place the study of variability in children's cognition more centrally in the field of cognitive development.

II. Assessing the Importance of Variability in Children's Thinking Recognizing the importance of variability in the study of cognitive development, or in psychology as a whole for that matter, is not an entirely new phenomenon. Throughout the 20th century, one of two opposing viewpoints within the discipline

4

Karl S. Rosengren and Gregory S. Braswell

of psychology focused on the static and universal and the other focused on the dynamic and diverse (Cronbach, 1957; Valsiner, 1984). The first viewpoint, referred to as the "typological mode of thinking" by Valsiner (1984) and the "experimental" mode by Cronbach (1957), has dominated psychological inquiry and has driven researchers to consider behavior and its causes in static and decontextualized terms. The primary aim of researchers adhering to this view is to minimize variability, which is usually dismissed as error or noise (Howe, 1994). This view of variability is also seen in much of the research in cognitive development. A. THE STABILITY OF COGNITIVE STRUCTURES

Inherent in many past and present conceptions of children's cognition is the assumption that underlying cognitive structures (naive theories, schemes, representations, etc.) are relatively stable over most intervals of development. That is, a child is assumed to use the same method, strategy, naive theory, and so on, to solve a particular kind of problem, or a set of related problems, during any given developmental period. One of the main assumptions of adherents of this view seems to be that the only kind of variability of much interest is variability attributable to the developmental process itself. From this perspective, variability attributable to characteristics of specific tasks is uninterestingmfor instance, a problem with methodology or experimental design or a problem of"experimental noise" around a relatively stable "signal" being obtained from the child. Indeed, variability attributable to task constraints and other "nondevelopmental" factors is often ignored or dismissed by researchers who instead focus on trying to capture what in their view is the central core or "essence" of development. Again, this failure to examine variability in thinking stems in part from the assumption that development occurs in an orderly, uniform progression from one relatively stable state of reasoning to another more sophisticated, but still relatively stable state. In many ways variability in performance has been viewed as a bad thing, something to eliminate from our observations or experiments, rather than as an object of study in its own right. As Cronbach (1957) wrote, "whatever your device, your goal in the experimental tradition is to get those embarrassing differential variables out of sight" (p. 674). Thus, if a child's performance is found to vary from day to day, or across related questions, the researcher often examines the methods, design, or task for flaws, assuming that the methods used are not capturing the true nature of the child. Another approach to removing this variability is to examine the data in terms of averages (Valsiner, 1984). By averaging, the observed phenomena are reduced to an abstracted state in which variability manifested within or between individuals is eliminated. The problem with this approach is that the real world is diverse and ever changing. Researchers are forced to build theories from their abstractions, which are inherently limited after being stripped of any initially observed variations (Valsiner,

Variability in Children's Reasoning

5

1984). According to Valsiner, the predominance of this view in psychology is derived largely from psychologists' desire to mimic the natural sciences, particularly physics, perhaps to gain credibility as a science. For instance, traditional theories of physics deal with idealized states and laws of nature. A dilemma arises, however, because psychological systemsmand biological onesmare "open," unlike physical systems, in that they are linked with environments that can affect them. Furthermore, physics deals with "homogeneous classes" in which members (e.g., electrons or photons) are highly similar and differences c a n be considered error; yet psychology and biology deal with "heterogeneous classes" in which members are quite different from one another because of the nature of open systems. The desire of psychologists to mimic physics in the pursuit of universal laws is accordingly not useful because physical and phychological systems are fundamentally dissimilar (Valsiner, 1984). B. INTER- AND INTRAINDIVIDUAL VARIABILITY

The alternative to the "experimental" mode discussed previously is known as the "variational" (Valsiner, 1984)or "correlational" (Cronbach, 1957)mode of thought, which emphasizes observations of inter- and intraindividual variability. As Cronbach (1957) wrote: "The correlational psychologist is in love with just those variables the experimenter left home to forget. He regards individual and group variations as important effects of biological and social causes" (p. 674). Researchers adhering to this paradigm shun the decontextualizing tendency of the typological paradigm and acknowledge the importance of interactions between individuals and the environmental contexts in which they behave. Another shortcoming of the typological mode, noted by Valsiner (1984), is that it is limited in its capacity to predict possible behavior because it is based on observations of behavior that have already happened. In contrast, the variational mode takes into consideration a range of possible behaviors that may be more predictive (i.e., more possibilities are acknowledged). The variational mode is particularly useful for developmental psychology in that its emphasis on dynamic processes may be more appropriate than a focus on static phenomena for understanding mechanisms of change (Valsiner, 1984). A number of individuals, most notably Siegler (1994, 1996) and Thelen and her colleagues (Smith & Thelen, 1993; Thelen & Smith, 1994) have begun to embrace this more variational mode of analysis, calling for an increased attention to issues of variability in children's thinking. Siegler (1994), for example, has argued that variability plays a crucial role in promoting cognitive growth and change. Likewise from a dynamic systems perspective, Thelen and Smith (1994) have argued that variability is a fundamental characteristic of complex systems. Indeed, numerous studies of highly complex systems, such as human postural control (Newell, van Emmerik, Lee, & Sprague, 1993) and a variety of human

6

Karl S. Rosengren and Gregory S. Braswell

physiological systems (Goldberger, Rigney, & West, 1990), suggest that healthy biological systems are typically more variable (i.e., more chaotic) than diseased ones. Almost surely, the human cognitive system (even that of a very young child) is a complex biological system. Thus, one might expect variability to be an inherent characteristic of cognition and its development, just as it is for complex biological systems in general. C. DEVELOPMENT VERSUS LEARNING

At another level of analysis, variability has often been dismissed as an aspect of learning and not development. Over the history of developmental psychology, many researchers have attempted to draw a sharp distinction between these two processes. For example, Piaget (cited in Ginsburg & Opper, 1988) viewed development as something that is endogenous to the child. Typically, development is characterized by such terms as universal, general, orderly, progressive, and unidirectional. Development has also been thought to describe changes that occur over a relatively long period of time and that are for the most part context independent. Thus, development has traditionally not been viewed as influenced by various task constraints. In some ways this characterization of development is synonymous with the concept of "maturation." The latter term is often thought to involve the genetic unfolding of various processes relatively independent of environmental inputs. In contrast to this traditional view of development, learning--in the narrow sense used by Piaget--has often been characterized as driven by forces exogenous to the child, as highly specific, and as greatly influenced by instructions and/or task constraints. Context and experiential effects have been more strongly associated with this description of learning than with development. Thus, learning, due to its context sensitivity and dependence on experience, is thought to lead to great variability across children of the same age, and even within a particular child when he or she is confronted with situations that draw on different knowledge bases. From this view, at any given point in time the child's developmental capabilities are viewed as highly stable and their learning capabilities are viewed as more unstable and variable. Many developmental psychologists have viewed learning in this sense as uninteresting and in the realm of educational researchers, saving the "pure" study of development for themselves. Yet, as Siegler (1996) has argued "Without... detailed mechanistic accounts, issues regarding the relation of learning and development are impossible to resolve (as the roundand-round nature of the learning/development debate so well illustrates)" (p. 22). We take a slightly stronger view and suggest that the dichotomy of learning versus development, though potentially of interest theoretically, is of little value in terms of describing how a child's thinking changes as a function of age and experience. That is, unless one adopts a very narrow and somewhat arbitrary view

Variability in Children's Reasoning

7

of learning, learning processes cannot be completely separated from developmental ones. Unfortunately, many developmentalists have attributed to learning all processes that occur over relatively short time periods or are highly context sensitivemand thus in the domain of educators rather than developmental psychologists. Like the learning-development debate we suggest this view is also misguided. D. NEGATIVE ASPECTS OF VARIABILITY

Of course, variability is not always a desirable trait in cognition, regardless of one's age. One disadvantage of variability is that it may lead to inefficient or ineffective performance. In some cases stereotyped behavior is appropriate and more effective than more variable behavior. For example, when the environment is stable a person may have no real need to display or even consider various alternative strategies because one particular strategy is all that is necessary for successful attainment of a goal (Valsiner, 1984). In addition, Goldin-Meadow, Alibali, and Church (1993) suggested that fewer mental resources are available to perform other tasks when the child is activating multiple strategies rather than considering only a single one. In their view, transitional knowledge states, which are thought to be characterized by high degrees of variability, are assumed to tax the child's cognitive abilities. E. POSITIVE ASPECTS OF VARIABILITY

Instead of viewing variability as undesirable noise, or merely due to uninteresting task effects, many researchers and theorists have begun to think that variability may serve a number of positive functions in cognitive development. These functions are quite varied and include (a) variability as an important indicator of transition states (Alibali & Goldin-Meadow, 1993), (b) variability as an important source of information about the environment to the child him- or herself, (c) variability as a source of information about the interaction between the child and the environment (Thelen & Smith, 1994), (d) variability as a driving force of development itself (Smith & Thelen, 1993), (e) variability as the origin of new adaptive behavior of any biological organism (Thelen & Smith, 1994), and (f) variability as an indicator that a child has an entire repertoire of strategies to draw from at any given time (Siegler, 1994, 1996). In many ways, these newer perspectives put variability in a more positive light, interpreting variability in reasoning ability as crucial to the child (and adult) in that it enhances one's flexibility and adaptive potential (Thelen & Smith, 1994). Just as in an evolutionary perspective where genetic variability allows species to adapt to changing environments, cognitive variability is seen as allowing children to adapt to the diverse environments they encounter (Siegler, 1996; Siegler & Ellis,

8

Karl S. Rosengren and Gregory S. Braswell

1996; Valsiner, 1984). A brief example is that in problem solving, certain strategies may be more appropriate than others in various contexts. Some strategies may be more beneficial when speed is important; other strategies may be more beneficial when accuracy is important (Siegler, 1996). The ability to draw from a repertoire of strategies thus may enable persons to maximize their performance in any given situation. As mentioned earlier, variability may also serve as an indicator of the individual's exploration of his or her environment. That is, when children are first faced with a novel task or situation, they may benefit from exploring the problem space. In studies by Kuhn (1995) and Schauble (1990), for example, children demonstrated high degrees of variability in strategy use as they explored computerized microworlds to discover the rules that operated in those artifical systems. Again, variability in this context is a good thing and enhances the adaptive capabilities of children and adults. An important point in this case is that variability may not lead to improved performance at the initial stages of attempting to solve the problem at hand because time is needed to explore the task environment in any useful way. However, this period of exploration is likely to lead to improvements in problem solving at some later point in time because the exploration is likely to lead to greater knowledge of the task and situation constraints that are likely to influence performance. This general discussion is meant to highlight the importance of theory in constraining the practice of developmental research. In many ways it is easy to say that children's thinking is variable, but what does this statement really mean? The meaning of variability, and especially variability in children's thinking, is tightly bound to one's particular theoretical and metatheoretical perspective (Overton, 1998). Notions of stability and variability are highly intertwined with assumptions about the nature of children's thinking and how it changes over the course of children's lives. Is children's thinking relatively coherent and consistent, stemming from stable underlying representations, or is children's thinking highly variable, greatly influenced by situation and task constraints? 1 We return to this discussion of different theoretical orientations with respect to the concept of variability in children's reasoning in a later section. But first, we examine various types of variability that may be found in children's thinking and its development.

III. Types of Variability in Children's Reasoning A child's thinking can be said to vary in many different ways and we are certainly not the first to address this issue. For example, one argument has been that children's 1Althoughwe only present two extremeviewshere, we acknowledgethat there are manyother ways of conceptualizing children's thinking in terms of issues of stability and variability.

Variability in Children's Reasoning

9

thinking is by nature heterogeneous, heavily influenced by context (DeLoache, Miller, & Pierroutsakos, 1998). Howe (1994) has also suggested that variability may in fact be a constant in development. In this section we examine different types of variability drawing on Siegler's (1994, 1996) discussion of variability in children's thinking. He cited five different ways in which children's thinking can vary. These include variability (a) across children of different ages, (b) across children of the same age, (c) within a child solving related problems, (d) within a child solving the same problem repeatedly, and (e) within a child within a single trial. In this section we examine each of these types of variability in turn, give some examples, and provide some analysis of the attention that each type has received from developmentalists. A. VARIABILITY ACROSS AGE

Variability across age has been the primary focus of most of the research in cognitive development. Indeed, mapping out differences in various cognitive functions across children of different ages might be viewed as the main focus of cognitive developmental research. Although describing developmental changes in children's cognition is obviously a necessary and important step in creating a comprehensive theory of cognitive development, as Siegler (1994) has pointed out, this focus does not often lead to a greater understanding of the process leading to the developmental differences obtained. Indeed, Siegler argued that "the emphasis in almost all cognitive developmental theories has been on identifying sequences of one-to-one correspondences between ages and ways of thinking or acting rather than on specifying how the changes occur" (p. 1). However, given the relative ease of collecting data with cross-sectional designs compared to longitudinal or microgenetic designs the cataloging of age differences in various cognitive functions is likely to continue for some time to come. Given that the major developmental changes in most areas of cognitive functioning can be found in any developmental psychology textbook, we will not spend any time cataloging the tremendous amounts of variability that occur as part of the acquisition of various developmental milestones in children's cognitive development. This approach of mapping out age-related changes in children's thinking has often led to somewhat heated debates about the specific age at which some cognitive capability is possessed by a child. For example, much research on the development of children's theory-of-mind has been focused on whether children 3g years old truly understand the concept of belief or whether this understanding does not develop until age 4 or 4g (Lillard & Flavell, 1992" Perner, 1989; Perner & Davies, 1991; Sodian, Taylor, Harris, & Perner, 1991; Wellman & Bartsch, 1988, 1989). Likewise, the domain of children's biological understanding contains much debate about whether biological reasoning is an early or later developing ability (Carey, 1985; Keil, 1989; Rosengren, Gelman, kalish, & McCormick, 1991; Solomon,

10

Karl S. Rosengren and Gregory S. Braswell

Johnson, Zaitchik, & Carey 1996; Springer & Keil, 1989, 1991). Although pinning down transition points in the developmental process is clearly important, the primary focus on age, with the implied underlying assumption that age is a valid metric of universal cognitive development, leads to at least three problems. First, this focus on age differences leads researchers to look for uniformity when variability may be much more the norm. Related to our earlier discussion, researchers focusing on age often attempt to categorize children's thinking into a coherent whole, assuming underlying stable structures (Reese & Overton, 1970). Second, the focus on age differences, as mentioned previously, shifts the focus away from the actual p r o c e s s of development and minimizes the role of context and experience. Again, this position reflects the assumption that researchers are capturing some purely developmental phenomenon that results from the child's own cognitive processes. A third problem, related to the second one, is that the focus on age-related changes tends to undervalue the complexity of cognitive development. 2 That is, if a cognitive ability, such as reasoning about the thoughts, desires, and intentions of others, develops in a relatively uniform manner regardless of the experiences an individual is confronted with, then the ability seems not altogether that complex. For example, if theory-of-mind reasoning is universal across cultures regardless of the nature of inputs, then only relatively simple, impoverished data and experiences are necessary for its development. This is clearly not the case. But in our view the focus on age differences and the debates over the specific age of acquisition of various cognitive abilities, fail to acknowledge the complexity of these achievements and of cognitive development more generally. Thus, although we should not stop collecting data on age-related changes in children's cognition, we need to focus on collecting data that enable us to investigate the processes of change and acknowledge that cognitive development involves a complex interaction between a maturing child and a complex and interesting environment. Debates about exactly what age children can be said to have a true understanding of a certain concept might be at least partially resolved by discarding the assumption of uniform stages of cognitive development.

B. VARIABILITYACROSS CHILDRENOF THE SAME AGE As any classroom teacher will acknowledge, we can find variability in the thinking of different children of the same age. This type of variability is often categorized under the heading of "individual differences" (Howe, 1994). This type of variability has received much less attention from cognitive developmentalists than 2We acknowledgethat the complexityof cognitive developmentdepends on one's worldview.In our view it is most sensible to characterizecognition and its developmentas complex and multifaceted.

Variability in Children's Reasoning

11

age-related variability. This relative lack of attention probably reflects the paradigmatic issues described earlier and the general notion that individual differences stem from learning and context effects and hence are less interesting to researchers studying cognitive development. More researchers in cognitive development have begun to focus thier attention on this type of variability, following the leads of Siegler (1994, 1996) and Thelen and Smith (1994). For example, researchers have begun to explore variability in reasoning by same-aged children in the biological domain (Hatano et al., 1993; Nguyen & Rosengren, 2001; Rosengren & Hickling, 2000; Rosengren, DeHart, Taylor, & Marzolf, 2001), in representational abilities (DeLoache et al., 1998), and in naive psychology (e.g., Lillard, 1998, 1999; Perner, Ruffman, & Leekman, 1994; Ruffman, Perner, Naito, Parkin, & Clements, 1998; Taylor & Carlson, 2000). Other researchers have begun to explore this type of variability within specific cognitive tasks. For example, performance on false-belief tasks has been found to be significantly influenced by family size. Specifically, children from larger families exhibit more advanced reasoning on theory-of-mind tasks than children with only one sibling (Perner et al., 1994). However, the benefits attributed to family size may stem primarily from influences by older rather than younger siblings (Ruffman et al., 1998). Other research on individual differences in theory-of-mind tasks, conducted by Taylor and Carlson (1997), showed that children with imaginary companions tend to reason at a more sophisticated level on these tasks than their peers without imaginary companions. Acknowledgment of the richness of this type of variability has led to the development of more comprehensive models of the development of theory-of-mind (TOM) reasoning (Lillard, 1998, 1999). Lillard (1998, 1999), for example, has studied the influences of cultural factors on the development of TOM reasoning. She argued that a variety of factors, including ontogenetic processes, the ability to introspect and analogize, and specific cultural experiences, influence the rate at which children acquire a TOM and influence the adult form of TOM. One might place cross-cultural differences in children's thinking within the general category of variability under discussion in this section. The initial focus of many cross-cultural studies seems to have been a search for universals of cognition and development. For example, research on the development of theory-of-mind reasoning in different cultures initially revealed more similarities than differences across cultures with same-aged individuals (e.g., Avis & Harris, 1991). In cases where cross-cultural differences have been found in the thinking of same-aged children, this variability has often been treated as less interesting than the cultural universals uncovered (e,g., Hatano et al., 1993). Specifically, in many instances very little emphasis has been placed on uncovering the nature of the cultural differences. For example, Hatano and his colleagues reported that children in Israel, Japan, and the United States were highly similar and accurate in their attribution

12

Karl S. Rosengren and Gregory S. Braswell

of many biological properties to humans. In this investigation, however, Israeli children, in contrast to Japanese and American children, were less likely to attribute to plants characteristics that are typical of all living things. Japanese children, in contrast to children from the other two countries, were more likely to attribute life properties to inanimate objects. Although these researchers attempted to provide some possible explanations for these cultural variations, neither they nor other researchers have explicitly investigated potential sources of this variability. Rather, an implicit goal of much of the cross-cultural research on cognitive development seems to have been to focus on cultural universalsmthose things at the "heart" of cognition and its development. An implicit assumption of this type of cross-cultural research is the notion that some type of cognitive constraint underlies the acquisition of various cognitive abilities (Atran, 1995; Hatano & Inagaki, 1996). In some instances, this assumption was explicitly stated. For example, Hatano and Inagaki (1996) suggested that underlying cognitive constraints lead to the acquisition of the core aspects of a naive biological theory. They also stated, however, that the exact manner in which naive biology becomes instantiated varies as a function of specific activities and experiences obtained in a particular culture. Much of the cross-cultural research on children's cognitive development has been fueled by the search for these underlying cognitive constraints that lead to cultural universals. This emphasis often leads researchers to downplay or ignore factors that lead to variability. For some time, researchers have also been investigating sources of variability in the area of cultural variations in children's academic performance, most often math and science performance (e.g., Stevenson et al., 1990). This focus stems in part from the well-established finding that children in the United States lag behind many of their same-aged peers in other countries (especially Asian ones) in many areas of academic performance (especially math and science). Within this domain of inquiry, the major focus has been on possible causes of crosscultural variability, with some researchers focusing on social and motivational factors and others examining the influence of language structure on cognitive development. For example, Stevenson and his colleagues (Chen & Stevenson, 1995; Stevenson, et al., 1990) found that one potential reason for the poorer performance of U.S. children compared to their counterparts in Japan and Taiwan is that parents in the latter two countries tend to be more involved and interested in their children's academic achievement. Typical Asian parents assume that effort, hard work, and practice will lead to success and therefore they emphasize effort more than ability. In contrast, typical American parents assume that ability is the key to success and therefore they emphasize ability training more than effort. Perhaps the most important point in these reports is that the source of variability is often viewed as external to the child. That is, under the surface of these studies seems to lurk the idea that all children have the potential to develop in the same manner as long as the culture provides the

Variability in Children's Reasoning

13

right information (i.e., the parents and the culture support the developmental process). A different perspective has been taken by Miller and his colleagues (Miller, Smith, Zhu, & Zhang, 1995), who have explored variability in number counting in same-aged children across a variety of cultures. They suggested that the observed cross-cultural variability in counting stems from an interaction between the child and the cultural symbols he or she is attempting to acquire. For example, they have found that children in the United States take longer to learn to use counting numbers and have greater difficulty in getting past the teens (numbers 11-19) than their same-aged peers in many Asian countries. Miller et al. suggested that this difference reflects the structure of the languages; specifically, counting numbers in English (as well as in most Western languages) are less regular than those in Chinese and other Asian languages. For example, in Chinese the teens are rendered as "ten-one," "ten-two," and so on. The irregular structure of the numbers in Western languages creates difficulties for children to discover the underlying symmetry of the base 10 counting system. Thus, the child's cognitive processes in the search for regularities interacts with the culturally imposed counting system and this interaction leads to variations in the cognitive performance across cultures. As can be seen, researchers are apparently becoming more interested in examining factors that influence variability in the reasoning of children of the same age, whether this variability is due to cultural practices, interactions between the child and symbol system, factors relating to the children, or some other series of factors. The majority of this research still, however, seems to be focused on factors that influence differences among groups of children rather than on describing the factors that influence the thinking of individual children. Thus, relatively few investigators examine factors that influence the cognitive development of particular children (an exception to this is research by Taylor & Carlson, 2000)--or provide us with predictive power relating to the cognitive outcomes of particular children. C. VARIABILITY IN A CHILD'S P E R F O R M A N C E ACROSS RELATED TASKS

The third type of variability described by Siegler (1994, 1996) captures the classic phenomenon that Piaget called "horizontal decalage." This type of variability is characterized by different levels of performance on tasks that are thought to require the same level of cognitive processing. For example, children solve certain conservation problems prior to others. From a traditional view this decalage should not occur because all conservation problems require the same sort of reasoning. We will explore this type of variability in more detail when we discuss the traditional Piagetian view of variability in section IV.A. Another way this type of variability might be construed is in terms of domain specificity. That is, children are often able to solve some problems that involve a certain type of knowledge, yet fail to solve other structurally similar problems

14

Karl S. Rosengren and Gregory S. Braswell

that draw on a different set of knowledge. Indeed, a relatively large body of evidence seems to suggest that children's thinking is quite variable across domains of knowledge. This type of variability in chi,klren's thinking across knowledge areas has been labeled as "domain specific" reasoning. Currently, a domain specific view of children's thinking has emerged as a powerful paradigm. Many researchers and theorists, for example, no longer attempt to provide global framework theories of cognitive development (exceptions are Case & Okamoto, 1996; Thelen & Smith, 1994). These researchers have begun to develop models or theories related to children's thinking within specific content areas, such as naive psychology (Flavell, Green, & Flavell, 1995; Wellman, 1991), naive biology (Atran, 1990, 1995; Carey, 1985), and naive physics (Carey & Spelke, 1994; Spelke, 1991). However, much debate has ensued over what exactly constitutes a domain (Hirschfeld & Gelman, 1994; Wellman & Gelman, 1998). We discuss this particular issue in Section IV.C. Although talking of a child's general cognitive capacity is no longer popular in many cognitive developmental forums, even researchers who argue that cognition is domain specific seem to assume a more or less one-to-one correspondence between underlying naive theories, mental representations, or mental models and children's problem-solving behavior. We discuss implications of acceptance of this notion when we discuss the role of theoretical approaches on interpretations of variability. D. VARIABILITY WITHIN A CHILD OVER REPEATED TRIALS

Variability within a child who is solving the same problem repeatedly has typically been ignored by developmentalists. Yet this type of variability is quite common. A child may, for example, be able to spell the word "potato" correctly one day but add an "e" the next day. Or the child may be able to solve a particular math problem, such as a three-digit addition or subtraction problem one day, and make an error the next. This type of variability is certainly seen by many parents and teachers as they observe children performing everyday and school-related cognitive tasks. Variability in performance on the same task over repeated trials or sessions is often downplayed and attributed to (a) the child's attention, (b) the child's motivation, or (c) some sort of learning or order effect. Regardless, when this type of variability is seen in a child, it is rarely if ever attributed to meaningful variation in the child's cognitive functioning. Children do appear incredibly sensitive to repeated questions. In some cases young children may become bored and uninterested if too many similar questions or problems are presented. Thus, researchers often rely on a single trial or a limited number of trials to assess children's thinking. An additional problem confronted by researchers is that children learn at an early age that an adult's repeating a question often means that the adult was dissatisfied

Variability in Children's Reasoning

15

with the initial answer (Siegal, 1997). For example, in a variety of cognitive tasks from conservation to appearance-reality, children tend to change their answer when asked the same (or similar) questions repeatedly (Siegal, 1997). One of the most thorough investigations of this type of variability was conducted by Siegler (1994, 1996). In a series of microgenetic studies of children's use of various addition strategies, such as summing the numbers, counting on the fingers, retrieval, and counting up from the larger addend, Siegler and his colleagues found that the strategy used varied from one trial to the next on about 30% of the paired problems. This variability was sometimes found when exactly the same problem was presented within a single session (Siegler & McGilly, 1989; Siegler & Shrager, 1984). Surprisingly, the children did not always adopt a more efficient strategy over sessions (as would be predicted by learning); rather, they were almost as likely to use the more sophisticated strategy in the initial encounter with the problem as they were in later encounters with the same problem. As we will discuss in greater detail in section IV.D, this type of variability in strategy use demonstrates that children have multiple strategies available at any given point. Thus, rather than development proceeding as the sequential replacement of strategies with more advanced ones, Siegler (1996) argues that we should view development as involving both the acquisition of new strategies and a shift in the likelihood that particular strategies will be used for a given type of problem. E. VARIABILITY WITHIN A CHILD WITHIN A SINGLE TRIAL

Like variability within a child across trials, variability within a child within a single trial has not received much attention, though for a very different reason. Specifically, this type of variability requires a focus on either problems that are quite complex and involved or ones that enable the researcher to focus on multiple modes of behavior. Much of the research on this type of variability has been conducted by GoldinMeadow and her colleagues (Goldin-Meadow, 1997; Goldin-Meadow et al., 1993; Perry, Church, & Goldin-Meadow, 1992). One of the main findings is that when children attempt to explain the answer to certain kinds of problems, such as Piagetian conservation tasks (Church & Goldin-Meadow, 1986) and number equivalence problems (Alibali & Goldin-Meadow, 1993), they often make gestures that do not match their verbal explanations. This gesture-speech mismatch was interpreted by Goldin-Meadow and her colleagues as an indicator of a transitional knowledge state. When children are in this transitional state, they are ready to advance to a higher level of understanding (Goldin-Meadow et al., 1993). Alibali and Goldin-Meadow suggested that during this transitional state the child has a number of different procedures for solving the problem and these procedures are activated at the same time. In many ways this notion is quite similar to Siegler's (1996) idea that at any given time a child is able to use a number of different strategies.

16

Karl S. Rosengren and Gregory S. Braswell

A slightly different case of variability within the same problem can be found in the research of Subbotsky (1985, 1990, 1992, 1997). In his research Subbotsky typically created situations in which children were asked to give verbal responses to a series of questions related to a particular phenomenon. For example, he has examined phenomena such as conservation of number, perception of length using the Mueller-Layer illusion, and beliefs in magical thinking. After obtaining a child's verbal response (asking which row of cookies is longer in a conservation task, which length is longer in the Mueller-Layer situation, or whether a box can magically transform a picture into a real object), Subbotsky either left the child to interact with the experimental stimuli (e.g., the magic box) or asked the child to perform some action (e.g., choose one of the rows of cookies). He has found that although children's verbal responses often seem quite logical and rational, their actual behavior reveals phenomenalistic or magical thinking. Specifically, even though children may say that the two rows of cookies have the same number of cookies, they will respond in a phenomenalistic manner, actually taking the row of cookies that looks longer due to wider spacing between the cookies. In the magic box study, Subbotsky found that when children are left alone with the "magic box" after hearing a story about how such a box can turn drawings into real objects when a magic word is uttered, the children often place pictures into the box and utter the magic words. The conclusion that Subbotsky drew from this series of studies is that children (and adults in certain situations) do not reason about the world in a consistently logical or rational manner. This logical mode may predominate when children are required to provide verbal judgments about unusual phenomena, but Subbotsky argued that this is only one form of reasoning that the child can draw upon. At other times, and within particular situations or contexts, a phenomenalistic or magical form of reasoning may predominate. Thus, according to Subbotsky, these different forms of reasoning "coexist" in the mind of both children and adults (Subbotsky, 2000).

IV. Theories and the Meaning of Variability in Children's Reasoning How should we address these various forms of variability in our theories of cognitive development? And should we? Before addressing these questions directly, we first turn to how various theoretical approaches within the field of cognitive development confront issues of variability. In this section we focus on how variability is viewed and the way in which assumptions regarding developmental processes in general influence the way in which variability in children's thinking is interpreted. One reason that variability in children's thinking has been virtually ignored by cognitive developmental researchers is that the majority of developmental theorists have embraced a view of development, owing much to Piaget, in which underlying cognitive processes are assumed to be relatively stable and fixed at any

Variability in Children's Reasoning

17

given point in the developmental progression. The notion that theories constrain the manner in which we consider data is not new (Kuhn, 1962). In this section we examine how variability in children's thinking is characterized by different theoretical approaches. We first consider Piaget's view of variability in children's thinking and then discuss sociocultural approaches, domain specific approaches, information processing accounts, and finally dynamic systems approaches. The key notion here is that different theoretical perspectives construe variability in children's thinking in very different ways. Some view variability as playing an important and central role in development; others either assume little variability in children's thinking or dismiss any importance to this variability. A. PIAGETIAN VIEW OF VARIABILITY

Piagetian theory is known primarily for its emphasis on universals in cognitive development. As Siegler and Ellis (1996) have noted, one of Piaget's primary goals was to define the "essences" of reasoning at different stages in development. Due to this focus on the underlying essences of cognitive development, the existence of variability in reasoning and any role variability might play in the process of cognitive development tended to be downplayed by Piaget. Furthermore, as Flavell (1963) noted, Piaget tended to gloss over both any differences among children within the same stage and any similarities among children across different stages of development. Flavell suggested that this stance was merely an unavoidable outcome of Piaget's attempts to isolate conceptually different strategies. However, as even a cursory reading of Piaget's work reveals, Piaget did not completely ignore the issue of variability. Indeed, Piaget's concepts of equilibration and horizontal decalage both relate to aspects of variability in children's reasoning.

1. Equilibration The concept of equilibration was used by Piaget to explain how cognitive growth is achieved through the competition and reconciliation of competing theories within the child (Siegler & Ellis, 1996). This process occurred within the child either endogenously or exogenously and was thought to be driven by interactions with others. For Piaget, interactions with individuals at the same level of cognitive development (i.e., peers, but ones with different perspectives) are among the major driving forces of developmental change (Rosengren, Behrend, & Perlmutter, 1993). What exactly equilibration is, or how it actually occurs, was never very well specified. Thus, although notions of variability seem implicit in the concept of equilibrationmhow these notions should be specified explicitly is unclear.

2. Horizontal Decalage More explicit recognition of cognitive variability can be found in the concept of horizontal decalage. This concept arose from the recognition that children

18

Karl S. Rosengren and Gregory S. Braswell

generally vary in their performance on tasks that should involve the same conceptual structures. For example, children in Piaget's concrete operational stage are often able to solve certain types of conservation problems prior to the time when they can solve other conservation problems. Specifically, children have been found to conserve mass prior to weight (Piaget & Inhelder, 1941). Another instance of horizontal decalage was observed in children's performance on class inclusion problems (Piaget, 1971). When preschool children are given a group of different flowers and are asked if there are more primroses or flowers, they are more likely to give the correct answer than if they were asked an analogous question about animals (perhaps because of differences in familiarity of the items). Solutions to both variations of the problem should involve the same cognitive structures (Flavell, 1963). To explain this type of variability (and in response to critics) Piaget formulated his theory of horizontal decalage. According to Piaget, cognitive structures characteristic of a stage can be used for some tasks and not others because of a decalage in the concepts needed to solve different versions of the tasks (Flavell, 1963). With regard to conservation problems, Piaget suggested that the hierarchical relationship of matter, weight, and volume with respect to one another is responsible for the decalage. That is, weight conservation is based on matter conservation, and volume conservation is based on weight conservation; thus, the acquisition of these abilities will be staggered (Piaget & Inhelder, 1941, as cited in Bovet, Domahidy-Dami, & Sinclair, 1982). Matter conservation is, according to Piaget, derived from object permanence and based on the "retrievability" of objects. In contrast, weight conservation is based on actions of carrying and holding that imply and go beyond matter conservation. Finally, volume conservation is based on understanding of matter and weight. A decalage is created because of this hierarchy of understanding, although the three characteristics are eventually integrated into the concept of density (Piaget, 1941, as cited in Bovet et al., 1982). These hierarchies of understanding typically were organized into a series of substages within the more general stage framework of Piaget's theory. This organization is seen clearly in Piaget's description of development during the sensorimotor stages (e.g., in the acquisition of object permanence), but it can be seen in other areas of Piaget's work as well (e.g., in the acquisition of the concept of "living thing"). Piaget also offered the more general notion of"resistances" to explain horizontal decalages in inclusion, conservation, and other cognitive tasks. The degree to which a phenomenon can be viewed logically determines its degree of resistance (Piaget, 1968, as cited in Bovet et al., 1982). In the case of the above-mentioned inclusion problems, Piaget (1971) speculated that the difference between using flowers versus animals may lie in the possibility that animals cannot be grouped like flowers on a table, but he only suggested this explanation as a possibility. Otherwise, he posited that "time lags are always due to an interaction between the

Variability in Children's Reasoning

19

person's structures on the one hand, and the resistances of the object on the other" (p. 11). Thus, some objects (e.g., flowers) possess less resistance than others (birds), just as some substances possess more friction than others in the domain of physics. In the case of conservation problems, weight may offer more "resistance" because it can be perceived as changing in some contexts, such as differences in exertion between pushing and pulling objects, and unchanging in other contexts, such as holding or supporting objects (Piaget & Garcia, 1971, as cited in Bovet et al., 1982). These resistances therefore were considered to drive horizontal decalages. This concept of resistance has a decidedly ad hoc flavor. 3. Cultural Variation In addition to decalages within specific domains, variability across cultures also had to be reckoned with. Although he originally posited a theory suggesting universality in cognitive development, Piaget (1972) later recognized that the rate of stage progression was found to differ across cultural groups. He offered three possible explanations for this type of variability. First, stage progressions may be faster in some cultural settings than in others due to the role of adult stimulation and stimulation afforded by the child's environment. Impoverished surroundings and lack of stimulation would likely slow development, although under the right sociocultural conditions Piaget thought that anyone could achieve the formal operations stage. Second, some individuals may not reach the formal operations stage due to different aptitudes. According to Piaget, at some point individuals develop their own particular aptitudes after passing through earlier stages of development, so that some individuals become better than others at artistic, practical, logical, mathematical, and other skills. The third explanation, preferred by Piaget, is a synthesis of the previous two. That is, everyone who develops normally achieves formal operations that are grounded in various domains that depend on particular aptitudes and "professional specializations," realized through apprenticeships or academic studies. Hence formal structures may be used in particular ways appropriate for the domains in which they arise; for example, law students may be good at logic that involves law issues but not as good at logic in other areas like physics (Piaget, 1972). As seen clearly in the cases of horizontal decalage and culture variability, Piaget was quite aware of issues of variability in children's thinking and he thought this variability was meaningful. However, equally clearly, Piaget's explanations maintain the notion of a one-to-one mapping between cognitive structures within the child and his or her performance on a particular task. B. SOCIOCULTURAL VIEW OF VARIABILITY

Vygotsky (1978) identified two kinds of development for a childmmaturation and learning. Maturation involves an individualistic, endogenous means of growth

20

Karl S. Rosengren and Gregory S. Braswell

and learning is social, requiring another person to teach the child. According to Vygotsky, adults and expert peers guide young children in advancing their problemsolving skills, allowing them to find solutions that they might not have arrived at on their own while providing them with a means for solving problems in the future. This type of learning experience need not be limited to formal teaching environments, but may be achieved through interactions with parents, older siblings, and more advanced children in play groups. In addition to dyad collaborations, however, cognitive strategies can be transmitted at the institutional level through books, maps, organizational conventions, academic training, and so on (Rogoff & Morelli, 1989; Wertsch, 1991). Because learning precedes maturation, a gap forms, known as the "zone of proximal development." This zone "is the distance between the actual developmental level as determined by independent problem solving and the level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peers" (Vygotsky, 1978, p. 86). The zone of proximal development provides a window into the future solitary capabilities of the child. Thus, variability in reasoning is manifest between the child's mental capabilities in solitary and collaborative situations; Valsiner (1984) considered the zone of proximal development as an example of "bounded variability." One criticism of Vygotsky's theory is that his analysis of mental processes was too general (Wertsch & Tulviste, 1992) in that he neglected the role of context in the transmission and use of psychological tools. According to Cole (1996), "A commitment to tool mediation as the fulcrum of thought logically requires the inclusion of context-level constraints: all tools must simultaneously conform to constraints arising from the activity they mediate and from the physical and mental characteristics of the human beings who use them. There is no universal, contextfree tool, independent of task and agent" (p.172). In an attempt to acknowledge this diversity, Wertsch (1991) suggested that a particularly fruitful approach may be to examine the multitude of cognitive strategies that can be used across and within cultures and individuals. Numerous modes of action may be suitable for a given situation; that is, a "heterogeneity" of tools may solve a particular problem. Related psychological tools used for a given context can be considered as forming what Wertsch called a "cultural tool kit." Some tools are more "privileged" than others in that some are more appropriate or effective depending on the context in which they are used (Wertsch, 1991). The ideas of Bakhtin, another Russian active at the beginning of the 20th century, have been used to extend the Vygotskian paradigm (Wertsch, 1991). Bakhtin (1981), whose interest lay in language, analyzed utterances in terms of voice, encompassing the speaker's point of view among other factors. Utterances are always couched in a social context, and they are always related to other utterances within conversation (i.e., they are "dialogic"). Thus, one must always look at who is producing utterances and who is the intended audience. In this way,

Variability in Children's Reasoning

21

a child's performance is viewed as highly context sensitive. Also according to Bakhtin's (1986) dialogic theory, utterances are created in the context of "social languages" and "speech genres." Speech specific to social strata such as age cohorts, professions, and social classes embodies social languages (e.g., jargon). Speech genres, in contrast, may be seen as different conventional frameworks for conservation (e.g., greetings, commands, informal conversation, and debates). We always utilize social languages and speech genres when we speak, and their use is always context-dependent. When one speaks, one is using many voices, including one's own voice and oftentimes a conventionalized voice derived from other speakers. This phenomenon is known as "ventriloquation"--"the process whereby one voice speaks through another voice or voice type in a social language" (Wertsch, 1991, p. 59). Thus, one will often use sayings or phrases taken from others, from popular culture, or from the particular social language being used. Hence, the utterance is both individual and social simultaneously. Although dialogic theory is more concerned with speech, these ideas can certainly be applied to other psychological processes such as perspective-taking ability (Fernyhough, 1996). Sociocultural approaches such as Vygotsky's clearly embrace some types of variability in children's reasoning and treat this variability as meaningful. For the most part, however, the focus of researchers adhering to this paradigm is on how children's reasoning varies as a function of context. One aspect of Vygotsky's theory that is in some way similar to notions of the dynamic systems approach (described in section IV.E) is the notion that different developmental processes proceed at different rates, yet interact and influence one another. For example, Vygotsky (1978) viewed cognition and language as developing somewhat independently at first, and later, after the initial stages of development, as interacting in a dynamic interplay, each influencing changes in the other. This complex interaction is quite similar to some concepts of dynamic systems theory. But prior to discussing that approach we examine two other theoretical approaches that have dominated the field of cognitive development. C. DOMAIN SPECIFICITY VIEW OF VARIABILITY

In recent years, many of the descriptions of children's thinking in terms of global properties, or general modes of reasoning, have been replaced by ones emphasizing domain specificity. This domain specificity or theory approach to cognitive development is relatively young, but research within this paradigm has grown to dominate much of the field of cognitive development. In particular, research under the rubric of "theory-of-mind" has become one of the major foci of many researchers. This theoretical approach, however, does not constitute a single theory in the manner that Piaget's theory dominated the field, but is actually comprised of a number of different theoretical viewpoints with some common assumptions. One area where these separate domain specific accounts differ is in how variable

22

Karl S. Rosengren and Gregory S. Braswell

they assume the process of cognitive development is. Prior to discussing issues of variability from this perspective, we briefly discuss a few of these accounts, outlining some of the basic assumptions of this general theoretical approach. The key assumption of the theory-theory approach is that children's (and adults') cognition (and its development) varies by domain. One explanation is that children have a number of foundational "theories" that are used to reason about distinctly different domains. The exact nature of these theories and what is actually meant by a"domain" are both quite controversial. Wellman and Gelman (1998) have categorized these approaches in three broad categories: modular approaches, theorytheory approaches, and expertise approaches (we discuss this last approach in the information processing section). Although these approaches share the common assumption that cognition and its development are greatly influenced by domain, they differ in how they define a cognitive domain, what they view as the basic metaphor of the child, what mechanisms are involved in cognitive processes, how they view the outcome of cognitive development in terms of variability, and what they view as innate. 1. Modular View

The modular approach generally views the child as a young adult, born with a set of innate biological constraints that lead to a developmental outcome that is highly fixed and constrained (Wellman & Gelman, 1998). From this perspective, a domain is defined by a particular type of module. Although language is the domain most commonly thought of from this modular perspective (e.g., Chomsky, 1988), other candidate domains include vision (Marr, 1982) and various cognitive processes (Karmiloff-Smith, 1992; Sperber, 1994) including theory-of-mind (Baron-Cohen, 1995; Leslie, 1994). A key assumption here is that modules function independent from one another, operating more or less in isolation (Wellman & Gelman, 1998). As mentioned earlier, this theoretical approach includes the assumption that cognitive development is highly channeled. As long as the necessary inputs are received and the system is not injured in some manner, a specific outcome will be achieved (e.g., fluency with language). For the most part, variability as a function of age or maturation is the only type of variability expected from this viewpoint. For example, biological constraints determine how language unfolds and lead to important changes as a function of age, but this process is assumed to be highly similar across different children. Additionally, within a particular age a child's language abilities are considered to be quite stable (though even the most strident proponent of this approach will likely acknowledge competence/performance issues). In a sense, this approach replaces Piaget's global one-to-one correspondence between age and a way of thinking with one-to-one correspondences within a particular domain. Thus, a child's underlying theory-of-mind is viewed as highly stable across a given time frame, regardless of context. Thus, researchers from this perspective attempt to get at the "essence" of these cognitive modules.

Variability in Children's Reasoning

23

2. Theory View Although sharing the key assumption of domain specificity, the theory-theory (Gopnik & Wellman, 1994) approach differs quite a bit from the modular account. As Wellman and Gelman (1998) suggested, the child might best be captured metaphorically by this perspective as an "alien," one whose reasoning is dramatically different from adults' reasoning. From this view, domains are defined by naive theories, with distinct causal-explanatory underpinnings. Across these domains children's reasoning is believed to be highly variable. However, variability is used here in a different sense. Variability is not viewed here in terms of levels of ability, but rather variability across domains is viewed as occurring in terms of the causal explanations that children provide and the particular entities that they use to reason about phenomena in a particular domain. For example, in the domain of physics, children reason about objects and forces that interact on them. In the domain of biology, children reason about living things and distinctly biological processes. In the domain of psychology, children reason about sentient beings and mental processes. Thus, reasoning across these domains varies as a function of the naive theory that is brought to bear in a particular situation. For the most part, researchers working within this theoretical approach assume highly stable outcomes, provided that children receive the necessary inputs. More specifically, children's theories are assumed to be activated by input that should be common to all individuals. Thus, for example, a naive psychology is thought to develop as long as an individual is in some sort of social situation (assuming of course no injury or insult to the cognitive system). Here again, we have mainly replaced Piaget's global one-to-one correspondences with domain specific ones. That is, children of a particular age are thought to reason in a fairly consistent and coherent manner. Before turning to the expertise approach to domain specificity, which we address in the information processing section, we discuss some of the ramifications of the modular and theory approach to domain specificity. One outcome of both of these approaches, similar to the outcome of Piaget's own endeavors (but with a domain specific flavor), has been an emphasis of researchers from these theoretical perspectives on documenting the particular age at which children can be characterized as exhibiting a particular level of competence in a given domain. However, as Fischer and Kennedy (1997) argued, "The actual behavior of real people does not fit conventional beliefs in monolithic competence! People act at multiple levels of skill, not only one level, even for a narrowly defined domain" (p. 120). In our view, this monolithic approach has led to some rather long-winded debates over the exact age at which a child can be said to have a particular competency. For example, within the theory-theory approach researchers have engaged in considerable debate over whether children can be said to have a theory-of-mind at age 1 3~ or later. Likewise, in the domain of children's biological reasoning, researchers have been debating for some time about whether young preschool children have

24

Karl S. Rosengren and Gregory S. Braswell

a biological form of reasoning that is distinct from either psychological reasoning or reasoning about social relations (Rosengren et al., 1991; Springer & Keil, 1989) or whether this differentiation occurs relatively late (Carey, 1985; Solomon et al., 1996). Again, much of the work in this field has been based on the assumption that most children at a specific age reason in a similar manner. In a sense, Piaget's global stages have been replaced with relatively static representations of children's thinking in particular domains. The influence of this assumption of one-to-one correspondence between mental representations and problem-solving behavior can be seen, for example, in the work of Vosniadou and Brewer (1992, 1994). In a series of fascinating and creative investigations they explored the types of mental models that children create when asked to reason about certain types of phenomena (e.g., the shape of the earth, the day-night cycle). They suggested that over the course of development children's mental models of the earth progress from initial models to synthetic models and then to scientific (or culturally accepted) models. These initial models are thought to be derived primarily from sensory data and have been found to be quite similar across cultures (Samarapungavan, Vosniadou, & Brewer, 1996; Vosniadou, 1994). The synthetic models emerge as the children combine their initial models with those taught by parents, teachers, and the culture at large. These synthetic models have been found to be much more variable both between individuals and cross culturally. Finally, if a child is provided enough data and education, he or she arrives at a model that is accepted by the community (in some cases this may be in fact the model accepted by scientists). Once again these models show a high level of similarity from individual to individual. The assumption is that at any one point a child will construct a single model to be used to solve various related problems. Based on this assumption a number of different types of models of the earth have been described (see Vosniadou & Brewer, 1992). Some of these, such as the rectangular earth, disk earth, flattened sphere, and sphere, do seem to capture a relatively coherent use of a single mental model or strategy. However, a number of the other mental models described, including the dual earth and hollow sphere models, might in reality be capturing the responses of children who are shifting between a number of different mental models in response to varying questions. Specifically, children who are characterized as holding the dual earth model, in which the earth is believed to be both flat like a pancake (so that we can stand on it without falling) and round like a ball (as it looks from the moon), could as easily be characterized as shifting from one model to another in an attempt to solve different problems. Vosniadou and Brewer (1992) described children, primarily first-graders, with mixed models of the earth, but one is left with the impression that these mixed models stem more from problems of categorization on the part of experimenters than from children shifting among a number of different models in response to various questions and task demands.

Variability in Children's Reasoning

25

D. INFORMATION PROCESSING VIEWS OF VARIABILITY

Researchers conducting research within this general paradigm often focus on children's acquisition of rules or strategies for solving particular problems. This approach is often conducted from a domain specific perspective, but many researchers from this general paradigm look for developmental changes in general cognitive functioning, such as speed of processing or memory abilities. Those who focus on domains of knowledge take quite a different perspective from the two approaches discussed earlier within the domain specificity section. In particular, they differ in what is considered a domain. Rather than defining a domain in terms of cognitive modules or naive theories, proponents of the information processing approach view domains as defined by particular sets of knowledge. Specifically, individuals' problem-solving performance is considered to vary based on their knowledge and experience. From this perspective, children's (and adults') performance will differ greatly based on their knowledge within a particular domain. Thus, because knowledge and experience vary greatly from child to child, and within a child across knowledge areas, this theoretical approach incorporates issues of variability to a much greater extent than the other approaches discussed so far. Researchers from this perspective are also more likely to be interested in examining individual differences in children's reasoning (e.g., Kerkman & Siegler, 1997). One manner in which the developing child has been characterized from an information processing approach is as acquiring more and more sophisticated and powerful rules with increased knowledge and experience (Siegler, 1983). Although the earlier view was that with development children replace ineffective rules or strategies with more effective ones, later research suggested that at any given time children have access to multiple rules or strategies for solving a problem (Siegler, 1996). Thus, rather than viewing development as the sequential acquisition of more powerful rules, Siegler (1996) characterized development as involving a shift in the distribution of strategies that might be brought to bear on a particular problem. Indeed, having multiple strategies available at any one time would seem quite adaptive, enabling children to explore the usefulness of various strategies as well as enabling them to shift easily in response to changing task demands. From this view, strategy use changes with age, but also with context. This notion of context sensitivity is crucial to dynamic systems approaches as well. E. DYNAMIC SYSTEMS VIEWS OF VARIABILITY

Unlike most of the other theoretical approaches to cognitive development discussed so far issues of variability are central to the dynamic systems approach. Thelen and Smith (1994), for example, noted that variability is a fundamental characteristic of complex systems. They have argued that variability is a major

26

Karl S. Rosengren and Gregory S. Braswell

(if not the main) driving force behind developmental change. Although this theoretical approach has primarily been used to describe developmental changes in the motor abilities of young children (Goldfield, 1995; Goldfield, Kay, & Warren, 1993; Thelen, 1996; Thelen & Spencer, 1998), it has been generalized to examine development in language and cognition as well (Smith, Thelen, Titzer, & McLin, 1999; van Geert, 1995). At the heart of this approach is the assumption that complex biological systems are composed of a complex interaction of subsystems. These subsystems interact in a complex and dynamic manner to lead the entire system to "self-organize" into a stable state or attractor (i.e., a stable form of behavior). The particular attractor, the form of the attractor, and its stability are determined by the interaction of the subsystems with each other. Researchers from this theoretical approach examine variability to assess whether a particular attractor (mode of behavior) is stable and whether the system is in a transition statemabout to shift to some other attractor state (a different mode of behavior). The environment or context is an additional factor that influences the manner in which the subsystems interact. These attractors are not stable in the sense of a Piagetian stage, but are seen as "dynamically" stablemnever actually at rest. Metaphorically, one can think of a series of molecules interacting. For example, under relatively low temperatures or high pressures water molecules will coalesce (self-organize) and form a solid, and under relatively moderate temperatures the molecules form a different dynamically stable state (liquid water). In terms of cognitive development, a child's level of performance can be modeled as a dynamically stable state arising from the interaction of various subsystems. The notion here is that emergence of a particular cognitive ability arises from the growth of a number of underlying properties. These underlying properties (or subsystems) grow at different rates and compete for limited resources. At a given point in time certain properties may serve as control variables or rate limiters. These variables define the level of the entire system; that is, they determine the particular attractor region of the entire system. In the water analogy, pressure and temperature may both serve as control variables, but they do so under different environmental conditions. Under normal conditions, temperature is the primary rate limiter; that is, the air temperature determines what state--solid, liquid, or g a s ~ t h e water will be in. However, in certain contexts, such as a sealed vessel, pressure can serve as a control variable. Specifically, by increasing the pressure in the vessel, the water can be forced to undergo a state transition from gas to liquid or liquid to solid. From this perspective, a particular cognitive ability should not be viewed in isolation, but should be examined as arising from an interaction of various subprocesses. For the most part, dynamic systems theory has been used in a metaphoric sense to describe changes seen in cognitive development. However, van Geert (1991, 1993, 1998) has employed this approach to develop mathematical models

Variability in Children's Reasoning

27

of cognitive growth. For example, van Geert (1993) modeled how the growth of a child's lexicon develops in relation to other aspects of language, specifically syntax. He has even recast both Piaget's model of assimilation-accommodation and Vygotsky's model of development in the zone of proximal development from a dynamic systems perspective (van Geert, 1995, 1998, 1999). In doing so, he found that hidden complexities (greater variability) are uncovered in the development of language and cognitive processes. Smith et al. (1999) have used a similar dynamic systems approach to reexamine the classic A-not-B error in Piagetian object permanence tasks. Rather than viewing this phenomenon as arising from representational, memory, spatial, or inhibitory factors, they suggested that the A-not-B error arises from a complex interaction of visual attention (gaze direction), task characteristics (similarity of hiding places), memory, and posture. Exploring the nature of the variability within each of these subcomponents leads to a better understanding of the behavior. Of the theoretical approaches discussed, only dynamic systems theory views variability as a truly central aspect of cognition and its development. However, this approach has made relatively little headway in the field of cognitive development, most likely because researchers lack familiarity with the theoretical underpinnings that come from the fields of physics and nonlinear dynamics and perhaps because some practitioners of dynamic systems theory emphasize the use of sophisticated modeling techniques. Indeed, some theorists argue that a true dynamic systems approach requires the application of nonlinear dynamics and mathematical modeling. This view may also deter many developmentalists from using the approach. We suggest that at this stage in the "development" of the theory, use of it in a metaphoric sense may be quite helpful in inducing researchers to take greater notice of the inherent variability within the cognitive system.

V. Methodological Issues Although theories constrain the way we collect and interpret data, research traditions lead us to choose particular experimental designs and statistical techniques that also structure the manner in which we examine children's thinking. As Fischer and Kennedy said, "Traditional research methods hide the nonlinear nature of development to produce the illusion of monotonic growth by minimizing variability. In this way social science researchers make action and thought seem constant or stable instead of variable" (1997, p. 120). In this section we briefly explore some of the traditional means of examining children's thinking, highlight some new trends that seem more appropriate for examining variability more closely, and make some suggestions about additional means for looking at variability. A central issue to be resolved when examining variability in children's reasoning is whether the variability observed is meaningful or represents either random noise

28

Karl S. Rosengren and Gregory S. Braswell

or problems with the experimental methodology. In this section we start with a discussion of detecting randomness, consider other issues that hinder assessment of whether children's reasoning is in fact variable, and turn to some alternative approaches. A. ASSESSING RANDOMNESS IN CHILDREN'S RESPONSES

Techniques for examining random responses are fairly well developed in the field of cognitive development. One of the most common procedures for uncovering random responses is to test children's performance against chance. Often this approach involves using t-tests to determine whether a child's performance is significantly different from chance, or it involves using binomial probability to examine the pattern of responses. Unfortunately, determining whether a response pattern is significantly different from chance is not always as easy at it appears. For example, in DeLoache's (1987, 1995) research on children's search for a hidden object in a real space and in a scale model of that space, no clear-cut way to determine chance responding is evident. More mathematically sophisticated techniques can be used to assess randomness in the data, but they require many more data points than are typically obtained in studies of cognitive development. For example, if one collects a time series of the frequency of some sort of behavioral response, one can use Fourier analysis to determine whether the data fit particular noise profiles or show a consistent pattern (Gottman, 1981). This technique requires a minimum of 128 data points, a number not likely to be seen in studies of cognitive development. B. ISSUES OBSCURING VARIABILITY IN CHILDREN'S REASONING

One problem with determining whether variability observed in children's reasoning is meaningful is that in many instances the data collected do not enable one to examine issues of variability to any great extent. As mentioned previously, studies in cognitive development typically rely on relatively few data points from any one child. Often fewer than two or three trials are given in studies of young children's reasoning. Part of the reason for this procedure is that young children become bored with any tasks that seem repetitive. Likewise, the attention span of young children does not often enable the researcher to collect data over more than a 10- to 20-minute session. Unfortunately, variability is hard to assess when only a few samples of data are obtained from any one child. Thus, if we are to examine issues of variability in children's reasoning we need to collect enough data to allow us to observe it when it is present. Increasing the number of items presented to young children is in many cases not feasible, but collecting data over multiple sessions is sometimes a reasonable approach.

Variability in Children's Reasoning

29

Another methodological technique that obscures issues of variability in children's reasoning is the cross-sectional design. Although cross-sectional studies are fast and efficient and they provide ample evidence of age-related differences in children's thinking, they mask variability. This masking is not an inherent problem in this design per se, but is related to the emphasis on mean differences in performance between children of different ages. The only measure of variability typically reported in most studies of cognitive development is the standard deviation. One problem with this reliance on standard deviations is that very different types of frequency distributions can lead to the same mean and standard deviation (see Gottman, 1981). Thus, by having relatively few samples and relying on means and standard deviations we may not capture the true nature of variability in children's reasoning. Another feature of much of the research in cognitive development that obscures issues of variability is the reliance on categorical data. In Siegler's (1994) research on counting, for example, children were categorized as using a limited set of different counting strategies. In studies of conceptual development, children are typically categorized as choosing from among two or more alternatives. For example, in Gelman and Markman's (1986) classic study, children were given two choices, one perceptually similar and one in the same taxonomic category. In many of these studies, the statistic reported is the mean number (and standard deviation) of children who use a particular strategy, or pick the perceptual, thematic, or taxonomic alternatives. This approach tends to reinforce the idea that children's thinking is relatively stable and consistent, in part because researchers typically discuss age-related differences in the dominant categorical response. As noted previously, Siegler has begun to examine how children employ a variety of strategies over trials and sessions. In the next section we discuss a method for empirically assessing variability in categorical data that goes beyond mean responses. A final methodological problem that obscures variability in children's thinking stems from the emphasis on domain specificity in children's thinking. The domain specific approach leads researchers to interpret performance on a given task as tapping into a specific, potentially isolated behavior. This interpretation leads researchers in this paradigm to downplay potential factors (either domain general or cross-domain influences) that might cause variation in the behavior of interest. That is, to the extent that we examine relatively narrow aspects of children's thinking, such as performance on a conservation task, object permanence, or a false-belief task, we tend to miss (or ignore) important sources of variability that might contribute to performance. C. HOW TO EXAMINE VARIABILITY IN CHILDREN'S THINKING

We have a number of suggestions for more fully examining variability in children's thinking. First, we suggest that researchers should strive to collect more

30

Karl S. Rosengren and Gregory S. Braswell

data from individual children. Although, young children's short attention span limits the number of responses that can be obtained from them in a given task, the number can be increased by collecting data over multiple sessions. Microgenetic approaches (Kuhn, 1995; Siegler and Crowley, 1991) fit this description. These approaches are clearly useful for examining transition points in children's reasoning, but we suggest that routinely collecting data over multiple sessions will enable us to have a better understanding of just how variable children's reasoning tends to be. Second, we suggest applying statistical techniques that enable us to quantify variability in categorical data. Two techniques that could be used are variability relative entropy (R; Schafer, 1980) or dispersion (Kirk, 1978). These techniques enable one to examine the extent to which data vary or are dispersed across categories. Relative entropy can vary from 0 (no variability--single-response type) to 1.0 (performance is distributed equally over all possible categories). Although no statistical procedures for testing the significance of different levels of relative entropy currently exist, we have adopted the convention of categorizing performance in terms of low levels of relative entropy (R < .34), moderate levels of relative entropy (.34 < R < .66), and high levels of relative entropy (R > .66). We have used this technique to examine variability in children's drawing strategies as a function of age (Braswell & Rosengren, 2001) and variability in children's reasoning about biological change (Rosengren et al., 2001). Relative entropy can be used to examine variability in responses across age (to pin down particular ages where transitions occur) or within particular children over trials. In Figure 1 we plot the relative entropy values based on data reported by Vosniadou and Brewer (1992) on mental models of the earth. First- and third-grade children in their study were characterized as using a number of different mental models including the sphere, flattened sphere, hollow sphere, dual earth, disk earth, and rectangular earth. As can be seen in the figure, high levels of relative entropy were found for both the first- and third-graders. This finding suggests that children in this age range are highly variable, with no single mental model predominating. In contrast, the relative entropy obtained for fifth-graders was quite low, suggesting that a single mental model (in this case the sphere) has emerged as the dominant one. As mentioned earlier, the relative entropy index can also be used to quantify variation in children's responses over sessions. In Figure 2, we plot the relative entropies for children's data presented in Figure 1 of Siegler's (1994) article on cognitive variability. Three different patterns can be seen in this figure. Child 1 exhibited a low level of relative entropy in the first block of sessions, indicating a high stable use of a particular strategy (in this case the sum strategy). In the second block of sessions the level of relative entropy jumped to the moderate level and by the third block the child's strategy use was characterized by high levels of entropy (no single strategy predominated). By the fifth block of sessions, the relative entropy values for this child were back in the moderate level and the child

Variability in Children's Reasoning

31

1.0

0.9

0.8

0.7 nQ,

0.6

c ILl

0.5

O t_

Q)

== 0,~ t~ ,.=.,

0.4

n0.3

0.2

0.0

First

Third

Fifth

Grade Level Fig. 1. Relative entropy based on data of children's mental models of the earth. (From Vosniadou, S., & Brewer, W. E (1992). Mental models of the earth." A study of conceptual change in child hood. Cognitive Psychology, 24, 535-585. Adapted by permission of Academic Press.)

had begun to settle into using a new strategy (the min strategy). In this case, the increase in relative entropy may indicate, in much the same way as a gesturespeech mismatch, that the child was in a transition state in which he or she might have benefited from instruction. In contrast to child 1, child 2 exhibited high levels of relative entropy across all sessions. Based on the percentage of strategy use, this child would be categorized as mostly using a retrieval strategy (the child used this strategy about 40% of the time over the 25 sessions observed); however, the relative entropy index indicates a greater degree of variability than the percentage measure would suggest. The third child depicted in Figure 2 exhibited yet a different pattern of relative entropy values. This child started out exhibiting high levels of entropy. In the third block of sessions the relative entropy values dropped, suggesting that the child adopted a dominant strategy (in this case the sum strategy). However, the relative entropy values jumped to high levels in the fourth block of sessions before once again dropping to a low level in the final block, in which a different strategy emerged

" ~ ~'~)

09

""

I"

p

..,,,

d

ii1 in.

,i1~,,~

0.8"

,"

0.7-

.~

,." ,',K

-

o=.0.6.

/

\

\

I:: 111

0.5-

B

m .,...

%

\

/

\ Ik

X \

J

IX: OA0.3-

\

-

\

/

\

4'

,,u

",

/

I

l_ ,.i..l

".

/

1

4)

-

,i,-

\

Child1

\

---...-I1--- C hil d 2 02-

............, ....

x \

-----Child3

x 0.1

"

0 , .............

..................

1

i ............................................

2

i

.........................................

3

i

............

4

!

............

5

l

.................

6

Sessions (in blocks of 5) Fig. 2. Relative entropy based on data of children's use of different strategies to solve addition problems. (From Siegler, R. S. (1994). Cognitive variability: A key to understanding cognitive development. Current Directions in Psychological Science, 1, 1-5. Adapted by permission, Blackwell Publishers.)

Variability in Children's Reasoning

33

(retrieval). This oscillation in relative entropy may also indicate that the child was in some sort of transition state, ready to move to a more advanced level of solving addition problems. We suggest that measures of variability in categorical data, such as the relative entropy index, are useful for a variety of reasons. First, in conjunction with frequency data, they can help researchers to examine issues of variability more clearly. Second, we suggest that this approach may help researchers to identify important transition points both across age and within children in children's reasoning. Finally, we suggest that this approach might also prove to be useful for examining individual differences in children's responses over time or trials. That is, relative entropy profiles of children solving problems over repeated sessions may help to identify children who might benefit most from instruction and the point at which instruction might be most effective for stimulating advances in their reasoning. If researchers are going to examine variability in children's reasoning, they will need to step back from the current emphasis on domain specificity in children's reasoning. One of the unfortunate outcomes of the emphasis on domain specificity in children's reasoning is a severe narrowing of the focus of our inquiries. Specifically, researchers have often focused solely on children's performance on such things as false-belief tasks, addition problems, or reasoning about inheritance. Although, children's reasoning probably varies by domain in very important ways (regardless of how "domain" is defined), we should acknowledge that the cognitive system is a complex one. Thus, we need to widen our focus and examine how cognitive processes are influenced by a whole host of factors, including noncognitive ones. We can see this influence in, for example, studies that have shown that attainment of certain motor milestones (sitting, reaching, locomoting) lead to changes in specific cognitive processes (object search, spatial representation, visual attention; Bertenthal & Campos, 1990). Although reasoning may develop at different rates and in somewhat different forms from domain to domain, the complexity and the integrative nature of humans imply that children's reasoning in a particular domain is unlikely to be completely encapsulated and uninfluenced by factors outside that domain. Although separating cognitive processes from motor or social processes often seems reasonable, both of the latter types of processes affect the nature of the environment in which the cognitive processes function. For example, being able to locomote expands children's spatial world and increases the demands for attention, as the children must learn to update their location as they move to new vantage points. Likewise, the ability to view the world from multiple vantage points enables children to create a more rich and elaborate representation of the spatial world, one that is not merely centered on their own bodies (Bertenthal & Campos, 1990). Thus, we strongly advocate the approach of Smith et al. (1999) who have examined how various behaviors, both motoric and cognitive, interact to lead to changes in infants' behavior with respect to the A-not-B error.

34

Karl S. Rosengren and Gregory S. Braswell

VI. Conclusions Variability clearly appears to be an inherent aspect of children's reasoning. Indeed, one can find variability in almost any aspect of cognitive behavior-provided that one is looking for it. We argue, as have Miller and Coyle (1999), Siegler (1994, 1996) and Thelen and Smith (1994), that children's thinking in most domains is characterized by variability and is highly context sensitive. We suggest that current views of children's thinking that emphasize domain specificity are flawed by the assumption that children of the same age reason in much the same way. Researchers who use this approach, with its emphasis on age-based similarities, have tended to ignore variability in children's reasoning, often treating variability as noise, and have treated development as a process that progresses in the same lock-step fashion across most children. Rather than characterizing children as having static theories within a particular domain, such as naive biology, naive physics, or naive psychology, or as using a single dominant mode of reasoning or problem solving at a particular age, researchers might more usefully characterize children as having dynamically stable modes of reasoning (see Rosengren & Hickling, 2000). In this context, dynamic means that children's reasoning changes with time and context. From a dynamic systems perspective, relatively stable modes of reasoning would be viewed as dynamically stable attractors within a larger attractor landscape (the range of possible behaviors). The stability of these attractors and the overall attractor landscape would also be viewed as shifting with age and context. At this point, no one can be sure whether a dynamic systems approach to cognitive development will have much success, although this theoretical approach has been successfully used to describe a variety of complex biological systems. The problem with the approach, as it is currently articulated by Thelen and colleagues, is that it fails to do justice to reasoning abilities that are usually described in terms of symbolic or representational abilities. For example, an issue that is not at all clear is whether a dynamic systems account can be constructed to provide an adequate explanation of the transitions in young children's search behaviors using scale models of real spaces (e.g., DeLoache, 1995; DeLoache, Miller, & Rosengren, 1997). The dynamic systems emphasis on dynamically stable states, multiple attractors, and context sensitivity does have substantial appeal in that it places variability in performance in a central role, as a signal of a transition state, as an indicator of stability of behavior, and as a driving force in the process of change itself. In our studies of children's reasoning, we need to identify meaningful transition points in cognitive development and study these points more extensively. Transition points can be determined to some extent through cross-sectional research, but to understand the processes involved in the changes that occur we will need to focus more intently on these periods with careful microgenetic studies designed to hone in on the processes underlying developmental change. In doing so, we need

Variability in Children's Reasoning

35

to explore variability carefully within and among individuals and do systematic manipulations of parameters that may alter the stability of the reasoning processes. To examine the most meaningful forms of variability we may need to place a greater emphasis on studying children's reasoning in their everyday life. Research on the natural language databases (e.g., Bartsch & Wellman, 1995) is perhaps a good starting place for this approach. Documenting influences on variability, the ways it changes, the ways it can be measured, and its role in mechanisms of change are just a few of the aims of a complete variational a c c o u n t m o n e that should serve as the underpinning of a more complete model of cognitive development. In summary, we believe that rather than collecting large amounts of data from relatively large numbers of children on very narrowly defined, specific cognitive tasks in a relatively short time frame, we need to collect more data across a variety of both cognitive and noncognitive tasks over longer time frames. Cognition by its nature is complex and multifaceted, therefore we must begin to examine cognitive abilities as arising from the dynamic interplay of many underlying p r o c e s s e s m rather than as a set of domain specific abilities that develop in a isolated manner. If we are to acknowledge the complexity of the cognitive system and its development we need to embrace the fact that reasoning cannot be divorced from processes that are not typically viewed as cognitive. Researchers need to acknowledge wholly that cognitive development is best captured as a complex, evolving system dynamically responding to a rich and vibrant environment. Only when we embrace the notion of variability as an important and inherent aspect of the cognitive system and its development will we be able to construct more adequate models of cognitive development.

ACKNOWLEDGMENTS We gratefully acknowledge Sarah Mangelsdorf and Hayne Reese for reading and commenting on an earlier version of this article. The writing of this article was completed while the first author was at the Free University, Amsterdam, on a sabbatical leave from the University of Illinois. During the writing of this article the first author was supported by Grant NSF SBR 97-20304 from the National Science Foundation, and the second author was funded by the Grant 1-5-62049 from the National Institutes of Health.

REFERENCES Alibali, M., & Goldin-Meadow, S. (1993). Gesture-speech mismatch and mechanisms of learning: What the hands reveal about a child's state of mind. Cognitive Psychology, 25, 468-523. Atran, S. (1990). Cognitive foundations ofnatural history. Cambridge, England: Cambridge University Press. Atran, S. (1995). Causal constraints on categories and categorical constraints on biological reasoning across cultures. In D. Sperber, D. Premack, & A. J. Premack (Eds.), Causal cognition: A multidisciplinary debate (pp. 205-233). New York: Oxford University Press.

36

Karl S. Rosengren and Gregory S. Braswell

Avis, J., & Harris, E L. (1991). Belief-desire reasoning among Baka children: Evidence for a universal conception of mind. Child Development, 62, 460-467. Bakhtin, M. M. (1981). The dialogical imagination: Four essays by M.M. Bakhtin (M. Holquist, Ed.; C. Emerson & M. Holquist, Trans.). Austin: University of Texas Press. Bakhtin, M. M. (1986). Speech genres and other late essays (C. Emerson & M. Holquist, Eds.; V. W. McGee, Trans.). Austin: University of Texas Press. Baron-Cohen, S. (1995). Mind blindness: An essay on autism and theory of mind. Cambridge, MA:MIT Press. Bartsch, K., & Wellman, H. M. (1995). Children talk about the mind. New York: Oxford University Press. Bertenthal, B.I., & Campos, J. J. (1990). A systems approach to the organizing effects of self-produced locomotion during infancy. Advances in Infancy Research, 6, 51-98. Bovet, M. C., Domahidy-Dami, C., & Sinclair, H. (1982). Conservations and decalage. Human Development, 25, 201-214. Braswell, G. S., & Rosengren, K. S. (2000). Variability in the development of graphic production. International Journal of Behavioral Development. 24, 153-166. Carey, S. (1985). Conceptual change in childhood. Cambridge, MA: MIT Press. Carey, S., & Spelke, E. (1994). Domain-specific knowledge and conceptual change. In L. A. Hirschfeld & S. A. Gelman (Eds.). Mapping the mind: Domain specificity in cognition and culture (pp. 169200). Cambridge, England: Cambridge University Press. Case, R., & Okamoto, Y. (1996). The role of central conceptual structures in the development of children's thought. Monographs of the society for Research in Child Development (Serial No. 246). Chen, C., & Stevenson, H. W. (1995). Motivation and mathematics achievement: A comparative study of Asian-American, Caucasian-American, and East Asian high school students. Child Development, 66, 1215-1234. Chomsky, N. (1988). Language and problems of knowledge. Cambridge, MA: MIT Press. Church, R. B., & Goldin-Meadow, S. (1986). The mismatch between speech and gesture as an index of transitional knowledge. Cognition, 23, 43-71. Cole, M. (1996). Cultural psychology: A once and future discipline. Cambridge, MA: Belknap Press. Cronbach, L. J. (1957). The two disciplines of scientific psychology. American Psychology, 12, 671684. Dansky, J. L. (1980). Make-believe: A mediator of the relationship between play and associative fluency. Child Development, 51, 31-36. DeLoache, J. S. (1987). Rapid change in the symbolic functioning of young children. Science, 238, 1556-1557. DeLoache, J. S. (1995). Early understanding and use of symbols: The model model. Current Directions in Psychological Science, 4, 109-113. DeLoache, J. S., Miller, K. E, & Pierroutsakos, S. L. (1998). Reasoning and problem solving. In D. Kuhn & R. Siegler (Eds.), Handbook of Child Psychology, 5th ed., Cognitive development (pp. 801-850). New York: Wiley. DeLoache, J. S., Miller, K. E, & Rosengren, K. S. (1997). The credible shrinking room: Very young children's performance with symbolic and nonsymbolic relations. Psychological Science, 8, 308313. Fernyhough, C. (1996). The dialogic mind: A dialogic approach to the higher mental functions. New Ideas in Psychology, 14, 47-62. Fischer, K. W., & Kennedy, B. P. (1997). Tools for analyzing the many shapes of development: The case of self in relationships in Korea. In E. Amsel & K. A. Renniger (Eds.), Change and development: Issues of theory, method, and application (pp. 117-152). Hillsdate, NJ: Erlbaum. Flavell, J. H. (1963). The developmental psychology of Jean Piaget. Princeton, NJ: Van Nostrand.

Variability in Children's Reasoning

37

Flavell, J. H., Green, E L., & Flavell, E. R. (1995). Young children's knowledge of thinking. Monographs of the Society for Research in Child Development, 60, (1, Serial No. 243). Gelman, S. A., & Markman, E. M. (1986). Categories and induction in young children. Cognition, 23, 183-209. Ginsburg, H. E, & Opper, S. (1988). Piaget's theory of intellectual development (3rd ed.). Englewood Cliffs, NJ: Prentice Hall. Goldberger, A. L., Rigney, D. R., & West, B. J. (1990). Chaos and fractals in human physiology. Scientific American, 262, 42-49. Goldfield, E. C. (1995). Emergent forms: Origins and early development of human action and perception. New York: Oxford University Press. Goldfield, E. C., Kay, B. A., & Warren, W. H. (1993). Infant bouncing: The assembly and tuning of action of systems. Child Development, 64, 1128-1142. Goldin-Meadow, S. (1997). When gestures and words speak differently. Current Directions in Psychological Science, 6, 138-143. Goldin-Meadow, S., Alibali, M. W., & Church, B. (1993). Transitions in concept acquisition: Using the hand to read the mind. Psychological Review, 100, 279-297. Gopnik, A., & Wellman, H. M. (1994). The theory. In L. A. Hirschfeld & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 257-293). New York: Cambridge University Press. Gottman, J. M. (1981). Time-series analysis for social scientists. Cambridge, England: Cambridge University Press. Hatano, G., & Inagaki, K. (1996). Young children's naive theory of biology. In J. Mehler & S. Franck (Eds.), Cognition on cognition: Cognition special series (pp. 153-170). Cambridge, MA: MIT Press. Hatano, G., Siegler, R. S., Richards, D. D., Inagaki, K., Stavy, R., & Wax, N. (1993). The development of biological knowledge: A multi-national study. Cognitive Development, 8, 47-62. Hirschfeld, L. A., & Gelman, S. A. (1994). Mapping the mind: Domain specificity in cognition and culture. New York: Cambridge University Press. Howe, M. L. (1994). Dynamics of cognitive development: A unifying approach to universal trends and individual differences. Learning and Individual Differences, 6, 365-377. Karmiloff-Smith, A. (1992). Beyond modularity. Cambridge, MA: MIT Press. Keil, E C. (1989). Concepts, kinds, and cognitive development. Cambridge, MA: MIT Press. Kirk, R. E. (1978). Introductory statistics. Pacific Grove, CA: BrooksCole. Kerkman, D. D., & Siegler, R. S. (1997). Measuring individual differences in children's addition strategy choices. Learning and Individual Differences, 9, 1-18. Kuhn, D. (1995). Microgenitic study of change: What has it told us ? Psychological Science, 6, 133-139. Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago: University of Chicago Press. Leslie, A. M. (1994). Tomm ToBy and agency: Core architecture and domain specificity in cognition and culture. In L. A. Hirschfeld & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 119-148). New York: Cambridge University Press. Lillard, A. (1998). Ethnopsychologies: Cultural variations in theories of mind. Psychological Bulletin, 123, 3-32. Lillard, A. (1999). Developing a cultural theory-of-mind: The CIAO approach. Current Directions in Psychological Science, 8, 57-61. Lillard, A., & Flavell, J. H. (1992). Young children's understanding of different mental states. Developmental Psychology, 28, 626-634. Marr, D. (1982). Vision. New York: Freeman. Miller, K. E, Smith, C. M., Zhu, J., & Zhang, H. (1995). Preschool origins of cross-national differences in mathematical competence: The role of number naming systems. Psychological Science, 6, 56-60.

38

Karl S. Rosengren and Gregory S. Braswell

Miller, E H., & Coyle, T. R. (1999). Developmental change: Lessons from microgenesis. In E. K. Scholnick, K. Nelson, S. A. Gelman, & P. H. Miller (Eds.), Conceptual development: Piaget's legacy (pp. 209-239). Mahwah, NJ: Erlbaum. Newell, K. M., van Emmerik, R. E. A., Lee, D., & Sprague, R. L. (1993). On postural variability. Gait Posture, 4, 225-230. Nguyen, S., & Rosengren, K. S. (2001). Biological and magical explanations for illness among European-American and Vietnamese-American children. Manuscript under review. Overton, W. E (1998). Developmental psychology: Philosophy, concepts, and methodology. In R. M. Lerner (Ed.), Handbook of Child Psychology, 5th ed., Theoretical Models of Human Development (pp. 107-185). New York: Wiley. Perner, J. (1989). Is "thinking" belief? Reply to Wellman Bartsch, Cognition, 33, 315-319. Perner, J., & Davies, G. (1991). Understanding the mind as an active information processor: Do young children have a "copy theory-of-mind"? Cognition, 39, 51-69. Perner, J., Ruffman, T., & Leekman, S. R. (1994). Theory-of-mind is contagious: You catch it from your sibs. Child Development, 65, 1228-1238. Perry, M., Church, R. B., & Goldin-Meadow, S. (1992). Is gesture-speech mismatch a general index of transitional knowledge? Cognitive Development, 7, 109-122. Piaget, J. (1941). Le m6canisme du d6veloppement mental et les lois du groupement des op6rations. Archives de Psychologie, Genkve, 28, 215-285. Piaget, J. (1955). Les stades du d6veloppement intellectuel de 1' enfant et de 1' adolescent. In E Osterrieth et al., Le problkme des stades en psychologie de l'enfant (pp. 33-113). Paris: Presses Universitaires de France. Piaget, J. (1968). Introduction. In E Laurendeau (Ed.), Les premikres notions spatiales de l' enfant. Neuch~tel: Delachaux et Niestle. Piaget, J. (1971). The theory of stages in cognitive development. In D. R. Green, M. E Ford, & G. B. Flamer (Eds.), Measurement and Piaget (pp.l-11). New York: McGraw-Hill. Piaget, J. (1972). Intellectual evolution from adolescence to adulthood. Human Development, 15, 1-12. Piaget, J., & Garcia, R. (1971). Les explications causales. Paris: Presses Universitaires de France. Piaget, J., & Inhelder, B. (1941). Le d~veloppement des quantit~s chez I' enfant. Neuch~tel: Delachaux et Niestle. Reese, H. W., & Overton, W. E (1970). Life-span developmental psychology: Research and theory. New York: Academic Press. Rogoff, B., & Morelli, G. (1989). Perspectives on children's development from cultural psychology. American Psychologist, 44, 343-348. Rosengren, K. S., Behrend, D., & Perlmutter, M. (1993). Parental influences on children's cognition. In R. Pasnak & M. Howe (Eds.), Emerging themes in cognitive development (pp. 103-127). New York: Springer-Verlag. Rosengren, K. S., DeHart, G., Taylor, M. G., & Marzolf, D. E (2001). Kittens & cats, caterpillars & butterflies: Young children's understanding of variation in parent-offspring resemblance. Manuscript under review. Rosengren, K. S., Gelman, S. A., Kalish, C., & McCormick, M. (1991). As time goes by: Children's early understanding of biological growth. Child Development, 62, 1302-1320. Rosengren, K. S., & Hickling, A. K. (2000). Metamorphosis and magic: The development of children's thinking about possible events and plausible mechanisms. In K. S. Rosengren, C. N. Johnson, & E Harris (Eds.), Imagining the impossible: Magical, scientific, and religious thinking in children (pp. 75-98). Cambridge, England: Cambridge University Press. Ruffman, T., Perner, J., Naito, M., Parkin, L., & Clements, W. A. (1998). Older (but not younger) siblings facilitate false belief understanding. Developmental Psychology, 34, 161-174. Samarapungavan, A., Vosniadou, S., & Brewer, W. E (1996). Mental models of the earth, sun, and moon: Indian children's cosmologies. Cognitive Development, 11, 491-521.

Variability in Children's Reasoning

39

Schafer, W. D. (1980). Assessment of dispersion in categorical data. Educational and Psychological Measurement, 40, 879-883. Schauble, L. (1990). Belief revision in children: The role of prior knowledge and strategies for generating evidence. Journal of Experimental Child Psychology, 49, 31-57. Siegal, M. (1997). Knowing children: Experiments in conversation and cognition (2nd ed.). Hove, England: Psychology Press. Siegler, R. S. (1985). Five generalizations about cognitive development. American Psychologist, 38, 263-277. Siegler, R. S. (1994). Cognitive variability: A key to understanding cognitive development. Current Directions in Psychological Science, 1, 1-5. Siegler, R. S. (1996). Emerging minds: The process of change in children's thinking. New York: Oxford University Press. Siegler, R. S., & Crowley, K. (1991). The microgenetic method: A direct means for studying cognitive development, American Psychologist, 46, 606-620. Siegler, R. S., & Ellis, S. (1996). Piaget on childhood. Psychological Science, 7, 211-215. Siegler, R. S., & McGilly, K. (1989). Strategy choice in children's time-telling. In I. Levin & D. Zakay (Eds.), Time and human cognition: A life span perspective. The Netherlands: Elsevier. Siegler, R. S., & Shrager, J. (1984). Strategy choices in addition and subtraction: How do children know what to do? In C. Sophian (Eds.), Origins of cognitive skills. Hillsdale, NJ: Erlbaum. Smith, L. B., & Thelen, E. (1993). A dynamic systems approach to the development of perception and action. Cambridge, MA: MIT Press. Smith, L. B., Thelen, E., Titzer, R., & McLin, D. (1999). Knowing in the context of acting: The task dynamics of the A-not-B error. Psychological Review, 106, 235-260. Smith, P. K., & Dutton, S. (1979). Play and training in direct and innovative problem solving. Child Development, 50, 830-836. Sodian, B., Taylor, C., Harris, P. L., & Perner, J. (1991). Early deception and the child's theory-of-mind: False trails and genuine markers. Child Development, 62, 468-483. Solomon, G. E. A., Johnson, S. C., Zaitchik, D., & Carey, S. (1996). Like father, like son: Young children's understanding of how and why offspring resemble their parents. Child Development, 67, 151-171. Spelke, E. S. (1991). Physical knowledge in infancy. In S. Carey & R. Gelman (Eds.), The epigenesis of mind: Essays on biology and cognition (pp.133-169). Hillsdale, NJ: Erlbaum. Sperber, D. (1994). The modularity of thought and epidemiology of representations. In L. A. Hirschfeld & S. A. Gelman (Eds.), Mapping the mind. New York: Cambridge University Press. Springer, K., & Keil, E (1989). On the development of biologically specific beliefs: The case of inheritance. Child Development, 60, 637-648. Springer, K., & Keil, E C. (1991). Early differentiation of causal mechanisms appropriate to biological and nonbiological kinds. Child Development, 62, 767-781. Stevenson, H. W., Lee, S., Chen, C., Stigler, J. W., Hsu, C., & Kitamura, S. (1990). Contexts of achievement: A study of American, Chinese, and Japanese children. Monographs of the Society for Research in Child Development, 55, (1 - 2, Serial No. 221). Subbotsky, E. V. (1985). Preschool children's perception of unusual phenomena. Soviet Psychology, 23, 91-114. Subbotsky, E. V. (1990). Phenomenal and rational perception of some object relations by preschoolers. Soviet Psychology, 28, 5-24. Subbotsky, E. V. (1992). Foundations of the mind: Children's understanding of reality. New York: Harvester Wheatsheaf. Subbotsky, E. V. (1997). Explanations of unusual events: Phenomenalistic causal judgments in children and adults. British Journal of Developmental Psychology, 15, 13-36. Subbotsky, E. V. (2000). Phenomenalistic perception and rational understanding in the mind of an

40

Karl S. Rosengren and Gregory S. Braswell

individual. In K. S. Rosengren, C. N. Johnson, & E Harris (Eds.), Imagining the impossible: Magical, scientific, and religious thinking in children (pp. 35-74). Cambridge, England: Cambridge University Press. Taylor, M., & Carlson, S. M. (1997). The relation between individual differences in fantasy and theoryof-mind. Child Development, 68, 436-455. Taylor, M., & Carlson, S. M. (2000). The influence of religious beliefs on parental attitudes about children's fantasy behavior. In K. S. Rosengren, C. N. Johnson, & P. Harris (Eds.), Imagining the impossible: Magical scientific, and religious thinking in children (pp. 247-268). Cambridge, England: Cambridge University Press. Thelen, E. (1996). Motor development: A new synthesis. Annual Progress in Child Psychiatry Child Development, 32-66. Thelen, E., & Smith, L. B. (1994). A dynamic systems approach to the development of cognition and action. Cambridge, MA: MIT Press. Thelen, E., & Spencer, J. P. (1998). Postural control during reaching in young infants: A dynamic systems approach. Neuroscience Biobehavioral Reviews, 22, 507-514. Valsiner, J. (1984). Two alternative epistemological frameworks in psychology: The typological and variational modes of thinking. Journal of Mind and Behavior. 5, 449-470. van Geert, P. (1991). A dynamic systems model of cognitive and language growth. Psychological Review, 98, 3-53. van Geert, P. (1993). A dynamic systems model of cognitive growth: Competition and support under limited resources. In L. B. Smith & E. Thelen (Eds.), A dynamic systems approach to development: Applications (pp. 265-331). Cambridge, MA: MIT Press. van Geert, P. (1995). Vygotskian dynamics of development. Human Development, 37, 346365. van Geert, P. (1998). A dynamic systems model of basic developmental mechanisms: Piaget, Vygotsky, and beyond. Psychological Review, 105, 634-677. van Greet, P. (1999). Vygotsky's dynamic systems. In P. Lloyd & C. Fernyhough (Eds.), Lev Vygotsky: Critical assessments: Future directions (Vol. 4, pp. 3-21). New York: Routledge. Vosniadou, S. (1994). Universal and culture-specific properties of children's mental models of the earth. In L. A. Hirschfeld & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 412-430). New York: Cambridge University Press. Vosniadou, S., & Brewer, W. E (1992). Mental models of the earth: A study of conceptual change in childhood. Cognitive Psychology, 24, 535-585. Vosniadou, S., & Brewer, W. E (1994). Mental models of the day/night cycle. Cognitive Science, 18, 123-183. Vygotsky, L. S. (1978). Mind in society. Cambridge, MA: Harvard University Press. Wellman, H. M. (1990). The child's theory of mind. Cambridge: MIT Press. Wellman, H. M., & Bartsch, K. (1988). Young children's reasoning about beliefs. Cognition, 30, 239277. Wellman, H. M., & Bartsch, K. (1989). Three-year-olds understanding of belief: A reply to Perner. Cognition, 33, 321-326. Wellman, H. M., & Gelman, S. A. (I 998). Knowledge acquisition in foundational domains. In D. Kuhn & R. S. Siegler (Eds.), Handbook of Child Psychology, Vol 2: Cognition, Perception, and Language (pp. 523-573). New York: John Wiley & Sons. Wertsch, J. V. (1991). Voices of the mind: A sociocultural approach to mediated action. Cambridge, MA: Harvard University Press. Wertsch, J. V., & Tulviste, P. (1992). L. S. Vygotsky and contemporary developmental psychology. Developmental Psychology, 28, 548-557.

FUZZY-TRACE THEORY: DUAL PROCESSES IN MEMORY, REASONING, AND COGNITIVE NEUROSCIENCE

C. J. Brainerd* 1 and V. E Reyna 2 IDEPARTMENT OF SPECIAL EDUCATION REHABILITATION, AND SCHOOL PSYCHOLOGY 2DEPARTMENTS OF SURGERY AND MEDICINE ARIZONA HEALTH SCIENCES CENTER UNIVERSITY OF ARIZONA TUCSON, ARIZONA 85721

I. INTRODUCTION II. M E M O R Y - R E A S O N I N G RELATIONS: NECESSITY, CONSTRUCTIVISM, AND DISSOCIATION A. EVIDENCE FOR NECESSITY B. EVIDENCE FOR CONSTRUCTIVISM III. DUAL-PROCESS CONCEPTIONS OF REASONING, MEMORY, AND M E M O R Y - R E A S O N I N G RELATIONS A. DUAL PROCESSES IN REASONING B. DUAL PROCESSES IN MEMORY C. M E M O R Y - R E A S O N I N G RELATIONS, REDUX IV. FOUR RESEARCH DOMAINS A. DEVELOPMENTAL COGNITIVE NEUROSCIENCE OF FALSE MEMORY B. STUDIES OF FALSE MEMORY IN BRAIN-DAMAGED PATIENTS C. DUAL-MEMORY PROCESSES IN JUDGMENT AND DECISION MAKING D. RETRIEVAL MECHANISMS IN CHILD AND ADULT RECALL V. SUMMARY REFERENCES

*Address for correspondence: C. J. Brainerd, Department of Special Education, Rehabilitation, and School Psychology, University of Arizona, Tucson, AZ 85721. E-Mail may be sent to brainerd @u.arizona.edu.

41 ADVANCESIN CHILDDEVELOPMENT AND BEHAVIOR,VOL.28

Copyright 9 2001 by AcademicPress. All rightsof reproductionin any formreserved. 0065-2407/01 $35.00

42

C.J. Brainerd and V. F. Reyna

I. Introduction The science of cognitive development is divisible into two broad fields of inquiry: (a) the ontogenesis of basic processes and capabilities (e.g., attention, memory, processing speed, inhibition) that support higher reasoning and (b) the ontogenesis of types of higher reasoning (e.g., decision making, deductive inference, judgment, problem solving). The paradigms that are central to research in the two fields are quite different, of course (e.g., free recall vs. mathematical problem solving), as are the theoretical assumptions that guide experimentation. At a deeper level, however, theories in both fields have long shared a core attributemnamely, they have been unitary theories. By "unitary theories," we mean theories in which an overriding developmental bottom line is posited; that is, a unifying theme or principle that characterizes what development is moving away from and what it is evolving toward. In the memory sphere, the unifying theme is that development moves away from recollections that are sketchy, distorted representations of experience and toward recollections that are detailed, veridical representations. In the reasoning sphere, the unifying theme is that development moves away from reasoning operations that are intuitive, qualitative, or heuristic toward reasoning operations that are analytical, computational, or logical. From the perspective of parsimony, unitary conceptions are the proper place to begin theory building. However, research in mainstream cognitive psychology has long since established that some basic facts of adult memory and reasoning pose serious difficulties for such conceptions. In adults, both memory and reasoning seem to be characterized by well-articulated dual systems that process different types of information in seemingly contradictory and incompatible ways. In the memory sphere, the same adults who exhibit highly veridical recall and recognition of experience also display powerful illusions of recollection for events that were never experienced (for a review, see Roediger, 1996). In the reasoning sphere, the same adults who provide analytical, logical solutions to certain problem-solving tasks also provide intuitive, heuristic solutions when the tasks are framed in slightly different ways (e.g., Kahneman, Slovic, & Tversky, 1982; Tversky & Kahneman, 1981, 1983). Such evidence poses two fundamental challenges to theories of cognitive development: first, to formulate testable theoretical models of these dual systems and, second, to spell out their developmental traj~,'tories. Fuzzy-trace theory (Brainerd & Reyna, 1990a; Reyna & Brainerd, 1995; Reyna, Lloyd, & Brainerd, 2001) is an attempt to respond to these challenges by integrating two disparate approaches to the study of memory (the constructivist and verbal-learning traditions) and by uniting two similarly disparate approaches to the study of reasoning (the intuitionist and logicist traditions). We survey the current status of fuzzy-trace theory in this article.

Dual Processes in Fuzzy-Trace Theory

43

In section II, we sketch two prominent interpretations of memory-reasoning relations in cognitive development that preceded fuzzy-trace theory, the necessity hypothesis and the constructivist hypothesis, and that were predicated on unitary conceptions of memory and reasoning. We summarize developmental findings that disconfirmed both interpretations and motivated fuzzy-trace theory as an alternative approach. In section III, we consider dual-process conceptions of reasoning, memory, and their relations, as posited in fuzzy-trace theory. Research that bears on these conceptions is also reviewed. In section IV, we examine four domains of research that have been active venues for evaluating the explanatory and predictive capabilities of fuzzy-trace theory: (a) developmental cognitive neuroscience studies of false memory; (b) studies of false memory in brain-damaged patients; (c) studies of judgment and decision-making errors in adults; and (d) studies of dual-retrieval processes in child and adult recall.

II. Memory-Reasoning Relations" Necessity, Constructivism, and Dissociation Fuzzy-trace theory was originally motivated by several considerations, some theoretical and others empirical (for a review, see Brainerd & Reyna, 1993a). A pivotal early factor was the accumulation of counterintuitive developmental data on the seemingly simple question, How is memory development related to the development of specific reasoning abilities? Until the mid- 1980s, there were two conventional answers. The first, which figured in both the Piagetian and information processing traditions (Bryant & Trabasso, 1971; Elkind, 1967; Smedslund, 1969), was the necessity hypothesis. Maintenance of veridical representations of problem information (e.g., the premises in transitive or conditional inference problems and the numerical values of sets in probability judgment problems) until reasoning operations could be completed was treated as an essential, though not sufficient, condition for valid reasoning--hence, "necessity" hypothesis. The constructivist hypothesis evolved from Bartlett's (1932) work in social psychology and characterized Piaget's later views on memory-reasoning relations (e.g., Piaget & Inhelder, 1973). Here, the core idea was that memory traces are not bedrocks for accurate reasoning but, rather, are themselves modified ("constructed") by reasoning operations (Brainerd & Reyna, 1995; Reyna & Kiernan, 1994). In other words, "The schemata of the memory are borrowed from the intelligence" (Piaget & Inhelder, 1973, p. 382). The necessity and constructivist hypotheses, though contradictory, were both supported by lines of evidence that were superficially compelling. Regarding necessity, for instance, some investigators used overlearning trials to ensure that children would remember the exact problem information that accompanies Piaget's concrete-operational reasoning tasks. In this procedure, before reasoning problems

44

C.J. Brainerd and V. E Reyna

are administered, children are trained on crucial background information until they perform perfectly on memory tests for that information. Overlearning of problem information produced dramatic improvements in children's reasoning accuracy (e.g., Trabasso, 1977). Regarding constructivism, recognition memory experiments with adults (Bransford & Franks, 1971) and children (Paris & Carter, 1973) seemed to demonstrate that participants remembered what they understood, not what they actually experienced. A key finding was that adults and children often failed to discriminate material that they had actually experienced (e.g., the statements "The cage is on the table" and "The cat is in the cage") from unpresented material that integrated the meaning of experience (e.g., "The cat is on the table"). This finding was said to show that (a) memory for experience takes the form of a unitary semantic code that lacks individuating surface details and that (b) responses to recognition probes depend on whether probes are consistent with the meaning of experience (i.e., "The cage is on the table" and the "The cat is on the table" are both true, based on the meaning of presented material). A. EVIDENCE FOR NECESSITY

In the late 1980s and early 1990s, both types of findings were cited in cognitive development textbooks (e.g., Bjorklund, 1989; Siegler, 1991) as establishing that, without deciding between the two hypotheses, accurate reasoning and accurate memory must be closely intermeshed. However, Brainerd and Kingma (1984) pointed out that such findings did not actually establish direct links between solutions to specific childhood reasoning problems (conservation, probability judgment, transitive inference, and so forth) and memory for the specific background facts that authorize those solutions. To close this gap, they investigated the relation between children's (5- to 8-year-olds') memory for individual premises on transitive inference problems ("John is taller than Jim" and "Jim is taller than Don") and their ability to make correct inferences from the premises ("John is tallest"). This required only the minor methodological adjustment of adding premise-memory probes to each transitive inference problem. The resulting experiments were focused on the necessity hypothesis, which predicts that if accurate premise memory is a precondition for valid reasoning, children will be more likely to make correct inferences when they can remember the premises for a problem than when they cannot. The necessity hypothesis also predicts that relations between performance on premise-memory tests and transitivity problems should be stronger when memory tests are administered before rather than after transitivity problems because children may begin to delete premise information from memory after reasoning is complete. Although Piaget (1970) had remarked some years before that children obviously cannot make transitive inferences without accurate memory for premises, Brainerd and Kingma's (1984) data did not agree. Across several variations of the transitive

Dual Processes in Fuzzy-Trace Theory

45

._I

< z

POSITIVE DEPENDENCY

O 0.6 I-D m

Z

O r,.) Z 0.2 cO

Z

9

9

9

9

9

m

I 9

m

, -0.2 <

z

O I-.D z -0.6 O o

I EGATIVE DEPENDENCY i ili l

<.6

0.6

0.65

0.7

0.75

0.8

0.85

UNCONDITIONAL INFERENCE PROBABILITY

0.9

Fig. 1. Statistical dissociation between the accuracy of children's performance on transitive inference problems and their performance on memory tests for premises. The plotted data are from experiments by Brainerd and Kingma (1984), Chapman and Lindenberger (1992), and Rabinowitz et al. (1994).

inference paradigm, performance on inference problems was statistically dissociated from performance on memory probes, and the relation was not affected by whether premise-memory probes preceded or followed inference problems. This pattern of memory-reasoning dissociation proved to be highly replicable (e.g., Chapman & Lindenberger, 1992; Rabinowitz, Grant, Howe, & Walsh, 1994). A visual summary of findings from several experiments appears in Figure 1, where the difference between the conditional probability of valid transitive inferences given accurate premise memory and the unconditional probability of valid transitive inferences without regard to premise memory has been plotted for several data sets. Of course, the necessity hypothesis demands that the difference should be large and positive, but the difference was consistently close to zero, indicating that children who did not remember the specific premises for a particular problem were just as likely to make valid inferences as children who remembered those premises. Necessity might still be preserved if transitive inference were an exception to the rule, and other reasoning paradigms might produce data that conform to prediction. To explore this possibility, Brainerd and Kingma (1985) conducted follow-up studies of three other childhood reasoning paradigms, class inclusion, conservation, and probability judgment. Again, memory probes for crucial background

C.J. Brainerd and V. E Reyna

46

< z 0.6 O I-.D Z O O Z 0.2 ._1

m

r El) Z . -0.2 < Z O I--a Z -O.6 O O

POSITIVE DEPENDENCY m -

9

_ n

n --

mm .m

---

9

_ mm

9

mm - -

NEGATIVE DEPENDENCY

<.25 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0165 0.7 0.75 0.8 0.85 0.9 UNCONDITIONAL INFERENCE PROBABILITY

Fig. 2. Statistical dissociation between the accuracy of children's performance on conservation, class-inclusion, and probability-judgment problems and their performance on memory testsfor problem information. The plotted data are from experiments by Brainerd and Kingma (1985).

information (e.g., the numerical values of the target sets in probability judgment) were added to the usual reasoning problems, and the modified tasks were administered to children in the 5- to 8-year-old range. These studies produced two principal findings. First, as can be seen in Figure 2, the dissociation pattern was clear-cut for these other reasoning tasks. Second, in addition to statistical dissociation, manipulations were identified that induced experimental dissociations between reasoning performance and memory performance. Indeed, two types of dissociative manipulations were identified, those that affected memory without affecting reasoning (e.g., making external stores of problem information continuously available during reasoning) and those that affected reasoning without affecting memory (e.g., visual illusions). These results were replicated in later studies (e.g., Brainerd & Reyna, 1993b, 1995). Further, these results are not aberrations of immature cognition because findings were analogous when experiments were performed with adult reasoning problems, such as decision framing (Reyna & Brainerd, 199 lb) and social judgment (Hastie & Park, 1986; Reyna, 1992). B. EVIDENCE FOR CONSTRUCTIVISM

Turning from necessity to constructivism, Reyna and Kiernan (1994) pointed out that constructivism made key predictions that went beyond predictions about

Dual Processes in Fuzzy-Trace Theory

47

relations between children's performance on memory tests versus reasoning problems. Specifically, memory for experience is supposedly based on a unitary semantic code that preserves the meanings of individual events but not their surface details. If so, performance on different memory tests that are consistent with the meaning of experience are all based on the same memory code. The constructivist hypothesis therefore predicts positive correlations between memory tests for material that was actually experienced (e.g., "The cat is in the cage") and memory tests for unpresented material because both are consistent with the meaning of presented material. Reyna and Kiernan studied this prediction in experiments with childeren in the 6- to 9-year-old range. Children were exposed to a series of three-sentence vignettes about familiar objects. Some of the statements specified familiar spatial relations between the objects (e.g., "The cat is in the cage"; "The cage is on the table"; "The cat has black fur") and some specified familiar magnitude relations (e.g., "The coffee is hotter than the tea"; "The tea is hotter than the cocoa"; "The cocoa is sweet"). Statistical relations were computed between memory tests for presented material and for unpresented material that preserved the meaning of experience. Contrary to constructivism, Reyna and Kiernan (1994) found that both comparisons yielded statistical dissociations. In addition, manipulations were identified that induced experimental dissociations between performance on different memory tests that, according to constructivism, tapped the same semantic code. For example, an overall pattern for spatial versus magnitude vignettes was that levels of acceptance were higher for presented sentences from spatial vignettes than for presented sentences from linear vignettes but the r e v e r s e was true for unpresented sentences that were consistent with the meaning of vignettes. Moreover, developmental improvements in performance on different types of memory tests that, according to constructivism, tapped the same semantic code were independent of each other. (According to fuzzy-trace theory, tests that tap verbatim memory should be independent of tests that tap inferences from gist.) Reyna and Kiernan's findings were replicated in later experiments that focused on children's memory for metaphorical statements (Reyna & Kiernan, 1995), for vignettes containing numerical information (Brainerd & Gordon, 1994), and for word lists (Marx & Henderson, 1996).

III. Dual-Process Conceptions of Reasoning, Memory, and Memory-Reasoning Relations Rethinking conventional ideas as to how reasoning operations process memories of problem information is essential to explain memory-reasoning dissocations. In fuzzy-trace theory, this rethinking takes the form of dual-process models of reasoning and its development (Brainerd & Reyna, 1990a, 1992, 1993a; Reyna &

C. J. Brainerd and V. F. Reyna

48

15 BLUE

15 BLUE

10 BLUE

10 YELLOW

10 YELLOW

15 YELLOW

5 RED

5 RED

5 RED

2 GREEN

2 GREEN

2 GREEN

A Which color will be drawn?

B

Which container would you pick to draw a blue token?

Fig. 3. One- and two-choice probability-judgment problems in which pairwise or global gist is sufficient for correct solution (from Reyna & Brainerd, 1994).

Brainerd, 1990, 1992, 1993) and of memory and its development (Brainerd & Gordon, 1994; Brainerd & Reyna, 1995; Reyna, 1992, 1995; Reyna & Brainerd, 1992, 1995). We discuss proposals about reasoning and memory separately. To preview the outcome of those proposals (see Section III.C), fuzzy-trace theory explains memory-reasoning dissociations on the grounds that preferred modes of processing on reasoning problems tap one of the dual-memory systems and preferred modes of processing on memory tests tap the other system. This distinction is not intrinsic, however. Reasoning tests typically require less precise representations (fuzzy gists) than memory tests (verbatim representations). A. DUAL PROCESSES IN REASONING

Three core principles of reasoning were formulated on the basis of developmental studies of memory-reasoning relations: gist extraction, fuzzy-to-verbatim continua, and fuzzy-processing preferences. For the sake of clarity, we exemplify all three with a familiar paradigm that produced some of the early findings of memory-reasoning dissociation, probability judgment. Illustrative problems are displayed in Figures 3 and 4. As can be seen, the pertinent background facts are the cardinal numbers of sets of objects (colored tokens in Figures 3 and 4) that compose spaces from which elements are randomly sampled. Reasoning problems consist of making predictions about which elements are most likely to result from

Dual Processes in Fuzzy-Trace Theory

49

if]if] DD

D F1 ff] A

DD DD B

Which container would you pick to draw a blue token? Fig. 4. A two-choice probability-judgment problem from Reyna and Brainerd (1994) in which pairwise and global gist produce errors.

random draws from different spaces (Figure 3, left) or making predictions about which spaces provide the best chances of obtaining specific elements on random draws (Figure 3, right, and Figure 4).

1. Reasoning Principles With respect to the gist-extraction principle, many results converge on the notion that when children or adults encode the background facts that accompany reasoning problems, the memories that are stored go beyond the surface form of those facts to include senses, patterns, and meanings that the facts instantiate. That is, although reasoners store memories of the background facts per se, they also use those facts as retrieval cues to access relevant global concepts in long-term memory and to assign local (i.e., task specific) episodic interpretations to them. For instance, the background facts in probability-judgment problems are discrete integers (6 red tokens and 2 blue tokens in Space B in Figure 4), but even preschoolers are found to store two further classes of gist memories (Brainerd, 1981; Reyna & Brainerd, 1994; Spinillo & Bryant, 1991): pairwise relations (e.g., "more blues than reds" for all of the sampling spaces in Figures 3 and 4, and "more blues on that side" for the sampling spaces in Figure 4) and global relations (e.g., "the blues are most" and "the greens are least" for the spaces in Figure 3, and "the blues are less than half" for the spaces in Figure 4). This process is the principle of gist extraction, although encoding also leads to the formation of verbatim memories. The gist-extraction principle means that reasoners are equipped with an assortment of memories of problem information and that they may therefore process memories other than those that represent the detailed surface content of that information. These memories will vary in their degree of specificity with respect to the

50

C. J. Brainerd and V. E Reyna

background facts. The notion of fuzzy-to-verbatim continua refers to hypothetical arrangements of the memories that compose such assortments, according to their inherent levels of specificity (e.g., on the left side of Figure 3, a possible fuzzyto-verbatim continuum is "blues are most" > "more blues than yellows" 15 blues/10 yellows/5 reds/2 greens). The third principle, the fuzzy-processing preference, refers to preferred modes of processing the contents of fuzzy-to-verbatim continua. In classical approaches to cognitive development, such as Piagetian, Wernerian, or information processing theories, advanced reasoning is equated with logico-mathematical operations that are performed on memories of the precise content of background facts. In probability judgment, for instance, Piaget proposed that such judgments are based on compensatory quantitative operations that compute ratios by processing the numerical values of individual sets (e.g., Inhelder & Piaget, 1958). However, data on both children and adults demonstrate that such reasoning is not prototypical. On the contrary, performance data for tasks as varied as class inclusion, conditional inference, decision framing, metaphorical reasoning, and transitive inference suggest that reasoning gravitates toward processing the gist of experience rather than its detailed content (for a review, see Reyna & Brainerd, 1991 a). Emprically, the working procedure (the fuzzy-processing preference) seems to be that reasoning is disposed to operate on the simplest and least precise representations that will get the job done (i.e., that will generate task-specific solutions). (Hereafter, reasoning operations that involve stepwise logico-mathematical processing of memories of the precise surface content of background facts will be referred to as analytical reasoning; reasoning operations that involve global, heuristic processing of gist memories of those facts will be referred to as intuitive reasoning.)

2. Evidence for the Principles At first glance, the notion that sophisticated adult reasoning is engineered to operate on fuzzy, gistified information (or that this preference becomes more pronounced with development) seems far fetched. If the ultimate aim of reasoning is to arrive at responses that are objectively correct, the opposite of the fuzzy-processing preference, a predisposition to process the most precise memory information available would seem to be a self-evident prerequisite. Upon closer inspection, however, two offsetting considerations emerge. First, in most problem-solving situations, correct responses do not demand such processing (Reyna & Brainerd, 1992). On the contrary, correct responses can also be generated, usually more rapidly and economically, by processing gist. This point is obvious for the two problems in Figure 3. On the one hand, both problems can be solved in the Piagetian formaloperational manner by performing ratio computations on the numerical values of the sets. Such computations would lead to the prediction that a blue token is most likely to be drawn in the problem on the left (because the relevant ratios are blue --.47, yellow --.31, red = .16, and green = .06) and the prediction that A has a better chance than B of yielding a blue token in the problem on the right (because

Dual Processes in Fuzzy-Trace Theory

51

TABLE I Some Advantages of the Fuzzy-Processing Preference Advantage

Explanation

Memory availability/ accessibility

Gist memories are more persistent over time than verbatim memories of precise details, so reasoning is engineered around the type of information that is most available.

Generalizability

Gist memories are applicable to a broader range of problem-solving contexts than memories of particularized problem facts.

Processing simplicity

The reasoning operations that process gist memories are less complicated, less effortful, less time consuming, and less error prone than the reasoning operations that process verbatim memories of precise details.

No necessary gain in accuracy

The increased potential for reasoning accuracy that is conferred by processing verbatim memories of details does not necessarily lead to actual increases in accuracy because of the increased potential for errors that are by-products of complicated, time-consumptive processing.

Parallelism

Gist memories are easy to store in parallel with verbatim memories of the surface form of background facts, by simply using those facts as retrieval cues to access well-learned concepts and patterns in long-term memory.

the relevant blue ratios are A = .47 and B = .31). On the other hand, ratio computations are unnecessary if crude gist memories are processed in the intuitive way: If "the blues are most" is processed by the intuitive principle that the largest set has the best chance of being drawn, one knows that a blue token is most likely to be drawn in the problem on the left. If "more blues in A than B" is processed by the intuitive principle that a set is more likely to be drawn wherever more exemplars of that set are found, one knows that A has a better chance than B of yielding a blue token. Second, beyond this, gist processing confers multiple cognitive advantages that redound to the superior accuracy of the fuzzy-processing preference. Reyna and Brainerd (1992) reviewed experimental evidence on such advantages, which is summarized in Table I. Although the content of gist memories is less precise than the content of verbatim memories, it can be seen in Table I that gist memories are better preserved over time, more widely accessible, more generalizable to different forms of reasoning, and easier to process during the course of generating solutions to problems. Hence, as we discuss in the sequel, advanced forms of reasoning seem to proceed on the basis of simple gist that apparently incorporates only the essential elements of a problem, which means that fuzzy-trace theory differs fundamentally from alternative theories in positing a computation-to-intuition shift with increasing development (e.g., Reyna & Ellis, 1994). Although advantages such as those listed in Table I lend credence to the precept that a fuzzy-processing preference would enhance accuracy in the long run, processing gist is not an infallible method of reasoning, and indeed, it leads to some ubiquitous errors and fallacies. In Figure 4, for instance, a dominant numerical gist

52

C.J. Brainerd and V. E Reyna

for children and adults is "B has more blues than A." If this gist is processed in the intuitive manner, the prediction would be that B has a better chance than A of yielding a blue token, which is incorrect. Correct prediction requires that the numbers of blue tokens in A and B be compared to the numerical values of all the elements in the sampling spaces (the denominators of probability ratios). This might also be done intuitively by estimating that 1 is closer to 3 than 2 is to 8 and that A therefore provides the better chance of yielding a blue token (Reyna & Brainerd, 1994). However, the correct prediction might also be generated analytically by computing the relevant ratios for the sampling spaces. Such processing, which takes denominators into account, either intuitively or analytically, is not usually observed before adolescence (Callahan, 1989; Offenbach, Gruen, & Caskey, 1984), and erroneous predictions based on gist processing that neglects denominators are often made by adults (Reyna, 1991; Reyna & Brainerd, 1993, 1994).

3. Developmental Change Fuzzy-trace theory is also distinguished from most earlier perspectives by the assumption that analytical processing of precise memory information and intuitive processing of gist both improve with age. Of course, the literature showing that analytical reasoning improves with age is vast, and any number of studies of the development of logical, mathematical, scientific, and spatial problem solving could be invoked in support of this contention. Much less is known about the development of intuitive reasoning, largely because of the hoary maxim promulgated in the writings of many theorists (e.g., Piaget, 1967; Werner, 1948) that intuitive thought is a cognitive analogue of the human appendix--a vestigial organ of intelligence that is supplanted by more advanced analytical abilities. For instance, Piaget often referred to the preoperational period as the period of intuitive intelligence (Berlyne, 1970), and Werner regarded intuitive thought as a transitional phase between the protocognition of infants and the advanced reasoning of adults (Langer, 1970). Findings generated by many childhood reasoning paradigms seem to support this maxim. For instance, Piaget's familiar concrete- and formal-operational problems (e.g., balance scale, conservation, class inclusion, perspective taking) are all tasks in which solutions based on some gistified representation of problem information are displaced by solutions that involve analytical exploration of the detailed content of that information. Although such findings have traditionally been viewed as demonstrating that analytical reasoning abilities increase with age, they do not establish that intuitive reasoning atrophies and is replaced by analytical reasoning because they involve an obvious confound: Opportunities to acquire the gist memories and heuristic operations that support intuitive reasoning on tasks such as Piaget's concrete- and formal-operational problems are available much earlier in life than opportunities to acquire the logico-mathematical operations that process the precise content of problem information. The result is what might be called an illusion of replacement

Dual Processes in Fuzzy-Trace Theory

53

in which intuitive solutions appear to be supplanted by analytical ones because the requisite specific knowledge for the latter was previously unavailable. If this confound were eliminated, intuitive reasoning might not be found to wane with age, on the contrary. Developmental studies of Tversky and Kahneman's (1981) influential decision-framing paradigm have confirmed this prediction. In the adult version of this paradigm, participants are told that a disease is about to strike the United States that is expected to kill 600 people, and participants are asked to choose between two treatment programs. In a gain frame, participants are told that Program A will save 200 people for sure and that Program B has a one-third chance of saving 600 people and a two-thirds chance of saving no one. In a loss frame, participants are told that 400 people will die for sure with Program C, and that with Program D there is a two-thirds chance that 600 people will die and a one-third chance that no one will die. Adults exhibit strong preferences for one or the other program, depending on whether the problem is framed positively or negatively. In the positive (gain-frame) version, the problem is phrased as lives saved by each program. In the negative (loss-frame) version, the problem is phrased as lives lost by each program. Participants show a marked preference for Program A in the gain-frame version (i.e., they are risk averse when it comes to saving lives), but they flip-flop and show a marked preference for Program D in the loss-frame version (i.e., they are risk seeking when it comes to losing lives). Quite sensibly from an intuitive (but not a logical) standpoint, participants prefer to be certain that people will live, but they prefer to avoid being certain that people will die. In studies of the representational basis for this framing illusion (Reyna & Brainerd, 1991), participants have been found to process different qualitative gists in the two frames, using the principle that risk is to be avoided for gains and is to be sought for losses. In the gain frame, Program A is gistified as "some people will be saved," and Program B is gistified as "some people will be saved or no one will be saved." In the loss frame, Program C is gistified as "some people will die," and Program D is gistified as "no one will die or some people will die." Developmental studies of decision framing have shown that such intuitive reasoning increases with age. Reyna and Ellis (1994) created a child-appropriate version of this paradigm by translating it into a task involving spinners and the winning of toys (gain frame) or the losing of toys (loss frame). Examples of childappropriate problems that were administered to preschoolers, second graders, and fifth graders are described in Table II. Reyna and Ellis reasoned that these problems eliminate the standard confound in which younger children are less likely to have acquired the knowledge that supports analytical thinking than the knowledge that supports intuitive thinking: In the long run, the two choices in the various problems shown in Table II do not differ in their win-loss chances, which supports the logically correct response of no preference between these alternatives. Consistent with this notion, Reyna and Ellis found that the standard intuitive > analytical age trend was reversed. Their youngest children, preschoolers, exhibited the logically

54

C. J. Brainerd and V. E Reyna

TABLE II Examples of Reyna and Ellis' (1994) Decision-Framing Problems for Children Features of problems

Description of features

Apparatus

Spinners with different amounts of winning (red) and losing (blue) area are shown to children to convey risk. Prizes (Superballs) are shown to children in transparent bags so that they know what is to be won and lost.

Gain-frame tasks

1. Option A - Win one Superball for sure. Option B = Spin the spinner with a 50% chance of winning two balls or winning nothing. 2. Option A = Win four Superballs for sure. Option B -- Spin the spinner with a 50% chance of winning eight balls or winning nothing. 3. Option A = Win 30 Superballs for sure. Option B -- Spin the spinner with a 50% chance of winning 60 balls or winning nothing.

Loss-frame tasks

1. Option C = Lose one of two Superballs for sure. Option D -- Spin the spinner with a 50% chance of losing two balls or keeping two balls. 2. Option C = Lose four of eight Superballs for sure. Option D - Spin the spinner with a 50% chance of losing eight balls or keeping eight balls. 3. Option C -- Lose 30 of 60 Superballs for sure. Option D -- Spin the spinner with a 50% chance of losing 60 balls or keeping 60 balls.

Note. Superballs are a commercial brand of plastic balls that are available in multiple sizes and colors.

ideal pattern of no preference between certain and risky alternatives in either the gain or loss frame. By second grade, however, the logically ideal pattern was no longer present. These children exhibited reverse framing for some problems. The logically ideal pattern was also absent among fifth graders. These children exhibited the standard framing pattern for some problems (i.e., risk avoidance in the gain frame coupled with risk seeking in the loss frame) and the reverse framing pattern for other problems. The standard framing pattern is consistent with assimilating quantitative differences across outcomes and fits the adult qualitative pattern discussed earlier (i.e., being certain to win something but avoiding a certain loss). Reyna and Ellis' (1994) data indicate that contrary to the maxim that intuitive reasoning withers and is supplanted by analytical reasoning, certain forms of intuitive reasoning increase with age. As noted earlier in connection with findings of memory-reasoning dissociation for transitive inference, one might argue that such data are limited to the decision-framing paradigm. However, other developmental studies have demonstrated the same age trend for two paradigms in which adult responses reflect intuitive rather than analytical reasoning" conjunction fallacies (Davidson, 1995) and representativeness (Jacobs & Potenza, 1991). Such results have stimulated a broad-based approach to the development of intuitive and analytical reasoning by Klaczynski and associates (Klaczynski, 2001; Klaczynski & Fauth, 1997; Klaczynski & Narasimham, 1998). To test the hypothesis of parallel development of intuitive and analytical reasoning at a global

Dual Processes in Fuzzy-Trace Theory

55

level, so that developmental patterns are not tied to particular tasks, Klaczynski has adopted a multitask factor analytic strategy, which has two principal features. First, the literature on judgment and decision making was searched to identify tasks that produce adult response patterns that are consistent with analytical reasoning and to identify further tasks that produce adult response patterns that are consistent with intuitive reasoning. Examples of tasks that were classified as analytical are conditional inference, contingency detection, and probability judgment. (An important complexity is that some versions of these tasks produce adult response patterns that are not consistent with analytical reasoning. We do not discuss this possibility here because those versions were not the ones studied by Klaczynski and associates.) Examples of tasks that were classified as intuitive are conjunction fallacies, decision framing, and hindsight bias. Second, a battery composed of tasks from both the analytical group and the intuitive group was administered to children, adolescents, and adults, and the resulting data were subjected to factor analysis. The hypothesis that there are dual-reasoning systems that develop in parallel (Reyna & Brainerd, 199 l a, 1992) leads to a straightforward prediction: Factor analysis should yield two factors, with tasks from the analytical group loading positively on one and tasks from the intuitive group loading positively on the other. This prediction was confirmed. Summing up, in the reasoning domain, fuzzy-trace theory posits that (a) children and adults store memories of the surface form of background facts (verbatim traces), but they also use those facts to access patterns and meanings in long-term memory (gist-extraction principle); (b) stored traces can be thought of as being arranged according to the level of specificity with which they represent problem information (fuzzy-to-verbatim continua); and (c) reasoning tends to operate on the most impoverished traces that will deliver solutions (fuzzy-processing preferences). One might wonder why reasoners would prefer to operate on less-precise gist memories, but several considerations (Table I) imply that this predisposition improves accuracy in the long run by engineering reasoning to process the types of memories that are most likely to be accessible. Contrary to the maxim that intuitive reasoning atrophies with development, fuzzy-trace theory assumes that both analytical and intuitive reasoning undergo developmental improvement. As development proceeds, reasoners tend to rely increasingly on improved gist representations that are globally accurate but that sometimes produce predictable local inaccuracies (Reyna, 2001; Reyna et al., 2001). B. DUAL PROCESSES IN MEMORY

As in cognitive-developmental theory, unitary conceptions have predominated in the mainstream memory literature. Global memory models--SAM (Gillund & Shiffrin, 1984), MINERVA (Hintzman, 1988), and TODAM (Murdock, 1982)-are examples of this tradition. Unlike cognitive-developmental theory, however,

56

C. J. Brainerd and V. F. Reyna

dual-process conceptions have also been prominent in certain segments of the memory literature. In studies of recognition memory, in particular, although signaldetection theory (a unitary conception) has been the most influential tradition (see Snodgrass & Corwin, 1988), an important dual-process approach began with work by Atkinson and associates (Atkinson & Juola, 1973, 1974; Atkinson & Wescourt, 1975; Juola, Fischler, Wood, & Atkinson, 1971). We summarize that approach before describing the dual-process assumptions of fuzzy-trace theory.

1. Recognition Memory In a recognition design, participants study a list of memory targets and then respond to a test on which some probes are targets and others are previously unpresented distractors. The test instructions are to accept targets and reject distractors. Atkinson and associates proposed that hits (acceptances of targets) can be based on either of two distinct mechanisms. The primary ground for acceptance was called familiarity. It was assumed that participants first evaluate global familiarity by accessing a continuous memory-strength scale of the sort that is posited in signal-detection theory. A probe is accepted (as old) if its scale value exceeds some high criterion, or it is rejected (as new) if its scale value falls below some low criterion. If a probe's scale value falls between the two criteria, a secondary list-search process is activated that scans participants' memory store of targets, yielding acceptance if the probe is located in the store and rejection if it is not. Mandler (1980) expanded this account by proposing that the familiarity process arises from perceptual-learning operations that store coherent traces by integrating the surface content of target presentations, and the list-search process (which he called retrieval or recall, but is usually called recollection nowadays) arises from elaborative encoding operations that access the semantic content of targets. As Jacoby (1991) later emphasized, a hallmark of this dual-process model is the prediction of experimental dissociations between the two processesmcircumstances in which manipulations change hit rates by affecting familiarity but not recollection or change hit rates by affecting recollection but not familiarity. Of course, predictions of dissociation can be tested only if familiarity and recollection can be separately measured. To do the needed measurements, Mandler (1980) used some types of memory tests (recall, in particular) as measures of recollection and other types of memory tests (stem and fragment completion, in particular) as measures of familiarity. Consistent with the prediction of familiarity-recollection dissociation, he found that manipulations such as intentional versus incidental learning improved performance on recollection measures without affecting performance on familiarity measures. He also reported that amnesiac patients showed large deficits on recollection measures coupled with normal performance on familiarity measures (but see Howe, Rabinowitz, & Grant, 1993, for discussion of the limitations of such evidence).

Dual Processes in Fuzzy-Trace Theory

57

Two further elaborations have been added to this dual-process view. One idea is that different phenomenologies accompany hits that are based on familiarity versus hits that are based on recollection. This idea evolved from two theoretical distinctions that were introduced during the 1980s, Tulving's (1985) rememberknow distinction and the implicit-explicit distinction of Schacter and associates (e.g., Graf & Schacter, 1987). The exact proposal is that global familiarity evaluations are accompanied by vague, nonspecific feelings of similarity between probes and studied material, and recollection is accompanied by conscious awareness of specific study events (earlier probe presentations echo in the mind's ear or flash in the mind's eye). Second, the notion that familiarity is the primary basis for hits and recollection is the secondary basis has been challenged by several findings. For example, after a hit, participants predominately report conscious awareness of the physical details of probe presentation during the study phase (the subjective experience that accompanies recollection) rather than unanchored feelings of similarity (Gardiner & Java, 1991). Further, when mathematical models are used to distinguish the respective contributions of recollection and familiarity to hit rates, a larger proportion of hits has been found to be due to recollection than to familiarity (Brainerd, Reyna, & Mojardin, 1999). Therefore, as Reyna and Kiernan (1994) concluded, in standard designs in which recognition tests are administered shortly after the study phase, recollection rather than familiarity appears to be the principal process supporting hits.

2. Fuzzy-Trace Theory's Dual-Process Model In fuzzy-trace theory, the dual-process approach to recognition memory has been expanded in two general waysmnamely, by further articulating the two processes and by extending those processes to recall. These expansions are accomplished through four principles derived from research on recognition and recall: (a) parallel verbatim-gist storage; (b) dissociated verbatim-gist retrieval; (c) memorial bases of conscious remembrance; and (d) identify and similarity judgments. We discuss the principles separately and conclude with an examination of developmental change. a. Parallel Verbatim-Gist Storage. Consistent with the gist-extraction principle of reasoning, a variety of evidence supports the notion that in memory experiments, children and adults store verbatim traces (representations of targets' surface forms and other item-specific information) and gist traces (representations of the semantic, relational, and elaborative properties in which targets participate) (for a review, see Reyna & Brainerd, 1995). Because verbatim and gist traces are alternative representations of the same targets, a reasonable working hypothesis is that they are related by serial dependencies. Thus, the classical view of verbatim and gist storage, as formulated in the psycholinguistic (e.g., Jarvella, 1971) and sentence recognition (e.g., Bransford and Franks, 1971) literatures, is that verbatim

58

C.J. Brainerd and V. F. Reyna

traces of targets are first stored in short-term memory, and when substantial verbatim information has accomulated, the gist of that information is extracted (e.g., Gehrig & Healy, 1983). Gist traces are then deposited in long-term memory and the verbatim traces from which they were extracted are erased, so that only gist memories remain a few minutes after study (e.g., Kitsch, Welch, Smalhofer, & Zimny, 1990). Fuzzy-trace theory's view of the relation between verbatim and gist storage is parallelism rather than serial dependency. Verbatim and gist traces are assumed to be stored in parallel, resulting in multiple dissociated representations of a target. Theoretically, parallel storage provides the basis for the fuzzy-to-verbatim continua that we discussed in connection with reasoning. However, this position was not motivated by a need for theoretical coherence but, rather, by several findings from memory experiments that disconfirmed the classical hypothesis of serial dependency between verbatim and gist storage (for reviews, see Brainerd & Reyna, 1998a; Reyna, 1992). Those findings suggest that the encoding of targets during the study phase of a recognition or recall experiment initiates parallel storage mechanisms. One mechanism deposits episodic traces of the exact surface forms of targets (e.g., "TABLE," "CHICAGO"), and the other deposits gist traces of the same targets by using them as retrieval cues to locate relevant concepts (e.g., "furniture," "city") in long-term memory and assigns local interpretations to them (e.g., "furniture -- TABLE," "city = CHICAGO"). Perhaps the most unambiguous lines of experimentation supporting parallel storage are ones that have demonstrated that targets' meanings are stored in memory before participants have finished processing targets' surface forms (for reviews, see Brainerd & Reyna, 1993b; Reyna & Brainerd, 1992). The word-superiority effect (Ankrum & Palmer, 1989) and the missing-letter effect (Moravcski & Healy, 1995) are examples. Concerning the former, when familiar words (e.g., COMPUTER) are presented at very short exposure durations (e.g., less than 100 ms), participants are able to identify the target itself (e.g., to respond correctly to "Was the word COMPUTER?") at shorter presentation durations than they are able to identify its constituent letters (e.g., to respond correctly to "Did the word contain the letter P?"). The other illustrative finding, the missing-letter effect, refers to the fact that when participants are asked to read a passage and circle all instances of a particular letter (e.g., t), they fail to circle some instances in which the letter occurs as part of common words (e.g., the). This effect shows that during reading, the processing of a word begins before the processing of its constituent letters, so that when the word is familiar, participants may process it so quickly that they move on without being aware of the words' precise surface form. Wallace and associates (Wallace, Stewart, & Malone, 1995; Wallace, Stewart, Shaffer, & Barry, 1998) have proposed a theory of parallel surface and semantic processing in word recognition experiments, a theory in which gist memories result from initial encoding of surface information, and verbatim memories are

Dual Processes in Fuzzy-Trace Theory

59

stored as surface processing is completed. Wallace and associates postulated that the encoding of targets during the study phase is divided into two stages, prerecognition (before a target has been identified as a word that the participant knows) and postrecognition (after a target has been identified). During the prerecognition stage, semantic information about possible candidate words is accessed, which allows the target to be identified as a specific word in the lexicon. Processing of the first few letters or phonemes of a target is the source of retrieval cues for semantic information. Once a target is identified, however, encoding becomes focused on fully processing the target's surface form. This two-stage model, in which gist memories are formed in parallel with the processing of surface information, leads to a nonobvious prediction about false alarms (incorrect acceptances) to distractors that share letters and phonemes with targets (e.g., false alarms to the distractors DOMINOES and PIONEER when DOMINEER was a target). According to the model, distractors that share early surface features (DOMINOES) are more likely to be falsely recognized than distractors that share late surface features (PIONEER). This prediction, which was confirmed in several experiments (Wallace et al., 1995, 1998), follows because (a) nontarget words will tend to be accessed during the prerecognition stage rather than the postrecognition stage and (b) those nontargets will be more likely to share initial surface features of targets than later surface features, given that access is driven by the processing of targets' initial surface features.

b. Dissociated Verbatim-Gist Retrieval. Available data (Brainerd & Poole, 1997; Reyna & Titcomb, 1997) appear to be consistent with the view that on immediate memory tests, hits are based predominately on retrieval of verbatim traces rather than gist traces. Reyna and Kiernan (1994) were the first to produce detailed findings on this point. As in section II, the 6- to 9-year-olds in their experiments studied lists of sentences that were organized into groups of three-sentence vignettes. On recognition tests, children responded to target probes, to true distractors that were consistent with target meanings, and to false distractors that violated target meanings. Reyna and Kiernan proposed that as long as verbatim and gist traces are both accessible, as they should be on immediate tests, targets will be better retrieval cues for their verbatim traces than for their gist traces, but meaning-preserving distractors will be better retrieval cues for targets' gist traces (because distractors, by definition, do not have verbatim traces). Although this idea is congruent with the encoding-specificity rule of retrieval (Tulving & Thomson, 1971), it leads to the counterintuitive prediction that, contrary to constructivism, children's correct recognition of targets will be dissociated from their false recognition of distractors with similar meaning. We saw in section II that this prediction was confirmed. Reyna and Kiernan (1994) also reasoned that if this dissociation is due to reliance on verbatim traces as the predominant basis for hits and reliance on gist

60

C. J. Brainerd and V. F. Reyna

traces as the predominant basis for meaning-based false alarms, dissociation should be replaced by associationmfor the identical sentencesmif children are instructed to focus solely on meaning when responding to recognition tests. This prediction, too, was confirmed in several comparisons. Thus, two patterns emerged from Reyna and Kiernan's experiments that are consistent with dissociated retrieval of verbatim and gist memories: dissociation of hits from false alarms to meaning-preserving distractors on immediate tests, coupled with association of hits and false alarms under conditions that discourage retrieval of verbatim memories. Both of these patterns have been replicated in developmental studies of on-line relations between verbatim and gist memories of numbers (Brainerd & Gordon, 1994), metaphors (Reyna & Kiernan, 1995), words (Brainerd, Reyna, & Kneer, 1995; Marx & Henderson, 1996), and stories (Ackerman, 1992, 1994). Although such findings suggest that the retrieval cues that are provided on memory tests (targets versus meaning-preserving distractors) determine whether verbatim or gist traces are accessed, another important factor is delay. Several types of findings converge on the conclusion that as time passes, verbatim traces become inaccessible more rapidly than gist traces (Gernsbacher, 1985; Kitsch et al., 1990; Murphy & Shapiro, 1994; Reese, 1977; Reyna, 1992). The implication is that on recognition tests, the immediate basis for hits (verbatim retrieval) fails more rapidly than that for false alarms to meaning-preserving distractors (gist retrieval). An obvious prediction is that the independence result for hit and falsealarm rates on immediate tests should be replaced by positive dependency on delayed tests. Reyna and Kiernan (1994) obtained confirmatory data on one-week delayed tests. Their results were later replicated on long-term retention tests for words (Brainerd, Reyna, & Kneer, 1995; Marx & Henderson, 1996). A second and less obvious prediction follows from differential forgetting rates for verbatim and gist memories: Because the memorial basis for false alarms is more stable than that for hits, false alarms will display considerable stability over time, and, under some conditions, they should be more stable than hits. Relevant to this point, Brainerd, Reyna, and Brandse (1995) reported that children's false alarms to words that preserved the meanings of targets were more persistent over a delay than hits were. Brainerd and Mojardin (1998) replicated this result in children using sentence lists, and Payne, Elie, Blackwell, and Neuschatz (1996) replicated it in adults using thematically related word lists. c. Memorial Bases of Conscious Remembrance. We mentioned earlier the assumption that recollection is accompanied by conscious remembrance of having previously studied a specific probe (feeling of remembering, explicit memory) and that familiarity is accompanied by nonspecific impressions of resemblance between a probe and studied material (feeling of knowing, implicit memory). In fuzzy-trace theory, a correlation is posited between these subjective experiences and the retrieval of verbatim and gist memories (Brainerd et al., 1999; Reyna,

Dual Processes in Fuzzy-Trace Theory

61

1998). Feelings of conscious remembrance are said to be predominantly the result of retrieving verbatim memories, and feelings of global resemblance are said to be predominantly the result of retrieving gist memories (Brainerd, Reyna, & Kneer, 1995; Reyna, 1998; but see Reyna & Lloyd, 1997). Verbatim retrieval supports feelings of conscious remembrance because verbatim traces are representations of actual surface content, leading to subjective impressions of reexperiencing targets' physical occurrence. Retrieval of gist memories supports feelings of nonspecific resemblance because gist traces are representations of semantic or other relational information, which leads to subjective impressions that targets of a generic type were presented (e.g., "I heard the name of some article of furniture"). When gist is strengthened, however, for example by repeated cuing of inferences that are never explicitly stated (e.g., Reyna, 1998), memories of semantic gist take on phenomenological vividness and intensity when they are retrieved at test. Although a correlation between verbatim retrieval and subjective experiences of conscious remembrance is assumed, certain findings suggest that phenomenology must be distinguished from memory representations. The findings in question show that under special conditions, feelings of vivid remembrance also result from retrieval of gist memories (Reyna, 1996a, 1998; Reyna & Lloyd, 1997). Brainerd and Reyna (1998b) identified three conditions that can combine to produce this effect: (a) retrieved gist memories are very strong because they have been repeatedly cued during the study phase (e.g., the targets TABLE, SIT, LEGS, COUCH, DESK, RECLINER, SOFA, WOOD, CUSHION, SWIVEL, STOOL, SITTING, ROCKING, and BENCH are all presented), (b) recognition probes are especially good retrieval cues for those strong gists (e.g., TABLE is a better retrieval cue for "furniture" gist than SWIVEL is), and (c) recognition tests encourage responding on the basis of meaning. When recognition probes are targets, one cannot unambiguously separate gist-based feelings of conscious remembrance from verbatim-based feelings. To demonstrate gist-based feelings of conscious remembrance, distractor probes that are exceptionally good retrieval cues for strong gist memories are used (i.e., they satisfy condition [b], although they were not presented). For the furniture illustration, CHAIR would be an example. Because distractors are not presented, verbatim traces were not stored. Yet, on tasks that measure conscious remembrance of probe presentations during the study phase (e.g., identifying the voices in which probes were read and identifying presentation positions in study lists), participants exhibit high levels of conscious remembrance of distractors such as CHAIR, levels that approach the corresponding ones for targets (e.g., Brainerd, Wright, Reyna, & Mojardin, 2001 ; Lampinen, Neuschatz, & Payne, 1998; Payne et al., 1996; Reyna, 1996b).

d. Identity and Similarity Judgments. Fuzzy-trace theory assumes that on immediate recognition tests, hits are predominantly categorical identity judgments. Empirically, support for this assumption comes from an experimental paradigm

62

C.J. Brainerd and V. F. Reyna

called conjoint recognition (Brainerds, Reyna, & Poole, 2000a; Brainerd, Stein, & Reyna, 1998; Brainerd et al., 1999, 2001). In this procedure, participants study a list of targets and respond to a recognition test consisting of targets (e.g., TABLE), gist-consistent distractors (e.g., FURNITURE), and unrelated distractors (e.g., ELEPHANT). Each participant responds to the recognition test under one of three types of instructions: verbatim-only (standard instructions to accept targets and reject distractors), gist-only (accept gist-consistent distractors and reject both targets and unrelated distractors), and verbatim + gist (accept targets and gist-consistent distractors but reject unrelated distractors). For our purposes, the relevant data pertain to the rates of target acceptance in the first two conditions, when those rates have been corrected for any differences in response bias that are induced by different instructions (i.e., acceptance rates are measured by signal-detection indexes such as A', d', or dE). Different mixes of identity and similarity judgments in the standard condition lead to different predictions about the rates of acceptance in the gist-only condition. (An identity judgment is an all-or-none judgment of exact probe-memory match, and a similarity, judgment is a graded judgement of probe-memory resemblance.) Consider three general scenarios. First, suppose that target acceptances in the standard condition are never based on similarity judgments; they are always outcomes of identity judgments or of response bias. If so, acceptance rates will not be significantly above chance (when response bias is equated) in the gist-only condition. Second, suppose that target acceptances in the standard condition are predominantly based on identity judgments but are sometimes based on similarity judgments and sometimes on response bias. If so, the acceptance rate will again drop in the gist-only condition, but owing to similarity judgments, it will be significant (when response bias is equated). Here, the acceptance rate in the gist-only condition (similarity judgments only) divided by the acceptance rate in the standard condition (identity judgments + similarity judgments) is a measure of the proportion of acceptances that are identity judgments. Third, suppose that target acceptances in the standard condition are predominately similarity judgments. If so, the acceptance rate will not drop in the gist-only condition. Data that bear on these three scenarios were reported by Brainerd et al. (1999). The participants studied lists of nouns and responded to tests comprised of targets, gist-preserving nouns, and unrelated nouns. Illustrative results (dL values) come from an experiment in which some targets were studied once, some were studied twice, and the gist-preserving distractors were synonyms (e.g., HILL and GOWN for the targets MOUNTAIN and DRESS, respectively). The data of this experiment support the first (identity judgment only) scenario. With the statistic dE, significant levels of target acceptance correspond to values greater that one. The computed values for the experiment were slightly (and nonsignificantly) less than one for both once-presented and twice-presented targets. Therefore, this result is consistent

Dual Processes in Fuzzy-Trace Theory

63

with Reyna and Kiernan's (1994) conclusion that target acceptance is predominantly an all-or-none (match-mismatch) judgment.

e. Developmental Change. Developmental change occurs in the quality and availability of intuitive and analytical reasoning and, similarly, developmental change occurs in the quality of both verbatim and gist traces, as well as in the ability to access them on memory tests. Indeed, in fuzzy-trace theory, developmental changes in verbatim and gist memory provide the memorial foundation for changes in intuitive and analytical reasoning (Brainerd & Reyna, 1993b; Reyna & Brainerd, 1993, 1995). Pertinent experimental evidence can be found in developmental studies in which the designs supply estimates of the strength of verbatim and gist memory. Standard findings in such studies are that verbatim and gist memory improve steadily between early childhood and young adulthood, although the quantitative details of age variability may differ markedly as a function of task demands, and that the improvements are independent of each other. Two well-known results from developmental studies of free recall are illustrative. First, both recall of individual target items (a measure of verbatim memory) and clustering of meaning-sharing items during recall (a measure of gist memory) improve during the preschool to young adult age years (for a review, see Bjorklund & Muir, 1988). Second, throughout early childhood, middle childhood, and early adolescence, when the bulk of the age improvements in these measures occur, participants' levels of item recall are uncorrelated with their levels of clustering (e.g., DeMarie-Dreblow, 1991). Some developmental studies have dealt with forgetting of verbatim and gist memories (for a review, see Brainerd, 1997). These studies provide separate estimates of the goodness of participants' memory for the exact surface form of targets and the goodness of their memory for the meaning of targets. The studies involved forgetting intervals of varying lengths, and participants ranged in age from early elementary school to late adulthood. Participants at all ages were first trained to criterion during an acquisition session, so that verbatim and gist memory were both at ceiling. Verbatim and gist memory were separately measured on delayed recall tests. The life-span pattern that emerged was that verbatim and gist retention both improved by roughly equal amounts during childhood, they also improved by roughly equal but larger amounts between late childhood and early adolescence, and verbatim retention declined while gist retention remained relatively constant between young and late adulthood. Several studies in the aging literature have provided converging evidence on the last point (e.g., Schacter, Koutstaal, Johnson, Gross, & Angell, 1997; Tun, Wingfield, Rosen, & Blanchard, 1998). Additional evidence indicates that, immediately after storage, the quality of verbatim and gist traces varies as a function of age and that the two types of variability are independent of each other. Experiments by Brainerd and Gordon (1994) are germane because memory tests were administered immediately after presentation

64

C.J. Brainerd and V. F. Reyna

of small amounts of target information, thus providing little opportunity for forgetting of either verbatim or gist memories. The participants were children in the 4- to 8-year-old range. Brief statements containing numerical information (e.g., "Farmer Brown owns 12 cows, 8 sheep, 7 pigs, 5 chickens, and 3 horses") were followed immediately by separate tests for verbatim number memory ("Does Farmer Brown own 7 pigs?"), and for pairwise numerical gist ("Does Farmer Brown own more sheep than chickens?"), and for global numerical gist ("Does Farmer Brown own more cows than anything else?"). After performance on all tests was corrected for age differences in response bias, the correct response probability on verbatim tests improved from .60 to .77. As with the gist measure in Reyna and Kiernan's experiments, the correct response probability on pairwise gist tests improved with age, from .56 to .82, and so did the correct response probability on global gist tests, from .58 to .79. C. MEMORY-REASONING RELATIONS, REDUX

In Section II, we reviewed findings of dissociation between the accuracy of children's reasoning and the accuracy of their memory for crucial problem information. In Section III, we have surveyed research that was stimulated by those findingsm specifically, research on dual processes in reasoning and research on dual processes in memory. Studies of the first type have generated some empirically grounded principles (gist extraction, fuzzy-to-verbatim continua, and fuzzy-processing preference) that exhibit developmental variability, and studies of the second type have spawned some corresponding principles (parallel verbatim-gist storage, dissociated verbatim-gist retrieval, the representational basis of conscious recollection, and identity judgments versus similarity judgments) that are also subject to developmental variability. An important by-product of these principles is that they provide a comprehensive explanation of memory-reasoning independence, which runs as follows. As we saw in Section II, the basic design of developmental studies of memoryreasoning relations involves administering standard childhood reasoning problems (probability judgment, transitive inference, etc.) accompanied by recognition or recall tests for the background facts that authorize solutions (numerical values of sets, premises stating magnitude relations between objects, etc.). Obviously, memory representations of the background facts will figure centrally in both types of performance, and the relation between tasks will depend on the relation between the respective types of representations. In this connection, research on dual reasoning and memory processes has produced three instructive generalizations. First, solving reasoning problems depends chiefly on the adequacy of children's gist memoriesmtheir ability to extract relevant patterns and meanings from background facts and to retrieve and process that information on reasoning problems. Second, responding accurately to memory tests depends primarily on the adequacy

Dual Processes in Fuzzy-Trace Theory

65

of children's verbatim memoriesmtheir ability to preserve a record of the surface content of background facts and to retrieve and process it when memory probes are administered. By themselves, these two generalizations would not lead one to expect memory-reasoning independence under the traditional assumption that verbatim and gist memory are closely interwined. Third, however, we have also seen (Section III) that verbatim and gist traces appear to be stored in parallel, rather than one being extracted from the other, and to be accessed by dissociated retrieval operations.

IV. Four Research Domains So far, fuzzy-trace theory has been presented as a set of theoretical principles that evolved from certain lines of experimentation. However, in several areas of current experimentation, these principles have been used to generate new predictions about memory and reasoning and to explain extant findings that are otherwise puzzling (e.g., Koriat, Goldsmith, & Pansky, 2000; Budson, Daffner, Desikan, & Schacter, 2001). We summarize work in four such areas: the developmental cognitive neuroscience of false memory, false memory in brain-damaged patients, judgment and decision making, and retrieval operations in child and adult recall. A. DEVELOPMENTAL COGNITIVE NEUROSCIENCE OF FALSE MEMORY

Broadly construed, the field of cognitive neuroscience deals with brain mechanisms that underlie memory, reasoning, and their development (for reviews, see Gabrieli, 1998; Schacter, Norman, & Koutstaal, 1998). The bulk of the literature is composed of five types of studies: (a) behavioral studies of normal adults (usually college students), (b) brain-imaging studies of normal adults, (c) neuropsychological studies of brain-damaged patients, (d) behavioral studies of normal child development, and (e) behavioral studies of aging. Behavioral studies of normal adults supply baseline data that are essential for interpreting findings from the other four types of experiments. In brain-imaging experiments, the concern is to identify patterns of brain activity that accompany specific forms of reasoning or remembering (e.g., Johnson, Nolde, Mather, Kounios, & Schacter, 1997). In studies of brain-damaged patients, the concern is to identify reasoning or remembering anomalies that accompany insults to specific brain loci (e.g., Melo, Winocur, & Moscovitch, 1999). Finally, behavioral studies of normal child development and aging form the important subdiscipline of developmental cognitive neuroscience (e.g., Reyna & Brainerd, 1998; Schacter, Kagan, & Leichtman, 1995), and they provide baseline data on developmental changes in behavior that may be grounded in ontogenetic changes in brain mechanisms.

66

C. J. Brainerd and V. E Reyna

The last three types of studies will concern us here because many studies of these types have dealt with false memory, which is directly related to fuzzy-trace theory. "False memory" means that an event reported as being remembered was in fact not experienced. False-memory studies that involve normal developmental samples (type d) are more numerous than the other two types of studies, and reviews of portions of that literature are available (Brainerd & Poole, 1997; Brainerd & Reyna, 1998a; Bruck & Ceci, 1997; Bruck, Ceci, & Melnyk, 1997; Ceci & Bruck, 1993, 1995; Howe, 2000; Quas, Qin, Schaaf, & Goodman, 1997; Reyna, 1995; Reyna & Lloyd, 1997). However, false-memory studies of elderly adults and braindamaged patients are now rapidly accumulating, and in 1999 a special issue of a leading cognitive neuroscience journal was devoted to work of this sort (Schacter, 1999). Many of these latter studies have revolved around the same theoretical principles that have been central in studies of normal child development. In this section, we confine attention to developmental cognitive neuroscience research on false memory. Research on false memory in brain-damaged patients is taken up in the next section.

1. Behavioral Studies of Normal Child Development Throughout most of its history, the study of memory development was synonymous with age variability in true memories. Indeed, memory was so thoroughly equated with true memory that the adjective "true" would have seemed utterly redundant until the mid-1990s. Thereafter, studies of children's false memories appeared with increasing frequency. The original impetus for such research came from outside the memory development literaturemspecifically, from court cases involving allegations of sexual abuse (see McGough, 1993). About a decade ago, cases began to surface in which witnesses reported highly detailed, credible memories of personal abuse experiences that were subsequently discredited by other facts. Two types of cases received the most attention. In one type (see Ceci & Bruck, 1995; Poole & Lamb, 1998), children, particularly preschoolers or younger elementary schoolers, reported memories of recent sexual abuse by adults, usually pursuant to questioning by police investigators, social workers, or therapists. In the other type of case (see Loftus, 1993, 1997), adults exhumed long-dormant memories of having been abused as children, often during the course of psychotherapy that had been initiated for other reasons. Following a two-year study of such cases, an American Psychological Association task force concluded that "It is possible to construct convincing pseudo-memories for events that never occurred" (Denton, 1994, p. 9). The response has been a flourishing developmental literature on false memory, which has yielded a number of probative effects, such as independence of true and false reports of the same events, long-term persistence of false memories, and false memory creation from mere memory testing. However, by common agreement (see Brainerd et al., 2000a; Koriat et al., 2000), this prolific literature has an important

Dual Processes in Fuzzy-Trace Theory

67

limitation: The accumulation of new empirical effects has far outdistanced the ability to explain them in theoretical terms. For instance, the continuation of the above-quoted statement is "although the mechanisms by which this occurs are not well understood" (Denton, 1994, p. 9). Ceci and Bruck (1993) offered an analogous conclusion in their review of children's memory suggestibility: "the exact mechanisms involved in producing distortions in young children's reports are still being debated" (p. 432). The explanatory impasse is not for want of memory theories. Rather, the impasse is rooted in the fact that classical memory theories evolved from research on true memory (Reyna & Titcomb, 1997). Consequently, they are not easily adapted to explaining how false memories could be formed from the same experiences that produce true memories, how true and false memories of the same experiences can coexist in the same person, and how these processes can vary with age (Koriat et al., 2000). Fuzzy-trace theory has been exploited to close this explanatory gap by extending some of the principles that we discussed earlier to false-memory reports. This extension has three elements: a taxonomy of the basic varieties of false reports, explanations of these phenomena, and new predictions that follow from those explanations. a. The Taxonomy. Reyna (1995) classified developmental studies of false memory along two simple dimensions: the types of tests that were used to measure false memories and the types of false memories that were measured. With respect to the first dimension, recall tests have sometimes been administered in adult studies (e.g., Roediger & McDermott, 1995), but recognition tests have been used overwhelmingly in developmental studies. Recognition tests are more sensitive than recall to the contents of memory, whether that content is true or false, particularly with children (Brainerd & Reyna, 1996). Of course, mixtures of target and distractor probes have been administered, but distractors have been further subdivided into meaning-consistent probes (e.g., ANIMAL or DOG if POODLE is a target) and meaning-inconsistent probes that supply the baseline measures of response bias that are necessary to determine whether levels of false memory are statistically reliable. Depending on the study, the latter probes have been either unrelated to the original target information or invalid statements of target meanings. For instance, if the targets are "The cat is in the cage," "The cage is on the table," and "The cat has black fur," the probe "Mother drove the children to school" would be an unrelated distractor and "The cat is under the table" would be an invalid distractor. Children can give two types of false reports, misses (incorrectly saying "no" to targets) and false alarms (incorrectly saying "yes" to distractors). The latter are of principal interest in developmental studies of false memory. In the other dimension of the taxonomy, two basic varieties of false memory have figured in developmental studies: spontaneous and implanted, which Binet (1900) originally called "autosuggestion" and "suggestion." The false-recognition

68

C. J. Brainerd and V. E Reyna

paradigm (e.g., Felzen & Anisfeld, 1971) is the standard procedure for inducing spontaneous false reports, and the misinformation paradigm (e.g., Ceci, Ross, & Toglia, 1987; Howe, 1991) is the standard procedure for implanting false reports. The key difference between the paradigms is the presence of an interference phase, interpolated between study and test, in misinformation designs. During this interference phase, children receive information indicating that certain events happened during the study phase that actually did not happen. Misinformation may be direct ("The girl wore a red hat") or indirect ("Do you remember that the girl wore a red hat?"). Significant levels of spontaneous false memory are present when false-alarm rates for meaning-consistent distractors exceed false-alarm rates for meaning-inconsistent distractors, which is called the false-recognition effect. Significant levels of implanted false memory are present when false-alarm rates are higher for meaning-consistent distractors that were subjects of misinformation (e.g., "Did the girl wear a red hat?") than for meaning-consistent distractors that were not (e.g., "Did the girl wear a raincoat?"), which is called the misinformation effe c t.

b. Explaining Spontaneous and Implanted False Memories. Reyna and Brainerd (1995) and Brainerd and Reyna (1998a) argued that the principles of parallel verbatim-gist storage, dissociated verbatim-gist retrieval, representational bases of conscious remembrance, and identity-similarity judgment could account for the false-recognition effect (i.e., spontaneous false reports) and the misinformation effect (i.e., implanted false reports) with the additional assumption that the distinction between verbatim-based judgments of identity/nonidentity and gistbased judgments of similarity applies to meaning-consistent distractors as well as to targets. First, note the obvious memory supports for false reports that are available in the false-recognition and misinformation paradigms. In the former, when meaning-consistent distractors are tested (e.g., ANIMAL or DOG if POODLE is a target), retrieval of gist memories of targets will support false alarms. In the latter, when meaning-consistent distractors are tested, retrieval of gist traces of targets, gist traces of misinformation, or verbatim traces of misinformation will all support false alarms (Reyna, 1995). According to the principle of parallel verbatim-gist storage, children deposit separate verbatim and gist traces of targets during the study phase of both paradigms, and they also deposit separate verbatim and gist traces of misinformation during the interference phase of misinformation experiments. According to the principle of dissociated verbatim-gist retrieval, a key factor that determines which traces will be accessed is the cues that are supplied by memory probes. We have seen that if verbatim and gist traces are both readily accessible, target probes seem to be better retrieval cues for their verbatim traces than for their gist traces. In false-recognition designs, meaning-consistent distractors seem to be better retrieval cues for gist traces than for verbatim traces (Reyna & Kiernan, 1994). In misinformation designs, meaning-consistent distractors (that

Dual Processes in Fuzzy-Trace Theory

69

were not presented during the study phase) seem to be better retrieval cues for verbatim traces of their presentation during the interference phase than they are for gist traces of their presentation or for gist or verbatim traces of earlier-presented targets (Brainerd & Reyna, 1998a). Thus, fuzzy-trace theory posits different memorial bases for hits and false alarms in false-recognition and misinformation designs. In the former, hits are predominately identity judgments provoked by the retrieval of verbatim traces of targets, and false alarms to meaning-preserving distractors are predominately similarity judgments provoked by the retrieval of gist traces of targets. (If gist memories are very strong and distractors are superior retrieval cues for those memories, gist retrieval may provoke illusory identity judgments about such distractors.) However, when meaning-preserving distractors induce retrieval of verbatim traces of their corresponding targets, the result will be nonidentity judgments ("I couldn't have heard DOG because I distinctly remember hearing POODLE"), supporting correct rejections rather than false alarms (Brainerd, Reyna, & Kneer, 1995; Clark & Gronlund, 1996). The misinformation effect is slightly more complex, owing to the presence of further memorial paths to false alarms for meaning-preserving distractors that have been presented as misinformation. Hits are again assumed to be predominately identity judgments provoked by the retrieval of targets' verbatim traces. False alarms to distractors, in contrast, are assumed to be predominately identity judgments provoked by the retrieval of verbatim traces of their presentation during the interference phase (Reyna & Titcomb, 1997). An important proviso is that at test, the origin of verbatim traces of misinformation is misattributed to the study phase or the origin is forgotten (Brainerd & Reyna, 1998a; Reyna & Lloyd, 1997; Reyna & Titcomb, 1997). If they are correctly attributed to the interference phase, the resulting identity judgment would authorize a correct rejection rather than a false alarm ("I heard 'The girl wore a red hat,' but I distinctly remember hearing it after the study phase was over"). In addition to verbatim-based identity judgments, gist-based similarity judgments, provoked by the retrieval of gist traces of targets or of misinformation, will support false alarms to meaning-consistent distractors, as will the illusions of identity that are sometimes induced by the retrieval of very strong gist traces (Reyna, 1998). As with the false-recognition effect, however, the retrieval of verbatim traces of these distractors' corresponding targets provokes nonidentity judgments supporting correct rejections ("The girl couldn't have worn a red hat because I distinctly remember hearing 'The girl wore a blue hat' during the study phase").

c. New Predictions. An immediate dividend of this account of children's spontaneous and implanted false memories is that it delivers several new predictions that can be exploited to assess its validity (Brainerd et al., 2000b, 2001). Because counterintuitive results provide the strongest tests of a theory, we summarize evidence

70

C. J. Brainerd and V. F. Reyna

on three predictions of this sort: variability in developmental trends, false memory persistence, and false-recognition suppression and reversal. 1. Variability in developmental trends. What does fuzzy-trace theory say about age changes in the frequency of false-memory reports? The standard view in the literature is that the frequency of such reports decreases with age because the improving accuracy of memory for actual experience allows for finer discriminations of true events from false ones and, in the case of misinformation, for better resistance to memory suggestion (e.g., Ceci & Bruck, 1995). Fuzzy-trace theory's predictions are more complex because its account of false memory is an opponent-processes model in which some factors increase false reporting while others decrease it. With respect to spontaneous false memories, we have just seen that gist-based judgments of similarity (and, occasionally, illusory identity; see Brainerd et al., 2001) support false alarms to meaning-consistent distractors, and verbatim-based judgments of nonidentity support correct rejections. Earlier, we saw that verbatim and gist memory both improve with age, from which a prediction follows that conflicts with the standard view of age changes in false reports: Developmental trends in the false-recognition effect will vary as a function of whether particular designs are more sensitive to age improvements in verbatim memory or to age improvements in gist memory (Brainerd & Reyna, 1998a). For simplicity, consider three hypothetical studies whose specific design features produce the following patterns for verbatim and gist memory: (a) Study 1: younger and older children's gist memories of target materials are comparable, but older children's verbatim memories are better; (b) Study 2: younger and older children's verbatim memories of target materials are comparable, but older children's gist memories are better; (c) Study 3: older children's verbatim and gist memories of target materials are both better than younger children's. The standard view of age changes in false reports would be confirmed in Study 1 because verbatim-based judgments will increase with age, and gist-based similarity and identity judgments will remain constant. However, counterintuitively, the opposite trend (false-recognition effects increase with age) will be present in Study 2 because gist-based similarity and identity judgments will increase with age while verbatim-based judgments will remain constant. Finally, Study 3 will produce no developmental trend if age improvements in verbatim and gist memory cancel each other out. Consider the same three studies in connection with the misinformation effect. We have just seen that false alarms to distractors are supported by verbatim memories of misinformation and by gist memories of targets and of misinformation, but that verbatim memories of targets support correct rejections. Developmental trends in Study 1 will obviously depend on whether verbatim memory for targets (the response is a correct rejection) or verbatim memory for misinformation (the response is a false alarm) is better, given the peculiarities of different designs. In the standard design, verbatim memory for targets would be expected to improve more because targets are usually presented in an earlier study session and their

Dual Processes in Fuzzy-Trace Theory

71

verbatim traces are therefore more difficult to preserve. Hence, age declines in misinformation effects would be expected. The prediction in Study 2 is that misinformation effects will increase with age because only gist memory is improving, and misinformation typically contradicts verbatim details but is consistent with global gist. Finally, Study 3 would fail to produce developmental trends if age improvements in verbatim memory for targets cancel out improvements in the memories that support false alarms. In short, fuzzy-trace theory predicts that developmental trends in false reports are not monolithic but rather will depend on specific a priori design factors (namely, factors that affect age variability in verbatim and gist memories). Hence, the possible outcomes are developmental increases, decreases, and invariances in false-recognition and misinformation effects as functions of those a priori design factors. Although detailed data on a range of design factors have yet to accumulate, Brainerd and Reyna (1998a) concluded that for the false-recognition effect at least, the overall empirical picture is consistent with fuzzy-trace theory's prediction rather than the standard view that the effect merely decreases with age. Although the largest group of studies that they surveyed revealed age decreases (e.g., Brainerd, Reyna, & Kneer, 1995; Reyna & Kiernan, 1994), some studies revealed age increases (e.g., Brainerd & Mojardin, 1998; Brown, Smiley, Day, Townsend, & Lawton, 1977), and others revealed age invariances (e.g., Felzen & Anisfeld, 1971; Paris & Carter, 1973). Similarly, in developmental studies of the misinformation effect, available data also favor fuzzy-trace theory's prediction rather than the standard view that the effect simply decreases with age. Here, Ceci and Bruck (1993) concluded that although most studies had identified age decreases, some studies had identified age invariances (e.g., Howe, 1991). Subsequent to Ceci and Bruck's review, Pezdek and Roe (1995, 1996) reported a series of experiments in which misinformation effects increased with age. 2. False memory persistence. In false-recognition designs, it is assumed in fuzzytrace theory that false alarms to meaning-consistent distractors are predominately similarity judgments provoked by the retrieval of targets' gist traces, and hits are predominately identity judgments provoked by the retrieval of targets' verbatim traces. As noted earlier, several lines of experimentation converge on the conclusion that, as time passes, verbatim memories become inaccessible more rapidly than gist. An important implication is that the memorial basis for initial true-memory reports (hits) is failing more rapidly than that for initial false-memory reports (false alarms), from which two surprising predictions fall out. First, although meaningpreserving distractors are never experienced, false alarms should exhibit good stability over time (because their memorial basis is quite stable). Second, under some conditions, such false alarms will be more stable than hits. The key condition that must be satisfied is self-evident: Initial false alarms must be (gist) memorybased (rather than based on nonmemorial factors such as response bias) about as often as hits are (verbatim) memory-based (Brainerd, Reyna, & Brandse, 1995;

72

C. J. Brainerd and V. E Reyna

Reyna, 1996a). These predictions about false-alarm stability have been confirmed in developmental studies with forgetting intervals of one week (Brainerd, Reyna, & Brandse, 1995), one month (Brainerd & Mojardin, 1998), and two years (Poole, 1995), and they have also been confirmed in adult studies with forgetting intervals of a few days (McDermott, 1996; Payne et al., 1996). 3. False-recognition suppression and reversal. The emphasis in most developmental studies of false memory has been on factors that increase levels of false reporting. However, some investigators (e.g., Brainerd, Reyna, & Kneer, 1995; Israel & Schacter, 1997; Schacter, Israel, & Racine, 1999) have observed that fuzzy-trace theory's opponent-processes explanation supplies a memory mechanism that should suppress false reportingmnamely, that when meaning-consistent distractors (e.g., ANIMAL, DOG) access verbatim traces of their corresponding targets (e.g., POODLE), the result is nonidentity judgments ("I couldn't have heard DOG because I distinctly remember hearing POODLE") that support correct rejections rather than false alarms. It follows that manipulations that enhance the accessibility of verbatim traces of targets, without enhancing the accessibility of other memories that support false alarms, will suppress false-recognition and misinformation effects. Counterintuitively, one should be able to reverse both the false-recognition effect (i.e., higher false-alarm rates for unrelated distractors than for meaning-consistent distractors) and the misinformation effect (i.e., higher false-alarm rates for distractors that are not misinformed than for distractors that are) under conditions that strongly favor retrieval of verbatim traces of targets (Brainerd, Reyna, & Kneer, 1995; Reyna, 1995). In both false recognition and misinformation designs, the predicted suppression and reversal effects have been detected under the expected conditions. In a review of false-recognition studies, Brainerd, Reyna, Wright, and Mojardin (2000b) noted that two approaches to increasing verbatim accessibility had been investigated: (a) test-phase priming of verbatim traces and (b) study-phase presentation formats that create robust verbatim traces of targets. Test-phase priming was used by Brainerd, Reyna, and Kneer (1995), Brainerd and Mojardin (1998), Israel and Schacter (1997), and Schacter et al. (1999). In Brainerd, Reyna, and Kneer's procedure, meaning-consistent distractors (e.g., DOG) were preceded by probes for their corresponding targets (e.g., POODLE). The rationale was that if target probes predominately cue the retrieval of verbatim traces, those traces will remain highly accessible when distractors are administered immediately thereafter. Consistent with this rationale, children's false-alarms to meaning-consistent distractors were suppressed by target priming, and, in some conditions, false-alarm rates for meaning-consistent distractors were below false-alarm rates for meaninginconsistent distractors. Some years earlier, MacLeod and Nelson (1976) obtained this reversal using a similar procedure. In the Brainerd and Mojardin, Israel and Schacter, and Schacter et al. experiments, the verbatim-priming procedure involved testing meaning-consistent distractors in oral or visual formats that closely

Dual Processes in Fuzzy-Trace Theory

73

resembled the oral or visual formats in which their corresponding targets had been studied. This procedure also suppressed children's and adult's false-alarms to meaning-consistent distractors. The other approach to increasing verbatim accessibility (studying targets in formats that create robust verbatim traces) was investigated by Israel and Schacter, Reyna & Kiernan (1994, 1995), and Schacter et al. (1999). Reyna and Kiernan's participants studied sentences that were presented in either normal literal language or in metaphorical language in which exact wording matters (robust verbatim traces). In the Israel and Schacter and Schacter et al. experiments, participants studied words that were presented either in standard printed form or as pictures (robust verbatim traces). In both instances, false-alarm rates for meaning-consistent distractors were suppressed in presentation conditions that were expected to produce robust verbatim traces. Reviews of studies with misinformation designs (Reyna, 1995; Reyna & Titcomb, 1997; Titcomb & Reyna, 1995) indicate that misinformation effects have been suppressed, and sometimes reversed, by manipulations that increase the accessibility of verbatim traces of target material. The manipulation for which the most extensive data are available is the timing of the interpolated interference phase. The finding that verbatim traces become inaccessible more rapidly than gist traces provides the rationale for this manipulation. In the standard design, the sequence of events is study phase, interference phase, memory tests. Suppose that memory tests are administered immediately after the interference phase, but that the interval between the study and interference phases is varied so that misinformation follows the study phase either immediately or after a delay (e.g., a few days). The clear prediction is that verbatim traces of targets will be more accessible when misinformation is immediate, and the misinformation effect will therefore be suppressed when misinformation is immediate. This prediction has received support in developmental studies (Reyna, 1995) and in adult studies (Belli, Windschitl, McCarthy, & Winfrey, 1992). In some cases, immediate misinformation has produced a reversal of the misinformation effect. For instance, Cassel and Bjorklund (1995) found that when adult participants viewed a film of a bicycle theft and responded immediately to an interview containing misinformation about events in the film, they displayed lower reporting of false events that were presented as misinformation than of false events that had not been presented as misinformation.

2. Behavioral Studies of Aging Although a few studies of implanted false memories in elderly adults have been reported (e.g., Adams-Price & Perlmutter, 1992), aging research has been focused almost exclusively on spontaneous false memoriesmspecifically, on false recognition and false recall of information that is consistent with the gist of experience. A favored procedure has been one that produces especially robust semantic false

74

C. J. Brainerd and V. F. Reyna

recognition in normal adultsmthe Deese/Roediger/McDermott paradigm (Deese, 1959; Roediger & McDermott, 1995). In this procedure, participants study several short lists of words whose members revolve around a core theme (e.g., NURSE, SICK, LAWYER, MEDICINE, HEALTH, HOSPITAL, DENTIST, PHYSICIAN, ILL). For each list, a critical word that is a semantic associate of all the targets does not appear on the list (DOCTOR in the sample list). Participants respond to recognition tests composed of targets (e.g., HEALTH), critical semantic associates (e.g., DOCTOR), and unrelated distractors (e.g., TELEVISION). The surprising outcome is that in normal adults, false-alarm rates for critical distractors approach hit rates for targets, and in most studies the differences between the two are not significant. The basal trend in aging studies is that false-recognition effects increase between young adulthood and late adulthood in paradigms in which young adults display very small effects (e.g., Hess, 1984; Rankin & Kausler, 1979). The same age trend has also been detected with procedures, such as the Desse/Roediger/McDermott paradigm, that produce higher levels of false recognition in young adults (Koutstaal & Schacter, 1997; Norman & Schacter, 1997), and these procedures have also revealed higher levels of false recall of critical distractors (Kensinger & Schacter, 1999; Norman & Schacter, 1997; Tun et al., 1998). The targets in all of these experiments had very familiar meanings, which casts doubt on one of the possible fuzzy-trace theory loci for this age trend. Specifically, age improvements in gist memory could hardly be responsible, leaving age declines in the accessibility of verbatim traces as a more likely explanation. This interpretation receives indirect support from the fact that when participants make rememberknow judgments about accepted targets, the incidence of remember judgments is usually lower in elderly adults than in younger adults (Schacter et al., 1998). More direct evidence comes from some experiments by Tun et al. (1998). These investigators found, first, that the usual Deese/Roediger/McDermott procedure produced higher false-alarm rates (and faster responses) to semantically related distractors (both critical and noncritical) in elderly adults than in young adults. If declines in the accessibility of verbatim traces are responsible, manipulations that encourage participants to respond on the basis of gist memory and not to access verbatim traces should selectively disadvantage younger participants (who access verbatim traces more readily). Consistent with this interpretation, Tun et al. found that aging differences in both levels of false recognition and speed of responding disappeared when experimental conditions encouraged gist access rather than verbatim access. However, experiments by Kensinger and Schacter (1999) suggest that lower levels of verbatim accessibility and higher levels of gist reliance may both contribute to elevated semantic false recognition in the elderly. Prior experiments with young adults demonstrated that repetition of targets at study strengthens fastfading verbatim traces, inoculating them against forgetting and making them more accessible on recognition tests (Brainerd, Reyna, & Kneer, 1995; Brainerd et al.,

75

Dual Processes in Fuzzy-Trace Theory

1999; Reyna & Titcomb, 1997). Repetition of Deese/Roediger/McDermott lists should therefore reduce high levels of false recognition of critical distractors because participants can more easily reject those distractors by accessing verbatim traces of targets. Previous experiments had shown that repetition decreased false recall of critical distractors by young adults (e.g., McDermott, 1996). Kensinger and Schacter found that repetition also decreased false recognition in young adults, but that elderly adults' elevated levels of false recognition were resistant: When Deese/Roediger/McDermott lists were studied 5 times, young adults' false-alarm rates declined from .90 (Trial 1) to below .5 (Trial 5), but elderly adults' false-alarm rates did not decline. To identify underlying changes in verbatim and gist memory across repetitions, Kensinger and Schacter computed two measures that involved the signal-detection statistic A' (Snodgrass & Corwin, 1988): (a) the discriminability of targets from critical distractors and (b) the discriminability of critical distractors from unrelated distractors. The first measure indexes verbatim accessibility because targets and critical distractors share the same gist, but verbatim traces were stored only for targets. The second measure indexes gist accessibility because verbatim traces were not stored for either critical or unrelated distractors, but critical distractors share the same gist. Results for these two measures appear by age level in Figure 5. For young adults, two trends were reliable: Verbatim access increased across list repetitions and gist access decreased. Together, these two trends would reduce false-alarm rates (because verbatim retrieval supports correct rejections and gist retrieval supports false alarms). For elderly adults, two trends were reliable: Verbatim access again increased across list repetitions, but gist access also increased. Thus, elderly participants' verbatim memories also benefited

;~ 0.9 1.1_

o 0.8 U.I

::30.7 _.J < >0.6~

[-

~

~

~ .........

. ~

.............

I

0.5 1

2

3

4

NUMBER OF PRESENTATIONS =

YOUNGVERBATIM ~ YOUNG GIST

5

ELDERLYVERBATIM

-El- ELDERLYGIST

Fig. 5. A' values for verbatim and gist retrieval in younger and elderly adults in Kensinger and Schacter (1999, Experiment 2).

76

C.J. Brainerd and V. E Reyna

from repetitions, but the beneficial effects of improved verbatim accessibility on false-alarm rates were canceled by increases in gist accessibility. B. STUDIES OF FALSE MEMORY IN BRAIN-DAMAGED PATIENTS

Studies of false memory in brain-damaged patients, like studies of elderly adults, have been focused on spontaneous rather than implanted false memories, and the Deese/Roediger/McDermott is again a favored paradigm. Two groups of studies are of particular interest in the present context: those in which brain-damaged patients displayed more robust false memory than normal participants and those in which brain-damaged patients exhibited reduced false memory, relative to normal participants. The first pattern has usually been obtained with patients who have ventromedial or posterior frontal lobe damage (Curran, Schacter, Norman, & Galluccio, 1997; Melo et al., 1999; Parkin, Bindschaedler, Harsent, & Metzler, 1996; Parkin, Ward, Bindschaedler, Squires, & Powell, 1999; Rapcsak, Reminger, Glisky, Kaszniak, & Comer, 1999). This pattern has also been obtained with split-brain patients who have received left hemispheric presentation of target material (Phelps & Gazzaniga, 1992). Patients with frontal lobe damage also show a condition that Parkin (1997) has termed pathological false recognition: very high false-alarm rates for unrelated distractors as well as for meaningpreserving distractors, suggesting generalized impariments in the ability to access verbatim traces of targets and to make identity or nonidentity judgments about recognition probes. Consistent with such an interpretation, false recognition of unrelated distractors drops sharply when verbatim prosthetics are supplied (e.g., increasing the perceptual dissimilarity between targets and distractors by administering distractors that differed markedly from targets in visual appearance). In split-brain patients, levels of false recognition of meaning-preserving distractors are greater when target material is presented to the left hemisphere than when it is presented to the right, but false recognition of unrelated distractors is the same. Fuzzy-trace theory provides three possible explanations: Gist storage is superior in the left hemisphere, verbatim storage is superior in the right hemisphere, or both. A series of experiments on a single split-brain patient, using a variety of manipulations of the surface and semantic content of target materials, favored the view that verbatim storage was superior in the right hemisphere and gist storage was superior in the left hemisphere (Metcalfe, Funnell, & Gazzaniga, 1995). However, these results may depend on visual presentation of target material (Bowden & Beeman, 1998). Studies of other types of patients, amnesiacs with damage to the medial temporal lobes, have produced conflicting findings. In false-recognition designs in which related distractors were composites of targets (e.g., HANDSTAND and SHOTGUN are targets and HANDGUN is a related distractor), amnesiacs with damage to the

Dual Processes in Fuzzy-Trace Theory

77

medial temporal lobes had higher false-alarm rates than normal participants (Kroll, Knight, Metcalfe, Wolfe, & Tulving, 1996), sometimes to the point that targets were not discriminated from distractors. Here, the obvious explanation is differences in the accessibility of verbatim traces: HANDGUN preserves the weapon meaning of SHOTGUN, which supports false alarms if verbatim traces of HANDSTAND and SHOTGUN cannot be accessed (Koutstaal, Schacter, Verfaellie, Brenner, & Jackson, 1999). In sharp contrast, when the Deese/Roediger/McDermott paradigm is used, amnesiacs' false-alarm rates to critical distractors are lower than normal participants' (Schacter et al., 1996). Fuzzy-trace theory again generates three possible explanations: differences in the accessibility of gist memories, differences in the accessibility of verbatim memories, or both. Schacter et al. concluded that their data favored the conclusion that relative to normal participants, the ability of some types of amnesiacs to store gist memories of the themes that are repeatedly cued by Deese/Roediger/McDermott lists is impaired. A finding supporting this conclusion is that semantic false-recognition effects are comparable in amnesiacs and controls with more conventional lists in which the targets do not repeatedly cue a single theme (Koutstaal et al., 1999). In work with Alzheimer's patients, Budson et al. (2001) have woven together the two threads of patients who exhibit elevated false-recognition effects versus patients who exhibit reduced effects. Budson et al. administered Deese/Roediger/ McDermott tasks to patients with probable Alzheimer's disease, to elderly adults, and to young adults. Alzheimer's patients, like amnesiacs with medial temporal lobe damage, display reduced false-recognition for Deese/Roediger/McDermott critical distractors (Balota et al., 1999). However, Budson et al. posited that such patients are deficient in both extraction of list gist and retention of verbatim traces of individual targets. To test this hypothesis, they measured variations in false-recognition effects as a function of repetition by presenting individual Deese/Roediger/McDermott lists 1-5 times. As noted earlier, with normal adults who readily store the gist of such lists, the net effect of repetition is to strengthen fast-fading verbatim memories, relative to gist memories, thus reducing false recognition. If Alzheimer's patients are deficient in gist storage, the obvious effect of repetition is to provide further opportunities to store the gist memories that cannot be deposited on a single study trial. If Alzheimer's patients are also deficient in verbatim retention, repetition will not strengthen their verbatim memories as much as it does controls' verbatim memories. These considerations imply increases in false-recognition effects as a function of repetition. Consistent with this expectation, opposite repetition effects were observed in Alzheimer's patients and non-Alzheimer's participants. In Alzheimer's patients, false-alarm rates for critical distractors increased from .69 (1 presentation) to .88 (5 presentations). Repetition did not affect the verbatim A' measure, and it increased the gist A' measure. In non-Alzheimer's participants, on the other hand, false-alarm rates for critical distractors decreased from .68 (1 presentation) to .42 (5 presentations) in

78

c.J. Brainerd and V. F. Reyna

young adults and from .67 to .50, respectively, in normal elderly adults. Also, repetition increased the verbatim A' measure and decreased the gist A' measure. C. DUAL-MEMORY PROCESSES IN JUDGMENT AND DECISION MAKING The psychology of j u d g m e n t and decision making covers a broad array of tasks that involve risk or uncertainty (Hogarth, 1980). For instance, prospective renters might be provided with information about several apartments that are available in their locale (e.g., square footage, proximity to bus lines), and they might have to choose the best apartment in the group (choice) or state how much rent they are willing to pay for each apartment (point estimation). Earlier, we introduced a classic decision-making task that has generated numerous studies, Tversky and K a h n e m a n ' s (1981) framing problem. In Table III, we present another classic j u d g m e n t task, an example of Tversky and Kahneman's (1983) conjuction fallacy problem. The generic finding for these and other tasks is that, their simplicity notwithstanding, participants' judgments and decisions are fallacious. In the framing problem, as we saw, adults show risk aversion when alternatives are framed as gains ("lives saved") and risk preference when alternatives are framed as losses ("lives lost"). Similarly, the conjunction fallacy problem in Table III fails to elicit from adults the logically appropriate response that a more inclusive set ("bank tellers") must be more (or equally) probable than a less inclusive set ("bank tellers who are feminists"). Instead, the less inclusive set is judged to be more probable when (a) the more inclusive sets refers to something that participants judge to be unlikely ("Linda is a bank teller") and (b) the less inclusive set is a proper subset involving a property that participants judge to be likely ("Linda is a feminist"). TABLE III Tversky and Kahneman's (1983) Conjunction Fallacy Problem Facts

Linda is 31-years-old, single, outspoken, and very bright. She majored in philosophy. As a student she was deeply concerned with issues of discrimination and social justice, and she also participated in antinuclear demonstrations.

Judgment

Rank the following items in terms of probability: Linda is a teacher in an elementary school. Linda works in a bookstore and takes Yoga classes. Linda is active in the feminist movement. Linda is a psychiatric social worker. Linda is a member of the League of Women Voters. Linda is a bank teller. Linda is an insurance salesperson. Linda is a bank teller and is active in the feminist movement.

Findings

Conjunction fallacy effect: Participants rank conjunctions, such as "bank teller and active in the feminist movement," as more probable than "bank teller" by itself, although the class of bank tellers includes feminist bank tellers.

Dual Processes in Fuzzy-Trace Theory

79

What do such demonstrations tell us about adult judgment and decision making? A possible answer is that these demonstrations involve artificial, contrived situations that rarely occur in everyday life and that, by comparison, naturalistic judgment and decision making are eminently adaptive (e.g., Gigerenzer, 1994). The traditional answer, however, which is exemplified by Tversky and Kahneman's (e.g., Kahneman, Slovic, & Tversky, 1982) heuristics-and-biases position, is that such demonstrations show that judgment and decision making are beset by foibles--more specifically, that they violate the basic axioms of prescriptive models of logic and rationality, such as those posited by Piaget (e.g., Piaget & Inhelder, 1969). This fundamental lack of rationality is traditionally ascribed to limitations of human information processing, especially memory limitations that constrain reasoning (e.g., Nisbett & Ross, 1980). That is, reasoning would be logical and rational if it could, but the excessive memory and processing demands imposed by judgment and decision making under conditions of uncertainty or risk force reasoning to take shortcuts, which are reflected in logically inappropriate responses. This hypothesis, though appealing, has proved unsatisfactory in three respects (Reyna & Brainerd, 1995). First, the hypothesis is not well specified at the level of process: Notions such as "information processing limitations" and "memory limitations" are crude, global constructs. Second, across a wide range of specific definitions of "memory limitations," the hypothesis is not explanatory at the level of the actual performance patterns that characterize individual tasks. Furthermore, although memory or processing demands may indeed derail logical reasoning, this outcome does not forecast the specific response patterns that are observed on framing and conjunction fallacy problems, for example. Few memory constraints exist in framing and conjunction fallacy tasks. Thus, memory limitations are unlikely explanations of why decisions are risk averse when equivalent options are framed as gains or why less-inclusive sets are judged as more probable than more inclusive sets. Third, the hypothesis has not fared well in experimental attempts to tie memory limitations directly to reasoning failures (see Reyna, 1992; Reyna & Brainerd, 1992). As we reported in section II, a number of studies have been conducted in which the relation between memory capacity (the ability to hold relevant facts in storage long enough to process them) and reasoning performance was examined. The denouement of those studies is that memory capacity apparently does not sharply constrain reasoning (because reasoning accuracy is usually dissociated from performance on memory tests for crucial background facts). In the same vein, tasks that normally produce illogical judgments or decisions will produce logical responses following minor changes that do not reduce overall memory or processing demands (Reyna & Brainerd, 1991 a). In the framing problem, for example, if the choices in the gain frame are changed to (a) saving 200 lives for certain versus (b) a one-third chance of saving 603 lives, participants now prefer risk (option b), which is the logically appropriate response (Reyna & Brainerd, 1995).

C. J. Brainerd and V. F. Reyna

80

Some of fuzzy-trace theory's principles have been implemented in judgment and decision-making studies in an attempt to improve our ability to explain illogical responses (e.g., Klaczynski & Narasimham, 1998; Kuhberger, 1995, 1998; Levin, Schneider, & Gaeth, 1998; Reyna & Brainerd, 1991b, 1994; Reyna & Ellis, 1994; Stone, Yates, & Parker, 1994; Wedell & Bockenholt, 1994; Wolfe, 1995). These studies have produced two general results that we describe in the remainder of this section: a general theory of the processing loci of reasoning failures and a model called inclusion illusions that shows how this theory can be used to provide an integrative explanation of illogical responses on several tasks that, on the surface, seem quite disparate (Reyna et al., 2001).

1. Graded Rationality: A General Processing Theory Rather than regard human reasoning as fundamentally rational or fundamentally irrational, Reyna et al. (2001) showed that fuzzy-trace theory's analysis of memory-reasoning relations could be exploited to identify distinct processing loci of reasoning failures in judgment and decision making (see also Reyna, 1991; Reyna & Brainerd, 1993, 1994). Importantly, some of these loci would represent higher levels of irrationality than others. Thus, by implementing this processing theory, absolute claims about rationality or irrationality are replaced by the notion of graded rationality, where different gradations are mapped with specific processing loci of reasoning failure. In this theory, the major loci of processing failure are: (a) lack of knowledge (competence) of relevant logical-reasoning principles; (b) verbatim memory interference with processing appropriate gist; (c) failure to encode the appropriate gist of a task; (d) interference between competing gists; (e) failure to retrieve relevant reasoning principles; and (f) processing interference. The gradations of rationality are: a (most irrational)

~b

>c

~d

~e

> f (least irrational)

a. Lack of Competence. Participants may not understand the relevant reasoning principle (e.g., the cardinal ordering rule or the ratio rule of probability); if so, they will not use the principle on a task that requires it. In normal, educated adults, this most irrational form of processing error occurs rarely, if ever: One of the features of judgment and decision-making tasks that makes illogical responses so disconcerting is that they tap reasoning principles that participants are thought to understand. Even in developmental studies, where lack of competence is a more realistic possibility, experiments that have parsed multiple loci of children's reasoning errors have typically found that competence failures do not predominate; children make reasoning errors because they process the wrong memory representations, or they fail to retrieve reasoning principles that they understand, or they fail to use those principles correctly once they are retrieved (Brainerd & Reyna, 1990b; Reyna, 1991).

Dual Processes in Fuzzy-Trace Theory

81

b. Verbatim Interference with Gist. The next, slightly less irrational, form of processing failure occurs when verbatim memory for task information interferes with gist processing. The fuzzy-processing preference specifies that reasoning is disposed to operate on gist and that "logical" responses usually involve extracting and processing the gist of problem information. As Reyna (1995) showed, paradoxically, the most accurate representations of problem informationmverbatim tracesmcan produce judgments and decisions that are less logical than those that result from gist processing. In general, such errors occur when thinking is too literal, as when reasoners mindlessly focus on the literal meanings of sentences ("Carol's stare was dark and thunderous"; "Frank's tie was loud") and fail to appreciate that Carol was angry and Frank's tie was garishly colored. This second type of processing error, like the first, is not very common in adult judgment and decision making (for a review, see Reyna et al., 2001). In developmental studies, it appears to be common in preschool children. For instance, Brainerd and Reyna (1993b) found that preschoolers display what they called a verbatim-exit bias~refusing to accept a logically sound inference that has some verbatim features that were not present in the problem information. To illustrate verbatim-exit bias, following the statements, "The bird is in the cage," "The cage is on the table," "The cat is under the table," preschoolers often rejected the inferences, "The bird is above the cat" and "The cat is below the bird," as being incorrect because the latter statements contain the new words "above" and "below." c. Failure to Encode Appropriate Gist. The remaining types of processing failures appear to be significant contributors to judgments and decisions in adults. The third type of failure occurs when reasoners fail to "get" the correct gist of a problem (though, unlike the second type of failure, this type does not occur because reasoners focus on verbatim content), and responses are therefore based on the wrong "take." To illustrate, von Winterfeldt and Edwards (1986) reported the case of an elderly professor facing mental decline from reduced blood flow to the brain who had to decide whether to undergo risky surgery to correct the problem. In making the eventual decision, the professor failed to gistify the problem correctly. Specifically, he failed to extract from the available facts that the likely alternative to surgery was life with some level of cognitive impairment; instead, he assumed that the alternatives were life without cognitive impairment or death. This type of error is far less irrational than the preceding two because it is made even by highly intelligent people when they fail to "get" the gist of pertinent facts. d. Interference between Competing Gists. In the fourth type of processing failure, participants extract the gist that is crucial for logical responses, but processing is usurped by some other, more salient, gist. Here, participants fix on gists that appear to deliver appropriate responses but that deliver inappropriate responses instead. Reyna and Brainerd (1993) provided an everyday example of this type of

82

C. J. Brainerd and V. F. Reyna

processing failure from an actual case of educational decision making. According to a news report, a survey of dangerous playground equipment had been conducted for purposes of deciding whether certain equipment should be removed from local playgrounds. However, in the report, a salient gistmthe probability that children who had accidents were playing on a certain type of equipment (e.g., slides, swings, monkey bars)mwas confused with the appropriate gist, the probability that children who were playing on each type of equipment had accidents. It was mistakenly concluded that swings were the most dangerous playground equipment because more injured children had been playing on swings than on anything else. However, swings were actually the safest equipment because the accident rate for children while playing on swings was lower than for any other equipment (see also Wolfe, 1995). e. Failure to Retrieve Relevant Reasoning Principles. Even when participants have encoded the correct gist, they may fail to access the reasoning principle that is required to process that gist. Piaget's class-inclusion problem, which is a simpler version of the conjunction fallacy, is a well-known case in point. In the basic task, the background facts consist of two or more disjoint sets of objects (e.g., 7 cows and 3 horses) that belong to a common superordinate set (10 animals). Children are well aware of the various gists in this task, including the critical one that every object is an animal, because the background information is continously available, and they respond appropriately to questions that indicate such understanding (e.g., Is there anything here that is not an animal?). Yet, when asked "Are there more cows or more animals?" the average child responds "more cows" until approximately age 10 (Winer, 1980). Adults make analogous errors on more difficult versions of this problem (e.g., Rabinowitz, Howe, & Lawrence, 1989) and on the related conjunction fallacy problem. Experimentation (e.g., Brainerd & Reyna, 1990b; Reyna, 1991) has suggested that retrieval failure is a major obstacle for younger children: When appropriate gists have been encoded in tasks that involve inclusion relations, those gists often fail to cue retrieval of the cardinal-ordering principle (the rule that regardless of the specific numbers involved, superordinate sets must contain more elements than any of their proper subsets). The studies that support this conclusion most directly are ones in which standard inclusion problems were presented, but participants were provided with more explicit retrieval cues for the cardinal-ordering principle (Brainerd & Reyna, 1990b, 1995). Such minor retrieval manipulations can cause reasoning accuracy to improve considerably (cf. Brainerd & Reyna, 1990b, Experiments 5 and 6). According to these same studies, one reason why retrieval fails is that problem statements imply that numerical comparisons are required ("Are there more cows or more animals? .... Which is more probable, that Linda is a bank teller or a feminist bank teller?"), but the cardinal-ordering rule is a qualitative principle that does not process specific numerical values.

Dual Processes in Fuzzy-Trace Theory

83

f Processing Interference. If the appropriate gists are encoded and the operations that process those gists are retrieved, processing may stall before it is successfully completed. Reyna (1995) noted that syllogistic reasoning problems are common examples of this type of failure. When adults are given premises of the form allAs are Bs ("all men are mortal") and of the form some Bs are Cs ("some men are Greeks"), other data indicate that they will extract the gist of these premises and retrieve the appropriate deductive-reasoning operations (Reyna, 1995). However, participants may fail to apply those operations coherently because they lose track of members of the classes (e.g., Bs that are not A, Cs that are not B). This failure leads to a class of syllogistic mistakes known as conversion errors: that all the Bs means that all Bs are As ("all mortals are men"), notwithstanding that all the necessary representational and processing components of correct reasoning are present. 2. Inclusion Illusions According to the general processing theory, illogical judgments and decisions arise from processing failures that are graded with respect to their intrinsic levels of irrationality. This theory can be used to generate detailed models that provide unifying explanations of performance patterns that are observed on seemingly disparate tasks. An example is provided by the inclusion illusions model of an important family of tasks (Reyna, 1991; Reyna & Brainerd, 1995; Reyna et al., 2001). This family consists of problems, some of whose characteristic errors have already been consideredmclass inclusion, conjunction fallacies, probability judgment, and syllogismsmplus three others that have also generated large literatures--conditional probability (e.g., Wolfe, 1995), disjunction bias (e.g., Tversky & Koehler, 1994), and Bayesian updating (e.g., Eddy, 1982). In conditional probability, the task is to estimate the likelihood of some event A given the presence of some event B (What is the likelihood of an accident, given that children are playing on swings? What is the likelihood of an accident, given that children are playing on slides?). In disjunction bias, participants estimate the probability of the same set of events when it is described in two ways: globally as a single class of events (What is the probability of accidental death?) versus disjunctively (What is the probability of accidental death in a car crash, by drowning, or by all other accidents?). The characteristic error is that the class of events is assigned a higher probability when it is described disjunctively than when it is described globally. Finally, in Bayesian updating, participants are given a prior probability, as well as information about test accuracy (miss and falsealarm rates) and a test result. Participants are asked for a posttest probability. The characteristic error is that the posttest probability fails to fully reflect the pretest probability. These tasks seem quite different, but they share a property that allows their characteristic errors to be brought under one umbrella. Specifically, they all involve

84

C.J. Brainerd and V. E Reyna

reasoning about classes that are related via inclusion relations. Adults as well as children have long been known to have difficulty making comparisons between classes that are related via inclusion relations (e.g., cows versus animals, horses versus animals) (Rabinowitz et al., 1989). Such comparisons induce a powerful mental illusion that hinders simultaneously keeping track of superordinate classes and their subclasses. When attention is focused on subclasses, superordinate classes recede into the mental background, and when attention is focused on superordinate classes, subclasses recede into the mental background (Reyna, 1991). This causes reasoners to make (incorrect) substitutions of comparisons between disjoint classes (e.g., cows versus horses), producing the familiar class-inclusion error in which subclasses are judged to contain more elements than superordinate classes. Reyna and Brainerd (1994) observed that the same difficulty crops up when reasoners compute ratios in which the numerator (e.g., 7 cows) is a subclass of the denominator (e.g., 7 cows + 3 horses = 10 animals)mthat is, with ratios of the form A (A § B). Proportions are familiar instances, and proportional reasoning is known to be difficult for children and adults (e.g., Hook & Cook, 1979; Moore, Dixon, & Haines, 1991; Spinillo & Bryant, 1991). Characteristic errors are the same as in class-inclusion reasoning: Responses are based on comparisons of subclasses (e.g., winners in Space A vs. winners in Space B, bank tellers vs. feminists, swing accidents vs. slide accidents) rather than on ratios of subclasses to superordinate classes {(e.g., [winners - (winners + losers)], [feminist bank tellers + (feminist bank tellers § nonfeminist back tellers)], [swing with accident - (swing with accident § swing without accident)] }. In Bayesian updating, as in other forms of probability judgment (see preceding discussion), reasoners will tend to base their judgments on A facts versus B facts, rather than on (A facts) - (A facts § B facts). Neglect of the denominator is also the underlying difficulty in the other two members of our family of tasks (Reyna et al., 2001). In both tasks, a difficulty is introduced that further encourages denominator neglect: Some of the classes that must be taken into account to identify the correct denominator are not specified, though they can be inferred. In disjunction bias, for example, the denominator ("death by accident") consists of many classes ("death by car crash," "death by drowning," "death by lightning strike," etc.), all of which are unspecified in the global version of the task but some of which are specified in the disjunctive version. To conclude, we have seen that in addition to furnishing a taxonomy of the loci of processing failure in judgment and decision making, the general processing theory generates local models to explain reasoning errors on particular tasks. To demonstrate this capability, we considered the inclusion illusions model, which explains the characteristic reasoning errors that occur in a family of superficially dissimilar tasks. The model is predicated on the observation that in all of these tasks, reasoning revolves around classes that are united via inclusion relations. Consequently, the processing difficulties that are triggered by set-subset relations can be exploited to explain reasoning errors.

Dual Processes in Fuzzy-Trace Theory

85

D. RETRIEVAL MECHANISMS IN CHILD AND ADULT RECALL

We have already considered extensions of fuzzy-trace theory's dual-process model of target recognition to false recognition of meaning-consistent distractors. We now consider extensions to recall. In the analysis of recognition, retrieved verbatim and gist memories are compared to verbatim and gist representations of probes, producing judgments of identity, nonidentity, and similarity, along with their associated phenomenologies. In recall, however, studied targets do not appear on memory tests as retrieval cues, which means that participants cannot compare retrieved memories with representations of physically present test probes (Bjorklund & Muir, 1988). Reyna and Brainerd (1995) and Payne et al. (1996) proposed dual-retrieval operations for recall that parallel those for recognition. The first operation, the faster and more accurate of the two, predominates at the start of a recall episode. This operation directly accesses verbatim traces of individual target presentations. When a verbatim trace is retrieved, the relevant surface information enters consciousness (the target echoes in the mind's ear or flashes in the mind's eye), which means that it can be recalled by merely reading that surface information out of consciousness, much as actors speak lines that are whispered to them by prompters or stenographers write what they hear or scribes copy what they see. Because direct access does not require that participants engage in complex semantic processing, it produces recall that is rapid, confident, and highly accurate. Despite these advantages, direct access has two fundamental limitations that necessitate a back-up operation: Verbatim traces become rapidly inaccessible as time passes, and verbatim traces are sensitive to output interference that accumulates during the course of recall (Brainerd, 1995; Brainerd & Reyna, 1993b; Dempster & Corkill, 1999; Reyna, 1995). The other retrieval operation, which is slower and less accurate, becomes increasingly important as recall proceeds. This operation r e c o n s t r u c t s studied material by processing gist traces of the material's meaning content. Because meaning content does not uniquely specify particular targets, reconstructive processing will generate multiple candidates for output, some of which may not be part of the studied material (if TABLE is studied, TABLE, CHAIR, FURNITURE, and SOFA may all be generated). Because multiple candidates are generated, a further judgment process is required to decide whether the meaning overlap between reconstructed candidates and stored gist traces is sufficient to warrant output (see Anderson & Bower, 1972, for a related discussion). Thus, although items that are retrieved via direct access can simply be read out of consciousness, items that are reconstructed must undergo a further judgment check that minimizes false recall of unpresented material. This distinction has been used to explain two classes of recall phenomena: cognitive triage effects (e.g., Brainerd, 1995; Reyna & Brainerd, 1995) and robust false recall (Lampinen et al., 1998; Payne & Elie, 1998; Payne et al., 1996).

86

C. J. Brainerd and V. F. Reyna

1. Cognitive Triage Effects Cognitive triage refers to surprising relations that have been detected, during free recall, between the difficulty of recalling an item (e.g., total errors to criterion, trial number of the last error) and the order in which the item is recalled (its output position in a free recall protocol). Classical one-process ("strength") theories make the commonsense prediction that the harder an item is, the later its output position will be (for a review, see Brainerd, Reyna, Howe, & Kevershan, 1991). Some years ago, however, Battig (1965; Battig, Allen, & Jensen, 1965) found that the early positions in recall protocols are occupied by difficult rather than easy targets: When items' recall positions on Trial 2 were separately computed for targets that had been recalled on Trial 1 ("easy") and for targets that had not been recalled ("difficult"), mean output position was earlier for the latter. In a series of experiments in our laboratory, we replicated this finding (which is called weak priority) and identified four other results that, taken together, form the cognitive triage pattern: (a) nonmonotonic emergence (when targets that are recalled on some later trial are assigned to graded difficulty categories based on their error-success history, the order of recall is typically harder, then easier, then harder; Brainerd et al., 1991); (b) criterion nonmonotonicity (when recall positions at criterion are plotted as a function of their complete error-success history, the harder-easier-harder ordering is again noted; Brainerd, Reyna, Harnishfeger, & Howe, 1993); (c) accuracy enhancement (total recall is improved by grouping output in terms of recall difficulty; Brainerd et al., 1993); (d) developmental sharpening (weak priority, nonmonotonic emergence, criterion nonmonotonicity, and accuracy enhancement all become more pronounced as development proceeds; Brainerd, Olney, & Reyna, 1993; Brainerd, Reyna, et al., 1993; Brainerd, Reyna, & Howe, 1990). Two of these results, criterion nonmonotonicity and developmental sharpening, can be seen in Figure 6. Reyna and Brainerd (1995) proposed that the individual components of the triage pattern fall out of the dual-retrieval model. Direct access, which processes interference-sensitive verbatim traces, predominates at the start of recall; reconstruction, which processes more stable gist traces, becomes more prominent as recall proceeds. Given that verbatim traces are highly sensitive to interference, the optimal order in which to access them is from weaker to stronger. (Because interference is more apt to impair retrieval of weaker verbatim traces, the total number of such traces that can be retrieved will be greater if weaker ones are accessed first, while accumulated levels of interference are low.) Iv. contrast, gist traces, which are more immune to the effects of interference, can be accessed in the intuitive stronger-to-weaker order. In this conception, free recall is a combination of two retrieval methods--early direct access (which processes verbatim traces in a weaker-to-stronger order) followed by increasing amounts of reconstructive retrieval (which processes gist traces in a stronger-to-weaker order). The phenomena of weak priority, nonmonotonic emergence, criterion nonmonotonicity, and

Dual Processes in Fuzzy-Trace Theory

87

10

Z

O rY 8 W

r~ o o 6 n,'

0 n,' n,' W Z < W

4

2

v

FIRST

w

SECOND

THIRD

FOURTH

RECALL POSITIONS OF ITEMS (QUARTILES) ii

~5YEARS

~8YEARS

~

11YEARS

Fig. 6. Criterion nonmonotonicity and developmental sharpening in children's free recall of word lists (data from Brainerd et al., 1990).

accuracy enhancement all follow: More difficult items have early output positions (weak priority) because recall begins with direct access (which retrieves traces in a weaker-to-stronger order), the overall output order is harder-easier-harder on later trials (nonmonotonic emergence) and at criterion (criterion nonmonotonicity) because traces are first processed in weaker-to-stronger order (direct access) and then in stronger-to-weaker order (reconstructive retrieval), and grouping output by item difficulty increases total recall (accuracy enhancement) because output from the more accurate direct access operation is maximized. Finally, developmental sharpening of these effects follows from other evidence showing developmental improvements in both gist-based reconstruction (Bjorklund & Muir, 1988) and direct access to verbatim traces (Brainerd, Reyna, & Kneer, 1995). 2. Robust False Recall As noted in section IV. B, Deese/Roediger/McDermott (DRM) lists produce robust false recall as well as high levels of false recognition, with critical distractors typically intruding in more than one-third of free recall protocols (e.g., Roediger & McDermott, 1995). Payne (Payne et al., 1996; Payne & Elie, 1997, 1998) used fuzzy-trace theory's dual-retrieval model to explain this finding and to generate several further predictions about it. According to Payne, targets can be recalled either by directly accessing their verbatim traces or by reconstructively processing

C. J. Brainerd and V. E Reyna

88

--- 1 ..1 1=1

~=

o n,. 0.8 z

o0.6 r,,,

~"0.4 Z tll >

,~0.2 ~

0

1

2

3

OUTPUT QUINTILE

4

5

Fig. 7. Positions of critical distractor intrusions in free recall (data from Payne et al., 1996).

gist memories of list themes, but reconstructive processing is the cause of false recall of semantic associates. Reconstructive processing produces critical distractor intrusions at high rates for two reasons. First, the fact that critical distractors are the most typical examples of list themes makes them more likely to be reconstructively generated than the targets themselves (see Brainerd & Reyna, 1998b). Second, critical distractors overlap so completely with list themes that once they are reconstructed, judgment checks are very likely to pass them on for output. This explanation leads to several predictions about false recall in the Deese/ Roediger/McDermott paradigm that have been studied by Payne (Payne et al., 1996; Payne & Elie, 1997, 1998), McDermott (1996), and Togila and associates (Toglia & Neuschatz, 1996, 1997; Toglia, Neuschatz, & Goodwin, 1999). An obvious prediction is that when critical distractors are falsely recalled, they will tend to be recalled in later output positions (when retrieval is likely to have shifted to reconstruction) rather than in earlier positions (when direct access predominates). To test this prediction, Payne et al. (1996) selected free recall protocols in which critical distractors intruded and plotted their output positions using Vincentized quintiles. The results that Payne et al. reported are displayed in Figure 7, where it can be seen that roughly half of the intrusions occurred in the final quintile of output. Three other predictions are that critical distractor intrusions should decrease as further study trials are provided, should increase under conditions of forced recall (requiring that participants output a minimum number of items that exceeds the number that they usually output), and should increase under conditions of repeated testing. We noted in Section IV.B that repetition seems to increase the accessibility of verbatim traces (in normal adults), which means that it should reduce false recall by reducing reliance on reconstructive processing of gist. McDermott (1996) obtained the predicted reductions in false recall. The other two predictions follow because forced recall increases reliance on reconstructive processing by ensuring that output must exceed the limits of verbatim access and because repeated testing

Dual Processes in Fuzzy-Trace Theory

89

will generate increased levels of interference (on later tests) that will increase reliance on gist processing by impairing access to verbatim traces. As expected, Payne et al. found 78% more intrusions with forced recall than with standard recall, and they found that intrusions increased by 30% across a series of three free-recall tests. Toglia and associates (Toglia & Neuschatz, 1996, 1997; Toglia et al., 1999) studied the effects of four other variables that are relevant to Payne's dual-retrieval account of robust false recall. First, as indicated earlier, target's meanings can be processed before their surface forms have been fully encoded. Therefore, the dual-retrieval model anticipates that intrusions of critical distractors will increase as exposure durations for targets decrease because shorter durations will impair the storage of detailed verbatim traces more than the storage of gist traces. Toglia and Neuschatz (1996) confirmed that intrusions of critical distractors increase as target exposure duration decreases. A second prediction is that study-phase instructions that direct participants to focus on the meaning of targets, rather than their surface form, should also increase false recall by increasing reliance on reconstructive retrieval (Reyna & Kiernan, 1994, 1995). Toglia et al. confirmed that such instructions did, indeed, elevate levels of false recall. Next, if participants study a single list composed of multiple DRM lists (e.g., Kensinger & Schacter, 1999), the core themes of individual lists will be more strongly cued if the targets that comprise each list are blocked together for study than if targets from different lists are mingled. Therefore, a third prediction is that levels of false recall should be higher following blocked presentation because of a stronger tendency to reconstructively process gist memories of lists' core themes. As predicted, Toglia et al. and Payne et al. found higher levels of false recall following blocked presentation than following random presentation. Finally, if participants recall orally, presenting Deese/Roediger/McDermott lists as pictures should impede direct access during recall, relative to the usual word presentation, because verbatim traces of pictures are not representations of phonetic information that can easily be vocalized. Again, as predicted, Toglia and Neuschatz (1997) found more critical distractor intrusions following DRM picture lists than following DRM word lists.

V. Summary Fuzzy-trace theory has evolved in response to counterintuitive data on how memory development influences the development of reasoning. The two traditional perspectives on memory-reasoning relationsmthe necessity and constructivist hypothesesmstipulate that the accuracy of children's memory for problem information and the accuracy of their reasoning are closely intertwined, albeit for different reasons. However, contrary to necessity, correlational and experimental dissociations have been found between children's memory for problem information

90

C.J. Brainerd and V. E Reyna

that is determinative in solving certain problems and their solutions of those problems. In these same tasks, age changes in memory for problem information appear to be dissociated from age changes in reasoning. Contrary to constructivism, correlational and experimental dissociations also have been found between children's performance on memory tests for actual experience and memory tests for the meaning of experience. As in memory-reasoning studies, age changes in one type of memory performance do not seem to be closely connected to age changes in the other type of performance. Subsequent experiments have led to dual-process accounts in both the memory and reasoning spheres. The account of memory development features four other principles: parallel verbatim-gist storage, dissociated verbatim-gist retrieval, memorial bases of conscious recollection, and identity/similarity processes. The account of the development of reasoning features three principles: gist extraction, fuzzy-to-verbatim continua, and fuzzy-processing preferences. The fuzzyprocessing preference is a particularly important notion because it implies that gist-based intuitive reasoning often suffices to deliver "logical" solutions and that such reasoning confers multiple cognitive advantages that enhance accuracy. The explanation of memory-reasoning dissociations in cognitive development then falls out of fuzzy-trace theory's dual-process models of memory and reasoning. More explicitly, in childhood reasoning tasks, it is assumed that both verbatim and gist traces of problem information are stored. Responding accurately to memory tests for presented problem information depends primarily on verbatim memory abilities (preserving traces of that information and accessing them when the appropriate memory probes are administered). However, accurate solutions to reasoning problems depend primarily on gist-memory abilities (extracting the correct gist from problem information, focusing on that gist during reasoning, and accessing reasoning operations that process that gist). Because verbatim and gist memories exhibit considerable dissociation, both during storage and when they are subsequently accessed on memory tests, dissociations of verbatim-based memory performance from gist-based reasoning are predictable. Conversely, associations are predicted in situations in which memory and reasoning are based on the same verbatim traces (Brainerd & Reyna, 1988) and in situations in which memory and reasoning are based on the same gist traces (Reyna & Kiernan, 1994). Fuzzy-trace theory's memory and reasoning principles have been applied in other research domains. Four such domains are developmental cognitive neuroscience studies of false memory, studies of false memory in brain-damaged patients, studies of reasoning errors in judgment and decision making, and studies of retrieval mechanisms in recall. In the first domain, the principles of parallel verbatim-gist storage, dissociated verbatim-gist retrieval, and identity/similarity processes have been used to explain both spontaneous and implanted false reports

Dual Processes in Fuzzy-Trace Theory

91

in children and in the elderly. These explanations have produced some surprising predictions that have been verified: false reports do not merely decline with age during childhood but increase under theoretically specified conditions; reports of events that were not experienced can nevertheless be highly persistent over time; and false reports can be suppressed by retrieving verbatim traces of corresponding true events. In the second domain, the same principles have been invoked to explain why some forms of brain damage lead to elevated levels of false memory and other forms lead to reduced levels of false memory. In the third domain, the principles of gist extraction, fuzzy-to-verbatim continua, and fuzzy-processing preferences have been exploited to formulate a general theory of loci of processing failures in judgment and decision making, cluminating in a developmental account of degrees of rationality that distinguishes more and less advanced reasoning. This theory has in turn been used to formulate local models, such as the inclusion illusions model, that explain the characteristic reasoning errors that are observed on specific judgment and decision-making tasks. Finally, in the fourth domain, a dual-process conception of recall has been derived from the principles of parallel verbatim-gist storage and dissociated verbatim-gist retrieval. In this conception, which has been used to explain cognitive triage effects in recall and robust false recall, targets are recalled either by directly accessing their verbatim traces and reading the retrieved information out of consciousness or by reconstructively processing their gist traces.

ACKNOWLEDGMENTS Preparation of this paper was supported by grants from the National Science Foundation (SBR9730143), the National Institutes of Health, the Department of Commerce, and the Academic and Managed Health Care Forum.

REFERENCES Ackerman, B. P. (1992). The sources of children's errors in judging causal inferences. Journal of Experimental Child Psychology, 54, 90-119. Ackerman, B. E (1994). Children's sourceerrorsin referentialcommunication.Journal of Experimental Child Psychology, 58, 432-464. Adams-Price, C., & Perlmutter, M. (1992). Eyewitness memory and aging research: A case study in everyday memory. In R. L. West & J. D. Sinnott (Eds.), Everyday memory and aging: Current research and methodology (pp. 246-258). New York: Springer-Verlag. Anderson, J. R., & Bower,G. H. (1972). Recognitionand retrievalprocessesin free recall.Psychological Review, 73, 97-123. Ankrum, C., & Palmer, J. ( 1989, November). The perception and memory of objects and their parts. Paper presented at the meeting of the Psychonomic Society, Atlanta, GA.

92

C. J. Brainerd and V. E Reyna

Atkinson, R. C., & Juola, J. E (1973). Factors influencing speed and accuracy in word recognition. In S. Kornblum (Ed.), Attention and performance IV. New York: Academic Press. Atkinson, R. C., & Juola, J. E (1974). Search and decision processes in recognition memory. In D. H. Krantz, R. C. Atkinson, R. D. Luce, & E Suppes (Eds.), Contemporary developments in mathematical psychology (Vol. 1): Learning, memory, and thinking. San Francisco, CA: Freeman. Atkinson, R. C., & Wescourt, K. T. (1975). Some remarks on a theory of memory. In E M. A. Rabbitt & S. Dornic (Eds.), Attention and performance V. London: Academic Press. Balota, D. A., Cortese, M. J., Duchek, J. M., Adams, D., Roediger, J. L., McDermott, K. B., & Yerys, B. E. (1999). Veridical and false memories in healthy older adults. Cognitive Neuropsychology, 16, 361-384. Bartlett, E C. (1932). Remembering: A study in experimental and social psychology. Cambridge, England: Cambridge University Press. Battig, W. E (1965). Further evidence that strongest free-recall items are not recalled first. Psychological Reports, 17, 745-746. Battig, W. E, Allen, M., & Jensen, A. R. (1965). Priority of free recall of newly learned items. Journal of Verbal Learning and Verbal Behavior, 4, 175-179. Belli, R. E, Windschitl, E D., McCarthy, T. T., & Winfrey, S. E. (1992). Detecting memory impairment with a modified test procedure: Manipulating retention interval with centrally presented event items. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 356-367. Berlyne, D. E. (1970). Children's reasoning and thinking. In P. H. Mussen (Ed.),Carmichael's manual of child psychology (Vol. 1, pp. 939-982). New York: Wiley. Binet, A. (1900). La suggestibilitY. Paris: Schleicher Feres. Bjorklund, D. E (1989). Children's thinking: Developmental and individual differences. Pacific Grove, CA: Brooks/Cole. Bjorklund, D. E, & Muir, J. E. (1988). Children's development of free recall memory: Remembering on their own. Annals of Child Development, 5, 79-123. Bowden, E. M., & Beeman, M. J. (1998). Getting the right idea: Semantic activation in the right hemisphere may help solve insight problems. Psychological Science, 9, 435-440. Brainerd, C. J. (1981). Working memory and the developmental analysis of probability judgment. Psychological Review, 88, 463-502. Brainerd, C. J. (1995). Interference processes in memory development: The case of cognitive triage. In E N. Dempster & C. J. Brainerd (Eds.), Interference and inhibition in cognition (pp. 105-139). San Diego: Academic Press. Brainerd, C. J. (1997). Development of forgetting, with implications for memory suggestibility. In N. L. Stein, E A. Ornstein, B. Tversky, & C. J. Brainerd (Eds.),Memoryfor everyday and emotional events (pp. 209-235). Hillsdale, NJ: Erlbaum. Brainerd, C. J., & Gordon, L. L. (1994). Development of verbatim and gist memory for numbers. Developmental Psychology, 30, 163-177. Brainerd, C. J., & Kingma, J. (1984). Do children have to remember to reason? A fuzzy-trace theory of transitivity development. Developmental Review, 4, 311-377. Brainerd, C. J., & Kingma, J. (1985). On the independence of short-term memory and working memory in cognitive development. Cognitive Psychology, 17, 210-247. Brainerd, C. J., & Mojardin, A. H. (1998). Children's false memories for sentences: Long-term persistence and mere testing effects. Child Development, 69, 1361-1377. Brainerd, C. J., Olney, C. A., & Reyna, V. E (1993). Optimization versus effortful processing in children's cognitive triage: Criticisms, reanalyses, and new data. Journal of Experimental Child Psychology, 55, 353-373. Brainerd, C. J., & Poole, D. A. (1997). Long-term survival of children's false memories: A review. Learning and Individual Differences, 9, 125-152.

Dual Processes in Fuzzy-Trace Theory

93

Brainerd, C. J., & Reyna, V. E (1988). Generic resources, reconstructive processing, and children's mental arithmetic. Developmental Psychology, 24, 324-334. Brainerd, C. J., & Reyna, V. E (1990a). Gist is the grist: Fuzzy-trace theory and the new intuitionism. Developmental Review, 10, 3-47. Brainerd, C. J., & Reyna, V. F. (1990b). Inclusion illusions: Fuzzy-trace theory and perceptual salience effects in cognitive development. Developmental Review, 10, 365-403. Brainerd, C. J., & Reyna, V. E (1992). Explaining "memory free" reasoning. Psychological Science, 3, 332-339. Brainerd, C. J., & Reyna, V. F. (1993a). Domains of fuzzy-trace theory. In M. L. Howe & R. Pasnak (Eds.), Emerging themes in cognitive development (Vol. 1, pp. 50-94). New York: Springer-Verlag. Brainerd, C. J., & Reyna, V. E (1993b). Memory independence and memory interference in cognitive development. Psychological Review, 100, 42-67. Brainerd, C. J., & Reyna, V. F. (1995). Autosuggestibility in memory development. Cognitive Psychology, 28, 65-101. Brainerd, C. J., & Reyna, V. F. (1996). Mere memory testing creates false memories in children. Developmental Psychology, 32, 467-476. Brainerd, C. J., & Reyna, V. E (1998a). Fuzzy-trace theory and children's false memories. Journal of Experimental Child Psychology, 71, 81-129. Brainerd, C. J., & Reyna, V. E (1998b). When things that never happened are easier to remember than things that did. Psychological Science, 9, 484-489. Brainerd, C. J., Reyna, V. F., & Brandse, E. (1995). A Children's false memories more persistent than their true memories? Psychological Science, 6, 359-364. Brainerd, C. J., Reyna, V. F., Harnishfeger, K. K., & Howe, M. L. (1993). Is retrievability grouping good for recall? Journal of Experimental Psychology: General, 122, 249-268. Brainerd, C. J., Reyna, V. F., & Howe, M. L. (1990). Children's cognitive triage: Optimal retrieval or effortful processing? Journal of Experimental Child Psychology, 49, 428-447. Brainerd, C. J., Reyna, V. F., Howe, M. L., & Kevershan, J. (1991). Fuzzy-trace theory and cognitive triage in memory development. Developmental Psychology, 27, 351-369. Brainerd, C. J., Reyna, V. F., & Kneer, R. (1995). False-recognition reversal: When similarity is distinctive. Journal of Memory and Language, 34, 157-185. Brainerd, C. J., Reyna, V. F., & Mojardin, A. H. (1999). Conjoint recognition. Psychological Review, 106, 160-179. Brainerd, C. J., Reyna, V. F., & Poole, D. A. (2000a). Fuzzy-trace theory and false memory: Memory theory in the courtroom. In D. F. Bjorklund (Ed.), False memory creation (pp. 93-127). Mahwah, N J: Erlbaum. Brainerd, C. J., Reyna, V. F., Wright, R., & Mojardin, A. H. (2000b). Recollection rejection: Editing false memories. Submitted for publication. Brainerd, C. J., Stein, L. M., & Reyna, V. F. (1998). On the development of conscious and unconscious memory. Developmental Psychology, 34, 342-357. Brainerd, C. J., Wright, R., Reyna, V. F., & Mojardin, A. H. (2001). Conjoint recognition and phantom recollection. Journal of Experimental Psychology: Learning, Memory, and Cognition 27, 307-327. Bransford, J. D., & Franks, J. J. (1971). The abstraction of linguistic ideas. Cognitive Psychology, 3, 193-209. Brown, A. L., Smiley, S. S., Day, J. D., Townsend, M. A. R., & Lawton, S. C. (1977). Intrusion of a thematic idea in children's comprehension and retention of stories. Child Development, 48, 1454-1466. Bruck, M., & Ceci, S. J. (1997). The nature of applied and basic research on children's suggestibility. In N. Stein, P. A. Ornstein, B. Tversky, & C. J. Brainerd (Eds.), Memory for everyday and emotional events (pp. 371-400). Mahwah, NJ: Erlbaum.

94

C. J. Brainerd and V. F. Reyna

Bruck, M., Ceci, S. J., & Melnyk, L. (1997). External and internal sources of variation in the creation of false reports in children. Learning and Individual Differences, 9, 289-316. Bryant, E, & Trabasso, T. (1971). Transitive inferences and memory in young children. Nature, 232, 457-459. Budson, A. E., Daffner, K. R., Desikan, R., & Schacter, D. L. (2001). When false recognition is unopposed by true recognition: Gist-based memory distortions in Alzheimer's disease. Neuropsychology. In press. Callahan, E (1989). Learning and development of probability concepts: Effects of computer assisted instruction and diagnosis. Unpublished doctoral dissertation, University of Arizona. Cassel, W. S., & Bjorklund, D. E (1995). Developmental patterns of eyewitness memory and suggestibility: An ecologically-based short-term longitudinal study. Law and Human Behavior, 19, 507-532. Ceci, S. J., & Bruck, M. (1993). Suggestibility of the child witness: A historical review and synthesis. Psychological Bulletin, 113, 403-439. Ceci, S. J., & Bruck, M. (1995). Jeopardy in the courtroom. Washington, DC: American Psychological Association. Ceci, S. J., Ross, D. F., & Toglia, M. E (1987). Suggestibility in children's memory: Psycholegal implications. Journal of Experimental Psychology: General, 116, 38-49. Chapman, M., & Lindenberger, U. (1992). Transitivity judgments, memory for premises, and models of children's reasoning. Developmental Review, 12, 124-163. Clark, S. E., & Gronlund, S. D. (1996). Global matching models of recognition memory: How the models match the data. Psychonomic Bulletin & Review, 3, 37-60. Curran, T., Schacter, D. L., Norman, K. A., & Galluccio, L. (1997). False recognition after a right frontal lobe infarction: Memory for general and specific information. Neuropsychological, 35, 1035-1049. Davidson, D. (1995). The representativeness heuristic and the conjunction fallacy in children's decision making. Merrill-Palmer Quarterly, 41, 328-346. Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in immediate recall. Journal of Experimental Psychology, 58, 17-22. DeMarie-Dreblow, D. (1991). Relation between knowledge and memory: A reminder that correlation does not imply causality. Child Development, 62, 484-498. Dempster, F. N., & Corkill, A. J. (1999). Interference and inhibition in cognition and behavior: Unifying themes for educational psychology. Educational Psychology Review, 11, 1-88. Denton, L. (1994, December). Interim report spells out five conclusions. APA Monitor, p. 9. Eddy, D. M. (1982). Probabilistic reasoning in clinical medicine: Problems and opportunities. In D. Kahneman, E Slovic, & A. Tuersky (Eds.), Judgment under uncertainty: Heuristics and biases. (pp. 249-267). New York: Cambridge University Press. Elkind, D. (1967). Piaget's conservation problems. Child Development, 38, 15-28. Felzen, E., & Anisfeld, M. (1971). Semantic and phonetic relations in false recognition of words by third- and sixth-grade children. Developmental Psychology, 3, 163-168. Gabrieli, J. D. E. (1998). Cognitive neuroscience of human memory. Annual Review of Psychology, 49, 87-115. Gardiner, J. M., & Java, R. I. (1991). Forgetting in recognition memory with and without recollective experience. Memory & Cognition, 19, 617-623. Gehrig, R. J., & Healy, A. E (1983). Dual processes in metaphor understanding: Comprehension and appreciation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, 667-675. Gernsbacher, M. A. (1985). Surface information loss in comprehension. Cognitive Psychology, 17, 324-363. Gigerenzer, G. (1994). Why the distinction between single-event probabilities and frequencies is important for psychology (and vice versa). In G. Wright & P. Action (Eds.), Subjective Probability (pp. 129-161). New York: Wiley.

Dual Processes in Fuzzy-Trace Theory

95

Gillund, G., & Shiffrin, R. M. (1984). A retrieval model for both recognition and recall. Psychological Review, 91, 1-67. Graf, E, & Schacter, D. L. (1987). Selective effects of interference on implicit and explicit memory for new associations. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13, 45-53. Hastie, R., & Park, B. (1986). The relationship between memory and judgment depends on whether the judgment task is memory-based or on-line. Psychological Review, 93, 258-268. Hess, T. M. (1984). Effects of semantically related and unrelated contexts on recognition memory of different-aged adults. Journal of Gerontology, 39, 444-451. Hintzman, D. L. (1988). Judgments of frequency and recognition memory in a multiple-trace memory model. Psychological Review, 95, 528-551. Hogarth, R. M. S. (1980). Judgment and choice: The psychology of decision. Chichester, England: Wiley. Hook, J. G., & Cook, T. D. (1979). Equity theory and the cognitive ability of children. Psychological Bulletin, 86, 429-445. Howe, M. L. (1991). Misleading children's story recall: Forgetting and reminiscence of the facts. Developmental Psychology, 27, 746-762. Howe, M. L. (2000). The fate of early memories. Washington, DC: American Psychological Association. Howe, M. L., Rabinowitz, E M., & Grant, M. J. (1993). On measuring (in)dependence of cognitive processes. Psychological Review, 100, 737-747. Inhelder, B., & Piaget, J. (1958). The growth of logical thinking from childhood to adolescence. New York: Basic Books. Israel, L., & Schacter, D. L. (1997). Pictorial encoding reduces false recognition of semantic associates. Psychonomic Bulletin & Review, 4, 577-581. Jacobs, J. E., & Potenza, M. (1991 ). The use of judgment heuristics to make social and object decisions: A developmental perspective. Child Development, 62, 166-178. Jacoby, L. L. (1991). A process dissociation framework: Separating automatic from intentional uses of memory. Journal of Memory and Language, 30, 513-541. Jarvella, R. J. ( 1971 ). Syntactic processing of connected speech. Journal of Verbal Learning and Verbal Behavior, 10, 409-416. Johnson, M. K., Nolde, S. E, Mather, M., Kounios, J., & Schacter, D. L. (1997). Test format can affect the similarity of brain activity associated with true and false recognition memory. Psychological Science, 8, 250-257. Juola, J. F., Fischler, I., Wood, C. T., & Atkinson, R. C. (1971 ). Recognition time for information stored in long-term memory. Perception & Psychophysics, 10, 8-14. Kahneman, D., Slovic, E, & Tversky, A. (1982). Judgment under uncertainty: Heuristics and biases. New York: Cambridge University Press. Kensinger, E. A., & Schacter, D. L. (1999). When true memories suppress false memories: Effects of aging. Cognitive Neuropsychology, 16, 399-415. Kitsch, W., Welch, D., Smalhofer, F., & Zimny, S. (1990). Sentence memory: A theoritical analysis. Journal of Memory and Language, 29, 133-159. Klaczynski, P. A. (2001). The development of adolescent judgment and decision making abilities: A two-process model of heuristic and analytical reasoning. Child Development. In Press. Klaczynski, E A., & Fauth, J. (1997). Developmental differences in memory-based intrusions and self-serving statistical reasoning biases. Merrill-Palmer Quarterly, 43, 539-566. Klaczynski, P. A., & Narasimham, G. (1998). Problem representations as mediators of adolescent deductive reasoning. Developmental Psychology, 34, 865-881. Koriat, A., Goldsmith, M., & Pansky, A. (2000). Toward a psychology of memory accuracy. Annual Review of Psychology, 51, 481-538.

96

C.J. Brainerd and V. E Reyna

Koutstaal, W., & Schacter, D. L. (1997). Gist-based false recognition of pictures in older and younger adults. Journal of Memory and Language, 37, 555-583. Koutstaal, W., Schacter, D. L., Verfaellie, M., Brenner, C., & Jackson, E. M. (1999). Perceptually based false recognition of novel objects in amnesia: Effects of category size and similarity to category prototypes. Cognitive Neuropsychology, 16, 317-341. Kroll, N. E. A., Knight, R. T., Metcalfe, J., Wolfe, E. S., & Tulving, E. (1996). Cohesion failure as a source of memory illusions. Journal of Memory and Language, 35, 176-196. Kuhberger, A. (1995). The framing of decisions: A new look at old problems. Organizational Behavior and Human Decision Processes, 62, 230-240. Kuhberger, A. (1998). The influence of framing on risky decisions. Organizational Behavior and Human Decision Processes, 75, 23-55. Lampinen, J. M., Neuschatz, J. S., & Payne, D. G. (1998). Memory illusions and consciousness: Examining the phenomenology of true and false memories. Current Psychology: Development, Learning, Personality, Social, 16, 181-224. Langer, J. (1970). Werner's theory of development. In P. H. Mussen (Ed.), Carmichael's manual of child psychology (Vol. 1, pp. 733-772). New York: Wiley. Levin, I. P., Schneider, S. L., & Gaeth, G. J. (1998). All frames are not created equal: A typology and critical analysis of framing effects. Organizational Behavior and Human Decision Processes, 76, 149-188. Loftus, E. E (1993). The reality of repressed memories. American Psychologist, 48, 518-537. Loftus, E. E (1997). Creating false memories. Scientific American, September, 277, 70-75. MacLeod, C. M., & Nelson, T. O. (1976). A nonmonotonic lag function for false alarms to associates. American Journal of Psychology, 89, 127-135. Mandler, G. (1980). Recognizing: The judgment of previous occurrence. Psychological Review, 87, 252-271. Marx, M. H., & Henderson, B. (1996). A fuzzy trace analysis of categorical inferences and instantial associations as a function of retention interval. Cognitive Development, 11, 551-569. McDermott, K. B. (1996). The persistence of false memories in list recall. Journal of Memory and Language, 35, 212-230. McGough, L. S. (1993). Child witnesses: Fragile voices in the American legal system. New Haven, CT: Yale University Press. Melo, B., Winocur, G., & Moscovitch, M. (1999). False recall and false recognition: An examination of the effects of selective and combined lesions to the medial temporal lobe/diencephalon and frontal lobe structures. Cognitive Neuropsychology, 16, 343-359. Metcalfe, J., Funnell, M., & Gazzaniga, M. S. (1995). Right-hemisphere memory superiority: Studies of a split-brain patient. Psychological Science, 6, 157-164. Moore, C. E, Dixon, J. A., & Haines, B. A. (1991). Components of understanding in proportional reasoning: A fuzzy set representation of developmental progressions. Child Development, 62, 441459. Moravcski, J. E., & Healy, A. E (1995). Effect of meaning on letter detection. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 82-95. Murdock, B. B., Jr. (1982). A theory for storage and retrieval of item and associative information. Psychological Review, 89, 609-626. Murphy, G. L., & Shapiro, A. M. (1994). Forgetting of verbatim information in discourse. Memory & Cognition, 22, 85-94. Nisbett, R. E., & Ross, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231-259. Norman, K. A., & Schacter, D. L. (1997). False recognition in young and older adults: Exploring the characteristics of illusory memories. Memory & Cognition, 25, 838-848.

Dual Processes in Fuzzy-Trace Theory

97

Offenbach, S. I., Gruen, G. E., & Caskey, B. J. (1984). Development of proportional response strategies. Child Development, 55, 963-972. Paris, S. G., & Carter, A. Y. (1973). Semantic and constructive aspects of sentence memory in children. Developmental Psychology, 9, 109-113. Parkin, A. J. (1997). The neuropsychology of false memory. Learning and Individual Differences, 9, 341-357. Parkin, A. J., Bindschaedler, C., Harsent, L., & Metzler, C. (1996). Verification impairment in the generation of memory deficit following ruptured aneurysm of the anterior communicating artery. Brain and Cognition, 32, 14-27. Parkin, A. J., Ward, J., Bindschaedler, C., Squires, E. J., & Powell, G. (1999). False recognition following frontal lobe damage: The role of encoding factors. Cognitive Neuropsychology, 16, 243265. Payne, D. G., & Elie, C. J. (1997, November). Comparing false memories for pictures and words. Paper presented at the meeting of the Psychonomic Society, Philadelphia, PA. Payne, D. G., & Elie, C. J. (1998, November). Repeated list presentation reduces false memories for pictures and words. Paper presented at the meeting of the Psychonomic Society, Dallas, TX. Payne, D. G., Elie, C. J., Blackwell, J. M., & Neuschatz, J. S. (1996). Memory illusions: Recalling, recognizing, and recollecting events that never occurred. Journal of Memory and Language, 35, 261-285. Pezdek, K., & Roe, C. (1995). The effect of memory trace strength on suggestibility. Journal of Experimental Child Psychology, 60, 116-128. Pezdek, K., & Roe, C. (1996). The suggestibility of children's memory for being touched: Planting, erasing, and changing memories. Submitted for publication. Phelps, E., & Gazzaniga, M. S. (1992). Hemispheric differences in mnemonic processing: The effects of left hemisphere interpretation. Neuropsychological, 30, 293-297. Piaget, J. (1967). The psychology of intelligence. New York: Random House. Piaget, J. (1970). Genetic epistemology. New York: Columbia University Press. Piaget, J., & Inhelder, B. (1969). The psychology of the child. New York: Basic Books. Piaget, J., & Inhelder, B. (1973). Memory and intelligence. New York: Basic Books. Poole, D. A. (1995). Strolling fuzzy-trace theory through eyewitness testimony (or vice versa). Learning and Individual Differences, 7, 87-94. Poole, D. A., & Lamb, M. E. (1998). Investigative interviews of children. Washington, DC: American Psychological Association. Quas, J. A., Qin, J., Schaaf, J. M., & Goodman, G. S. (1997). Individual differences in children's and adults' suggestibility and false event memory. Learning and Individual Differences, 9, 359-390. Rabinowitz, E M., Grant, M. J., Howe, M. L., & Walsh, C. (1994). Reasoning in middle childhood: A dynamic-model of performance on transitivity tasks. Journal of Experimental Child Psychology, 58, 252-288. Rabinowitz, F. M., Howe, M. L., & Lawrence, J. A. (1989). Class inclusion and working memory. Journal of Experimental Child Psychology, 48, 379-409. Rankin, J. S., & Kausler, D. H. (1979). Adult age differences in false recognition. Journal of Gerontology, 34, 58-65. Rapcsak, S. Z., Reminger, S. L., Glisky, E. L., Kaszniak, A. W., & Comer, J. E (1999). Neuropsychological mechanisms of false facial recognition following frontal lobe damage. Cognitive Neuropsychology, 16, 267-292. Reder, L. M., & Ritter, E E. (1992). What determines initial feeling of knowing--familiarity with question terms, not with the answer. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 435-451.

98

C. J. Brainerd and V. E Reyna

Reyna, V. E (1991). Class inclusion, the conjunction fallacy, and other cognitive illusions. Developmental Review, 11, 317-336. Reyna, V. E (1992). Reasoning, remembering, and their relationship: Social, cognitive, and developmental issues. In M. L. Howe, C. J. Brainerd, & V. E Reyna (Eds.), Development of long-term retention (pp. 103-127). New York: Springer-Verlag. Reyna, V. E (1995). Interference effects in memory and reasoning: A fuzzy-trace theory analysis. In F. N. Dempster & C. J. Brainerd (Eds.), New perspectives on interference and inhibition in cognition (pp. 29-61). San Diego: Academic Press. Reyna, V. E (1996a). Conceptions of memory development, with implications for reasoning and decision making. Annals of Child Development, 12, 87-118. Reyna, V. F. (1996b, November). Repetition dissociates verbatim and gist memory for narratives. Paper presented at the meeting of the Psychonomic Society, Chicago, IL. Reyna, V. E (1998). Fuzzy-trace theory and false memory. In M. Intons-Peterson & D. Best (Eds.), Memory distortions and their prevention (pp. 15-27). Hillsdale, NJ: Erlbaum. Reyna, V. E (2001). Data, development, and dual processes in rationality. Brain and Behavioral Sciences. In press. Reyna, V. E, & Brainerd, C. J. (1990). Fuzzy processing in transitivity development. Annals of Operations Research, 23, 37-63. Reyna, V. E, & Brainerd, C. J. (1991a). Fuzzy-trace theory and the acquisition of scientific and mathematical concepts. Learning and Individual Differences, 3, 27-60. Reyna, V. F., & Brainerd, C. J. (199 lb). Fuzzy-trace theory and framing effects in choice: Gist extraction, truncation, and conversion. Journal of Behavioral Decision Making, 4, 249-262. Reyna, V. E, & Brainerd, C. J. (1992). A fuzzy-trace theory of reasoning and remembering: Patterns, paradoxes, and parallelism. In A. Healy, S. Kosslyn, & R. Shiffrin (Eds.), From learning processes to cognitive processes: Essays in honor of William K. Estes (pp. 235-259). Hillsdale, NJ: Erlbaum. Reyna, V. E, & Brainerd, C. J. (1993). Fuzzy memory and mathematics in the classroom. In R. Logie & G. Davies (Eds.), Everyday memory (pp. 91-119). Amsterdam: North-Holland. Reyna, V. E, & Brainerd, C. J. (1994). The origins of probability judgment: A review of data and theories. In G. Wright & P. Acton (Eds.), Subjective Probability (pp. 239-272). New York: Wiley. Reyna, V. E, & Brainerd, C. J. (1995). Fuzzy-trace theory: An interim synthesis. Learning and Individual Differences, 7, 1-75. Reyna, V. E, & Brainerd, C. J. (1998). Fuzzy-trace theory and false memory: New frontiers. Journal of Experimental Child Psychology, 71, 194-209. Reyna, V. E, & Ellis, S. C. (1994). Fuzzy-trace theory and framing effects in children's risky decision making. Psychological Science, 5, 275-279. Reyna, V. E, & Kieman, B. (1994). The development of gist versus verbatim memory in sentence recognition: Effects of lexical familiarity, semantic content, encoding instructions, and retention interval. Developmental Psychology, 30, 178-191. Reyna, V. E, & Kieman, B. (1995). Children's memory and interpretation of psychological metaphors. Metaphor and Symbolic Activity, I0, 309-331. Reyna, V. E, & Lloyd, E (1997). Theories of false memory in children and adults. Learning and Individual Differences, 9, 95-124. Reyna, V. E, Lloyd, E J., & Brainerd, C. J. (2001). Memory, development, and rationality: An integrative theory of judgment and decision making. In s. Schneider & J. Shanteau (Eds.), Emerging perspectives on decision research. New York: Cambridge University Press. In press. Reyna, V. F., & Titcomb, A. L. (1997). Constraints on the suggestibility of eyewitness testimony: A fuzzy-trace theory analysis. In D. G. Payne & E G. Conrad (Eds.), A synthesis of basic and applied approaches to human memory (pp. 157-174). Mahwah, NJ: Erlbaum. Roediger, H. L. (1996). Memory illusions. Journal of Memory and Language, 35, 76-100.

Dual Processes in Fuzzy-Trace Theory

99

Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented on lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 803-814. Schacter, D. L. (1999). The cognitive neuroscience of false memories: An introduction. Cognitive Neuropsychology, 16, 193-195. Schacter, D. L., Israel, L., & Racine, C. (1999). Suppressing false recognition in younger and older adults: The distinctiveness heuristic. Journal of Memory and Language, 40, 1-24. Schacter, D. L., Kagan, J., & Leichtman, M. D. (1995). True and false memories in children and adults: A cognitive neuroscience perspective. Psychology, Public Policy, and Law, 1, 411-428. Schacter, D. L., Koutstaal, W., Johnson, M. K., Gross, M. S., & Angell, K. A. (1997). False recollection induced by photographs: A comparison of older and younger adults. Psychology and Aging, 12, 203-215. Schacter, D. L., Norman, K. A., & Koutstaal, W. (1998). The cognitive neuroscience of constructive memory. Annual Review of Psychology, 49, 289-318. Siegler, R. S. (1991). Children's thinking. Englewood Cliffs, NJ: Prentice-Hall. Smedslund, J. (1969). Psychological diagnostics. Psychological Bulletin, 71, 237-248. Snodgrass, J. G., & Corwin, J. (1988). Pragmatics of measuring recognition memory: Applications to dementia and amnesia. Journal of Experimental Psychology: General, 117, 34-50. Spinillo, A. G., & Bryant, P. (1991 ). Children's proportional judgments: The importance of half. Child Development, 62, 427-440. Stone, E. R., Yates, J. E, & Parker, A. M. (1994). Risk communication: Absolute versus relative expressions of low-probability risks. Organizational Behavior and Human Decision Processes, 60, 387-408. Titcomb, A. L., & Reyna, V. E (1995). Memory interference and misinformation effects. In F. N. Dempster & C. J. Brainerd (Eds.), Interference and inhibition in cognition (pp. 263-294). San Diego: Academic Press. Toglia, M. E, & Neuschatz, J. S. ( 1996, November). False memories: Where does encoding opportunity fit into the equation? Paper presented at the meeting of the Psychonomic Society, Chicago, IL. Toglia, M. P., & Neuschatz, J. S. (1997, November). False memories: Are pictures really worth a thousand words? Paper presented at the meeting of the Psychonomic Society, Dallas, TX. Toglia, M. P., Neuschatz, J. S., & Goodwin, K. A. (1999). Recall accuracy and illusory memories: When more is less. Memoo; 7, 233-256. Trabasso, T. (1977). The role of memory as a system in making transitive inferences. In R. V. Kail, Jr. & J. W. Hagen (Eds.), Perspectives on the development of memory and cognition (pp. 333-366). Hillsdale, N J: Erlbaum. Tulving, E. (1985). Memory and consciousness. Canadian Psychologist, 26, 1-12. Tulving, E., & Thomson, D. M. (1971). Retrieval processes in recognition memory: Effects of associative context. Journal of Experimental Psychology, 87, 116-124. Tun, E A., Wingfield, A., Rosen, M. J., & Blanchard, I. (1998). Older adults show greater susceptibility to false memory than young adults: Temporal characteristics of false recognition. Psychology and Aging, 13, 230-241. Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211, 453-458. Tversky, A., & Kahneman, D. (1983). Extensional versus intuitive reasoning: The conjunction fallacy in probability judgment. Psychological Review, 91, 293-315. Tversky, A., & Koehler, D. J. (1994). Support theory: A nonextensional representation of subjective probability. Psychological Review, 101, 547-567. von Winterfeldt, D., & Edwards, W. (1986). Decision analysis and behavioral research. Cambridge, England: Cambridge University Press. Wallace, W. R, Stewart, M. T., & Malone, C. P. (1995). Recognition memory errors produced by

C.J. Brainerd and V. E Reyna

100

implicit activation of word candidates during the processing of spoken words. Journal of Memory

and Language, 34, 417-439. Wallace, W. P., Stewart, M. T., Shaffer, T. R., & Barry, C. R. (1998). Are false recognitions influenced by prerecognition processing? Journal of Experimental Psychology: Learning, Memory, and Cognition, 24, 284-298. Wedell, D. H., & Bockenholt, U. (1994). Contemplating single versus multiple encounters of a risky prospect. American Journal of Psychology, 107, 499-518. Wemer, H. (1948). Comparative psychology of mental development. New York: International Universities Press. Winer, G. A. (1980). Class-inclusion reasoning in children: A review of the empirical literature. Child Development, 51, 309-328. Wolfe, C. (1995). Information seeking on Bayesian conditional probability problems: A fuzzy-trace theory account. Journal of Behavioral Decision Making, 8, 85-108.

R E L A T I O N A L F R A M E THEORY: A P O S T - S K I N N E R I A N A C C O U N T OF H U M A N L A N G U A G E A N D C O G N I T I O N

Yvonne Barnes-Holmes, ~ Steven C. Hayes, 2 D e r m o t Barnes-Holmes, 1 and Bryan Roche t 1DEPARTMENT OF PSYCHOLOGY NATIONAL UNIVERSITY OF IRELAND, MAYNOOTH KILDARE, IRELAND 2DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF NEVADA, RENO RENO, NEVADA 89557

I. I N T R O D U C T I O N II. R E L A T I O N A L F R A M E T H E O R Y AS A N A P P R O A C H TO L A N G U A G E A N D C O G N I T I O N A. THE D E F I N I N G C H A R A C T E R I S T I C S OF DERIVED RELATIONAL RESPONDING B. M U L T I P L E S T I M U L U S R E L A T I O N S A N D THE T R A N S F O R M A T I O N OF F U N C T I O N S IN H U M A N L A N G U A G E C. R E L A T I O N S A M O N G R E L A T I O N S : A N A L O G I E S , M E T A P H O R S , A N D STORIES D. T H I N K I N G , P R O B L E M S O L V I N G , A N D P R A G M A T I C V E R B A L A N A L Y S I S E. U N D E R S T A N D I N G A N D VERBAL R E G U L A T I O N E SELF A N D S E L F - D I R E C T E D RULES G. P S Y C H O L O G I C A L D E V E L O P M E N T H. E D U C A T I O N III. C O N C L U S I O N REFERENCES

I. Introduction Language and cognition are important domains in the discipline of psychology, and they are often the primary focus in the study of psychological development. In most introductory textbooks, for example, one will find sections or chapters headed "Language Development" and "Cognitive Development." In both domains, the mainstream approach within developmental psychology has been to identify relatively invariant patterns of psychological change as they unfold during the 101

ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR,VOL.28

Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.

0065-2407/01 $35.00

102

Yvonne B a r n e s - H o l m e s et al.

lifetime of the individual. Chomsky (1959) and Pinker (1991) both provide wellknown examples of this approach in the area of language, and Piaget (1967) does in the area of cognition. In contrast, behavioral psychologists tend to resist the idea of invariant sequences in language and cognitive development, and they try to show how these complex behaviors result from basic behavioral processes (e.g., Fischer, 1980). The behavioral view has had limited impact on mainstream developmental psychology. Why has behavioral psychology had such a limited role in this domain? We believe that one major contributing factor is that behavior analysis has yet to offer a viable account of human language and cognition that appears capable of handling many of the complexities and subtleties of these two domains. Perhaps the most famous (or infamous) behavioral attempt at dealing with human language was Skinner's (1957) Verbal Behavior, which was self-consciously and explicitly derived from experimental work conducted almost exclusively with nonhuman species. With its emphasis on principles wrought from nonhuman research, the book was perhaps an easy target for nonbehavioral language researchers, and it was roundly condemned for failing to address the key properties of human language (e.g., Chomsky, 1959). Even within the behavioral community the book has not gone unchallenged (e.g., Hayes & Hayes, 1992; see also Barnes-Holmes, Barnes-Holmes, & Cullinan, 2000). Specifically, some have argued that the directcontingency account of human verbal behavior presented by Skinner could never capture the symbolic and emergent qualities that define human language and cognition. Within behavioral psychology, a major empirical source of this challenge has been the growing literature on derived relational responding or relational framing. Our purpose in this article is to provide a behavioral approach to this literature that, we believe, forms the basis of a new account of language and cognition (see Hayes, Barnes-Holmes, & Roche, 2001, for a booklength treatment upon which this article is based). The account, Relational Frame Theory (RFT), is firmly rooted in the philosophical contextualism of behavior analysis and draws on many of the key behavior analytic concepts and principles. Nevertheless, RFT extends well beyond the direct-contingency account of human language and cognition presented by Skinner. Relational frame theory, like any behavioral account, is inherently developmental in orientation. A behavioral perspective is focused on streams of organismenvironment interactions, considered both historically and situationally, and therefore change is assumed and expected. The challenge for the behavioral theorist has more to do with constancy. Given the important role attributed to the stream of an individual's history, why would an order and progression in development be found? Two answers are typically given. First, a child's environment, and in particular the social context, provides a great deal of teaching in an orderly way. Second, some behaviors may be more complex than others: A child cannot run before walking. Relational frame theory adopts both of these views, but what is new is

Relational Frame Theory

103

the detailed analysis of human language and cognition that specifies why some cognitive tasks are more complex than others. In the first half of this article we address the basic features of relational frame theory. In the second half we turn our attention to a few psychological phenomena that we believe can be explained by these processes.

II. Relational Frame Theory as an Approach to Language and Cognition From the perspective of RFT, both language and cognition are products of a core behavioral process termed arbitrarily applicable relational responding. In the next subsection we outline basic relational concepts and attempt to show how they may be used to interpret some of the basic properties of human language. In subsequent sections we demonstrate how the basic relational concepts of RFT may be used to analyze more complex features of language and cognition, such as analogy, metaphor, narrative, problem solving, and self-generated instructional control. A. THE DEFINING CHARACTERISTICS OF DERIVED RELATIONAL RESPONDING

1. Definitions Relating means to respond to one event in terms of another. Most living organisms, given the appropriate training, are capable of responding to nonarbitrary relations among the physical properties of two or more stimuli. For example, adult rhesus monkeys can be trained to select the taller of two stimuli (which differ only in terms of height) when presented stimuli of varying heights. When later presented with a previously "correct" stimulus and a novel taller stimulus, the monkeys will select the novel stimulus, indicating that they are responding based on relational rather than absolute properties of the stimuli (Harmon, Strong, & Pasnak, 1982). When relational responding itself is abstracted and brought under the control of contextual features other than solely the formal properties of the stimuli, the behavior is defined as arbitrarily applicable relational responding. Such responses can be brought to bear on any stimuli encountered in the appropriate relational context. Such responding is arbitrarily applicable because in some contexts it is under the control of cues that can be modified entirely on the basis of social whim. In an educational game, for example, a child might be told that to solve a series of logical problems, "more means less, and less means more." This example illustrates how relational functions can be applied in a purely arbitrary manner. For the behavior of relating to be abstracted, the organism must learn to discriminate between the relevant features of the task (responding to one event in terms of another based on a contextual cue) and the irrelevant features (the actual physical properties of the related objects). The explicit training of symmetrical relations between words and their referents seen in early language training is a

104

Yvonne Barnes-Holmes et al.

good example. For a young child learning to name objects, early language training consists of a wealth of name-object and object-name interactions across an extensive range of objects and names. In essence, a symmetrical relation between the object name and the object itself is being directly trained (i.e., given name of object, select object and given object, select name of object). Reinforcement for such bidirectional responding is rich in a naturally occurring language training history. This kind of symmetrical responding occurs only in certain contexts, and a variety of cues indicate the task at hand, including the use of phrases such as "what's that?" and the juxtaposition of objects and words. A core claim in RFT is that if name-object relations reliably predict object-name relations in certain contexts, this extensive training history may be brought to bear and provide a basis of generalized symmetrical responding. As a result, explicit training in a new name-object relation may result in a derived object-name relation. If this generalized responding is possible with relations of identity it should also be possible with many other kinds of relations, and thus a generic set of terms is needed to describe and to define these performances. Arbitrarily applicable relational responding has three defining characteristics. Mutual entailment describes the fundamental bidirectionality of relational responding. Arbitrary stimulus relations are always mutual: If A is related to B, then B is related to A. If the first relation is specified, the second is entailed: thus the term "mutual entailment." Although these relations are bidirectional, they are not always symmetrical, for example, if A is larger than B, then B is smaller than A. Arbitrarily applicable relational responding is brought to bear on the situation by contextual stimuli beyond solely the physical properties of the relata themselves (to avoid confusion, the Latin term relata is used to describe related events since the English term, relates, is obscure and has much more dominant meanings). Thus, mutual entailment applies when, in a given context, A is related in a characteristic way to B, and as a result, in that context, B is related in another characteristic way to A. Combinatorial entailment refers to a derived stimulus relation in which two or more stimulus relations (trained or derived) mutually combine. For example, combinatorial entailment applies when, in a given context, if A is related to B and A is related to C, then B and C are mutually related in that context. For example, if hot is the opposite of cold, and caliente is the opposite of cold, then a derived mutual relation between hot and caliente is entailed (in this case, caliente and hot are the same). Combinatorial entailment must be described specifically, but (unlike mutual entailment) the derived relations may be much less precise than the original relations. For example, if A is different from B and B is different from C, we cannot say what the precise relation is between A and C and between C and A. Due to its complexity and training history, combinatorial entailment will probably emerge somewhat later in language training than mutual entailment. The term transformation of stimulus function is employed when the functions of a given stimulus are modified in accordance with derived relations with other

Relational Frame Theory

105

stimuli. Transformation of function is an important feature of relational responding because it enables derived relations to have psychological content. For example, if a person is trained to select stimulus B as the "opposite" of stimulus A, and if A is given a conditioned punishing function, such as by pairing it with a loss of points, B should have a reinforcing function (without this function having to be directly trained), because it is "opposite" to the punishing A stimulus (Dymond & Barnes, 1995; Roche & Barnes, 1997; Roche, Barnes-Holmes, Smeets, BarnesHolmes, & McGeady, 2000). Derived functions of this kind have been observed with conditioned reinforcing functions (Hayes, Brownstein, Devany, Kohlenberg, & Shelby, 1987; Hayes, Kohlenberg, & Hayes, 1991), discriminative functions (Hayes et al., 1987), elicited conditioned emotional responses (Dougher, Auguston, Markham, Greenway, & Wulfert, 1994), extinction functions (Dougher et al., 1994), and self-discrimination functions (Dymond & Barnes, 1994). The transformation of stimulus functions must come under contextual control, just as the relational response itself is controlled by context. An example would involve any two stimuli, such as the word "apple" and an actual apple, that participate in a relational class of equivalence. An apple, of course, has several stimulus functions, including perceptual functions of taste, texture, and sight. When a person says "picture an apple," most verbally able individuals would "see" an apple in the absence of the real object. In the language of RFT, this effect can be explained as follows: Actual apples have visual perceptual functions. The word "apple" and actual apples are in an arbitrarily applicable stimulus relation (in this case, an equivalence relation or "frame of coordination"). The words "picture an" are a context in which visual functions are actualized in terms of the underlying relation. In another context (e.g., "imagine tasting an ..."), other functions (e.g., taste) could be actualized. Contextual cues, therefore, not only establish and indicate the particular conditions under which relational activity occurs, they also specify which functions should be transformed (or transferred). In the early stages of language training, many "inappropriate" functions of stimuli participating in a derived stimulus relation will likely be transformed before reinforcement contingencies are able to establish contextual control. A young child told that "Uncle George is a snake" may be frightened that Uncle George will bite or be confused because Uncle George does not look like a snake. Eventually, however, the child (through differential reinforcement from the verbal community) will learn that in the context of saying someone is an animal, the perceptual functions should not transfer to the person unless additional cues are offered (e.g., the words "looks like"). The term relational frame designates a specific kind of arbitrarily applicable relational responding (Hayes & Hayes, 1989) that has three sets of properties. (a) It shows the contextually controlled qualities of mutual entailment, combinatorial entailment, and transformation of stimulus functions; (b) it results from a history of relational responding relevant to the contextual cues involved; and

106

Yvonne Barnes-Holmes et al.

(c) it is not based solely on direct nonrelational training with regard to the particular stimuli of interest, nor is it based solely on nonarbitrary characteristics of either the stimuli or the relation between them. Although the term "frame" is employed, a relational frame is always an actionm"framing events relationally." Arbitrarily applicable relational responding is the generic name for behavior of this kind; a relational frame is a specific type of such responding.

2. Types of Relational Frames Stimulus events can be related to one another in a variety of ways. These different relations are kinds of relational frames. The frame of coordination is perhaps the most common type of relational responding. This frame incorporates the relation of identity, sameness, or similarity. Naming is an example of the frame of coordination at its simplest. This relational frame is probably the first to be abstracted sufficiently to enable its application to become arbitrary, in part because it is the only relation in which derived and trained relations are the same, regardless of the size of the stimulus network. A frame of coordination is involved in an equivalence class: a set of three or more stimuli mutually related via frames of coordination (e.g., an actual apple, and the spoken and printed word "apple"). Another family of relational frames is that of opposition. In natural language use, this kind of relational responding involves an abstracted dimension along which events can be ordered and distinguished in equal ways from a reference point. Along the verbally abstracted dimension of temperature, for example, cool is the opposite of warm, and cold is the opposite of hot. The specific relational frame of opposition typically (but not necessarily) implicates the relevant dimension (e.g., "pretty is the opposite of ugly" is relevant to appearance). Opposition should emerge after coordination because the combinatorially entailed relation in frames of opposition includes frames of coordination (e.g., if hot is the opposite of freezing and cold is the opposite of hot, then cold is the same as freezing). Distinction involves responding to one event in terms of its differences from another, typically also along a particular dimension. Like a frame of opposition, this frame implies that responses to one event are unlikely to be appropriate in the case of the other, but unlike opposition, the nature of an appropriate response is typically not defined. If I am told only, for example, "this is not warm water," I do not know whether the water is ice cold or boiling hot. The family of comparative relational frames is involved whenever one event is responded to in terms of a quantitative or qualitative relation along a specified dimension with another event. Many specific subtypes of comparison exist (e.g., bigger/smaller, faster/slower, better/worse). Although each subtype may require its own history, the family resemblance may allow the more rapid learning of successive members. The different members of this family of relations are defined in part by the dimensions along which the relation applies (e.g., size, attractiveness, speed). Comparative frames may be made more specific by

Relational Frame Theory

107

quantification of the dimension along which a comparative relation is made. For example, the statement "A is twice as fast as B and B is twice as fast as C" allows a precise specification of the relation within all three pairs of elements in the network. Hierarchical relations or hierarchical class memberships have the same diodelike quality of frames of comparison, but the combinatorially entailed relations differ. For example, if Bob is the father of Dave and Barb, then Dave and Barb are known to be siblings. If Bob is taller than both Dave and Barb, however, the relative heights of Dave and Barb are unknown. Hierarchical relations are essential to many forms of verbal abstraction. Other families of relations include spatial relations such as over/under and front/back, temporal relations such as before/after, and causal~contingency frames such as " i f . . . then." Deictic relations are those that specify a relation in terms of the perspective of the speaker such as left/right; I/you (and all of its correlates, such as "mine"); here/ there; and now/then (see Barnes & Roche, 1997; Hayes, 1984). Some relations may or may not be deictic, such as front/back or above/below, depending on the perspective applied. For example, the sentence "The back door of my house is in front of me" contains both a spatial and deictic form of "front/back." Deictic relations seem to be a particularly important family of relational frames that may be critical for perspective-taking. An example is the three frames of I and YOU, HERE and THERE, and NOW and THEN (by convention we will place relational terms in all capitals when a device is needed to avoid reading the terms normally). These frames are unlike the others mentioned previously in that they do not appear to have any formal or nonarbitrary counterparts. Coordination, for instance, is based on formal identity or sameness, and "bigger than" is based on relative size. Temporal frames are particularly verbal in that they are based on the nonarbitrary experience of change, but the dimensional nature of that experience (past, present, future) must be verbally constructed. Frames that depend on perspective cannot be traced to formal dimensions in the environment at all; instead, the relationship between the individual and other events serves as the constant variable upon which these frames are based. Various relational frames can be combined into relational networks, which in turn can be related to other relational networks. For example, if one equivalence class is the opposite of another equivalence class, then each member of the first class should be the opposite of all members of the second and vice versa. These relations can continue to expand to virtually any level of complexity. For example, many relations surround a given word, such as "car." It is part of many hierarchical classes, such as the class "noun" and the class "vehicles." Other terms are in a hierarchical relation with it, such as "windshield" and "wheel." It enters into many comparisons: "it is faster than a snail, bigger than a breadbox, heavier than a book." It is the same as "automobile," but different from a "house." The participation of the word "car" in these relations is part of the training required for

108

Yvonne Barnes-Holmes et al.

the verbal community to use the stimulus "car" in the way that it does. This complexity is managed by contextual control, and thus not all derived relations are psychologically present to the same degree when the word is used. 3. Defining Verbal Events According to relational flame theory verbal behavior is the action of flaming events relationally. Speakers speak with meaning when they flame events relationally in the act of producing sequences of stimuli. If the same formal stimuli are produced but not because the speaker has flamed events relationally (e.g., when a parrot repeats what was said), then according to RFT, no verbal behavior is involved. Listeners listen with understanding when they flame events relationally (and in that sense the listener can understand both the person and the parrot). Verbal stimuli are not the products of speakers in a formal or topographical way: rather, they are stimuli that have their effects because they participate in relational flames. B. MULTIPLE STIMULUS RELATIONS AND THE TRANSFORMATION OF FUNCTIONS IN HUMAN LANGUAGE

The relational frame approach to language contains an apparent paradox. The concept of derived stimulus relations is simple at the level of process, and yet it gives rise to amazingly complex human performances. In other words, only the result of deriving stimulus relations is complex, not the process itself. Indeed, highly complex patterns of responding have already been demonstrated in a range of laboratory studies of relational framing activity (e.g., Barnes & Keenan, 1993; Dymond & Barnes, 1995, 1996; Roche & Barnes, 1996, 1997; Roche et al., 2000; Steele & Hayes, 1991). Relational frame theory applies this complexity directly to some of the more challenging aspects of the study of language and cognition. Relational frame theory provides a naturalistic, functional-analytic approach to verbal behavior. Words and symbols participate in relational networks, and the relational frames that participate in these networks help to establish the meaning or psychological functions of the network for the language user. The simple sentence "This is a cup" illustrates three points about this network. First, the word "cup" participates in a frame of coordination with the actual cup to which the speaker is referring. Second, the phrase "This is a" may participate in a frame of coordination with other contextual cues that control the frame of coordination itself (e.g., "same as," "goes with," "equivalent to"). Third, the word sequence in the sentence establishes grammatical relational frames, so that the listener responds appropriately to the statement. For instance, the phrase "This is a cup" is not the same as the question "Is this a cup?" The same words are used in each sentence, but the two sentences will typically have different effects upon the listener. In certain contexts, the spoken word "cup" may function in exactly the same way as either "This is a cup" or "Is this a cup?" The function of the single word "cup" in this example will

Relational Frame Theory

109

be determined by a range of possible contextual cues, such as the conversational context in which the word is uttered, the facial expression of the speaker, the vocal inflection, and so forth. From the perspective of RFT, single words, grunts, raised eyebrows, a frown, or virtually any discrete event may function as part of a relational network if the historical and current context supports the relevant relational (and thus verbal) functions.

1. Complete and Coherent Networks The concept of a relational network provides a way to approach the organization of larger language units in everyday terms, such as sentences, paragraphs, chapters, stories, and trilogies. Relational networks can be more or less complete. By "complete" we mean the degree to which the events in the network, and the network itself, serve as a context for relational activity. At the lowest level, a network is complete if terms (serving as contextual cues) set the occasion for relational activity necessary to specify a relation between all of the events in the network. This level corresponds closely to the commonsense notion of a sentence, and thus one could say that, in terms of RFT, the lowest level of a complete relational network is a sentence (a similar view can be found in Place, 1998). According to this view, a sentence does not have to make sense to be classified as a sentence, but it does have to form a complete network in the sense just described. This view easily incorporates distinctions between complete sentences, meaningful sentences, nonsense sentences, and nonsentences. For example, the statement"This cup is a . . . " would normally be defined as an incomplete sentence. According to RFT, it is also an incomplete network because the contextual cue "is a" evokes a relational response that cannot be completed (because the frame of coordination is to be applied to the cup and something that as yet remains unspecified). A point worth remembering is that meaningful networks are generally complete, but they need not be in a formal sense. The relational frame theory of language thus described can also account for nonsense sentences, which virtually by definition have never occurred. The famous nonsense sentence "Colorless green ideas sleep furiously" provides one example. In the language of RFT, this sentence is a complete relational network, because the syntactical structure and the terms themselves serve appropriate relational functions. In other words, clear hierarchical relations can be found between "colorless," "green," and "ideas" and between "furiously" and the function word "sleep." However, the relational network is nonsense because it is composed of relations that are almost never found in the relational networks that operate in the natural language community, and thus few functions are transformed through the network (except, perhaps, in the context of poetry). For instance, "colorless" and "green" would normally participate in frames of opposition or difference, but in the foregoing nonsense sentence "colorless" is an attribute of the color "green." Sentences without a contextual function are purely arbitrary or meaningless, but

110

Yvonne Barnes-Holmes et al.

if they provide both a proper relational context and fulfillment of that relational response they are complete. Compare the above-mentioned example with the nonsentence "Jockstrap, monkey, dishwasher." Although verbal events are presented in this example, neither the sentence itself nor the larger verbal context provides contextual cues that would specify the relations to be derived from the words (i.e., we do not know what the relation is between the monkey and the dishwasher). In RFT terms, this is a nonsentence because it is not a complete relational network. By adding contextual cues to specify relations among the words, however, the nonsentence may become a meaningful sentence. For example, "A Jockstrap IS A monkey IN A dishwasher" is a coherent network because the relations among the terms are specified, and thus relations among all of the elements can be derived. Similarly, the three words can become a coherent network by placing before them a sentence that describes how they are related, such as "List your three favorite possessions." From this perspective, we can also consider the completeness of larger units of language. A network can be complete in a local sense, but not in a larger sense, because previous verbal material or the general verbal context specifies that a larger relational network is being formed. This network requires certain features to be complete. Suppose a parent says to a young child: "Let me tell you a story. Once upon a time there was a king who went on a quest to find a magical ring." If the parent then says nothing more, a child with even preschool verbal abilities will quickly demand "tell me the rest of the story !" In this case, the child's history with the larger relational networks called "stories" combine with cues, such as the parent's first sentence or the first phrase of the second sentence, to establish a relational function. "There was a king who went on a quest to find a magical ring" is a complete sentence because a relational context at that level is fulfilled, but it is not a complete story because the larger relational network should specify how the quest turned out. In this particular case, it does not. C. RELATIONS AMONG RELATIONS: ANALOGIES, METAPHORS, AND STORIES

Relational frame theory can take the foregoing analysis further in tackling more complex verbal styles such as analogy, metaphor, stories, allegory, anecdote, parable, humor, and the everyday use of examples. Indeed, the analysis of analogy, metaphor, and parable is believed by some to be an important test of any theory of language and cognition (Ortony, 1993).

1. Analogy Analogical abilities are commonly used as a metric of intelligent behavior (Sternberg, 1977) and to predict academic success (e.g., in the Graduate Record Examination), and they are believed to be prominent in scientific and mathematical abilities (e.g., Polya, 1954) and problem solving more generally (e.g., Brown, 1989).

Relational Frame Theory

111

Lipkens (1992) offered a relational frame account of analogy, in which analogy was said to involve two relational networks. In this view, two networks are deemed to be analogous (i.e., related analogically) if the trained or derived relations in one network are placed in a frame of coordination with the trained or derived relations in the other network (for empirical evidence see Barnes, Hegarty, & Smeets, 1997; Lipkens, 1992). A relevant example is as follows. A person is directly trained in a given context to relate A to B in a particular way and to relate C to D in the same way. The trained relation between A and B in this instance is analogous to the trained relation between C and D. For example, after being taught that A is opposite to B and C is opposite to D, an analogous relation can be derived between the relation "A is opposite to B" and the relation "C is opposite to D." At the simplest level of analogy, we can describe a derived analogous relation between derived relations of mutual entailment. That is, given the same trained relation between A and B and between C and D, the relation of mutual entailment between B and A is analogous to the relation of mutual entailment between D and C. For example, after being taught that A is opposite to B and C is opposite to D, an analogous relation can be derived between the relation "B is opposite to A" and the relation "D is opposite to C." At a more complex level, with more than one relation in each system, derived analogous relations between derived relations of combinatorial entailment can also occur. An illustrative example is as follows. The relations of combinatorial entailment between A and C are analogous to the relations of combinatorial entailment between D and E At the same time the relations between C and A are analogous to the relations between F and D. A person, for example, may be directly trained that A is the same as B and B is the opposite of C and also that D is the same as E and E is the opposite of E This person might well derive an analogy between the relations "C is the opposite of A" and "F is the opposite of D" and vice versa (i.e., "A is the opposite of C" and the relation "D is the opposite of F"). As we have described previously (see Section II.A.1), the transformation of stimulus functions is the characteristic that gives arbitrarily applicable relational responding its psychological significance. In terms of analogy, therefore, we can also extend this notion to include a transfer of function of one relation to another relation.

2. Metaphor In this subsection, we consider the more complex process of metaphor. As noted earlier, analogy does not require nonarbitrary relations among the stimuli used (although such relations are often involved). Unlike analogy, however, metaphor must involve forms of relational control based in part on the verbally abstracted nonarbitrary properties of some events in the relational network. The following metaphor, for example, is often used in clinical practice, "Struggling with anxiety is like struggling in quicksand." This metaphor is used to enable the client to see

112

Yvonne Barnes-Holmes et al.

that struggling with anxiety (i.e., trying really hard to escape feelings of anxiety) serves only to make the anxiety worse in the same way that struggling in quicksand serves only to make you sink more quickly. The relational network suggested by this example might be denoted as follows: "anxiety" (A) is to "psychological struggle" (B) as "quicksand" (C) is to "physical struggle" (D). What distinguishes this metaphorical example from the analogical examples described previously is that the relational network involves a formal relation between two apparently very different events. This formal relation is important because it transforms the functions of the related events for the listener. Specifically, deriving a relation between anxiety and quicksand could help the listener to discriminate that struggling in either case leads to formally similar effects. The functions of "anxiety" are therefore said to be transformed, and the anxious client might derive important cause/effect relations between these events, such as "struggling with anxiety will only make my situation worse." In this way, the original anxiety problem may be altered in view of the new relational network.

3. Storytelling Relational frame theory approaches an understanding of storytelling in a manner similar to metaphor. Stories are generally less complex than metaphors primarily because stories carefully establish the relational network in one domain rather than relying as much on the existing verbal relations of the listener. Stories are built out, piece by piece, sometimes over many hundreds of pages or hours of telling. After they are built, they can readily be applied metaphorically. For example, a biblical story is used metaphorically every time someone is called a "Good Samaritan" (of course, this phrase may be used without knowledge of its biblical origin). This process difference between stories and metaphors is not absolute, because any story relies on unstated contextual and verbal features to be understood. In fact, no story is truly built ab initio, because all language requires a larger context. The story of the "Good Samaritan," for example, is much better understood when it is made clear how much the Samaritans loathed the Jews. Nevertheless, stories are usually less demanding of a listener in that the listener will be brought along, step by step. Stories are cognitively simpler than metaphors for another reason. They contain or develop many nonarbitrary formal properties within the domain of the story via derived stimulus relations and the transformation of stimulus functions that they enable. Even a very young child will attend for a long time to a good story because the development of the relational network in the story itself establishes intense, coherent, and direct stimulus functions. In a good story, the child can "imagine" the prince and the princess, or the scary dragon, or the evil witch. These functions cohere with the story. Because of this property, stories are said to be "concrete"; lacking this property, analogies and metaphors are relatively "abstract."

Relational Frame Theory

113

That is, stories have reinforcing value based merely on the properties of developing complex relational networks and transforming emotional and other psychological functions in terms of these relations. More informally, such relational responding allows the listener to experience events that would be reinforcing had they actually occurred (e.g., defeating the scary dragon and winning the love of the princess). Contacting the value of analogy and metaphor appears to require more complex and subtle forms of relational responding than that involved in following a simple or obvious story. 4. Humor A humorous story or joke involves stimulus networks which themselves serve as contextual cues for specific kinds of relations. In a "shaggy-dog" story, for example, the tale appears to go on endlessly. Although cues are provided that would lead the listener to respond to the telling as if it were a story (i.e., with a beginning, middle, and end), the story is told in excruciating detail with constant asides and irrelevant tangents, until the listener realizes that the story has no end, it is not truly about anything, and that in itself is the joke. A slightly different process is basic to many other forms of humor in which the relational networks are complete, meaningful, and coherent, but they are incongruous. One of several types of incongruity can be observed in humor in which the story appears to be congruous until the punch line. Salient cues are provided that lead the listener to predict that the network is being completed in a particular fashion. The story involves a particular topic, approached in a particular w a y - - o r so it seems until the punch line. Suddenly and unexpectedly (thus the metaphor of a "punch"), the network collapses into incoherence, only to re-form an instant later in an unusual and often ridiculous and incongruent way. It becomes obvious that the listener "should have seen it coming" (i.e., that the dominant cues should not have been relied upon so thoroughly). According to RFT, riddles and word puzzles rely on the same verbal properties as humor. In both cases, the contextual features needed to solve a riddle are subtle or nondominant. The difference is that the network achieved and the process of achieving it are more clever than incongruous. For example, "if you saw a basketball team chasing a baseball team, what time must it be?" Answer: "Five after nine." The number of players on a team is a nondominant feature, but it is made relevant by the request to supply a time, since times are usually stated as numbers, and with a few nonnumerical words such as "until" or "after." The relation between "chasing" and "after" supports the solution process further. The ability to solve riddles and word puzzles has long been taken to be a sign of learning, intelligence, or creativity. This apparent relationship makes sense if one thinks of the process of puzzle solution as having to do with both the degree of elaboration of relational networks, and the subtlety and flexibility of contextual control over relational responding.

114

Yvonne Barnes-Holmes et al.

5. Conclusions The RFT account of analogy, metaphor, stories, and humor consists of relating networks of stimulus relations. This process is a critical attribute of the RFT account of language, because it appears to explain two important hallmarks that underpin the generativity, richness, and complexity of language, as commonly described. First, the ability to relate networks of stimulus relations is extraordinarily efficient and generative. Hundreds or thousands of existing stimulus relations in one domain can be brought suddenly to bear on another domain and can generate a myriad of derived relations as a result. Second, relating relations is essential in the establishment of abstract verbal terms, precisely because what is meant must be abstracted from a larger set of phenomena. From the perspective of RFT, relating relational networks to relational networks may also lie at the heart of what is commonly referred to as "remembering." Someone who is remembering is essentially telling a story. According to RFT, elaborated networks can be retained well because the entire network can often be derived when some of the trained or derived relations have become weak. D. THINKING,PROBLEM SOLVING,AND PRAGMATICVERBALANALYSIS

1. Pragmatic Verbal Analysis: The Nonarbitrary Application of Relational Frames Truly nonarbitrary stimulus relations are hard to distinguish from verbal relations that are applied to the nonarbitrary world. An illustrative example is as follows. Two clear cups with milk are placed side by side. They are identical in shape but one is physically larger than the other. A child who is asked "which one has more?" points to the larger of the two cups. Although the pointing may look like a nonverbal event, we do not know if it is verbal or not. We would predict, for example, that a well-trained nonhuman with the relevant history with differentsized objects and the nonarbitrary stimulus relation between these particular two objects, would be able to accomplish this type of task. However, according to RFT, a given act that could be nonverbal (as would be the case with the nonhuman subject), is not necessarily so. Even in this simple example, the child may be applying a relational frame. Specifically, selecting the larger cup in this instance would be deemed to be verbal if the behavior is in a class with behaviors that are controlled in other contexts by arbitrary contextual cues. If so, we would say that the act of relating the two cups is an arbitrarily applicable relational response that is not arbitrarily applied in this instance. When relational frames are applied nonarbitrarily, the relevant history involves both arbitrary training and nonarbitrary features of the environment. This type of verbal behavior is called pragmatic verbal analysis in RFT; this phrase thus refers to a form of verbal analysis of the nonarbitrary world that involves acting upon the world verbally and having the world serve verbal functions as a result.

Relational Frame Theory

115

According to RFT, the majority of what is considered nonarbitrary in human affairs is in fact verbally involved as a result of pragmatic verbal analysis. Any physical dimension that one can speak about meaningfully and can understand is, in RFT terms, functioning verbally for these speakers and listeners. The nonarbitrary environment can, therefore, become involved in arbitrarily applicable relational responding in two ways. First, nonarbitrary features or relations may set the occasion for verbal relations, but they do not themselves enter into relational frames. For example, a very young child might point reliably to the larger of two toys when asked "which one is bigger," without the word "bigger" participating in relational frames with the relevant formal dimension or other words. The same child would, for example, be unable to answer questions such as "what is another word for bigger." The functions of the physical environment (i.e., "bigger than") that were responsible for the relational responses are not transformed, and thus cannot themselves be defined as verbal. Second, nonarbitrary features can set the occasion for relational responding of a particular kind because the features themselves are functioning verbally. In that case, the nonverbal environment actually becomes verbal to a degree, and verbal relations act upon it. Let us return to the preceding example. The larger of the two cups containing milk described now has extremely thick sides, such that only a thin strand of milk can be seen extending down the center of the cup. In other contexts the child has placed "milk" and "heavy" into a relational frame of "difference" (i.e., "milk is not heavy"). Subsequently, the child is given the choice of the two cups again and is asked "which one has more milk?" and later "which one is more heavy." If the child points to the small cup in answer to the first question and the large cup in answer to the second, we may begin to conclude that the child's original performance was regulated by verbally abstracted nonarbitrary features of the choice situation.

2. Verbal Abstraction of Stimuli From the perspective of RFT, the freedom from co-occurring stimulus properties (e.g., nonarbitrary features of the environment) increases enormously because stimulus attributes participate in networks of relational frames and in particular hierarchical relations such as "attribute of" or "part of." This outcome in turn allows verbally abstracted attributes to enter into other verbal relations or to serve as verbal relational cues. The importance of hierarchical relations to abstraction is emphasized explicitly in a range of early preschool activities. A Sesame Street task, for example, presents to a child four objects, three of which are identical, and asks the child "one of these things does not belong; can you tell which thing is not like the others?" Tasks such as these train identity matching and exclusion, and they support the development of frames of coordination and of distinction. A more complex task can be constructed in which all four objects are different, but three share a common attribute. This kind of training builds on the earlier training

116

Yvonne Barnes-Holmes et al.

by adding a hierarchical relation and using it as a relational cue for the derivation of frames of coordination among the three similar objects. From the perspective of RFT, this procedure amounts to a kind of relational abstraction training and should greatly enhance the impact of verbal relations on the ability to abstract features of the environment. As verbal humans interact with their environment, events gain an ever increasing number of verbal functions of this kind. Specifically, pragmatic verbal analysis supports more complex forms of interaction with the natural environment that are "purposive" or "intentional" in the verbal sense of these terms (see section II.A.3). In other words, patterns of verbal behavior are sequenced to produce certain consequences as verbally conceived. The distinction between different forms of pragmatic verbal analysis comes from the precise use to which this verbal analytic process is put. Thinking and problem solving serve as useful examples of these verbal processes.

3. Thinking Thinking involves the purposeful action of pragmatic verbal analysis and thus is both situated in the practical environment and controlled by direct or verbally constructed consequences. The term "thinking" tends to be used when the specific performances involved are not already established. Thinking is reflective in the sense that verbal responses are often made to verbal and nonverbal responses in the process. That is, steps are taken in the verbal analysis of a situation that serve as the occasion for additional verbal analyses. The term is more likely to be invoked (though not always) when the verbal activities involved are private, that is, observable only by the agent that performs them. Relational frame theory thus defines thinking in a verbal sense as a reflective behavioral sequence, often private, of pragmatic verbal analysis that transforms the functions of the environment to lead to novel, productive acts. In this definition, thinking is said to transform the environment, but this transformation of stimulus functions is not random. It is an action controlled by its consequences, both verbal and nonverbal, in a situation in which relatively novel productions are required. In essence, this definition treats thinking as a particular kind of verbal problem solving, a related topic to which we now turn.

4. Problem Solving Relational frame theory offers an account of problem solving similar to that provided for thinking. In behavioral terms, we can think of having a problem as lacking the behavior that occasions available reinforcement (Skinner, 1953) or as the presentation of situations in which behavioral obstacles are to be avoided or removed (see Reese, 1994, for an excellent analysis of various traditional models of problem solving). From a relational frame perspective, verbal problem solving can be defined as framing events relationally under the antecedent and consequential

Relational Frame Theory

117

control of an apparent absence of effective actions. Stated another way, the absence of effective actions is an antecedent for relational activity that is itself oriented toward the establishment of such actions. Problem solving so defined is usually (but not always) a form of pragmatic verbal analysis. Some problems are almost entirely verbal, for example, solving an abstract mathematical or logical problem. Problems of this nature often begin and remain verbal and arbitrary. Most problems, however, involve the stimulus functions of the nonarbitrary environment. In these cases, verbal problem solving can be said to be pragmatic verbal analysis that changes the behavioral functions of the environment under the antecedent and consequential control of an apparent absence of effective action. According to this definition, all forms of problem solving involve detection of a problem state, namely, the absence of effective action in a given situation. However, problem-solving strategies differ in the degree to which the problem solution is verbally discriminated. For example, strategic problem solving is said to occur when solutions have been identified verbally. In other words, the desired goal or purpose has been placed in a relational frame and can be compared to various verbal networks that specify the current situation and the steps that need to be taken to reach the verbally constructed goal. This activity is inherently metaphorical in the sense described in Section II.C.2, because entire sets of derived stimulus relations are being related to one another on the basis of physical properties, dimensions, or relations and their transformation. Valuative problem solving, in contrast, applies when the lack of effective action is verbally accessible, but not to a degree that would constitute effective action. This type of problem solving involves the use of relational frames to contact possible outcomes to select from among them. In major life decisions, such as employment and marriage, the problem may be more about the selection among possible consequences as it is the means of reaching consequences that are selected. Faced with problems of this kind, a list of pros and cons may be constructed, formally or informally, to try to abstract features of the situation that might be contacted. In this case the "contact" is through the transformation of stimulus functions through a network of derived stimulus relations.

5. The Utility of Pragmatic Verbal Analysis Pragmatic verbal analysis is an important part of verbal behavior because it is useful for many reasons. First, it enables us to delay and reduce the possibility of impulsive actions, because of the intrusion of functions that are transformed and the increased predictability of temporally extended consequences. Second, it enables us to deal with problems before they are experienced and thus to work out verbal solutions. Third, it allows us to develop overarching or general verbal strategies that are particularly useful in instances of ambiguity. Fourth, pragmatic verbal analysis can be applied to one's own behavior. In the case of problem-solving

118

Yvonne Barnes-Holmes et al.

strategies, for example, responding to one's responses may contribute to evaluating the success or failure of behavioral efforts. Fifth, by far the most important adaptive function of pragmatic verbal analysis is the construction of verbal rules that allow for stimulus and response functions to be augmented or diminished, for possible outcomes to be detected, and for lengthy behavioral sequences to be performed with regard to those possible consequences. This issue is discussed in the next section. E. UNDERSTANDING AND VERBAL REGULATION

From the perspective of relational frame theory, pragmatic verbal analysis and relational networks facilitate verbal understanding and verbal regulation, in part by the construction of verbal rules. Relational frame theory considers an analysis of rules and rule-governed behavior as important for establishing a technical definition of understanding and verbal control.

1. Understanding Verbal Events According to RFT, words and other events (e.g., facial expressions) function as verbal stimuli because they participate in relational frames. Verbally governed behavior is simply behavior governed by verbal stimuli. A listener's "understanding" is synonymous with the set of derived and specified relations that results for the listener from verbal stimuli. The important point is that the functional source of stimulus control is the feature that defines the listener's performance as verbally governed, rather than the form of the behavior or the form or social source of its antecedent. Hayes and Hayes (1989) simply defined rule-governed behavior as "behavior controlled by antecedent verbal stimuli" (see also Zettle & Hayes, 1982). Verbal antecedents that are complex have the commonsense connotations of the term "rule" and for that reason are more likely to be called "rules" than "stimuli." Hayes, Gifford, and Hayes (1998) provide the following example. A person says, "I'm going on vacation in two weeks and will be gone for a month. If you water and mow my lawn each week that I am gone, the following month I will pay you $100." This is a thoroughly specified contingency. It alters the functions of calendar time, the grass, and the implements needed to water and mow the lawn. It specifies all the major elements of a contingency: a temporal antecedent, a topographical form and the context within which it should occur, and the nature and delay of a consequence. The contingencies that are specified could not be effective through direct training, in part because greatly delayed consequences are simply not effective in the absence of verbal rules. According to RFT, understanding this rule requires first the examination of the specific relational frames and the cues that occasioned them, and then it requires the functions of the events that are transformed in terms of these relations and

Relational Frame Theory

119

the cues that occasioned these transformations. Several core relational frames seem necessary for understanding this rule. Coordination exists between some of the terms (e.g., grass) and classes of physical events. Before-after relational frames, made more specific by numerical temporal terms, are used to specify a temporal antecedent and a consequence (e.g., begin mowing after 2 weeks). Ifthen relational frames are used to specify the contingent relations (e.g., if you mow and water weekly for 4 weeks, then you will receive $100). Terms like "mow" alter the behavioral functions of the grass, and the transformation of stimulus functions provides these actions and contexts with some of the features of the specified consequence (e.g., approach). Clearly, many rules are more complex than this example. However, in the language of RFT, complex rules are not different at the level of processmthey are simply examples of more complex relational networks and transformations of function. As the listener follows a rule, the coordination between the original relational network as understood and the relations sustained among the specified events when the rule is followed provide an ongoing source of control over behavior. Stated loosely, the person following a rule can detect that the rule is being followed (or not), because what is being done corresponds (or not) with the rule. This monitoring does not require that the rule itself be restated, although it often is restated. Rather, the elements specified in the rule are actualized in a particular manner, and these events themselves form a relational network that corresponds (or not) to the original relational network. If the person exposed to the lawn-mowing rule in the example described above mows the lawn two and a half weeks after the rule was provided, he or she, as the listener, will be able to detect that the rule is being followed, based on a verbal comparison between what was done and what was asked. Whenever a frame of coordination between two such networks serves as a source of control over behavior, in RFT terms the behavior can be defined as rule-governed. Most complex examples of verbal control, in which intricate verbal sequences control the behavior of listeners across large temporal gaps, seem to have this feature. However, when verbal antecedents are much simpler, and when correspondence between the rule and relevant behavior is not part of the source of behavioral control, RFT would adopt the simple concept of "verbal stimulus" rather than the concept of a "rule." In RFT, a clear distinction is made between nonverbal and verbal regulation. Rule-governed behavior is a subset of verbal regulation. The term becomes more likely to be used when the verbal antecedent is a relational network or a comparison of such networks, and especially when comparison between a verbal antecedent and the verbal construction of ongoing events is part of the source of control over behavioral regulation. Rule governance is also more likely to be used when the nonarbitrary features of the environment are abstracted and transformed and when the verbal network is generally applicable.

120

Yvonne Barnes-Holmes et al.

2. Verbal Regulation and Rule Following Although the newly acquired verbal functions of the previously nonverbal environment allow a rule to control behavior in contexts that are sometimes radically different from those in which the rule was presented, actual verbal regulation of overt behavior does not always occur. In other words, for a number of reasons, rules may not be followed even when they are understood (in what follows, we assume that any consequence specified in a rule functions as either a reinforcer or punisher for the listener). Relational frame theory approaches this issue by analyzing the contextual features and content of an instance of rule following. First, a particular instance of not following a rule may be due to insufficient control by nonverbal contingencies (e.g., the behavior proscribed by the rule may not be present in the behavioral repertoire). Second, the extent of rule following may depend on the credibility of the speaker, one of whom (e.g., a policeman) may be more likely to produce verbal formulas that are predictive than others. Credibility can be acquired directly or verbally, and generalization of credibility from one speaker to another can be through formal or relational means. Speakers who are said to be intelligent, for example, will likely be more effective in producing rule following because these very verbal relations are seemingly predictive of rules that are accurate. Third, rule following also depends in part on the speaker's authority and ability to mediate reinforcement. For compliance in particular, the capacity to mediate consequences to the listener will predict rule following (e.g., young children will often follow rules provided by parents, but not by babysitters). Fourth, the plausibility of the message will also determine whether or not the message is followed (cf., Hovland, Lumsdaine, & Sheffield, 1949). Plausibility is undermined by the emergence of relations of distinction and opposition between the stated rule and other current and relevant relational networks actualized by the terms in the rule. Nowadays, an average listener will have been exposed to many previous rules of the form, "Cigarette smoking causes cancer" from highly credible sources (e.g., national medical leaders). Because "smoke" and "cigarettes" participate in frames of coordination or hierarchical class membership with words such as "cancer," "heart disease," and so forth, the frame of coordination specified in the rule "Smoke more cigarettes and you will live a longer and healthier life" cannot occur without a radical change in the existing relational network. Fifty years ago the case was different: stimulus networks were being elaborated, not contradicted. Given the protracted history of exposure to the earlier rules, they are more likely to be followed than the implausible rule.

3. Verbal Communication Thus far we have tended to focus on the verbal behavior of the listener, but a complete analysis of understanding and verbal regulation requires that these be considered in the wider context of verbal communication (i.e., communication that involves relational frames). More specifically, we need to address the relationship

Relational Frame Theory

121

between speakers and listeners. Speaking with meaning and listening with understanding both involve arbitrarily applicable derived relational responding. Verbal communication from the speaker's point of view becomes verbally purposive (e.g., involving if-then relational frames) by being used to produce verbally known effects in a listener (i.e., the effects participate in relational frames). A salesperson, for instance, might adjust his or her sales pitch depending on the relational frames in which the listener participates (e.g., a prospective buyer who participates in a frame of coordination with "friendly"). In effect, the distinctions among types of verbal communication have to do with the specific purposes, contexts, and listeners involved. A parent teaching a child to talk will behave differently from a teacher teaching a child to do math or a lover trying to establish an intimate relationship. The type of communication used will differ profoundly if the listeners participate in frames of coordination or hierarchy with terms such as friendly, intelligent, psychotic, mean, stupid, and so on. Of course, such verbal repertoires need not necessarily be coordinated, and for young children they are almost certainly not (i.e., children have to learn how to manipulate listeners purposively by responding in accordance with relational frames). E

SELF AND SELF-DIRECTED RULES

To provide a relational frame analysis of self-directed rules, we need first to address the concept of self.

1. The Concept of Self Derived relational responding makes verbal self-knowledge important and useful (Hayes & Gifford, 1997; Hayes & Wilson, 1993). Verbal reports of one's own behavior or of the contingencies controlling it can alter the functions of both. The temporal delays experienced by a young child provide an example. Selfinstructions can reduce the effect of temporal delays as a young child develops (Bentall & Lowe, 1987). A child, for example, who is able to place "waiting" into a conditional (if-then) relational frame with "getting more" will probably, via the frame of opposition, place "not waiting" in an if-then frame with "getting less." Getting a reward after not waiting may now be less reinforcing because by derivation of a relation, the consequence received after not waiting will be in a comparative relational frame with the verbally known alternative that would have been received had the child waited. If put into verbal form, this process could be expressed as something like "this is no goodmthis is less than I would have gotten." Of course, the child does not necessarily have to say such a thing for the behavior of waiting to occur, although about the time that relational responding is sufficiently strong to have these effects, statements of this kind will also be likely. From the perspective of RFT, the relational (cognitive) processes that allow children to regulate their own impulsive responding in this way are:

122

Yvonne B a r n e s - H o l m e s et al.

a. Coordination and temporal/causal relations, at a minimum, b. the transformation of stimulus functions in terms of these relations which allows uncontacted "consequences" to have stimulus functions, c. a sufficient history of rule following that has previously led to successful contact with verbally constructed consequences, d. comparative relational responding that can lead to the next item, that is, e. the reduction in the relative value of immediate, small, contacted consequences over larger, delayed, verbally contacted ones via a transformation of stimulus functions, and f. rule following based on this verbal comparison. From the perspective of RFT, only verbal self-knowledge of this kind will enable the individual to predict and control his or her own behavior because it leads to self-rule following.The downside of this same relational process is that derived relational responding also makes verbal self-knowledge emotional and difficult (Hayes & Gifford, 1997; Hayes & Wilson, 1993). For example, persons who have experienced a traumatic event seem to reexperience the aversiveness of the event in reporting it. This effect is not surprising because the verbal report can carry with it some of the functions of the original trauma. Without verbal relations, this effect would not occur.

2. Perspective-Taking According to RFT, perspective-taking frames appear to be essential in the verbal construction of self. Three of these frames are of particular importance: the frames of I and You, Here and There, and Now and Then (Barnes & Roche, 1997; Hayes, 1984). As argued in section II.A.2, these frames are unlike most of the other relational frames in that they do not appear to have formal or nonarbitrary counterparts. The physical properties involved in these frames are abstracted only in the context of relational frames. Abstraction of an individual's perspective on the world, and that of others, requires a combination of a sufficiently well-developed relational repertoire and an extensive history of multiple exemplars that take advantage of that repertoire. Each time someone is asked a question such as, "How are you?" the physical environment and its formal properties will likely be different. The only constant across all such questions is the relational properties of I versus You, Here versus There, and Now versus Then. These properties appear to be abstracted through learning to talk about one's own perspective in relation to other perspectives. For example, I is always from this perspective here, not from someone else's perspective there. The establishment of these three relational frames gives rise to responding to a range of relational networks as follows: I-HERE-NOW I-THERE-NOW

YOU-HERE-NOW YOU-THERE-NOW

I-HERE-THEN I-THERE-THEN

YOU-HERE-THEN YOU-THERE-THEN

Relational Frame Theory

123

In everyday discourse, many phrases may be constructed from these eight relational networks; "I am here now, but you were here then" or "You were there then, but I'm here now" or "You and I are both here now, but I was here then." In actual conversation, of course, the phrases would often include or substitute words coordinated with particular individuals, places, and times. The actual words that function as relational cues are not important (as is the case for all relational framing), the generalized relational activity (similar to what psycholinguists might speak of as "deep structure") is what matters. Thus, the English speaking community probably normally establishes perspective-taking relational functions in words such I, YOU, HERE, THERE, NOW, and THEN.

3. The Three Selves In combination with an elaborated relational repertoire, RFT would argue that perspective-taking can establish three types of self: (a) self as the content of verbal relations, (b) self as an ongoing process of verbal relations, and (c) self as the context of verbal relations (Hayes, 1995). Stated another way, verbal relations can lead to a conceptualized self, a knowing self, and a transcendent or conscious self. a. Self as ContentmThe Conceptualized Self Perspective-taking frames allow each of us to derive relations between our ongoing unified stream of behavior and a panoply of categorical concepts. We can evaluate, interpret, predict, explain, rationalize, and otherwise interact verbally with our own behavior. As soon as we can interact with ourselves verbally (in terms of the frames of I-YOU, HERE-THERE, and NOW-THEN), we begin to form a "conceptualized self." In constructing a conceptualized self, most of us try verbally to make sense of ourselves, and to put our own histories and tendencies into a coherent relational network. In short, self as content refers to the descriptive and evaluative relational networks that we construct HERE and NOW when talking about I or ME (or the behaviors of I or ME) located THERE and THEN. A middle-aged man might say (HERE and NOW), for example, that he (I) is a really bad son because he did not visit his mother before she died (THERE and THEN), but his sister (YOU or not I) is a good daughter because she visited regularly. b. Self as Verbal ProcessmThe Knowing Self To understand the second type of self, we need to focus on the training that goes on in forming frames of coordination between the fuzzy set of bodily sensations, behavioral predispositions, thoughts, and environmental situations and on a name for an "emotion." Furthermore, we need to address the training involved in learning to talk about these frames of coordination in terms of the perspective-taking frames (e.g., "Oh no, I have that pounding feeling again in my chest [HERE and N O W ] m I ' m going to panic" [THERE and THEN]). According to RFT, humans will have emotions that nonverbal organisms do not have, because emotions are in large part

124

Yvonne Barnes-Holmes et al.

verbally constructed through complex relational networks (e.g., anxiety is not simply fear). From this perspective, therefore, human emotion generally can be viewed as a complex set of events related together through perhaps complex relational networks.

c. Self as ContextmThe Transcendent Self The third aspect of self is perhaps the most difficult to explain. As indicated earlier, once the perspective-taking frames are established in the behavior of a particular person they become an inherent property of most verbal events for that person. Whenever the person talks to someone else, the talking will be from the perspective of I located HERE and NOW about events that occurred THERE and THEN. This inherent property of verbal events serves an important social and communicative function. If I ask you what you did yesterday, for example, I have to be able to trust that the report that is made is from a perspective or point of view that is consistent and predictable. In other words, the speaker is always I (not YOU), located HERE and NOW, and what is spoken about is always located THERE and THEN. Self as context refers to the I who is always doing the discriminating (located HERE and NOW). If I ask many, many questions of a person, the only thing that will be consistent is not the content of the answer, but the context from which the answer occurs. "I, HERE and NOW" is the self that is left behind when all the content differences are subtracted out. For this reason, this sense of perspective or self as a place from which things occur does not change once it emerges (around the age of 3 years). A noteworthy point is that in describing these three senses of the term "self," we have focused only on the psychological "self" from the point of view of the person of interest. Clearly, the term self has many other senses, including self as a biological organism and self as an integrated behavioral repertoire, to name but two, although these selves do not appear to be psychological selves. 4. Self-Directed Rules Having analyzed the self, we are now ready to deal with the issue of self-rules. In section II.D.5, we argued that applying pragmatic verbal analysis to one's own behavior is useful for permitting greater self-control and for allowing an individual to evaluate the success or failure of behavioral efforts. The most important adaptive function of pragmatic verbal analysis, however, is the construction of verbal rules that allow functions to be augmented or diminished, possible outcomes to be detected, and lengthy behavioral sequences to be performed with regard to those possible consequences. From the perspective of RFT, this is the point at which self-directed rules enter the picture.

5. Self-Directed Rules and Problem Solving According to RFT, the average verbally sophisticated human produces vast numbers of self-directed rules during each day, some simple and perhaps facile, such

Relational Frame Theory

125

as "I must go home and feed the cat," and some more complex and fundamental, such as "I must do something meaningful with my life, and so I must make the right decision now about which career I should choose to follow, but in doing so I should be careful not to sacrifice my personal interests and intimate relationships to that career." The cat-feeding self-rule is relatively simple. The problem is purely strategic because the outcome is specified, and presuming that all of the terms in the network possess relatively precise behavioral functions, the problem devolves into a simple matter of when and how to get home, whether cat food is available there, and so on. Such limited strategic problems involve self-rules, but the role of the self per se is limited. A spouse could just as readily specify the steps to be taken: "Take the car home on your lunch break and feed the cat, using the cat food in the refrigerator." The issues of self enter into the self-rule in this case primarily in the form of the conditions under which the self-rule is generated and of the larger patterns of behavior in which the specific problem participates. For example, the person may have noticed a feeling of guilt that the cat was home hungry; perhaps the cat was purchased to help the children learn to care for others and not feeding the cat properly would present a bad model to the children. Such neglect might conflict with the person's self-beliefs, such as "I'm a kind person." Such factors would involve the self in several ways. At the level of process, becoming verbally aware of thoughts about the cat, of feelings of guilt about the cat, or of values regarding child rearing is part of the ongoing process of self-knowledge. Actually going home to feed the cat may involve a conceptualized self ("I am never cruel to animals"). If the present analysis is correct, both of these forms of self-knowledge are dependent upon the contextual clarity that this self is known from a given perspective or point of view. The career choice rule is more complex. From the perspective of RFT, it involves the frames of IF-THEN, I-YOU, HERE-THERE, and NOW-THEN: IF right decision (HERE and NOW), THEN I will be fulfilled in my career and personal life (THERE and THEN). In this rule, however, some of the terms in the relational network do not possess precisely controlled behavioral functions and the outcome of the problem-solving activity is not known--for example, what exactly is the "right decision," and what would the speaker define as "something meaningful"? As a result, this rule generates a type of valuative problem, with strategic problems to be solved once an outcome goal is clearer. The problem that emerges from the complex rule is valuative, because additional relational framing is needed to contact possible outcomes, so as to select among them. For example, a list of pros and cons may be constructed to try to abstract features of the situation that might be contacted (e.g., "in thinking about my future career, being a doctor would be well paid, but the hours are often long and unsociable"). Iterative and metaphorical processes may also be employed to amplify the behavioral effects of different verbally constructed futures (e.g., "if

126

Yvonne B a r n e s - H o l m e s et al.

being a doctor was an animal, what animal would it be?"). No matter how the person attempts to solve this valuative problem, the difficulties involved may be better understood in terms of the three selves outlined in Section II.E3. If the person simply chooses a career based largely on self as content, little or no contact is being made verbally with many of the important consequences of that decision. The person in the earlier example may decide to train to be a doctor because she or he thinks "doctors are good people and so becoming a doctor will make me a good person." In this case, the verbal construction of self (as content) as a good person dominates the decision, rather than the verbally constructed future of what being a doctor would actually be like. As a result, the person may well find later that he or she does not like being a doctor, thus having wasted many years training. A decision based solely on self as process might also be problematic. In this case, the person might verbally construct a career as a doctor and decide that "I feel like I would enjoy it." If the decision is made on this basis alone, the person may start training to be a doctor and then give up "impulsively" if at some point being a doctor no longer feels good. In short, when a person makes a "good decision," all three types of self seem likely to play a role. Self as context broadens the scope of the stimulus control, so that neither self as content nor self as process obtains absolute control over the final decision. In effect, self as context provides a psychological space in which the person can contact self as content (e.g., I have good eyesight and a steady hand, so I might make a good surgeon), and self as process (e.g., I really enjoy reading medical textbooks, so I might enjoy being a doctor). In this sense, a more balanced decision can be made in which a broader range of relevant issues are factored into the final choice. 6. Verbal Communication Revisited: The Verbal Other In early childhood, as verbal relations are just forming, the action of the listener as it impacts on that of the speaker is not verbally known from the point of view of the speaker. Over time, however, as verbal behavior develops, the listener as verbally constructed by the speaker provides relational cues and serves relational functions for the speaker, along with other contextual features of the communicative episode (e.g., the purpose of the communication). The actual behavior of the listener is also verbally construed, and in extended interactions these verbal relations enter into the ongoing verbal construction of the listener. All these verbal actions participate in the regulation of the behavior of the speaker. Perspective-taking leads to both the development of the self and an elaboration of the verbal other. Perspective-taking can establish three types of verbal other: (a) other as verbal relations about the stable content of the other, (b) other as verbal relations about the ongoing process of the other, and (c) other as verbal relations about the context of verbal relations in the other. Stated another way, verbal relations can lead to a conceptualized other, a knowing other, and a transcendent or conscious other.

Relational Frame Theory

127

The conceptualized other is the normal verbal construction of the listener. In most verbal interactions, the conceptualized other serves, to some degree, as a relational cue for the speaker's behavior. The knowing other is more fluid because it is based on a moment to moment construction of reactions of the other. This sense of the other emerges in many conversations, especially with friends or others who are willing to share their reactions openly, or with those who one knows well enough to "read" their expressions and gestures. This fluidity is relevant even to public speakers, who are said to be able to "read" their audiences. A sense of the transcendent other is relatively uncommon, occurring most often in religious, intimate, or therapeutic relations. This sense occurs when the speaker is psychologically connected to the listener as a purely conscious person. In this aspect, the speaker and listener are one, because "HERE and NOW" is imputed to be a singular event (i.e., one cannot be HERE and NOW, simultaneously, at different times and places). Perhaps for this reason, the sense of self as context is associated with a sense of the transcendent othermthe two go hand in hand.

7. Summary In Sections II.A-F, we outlined the defining features of relational frame theory at the level of process. Globally, RFT stands on three pillars: relational frames, relational networks, and abstraction from and transformation of the nonarbitrary environment. Each of these pillars is implicit in the others and all are developing simultaneously, but what is barely seen at one level becomes dominant at another. Archetypal examples of these three pillars are naming, storytelling, and problem solving, respectively. In the rest of this article, we will attempt to explore some of the implications of relational frame theory for development and education (for its implications for other areas, including social behavior, psychopathology, and spirituality, see Hayes et al., 2001). Although substantial empirical evidence supports the major points made in the first part of this article, the same cannot be said for the next part, which is more interpretive and speculative. As a pragmatic theory, if RFT cannot lead to interesting and useful outcomes in important domains of human concern obviously relevant to human language and cognition, then it is functionally false. Nevertheless, the specific ways that relational frame theory might be applied vary and must be worked out over time. Thus, our aim here is simply to show that it is plausible that RFT might lead to new things in important domains and to suggest possible research avenues to explore. G. PSYCHOLOGICAL DEVELOPMENT

From the perspective of RFT, relational frames enter into virtually all developmental changes and become a defining feature of complex human behavior and psychological maturity. In the current section, we present RFT analyses of

128

Yvonne B a r n e s - H o l m e s et al.

the broad categories normally found in the traditional developmental literature: cognitive, language and emotional development.

1. Cognitive Development: Intelligence According to RFT, a small number of psychological processes are sufficient to yield the full gamut of cognitive skills (deductive reasoning, inductive reasoning, analogy, etc.). Across various intelligence tests, certain main verbal factors appear to be relatively consistent, including vocabulary, reading comprehension, the use of analogies, and verbal problem solving. In the language of RFT, the traditionally distinct verbal and numerical intelligences are both verbal, although they are distinct in several meaningful ways. We will consider each of these types of intelligence in turn. a. Verbal Intelligence. Vocabulary is the primary factor in verbal intelligence. According to RFT, vocabulary is the context in which relational frames originally emerge and from which elaborated networks of verbal content develop, incorporating every relational frame in common use. Thus, persons with a highly elaborated vocabulary will tend also to have highly elaborated relational repertoires. Consequently, a task such as learning to spell is far less relationally rich than learning word meanings; not surprisingly, therefore, spelling performance correlates less with overall levels of intellectual behavior than does vocabulary, even though both tasks involve verbal material. Other more complex aspects of verbal intelligence, such as reading comprehension and the use of analogies and metaphors, require that an individual relate relational networks with flexibility and relative ease. Specifically, these aspects require the learner to elaborate entire networks of stimulus relations quickly and to bring them under increasingly subtle forms of contextual control, to transform stimulus functions through entire networks, and to abstract features of the natural environment that will support and sustain relational responding. Although these relational performances are largely abstract, they are dependent on a great deal of verbal content, and a natural developmental transition is based on acquisition of that content. All these relational abilities come together in verbal problem solving, thereby emphasizing the verbal abstraction of features in the environment and the transformation of the functions of the environment. According to RFT, verbal problem solving should be strongly correlated with other forms of relational responding, and indeed that correlation is typically obtained. From the perspective of RFT, therefore, the "global g" that underlies many intelligence tests may not be some innate biological readiness, but rather it may be the core acquisition of relational responding. b. Numerical Intelligence. Number-series problems are a common component of tests of numerical intelligence (e.g., "1,7, 13, 19, 25 . . . . What is the next number

Relational Frame Theory

129

in this series?"). According to RFT, a history of arbitrarily applicable relational responding is necessary before this type of problem can be solved correctly. The proper relational performance is acquired by presenting a child multiple exemplars of given problem types (e.g., with solutions based on other relations, such as "plus 7" instead of "plus 6" as in the answer to the example). In mathematics, the range of quantitative comparative relations is large, but the set of core relational skills is not. Mathematical relations are highly precise and focused, and the contextual control over relational performances is highly specified; thus mathematical relations depend to a relatively limited extent on verbal content. Higher mathematics, however, has more of the features of abstract verbal skills because the contextual cues controlling the numerical relational response become more subtle and the range of specific responses becomes more varied. Nevertheless, higher mathematical skills differ from content-dominated verbal skills, because the relational frames involved are essentially "empty" and almost completely arbitrary. The quality of arbitrariness at a higher level of analysis forms a barrier to successful acquisition of mathematical relations for most children, but it removes barriers of another kind. Once mathematical relations come under abstract and arbitrary contextual control, amplification and elaboration of these skills is not dependent on significant verbal content. This relative freedom from verbal knowledge may explain why mathematics is one of those few areas (along with music, logic, or art) in which child prodigies periodically appear. These areas require highly developed abstract relational abilities, not content knowledge about real-world domains. Child prodigies in history, chemistry, or psychology, for example, do not appear because the importance of the arduously acquired content of relational networks in these areas is too great.

2. Language Development Relational frame theory analyses can be used to account for a number of specific linguistic features that characterize language development. We will briefly consider each of these in turn. a. Production of Novel Utterances. Children from a young age show productivity in their language acquisition, in terms of producing novel utterances and dealing effectively with novel words (e.g., Berko, 1958). According to RFT, some element of relational framing is involved in the production of novel nonsense words, facilitating the production of large numbers of derived stimulus relations, which have not been established explicitly. b. Speech Errors. Language development is characterized by certain types of speech errors, including the use of incorrect prepositions in conjunction with verbs--"can I have any reading behind dinner" (example from Pinker, 1990). Relational frame theory would account for this phenomenon by suggesting that children develop somewhat loosely constrained relational frames in the early stages of

130

Yvonne B a r n e s - H o l m e s et al.

language development. For instance, prepositions denoting temporal events (e.g., before or after) and spatial events (e.g., in front of or behind) are of the same family of comparative relations and may initially participate in a more general relational frame. Due to the loosely constrained nature of these relational frames, errors of the type cited here are likely to occur. As language skills are shaped up through ongoing interactions with the verbal community, these relational frames become increasingly refined so that prepositions denoting temporal events, for example, such as "before" and "after" come to participate in more specific temporal relational frames of comparison, and in turn enter into relational frames of difference with prepositions denoting spatial comparatives, such as "behind" and "in front of" (see Barnes & Roche, 1997). Thus, the relevant relational frame will be brought to bear by the context of a particular sentence. This process of contextual refinement of relational frames into elaborate networks of relations will result in the gradual elimination of so-called "conceptual errors" in sentence construction. Furthermore, once relational frames are acquired, two of the main consequences that maintain them are sense making and coherence. From the perspective of RFT, therefore, children do not require consistent feedback on grammatically incorrect sentences for grammatically correct adult speech to develop. c. Comprehension versus Production. Comprehension precedes production at all stages of language acquisition. Many words (e.g., "no") are understood well before they are produced, and as adults we understand many more words than we use. According to RFT, relational framing is as readily established for a listener as it is for a speaker. In one example, a child sees Daddy and is told "that's Daddy." At a subsequent point, the child's orienting toward Daddy upon hearing the name may instantiate mutual entailment and could be rewarded in the same way as if the child had said the word "Daddy." If a child is told at a later point "This is an orange, and an orange is a type of fruit," he or she may now be able to get oranges when asked to "go get some fruit." As these relational repertoires of coordination and hierarchy begin to combine, both kinds of performances help establish a relational response, and responses in one domain will generalize to the other. Furthermore, the acquisition of meaning for hearing single instances is not surprising once relational frames have been acquired. d. U-Shaped Development. One of the key features observed in the development of grammar in young children is its U-shaped development (see Pinker, 1999). This is a characteristic of certain aspects of natural language acquisition, whereby initially good performance is replaced by incorrect performance, and then final good performance is achieved. The acquisition of the irregular past tense of the verb "eat" provides an example. Across time many children use the past-tense forms "he ate it"m"he eated it"m"he ate it" in that order. According to RFT, grammatical errors come in part from overgeneralization and inappropriate contextual

Relational Frame Theory

131

control over relational frames and the transformation of functions through them. High-frequency verbs, which are undoubtedly learned before low-frequency verbs, tend to have irregular past-tense forms (go-went, see-saw, eat-ate, etc.), and therefore the frame of coordination for each high-frequency verb is applied to a specific past-tense verb. Later, the child learns the regular form (-ed) but due to inappropriate contextual control this regular form is also coordinated with the irregular verbs. This inappropriate contextual control oversimplifies conventional (grammatical) relational networks, but eventually the appropriate forms of contextual control are established. For example, the frames of coordination within which the verbs participate (i.e., the verb meanings) come to function as contextual cues. Thus the frame of coordination between some verbs and irregular past-tense forms are reestablished, whereas all other verbs continue to participate in a frame of coordination with the regular past-tense form. Thus, errors may not occur, then begin to occur, and then fade away because relational processes may gradually emerge which make errors possible, but then fade as these processes come under better refinement and control. In short, U-shaped development is observed as the child's speech shifts from being based on minimal forms of relational framing, to relatively limited forms of relational framing, and then finally to complex forms of framing activity.

3. Emotional Development For RFT, healthy emotional development involves learning relational networks that contain emotional terms, and learning to respond in accordance with them in ways that overlap significantly with similar networks operating in the wider verbal community. A relational frame analysis of understanding the emotions of self and others relies heavily upon the concept of the three selves described in section II.F.3. a. Understanding the Emotions of the Self According to RFT, as limited forms of relational framing are established in the behavioral repertoire of the young child, the bidirectional transfer of functions between self as content and self as process begin to emerge. An example of a transfer of functions from content to process might occur when a young child is told that a family relative has died and that the child should be "sad." As a result, the child might actually experience some of the psychological properties of sadness that otherwise would not have occured without the content statement, "When a relative dies I should be sad." According to RFT, emotional self-regulation involves the application of pragmatic verbal analysis to emotional states. A relatively simple relational network that a child might construct as a means of dealing with fear of the dark provides an example, "If I am afraid of the dark, then I'll close my eyes and pretend I'm at the seaside." According to RFT, particular classes of pragmatic verbal analysis will either strengthen or weaken depending upon how successful they are in producing appropriate forms of emotional control (e.g., reducing fear when alone in

132

Yvonne Barnes-Holmes et al.

the dark). In addition to regulating their own emotional states, children also have to learn to regulate displays of their emotions (Miller & Sperry, 1987). Relational frame theory would define emotional display rules as a class of self-directed rule control. This rule control will have been established by rules provided by significant others with well-worn phrases such as "Big boys don't cry." Obviously, the ability to control emotional expression will likely increase with the development of more complex forms of relational framing activity. This advanced form of relational framing with respect to emotions brings us to what some researchers have described as emotional intelligence. b. Emotional Intelligence. According to RFT, emotional intelligence begins with simple elicited imitation, such as a child crying when she or he sees another child crying (i.e., global empathy; see Hoffman, 1982). At a more complex level, an increasing number of emotional terms coordinate with emotional states and increased development in the perspective-taking frames (i.e., egocentric empathy and empathy for another's feelings). At a more complex level still, the child is capable of complex forms of relational framing in which she or he can create a complex relational network (i.e., a mininarrative) concerning the events in another's life and then coordinate that relational network with an emotional term or other network (i.e., empathy for another's life condition). H. EDUCATION Hayes (1994) argued that overarching relational skills can be taught, and that the subsequent improvement in relational responding should lead to improved abilities in areas of cognition such as language and intelligence. Relational frame theory explicitly emphasizes the role of multiple-exemplar training in the teaching of relational skills. Examples are given in the following subsections. As will be seen, according to RFT, identifying the core behavioral units involved in cognitive skills, however simple or complex, and targeting them with appropriate multipleexemplar training, should lead to significant improvements in the methods used in many educational settings. 1. Mainstream Education According to RFT, explicit training with multiple exemplars is required for a child to abstract any relational frame sufficiently to allow the child to apply it arbitrarily to almost any event. Multiple-exemplar training is an inherent component of normal educational practice. Much of education involves presenting children clusters of tasks that are grouped by content (adding two numbers together, identifying the adverb in numerous sentences, etc.). From the perspective of RFT, relational skills are incidentally trained in these tasks, even though these practices were not designed specifically for this purpose. Relational frame theory is more concerned

Relational Frame Theory

133

with the cognitive skills that are being abstracted rather than with the content of the material on which this abstraction is based. Concern for these core psychological processes lies at the heart of RFT, and one of its key research aims is to harness these skills in the form of effective interventions that will bring about improved educational achievement. The RFT approach to education, therefore, is twofold. First, relational skills lie at the heart of a wide range of cognitive abilities that form the basis of educational achievement. Second, multiple-exemplar training is critical to the effective education of these cognitive skills. Some evidence has already been obtained to support the principle that multipleexemplar training plays a vital role in simple forms of derived relational responding such as the transformation of functions in accordance with symmetry (BarnesHolmes, Barnes-Holmes, Roche, & Smeets, 2001 a). Other evidence supports the role of multiple-exemplar training in complex forms of derived relational responding, such as responding in accordance with the relational frames of more-than, lessthan, and opposite (Barnes-Holmes, Barnes-Holmes, Roche, & Smeets, 2001b). This research may have important implications for how we teach cognitive abilities to young children in the course of normal educational practice. For example, a great many early mathematical tasks involve numerous exemplars of responding in accordance with items that are the same as, more than, or less than each other ("If these two coins look the same, but this one is worth more than this one, which coin would you choose?"). The more than/less than study referred to here, for example, demonstrated that for a number of children, training in the relational frame of "same" facilitated responding in accordance with opposite. These findings might significantly improve the focus and success of common educational practice, for example, by training children to respond in accordance with same before commencing training in the more complex relations of more than and less than. A number of other possibly important research questions emerge. What are the key relational frames that underpin specific areas of educational achievement, such as reading or arithmetic? Does multiple-exemplar training in nonarbitrary relations facilitate or hinder arbitrarily applicable relational responding? What is the minimum number of exemplars required for effective application of a particular relational frame? As long as these and many other important research questions remain, we will continue to wonder why education works and how it can be improved.

2. Special Education Relational frame theory may have a useful contribution to make to special education, for example, with developmentally delayed populations. Cognitive perspective-taking, commonly referred to as "Theory of Mind," is an essential feature of remedial educational programs for autistic children (Howlin, Baron-Cohen, & Hadwin, 1999). From the perspective of RFT, these special educational practices incidentally employ multiple exemplars that establish or facilitate relational

134

Yvonne B a r n e s - H o l m e s et al.

skills. According to RFT, cognitive perspective-taking involves increasingly complex contextual control of the relational frames of I-YOU, HERE-THERE, and NOW-THEN, and various combinations of these frames. According to RFT, a more effective means of establishing these repertoires would be to target these relational frames directly and thereby focus explicitly on the largely verbal nature of these skills. In one RFT research program, interventions consisting of training in the three perspective taking frames established repertoires of cognitive perspective-taking ab initio in young children (Barnes-Holmes, Barnes-Holmes, Roche, & Smeets, 2001 c). This finding has important implications for special education with autistic children. In this study, for example, the I-YOU perspective-taking frame appears to be an essential prerequisite to the HERE-THERE frame. These findings might contribute to the education of autistic children in that they suggest that these children should be trained to respond to the I-YOU frame before commencing training in the more complex relations of HERE-THERE and NOW-THEN. Deficiency in cognitive perspective-taking is a critical obstacle that impedes both the educational progress and the quality of life generally for autistic populations, and therefore warrants intense empirical and educational attention.

3. Higher Level Education: Logic Relational frame theory predicts that multiple-exemplar training should be useful in establishing the most highly complex forms of relational responding, in areas such as logic and abstract mathematics. According to RFT, logic is essentially a relational activity that involves the derived transformation of functions in accordance with multiple stimulus relations. Let us consider deduction and, in particular, focus on the errors that nonlogicians often make when confronted with examples of reasoning with conditionals. When faced with tasks or examples that require reasoning with conditionals, individuals usually make one of four deductions. First, individuals will often correctly make the modus ponens deduction: If there is a square then there is a circle; there is a square; therefore, there is a circle. Second, many individuals often fail to make the modus tollens deduction: If there is a square then there is a circle; there is not a circle; therefore, there is not a square (many report that nothing follows; see Evans, 1982; Wason & Johnson-Laird, 1972). Third, denying the antecedent is another logical error that people often make when reasoning with conditionals: If there is a square then there is a circle; there is not a square; therefore, there is not a circle (logically, there could be circle). Fourth, affirming the consequent is another logical error commonly made: If there is a square then there is a circle; there is a circle; therefore, there is a square (logically, there may not be a square). From a relational frame perspective, these examples of logical and illogical reasoning may be explained, in large part, by focusing on the verbal histories that established the functions of the contextual cues contained within the premises and

Relational Frame Theory

135

conclusions. The two cues that seem most relevant for RFT are the words "if" and "then," and the key question concerns the relational functions that have been established for these words in the history of the reasoner. According to RFT, if and then often function as contextual cues that cause a listener to respond to the term that directly follows "then" as completely dependent upon the term that directly follows "if," although according to strict logic the dependency is not complete. For example, a child might be told by a parent; "if you tidy your room, then you will get some chocolate," implying that getting chocolate depends entirely on tidying the room. Parents would not normally use this if-then statement to indicate that the child might get the chocolate even if the room is not tidied, although according to strict logic, this possibility is implied by the if-then statement (see the third deduction). Similarly, parents would not wish to imply that eating chocolate might occur in the absence of a tidied room, although this again is implied according to strict logic (see the fourth deduction). These types of errors in strict logic, therefore, are entirely predictable given the common history to which most of us have been exposed. Of course, not all reasoners produce the typical errors associated with the latter three previously described deductions. To explain this outcome, RFT would again look to the verbal history and current context of the reasoner. The verbal history of a sophisticated reasoner may have established multiple relational functions for the cues "if' and "then," such that in some sentences these cues establish a relation of complete dependency, and in other sentences they do not. The nature of the relational functions that are actualized within a particular instance of reasoning will, therefore, determine whether or not the reasoner reasons correctly. For example, persons who have a common history will have the common (mis)understanding of "if-then" and therefore should use it in dealing with conditionals in everyday contexts. However, if they have also been trained in formal logic, they should use the formal meaning of "if-then" in dealing with conditionals that are posed as logical problems, that is, presented in the context of logic.

III. Conclusion Based on a relatively small array of psychological and behavioral processes, RFT allows even fairly complex verbal events, such as metaphor and humor, to be approached behaviorally. From this perspective most human actions are verbal or cognitive in part, as the functions of the nonarbitrary environment are transformed by way of derived stimulus relations. This property means that traditional behavioral interpretations of virtually all complex forms of human behavior must be reworked. Therefore, RFT is appropriately characterized as a post-Skinnerian behavior analytic theory. If RFT is correct, many important topics in behavioral psychology will need to be reworked substantially, including the nature of verbal

136

Yvonne Barnes-Holmes et al.

behavior and of cognition, the nature of self, the nature and role of self-knowledge, the nature of problem solving, the nature of intelligence, and the focus of education. Behavioral psychologists may or may not be willing to take steps of this kind, even though the empirical basis for the approach we are presenting came from behavioral psychology. The fact remains that behavioral approaches to language and cognition have not been very successful empirically, however. We believe that RFT offers a coherent alternative worth exploring.

REFERENCES Barnes, D., Hegarty, N., & Smeets, E M. (1997). Relating equivalence relations to equivalence relations: A relational framing model of complex human functioning. The Analysis of Verbal Behavior, 14, 57-83. Barnes, D., & Keenan, M. (1993). Concurrent activities and instructed human fixed-interval performance. Journal of the Experimental Analysis of Behavior, 59, 501-520. Barnes, D., & Roche, B. (1997). A behavior-analytic approach to behavioral reflexivity. The Psychological Record, 47, 543-572. Barnes-Holmes, D., Barnes-Holmes, Y., & Cullinan, V. (2000). Relational frame theory and Skinner's Verbal Behavior: A possible synthesis. The Behavior Analyst, 23, 69-84. Barnes-Holmes, Y., Barnes-Holmes, D., Roche, B., & Smeets, P. M. (2001a). Exemplar training and a derived transformation of function in accordance with symmetry. The Psychological Record. In press. Barnes-Holmes, Y., Barnes-Holmes, D., Roche, B., & Smeets, P. M. (2001b). Testing and training relational responding in accordance with "more-than," "less-than," and opposite in young children. In preparation. Barnes-Holmes, Y., Barnes-Holmes, D., Roche, B., & Smeets, P. M. (2001c). Analyzing the development of complex perspective-taking in young children. In preparation. Bentall, R. P., & Lowe, C. E (1987). The role of verbal behavior in human learning. III. Instructional effects in children. Journal of the Experimental Analysis of Behavior, 47, 177-190. Berko, J. (1958). The child's learning of English morphology. Word, 14, 150-177. Brown, A. L. (1989). Analogical learning and transfer: What develops? In S. Vosniadou and A. Ortony (Eds.), Similarity and analogical reasoning (pp. 369-412). Cambridge, England: Cambridge University Press. Chomsky, N. (1959). A review of B. E Skinner's Verbal Behavior. Language, 35, 26-58. Dougher, M. J., Auguston, E. M., Markham, M. R., Greenway, D. E., & Wulfert, E. (1994). The transfer of respondent eliciting and extinction functions through stimulus equivalence classes. Journal of the Experimental Analysis of Behavior, 62, 331-351. Dymond, S., & Barnes, D. (1994). A transfer of self-discrimination response functions through equivalence relations. Journal of the Experimental Analysis of Behavior, 62, 251-267. Dymond, S., & Barnes, D. (1995). A transformation of self-discrimination response functions in accordance with the arbitrarily applicable relations of sameness, more-than, and less-than. Journal of the Experimental Analysis of Behavior, 64, 163-184. Dymond, S., & Barnes, D. (1996). A transformation of self-discrimination response functions in accordance with the arbitrarily applicable relations of sameness and opposition. The Psychological Record, 46, 271-300. Evans, J., St. B. T. (1982). The psychology of deductive reasoning. London: Routledge and Kegan Paul. Fischer, K. W. (1980). A theory of cognitive development: The control and construction of hierarchies of skills. Psychological Review, 87, 477-531.

Relational Frame Theory

137

Harmon, K., Strong, R., & Pasnak, R. (1982). Relational responses in tests of transposition with rhesus monkeys. Learning and Motivation, 13 (4), 495-504. Hayes, S. C. (1984). Making sense of spirituality. Behaviorism, 12, 99-110. Hayes, S. C. (1994). Relational frame theory: A functional approach to verbal events. In S. C. Hayes, L. J. Hayes, M. Sato, and K. Ono (Eds.), Behavior analysis of language and cognition, (pp. 9-30). Reno, NV: Context Press. Hayes, S. C. (1995). Knowing selves. The Behavior Therapist, 18, 94-96. Hayes, S. C., Barnes-Holmes, D., & Roche, B. (2001). Relational Frame Theory: A post-Skinnerian account of language and cognition. New York: Plenum Press. Hayes, S. C., Brownstein, A. J., Devany, J. M., Kohlenberg, B. S., & Shelby, J. (1987). Stimulus equivalence and the symbolic control of behavior. Mexican Journal of Behavior Analysis, 13, 361-374. Hayes, S. C., & Gifford, E. V. (1997). The trouble with language: Experiential avoidance, rules, and the nature of verbal events. Psychological Science, 8, 170-173. Hayes, S. C., Gifford, E. V., & Hayes, G. J. (1998). Moral behavior and the development of verbal regulation. The Behavior Analyst, 21, 253-279. Hayes, S. C., & Hayes, L. J. (1989). The verbal action of the listener as a basis for rule-governance. In S. C. Hayes (Ed.), Rule-governed behavior: Cognition, contingencies, and instructional control (pp. 153-190). New York: Plenum. Hayes, S. C., & Hayes, L. J. (1992). Verbal relations and the evolution of behavior analysis. American Psychologist, 47, 1383-1395. Hayes, S. C., Kohlenberg, B. K., & Hayes, L. J. (1991). The transfer of specific and general consequential functions through simple and conditional equivalence classes. Journal of the Experimental Analysis of Behavior, 56, 119-137. Hayes, S. C., & Wilson, K. G. (1993). Some applied implications of a contemporary behavior-analytic account of verbal events. The Behavior Analyst, 16, 283-301. Hoffman, M. L. (1982). Development of prosocial motivation: Empathy and guilt. In N. Eisenberg (Ed.), The development ofprosocial behavior (pp. 281-314). New York: Academic Press. Hovland, C. I., Lumsdaine, A. A., & Sheffield, E D. (1949). Experiments on mass communication. Princeton, N J: Princeton University Press. Howlin, P., Baron-Cohen, S., & Hadwin, J. (1999). Teaching children with autism to mind-read: A practical guide. Chichester: Wiley. Lipkens, G. (1992). Analogical reasoning as arbitrarily applicable relational responding. Unpublished manuscript. University of Nevada, Reno. Miller, P. J., & Sperry, L. L. (1987). The socialization of anger and aggression. Merrill-Palmer Quarterly, 33, 1-31. Ortony, A. (1993). Metaphor and thought. Cambridge, England: Cambridge University Press. Piaget, J. (1967). Six psychological studies. New York: Vintage. Pinker, S. (1990). Language acquisition. In D. N. Osherson and H. Lasnik (Eds.), An invitation to cognitive science, Vol. I. Language (pp. 199-241). Cambridge, MA: MIT Press. Pinker, S. (1991). Rules of language. Science, 253, 530-535. Pinker, S. (1999). Words and Rules: The Ingredients of Language. New York: Basic Books. Place, U. T. (1998). Sentence and sentence structure in the analysis of verbal behavior. Analysis of Verbal Behavior, 15, 131-133. Polya, G. (1954). Mathematics and plausible reasoning: Volume 1. Induction and analogy in mathematics. Princeton, N J: Princeton University Press. Reese, H. W. (1994). Cognitive and behavioral approaches to problem solving. In S. C. Hayes, L. J. Hayes, M. Sato, and K. Ono (Eds.),Behavior analysis oflanguage and cognition (pp. 197-258). Reno, NV: Context Press. Roche, B., & Barnes, D. (1996). Arbitrarily applicable relational responding and sexual categorization: A critical test of the derived difference relation. The Psychological Record, 46, 451-475.

138

Yvonne Barnes-Holmes et al.

Roche, B., & Barnes, D. (1997). A transformation of respondently conditioned function in accordance with arbitrarily applicable relations. Journal of the Experimental Analysis of Behavior, 67, 275-301. Roche, B., Barnes-Holmes, D., Smeets, P. M., Barnes-Holmes, Y., & McGeady, S. (2000). Contextual control over the derived transformation of discriminative and sexual arousal functions. The Psychological Record, 50, 267-292. Skinner, B. E (1953). Science and human behavior. New York: Free Press. Skinner, B. E (1957). Verbal behavior. New York: Appleton-Century-Crofts. Steele, D. L., & Hayes, S. C. (1991). Stimulus equivalence and arbitrarily applicable relational responding. Journal of the Experimental Analysis of Behavior, 56, 519-555. Sternberg, R. J. (1977). Intelligence, information processing, and analogical reasoning: The componential analysis of human abilities. Hillsdale, NJ: Erlbaum. Wason, P. C., & Johnson-Laird, P. N. (1972). Psychology of reasoning: Structure and content. Cambridge, MA: Harvard University Press. Zettle, R. D., & Hayes, S. C. (1982). Rule-governed behavior: A potential theoretical framework for cognitive-behavior therapy. In P. C. Kendall (Ed.), Advances in cognitive-behavioral research and therapy (pp. 73-118). New York: Academic Press.

THE CONTINUITY OF DEPRESSION ACROSS THE ADOLESCENT TRANSITION

Shelli Avenevoli* DEPARTMENT OF EPIDEMIOLOGY AND PUBLIC HEALTH YALE UNIVERSITY SCHOOL OF MEDICINE NEW HAVEN, CONNECTICUT 06510

Laurence Steinberg DEPARTMENT OF PSYCHOLOGY TEMPLE UNIVERSITY PHILADELPHIA, PENNSYLVANIA 19122

I. INTRODUCTION II. DISORDER, SYNDROME, AND MOOD A. CONCEPTUAL AND METHODOLOGICAL DIFFERENCES B. DIFFERENCES IN PREVALENCE RATES ACROSS DEFINITIONS OF DEPRESSION C. DIFFERENCES IN ESTIMATES OF STABILITY ACROSS DEFINITIONS OF DEPRESSION III. STABILITY AND CONTINUITY A. HOMOTYPIC AND HETEROTYPIC CONTINUITY B. PRELIMINARY EVIDENCE OF HETEROTYPIC CONTINUITY OF DEPRESSIVE SYMPTOMS C. INTERNALIZED NEGATIVE AFFECTIVITY AND HETEROTYPIC CONTINUITY D. SUMMARY IV. GENETIC, BIOLOGICAL, COGNITIVE, AND SOCIAL INFLUENCES ON CONTINUITY A. GENETIC INFLUENCES B. BIOLOGICAL INFLUENCES C. COGNITIVE INFLUENCES D. S O C I A L - C O N T E X T U A L INFLUENCES E. SUMMARY V. CONCLUSIONS REFERENCES *Address for correspondence: Shelli Avenevoli, Genetic Epidemiology Research Unit, Yale University School of Medicine, 40 Temple Street, Suite 7B, New Haven, CT 06510 or at

shelli.avenevoli @yale.edu. 139 ADVANCESIN CHILDDEVELOPMENT AND BEHAVIOR,VOL.28

Copyright 9 2001 by AcademicPress. All rightsof reproductionin any formreserved. 0065-2407/01 $35.00

140

Shelli Avenevoli and Laurence Steinberg

I. Introduction Central to a developmental perspective on human behavior is the notion that development is lawfulm"that there is a common general course of development ... [and] a coherence to the course of each individual's development" (Sroufe & Rutter, 1984, p. 21). Following from this basic tenet, the scholarly pursuits of developmentalists focus on the continuities of traits, behaviors, and processes across developmental periods (Carlson & Garber, 1986). For decades, the direct examination of continuity over time and across important developmental transitions has led to our understanding of the development, the stability, and the changing manifestation of particular traits and capacities. For example, studies of continuity have informed the field about the course of personality development and about the stability and change that characterizes this evolution (Caspi, 1998; Caspi & Bem, 1990). Moreover, the underlying assumption of continuity that pervades most developmental research has lead to our more general understanding of adaptive functioning throughout the life span. The concept of continuity is relevant to the study of nonnormative processes as well. Conceptualized within a developmental psychopathology framework, the study of continuity allows for the examination of developmental pathways leading from adaptation or maladaptation at one point in the life span to adaptation or maladaptation at a later period. Moreover, a developmental approach to the study of psychopathology recognizes multiple pathways of adaptation or maladaptation based on individual differences in experience and proneness to disorder (e.g., Cicchetti & Rogosch, 1996). This framework acknowledges that behavioral manifestations of a particular type of psychopathology may vary in different developmental periods and that age-related differences in cognitive, affective, and behavioral aspects of psychopathology may reflect emergent capacities in cognition, language, and socioemotional development (e.g., Sroufe & Rutter, 1984; Cicchetti & Schneider-Rosen, 1986). Accordingly, a developmental approach can be used to uncover and examine qualitative links among meaningful periods of the life span, thereby informing the field about continuity and early developmental trajectories leading to a particular outcome. Studies of the continuity of psychopathology have focused on the stability of behaviors (e.g., Dumas, Neese, Prinz, & Blechman, 1996; Fergusson, Horwood, & Lynskey, 1995; Verhulst & van der Ende, 1992), the changing manifestations of underlying problems (e.g., Moffit, 1993; Patterson, 1992), and the developmental course leading from adaptation to maladaptation (e.g., Patterson, DeBaryshe, & Ramsey, 1989). Although work in this area has provided valuable information on the developmental trajectories of maladaptive behaviors, extant research is limited in its focus on externalizing problems such as delinquent and other acting-out behaviors, rather than internalizing problems such as depression and anxiety. In contrast, in this article we focus on the continuity of depression, one of the most commonly observed internalizing problems.

Continuity of Depression

141

Depression is an especially salient case of maladaptation that warrants examination from a developmental perspective. By virtually any indicator, depression is one of the most significant mental health problems in contemporary society, as exemplified by its high prevalence rates, potential chronicity, and resultant impairment. Clinical depression is the most highly reported mental health condition in both mental health and medical settings (Blacker & Clare, 1987), and the prevalence of depression is increasing in the general population (Lewinsohn, Rohde, Seeley, & Fischer, 1993). Moreover, the age of onset is decreasing with each generation (Kovacs, cited in Lewinsohn et al., 1993). Although the continuity of depression can be traced across any developmental transition (e.g., childhood to adolescence, adolescence to adulthood, adulthood to older adulthood), the transition from childhood to adolescence is a particularly important developmental period for the study of the continuity of depression for many reasons. First, adolescence appears to be marked by a rise in depressive mood, symptoms, and disorder. Second, gender differences in the prevalence of depression emerge during adolescence (Nolen-Hoeksema & Girgus, 1994; Petersen et al., 1993). Prior to adolescence, prevalence rates of depression are low, and gender differences appear to be nonexistent. However, during adolescence prevalence rates of depression increase, especially for girls, and continue to rise into adulthood. Findings of increased prevalence and emergent gender differences regarding depression suggest something particular to adolescence or to the events associated with the period that contribute to the "onset" of depression or to the "differential manifestation" of symptoms (i.e., adolescents may express behaviors that are more recognizable as depressive symptoms) during this era. More importantly, however, the examination of depression across the transition from childhood to adolescence is especially salient given the many biological, social, psychological, and cognitive changes that take place during this time. This transition is also significant because of its implications for early detection and prevention. Research that identifies early manifestations of depression in childhood may eventually lead to efforts to prevent depression in adolescence and adulthood. Because early-onset depression is particularly oppressive--depressed children and adolescents are at increased risk for more frequent and severe episodes of depression (Kovacs, 1996) and their social relationships and overall functioning remain markedly impaired even after recovery (Puig-Antich et al., 1985; Rao et al., 1995)--the identification of depression prior to adolescence is especially critical for prevention. Three shortcomings have limited progress in understanding the continuity of depression and other internalizing problems. First, research on depression has seldom used prospective, longitudinal designs that examine continuity and change across important developmental transitions. Second, research on the continuity of depression has been hindered by inconsistent definitions of depression; as we shall discuss, some studies examine depressive symptomatology, others examine depressive syndromes, and still others examine depressive disorder (see Compas,

142

Shelli Avenevoli and Laurence Steinberg

Ey, & Grant, 1993). Third, many researchers have erroneously equated continuity with behavioral stability, thereby underestimating the true extent and nature of continuity in depression. In the current article, we use a developmental psychopathology framework to review existing research on the continuity of depression across the important developmental transition from childhood to adolescence. We organize existing findings on the continuity of depression, highlight gaps in the literature, and suggest ways in which to improve upon previous research. In Section II, we present multiple conceptualizations of depression and show that these different conceptualizations result in varying estimates of prevalence and continuity. In Section III, we outline the different types of the continuity of depression that may be observed during the transition from childhood to adolescence, and in Section IV we discuss the genetic, biological, cognitive, and social influences that may explain continuity of depression. Finally, in Section V, we propose a research approach to examining the continuity of depression for future investigations.

II. Disorder, Syndrome, and Mood We noted earlier that one reason for the lack of sufficient knowledge on the continuity of depression concerns problems in definition. The term "depression" describes a variety of related phenomena in the psychological literature (Angold, 1988), including depression as a clinical disorder, depression as a syndrome, and depression as a symptom (i.e., depressed mood or affect). Although these conceptualizations represent different approaches to the measurement of depression as well as varying levels of impairment, research on the continuity of depression has been hindered by a failure to clearly articulate what is meant by depression, to recognize that estimates of continuity may vary with the definition of depression, and to incorporate the three conceptualizations of depression into a more global assessment of depression. A. CONCEPTUAL AND METHODOLOGICAL DIFFERENCES

In an attempt to remedy the confusion over what is meant by "depression," Compas et al. (1993) and others (Angold, 1988) highlight the need to distinguish among three different operationalizations: depressive disorder, depressive syndromes, and depressed mood. As a disorder, depression is conceptualized as a diagnostic category based on an already-identified grouping of symptoms (e.g., those established by the American Psychiatric Association or the World Health Organization) and the disability or impairment caused by these symptoms (Compas et al., 1993). Important examples are diagnoses of major depressive disorder (MDD) and dysthymic disorder.

Continuity of Depression

143

(Bipolar disorder does not represent "pure" depression and, consequently, will not be discussed). For a child or adolescent to be diagnosed with MDD according to DSM-IV (APA, 1994), he or she must exhibit five or more of the following symptoms: depressed or irritable mood, significantly decreased interest in most activities, weight/appetite loss or gain, sleep problems, psychomotor agitation or retardation, fatigue or loss of energy, excessive feelings of worthlessness or guilt, difficulty concentrating or making decisions, and repeated thoughts of death, suicidal ideation, or suicidal plans and/or attempts. Depressed or irritable mood or decreased interest in pleasurable activities for most of the day, almost every day, are required for diagnoses. Furthermore, symptoms must exist for at least a two-week interval, must cause clinically significant impairment of functioning, and may not be accounted for by another psychiatric disorder. Criteria for diagnoses of dysthymic disorder for children and adolescents requires a period of at least one year of daily depressed or irritable mood. The child cannot be symptom free for more than 2 months, and there cannot be an episode of MDD during the first year of dysthymia. In addition, dysthymic disorder diagnosis includes at least two of the following symptoms: eating problems, sleeping problems, lack of energy, low self-esteem, difficulty concentrating or making decisions, and feelings of hopelessness. Measures typically used in studies of depressive disorders are structured or semistructured interviews, such as the Diagnostic Interview for Children (DISC) and the Schedule for Affective Disorders and Schizophrenia for School-Age Children (K-SADS). Although the most commonly used measures of depression have been constructed for use with children and/or adolescents, the diagnostic criteria on which they are based have been established mainly on samples of adults. If, in fact, the manifestation of depression changes between childhood and adolescence, or between adolescence and adulthood, prevalence estimates among youth that are based on adult-derived diagnostic criteria may be faulty. The second approach to the study of depression defines depression as a syndrome--a constellation or pattern of behaviors and emotions that represent depression (Compas et al., 1993). Unlike the disorder approach, which relies on predetermined diagnostic categories, syndromes, as defined by Compas et al. (1993), are typically derived empirically and measured through questionnaires completed by respondents (e.g., Youth Self Report) or other knowledgeable informants (e.g., Child Behavior Checklist, Teacher Report Form). To date, much of the literature on depressive syndromes reflects the empirical work of Achenbach and colleagues. Through parent, teacher, and adolescent reports, Achenbach (1991 a, 1991b, 1991c) has reliably identified a cluster of symptoms that reflect an anxious/depressed syndrome. These include the following: cries a lot; worries; fears impulses; acts perfectionistic; feels lonely, unloved, persecuted, worthless, nervous, fearful, guilty, self-conscious, suspicious, and sad. Other researchers have employed the Child Depression Inventory (CDI) and the Center for Epidemiologic

144

Shelli Avenevoli and Laurence Steinberg

Studies of Depression Scale (CES-D) to examine syndromes of depression. These scales were not derived empirically, and they are composed of depressive symptoms from the D S M - I I I (APA, 1980) and the Research Diagnostic Criteria (RDC; Spitzer, Endicott, & Robins, 1978a, 1978b). Many of these scales have identified a threshold above which scores are considered clinically significant. The third approach to the study of depression measures depressed m o o d or affectmperiods of sadness, unhappiness, or dysphoric mood. Unlike depressive disorders or syndromes, depressed mood is often normative and has no conditions on the length of time it may last. Depressed affect is common to all individuals at some point in their life and is a natural reaction to certain events, such as the loss of a loved one (Compas et al., 1993). Depressed mood is most often assessed with single-item or multi-item self-report measures. Although there are numerous measures of this sort, depressed mood is typically derived from checklists of emotions and symptoms, which employ Likert-type scales that instruct participants to rate degree of symptom severity or degree to which each symptom applies to them. Some existing scales (e.g., CDI) also include non-mood-related items reflective of diagnostic symptoms of depression (e.g., eating disturbance, somatic complaints). Subsets of items from these scales have been identified as indices of depressed affect (Weiss et al., 1991). B. DIFFERENCESIN PREVALENCERATES ACROSSDEFINITIONSOF DEPRESSION Until recently, little had been known about the prevalence of childhood and adolescent depression. Previously, prevalence rates for depression were based on studies with small, unrepresentative, or primarily clinical samples of children and adolescents (Gotlib & Hammen, 1992). Although more recent studies are based on larger, more representative, community samples, there is wide variation in estimates of prevalence across studies (primarily attributable to methodological inconsistencies, such as use of different assessment instruments and sampling frames, across studies; Merikangas & Angst, 1995). In two reviews of the largest and most comparable epidemiologic studies, prevalence rates for major depression among school-age children ranged from less than 1 to 3%. Rates among adolescents ranged from 1 to 8% (Hammen & Rudolph, 1996; Birmaher et al., 1996), with a study of older adolescents (15- to 18-year-olds) reporting an estimate of 3% (Lewinsohn, Rohde, & Seeley, 1998). Estimates of prevalence for dysthymic disorder follow this same general developmental pattern, but are typically lower overall (Fleming & Offord, 1990; Hammen & Rudolph, 1996), although one study of a sample of adolescents reported a rate of 8% (Kashani et al., 1987). The prevalence of the anxious/depressed syndrome identified by Achenbach and colleagues is estimated at 5% in the normal population of both children and adolescents. "Because this approach is empirically based, it is recognized that any

Continuity of Depression

145

criterion used to establish prevalence is to some degree arbitrary" (Compas et al., 1993, p. 328). However, prevalence rates are based on the 95th percentile, which was established as the cutoff point to represent the clinical range for Achenbach's syndrome (Achenbach, 1991 c, 1991 d, 1991 e). The 95th percentile minimizes both false positives (specificity) and false negatives (sensitivity) in discriminating referred and nonreferred youth. The prevalence estimate for depressive syndrome derived from the CES-D was also reported at 5% in one study of adolescents (Garrison et al., 1989). Studies that have examined prevalence rates of depressed mood in community samples of children and adolescents suggest that depressed affect is more prevalent than both depressive disorder and the anxious/depressed syndrome. In the few studies of community samples of preadolescent children, prevalence rates ranged from approximately 10 to 18% (e.g., Kashani & Simonds, 1979; Rutter, 1986b). In contrast to the paucity of research on depressed mood in childhood, depressed mood in adolescence has been examined in numerous studies. Although it is difficult to distinguish prevalence of depressive syndromes measured on depression scales from prevalence of depressed affect measured on some of the same scales, a review by Petersen, Compas, and Brooks-Gunn (1992) suggests that prevalence rates of depressed affect among adolescents range from about 21 to 50%. In his Isle of Wight study, Rutter (1986a, 1986b) reported that rates of depressed mood sharply increased from 10-12% when participants were 10-years-old to 45% when they were 14- to 15-years-old. Not surprisingly, the different operationalizations of depression yield varying estimates of prevalence. Rates of depressive disorder appear to be the lowest and rates of depressed mood appear to be the highest. C. DIFFERENCES IN ESTIMATES OF STABILITY ACROSS DEFINITIONS OF DEPRESSION

The different definitions and ways of measuring depression also result in divergent estimates of the continuity of depression. In this section, we briefly review studies that have examined the continuity of depressive disorder, the anxious/ depressed syndrome, and depressive symptoms. Unfortunately, only a limited amount of empirical work has directly examined the continuity of depression from childhood to adolescence. In addition, as mentioned previously, studies of continuity have been limited primarily to examinations of behavioral stability. Therefore, the literature reviewed in this section focuses on studies assessing the stability of depression over time. The stability of depressive disorder is typically examined in terms of rates of recurrence, because clinical depression is generally episodic rather than unremitting over long periods of time. It is widely accepted that individuals who have experienced major depressive episodes are at increased risk for recurrence

146

Shelli Avenevoli and Laurence Steinberg

(Kovacs, 1996; Lewinsohn et al., 1994). Children diagnosed with depression are likely to experience recurrences of depression within a few years (e.g., Garber, Kriss, Koch, & Lindholm, 1988; Lewinsohn, Hops, Roberts, Seeley, & Andrews, 1993). For example, Asarnow et al. (1988) found that 35% of depressed children were rehospitalized within 1 year of discharge, and 45% were rehospitalized within 2 years. Similarly, Kovacs et al. (1984) reported that 26% of a sample of children who had recovered from major depression experienced a second episode within 1 year, and 40% experienced a second episode within 2 years. In a review of longitudinal studies of clinical and epidemiological samples of children and adolescents, B irmaher et al. (1996) report a cumulative probability of recurrence of 40% by 2 years and 70% by 5 years. Although they did not assess the stability of depression specifically, McGee and colleagues (McGee, Feehan, Williams, & Anderson, 1992) report on the stability of internalizing disorder (mood and/or anxiety disorder) from early adolescence to middle adolescence. They found that girls who had an internalizing disorder at age 11 were six times more likely to exhibit an internalizing disorder at age 15. Studies on the stability of depressive syndromes have been based primarily on the anxious/depressed syndrome as measured via responses to the Child Behavior Checklist (CBCL), Teacher Report Form (TRF), and/or Youth Self-Report (YSR). In one study, McConaughy, Stanger, and Achenbach (1992) performed a 3-year reassessment of a nationally representative sample of American children first assessed at ages 4 to 16 (Achenbach, Howell, Quay, & Connors, 1991). Averaged across age and gender groups, stability of parents' ratings of children's problems over a 3-year period was moderate to large for the anxious/depressed syndrome. Individuals who scored in the deviant range (i.e., >95th percentile on parent's reports of symptoms) of the anxious/depressive syndrome at the beginning of the study were more than nine times more likely to be in the deviant range 3 years later. This suggests that children meeting criteria for the anxious/depressed syndrome were at significant risk for continued disturbance. In a 6-year follow-up of the same nationally representative sample, Achenbach, Howell, McConaughy, and Stanger (1995) reported on the relations among the 1986, 1989, and 1992 anxious/depressed syndrome scores based on different combinations of parent, teacher, and self-reports. Stability of the anxious/depressed syndrome over a 3-year period ranged from moderate to large, depending on the informant and the gender of the child, even after taking into account other syndromes, family variables, and stressful experiences. Stability estimates over the 6-year period were smaller and accounted for only a small percentage of the variance (1 and 2% in girls and boys, respectively). However, the moderate correlation and predictive power (even after controlling for other relevant factors) suggests considerable continuity in the anxious/depressed syndrome over the 6-year period. The few studies that have examined depressive mood or symptoms also have found high stability over brief periods of time. For example, Cole, Martin, Powers,

Continuity of Depression

147

and Truglio (1996) reported high stability of a latent construct of depressed mood (based on self-, teacher-, parent-, and peer-report) over a 6-month period for both third graders and sixth graders. Similarly, studies examining 1-week to 6-month intervals on the Children's Depression Inventory (CDI) have yielded large correlations (Finch, Saylor, Edwards, & McIntosh, 1987; Kandel & Davies, 1986). Longitudinal studies spanning longer periods of time yield low-to-moderate stabilities and moderate predictive power of early depressive symptoms. Longitudinal analyses of adolescents' depression scores on the CES-D indicate moderate 1- and 2-year stability and good predictability (previous years' CES-D scores accounted for 12-20% of the variance in current scores; Garrison, Jackson, Marsteller, McKeown, & Addy, 1990). Longitudinal analyses of adolescents' scores on the CDI indicate moderate-to-high stability of depressive symptoms across consecutive years and lower stability across longer intervals (Devine, Kempton, & Forehand, 1994). One study suggests moderate stability of depressive symptoms during the transition from childhood to adolescence. Nolen-Hoeksema, Girgus, and Seligman (1992) reinterviewed children on the CDI every 6 months from the third grade through the seventh grade. CDI scores were moderately stable over a 1- to 4-year period, including the transition from childhood to adolescence. No known studies have reported the stabilities of individual symptoms (e.g., sleep disturbance, anhedonia, etc.) over time o r across developmental transitions. In sum, depressive disorders, syndromes, and mood appear to be moderately stable over time. In fact, the stability of depression is a relatively robust finding that spans multiple research designs, age groups, samples, and measures of depression. However, the degree to which "depression" is stable depends on the definition of depression employed. Based upon the data presented here, stability over periods more than 1 or 2 years appear to be highest for depressive disorder and syndrome and lowest for depressed mood (and total symptoms as assessed on self-report measures). As the time interval between assessments increases, stability rates for both depressive syndrome and depressed mood decrease. However, the variability in stability rates is attributable not only to the conceptual differences in the three approaches to depression, but also to the inconsistent way in which stability is defined and measured over time for the various types of depression. In particular, depressive disorder, by its very nature, is a generally episodic condition. As a result, the stability of depressive disorder is typically measured in terms of whether there has e v e r been a recurrent episode of depression within a specified amount of time (e.g., the last 10 years). In contrast, the time interval for the stability of depressive syndrome and mood typically focuses on an individual's present or very recent (e.g., within 6 months) emotional state and symptoms (i.e., focusing on a particular point in time). The findings of varying estimates of continuity across the three main conceptualizations of depression (e.g., disorder, syndrome, and symptom) suggest a need

148

Shelli Avenevoli and Laurence Steinberg

to distinguish among the different constructs in future discussions of continuity. To date, most research on the continuity of depression has focused on only one type of depression within any given sample. To advance the study of the continuity of depression, future research must not only acknowledge the different conceptualizations of depression but also incorporate these different conceptualizations of depression into assessments of depression across the life course.

III. Stability and Continuity Previous research on the continuity of depression is also limited by an almost exclusive focus on the stability of the same symptom or constellation of symptoms over time, which represents homotypic continuity. Examinations limited to behavioral stability are inherently nondevelopmental because a fundamental premise of the developmental approach is the notion that "continuity lies not in isomorphic behaviors over time but in lawful relations to later behavior, however complex the link" (Sroufe & Rutter, 1984, p. 21). A developmental approach not only argues for the examination of continuity across developmental periods but also acknowledges that emergent capacities contribute to differences in the experience and expression of depression during different developmental periods. As suggested by this approach, a more comprehensive understanding of continuity demands that attention be paid to both homotypic continuity and heterotypic continuity (continuity of the underlying construct). A. HOMOTYPIC AND HETEROTYPIC CONTINUITY

1. Homotypic Continuity Homotypic continuity--also called phenotypic continuity or behavioral stability --refers to stabilities in the same or similar behavioral responses over time. For example, homotypic continuity of depression is evidenced by the persistence of the disorder, syndrome, or symptoms (e.g., dysphoric mood, lack of appetite, sleep problems, etc.) over time or across developmental periods. Kagan and Moss (1962) argue that homotypic continuity is rare, especially during the first 10 years of life or over long periods of time. Homotypic continuity is typical in adulthood or over short periods of time. While studies of behavioral stability during childhood and adolescence suggest moderate stability of depression over time, there is also evidence of some discontinuity or heterotypic continuity across time and perhaps across developmental transitions.

2. Heterotypic Continuity Heterotypic continuity refers to the continuity of an underlying construct that is manifested differentially over time or across developmental periods. Caspi and

Continuity of Depression

149

Bem (1990) suggest that this type of continuity refers to a conceptual continuity, and they define it as the continuity of an inferred attribute presumed to underlie diverse phenotypic behaviors. In terms of depression, heterotypic continuity suggests that depressive symptomatology is a manifestation of a more general, underlying pathology and that the external manifestation of this underlying pathology--whether labeled depression or something else--may vary with developmental stage. Although difficult to measure directly, heterotypic continuity is more plausible from a developmental perspective than is homotypic continuity because the former acknowledges the effects of emergent capacities on past, present, and future behavior. Unlike homotypic continuity, heterotypic continuity is likely to be observed during any developmental period and is especially salient during the first decade of life (Kagan & Moss, 1962). Kagan (1980) suggests that at critical junctures in development, behavior and processes are reorganized, resulting in the differential manifestation of earlier behavior. For example, he mentions the critical junctures when language first appears (between the ages of 18 and 24 months) and when change in cognitive functioning occurs (between the ages of 5 and 7 years). In accordance with Kagan's reasoning, the transition into adolescence also represents a critical juncture in which emergent cognitive capacities, social competencies, and biological changes influence the reorganization of behavior. B. PRELIMINARY EVIDENCE OF HETEROTYPIC CONTINUITY OF DEPRESSIVE SYMPTOMS

In addition to changes in nearly every domain of individual functioning, the period of adolescence is marked by change in multiple domains of social context. The many changes associated with the transition into adolescence alter the way that adolescents view and experience the world as well as how they respond to it. In the same way that these changes influence adolescent normative behavior, they also influence the experience and expression of psychopathology during adolescence, highlighting the importance of the transition between childhood and adolescence as a critical period for examining the continuity of depression. Considering the many changes that take place during the transition from childhood to adolescence, it is reasonable to assume that the manifestation of depression may differ, at least somewhat, during childhood and adolescence. Some empirical work has compared the similarity of frequency, severity, and types of symptoms of depression across different developmental and age groups. However, nearly all of this research has been cross-sectional and based on clinical samples presenting with depression or related problems or disorders. Findings from cross-sectional studies suggest a number of differences in the expression of depression from childhood to adolescence. One general finding is that preadolescent children are more likely to report somatic complaints than are

150

Shelli Avenevoli and Laurence Steinberg

adolescents (Kashani, Rosenberg, & Reid, 1989; Ryan et al., 1987; Weiss et al., 1992). In addition, depressed children exhibit greater depressed appearance (i.e., sad expression; Carlson & Kashani, 1988), less dysphoria, and less hopelessness (Ryan et al., 1987) than depressed adolescents. Depressed adolescents exhibit a significantly higher rate of hypersomnia than depressed children (Kovacs & Gatsonis, 1989; Mitchell, McCauley, Burke, & Moss, 1988; Ryan et al. 1987) and more anhedonia, hopelessness, weight loss, and use of alcohol and illicit drugs, as well as greater lethality of suicide attempts (Garber, 1984; Ryan et al., 1987). Suicidal behavior is usually rare in the general population before the ages of 12-14 years (Rutter, 1986b). In addition to changes in their presence or intensity, the manner in which symptoms of depression cluster together may also vary with age or developmental level. In one of the earliest studies comparing clusters of symptoms across age groups, McConville, Boag, and Purohit (1973) differentiated three groups of depressed 6- to 13-year-olds on type of symptoms. Symptoms of the "affectual type" (e.g., feelings of sadness, helplessness, loneliness) were found primarily among the 6- to 8-year-olds. Symptoms of the "self-esteem type" (e.g., feelings of hopelessness, of being disliked, of negative self-esteem) were found primarily among the 8to 10-year-olds. Symptoms of the "guilt type" (e.g., feelings of being wicked or hated, suicidal ideation) were found primarily among 10- to 13-year-olds. Garber (1984) also examined symptomatology in a sample of 137 seven- to thirteen-year-old girls. Nearly half of the girls in her sample manifested a depressive syndrome according to DSM-III but were not necessarily diagnosed with a Major Affective Disorder. Garber divided her participants into three age periods representing prepubertal children, transitional children, and adolescents: 7-8 years, 9-11 years, and 12-13 years, respectively. Self-reported level of depression did not vary by age, but types of symptoms did. Many of the symptoms that were rated higher (i.e., more severe) with age are of the "endogenous" (biological) subtype. In comparing endogenous symptoms (pervasive loss of interest, sleep problems, appetite problems, guilt, psychomotor agitation, and hypoactivity) across age groups, the oldest group was significantly different from the other two. In addition, Garber (1984) reported that hopelessness, irritability, fatigue, problems at school, difficulty concentrating, depressed feelings and low self-esteem all increased with age, whereas morbid ideation and weeping tended to decrease with age. Her work is an improvement over other previously mentioned studies in that it begins to address differences in depressive symptomatology over the transitional period from childhood to adolescence. However, this work is still limited by its cross-sectional nature and its inability to confirm that earlier manifestations of depression predict later manifestations of depression. In a previous study, we examined cross-sectional differences in depressive symptoms in a sample that spanned childhood and adolescence (Avenevoli, 1998; Avenevoli & Silberg, 1999). However, unlike the Garber study which included

Continuity of Depression

151

a sample of girls referred for educational, emotional, or behavioral evaluations, this study was conducted on an epidemiological sample exhibiting general population rates of depression and other disorders. The data were derived from the Virginia Twin Study of Adolescent Behavioral Development (e.g., Meyer, Silberg, Simonoff, Kendler, & Hewitt, 1996). The Child and Adolescent Psychiatric Assessment (CAPA) interview (Angold, Cox, Predergast, Rutter, & Simonoff, 1994) was used to obtain depressive and anxiety symptomatology among the children. The participants, 995 eight- to sixteen-year-olds, completed interviews in their homes at baseline and approximately 15 months later. To assess crosssectional differences, the study sample was divided by age into three groups representing three developmental periods--preadolescence (8-10 years), transitional (11-13 years), and adolescence (14-16 years)--chosen based on the normative age range for the most influential changes associated with adolescence (e.g., transition from junior high school to high school). Among girls, adolescents reported more intense symptoms of depression than transitional youth or preadolescents, virtually regardless of the specific symptom assessed. In particular, adolescents reported significantly more intense depressed mood, tearfulness, anergia, indecisiveness, loneliness, and hypersomnia than both transitional youth and preadolescents. Adolescent girls also reported significantly more intense agitation, irritability, anhedonia, boredom, hopelessness, suicidal thoughts, and reduced appetite than preadolescents. For other symptoms (loss of interest, inefficient thinking, guilt, etc.), group differences were not significant, although a similar trend was evident. Among boys, adolescents reported greater intensity of most symptoms than did preadolescents. Boys' reports of the intensity of symptoms on the CAPA indicated that transitional and adolescent boys experienced significantly more intense depressed mood, tearfulness, irritability, anergia, anhedonia, loneliness, and guilt and a greater increase in appetite than did preadolescent boys. Intensity was greatest for transitional and preadolescent boys for some symptoms. In particular, transitional boys reported the greatest intensity of self-depreciation, reduced appetite, and excessive appetite, and preadolescent boys reported the greatest intensity of thoughts of death. Once developmental differences in symptoms were examined cross sectionally, it was of interest to examine longitudinal changes in symptoms within developmental groups. Longitudinal analyses yielded many fewer significant findings but confirmed some of the differences found in cross- sectional analyses. In particular, the symptoms of hopelessness, irritability, and reduced appetite showed significant change over time among transitional and adolescent girls. Among boys, irritability, helplessness, and hypersomnia exhibited more change in the transitional and adolescent groups. Three different explanations may account for inconsistencies between crosssectional and longitudinal findings in this study. First, a significant amount of

152

Shelli Avenevoli and Laurence Steinberg

change may not occur in the same individuals over the transition from childhood to adolescence. Indeed, cross-sectional differences may represent cohort effects as well as developmental change. However, the cross-sectional differences found in the current study mimic those found in other cohorts, making this explanation unlikely. Second, the lack of significant longitudinal findings in the current study may be attributable to the age groups used. Although the age groups were meant to approximate specific developmental periods (preadolescence, transition into adolescence, and adolescence), age is not always the best indicator of developmental level as change occurs at different rates, within different domains, among different individuals. During the transition into adolescence, individuals may be more appropriately grouped by pubertal status, cognitive competence, or some other developmental milestone. Third, the small number of developmental changes found in the longitudinal analyses may be attributable to the short time interval (i.e., 15 months) representing the transition from childhood to adolescence in the current study. A greater length of time may be needed to find the differences evidenced in cross-sectional analyses. Nonetheless, the findings from this study suggest important changes in symptoms of depression across the adolescent transition. However, it is important to note that none of the studies reviewed in this section directly tests the continuity of the underlying phenomenon of depression. Additional work is necessary to determine whether earlier manifestations of depression predict later manifestations. Although firm conclusions cannot yet be made regarding the heterotypic continuity of depression, preliminary evidence of developmental differences and changes in symptoms suggest such a search is warranted. C. INTERNALIZED NEGATIVE AFFECTIVITY AND HETEROTYPIC CONTINUITY

Although they are informative, the studies just described are limited in that they did not examine changes in the manifestation of depression that include symptoms not currently defined as, but related to, classically identified symptoms of depression. A more thorough understanding of the continuity of depression and the differential manifestation of depression across developmental periods demands further exploration into the relation among depressive symptoms (as currently defined) and symptoms that are theoretically and empirically related to depression during different developmental periods. One promising avenue for a more thorough examination of the heterotypic continuity of depression involves the notion of internalized distress. Indeed, current debate in the psychiatric and psychological literatures centers on the issue of whether depression and anxiety represent distinct entities (with some overlapping features), or whether they represent different manifestations of a single, underlying construct (i.e., internalized distress). Historically, anxiety and depression have been regarded primarily as separate emotional problems. As a consequence, divergent

Continuity of Depression

153

literatures on the etiology and clinical course of the two disorders have emerged. Depression and anxiety have also warranted different diagnostic labels, assessment techniques, and treatments (King, Ollendick, & Gullone, 1991). Proponents of the unitary construct approach argue that it is not useful to distinguish between anxiety and depression given the considerable overlap in symptoms, family history, and etiology. Instead, they suggest that depression and anxiety represent a unitary construct of common emotions, cognitions, and biology. Some, like Watson and Clark (1984; following the work of Eysenck, 1960, 1986) use the term "negative affectivity" to characterize a global construct of emotional distress that includes both anxiety and depression. Others have used terms like "internalizing" or "overcontrolled" to describe factors that include problems related to depression, anxiety, social withdrawal, and somatic complaints (Achenbach & Edelbrock, 1978). Research and theory on children addressing the distinct entity versus common pathology debate has provided information to suggest that depression and anxiety are so strongly related that they may represent different manifestations of the same underlying pathology, or that they at least share common etiological roots. More importantly and most relevant to the present study, some research suggests that changes in the manifestation of internalized distress may be developmental in nature (e.g., Kovacs, 1990; Kovacs & Devlin, 1998).

1. Covariation and Comorbidity Evidence to suggest that depression and anxiety share a common underlying process stems partly from the high rates of comorbidity of depression and anxiety and the co-occurrence of their symptoms. In the study of children and adolescents diagnosed with depressive disorder, "comorbidity is the rule rather than the exception" (Gotlib & Hammen, 1992). Generally, a wide range of disorders, including conduct disorder, oppositional disorder, and alcohol and drug abuse, have demonstrated comorbidity with depressive disorder in both children and adolescents, but particular attention has been paid to anxiety disorders, which typically exhibit the highest rates of comorbidity with depression. In a review of studies incorporating both clinical and community samples, comorbidity of anxiety (including separation anxiety, generalized anxiety disorder, severe phobias, and obsessive-compulsive disorder) ranged from 30 to 75% in cases of depression (Kovacs, 1990). For example, in one community study of 11-year-olds sampled from the general population in New Zealand, Anderson, Williams, McGee, and Silva (1987) reported that 71% of children diagnosed depressed were concurrently diagnosed anxious. At age 15, 33% of depressives were concurrently diagnosed with anxiety, and at age 18, 50% had comorbid anxiety (McGee et al., 1990). Kashani et al. (1987) reported a rate of 75% comorbidity between depression and anxiety in their community sample of 14- to 16-year-olds. Thus, although estimates of comorbidity vary across studies (partly reflecting

154

Shelli Avenevoli and Laurence Steinberg

differences in diagnostic systems and in the approach to gathering clinical data) the results converge on the conclusions of substantial comorbidity of depression and anxiety. Studies that focus on anxiety as the diagnosis of interest in clinical samples report similar rates of comorbidity between anxiety and depression. According to the Kovacs (1990) review, 33-100% of indexed anxious children and adolescents merit an additional and concurrent diagnosis of depression. Strauss, Last, Hersen, and Kazdin (1988) reported that 28% of 5- to 17-year-old outpatients at an anxiety disorders clinic were also diagnosed with a depressive disorder. Also, in a study of school phobics, Bernstein and Garfinkel (1986) reported that 50% were depressed. Many investigators have also examined the relation between depressed mood or symptoms and anxiety symptoms using self-report measures with children and adolescents. Overall, studies that rely on a single method as well as those that employ multiple methods and multiple informants report substantial correlations among depressive and anxiety symptoms on self-report measures, behavioral scales, and rating scales (see Finch, Lipovsky, & Casat, 1989). In one study, children who reported themselves to be both depressed and anxious were also identified as depressed and anxious by their teachers and peers (Strauss, Forehand, Frame, & Smith, 1984). Some researchers argue that high rates of co-occurrence can be explained by item overlap on self-report scales of depression and anxiety and symptom overlap in diagnostic systems. Indeed, depression and anxiety disorders have many symptoms in common, and this may contribute to their apparent comorbidity. Similarly, the poor discriminant validity of depression scales and anxiety scales may well contribute to artificially high correlations between these measures (Brady & Kendall, 1992; Finch et al., 1989). However, poor discriminant validity is not a sufficient explanation for levels of comorbidity, because the relation between depression and anxiety remains even when nonoverlapping symptoms and criteria are examined (Angold, Costello, & Erkanli, 1999; Watson & Kendall, 1989). In addition, discriminant analyses have failed to detect natural boundaries between the disorders of depression and anxiety (see Cloninger, Martin, Guze, & Clayton, 1990). Moreover, the high correlation between anxious and depressive states is quite robust and generalizes across clinical and nonclinical samples, children and adults, different informants, and various types of assessment (Angold Costello, & Erkanli, 1999). 2. Failure to Isolate a Pure Syndrome Further evidence of the overlap between depression and anxiety follows from two converging lines of work that have failed to empirically isolate a discrete depressive syndrome for children and adolescents. In one attempt to derive an empirically based syndrome of depression, Quay (1986) identified a factor labeled anxiety-withdraw-dysphoria, which included symptoms of depression as well as

Continuity of Depression

155

anxiety. Similarly, in a principal-components analysis of reports on the CBCL, YSR, and TRF on a normative sample of children ranging in age from 4 to 16 years, Achenbach, Connors, Quay, Verhulst, and Howell (1989) failed to derive separate syndromes for depression and anxiety. Of the eight syndrome scales that resulted from factor analyses, one syndrome--the Anxious/Depressed Syndrome-incorporated symptoms of both depression and anxiety (see previous section for a list of symptoms). In subsequent work employing a confirmatory factor approach to the identification of latent constructs, evidence for distinct constructs of anxiety and depression was found (Chorpita, Albano, & Barlow, 1997). Studies in which confirmatory factor analysis were used provide a direct test of the hypothesis that depression and anxiety overlap; however, a shortcoming of these studies is that most were conducted with clinical samples, not the representative samples used by Quay and Achenbach. Differences in findings may be explained by this differing methodology such that symptoms of anxiety and depression may be more discriminable among youths with diagnoses of either anxiety or depression than those with no diagnosis (Gurley, Cohen, Pine, & Brook, 1996).

3. Specificity Aside from common symptoms and high rates of covariation, depression and anxiety also appear to share other attributes. Depressed children and anxious children experience many of the same cognitive processes, such as negative attributional style (Asarnow & Bates, 1988; Finch et al., 1989), negative self-evaluations (Asarnow & Bates, 1988; Zatz & Chassin, 1985), and cognitive errors (Leitenberg, Yost, & Carroll-Wilson, 1986). Depressed and anxious children also experience the same affective states, including negative affect and irritability. Depression and some forms of anxiety share common biological correlates (see Finch et al., 1989). They may benefit from the same treatments; however, this evidence is tentative and much more research is needed to draw firm conclusions (Clark, 1989; Stark, Reynolds, & Kaslow, 1987). Moreover, in family studies, disorders of depression and anxiety seem to be interchangeable across generations. In a review of seven high-risk studies of children of parents with either MDD or anxiety disorders, Weissman (1990) reported that children of depressed parents are as equally likely to become anxious as they are to become depressed (e.g., Weissman et al., 1987). Similarly, children of parents with generalized anxiety disorder or panic disorder are as likely to be diagnosed depressed as they are to be diagnosed anxious (see Weissman, 1990). Despite their similarities, other factors appear to be specific to either depression or anxiety. One of the most widely cited differences between anxiety and depression is the presence of low positive affect (see Watson & Kendall, 1989). Although high negative affectivity is associated with both depression and anxiety, low positive affectivity is associated only with depression (see Lonigan, Carey, & Finch, 1994,

156

Shelli Avenevoli and Laurence Steinberg

for the only known study of children and adolescents). In contrast, physiological hyperarousal is characteristic only of anxiety (Clark, Steer, & Beck, 1994). Content-specific cognitions are also believed to distinguish between symptoms of depression versus anxiety. According to Beck (1976), depressed individuals typically exhibit cognitive structures characterized by past-oriented, absolute, and loss-oriented thoughts, whereas anxious individuals report future-oriented, relative, and danger-oriented thoughts. Depressive cognitions are more highly related to depressive symptoms and anxiety cognitions are more highly related to anxiety symptoms (e.g., Clark, Beck, & Brown, 1989; Greenberg, Vasquez, & Alloy, 1988). Both affective (low positive affect) and cognitive components are needed to discriminate between depressive versus anxiety symptoms (Jolly, Dyck, Kramer, & Wherry, 1994). Unfortunately, most of this work has been conducted on adult samples, and the generalizability of these findings to children and adolescents is currently unknown.

4. DevelopmentalProgression of Anxiety and Depression A number of diverse theories and perspectives attempt to explain the strong relation between anxiety and depression. For example, Beck's theory of cognitive schemata proposes that both anxious and depressed individuals experience cognitive distortions of life events (Beck, 1976; Clark et al., 1989). Berenbaum, Kerns, and Taylor (1990) propose that anxiety and depression share the same underlying neuropsychological systems. Watson and Clark (1984) suggest that the two disorders share the same underlying affective state. A more recent perspective, the Helplessness-Hopelessness Perspective on Depression, provides a sophisticated model of the chain of events (involving stressful events and attributions) leading from anxiety to depression (Alloy, Kelly, Mineka, & Clements, 1990). This perspective is able to account for many of those aspects of comorbidity between anxiety and depression described in the adult psychiatric literature, including the sequential relation between anxiety and depressive disorders, the differential comorbidity between the disorders, the infrequency of pure depression, and the unique and overlapping symptoms of depression and anxiety (Alloy et al., 1990). This perspective suggests that anxiety results from a sense of helplessness and depression results from a sense of helplessness turning to hopelessness (i.e., a shift from the uncertainty of future outcomes to the certainty that future outcomes will be negative; Alloy et al., 1990). Although empirical research provides support for the HelplessnessHopelessness perspective, as well as other explanations of the comorbidity between anxiety and depression, most of these theories are based on findings in the adult literature on comorbidity and on studies that focus only on disorder (i.e., they consider overlap in diagnoses but not overlap in symptomatology). Although these theories might be extended successfully to children, or to the study of depressive syndromes or symptoms, developmental differences in the manifestation of

Continuity of Depression

157

depression may instead help explain the relation between anxiety and depression (Hammen & Compas, 1994). Given the overwhelming evidence that depression and anxiety are highly related, and given that depression likely manifests itself differently across development, the symptoms typically associated with anxiety disorders may represent an earlier developmental manifestation of depression or a more general pathology underlying different internalizing disorders. To date, no known studies have attempted to address this question directly. However, in depressed adolescents with comorbid anxiety, the onset of anxiety has been shown to precede the onset of depression (Kovacs, Gatsonis, Paulauskas, & Richards, 1989; Rohde, Lewinsohn, & Seeley, 1991). Furthermore, in these adolescents, anxiety often is present in childhood even though depression does not emerge until adolescence (Kovacs et al., 1989). In a sample of children and adolescents followed for 3 years, Cole, Peeke, Martin, Truglio, and Seroczynski (1998), reported that high levels of anxiety symptoms predicted high levels of depressive symptoms at subsequent points in time, even after controlling for prior level of depressive symptoms. Studies assessing the average age of onset of anxiety and depressive disorders suggest not only a temporal relation between anxiety and depression, but a relation that depends upon developmental level. According to a review by Kovacs and Devlin (1998), anxiety disorders typically appear sometime during early to middle childhood (average age of onset across studies ranges from 6.7 to 11.3 years). Depressive disorders, in contrast, typically surface during early to middle adolescence (range -- 10.8 to 14.9 years). This developmental progression is consistent with theoretical propositions on the expression of distress (Bowlby, 1973), animal models in response to stress (Suomi, 1991), and what we know about the development of emotions and emotional regulation in childhood and adolescence (Saarni, Mumme, & Campos, 1998). Further support for the notion that anxiety may be a developmental antecedent of depression is found in studies showing that depressive and anxiety symptoms tend to be more highly correlated among younger children than adolescents. In a multitrait, multimethod study of the relation between symptoms of anxiety and depression, for example, Cole, Truglio, and Peeke (1997) found that depression and anxiety were virtually indistinguishable constructs in a group of third grade "normal" children but were distinct (although highly related) constructs in a group of sixth grade "normal" adolescents. The developmental relevance of such findings are somewhat inconclusive because they are cross-sectional in nature and do not explicitly address the issue of developmental progression. However, these findings certainly warrant additional research that examines the progression of depression and its relation to anxiety across development. Application of a developmental perspective to the Helplessness-Hopelessness theory further corroborates the developmental relation between anxiety and depression. As children become adolescents who are more able to experience and

158

Shelli Avenevoli and Laurence Steinberg

understand complex cognitions like hopelessness, they are more likely to move from anxiety to depression. Moreover, as the emotions and cognitions associated with symptoms of depression and anxiety differentiate and become more complex with age (e.g., Weiner & Graham, 1988), individuals are more likely to experience pure affective states like depression. Yet, although they become more capable of experiencing pure depression, the reciprocal processes of integration and consolidation of emotions and cognitions operating during adolescence support the finding that comorbidity between anxiety and depression is also evident in adulthood. D. SUMMARY

In the psychiatric and psychological literature, the association between depression and anxiety (whether disorder or symptoms) is strong and robust. In light of findings that depression and anxiety are highly related and that their relation may be developmental in nature, research on the continuity of depression ought to consider that they are heterotypically continuous. Indeed, as we suggested previously, extant research on the continuity of depression has been hampered by a lack of attention to the changing manifestation of pathology (i.e., heterotypic continuity) over time and across development. Future research may benefit from the examination of the relation between anxiety and depression across different stages of development.

IV. Genetic, Biological, Cognitive, and Social Influences on Continuity A large literature has documented genetic, biological, cognitive, and psychosocial correlates of child and adolescent depression. Much less is known, however, about factors that may contribute to the continuity of depression during adolescence. This section briefly reviews findings on factors associated with depression in childhood and adolescence and presents evidence (often preliminary) regarding processes that may account for continuity of depression. At issue is whether individual differences in changes that occur in adolescence may help to explain patterns of continuity and discontinuity in depression. It should be noted that most evidence on continuity focuses on factors that contribute to the stability of depression, as little is currently known about the factors that influence heterotypic continuity. A. GENETIC INFLUENCES

Family and twin studies provide consistent evidence that MDD has a strong genetic component (Kovacs & Devlin, 1998), particularly for childhood-onset

Continuity of Depression

159

depression (Neuman, Geller, Rice, & Todd, 1997). Children of depressed parents are three times more likely to have an episode of major depression than children whose parents are not depressed (Birmaher et al., 1996). Risk is even greater when both parents exhibit a mood disorder (Merikangas, Weissman, Prusoff, & John, 1988). Genetic factors also play a large role in the recurrence and stability of depression. In one twin study, O'Connor, Neiderhiser, Neiderhiser, Hetherington, & Plomin, (1998) found that the stability in depressive symptoms over a 3-year period was explained primarily by genetic influence; Silberg et al. (1999) found similar results among girls but not boys. Although a strong genetic component is implicated, O'Connor et al. (1998) warn that it is premature to accept the conclusion that the identified influence of heritability is, in fact, purely genetic, given that reports suggest strong and pervasive gene-environment correlations (e.g., evocation of stressful events based upon a genetic predisposition) and in view of the fact that most heritability estimates fail to take these correlations into account. Consistent with this, Silberg et al. (1999) found that individuals who inherited a genetic predisposition for depression also inherited a tendency to experience negative life events. B. BIOLOGICAL INFLUENCES

Many of the biological changes associated with adolescence are linked to mood changes and the expression of depression. In particular, puberty-related hormonal changesmincluding rises in levels of adrenal and gonadal hormones during the pubertal yearsmare related to increases in depressed affect and MDD during this period. Negative affective tone has been linked to adrenal androgen levels among boys and to follicle-stimulating hormone levels (FSH) among girls (Nottelman, Inoff-Germain, Susman, & Chrousos, 1990; Susman et al., 1985). In addition, levels of testosterone and estradiol are both strongly predictive of concurrent depressive disorder (Angold, Costello, Erkanli, & Worthman, 1999). Although all adolescents experience hormonal changes during puberty, not all adolescents become depressed. Most research has suggested that this can be explained, at least in part, by examining whether and how hormonal change at puberty provokes changes in the adolescent's interpersonal world. For instance, studies of the relation between pubertal timing and depression indicate that early maturing girls experience more negative affect and internalizing problems than do their average- or later-maturing peers (Graber, Lewinsohn, Seeley, & Brooks-Gunn, 1997; Petersen, Sarigiani, & Kennedy, 1991). It has been hypothesized that this is due in part to the impact of puberty on adolescents relations with same- and opposite-sex peers (Brooks-Gunn & Reiter, 1990; Stattin & Magnusson, 1990). Research also shows that pubertal maturation may interact with negative life events and psychosocial difficulties to explain increases in depression in

160

Shelli Avenevoli and Laurence Steinberg

adolescence, particularly among girls. Brooks-Gunn and Warren (1989), for example, found that negative life events, and the interaction between negative events and estrogen levels, accounted for much more of the variance in depressed affect than did variation in estrogen levels alone. Similarly, there is evidence that risk for depressed affect is greatest when puberty and school change, a stressful event for some adolescents, occur simultaneously (Petersen et al., 1991). Other researchers have developed more complex models for the interactive effects of biology and the environment. For example, Cyranowski, Frank, Young, and Shear (2000) argued that adolescent girls with high levels of anxiety and poor coping skills are especially prone to depression when they simultaneously experience high levels of affiliative need (hypothesized to be caused by an increase in oxytocin during puberty) and threats to their sense of attachment security (associated with normative changes in parent and peer relations during adolescence). Although the biological changes of puberty have been linked to concurrent depressed affect and, in some cases, the onset of depression, there is only indirect evidence regarding their relation to the stability or continuity of depression over the course of adolescence. This may be due to the fact that changes in affective experience in early adolescence may be more strongly related to fluctuations, rather than absolute changes, in hormone levels. Indeed, the effects of hormones on mood appear to be strongest early in puberty when hormonal levels are highly variable and characterized by relatively rapid oscillations (Buchanan, Eccles, & Becker, 1992). Rapid increases in hormone levels seem to be associated with increased irritability in boys and depression in girls, but these effects lessen as hormonal levels stabilize (Buchanan et al., 1992). Thus, it is not clear how pubertal change is related to the long-term stability or recurrence of depression among adolescents. Other biological factors in adolescence have been linked to depression as well. Although findings are less consistent in studies of children and adolescents than in studies of adults, research suggests that depression is related to dysregulation of neurotransmitters, sleep problems, and other neuroendocrine processes (e.g., Dahl et al., 1992). Additionally, there is preliminary evidence that the association between these factors and depression becomes more evident across the adolescent period (Dahl & Ryan, 1996). Moreover, Kovacs and Devlin (1998) highlight that impairment in the ability to regulate negative affect, a trait that has at least some of its roots in aspects of temperament that are stable and observable long before adolescence, may be important for understanding the development of internalizing disorders and may explain the predictive validity of anxiety to depression. Individuals who manifest an internalizing disorder early in life may be at risk for depression during adolescence due to this underlying temperamental predisposition toward negative affectivity. Studies of individual differences in these biological factors may partially explain the changing manifestation of depression across the adolescent transition. Because pubertal changes affect sleep and eating patterns during these transitional years

Continuity of Depression

161

(Dahl, 1998), certain vegetative and body-related symptoms of depression may also change in salience. For example, heightened awareness of body image may lead girls to eat less than they normally would and boys to eat more (both symptoms of depression). Also, changes in the sleep-wake cycle during early adolescence (Carskadon, Vieria, & Acebo, 1993) suggest that hypersomnia may increase during the adolescent years. More research is needed to determine whether these appetitive and sleep-related changes are causative, or merely indicative, of depression during adolescence. C. COGNITIVE INFLUENCES

Numerous studies have indicated an association between depression among children and adolescents (both in the normal population and in clinical settings) and a variety of cognitive factors: negative explanatory style/depressive attributional style (e.g., Seligman et al., 1984), tendency toward negativistic interpretations (e.g., Moyal, 1977), cognitive distortions (e.g., Haley, Fine, Marriage, Moretti, & Freeman, 1985), hopelessness (e.g., McCauley, Mitchell, Burke, & Moss, 1988), and maladaptive attitudes (e.g., Nolen-Hoeksema, Girgus, & Seligman, 1986). Moreover, research has consistently shown that depressed children display more cognitive errors, endorse more negative attributions, have greater external locus of control, and have lower self-esteem than do nondepressed children (Tems, Stewart, Skinner, Hughes, & Emslie, 1993). Like biological factors, cognitive influences also act as predispositional factors that interact with exposure to stressful events to predict depression and depressed affect (Garber & Hilsman, 1992). Because adolescence is a time of dramatic change in cognitive abilities (Keating, 1990), it is of interest to determine whether changes in depression and depressed affect at adolescence are related to changes in cognitive abilities at this point in development. Despite the consistent finding that the presence of negative cognitions distinguishes depressed individuals from those who are not, negative cognitions have not been shown conclusively to contribute to the onset or stability of depression in adolescence (or, for that matter, in childhood or adulthood). Many studies comparing depressed individuals to remitted depressives have concluded that distorted cognitions, negative self-schemas, or dysfunctional attitudes are "mood-state-dependent concomitants of depressive symptoms" (Hammen, 1991). The adult literature suggests that negative cognitions normalize or dissipate entirely after recovery (Barnett & Gotlib, 1988; Dohr, Rush, & Bernstein, 1989). Similarly, a short-term longitudinal study of depressed children indicated that depressogenic cognitions decreased after inpatient treatment (Terns et al., 1993). And two studies of children showed that remitted depressives did not differ from nondepressed patients on levels of global self-worth, self-concept, and negative attributional style, although depressed children did (Asarnow & Bates, 1988; McCauley et al., 1988). In contrast, however,

162

Shelli Avenevoli and Laurence Steinberg

another study reported that, although remitted adolescents differed from currently depressed adolescents, both groups had more negative cognitions and a more negative attributional style than did adolescents with no prior history of depression (Gotlib, Lewinsohn, Seeley, Rohde, & Redner, 1993). Studies attempting to determine whether cognitive style assessed at one point in time predicts subsequent depressive symptoms have also shown mixed results. Nolen-Hoeksema, Girgus, and Seligman (1986) reported that both explanatory style and depressive symptoms were relatively stable over a 1-year period and that explanatory style predicted subsequent depression, even after controlling for initial depression scores. Seligman et al. (1984) also reported that attributional style predicted children's depressive symptoms 6 months after initial assessment. However, a comprehensive review of these and related studies suggests a predictive relationship between attributional style and depressed affect, but not clinical depression (Barnett & Gotlib, 1988). A study did find that adolescents with elevated levels of both dysphoric mood and dysfunctional thinking were at increased risk for recurrence of MDD (Lewinsohn, Allen, Seeley, & Gotlib, 1999). In sum, little is known about how developmental and individual differences in cognitive abilities at puberty, and, in particular, how the development of a more abstract, differentiated, and organized self-concept, contribute to the differential manifestation of depressive symptoms across the adolescent transition. Less is known regarding whether individual differences in tendencies toward depressogenic cognitions are stable enough to explain individual differences in the continuity of depression. In view of the prominence accorded cognitive theories of depression in current thinking about the disorder, it seems to us that this is an important area for future research. D. S O C I A L - C O N T E X T U A L INFLUENCES

Exposure to stressful life events is one of the most widely studied risk factors for depressive symptoms and disorder. Numerous studies of clinical and community samples of children and adolescents have reported a significant relation between stressful life events and depression (e.g., Williamson, Birmaher, Anderson, A1Shabbout, & Ryan, 1995). Because adolescence is a time of important changes in social context and social relationships, it is reasonable to ask whether developmental or individual differences in exposure to stress during adolescence are related to developmental or individual differences in depression. Many events associated with adolescence may be stressful for some adolescents. For instance, new capacities for autonomy and intimacy bring new challenges, including changes in family dynamics (Larson & Richards, 1991; Smetana, 1995; Steinberg, 1988), friendships (Bukowski, Gauze, Hoza, & Newcomb, 1993), and romantic relationships. Moreover, the transition from junior high school to high school proves stressful to some (Isakson & Jarvis, 1999). Other stressful events associated with the adolescent period are linked to increases in depressive affect

Continuity of Depression

163

and depressive symptoms, including problems with family members (Kaslow et al., 1994), poor grades in school (Hilsman & Garber, 1995), and low popularity (Jacobsen, Lahey, & Strauss, 1983). Combined with a genetic/biological or cognitive predisposition, negotiating these life changes may trigger depression in adolescents. While a number of factors seem to increase the risk for depression, close personal relations, good dyadic family relationships, and general social support buffer the impact of stress during adolescence (e.g., Petersen et al., 1991). Like prior examinations of continuity, studies assessing psychosocial stress as a predictor of the stability of depression have yielded mixed results. Many studies suggest that accumulation of negative life events is associated with subsequent depressed mood, syndromes, and disorder (e.g., Compas, Howell, Phares, Williams, & Giunta, 1989). Ge, Lorenz, Conger, Elder, & Simons (1994) reported that the level of depressive symptoms across a 4-year period was related to the level of life events along this same trajectory. In contrast, Lewinsohn et al. (1999) argue that stressors are more important in predicting the onset of MDD than subsequent episodes of MDD. Hammen (1991) has proposed a model of stress generation that may help explain continuity of depression over adolescence. Whereas most stress research has focused on the effects of stress on depression, Hammen proposes that stressful events may be generated by the depressed individual, which may trigger subsequent experiences of depression. Studies have shown that depressed children generate conflicts with parents and peers (e.g., Adrian & Hammen, 1993). Unfortunately, most studies of stress and depression in child and adolescent samples are crosssectional and cannot inform tests of Hammen's model. E. SUMMARY

Although the factors discussed in this section are by no means exhaustive, this section highlights some of the main genetic, biological, cognitive, and contextual influences that have been linked, or suggest links, to depression and its continuity over time. Extant studies suggest that genetic predisposition is consistently linked to recurrence of depression. Findings concerning the roles of puberty, cognitive change, and the main social stressors of adolescence are less consistent. Although studies have begun to examine how these factors and related processes explain continuities and discontinuities in depressed affect and depression, there is a need for additional research that goes beyond extant studies of factors related to the initial onset and concurrent symptoms of depression.

V. Conclusions Our review makes clear that any comprehensive study of the continuity of depression must address three separate issues. First, research on the continuity of

164

Shelli Avenevoli and Laurence Steinberg

depression must appropriately distinguish among depression as a disorder, syndrome, and symptom. The different conceptualizations of depression found in the literature on children and adolescents vary not only by definition but also in the prevalence and continuity estimates they provide. To develop a more thorough understanding of the continuity of depression, it is necessary to examine multiple conceptualizations of depression in any analysis of continuity. The examination of depression as a syndrome is particularly important for the study of continuity, because syndromes of depression can be derived empirically and, thus, are more likely to accurately reflect the depressive states experienced by children and adolescents. With this approach, researchers may discover that certain symptoms, not always associated with depressive disorder, are in fact related to depressive syndrome at different ages. For example, crying or anxiety symptoms may be identified as symptoms of depression in children but not in adolescents. In contrast to the syndrome level approach, there is less opportunity to discover developmental change when depressive disorder is the focus of attention, because depression at the level of disorder is based on criteria that are not solely data driven (Achenbach, 1991 a) and that are principally based on adult symptomatology. (However, it is important to note that depressive disorders have a strong clinical basis and should not be disregarded). In addition, the study of depression at the syndrome level allows for the examination of changes in patterns of relations among symptoms across different ages (e.g., whether the strength of the correlations among anxiety and different symptoms of depression vary with development). In addition to syndrome-level analyses, research on the continuity of individual symptoms over time is also important. Although focusing on individual symptoms does not permit the identification of syndromal patterns, examining the stability of individual symptoms over time and across developmental periods nevertheless contributes to our understanding of how elements of depressive syndromes change over time and how individuals express this underlying pathology during different periods of the life span. Previous research has addressed the issue of changing manifestation of depressive symptoms only through cross-sectional and very shortterm prospective analyses. A second criterion for comprehensive research on the continuity of depression is the examination of both homotypic and heterotypic continuity of diagnosis and symptoms. It is critical to examine both types of continuity in the course of depression because examinations that incorporate heterotypic continuity are likely to lead to the discovery of relations among symptoms not currently viewed as depressive symptoms, thereby contributing to information on the nosology of depression. Studies of continuity that focus exclusively on diagnosed depressive disorder, for example, limit estimates of continuity to observations of homotypic stability, which may greatly underestimate the extent to which depression actually is continuous over time.

Continuity of Depression

165

Finally, research on the continuity of depression m u s t be prospective, longitudinal, and d e v e l o p m e n t a l in nature (i.e., it m u s t span d e v e l o p m e n t a l transitions). W h e r e a s cross-sectional and retrospective designs can indirectly i n f o r m questions of continuity, only longitudinal designs that take into account the concepts of h o m o typic and heterotypic continuity and the multiple c o n c e p t u a l i z a t i o n s of depression can lead to a clearer u n d e r s t a n d i n g of h o w depression unfolds over time. L o n g i tudinal studies that follow the same individuals across i m p o r t a n t d e v e l o p m e n t a l transitions can directly assess how depression unfolds across d e v e l o p m e n t and w h e t h e r the manifestation of depression varies with d e v e l o p m e n t a l level. Given what we k n o w about depression and about child and adolescent d e v e l o p m e n t , the adolescent transition stands out as a particularly i m p o r t a n t period for research on the continuity of depression and other p r o b l e m s of internalization.

ACKNOWLEDGMENT Preparation of this article was supported by the John D. and Catherine T. MacArthur Foundation Research Network on Psychopathology and Development.

REFERENCES Achenbach, T. M. (1991a). The derivation of taxonomic constructs: A necessary stage in the development of developmental psychopathology. In D. Cicchetti (Ed.) Rochester symposium on developmental psychopathology: Models and integrations (Vol. 3, pp. 43-74). Hillsdale, NJ: Erlbaum. Achenbach, T. M. ( 1991b). Integrative guide for the 1991 CBCL/4-18, YSR, and TRF Profiles. Burlington, VT: University of Vermont, Department of Psychiatry. Achenbach, T. M. ( 1991c). Manual for the Child Behavior Checklist and 1991 Profile. Burlington, VT: University of Vermont, Department of Psychiatry. Achenbach, T. M. (1991d). Manual for the Teacher Report Form and 1901 Profile. Burlington, VT: University of Vermont, Department of Psychiatry. Achenbach, T. M. ( 1991e). Manu.T1for the Youth Self-Report and 1991 Profile. Burlington, VT: University of Vermont, Department of Psychiatry. Achenbach, T. M. Edelbrock, C. (1978). The classification of child psychopathology: A review and analysis of empirical efforts. Psychological Bulletin, 85, 1275-1302. Achenbach, T. M., Connors, C. K., Quay, H. C., Verhulst, E C., & Howell, C. T. (1989). Replication of empirically derived syndromes as a basis for taxonomy of child/adolescent psychopathology. Journal of Abnormal Child Psychology, 17, 299-323. Achenbach, T. M., Howell, C. T., McConaughy, S. H., & Stanger, C. (1995). Six-year predictors of problems in a national sample of children and youth: I. Cross-informant syndromes. Journal of the American Academy of Child and Adolescent Psychiatry, 34 (3): 336-347. Achenbach, T. M., Howell, C. T., Quay, H. C., & Connors, C. K. (1991 ). National survey of competencies and problems among 4- to 16-year-olds: Parents' reports for normative and clinical samples. Monographs of the Society for Research on Child Development, 56, (3, Serial No. 225). Adrian, C., & Hammen, C. (1993). Stress exposure and stress generation in children of depressed mothers. Journal of Consulting and Clinical Psychology, 61, 354-359.

166

Shelli Avenevoli and Laurence Steinberg

Alloy, L. B., Kelly, K. A., Mineka, S., & Clements, C. M. (1990). Comorbidity of anxiety and depressive disorders: A helplessness-hopelessness perspective. In J. D. Maser & C. R. Cloninger (Eds.), Comorbidity ofmood and anxiety disorders (pp. 499-543). Washington, DC: American Psychiatric Press. American Psychiatric Association (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC. American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC. Anderson, J. C., Williams, S., McGee, R., & Silva, E (1987). DSM-III disorders in preadolescent children. Archives of General Psychiatry, 44, 69-76. Angold, A. (1988). Childhood and adolescent depression: I. Epidemiological and aetiological aspects. British Journal of Psychiatry, 152, 601-617. Angold, A., Costello, E. J., & Erkanli, A. (1999). Comorbidity. Journal of Child Psychology, Psychiatry, and Allied Disciplines, 40, 57-87. Angold, A., Costello, E. J., Erkanli, A., & Worthman, C. M. (1999). Pubertal changes in hormone levels and depression in girls. Psychological Medicine, 29, 1043-1053. Angold, A., Cox, A., Predergast, M., Rutter, M., & Simonoff, E. (1994). The Child and Adolescent Psychiatric Assessment: Information Packet. Unpublished manuscript, Duke University. Asarnow, J. R., & Bates, S. (1988). Depression in child psychiatric inpatients: Cognitive and attributional patterns. Journal of Abnormal Child Psychology, 16, 601-615. Asarnow, J., Goldstein, M., Carlson, G., Perdue, S., Bates, S., & Keller, J. (1988). Childhood-onset depressive disorders: A follow-up study of rates of rehospitalization and out-of-home placement among child psychiatric inpatients. Journal of Affective Disorders, 15, 245-253. Avenevoli, S. (1998). The continuity of depression from childhood to adolescence (Doctoral dissertation, Temple University, 1998). Dissertation Abstracts International, 59, 5596. Avenevoli, S., & Silberg, J. L. (1999). Developmental Differences in Depressive Symptomatology from Childhood to Adolescence. Poster presented at the Biennial Meeting of the Society for Research on Child Development, Albuquerque, NM. Barnett, E A., & Gotlib, I. H. (1988). Psychosocial functioning and depression: Distinguishing among antecedents, concomitants, and consequences. Psychological Bulletin, 104, 97-126. Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New York: International Universities Press. Berenbaum, S. A., Kerns, K., & Taylor, M. A. (1990). A neurological perspective on comorbid disorders. In J. D. Maser & C. R. Cloninger (Eds.), Comorbidity ofmood and anxiety disorders (pp. 499-543). Washington, DC: American Psychiatric Press. Bernstein, G., & Garfinkel, B. (1986). School phobia: The overlap of affective and anxiety disorders. Journal of the American Academy of Child Psychiatry, 25, 235-241. Birmaher, B., Ryan, N. D., Williamson, D. E., Brent, D. A., Kaufman, J., Dahl, R. E., Perel, J., & Nelson, B. (1996). Childhood and adolescent depression: A review of the past 10 years. Part I. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 1427-1439. Blacker, C. V. R., & Clare, A. W. (1987). Depressive disorder in primary care. British Journal of Psychiatry, 150, 737-751. Bowlby, J. (1973). Attachment and Loss, Vol. 2: Separation. NY: Basic Books. Brady, E. U., & Kendall, E C. (1992). Comorbidity of anxiety and depression in children and adolescents. Psychological Bulletin, 111, 244-255. Brooks-Gunn, J., & Reiter, E. (1990). The role of pubertal processes. In S. Feldman & G. Elliot (Eds.), At the threshold: The developing adolescent (pp. 16-23). Cambridge, MA: Harvard University Press. Brooks-Gunn, J., & Warren, M. E (1989). Biological and social contributions to negative affect in young adolescent girls. Child Development, 60, 40-55.

Continuity of Depression

167

Buchanan, C. M., Eccles, J. S., • Becker, J. B. (1992). Are adolescents victims of raging hormones: Evidence for activational effects of hormones on moods and behavior at adolescence. Psychological Bulletin, 111, 62-107. Bukowski, W., Gauze, C., Hoza, B., & Newcomb, A. (1993). Differences and consistency between same-sex and other-sex peer relationships during early adolescence. Developmental Psychology, 29, 255-263. Carlson, G. A., & Garber, J. (1986). Developmental issues in the classification of depression in children. In M. Rutter, C. E. Izard, & P. B. Read (Eds.), Depression in young people: Developmental and clinical perspectives (pp. 399-434). New York: Guilford Press. Carlson, G. A., & Kashani, J. H. (1988). Phenomenology of major depression from childhood through adulthood: Analysis of three studies. American Journal of Psychiatry, 145, 1222-1225. Carskadon, M., Vieria, C., & Acebo, C. (1993). Association between puberty and delayed phase preference. Sleep, 16, 258-262. Caspi, A. (1998). Personality development across the life course. In N. Eisenberg (Vol. Ed.), Social, emotional, and personality development: Volume 3. Handbook of Child Psychology (W. Damon, Ed.). New York: Wiley. Caspi, A., &Bem, D. (1990). Personality continuity and change across the life course. In L. A. Pervin (Ed.), Handbook of personality theory and research (pp. 549-575). New York: Guilford Press. Chorpita, B. F., Albano, A. M., & Barlow, D. H. (1997). The structure of negative emotions in a clinical sample of children and adolescents. Journal of Abnormal Psychology, 107, 74-85. Cicchetti, D., & Rogosch, F. (1996). Equifinality and multifinality in developmental psychopathology. Development and Psychopathology, 8, 597-600. Cicchetti, D., & Schneider-Rosen, K. (1986). An organizational approach to childhood depression. In M. Rutter, C. Izard, & P. Read (Eds.), Depression in young people: Developmental and clinical perspectives (pp. 71-134). NY: Guilford Press. Clark, D. A., Beck, A. T., & Brown, G. (1989). Cognitive mediation in general psychiatric outpatients: A test of the content-specificity hypothesis. Journal of Personality and Social Psychology, 56, 958-964. Clark, D. A., Steer, R. A., & Beck, A. T. (1994). Common and specific dimensions of self-reported anxiety and depression: Implications for the cognitive and tripartite models. Journal of Abnormal Psychology, 103, 634-654. Clark, L. A. (1989). The anxiety and depressive disorders: Descriptive psychopathology and differential diagnosis. In P. C. Kendall & D. Watson (Eds.), Anxiety and depression: Distinctive and overlapping features (pp. 83-129). San Diego: Academic Press. Cloninger, C. R., Martin, R. L., Guze, S. B., & Clayton, E J. (1990). The empirical structure of psychiatric comorbidity and its theoretical significance. In J. D. Maser & C. R. Cloninger (Eds.), Comorbidity of mood and anxiety disorders (pp. 439-462). Washington, DC: American Psychiatric Press. Cole, D. A., Martin, J. M., Powers, B., & Truglio, R. (1996). Modeling causal relations between academic and social competence and depression: A multitrait-multimethod longitudinal study of children. Journal of Abnormal Psychology, 105, 258-270. Cole, D. A., Peeke, L. G., Martin, J. M., Truglio, R., & Seroczynski, A. D. (1998). A longitudinal look at the relation between depression and anxiety in children and adolescents. Journal of Consulting and Clinical Psychology, 66, 451-460. Cole, D. A., Truglio, R., & Peeke, L. (1997). Relation between symptoms of anxiety and depression in children: A multitrait-multimethod-multigroup assessment. Journal of Consulting and Clinical Psychology, 65, 11O- 119. Compas, B. E., Ey, S., & Grant, K. E. (1993). Taxonomy, assessment, and diagnosis of depression during adolescence. Psychological Bulletin, 114, 323-344. Compas, B. E., Howell, D. C., Phares, V., Williams, R. A., & Giunta, C. (1989). Risk factors for

168

Shelli Avenevoli and Laurence Steinberg

emotional/behavioral problems in young adolescents: A prospective analysis of adolescent and parental stress and symptoms. Journal of Consulting and Clinical Psychology, 57, 732-740. Cyranowski, J. M., Frank, E., Young, E., & Shear, K. M. (2000). Adolescent onset of the gender difference in lifetime rates of major depression: A theoretical model.Archives of General Psychiatry, 57, 21-27. Dahl, R. E. (1998). The regulation of sleep and arousal: Development and psychopathology. In M. E. Hertzig & E. A. Farber (Eds.), Annual progress in child psychiatry and child development (Vol. 7, pp. 3-28). Bristol, PA: Brunner/Mazel. Dahl, R. E., & Ryan, N. D. (1996). The psychobiology of adolescent depression. In D. Ciccheti & S. L. Toth (Eds.), Rochester Symposium on developmental psychopathology, Vol. 7: Adolescence: Opportunities and challenges (pp. 197-232). Rochester NY: University of Rochester Press. Dahl, R. E., Ryan, N. D., Williamson, D. E., Ambrosini, P. J., Rabinovich, H., Novacenko, H., Nelson, B., & Puig-Antich, J. (1992). Regulation of sleep and growth hormone in adolescent depression. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 615-621. Devine, D., Kempton, T., & Forehand, R. (1994). Adolescent depressed mood and young adult functioning: A longitudinal study. Journal of Abnormal Child Psychology, 22, 629-640. Dohr, K. B., Rush, J. A., & Bernstein, I. H. (1989). Cognitive biases and depression. Journal of Abnormal Psychology, 98, 263-267. Dumas, J. E., Neese, D. E., Prinz, R. J., & Blechman, E. A. (1996). Short-term stability of aggression, peer rejection, and depressive symptoms in middle childhood. Journal of Abnormal Child Psychology, 24, 105-119. Eysenck, H. J. (1960). The structure of human personality (2nd ed.). London: Methuen. Eysenck, H. J. (1986). A critique of contemporary classification and diagnosis. In T. Millon & G. L. Klerman (Eds.), Contemporary directions in psychopathology: Toward the DSM-IV (pp. 73-98). New York: Guilford Press. Fergusson, D. M., Horwood, L. J., & Lynskey, M. T. (1995). The stability of disruptive childhood behaviors. Journal of Abnormal Child Psychology, 23, 379-396. Finch, A. J., Lipovsky, J. A., & Casat, C. D. (1989). Anxiety and depression in children and adolescents: Negative affectivity or separate constructs? In P. C. Kendall & D. Watson (Eds.), Anxiety and depression: Distinctive and overlapping features (pp. 171-202). San Diego: Academic Press. Finch, A. J., Saylor, C. E, Edwards, G. L., & Mclntosh, J. A. (1987). Children's Depression Inventory: Reliability over repeated administrations. Journal of Clinical Child Psychology, 16, 339-341. Fleming, J. E., & Offord, D. R. (1990). Epidemiology of childhood depressive disorders: A critical review. Journal of the American Academy of Child and Adolescent Psychiatry, 29 (4), 571-580. Garber, J. (1984). The developmental progression of depression in female children. New Directions for Child Development, 26, (Dec) 29-58. Garber, J., & Hilsman, R. (1992). Cognition, stress, and depression in children and adolescents. Child and Adolescent Psychiatric Clinics of North America, 1, 129-167. Garber, J., Kriss, M. R., Koch, M., & Lindholm, L. (1988). Recurrent depression in adolescents: A follow-up study. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 4954. Garrison, C. Z., Jackson, K. L., Marsteller, E, McKeown, R., & Addy, C. (1990). A longitudinal study of depressive symtomatology in young adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 29, 581-585. Garrison, C. A., Schluchter, M. D., Schoenbach, V. J., & Kaplan, B. K. (1989). Epidemiology of depressive symtoms in young adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 28, 343-351. Ge, X., Lorenz, E O., Conger, R. D., Elder, G. H., & Simons, R. L. (1994). Trajectories of stressful life events and depressive symptoms during adolescence. Developmental Psychology, 30, 467-483.

Continuity of Depression

169

Gotlib, H., & Hammen, C. L. (1992). Psychological aspects of depression: Toward a cognitiveinterpersonal integration. Chichester, England: Wiley. Gotlib, I. H., Lewinsohn, E M., Seeley, J. R., Rohde, E, & Redner, J. E. (1993). Negative cognitions and attributional style in depressed adolescents: An examination of stability and specificity. Journal of Abnormal Psychology, 102, 607-615. Graber, J. A., Lewinsohn, E M., Seeley, J. R., & Brooks-Gunn, J. (1997). Is psychopathology associated with the timing of pubertal development? Journal of the American Academy of Child andAdolescent Psychiatry, 36, 1768-1776. Greenberg, M. S., Vasquez, C. V., & Alloy, L. B. (1988). Depression versus anxiety: Differences in self- and other-schemata. In L. B. Alloy (Ed.), Cognitive processes in depression (109-142). New York: Guilford Press. Gurley, D., Cohen, P., Pine, D. S., & Brook, J. (1996). Discriminating depression and anxiety in youth: A role for diagnostic criteria. Journal of Affective Disorders, 39, 191-200. Haley, G. M., Fine, S., Marriage, K., Moretti, M. M., & Freeman, R. J. (1985). Cognitive bias and depression in psychiatrically disturbed children and adolescents. Journal of Consulting and Clinical Psychology, 53, 535-537. Hammen, C. (1991). Generation of stress in the course of unipolar depression. Journal of Abnormal Psychology, 100, 555-561. Hammen, C., & Compas, B. E. (1994). Unmasking masked depression in children and adolescents: The problem of comorbidity. Clinical Psychology Review, 14, 585-603. Hammen, C., & Rudolph, K. D. (1996). Childhood depression. In E. J. Mash & R. A. Barkley (Eds.), Child psychopathology (pp. 153-195). New York: Guilford Press. Hilsman, R., & Garber, J. (1995). A test of the cognitive diathesis-stress model of depression in children: Academic stressors, attributional style, perceived competence, and control. Journal of Personality and Social Psychology, 69, 370-380. Isakson, K., & Jarvis, E (1999). The adjustment of adolescents during the transition into high school: A short-term longitudinal study. Journal of Youth and Adolescence, 28, 1-26. Jacobsen, R. H., Lahey, B. B., & Strauss, C. C. (1983). Correlates of depressed mood in normal children. Journal of Abnormal Child Psychology, 11, 29-39. Jolly, J. B., Dyck, M. J., Kramer, T. A., & Wherry, J. N. (1994). Integration of positive and negative affectivity and cognitive content-specificity: Improved discrimination of anxious and depressive symptoms. Journal of Abnormal Psychology, 103, 544-552. Kagan, J. (1980). Perspectives on continuity. In O. Brim & J. Kagan (Eds.), Constancy and change in human development. Cambridge MA: Harvard University Press. Kagan, J., & Moss, H. A. (1962). Birth to maturity. New York: Wiley. Kandel, D. B., & Davies, M. (1986). Adult sequelae of adolescent depressive symptoms. Archives of General Psychiatry, 43, 255-262. Kashani, J. H., Beck, N. C., Hoeper, E. W., Fallahi, C., Corcoran, C. M., McAllister, J. A., Rosenberg, T. K., & Reid, J. C. (1987). Psychiatric disorders in a community sample of adolescents. American Journal of Psychiatry, 144, 584-589. Kashani, J. H., & Carlson, G. A. (1987). Seriously depressed preschoolers. American Journal of Psychiatry, 144, 348-350. Kashani, J. H., Rosenberg, T. K., & Reid, J. C. (1989). Developmental perspectives in child and adolescent depressive symptoms in a community sample. American Journal of Psychiatry, 146, 871-875. Kashani, J. H., & Simonds, J. E (1979). The incidence of depression in children. American Journal of Psychiatry, 136, 1203-1205. Kaslow, N. J., Deering, C. G., & Racusin, G. R. (1994). Depressed children and their families. Clinical Psychology Review, 14, 39-59. Keating, D. (1990). Adolescent thinking. In S. S. Feldman & G. R. Elliot (Eds.), At the threshold: The developing adolescent (pp. 54-89). Cambridge, MA: Harvard University Press.

170

Shelli Avenevoli and Laurence Steinberg

King, N. J., Ollendick, T. H., & Gullone, E. (1991). Negative affectivity in children and adolescents: Relations between anxiety and depression. Clinical Psychology Review, 11, 441-459. Kovacs, M. (1990). Comorbid anxiety disorders in childhood-onset depressions. In J. D. Maser & C. R. Cloninger (Eds.), Comorbidity of mood and anxiety disorders (pp. 272-281). Washington, DC: American Psychiatric Press. Kovacs, M. (1996). Presentation and course of major depressive disorder during childhood and later years of the lifespan. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 705-715. Kovacs, M., & Devlin, B. (1998). Internalizing disorders in childhood. Journal of Child Psychology, Psychiatry, and Allied Disciplines, 39, 47-63. Kovacs, M., Feinberg, T. C., Crouse-Novak, M. A., Paulauskas, S. L., Pollack, M., & Finklestein, R. (1984). Depressive disorders in childhood: II. A longitudinal study of the risk for a subsequent major depression. Archives of General Psychiatry, 41, 643-649. Kovacs, M., & Gatsonis, C. (1989). Stability and change in childhood-onset depressive disorders: Longitudinal course as a diagnostic validator. In L. N. Robins & J. E. Barrett (Eds.) The validity of psychiatric diagnosis (pp. 57-75), New York: Raven Press. Kovacs, M., Gatsonis, C., Paulauskas, S. L., & Richards, C. (1989). Depressive disorders in childhood: IV. A longitudinal study of comorbidity with and risk for anxiety disorders. Archives of General Psychiatry, 46, 776-782. Larson, R., & Richards, M. H. (1991). Daily companionship in late childhood and early adolescence: Changing developmental contexts. Child Development, 62, 284-300. Leitenberg, H., Yost, L. W., & Carroll-Wilson, M. (1986). Negative cognitive errors in children: Questionnaire development, normative data, and comparisons between children with and without selfreported symptoms of depression, low self-esteem, and evaluation anxiety. Journal of Consulting and Clinical Psychology, 54, 528-536. Lewinsohn, P. M., Allen, N. B., Seeley, J. R., & Gotlib, I. H. (1999). First onset versus recurrence of depression: Differential processes of psychosocial risk. Journal of Abnormal Psychology, 108, 483-489. Lewinsohn, P. M., Hops, H., Roberts, R. E., Seeley, J. R., & Andrews, J. A. (1993). Adolescent psychopathology: I. Prevalence and incidence of depression and other DSM-H-R disorders in high school students. Journal of Abnormal Psychology, 102, 133-144. Lewinsohn, P. M., Roberts, R. E., Seeley, J. R., Rohde, P., Gotlib, I. H., & Hops, H. (1994). Adolescent psychopathology: II. Psychosocial risk factors for depression. Journal of Abnormal Psychology, 103, 302-315. Lewinsohn, P. M., Rohde, P., & Seeley, J. R. (1998). Major depressive disorder in older adolescents: Prevalence, risk factors, and clinical implications. Clinical Psychology Review, 18, 765-794. Lewinsohn, P. M., Rohde, P., Seeley, J. R., & Fischer, S. A. (1993). Age-cohort changes in the lifetime occurrence of depression and other mental disorders. Journal of Abnormal Psychology, 102, 110-120. Lonigan, C. J., Carey, M. P., & Finch, A. J., Jr. (1994). Anxiety and depression in children and adolescents: Negative affectivity and the utility of self-reports. Journal of Consulting and Clinical Psychology, 62, 1000-1008. McCauley, E., Mitchell, J. R., Burke, P., & Moss, S. (1988). Cognitive attributes of depression in children and adolescents. Journal of Consulting and Clinical Psychology, 56, 903-908. McConaughy, S. H., Stanger, C., & Achenbach, T. M. (1992). Three-year course of behavioral/ emotional problems in a national sample of 4- to 16-year-olds: I. Agreement among informants. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 932-940. McConville, B. J., Boag, L. C., & Purohit, A. P. (1973). Three types of childhood depression. Canadian Psychiatric Association Journal 18, 133-138. McGee, R., Feehan, M., Williams, S., & Anderson, J. (1992). DSM-III disorders from age 11 to age 15 years. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 50-59.

Continuity of Depression

171

McGee, R., Feehan, M., Williams, S., Partridge, F, Silva, E, & Kelly, J. (1990). DSM-IIIdisorders in a large sample of adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 29, 611-619. Merikangas, K. R., & Angst, J. (1995). The challenge of depressive disorders in adolescence. In M. Rutter (Ed.), Psychosocial disturbances in young people: Challenges for prevention (pp. 131-165). New York: Cambridge University Press. Merikangas, K. R., Weissman, M. M., Prusoff, B. A., & John, K. (1988). Assortative mating and affective disorders: Psychopathology in offspring. Psychiatry, 51, 48-57. Meyer, J. M., Silberg, J. L., Simonoff, E., Kendler, K. S., & Hewitt, J. K. (1996). The Virginia TwinFamily Study of Adolescent Behavioral Development: Assessing sample bases in demographic correlates of psychopathology. Psychological Medicine, 26, 1-15. Mitchell, J., McCauley, E., Burke, P. M., & Moss, S. J. (1988). Phenomenology of depression in children and adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 12-20. Moffit, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674-701. Moyal, B. R. (1977). Locus of control, self-esteem, stimulus appraisal, and depressive symptoms in children. Journal of Consulting and Clinical Psychology, 45, 951-952. Neuman, R. J., Geller, B., Rice, J. E, & Todd, R. D. (1997). Increased prevalence and earlier onset of mood disorders among relatives of prepubertal versus adult probands. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 466-473. Nolen-Hoeksema, S., & Girgus, J. S. (1994). The emergence of gender differences in depression during adolescence. Psychological Bulletin, 115, 424-443. Nolen-Hoeksema, S., Girgus, J. S., & Seligman, M. E. P. (1986). Learned helplessness in children: A longitudinal study of depression, achievement, and explanatory style. Journal of Personality and Social Psychology, 51, 435-442. Nolen-Hoeksema, S., Girgus, J. S., & Seligman, M. E. E (1992). Predictors and consequences of childhood depressive symptoms: A 5-year longitudinal study. Journal of Abnormal Psychology, 101, 405-422. Nottelmann, E. D., Inoff-Germain, G., Susman, E. J., & Chrousos, G. E (1990). Hormones and behavior at puberty. In J. Bancroft & J. M. Reinisch (Eds.), Adolescence and puberty (pp. 88-123). New York: Oxford University Press. O'Connor, T. G., Neiderhiser, J. M., Neiderhiser, D. R., Hetherington, E. M., & Plomin, R. (1998). Genetic contributions to continuity, change, and co-occurrence of antisocial and depressive symptoms in adolescence. Journal of Child Psychology and Psychiatry, 39, 323-336. Patterson, G. R. (1992). Developmental changes in antisocial behavior. In R. D. Peters, R. J. McMahan, & V. L. Quinsey (Eds.), Aggression and violence throughout the lifespan (pp. 52-82). Newbury Park, CA: Sage. Patterson, G. R., DeBaryshe, B. D., & Ramsey, E. (1989). A developmental perspective on antisocial behavior. American Psychologist, 44, 329-335. Petersen, A. C., Compas, B. E., & Brooks-Gunn, J. (1992). Depression in adolescence: Current knowledge, research directions, and implications for programs and policy. Washington, DC: Carnegie Council on Adolescent Development. Petersen, A. C., Compas, B. E., Brooks-Gunn, J., Stemmler, M., Eye, S., & Grant, K. E. (1993). Depression in adolescence. American Psychologist, 48, 155-168. Petersen, A. C., Sarigiani, P. A., & Kennedy, R. E. (1991). Adolescent depression: Why more girls? Journal of Youth and Adolescence, 20, 247-271. Puig-Antich, J., Lukens, E., Davies, M., Goetz, D., Brennan-Quattrock, J., & Todak, G. (1985). Psychosocial functioning in prepubertal major depressive disorders. I, II. Archives of General Psychiatry, 44, 81-89.

172

Shelli Avenevoli and Laurence Steinberg

Quay, H. C. (1986). Classification. In. H. C. Quay & J. S. Werry (Eds.), Psychopathological disorders of childhood. New York: Wiley. Rao, U., Ryan, N. D., Birmaher, B., Dahl, R. E., Williamson, D. E., Kaufman, J., Rao, R., & Nelson, B., et al. (1995). Unipolar depression in adolescents: Clinical outcome in adulthood. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 566-578. Rohde, E, Lewinsohn, P. M., & Seeley, J. R. (1991). Comorbidity of unipolar depression: II. Comorbidity with other mental disorders in adolescents and adults. Journal of Abnormal Psychology, 54, 653-660. Rutter, M. (1986a). Depressive feelings, cognitions, and disorders: A research postscript. In M. Rutter, C. E. Izard, & E B. Read (Eds.), Depression in young people: Developmental and clinical perspectives (pp. 491-519). New York: Guilford Press. Rutter, M. (1986b). The developmental psychopathology of depression: Issues and perspectives. In M. Rutter, C. E. Izard, & E B. Read (Eds.), Depression in young people: Developmental and clinical perspectives (pp. 3-30). New York: Guilford Press. Ryan, N. D., Puig-Antich, J., Ambrosini, P., Rabinovich, H., Robinson, D., Nelson, B., Iyengar, S., & Twomey, J. (1987). The clinical picture of major depression in children and adolescents. Archives of General Psychiatry, 44, 854-861. Saarni, C., Mumme, D. L., & Campos, J. J. (1998). Emotional development: Action, communication, and understanding. In N. Eisenberg (Vol. Ed.), Social, emotional, and personality development: Volume 3. Handbook of Child Psychology (W. Damon, Ed.). New York: Wiley. Seligman, M. E. E, Peterson, C., Kaslow, N. J., Tanenbaum, R. L., Alloy, L. B., & Abramson, L. Y. (1984). Attributional style and depressive symptoms among children. Journal of Abnormal Psychology, 93, 235-238. Silberg, J., Pickles, A., Rutter, M., Hewitt, J., Simonoff, E., Maes, H., Carbonneau, R., Murrelle, L., Foley, D., & Eaves, L. (1999). The influence of genetic factors and life stress on depression among adolescent girls. Archives of General Psychiatry, 56, 225-232. Smetana, J. G. (1995). Parenting styles and conceptions of parental authority during adolescence. Child Development, 66, 299-316. Spitzer, R. L., Endicott, J., & Robins, E. (1978a). Research diagnostic criteria: Rationale and reliability. Archives of General Psychiatry, 35, 773-782. Spitzer, R. L., Endicott, J., & Robins, E. (1978b). Research diagnostic criteria for a selected group of functional disorders (3rd ed.). New York: New York State Psychiatric Institute. Sroufe, A., & Rutter, M. (1984). The domain of developmental psychopathology. Child Development, 55, 17-29. Stark, K. D., Reynolds, W. M., & Kaslow, N. J. (1987). A comparison of the relative efficacy of self-control therapy and behavioral problem-solving therapy for depression in children. Journal of Abnormal Child Psychology, 15, 91-113. Stattin, H., & Magnusson, D. (1990). Pubertal maturation in female development. Hillsdale, NJ: Erlbaum. Steinberg, L. (1988). Reciprocal relation between parent-child distance and pubertal maturation. Developmental Psychology, 24, 122-128. Strauss, C. C., Forehand, R., Frame, C., & Smith, K. (1984). Characteristics of children with extreme scores on the Children's Depression Inventory. Journal of Clinical Child Psychiatry, 13, 227231. Strauss, C. C., Last, C. G., Hersen, M., & Kazdin, A. E. (1988). Association between anxiety and depression in children and adolescents with anxiety disorders. Journal of Abnormal Child Psychology, 16, 57-68. Suomi, S. J. (1991). Adolescent depression and depressive symptoms: Insights from longitudinal studies with rhesus monkeys. Journal of Youth and Adolescence, 20, 273-287. Susman, E. J., Nottelmann, E. D., Inoff-Germain, G. E., Dorn, L. D., Cutler, G. B., Loriaus, D. L.,

Continuity of Depression

173

& Chrousos, G. E (1985). The relation of relative hormonal levels and physical development and socio-emotional behavior in young adolescents. Journal of Youth and Adolescence, 14, 245-264. Tems, C. L., Stewart, S. M., Skinner, J. R., Hughes, C. W., & Emslie, G. (1993). Cognitive distortions in depressed children and adolescents: Are they state dependent or traitlike? Journal of Consulting and Clinical Psychology, 22, 316-326. Verhulst, F. C., & van der Ende, J. (1992). Six-year developmental course of internalizing and externalizing problem behaviors. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 924-930. Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversive emotional states. Psychological Bulletin, 96, 465-490. Watson, D., & Kendall, E C. (1989). Common and differentiating features of anxiety and depression: Current findings and future directions. In E C. Kendall and D. Watson (Eds.), Anxiety and depression: Distinctive and overlapping features (pp. 493-508). San Diego: Academic Press. Weiner, B., & Graham, S. (1988). An attributional approach to emotional development. In C. E. Izard, J. Kagan, & R. B. Zajonc (Eds.), Emotions, cognitions, and behavior (pp. 167-191). New York: Cambridge University Press. Weiss, B., Weisz, J. R., Politano, M., Carey, M., Nelson, W. M., & Finch, A. J. (1991). Developmental differences in the factor structure of the Children's Depression Inventory. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 3, 38-45. Weiss, B., Weisz, J. R., Politano, M., Carey, M., Nelson, W. M., & Finch, A. J. (1992). Relations among self-reported depressive symptoms in clinic-referred children versus adolescents. Journal of Abnormal Psychology, 101, 391-397. Weissman, M. M. (1990). Evidence for comorbidity of anxiety and depression: Family and genetic studies of children. In M. J. Maser & C. R. Cloninger (Eds.), Comorbidity of mood and anxiety disorders (pp. 349-365). Washington, DC: American Psychiatric Press. Weissman, M. M., Gammon, G. D., John, K., Merikangas, K. R., Warner, V., Prusoff, B. A., & Sholomskas, D. (1987). Children of depressed parents: Increased psychopathology and early onset of major depression. Archives of General Psychiatry, 44, 847-853. Williamson, D. E., Birmaher, B., Anderson, B. P., A1-Shabbout, M., & Ryan, N. D. (1995). Stressful life events in depressed adolescents: The role of dependent events during the depressive episode. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 591-598. Zatz, S., & Chassin, L. (1985). Cognitions of test-anxious children under naturalistic test-taking situations. Journal of Consulting and Clinical Psychology, 53, 393-401.

This Page Intentionally Left Blank

THE TIME OF OUR LIVES" SELF-CONTINUITY IN NATIVE AND NON-NATIVE YOUTH

Michael J. Chandler DEPARTMENT OF PSYCHOLOGY THE UNIVERSITY OF BRITISH COLUMBIA VANCOUVER, BRITISH COLUMBIA V6T lZ4, CANADA

I. INTRODUCTION II. NARRATIVITY AND ESSENTIALISM WRIT LARGE A. THE NARRATIVIST-ESSENTIALIST DICHOTOMY: A PRIMER B. THE PARADOX OF PERSONAL PERSISTENCE AND CHANGE C. ESSENTIALISTS ARE FROM MARS, NARRATOLOGISTS ARE FROM VENUS: A HIGH-ALTITUDE OVERVIEW III. THE ESSENTIALIST TRADITION: STRICT ESSENTIALISM, NOMINAL ESSENTIALISM, NEO-ESSENTIALISM, AND ANY OTHER ESSENTIALISM YOU CAN THINK OF A. WHO WERE ALL OF THESE CLASSICAL ESSENTIALISTS ANYWAY, AND WHAT WERE THEY AFTER? B. IDENTITY, SURVIVAL, AND PERSISTENCE C. PSYCHOLOGICAL ESSENTIALISM AS A ONE-TRICK PONY D. THE DOWNSIDE OF ESSENTIALISM, PSYCHOLOGICAL AND OTHERWISE IV. THE NARRATIVE TRADITION AS ANOTHER SINGULAR POSSIBILITY A. AUTHORS LIVING AND DEAD B. GOOD THINGS TO SAY ABOUT NARRATIVITY C. NARRATIVITY PLAYS TO CRITICAL REVIEWS D. NARRATIVE THIS, ESSENTIAL THAT V. A THEORY IN SEARCH OF EVIDENCE A. A DATA SUMMARY B. IN SEARCH OF A METHOD: A BILDUNGSROMAN FOR EVERY OCCASION C. A TYPOLOGY OF ALTERNATIVE SELF-CONTINUITY WARRANTS VI. REAL EVIDENCE A. FALSE STARTS: THE EARLY TOOL-USING STAGE B. GETTING FROM SELF-CONTINUITY TO YOUTH SUICIDE C. CULTURE AS A SET-POINT IN THE CHOICE BETWEEN NARRATIVIST AND ESSENTIALIST SELF-CONTINUITY WARRANTING PRACTICES

175 ADVANCESIN CHILDDEVELOPMENT AND BEHAVIOR,VOL.28

Copyright 9 2001 by AcademicPress. All rightsof reproductionin any formreserved. 0065-2407/01 $35.00

Michael J. Chandler

176

D. CULTURAL CONTINUITY AS A PROTECTIVE FACTOR AGAINST SUICIDE IN FIRST NATIONS YOUTH VII. SUMMARY OF CONCLUSIONS: FIVE EASY PIECES REFERENCES

There is but one truly serious philosophical problem and that is suicide. Judging whether life is or is not worth living amounts to answering the fundamental question of philosophy. u Albert Camus, The Myth of Sisyphus (1955)

I. Introduction The account that follows ends as a progress report on an ongoing program of research--a study sequence currently aimed at searching out developmental similarities and differences in the ways that Canada's culturally mainstream and Native (or "First Nations") youth undertake to sustain the necessary conviction that, despite overwhelming evidence of ongoing personal change, they nevertheless go on being the selfsame, numerically identical individuals through thick and thin. Although these data are meant to contribute to a better understanding of what is common, and what is not, about the process of identity formation in these different age and cultural groups and to manage, in the process, to afford certain insights about the largely intractable problem of youth suicide, they are nevertheless a poor place to begin. Before undertaking to make the case that Native and non-Native youth ordinarily exploit different developmental pathways toward a mature sense of personal persistence and before even trying to unpack the key claim that, although rank-and-file Canadian youth tend to be closet essentialists, aboriginal youth are by and large some sort of "folk narratologists"--before any of that--it is important to first clarify a whole raft of other heady and sometimes heartstopping things. The very least of these concern exactly who and what were measured and how. Much harder than that is a range of conceptually prior questions of the following sort. What, after all, is personal persistence and why should we (why should anybody) care? How, in the fullness of time, could anything (ourselves included) ever be understood to be "identical" with itself? How high would the personal and social costs be if we tried and failed to imagine earlier and later installments of ourselves as other than instances of one selfsame and continuous person? Just how long is this list of unanswered questions anyway?

Self-Continuity in Youth

177

What follows (by way of an attempt to address at least some of these questions) comes in two large parts. The first of these (the remedial-reading part) pilfers shamelessly from intellectual history, searching for alternative solutions to the perennial problem of personal persistence, and comes up, though not entirely empty handed, at least decidedly light. By way of a quick preview, and just in the hope that it may see you through a lot of potentially aggravating details, the alternatives that are uncovered boil down to just two. You get to suppose, as will be argued, that you go on being yourself either (a) because you are persuaded that you have successfully ferreted out some essential part of yourself that has somehow hidden out from the ravages of time or, failing that, (b) because, having abandoned as foolhardy all faint hopes of turning up any such timeless or preduring parts, your struggles toward otherwise gluing together the admittedly different time-slices of your life just happen to make such a compelling story that it salvages your identity in the telling. Darkly, as things turn out, if you fail to come up with (or if your culture fails to provide you with) one or the other of these identity preserving heuristics, then suicide, as I will go on to describe, paradoxically becomes a live option. By way of a trailside warning, getting to identity development and youth suicide by way of the death of the author and the demise of classical essentialism involves a path that is often torturous and sometimes a bit macabre, and so your forbearance-a whole sheaf of pages worth of your forbearance--is sought in advance. As if that were not enough, the later-arriving parts of this account, though hopefully less otherworldly and generally built much closer to the ground, end up repeating some of the same toplofty things about essentialism and narrativism considered earlier, but work to say them more concretely, and about real young people, held out as representative of different age groups and "cultures" and mental-health statuses. In all of these particular ways, however, I hope to make the same point: that developing some workable way of imaginatively surviving timemsome culturally sanctioned means of warranting a sense of personal continuitymis usual, developmentally orderly, and life-affirming, and failing at this task is just the opposite.

II. Narrativity and Essentialism Writ Large A. THE N A R R A T I V I S T - E S S E N T I A L I S T DICHOTOMY: A PRIMER

Before attempting to conjure what amounts to a 2000-year-old witch's brew of conflicting opinion about what might count as appropriate criteria for individual identity and personal persistence into something more closely resembling what will be elaborated here as a dichotomy between essentialistic solutions to the problem of self-continuity and their opposite narrativistic number, something first needs to be said about why personal persistence is so regularly imagined to matter. The

178

Michael J. Chandler

reason, in short, is this: being seen to remain selfsame across the various phases of our temporal existence is simply a constitutive condition for being recognized as any sort of person at all (Chandler, 2001; Chandler, Lalonde, & Sokol, 2001; Lewis & Ferrari, 2001). Not only is personal persistence definitional of selves, but what Flanagan (1996, p. 65) has called our self-imposed "one self to a customer rule" is such a fundamental part of our legal, political, and theological conceptions of moral responsibility (Rorty, 1987), and so necessary a feature of our ordinary commonsense conception of personhood, that any account of selves that did not make adequate provision for somehow understanding each of us as necessarily continuous in time would strike us as fundamentally nonsensical (Luckman, 1979). All of this is thought to follow for the reasons that if we could not count (or reidentify) one another as the same continuous and numerically identical individuals, then not only would much of formal logic be left in tatters, but social life as we understand it would cease to exist. The concept of moral duty would be emptied of meaning (Rorty, 1973), grounds would no longer exist for allocating legal responsibility and liability (Whittaker, 1992), and all prospects of looking forward to our own just desserts (Unger, 1975, pp. 55-59) would fly out the same window without some way of counting the earlier and later manifestations of a life as belonging "timelessly" to one and the same continuant (van Inwagen, 1990). How, for example, could there be a heaven or hell, where those with a history of good and evil are meant to languish, if it were not possible to describe ways in which each of us owns our past (Flanagan, 1996, p. 65)? How could St. Peter ever hope to resolve the accounting nightmare that would result if the number of different souls lined up at the Pearly Gates were not automatically equal to the number of people born into the world? For all of these reasons, and more, personal persistence is ordinarily (perhaps even universally) understood to be an immanent providence at work in the whole of human affairs (Shotter, 1984, pp. 37-40), and thus a necessary prerequisite for any recognizably decent human social world. As such, the requirement that individuals possess what Taylor (1991, p. 27) has called a bare reflective awareness of themselves as continuing persons is not simply an elective "feature" of selves, but a necessary and "constitutive condition" of their coming into being (Habermas, 1985). On these and related grounds, a very long list of contemporary philosophers (e.g., Cassirer, 1923; Harr6, 1980; Hirsch, 1976; MacIntyre, 1977; Parfit, 1971; Perry, 1976; Rorty, 1976; Taylor, 1991; Wiggins, 1980), a similar complement of touchstone psychological theorists (e.g., Erikson, 1968; James, 1902; Piaget, 1968), and--perhaps more surprisingly--even a strong contingent of otherwise apparently committed cultural relativists (e.g., Geertz, 1973; Hallowell, 1976; Shweder & Bourne, 1984) have all come to agree that selves of every conceivable stripe require being understood as capable of at least subordinating change, and that a sense of personal persistence is a "universal in the human experience" (Levine & White, 1986, p. 38), and so qualifies as "ubiquitous to all of human

Self-Continuity in Youth

179

kind" (Harr6, 1980, p. 397). You rarely get more of a concert of opinion than this. Discounting the occasional scorched-earth deconstructionist, for whom the idea of an integral enduring self seems as necessarily empty and suspect as does everything else (Shusterman, 1992, p. 248), more or less everyone appears prepared to agree with Cassirer (1923, p. 324) that any self that was not somehow understood to be persistent would simply fail to qualify as a recognizable instance of what selves are standardly taken to be. That is the good news. Where all of this happy agreement quickly begins to break down is in working out howmin the concretem this necessary feat of surviving time is imagined possible. B. THE PARADOX OF PERSONAL PERSISTENCE AND CHANGE

Presumably convinced, then, by the absurdity of the consequences to which its rejection would lead, you are prepared, I trust, to agree that the fundamental logic of individual identity necessarily understands persons as being somehow sufficiently persistent that earlier and later versions of themselves can and somehow must be counted as at least equivalent, if not precisely identical, instances of one and the same continuing person. All this and more said in favor of sameness is, of course, only half of a matched pair. Here is the other shoe. Life, whatever else it is, is a breakneck, and more often than not "whirl is king" (Schlesinger, 1977, p. 279). Clearly, our bodies change, our beliefs and desires (along with our projects and our commitments and our interpersonal relationships) all change, often seemingly beyond all recognition. We are, in short, like sharks, awash in the temporal flux, and any failure to keep moving is just another way of dying (Chandler, 2001). If something like the foregoing were not so, if selfhood was not vectored, and if change did not lie at the heart of subjectivity (Gallagher, 1998, p. 9), then we could make sense of neither the experience of innovation in the lives of individuals, nor novelty in the flow of human history (Unger, 1975, p. 56). Change, somewhat ironically, appears then to be as much a permanent fixture of our existence as does persistence. There you have it, the interpretive predicament that has troubled so very many for so long. Selves are required to embody both permanence and change simultaneously (Fraisse, 1963, p. 292)--a paradox whose hoped-for resolution has spawned a range of potential solution strategies so absurd that, to paraphrase Bertrand Russell (1959), "only very learned persons could possibly adopt them" (p. 149). As it is, the really strong solutions to this problemmthe ones capable of being spoken out clearly--were used up at the very beginning. No one since Parmenides and Zeno has seriously doubted that change is real, and next to no one beyond Heracleitus and Cratylus has boldly believed that absolutely everything is change and polysemic flux (Reese, 1991, pp. 200-201; Ring, 1987, pp. 66-68). With the really decisive possibilities already taken, and with straddling the middle always an awkward stance, the ambivalent alternatives remaining have left most contributors

180

Michael J. Chandler

to this vast literature dressing either slightly to the right or the left. Somemthe taxonomists, the neo-Darwinians, and the psychological essentialists--have gone on championing some modified version of essential sameness, while typically smuggling bits of difference in through the back door (Overton, 1998). Othersm the narrativists, hermeneuticists, and social constructivistsmhave typically leaned the other way, hedging their heavy-money bets on change by investing interest in those perhaps less reliable sorts of permanence that can be leached out of the connections binding together the distinctive time-slices of peoples' lives. Because of all of this backing and filling, a certain current lack of fidelity, with its attendant confusions, naturally reigns. What follows is a sometimes overly procrustean effort to get everyone reassigned to their own respective beds, hopefully with a minimum of body parts left over. C. ESSENTIALISTS ARE FROM MARS, NARRATOLOGISTS ARE FROM VENUS: A HIGH-ALTITUDE OVERVIEW

Although the business of characterizing whole intellectual traditions is, at best, a suspect profession, it has been widely practiced and it provides considerable precedent for imagining that, if not the whole of human thought, at least those parts of it devoted to grasping the paradox of sameness within change, conveniently divide themselves into two fundamentally distinct and irreducible modes of understanding. Something like the dichotomy that is in mind here is contained in Kant's and Hegel's now classical distinction between Vernunft and Verstand (see Overton, 1991, pp. 68-69), and the fissures that it has cut, particularly in continental European thought. As William James put it, "To say that all human thinking is essentially of two kinds.., reasoning on the one hand, and narrative, descriptive, contemplative thinking on the other.., is to say only what every reader's experience will corroborate" (cited in Bruner, 1986, epigraph). Certainly this would prove to be the case if one happened to be reading Bruner's (1986) own "dual landscape" account of the proposed differences between so-called "paradigmatic," "propositional," "argumentative," "logico-scientific" thought, on the one hand, and his "story telling," or "narrative" version of thinking, on the other. Much the same conclusion is afforded by Mandler's (1984, pp. 1-31) related dichotomy between "taxonomic" and "schematic" thought, as it also is, for example, by the still more recent distinction proposed by Nisbett and his colleagues (Choi, Nisbett, & Norenzayan, 1999; Norenzayan, Choi, & Nisbett, 1999; Ross & Nisbett, 1991) between "analytic" and "holistic" thought. The first of this matched pair, the one having to do with supposed entities and essences, is generally said to be foundational and metaphysical and is regularly viewed as being aligned with Plato and "the philosophers" and against the practitioners of "mere" rhetoric (Ring, 1987, pp. 84-87). By report, it is aimed at the universal or transcendental, as opposed to the historic or simply local or indigenous

Self-Continuity in Youth

181

(Habermas, 1985), and is always more monistic and individualistic than dialogical or collectivistic or multivoiced. Its mode of operation is said to be formal, logical, deductive, and rational (Hermans, 1996), as opposed to intuitive, inductive, and relational, and it is aimed at universal truth rather than verisimilitude, believability, or sincerity (Lightfoot, 1999). The second, more narratively based of these alternative pathways of thought is said to lead to a different intellectual landscape, one that is more extrinsic than intrinsic (Berzonsky, 1993), more centered upon subjects than their objects, more involved with process than structure, with becoming than being, and with the chronology of sequence, rather than the a chronology of models (Ricoeur, 1985). It is held to be more about extensions than intentions, characters than identities (Lightfoot, 1999), and it is said to be discursive and sequential, rather than timeless and substantial. It is unavoidably temporal, as opposed to a temporal; pragmatic, instead of encyclopedic; more relational than individualistic (Overton, 1998); and always asks "who" (Rorty, 1973), rather than "what" (Wiggins, 1980, chap. 3). Its memories are episodic instead of semantic (Tulving, 1983, pp. 9-10); its selves are ephemeral and unbounded, rather than bounded (Geertz, 1975), or natural, or embodied (Holland, 1997); and it is always seen to be socially constructed, as opposed to merely discovered (Waterman, 1984).

1. Falling between the Cracks of Received Intellectual Traditions Although these and other similarly rhetorical pronouncements succeed well enough in spelling out a rough bill of particulars--a loose accounting strategy of the sort that would allow certain historical figures to be collected together under the respective banners of either narrativism or essentialism--the list of things that such a trademarking would still leave unclear is even longer. Who, for example, or what, exactly, is meant to own one or the other of these distinctive modes of ordering experience: entire epochs; whole intellectual traditions; the East versus the West; only one's theoretical rivals; or, perhaps, those particular individuals whose close-to-the-ground cognitive habits are the topic of someone's empirical research? Does the predisposition to rely more or less exclusively on one or the other of these interpretive modes reflect differences in ability, or perhaps only commitment, and could they both sometimes or always be exploited under the right circumstances (Norenzayan, 1999)? Is essentialism really a by-product of our Western European and American "Judeo-Graeco-Roman-Christian-RenaissanceEnlightenment-Romanticist tradition" (Rorty, 1987, p. 57) and, if so, where were all of those budding narratologists and closet aficionados of what I will go on to describe as the storytelling tradition hiding out that allowed them to escape the metaphysical plague that Plato, Descartes, and Kant are said to have visited upon all of the rest of us (Murray, 1993)? Are such thoughts perhaps a symptom of the breakdown of the medieval order, or a product of the Enlightenment, or an offshoot of Romanticism, or something else entirely (Lightfoot, 1999)? Are these

182

Michael J. Chandler

"more or less accurate maps of the same territory," or are they "different maps that bespeak a difference in territory" (Holland, 1997, p. 165)? Is the self, or subject, then, merely the result of this or that discursive practice, or is it a substantial entity that has ontological priority over practice, some originator of epistemic meaning that authors the stories we tell about ourselves (Kerby, 1991, pp. 101-108)? Is everybody simply babbling? A useful first step toward addressing certain of these unanswered questions is to take a more careful look at the so-called essentialist and narrative traditions, or habits of mind, and to review some part of what is commonly said about them, both by their advocates and their detractors. "Useful," of course, can be made to cover a multitude of sins. The promissory note, at least in this case, is that considering these alternatives in more detail is worth the effort it requires for the reason that, in the course of their own development, young persons actually draw upon, or perhaps reinvent, something very much like these intellectual traditions in working out the conditions of their own persistent identity. One, but only one, of the ways that they do so is to imagine that they harbor some a temporal (sometimes even transcendental) core of persistent sameness. They are not, of course, the first to have entertained such "essentialist" thoughts. Here are some others.

Ill. The Essentialist Tradition: Strict Essentialism, Nominal Essentialism, Neo-essentialism, and Any Other Essentialism You Can Think Of A. WHO WERE ALL OF THESE CLASSICAL ESSENTIALISTS ANYWAY, AND WHAT WERE THEY AFTER?

Essentialismmthe metaphysical doctrine that holds that some or all physical objects necessarily possess a timeless, definitive core; some material or transcendental center that is immune to natural or historical changemhas been the subject of intense philosophical debate for more than 2000 years (Brockelman, 1985; Shalom, 1985). The good news--given the fact that the battle grounds on which most of these debates have been fought are "littered with the corpses of centuries worth of combatants" (Rorty, 1987, p. 65)mis that (as luck would have it) surprisingly little of this long controversy turns out to have any direct relevance to our present purposes, and so we are largely free to go on acting as though most of it never happened. This is currently as true as it is not only because next to no one still seriously believes that things really do have some God-given essence that makes them identical with themselves and no other, but also because the reasons that drove many of our intellectual history's best minds to think differently, and to overly agonize about such things, are not our reasons. They (where "they" refers to just about everyone from Aristotle to Leibniz) were, as Hirsch (1976, p. 2)

Self-Continuity in Youth

183

pointed out, primarily interested in what constitutes being a thing of some particular category or kind, and so they scarcely gave a fig about the possibility of some sort of individual essence that might potentially set you enduringly apart from me, or keep either of us selfsame over time. Rather than simply being passively uninterested, all but a few of these pivotal thinkers were, in fact, actively hostile toward the possibility of such individual essences, whose prospect they tended to regard as either trivial or vanishing. Classical accounts of identity were instead about timeless, all-or-nothing matters, such as whether "a" is identical with "a," or whether Constantinople just is Istanbul, and not about matters of persons or equivalence or degrees. By these lights, only objects that exist simultaneously in some frozen, timeless world of pristine logic are seen to hold out any serious hope of being identical, or at least identical in just that prissy way that made matters of strict identity of special interest to classical thinkers and almost no one else. When questions about the possibility of particular or individual essences did arise (as they did, for example, for Leibniz and Duns Scotus before him), they typically had to do with such disputed matters as whether angels, and other disembodied beings, had anywhere to keep their essences, and, if so, whether only an omniscient being could actually see them and had more or less nothing to do with such pedestrian matters as your personal persistence or mine. The classic essentialists were after bigger fish. What did sometimes motivate the interests of these classical thinkers in working out what might constitute at least the category or species essence of humankind was the ambition for some better legitimization of their own choices among competing moral and political ideologies, by grounding them in a knowledge of what was "essential" to the concept of the human person (Rorty, 1987). If you had some sweeping idea about how things should go in these larger arenas, and if you could show that your pet idea fit like a hand into the particular glove of essential human nature then your idea was simply better than some other competing thought that was less well tailored to such purposes. Your idiosyncrasies and mine--the things that might make us distinct and might account for our persistence--simply had no standing in such debates. Even when empiricists such as Locke and Hume eventually took aim at the classic notion of essence, insisting, as did Hume, that, after all, "nothing [exists] but diversity, diversity abstract and absolute" (Gallagher, 1998, p. 11), their targets were not our puny grounds for self-continuity, but rather the pretensions of their own professional adversaries. In short, if you are prepared to overlook the past 100 years or so, those whose names are likely to come to mind never really bothered to exercise themselves very much about what sort of essence may account for your just going on being you. Nor, despite much wailing and gnashing of teeth on the part of contemporary narratologists (many of whom continue to count coup on the heads of essentialists long since dead and buried), has much changed about all of this in the interim. Rather, like Locke, most contemporary philosophers, or at least most contemporary

184

Michael J. Chandler

analytic philosophers, continue to regard everything beyond a mere nominalist account of essence as a philosophical dead end. And even if they did not, few in the business of exploring matters of selfhood would care. Classical essentialism simply fails to "touch the mustache" of those identity issues that bear on the problem of personal persistence. True, the past 40 years or so have witnessed a certain resurgence of neo-essentialism among the so-called "friends of possible worlds" (e.g., Kripke, Montague, and Putnam), but even here, among those most concerned to reemploy the notion of "natural kinds" as a vehicle for accessing questions of individual essences, doubts persist as to whether persons are in fact "natural kinds" at all (see Putnam, 1982). As Rorty (1987) has put it, persons are, if anything, "that natural kind whose natural kind is not fixed" (p. 69). The takehome message, then, is that all of those times you spent flagellating yourself for not having paid more serious attention in philosophy class, and all those occasions on which you swore to go back and do it right, were wasted times and wasted occasions. Or at least this is true if your guilt was motivated by an ambition to get clear about classical notions of essentialism in order to do a better job of understanding the process of human identity formation. You need not have been so hard on yourself. B. IDENTITY, SURVIVAL, AND PERSISTENCE

Even if you are inclined to accept this ungenerousmeven throwawaymreading of the possible relevance of the classical identity literature, you might well ask whether something, anything, could possibly be salvaged from philosophy's ongoing pursuit of these matters and brought to bear on our own efforts to understand the psychology of personal persistence. At least two such promising prospects can be identified. One of these is about so-called "psychological essentialism" and will be addressed in the next subsection. The other is about what have broadly come to be considered matters of "survival" (Parfit, 1971)ma notion that already informs much of what has and will be said here on the subject of personal identity. As Parfit (1971) and a long train of others since (e.g., Hirsch, 1976, p. 64; Wiggins, 1980, chap. 3) have come to see, the term "identity," when applied to the unfolding lives of real persons, is a serious misnomer that deserves to be abandoned in favor of different and more appropriate talk about "persistence" or "survival." On this view, identity, which is always all-or-nothing, is not in itself an important relation. Most of the relations that are seen to matter, at least in the case of personal persistence, are relations of "psychological continuity" and are ordinarily measured in degrees. Importantly, we are not, on this account, psychologically continuous because we are more or less identical with earlier and later instantiations of ourselves. Rather, it is the other way around. As Parfit (1971) argued, "judgments of personal identity.., derive their importance from the fact that they imply psychological continuity" (p. 12), and not the reverse. What this different way of talking and thinking

Self-Continuity in Youth

185

accomplishes is to help free us from the felt obligation to explain our survival as ongoing persons by continually ferreting out some new candidate essence that is imagined to be stubbornly identical with itself. When warranting one's survival is what is seen to matter, then attempting to conjure up some putative essence that refuses to change is, of course, always a potential option, and one that can often be made to work. At the same time, however, such essentialist moves ought not to be seen as the only game in town. Rather, they are perhaps best viewed as only one option from a potentially larger bag of tricks useful in performatively vouchsafing, or argumentatively redeeming, one's claims about personal persistence. Still, as the following section is meant to make clear, the "discovery" of some material or transcendental thing that supposedly explains identity by managing to stand outside of time continues to be marketed as the one and only trick in the books by the advocates of so-called "psychological essentialism." C. PSYCHOLOGICAL ESSENTIALISM AS A ONE-TRICK PONY

The rehabilitative efforts of various neo-essentialists notwithstanding, fully fledged essentialisms of the classical sort practiced by everyone from Aristotle to Leibniz and beyond have, as was already indicated, come to be more or less universally regarded as "bad metaphysics," and, in consequence, have largely disappeared off the radar screens of most contemporary philosophers and social scientists. All of this could easily lead you to the mistaken idea that you are free to dismiss the whole confusing enterprise as a failed experiment. Instead, at least two (perhaps related) reasons should lead you to think otherwise. According to the first of these, our conceptual and linguistic traditions of relentlessly separating the necessary from the accidental has left traces of essentialism so deeply sedimented into our worldview that we can scarcely think a thought without appealing to it (e.g., Mayr, 1991, pp. 40-42). Something like this "indelible stain" hypothesis seems at work in the writings of Bruner, Nisbett, Mandler, and others cited earlier, and is generally seen to follow from much that is still to be said about the importance of history and culture in shaping our worldviews in general and our conceptions of selfhood in particular. The second reason to be slow to disown any lingering interest you might have in classical notions of essentialism is that they will actually prove useful in orienting you to a whole new subdiscipline within psychology devoted to the topic of socalled "psychological essentialism." Practitioners of this new cognitive science trade are typically quick to attempt to put distance between their own enterprise and that of more classical essentialists by claiming to dismiss, or to be uninterested in or agnostic about, all metaphysical questions having to do with how the world actually "is," insisting instead that they are concerned only with addressing the psychological question of whether ordinary people ordinarily believe (or proceed as if they believe) that certain kinds of objects or substances have something like

186

Michael J. Chandler

a constitutive, though often unknown, essence (Gelman, 1999). As Medin (1989), one of the progenitors of this new subdiscipline, has put it, though essentialism is almost certainly "bad metaphysics.., it may prove to be good epistemology," and for better or for worse, people just do, he claims, commonly "adopt an essentialist heuristic" (p. 1476)--a heuristic that is said to prompt adults, and children of a certain age, to disregard external appearance by "seeking deeper levels of analysis and grasping nonobvious essences" (Gelman & Wellman, 1990, p. 214). As such, psychological essentialism tends to be promoted (e.g., Gelman, 1999; Gelman, Coley, & Gottfried 1994; Medin& Ortony, 1989), not as a commentary on the relations between minds and worlds, but as a "plausible candidate for an important cognitive predisposition" (Gelman & Wellman, 1990, p. 242). Related thoughts about related matters appear to have arisen more or less simultaneously in diverse quarters. Atran (1994), p. 334), for example, reported that essentialist beliefs form an ordinary part of the folk-biology of adults across a wide variety of cultures--a conclusion that Gelman and her colleagues (Gelman, 1999; Gelman et al., 1994; Gelman & Wellman, 1990) argued is strikingly similar to their own claims about essentialism in the domain of folk-psychology. Ross and Nisbett (1991) similarly discussed what they referred to as "lay dispositionalism"ma like-minded and (they argued) cross-culturally widespread tendency on the part of most adults to suppose that something like hidden essences accounts for cross-situational regularities in people's behavior.

1. Killing Off the Fictive "Child-as-Externalist" Tradition Although the purposes to which these contemporary thoughts about psychological essentialism have been put are rather diverse, much of the considerable energy expended on this enterprise has been given over to a program of thought reform meant to disabuse us of what Gelman and Wellman (1990) characterized as Piaget's mistaken "child-as-externalist position" (p. 214). On this account, Piaget (e.g., 1951) and others (e.g., Aboud & Ruble, 1987; DeVries, 1969; Guardo & Bohan, 1971; Mohr, 1978) of this same supposedly withholding stripe are understood to have promoted the view that infants and preschoolers are invariably preoccupied with mere surface appearances and generally lack those capacities for abstraction that would allow them to penetrate beneath the surface and to imagine "realities" that go beyond mere appearance. Forgetful, perhaps, of the lessons Piaget meant to teach through the study of object permanence, failures in conservation are held up as paradigm cases of the preschooler's alleged snakelike fascination with the mere surface structure of transformations. Setting themselves against all such supposed dismissals of the natural talents of the very young, most contributors to the sizable literature on pyschological essentialism have been hard at work demonstrating that, by focusing upon what identity is rather than by concentrating on whether identity changes under certain transformations, earlier theorists such as Piaget inadvertently worked to "obscure young children's real

Self-Continuity in Youth

187

knowledge" (Gelman & Wellman, 1990, p. 214). Here the well-known work of Keil (1989) (who, it should be noted, worked with school-age and not preschool children) is cited approvingly as showing that, at least in the case of so-called "natural kinds," young children ordinarily discount a variety of radical (and often macabre) surface transformations in favor of inferences about the underlying nature, or category essence, of things (Gelman et al., 1994). Employing various "problem cases" of the science fiction sort so dear to the hearts of analytic philosophers, Johnson (1990) similarly concluded that, at least by early school age, children come to view individual identity as residing beneath the surface, and perhaps localizable in the brain. That is, like the work of Keil (1989) and Gelman and Wellman (1990), Johnson's findings are widely interpreted as supporting the general conclusion that by a surprisingly early age, young preschool children "understand that insides are typically more important than outer surfaces for an objects identity.., reflecting an understanding of inborn essence and a basic predisposition toward psychological essentialism" (p. 213). Others (e.g., Spelke, 1985; Spelke, Kestenbaum, Simons, & Wein, 1995; Xu & Carey, 1996), bolder still, have invoked the much feared "preference-for-novelty" paradigm as a means of promulgating doubts as to whether children, no matter how young, actually understand absolutely nothing of object persistence. How children actually accomplish these feats is never made clear, but a safe assumption seems to be that they do not do it by constructing the sort of narrative structure that is had in mind by the usual enemies of essentialism, psychological or otherwise. 2. Kind versus Individual Identity

Although much about the large body of research on psychological essentialism is obviously instructive, whether it succeeds in addressing the question (our question) of how persons think about their own personal persistence across the kinds of transformations that time has in store is a good deal less clear than has been widely assumed. This conclusion follows for largely the same reasons that have served to call into question the relevance of more classical philosophical accounts of essentialism per se. The problem, for present purposes, is that the bulk of empirical research concerned with young children's early intuitions about psychological essentialism is once again concerned with matters of "kind" persistence. What Keil (1989), for example, actually tried to accomplish, what DeVries (1969) earlier tried to find out, and what Gelman and Wellman and their colleagues (e.g., Gelman et al., 1994; Gelman & Wellman, 1990) have worked so systematically to specify, is the age at which young persons first come to intuit something about "species identity," of "category essence," and what would be needed to move this or that "natural kind" (e.g., a skunk or a cat) across some standard ontological boundary, such that it becomes a thing of a different kind. Of course, nothing is wrong with pursuing such taxonomic questions, but whether the data that such studies generate have any direct bearing on the separate issue of individual

188

Michael J. Chandler

identities and their cross-temporal persistence is far from assured. When DeVries, for example, put dog and rabbit masks on the classroom cat and asked 3- to 9-yearolds to reidentify the animal, the right answer was taken to be simply "cat," any cat, and not "Tabby" or "our cat" or the very same God-given cat that it was before all of the hocus-pocus. Similarly, when Keil proposed painting out the stripes on skunks or surgically "removing their insides," what he was after was whether children of a particular age judged the resulting specimen to be still a skunk, and not whether it emerged from all of these insults with its personal identity intact. Remember William James? Remember that crab of his acquaintance that, on hearing itself referred to as a crustacean, rose upon its hind legs and shouted "I am no such thing, I am MYSELF, MYSELF alone." (James, 1902, Lecture 1, p. 10.) Well, this is not like that. Rather, much as was true in the case of more classical accounts of strict identity per se, the bulk of available studies on psychological essentialism are concerned with matters of category membership and so are largely irrelevant to the question of personal persistence that motivates this article. As if to prove the rule, however, an important, though small handful of studies are exceptions to the otherwise standing generalization that psychological essentialism is all about species rather than individual identity (e.g., Gutheil & Rosengren, 1996; Hall, 1998; Rosengren, 1999). Drawing on trace elements of earlier concerns about young children's first thoughts on personal persistence (e.g., Aboud & Ruble, 1987; Guardo & Bohan, 1971; Mohr, 1978; Oppenheimer, Mur, Koeman, & Chandler, 1984; Piaget, 1968), Rosengren and his colleagues have, for example, demonstrated that even preschoolers tacitly employ essentialistlike ideas regarding the "behavioral personalities" of particular cats and dogs, all in an effort to account for the persistence of certain of their idiosyncratic habits. These early expressions of what, in another context, Ross and Nisbett (1991) called "lay dispositionalism," are read by Gutheil and Rosengren (1996, p. 495) as suggesting that by 4 years of age, children already possess a "substantial, if incomplete understanding of individual stability." More recently Hall (1998) showed that when adjudicating the question of whether individual things (in this case, nonhuman things) go on being themselves despite the gradual replacement of all of their parts, even preschoolers care whether the subjects of such transformations happen to be artifacts or living things. What is particularly promising about this line of inquiry is that it raises the prospect that children's growing conception of both individual and species identity, while distinct, may prove to turn upon related cognitive mechanisms. D. THE DOWNSIDE OF ESSENTIALISM, PSYCHOLOGICAL AND OTHERWISE

To the credit of recent students of psychological essentialism, due in no small measure to their ruggedly antimetaphysical or even postmetaphysical stance, they have succeeded in rehabilitating the topic of identity maintenance--a subject once

Self-Continuity in Youth

189

banished by mainstream psychology to the cloudland of unobservable and irrelevant abstraction (Sass, 1988, p. 251). Their success has nevertheless come with a price. The hidden cost of newly reimagining that both children and adults are, by their very "nature," born and bred epistemic (if not ontological) essentialists, I want to suggest, is that it serves to perpetuate the pernicious and perhaps ethnocentric myth that essentialism is simply the only and necessarily right way to approach the problem of species and (by implication) personal persistence. Much of Western intellectual history can be read as a search for the foundational subject, leading, as Kerby (1991, p. 5) pointed out, either to a mysticism, wherein the existence of the self is assumed but cannot be demonstrated, or to skepticism concerning the whole enterprise. Of course, contemporary champions of psychological essentialism are more like Kant than most of his predecessors in imagining that being an essentialist is only a natural feature of the knowing process, and not an ontological obligation imposed by the nature of substances themselves, but for most psychological purposes this distinction is moot. In either case the job of securing a sense of personal persistence amounts to the same thing. We either have or just naturally suppose we have some unique or species-general, but typically hidden, property or essence. As such, the proper measure of our maturity is assumed to be found in the degree to which we succeed in waving off the mere surface structure of things on our way to envisioning this more essential core. We, along with Keil's skunks, Rosengren's or De Vries' cats and dogs, and, as Rorty (1987) put it, particular patches of pachysandra and the Mediterranean basin, each belong to our own natural kind, and it is assumed to be every young child's responsibility to "discover," or perhaps imaginatively invent, whatever foundational stuff is thought to be necessary to ensure our individual and categorical persistence through thick and thin. "It seems to me" said Schwartz (1979), summing up such views for all who subscribe to them, "that people observing the developmental stages in animals, or the changes in form of, say, water, would naturally be led to assume the existence of some underlying trait that makes some stuff or thing to be of a kind . . . . It seems to me that there must be such an essence or my worldview is radically false" (p. 315). Schwartz perhaps went too far. All he really needed to admit is that, stripped of some perhaps culture-bound commitment to essentialism, or even psychological essentialism, his worldview would need to be radically different. Were he to make such a conceptual leap, his landing place would not be otherwise unoccupied. Rather, he would likely end up surrounded by all of those numerous other narratologists, presentist historiographers, and social constructivists who together imagine that selves and their persistence in time are not rooted in some real or imagined essence, merely awaiting discovery, but are due instead to the fact that being a self is less like being a substance and more like being what Dennett (1992) called a "center of narrative gravity." Trying to get clear about this alternative vision is the central business of the next section.

190

Michael J. Chandler

IV. The Narrative Tradition as Another Singular Possibility You are, I trust, still out there in this forest of words waiting for the sound of the second shoe dropping. Its sound, as it turns out, is regularly heard (by literary criticists, presentist historiographers, and social constructivists of various stripes) to have a performative tone and all of the natural cadence of a story being told (Ricoeur, 1980). Here is the question for which "a story" is the narratologists' standard answer. If the container/substance view of self (Holland, 1997) is simply mistaken on its face, and if the prospect of finding some enduring metaphysical something, some foundational core of selfhood standing behind our always shifting ways of being is really as bankrupt as in-house audits have recently suggested, then we need to ask, along with Dilthey (1962), "what is it then which, in the contemplation of one's life, links the parts to the whole?" His answer, along with that of just about everyone who subscribes to what is being billed here as the "narrative option," is that "It is the fact that the connectedness of life can only be understood through meaning" (pp. 201-202)--meaning that can be conferred only by linking up what would otherwise be disparate time-slices of one's existence into some overarching narrative structure (Bruner, 1986, chap. 2; Nicolopoulou, 1997). The broad prospect that the "natural way human lives are given unity and continuity is through narratives, can," as Peacock and Holland (1993) meant to show, "be traced through many fields, from the cultural-historical school of psychology and related approaches (Volosinov, 1986; Vygotsky, 1987), to literary criticism (Bakhtin, 1981; Luk~cs, 1914), to psychiatry (Lifton's "protean man" [Lifton, 1970], and Lacan's tack that the deep self is merely linguistic [Lacan, 1968]), and to anthropological debates about the real versus constructed emphasis in self" (p. 368). Much the same case is made by McAdams (1997), who argued that philosophers and psychologists, in growing numbers, have also come to insist that "the predominate way that human lives are comprehended is through narrative (Bruner, 1990; Charme, 1984; Cohler, 1982; Cohler & Cole, 1994; Hermans & Kempen, 1993; Howard, 1991; Kerby, 1991; Kotre, 1984; Linde, 1990; MacIntyre, 1981; McAdams, 1984, 1985, 1990, 1993; Polkinghorne, 1998; Singer, 1995; Singer & Salovey, 1993)" (p. 63). Given all of these cooks stirring this same pot, you will not be surprised to learn that we have several different ways of telling ourselves the story about why selves are simply stories, and of explaining why it took until roughly now to get clear about such a simple fact of the matter. One such putative explanation, perhaps the hardest of the lot to swallow, goes something like this. Before the scales somehow mysteriously dropped from our eyes, we were all, somehow, simply misled by the ancient philosophers, who collectively managed to persuade us that "being was given once and for all, complete and perfect, in an ultimate system of essences" (Schlesinger, 1977, p. 271). On this "as-fate-would-have-it" account, essentialism

Self-Continuity in Youth

191

is something not unlike a historically induced nightmare, a delusion from which, of late, we have happily begun to awaken. Others (e.g., Miller, 1996), relying less on happenstance, have supposed that, being "naturally" less dependent upon specialized forms of enculturation, narrative accounts of the self have always been present in all societies and were merely driven out of awareness--at least professional awareness~by certain "probabilistic modes of reasoning.., unique to only a handful of 'modern' [read "Western"] cultures" (p. 113); cultures such as our own. We do not, from the point of view of most contemporary narratologists, actually need to choose between these speculative possibilities. Rather, however earlier generations might have come to miss it, the serious prospect that selves really are no more or less than the narrative embodiment of lives told (Spence, 1983) has, of late, come to be broadly accepted, or at least intensively discussed, by 20th century literary critics (e.g., Frye, 1957), by historians (e.g., Mink, 1978; White, 1973), and by philosophers and psychological theorists (e.g., Gergen & Gergen, 1986; MacIntyre, 1981; McAdams, 1997; Sarbin, 1986) of otherwise broadly different stripes (Carr, 1988, p. 5). A. AUTHORS LIVING AND DEAD

What, perhaps, best sets this collection of strange bedfellows apart are the different views that they take regarding the vital signs of authors, and, in particular, whether we should imagine them to be living or dead. On the one sidemon the side of imagining a real "auto" in "autobiography"mare those who agree with Nietzsche in imagining that if a story is being told, then it must include a "doer of the deed" (Kerby, 1991, p. 14). Those whose writing gives voice to this view may be just as convinced, as are their more extravagant, dyed-in-the-wool narrativist counterparts, that we are "storytelling animals" (MacIntyre, 1981, p. 201), that each of us "configures our own historical unity" (Polkinghorne, 1988, p. 183), and that we make sense of our otherwise scattered lives by fashioning stories meant to integrate our reconstructed past, present, and anticipated future (McAdams, Diamond, de St. Aubin, & Mansfield, 1997), but they still, nevertheless, manage, in the process of doing so, to imagine the existence of a self operating outside of these acts--an actual doer of these narrative deeds (Kerby, 1991, p. 14). On this account, narrativization may very well serve to dress up reality in ways that satisfy our need for coherence and do so in ways that can be usefully seen to conform to this or that literary structure or genre, but we, as the architects of record of our autobiographies, are not imagined to be coextensive with them and are most decidedly not dead. The opposite corner in this debate is occupied by altogether more committed social constructivists (e.g., Gergen & Gergen, 1986; Harr6, 1989) who regard the self not as the source of personal narratives, but as their product, effectively dissolving personal history into a species of literature. On this "topological thesis" (Zagorin,

192

Michael J. Chandler

1999), self narratives are neither the simple consequence of somehow imposing art onto life (Kerby, 1991, p. 41), nor the inventions of some deep structure rooted in an otherwise independent mind, but they are seen instead to arise directly out of the fact that the social and material conditions of human existence themselves have a narrative structure. On such typically postmodern views, not only is God dead, but so too are selves. With no more prospective authors left standing, the remaining possibility (or so it is argued) is that the stories that are "us" must somehow have written themselves. As you might imagine, not every card-carrying narratologist is prepared to go quite this far around the interpretive turn. To claim that we are no more or less than authorless fictions (Kerby, 1991, p. 34) is, according to Zagorin (1999), for example, to become "lost in the tropics of discourse" (p. 23) and to give narrativity a bad name. Even if "it is in telling our own stories that we give ourselves our identity," and even supposing that "our own existence cannot be separated from the accounts we give of ourselves" (Ricoeur, 1985, p. 214), many contemporary narratologists appear unprepared to accept the full equating of lives and stories. Stories, as Mink (1969) pointed out "are not lived but told. Life [or at least a life still in the making] has no beginnings, middles, or ends" (p. 557). Nor, as many others (e.g., Carr, 1986, pp. 98-99) pointed out, are our own stories amenable to just any telling. Self-deception, and repression, and lying through one's teeth actually do happen. Not only does our past have a certain fixedness that needs to be honored, but on this more sober account, it must be squared with the rest of what has previously been agreed to in the stories of others who share common experience with us. These and other qualifications notwithstanding, however, the likely prospect that self narratives are "poetic orderings" (Ricoeur, 1980) that have been "transferred from art to life" (Mink, 1969, p. 557) in no way disqualifies them as an alternative and ready means of drawing useful connective threads of accountability (of continuity and persistence) through our lives. In short, whether the story we tell about our lives is, or is only about, the self may matter less than whether, as Giddens argued (1991), we are "able to keep particular narratives going" (p. 54). We could, of course, attempt to bridge our past, present, and future lives by doggedly imagining the persistence of some essential thing, we perhaps know not what, but there is an alternative. What Flanagan (1996) called "narrative connectedness" at least holds out the possibility of grounding personal persistence without essence (Putnam, 1988, p. xv). B. GOOD THINGS TO SAY ABOUT NARRAT1VITY

The point in drawing attention to this alternative is definitely not, however, to somehow persuade you that essentialist accounts of personal persistence ought to be set aside in favor of some putatively superior narratively-based alternative.

Self-Continuity in Youth

193

Rather, the whole point of the present exercise is to double, rather than cut in half, the number of available options for thinking about self persistence. Still, it is, nevertheless, also true that such narrative approaches do get us to places that cannot be easily reached using essentialist alternatives. One of these is obviously the place where the considerable resources of literary and historical analyses are kept. Literary analysts such as Bakhtin (1986), Propp, Wagner, Scott, & Dundes (1968), Frye (1957) and Rorty (1973) have, for example, formulated conceptual and procedural means of exploring plot structures and folk/literary genres that have been put to especially good use by psychologists concerned with the study of selves (e.g., Chandler & Sokol, in press; Feldman, Bruner, Kalmar, & Renderer, 1993; Lightfoot, 1994, 1999; Nicolopoulou, 1997; Sutton-Smith, 1986). Similarly, and because stories are meant for the ears of listeners, narrativity succeeds where essentialism fails, in achieving a more responsible degree of dialogicality (Hermans, 1996; Hermans, Kempen, & Van Loon, 1992; Lewis & Ferrari, 2001; Sarbin, 1986, 1993). That is, because the persistence of characters (as opposed to stand-alone selves) turns primarily on their abiding fidelities and sincerities, the possibility is opened up that self-continuity may be more about being counted upon by other, than being counted equal to one's self and no other (Lightfoot, 1999). Finally, and perhaps more important than any of their other consequences, narrative accounts of the self open a window onto the possibility of productive cultural comparisons. That is, in contrast to traditional essentialist views that promote a bounded (Geertz, 1975) and situationally impermeable conception of the self (Shweder & Bourne, 1984), narratively based accounts can be seen to be more or less automatically open to the possibility that different cultures might make available for use different narrative frameworks for organizing the temporal flow of one's experience. In view of the now obvious fact that the world's cultures do evidence a "huge variety of discourses, practices, concepts, means, and modalities of the self' (Holland, 1997, p. 163), any accounting system that includes easy provisions for accommodating to this evidence gains a decided advantage. Nor is the cause of essentialism helped by the fact that the criteria it proposes for personal persistence turn out to be, more or less, duplicates of the ideals of the Western Enlightenment (Giddens, 1991, p. 117). More to the point than any such facile charge of "Eurocentrism," however, is the fact that narrative accounts offer, in a way that essentialism cannot, a common way of understanding both cultural and individual continuity in the same breath. As Feldman and her colleagues (Feldman et al., 1993) have worked to show, access to shared narratives is not only a viable way of maintaining the persistence of individuals, but is similarly a key to the maintenance of continuity and identity in whole subcultural groups. For such reasons, as Dobert, Habermas, and Nunner-Winkler (1987) meant to show, extreme cases of loss of individual identity (as in the definite leave-taking we call suicide) can be understood as "just as much the symptoms of a ruptured, distorted form of life context in the social group as they are signs of the course of individual illness"

194

Michael J. Chandler

(p. 277). No essentialist view, which sets the self contrastively both against other selves and against its own social and historical background, could foresee this culture-identity relation or anticipate the strong results (to be presented in Section VI) that directly link youth suicide to various markers of cultural deconstruction in First Nations communities. C. NARRATIVITY PLAYS TO CRITICAL REVIEWS

All the good things so far said about narrativity are obviously meant to be understood as being for the best. Being more dialogical, more welcoming to meaningful cross-cultural comparisons, and more open to the conceptual and methodological contributions of literary critics and intellectual historians can hardly be other than a good thing. Unfortunately, as matters now stand, such narratological views also have a downside, and this is it. What is perhaps most dishearteningly familiar about this otherwise fundamentally good bet concerning selves and their temporal persistence is its all too familiar presumptive exclusivity. With an oracular certainty matched only by the absolute convictions of your typically committed essentialist, champions of the new narratology also regularly see themselves as having been gifted with their own singular revelation. On this account, self narratives are not simply one among other potential symbolic devices capable of being brought to bear on the construction and defense of one's identity. Rather, on this view "there [simply] is no preexisting essence or self to be discovered . . . . People must literally make up a story about who they think they are" (Berzonsky, 1993, p. 177, emphasis added). "Life [just] is inherently a narrative structure" (Kerby, 1991, p. 40, emphasis added) or "narrative construction" (Bruner & Kalmar, 1998); telling a story automatically is the "essential genre" of self represented identity (Flanagan, 1996, p. 23); the "natural"' (Maclntyre, 1984, p. 192), "native tongue" (Mink, 1978) of the self; so is seen to be "the only way to account for a specific person" (Weintraub, 1975, p. 847, emphasis added). On the off chance that narrativity should turn out not to be the only conceivable way to understand one's identity, it is at least held out to be the "best" or the most "privileged" (Kerby, 1991, p. 4, emphasis added) way of doing so, all because the self is seen to be altogether "more narrative in nature than logical" (Bruner & Kalmar, 1998, p. 314). It is, we are told, "the form of hermeneutic expression in which human action is understood and made meaningful" (Polkinghorne, 1988, p. 145), and so it deserves being counted as the "primary embodiment of our understanding of the world, of experience, and ultimately ourselves" (Kerby, 1991, p. 3). In short, narrativity is regularly held out, by converts to what threatens to become the new dogma, to be the one and only right sort of dynamic glue with which to hold together all of the admittedly distinct time-slices that, taken together, turn the mere chronology of a life into a coherent and diachronically unified self.

Self-Continuity in Youth

195

D. NARRATIVE THIS, ESSENTIAL THAT

Clearly you have a right to be disappointed. The best minds, it would seem, are once again at loggerheads. This is how things are alleged to be. Adults, and even children, have psychological essentialism bred into their bones. Rightly or wrongly, they are just naturally inclined to suppose that things such as ourselves (i.e., things of a natural kind) are invariably characterized by some enduring essence that accounts for their persistence in time. No, wait a minute, that isn't right, it's just the opposite. It is narrativity that is the only, natural, primary, privileged, best, native tongue of the self. But if so, what are we to make of all the evidence which so clearly shows that when given half a chance, even very young children just go for essentialism every time? If we take all of that in, are we then meant, as a consequence, simply to turn a blind eye to the whole literary canon? Did everyone from Homer to Dostoyevsky actually get it all wrong? Do the meanings of selfhood and the continuities of our lives not reside in the stories we tell about ourselves? Are you confused yet? Is it OK to go on being this schizophrenic, and, if not, on what grounds are we to choose between these stark and exclusionary alternatives? Are such choices really necessary? Is this unmediated confornation between a "pure transcendentalism ... burdened as it is with our consciousness of the fallibility of human knowledge" and a "pure historicism . . . that violates our need for consistency" (Habermas, 1985, p. 193) as good as it gets? Although not exactly meant as an answer to any or all of these contrary questions, the fact of having listed them out altogether is obviously meant to be cautionary-to signal a concern that we appear to be somehow going about all of this in the wrong way. What is wrong, I want to suggest, with much of the available commentary about individuality and selfhood and personal persistence, is that it is typically offered up as a descriptive means of somehow characterizing our ongoing singularity. For some, a self boils down to an individual essence. Others see ongoing you and ongoing me as being better described as a story. Incommensurable lines of evidence in support of both of these possibilities ghost past one another without touching a hair on either head. Everything comes down to a choice, where choosing seems little more than a matter of blind liking. "I can describe anything better than you c a n m N o you can't--Yes I can," all said tunefully until, in exasperation, our audience simply switches to another channel. This can scarcely be the way to go. Here is an alternative--a stab at what is hopefully a better direction. On the logic that we have no choice, short of falling into incoherence, but to regard ourselves as one and the same persistent and numerically identical individuals through time, and in light of the fact that profound personal changes are inevitable, circumstances (both public and private) naturally and regularly arise under which we must offer up some guarantee, or provide some followable reason, as to why we deserve to be taken as precisely that selfsame and continuous person we hold ourselves out to be. Consequently, claims, at least individual first-person claims, about identity and

196

Michael J. Chandler

persistence are perhaps not ordinarily meant to be taken as valid ways of literally describing anything. Selves, as Habermas (1991) argued, may actually have no descriptive or elocutionary meaning. Instead, our usual efforts to drag up some presumably persistent and essential parts of ourselves, or to point out otherwise this or that narrativelike connection that allegedly ties together the various episodes of our lives, are, I mean to suggest, better understood as performative bids to be recognized as just that particular ongoing person we imagine ourselves to be. Under these usual circumstances, pointed questions about our identity, or about how we qualify as one and the same diachronically continuous person, need to be, and typically are, seen as adversarial questions, and our answers are standardly offered up as arguments for the defense, meant to vouchsafe, or argumentatively redeem our disputed claims for distinctiveness and persistence in the face of lurking evidence to the contrary. In short, to talk of one's essential attributes, or to drag out one's life story, is just to announce whatever credentials we imagine will be accepted as a legitimate guarantee of one's right to be judged as distinctively one's self, or otherwise personally persistent through thick and thin. The prospects of winning such rhetorical arguments turn, then, more on verisimilitude than on truth, and rather than being calculable in strictly monological terms, they depend most decisively on what one's audience (one's culture, if you will) is likely to find most persuasive. If something like the above-mentioned scenario can be made to count as a fair description--that is, if, when pressed for personal details, what we typically have to say about ourselves is rather more like jumping to our own defense--then a number of other things ought to follow. The first of these is that we would be within our own rights to expect that people, in general, and our research participants, in particular, would know what we are on about with our questions concerning personal persistence. Not that they all necessarily spend a lot of private time ruminating about their place in the temporal flow, but, rather, when it is put to them point blank, they ordinarily rise to such provocative occasions by taking up the implicit challenge thrown down. In short, they recognize that their job is to build a case that, all evidence or suspicions to the contrary, they are in fact more or less the same person they had been before wear and tear began throwing everything into doubt. Second, should circumstances turn in such a way that they felt that they had no such defense (i.e., if they tried and failed to come up with some acceptable warrant for their own personal persistence), they might be expected to take this all rather badly and as a sign that something was terribly amiss. That is, they would worry, just as those around them would surely worry, that such a failure to argumentatively redeem themselves (Habermas, 1991) would throw their past into doubt and jeopardize their future prospects. Finally, we would have every right to expect that when confronted with evidence of radical personal change, people would (within the limits of their ability) ordinarily do the done thing by exploiting, as a first line of defense, whatever culturally approved self-continuity warranting

Self-Continuity in Youth

197

strategy was seen to work best in their own venue. As it turns out, and as I mean to go on now to try and make clear, all of these expectations have been more or less borne out in certain data we have collected.

V. A Theory in Search of Evidence Here we are, however many pages later, ready to line out the details of how a collection of young persons (as opposed, for example, to standing armies of narrativistically- and essentialistically-minded philosophers or historians or warring psychologists) actually respond when urged to comment upon their own and others' possible persistence in time. So far we have responses from about 400 of them, representative of different age and ability groups, some of whom were actively suicidal and others not. Some (nearly a half) were drawn from the aboriginal cultures of Canada's West Coast First Nations communities; the rest were intended to be more representative of Canada's cultural mainstream. All of these respondents (adolescents for the most part) have been interviewed in rather elaborate detail about how they understand their own and others' personal persistence, or diachronic identity, all in the face of the unavoidable facts that their bodies have changed and their beliefs and desires (along with their projects, commitments, and interpersonal relationships) have all changed, often seemingly beyond all recognition. Before going on to say who, in particular, these respondents were, and exactly how they were "tested," I think it is probably best, although perhaps dramaturgically unwise, to give the whole game away by stating at the outset that neither committed psychological essentialists nor dyed-in-the-wool narrativists are likely to take much pleasure in these findings. On the whole, the young people whom my research colleagues and I have interviewed are not, as champions of both of these exclusionary accounts would have it, uniquely committed to a view of themselves that makes personal persistence the exclusive consequence of either some foundational essence or some narrative structure. Rather, among all of these participants are individuals whose personal views fit comfortably and well into each of these dogmatic alternatives. Nor have individual participants generally proven to be simply all over the map, haphazardly expressing essentialist views at one moment and narrativist views at another. The facts are more compelling than that. A. A DATA S U M M A R Y

Here are five such points of interest, each of which, in its own turn, will require some amount of further elaboration. First, of the 400 plus adolescents so far interviewed, all but a very few (20, to be exact) were absolutely convinced that they had good reasons (essentialist reasons; narrativist reasons) to back their own hoped-for

198

Michael J. Chandler

conclusion that they were, despite everything, still continuous in time (e.g., Ball & Chandler, 1989; Chandler & Ball, 1990; Chandler, Boyes, Ball, & Hala, 1987; Chandler et al., 2001). Second, with what amounts to a single exception, all of the young people who failed to come up with any way whatsoever of argumentatively redeeming their necessary claims for self-continuity were actively suicidal and had recently taken some definite leave-taking step meant to end their own lives (Ball & Chandler, 1989). Third, although effectively everyone else regularly, and, for the most part, consistently, mounted some version of either an essentialist or a narrativist self-continuity warranting strategy, their attempts to do so proved to be more or less adequate or complex. Younger, as opposed to older and cognitively more mature, persons generally tended to approach the job of vouchsafing their own and others' personal persistence in more concrete and particularistic ways. Fourth, the distribution of narratively based and essentialist self-continuity warranting practices observed across this large and heterogeneous sample was strongly determined by the culture of origin of these participants (Chandler et al., 2001). By a measure of better than four to one, culturally mainstream Canadian youth proved to be committed essentialists, convinced that, lurking beneath the variable surface structure of their manifestly different ways of being, there is some more or less abstract, but fully founded essential, something that they believe guarantees that they will go on being themselves through thick and thin. By contrast, First Nations youth, in almost equal but opposite proportion, proved to be just as committed to the competing view that, rather than depending upon the persistence of some privately owned essential something hidden away from time, their own and others' self-continuity is guaranteed by the fact that there is always some story (often multiply authored) that succeeds in gluing together the distinctive time-slices of their lives. Of course, hints of both of these self-warranting strategies are sometimes evident in the protocols of individual respondents in either group, but very clearly, most have a decided way of approaching the problem of personal persistence, and this way is largely decided by their membership in one or the other of these cultural groups. Fifth, epidemiological data now in hand (Chandler & Lalonde, 1998) strongly suggest that the degree to which individual youth succeed or fail in working out some acceptable means of resolving questions about their own and others' personal persistence is heavily influenced by the degree of "cultural continuity" to be found within their own communities. At least dramatic results are available that demonstrate that the rate of youth suicide (our proxy marker for the loss of personal self-continuity) is some 200 to 400 times higher in those First Nations communities that have been least successful in providing their members with a strong sense of connectedness to their own cultural past. The rest of this report amounts to a clarification of and elaboration upon these key findings. Because the data to be reported were collected over more than a 10-year period, and because the ways in which these matters have been thought about have also evolved over the interim, the best account of what was done is

Self-Continuity in Youth

199

not a strictly chronological account. Rather, the preferred way to begin is with what my research colleagues and I currently hold to be our best understanding of these matters, with provisos and apologies for less-well-thought-through beginning efforts salted in along the way. Given all of this, and in light of what could only be your need to hear, finally, about something very down to earth, perhaps the best place to begin is with a hands-on account of exactly how we have gone about collecting information about young people's beliefs about personal persistence. B. IN SEARCH OF A METHOD: A BILDUNGSROMAN FOR EVERY OCCASION

If you want to know in some detail how young persons of various ages and cultures actually think about personal persistence and how, in particular, they undertake to warrant their own and others' self-continuity through time, the best procedure is probably not simply to ask them straight out. Of course, the work to be reported here is predicated on the organizing assumption that each and every one of us is obliged (by the logic of identity relation, by the semantics of selfhood, by the moral obligations of social life, and by our collective hopes of eventually inheriting our own just desserts) to understand our selves and others as diachronically continuous people. Nevertheless, young people tend to look at you rather strangely when asked up-front to supply their personal criteria for persistent identity or to list out their own best warrants for self-continuity. Having discovered this fact early on, my research colleagues and I (i.e., Chandler et al., 1987) hit upon a rather more roundabout but otherwise more productive alternative. This procedure turned on the fact that what has, of late, come to be regarded as serious literature is most often about matters of character development and how strong characters typically start out one way and then, through a chain of more or less followable transformations, typically end up being some often radically different, but understandably related version of their own former selves. In this century, at least, such Bildungsromane (or stories of character development) have come to be taken as something of a gold standard for successful literature. Like life, they cannot be read and properly understood without bringing to bear some interpretive means of rationalizing how someone who managed to become so different could still legitimately count as more or less the same individual person. As we have learned from scholarship on "reader-response criticism" (e.g., Tompkins, 1986), the glue that we as individual consumers introduce to cement the parts of novels together need not necessarily be the same brand their authors had in mind. Still, one way or another, supplying some such glue (narrative glue, essentialist glue) is expected of us, or else the story in question fails to be a story about one person. Given all of these things, a promising approach seems to require research participants to read such works of fiction, while the researchers stand back and see what sort of glue they chose to apply. What would come out of such readerly attempts

200

Michael J. Chandler

at meaning making, we assumed would likely tell us something useful about how participants understood the connectedness in the lives of others. One couldmand we didmalso ask volunteers similar questions about their understanding of their own attempts at autobiography. By extension, we reasoned that by such efforts our participants would speak more or less directly about their own views of the basis of their own personal persistence. Doing a complete job of this would, understandably, be excessive, or would at least have drawn too heavily on the usual allotment of good will with which volunteer research participants ordinarily come equipped. What we did instead was to ask our young research participants to read along with us two or more "Classic Comic Books" that we had further stripped down to some 20 pages or less. We did this, for example, with Dickens' A Christmas Carol and with Victor Hugo's Les Miserables, and we further helped our young volunteers along by supplying audiotapes meant to spare them the problematic necessity of reading anything at all. Having gone over these materials, we asked them to describe Ebenezer Scrooge and Jean Valjean at the beginnings of their respective stories and then to offer up an account of these "same" characters as they appeared at the end. Their central task was to say whether these stories were about one and the same individual from beginning to end and, if so, to explain how characters who acted and appeared so different could be fairly judged to be the selfsame individuals. We also asked them to characterize themselves at two remote periods in their own lives (now and 5- or 10-years ago) and to answer the same questions about identity and persistence on their own behalf. Their often lengthy answers to all of these questions became the stuff out of which we sought to understand and characterize their various ways of warranting claims of personal persistence in self and other. The analytic upshot of all of these efforts was a coding s c h e m e ~ a typology of alternative self-continuity warrants, to be outlined in the next section. Before that, however, something quick needs to be said about how this initial procedure was elaborated along the way. First, one of the changes that did happen grew out of a concern that requiring subjects to read anything, even the stripped down "Classic Comic Books" we had fashioned, might serve to get us off on the wrong foot, or otherwise obscure the best abilities and understanding of some of our Native and non-Native participants, for whom reading was not always their strongest suit. In an effort to get past this eventuality, and in tribute to the fact that even comic books are no longer "hot" media, we also brutally edited down filmed versions of these same and other stories, hoping to allow participants who have learned most of what they know from television to participate on an equal footing. Second, and out of concern that the Western European "Judeo-Graeco-RomanChristian-Renaissance-Enlightenment-Romanticist" (Rorty, 1987, p. 57) "bias" that some would ascribe to Dickens' and Hugo's classic stories might work to the disadvantage of our young First Nations participants, we scouted out counterpart

Self-Continuity in Youth

201

Native stories and films that were presented in similar formats and in all conceivable counterbalanced orders. That is how the data were collected. Here is how they were eventually coded. c. A TYPOLOGYOF ALTERNATIVESELF-CONTINUITYWARRANTS What has gradually become clear, after countless hours of typing and pouring over the transcripts of our several hundred participants, is that not only do our young participants typically commit themselves to either a narrativist or an essentialist course, but also they do so with varying degrees of complexity that are loosely related to their present chronological age and level of cognitive maturity. This fact necessitated that we try to keep in mind not only what might count as evidence that our participants approached our various "stimulus materials" with either an essentialist or a narrativist orientation, but also to work out what these orientations would look like through the eyes of the very young. That is, we needed to think about and articulate a scoring scheme that would allow us to characterize juvenile steps that might be taken along the developmental path toward a more fully fledged commitment to an ordinary adult folk conception of essentialism, on the one hand, and narrativism, on the other. The consequence of these efforts has been a coding scheme that recognizes five more or less complex, or adequate, or mature versions of narrativism and essentialism that, briefly put, go roughly as follows.

1. Essentialist Self-Continuity Warranting Strategies In the case of those of our participants who have evidently adopted an essentialist stance, their typical way of responding to the sharply different ways that they or others "appear" at earlier and later moments in their life course was to take the notion of mere "appearances" with special seriousness, by waving off all supposedly more superficial differences in favor of something else that was argued not to have changed at all. With such bits of supposedly enduring something comfortably in hand, all competing evidence of personal change is simply ignored, or discounted, or trivialized as being only partial, or presentational, or hardly to the point of who one claims to be. The several different ways in which our young research participants have articulated such views have supported our efforts to distinguish what are, so far, five distinctive versions of this general essentialist strategy.

a. Simple Inclusion Arguments. The most primitive of these, which we have labeled Simple Inclusion Arguments, are predicated on a picture of selfhood in which each of us is imagined to be no more than a transient, loosely federated, additivelike assemblage of juxtaposed autobiographical bits and pieces that have been haphazardly collaged on and can be just as easily sloughed off. Random pieces of this sort of accidental firmament, part of this loosely conceived main, are seen simply to come and go without remarkable consequences to one's personal

202

Michael J. Chandler

persistence, so long as the remnants include at least one atomic fact that can still be pointed to as the guarantor of one's persistence through time (Chandler, 2001). Respondents who score at this Simple Inclusion level are quick to grant you that they, or Ebenezer Scrooge, or Jean Valjean, have changed in all of the ways pointed out, but happily rest their case for enduring identity on the persistent fact that their name is the same, or that they continue to live at the same address.

b. Topological Self-Continuity Warrants. Rather than merely turning a blind eye to personal change, other of our somewhat older and more mature subjects adopted a different and more Topological conception of their own and other's identities, according to which the self is envisioned as a kind of empty surface structure (not unlike one of those hollow polyhedronic desk calendars with a different facet for every month) that lends itself to being differently viewed from different vantages. Participants scored at this second level are quick to acknowledge apparent changes in themselves and others, but simply discount them as only apparent or presentational, insisting instead that although one or another aspect of one's personality may be shown to be temporarily thrown into eclipse, real change is impossible and amounts to no more than what Shotter (1984) called a spatial repositioning of parts. The idea that some have an angel on one shoulder and a devil on the other is a familiar variation on this topological theme. c. Epigenetic Self-Continuity Warrants. A third variation on this same general essentialistic theme involves subscribing to a kind of Epigenetic or "preformistic" account of character development that, in contrast to the more static representations of the preceding two categories, makes some modest provisions for the diachronic workings of time. What it still fails to do, however, is to allow for the possibility of true novelty. Rather, on this Epigenetic account, the passage of time is thought to simply reveal previously obscured, but still always nascent and preordained aspects of the self--fully canalized parts that have been quietly waiting in the wings for their proper moment of ascendancy. On this account, the psychological changes that typically unfold in adolescence are simply seen to be the loose procedural equivalent of acquiring one's secondary sexual characteristics or the emergence of one's grown-up teeth. d. Frankly "Essentialistic" Self-Continuity Warrants. The fourth in this list of increasingly more complex essentialist solutions to the problem of personal persistence represents an important advance, relative to simple Inclusion, Topological, and Epigenetic arguments, in that it actively subsumes change instead of simply overlooking or frankly denying its existence. The key move in this progressive step toward fully fledged essentialism is the introduction of a new hierarchicalized conception of one's personal attributes that allows some aspects of the self to be

Self-Continuity in Youth

203

seen as more foundational, while others are written off as mere window dressing. Given this distinction between those supposedly deeper lying and definitive things that are thought to form the subterranean and productive core of one's identity, and all of those endless concrete variations that make up the phenotypic surface structure of one's outward life, any change that can be made to fit within the second and more superficial of these categories can be easily discounted as being really beside the point. The paradigm instance of such a warranting strategy might sound something like: True, I once looked like or acted like that, and now look or act like this, but when viewed with a more discerning eye, all of these seemingly radical changes (changes that might otherwise be easily imagined to throw my sense of personal persistence into a tailspin) are easily seen to be mere epiphenomena, and simply fail to touch my stubbornly persistent, rock-bottom self--a self that is capable of productively paraphrasing itself through endless surface variations.

e. Self as Theory (a Hybrid). Finally, among the oldest and the most cognitively mature of the young essentialists so far tested, a very few have expressed some degree of discomfort with the serious prospect of having a core personality "given once and for all, complete and perfect, in an ultimate system of essences" (Schlesinger, 1977, p. 271). What such respondents seem to intuit vaguely is that absolute immutability of the self is too high a price to pay for protection against the threat of ongoing change. Rather, they express a certain disenchantment with fixity and, like certain contemporary developmental theorists (e.g., Berzonsky, 1993; Moshman, 1998), attempt to amend their earlier claims about enduring self structures by marking them as somehow provisional or theorylike. Participants who were scored at this level tended to hold out alternative views about their own or others' personality or character and offered evidence in support of these competing possibilities. 2. Narratively Based Self-Continuity Warranting Strategies In direct contrast to the hundreds of our young respondents who proceeded more essentialistically, by trying to locate some island of sameness in a sea of difference, others, in almost equal numbers, have been found to practice a more narrativistic strategy. Instead of dismissing as irrelevant all of those parts of themselves that have had the temerity to change, they dismiss nothing, apparently assuming, along with Daniel Webster, that the self (like the devil) is to be found in the details. As a consequence, respondents of this general narrativistic stripe found the point of all of our questions less compelling than did their essentialist counterparts, who generally interpreted the demonstration that they or others appeared radically different at two points in time as a direct intellectual challenge to go out and dig up something more enduring. Rather, perhaps because they were more inclined to "take time seriously" (Schlesinger, 1977, p. 271), and to regard themselves as "beings in time" (Flanagan, 1996, p. 67), they typically began by simply repeating

204

Michael J. Chandler

what they took to be key parts of the stories and then appeared puzzled that more was expected of them. What, under duress, most of them eventually did, however, was to retell some kind of story. What distinguishes their efforts is that not all of them made an equally good job of it or even seem to have the same idea about what a story is. The least persuasive of these participants, coded at Level I in our eventual fivepoint coding scheme, evidently had not yet realized that, as E. M. Forster (1954) cautioned, doggedly tracking a chronology "is not the same thing as having a plot" (p. 51). Rather, they simply seemed to go on the assumption that, as Whitehead famously put it, life is merely "one damned thing after another" (Gallagher, 1998, p. 87). Like the lives of Valjean and Scrooge, their lives, as they saw it, originally had them doing this, and later had them doing, or looking, or acting like that. Usually older or otherwise more cognitively mature subjects scored in our coding scheme as being at Level II, generally worked toward creating more foll ow ab le , if still somewhat run-on, stories that had some of the features of what Rorty (1976) and Lightfoot (1997) have characterized as more "Picaresque" tales (think Don Quixote), or even still earlier "Medieval Romances" (imagine Lancelot and his obsessional quest for the Holy Grail). In all of these cases, such respondents seemed to appreciate that the episodes of their own and others' lives needed to be somehow "related," but were generally satisfied to meet this requirement by stringing together key events much like threading beads on a string. That is, they and others were typically understood to possess only that kind of functional identity (Rorty, 1976) that works merely to detail what one does (Chandler, 2001). Beyond these early and essentially plotless stories, we could further distinguish three distinctive levels of functionality. In the first of these (Level III), the past and everything else generally was understood to be stitched together through some oneway relation of cause-and-effect. That is, according to respondents who practiced this kind of self-understanding, the persons we and others eventually become are the inevitable consequence of those events (however accidental) that have set life on its present unwavering course. What needs to be seen as a mark of progress in causal accounts of this "leading-to" variety is that the relations they specify are other than entirely fortutious. Their equally obvious shortcoming is that they are unremittingly fatalisticmone's die is simply cast by a controlling present and a determining past, and nothing can be done but live inside whatever iron cage circumstance builds around us. For something like the reasons just outlined, others of our respondents (those scored at Level IV) generally worked to "liberalize" the obligatory past-to-present nature of more explicitly causal accounts by imagining that they could somehow still shake the dead hand of a determining past. One such escape route lay in the direction of imagining that their first attempt to "discover" that the true plotline dictating the course of their lives was somehow mistaken in ways that were made

Self-Continuity in Youth

205

plain as a result of different and more up-to-date information. Such accounts typically began with some version of "I used to think that . . . . but now I realize that . . . . " In short, these respondents still viewed the process of finding the truth about the storyline of their lives as an act of discovery, but one open to errors of ignorance. Finally (at what was coded as Level V), the discursive arguments offered up in favor of this or that account of one's life were no longer understood as acts of discovery at all, but were seen, rather, as only the latest in a perhaps endless series of possible interpretations or "reenplotments" (Polkinghorne, 1988, p. 182) that succeed, when they do, not by getting things exactly right, but rather by more effectively incorporating not only the current details of one's life, but also earlier though now outmoded attempts to accomplish precisely the same thing (Ricoeur, 1983). As a consequence, the gap that comes to divide Level V Essentialist efforts at "theory revision," and more narratively constructed attempts to redraft, or reenplot the stories of our lives, threatens to vanish.

VI. Real Evidence These different methods and scoring possibilities (only some of which were available at the beginning of this extended research effort) have allowed us to explore a number of different aspects of young people's maturing understanding of their own and others' personal persistence. Four of these research efforts will be briefly discussed here. First, one of these (Chandler et al., 1987), the earliest of the lot, is basically descriptive and normative in character, and details efforts simply to characterize the usual thoughts of mainstream Canadian youth about matters of personal sameness, and to relate their responses to an early version of our story-based assessment procedure to standard Piagetian tests of concrete and formal operational thought. The second (Ball & Chandler, 1989; Chandler & Ball, 1990) concerns the relevance of these measures to a better understanding of youth suicide. The third group of these studies is cross-cultural in character, and addresses the question of whether the previously discussed essentialistically and narratively based approaches to the problem of personal persistence are in any way culturally determined. Finally, I will report on an epidemiological study (Chandler & Lalonde, 1998) that is aimed at relating measures of "cultural continuity" to youth suicide rates in Native and non-Native communities. All of this is, of course, followed by a quick summing up, meant in this case to only briefly repeat the refrain that essential and narrative approaches to the paradox of personal sameness within change are not alternative possibilities to choose between, but are instead cultural resources that are differently exploited by young persons living in different contexts.

206

Michael J. Chandler

A. FALSE STARTS: THE EARLY TOOL-USING STAGE

The first of what became a half-dozen studies intended to detail the developmental course and consequences of young people's thoughts about their own and others' personal persistence (Chandler et al., 1987) was meant primarily as a way of "proofing" an earlier version of the measurement and coding procedures outlined above. As events have turned, some of the interpretive differences that do separate this first look at self-continuity from other more recent efforts are sufficiently large that direct point-by-point comparisons of these findings are not always possible. In particular, my research colleagues and I were initially held sufficiently hostage by underexamined commitments to our own in-house brand of essentialism that we were ill-prepared to see the possibility that narrativist and essentialist approaches to the problem of sameness within change might reasonably compete on an equal footing at every level of developmental maturity. Instead, although our data quickly convinced us that something like what we have gone on calling Simple Inclusion, Topological, and frankly Essentialist self-continuity warranting practices were everywhere to be found in our data, we also noted that only a small fraction of our research participants (approximately 15%) chose to respond by describing their own past as "causally" related to their present, or by otherwise emphasizing possible narrativelike connections bridging earlier and later manifestations of their own and others' lives. Furthermore, we regularly found that in this particular sample of culturally mainstream youth, all of those who adopted anything like a narratively based argument were among the oldest, and most cognitively mature of those being tested. As a consequence, we were led, we now see quite mistakenly, to conclude that this response style was the exclusive province of only those respondents who had tried out essentialism and somehow found it wanting. That is, we were prompted by our data, and perhaps by our own prejudices, to regard narrative views of the self, not as a fair-minded though locally disfavored potential alternative to essentialism, but rather as its natural developmental sequel. Although this now evident misreading of the evidence followed us into the next phase of this work (i.e., into the subsequent studies on adolescent suicide), the fact that so few of our own culturally mainstream subjects actually said or did anything of a narrative nature coincidentally served to empty this interpretive error of much of its practical significance. That is, except for a rump-group of more than usually promising respondents, the large bulk of our study group just were amateur essentialists of one kind or another, and automatically assuming that this is how everyone begins simply made no actionable difference. All of this getting off on the wrong foot aside, our first study of 50 ordinary 10- to 15-year-old Canadian youths did yield two general findings that continue to have current relevance. One of these was the discovery that requiring research participants to begin by commenting on continuities and discontinuities in the lives of various comic book characters did serve the intended purpose of confronting them with clear examples of radical character

Self-Continuity in Youth

207

transformation, and in opening up the topic of personal persistence in ways that they later judged to have relevance to their own lives. In short, our assessment procedure was found to work, and to distribute our participants across the various coding categories that had been envisioned. Second, because all of these participants had been previously screened using standard measures of operational thought, we could also test various hypotheses about the cognitive competencies that are prerequisite to the use of one or another sort of self-continuity warranting practice. Perhaps not surprisingly, the key finding to emerge from this analysis was that only subjects who otherwise met the criteria of being able to reason in formal operational ways ever succeeded in inventing accounts of personal persistence that got beyond the level of "Simple Inclusion" or "Topological" self-continuity warranting practices, or ever envisioned the possibility that by regarding the self as a structure with depth or hierarchy, sameness could be entertained without altogether denying or discounting change. B. GETTING FROM SELF-CONTINUITY TO YOUTH SUICIDE

If, as advertised, self-continuity not only guarantees that we experience ourselves as the rightful owners of our own past, but also works to connect us to our own as yet unrealized future, then any glitch that threatens to cost us an affinity with the individual we once were, or an investment in the person we are en route to becoming, is a dangerous glitch indeed. That is, self-continuity is not only the glue that binds us to our own past, but it is the procedural stuff that ensures that we have some serious stal~e in our unfolding future. Why is it, you might well ask, that when we suffer the various slings and arrows of outrageous fortune, we do not ordinarily shuffle off this mortal coil without so much as a backward glance? The answer--the rub--is that the dead person would be us, wonderful us. All of that is easy to understand. When it comes to the possibility of dying at our own hand, we are not like high-altitude bombardiers, who succeed in obliterating unknown others presumably without emotional consequences, or like perpetrators of some drive-by shooting, who kill coldly and anonymously. We ordinarily have a serious stake in the person we are en route to becoming. We want and expect our own just desserts. Who, without confidence that they are well on their way to becoming the person they imagine that they will eventually be, would ever subject themselves to diets, or to exercise regimens, or to the humiliations of higher education, if not for the fact that the thin, buff, and well-informed person that they are working so hard to become would eventually prove to be dear old them? All of this is why we find it so incredible that youth, with all of life's potential sweetness full upon their lips, should actually be the ones who undertake to kill themselves out of all proportion, and in numbers so high that they can scarcely be believed. How could they do it, we want to know? How could all of those young people, many of whom could scarcely kill a fly, actually take those cruel steps to end their own

208

Michael J. Chandler

lives? But the truth is that they do it in numbers that are wildly out of keeping with their proportionate representation in the general population. By various reports (e.g., Harkavy Friedman, Asnis, Boeck, & DiFiore, 1987; Ross, 1985; Rubenstein Heeren, Housman, Rubin, & Stechler, 1988) the rate of suicidal behaviors among adolescents is some 20 to 200 times greater than that of any other age group. We hear, of course, that adolescent life is extremely stressful, but that will hardly do as an explanation. You, who are reading this, are also often stressed, and likely stressed in ways that you could scarcely have imagined in your youth. Some say that the hormonal ups and downs of puberty are the problem, and that they regularly occasion depressions which are themselves the explanation for the high incidence of youth suicides. But, you wonder, what about my hormones, and are their depressions really all that different from yours or mine? Being somewhat depressed, we are told, is common as clay, and often a mark of good intelligence, and yet suicide is rare at any age. Do stressors of almost every description not multiply with the obligations of maturity? But suicide shows no such escalating picture. If you survive your adolescence, everything else being equal, the chances that you will undertake to end your own life drop precipitously. All of this is one of the big puzzles, and what ordinarily pass for explanations will rarely do. Here is a different, but possible explanation that is meant to apply in at least some of these puzzle cases. The reason that certain young people are able to contemplate their own death, and are often so quick to take steps to bring it about, is that, for them, there often is no "rub." That is, although you and I are perfectly convinced that the person who would remain dead at the end of any of our own successful attempts at suicide would be decidedly us, adolescents are often less sure. Unlike ourselves, who are settled in our ways of warranting self-continuity, young people are typically in the midst of trying on and discarding one feeble self-continuity warranting strategy after another, and like some soft-bellied thing that has just shed its old skin, they are sometimes caught out naked and without visible means of supporting their own personal persistence. Or, otherwise put, it is because in so quickly climbing the steps of their own identity development, they are frequently caught out with both feet off the ground, that they are, in consequence, so easily thrown off balance, and end up taking their own lives over lost loves and childish failures that (should they survive them) would often produce, in retrospect, little more than a wry smile. If something like this is true, then no matter how many philosophers assure us that continuity is the universal constitutive condition of selfhood, there will always be adolescents who contradict them by losing the thread of their own persistence, and with it, any stake in their own future. On this same logic, seriously suicidal adolescents--that is, those who have seriously attempted, but inadvertently failed to kill themselves--should also be uniquely characterized by being just those young persons who currently lack any workable way of effectively warranting their own personal persistence in time.

Self-Continuity in Youth

209

1. A Comparison between Suicidal and Nonsuicidal Youth This is precisely the search pattern that some of my colleagues and I took up (Ball & Chandler, 1989; Chandler & Ball, 1990). What we did, in short, was to test 20 actively suicidal and psychiatrically hospitalized adolescents, along with equivalent groups of their nonsuicidal age mates (both in and out of acute psychiatric hospitals), all in an effort to see what, if any, self-continuity warranting strategies they happened to have at their disposal. The answer was very clear. Although the members of our psychiatric and nonpsychiatric control groups were all over the map in terms of their preferred way of arguing in favor of their own temporal persistence, they all (with one exception) had some way, usually an essentialist way, of projecting themselves forward and backward in their own life history. By contrast, 85% of the actively suicidal subjects simply failed to find any way of understanding themselves as continuous in time. They often remembered and discounted earlier ideas they had once had on the subject, but for now, they came up entirely empty-handed. They simply could find or offer no personally convincing reasons that the person they now were, had earlier been, or would eventually become could be meaningfully counted as one and the same numerically identical individual. They were not, as you might have suspected, more depressed than the nonsuicidal but psychiatrically hospitalized counterparts, nor were they otherwise more profoundly ill. They simply could no longer find what they took to be acceptable grounds for imagining themselves as continuous in time. This finding is significant, if one may say so about one's own work. Not just statistically significant (which it was), but significant in the human sense. For the reasons outlined previously, suicidal adolescents and their appropriately matched nonsuicidal counterparts are typically found to be as indistinguishable as your usual peas in a pod. These data are not like that. Such findings, attached as they are to an account of the process of adolescent identity formation, help to make sense of the joint facts that teenagers both take more steps to end their own lives than do members of any other age group, and that, when unsuccessful, they ordinarily disappear back into the general population of those who would not kill themselves if their lives depended on it. C. CULTURE AS A SET-POINT IN THE CHOICE BETWEEN NARRATIVIST AND ESSENTIALIST SELF-CONTINUITY WARRANTING PRACTICES

Our next move in this extended study sequence (Chandler et al., 2001) was an attempt to describe, and to understand, the consequences of possible differences in the self-continuity warranting practices of various cultural groups. Other moves could have been made, but a long leap-to-mind list of convergent reasons recommends the reasonableness of this particular choice of study problems. The first one of these is that, although self-continuity is widely acknowledged as a necessary design feature, or constitutive condition, for selfhood in all possible

210

Michael J. Chandler

worlds (Chandler, 2001) and so potentially deserves its place in that small bag of transcultural things belonging to all people in all times (Harr6, 1980), it is also an uncontested fact that what selves are taken to be also varies dramatically across cultures (e.g., Holland, 1997; Lillard, 1998). Given that both of these things are true at once, we have good reasons to hope that coming to some better understanding of how personal persistence is differently negotiated in different cultures is a potentially important key in resolving this apparent paradox. Second, a frequent argument, particularly by those who see narrativity as the natural language of the self (Mink, 1969), is that essentialism in all of its forms is best regarded as just another of those unfortunate by-products of Western European history's chronic preoccupation with all things hierarchical, phallocentric, and abstract. On this account, essentialism is seen to be suspiciously entangled with the power politics of empire and its decline, and it has little to do with the prevailing narrative frame that, according to many ethnographers (e.g., Ewing, 1990), is used by most of the world's cultures as a primary means of interpreting human experience. As Hirschfeld (1995) has argued, for example, "in many, perhaps most, cultures there is a marked absence of discourse that explains human behavior in terms of transsituationally stable.., properties captured by explanations of trait or disposition" (p. 315). Although this is not the place to attempt a serious rehashing of this increasingly shopworn universalist-culturalist debate within anthropology and cultural psychology (see Holland, 1997), even a superficial reading of the controversy reveals clearly enough that a more than ample precedent exists (e.g., Fiske, Kitayama, Markus, & Nisbett, 1998; Markus & Kitayama, 1991; Miller, 1996) for imagining that young persons from industrialized and nonindustrialized, or Euro-American versus "other," non-Western cultures, may differ in their predisposition to favor either narrativist or essentialist self-continuity warranting strategies. What has been said so far about the likely role of ethnographic factors in shaping the course of identity development is perhaps enough to recommend a decision to go on to do the hard work of searching out possible cross-cultural differences in the self-continuity warranting practices of different ethnic or cultural group. It is not enough, however, to settle the question of where, in particular, one ought to begin lookingma choice that, given Canada's unprecedented cultural diversity, is minimally constrained. In the end, the comparison settled upon was between culturally mainstream Canadian youth and their Native or First Nations counterparts, drawn from two more and less remote aboriginal communities. To date, we have systematically administered our previously described self-continuity warranting measures to a total of 120 twelve- to eighteen-year-old Native adolescents, along with a reasonably comparable group of non-Native youth. Availability (although scarcely "easy" availability) was clearly a factor in this choice, but the decision to focus on a comparison between Native and non-Native youth was driven by more than simple feasibility. If, as certain sectors of the

Self-Continuity in Youth

211

contemporary literature in cultural psychology urge, "individualism versus collectivism" is a principal dichotomy at work in distinguishing contemporary world cultures (e.g., Markus & Kitayama, 1991; Miller, 1996), then our own Native and non-Native comparison groups hold out a reasonable promise of ensuring variability on this much debated dimension (for a critique of the individualismcollectivism dichotomy, see Turiel & Wainryb, 1994). Similarly, if the dimension of relevance is industrialized, or even Euro-American, versus nonindustrialized or "other," then the inclusion of aboriginal cultures, whose sustained contact with Western intellectual traditions is scarcely more than 100 years old, is again a defensible choice. Finally, because our own earlier research (e.g., Ball & Chandler, 1989; Chandler & Ball, 1990) indicated that youth suicide can be partially understood as a consequence of a loss of individual self-continuity, and because the rate of youth suicide in certain First Nations communities is known to be alarmingly high--higher, perhaps, than that of any other identifiable cultural group in the world--(Kirmayer, 1994), we had grounds for hope that such a program of study might contribute usefully to a better understanding of this tragic social problem. For all of the reasons detailed earlier, then, my research colleagues and I sought and eventually obtained the cooperation of two aboriginal communities. One of these is located on the apron of a major urban center, a short walk from the nearest shopping mall. The other occupies a village located on an island 30 miles of open sea beyond the last road on Canada's rugged West Coast. Although the full details of a point-by-point comparison between the selfcontinuity warranting practices of young persons from both of these Native communities, along with those of their non-Native counterparts, are not yet available, the bold outlines of these results are already clear enough. The members of the non-Native comparison group, like other culturally mainstream youth tested before them, strongly favored essentialist self-continuity warranting strategies. That is, more than 80% of them regularly rooted their convictions that they are diachronically continuous on the presumption of some persistent, essential aspect or feature of their identity, situated at some more or less subterranean level of internality, that manages to stand apart from the ravages of time and works to vouchsafe their own personal persistence (Chandler et al., 2001). What tends to change in the course of their ontogenetic development, this research again suggests, is the level of abstraction or complexity at which the structural claims of these 12- to 18-year-olds are pitched. The youngest of this new group almost always relied on some Simple Inclusion or Topological form of essentialistic self-continuity warranting, whereas older adolescents more often tended to employ "Epigenetic" strategies, or to invent an identity structure with sufficient depth to allow for the possibility that changes in their outward way of being could be easily written off as merely "superficial" changes in appearance, changes that did not reach the fixed, essential core of their identity. As before, only a few of the oldest among this new comparison group ended up calling even this classical essentialist notion into question by suggesting

212

Michael J. Chandler

that their ideas about having some "true" or fixed personality structure were, after all, only "theories" and therefore at best provisional. By contrast, the young First Nations participants in this study only occasionally attempted to warrant their own or others' self-continuity by alluding to the existence of some hidden essential something. Instead, roughly two thirds of them undertook to frame some story about themselves, or to provide an account of the films or comic books they had been shown, that was meant to link up their own or others earlier and later ways of being through some sort of narrative structure. Of course, most of these young respondents did not actually succeed in producing anything that would properly qualify as fully fledged Bildungsromane in which later ways of being necessarily grew out of, or interpretively subsumed, what had gone before. Rather, the large bulk of their efforts more often took the form of simple chronologies (Level I, in the narrative coding scheme outlined earlier) or "picaresque" and "Medieval romances" (Level II) that simply strung the events of their lives (and those of the story characters) together like so many beads on a string. Others came closer still to having generated plots of their own, by reeling off putative chains of cause and effect that were meant to explain how they began one way and ended up another (Level III). Some did succeed, however, in adopting a more genuinely interpretive stance toward their lives, either by describing their efforts to "find" the hidden plotline in their still short autobiographies (Level IV) or by characterizing themselves as being involved in what was taken to be the never ending job of revising their life stories (Level V). Some cases were obvious crossovers (i.e., again some 15% of the culturally mainstream respondents adopted a narrativist stance, and roughly a third of the Native participants were committed essentialists), but on the whole the data offer clear support for the view that cultural factors do play an important role is determining the likelihood that one will exploit either narrativistic or essentialistic strategies in attempting to solve the problem of one's own and others' persistence in time. Not everything, of course, worked as planned. As became evident when a forest of satellite dishes greeted the research team as it approached our "remote" testing site, simple physical distance is no longer a reliable metric of how culturally "isolated" a given community might be. Nor were there evident psychometric differences between the young volunteers from the two Native communities chosen. What is perhaps more remarkable is that, however wired or plugged into Euro-American culture, on the subject of identity formation First Nations youth are still altogether more like one another and less like their culturally mainstream counterparts than anything about their often high-tech lives would cause you to anticipate. This particular fact of the mattermthe fact that First Nations youth are characteristically committed to an understanding of their own self-continuity that is rooted more in narrativity than in the possibility of essential samenessmwhen related to the grisly evidence that Native youth take their own lives at a rate that is often hundreds of times greater than their non-Native counterparts, suggests a

Self-Continuity in Youth

213

hypotheses that my research colleagues and I have worked to evaluate. Given, as we have shown, that youth suicide is strongly associated with failures in selfcontinuity, and given, as we have shown, that First Nations youth are especially reliant upon a narrative-based form of self-understanding, and given, as everyone already knows, that the collective cultural narratives of aboriginal peoples have been forbidden and criminalized and turned into a laughingstock, then what is the chance that all of these matters are unrelated? The final research to be summarized here is meant as a way of addressing this question. D. CULTURAL CONTINUITY AS A PROTECTIVE FACTOR AGAINST SUICIDE IN FIRST NATIONS YOUTH

Insisting, as has been done in these pages, that there is not some blind scale of justice upon which to weigh essentialist and narrativist claims about personal persistence is not, it needs to be said, the same thing as imagining that both are equally effective in dealing with every possible sort of radical change that circumstance might throw in one's path. Of course we all are, to one degree or another, living through a time of dramatic historical and cultural change that deeply undermines our collective and individual efforts to ground our identity. "May your children live in interesting times" is, consequently, a curse whose meaning is all too clear-cultural continuity and personal continuity are not names for two different things, but different names of adjacent parts of the same thing. Under circumstances such as our own, when "whirl is king" (Schlesinger, 1977, p. 279), essentialism (although not without alienating costs of its own) may well be seen to confer a certain short-run advantage by making the place to which one turns for sameness a private (as opposed to public) and a deep lying (as opposed to an interactive) place, a place where not only time, but nothing and nobody can easily get at you. Narrativistic accounts, by contrast, would seem to fair best during those relatively flat places in cultural history that provide enough wind-up room for your story and the stories of those around you to work themselves into a common groove. In such shared circumstances the faltering identities associated with adolescence are comfortably back-stopped by the fact that (should you happen to lose its thread momentarily) the plotline of your life and mine is already a familiar story, known and agreed upon by all. What, given all of this, would not be good would be to be someone (particularly an adolescent someone, with built-in identity problems) who happens to come equipped with both a narrative appreciation of the problem of personal persistence and a culture on the brink of extinction. This, of course, is exactly the fate of contemporary First Nations youth, whose typical approach to managing problems of personal persistence is predominately narratively based. After apparently thousands of years of adaptive success, their culture's religious practices were criminalized, their language was forbidden, the right of parents to rear and educate

214

Michael J. Chandler

their children was suspended, all as a matter of official policy of the Canadian government. And, as should surprise no one, their rate of suicide is clocked as being dramatically higher (three to five times higher) than that of perhaps any other cultural group in the world (Kirmayer, 1994). The story, however, still has some surprises left. Here are two of them. The first of these is important because of its extravagance. Many people know about the high rate of suicide among aboriginal groups, and few if any would be surprised if told that the rate of youth suicide is marginally higher in some Native communities than in others. Such marginal differences would be expected even if, against all reason, the real risks of dying at one's own hand were to arise as the simple and sovereign result of merely being a First Nations person. Given the small numbers involved, a difference of two, or even tenfold, would not seem a statistical impossibility. Our data are not like that. As it is, more than half of the almost 200 bands in British Columbia experienced no youth suicides during the reporting period, while others suffered suicide rates more than 800 times the Canadian national average. Even at the level of band councils, which aggregate between 5 and 15 separate bands, some 90% of the known suicides were found to occur in less than 10% of these extended communities. The simple truth to be found in all of this extreme variability is that the high rate of suicide said to characterize First Nations youth as a whole is an arithmetic fiction, and cannot be explained simply by pointing to whatever factors might serve to distinguish aboriginal from nonaboriginal people. Fact two--surprise twompresupposes the first, and is meant as a partial answer to the question of w h y the youth suicide rate is so high in some First Nations communities and so low in others. Clearly, some bands are successful in insulating their young from suicide in ways that others are not, and it hardly puts too fine a point on it to say that figuring out how they manage to do so is a very real matter of life and death. In light of the strong links hypothesized between individual and cultural continuity, our own search for such protective factors took the form of attempting to identify possible markers that might index community-based efforts to reconstruct a form of cultural continuity previously swept away by a decades old public policy of "ethnic cleansing." To this end we specified six readily available proxy measures of cultural continuity including evidence of efforts (a) to secure aboriginal title to traditional lands; (b) to achieve a measure of self-government; (c) to secure some degree of community control over educational, (d) health, and (e) police and fire protection services; and (f) to establish within their communities certain officially recognized "cultural facilities" to help preserve and enrich their cultural lives. The results of these efforts to build some quantitative measure of cultural continuity are unambiguous. In every band or band council in which all six of these "protective factors" were in place, no youth suicides were reported during the

Self-Continuity in Youth

215

5-year study window. By contrast, communities that lacked evidence of having taken only one or some of these steps toward the preservation of their cultural heritage experienced suicide rates 5 to 100 times the national average. The strong conclusion supported by these findings is that the frequency with which the young First Nations persons kill themselves appears to be a direct function of a series of factors that, seen in the aggregate, represents community-based efforts to reconstitute a form of cultural continuity.

VII. Summary of Conclusions: Five Easy Pieces Without even attempting to revisit all of the diverse matters touched on in this long account, I would offer the following short-list of conclusions that would be good for the reader to take away. First, I hope the case has been made that recourse to some conceptual means of preserving a sense of personal persistence is a parameter of self-understanding common to all human cultures. Second, notwithstanding a multitude of exclusionary claims to the contrary, neither essentialism nor narrativity qualifies as the natural language of the self. Instead, both of these irreducibly different self-continuity warranting strategies remain as exploitable resources, available for use in accomplishing the performative task of justifying selfsameness in the face of inevitable change. Third, those adolescents who do not successfully enlist either of these selfcontinuity warranting strategies suffer, as a natural consequence, a loss of connectedness to their own future and are thereby placed at special risk for suicide. Fourth, different cultures promote different approaches to the problem of personal persistence, with essentialist strategies more favored among those young persons who are products of Euro-American cultures, and narrative strategies more favored among aboriginal adolescents. Fifth, individual and cultural continuity are strongly linked, such that First Nations communities (and perhaps other groups) that are excluded from the mainstream culture and that fail to preserve their heritage culture are dramatically less successful in insulating their youth against the risks of suicide.

REFERENCES Aboud, F. E., & Ruble, D. N. (1987). Identity constancy in children: Development processes and implications. In T. Honess & K. Yardley (Eds.), Self and identity: Perspectives across the lifespan (pp. 95-107). Boston: Routledge & Kegan Paul. Atran, S. (1994). Core domains versus scientific theories: Evidence from systematics and Itza-Maya folkbiology. In L. A. Hirschfeld & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 316-340). New York: Cambridge University Press.

216

Michael J. Chandler

Bakhtin, M. M. (1981). The dialogic imagination: Four essays by M. M. Bakhtin (M. E. Holquist, Ed., C. Emerson & M. Holquist, Trans.). Austin, TX: University of Texas Press. Bakhtin, M. M. (1986). Speech genres and other late essays (C. Emerson & M. Holquist, Eds.; V. W. McGee, Trans.). Austin, TX: University of Texas Press. Ball, L., & Chandler, M. J. (1989). Identity formation in suicidal and non-suicidal youth: The role of self-continuity. Development and Psychopathology, 1, 257-275. Berzonsky, M. (1993). A constructivist view of identity development: People as postpositivist selftheorists. In J. Kroger (Ed.), Discussions on ego identity (pp. 169-203). Hillsdale, NJ: Erlbaum. Brockelman, P. (1985). Time and self New York: Crossroads. Bruner, J. S. (1986). Actual minds, possible worlds. Cambridge, MA: Harvard University Press. Bruner, J. S. (1990). Acts of meaning. Cambridge, MA: Harvard University Press. Bruner, J. S., & Kalmar, D. A. (1998). Narrative and metanarrative in the construction of the self. In M. D. Ferrari & R. J. Sternberg (Eds.), Self-awareness: Its nature and development (pp. 301331). New York: Guildford. Camus, A. (1955). The myth of Sisyphus and other essays. London: Hamish Hamilton. Carr, D. (1986). Time, narrative and history. Bloomington, IN: Indiana University Press. Cassirer, E. (1923). Substance and function. Chicago: Open Court. Chandler, M. J. (2001). Surviving time: The persistence of identity in this culture and that. Culture and Psychology, 6, 209-231. Chandler, M. J., & Ball, L. (1990). Continuity and commitment: A developmental analysis of the identity formation process in suicidal and non-suicidal youth. In H. Bosma & S. Jackson (Eds.), Coping and self-concept in adolescence (pp. 149-166). New York: Springer-Verlag. Chandler, M., Boyes, M., Ball, L., & Hala, S. (1987). The conservation of selfhood: A developmental analysis of children's changing conceptions of self-continuity. In T. Honess & K. Yardley (Eds.), Self and identity: Perspectives across the lifespan (pp. 108-133). Boston: Routedge & Kegan Paul. Chandler, M. J., & Lalonde, C. (1998). Cultural continuity as a hedge against suicide in Canada's First Nations. Transcultural Psychiatry, 35, 191-219. Chandler, M. J., Lalonde, C. E., & Sokol, B. (2001). Continuities of selfhood in the face of radical development and cultural change. In L. Nucci, G. Saxe, & E. Turiel (Eds.), Culture, thought, and development (pp. 65-84). Mahwah, NJ: Erlbaum. Chandler, M. J., & Sokol, B. (in press). Level this, level that: The place of culture in the construction of the self. In C. Raeff & J. B. Benson (Eds.), Culture and development: Essays in honor of Ina Uzgiris. New York: Routledge. Charme, S. T. (1984). Meaning and myth in the study of lives: A Sartrean perspective. Philadelphia: University of Pennsylvania Press. Choi, I., Nisbett, R. E., & Norenzayan, A. (1999). Causal attribution across cultures: Variation and universality. Psychological Bulletin, 125, 47-63. Cohler, B. J. (1982). Personal narrative and the life course. In P. B. Balthes & O. G. Brim, Jr. (Eds.), Life-span development and behavior (Vol. 4, pp. 205-241). New York: Academic Press. Cohler, B. J., & Cole, T. R. (1994). Studying older lives: Reciprocal acts of telling. Paper presented at the meeting of the Society for Personology, Ann Arbor, MI. Dennett, D. C. (1992). The self as a center of narrative gravity. In E S. Kessel & E M. Cole (Eds.), Self and consciousness: Multiple perspectives (pp. 103-115). Hillsdale, NJ: Erlbaum. DeVries, R. (1969). Constancy of generic identity in the years three to six. Monographs of the Society for Research in Child Development, 34 (3, Serial No. 127), 56-60. Dilthey, W. (1962). Pattern and meaning in history: Thoughts on history and society (H. P. Rickman, Ed.). New York: Harper & Row. Dobert, R., Habermas, J., & Nunner-Winkler, G. (1987). The development of the self. In J. M. Broughton (Ed.), Critical theories of psychological development (pp. 275-301). New York: Plenum.

Self-Continuity in Youth

217

Erikson, E. H. (1968). Identity: Youth and crisis. New York: Norton. Ewing, K. P. (1990). The illusion of wholeness: Culture, self and the experience of inconsistency. Ethos, 18, 251-278. Feldman, C., Bruner, J., Kalmar, D., & Renderer, B. (1993). Plot, plight, and dramatism: Interpretation at three ages. Human Development, 36, 327-342. Fiske, A. E, Kitayama, S., Markus, H. R., & Nisbett, R. E. (1998). The cultural matrix of social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook ofsocial psychology (4th ed., Vol. 2, pp. 915-981). Boston: McGraw-Hill. Flanagan, O. (1996). Self expressions: Mind, morals and the meaning of life. New York: Oxford University Press. Forster, E. M. (1954). Aspects of the novel. New York: Harcourt, Brace & World. (Original work published 1927.) Fraisse, E (1963). The psychology of time. New York: Harper & Row. Frye, N. (1957). Anatomy of criticism: Four essays. Princeton, NJ: Princeton University Press. Gallagher, S. (1998). The inordinance of time. Evanston, IL: Northwestern University Press. Geertz, C. (1973). The interpretation of cultures: Selected essays. New York: Basic Books. Geertz, C. (1975). On the nature of anthropological understanding. American Scientist, 63, 47-53. Gelman, S. A. (1999). The role of essentialism in children's concepts. In H. W. Reese (Ed.), Advances in child development and behavior (Vol. 27, pp. 55-98). San Diego: Academic Press. Gelman, S. A., Coley, J. D., & Gottfried, G. M. (1994). Essentialist beliefs in children: The acquisition of concepts and theories. In L. A. Hirschfield & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 341-365). New York: Cambridge University Press. Gelman, S., & Wellman, H. (1990). Insides and essences: Early understandings of the nonobvious. Cognition, 38, 213-244. Gergen, K. J., & Gergen, M. M. (1986). Narrative from and the construction of psychological science. In T. R. Sarbin (Ed.), Narrative psychology: The storied nature of human conduct (pp. 22-44). New York: Praeger. Giddens, A. (1991). Modernity and self-identity: Self and society in the late modern age. Stanford, CA: Stanford University Press. Guardo, C. J., & Bohan, J. B. (1971). The development of a sense of self-identity in children. Child Development, 42, 1909-1921. Gutheil, G., & Rosengren, K. S. (1996). A rose by any other name: Preschoolers' understanding of individual identity across name and appearance changes. British Journal of Developmental Psychology, 14, 477-498. Habermas, J. (1985). Questions and counterquestions (J. Bohman, Trans.). In R. J. Bernstein (Ed.), Habermas and modernity (pp. 192-216). Cambridge, MA: MIT Press. Habermas, J. (1991). An intersubjective concept of individuality. Journal of Chinese Philosophy, 18, 133-141. Hall, D. G. (1998). Continuity and the persistence of objects: When the whole is greater than the sum of the parts. Cognitive Psychology, 37, 28-59. Hallowell, A. I. (1976). Contributions to anthropology: Selected papers of A. Irving Hallowell. Chicago: University of Chicago Press. Harkavy Friedman, J., Asnis, G., Boeck, M., & DiFiore, J. (1987). Prevalence of specific suicidal behavior in a high school sample. American Journal of Psychiatry, 144, 1203-1206. Harr6, R. (1980). Social being: A theory for social psychology. Lanham, NJ: Littlefield Adams. Harr6, R. (1989). The self as a theoretical concept. In M. Krausz (Ed.), Relativism, interpretation and confrontation (pp. 387-417). South Bend, IN: University of Notre Dame Press. Hermans, H. J. (1996). Voicing the self: From information processing to dialogical interchange. Psychological Bulletin, 119, 31-50. Hermans, H. J., & Kempen, H. J. (1993). The dialogical self New York: Academic Press.

218

Michael J. Chandler

Hermans, H. J., Kempen, H. J., & Van Loon, R. J. (1992). The dialogical self: Beyond individualism and rationalism. American Psychologist, 47, 23-33. Hirsch, E. (1976). The persistence of objects. Philadelphia: University City Science Center. Hirschfeld, L. A. (1995). Anthropology, psychology, and the meanings of social causality. In D. Sperber & D. Premack (Eds.), Causal cognition: A multidisciplinary debate (pp. 313-350). New York: Oxford University Press. Holland, D. (1997). Selves as cultured: As told by an anthropologist who lacks a soul. In R. D. Ashmore & L. J. Jussin (Eds.), Self and identity: Fundamental issues (pp. 160-190). New York: Oxford University Press. Howard, G. S. (1991). Culture tales: A narrative approach to thinking, cross-cultural psychology, and psychotherapy. American Psychologist, 46, 187-197. James, W. (1902). Varieties of religious experience. New York: Modern Library. Johnson, C. N. (1990). If you had my brain, where would I be? Children's understanding of the brain and identity. Child Development, 61, 962-972. Keil, E C. (1989). Concepts, kinds, and cognitive development. Cambridge, MA: MIT Press. Kerby, A. E (1991). Narrative and the self Bloomington, IN: Indiana University Press. Kirmayer, L. J. (1994). Suicide among Canadian aboriginal peoples. Transcultural Psychiatric Research Review, 31, 3-58. Kotre, J. (1984). Outliving the self" Generativity and the interpretation of lives. Baltimore, MD: Johns Hopkins Press. Lacan, J. (1968). The language of the self" The function of language in psychoanalysis. Baltimore, MD: Johns Hopkins Press. Levine, R. A., & White, M. I. (1986). Human conditions: The cultural basis of educational development. New York: Routledge & Kegan Paul. Lewis, M., & Ferrari, M. in press. Cognitive-emotional self-organization in personality development and personal identity. In H. A. Bosma & E. S. Kunnen (Eds.), Identity and emotions: A selforganizational perspective. Cambridge, England: Cambridge University Press. Lifton, R. J. (1970). Boundaries: Psychological man in revolution. New York: Random House. Lightfoot, C. (1994, June). Literary forms of the hero, and adolescents' narratives of risk-taking. Chicago, IL. Paper presented at the meeting of the Jean Piaget Society, Lightfoot, C. (1997). The culture of adolescent risk-taking. New York: Guilford. Lightfoot, C. (1999, June). Between self-continuity and self-transformation: Adolescents' conceptions of the quest for identity. In C. E. Lalonde (Chair), The persistence of identity: Developmental and cross-cultural explorations. Mexico City, Mexico. Symposium conducted at the meeting of the Jean Piaget Society, Lillard, A. (1998). Ethnopsychologies: Cultural variations in theories of mind. Psychological Bulletin, 123, 3-32. Linde, C. (1990). Life stories: The creation of coherence. Palo Alto, CA: Institute for Research on Learning. Luckman, T. (1979). Personal identity as an evolutionary and historical problem. In M. von Cranach (Ed.), Human ethology: Claims and limits of a new discipline (pp. 56-74). New York: Cambridge University Press. Luk~cs, G. (1914). Zur Soziologie des modernen Dramas. Archivfiir Sozialwissenschaft und Sozialpolitik, 38, 303-345. MacIntyre, A. (1977). Epistemological crises, dramatic narrative and the philosophy of science. The Monist, 60, 453-472. MacIntyre, A. (1981). After virtue: A study in moral theory. Notre Dame, IN: University of Notre Dame Press. Mandler, J. M. (1984). Stories, scripts, and scenes: Aspects of schema theory. Hillsdale, NJ: Erlbaum. Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224-253.

Self-Continuity in Youth

219

Mayr, E. (1991 ). One long argument: Charles Darwin and the genesis of modern evolutionary thought. Cambridge, MA: Harvard University Press. McAdams, D. E (1984). Love, power, and images of the self. In C. Z. Malatesta & C. E. Izard (Eds.), Emotion in adult development (pp. 159-174). Beverly Hills, CA: Sage. McAdams, D. E (1985). Power, intimacy, and the life story: Personological inquiries into identity. New York: Guildford Press. McAdams, D. E (1990). Unity and purpose in human lives: The emergence of identity as a life story. In A. I. Rabin, R. A. Zucker, R. A. Emmons, & S. Frank (Eds.), Studying persons and lives (pp. 148-200). New York: Springer. McAdams, D. P. (1993). The stories we live by: Personal myths and the making of the self New York: William Morrow. McAdams, D. E (1997). The case for unity in the (post)modern self: A modest proposal. In R. D. Ashmore & L. J. Jussim, Self and identity: Fundamental issues (pp. 46-78). New York: Oxford University Press. McAdams, D. E, Diamond, A., de St. Aubin, E., & Mansfield, E. (1997). Stories of commitment: The psychosocial construction of generative lives. Journal of Personality and Social Psychology, 72, 678-694. Medin, D. L. (1989). Concepts and conceptual structure. American Psychologist, 44, 1469-1481. Medin, D. L., & Ortony, A. (1989). Psychological essentialism. In S. Vosniadou & A. Ortony (Eds.), Similarity and analogical reasoning (pp. 179-195). New York: Cambridge University Press. Miller, J. G. (1996). Theoretical issues in cultural psychology. In J. W. Berry, Y. H. Poortinga & J. Pandley (Eds.), Handbook of cross-cultrual psychology: Vol. 1. Theory and method (pp. 85128). Boston: Allyn & Bacon. Mink, L. O. (1969). History and fiction as modes of comprehension. New Literary History, 1, 541-558. Mink, L. O. (1978). Narrative form as a cognitive instrument. In R. H. Canary & H. Kozicki (Eds.), The writing of history: Literary form and historical understanding (pp. 129-150). Madison, WI: University of Wisconsin Press. Mohr, D. M. (1978). Development of attributes of personal identity. Developmental Psychology, 14, 427-428. Moshman, D. (1998). Identity as a theory of oneself. Genetic Epistemologist, 26, 1-14. Murray, D. W. (1993). What is the Western concept of the self? On forgetting David Hume. Ethos, 21, 3-23. Nicolopoulou, A. (1997). Worldmaking and identity formation in children's narrative play-acting. In B. D. Cox & C. Lightfoot (Eds.), Sociogenetic perspectives on internalization (pp. 157-187). Mahwah, NJ: Erlbaum. Norenzayan, A. (1999, December). Formal versus intuitive thought: The cultural psychology of reasoning. Vancouver, Canada. Paper presented at colloquium, Department of Psychology, University of British Columbia, Norenzayan, A., Choi, I., & Nisbett, R. E. (1999). Eastern and Western perceptions of causality for social behavior: Lay theories about personalities and situations. In D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding and overcoming group conflict (pp. 239-272). New York: Sage. Oppenheimer, L., Mur, Y., Koeman, H., & Chandler, M. (1984). Children's conceptions of selfhood. In K. E. Grossman & P. Lutkeuhus (Eds.), Bericht uber die achte tagung in entwicklungspsychologie (pp. 432-433). Regensburg, Germany: University of Regensburg Press. Overton, W. E ( 1991). Metaphor, recursive systems, and paradox in science and developmental theory. In H. W. Reese (Ed.), Advances in child development and behavior (Vol. 23, pp. 59-71 ). San Diego: Academic Press. Overton, W. E (1998). Developmental psychology: Philosophy, concepts, and methodology. In W. Damon (Series Ed.) & R. M. Lerner (Vol. Ed.), The handbook of child psychology: Vol. 1. Theoretical models of human development (5th ed., pp. 107-188). New York: Wiley.

220

Michael J. Chandler

Parfit, D. (1971). Personal identity. 80, 3-27. Peacock, J. L., & Holland, D. C. (1993). The natural self: Life stories in process. Ethos, 21, 367-383. Perry, J. (1976). The importance of being identical. In A. O. Rorty (Ed.), The identities ofpersons (pp. 67-90). Berkeley: University of California Press. Piaget, J. (1951). Play, dreams, and imitation in childhood. New York: Norton. Piaget, J. (1968). On the development of memory and identity (E. Duckworth, Trans.). Barre, MA: Clark University Press. Polkinghorne, D. (1988). Narrative knowing and the human sciences. Albany, NY: SUNY Press. Propp, V., Wagner, L. A., Scott, L., & Dundes, A. (1968). Morphology of the folktale. Austin: University of Texas Press. Putnam, H. (1982). Natural kinds and human artifacts. Mind, 91, 418-419. Putnam, H. (1988). Representation and reality. Cambridge, MA: MIT Press. Reese, H. W. (1991). Contextualism and developmental psychology. In H. W. Reese (Ed.), Advances in child development and behavior (Vol. 23, pp. 187-230). San Diego: Academic Press. Ricoeur, P. (1980). Narrative time. Critical Inquiry, 7, 165-186. Ricoeur, P. (1983). Can fiction narratives be true? Analecta Husserliana, 14, 3-19. Ricoeur, P. (1985). History as narrative and practice. Philosophy Today, 29, 213-222. Ring, M. (1987). Beginning with the pre-Socratics. Mountain View, CA: Mayfield. Rorty, A. O. (1973). The transformations of persons. Philosophy, 48, 261-275. Rorty, A. O. (1976). A literary postscript. In A. O. Rorty (Ed.), The identities ofpersons (pp. 301-323). Berkeley: University of California Press. Rorty, A. O. (1987). Persons as rhetorical categories. Social Research, 54, 55-72. Rosengren, K. S. (1999, June). The role of essentialism in children's understanding of identity over change. In C. E. Lalonde (Chair), The persistence of identity: Developmental and cross-cultural explorations. Mexico City, Mexico. Symposium conducted at the meeting of the Jean Piaget Society, Ross, C. P. (1985). Teaching children facts of life and death: Suicide prevention in the schools. In M. L. Peck, N. L. Faberow, & R. E. Litman (Eds.), Youth Suicide (pp. 147-169). New York: Springer. Ross, L., & Nisbett, R. E. (1991). The person and the situation: Essential contributions of social psychology. New York: McGraw-Hill. Rubenstein, J. L., Heeren, T., Housman, D., Rubin, C., & Stechler, G. (1988, March). Suicidal behavior in "normal" adolescents: Risk and protective factors. Alexandria, VA. Paper presented at the biennial meeting of the Society for Research in Adolescence, Russell, B. (1959). My philosophical development. New York: Simon & Schuster. Sarbin, T. R. (1986). 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, L. J. Hayes, H. W. Reese, & T. R. Sarbin (Eds.), Varieties of scientific contextualism (pp. 51-65). Reno, NV: Context Press. Sass, L. A. (1988). Humanism, hermeneutics, and the concept of the human subject. In S. B. Messer, L. A. Sass, & R. L. Woolfolk (Eds.), Hermeneutics and psychological theory: Vol. 2. Interpretive perspectives on personality, psychotherapy, and psychopathology (pp. 222-271). New Brunswick, NJ: Rutgers University Press. Schlesinger, A. (1977). The modern consciousness and the winged chariot. In B. Gorman and A. Wessman (Eds.), The personal experience of time (pp. 268-288). New York: Plenum. Schwartz, S. P. (1979). Natural kind terms. Cognition, 7, 301-315. Shalom, A. (1985). The body-mind conceptual framework and the problem of personal identity. Atlantic Highlands, NJ: Humanities Press. Shotter, J. (1984). Social accountability and selfhood. Oxford: Basil Blackwell.

Self-Continuity in Youth

221

Shusterman, R. (1992). Pragmatist aesthetics: Living beauty, rethinking art, Cambridge, MA: Blackwell. Shweder, R. A., & Bourne, L. (1984). Does the concept of the person vary cross-culturally? In R. A. Shweder & R. A. Levine (Eds.), Culture theory: Essays on mind, self and emotion (pp. 158-199). Cambridge, England: Cambridge University Press. Singer, J. A. (1995). Seeing one's self: Locating narrative memory in a framework of personality. Journal of Personality, 63, 429-457. Singer, J. A., & Salovney, E (1993). The remembered self New York: Free Press. Spelke, E. S. (1985). Preferential-looking methods as tools for the study of cognition in infants. In G. Gilbert & N. A. Krasnegor (Eds.), Measurement of audition and vision in the first year of post-natal life: A methodological overview (pp. 323-363). Norwood, NJ: Ablex. Spelke, E. S., Kestenbaum, R., Simons, D. J., & Wein, D. (1995). Spatiotemporal continuity, smoothness of motion and object identity in infancy. British Journal of Developmental Psychology, 13, 113-142. Spence, D. P. (1983). Narrative persuasion. Psychoanalysis and Contemporary Thought, 6, 457-481. Sutton-Smith, B. (1986). Children's fiction making. In T. R. Sarbin (Ed.), Narrative psychology: The storied nature of human conduct (pp. 67-90). New York: Praeger. Taylor, C. (1991). The malaise of modernity. Concord, Canada: Anansi. Tompkins, J. P. (Ed.). (1986). Reader-response criticism: From formalism to post-structuralism. Baltimore: Johns Hopkins University Press. Tulving, E. (1983). Elements of episodic memory. New York: Oxford University Press. Turiel, E., & Wainryb, C. (1994). Social reasoning and the varieties of social experiences in cultural contexts. In H. W. Reese (Ed.),Advances in child development and behavior (Vol. 25, pp. 289-326). San Diego: Academic Press. Unger, R. M. (1975). Knowledge and politics. New York: Free Press. van Inwagen, P. (1990). Four-dimensional objects. Nous, 24, 245-255. Volosinov, V. N. (1986). Marxism and the philosophy of language. Cambridge, MA: Harvard University Press. (Original work published 1929) Vygotsky, L. S. (1987). The collected works of L. S. Vygotsky: Vol. 1, Problems of general psychology. (R. W. Reiber & A. S. Carton Eds. & N. Minick, Trans.). New York: Plenum. Waterman, A. S. (1984). Identity formation: Discovery or creation? Journal of Early Adolescence, 4, 329-341. Weintraub, K. J. (1975). Autobiography and historical consciousness. Critical Inquiry, June 1975, 821-848. White, H. (1973). Metahistory. Baltimore: Johns Hopkins University Press. Whittaker, E. (1992). The birth of the anthropological self and its career. Ethos, 20, 191-219. Wiggins, D. (1980). Sameness and substance. Cambridge, MA: Harvard University Press. Xu, F., & Carey, S. (1996). Infants' metaphysics: The case of numerical identity. Cognitive Psychology, 30, 111-153. Zagorin, E (1999). History, the referent, and narrative: Reflections on postmodernism now. History and Theory, 38, 1-24.

This Page Intentionally Left Blank

Author Index

Barry, C., 58 Bartsch, K., 9, 35 Bates, S., 155, 161 Battig, W., 86 Beck, A., 156 Becker, J., 160 Beeman, M., 76 Behrend, D., 17 Belli, R., 73 Bem, D., 140, 148-149 Bentall, R., 121 Berenbaum, S., 156 Berko, J., 129 Bernstein, G., 154 Bernstein, I., 161 Berzonsky, M., 181, 194, 203 Bindschaedler, C., 76 Birmaher, B., 144, 146, 159, 162 Bjorklund, D., 44, 63, 73, 85, 87 Blacker, C., 141 Blackwell, J., 60 Blanchard, I., 63 Blechman, E., 140 Boag, L., 150 Bockenholt, U., 80 Boeck, M., 208 Bohan, J., 186, 188 Bourne, L., 178, 193 Bovet, M., 18 Bowden, E., 76 Bower, G., 85 Bowlby, J., 157 Boyes, M., 198 Brady, E., 154 Brainerd, C., 41-91 Bransford, J., 44, 57 Braswell, G., 30 Brenner, C., 77 Brewer, W., 24, 30 Brockelman, E, 182 Brook, J., 155 Brooks-Gunn, J., 145, 159, 160

A Aboud, E, 186, 188 Acebo, C., 161 Achenbach, T., 143, 145, 146, 153, 155, 164 Ackerman, B., 60 Adams-Price, C., 73 Addy, C., 147 Adrian, C., 163 Albano, A., 155 Alibali, M., 7, 15 Allen, M., 86 Allen, N., 163 Alloy, L., 156 A1-Shabbout, M., 162 Anderson, B., 162 Anderson, J., 85, 146, 153 Andrews, J., 146 Angell, K., 63 Angold, A., 142, 151, 154, 159 Anisfeld, M., 68, 71 Ankrum, C., 58 Asarnow, J., 146, 155, 161 Asnis, G., 208 Atkinson, R., 56 Atran, S., 12, 186 Auguston, E., 105 Avenevoli, S., 150 Avis, J., 11

B

Bakhtin, M., 20, 21, 190 Ball, L., 198,205,209, 211 Balota, D., 77 Barlow, D., 155 Barnes, D., 105, 107, 108, 122, 130 Barnes-Holmes, D., 102, 105, 133, 134 Barnes-Holmes, Y., 102, 105, 133, 134 Barnett, E, 161,162 Baron-Cohen, S., 133 223

224

Author Index

Brown, A., 71, 110 Brown, G., 156 Brownstein, A., 105 Bruck, M., 66, 67, 70, 71 Bruner, J., 180, 185, 190, 193, 194 Bryant, E, 43, 49, 84 Buchanan, C., 160 Budson, A., 65, 77 Bukowski, W., 162 Burke, E, 150, 161

C Callahan, E, 52 Campos, J., 157 Camus, A., 176 Carey, M., 155 Carey, S., 2, 10, 14 Carlson, G., 140, 150 Carlson, S., 11, 13 Carr, D., 191,192 Carroll-Wilson, M., 155 Carskadon, M., 161 Carter, A., 44, 71 Casat, C., 154 Caskey, B., 52 Caspi, A., 140, 148-149 Cassel, W., 73 Cassirer, E., 178, 179 Ceci, S., 66, 67, 68, 70, 71 Chandler, M., 178, 179, 188, 193, 198, 199, 202, 204, 205,206, 209, 210, 211 Chapman, M., 45 Chassin, L., 155 Chen, C., 12 Choi, I., 180 Chomsky, N., 101, 102 Chorpita, B., 155 Chrousos, G., 159 Church, R., 7, 15 Cicchetti, D., 140 Clare, A., 141 Clark, D., 156 Clark, L., 153, 155, 156 Clark, S., 69 Clayton, E, 154 Clements, C., 156 Clements, W., 11

Cloninger, C., 154 Cohen, E, 155 Cole, D., 146, 157 Cole, M., 20 Coley, J., 186 Compas, B., 141,142, 143, 145, 157, 163 Conger, R., 163 Connors, C., 146, 155 Cook, T., 84 Corkill, A., 85 Corwin, J., 56, 75 Costello, E., 154, 159 Cox, A., 151 Cronbach, L., 4, 5 Crowley, K., 30 Cullinan, V., 102 Curran, T., 76 Cyranowski, J., 160

D

Daffner, K., 65 Dahl, R., 160, 161 Davidson, D., 54 Davies, G., 9 Davies, M., 147 Day, J., 71 DeBaryshe, B., 140 DeLoache, J., 2, 9, 28, 34 DeMarie-Dreblow, D., 63 Dempster, E, 85 Dennett, D., 189 Denton, L., 66, 67 Desikan, R., 65 de St. Aubin, E., 191 Devany, J., 105 Devine, D., 147 Devlin, B., 153, 157, 160 DeVries, R., 187, 188 Diamond, A., 191 DiFiore, J., 208 Dilthey, W., 190 Dixon, J., 84 Dobert, R., 193 Dohr, K., 161 Domahidy-Dami, C., 18 Dougher, M., 105

Author Index

Freeman, R., 161 Frye, N., 191,193 Funnell, M., 76

Dumas, J., 140 Dundes, A., 193 Dyck, M., 156 Dymond, S., 105, 108

Eccles, J., 160 Edwards, G., 147 Elder, G., 163 Elie, C., 87, 88 Elkind, D., 43 Ellis, S., 7-8, 17, 51, 53, 54, 80 Emslie, G., 161 Endicott, J., 143 Erikson, E., 178 Erkanli, A., 154, 159 Evans, J., 134 Ewing, K., 210 Ey, S., 142 Eysenik, H., 153

F

Fauth, J., 54 Feehan, M., 146 Feldman, C., 193 Felzen, E., 68, 71 Fergusson, D., 140 Fernyhough, C., 21 Ferrari, M., 178, 193 Finch, A., 147, 154, 155 Fine, S., 161 Fischer, K., 23, 27, 102 Fischer, S., 141 Fischler, I., 56 Fiske, A., 210 Flanagan, O., 178, 192, 194, 203 Flavell, E., 14 Flavell, J., 9, 14, 17, 18 Fleming, J., 144 Forehand, R., 147, 154 Forster, E., 204 Fraisse, P., 179 Frame, C., 154 Frank, E., 160 Franks, J., 44, 57

Gabrieli, J., 65 Gaeth, G., 80 Gallagher, S., 179, 204 Galluccio, L., 76 Garber, J., 140, 146, 150, 161 Garcia, R., 19 Garfinkel, B., 154 Garrison, C., 145, 147 Gatsonis, C., 150, 157 Gauze, C., 162 Gazzaniga, M., 76 Ge, X., 163 Geertz, C., 178, 181,193 Gehrig, R., 58 Geller, B., 159 Gelman, S., 2, 9, 14, 22, 23, 29, 186, 187 Gergen, K., 191 Gergen, M., 191 Gernsbacher, M., 60 Giddens, A., 193 Gifford, E., 118, 121, 122 Gigerenzer, G., 79 Gillund, G., 55 Girgus, J., 141,147, 161,162 Giunta, C., 163 Goldberger, A., 6 Goldfield, E., 26 Goldin-Meadow, S., 7, 15 Goldsmith, M., 65 Goodman, G., 66 Gopnik, A., 23 Gordon, L., 47, 48, 63 Gotlib, H., 144, 153, 161,162 Gottfried, G., 186 Gottman, J., 28, 29 Graber, J., 159 Graf, P., 57 Grant, K., 142 Grant, M., 45, 56 Green, E, 14 Greenberg, M., 156 Greenway, D., 105

225

226

Author Index

Gronlund, S., 69 Gross, M., 63 Gruen, G., 52 Guardo, C., 186, 188 Gullone, E., 153 Gurley, D., 155 Gutheil, G., 188 Guze, S., 154

Habermas, J., 178, 181,193, 194, 195 Hadwin, J., 133 Haines, B., 84 Hala, S., 198 Haley, G., 161 Hall, D., 188 Hallowell, A., 178 Hammen, C., 144, 153, 157, 161,163 Harkavy Friedman, J., 208 Harmon, K., 103 Harnishfeger, K., 86 Harre, R., 178, 179, 191,210 Harris, P., 9, 11 Harsent, L., 76 Hastie, R., 46 Hatano, G., 11, 12 Hayes, L., 105, 118 Hayes, S., 102, 105, 108, 121,122, 127, 132 Healy, A., 58 Heeren, T., 208 Henderson, B., 47, 60 Hermans, H., 181, 190, 193 Hersen, M., 154 Hess, T., 74 Hetherington, E., 159 Hickling, A., 11, 34 Hilsman, R., 161 Hintzman, D., 55 Hirsch, E., 178, 182, 184 Hirshfeld, L., 14, 210 Hoffman, M., 132 Hogarth, R., 78 Holland, D., 181,182, 190, 193,210 Hook, J., 84 Hops, H., 146 Horwood, L., 140 Housman, D., 208 Hovland, C., 120

Howe, M., 4, 9, 10, 45, 56, 66, 71, 82, 86 Howell, C., 146, 155 Howell, D., 163 Howlin, P., 133 Hoza, B., 162 Hughes, C., 161

Inagaki, K., 12 Inhelder, B., 18, 43, 50, 79 Inoff-Germain, G., 159 Isakson, K., 162 Israel, L., 72, 73

J

Jackson, E., 77 Jackson, K., 147 Jacobs, J., 54 Jacobsen, R., 163 Jacoby, L., 56 James, W., 178, 180, 188 Jarvella, R., 57 Jarvis, E, 162 Jenson, A., 86 John, K., 159 Johnson, C., 187 Johnson, M., 63, 65 Johnson, S., 10 Johnson-Laird, E, 134 Jolly, J., 156 Juola, J., 56

K

Kagan, J., 65, 148, 149 Kahneman, D., 42, 53, 78, 79 Kalish, C., 9 Kalmar, D., 193, 194 Kandel, D., 147 Kashani, J., 144, 145, 150, 153 Kaslow, N., 155, 163 Kausler, D., 74 Kay, B., 26 Kazdin, A., 154 Keating, D., 161

227

Author Index

Keenan, M., 108 Keil, F., 9, 10, 24, 187, 188 Kelly, K., 156 Kempen, H., 190, 193 Kempton, T., 147 Kendall, E, 154, 155 Kennedy, B., 23, 27 Kennedy, R., 159 Kensinger, E., 74, 75, 89 Kerby, A., 182, 189, 190, 191,192, 194 Kerkman, D., 25 Kerns, K., 156 Kesterbaum, R., 187 Kevershan, J., 86 Kiernan, B., 43, 46, 47, 57, 59, 60, 68, 70, 71, 73, 89, 90 King, N., 153 Kingma, J., 44, 45 Kirmayer, L., 211, 214 Kitayama, S., 210, 211 Kitsch, W., 58, 60 Klaczynski, E, 54, 55, 80 Kneer, R., 60, 61, 69, 71, 72, 74, 87 Koch, M., 146 Koehler, D., 83 Koeman, H., 188 Kohlenberg, B., 105 Koriat, A., 65, 66 Kounios, J., 65 Koutstaal, W., 63, 65, 74, 77 Kovacs, M., 141,146, 150, 153, 154, 157, 160 Kramer, T., 156 Kriss, M., 146 Kuhberger, A., 80 Kuhn, T., 8, 17, 30

Lahey, B, 163 Lalonde, C., 178, 198, 205 Lamb, M., 66 Lampinen, J., 61, 85 Larson, R., 162 Last, C., 154 Lawrence, J., 82 Lawton, S., 71 Lee, D., 5 Leekman, S., 11 Leichtman, M., 65

Leitenberg, H., 155 Levin, I., 80 Levine, R., 178 Lewinsohn, E, 141,144, 146, 157, 159, 162, 163 Lewis, M., 178, 193 Lifton, R., 190 Lightfoot, C., 181, 193,204 Lillard, A., 9, 11,210 Lindenberger, U., 45 Lindholm, L., 146 Lipkens, G., 111 Lipovsky, J., 154 Lloyd, F., 42, 61, 69 Loftus, E., 66 Lonigan, C., 155 Lorenz, E, 163 Lowe, C., 121 Luckman, T., 178 Lumsdaine, A., 120 Lynskey, M., 140

M

Maclntyre, A., 178, 191,194 MacLeod, C., 72 Magnusson, D., 159 Malone, C., 58 Mandler, G., 56 Mandler, J., 180, 185 Mansfield, E., 191 Markham, M., 105 Markus, H., 210, 211 Marriage, K., 161 Marsteller, E, 147 Martin, J., 146, 157 Martin, R., 154 Marx, M., 47, 60 Mather, M., 65 Mayr, E., 185 McAdams, D., 190, 191 McCarthy, T., 73 McCauley, E., 150, 161 McConaughy, S., 146 McConville, B., 150 McCormick, M., 9 McDermott, K., 67, 72, 74, 75, 87, 88 McGeady, S., 105 McGee, R., 146, 153

228 McGough, 1., 66 McIntosh, J., 147 McKeown, R., 147 McLin, D., 26 Medin, D., 186 Melo, B., 65, 76 Merikangas, K., 159 Metcalfe, J., 76 Metzler, C., 76 Meyer, J., 151 Miller, J., 191,210, 211 Miller, K., 9, 13, 34 Miller, P., 132 Mineka, S., 156 Mink, L., 191,192, 194, 210 Mitchell, J., 150, 161 Moffitt, T., 140 Mohr, D., 186, 188 Mojardin, A., 57, 60, 61, 71, 72 Moore, C., 84 Moravcski, J., 58 Morelli, G., 20 Moretti, M., 161 Moscovitch, M., 65 Moss, H., 148, 149 Moss, S., 150, 161 Moyal, B., 161 Muir, J., 63, 85, 87 Mumme, D., 157 Mur, Y., 188 Murdock, B., 55 Murphy, G., 60 Murray, D., 181

Naito, M., 11 Narasimham, G., 54, 80 Neiderhiser, D., 159 Neiderhiser, J., 159 Nelson, T., 72 Nesse, D., 140 Neuman, R., 159 Neuschatz, J., 60, 61 Newcomb, A., 162 Newell, K., 5 Nicolopoulou, A., 190, 193 Nisbett, R., 180, 185, 186, 188, 210 Nolde, S., 65

Author Index

Nolen-Hoeksema, S., 141,147, 161, 162 Norenzayan, A., 180, 181 Norman, K., 65, 74, 76 Nottelman, E., 159 Nunner-Winkler, G., 193

O'Connor, T., 159 Offenbach, S., 52 Offord, D., 144 Ollendick, T., 153 Olney, C., 86 Oppenheimer, L., 188 Ortony, A., 110, 186 Overton, W., 8, 10, 180, 181

Palmer, J., 58 Pansky, A., 65 Parfit, D., 178, 184 Paris, S., 44, 71 Park, B., 46 Parker, A., 80 Parkin, A., 76 Pasnak, R., 103 Patterson, G., 140 Paulauskas, S., 157 Payne, D., 60, 61, 72, 85, 87, 88 Peacock, J., 190 Peeke, L., 157 Perlmutter, M., 17, 73 Perner, J., 9, 11 Perry, J., 178 Perry, M., 15 Peterson, A., 141,145, 159, 160, 163 Pezdek, K., 71 Phares, V., 163 Phelps, E., 76 Piaget, J., 18, 19, 27, 43, 44, 50, 52, 79, 101-102, 178, 186, 188 Pierroutsakos, S., 9 Pine, D., 155 Pinker, S., 101, 129 Plomin, R., 159 Polkinghorne, D., 190, 191, 194, 205

Author Index

Polya, G., 110 Poole, D., 59, 62, 66, 72 Potenza, M., 54 Powell, G., 76 Powers, B., 146 Predergast, M., 151 Prinz, R., 140 Propp, V., 193 Prusoff, B., 159 Puig-Antich, J., 141 Purohit, A., 150

Q Qin, J., 66 Quas, J., 66 Quay, H., 146, 154, 155

Rabinowitz, F., 45, 56, 82, 84 Racine, C., 72 Ramsey, E., 140 Rankm, J., 74 Rao, U., 141 Rapcsak, S., 76 Reese, H., 10, 60, 116, 179 Reid, J., 150 Reiter, E., 159 Reminger, S., 76 Renderer, B., 193 Reyna, V., 41-91 Reynolds, W., 155 Rice, J., 159 Richards, C., 157 Richards, M., 162 Ricoeur, E, 181, 190, 192, 205 Rigney, D., 6 Ring, M., 179, 180 Roberts, R., 146 Robins, E., 143 Roche, B., 102, 105, 107, 108, 122, 130, 133, 134 Roe, C., 71 Roediger, H., 67, 74, 87 Rogoff, B., 20 Rogosch, E, 140 Rohde, E, 141,144, 157, 162

229

Rorty, A., 178, 181,182, 183, 184, 189, 193, 200, 204 Rosen, M., 63 Rosenberg, T., 150 Rosengren, K., 9, 11, 17, 24, 30, 34, 188 Ross, C., 208 Ross, D., 68 Ross, L., 180, 186, 188 Rubenstein, J., 208 Rubin, C., 208 Ruble, D., 186, 188 Rudolph, K., 144 Ruffman, T., 11 Rush, J., 161 Russell, B., 179 Rutter, M., 140, 145, 148, 150, 151 Ryan, N., 150, 160, 162

Saarni, C., 157 Sarbin, T., 191 Sarigiani, E, 159 Sass, L., 189 Saylor, C., 147 Schaaf, J., 66 Schacter, D., 57, 63, 65, 72, 73, 74, 75, 76, 77, 89 Schafer, W., 30 Schlesinger, A., 179, 190, 203, 213 Schneider, S., 80 Schneider-Rosen, K., 140 Schwartz, S., 189 Scott, L., 193 Seeley, J., 141,144, 146, 157, 159, 162, 163 Seligman, M., 147, 161,162 Seroczynski, A., 157 Shaffer, T., 58 Shalom, A., 182 Shapiro, A., 60 Shear, K., 160 Sheffield, E, 120 Shelby, J., 105 Shiffrin, R., 55 Shotter, J., 178, 202 Shusterman, R., 179 Shweder, R., 178, 193 Siegler, R., 2, 3, 5, 6, 7-8, 9, 11, 13, 15, 17, 25, 29, 30, 34, 44

230

Silberg, J., 150, 151,159 Silva, E, 153 Simonds, J., 145 Simonoff, E., 151 Simons, D., 187 Simons, R., 163 Sinclair, H., 18 Skinner, B., 102, 116 Skinner, J., 161 Slovic, E, 42, 79 Smalhofer, E, 58 Smedslund, J., 43 Smeets, E, 105, 133, 134 Smetana, J., 162 Smiley, S., 71 Smith, C., 13 Smith, L., 3, 5, 7, 11, 14, 25, 26, 27, 33, 34 Snodgrass, J., 56, 75 Sodian, B., 9 Sokol, B., 178 Solomon, G., 9-10 Spelke, E., 14, 187 Spencer, J., 26 Sperry, L., 132 Spinillo, A., 49, 84 Spitzer, R., 143 Sprague, R., 5 Springer, K., 10, 24 Squires, E., 76 Sroufe, A., 140, 148 Stanger, C., 146 Stark, K., 155 Stattin, H., 159 Stechler, G., 208 Steele, D., 108 Steer, R., 156 Stein, L., 62 Steinberg, L., 162 Sternberg, R., 110 Stevenson, H., 12 Stewart, M., 58 Stewart, S., 161 Stone, E., 80 Strauss, C., 154, 163 Strong, R., 103 Subbotsky, E., 16 Suomi, S., 157 Susman, E., 159 Sutton-Smith, B., 193

Author Index

T Taylor, C., 9, 178 Taylor, M., 11, 13, 156 Tems, C., 161 Thelen, E., 3, 5, 7, 11, 14, 25, 26, 34 Thomson, D., 59 Titcomb, A., 59, 67, 69, 73, 75 Titzer, R., 26 Todd, R., 159 Toglia, M., 68, 88, 89 Tompkins, J., 199 Townsend, M., 71 Trabasso, T., 43, 44 Truglio, R., 147, 157 Tulving, E., 57, 59, 181 Tulviste, P., 20 Tun, P., 63, 74 Turiel, E., 211 Tversky, A., 42, 53, 78, 79, 83

U Unger, R., 178, 179

V Valsiner, J., 4, 5, 7 van der Ende, J., 140 van Emmerik, R., 5 van Geert, P., 26-27 Van Loon, R., 190, 193 Vasquez, C., 156 Verfaellie, M., 77 Verhulst, E, 140, 155 Vieria, C., 161 Vosniadou, S., 2, 24, 30 Vygotsky, L., 19-20, 21, 27, 190

W Wagner, L., 193 Wainryb, C., 211 Wallace, W., 58, 59 Walsh, C., 45 Ward, J., 76

Author Index

Warren, M., 160 Warren, W., 26 Wason, R, 134 Waterman, A., 181 Watson, D., 153, 154, 155 Wedell, D., 80 Weintraub, K., 194 Weiss, B., 144, 150 Weissman, M., 155, 159 Welch, D., 58 Wellman, H., 2, 3, 9, 14, 22, 23, 35, 186, 187 Wem, D., 187 Wemer, H., 52 Wertsch, J., 20, 21 Wescourt, K., 56 West, B., 6 Wherry, J., 156 White, H., 191 White, M., 178 Whittaker, E., 178 Wiggins, D., 178, 181, 184 Williams, R., 163 Williams, S., 146, 153 Williamson, D., 162 Wilson, K., 121, 122 Windschitl, E, 73 Winer, G., 82 Winfrey, S., 73 Wingfield, A., 63

Winocur, G., 65 Wolfe, C., 80, 82, 83 Wood, C., 56 Worthman, C., 159 Wright, R., 61, 72 Wulfert, E., 105

Xu, E, 187

Yates, J., 80 Yosi, L., 155 Young, E., 160

Zagorin, E, 191-192 Zaitchik, D., 10 Zatz, S., 155 Zettle, R., 118 Zhang, H., 13 Zhu, J., 13 Zinny, S., 58

231

This Page Intentionally Left Blank

Subject Index t following page number indicates a table. f following page number indicates a figure. verbal intelligence, 128 unitary theories for, 42 Cognitive neuroscience, of false memory, 65-76 Cognitive schemata, Beck's theory of, 156 Cognitive triage effect, in recall, 86-87 Conjunction Fallacy Problem, 78, 78t Constructivism, view of self in, 191-192 Constructivist hypothesis, and memory-reasoning relations, in fuzzy-trace theory, 43-44 defined, 47 evidence for, 46-47 Continuity concept of, 140 of depression, see Depression, continuity of of psychopathology, 140 Cross-cultural differences in variability of children's reasoning, Piaget's theory, 19 personal persistence and, 205,209-213 Cultural continuity, as protection against suicide, 205, 213-215

Adolescents continuity of depression anxiety and, 152-158 cross-sectional analysis, 149-150 in boys, 151 in girls, 151 longitudinal analysis, 151-152 transition in, 141 depression and, 139-165 Alzheimer's patients, false memory in, 77-78 Analogy, in relational frame theory, 110-111 Anxiety adolescent depression and, 152-158 covariation and comorbidity, 153-154 developmental progression of anxiety and depression, 156-158 failure to isolate a pure syndrome, 154-155 specificity in, 155-156

Beck's theory of cognitive schemata, 156 Behavioral psychology, relational frame theory and, 102 Bildungsromane, 199

Decision making, fuzzy-trace theory and, 78-84 Decision-framing paradigm for adults, 53 for children, 53-54, 54t Depression, 141 anxiety and, 152-158 covariation and comorbidity, 152-154 developmental progression of anxiety and depression, 156-158 failure to isolate a pure syndrome, 154-155 specificity in, 155-156 as a disorder, 142-143

Canadian youth, personal persistence in, 197-215, see also Personal persistence, in native and non-native youth Cognitive development, see also Reasoning complexity of, 10 in relational frame theory, 101-102 Piaget's theory of, 2 relational frame theory for, 128-129 numerical intelligence, 128-129

233

234

Depression (continued) research on, 164 stability of, 145-146 as a mood, 144 prevalence of, 145 research on, 164 stability of, 146-147 as a syndrome, 143-144 prevalence of, 144-145 research on, 164 stability of, 146 continuity of, 139-165, 141,148-149 heterotypic, 148-149 internalized negative affectivity and, 152-158 preliminary evidence of, 149-152 homotypic, 148 in adolescents anxiety and, 152-158 cross-sectional analysis, 149-150 in boys, 151 in girls, 151 longitudinal analysis, 151-152 influences on biological, 159-160 cognitive, 161-162 genetic, 158-159 social-contexual, 162-163 research on, 141-142 research suggestions, 163-165 consideration of continuity type, 164 consideration of depression type, 164 longitudinal analysis, 165 stability vs. continuity, 148 Helplessness-Hopelessness theory in, 156-158 prevalence of, 144-145 stability of, 145-148 Dissociated verbatim-gist retrieval distractors, 59-60 gist traces in, 59-60 verbatim traces in, 59-60 Domain specificity, variability in children's reasoning and, 2-3, 13-14, 21-24, 29 modular view, 22 theory view, 23-24 Dynamic systems perspective, of children's reasoning, 5-6, 25-27, 34

Subject Index

Education, relational frame theory in higher level education, 134-135 mainstream, 132-133 special education, 133-134 Emotional development, relational frame theory for, 131-132 emotional intelligence, 132 understanding the emotions of self, 131-132 Equilibration, 17 Essentialism, 182-189 defined, 182 down side of, 188-189 personal persistence in, 179-180 change and, 179-180 survival and, 184-185 psychological, 185-188 variations in youth categories, 201-203 cultural differences, 209-213 vs. narrativity, 177-182, 195-197

False memory cognitive neuroscience of, 65-76 false-recognition paradigm in, 67-69, 71-73 fuzzy-trace theory and, 70-71 implanted, 67-69 in brain damaged patients, 76-78 Alzheimer's patients, 77-78 frontal lobe damage, 76 split-brain, 76 in children, 66-73 in elderly, 73-76 misinformation paradigm in, 68-71, 73 persistence of, 71-72 sexual abuse and, 66 spontaneous, 67-69 False recall, robust, effects in recall, 87-89 False-recognition paradigm, in false memory, 67-69, 71-73 Familiarity, in recognition memory, 56-57 Fuzzy-processing preference advantages of, 51, 51 t defined, 50, 55 disadvantages of, 51-52 Fuzzy-to-verbatim continua, defined, 50, 55

Subject Index

Fuzzy-trace theory, 41-91 concepts, 43, 55, 64-65, 89-91 dual processes in, 47-65, 90 memory, 55-64 developmental change, 63-64 dissociated verbatim-gist retrieval, 59-60 identity and similarity judgments, 61-63 memorial bases of conscious remembrance, 60-61 parallel verbatim-gist storage, 57-59 recognition memory, 56-57 reasoning, 48-55 developmental change, 52-55 evidence for principles, 50-52 fuzzy-processing preference, 50 fuzzy-to-verbatim continua, 50 gist-extraction principle, 49-50, 57 principles for, 49-50 false memory and, 70-71, 91 memory-reasoning relations, 43-47, 89-90 constructivist hypothesis, 43-44 defined, 47 evidence for, 46-47 necessity hypothesis, 43-46 defined, 44 evidence for, 44-46 transitive inference paradigm, 44-46, 45f research domains, 65-91 false memory behavioral studies of aging, 73-76 behavioral studies of child development, 66-73 brain damaged patients, 76-78 cognitive neuroscience of, 65-76 judgment and decision making, 78-84 graded rationality, 80-83 inclusion illusions, 83-84 retrieval mechanisms in recall, 85-89 cognitive triage effects, 86-87 robust false recall, 87-89

Gist retrieval, adults compared to older adults, 74-76, 75f Gist traces dissociated verbatim-gist retrieval, 59-60

235

forgetting of, 63 memorial bases of conscious remembrance, 60-61 parallel verbatim-gist storage, 57-59 Gist-extraction principle, reasoning and, defined, 49-50, 55 Graded rationality, in judgment, 80-83

Helplessness-Hopelessness theory, in depression, 156-158 Horizontal decalage, 13-14, 17-19 conservation problems, 18 hierarchies of understanding, 18 resistances in, 18-19 Humor, in relational frame theory, 113

Identity information, see Personal persistence Identity, kind vs. individual, 187-188 Inclusion illusions, in judgment errors, 83-84 Individual differences, in children's reasoning, 10-13 Internalized distress, adolescents depression and, 152-158

Judgment fuzzy-trace theory and, 78-84 graded rationality in, 80-83 identity and similarity, memory and, 61-63 inclusion illusions in, 83-84

Language development in relational frame theory, 101-102 relational frame theory for, 129-131 comprehension vs. production, 130 production of novel utterances, 129 speech errors, 129-130 u-shaped development, 130-131

236

Subject Index

List-search process, in recognition memory, 56-57 Logic, relational frame theory and, 134-135 affirming the antecedent, 134 denying the antecedent, 134 modus ponens, 134 modus tollens, 134

M

Memory conscious, memorial bases of, 60-61 dual processes in, 55-64 developmental change, 63-64 dissociated verbatim-gist retrieval, 59-60 identity and similarity judgments, 61-63 memorial bases of conscious remembrance, 60-61 parallel verbatim-gist storage, 57-59 recognition memory, 56-57 explicit, defined, 60 false, see False memory implicit, defined, 60 recognition, 56-57 familiarity, 56-57 list-search process, 56-57 recollection, 56-57 unitary theories for, 42 Memory-reasoning relations dual processes in, 47-65, see also Fuzzy-trace theory, dual processes in fuzzy-trace theory, 43-47, 89-90 constructivist hypothesis, 43-44 defined, 47 evidence for, 46-47 necessity hypothesis, 43-46 defined, 44 evidence for, 44--46 transitive inference paradigm, 44-46, 45f Metaphor, in relational frame theory, 111-112 Misinformation paradigm, in false memory, 68-71, 73

N Narrativity, 190-197 negative aspects of, 194-195 personal persistence in, change and, 179-180

positive aspects of, 192-194 variations in youth categories, 203-205 cultural differences, 209-213 view of self and, 191 vs. essentialism, 177-182, 195-197 Native youth, personal persistence in, 197-215, see also Personal persistence, in native and non-native youth Necessity hypothesis, and memory-reasoning relations, in fuzzy-trace theory, 43-46 defined, 44 evidence for, 44-46 transitive inference paradigm, 44-46, 45f Negative affectivity adolescents depression and, 152-158 defined, 153

P

Parallel verbatim-gist storage gist traces, 57-59 parallelism in, 58-59 verbatim traces, 57-59 Personal persistence concepts, 177 essentialism in, 179-180 in native and non-native youth, 197-215 cultural continuity as protection against suicide, 205, 213-215 cultural differences, 205,209-213 data categorization, 201-205 data collection, 199-201 essentialist self-continuity, 201-203 key findings, 197-198, 215 narratively based self-continuity, 203-205 youth suicide, 198, 205,207-209 narrativity in, 179-180 narrativity vs. essentialism in, 177-182 Piaget's theory variability in children's reasoning, 2, 17-19 cultural differences, 19 equilibration, 17 horizontal decalage, 17-19 view of child development, 6 Problem solving self-directed rules and, 124-126 strategic, 117 valuative, 117

Subject Index

Proximal development, zone of, 20 Puberty, depression and, 159-160

Reasoning dual processes in, 48-55 developmental change, 52-55 evidence for principles, 50-52 fuzzy-processing preference, 50 fuzzy-to-verbatim continua, 50 gist-extraction principle, 49-50, 57 principles for, 49-50 intuitive, development of, 52-55 unitary theories for, 42 variability in, 1-35 assessing importance of, 3-8 development vs. learning, 6-7 inter- and intraindividual variability, 5-6 negative aspects of variability, 7 positive aspects of variability, 7-8 stability of cognitive structures, 4-5 cross-cultural differences and, 11-13 domain specificity in, 2-3, 13-14, 29 methodological issues, 27-34 issues obscuring, 28-29 randomness, 28 Piaget's theory, 2 research suggestions, 29-33, 35 data collection, 29-30 use of relative entropy, 30-31, 3 lf, 32f, 33 theories and meaning, 16-27 domain specificity view, 21-24 dynamic systems perspective, 5-6, 25-27, 34 information processing view, 25 Piaget's view, 17-19 sociocultural view, 19-21 types of, 8-16 across age, 9-10 across children of same age, 10-13 in a child's performance across related tasks, 13-14 within a child over repeat trials, 14-15 within a child within a single trial, 15-16 Recall, see also Recollection cognitive triage effects in, 86-87

237

false robust, 87-89 process of, 85 retrieval mechanisms in, 85-89 cognitive triage effects, 86-87 robust false recall, 87-89 Recollection, see also Recall in recognition memory, 56-57 Relational frame theory, 101-136 and psychological development, 127-132 cognitive development, 128-129 emotional development, 131-132 language development, 129-131 concepts, 102, 127, 135-136 arbitrarily applicable, 103 arbitrarily applicable relational responding, 103-104 bidirectionality, 104 combinatorial entailment, 104 mutual entailment, 104 relata, 104 relational frames, 105-106 types of, 106-108 transformation of stimulus, 104-105 verbal events, 108 in education, 132-135 higher level education, 134-135 mainstream, 132-133 special education, 133-134 multiple stimulus relations, 108-110 complete and coherent networks, 109-110 pragmatic verbal analysis, 114-115 utility of, 117-118 problem solving, 116-117 relations among relations, 110-114 analogy, 110-111 humor, 113 metaphor, 111-112 storytelling, 112-113 self and self-directed rules, 121-127 concept of self, 121-122 perspective taking, 122-123, 126 self-directed rules, 124 self-directed rules and problem solving, 124-126 types of self, 123-124 conceptualized self, 123 knowing self, 123-124 transcendent self, 124 verbal other types, 126-127

238

Subject Index

Relational frame theory (continued) thinking, 116 understanding and verbal regulation, 118-121 understanding verbal events, 118-119 verbal communication, 120-121 verbal regulation and rule following, 120 verbal abstraction of stimuli, 115-116 Relational frames, in relational frame theory, 105-108 comparative, 106-107 coordination, 106 deictic relations, 107 distinction, 106 hierarchical relations, 107 opposition, 106 Relative entropy, in research on variability of children's reasoning, 30-31, 31f, 32f, 33 Retrieval mechanisms, in recall, 85-89 cognitive triage effects, 86-87 robust false recall, 87-89

Self-continuity, see Personal persistence Sexual abuse, false memory and, 66, see also False memory Sociocultural view, variability in children's reasoning, 19-21 learning, 20-21 maturation, 19-20 Special education, relational frame theory in, 133-134 Storytelling, in relational frame theory, 112-113 Suicide, youth cultural continuity as protection against, 205, 213-215 personal persistence and, 205,207-209 Survival, personal persistence and, 184-185

Transition, in adolescents, 141 depression and, 139-165, see also Depression, continuity of Transitions points, in variability of children's reasoning, 34-35 Transitive inference paradigm, in memory-reasoning relations, 44-46, 45f

Unitary theories of memory development, 42 of reasoning development, 42

Variational mode, for variation in children's reasoning, 5-6 Ventriloquation, 20 Verbatim retrieval, adults compared to older adults, 74-76, 75f Verbatim traces dissociated verbatim-gist retrieval, 59-60 forgetting of, 63 memorial bases of conscious remembrance, 60-61 parallel verbatim-gist storage, 57-59 Verbatim-gist retrieval, dissociated distractors, 59-60 gist traces in, 59-60 verbatim traces in, 59-60 Verbatim-gist storage, parallel gist traces, 57-59 parallelism in, 58-59 verbatim traces, 57-59

Youth suicide cultural continuity as protection against, 205, 213-215 personal persistence and, 205,207-209 Target acceptances, in memory and judgment, 62-63 Theory-of-mind tasks (TOM) in special education, 133-134 research on, 11, 21

Zone of proximal development, 20

Contents of Previous Volumes

Volume 1

Responses of Infants and Children to Complex and Novel Stimulation Gordon N. Cantor Word Associations and Children's Verbal Behavior David S. Palermo Change in the Stature and Body Weight of North American Boys during the Last 80 Years Howard V. Meredith Discrimination Learning Set in Children Hayne W. Reese Learning in the First Year of Life Lewis P. Lipsitt Some Methodological Contributions from a Functional Analysis of Child Development Sidney W. Bijou and Donald M. Baer The Hypothesis of Stimulus Interaction and an Explanation of Stimulus Compounding Charles C. Spiker The Development of "Overconstancy" in Space Perception Joachim F. Wohlwill Miniature Experiments in the Discrimination Learning of Retardates Betty J. House and David Zeaman

Social Reinforcement of Children's Behavior Harold W. Stevenson Delayed Reinforcement Effects Glenn Terrell A Developmental Approach to Learning and Cognition Eugene S. Gollin Evidence for a Hierarchical Arrangement of Learning Processes Sheldon H. White Selected Anatomic Variables Analyzed for Interage Relationships of the Size-Size, Size-Gain, and Gain-Gain Varieties Howard V. Meredith AUTHOR INDEX--SUBJECT INDEX

Volume 3

Infant Sucking Behavior and Its Modification Herbert Kaye The Study of Brain Electrical Activity in Infants Robert J. Ellingson Selective Auditory Attention in Children Eleanor E. Maccoby Stimulus Definition and Choice Michael D. Zeiler Experimental Analysis of Inferential Behavior in Children Tracy S. Kendler and Howard H. Kendler Perceptual Integration in Children Herbert L. Pick, Jr., Anne D. Pick, and Robert E. Klein Component Process Latencies in Reaction Times of Children and Adults Raymond H. Hohle

AUTHOR INDEX--SUBJECT INDEX

Volume 2

The Paired-Associates Method in the Study of Conflict Alfred Castaneda Transfer of Stimulus Pretraining to Motor PairedAssociate and Discrimination Learning Tasks Joan H. Cantor The Role of the Distance Receptors in the Development of Social Responsiveness Richard H. Waiters and Ross D. Parke

AUTHOR INDEX--SUBJECT INDEX 239

240

Contents o f Previous Volumes

Volume 4

Developmental Studies of Figurative Perception David Elkind The Relations of Short-Term Memory to Development and Intelligence John M. Belmont and Earl C. Butterfield Learning, Developmental Research, and Individual Differences Frances Degen Horowitz Psychophysiological Studies in Newborn Infants S. J. Hutt, H. G. Lenard, and H. F. R. Prechtl Development of the Sensory Analyzers during Infancy Yvonne Brackbill and Hiram E. Fitzgerald The Problem of Imitation Justin Aronfreed AUTHOR INDEXuSUBJECT INDEX

Volume 5

The Development of Human Fetal Activity and Its Relation to Postnatal Behavior Tryphena Humphrey Arousal Systems and Infant Heart Rate Responses Frances K. Graham and Jan C. Jackson Specific and Diversive Exploration Corinne Hutt Developmental Studies of Mediated Memory John H. Flavell Development and Choice Behavior in Probabilistic and Problem-Solving Tasks L. R. Goulet and Kathryn S. Goodwin AUTHOR INDEX--SUBJECT INDEX

Volume 6

Incentives and Learning in Children Sam L. Witryol Habituation in the Human Infant Wendell E. Jeffrey and Leslie B. Cohen Application of Hull-Spence Theory to the Discrimination Learning of Children Charles C. Spiker

Growth in Body Size: A Compendium of Findings on Contemporary Children Living in Different Parts of the World Howard V. Meredith Imitation and Language Development James A. Sherman Conditional Responding as a Paradigm for Observational, Imitative Learning and VicariousReinforcement Jacob L. Gewirtz AUTHOR INDEX--SUBJECT INDEX

Volume 7

Superstitious Behavior in Children: An Experimental Analysis Michael D. Zeiler Learning Strategies in Children from Different Socioeconomic Levels Jean L. Bresnahan and Martin M. Shapiro Time and Change in the Development of the Individual and Soc!ety Klaus F. Riegel The Nature and Development of Early Number Concepts Rochel Gelman Learning and Adaptation in Infancy: A Comparison of Models Arnold J. Sameroff AUTHOR INDEX--SUBJECT INDEX

Volume 8

Elaboration and Learning in Childhood and Adolescence William D. Rohwer, Jr. Exploratory Behavior and Human Development Jum C. Nunnally and L. Charles Lemond Operant Conditioning of Infant Behavior: A Review Robert C. Hulsebus Birth Order and Parental Experience in Monkeys and Man G. Mitchell and L. Schroers Fear of the Stranger: A Critical Examination Harriet L. Rheingold and Carol O. Eckerman

Contents of Previous Volumes

Applications of Hull-Spence Theory to the Transfer of Discrimination Learning in Children Charles C. Spiker and Joan H. Cantor

AUTHOR INDEX--SUBJECT INDEX

Volume 9

Children's Discrimination Learning Based on Identity or Difference Betty J. House, Ann L. Brown, and Marcia S. Scott

Two Aspects of Experience in Ontogeny: Development and Learning Hans G. Furth

The Effects of Contextual Changes and Degree of Component Mastery on Transfer of Training Joseph C. Campione and Ann L. Brown

Psychophysiological Functioning, Arousal, Attention, and Learning during the First Year of Life Richard Hirschman and Edward S. Katkin

Self-Reinforcement Process in Children John C. Masters and Janice R. Mokros

AUTHOR INDEX--SUBJECT INDEX

241

Volume 11

The Hyperactive Child: Characteristics, Treatment, and Evaluation of Research Design Gladys B. Baxley and Judith M. LeBlanc

Peripheral and Neurochemical Parallels of Psychopathology: A Psychophysiological Model Relating Autonomic Imbalance to Hyperactivity, Psychopathy, and Autism Stephen W. Porges

Constructing Cognitive Operations Linguistically Harry Beilin

Operant Acquisition of Social Behaviors in Infancy: Basic Problems and Constraints W. Stuart Millar

Mother-Infant Interaction and Its Study Jacob L. Gewirtz and Elizabeth F. Boyd

Symposium on Implications of Life-Span Developmental Psychology for Child Development: Introductory Remarks Paul B. Baltes

Theory and Method in Life-Span Developmental Psychology: Implications for Child Development Aletha Huston-Stein and Paul B. Baltes

The Development of Memory: Life-Span Perspectives Hayne W. Reese

Volume 10

Current Trends in Developmental Psychology Boyd R. McCandless and Mary Fulcher Geis

The Development of Spatial Representations of Large-Scale Environments Alexander W. Siegel and Sheldon H. White

Cognitive Perspectives on the Development of Memory John W. Hagen, Robert H. Jongeward, Jr., and Robert V. Kail, Jr.

The Development of Memory: Knowing, Knowing About Knowing, and Knowing How to Know Ann L. Brown

Cognitive Changes during the Adult Years: Implications for Developmental Theory and Research Nancy W. Denney and John C. Wright

Social Cognition and Life-Span Approaches to the Study of Child Development Michael J. Chandler

Life-Span Development of the Theory of Oneself: Implications for Child Development Orville G. Brim, Jr.

Implications of Life-Span Developmental Psychology for Childhood Education Leo Montada and Sigrun-Heide Filipp

AUTHOR INDEX--SUBJECT INDEX

Developmental Trends in Visual Scanning Mary Carol Day

The Development of Selective Attention: From Perceptual Exploration to Logical Search John C. Wright and Alice G. Vlietstra

AUTHOR INDEX--SUBJECT INDEX

Volume 12

Research between 1960 and 1970 on the Standing Height of Young Children in Different Parts of the World Howard V. Meredith

242

Contents of Previous Volumes

The Representation of Children's Knowledge David Klahr and Robert S. Siegler

Chromatic Vision in Infancy Marc H. Bornstein

Developmental Memory Theories: Baldwin and Piaget Bruce M. Ross and Stephen M. Kerst

Child Discipline and the Pursuit of Self: An Historical Interpretation Howard Gadlin

Development of Time Concepts in Children William J. Friedman

AUTHOR INDEX--SUBJECT INDEX

Volume 13

Coding of Spatial and Temporal Information in Episodic Memory Daniel B. Berch

A Developmental Model of Human Learning Barry Gholson and Harry Beilin

The Development of Discrimination Learning: A Levels-of-Functioning Explanation Tracy S. Kendler

The Kendler Levels-of-Functioning Theory: Comments and an Alternative Schema Charles C. Spiker and Joan H. Cantor

Commentary on Kendler's Paper: An Alternative Perspective Barry Gholson and Therese Schuepfer

Reply to Commentaries Tracy S. Kendler

On the Development of Speech Perception: Mechanisms and Analogies

Sibship-Constellation Effects on Psychosocial Development, Creativity, and Health Mazie Earle Wagner, Herman J. P. Schubert, and Daniel S. P. Schubert

The Development of Understanding of the Spatial Terms Front and Back Lauren Julius Harris and Ellen A. Strommen

The Organization and Control of Infant Sucking C. K. Crook

Neurological Plasticity, Recovery from Brain Insult, and Child Development lan St. James-Roberts

AUTHOR INDEX--SUBJECT INDEX

Volume 15

Visual Development in Ontogenesis: Some Reevaluations Jiiri Allik and Jaan Valsiner

Binocular Vision in Infants: A Review and a Theoretical Framework Richard N. Aslin and Susan T. Dumais

Validating Theories of Intelligence Earl C. Butterfield, Dennis Siladi, and John M. Belmont

Cognitive Differentiation and Developmental Learning William Fowler

Children's Clinical Syndromes and Generalized Expectations of Control Fred Rothbaum

AUTHOR INDEX--SUBJECT INDEX

Peter D. Eimas and Vivien C. Tartter

The Economics of Infancy: A Review of Conjugate Reinforcement Carolyn Kent Rovee-Collier and Marcy J. Gekoski

Human Facial Expressions in Response to Taste and Smell Stimulation Jacob E. Steiner

AUTHOR INDEX--SUBJECT INDEX

Volume 14

Development of Visual Memory in Infants John S. Werner and Marion Perlmutter

Volume 16

The History of the Boyd R. McCandless Young Scientist Awards: The First Recipients David Palermo

Social Bases of Language Development: A Reassessment Elizabeth Bates, Inge Bretherton, Marjorie Beeghly-Smith, and Sandra McNew

Perceptual Anisotrophies in Infancy: Ontogenetic Origins and Implications of Inequalities in Spatial Vision Marc H. Bornstein

Contents o f Previous Volumes

Concept Development Martha J. Farah and Stephen M. Kosslyn Production and Perception of Facial Expressions in Infancy and Early Childhood Tiffany M. Field and Tedra A. Walden Individual Differences in Infant Sociability: Their Origins and Implications for Cognitive Development Michael E. Lamb The Development of Numerical Understandings Robert S. Siegler and Mitchell Robinson AUTHOR INDEX--SUBJECT INDEX

Volume 17

The Development of Problem-Solving Strategies Deanna Kuhn and Erin Phelps Information Processing and Cognitive Development Robert Kail and Jeffrey Bisanz Research between 1950 and 1980 on Urban-Rural Differences in Body Size and Growth Rate of Children and Youths Howard V. Meredith Word Meaning Acquisition in Young Children: A Review of Theory and Research Pamela Blewitt Language Play and Language Acquisition Stan A. Kuczaj H The Child Study Movement: Early Growth and Development of the Symbolized Child Alexander W. Siegel and Sheldon H. White AUTHOR INDEX--SUBJECT INDEX

Volume 18

The Development of Verbal Communicative Skills in Children Constance R. Schmidt and Scott G. Paris Auditory Feedback and Speech Development Gerald M. Siegel, Herbert L. Pick, Jr., and Sharon R. Garber Body Size of Infants and Children around the World in Relation to Socioeconomic Status Howard V. Meredith Human Sexual Dimorphism: Its Cost and Benefit James L. Mosley and Eileen A. Stan

243

Symposium on Research Programs: Rational Alternatives to Kuhn's Analysis of Scientific Progress--Introductory Remarks Hayne W. Reese, Chairman World Views and Their Influence on Psychological Theory and Research: Kuhn-Lakatos-Laudan Willis F. Overton The History of the Psychology of Learning as a Rational Process: Lakatos versus Kuhn Peter Barker and Barry Gholson Functionalist and Structuralist Research Programs in Developmental Psychology: Incommensurability or Synthesis? Harry Beilin In Defense of Kuhn: A Discussion of His Detractors David S. Palermo Comments on Beilin's Epistemology and Palermo's Defense of Kuhn Willis F. Overton From Kuhn to Lakatos to Laudan Peter Barker and Barry Gholson Overton's and Palermo's Relativism: One Step Forward, Two Steps Back Harry Beilin AUTHOR INDEX--SUBJECT INDEX

Volume 19

Response to Novelty: Continuity versus Discontinuity in the Developmental Course of Intelligence Cynthia A. Berg and Robert J. Sternberg Metaphoric Competence in Cognitive and Language Development Marc Marschark and Lynn Nall The Concept of Dimensions in Developmental Research Stuart I. Offenbach and Francine C. Blumberg Effects of the Knowledge Base on Children's Memory Strategies Peter A. Ornstein and Mary J. Naus Effects of Sibling Spacing on Intelligence, Interfamilial Relations, Psychosocial Characteristics, and Mental and Physical Health Mazie Earle Wagner, Herman J. P. Schubert, and Daniel S. P. Schubert

244

Contents of Previous Volumes

Infant Visual Preferences: A Review and New Theoretical Treatment Martin S. Banks and Arthur P Ginsburg AUTHOR INDEXwSUBJECT INDEX

Volume 20

Variation in Body Stockiness among and within Ethnic Groups at Ages from Birth to Adulthood Howard V. Meredith The Development of Conditional Reasoning: An Iffy Proposition David P O'Brien Content Knowledge: Its Role, Representation, and Restructuring in Memory Development Michelene T. H. Chi and Stephen J. Ceci Descriptions: A Model of Nonstrategic Memory Development Brian P. Ackerman Reactivation of Infant Memory: Implications for Cognitive Development Carolyn Rovee-Collier and Harlene Hayne Gender Segregation in Childhood Eleanor E. Maccoby and Carol Nagy Jacklin Piaget, Attentional Capacity, and the Functional Implications of Formal Structure Michael Chapman INDEX

Volume 21

Social Development in Infancy: A 25-Year Perspective Ross D. Parke On the Uses of the Concept of Normality in Developmental Biology and Psychology Eugene S. Gollin, Gary Stahl, and Elyse Morgan Cognitive Psychology: Mentalistic or Behavioristic? Charles C. Spiker Some Current Issues in Children's Selective Attention Betty J. House Children's Learning Revisited: The Contemporary Scope of the Modified Spence Discrimination Theory Joan H. Cantor and Charles C. Spiker

Discrimination Learning Set in Children Hayne W. Reese A Developmental Analysis of Rule-Following Henry C. Riegler and Donald M. Baer Psychological Linguistics: Implications for a Theory of Initial Development and a Method for Research Sidney W. Bijou Psychic Conflict and Moral Development Gordon N. Cantor and David A. Parton Knowledge and the Child's Developing Theory of the World David S. Palermo Childhood Events Recalled by Children and Adults David B. Pillemer and Sheldon H. White INDEX

Volume 22

The Development of Representation in Young Children Judy S. DeLoache Children's Understanding of Mental Phenomena David Estes, Henry M. Wellman, and Jacqueline D. Woolley Social Influences on Children's Cognition: State of the Art and Future Directions Margarita Azmitia and Marion Perlmutter Understanding Maps as Symbols: The Development of Map Concepts Lynn S. Liben and Roger M. Downs The Development of Spatial Perspective Taking Nora Newcombe Developmental Studies of Alertness and Encoding Effects of Stimulus Repetition Daniel W. Smothergill and Alan G. Kraut Imitation in Infancy: A Critical Review Claire L. Poulson, Leila Regina de Paula Nunes, and Steven E Warren AUTHOR INDEXmSUBJECT INDEX

Volume 23

The Structure of Developmental Theory Willis E Overton

Contents of Previous Volumes

Questions a Satisfying Developmental Theory Would Answer: The Scope of a Complete Explanation of Development Phenomena Frank B. Murray The Development of World Views: Toward Future Synthesis? Ellin Kofsky Scholnick Metaphor, Recursive Systems, and Paradox in Science and Developmental Theory Willis E Overton Children's Iconic Realism: Object versus Property Realism Harry Beilin and Elise G. Pearlman The Role of Cognition in Understanding Gender Effects Carol Lynn Martin Development of Processing Speed in Childhood and Adolescence Robert Kail Contextualism and Developmental Psychology Hayne W. Reese Horizontality of Water Level: A Neo-Piagetian Developmental Review Juan Pascual-Leone and Sergio Morra AUTHOR INDEX--SUBJECT INDEX

Volume 24

Music and Speech Processing in the First Year of Life Sandra E. Trehub, Laurel J. Trainor, and Anna M. Unyk Effects of Feeding Method on Infant Temperament John Worobey The Development of Reading Linda S. Siegel Learning to Read: A Theoretical Synthesis John P. Rack, Charles Huhne, and Margaret J. Snowling Does Reading Make You Smarter? Literacy and the Development of Verbal Intelligence Keith E. Stanovich Sex-of-Sibling Effects: Part I. Gender Role, Intelligence, Achievement, and Creativity Mazie Earle Wagner, Herman J. P. Schubert, and Daniel S. P. Schubert The Concept of Same Linda B. Smith

245

Planning as Developmental Process Jacquelyn Baker-Sennett, Eugene Matusov, and Barbara Rogoff AUTHOR INDEX--SUBJECT INDEX

Volume 25

In Memoriam: Charles C. Spiker (1925-1993) Lewis P. Lipsitt Developmental Differences in Associative Memory: Strategy Use, Mental Effort, and Knowledge Access Interactions Daniel W. Kee A Unifying Framework for the Development of Children's Activity Memory Hilary Horn Ratner and Mary Ann Foley Strategy Utilization Deficiencies in Children: When, Where, and Why Patricia H. Miller and Wendy L. Seier The Development of Children's Ability to Use Spatial Representations Mark Blades and Christopher Spencer Fostering Metacognitive Development Linda Baker The HOME Inventory: Review and Reflections Robert H. Bradley Social Reasoning and the Varieties of Social Experiences in Cultural Contexts Elliot Turiel and Cecilia Wainryb Mechanisms in the Explanation of Developmental Change Harry Beilin AUTHOR INDEX--SUBJECT INDEX

Volume 26

Preparing to Read: The Foundations of Literacy Ellen Bialystok The Role of Schemata in Children's Memory Denise Davidson The Interaction of Knowledge, Aptitude, and Strategies in Children's Memory Performance David E Bjorklund and Wolfgang Schneider Analogical Reasoning and Cognitive Development Usha Goswami

246

Contents of Previous Volumes

Sex-of-Sibling Effects: A Review Part II. Personality and Mental and Physical Health Mazie Earle Wagner, Herman J. P. Schubert, and Daniel S. P. Schubert Input and Learning Processes in First Language Acquisition Ernst L. Moerk AUTHOR INDEX--SUBJECT INDEX

Volume 27

From Form to Meaning: A Role for Structural Alignment in the Acquisition of Language Cynthia Fisher

ISBN 0-12-009728-1 90051

The Role of Essentialism in Children's Concepts Susan A. Gelman Infants' Use of Prior Experiences with Objects in Object Segregation: Implications for Object Recognition in Infancy Amy Needham and Avani Modi Perseveration and Problem Solving in Infancy Andrka Aguiar and Renke Baillargeon Temperament and Attachment: One Construct or Two? Sarah C. Mangelsdorf and Cynthia A. Frosch The Foundation of Piaget's Theories: Mental and Physical Action Harry Beilin and Gary Fireman AUTHOR INDEX--SUBJECT INDEX