DEVELOPMENTAL ORIGINS OF AGGRESSION
This page intentionally left blank
DEVELOPMENTAL ORIGINS OF AGGRESSION
Edited ...
103 downloads
1800 Views
2MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
DEVELOPMENTAL ORIGINS OF AGGRESSION
This page intentionally left blank
DEVELOPMENTAL ORIGINS OF AGGRESSION
Edited by RICHARD E. TREMBLAY WILLARD W. HARTUP JOHN ARCHER
THE GUILFORD PRESS New York London
© 2005 The Guilford Press A Division of Guilford Publications, Inc. 72 Spring Street, New York, NY 10012 www.guilford.com All rights reserved No part of this book may be reproduced, translated, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission from the Publisher. Printed in the United States of America This book is printed on acid-free paper. Last digit is print number: 9 8 7 6 5 4 3 2 1
Library of Congress Cataloging-in-Publication Data Developmental origins of aggression / edited by Richard E. Tremblay, Willard W. Hartup, John Archer. p. cm. Includes bibliographical references and index. ISBN 1-59385-110-3 (hardcover : alk. paper) 1. Aggressiveness. 2. Aggressiveness—Longitudinal studies. 3. Child psychology. I. Tremblay, Richard Ernest. II. Hartup, Willard W. III. Archer, John, 1944– BF575.A3D45 2005 155.2′32—dc22 2004025158
In memory of Joan McCord (1930–2004), colleague, role model, and friend
(left to right): Richard E. Tremblay, Joan McCord, Willard W. Hartup, and John Archer.
This page intentionally left blank
About the Editors
About the Editors
Richard E. Tremblay, PhD, is Professor of Pediatrics, Psychiatry, and Psychology at the University of Montreal and Director of the Inter-University Research Unit on Children’s Psychosocial Maladjustment. He is also Director of the Centre of Excellence for Early Childhood Development, Canada Research Chair in Child Development, Molson Fellow of the Canadian Institute for Advanced Research, Fellow of the Academy of Experimental Criminology, and Fellow of the Royal Society of Canada. For over 20 years, he has conducted a program of longitudinal and experimental studies addressing the physical, cognitive, emotional, and social development of children from conception onward to understand the development and prevention of antisocial behavior. Willard W. Hartup, EdD, is Regents’ Professor Emeritus and former Director of the Institute of Child Development at the University of Minnesota. Dr. Hartup has spent many years researching friendship and peer relations in child development, antipathies and their significance, and conflict and aggression in childhood and adolescence. He obtained an EdD degree from Harvard University, and has received the G. Stanley Hall Award for Distinguished Contributions to Developmental Psychology from the American Psychological Association and Distinguished Scientific Contribution Awards from both the Society for Research in Child Development and the International Society for the Study of Behavioural Development. John Archer, PhD, is Professor of Psychology at the University of Central Lancashire, Preston, United Kingdom. He received a PhD from the University of Bristol in 1970. A Fellow of the British Psychological Society and President of the International Society for Research on Aggression, his research is concerned with human aggression, grief and loss, and sex differences. Dr. Archer is also author of several books, including Sex and Gender (with Barbara Lloyd) and The Nature of Grief; numerous book chapters; and over 100 articles in refereed journals covering psychology, medicine, and biology. In recent years, he has published a number of meta-analytic reviews on topics connected with sex differences in aggression. vii
This page intentionally left blank
Contributors
Contributors
John Archer, PhD, Department of Psychology, University of Central Lancashire, Preston, United Kingdom Chawki Benkelfat, MD, Department of Psychiatry, McGill University, Montreal, Quebec, Canada Michel Boivin, PhD, Research Unit on Children’s Psychosocial Maladjustment, School of Psychology, Laval University, Quebec City, Quebec, Canada Mara Brendgen, PhD, Department of Psychology, University of Quebec at Montreal, Montreal, Quebec, Canada Sylvana Côté, PhD, School of Psychoeducation, University of Montreal, Montreal, Quebec, Canada Ginette Dionne, PhD, School of Psychology, Laval University, Quebec City, Quebec, Canada Joseph L. Flanders, BSc, Department of Psychology, McGill University, Montreal, Quebec, Canada Afra Foroud, MSc, Canadian Centre for Behavioural Neuroscience, Department of Psychology and Neuroscience, University of Lethbridge, Lethbridge, Alberta, Canada Uberto Gatti, MD, Section of Criminology and Forensic Psychiatry, University of Genoa, Genoa, Italy Paul L. Gendreau, PhD, School of Psychoeducation, University of Montreal, Montreal, Quebec, Canada Willard W. Hartup, EdD, Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Dale F. Hay, PhD, School of Psychology, Cardiff University, Cardiff, Wales, United Kingdom ix
x
Contributors
D. Lynn Homish, BSc, Western Psychiatric Institute and Clinic, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania David Joubert, MA, MCA, Department of Psychology, University of Quebec at Montreal, Montreal, Quebec, Canada Eric Lacourse, PhD, Department of Sociology, University of Montreal, Montreal, Quebec, Canada Rolf Loeber, PhD, Western Psychiatric Institute and Clinic, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania Tania Mazzarello, BSc, Department of Psychology, University of Quebec at Montreal, Montreal, Quebec, Canada Daniel S. Nagin, PhD, H. J. Heinz III School of Public Policy and Management, Carnegie Mellon University, Pittsburgh, Pennsylvania Tomáë Paus, MD, PhD, Departments of Neurology and Neurosurgery, and Psychology, McGill University, Montreal, Quebec, Canada Sergio M. Pellis, PhD, Canadian Centre for Behavioural Neuroscience, Department of Psychology and Neuroscience, University of Lethbridge, Lethbridge, Alberta, Canada Vivien C. Pellis, PhD, Canadian Centre for Behavioural Neuroscience, Department of Psychology and Neuroscience, University of Lethbridge, Lethbridge, Alberta, Canada Daniel Pérusse, PhD, Department of Anthropology, University of Montreal, Montreal, Quebec, Canada Jordan B. Peterson, PhD, Department of Psychology, University of Toronto, Toronto, Ontario, Canada Robert O. Pihl, PhD, Departments of Psychology and Psychiatry, McGill University, Montreal, Quebec, Canada François Poulin, PhD, Department of Psychology, University of Quebec at Montreal, Montreal, Quebec, Canada Elisa Romano, PhD, Department of Psychology, University of Ottawa, Ottawa, Ontario, Canada Jean Richard Séguin, PhD, Research Unit on Children’s Psychosocial Maladjustment, Department of Psychiatry, University of Montreal, Montreal, Quebec, Canada Stephen J. Suomi, PhD, Laboratory of Comparative Ethology, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland Richard E. Tremblay, PhD, Departments of Pediatrics, Psychiatry, and Psychology, University of Montreal, Montreal, Quebec, Canada
Contributors
Tracy Vaillancourt, PhD, Department of Psychology, McMaster University, Hamilton, Ontario, Canada Stephanie H. M. Van Goozen, PhD, School of Psychology, Cardiff University, Cardiff, Wales, United Kingdom Frank Vitaro, PhD, Research Unit on Children’s Psychosocial Maladjustment, School of Psychoeducation, University of Montreal, Montreal, Quebec, Canada Philip David Zelazo, PhD, Department of Psychology, University of Toronto, Toronto, Ontario, Canada Mark Zoccolillo, MD, Department of Psychiatry, Montreal Children’s Hospital, Montreal, Quebec, Canada
xi
This page intentionally left blank
Preface
Preface
It has been over 30 years since the publication of a book giving a detailed overview of the state of knowledge on the developmental origins of aggressive behavior (de Wit & Hartup, 1974). Much has changed during this time. Most of the developmental work published in the 1974 book was based on crosssectional studies of small samples of school-age children. The focus of attention was on aggressive events and the situations that instigate them. Over the past 30 years, investigators have conducted longitudinal studies of large samples of children assessed repeatedly from birth to adulthood. Results from these studies are telling unexpected stories. For example, children appear not to be learning to use physical aggression as they grow older; rather they appear to be learning not to use physical aggression. As expected, girls use physical aggression less often than boys from infancy to adolescence; however, they use another form of aggression (indirect aggression) more often than boys from early childhood to adulthood. Such findings are clearly of interest for our general understanding of human nature. Philosophers have argued for centuries on the origins of antisocial behavior. Rousseau defended the thesis that humans are created good and become evil through the influence of society, and Hobbes argued that a wicked person was simply a child who had not grown up. In 1973, Albert Bandura wrote, “People are not born with preformed repertoires of aggressive behavior; they must learn them in one way or another” (p. 61). The findings from research on the development of aggressive behavior are also extremely important for the prevention and treatment of violent behavior. Violence is still a serious problem in our modern societies. To help children learn alternatives to violent behavior, we need to understand the development of aggression and its alternatives. Every professional who works with children and adolescents needs to understand the developmental origins of aggression. Technological progress over the past 30 years has enabled investigators to combine longitudinal studies of large samples of children over their lifespans with molecular assessments of their genetic material and assessments of xiii
xiv
Preface
hormones and neuromodulators through blood and saliva samples and brain imaging. More sophisticated statistical techniques provide researchers with tools to make better use of twin samples and repeated measurements over long periods of time. This volume provides a timely overview of the best research and the key questions at the beginning of a new era, which will lead to a synthesis of genetic, brain, behavior, and environmental research.
REFERENCES Bandura, A. (1973). Aggression: A social learning analysis. New York: Holt. de Wit, J., & Hartup, W. W. (Eds.). (1974). Determinants and origins of aggressive behavior. The Hague: Mouton.
Contents
Contents
I. INTRODUCTION 1
The Development of Aggression: Where Do We Stand?
3
Willard W. Hartup
II. THE DEVELOPMENT OF AGGRESSION IN ANIMALS AND HUMANS 2
Subtypes of Aggression in Humans and Animals
25
Paul L. Gendreau and John Archer
3
Play Fighting: Aggression, Affiliation, and the Development of Nuanced Social Skills
47
Sergio M. Pellis, Vivien C. Pellis, and Afra Foroud
4
Genetic and Environmental Factors Influencing the Expression of Impulsive Aggression and Serotonergic Functioning in Rhesus Monkeys
63
Stephen J. Suomi
5
The Developmental Origins of Physical Aggression in Humans
83
Richard E. Tremblay and Daniel S. Nagin
6
The Beginnings of Aggression in Infancy
107
Dale F. Hay
7
Play and the Regulation of Aggression
133
Jordan B. Peterson and Joseph L. Flanders
8
Indirect Aggression among Humans: Social Construct or Evolutionary Adaptation?
158
Tracy Vaillancourt
9
Proactive and Reactive Aggression: A Developmental Perspective
178
Frank Vitaro and Mara Brendgen
10
Homicide, Violence, and Developmental Trajectories Rolf Loeber, Eric Lacourse, and D. Lynn Homish
xv
202
xvi
Contents
III. DETERMINANTS OF AGGRESSION 11
Genetics and the Development of Aggression
223
Daniel Pérusse and Paul L. Gendreau
12
Mapping Brain Development and Aggression
242
Tomáë Paus
13
Neuromodulators in the Development and Expression of Inhibition and Aggression
261
Robert O. Pihl and Chawki Benkelfat
14
Hormones and the Developmental Origins of Aggression
281
Stephanie H. M. Van Goozen
15
Executive Function in Early Physical Aggression
307
Jean Richard Séguin and Philip David Zelazo
16
Language Development and Aggressive Behavior
330
Ginette Dionne
17
The Intergenerational Transmission of Aggression and Antisocial Behavior
353
Mark Zoccolillo, Elisa Romano, David Joubert, Tania Mazzarello, Sylvana Côté, Michel Boivin, Daniel Pérusse, and Richard E. Tremblay
18
Peer Relationships and the Development of Aggressive Behavior in Early Childhood
376
Michel Boivin, Frank Vitaro, and François Poulin
19
Social Capital and Physical Violence
398
Uberto Gatti and Richard E. Tremblay
20
Sex Differences in Aggressive Behavior: A Developmental and Evolutionary Perspective
425
John Archer and Sylvana Côté
IV. CHALLENGES FOR THE FUTURE 21
The Developmental Origins of Aggression: Where Are We Going?
447
Richard E. Tremblay and Sylvana Côté
Index
465
Part I INTRODUCTION
This page intentionally left blank
The Development of Aggression INTRODUCTION
1 The Development of Aggression Where Do We Stand? W ILLARD W. H ARTUP
Three changes have occurred in research on the development of aggression. First, both theoretical and empirical attention have shifted from the aggressive act, on one hand, to the development of aggressive individuals, on the other; that is, from aggressive events and the situations that instigate them to aggressive children assessed within ecological systems. Investigators no longer examine the effects of exposure to aggressive models (Bandura & Walters, 1963), for example, but rather the developmental consequences of reciprocities in social interaction involving individuals, bilateral and transactional processes in family and peer interaction, gene–environment permutations, and contextual components (including exposure to aggression and violence in the community). A second change is the emerging synthesis between research on aggression and research on antisocial behavior (Tremblay, 2000). Everyone knows that aggression and antisocial behavior are overlapping but not identical constructs. Nevertheless, because aggressive behavior is relatively stable as an individual characteristic, and because stable aggression predicts antisocial behavior during adolescence and adulthood for males at least (Moffitt, Caspi, Rutter, & Silva, 2001), it no longer seems wise to separate these terms in thinking about social development and adaptation. Third, a trend toward more developmentally oriented studies of aggression has emerged. What was known about aggression in childhood and adolescence in the late 1960s was quite nondevelopmental: The relevant work dealt with contingencies in observation or social interaction that increased aggression in the child, usually in constrained situations and with little attention to age differences. Normative studies encompassed only a few years (in early 3
4
INTRODUCTION
childhood), and the events associated with individual differences in the real world were not much remarked on (see Feshbach, 1970). To be concerned with development, however, is to be concerned with continuities and discontinuities in harm-doing activity across the lifespan, especially during the years before maturity, along with changes across time in the morphology and functioning of aggressive behavior, the reorganization of various mental structures as these involve aggression, and the role of violence in the development of attitudes toward oneself and others (Hartup & DeWit, 1974). By the 1980s, new ideas were percolating through the entire field of social and emotional development even though long-term longitudinal data were scarce. Studies linking aggression in early childhood with aggression in adolescence and adulthood were extremely rare; even cross-sectional studies were not abundant. But the zeitgeist was shifting. Ross Parke and Ronald Slaby (1983) knew it, and called their chapter for the Handbook of Child Psychology, “The Development of Aggression.” Ahead of its time, this chapter led the way to fundamental changes in research and research strategy highlighting developmental processes in the emergence of aggression and antisocial behavior. Nowadays, the major studies of aggression and antisocial behavior are definitely developmental, and most use longitudinal data to search for determinants, dynamics, and outcomes of these behaviors. Against the background of these comments, several issues are now discussed in terms of how things have changed and what we might expect in the near future in understanding the development of aggression: First, what changes? Three issues are considered: (1) construct selection, (2) normative change versus individual differences, and (3) child development versus lifespan perspectives. Second, what are the determinants of developmental change in aggression and antisocial behavior? Again, three issues are addressed: (1) nativism versus empiricism, (2) socialization and relationships, and (3) the ecological context.
WHAT CHANGES? Aggression, like most human activity, changes with age both in its incidence and in the processes leading to its instigation. But which changes are meaningful? Not every change in language and cognitive activity at every level of analysis is meaningful, nor is every change in aggression and antisocial behavior. Our first task, then, is to identify those changeworthy behaviors that have some significance in behavioral adaptation across time.
Construct Selection One must proceed cautiously through the thicket of constructs available in the current aggression literature. Merely specifying a workable construct that can
The Development of Aggression
5
be called “aggression” is difficult. Saying that aggression is “intentional harm doing” or “harm doing for its own sake” may be reasonable for the ordinary user of the English language, but it is fraught with difficulties for scientists who want to describe relevant phenomena with precision: Intentions cannot be observed easily; instigating conditions are difficult to separate from the structure of the aggressive act; outcomes are difficult to specify; and we can’t always tell whether the act has been aversive to the victim (Bandura, 1973; Hartup & De Wit, 1974; Coie & Dodge, 1998; Tremblay, 2000). Moreover, similar measures across age (if not equivalent ones) must be used in many forms of developmental analysis (e.g., specifying developmental trajectories and pathways), and one cannot always count on this equivalence. Researchers have made much progress by becoming more global in defining aggression and antisocial behavior, that is, by making increasingly greater use of global constructs drawn from natural language. Many investigators rely on ordinary observers (parents, teachers, and children) to tell them when aggression occurs and what some of its implications are. For example, elementary school children are asked to tell us which classmates “start fights.” When they answer, the reliability and validity of these nominations turn out to be substantial even though we do not know the circumstances under which the target child may start fights, the behaviors used in fighting, the consequences to the child’s companions, or the goals achieved by the target child. At certain times, however, these “down and dirty” measurement strategies will not work, that is, for determining whether babies “start fights” or whether 3-year-old and 13-year-old aggression reflect the same functional dynamics. Particulars about instigation, motor patterns, and consequences (functions) are needed in those instances even if it is difficult to obtain them. One endeavor needs to be encouraged strongly: construct selection at intermediate levels of conceptual complexity. First, certain older constructs need to be maintained. Consider that physical and verbal aggression have been distinguished from one another for many years (Goodenough, 1931), and several examples will show that this distinction remains useful: 1. Neither physical aggression nor most forms of antisocial behavior occur as frequently among females as among males, and gender differences in aggression cannot be understood without taking this into consideration. 2. Physical aggression among females during childhood does not predict adolescent antisocial behavior as it does for males (Broidy et al., 2003), and we do not completely understand this. Other midlevel constructs must be preserved, for example, violent versus nonviolent aggression, instrumental versus hostile aggression, and proactive versus reactive aggression. Bullying and victimization are also uniquely useful constructs. It may be time to resurrect other old constructs such as prosocial versus antisocial aggression; this one has languished since Robert Sears and
6
INTRODUCTION
his colleagues could not find differential correlates in the doll play of kindergarten children more than 40 years ago (Maccoby, 1992). But we must put this old wine in new bottles. New midlevel constructs are badly needed. Relational (or indirect) aggression has evoked considerable interest in the last decade (Crick et al., 1998; Vaillancourt, Chapter 8, this volume) Research shows that relational aggression is more frequent among females than among males, although its sequelae are not remarkably different (Crick et al., 1998; Vaillancourt, Chapter 8, this volume). It is not really surprising that harm doing may occur in normatively different ways in males and females; it is only surprising that it took us so long to realize this. Natural language may or may not furnish us with the complete conceptual armamentarium needed for research on aggression and antisocial behavior. As a consequence, the research community needs to look beyond the dictionary for constructs that will invigorate their work.
Normative Change or Individual Differences? Once upon a time, the following item appeared regularly on the exams for my graduate course: “Discuss the similarities and differences between the study of normative issues in developmental psychology and individual differences: Demonstrate how each of these issues informs the other about developmental process.” Historically, developmental psychology has been much concerned with both issues, with students of cognitive development emphasizing normative continuities and change rather than the emergence of individual differences (except for investigators interested in the IQ); students of social and personality development, however, have been interested mainly in individual differences. Normative change, of course, refers to age-related changes that are universal or characteristic of major population groups rather than developmental variations within or between groups. Aggression researchers, especially earlier ones, were interested in normative questions ranging from the ontogeny of anger regulation (Goodenough, 1931), to delinquency (Glueck & Glueck, 1950), to verbal mediation in the child’s response to aggressive modeling (Coates & Hartup, 1969). Contemporary researchers remain interested in normative assessment of aggression, as revealed in Figure 1.1 which shows data on hitting, biting, and kicking in a Canadian sample (Tremblay et al., 1996). These behaviors peak at about 27 months for both boys and girls, with slightly more boys than girls showing them “sometimes” or “frequently.” Frequencies decline steadily thereafter and are not often evident (but not completely absent) by age 12. These data represent one of the most comprehensive normative reports on physical aggression existing anywhere. Normative studies, however, are increasingly difficult to distinguish from differential studies of aggression. Normative and differential research on ag-
The Development of Aggression
7
FIGURE 1.1. Hitting, biting, and kicking (boys and girls ages 2–11 years). From Tremblay et al. (1996). Statistics Canada information is used with the permission of the Minister of Industry, as Minister responsible for Statistics Canada. Information on the availability of the wide range of data from Statistics Canada can be obtained from Statistics Canada’s Regional Offices, its website at www.statcan.ca, and its toll-free access number 1-800-2631136.
gression have blurred together, underscoring something we have known all along but have chosen to ignore: Most normative changes in aggression are constrained by variations among individuals. Recent studies show dramatically that individual-to-individual variations in developmental pathways are ubiquitous. Some of these developmental variations are associated with gender, social class, and ethnicity. Indeed, one can describe many relevant differences between the sexes in developmental trajectories, even though the basic processes by which children acquire certain types of aggression may not differ between boys and girls (Moffitt et al., 2001). Even among children of a single race or gender, however, normative changes encompass a range of individual differences. Nowhere is this conclusion more clearly revealed than in a recent manuscript that brings together data from six well-known longitudinal studies (Broidy et al., 2003). In every instance, semiparametric analyses reveal multiple types of developmental change. For example, four trajectories in physical aggression can be identified within a male sample from Montreal studied by Nagin and Tremblay (1999; Tremblay & Nagin, Chapter 5, this volume). Two are similar to those identified in a number of longitudinal studies: (1) high levels of aggression among a relatively small group of boys that persist across childhood and into adolescence and (2) low aggression that persists across childhood and characterizes a much larger group. Two other trajectories were also identified in these data:
8
INTRODUCTION
(3) one that begins with high levels of aggression and declines steadily as time goes on (boys sometimes called “desistors” or “recovered” aggressors) and (4) one that begins at more moderate levels and also declines. One other example included in this six-site investigation, the so-called Dunedin (New Zealand) studies of Moffitt and her associates (Moffitt, Caspi, Rutter, & Silva, 1996; Broidy, et al., 2003), revealed three distinct trajectories instead of four: (1) a small group of consistently aggressive boys, (2) a large group of consistently nonaggressive boys, and (3) another stable group of moderately aggressive boys. No cluster of decliners was identified in this study. The remaining studies in the investigation show either three- or fourcluster patterns, with only one (based on a sample from the United States) showing a trajectory of physical aggression that increases across childhood (Broidy et al., 2003). Developmental trajectories derived from longitudinal studies of girls are similar to those obtained with boys but also different: For example, four distinctive patterns were obtained in another Quebec study that are similar to those obtained for boys, but with the following exceptions: (1) Scores were decidedly lower than those for boys across the board, and (2) the desistors among the females declined more sharply across childhood into adolescence than those among the boys (see Broidy et al., 2003). Once again, a small group of chronic aggressors emerged, as well as a larger group of consistently nonaggressive girls. Among the girls in the Dunedin (New Zealand) sample, again, the trajectories are much like the boys’, except scores are lower and the decline in two groups is a bit greater (see Broidy et al., 2003). Normatively, then, the existence of two particular trajectories seems ubiquitous—a high level of aggression that is continuous throughout childhood and into adolescence and a low level of aggression that is equally consistent across age. Other patterns, including those characterized by aggression that declines during childhood and those typified by increases around the time of puberty, occur in some data sets but not others. No one is entirely certain as to why these sample-to-sample variations occur in the types of trajectories identified. Moreover, clear evidence does not yet establish the reasons for the variations that exist. Some investigators believe that persistent and high levels of aggression across childhood (much more common in boys than in girls) may have biological determinants, whereas “adolescence-limited” aggression may be derived from social experience, especially experience with aggressive peers (Moffitt et al., 2001). Other investigators, however, believe that dysfunctional family relationships in early childhood differentiate persistently aggressive boys from those showing other developmental patterns (Aguilar, Sroufe, Egeland, & Carlson, 2000). These questions are now at center stage in the research on the development of aggression. One conclusion, however, cannot be questioned: Normative change in aggressive behavior must be evaluated in the context of individual differences. The obverse is also true, namely, that individual differences must be evaluated in the context of normative change.
The Development of Aggression
9
Developmental trajectories in aggression development also differ according to characteristics of the individuals involved. For example, my own work with colleagues (Haselager, Cillessen, Van Lieshout, Riksen-Walraven, & Hartup, 2002; Hartup, Cillessen, Haselager, Scholte, & Van Lieshout, 2002a) shows a connection between developmental trajectories in aggression and sociometric status among boys. We examined, first, trajectories in composite measures of peer-reported and self-reported aggression between the ages of 6 and 11 among rejected boys (that is, boys who were peer rejected at a relatively early age). Groups emerging in a cluster analysis somewhat resembled the normative samples from Montreal and other sites (see Figure 1.2): Three more-or-less stable groups (highly aggressive, moderately aggressive, and persistently nonaggressive), as well as one that showed marked decreases in aggressive activity across the 5 years studied (a desistor subgroup). Discriminant function analysis shows that the highly stable aggressive group differed from the other three in social maladaptation, again consistent with the results of numerous studies. The only distinctive characteristic of the children who declined in aggression is that they also increased in cooperation over the period and were better liked by their peers as time went on. Shown in the Figure 1.3 are trajectories in aggressive behavior that emerged when we examined longitudinally a group of 87 boys who were all popular at age 6 (fair-haired boys, we might call them). Cross-sectional cluster analyses (not shown here) revealed differences among these popular children even at this age: Some were more aggressive and less cooperative than others. Two groups, however, emerged from a cluster analysis of the longitudinal
FIGURE 1.2. Developmental pathways of aggressive behavior in four clusters of initially rejected boys. From Haselager, Cillessen, Van Lieshout, Riksen-Walraven, and Hartup (2002). Copyright 2002 by the American Psychological Association. Reprinted by permission.
10
INTRODUCTION
FIGURE 1.3. Developmental pathways of aggression in three initially popular clusters. From Hartup, Cillessen, Haselager, Scholte, and Van Lieshout (2002a).
data—one group composed of individuals who were consistently nonaggressive and another group of individuals (15% of the subjects) who changed decisively in the direction of greater aggressiveness. These increasingly aggressive children were distinctive. Concomitant with their increasing aggression were decreases in cooperativeness and increases in peer rejection. Most interesting, though, are data available for 47 of these boys when they were 14 years of age: Discriminant function analysis identified two functions that successfully discriminated these boys from the other groups: (1) they were more sociable, were less inhibited, and achieved more poorly in school and (2) they were less well liked and more likely to engage in aggression, as well as in more bullying and victimization, than members of the other groups. These results suggest that synergies involving sociometric status and aggressive trajectories must be taken into account in order to describe, on one hand, “model” boys (stable, nonaggressive, popular boys) and, on the other hand, “toughs” (popular kids who are increasingly antisocial and not turned on by school) (Rodkin, Farmer, Pearl, & Van Acker, 2000). The message is clear: Normative change in aggression and antisocial behavior is meaningful only in the context of differential analysis. And the reverse may also be true: individual differences can be understood only in the context of normative analysis. Consideration of these combined forces is necessary to understanding both the development of aggression and social adaptation more generally.
Child Development versus Lifespan Perspectives Owing mainly to child-oriented societal forces, over most of the last century primary attention was given to developmental changes occurring in childhood
The Development of Aggression
11
rather than to changes occurring in adolescence or adulthood. Most studies of aggression and its development at the beginning of the child development movement were devoted to young children (Goodenough, 1931; Dawe, 1934); juvenile delinquency was studied among adolescents by sociologists and psychiatrists (Thrasher, 1927; Glueck & Glueck, 1934), but other aspects of adolescent aggressive behavior were largely ignored. These truncated perspectives dominate developmental science even today, although they have been modified in various ways—i.e., the doctrine of early experience is questioned in important aspects (Schaffer, 2002) and lifespan perspectives are promulgated as frameworks for examining developmental changes (Baltes, Lindenberger, & Staudinger, 1998). The bias toward childhood studies of behavioral development stems from theoretical/scientific sources as well as the nature of the early child welfare movement in the United States and elsewhere (Hartup, Johnson, & Weinberg, 2001). First, developmental science was heavily influenced in the early 20th century by psychoanalysis, on one hand, and learning theories, on the other, including their assumptions that development during the early years has pervasive and long-lasting consequences in both cognitive and social development. At the same time, social activists came to believe that the road to general societal progress begins with changes made in the lives of children and families, especially during a child’s early years. Parent education, social welfare services, and early childhood education were expanded greatly in many countries to assist in this effort. Services to adolescents were also improved, especially those relating to juvenile delinquency, but community welfare in the larger sense did not encompass lifespan considerations. Meanwhile, criminologists were discovering that childhood antecedents account for a large amount of variance in adolescent and adult crime (Glueck & Glueck, 1950; McCord, 1979). Clearly, short-term studies of children or adolescents were not going to be able to furnish an adequate basis for social intervention. More recent investigators utilize increasingly longer-term studies of aggression and antisocial behavior. Although infancy and early childhood are not encompassed by most of these studies, longitudinal analysis beginning in childhood and ending in early adulthood teaches us a great deal. We are in debt to David Farrington, Rolf Loeber, Robert Cairns, Leonard Eron and Rowell Husemann, John Coie, Joan McCord, and many others for work like this. It has been difficult to extend these studies through the life course, beginning in infancy and ending in middle or old age. Moffitt et al.’s (2001) work is notable for the fact that the data set begins with measurements at age 3, and newer studies (e.g., Tremblay, 2000) begin prenatally and remain open-ended. Studies ranging from middle childhood to young adulthood are still more common than those covering a longer term, and cradle-to-grave studies do not exist. Nevertheless, the message of lifespan developmentalists is being taken with increasing seriousness: Understanding adaptational problems requires developmental perspectives that extend beyond childhood and adolescence,
12
INTRODUCTION
and the coherence of human development, whether outcomes are good or bad, can be assessed only in terms of both growth and decline in behavioral function. Must we wait for geriatric data on the Montreal subjects or the Dunedin sample before studying the decline in aggression and antisocial behavior that is known to mark middle and old age? Probably not. Overlapping longitudinal studies—some beginning in adolescence, some beginning later—can unravel the conditions responsible for decline in antisocial behavior with age. One can argue, then, that normative change must be studied not only within the context of individual differences, but also within the context of the life course.
WHAT ARE THE DETERMINANTS OF BEHAVIOR CHANGE? Our ultimate goal is to identify the conditions responsible for normative development of aggression within the context of individual differences. As everyone knows, explanation in the behavioral sciences is extraordinarily difficult. Most developmental change is multidetermined, and causation is buried deeply within the genome and the mind. Moreover, the time separating “cause” from “effect” in human development is frequently decades, not moments, and change within one behavioral domain is entangled with changes in others. One cannot synthesize everything that has been written about the determinants of aggression in the space available here. Instead, several “tensions” or “dialectics” remaining in this field are discussed.
Nativism versus Empiricism The notion that human beings begin life with “innate ideas” and “inborn proclivities” appears in our most ancient religious and philosophical texts. Also expressed in classical philosophical writing is extreme empiricism, the argument being that nearly everything about human behavior—including aggression—derives from the individual’s commerce with the world. Most modern writers, though, view the development of aggression (like many other traits) as constrained by intrinsic structures (these may include the physiological and neural structures necessary to arousal, anger, learning, attention, language, and emotional regulation). The structures in question are species-wide and both determine the success of individual adaptation and make individual differences possible. Although relatively little is known about the processes through which these innate structures are deployed in behavioral development, especially in social and emotional behavior, the evidence suggests that relationships with others (including caretakers, friends, and enemies) are important vectors. The task, then, is to specify both the intrinsic and experiential contributions to developmental acquisitions and the manner in which they “synthesize” or “interact.”
The Development of Aggression
13
Recent studies in molecular genetics suggest that allelic variations may commonly interact with social and relationship experience in determining the aggressive phenotype. Suomi (Chapter 4, this volume) discovered that variations in the serotonin transporter gene (5-HTT) interact with rearing conditions to affect a variety of outcomes, including aggression and position in the dominance hierarchy, among rhesus monkeys. So-called “short” allelic variations are associated with nonaggressive maladaptations—but for peer-reared individuals and not mother-reared individuals. In other words, attachment to a maternal figure moderates the effects of the genetic variations. The findings also partially explain why peer rearing does not produce extremely aggressive animals in all cases. In recent work with human beings, Caspi et al. (2002) discovered an interaction between parental maltreatment and levels of monoamine oxidase A (MAOA) genetic expression (MAOA is a gene located on the X chromosome that encodes an enzyme that metabolizes various neurotransmitters, rendering them inactive). Maltreated boys in a large longitudinal sample possessing the genotype associated with high levels of MAOA expression were less likely to have antisocial problems in the course of their development than those with genotypes associated with low levels of MAOA expression. Results of this study help to account for the fact that, although harsh parental discipline and maltreatment of young boys is related to aggressive and antisocial outcomes, not all maltreated children grow up to victimize others. In addition to these studies, process-oriented prospective studies are needed (beginning earlier than most existing studies) to explain the manner in which so-called biological variations contribute to individual differences in aggressiveness and antisocial behavior. One hopes that investigators will not simplistically consider factors like maternal IQ, maternal age, and maternal social adjustment as well as child IQ, child memory, attention disabilities, and child temperament to be exclusively biological indicators (Moffitt et al., 2001), nor attachment histories to be exclusively social indicators. These constructs are only the roughest indications of intrinsic or social structures, respectively, and the manner in which they work. We need new expertise in psychoendocrine measurement in infancy and beyond, temperament, emotional processes in infancy and early childhood (Thompson, 1998), and behavioral regulation in early relationships between the child and his or her caretakers. Implications of the neuropsychological history for the child’s social and emotional history must be studied along with experience. Overall, refinements in methods for studying psychological processes in the development of aggression are just as important as advances in gene mapping.
Socialization: From Unilateral to Bilateral Constructions Although there were many indications in earlier writings that children are “socializing creatures” as well as “socialized ones” (Baldwin, 1897), a consensus concerning a two-way or bilateral view of childhood socialization was
14
INTRODUCTION
reached only recently (Bell, 1968). The notion that social adaptation is the outcome of interaction between the child and others, rather than some sort of one-way social-molding process, is now accepted by almost everybody.
Parent–Child Relationships With respect to the development of aggression, one does not find many studies that deal directly with social interaction or social relationships within the family. Most commonly, investigators enter measures like the following into their regression models as predictors of child or adolescent outcomes: (1) child characteristics—temperament, intelligence, troublesomeness, memory and attention problems, impulsivity, and personality and (2) parent characteristics— warmth versus negativity, family coercion and inconsistent discipline, punishment and punitiveness, along with abusive parenting (Coie & Dodge, 1998). The ensuing regressions allow us to examine main and interaction effects among these child and maternal characteristics, but not the social exchanges between parents and children from which changes over time actually emanate. One of the most brilliant achievements of the last century is the “performance theory” formulated and tested by Gerald Patterson and his associates. This socialization model specifies the parameters and contingencies occurring in family interaction that instantiate troublesomeness among young boys, maintain it, and supply the basis for the attitudes that characterize these same boys in their early interactions with other children outside the home. These relationships, in turn, involve them in increasingly violent behavior (Patterson, Reid, & Dishion, 1992). In other studies, certain mileage can be gained by using global measures of family conflict, maternal punitiveness, and harsh discipline, but the “performance model” makes clear that such measures are only proxies for measures of social interaction. One concludes that microanalytic approaches to socialization studies of the development of aggression are needed as well as macroanalytic ones. Most parent–child measures also do not pass muster as relationships constructs. Relationships are usually defined as aggregates of interaction between individuals who have formed expectations about one another that guide interpersonal behavior over time (Hinde, 1997). It so happens that relationships are the social contexts within which most development occurs: (1) basic skills in language and communication, (2) competence in coping with both social and nonsocial problems, and (3) the attitudes and abilities needed for constructing other relationships. Relationships theory, however, has not penetrated very deeply into research on the development of aggression. Parental punitiveness or harsh discipline (common measures) are parent characteristics, not dimensions of the parent–child relationship. Similarly, maternal warmth does not describe a relationship between mother and child. Exactly what this omission means for research is not really known. Diana Baumrind (1971) had considerable success earlier in describing permissive relationships between parents and children and showing that these were associated with early ag-
The Development of Aggression
15
gression. Some investigators have consistently reported cross-time correlations between disorganized attachment patterns in infancy and early childhood with manifestations of aggression and antisocial behavior in middle childhood and adolescence (Renken, Egeland, Marvinney, Mangelsdorf, & Sroufe, 1989). Others report similar findings, especially among vulnerable children, for example, children who live in chaotic, poor families with a single parent (see Coie & Dodge, 1998). And, in the Dunedin data set (Moffitt et al., 2001), a relationships measure (“relationship with parents,” based on items like trust, communication, and lack of alienation between adolescents and their parents) turns out to be a better predictor of adolescent aggression (–.44, –.47) than measures of harsh discipline (.19, .27). No one would argue that relationship constructs are the only ones that account for significant variance in adolescent aggression. We have sufficient evidence to argue, however, that relationships measurement definitely belongs in studies of the family antecedents of aggression, especially as we look forward to research dealing with gene–environment interactions.
Peer Relationships Until recently, the significance of peer relations in the development of aggression and antisocial behavior was considered to be mostly a unidirectional process. That is, the prevailing view was that kids get into trouble because they are exposed to the exhortations and examples of deviant peers and are rejected by better-adjusted associates, an occurrence that lowers self-esteem and restricts access to well-socialized companions. Sometimes the interaction between child characteristics (e.g., earlier disruptiveness or troublesomeness) and peer experience (e.g., social rejection by other children) is examined in relation to aggressive outcomes. In most instances, significant variance in child or adolescent aggression flows additively from both child sources (including earlier aggression or compliance difficulties) and the attitudes and behaviors of other children (Coie & Dodge, 1998). Additional variance in the development of aggression, however, may be traced to close relationships with other children, including (1) friends, (2) enemies, and (3) bully–victim relationships (see Boivin, Vitaro, & Poulin, Chapter 18, this volume).
Friends Friendship dyads, in general, are neither more nor less aggressive or antisocial than “neutral” dyads (Hartup, Verhoeven, DeBoer, Scholte, & Van Lieshout, 2002b). Among individual children and adolescents, those who have friends are not as troubled as those who do not, but “not having friends” does not predict aggression very well either concurrently or over time (Hartup, 1996). Friends, however, are more similar to one another in both aggressiveness and antisocial behavior than nonfriends (Haselager, Hartup, Van Lieshout, &
16
INTRODUCTION
Risken-Walraven, 1998; Hartup et al., 2002b), a situation that can also be observed in social networks or cliques (Cairns & Cairns, 1994). Moreover, these similarities reflect more than “birds of a feather flock together.” The day-today interaction that occurs between antisocial friends actually increases their antisocial behavior over time, especially among children already identified as aggressive and rejected (Dishion, 1990; Tremblay, Masse, Vitaro, & Dobkin, 1995). Most likely, this happens because the interaction between aggressive children and their friends is more contentious and conflict-ridden than interaction between nonaggressive friends (Dishion, Andrews, & Crosby, 1995). In addition, overtly aggressive children are not notably intimate with one another nor as exclusive in their relationship attitudes as their nonaggressive counterparts (Grotpeter & Crick, 1996). Thus, aggressive friends are aggressive risks. Friendship quality also contributes significant variance to the development of aggression in childhood and adolescence. In general, unharmonious friendships exacerbate the individual child’s problems in school, create adjustment difficulties, and increase aggressiveness. In many instances, however, friendship quality has a greater effect on children who are at risk in social development than on those who are not. For example, Poulin, Dishion, and Haas (1999) found significant interaction effects between boys’ antisocial behavior (delinquency) at age 13–14 and friendship quality at that same age (harmoniousness) in relation to antisocial behavior one year later, at age 14–15. The greatest effects of low-quality friendships on subsequent delinquency occurred when baseline delinquency levels were high rather than low. In other words, the friendship construct was more strongly related to developmental outcome among vulnerable individuals than among less vulnerable ones. Overall, friends may or may not support good developmental outcome. Empirical studies tell us that aggressive friends (or unharmonious relationships) are risk factors inasmuch as the children are not well socialized, instigate aggressive behavior (e.g., train one another in deviancy), and do not instigate effective regulatory controls. In addition, friendships that are not harmonious instigate increases in aggressive behavior over time, possibly for some of the same reasons, and mainly among children who are at risk (Boivin et al., Chapter 18, this volume).
Enemies Social networks also include relationships that are not based on social attraction but, rather, are rooted in antipathy, animosity, and enmity. The term “mutual antipathies” describes this relationships category; the word encompasses “being enemies” as well as other relationships maintained on the basis of social aversion. Recent studies (Hartup et al., 2002b) show that differences in aggression and antisocial behavior between adolescents who do not like each other are greater than differences between nonenemies (individuals in
The Development of Aggression
17
neutral dyads). This confirms the “repulsion hypothesis,” which states that individuals dislike those who are different from themselves (Rosenbaum, 1986). Adolescents in mutual antipathies are also substantially more antisocial and aggressive than in “control” or “neutral” dyads, as well as more internalizing and victimized by others. These differences are consistent with earlier analyses of individual children and adolescents in this same data set, which suggest that involvement in mutual antipathies is a risk marker. That is, both children and adolescents who are involved in mutual antipathies show more fighting and bullying, social ineffectiveness, and victimization than those who are not so involved. Peer rejection was a covariate in these analyses, so the variance associated with mutual antipathies is unique. These results characterize both boys and girls who are involved in same-sex antipathies, and boys, but not girls, who are involved in mixed-sex antipathies (Abecassis, Hartup, Haselager, Scholte, & Van Lieshout, 2002). Results from several other studies of mutual antipathies and their correlates are not entirely consistent, although these results have been replicated in three instances (see Hodges & Card, 2003).
Bully–Victim Relationships Some bullies and victims dislike each other, some are neutral toward one another, and some are friends. The correlation between bully–victim status and friendship attraction is negative but moderate in magnitude (Perry, Finnegan, Hodges, Kennedy, & Malone, 1993). Although the literature on bullying and victimization is substantial, most of the relevant studies are monadically oriented, that is, centered on questions such as “What makes a bully?” or “What makes a victim?” rather than on ongoing bully–victim relationships. Great progress has been made in understanding bullying and victimization during the last 30 years (see Rigby, 2002), but relationships-oriented ideas are only slowly making their way into this area. As it turns out, both friends and enemies bear on these situations: For example, increases in victimization are inversely related to the number of externalizing problems evinced by friends, suggesting that victims with externalizing friends may well be protected by them (Hodges & Perry, 1999). A significant relation also exists between characteristics of children’s enemies and victimization: The aggressiveness, physical strength, and victimization experienced by a child’s enemy uniquely predicts the child’s own victimization (Card & Hodges, 2003). Much more, however, remains to be learned about relationships between bullies and victims, their varieties, and their vicissitudes.
Relations among Relationships The social networks of most children and adolescents include friends, neutral associates, enemies, and sometimes bullies or victims. Do these relationship experiences moderate one another in the development of aggression? Does in-
18
INTRODUCTION
volvement in one relationship affect correlates of the other? Recently, my colleagues and I (Hartup et al., 2002b) examined adolescent friendship dyads as a function of whether one or both individuals is also involved in a mutual antipathy (i.e., has an enemy). We also asked whether antipathy dyads vary in psychological adjustment as a function of whether one or both individuals has a mutual friend. First, we gathered together all dyads in a large data set in which the adolescents (14-year-olds) mutually nominated each other as “someone you especially like,” as well as all dyads in which the participants mutually nominated each other as “someone you don’t like at all.” Results show that friends with mutual antipathies are both more antisocial and more aggressive than friends without mutual antipathies or those in which only one child has an enemy. Peer victimization and internalizing behavior are also greater among friends in which one or both children have mutual antipathies. Second, we examined pairs of adolescents who did not like one another as well as whether either adolescent had a mutual friend. Antisocial behavior is less frequent among the adolescent enemies who also have friends than among those who do not. Peer victimization is lower, too, when one or both enemies have a friend. But differences are not significant for aggressive behavior according to whether enemies also have friends. Overall, then, friends are more antisocial when they also have enemies; concomitantly, enemies have lower antisocial behavior scores when they also have friends. Aggressiveness, however, is not moderated in the same way: Friends with enemies are more aggressive than friends without, but friendships do not appear to be strong enough to lessen the aggressiveness that accompanies involvement in a mutual antipathy.
Does Context Make a Difference? Currently, the most advanced ideas being pursued about contextualism and development do not include the notion that person and environment can be conceived dualistically. Gone are notions that persons cause environmental change and vice versa. Instead, the modern view treats the person and the environment as integrated or fused; neither is an entity that can be described sensibly without reference to the other, and development is the story of change over time in this integrated system. One must consider development to be characteristic of biological and social systems, not an attribute of either individuals or contexts. Conventional (dualistic) contextualism is well represented in research on the development of aggression. Cross-national studies show differences in developmental trajectories (Broidy et al., 2003); cross-cultural studies, the same (Whiting & Whiting, 1975). There is also clear evidence that ecological factors like neighborhood violence have a great deal to do with the early onset of aggressive behavior. These scenarios, however, are not simple. As Schwartz and colleagues (Schwartz, Gorman, Toblin, & Abou-ezzedine, 2003) demon-
The Development of Aggression
19
strate, exposure to community violence is consistently related to childhood aggression only among children who are involved in a relatively large number of mutual antipathies with other children. Once again, a relationship factor increases risk in the development of aggression among children who are already vulnerable. These results move us a step or two from dualistic notions about individuals and the community because, in this case, relationship systems are seen to moderate the effects of external conditions. Community and societal variables will remain in research on the development of aggression, but notions about the “environment” need to be rethought in terms of ecological systems theory (Gatti & Tremblay, Chapter 19, this volume).
CONCLUSION Conclusions to this survey of research on aggression and its development are relatively straightforward: Conceptualization and empirical studies dealing with the development of aggression need to organized as a series of “nests.” 1. Normative change needs to be nested theoretically in the context of individual differences and within long-term (i.e., lifespan) perspectives. 2. The development of aggression can be understood only when the earliest genetic expressions are examined as nested in early social experience, especially individual social histories and close relationships. 3. The aggressive act must be considered as nested within a broad and complex ecology of harm doing and antisocial behavior. We stand at the edge of the nest, then, regarding what we know about the development of aggression.
REFERENCES Abecassis, M., Hartup, W. W., Haselager, G. J. T., Scholte, R. J. H., & Van Lieshout, C. F. M. (2002). Mutual antipathies and their significance in childhood and adolescence. Child Development, 73, 1543–1556. Aguilar, B., Sroufe, L. A., Egeland, B., & Carlson, E. (2000). Distinguishing the earlyonset-persistent and adolescent-onset antisocial behavior types: From birth to 16 years. Development and Psychopathology, 12, 109–132. Baldwin, J. M. (1897). Social and ethical interpretations in mental development. New York: Macmillan. Baltes, P. B., Lindenberger, U., & Staudinger, U. M. (1998). Life-span theory in developmental psychology. In W. Damon (Series Ed.), R. M. Lerner (Vol. Ed.), Handbook of child psychology: Vol. 1. Theoretical models of human development (5th ed., pp. 1029–1144). New York: Wiley. Bandura, A. (1973). Aggression: A social learning analysis. Englewood Cliffs, NJ: Prentice-Hall.
20
INTRODUCTION
Bandura, A., & Walters, R. H. (1963). Social learning and imitation. New York: Holt, Rinehart & Winston. Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology Monographs, 4(1), 1–103. Bell, R. Q. (1968). A reinterpretation of direction of effects in studies of socialization. Psychological Review, 75, 81–95. Broidy, L. M., Nagin, D. S., Tremblay, R. E., Bates, J. E., Brame, B., Dodge, K. A., Fergusson, D., Horwood, J. L., Loeber, R., Laird, R., Lynam D. R., Moffitt, T. E., Pettit, G. S., & Vitaro, F. (2003). Developmental trajectories of childhood disruptive behaviors and adolescent delinquency: A six site, cross-national study. Developmental Psychology, 39(2), 222–245. Cairns, R. B., & Cairns, B. D. (1994). Lifelines and risks. Cambridge, UK: Cambridge University Press. Card, N. A., & Hodges, E. V. E. (2003) Parent–child relationships and enmity with peers: The role of avoidant and preoccupied attachment. In E. V. E. Hodges & N. A. Card (Eds.), Enemies and the darker side of peer relations (pp. 23–37). San Francisco: Jossey-Bass. Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., Taylor, A., & Poulton, R. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297, 851–854, Coates, B., & Hartup, W. W. (1969). Age and verbalization in observational learning. Developmental Psychology, 1, 556–562. Coie, J. D., & Dodge, K. A. (1998). Aggression and antisocial behavior. In W. Damon (Series Ed.), N. Eisenberg (Vol. Ed.), Handbook of child psychology: Vol. 3. Social, emotional, and personality development (pp. 779–862). New York: Wiley. Crick, N. R., Werner, N. E., Casas, J. F., O’Brien, K. M., Nelson, D. A., Grotpeter, J. K., & Markon, K. (1998). Childhood aggression and gender: A new look at an old problem. In D. Bernstein (Ed.), Nebraska Symposium on Motivation: Vol. 45. Gender and motivation. Lincoln: University of Nebraska Press. Dawe, H. C. (1934). An analysis of two hundred quarrels of preschool children. Child Development, 5, 139–157. Dishion, T. J. (1990). The peer context of troublesome child and adolescent behavior. In P. Leone (Ed.), Understanding troubled and troublesome youth (pp. 128–153). Newbury Park, CA: Sage. Dishion, T. J., Andrews, D. W., & Crosby, L. (1995). Anti-social boys and their friends in early adolescence: Relationship characteristics, quality, and interactional process. Child Development, 66, 139–151. Feshbach, S. (1970). Aggression. In P. H. Mussen (Ed.), Carmichael’s manual of child psychology (Vol. 2, pp. 159–259). New York: Wiley. Glueck, S., & Glueck, E. T. (1934). One thousand juvenile delinquents. Cambridge, MA: Harvard University Press. Gleuck, S., & Glueck, E. T. (1950). Unraveling juvenile delinquency. Cambridge, MA: Harvard University Press. Goodenough, F. L. (1931). Anger in young children. Minneapolis: University of Minnesota Press. Grotpeter, J. K., & Crick, N. R. (1996). Relational aggression, overt aggression, and friendship. Child Development, 67, 2328–2338. Hartup, W. W. (1996). The company they keep: Friendships and their developmental significance. Child Development, 67, 1–13.
The Development of Aggression
21
Hartup, W. W., Cillessen, A. H. N., Haselager, G. J. T., Scholte, R. J., & Van Lieshout, C. F. M. (2002a). Heterogeneity among popular boys: Subtypes and developmental trajectories. Unpublished manuscript, University of Minnesota. Hartup, W. W., & De Wit, J. (1974). The development of aggression: Problems and perspectives. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 595–620). The Hague: Mouton. Hartup, W. W., Johnson, A., & Weinberg, R. A. (2001). The Institute of Child Development: Pioneering in science and application. In W. W. Hartup & R. A. Weinberg (Eds.), Child psychology in retrospect and prospect: The 32nd Minnesota Symposium on Child Psychology (pp. 227– 257). Mahwah, NJ: Erlbaum. Hartup, W. W., Verhoeven, M., DeBoer, R., Scholte, R. J. H., & Van Lieshout, C. F. M. (2002b, August). Heterogeneity of mutual friendships and mutual antipathies: A cross-sectional study. Paper presented at the biennial meeting of the International Society for the Study of Behavioural Development, Ottawa (Canada). Haselager, G. J. T., Cillessen, A. H. N., Van Lieshout, C. F. M., Riksen-Walraven, J. M. A., & Hartup, W. W. (2002). Heterogeneity of social behavior in peer rejected boys across middle childhood: Developmental pathways and their correlates. Developmental Psychology, 38, 446–456. Haselager, G. J. T., Hartup, W. W., Van Lieshout, C. F. M., & Riksen-Walraven, J. M. A. (1998). Similarities between friends and nonfriends in middle childhood. Child Development, 69, 1198–1208. Hinde, R. A. (1997). Relationships: A dialectical perspective. Hove, UK: Psychology Press. Hodges, E. V. E., & Card, N. A. (Eds.) (2003). Enemies and the darker side of peer relations. San Francisco: Jossey-Bass. Hodges, E. V. E., & Perry, D. G. (1999). Personal and interpersonal antecedents and consequences of victimization by peers. Journal of Personality and Social Psychology, 76, 677–685. Maccoby, E. E. (1992). The role of parents in the socialization of children: An historical overview. Developmental Psychology, 28, 1006–1017. McCord, J. (1979). Some child-rearing antecedents of criminal behavior in adult men. Journal of Personality and Social Psychology, 37, 1477–1486. Moffitt, T. E., Caspi, A., Dickson, N., Silva, P., & Stanton, W. (1996). Childhood-onset versus adolescent-onset antisocial conduct problems in males: Natural history from ages 3 to 18 years. Development and Psychopathology, 8, 399–424. Moffitt, T. E., Caspi, A., Rutter, M., & Silva, P. A. (2001). Sex differences in antisocial behavior. Cambridge, UK: Cambridge University Press. Nagin, D., & Tremblay, R. E. (1999). Trajectories of boys’ physical aggression, opposition, and hyperactivity on the path to physically violent and nonviolent juvenile delinquency. Child Development, 70(5), 1181–1196. Parke, R. D., & Slaby, R. (1983). The development of aggression. In P. H. Mussen (Series Ed.), E. M. Hetherington (Vol. Ed.), Handbook of child psychology: Vol. 4. Socialization, personality, and social development (pp. 547–641). New York: Wiley. Patterson, G. R., Reid, J., & Dishion, T. J. (1992). Antisocial boys. Eugene, OR: Castalia Press. Perry, D. G., Finnegan, R. A., Hodges, E. V. E., Kennedy, E., & Malone, M. (1993, August). Aggressive and victimized children’s relations with parents and peers. Paper presented at the annual meeting of the American Psychological Association, Toronto.
22
INTRODUCTION
Poulin, F., Dishion, T. J., & Haas, E. (1999). The peer influence paradox: Friendship quality and deviancy training within male adolescent friendships. Merrill-Palmer Quarterly, 45, 42–61. Renken, B., Egeland, B., Marvinney, D., Mangelsdorf, S., & Sroufe, L. A. (1989). Early childhood antecedents of aggression and passive-withdrawal in early elementary school. Journal of Personality, 57, 257–281. Rigby, K. (2002). Bullying in childhood. In P. K. Smith & C. H. Hart (Eds.), Blackwell handbook of childhood social development (pp. 549–568). Oxford, UK: Blackwell. Rodkin, P. C., Farmer, T. W., Pearl, R., & Van Acker, R. (2000). Heterogeneity of popular boys: Antisocial and prosocial configurations. Developmental Psychology, 36, 14– 24. Rosenbaum, M. E. (1986). The repulsion hypothesis: On the nondevelopment of relationships. Journal of Personality and Social Psychology, 51, 1156–1166. Schaffer, H. R. (2002). The early experience assumption: Past, present, and future. In W. W. Hartup & R. K. Silbereisen (Eds.), Growing points in developmental science (pp. 24–46). Hove, UK: Psychology Press. Schwartz, D., Gorman, A. H., Toblin, R., & Abou-ezzedine, T. (2003). Mutual antipathies in the peer group as a moderating factor in the association between community violence exposure and psychosocial maladjustment. In E. V. E. Hodges and N. A. Card (Eds.), Enemies and the darker side of peer relations (pp. 39–54). San Francisco: Jossey-Bass. Thompson, R. A. (1998). Early sociopersonality development. In W. Damon (Series Ed.), N. Eisenberg (Vol. Ed.), Handbook of child psychology: Vol. 3. Social, emotional, and personality development (pp. 25–104). New York: Wiley. Thrasher, F. M. (1927). The gang. Chicago: University of Chicago Press. Tremblay, R. E. (2000). The development of aggressive behaviour during childhood: What have we learned in the past century? International Journal of Behavioral Development, 24, 129–141. Tremblay, R. E., Boulerice, B., Harden, P. W., McDuff, P., Perusse, D., Pihl, R. O., & Zoccolillo, M. (1996). Do children in Canada become more aggressive as they approach adolescence? In Human Resources Development Canada & Statistics Canada (Eds.), Growing up in Canada: National longitudinal survey of children and youth (Catalogue 89-550, pp. 127–137). Ottawa: Statistics Canada. Tremblay, R. E., Masse, L. C., Vitaro, F., & Dobkin, P. L. (1995). The impact of friends’ deviant behavior on early onset of delinquency: Longitudinal data from 6 to 13 years of age. Development and Psychopathology, 7, 649–667. Whiting, B. B., & Whiting, J. W. M. (1975). Children of six cultures. Cambridge, MA: Harvard University Press.
Part II THE DEVELOPMENT OF AGGRESSION IN ANIMALS AND HUMANS
This page intentionally left blank
THE DEVELOPMENT Subtypes of Aggression OF AGGRESSION
2 Subtypes of Aggression in Humans and Animals P AUL L. G ENDREAU and J OHN A RCHER
From the elementary fencing behavior of fruit flies to the lunging attack of mice or rats, and from the biting of dogs to the more subtle form of indirect aggression in humans, the expression of aggressive behavior becomes increasingly sophisticated and diversified as animals increase in neural complexity. Comparable transformation is observed during the course of human ontogeny. Children first bite, hit, push, and kick (Tremblay et al., 1999); then gradually, they may come to choose more indirect, socially oriented forms of aggression such as spreading negative rumors about a classmate or attempting to alter relationships (Björkqvist, Lagerspetz, & Kaukiainen, 1992). The expression of aggression undoubtedly reaches a peak of complexity in adulthood, not only in terms of expression but also with regard to its multiple causes and consequences (Loeber & Day, 1997). The fact that aggression is not a unitary or homogeneous phenomenon is clearly not in dispute among scientists. More challenging, however, is the need to provide a definition of aggression that would satisfy researchers from most (or, it is hoped, all) spheres of expertise (Archer & Browne, 1989). Some authors have argued that this is an impossible mission because the concept of aggression involves too many variables and determinants and, more particularly, a significant part of social and moral judgment (e.g., Johnson, 1972). Others avoid using the term aggression because of its diffuse meaning (e.g., Patterson & Cobb, 1973), or declare that reaching a consensus on a definition is not necessary to perform a meticulous analysis of aggression (e.g., Cairns, 1979). Finally, some researchers have maintained that aggression is essentially an inadequate concept that ought to be replaced by more precise terminology (e.g., coercive control; Tedeschi & Felson, 1994; Felson, 2002). 25
26
THE DEVELOPMENT OF AGGRESSION
Before reaching a potential agreement about a definition, a first step is to describe and classify the different types of aggression. As Moyer (1968) argued, “progress in understanding the general phenomenon of aggressive behavior can only be made when the various aggressions are carefully and operationally defined” (p. 65). Hence, several attempts have been made to sort the different forms of aggression into discrete categories. Various taxonomies have been offered for both animal and human aggression, based on the similarity of behavioral expression, contextual characteristics of the eliciting stimulus, functional significance, motivational underpinning, or underlying neurophysiological mechanisms. In this chapter we review the diverse forms of aggression in both animals and humans in an attempt to integrate the current knowledge into a coherent theoretical and practical framework. We argue that a better understanding of the concept of aggression requires a systematic examination of the various proximal and distal forces that induce, facilitate, or maintain the development and expression of the different subtypes of aggression in both humans and animals. Comparing the similarities and dissimilarities of these multiple forces in a variety of species, and among individuals within a given species, is the only possible strategy for formulating a universal taxonomy of aggression.
COMPARING DIFFERENT LEVELS OF PHYLOGENETIC AND ONTOGENETIC ORGANIZATION The Risk of Anthropomorphism and Zoomorphism To recognize the phylogenetic sophistication of aggression, it is important to discriminate between subtypes of aggression that are unique to humans and those that are also present in other species. The ontogenetic sophistication of aggression will become manifest only by differentiating subtypes of aggression that are expressed during a specific developmental period from those that occur throughout ontogeny. Distinctions should be articulated behaviorally, functionally, and neurobiologically. At the behavioral level, a number of aggression-related behaviors are strikingly similar among many species, even in those that are extremely distant in phylogenetic terms. For instance, a “boxing” stance (an upright posture observed during fighting) is shown by fruit flies, rats, and humans. Although the behavioral repertoire of flies and rodents is more complex than one would expect (Chen, Lee, Bowens, Huber, & Kravitz, 2002; Barnett & Marples, 1981; Gendreau, Gariépy, Petitto, & Lewis, 1997), it does not reach the range of human behavioral expression. Nevertheless, one may ask how is it possible that such unrelated species exhibit similar forms of aggression. What do flies, rodents, and humans have in common besides possessing pairs of limbs and being able to display a similar motor-expressive pattern? At the functional level, fruit flies and rats do not fight for money or pride, but both can fight for sexual access and territoriality, as humans sometimes do. At the neurobiological level, the divergence is con-
Subtypes of Aggression
27
siderable. The rudimentary nervous system of a fruit fly is made up of a mere 250,000 neurons, and the brain of an adult rat weighs more or less 2 g. With its 100 billion neurons, the 1.4-kg human brain is at a different level of complexity. Nevertheless, rats and humans share similar brain structures and pathways, and, like fruit flies, they can develop addiction to cocaine, nicotine, and alcohol (Bainton, Tsai, Singh, Moore, Neckameyer, & Heberlein, 2000), suggesting a certain degree of neurobiological similarity. Cross-species comparisons call for differentiating behavioral and neurophysiological characteristics that are homologous (i.e., having a common evolutionary origin) from those that are simply analogous (i.e., likeness in function but not in evolutionary origin). When analyzing behavioral and/or physiological similarity between species, we run the risk of anthropomorphism, that is, viewing animals as having human-like qualities, or zoomorphism, that is, mistakenly perceiving human behavior as the mirror image of animal behavior (Cairns, 1979). For instance, contrary to the controversial and sensationalist opinion of authors in the past (e.g., Ardrey, 1966), predatory behavior in animals should not be equated with hunting or warfare in humans. Although these behaviors may share some common neuroevolutionary processes, the function of predatory behavior is to supply food, thus enhancing the probability of survival. Human hunting, at least in modern society, is often motivated only by the pleasure that the activity provides, and not by the necessity to gain food. Predation targets other species, whereas warfare is directed at other humans. Likewise, matching children’s and adults’ behavior on the basis of their expressive similarity, something one could adventurously call “adultomorphism” or “pedomorphism,” depending on the direction of the comparison, is not uncommon. A child hitting another child in kindergarten and an adult hitting a colleague at work are events that—albeit similar in their expression and to some extent in their immediate consequences (i.e., physical hurt)—clearly differ in terms of antecedents and long-term consequences. In sum, any definition of aggression that does not address the issue of phylogenetic and ontogenetic similarity and contrast encourages a static or reified view of a complex phenomenon.
Early Taxonomies of Aggression Animal models of aggression provide a strong conceptual base for approaching the study of human aggression. Moyer (1968) established seven categories of aggressive behavior based on their functional values and stimulus-bound characteristics: predatory aggression, intermale aggression, fear-induced aggression, irritable aggression, territorial defense, maternal aggression, and instrumental aggression. A possible eighth category, sex-related aggression, was also suggested. As Moyer himself mentioned five years later, however, “definition of the kinds of aggression on that basis alone now appears too restrictive: the kinds of aggression vary on a number of different dimensions,
28
THE DEVELOPMENT OF AGGRESSION
and all of them must be considered in the definition of each kind” (Moyer, 1973, p. 12). The dimensions that Moyer referred to were the specificity of the eliciting stimulus, the presence or absence of an emotional display, the sex of the attacker, and the neurophysiological correlates. With these four “dimensions” in mind, Moyer (1973) finally rejected territorial defense as a subtype of aggression. Although Moyer’s taxonomy was primarily targeting animal aggression, reference was also made to human behavior; however, the uniqueness and complexity of human aggression was not specifically addressed. The use of the categories to describe and differentiate human aggression undoubtedly leans toward zoomorphism. A similar comment can be made regarding Wilson’s sociobiological taxonomy (Wilson, 1980). Although most of the subtypes of aggression (i.e., territorial, dominance, sexual, parental disciplinary, weaning, moralistic, predatory, and antipredatory aggression) were meant to be theoretically applicable to both animals and humans, there are obvious exceptions. Weaning aggression, for example—that is, occurring when parents “gently attack” their offspring to make them stop begging for food—is clearly more relevant to animal behavior. The opposite can be said of moralistic aggression, which for Wilson (1980) represents advanced forms of reciprocal altruism that reduce the manifestation of aggression via rules, codes of punishment, or enforced conformity. In conjunction with the risk of zoomorphism and/or anthropomorphism, a major problem with Moyer’s and Wilson’s taxonomies is that they are based on an eclectic set of criteria. These involve a specific target (e.g., prey, predator, another male), a specific context (e.g., territorial, maternal, weaning), a specific function (e.g., to dominate, to obtain sexual access) or the concomitant occurrence of an emotional state (e.g., fear, anger), or a learning process (e.g., instrumental, moralistic). As argued by Archer and Browne (1989), one should try to be more explicit, and perhaps more restrictive, in terms of which criteria are used to divide aggression into distinct classes. Archer (1988) proposed a simpler classification on the basis of function by distinguishing forms of aggression that are “competitive,” “protective,” and “parental.” Parental aggression can be viewed as an extension of protective aggression, involving the protection of those recognized as genetically related and dependent. Predation was omitted from this scheme because, as argued in an earlier article (Archer, 1976), Moyer (1968) actually presented good evidence for the neural and motivational separation of the two forms of behavior. The argument was that the motivational systems underlying aggression and predation had evolved independently of one another; a good example is provided by cats, who show very different types of behavior and emotional states when fighting one another and when stalking prey. However, it is undoubtedly the case that the systems controlling predation and aggression are more related in some cases, for example, in the case of mouse killing by rats (e.g., Karli, 1956), in cannibalism in chimpanzees (Bygott, 1972), and more widely in the case of male infanticide. In humans, the link between hunting and aggression may be more complex.
Subtypes of Aggression
29
At about the same time that Moyer proposed his taxonomy and suggested four dimensional criteria to discriminate subtypes of aggression in animals, similar attempts were made to categorize the different forms of aggression in humans. Buss (1961) suggested that aggression be dichotomized along three behavioral dimensions: physical–verbal, active–passive, and direct–indirect. In 1969, Pulkkinen formulated a bidimensional (and bicriteria) model characterizing human aggression (see Pulkkinen, 1987). A first dimension was related to the expression of aggression, ranging from indirect forms to more direct ones. A second dimension concerned the defensive–offensive dichotomy, which was determined by the presence or absence of proximal threatening stimuli. Although these models have the merit of being parsimonious, they are not truly pertinent to animal aggression, as indirect forms of aggression are exclusive to humans and, possibly, nonhuman primates (Campbell, 1999). In addition, a model depicting behavioral or contextual features in terms of a continuum (poles or axes) may not be appropriate for some distinctions, because it postulates quantitative differences between the different forms of aggression. An aggressive act can fluctuate in intensity, but it cannot be half direct and half indirect.
A MULTILEVEL ANALYSIS OF AGGRESSION IN ANIMALS AND HUMANS Aggression is a dynamic, multifaceted social-emotional process that calls for a careful examination of its various antecedents, expressions, and consequences (or functions). The adoption of a dynamic and multifactorial perspective when investigating the foundations of aggression and other types of social behavior in humans and animals was ardently advocated by Cairns (1979). Cairns stressed the importance of using adequate criteria in order to generalize from one phylogenetic level to another and from one developmental stage to another. Observing a similar feature in two or more species (e.g., boxing) or in a child and an adult (e.g., hitting) is not enough to conclude that these behaviors serve the same function and derive from the same proximal or distal antecedents. This principle has a significant impact on how aggression is ultimately defined and how the different forms of aggression can be grouped. Figure 2.1 summarizes the major domains or themes of investigation in aggression research. This is an extended representation of Cairns’s model of what he called “polythetic analysis,” that is, the examination of aggression from multiple conceptual angles. The first three domains encompass the proximal (eliciting context and neurophysiological mechanisms) and the more distal antecedents of aggression, which are perceptible only through a longitudinal, developmental analysis. The fourth domain relates to the expression of aggression, whereas the fifth and the sixth cover the consequences of aggression to others (harm) and to self and society (social/emotional function), respectively. One must recognize that there are important differences between, as well
30
THE DEVELOPMENT OF AGGRESSION
FIGURE 2.1. Polythetic analysis of aggression. Based on Cairns (1979).
as within, species in terms of variables that influence the development and expression of aggression. Nevertheless, this thematic framework probably covers most subtypes of aggression found in the scientific literature and can be generally applied to both animal and human research. Obviously, dichotomizing aggression as direct versus indirect or as producing physical versus mental harm may be more appropriate and applicable to human behavior than to rat or fruit fly aggression. It is also important to note that even if most subtypes of aggression have been established on the basis of a single criterion, it is always possible to relate them to other criteria. Therefore, overlap between criteria is clearly the rule rather than the exception. For example, parental aggression in animals serves an obvious long-term function (i.e., survival of offspring), but it can also be described in terms of antecedents, either proximal (e.g., perception of a prey, underlying neurophysiological mechanism) or more distal (e.g., genetic predisposition). Common sense would dictate that we follow the temporal sequence of aggression, beginning with its antecedents, then considering its expression, and finishing with its consequences for others and the self. We decided to go the opposite way, starting with the first and most obvious outcome of aggression (its consequences for others and self) and then considering elements that necessitate a more thorough examination of the problem (e.g., distal antecedents and developmental issues). This upward presentation of the different components of aggression ends with a provisional integrative representation of the most significant subtypes of aggression and their underlying processes.
Subtypes of Aggression
31
SUBTYPING AGGRESSION ON THE BASIS OF ITS CONSEQUENCES ON OTHERS AND SELF Based on Harm and Injury Harm or injury to others is the foremost indicator that an aggressive act has occurred. This is perhaps why it has been a common criterion for qualifying aggression (Berkowitz, 1981; Rule, 1974). In its most simplified form, aggression has been defined as the delivery of noxious or painful stimuli to another individual (Buss, 1961) or as a response resulting in injury (Rule, 1974) or as “damage or destruction of some goal entity” (Moyer, 1968). Some authors specified that the targeted individual ought to “be motivated to avoid such treatment,” in order to exclude cases of sadomasochism (Baron, 1977). Research has traditionally focused on physical harm, but with increasing frequency, has been including mental or psychological harm as a potential consequence of aggression, as in the case of indirect aggression or persistent bullying. A number of issues have been raised concerning the validity of harm as a criterion for assessing or qualifying aggression. First, as potential harm, either physical or psychological, seems inherent to all forms of aggression, it has little discriminatory value. In addition, assessment of harm is not only contingent upon the nature of the aggressive action (a hit by a 2-year-old may not hurt an adult), but it is also conditional on the victim’s sensibility. Some individuals may be hypersensitive to external stimulation and may be more prone to feel physically and/or mentally harmed. Assessment of harm is indeed a highly subjective matter. Even from a behaviorist perspective that targets the observable behavior of the aggressor (delivery of noxious stimuli), a judgment must be made to determine whether pain or harm has been inflicted or not. Furthermore, the harm criterion has little value for subtyping aggression in animals. Automated bite-recording devices to measure the intensity of noxious delivery by aggressive primates and rodents have been designed (Ulrich, Dulaney, Arnett, & Mueller, 1973) but the devices turned out to be too impractical and restrictive (Knutson, 1973).
Based on Intent, Motivation Descriptive models of aggression have often emphasized the cognitive/motivational antecedent of aggression. An influential paper by Feshbach (1964) was critical in establishing two major types of aggression based on harm, or, more specifically, on the motivation to harm. If injury (to a person or object) was the primary goal of the action (pleasure or satisfaction following injury being the main reward), it was labeled hostile aggression. If injury was not the main purpose and the action was executed for reward other than the pleasure of injuring, then it was termed instrumental aggression. More specifically, Feshbach (1964) defined instrumental aggression as any act that produces harm and that “is directed toward the achievement of nonaggressive goals.” These
32
THE DEVELOPMENT OF AGGRESSION
“nonaggressive” goals include getting attention and acquiring an object or a resource. Although differentiating aggression based on the presence of hostile intention has been a dominant and relatively valid conceptual dichotomy in aggression research (Atkins, Stoff, Osborne, & Brown, 1993), it raises theoretical problems. First, aggression can be both hostile and instrumental (Bushman & Anderson, 2001; Hartup & De Wit, 1974). Second, labeling aggression as hostile is rather tautological. Are there nonhostile, friendly forms of aggression? This semantic glitch should not hinder the fact that some people may gain a considerable satisfaction or pleasure from hitting and injuring someone, whereas others may benefit more from the social or materialistic consequences of their aggressive action. It is important to note that getting pleasure from producing harm to someone may not be the most common motivational antecedent of aggression. Indeed, for many explanatory models (e.g., Archer, 1976; Berkowitz, 1993; Dollard, Doob, Miller, Mowrer, & Sears, 1939), bringing an end to a situation that is annoying and removing an irritant or a discrepancy from what is expected are frequent motivational precursors of aggression. Nonetheless, both pleasure-motivated aggression and relief-motivated aggression imply an emotional outcome or reward that is either pleasurable (positive reinforcement) or a release from a previously noxious state (negative reinforcement). Whether we call Feshbach’s categories of aggression hostile versus instrumental, annoyance-motivated versus incentive-motivated (Zillman, 1978), or emotionally rewarding versus materialistically/socially rewarding, does not really matter if an unambiguous definition is provided. Researchers with a social-interactionist perspective (Tedeschi & Felson, 1994) view aggression (which they term “coercive power”) as being motivated by interpersonal goals—to control others, to maintain justice, or to defend social identity. From the perspective of the actor, all aggression is therefore instrumental in that it pursues one of these social goals. However, these situations can just as readily be viewed as removing an irritant (i.e., negative reinforcement), and it may be confusing to regard such situations—which inevitably involve anger—as instrumental in nature. For the purposes of clarity, it is probably better to restrict the term “instrumental” to cases involving positive reinforcement of aggression by a reward unconnected with activation of the aggression system. This would highlight the parallels between such human cases as robbery and rape (Felson, 2002) and the use of conventional reinforcers such as food or water to facilitate animal aggression (e.g., Ulrich, Johnston, Richardson, & Wolff, 1963). Feshbach (1971) and Rule (1974) proposed a different taxonomy by subdividing instrumental aggression into personally motivated aggression and socially motivated aggression. Personally motivated aggression is different from hostile aggression in that the primary goal is not to hurt but to obtain reward via inflicting injury. Yet aggression is said to be socially motivated when the primary goal is to gain a social advantage. Even in the authors’ views, however, the distinction between these subtypes of aggression was problematic,
Subtypes of Aggression
33
because of the difficulty in assessing motivation, intent, or feeling. Sears (1961) suggested the term prosocial aggression to distinguish aggressive behavior considered as socially and morally acceptable. However, to the same extent that the term hostile aggression is certainly a truism, the expression “prosocial aggression” appears to be an oxymoron. Naming an aggressive act as prosocial or antisocial depends too much on individual viewpoints to be a useful categorization criterion. Physically punishing a child to prevent the occurrence of a life-threatening situation (e.g., crossing a high-traffic road) may have a justifiable intention and long-term benefits. From the child’s perspective, however, the immediate consequences of being spanked by a parent or being slapped by a peer may not differ much. Moreover, any act of terrorism or war can be considered legitimate and morally acceptable from the aggressor’s perspective. From the victim’s point of view, the immediate and harmful consequences do not depend on how good the intention was. It has been claimed that behavior should be at once intentional (motivation to injure) and harmful to be classified as being aggressive (Berkowitz, 1993; Dollard et al., 1939). As with the nature and intensity of harm, however, intent is a concept that is difficult to prove and easy to deny (Loeber & Hay, 1997). It also imposes serious limitations in investigating aggression in children or animals (Tremblay, 2000). For these reasons, although intent is certainly central to the definition of aggression, it may not be an adequate criterion in practical terms for establishing a taxonomy of aggression applicable to both animals and humans (children and adults). Intention may be also problematic in cases of indirect aggression where the perpetrator seeks to cover up his or her actions.
SUBTYPING AGGRESSION ON THE BASIS OF ITS EXPRESSION Categorizing aggression on the basis of its behavioral expression has been a popular criterion as it circumvents the pitfalls of intrinsic notions such as instinct, motivation, drive, or intent to harm. Unfortunately, too many behavioral responses that have little in common have been subsumed under the label of aggression. Display of aggressive behavior is species-typical and agedependent, as it is manifestly constrained by morphogenetic, neurophysiological, and maturational factors unique to each species and developmental stage. Topographical descriptions of aggressive interactions in animals have been performed systematically (Barnett & Marples, 1981; Grant & Mackintosh, 1963; see also Pellis, Pellis, & Foroud, Chapter 3, this volume). Although attempts have been made to extend the ethological approach in children (Blurton-Jones, 1967; McGrew, 1972) much less has been achieved regarding the operationalization of specific aggressive behavior in humans (Knutson, 1973; Tremblay, 2000). Buss (1961) was among the first to offer a behavioral taxonomy of human aggression by dichotomizing it along the physical–verbal, active–passive, and direct–indirect dimensions. The active–passive dimension is
34
THE DEVELOPMENT OF AGGRESSION
debatable, however, because it is difficult to determine whether behavioral inactivity (e.g., not helping someone in need) was intended or not. The physical– verbal and the direct–indirect dimensions are more defendable. They are more useful when applied to humans, although animals often begin aggressive encounters with threat displays, including vocalizations (Archer, 1988). In animals, direct aggression is straightforward, involving bodily contact such as biting, hitting, or pushing. Some behavioral elements that do not involve any physical contact, such as threat and thrust (Grant & Mackintosh, 1963), are clearly direct and aggressive, as their objective is to intimidate and they are often precursors of genuine attack. In humans, Blurton-Jones (1972) also labeled as aggressive a number of nonphysical behaviors that were temporally related to other more obvious aggressive actions. In humans, however, the expression of aggression is far more complex and includes a broader assortment of direct and indirect actions. Direct forms of aggression comprise physical assault and a range of verbal behaviors that may be hostile in content or in tone. These verbal behaviors may occur with or without physical attack. In humans, direct aggression is somehow equivalent to overt aggression (Crick, 1996). Indirect aggression is at the other end of the behavioral dichotomy, consisting of actions aimed at harming others that involve some kind of social intermediary between the aggressor and the victim. The course of action is often delayed and subtle, as there is no physical contact. The psychological effects, however, may persist over a long period. This can be achieved through gossiping and ostracizing, which ultimately may damage peer relationships or social status, or through actions aiming at reducing accessibility to resources. Research on indirect aggression has grown substantially in the last few years, as this may be the most prevalent form of aggression during adolescence and adulthood (Björkqvist et al., 1992). It is the only form of aggression that may be more prevalent in females than in males (Campbell, 1999). Indirect aggression overlaps with two further categories, relational (Crick, Grotpeter, & Bigbee, 2002) and social (Galen & Underwood, 1997) aggression. Relational aggression emphasizes damage to relationships as a way of harming someone. As this can include face-to-face statements such as “I’m not friends with you anymore,” relational aggression is therefore not always direct (Coyne, Archer, & Eslea, 2004). Social aggression is a wider category that includes both relational and indirect forms, along with expressions such as negative facial expressions or body movements.
SUBTYPING AGGRESSION ON THE BASIS OF ITS ANTECEDENTS Proximal Contextual and Emotional Elicitors Proximal antecedents, either contextual or emotional, are the most frequently used criteria for discriminating subtypes of aggression, particularly in animals. A wide range of social and nonsocial contextual/situational variables have
Subtypes of Aggression
35
been used to elicit aggressive behavior in animals (Archer, 1976, 1988). Social elicitors include the presence of a prey or exposure to a same-species, samesex conspecific either in the home cage (resident–intruder paradigm) or in a novel environment. These social encounters are often combined with prior or concomitant experimental manipulations (e.g., social isolation, food deprivation, electric shock) that alter the neurobehavioral state of the potential aggressor. These predisposing factors, which influence the potency of the elicitors (Hinde, 1974), represent an important issue in aggression research. A decisive factor is the presence or absence of a threatening elicitor. When a situation is clearly threatening or perceived as threatening, an arsenal of behavioral responses is put in motion to protect the individual from pain, injury, and, possibly, death. Responses to threat range from harmless reactions, such as heightened immobility (i.e., freezing, crouching), escape, and holding off the threatening assailant with stretching of the upper limbs (boxing), to fierce attack and delivery of noxious stimuli. This constellation of responses has been traditionally subsumed under the label “agonistic behavior” (Scott & Fredericson, 1951). The notion of aggression as a response to potential threat to an animal’s welfare is central to theories that conceive of aggression as an aversion, whereby the animal is motivated to avoid a situation. This contrasts with an appetite, whereby the animal is motivated to approach or seek out a situation (Craig, 1928). Aversion-based theories seek to identify the mechanism through which animals or humans detect threats, and they include the frustration– aggression hypothesis (Dollard et al., 1939), the discrepancy model of animal aggression (Archer, 1976), and the theory of threatened egotism (Baumeister, Smart, & Boden, 1996). Pain is another “emotional” antecedent of aggressive behavior. Animals exposed to painful stimulation (e.g., electrical shocks, sudden heat, bite) will start fighting against each other or will attack an inanimate object if tested alone (Archer, 1989/1990; Ulrich & Azrin, 1962). Blanchard and Blanchard (1981) have convincingly argued that such pain-induced aggression is analogous to fear-induced defensive aggression. Archer (1989/1990) suggested that, although this is the case in laboratory setups involving repeated painful shocks, an isolated single painful stimulus may under natural conditions evoke anger-induced offensive aggression. Overall, aggressive behavior can be generated by a variety of external stimuli that are threatening or painful, and these stimuli activate, to different degrees, both the peripheral and the central nervous systems. In pain-induced aggression, activation of the nociceptors and/or thermal receptors sets in motion neural pathways that control the expression of aggressive behavior, and, in this case, the threshold for fear is easily reached. Then, determining whether aggression is mediated by fear or anger will depend on which brain regions are activated during attack. Under threatening circumstances, aggression may be offensive or defensive (Blanchard & Blanchard, 1989), that is, anger or fear motivated. In humans, responses to provocation broadly similar to threats in animals have
36
THE DEVELOPMENT OF AGGRESSION
been termed reactive aggression. Thus, reactive aggression is an impulsive, negatively valenced act displayed in response to a threat or provocation (Dodge, & Coie, 1987; Vitaro, Gendreau, Tremblay, & Oligny, 1998). It stems from the frustration–aggression hypothesis that viewed aggression as a “primordial reaction . . . whenever pleasure-seeking or pain-avoiding behavior is blocked” (Dollard et al., 1939, p. 21). Interestingly, it is not so much contextual stimuli per se that determine whether aggression will be labeled as reactive, but the perception that the individual has and what he or she makes of it. Indeed, an interesting finding in human research is that reactively aggressive children more readily perceive an ambiguous situation as threatening or hostile (Dodge & Coie, 1987; Vitaro, Brendgen, & Tremblay, 2002). This attributional/perceptual bias and behavioral hypersensitivity to what typically should be mild, nonthreatening stimulation is reminiscent of what is observed in animals after prolonged social isolation (Gendreau, Gariépy, Petitto, & Lewis, 1998) or after successive defeat experiences (Keeney & Hogg, 1999). These animal paradigms may be more relevant to human reactive aggression than paradigms using clear life-threatening conditions. Conversely, proactive aggression occurs with more forethought and does not seem to be associated with any apparent proximal elicitor. It is more controlled, more premeditated, and less emotionally reactive. It can be understood in terms of social learning, as a result of previous external reinforcement (Bandura, 1973; Patterson & Cobb, 1973). Discriminant and convergent validity of these subtypes of aggression has been provided by an increasing number of empirical reports (Dodge & Coie, 1987; Poulin & Boivin, 2000; Pulkkinen, 1996; Vitaro et al., 1998, 2002). Contrary to the dichotomy hostile–instrumental, which is determined by the nature of the motivation (goal is to hurt vs. goal is to gain), reactive and proactive forms of aggression are differentiated by the presence of a prior provocative event and the short delay before enactment. The distinction between reactive and proactive aggression, however, is problematic, as it depends on no provocation being identified, and what constitutes a provocation varies considerably between individuals. Looking at someone the wrong way may be classed as a provocation in a subculture where everyone is on the lookout for signs of disrespect. Having said this, the category proactive aggression (i.e., aggression without obvious provocation or threat) may alert us to cases, primarily among young men, where fights are sought for no apparent reason other than to inflict a physical defeat on another person.
Brain Mechanisms The use of psychic or inner states to qualify aggression is not necessary when examining the direct and observable effects of lesions and stimulations of specific areas of the brain. The brains of humans and other mammals share many structures, pathways, and neurochemical properties, reflecting the long and progressive transformation of the brain during evolution. In the last four de-
Subtypes of Aggression
37
cades an increasing body of evidence has accumulated describing the role of specific neural circuits in the expression of different types of aggression. Numerous sites within the neocortex and subcortical structures have been identified as central in mediating aggressive behavior. Moyer (1968, 1973) was among the first to relate specific neural circuitries and distinctive neuroendocrinological status to subtypes of aggression. Particular attention was given to the amygdaloid complex, the various hypothalamic nuclei, and other parts of the so-called limbic system (e.g., septum, cingulate cortex). Moyer (1973) argued that each of the subtypes of aggression that he had previously identified (except for instrumental aggression) had its own set of neural circuits. Functions associated with the expression of specific forms of aggressive behavior have indeed been localized in the brain (Valzelli, 1981; Panksepp, 1998). As Panksepp pointed out, however, there are more subtypes of aggression based on the proximal contextual antecedents than there are subtypes based on neural processes. In other words, similar brain circuitry may allow different types of aggression to be expressed; what differs are the eliciting circumstances and the perceptual processing. As we will see, this has a significant impact on our attempt to produce a parsimonious taxonomy of aggression. Panksepp (1998) denoted distinct neural circuitry for no more than three subtypes of aggression in the brains of rats and cats: predatory aggression, affective or rage-like aggression (which may coincide with defensive aggression identified by Blanchard & Blanchard, 1989); and intermale aggression (what the Blanchards called offensive aggression). So-called predatory aggression (or quiet-biting attack in the laboratory) can be generated by stimulation of the dorsolateral hypothalamus and involves activation of the ventral part of the periaqueductal gray. It is believed to be primarily mediated by what Panksepp (1998) termed the seeking or the appetitive motivational system of the brain. If so, this contrasts with the emphasis on aversive motivation in most theories of aggression (see above). This system involves not only the lateral hypothalamus but also most structures that are innervated by neurons releasing dopamine and serotonin. Abnormal development of the prefrontal cortex has been associated with antisocial personality disorder (Raine, Lencz, Bihrle, LaCasse, & Colletti, 2000), and people diagnosed with this disorder have been shown to display more proactive forms of aggression (Chase, O’Leary, & Heyman, 2001). This would be consistent with their seeking out situations in which to display aggressive actions. In contrast, the affective subtype of aggression is mediated by different and more primitive (Archer, 1988) interactive neural circuits, what Panksepp (1998) termed the fear system and the rage system. Affective aggression involves the ventrolateral–medial hypothalamus, several amygdaloid nuclei, and the dorsal part of the periaqueductal gray, to cite just a few (Gregg & Siegel, 2001). It is important to note that this subcortical circuitry operates relatively independently of neocortical input. The neurophysiological basis of intermale (or intrasexual) aggression has not been well established. Panksepp (1998) mentioned that the brain circuitry of intermale aggression may interact with
38
THE DEVELOPMENT OF AGGRESSION
both the seeking and rage systems, but is somehow relatively independent. The evidence for this is clearly not overwhelming, but Panksepp (1998) reported that brain lesions that impair predatory and affective aggression do not affect intermale aggression. Obviously, this is an extremely simplified account of very complex brain mechanisms involving many more structures and neurotransmitter systems (see Gregg & Siegel, 2001). Although primitive forms of aggression, such as those elicited by stimuli perceived as painful, threatening or annoying, may be processed through brain circuitry highly similar to that of other species, the sophistication and diversification of aggression in humans necessarily implies the existence of brain processes unique to our species. Our neocortex, which does not complete its maturation before adulthood (Giedd et al., 1999), plays an important inhibitory role in more primitive brain structures, therefore controlling the expression of social/emotional reactivity. The gradual transformation of the brain during development may explain the progressive shift in the expression of aggression from early childhood to adulthood, that is, direct, reactive, and physical forms of aggressive behavior being gradually substituted by more indirect, controlled, and nonphysical forms (see Vaillancourt, Chapter 8, this volume).
Developmental Issues What differentiates humans from animals is not only the increased complexity of the human brain, the advanced cognitive skills and social world, but the longer developmental period that is necessary for organizing behavior (Cairns, 1979; Hinde, 1974). As mentioned in the preceding section, the expression of aggression changes throughout ontogeny, with physical aggression reaching its peak in early childhood (Tremblay et al., 1999) and relationally oriented forms of aggression gradually emerging in late childhood/early adolescence (Björkqvist et al., 1992). Hartup (1974) observed that aggression develops from being first object-oriented (instrumental) to being more person-oriented (hostile) aggression (see also Caplan, Vespo, Pedersen, & Hay, 1991). Hinde (1992) proposed one of the few categorizations of aggression exclusive to childhood. He discriminated instrumental (or specific) aggression (aiming at gaining or retrieving an object or situation), teasing aggression (aggression unrelated to acquiring a specific object or situation), defensive aggression (in response to an attack), and game aggression (that results from rough-and-tumble play). Hinde suggested that childhood instrumental and defensive aggression were somehow equivalent to reactive–impulsive aggression as observed in adulthood. Teasing aggression had its adult counterpart under the label of “spontaneous aggression.” No further distinction between impulsive and spontaneous aggression and no adult correspondence for game aggression were provided. A social–developmental analysis of the functional and expressive characteristics of aggression from early childhood to adulthood is indispensable for
Subtypes of Aggression
39
our understanding of the different ontogenetic pathways to more sophisticated forms of aggression. Hinde’s distinction between subtypes of aggression in childhood was clearly a step toward establishing a developmental topology of aggression. One may question, however, the need to use different taxonomical labels for children and adult aggression, especially when there is correspondence between the different subtypes. In addition, characterizing one form of aggression as impulsive and another as spontaneous is confusing. Finally, game aggression (or rough-and-tumble play) is clearly motivationally distinct from other forms of aggression (Blurton-Jones, 1972; Smith, 1974). In the end, we need to agree on a general taxonomy that is applicable throughout development, not only to a specific period. Once this is achieved, the developmental trajectory of the different subtypes of aggression can be investigated. The onset and stability of aggression during ontogeny has been an important issue in developmental and clinical sciences. Moffitt (1993) proposed a developmental taxonomy of conduct disorders based on the onset and persistence of antisocial behavior (which relates to aggression). Specifically, conduct disorders could be described as life-course persistent (behavioral problems start during childhood and persist throughout childhood up to adolescence and adulthood) or as adolescence-limited (when problems emerge and end during adolescence) Discriminating between early- versus late-onset conduct disorders has been a valuable clinical classification (American Psychiatric Association, 1994). The relationship between individual development and the transformation of aggression over time in boys and girls, in both expression and function, is central to the science of aggression. We now recognize the gender-dependent development and use of physical aggression (Tremblay et al., 1999) and the gradual emergence of hostile (Caplan et al., 1991; Hartup, 1974) and indirect (Björkqvist et al., 1992) aggression. Similarly, reactive aggression seems to appear first (as tantrums) or may be a more prevalent form of aggression early on (Loeber & Hay, 1997). However, children prone to display proactive forms of aggression may be more at risk of exhibiting delinquent activities in adolescence (Vitaro et al., 2002). Finally, recent evidence suggests that reactive and proactive aggression in boys and girls may have different etiological antecedents (Connor, Steingard, Anderson, & Melloni, 2003). Many important issues remain to be investigated regarding the onset and developmental trajectory of the different forms of aggression across gender and sociocultural background.
AN INTEGRATIVE MODEL OF AGGRESSION The goal of this chapter is to provide a summary of conceptual and theoretical issues related to the categorization of different subtypes of aggression in humans and animals. Finding an appropriate taxonomy seems to be a recurrent problem in aggression research. A taxonomy is, by definition, a classification based on similarities of a specified characteristic. Suggesting a taxonomy for
40
THE DEVELOPMENT OF AGGRESSION
subtypes of aggression based on a single criterion or for a single species is one thing, but conceiving one that encompasses the multiple ontogenetic and phylogenetic aspects of this phenomenon is more challenging. Aggression is not a behavior, or a set of genes, or a brain mechanism. It is a complex social phenomenon that conveys so many meanings, takes so many forms, results from so many proximal and distal antecedents, and has so many consequences that a universal taxonomy seems an unachievable puzzle. It has identifiable underlying brain processes, but those are dependent upon context, emotional state, and previous learning experience (reinforcement). It is important to note that aggression is a phenomenon in constant transformation, both phylogenetically and ontogenetically. Hinde (1974) summarized well the challenge of elaborating a taxonomy for subtypes of aggression in both humans and animals: “The range validity of any generalizations we make is inversely related to their precision. As more diverse phenomena are included within our category of aggressive behavior, our generalizations inevitably become less precise” (p. 4). The primary problem with respect to categorizing aggression into different subtypes is to select a parsimonious set of criteria that integrate most forms of aggression. In this regard, some criteria, despite being relevant to our better understanding of aggression, do not appear to be essential for establishing a categorization. Harm to others, for instance, either physically or psychologically induced, may be the first, more expeditious consequence of aggression, but it has little taxonomic value. Therefore, we settled on a simplified taxonomic model that encompasses the most significant and discriminative features of aggression (Figure 2.2). These include the antecedent (to determine whether aggression is proactive or reactive), the expression (to determine whether aggression is direct or indirect), and the function (to determine whether aggression was produced only to hurt or to benefit socially). Other important elements of aggression, such as distal antecedents (past experience, biological or genetic predisposition), aggressor’s emotional–perceptual bias, basic neurobiological mechanisms, and learning, were also integrated. The first step in identifying the form of aggression should be to determine whether there is a proximal contextual elicitor. If there is no apparent proximal antecedent, then aggression is proactive. This form of aggression is primarily controlled by the neocortex, more particularly the prefrontal cortex and the descending corticolimbic pathways. When a proximal antecedent can be identified, then aggression is said to be reactive. This behavioral output involves various motivational/emotional states (e.g., fear, pain, anger, or annoyance) that are under the control of subcortical, primarily limbic, structures. In reactive aggression, perception of the eliciting stimulus is the key for judging the pathological nature of the response. Hypersensitivity to certain contexts and stimulations may indicate heightened anxiety or heightened fearfulness (paranoia). Both proactive and reactive aggression can be expressed directly or indirectly, and they can target either the source of stimulation or a substitute person/object (displaced aggression). Then the aggressive action may pro-
Subtypes of Aggression
41
FIGURE 2.2. A simplified taxonomical model of aggression based on context, neuroaffective processes, consequences on self, and learning.
vide a pleasurable reward from harming (hostile aggression), a social/materialistic reward (instrumental aggression), or both. Following these consequences, learning (reinforcement) will take place and certainly influence individual development.
CONCLUDING REMARKS Aggression is always a timely topic. The news media provide an incessant flow of information on the shocking nature of human behavior. We were just finishing this chapter when we heard the news that three teenage girls were facing charges of attempted murder for allegedly trying to poison a classmate by adding copper sulfate, a highly toxic substance, to her drink. What aggressive category does poisoning fit into? It undoubtedly relates to physical harm, but without a clear, direct physical action. It is a physical action but with an indirect, concealed behavioral twist. Although it is a planned, proactive-like aggressive action, it is likely a retaliatory response to some prior event or situation. This is a good example of the difficulty in fitting every act of aggression into a definite category or model. Do we need more research to produce a better taxonomy? Reaching a consensus on this issue may not be as necessary as continuing our thorough analysis of the wide range of proximal and distal
42
THE DEVELOPMENT OF AGGRESSION
forces that influence the development and expression of aggression (Cairns, 1979). It is important to refine our knowledge of the many genetic, biological, and social factors that induce, facilitate, or maintain aggression throughout development. Eventually, a universally agreed-upon taxonomy, at once parsimonious and comprehensive and integrating both animal and human aggression, may emerge.
REFERENCES American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Archer, J. (1976). The organization of aggression and fear in vertebrates. In P. P. G. Bateson & P. Klopfer (Eds.), Perspectives in ethology (Vol. 2, pp. 231–298). New York: Plenum Press. Archer, J. (1988). The behavioural biology of aggression. Cambridge, UK: Cambridge University Press. Archer, J. (1989/1990). Pain-induced aggression: An ethological perspective. Current Psychology: Research and Reviews, 8, 298–306. Archer, J., & Browne, K. (1989). Concepts and approaches to the study of aggression. In J. Archer & K. Browne (Eds.), Human aggression: Naturalistic approaches (pp. 3– 24). London: Routledge. Ardrey, R. (1966). The territorial imperative. New York: Atheneum. Atkins, M. S., Stoff, D. M., Osborne, M. L., & Brown, K. (1993). Distinguishing instrumental and hostile aggression: Does it make a difference? Journal of Abnormal Child Psychology, 21, 355–365. Bainton, R. J., Tsai L. T., Singh, C. M., Moore, M. S., Neckameyer, W. S., & Heberlein, U. (2000). Dopamine modulates acute responses to cocaine, nicotine and ethanol in Drosophila. Current Biology, 10, 187–194. Bandura, A. (1973). Social learning theory and aggression. In J. F. Knutson (Ed.), The control of aggression: Implications from basic research (pp. 201–250). Chicago: Aldine. Barnett, S. A., & Marples, T. G. (1981). The “threat posture” of wild rats: A social signal or an anthropomorphic assumption? In P. F. Brain & D. Benton (Eds.), Multidisciplinary approaches to aggression research (pp. 39–52). Amsterdam: Elsevier/North Holland Biomedical Press. Baron, R. A. (1977). Human aggression. New York: Plenum Press. Baumeister, R. F., Smart, L., & Boden, J. M. (1996). Relation of threatened egotism to violence and aggression: The dark side of high self-esteem. Psychological Review, 103, 5–33. Berkowitz, L. (1981). The concept of aggression. In P. F. Brain & D. Benton (Eds.), Multidisciplinary approaches to aggression research (pp. 3–15). Amsterdam: Elsevier/ North Holland Biomedical Press. Berkowitz, L. (1993). Aggression: Its causes, consequences and control. New York: McGraw-Hill. Björkqvist, K., Lagerspetz, K. M. J., & Kaukiainen, A. (1992). Do girls manipulate and boys fight?: Developmental trends in regard to direct and indirect aggression. Aggressive Behavior, 18, 117–127.
Subtypes of Aggression
43
Blanchard, D. C., & Blanchard, R. J. (1989). Experimental animal models of aggression: What do they say about human behavior? In J. Archer & K. Browne (Eds.), Human aggression: Naturalistic approaches (pp. 94–121). London: Routledge. Blanchard, R. J., & Blanchard, D. C. (1981). The organization and modelling of animal aggression. In P. F. Brain & D. Benton (Eds.), The biology of aggression (pp. 529– 562). Rockville, MD: Sijhoff & Noordhoff. Blurton-Jones, N. G. (1967). An ethological study of some aspects of social behaviour of children in nursery school. In D. Morris (Ed.), Primate ethology (pp. 347–368). Chicago: Aldine. Blurton-Jones, N. G. (1972). Categories of child–child interaction. In N.G. Blurton Jones (Ed.), Ethological studies of child behaviour (pp. 97–127). London: Cambridge University Press. Bushman, B. J., & Anderson, C. A. (2001). Is it time to pull the plug on the hostile vs. instrumental aggression dichotomy? Psychological Review, 108, 273–279. Buss, A. H. (1961). The psychology of aggression. New York: Wiley. Bygott, J. D. (1972). Cannibalism among wild chimpanzees. Nature, 238, 410–411. Cairns, R. B. (1979). Social development: The origins and plasticity of interchanges. San Francisco: Freeman. Campbell, A. (1999). Staying alive: Evolution, culture, and women’s intrasexual aggression. Behavioral and Brain Sciences, 22, 203–252. Caplan, M., Vespo, J. E., Pedersen, J., & Hay, D. F. (1991). Conflict and resolution in small groups of one- and two-year-olds. Child Development, 62, 1513–1524. Chase, K. A., O’Leary, K. D., & Heyman, R. E. (2001). Categorizing partner-violent men within the reactive-proactive typology model. Journal of Consulting and Clinical Psychology, 69, 567–572. Chen, S., Lee, A. Y., Bowens, N. M., Huber, R., & Kravitz E. A. (2002). Fighting fruit flies: A model system for the study of aggression. Proceedings of the National Academy of Science USA, 99, 5664–5668. Connor, D. F., Steingard, R. J., Anderson, J. J., & Melloni, R. H., Jr. (2003). Gender differences in reactive and proactive aggression. Child Psychiatry and Human Development, 33, 279–294. Coyne, S. M., Archer, J., & Eslea, M. (2004). We’re not friends anymore! Unless . . . ”: The frequency and harmfulness of relational, indirect and social aggression. Manuscript submitted for publication. Craig, W. (1928). Why do animals fight ? International Journal of Ethics, 31, 264–278. Crick, N. R. (1996). The role of overt aggression, relational aggression, and prosocial behavior in the prediction of children’s future social adjustment. Child Development, 67, 2317–2327. Crick, N. R., Grotpeter, J. K., & Bigbee, M. A. (2002). Relationally and physically aggressive children’s intent attributions and feelings of distress for relational and instrumental peer provocations. Child Development, 73,1134–1142. Dodge, K. A., & Coie, J. D. (1987). Social-information-processing factors in reactive and proactive aggression in children’s peer groups. Journal of Personality and Social Psychology, 53, 1146–1158. Dollard, J., Doob, L. Miller, N., Mowrer, O., & Sears, R. (1939). Frustration and aggression. New Haven, CT: Yale University Press. Felson, R. B. (2002). Violence and gender reexamined. Washington, DC: American Psychological Association.
44
THE DEVELOPMENT OF AGGRESSION
Feshbach, S. (1964). The function of aggression and the regulation of aggressive drive. Psychological Review, 71, 257–272. Feshbach, S. (1971). Dynamics and morality of violence and aggression. American Psychologist, 26, 281–292. Galen, B. R., & Underwood, M. K. (1997). A developmental investigation of social aggression among children. Developmental Psychology, 33, 589–600. Gendreau, P. L., Gariépy, J. L., Petitto, J. M., & Lewis, M. H. (1997). D1 dopamine receptor mediation of social and nonsocial emotional reactivity in mice: Effects of housing and strain difference in motor activity. Behavioral Neuroscience, 111, 424–434. Gendreau, P. L., Petitto, J. M., Gariépy, J. L., & Lewis, M. H. (1998). D2-like dopamine receptor mediation of social-emotional reactivity in a mouse model of anxiety: Strain and experience effects. Neuropsychopharmacology, 18, 210–221. Giedd, J. N., Blumenthal, J., Jeffries, N. O., Castellanos, F. X., Liu, H., Zijdenbos, A., Paus, T., Evans, A. C., & Rapoport, J. L. (1999). Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neuroscience, 2, 861– 863. Grant, E. C., & Mackintosh, J. H. (1963). A comparison of the social postures of some common laboratory rodents. Behaviour, 21, 246–259. Gregg, T. R., & Siegel, A. (2001). Brain structures and neurotransmitters regulating aggression in cats: Implications for human aggression. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 25, 91–140. Hartup, W. W. (1974). Aggression in childhood: Developmental perspectives. American Psychologist, 29, 336–341. Hartup, W. W., & De Wit, J. (1974). The development of aggression: Problems and perspectives. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 595–620). The Hague: Mouton. Hinde, R. A. (1974). The study of aggression: Determinants, consequences, goals and functions. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 3–27). The Hague: Mouton. Hinde, R. A. (1992). Some complexities in aggressive behaviour. In A. Fraczek & H. Zumkley (Eds.), Socialization and aggression (pp. 3–10). Berlin: Springer-Verlag. Johnson, R. N. (1972). Aggression in man and animals. Philadelphia: Saunders. Karli, P. (1956). The Norway rat’s killing response to the while mouse: An experimental analysis. Behaviour, 10, 81–103. Keeney, A. J., & Hogg, S. (1999). Behavioural consequences of repeated social defeat in the mouse: Preliminary evaluation of a potential animal model of depression. Behavioral Pharmacology, 10, 753–764. Knutson, J. F. (1973). Aggression as manipulatable behavior. In J. F. Knutson (Ed.), The control of aggression: Implications from basic research. (pp. 253–295). Chicago: Aldine. Loeber, R., & Hay, D. (1997). Key issues in the development of aggression and violence from childhood to early adulthood. Annual Review of Psychology, 48, 371–410. McGrew, W. C. (1972). An ethological study of children’s behaviour. London: Academic Press. Moffitt, T. E. (1993). Adolescent-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674–701. Moyer, K. E. (1968). Kinds of aggression and their physiological basis. Communications in Behavioral Biology, 2, 65–87.
Subtypes of Aggression
45
Moyer, K. E. (1973). The physiological inhibition of hostile behavior. In J. F. Knutson (Ed.), The control of aggression: Implications from basic research (pp. 9–38). Chicago: Aldine. Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal emotions. New York: Oxford University Press. Patterson, G. R., & Cobb, J. A. (1973). Stimulus control for classes of noxious behavior. In J. F. Knutson (Ed.), The control of aggression: implications from basic research (pp. 145–199). Chicago: Aldine. Poulin F., & Boivin, M. (2000). Reactive and proactive aggression: Evidence of a two-factor model. Psychological Assessment, 12, 115–122. Pulkkinen, L. (1987). Offensive and defensive aggression in humans: A longitudinal perspective. Aggressive Behavior, 13, 197–212. Pulkkinen, L. (1996). Proactive and reactive aggression in early adolescence as precursors to anti- and prosocial behavior in young adults. Aggressive Behavior, 22, 241– 257. Raine, A., Lencz, T., Bihrle, S., LaCasse, L., & Colletti, P. (2000). Reduced prefrontal gray matter volume and reduced autonomic activity in antisocial personality disorder. Archives of General Psychiatry, 57, 119–127. Rule, B. G. (1974). The hostile and instrumental functions of human aggression. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 121–141). The Hague: Mouton. Scott, J. P., & Fredericson, E. (1951). The causes of fighting in mice and rats. Physiological Zoology, 24, 273–309. Sears, R. R. (1961). Relation of early socialization experiences to aggression in early childhood. Journal of Abnormal and Social Psychology, 63, 466–492. Smith, P. K. (1974). Aggression in a preschool playgroup: Effects of varying physical resources. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 97–105). The Hague: Mouton. Tedeschi, J. T., & Felson, R. B. (1994). Violence, aggression, and coercive actions. Washington, DC : American Psychological Association. Tremblay, R. E. (2000). The development of aggressive behaviour during childhood: What have we learned in the past century? International Journal of Behavioral Development, 24, 129–141. Tremblay, R. E., Japel, C., Pérusse, D., Boivin, M., Zoccolillo, M., Montplaisir, J., & McDuff, P. (1999). The search for the age of “onset” of physical aggression: Rousseau and Bandura revisited. Criminal Behavior and Mental Health, 9, 24– 39. Ulrich, R. E., & Azrin, N. H. (1962). Reflexive fighting in response to aversive stimulation. Journal of the Experimental Analysis of Behavior, 5, 511–520. Ulrich, R. E, Dulaney, S., Arnett, M., & Mueller, K. (1973). An experimental analysis of nonhuman and human aggression. In J. F. Knutson (Ed.), The control of aggression: Implications from basic research (pp. 79–111). Chicago: Aldine. Ulrich, R. E., Johnston, M., Richardson, J., & Wolff, P. (1963). The operant conditioning of fighting behaviour in rats. Psychological Record, 13, 465–470. Valzelli, L. (1981). Psychobiology of aggression and violence. New York: Raven Press. Vitaro, F., Brendgen, M., & Tremblay, R. E. (2002). Reactively and proactively aggressive children: Antecedents and subsequent characteristics. Journal of Child Psychology and Psychiatry, 43, 495–505.
46
THE DEVELOPMENT OF AGGRESSION
Vitaro, F., Gendreau, P. L., Tremblay, R. E., & Oligny, P. (1998). Reactive and proactive aggression differentially predict later conduct problems. Journal of Child Psychology and Psychiatry, 39, 377–385. Wilson, E. O. (1980). Sociobiology: The abridged edition. Cambridge, MA: Harvard University Press. Zillman, D. (1978). Hostility and aggression. Hillsdale, NJ: Erlbaum.
THEFighting Play DEVELOPMENT OF AGGRESSION
3 Play Fighting Aggression, Affiliation, and the Development of Nuanced Social Skills S ERGIO M. P ELLIS, V IVIEN C. P ELLIS, and A FRA F OROUD
In traditional New Guinean society, if two men—strangers to each other— were to meet on a path between their respective villages, they would then likely, if they wished to avoid bloodshed, begin to list all the people to whom they were related in the hope of discovering some common ground between them (Diamond, 1997). Although having the skill to win in a potentially aggressive encounter is important, this New Guinean practice illustrates that having the skill to avoid unnecessary aggression may be just as important. It is becoming clear that aggression is but a part of a larger tool kit for gaining advantage in a complex social world. This appears to be true not just for humans, but for many social mammals (de Waal & Aureli, 2000). In such a world, knowing who and when to fight becomes important, and thus so does having the cognitive skills to gather the information required to make the appropriate decisions. Placing aggression into a broader social context has implications for the understanding of the developmental processes that shape aggression. The problem lies not only in learning how and when to use aggression, but also in knowing how to assess the situation so as to determine whether a nonaggressive approach would be a more suitable solution. In this chapter we show that play fighting can provide a vehicle for the development of such skills. Play fighting resembles serious fighting in that the partners compete for access to some advantage. Most often, this advantage takes the form of biting or striking another individual on a particular body target (Aldis, 1975). The targets competed over during play fighting can be those that in adulthood are 47
48
THE DEVELOPMENT OF AGGRESSION
used in conspecific aggression, sex, or predation. Despite these differences in targets, all forms of play fighting have play-typical rules of attack and defense in common (Pellis & Pellis, 1998a). In contrast, during conspecific aggressive encounters, in the early stages of sexual encounters and in predatory encounters, the opponents follow the rules of attack and defense typical of serious fighting. For play fighting to remain playful, it needs to follow the 50:50 rule (Aldis, 1975; Altmann, 1962). That is, each pairmate has to win close to 50% of the playful encounters. Such a win:loss ratio requires that the rules of attack and defense differ from those that apply to serious fighting (Pellis & Pellis, 1998a). When an attack is launched during a serious fight, the attacking animal has to guard against retaliation from the opponent. Typically, in order to do so, the attacker simultaneously incorporates some defensive tactic into its attack, which limits the defender’s opportunity to counterattack (Pellis, 1997). Similarly, when defending itself against a serious attack, the defender uses an intensity in its defense that reduces the likelihood of a successful penetration by the attacker (Pellis, 1997). In contrast, during play fighting, the attacker rarely incorporates defensive maneuvers in its attack, thus facilitating successful counterattacks by the defender (Pellis & Pellis, 1998a). In addition, when defending itself against a playful attack, the defender will use an intensity of defense that is lower than that in serious fights, increasing the likelihood of a successful contact by the attacker (Pellis & Pellis, 1998a). That is, in play fighting, the tactics of attack and defense are decoupled so as to ensure that each animal can successfully contact the partner. This contrasts sharply with serious fighting, in which an animal may deliver injurious, or even lethal, attacks if the opponent leaves its target exposed (Blanchard, Blanchard, Takahashi, & Kelly, 1977; Geist, 1971; Pellis & Pellis, 1992). In the following text, we show that given its rule structure, play fighting can provide fertile ground for learning a variety of social skills. Detailed studies of the development of play fighting in rats (Pellis, 2002b) provide a model for understanding how play fighting may be used for sharpening the social skills needed to straddle the subtle divide between competition and cooperation effectively (Pellis, 2002a).
PLAY FIGHTING IN RATS1 In rats, play fighting involves attack and defense of the nape (Figure 3.1). Successful contact with the nape involves gently rubbing the nape with the nose (Pellis & Pellis, 1987; Siviy & Panksepp, 1987). Nape contact is a target typical of adult sexual encounters (Pellis, 1988, 1993), whereas agonistic attacks involve bites directed at the lower dorsum or the face (Blanchard et al., 1977; Pellis, 1997). To defend themselves against playful nape contact, rats use the same tactics of defense that are used in serious aggression, but modify them to match both the differences in targets (Pellis & Pellis, 1987) and the differences
Play Fighting
49
FIGURE 3.1. A sequence of play fighting is shown for a pair of 31-day-old rats. Note that the attacking rat (on the left), lunges toward its partner’s nape. The defending rat avoids nape contact by turning to supine. From this on-top/on-bottom position, both rats make attempts to contact each other’s napes. Adapted from Pellis and Pellis (1987). Copyright 1987 by Wiley. Adapted by permission.
in the rules of execution of those tactics (Pellis & Pellis, 1998a). Even though most common in the juvenile phase (Meaney & Stewart, 1981; Panksepp, 1981; Thor & Holloway, 1984), play fighting in rats persists well into young adulthood (Adams & Boice, 1983, 1989; Pellis & Pellis, 1991b). Indeed, play fighting between adults occurs among many mammalian species, in both sexual and nonsexual encounters (Pellis & Iwaniuk, 1999b, 2000). However, whether the context is sexual or nonsexual, such play fighting appears to be used to assess and manipulate the other animal, be it human or nonhuman (Breuggeman, 1978; Pellis & Iwaniuk, 1999b). When it occurs in subadults and adults, play fighting appears to be used in two general contexts—for social bonding and for social testing (Pellis & Pellis, 1996; Pellis & Iwaniuk, 1999b, 2000; Smith, Fantella, & Pellis, 1999). Social bonding may include courtship and pair bonding, as well as nonsexual social affiliation. Social testing most typically involves some form of jostling for social status. The distribution of these functions vary across species, with some species using them only for affiliation, some using them only for dominance testing, and some using them for both (Pellis, 2002a). In rats, play fighting is used for both these purposes (Pellis, Pellis, & McKenna, 1993; Smith et al., 1999). It should be noted that in rats, as in many other mammals, play fighting is more frequent among males (Pellis & Pellis, 1990, 1997), although preferences for which sex is the most attractive as a play partner may change during
50
THE DEVELOPMENT OF AGGRESSION
development (Meaney & Stewart, 1981). With regard to the social uses of play fighting, only male rats show the context-specific modulation of content that is consistent with such uses (Pellis, Field, Smith, & Pellis, 1997). In some other species of mammals, females also appear to use play fighting for social assessment and manipulation (Pellis, 2002a; Pellis & Iwaniuk, 2000). For the present purposes, in this discussion of rats, males are the focus of consideration. Play fighting in rats has several features useful for the analysis of the development of social skills. First, the difference in targets between play fighting and serious fighting makes it easy to differentiate between these two forms of fighting. This is very useful for identifying cases where play fights escalate into serious fights. By using a switch in target as a marker, it was found that playful encounters escalate into serious ones if one of the partners breaks the playtypical rules of attack and defense. That is, if one of the pair stops allowing the partner an opportunity for reciprocal contact, the partner may switch to an aggressive attack (Pellis & Pellis, 1998a). Second, adults use play fighting in clearly functional contexts, allowing us to identify the social skills needed and so evaluate play fighting earlier in development to ascertain whether the opportunity to acquire such skills is present. To explain what those skills entail, we describe the social uses of play fighting among adult male rats. This understanding then sets the stage for interpreting the developmental changes in play fighting present in this species.
THE ADULT USES OF PLAY FIGHTING Within a colony, adult male rats form a dominance hierarchy (Blanchard & Blanchard, 1990; Calhoun, 1962; Flannery & Lore, 1977) in which subordinates initiate more playful encounters with dominants than they do with each other (Pellis et al., 1993). When playfully contacted by a dominant, subordinates are more likely to roll over onto their backs, a response typical of juveniles (Panksepp, 1981; Pellis & Pellis, 1987), whereas when contacted by another subordinate, they are more likely to remain standing and push against the attacker with their flanks (Pellis et al., 1993), a response typical of a dominant male (Pellis & Pellis, 1991a, 1992). In the absence of dominance relationships, postpubertal males are more likely to stand and push than to roll over to supine (Pellis & Pellis, 1990; Smith, Field, Forgie, & Pellis, 1996; Smith, Forgie, & Pellis, 1998, Smith et al.,1999). Thus, the increased soliciting of playful contact by subordinates toward dominants, and the juvenile-like response toward those dominants, suggests that the play fighting is used as a means of maintaining “friendly” relations with dominants (Pellis et al., 1993). However, not all subordinates are equally obsequious. The greater the dominance asymmetry between pairmates, the more juvenile-like the playful responses by the subordinate (Pellis & Pellis, 1992).
Play Fighting
51
Indeed, it is those pairmates that are the least asymmetrical in their play relations that are the most likely to escalate the playful encounter into a serious fight (Pellis & Pellis, 1991a). Similarly, when male rats that are unfamiliar with each are placed into a neutral arena, it is the dominant–dominant combination that is the most likely to escalate into a serious fight (Smith et al., 1999). Close inspection of these escalations suggests that the play fighting preceding the serious fight is “rougher.” These observations also suggest that when animals are testing each other in regard to their opponent’s capability to maintain or gain a position of superior dominance, the play can then be escalated into a quasi-aggressive intensity. Therefore, there are two extremes in the style of play that are available to an individual—a gentler form, which is seen when a subordinate is maintaining familiarity with a dominant, and a rougher form, which is seen when one rat is probing another one for its weaknesses. As previously described, these differences in play intensity among adult male rats can be explicitly converted into formal rules of engagement and can provide a rat with the basis for judging how such play can be used to assess and manipulate its partner. For example, when a subordinate rat uses play fighting as a means of social bonding with the dominant male in the colony, it bends the 50:50 rule in the dominant’s favor, whereas when a subordinate uses play fighting to probe the dominant for weakness, or when it encounters an unfamiliar rat, it bends the rule in its own favor. In the latter case, the rat can assess how much of a deviation from equality the opponent will tolerate before retaliating aggressively. A simple rule of thumb may operate in such a scenario. If the dominant or unfamiliar rat tolerates the bending of the 50:50 rule in the opponent’s favor, the opponent may then bend the rule further, until the relative status of the pair is reversed. However, if the dominant starts to respond forcefully, there is still time for the opponent to back down before the encounter escalates into a serious fight. Many primates and carnivores that have affiliative signals that can be used to indicate playful intention may have a capacity for greater flexibility along this playful–serious gradient, as such signals can be used to diffuse unwanted escalations (see Bekoff, 1995; Boulton, 1994; Pellis & Pellis, 1996, 1997). What all these social uses of play fighting have in common is that they involve assessment and manipulation of the partner. As such, they are consistent with evolutionary game theory models that show that if an animal can assess the opponent’s fighting ability relative to its own, potentially damaging fights can be avoided (Parker, 1974). The benefits derived from such assessment could thus have been the selective advantage for the evolution of play fighting as a social tool in rats and other mammals. What the present analysis offers is a window into the causal mechanisms that make such an assessment possible. Furthermore, given the cognitive demands associated with such functions, the question arises as to whether play in the juvenile phase is organized so as to facilitate the development of these skills.
52
THE DEVELOPMENT OF AGGRESSION
IMPLICATIONS FOR OUR UNDERSTANDING OF THE PLAY OF JUVENILES All muroid rodents studied thus far appear to have a form of play fighting that resembles the species-typical patterns of adult precopulatory behavior (Pellis, 1993). The resemblance is in both the body targets attacked and in the defensive maneuvers that are used to block these attacks. During postweaning development, these playful interactions grow increasingly like the adult sexual encounters (Pellis & Pellis, 1998b). The rat, however, exhibits a strikingly different developmental pattern. Although in the rat the target competed over during play fighting is the same as for sexual encounters (Pellis & Pellis, 1987), the pattern of defense differs. During sexual encounters the female rat primarily evades the male’s contact, while during play fighting the attacked partner does so in only 20–25% of contacts (Pellis & Pellis, 1987). Moreover, the form of the contact promoting defense also differs. During precopulatory encounters, female rats rarely flip over onto their backs ( 4.0) are superimposed on axial (capsule) and sagittal (arcuate) sections through the magnetic resonance image of a single subject. The images depict the exact brain locations that showed statistically significant correlations between white matter density and the subject’s age (n = 111; age: 4–17 years). From Paus et al. (1999). Copyright 1999 by the American Association for the Advancement of Science. Reprinted by permission.
246
DETERMINANTS OF AGGRESSION
FIGURE 12.3. Principles of functional MRI (the blood oxygenation-level dependent signal). Courtesy of Dr. Bruce Pike, Montreal Neurological Institute.
EEG and MEG Compared with fMRI, more direct and real-time measurements of electrical and magnetic signals generated by brain tissue are recorded by EEG and MEG (Niedermeyer & Lopes da Silva, 1999). EEG and MEG are the methods of choice for determining when changes in neural activity occur. The main sources of these signals are intra- and extracellular currents, or field potentials, elicited by the activation of excitatory and inhibitory synapses. Spatial and temporal summation is necessary, however, to generate signals strong enough to be detected outside the head. Such a summation occurs most often during simultaneous excitatory inputs onto apical dendrites of pyramidal cells; the apical dendrites are for the most part oriented in parallel with each other and are perpendicular to the cortical surface. EEG detects field potentials regardless of their orientation relative to the skull. In contrast, MEG can measure only magnetic fields perpendicular to the skull. Such fields are generated by tangential current dipoles and, because of the orientation of pyramidal cells and their apical dendrites in the cortex, reflect primarily postsynaptic activity occurring in the cerebral sulci. Analysis of EEG and MEG data for brain mapping purposes focuses typically on two domains: (1) event-related potentials and (2) event-related synchronization or desynchronization of brain activity.
MRI STUDIES OF BRAIN DEVELOPMENT Brain weight reaches adult values (about 1.45 kg) between 10 and 12 years of age. The fastest growth occurs during the first three years of life, so that by the
Mapping Brain Development and Aggression
247
age of 5 years the infant’s brain weighs about 90% of the adult value (Dekaban, 1978). Clearly, the changes in brain morphology in childhood and adolescence are more subtle that those in the first four years of life (see Paus et al., 2001, for a review of MR studies carried out during infancy). Qualitative evaluation of MR images is of little value at this point, and the ability to obtain quantitative measurements is of the essence if we are to detect brain maturation during this “late” period of brain development. Several approaches have been used to obtain such measurements, including (1) semiautomatic or automatic classification of brain tissue and the subsequent “count” of image elements classified as a particular tissue type (e.g., white matter), (2) manual outlining of a structure of interest (e.g., the corpus callosum) on an MR image and the subsequent calculation of its volume/area, (3) voxel-wise analysis of local growth using deformation fields, and (4) voxel-wise statistical analysis of white matter “density.” Classification of brain tissue into grey and white matter and cerebrospinal fluid allowed several investigators to calculate the absolute and/or relative brain volume occupied by white matter. In the most extensive study to date, Giedd et al. (1999a) reported age-related changes in 145 children and adolescents (age 4–20 years) in volumes of grey and white matter of the frontal, parietal, temporal, and occipital lobes; the study used a cross-sectional (n = 145) and longitudinal (65/145 subjects) design. Volumes were quantified by combining a technique using an artificial neural network to classify tissues based on voxel intensity, with a technique performing nonlinear registration to a template brain for which the four lobes had been manually defined (Collins, Zijdenbos, Baare, & Evans, 1999). A significant increase in the absolute volume of white matter was found in this study, with the volume increasing steadily across ages 4–22 by about 12%; this increase was similar in the four different lobes but was steeper in boys than in girls. These recent findings are consistent with the previous observations of other authors, including the agerelated increase in the cerebral white matter : grey matter ratio (Jernigan & Tallal, 1990; 8–10 years [n = 9] vs. 25–39 years [n = 15]); the increase in absolute and relative volumes of “cortical” white matter (Pfefferbaum et al., 1994; 3 months–30 years [n = 88]); and increase in the relative volume of cerebral white matter (Reiss, Abrams, Singer, Ross, & Denckla, 1996; 5–17 years [n = 85]). In the Pfefferbaum et al. (1994) cross-sectional study, cortical white matter continued to increase from birth to about 20 years, after which it leveled off. In the majority of the aforementioned studies, age-related increases in white-matter volume are accompanied by decreases in grey-matter volume. This observation is best illustrated in a study by Steen, Ogg, Reddick, and Kingsley (1997; 4–10 years [n = 19], 10–20 years [n = 31], 20–30 years [n = 20]), who measured T1 relaxation times and used these to classify all voxels into those containing either “pure” grey or white matter; the number of white-matter and grey-matter voxels increased and decreased with age, respectively, with the inverse relationship between the number of “white” and
248
DETERMINANTS OF AGGRESSION
“grey” voxels (r = .64; p < .0001) (Steen et al., 1997). This finding raises the possibility that a decrease in (cortical) grey-matter volumes may, at least in part, reflect intracortical myelination (and the resulting partial volume effects) rather than loss of grey-matter tissue due to “pruning.” Recent observations by Zilles and collaborators are consistent with this view; using high-resolution MR imaging and post-mortem histological analysis of the (adult) human cerebral cortex, they could explain up to 80% of variance in T1 relaxation times by differences in myeloarchitecture in deep cortical layers (Zilles et al., unpublished observations). These changes in the volume of white matter, calculated using various tissue-classification algorithms, reflect, to a great extent, subtle age-related changes in relaxation times. Steen et al. (1997; see earlier for sample size and age) measured T1 relaxation times in nine regions of interest (ROIs), including the four placed in the white matter: (1) the genu of the corpus callosum, (2) frontal white matter, (3) optic radiation, and (4) occipital white matter. The mean values of T1 decreased from childhood (4–10 years) to adulthood (20– 30 years) in all four white-matter structures. Significant decreases were also observed in comparing children with adolescents (10–20 years) in all whitematter structures except the genu of the corpus callosum, and in comparing adolescents with young adults in all regions except occipital white matter. Transverse relaxation times (T2) were measured by Hassink, Hiltbrunner, Muller, and Lutschg (1992; 8–10 years [n = 9] and 24–25 years [n = 8]) in eight ROIs, including the frontal subcortical white matter, the corpus callosum and the genu of the internal capsule. Although limited by a small sample, significant decreases in T2 values were found from childhood to adulthood in all white-matter regions but the corpus callosum. These two quantitative studies clearly illustrate significant age-related changes in white matter, most likely caused by small but consistent increases in the degree of myelination. The volumetric assessment of age-related changes in white matter is often a global rather than a regional one. This is due to the absence of readily detectable boundaries between different fiber tracts constituting the white matter of the four cerebral lobes. This is not the case for the corpus callosum, the largest fiber tract in the human brain interconnecting the cerebral cortex of the left and right hemispheres. The corpus callosum can be easily delineated on a midsagittal section and, therefore, its area can be measured with great precision. Even though the corpus callosum acquires an “adult” appearance by about the age of 1 year (Barkovich & Kjos, 1988), its growth continues until early adulthood. In one of the first MR-based morphometric studies, Pujol, Vendrell, Junque, Marti-Vilalta, and Capdevila (1993) measured the area of the corpus callosum in adolescents (11–19 years, n = 14), young adults (20–29 years, n = 45) and older adults (30–61 years, n = 31) using a longitudinal design; two scans were acquired in each subject with an average between-scan interval of 2 years. Not surprisingly, the largest growth in the 2-year period was observed in the youngest group (an increase of about 10%). But the area of the corpus callosum increased significantly even in the group of young
Mapping Brain Development and Aggression
249
adults (24–25 years, an increase of about 5% over 2 years); no change was observed from the age of 27 years on. In a cross-sectional study, Giedd et al. (1999a; 4–18 years, n = 114) confirmed the significant age-related increase in the total area of the corpus callosum. They were also able to demonstrate that most of this growth occurs in the posterior half of the corpus callosum and, in particular, in the splenium; no growth was observed in the genu. These findings were corroborated by the same authors, who followed up their original sample and rescanned a large number of the subjects (n = 75) at least once 2 years later (Giedd et al., 1999b). Again, the growth of the splenium far outperformed that of any other subdivision of the corpus callosum. This finding is somewhat surprising in light of the caudal-to-rostral time sequence of myelination of this structure after birth and the presumed late development of the prefrontal cortex, a region interconnected through the (rostral) genu of the corpus callosum. Further studies are necessary to establish whether changes in the splenium reflect possible ongoing maturation of the inferior temporal and occipital cortices, the main sources of the callosal fibers in the splenium (Pandya & Seltzer, 1986). The rostro–caudal wave of growth of the corpus callosum was recently demonstrated by a computational analysis of so-called deformation fields (Thompson et al., 2000; 3–15 years, n = 6). The 3-D deformation fields specify, at each voxel, the vector of forces applied to bring the local anatomy of the subject’s brain in alignment with that of the template brain. Thompson et al. (2000) rescanned each subject within 1–4 years and, by subtracting local deformation fields obtained in each subject at Time 1 and Time 2, were able to visualize changes in the local growth of the corpus callosum. Consistent with the previous findings, changes in the genu were observed in the youngest subject (scanned at 3 and 6 years of age), whereas changes in the splenium were most pronounced in rescanning during adolescence (Time 1: 9 years, Time 2: 13 years). In order to detect more subtle variations, however, group analysis of deformation fields and its statistical evaluation will be necessary. Chung et al. (2001) have developed a novel statistical analysis of local growth that will allow investigators to evaluate the statistical significance of age-related changes in deformation fields throughout the brain. Subtle regional variations in white matter can also be evaluated using a voxel-by-voxel analysis of the images. This approach borrows from the concepts developed in the context of functional neuroimaging: It is based on the use of standardized stereotaxic space and voxel-based statistics. In a study of age-related changes in white matter, we (Paus et al., 1999; 4–18 years, n = 111) have observed significant age-related changes in white-matter “density” in the posterior limb of the internal capsule and in the left arcuate fasciculus; the former is likely to contain fibers connecting the motor cortex and spinal cord, and the latter those connecting the anterior (Broca’s) and posterior (Wernicke’s) language areas. The analysis of images obtained in 111 children and adolescents included the following steps: (1) nonlinear transformation of images into standardized stereotaxic space to remove global and local differ-
250
DETERMINANTS OF AGGRESSION
ences in the size and shape of the individual brains, (2) classification of brain tissue into white matter, grey matter, and cerebrospinal fluid, (3) blurring of white-matter binary masks to generate 3-D maps of white-matter “density,” and (4) correlation between voxel values of white-matter density and the subject’s age. Such voxel-based analyses of age-related variations in white- and grey-matter (Sowell et al., 1999) densities complement the volumetric approach in allowing for subtle local differences to emerge in regions that may not be delineated as a single volume. However, relatively large numbers of individuals and rather conservative statistical criteria need to be applied to separate signal from noise in a reliable fashion. Overall, there is ample evidence that white matter continues to mature during childhood and adolescence, increasing its volume and becoming more myelinated. Most of the changes are occurring throughout this period of development, with no pronounced differences in the rate of maturation during puberty. The absence of such developmental “spurs” could, however, reflect the lack of high-density longitudinal data for this period of life. Several largescale longitudinal MR studies of brain maturation are under way at the moment; these are likely to provide important insights regarding the structure– function relationship during infancy, childhood, and adolescence in healthy individuals.
BRAIN MAPPING STUDIES OF AGGRESSION Most of the literature on the neural underpinnings of aggressive behavior reflects a theoretical framework that views aggression in the context of antisocial behavior, such as violence and delinquency, rather than as one of the essential (adaptive) motivational states, such as hunger or maternal instinct (for recent criticism, see Tremblay, 2003). It is therefore not surprising that the vast majority of previous studies are based on two methodological approaches: (1) clinical studies of overt aggressive behavior in patients with brain lesions and (2) neuroimaging studies of brain abnormalities in individuals who had committed acts incompatible with societal norms, such as murderers. In the first part of this section, I review this literature and, given its emphasis on the frontal lobes, I also provide a brief synopsis of the structure and function of this part of the primate cerebral cortex. In the second part, I suggest an alternative approach for studying aggression and its elements in healthy subjects; this approach is based on a theoretical framework that views aggression as a social interaction between familiar individuals and, consequently, includes conflict resolution (reconciliation) as an integral part of aggressive behavior (De Waal, 2000). The search for a neural substrate of aggression has a long history that started with clinical observations of patients with brain lesions. In the 19th century, several reports described individuals who had had an injury to the frontal lobes and, subsequently, showed changes in their “personality.” The
Mapping Brain Development and Aggression
251
most famous example is Phineas Gage, who suffered a large penetrating injury to the ventromedial and orbitofrontal cortex, after which he became fitful, irreverent, profane, indifferent, inpatient, obstinate, and unable to plan (Harlow, 1868; Damasio, Grabowski, Frank, Galaburda, & Damasio, 1994). In the 20th century, several studies of war veterans confirmed a link between lesions to frontal lobes and an increased risk of aggressive behavior (e.g., Grafman et al., 1996, reviewed in Brower and Price, 2001; Kleist, 1934, cited in Brower & Price, 2001; Lishman, 1968; Virkkunen, Nuutila, & Huusko, 1976). In the majority of these studies, the injury occurred in young male adults, who were studied many years after their injuries. Several case studies also described a high incidence of aggressive behavior in adults who suffered lesions to the frontal cortex early (before the age of 8 years) in their lives (Anderson, Bechara, Damasio, Tranel, & Damasio, 1999; Eslinger, Grattan, Damasio, & Damasio, 1992; Price, Daffner, Stowe, & Mesulam, 1990). Together, these lesion studies suggest that an intact frontal cortex is important for adequate control of aggressive behavior in adulthood; the early-lesion cases suggest that remaining brain structures cannot compensate for the loss of brain tissue in childhood. Which parts of the frontal lobes are critical in this respect? A reconstruction of the skull injury suffered by Phineas Gage led Damasio and colleagues to conclude that “the lesion favored the ventromedial region of both frontal lobes while sparing the dorsolateral” (Damasio et al., 1994). The Vietnam Head Injury Study found a similar association between aggressive behavior and lesions in the orbital and medial aspects of the frontal lobes (Grafman et al., 1996). To understand the significance of this pattern, a brief review of the organization of the primate frontal cortex can be helpful here. The frontal cortex constitutes about 35% of all cerebral cortex in humans and great apes (Semendeferi, Lu, Schenker, & Damasio, 2002). It contains the primary motor cortex, premotor cortex (dorsal and ventral premotor cortex on the lateral convexity; supplementary motor area (SMA) and the cingulate motor areas (CMA) on the medial convexity), and the prefrontal cortex. The prefrontal cortex is further subdivided into the mid-dorsolateral and mid-ventrolateral frontal cortex, the frontopolar cortex, the orbitofrontal cortex, and the mesial frontal cortex; the latter also includes the anterior cingulate cortex surrounding the anterior part of the corpus callosum. Each of these gross morphological subdivisions consists of a number of areas that differ in the exact arrangement of cells in the six cortical layers (cytoarchitecture; see Figure 12.4) and in their connectivity with other cortical and subcortical regions. These anatomical differences determine to a great extend what type of information is being processed in a given region, and which other regions share this information. Such regional specialization is reflected, in turn, in the specific contributions made by a given “node” of a particular neural network in support of a specific behavior. Given the structural heterogeneity of the frontal cortex and the presence of extensive connectivity of frontocortical regions with other parts of the cere-
252
DETERMINANTS OF AGGRESSION
FIGURE 12.4. Schematic diagrams of the lateral (left), medial (middle), and inferior (right) surfaces of the human frontal lobe to illustrate its cytoarchitectonic parcellation. From Petrides and Pandya (1994). Copyright 1994 by Elsevier. Reprinted by permission.
bral cortex, most prominently with the temporal and parietal cortex, it is not surprising that the frontal cortex supports a multitude of motor, sensory, and cognitive functions. For example, the premotor areas on the medial convexity may be essential for sequencing behavior (SMA) and willed initiation and suppression of movement (CMA; reviewed in Paus, 2001). The mid-dorsolateral and mid-ventrolateral frontal cortex play an important role in working memory and retrieval, respectively (Petrides, 1996). Maturation of the mid-dorsolateral frontal cortex underlies the emergence of working memory during infancy (e.g., A-not-B test) and a further increase in working memory and planning during childhood. Maturations of the lateral prefrontal and/or cingulate cortices may be critical for the dramatic increase in the child’s ability to suppress external interference observed just before the onset of puberty (e.g., suppression of reflexive eye movements; Munoz, Broughton, Goldring, & Armstrong, 1998; Paus, Babenko, & Radil, 1990). Maturation of the orbitofrontal, mediofrontal, and cingulate cortex may be associated with the development of delayed gratification and other aspects of intentional behavior, including the control of motivation states such as aggression. Underlying the latter control of motivational states are extensive connections between the orbitofrontal/cingulate cortex and the amygdala (Barbas, 1995; Morecraft & Van Hoesen, 1998). The second “classical” approach to studies of the neural underpinnings of aggressive behavior is that of imaging the brains of individuals who show different forms of violent or antisocial behavior; unlike the neurological patients described earlier, these individuals have not suffered any known external damage to their brains. Using structural MR imaging, two recent studies described reductions of the volume of grey matter in different parts of the frontal lobes in aggressive patients with temporal lobe epilepsy (Woermann et al., 2000) and individuals with antisocial personality disorder (Raine, Lencz, Bihrle, LaCasse, & Colletti, 2000); the latter study described a 11% reduction in the volume of the orbitofrontal cortex in 21 community volunteers with an-
Mapping Brain Development and Aggression
253
tisocial personality; like changes were not found in two different control groups. Using functional imaging, Raine, Buchsbaum, and LaCasse (1997) described hypometabolism in the prefrontal cortex in 41 individuals charged with murder or manslaughter; in subsequent reanalyses of this sample, Raine and colleagues found that this hypometabolism was present only in those individuals who did not have history of psychosocial deprivation (Raine, Pihl, Stoddard, Bihrle, & Buchsbaum, 1998b) and those who committed impulsive/ emotional rather than purposeful aggression (Raine et al., 1998a). Similar findings were obtained in other clinical populations (reviewed in Brower & Price, 2001). Given the above lesion and imaging findings, we may ask: What is happening in the human brain during early infancy and childhood that “enables” aggressive behavior, which arguably peaks between 2 and 4 years of age (Tremblay, 2003; Tremblay et al., 2004) and then provides adaptive mechanisms, allowing its expression and resolution in a socially acceptable manner? Admittedly, we know too little about structural and functional development of the human brain to answer these questions at this point. It is clear that the most dramatic phase of brain growth takes place during the first 2 years of life: brain weight more than doubles, the overall volume of white matter and the degree of myelination increases sharply, with the overall increases in grey matter being less dramatic (e.g., Matsuzwa et al., 2001; Paus et al., 2001; Utsunomia, Takano, Okazaki, & Mitsudome, 1999). It is important to note that some interesting regional differences are beginning to emerge. For example, the volume of grey matter constituting the hippocampal formation shows a fast increase during the first 2 years of life and continues to increase slowly during childhood and adolescence (Utsunomyia et al., 1999). However, the volume of prefrontal grey matter appears to increase slowly until the age of 8 years, followed by rapid growth between 8 and 14 years (Kanemura, Aihara, Aoki, Araki, & Nakazawa, 2003). Although these patterns are based on cross-sectional studies of a relatively small number of infants and children, they provide a preview of what we may learn from several large-scale MRbased developmental studies that are under way. Returning to our question, it is possible that the development of structures in the mesial temporal lobes, such as the hippocampus and amygdala, underlie the initial development of aggressive behavior, whereas the relatively late development of cortical regions, such as the prefrontal and parietal cortex, enables adaptive regulation of aggression in the social context. Let us now examine some of the processes, and their neural substrates, that may be important for the latter. As suggested in the opening paragraph of this section, most acts of aggression take place in the context of social interactions among familiar individuals, be it at home, daycare, school, or the workplace. As with any other forms of communication, each side of the conflict emits and receives a number of cues, both verbal and nonverbal, that are essential components of the overall dynamics of their (aggressive) interaction. I suggest that the study of the neural substrates of such interactions, and their normal development, will
254
DETERMINANTS OF AGGRESSION
benefit from the use of the experimental paradigms utilized over the past decade by cognitive neuroscientists interested in the brain mechanisms of interpersonal communication carried by gaze, face, body, and voice. Over the past several years a number of brain-mapping studies carried out with PET, fMRI, and EEG identified a key set of brain regions involved in the processing of “biological motion” generated by the movements of the eyes, mouth, and body. In such studies the subjects view stimuli such as eye and mouth movements (e.g., Puce, Allison, Bentin, Gore, & McCarthy, 1998), hand movements/ actions (e.g., Beauchamp, Lee, Haxby, & Martin, 2002; Decety et al., 1997; Grezes, Costes, & Decety, 1999), or body movements conveyed by point lights attached to the body (e.g., Bonda, Petrides, Ostry, & Evans, 1996). Among other brain regions, biological motion leads consistently to the increases in neural activity along the superior temporal sulcus, (STS) (Figure 12.5; reviewed in Allison, Puce, & McCarthy, 2000); the exact location may depend on the moving body part (e.g., mouth vs. eyes vs. hands) and the source of movement (tool vs. hand; Beauchamp et al,. 2002). Although most of the previous studies of biological motion used emotionally “neutral” stimuli, several investigators observed changes in brain activity in structures involved in processing emotional salience (amygdala) or reward value (orbitofrontal cortex). As suggested by Allison et al. (2000), feedforward and feedback interactions between the STS and the amygdala may be critical for the discrimination of various facial expressions and for the
FIGURE 12.5. Regions along the left (a) and right (b) superior temporal sulcus of the human brain “activated” by the movement of various parts of the body. From Allison, Puce, and McCarthy (2000). Copyright 2000 by Elsevier. Reprinted by permission.
Mapping Brain Development and Aggression
255
attentional enhancement of the neural response to socially salient stimuli. Consistent with such an “amplification” mechanism, Kilts, Egan, Gideon, Ely, and Hoffman (2003) observed significantly stronger neural response to dynamic, as compared with static, facial expressions of anger in both the STS and the amygdala. Developmentally, the basic aspects of face perception are in place shortly after birth (Goren, Sarty, & Wu, 1975), but both the quantity and quality of face processing continues to increase all the way through adolescence (e.g., Carey, 1992; McGivern, Andersen, Byrd, Mutter, & Reilly, 2002; Taylor, McCarthy, Saliba, & Degiovanni, 1999); perhaps one of the structural correlates of such continuing maturation of face processing in an increase in white matter density along the occipitotemporal pathway (Watkins et al., 2002). Seeing and hearing signals of upcoming aggression is often enough to prevent it; in the monkey, this can be achieved by quickly averting the gaze (e.g., Capitanio, 2002). The frontal cortex and its various subregions (see the preceding discussion) are likely involved in evaluating the overall context and significance of the detected visual (eyes, face) and auditory (vocalization) displays of aggression and in generating the appropriate behavioral response. This response could, nonetheless, be an aggressive act. If this is the case, what happens next? In many primate species, aggression is often followed by “reconciliation” or “conflict resolution” (Aureli & De Waal, 1997). Rather than viewing aggression as serving a “spacing” function, De Waal argues that “individuals try to ‘undo’ the social damage inflicted by aggression; hence, they will actively seek contact, specifically with former opponents” (De Waal, 2000). Several studies are beginning to document different forms of reconciliatory behavior in children observed in naturalistic settings (e.g., Ljungberg, Westlund, & Lindqvist Forsberg, 1999). Overall, careful studies of the neural substrate underlying various aggression-relevant elements of social cognition will undoubtedly further our understanding of human aggression, its normal development, and its socially unacceptable expressions.
CONCLUSIONS Structural and functional neuroimaging is making strides in elucidating the structure–function relationships underlying aggression, as well as providing novel information on brain maturation in children and adolescents. Previous research has focused on studies of adult individuals and overt aggressive behavior; lesions to the frontal lobes seem to increase the incidence of certain elements of aggressive behavior, whereas having aggression-relevant behavioral traits appears to correlate with functional or structural abnormalities in this brain region. Very little is known, however, about the relationship between aggression and the individual’s ability to detect and interpret nonverbal cues carried by the gaze, the face, and the body in a social context. I
256
DETERMINANTS OF AGGRESSION
suggest that the study of brain–behavior relationships in this domain would be very helpful when combined, for example, with a detailed assessment of the perceptual and cognitive skills of individuals with different developmental histories of aggressive behavior. Finally, it is likely that future research on the role of the frontal lobes in aggression will shift from studying the actual act of aggression to investigating the brain substrate of reconciliation and conflict resolution. Although neuroimaging is a valuable tool in this endeavor, one should not forget about its limitations. For example, although T1-weighted MR images provide an impressive contrast between the (cortical) gray and white matter, we are not sure about the biological source of signal differences between various clinical or age groups; many factors may influence the signal, including the density of neurons, glia, or vascularization, or the degree of (intracortical) myelination. Functional imaging poses different challenges; for example, are age-related differences in the fMRI signal the cause or consequence of age-related changes in a given behavior? In other words, how do we make sure that the “stimulus” is held constant across different populations so that we can evaluate the magnitude of neural response to the same probe, rather than use the fMRI signal as a mere correlate of different behavior? Finally, structure–function correlations observed with structural and functional neuroimaging are just that, correlations. Other experimental approaches must be used to confirm whether a given region is necessary for a given behavior; such an inference can be made only when irreversible (lesion) or reversible (electrical or pharmacological stimulation) manipulations lead to the predicted behavioral consequences. Although some of these techniques can be used in (adult) human subjects, the ultimate tests would require animal models.
ACKNOWLEDGMENTS My research is supported by the Canadian Institutes of Health Research, the Canadian Foundation for Innovation, and the National Science and Engineering Research Council of Canada. I thank Drs. Richard Tremblay and Gabriel Leonard for their comments on the early version of this chapter.
REFERENCES Allison, T., Puce, A., & McCarthy, G. (2000). Social perception from visual cues: Role of the STS region. Trends in Cognitive Science, 4, 267–278. Anderson, S. W., Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (1999). Impairment of social and moral behavior related to early damage in human prefrontal cortex. Nature Neuroscience, 2, 1032–1037. Aureli, F., & De Waal, F. B. (1997). Inhibition of social behavior in chimpanzees under high-density conditions. American Journal of Primatology, 41, 213–228.
Mapping Brain Development and Aggression
257
Barbas, H. (1995). Anatomic basis of cognitive-emotional interactions in the primate prefrontal cortex. Neuroscience and Biobehavioral Reviews, 19, 499–510. Barkovich, A. J., & Kjos, B. O. (1988). Normal postnatal development of the corpus callosum as demonstrated by MR imaging. American Journal of Neuroradiology, 9, 487–491. Beauchamp, M. S., Lee, K. E., Haxby, J. V., & Martin, A. (2002). Parallel visual motion processing streams for manipulable objects and human movements. Neuron, 34, 149–159. Bonda, E., Petrides, M., Ostry, D., & Evans, A. (1996). Specific involvement of human parietal systems and the amygdala in the perception of biological motion. Journal of Neuroscience, 16, 3737–3744. Brower, M. C., & Price, B. H. (2001). Neuropsychiatry of frontal lobe dysfunction in violent and criminal behaviour: a critical review. Journal of Neurology, Neurosurgery, and Psychiatry, 71, 720–726. Capitanio, J. P. (2002). Sociability and responses to video playbacks in adult male rhesus monkeys (Macaca mulatta). Primates, 43, 169–177. Carey, S. (1992). Becoming a face expert. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, 335, 95–102. Chung, M. K., Worsley, K. J., Paus, T., Cherif, C., Collins, D. L., Giedd, J. N., Rapoport, J. L., & Evans, A. C. (2001). A unified statistical approach for deformation-based morphometry. NeuroImage, 14, 595–606. Collins, D. L., Zijdenbos, A. P., Baare, W. F. C., & Evans, A. C. (1999). Animal + insect: Improved cortical structure segmentation. In A. Kuba, M. Samal, & A. Tod (Eds.), Proceedings of the 16th International Conference on Information Processing in Medical Imaging (IPMI) (pp. 210–223). New York: Springer-Verlag. Damasio, H., Grabowski, T., Frank, R., Galaburda, A. M., & Damasio, A. R. (1994). The return of Phineas Gage: Clues about the brain from the skull of a famous patient. Science, 264, 1102–1105. Decety, J., Grezes, J., Costes, N., Perani, D., Jeannerod, M., Procyk, E., Grassi, F., & Fazio, F. (1997). Brain activity during observation of actions. Influence of action content and subject’s strategy. Brain, 120, 1763–1777. Dekaban, A. S. (1978). Changes in brain weights during the span of human life: Relation of brain weights to body heights and body weights. Annals of Neurology, 4, 345– 356. De Waal, F. B. (2000). Primates: A natural heritage of conflict resolution. Science, 289, 586–590. Eslinger, P. J., Grattan, L. M., Damasio, H., & Damasio, A. R. (1992). Developmental consequences of childhood frontal lobe damage. Archives of Neurology, 49(7), 764–769. Giedd, J. N., Blumenthal, J., Jeffries, N. O., Castellanos, F. X., Liu, H., Zijdenbos, A., Paus, T., Evans, A. C., & Rapoport, J. L. (1999a). Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neuroscience, 2, 861–863. Giedd, J. N., Blumenthal, J., Jeffries, N. O., Rajapakse, J. C., Vaituzis, A. C., Liu, H., Berry, Y. C., Tobin, M., Nelson, J., & Castellanos, F. X. (1999b). Development of the human corpus callosum during childhood and adolescence: A longitudinal MRI study. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 23, 571– 588. Goren, C. C., Sarty, M., & Wu, P. Y. (1975). Visual following and pattern discrimination of face-like stimuli by newborn infants. Pediatrics, 56, 544–549.
258
DETERMINANTS OF AGGRESSION
Grafman, J., Schwab, K., Warden, D., Pridgen, A., Brown, H. R., & Salazar, A. M. (1996). Frontal lobe injuries, violence, and aggression: A report of the Vietnam Head Injury Study. Neurology, 46, 1231–1238. Grezes, J., Costes, N., & Decety, J. (1999). The effects of learning and intention on the neural network involved in the perception of meaningless actions. Brain, 122, 1875–1887. Harlow, J. M. (1868). Recovery from the passage of an iron bar through the head. Publications of the Massachusetts Medical Society, 2, 327–347. Hassink, R. I., Hiltbrunner, B., Muller, S., & Lutschg, J. (1992). Assessment of brain maturation by T2–weighted MRI. Neuropediatrics, 23, 72–74. Jernigan, T. L., & Tallal, P. (1990). Late childhood changes in brain morphology observable with MRI. Developmental Medicine and Child Neurology, 32, 379–385. Kanemura, H., Aihara, M., Aoki, S., Araki, T., & Nakazawa, S. (2003). Development of the prefrontal lobe in infants and children: A three-dimensional magnetic resonance volumetric study. Brain and Development, 25(3), 195–199. Kilts, C. D., Egan, G., Gideon, D. A., Ely, T. D., & Hoffman, J. M. (2003). Dissociable neural pathways are involved in the recognition of emotion in static and dynamic facial expressions. NeuroImage, 18, 156–168. Lishman, W. A. (1968). Brain damage in relation to psychiatric disability after head injury. British Journal of Psychiatry, 114, 373–410. Ljungberg, T., Westlund, K., & Lindqvist Forsberg, A. J. (1999). Conflict resolution in 5year-old boys: Does postconflict affiliative behaviour have a reconciliatory role? Animal Behavior, 58, 1007–1016. Matsuzawa, J., Matsui, M., Konishi, T., Noguchi, K., Gur, R., Bilker, W., & Miyawaki, T. (2001). Age-related volumetric changes of brain gray and white matter in healthy infants and children. Cerebral Cortex, 11, 335–342. McGivern, R. F, Andersen, J., Byrd, D., Mutter, K. L., & Reilly, J. (2002). Cognitive efficiency on a match to sample task decreases at the onset of puberty in children. Brain and Cognition, 50, 73–89. Morecraft, R. J., & Van Hoesen, G. W. (1998). Convergence of limbic input to the cingulate motor cortex in the rhesus monkey. Brain Research Bulletin, 45, 209–232. Munoz, D. P., Broughton, J. R., Goldring, J. E., & Armstrong, I. T. (1998). Age-related performance of human subjects on saccadic eye movement tasks. Experimental Brain Research, 121, 391–400. Niedermeyer, E., & Lopes da Silva, F. (Eds.). (1999). Electroencephalography. Basic principles, clinical applications, and related fields (4th ed.). Baltimore: Williams & Wilkins. Pandya, D. N., & Seltzer, B. (1986). The topography of commissural fibers. In F. Lepore, M. Ptito, & H. Jasper (Eds.), Two hemispheres, one brain: Functions of the corpus callosum (pp. 47–73). New York: Liss. Paus, T. (2001). Primate anterior cingulate cortex: Where motor control, drive and cognition interface. Nature Reviews Neuroscience, 2, 417–424. Paus, T. (2003). Principles of functional neuroimaging. In R. B. Schiffer, S. M. Rao, & B. S. Fogel (Eds.), Neuropsychiatry (2nd ed., pp. 63–90). New York: Lippincott, Williams & Wilkins. Paus, T., Babenko, V., & Radil, T. (1990). Development of an ability to maintain verbally instructed central gaze fixation studied in 8 to 10 year children. International Journal of Psychophysiology, 10, 53–61. Paus, T., Collins, D. L., Evans, A. C., Leonard, G., Pike, B., & Zijdenbos, A. (2001). Mat-
Mapping Brain Development and Aggression
259
uration of white matter in the human brain: A review of magnetic-resonance studies. Brain Research Bulletin, 54, 255–266. Paus, T., Zijdenbos, A., Worsley, K., Collins, D. L., Blumenthal, J., Giedd, J. N., Rapoport, J. L., & Evans, A. C. (1999). Structural maturation of neural pathways in children and adolescents: In vivo study. Science, 283, 1908–1911. Petrides, M. (1996). Lateral frontal cortical contribution to memory. Seminars in the Neurosciences, 8, 57–63. Petrides, M., & Pandya, D. N. (1994). Comparative architectonic analysis of the human and the macaque frontal cortex. In F. Boller & J. Grafman (Eds.), Handbook of neuropsychology (Vol. 9, pp. 17–58). Amsterdam: Elsevier Science. Pfefferbaum, A., Mathalon, D. H., Sullivan, E. V., Rawles, J. M., Zipursky, R. B., & Lim, K. O. (1994). A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood. Archives of Neurology, 51, 874–887. Price, B. H., Daffner, K. R., Stowe, R. M., & Mesulam, M. M. (1990). The compartmental learning disabilities of early frontal lobe damage. Brain, 113, 1383–1393. Puce, A., Allison, T., Bentin, S., Gore, J. C., & McCarthy, G. (1998). Temporal cortex activation in humans viewing eye and mouth movements. Journal of Neuroscience, 18, 2188–2199. Pujol, J., Vendrell, P., Junque, C., Marti-Vilalta, J. L., & Capdevila, A. (1993).When does human brain development end? Evidence of corpus callosum growth up to adulthood. Annals of Neurology, 34, 71–75. Raine, A., Buchsbaum, M., & LaCasse, L. (1997). Brain abnormalities in murderers indicated by positron emission tomography. Biological Psychiatry, 42, 495–508. Raine, A., Lencz, T., Bihrle, S., LaCasse, L., & Colletti, P. (2000). Reduced prefrontal gray matter volume and reduced autonomic activity in antisocial personality disorder. Archives of General Psychiatry, 57, 119–127. Raine, A., Meloy, J. R., Bihrle, S., Stoddard, J., LaCasse, L., & Buchsbaum, M. S. (1998a). Reduced prefrontal and increased subcortical brain functioning assessed using positron emission tomography in predatory and affective murderers. Behavioral Sciences and the Law, 16, 319–332. Raine, A., Pihl, D., Stoddard, J., Bihrle, S., & Buchsbaum, M. (1998b). Prefrontal glucose deficits in murderers lacking psychosocial deprivation. Neuropsychiatry, Neuropsychology, and Behavioral Neurology, 11(1), 1–7. Reiss, A. L., Abrams, M. T., Singer, H. S., Ross, J. L., & Denckla, M. B. (1996). Brain development, gender and IQ in children. A volumetric imaging study. Brain, 119, 1763–1774. Semendeferi, K., Lu, A., Schenker, N., & Damasio, H. (2002). Humans and great apes share a large frontal cortex. Nature Neuroscience, 5, 272–276. Sowell, E. R., Thompson, P. M., Holmes, C. J., Batth, R., Jernigan, T. L., & Toga, A. W. (1999). Localizing age-related changes in brain structure between childhood and adolescence using statistical parametric mapping. NeuroImage, 9, 587–597. Steen, R. G., Ogg, R. J., Reddick, W. E., & Kingsley, P. B. (1997). Age-related changes in the pediatric brain: Quantitative MR evidence of maturational changes during adolescence. American Journal of Neuroradiology, 18, 819–828. Taylor, M. J., McCarthy, G., Saliba, E., & Degiovanni, E. (1999). ERP evidence of developmental changes in processing of faces. Clinical Neurophysiology, 110, 910–915. Thompson, P. M., Giedd, J. N., Woods, R. P., MacDonald, D., Evans, A. C., & Toga, A. W. (2000). Growth patterns in the developing brain detected by using continuum mechanical tensor maps. Nature, 404, 190–193.
260
DETERMINANTS OF AGGRESSION
Tremblay, R. E. (2003). Why socialization fails: The case of chronic physical aggression. In B. B. Lahey, T. E. Moffitt, & A. Caspi (Eds.), Causes of conduct disorder and juvenile delinquency (pp. 182–224). New York: Guilford Press. Tremblay, R. E., Nagin, D. S., Séguin, J. R., Zoccolillo, M., Zelazo, P., Boivin, M., Pérusse, D., & Japel, C. (2004). Physical aggression during early childhood: Trajectories and predictors. Pediatrics, 114, e43–e50. Utsunomiya, H., Takano, K., Okazaki, M., & Mitsudome A. (1999). Development of the temporal lobe in infants and children: Analysis by MR-based volumetry. American Journal of Neuroradiology, 20, 717–723. Virkkunen, M., Nuutila, A., & Huusko, S. (1976). Effect of brain injury on social adaptability: Longitudinal study on frequency of criminality. Acta Psychiatrica Scandinavica, 53, 168–172. Walsh, V., & Cowey, A. (2000). Transcranial magnetic stimulation and cognitive neuroscience. Nature Reviews. Neuroscience, 1, 73–79. Watkins, K. E., Paus, T., Mangin, J.-F., Zijdenbos, A. P., Collins, D. L., Lerch, J., Worsley, K. J., Blumenthal, J., Giedd, J., Rappoport, J., & Evans, A. C. (2002). Maturation of white matter tracts during adolescence: A longitudinal MRI study. Paper presented at the 8th International Conference on Functional Mapping of the Human Brain, June 2–6, Sendai, Japan. Available on CD-Rom in NeuroImage 16(2). Woermann, F. G., Van Elst, L. T., Koepp, M. J., Free, S. L., Thompson, P. J., Trimble, M. R., & Duncan, J. S. (2000). Reduction of frontal neocortical grey matter associated with affective aggression in patients with temporal lobe epilepsy: An objective voxel by voxel analysis of automatically segmented MRI. Journal of Neurology, Neurosurgery and Psychiatry, 68, 162–169.
DETERMINANTSinOF Neuromodulators Aggression AGGRESSION
13 Neuromodulators in the Development and Expression of Inhibition and Aggression R OBERT O. P IHL and C HAWKI B ENKELFAT
Lack of behavioral control/regulation is known to predispose to aggression and substance abuse/dependency (Pihl, Peterson, & Finn, 1990). Conversely, the ability to inhibit/regulate behavior is a basic characteristic of healthy development. It is the assumption of this chapter that much of inhibition represents learned behavior. For example, a lack of inhibition is fundamental in impulsive–reactive aggression, and its frequency and form at various ages is affected by how biology and experience alter particular neuromechanisms involved in behavioral control. Tremblay et al., (1996), among others, have illustrated how aggression typically decreases with age, steadily declining from a peak around age 2. Clearly, the form of this behavior changes with age and sex, with the declining frequency likely representing the incorporation by the individual of societal restrictions and thus behavioral control. To drastically restate Rousseau, we begin not as noble savages but as savages who may or may not be “nobilized.” Thus, how the ability to inhibit develops and what mechanisms are operative is a preeminent developmental question. The purpose of this chapter is to present information about various neuromodulators that fuel brain development and functioning as it pertains to inhibition and aggression. Foremost among these are the neurotransmitters serotonin, dopamine, norepinephrine, and gamma-aminobutyric acid (GABA), as well as their interactions and the myriad of factors that modulate their activity. We first state some important caveats and then review the evidence supporting the involvement of these neurotransmitters and their interactions in 261
262
DETERMINANTS OF AGGRESSION
inhibition. Finally, we examine how these systems change with development and specifically affect the development and functioning of the prefrontal cortex and its putative “executive control.”
A FEW ESSENTIAL CAVEATS That neuromodulators affect inhibition and aggression is a given. The specification of which neuromodulators play the key roles, and how, is, however, a different matter. Generalizations and conclusions must be qualified by at least four caveats. First, although we discuss various neurotransmitters and their respective systems as separate entities, nothing in the brain functions independently. The serotonin system, for example, interacts with most other major systems affecting behaviors ranging from aggression to sleep. In fact, the conclusion that the serotonin system’s main role is to modulate the activity of other systems has much support. Further, pharmacological studies, particularly, from which many conclusions regarding how these systems operate are drawn, rarely impact one system, one set of receptors, and so forth. Thus, findings often presented as specific and clear are in fact suspect. Second, although we argue for particular functions of certain neurotransmitters, these functions are neither static nor stable. For example, GABA, well recognized as the major inhibitory neurotransmitter in the brain, has been shown to be a major stimulus for neuronal excitation in the developing brain (McCarthy, Auger, & Perrot-Sinal, 2002). It is only with brain maturation that the more traditional function of neuronal inhibition becomes dominant. Facts like this, of course, raise the general issue of plasticity. Nelson (1999) has analogized neural plasticity as a “subtle but orchestrated dance that occurs between the brain and the environment” (p. 42). The brain shapes how one reacts to the environment, which in turn shapes how the brain affects anatomical, neurochemical, and metabolic change. As Nelson notes, these changes can be both adaptive and maladaptive for the organism, and for our purposes this phenomenon greatly complicates discussion of generic neurochemical systems. Third, consideration of sex differences in the functioning of neurochemicals is rare, but is important. Such differences are especially relevant, as a wealth of studies have reported significant differences between men and women in levels and types of aggression (see Archer & Lloyd, 2002). Further, laboratory studies of alcohol-related aggression typically display a much greater effect for men than women (Hoaken & Pihl, 2000). How hormones may alter gene expression, and thus subsequent brain development, particularly during adolescence, may have direct ramifications for sex differences in mood disorders and aggression, as well as in personality traits like introversion and sensation seeking. The frontal and temporal corticies, as well as the amygdala and hippocampus in particular, reflect significant postpubertal
Neuromodulators in Aggression
263
development. Walker (2002) has recently pointed out the significance of this time for brain development, which is reflected in the fact that the heritability of some behaviors and disorders increase with age, particularly during adolescence. Fourth, any attempt to delineate the neurochemical precursors of inihibition and aggression implies that we have a clear definition and subsequent measures for these phenomena. We have previously examined the relationship between self-report and behavioral measures of inhibition and have found it to be low to nonexistent (Helmers, Young, & Pihl, 1995). Similarly, definitions of aggression proliferate, as do measures, both lacking, respectively, in operationism and validity.
EVIDENCE FOR THE DIFFERENTIAL INVOLVEMENT OF NEUROTRANSMITTERS IN DISINHIBITION Serotonin The contention that altered serotonergic neurotransmission might play an important role in the pathophysiology of impulse-control disorders/aggressivity is not new. Using measurements in lumbar cerebrospinal fluid (CSF) of the main metabolite of serotonin (5-HT), 5-hydroxyindoleacetic acid (5-HIAA), investigators in the 1970s and 1980s proposed theoretical models implicating low serotonergic neurotransmission as a critical biological substrate for behavioral dyscontrol, often associated with impulsive, suicidal, and/or aggressive behaviors. These observations were fully consistent at the time with the findings that neurochemical lesions of the ascending serotonergic fibers were reported to facilitate aggressive behavior in rodents and with the subsequent discovery that tricyclic antidepressants enhanced serotonergic neurotransmission in rodents. Some researchers were then prompt to close the circle and suggest that these medications were effective in preventing suicide and attenuating aggressivity–impulsivity through their purported ability to correct for a neurochemical imbalance. This model of a deficit in serotonergic neurotransmission being primary in the causation of impulsivity and/or aggressivity, and predictive of therapeutic response to drugs shown to enhance serotonergic neurotransmission, is now obsolete. For example, not all impulsive–aggressive patients show these abnormalities and benefit from drugs enhancing serotonergic neurotransmission. Further, several drugs devoid of major effects on serotonergic neurotransmission are known to be effective in controlling aggressive behavior/impulsivity. The categorical view has gradually been replaced by a dimensional approach, that is, that low 5-HT explains in part the vulnerablity to impulsivity across various disorders. Hence, the thinking about this system has progressively evolved, which is the case for all systems described in this chapter, from an absolute deficit to a high-risk model. In other words, low serotonin neurotransmission is now thought to operate as a biological risk factor, resulting from genetic inheritance and/or early environmen-
264
DETERMINANTS OF AGGRESSION
tal insult, neither sufficient nor necessary, yet when combined with other salient genetic factors and/or environmental stressors, may play an important role in the triggering and maintenance of disorders of impulse control and aggressive behavior(s). Most indices suggest that 5-HT metabolism is a relatively stable trait. Over time (months to years), CSF concentrations of 5-HIAA are highly correlated in studies in both human and nonhuman primates. This stability of CSF 5-HIAA concentrations likely reflects genetic contributions. Evidence for the existence of a low 5-HT neurotransmission trait is manifold; 5-HT1B receptor knockout mice are more aggressive (Hen, 1996); aggressive–impulsive primates followed longitudinally in their own ecosystem for years have stable low CSF 5-HIAA over time (Higley et al., 1996), and low CSF 5-HIAA has also been found in patients who reattempted suicide within 5 years as compared with those who did not (Roy, Pickar, De Jong, & Karoum, 1989). Further, it has been reported that newborns from families with antisocial personality disorder show lower CSF 5-HIAA concentrations than those from families devoid of a history of antisocial behavior (Constantino, Morris, & Murphy, 1997), and low central 5-HT neurotransmission in probands has been associated with an increased risk for impulsive aggression in their firstdegree relatives (Coccaro, Kavoussi, Hauger, Cooper, & Ferris, 1998). We have shown that subjects at genetic risk for mood and/or impulse control disorders are more vulnerable to the effects of transient lowering of serotonergic neurotransmission by means of dietary manipulations than age- and sexmatched controls devoid of family history for depression, substance abuse, and/or violent behavior (Benkelfat, Le Marquand, & Pihl, 1994; LeMarquand, Benkelfat, Pihl, Palmour, & Young, 1999). Impulsive aggression, including impulsive suicidal behavior, has been shown to be associated with specific polymorphisms of tryptophan hydroxylase (TPH) (the rate-limiting enzyme for 5-HT synthesis) and the serotonin transporter (5-HTT) (the protein responsible for the neuronal reupake of 5-HT and responsible for the recycling of up to 85% of synaptic 5-HT) (Nielsen et al., 1998). In addition, postmortem brain studies and in vivo imaging studies have demonstrated greater 5-HT1A receptor binding in females as compared with males (Parsey et al., 2002). Accordingly, there are several reports of an inverse relationship between aggression and 5-HT1A receptors (Cleare & Bond, 2000). Although the underlying cause for this sex difference is unclear, sex steroids are known to modulate 5-HT1A binding and it has been suggested that estradiol increases 5-HT1A binding (Biegon & McEwan, 1982). Furthermore, several lines of nonhuman evidence exist that suggest a functional difference in 5-HT1A receptors between males and females. For example, 5-HT transporter knockout mice have lower 5-HT1A receptor binding, which is more extensive in females than in males (Li, Wichems, Heils, Lesch, & Murphy, 2000). Finally, in the hippocamus, 5-HT and/or 5-HIAA concentrations are higher in female rats than in males (Haleem, Kennett, & Curzon, 1990). These serotonergic differ-
Neuromodulators in Aggression
265
ences may help to explain the higher incidence of mood disorders and depression among females, being twice that of males, and the lower reported overall incidence of overt female aggression. The concept of a low 5-HT trait being a predisposing factor for abnormal behavior has generated considerable interest and lots of supporting indirect evidence. Yet this concept is still much of a working hypothesis. Most of the findings establishing a link between low 5-HT neurotransmission and impulsivity and/or aggressive behavior were obtained in psychiatric patients, collected from nonepidemiological samples, studied as adults typically once, and compared with controls. In general, no information could be inferred from this type of design as to whether the purported abnormalities are trait- or state-related. Moreover, studies of clinical samples, though essential, have limitations of their own: sample heterogeneity, small sample size, not representative, retrospective, confounded by uncontrolled random environmental events (i.e., drugs), and so on. Further, there is no consensus of the operational definition of the corresponding behavioral phenotype for the proposed trait with the clinical syndrome complex, combining symptoms of mood liability/ explosiveness-irritability, increased impulsivity, quarrelsomeness, a predisposition for binge eating and substance abuse, with unclear boundaries, and with character cluster B pathology. More germane is the consideration of involvement and interaction of 5-HT with other neurotransmitters.
Dopamine The relationship between 5-HT and dopamine (DA) is particularly relevant. In a study in which microdialysis probes were used to measure these neurotransmitters in the nucleus accumbens (NAC) and medial prefrontal cortex (MPFC) in rats during intruder-precipitated aggressive behavior, it was found that DA levels increased significantly in the NAC and 5-HT decreased 80% in the prefrontal cortex (PFC; Van Erp & Miczek, 2000). Increased DA levels in the NAC have also been linked to many motivational and motor functions, including drug taking, food-reinforced behavior, sexual behavior, foot shock, social defeat, and attention to novel environmental stimuli. This range of behaviors reflects the fact that DA ascending pathways from this area project to and innervate more than 20 different structures, thus reflecting general rather than specific functions, and, in effect, neuromodulating various areas of incentive motivation (Le Moal & Simon, 1991). In particular, this system responds to cues of incentive reward, including novelty, receiving differential afferents from many sources, including the basolateral amygdala, the extended amygdala, the hippocampus, and the prefrontal medial orbital cortex. DA’s activity is related to its activation of two families of DA receptors: the D1 and D2 subtypes. Both are linked to G-protein activation, thus their signals are generally slow and long lasting rather than fast (Nieoullon, 2002). D1 receptors appear to underlie DA’s tonic activity, whereas D2 receptors underlie
266
DETERMINANTS OF AGGRESSION
DA’s phasic activity. A tonic response implies that the neurons fire continuously throughout the duration of a stimulus, which appears to relate to DA’s ability to increase the stability of maintained representations. Stabilization involves increasing the signal-to-noise ratio of background versus evoked activity in the prefrontal cortex (Cohen, Braver, & Brown, 2002) D1 receptors appear to interact with other receptors, especially fast-signaling ones such as glutamate (Nieoullon, 2002), and have been implicated in the working memory functions of the dorsolateral PFC. Conversely, a phasic response implies that the neurons fire briefly at a particular phase of the stimulus (usually the onset) and likely relates to DA’s role as a gating signal. The DA system is responsible for indicating when new inputs should be encoded and maintained, or when currently maintained representations should be updated in response to salient, reward-predicting information. DA’s phasic response appears to involve activation of D2 receptors (Cohen et al., 2002). These receptors have been implicated in the reward circuits of the brain (Blum, Cull, Braverman, & Comings, 1996), as well as in the PFC’s regulation of the subcortical DA system. The subcortical DA system is involved in signaling reward-predicting information and, as such, is involved in motivational aspects of behavior. Thus, increases in DA transmission lead to more information being allowed to enter the “motive circuit” (ventral tegmental area–accumbens–ventral pallidum), so that this information becomes more motivationally relevant (Spear, 2000). In seeking out reward-related information, the DA system helps the organism to adapt to its present situation by facilitating the initiation of new responses and switching between alternate responses (Spoont, 1992). Serotonin and GABA appear to modulate DA’s reward-related activity. Blum et al. (1996) have proposed a “cascade theory” of the reward circuitry of the brain, implicating dopamine, serotonin, and GABA. In particular, they propose that the release of serotonin in the hypothalamus causes a release of enkephalin (an opioid peptide) in the ventral tegmental area, which, in turn, inhibits GABAergic neurons in the substantia nigra. This disinhibits DA neurons in the ventral tegmental area, allowing them to release DA in the nucleus accumbens and hippocampus. The released DA then activates D2 receptors in these regions, producing rewarding effects. Deficits in this reward pathway lead to what Blum et al. (1996) refer to as the “reward deficiency syndrome,” characterized by addictive, impulsive, and compulsive behavior. Thus, hypoactivity of the limbic DA system may be related to reduced levels of 5-HT or increased levels of GABA, whereas hyperactivity of this system should be associated with increased levels of 5-HT and decreased levels of GABA. As stated previously, reduced 5-HT has been associated with disorders of impulsivity, addiction, and aggression. Furthermore, GABA agonists have been suggested as a possible treatment for drug addiction, and aggression (Eichelman, 1990). In particular, GABA appears to dampen DA neurons via inhibitory GABAB receptors. Stimulation of GABAB receptors in the ventral tegmental area has been shown to dampen DA release in the nucleus accumbens produced by abused drugs
Neuromodulators in Aggression
267
(Cousins, Roberts, & De Wit, 2002), thereby reducing the rewarding effects of these drugs. Sex differences also exist in the regulation of dopaminergic neurotransmission. There is substantial evidence, primarily from nonhumans, that gonadal hormones modulate behavioral and neurochemical indices of activity in the striatum and nucleus accumbens. For example, female rats have higher striatal dopamine and homovanillic acid (HVA) levels than males, and aging reduces DA and HVA in males but not in females. Endogenous ovarian hormones, but not testicular hormones, modulate extracellular striatal DA concentrations in rats (Xiao & Becker, 1994). These differences may be related to the observations that female rats appear to be more sensitive than males to the toxic and reinforcing effects of psychostimulants that increase the DA system, an idea that is supported by research on gender differences in the behavioral response to cocaine, in which males and females do not differ in cocaine metabolism but female rats exhibit greater locomotor activation in response to cocaine than do male rats (Bowman et al., 1999). A contrasting view of serotonin’s interaction with dopamine has also been proposed. Spoont (1992) has suggested that 5-HT constrains DA’s reward-related activity, such that the organism’s behavioral output becomes more controlled. Thus, 5-HT’s role would be to increase the threshold by which exogenous stimuli become motivationally relevant, by decreasing the signal-to-noise ratio. Indeed, 5-HT has been shown to decrease the signal-tonoise ratio by inhibiting spontaneous cell activity to a lesser extent than specific evoked responses (Tassin, 1992). This would suggest that, through 5HT’s activity, a greater evoked response would be required in order for a stimulus to cause the cell to fire. This would ensure that only signals of sufficient strength or of sufficient relevance to the organism are able to interrupt current information flow and would prevent overshoots in the system by constraining the intensity of signals once they gain access to the system (overshoots in the system are associated with an increased magnitude of response and decreased sensitivity to cues of suppression). Supporting evidence for this view is drawn from studies showing decreases in 5-HT associated with increased locomotor activity and increased exploratory behavior signifying a release of normally inihibited operant behavior in diverse contexts, as well as increased food intake, sexual behavior, and aggression. Furthermore, behavioral evidence showing serotonergic modulation of the startle reflex is cited, with 5-HT decreasing the signal-to-noise ratio, thereby protecting the animal from interference from nonsalient environmental cues. Spoont (1992) further argues that DA normally functions to initiate new responses as well as to switch between alternative responses (a role that would appear to be related to DA’s rewardsignaling activity), which 5-HT constrains by making it more difficult for signals to gain access to the system. A similar pathway could be postulated to explain the effects of decreased serotonin on operant behavior, food intake, sexual behavior, and aggression. Furthermore, in considering the evidence that 5-HT normally constrains operant behavior and protects the organism from
268
DETERMINANTS OF AGGRESSION
interference from nonsalient environmental cues, as well as evidence that decreasing 5-HT increases the propensity to switch between alternative responses while decreasing the integrity of those responses, one could argue that these are indicators of 5-HT’s effect on PFC function. As the section“Executive Control” discusses further, 5-HT appears to enhance PFC function, although the effect may depend on the particular characteristics of the situation in which behavior is assessed.
Norepinephrine The noradrenergic system serves many different functions, all of which are known to affect behavior profoundly. One comprehensive review argued that noradrenaline (NA) centrally and indirectly affects aggression on three different levels: the hormonal system, the sympathetic nervous system, and the central nervous system, in different, but functionally synergistic ways (Haller, Makara, & Kruk, 1998). All three components react very rapidly to environmental changes, catecholamine hormones being amongst the fastest to react. Behavior, of course, may involve a very rapid response to environmental challenges and engages the whole organism. Behavior, including aggressive behavior, is also energy dependent and therefore requires the rapid mobilization of energy stores. The peripheral catecholamines are important in ensuring the “energy” of behavior, whereas central noradrenaline is activated by offensive and defensive aggression in males (Higley et al., 1992). As such, external stimuli anticipating a demanding behavior or the demanding behavior itself activates the central nervous system noradrenergic system. Thus, even if NA would have no effect on mechanisms directly involved in aggression, it would nonetheless have profound influence on the expression of aggressive behavior through its indirect effects on other mechanisms vital during social challenges such as arousal and attention, pain sensitivity, olfaction, and memory. The NA neurotransmitter system has long been implicated as a modulator of normal arousal, vigilance, and engagement with the environment. It has also been shown that agents that increase brain noradrenaline improve attention in rats (Sirvio et al., 1993). Furthermore, pain sensitivity, beyond opiodergic control, is decreased by medullar descending noradrenergic pathways, and changing pain perception has an obvious function in aggressive behavior. The NAergic system is also involved in memory/recognition processes and in learning, mainly when olfaction is also involved. In summary, various socially relevant stimuli elicit a central release of norepinephine, which in turn promotes arousal and shifts attention toward relevant stimuli, induces analgesia, and enhances olfaction, learning, and memory. Such specific changes may together significantly change an animal’s reaction to social challenges. In general, studies have demonstrated an increase in aggressive behavior when norepinephine (NE) activity surges, and a decrease in aggression when NE activity is decreased. Increased NE function is positively correlated with the number of shock-induced aggressive episodes in rodents (Stolk, Connor,
Neuromodulators in Aggression
269
Levine, & Barchas, 1974), and pharmacological agents that increase NE function also increase shock-induced fighting in rodents. Mice that have been isolated for prolonged periods of time show increased aggressive behavior when given desipramine (which increases NE), an effect that is not observed when they are pretreated with a selective NE neurotoxin (Matsumoto, Ojima, & Watanabe, 1995). Stimulating postsynaptic noradrenergic receptors in cats’ hypothalami facilitates aggression; in humans, clinical treatment with tricyclic/ monoamine oxidase inhibitor antidepressants (which increase NE function) is associated with agitation and irritability, particularly in patients with borderline personality disorder. Thus, NE seems to increase/facilitate aggression. However, there are also contradictory studies. For example, positive correlations have been reported between the CSF concentration of NE’s major metabolite, 3-methoxy-4-hydroxyphenylglycol (MHPG), and life history of aggressive events in men confined to a military forensic unit (Brown, Goodwin, Ballenger, Goyer, & Major, 1979), and to impulsivity ratings in humans (Roy et al., 1989). Yet Virkkunen, Nuutila, Goodwin, and Linnoila (1987) have found lower CSF MHPG in violent offenders and in impulsive arsonists as compared with healthy volunteers, and a study by Coccaro et al. (1998) reported a negative correlation between plasma MHPG and life history of aggression in personality-disordered patients. The most extensive review on the relationship between norepinephrine and aggression offers an explanation for this discrepancy in results by suggesting that NA’s contradictory effects on aggressive behavior depends on what kind of aggressive paradigm is used (Haller et al., 1998). According to Haller and colleagues (1998), as far as isolation-induced or social (rank associated) aggression is concerned, a reduction in NA is followed by a parallel inhibition of aggressive behavior, but NA depletion appears to increase shock-induced and chemically elicited aggression. In these latter instances, the increase in aggression after NA depletion may be caused by parallel changes in other neurotransmitter systems, such as dopamine and/or serotonin and/or alterations in reactivity to environmental challenges. Haller and colleagues (1998) also conclude that when noradrenaline is in excess (i.e., when release is stimulated or reuptake is blocked), it influences aggression biphasically, such that a slight activation increases aggression whereas a strong activation decreases it.
GABA Similarly elusive is the role of gamma-aminobutyric acid (GABA) in aggression. It is well known that GABA serves as an inhibitory neurotransmitter at many central nervous system synapses. Yet increased GABAA receptor activation by alcohol suggests that GABA transmission can facilitate aggression (Miczek, De Bold, van Erp, & Tomatzky, 1997). In addition, one study surprisingly showed that when GABA was injected into medial hypothalamic defensive rage sites in cats, response thresholds were actually lowered following drug infusion (Nakao, Tashiro, Kono, & Araki, 1979). However, interven-
270
DETERMINANTS OF AGGRESSION
tions that increase GABA receptor activation nonselectively do inhibit aggression. For example, microinjections of the GABA agonist muscimol into defensive rage sites within the dorsal periaqueductal gray matter inhibits aggression (Siegel, Roeling, Gregg, & Kruk, 1999). Similarly, aggressive mice have lower brain GABA concentrations as compared with nonaggressive mice, and agents that inhibit GABA metabolism dose-dependently decrease fighting in conjunction with increases in brain GABA (Simler, Puglisi-Allegra, & Mandel, 1982). Furthermore, both clonidine and lithium, which have anti-aggressive behavioral effects in some patients, have reportedly increased peripheral measures of GABA (Kemph, De Vane, Levin, Jarecke, & Miller, 1993). A study by Bjork et al. (2001) found that plasma GABA levels were lower in first-degree relatives of patients with major depressive disorder—who are also shown to have low plasma GABA—as compared with controls, thus indicating that plasma GABA is under genetic control. This study went on to demonstrate that plasma GABA levels and self-reported aggressiveness correlated negatively, but that this was more specific to subjects with a family history of depression.
DEVELOPMENT At each of the developmental stages, from genes to old age, the functioning and density of these neurotransmitters critical to behavioral control are operative. For example, regarding dopamine, a recent review (Di Maio, Grizenko, & Joober, 2003) has detailed the involvement of dopamine genes in attentiondeficit/hyperactivity disorder (ADHD). These authors marshal considerable evidence to support the strong involvement of the dopamine transporter gene (SLC3A6) and the dopamine receptor 4 gene (DRD4) in ADHD. A mouse with the dopamine transporter gene knockout has been developed, with the resultant animals displaying spontaneously increased activity and higher reactions to stress, which are significantly calmed by the administration of stimulants. This review further details how extra repeat copies of this gene have been reported in diagnosed ADHD populations. A similar repeat polymorphism has been reported for the alleles of the DRD4 gene, with particular focus on the 7 repeat allele. In a recent meta-analysis (Faraone, Doyle, Mick, & Biederman, 2001) it was reported that an odds ratio of 1.9 existed between the repeat allele and ADHD for family-based studies, suggesting a small but significantly increased risk. Another study of DA receptor density in 16 mice strains found threefold differences for the nucleus accumbens and a sevenfold difference among strains for the caudate-putamen between strains, demonstrating remarkable genetic variability (Jones, Hou, Cook, & Melloni, 1996). Excess dopamine prenatally, such as occurs when a mother is cocaine dependent, likely results in abnormalities, which studies suggest affect the differentiation of neurotransmitter systems, cell function, migration, and neuronal growth. Cocaine levels in the fetal brains of rats have been found to be greater than the 100% of those in the maternal blood. As shown in longitudinal stud-
Neuromodulators in Aggression
271
ies, infants born to cocaine-using mothers typically display poor behavior regulation, habituation, and orientation (Chasonoff et al., 1998). At age 6 this is seen in difficulty in managing impulses, frustration, and arousal. Studies of monkeys have shown that this prenatal exposure results in reduced expression of the rate-controlling enzyme in dopamine synthesis, as well as an elevation in dopamine receptors, particularly in frontal brain areas (Ronnekleiv, Fang, Choi, & Chai, 1998). Similar behavioral profiles to that produced by prenatal cocaine insult have been reported for many other drugs, with alcohol predominating. Prenatal alcohol exposure affects the functioning of many neurotransmitters and neuromodulators. Recent work focusing on serotonin has shown that alcohol fed to pregnant mice resulted in a 20–30% reduction in serotonin neurons (Zhou, Sari, Zhang, Goodlett, & Li, 2001) and a decrease in the density of serotonin fibers, particularly in the forebrains of the offspring (Sari, Powrozek, & Zhou, 2001). These effects likely permanently decrease serotonergic functioning in the mature animal, thus affecting behavioral regulation. The effect of prenatal malnutrition on the developing brain is another example of altered neurochemical function that has been extensively reviewed (Galler & Tonkiss, 1998). Depending on the time, type, and severity of malnutrition, the basic cytoarchitectine of the brain can be affected. Protein deficiencies in particular are significant, as proteins are precursors of enzymes, peptide hormones, and neurotransmitters. Morgane et al. (1993) have demonstrated in rats long-term inhibition in certain hippocampal GABA interneuron systems following prenatal protein malnutrition, which affects a myriad of other systems and is possibly responsible for many attentional and learning problems. In terms of neonates, rearing rats in isolation produces a behavioral syndrome that includes hyperactivity and an increased response to reinforcers, response perseveration, decreased pain thresholds, and a heightened response to dopamine agonists. Further, basal levels of dopamine are higher, and a down regulation of D2 receptors in the nucleus accumbens is affected (Hall et al., 1998). In the turbulent years of adolescence there is a concomitant shift in dopaminergic activity from subcortical to cortical areas. As one further ages, so does neurotransmitter functioning. In a PET study of individuals ages 24– 86 correlations were conducted between the density of D2 receptors and neuropsychological test performance. It was found that D2 receptor availability in the caudate and putamen declined with age and was correlated with decreased performance on a motor task, a measure of abstraction, and a response inhibition task (Volkow et al., 1998). As the preceding examples illustrate, throughout childhood, adolescence, and even into old age, changes occur in the circuitry of the brain that likely account for age-related cognitive and behavioral changes. For example, in adolescence there is a massive loss of synapses in the neocortex, with the most intense period of pruning occurring between 7 and 16 years of age. Although this pruning occurs for many different kinds of receptors, the majority of syn-
272
DETERMINANTS OF AGGRESSION
apses lost are those that are excitatory in nature. Such pruning occurs throughout the brain, including the PFC and limbic regions. During adolescence there is also an increase in DA fiber density in the PFC, which may be compensated for later in adolescence by a decline in DA synthesis and/or turnover in this region. Developmental changes also occur in the GABAergic system. Studies in rats have shown increased responsiveness of the GABAB system and of the GABA/benzodiazepine receptor complex from adolescence to adulthood, but a decrease in responsiveness of the cortical GABAA system. Based on studies of knockout mice, it has been suggested that dysfunctional GABAA receptors lead to increased emotional reactivity (Mohler, Fritschy, Benke, Benson, & Luscher, 1996). In light of GABAB’s previously mentioned involvement in the reward circuit of the brain (in particular, dampening DA release produced by drugs), it could be inferred that this gradual increase in responsiveness of the GABAB system may reduce one’s susceptibility to the rewarding effects of certain drugs. Thus, it may be postulated that from childhood into adulthood, there is a decrease in the responsiveness to cues of reward. With respect to the serotonergic system, 5-HT binding appears to be lower during adolescence than during childhood or adulthood (Spear, 2000). Considering the myriad studies linking low serotonin levels to disorders of behavioral control and impulsivity, this lowered serotonin binding during adolescence may account for the impulsive behavior often seen at this developmental period. As concerns the relative balance between subcortical and cortical DA systems, there appears to be a shift toward greater predominance of cortical DA during early adolescence. With a peak in cortical DA activity at this time, a concomitant loss in excitatory drive to the cortex may result. Both rat and human studies have found a decrease of one-third to one-half in dopamine receptors in the striatum during adolescence (Seeman, Bzowej, Guan, & Bergeron, 1987). Spear (2000) has postulated that this adolescent-associated loss in excitatory drive to the cortex may create a mini-“reward–deficiency syndrome.” Thus, too little DA stimulation caused by overactivity of the PFC may lead to a state of underarousal. It is further thought that this underarousal may account for the increase in impulsive behavior often seen during adolescence. In addition, these mesocortical DA projections are seen as more sensitive to activation by stressors than are mesolimbic DA systems. Consequently, stress may exacerbate this adolescent-associated mini-reward– deficiency syndrome by increasing the extent to which cortical DA inhibits excitatory drive to the cortex. Adolescence is also a time when there is a shift toward greater efficiency of processing information. MRI studies have shown an inverse relationship between cortical grey matter loss and brain growth in the frontal cortex of adolescents, suggesting that the synaptic losses may be concomitant with increased myelination. Thus, maturation of the frontal cortex seems to involve a shift toward fewer and faster connections, a change that presumably increases the speed and efficiency of processing. These changes involve use-dependent
Neuromodulators in Aggression
273
patterns of connectivity, implying that experience plays an important role in the development of the PFC.
EXECUTIVE CONTROL Russian neuropsychologist Luria (1980) conceptualized the operation of the PFC more wholistically and realistically than the current focus on specific functions and their localization. Conceptually, he saw working memory, attentional control, verbal organization, and so forth, operating collectively to ascribe meaning to stimuli, in effect creating the context for possible responses. In this view, the frontal lobes are seen as responsible for planning, controlling, and verifying behavior in the presence of goals, working within a context and providing control over the more automatic subcortical systems. As reviewed earlier, the major neurotransmitters are integral to the development and functioning of the PFC. Dopamine is highly concentrated in the prefrontal cortex, in fact, more concentrated than in any other cortical region. One study showed that as infant monkeys improved their performance on a delayed response task requiring inhibition, the level of dopamine in the prefrontal cortex increased, as did the density of dopamine receptors (Brown & Goldman, 1977). Further, lesioning of the dorsolateral prefrontal cortex interferes greatly with performance on delayed response tasks, as does blocking dopamine receptors in these areas. On these tasks delay of responding is critical, which is why this area has been related to sustained attention and working memory. What is also noteworthy is that the pattern of responses in lesioned and pharmacologically manipulated animals is not random. Specifically, what occurs is a perseveration of the previous response. On tasks requiring memory with inhibition, little is known about dopaminergic functioning in children between the ages of 3 and 6. However, Diamond (2000) has been studying children who have been treated early and continuously for phenylketonuria, and who thus have reduced levels of dopamine in the prefrontal cortex, but have otherwise normal brains. These children typically reflect cognitive deficits and score in a similar IQ range to children with prefrontal damage (Diamond, 2000). Treated phenylketonuric children have a reduced level of the dopamine precursor tyrosine in the brain, which is particularly significant for the prefrontal cortex because it has a higher rate of dopamine turnover than any other brain area. The nature of the deficits in these treated children involve deficits in working memory and in inhibitory control. These deficits are seen even when IQ, gender, and other variables are controlled and are evident in all age groups: infants, toddlers, and young children. It is also notable that it is specifically deficits in working memory and inhibitory tasks, rather than other cognitive measures, that reflect the determined deficiencies. Cognitive and behavioral inhibition have long been linked to DA activity in the dorsolateral and ventrolateral PFC. Braver et al. (2001) have sug-
274
DETERMINANTS OF AGGRESSION
gested that one of the key components in the ability to exert control over thoughts and actions is context processing, using subsets of representations within working memory that govern how other representations are used. Thus, in order to understand behavioral control and aggression, it is helpful to examine the mechanisms underlying working memory. Within the PFC, DA enhances working memory by modulating excitatory neurotransmission. D1 receptors, being colocalized with glutamate receptors on the spines of pyramidal neurons in the PFC, appear to enhance working memory by inhibiting excitatory input by activating the neuron. Thus, too little DA may lead to cognitive dysfunction through excessive stimulation, whereas too much DA may lead to deficits through insufficient stimulation (GoldmanRakic, 1999). This would be consistent with the theory of schizophrenia, which postulates that hypodopaminergic activity in the frontal cortex (along with hyperdopaminergic activity in the limbic system) leads to overstimulation and disorganized thought. This is also consistent with Arnsten’s (2000) finding that, in the PFC, DA has an inverted U-shaped dose–response curve, whereby either very high or very low levels of D1 stimulation impair cognitive function. The activity of the subcortical DA system appears to be partially under the control of the PFC. Deutch (1992) has proposed that DA tonically inhibits corticostriatal projection neurons, such that glutamate’s release in the striatal complex, in turn, inhibits DA from being released in the subcortical site. This hypothesis implies that diminishing the activity of DA in the PFC results in increased subcortical DA being released. Indeed, levels of DA activity in the PFC appear to be inversely related to DA release in subcortical regions (Spear, 2000). For example, it has been proposed that in children with ADHD, reduced activity in the PFC may lead to low tonic DA activity in limbic regions. This low tonic stimulation of DA autoreceptors may lead to high phasic activity in the nucleus accumbens and other subcortical sites. Thus, reduced PFC activity leads to low tonic DA activity, which leads to hyperresponsiveness of the DA system to reward-related cues. This model would account for the beneficial effects of stimulants in treating ADHD. DA’s inhibitory action may be mediated through direct activation of D2 receptors, or may occur indirectly by facilitating GABA release from interneurons. D2 receptors have been implicated in a number of addictive, impulsive, and compulsive disorders (Blum et al., 1996), and drugs that enhance GABAergic function have been used in clinical settings to control aggressive behavior. GABA and 5-HT have also been implicated in working memory in the PFC. GABAergic interneurons stabilize representations by inhibiting pyramidal neurons with opposite “best directions” (Goldman-Rakic, 1999). Thus, by inhibiting contradictory signals, the GABAergic system provides feedforward inhibition, thereby stabilizing maintained representations and preventing distractors from disturbing the system. In fact, postmortem and other biological studies have supported the suggestion of a defective GABAergic
Neuromodulators in Aggression
275
system in the frontal cortex of patients with schizophrenia (Blum & Mann, 2002). With respect to the 5-HT system, Williams, Srinivas, & GoldmanRakic, 2002) have found that in rhesus monkeys, 5-HT enhanced PFC function via activation of the 5-HT2A receptor, which facilitates both pyramidal cells and interneurons. In particular, direct facilitation of pyramidal cells, as well as indirect feedforward inhibition of pyramidal cells via GABA interneurons, provides tonic facilitation of the neuronal network (increased signal-to-noise ratio). However, in light of behavioral findings linking 5-HT to deficits in PFC function (Sasaki-Adams & Kelley, 2001), 5-HT’s facilitory effects may apply to only very constrained situations. Thus, 5-HT, by increasing the signal-to-noise ratio, may improve PFC function when the organism is highly motivated and when only task-relevant stimuli are present in the environment. However, in more realistic situations where there are many possible behaviors in response to a wide variety of stimuli, 5-HT’s facilitory effect may impair PFC function by increasing the amount of “noise” entering the system. NE has also been implicated in the functioning of the PFC. It appears to exert its influence through activation of alpha-1, alpha-2, and beta-noradrenergic receptors, all of which have been located in the PFC, and may also act through D4 receptors. Studies with rats, humans, and primates have indicated opposing effects of alpha-1 and alpha-2 receptor stimulation. Activation of alpha-2A subtype appears to enhance PFC functioning, as indexed by improvements in working memory tasks, response inhibition, and planning (Arnsten, 2000). Conversely, activation of alpha-1 receptors appears to impair performance on tasks subserved by the PFC. Furthermore, there appears to be an optimal level of NE, as NE exhibits higher affinity for alpha-2 than for alpha-1 adrenoceptors. Thus, moderate levels of NE (for example, as induced by stress) engage alpha-1 receptors and impair PFC function (Arnsten, 2000). Furthermore, there are NE–DA interactions, for example, cortical alpha-1 adrenergic and D1 receptors may be located on the same cells, and “the message induced by the stimulation of cortical D1 receptors is inhibited by alpha-1 adrenergic receptors” (Tassin, 1992, p. 145). NE’s functioning in the PFC also seems to be consistent with clinical research. Administration of beta-adrenergic receptor antagonists appears to attenuate aggressiveness in organically impaired children and adults, as well as in schizophrenic patients prone to violent outbursts. Moreover, lithium (which describes the functional availability of NE) reduces aggressiveness in personality-disordered adults and conduct-disordered children. Furthermore, one study found a positive correlation between CSF levels of the NE metabolite (MHPG, 3-methoxy-4-hydroxyphenylglycol) and a life history of aggressive behavior (Eichelman, 1990). Thus, in individuals with unusually high levels of NE, PFC function may be impaired by increased activation of alpha-1 adrenoceptors, which may lead to impaired behavioral control. In these populations, decreasing available NE seems to reduce aggressiveness, perhaps by reducing activation of alpha-1 receptors.
276
DETERMINANTS OF AGGRESSION
CONCLUSION The development of behavioral regulation is essential to the control of impulsivity and most forms of aggressive behavior. This development requires the normative functioning of a wide range of neurochemical activities that affect and are affected by genetic variation and by pre- and postnatal experience. These interactions shape the brain and its functioning in that orchestrated dance between brain and environment. Although many neuromodulators are active participants, featured performers are serotonin, dopamine, norepinephrine, and GABA. Low serotonergic neurotransmission has been repeatedly linked to impulsive–aggressive behaviors. The mechanisms of the linkage, however, remain at issue, with enhanced vulnerability most likely associated with serotonin’s interaction with dopamine—specifically, hypoactivity at the limbic level, an involvement not unlike that of GABA’s. With reduced serotonin and GABA activity at this level weak, even irrelevant signals produce system activation. The inhibiting potential of the prefrontal cortex can limit this overactivation, again, through neurochemical functioning. Prefrontal serotonin levels, which appear reduced during adolescence, seem germane specifically to situations where highly motivated and task-relevant stimuli are present. Dopamine’s role, which is initiated when limibically active, appears inhibiting when active in the prefrontal cortex. This brain area basic to the control of behavior is a structure that develops through adolescence, is programmed by experience but fueled by the interaction of basic brain neuromodulators. Finally, an unmentioned but most significant caveat is relevant. That is, with the remarkable development of new technologies and daily flurry of scientific study, “all of the above is subject to change.”
REFERENCES Archer, J., & Lloyd, B. (2002). Sex and gender. London: Cambridge University Press. Arnsten, A. F. T. (2000). Genetics of childhood disorders: XVIII. ADHD, Part 2: Norepinephrine has a critical modulatory influence on prefrontal cortical function. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 1201– 1203. Benkelfat, C., LeMarquand, D., & Pihl, R. O. (1994). Serotonin and alcohol intake, abuse, and dependence: Clinical evidence. Biological Psychiatry, 1(36), 327–337. Biegon, A., & McEwan, B. S. (1982). Modulation by estradiol of serotonin receptors in brain. Journal of Neuroscience, 2, 199–205. Bjork, J. M., Moeller, F. G., Kramer, G. L., Kram, M., Suris, A., Rush, A. J., & Petty, F. (2001). Plasma GABA levels correlate with aggressiveness in relatives of patients with unipolar depressive disorder. Psychiatry Research, 101, 131–136. Blum, B. P., & Mann, J. J. (2002). The GABAergic system in schizophrenia. International Journal of Neuropsychopharmacology, 5(2), 159–179. Blum, K., Cull, J. G., Braverman, E. R., & Comings, D. E. (1996). Reward deficiency syndrome. American Scientist, 84, 132–145.
Neuromodulators in Aggression
277
Bowman, B., Vaughan, S., Walker, Q., Davis, S., Little, P., Scheffler, N., Thomas, B., & Kuhn, C. (1999). Effects of gender and gonadectomy on cocaine metabolism in the rat. Journal of Pharmacology and Experimental Therapeutics, 297, 114–120. Braver, T. S., Barch, D. M., Keys, B. A., Cohen, J., Taylor, S. F., Carter, C. S., Kaye, J. A., Janowsky, J. S., Yesavage, J. A., & Mumenthaler, M. S. (2001). Context processing in older adults: Evidence for a theory relating cognitive control to neurobiology in healthy aging. Journal of Experimental Psychology: General, 130(4), 746–763. Brown, G. L., Goodwin, F. K., Ballenger, J. C., Goyer, P. F., & Major, L. F. (1979). Aggression in humans correlates with cerebrospinal fluid amine metabolites. Psychiatry Research, 1, 131–139. Brown, R. M., & Goldman, P. S. (1977). Catecholamines in neocortex of rhesus monkeys: Regional distribution and ontogenetic development. Brain Research, 168, 133–150. Chasnoff, I., Anson, A., Hatcher, R., Stenson, H., Iankea, K., & Randolph, L. (1998). Prenatal exposure to cocaine and other drugs. Annals of the New York Academy of Science, 846, 314–328. Cleare, A. J., & Bond, A. J. (2000). Ipsapirone challenge in aggressive men shows an inverse correlation between 5-HT1A receptor function and aggression. Psychopharmacology, 148, 344–349. Coccaro, E. F., Kavoussi, R. J., Hauger, R. L., Cooper, T. B., & Ferris, C. F. (1998). Cerebrospinal fluid vasopressin levels correlate with aggression and serotonin function in personality disordered subjects. Archives of General Psychiatry, 55, 708– 714. Cohen, J. D., Braver, T. S., & Brown, J. W. (2002). Computational perspectives on dopamine function in prefrontal cortex. Current Opinion in Neurobiology, 12, 223– 229. Constantino, J. N., Morris, J. A., & Murphy, D. L. (1997). CSF 5-HIAA and family history of antisocial personality disorder in newborns. American Journal of Psychiatry, 154, 1771–1773. Cousins, M. S., Roberts, D. C. S., & De Wit, H. (2002). GABAB receptor agonists for the treatment of drug addiction: A review of recent findings. Drug and Alcohol Dependence, 65(3), 209–220. Deutch, A. Y. (1992). The regulation of subcortical dopamine systems by the prefrontal cortex: Interactions of central dopamine systems and the pathogenesis of schizophrenia. Journal of Neural Transmission, 36, 61–89. Diamond, A. (2000). Close interrelation of motor development and cognitive development and of the cerebellum and prefrontal cortex. Child Development, 71, 44–56. Di Maio, S., Grizenko, N., & Joober, R. (2003). Dopamine genes and attention-deficit hyperactivity disorder: A review. Journal of Psychiatry and Neuroscience, 28, 27– 38. Eichelman, B. S. (1990). Neurochemical and psychopharmacologic aspects of aggressive behavior. Annual Review of Medicine, 41, 149–158. Faraone, S. V., Doyle, A. E., Mick, E., & Biederman, J. (2001). Meta-analysis of the association between the 7–repeat allele of the dopamine D-sub-4 receptor gene and attention deficit hyperactivity disorder. American Journal of Psychiatry, 158, 1052– 1057. Galler, J. R., & Tonkiss, J. (1998). The effects of prenatal protein malnutrition and cocaine on the development of the rat. Annals of the New York Academy of Sciences, 846, 29–39.
278
DETERMINANTS OF AGGRESSION
Goldman-Rakic, P. S. (1999). The “psychic” neuron of the cerebral cortex. Annals of the New York Academy of Sciences, 868, 13–26. Haleem, D. J., Kennett, G. A., & Curzon, G. (1990). Hippocampal 5–hydroxytryptamine synthesis is greater in female rats than in males and more decreased by the 5-HT1A agonist 8-OH-DPAT. Journal of Neural Transmission, 79, 93–101. Hall, F., Wilkinson, L., Humby, T., Inglis, W., Kendall, D., Marsden, C., & Robbins, T. (1998). Isolation rearing in rats: Pre- and postsynaptic changes in striatal dopaminergic systems. Pharmacology, Biochemistry and Behavior, 59, 859–872. Haller, J., Makara, G., & Kruk, M. (1998). Catecholaminergic involvement in the control of aggression: Hormones, the peripheral sympathetic, and central noradrenergic systems. Neuroscience and Biobehavioral Reviews, 22(1), 85–97. Helmers, K., Young, S., & Pihl, R. O. (1995). Assessment of measures of impulsivity in healthy male volunteers. Personality and Individual Differences, 19, 927–935. Hen, R. (1996). Mean Genes, Neuron, 16, 17–21. Higley, J., Mehlman, P., Taub, D., Higley, S., Suomi, S., & Linnoila, M. (1996). Stability of interindividual differences in serotonin function and its relationship to severe aggression and competent social behavior in rhesus macaque females. Neurospsychopharmacology, 7, 67–76. Higley, J., Mehlman, P., Taub, D., Higley, S., Suomi, S., Vickers, J., & Linnoila, M. (1992). Cerebrospinal fluid monoamine and adrenal correlates of aggression in free-ranging rhesus monkeys. Archives of General Psychiatry, 49, 436–441. Hoaken, P. N. S., & Pihl, R. O. (2000). The effects of alcohol intoxication on aggressive responses in men and women. Alcohol and Alcoholism, 35, 471–477. Jones, B., Hou, X., Cook, M., & Melloni, N. (1996). Effect of exposure to novelty on brain monoamines in C57BL/6 and DBA/2 mice. Physiology and Behavior, 59, 361–367. Kemph, J. P., DeVane, C. L., Levin, G. M., Jarecke, R., & Miller, R. L. (1993). Treatment of aggressive children with clonidine: Results of an open pilot study. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 577–581. LeMarquand, D., Benkelfat, C., Pihl, R., Palmour. R., & Young, S. (1999). Behavioral disinhibition induced by tryptophan depletion in nonalcoholic young men with multigenerational family histories of paternal alcoholism. American Journal of Psychiatry, 15, 1771–1779. Le Moal, M., & Simon, H. (199l). Mesocorticolimbic dopaminergic network: Functional and regulatory roles. Physiological Review, 71, 155–234. Li, Q., Wichems, C., Heils, A., Lesch, K. P., & Murphy, D. L. (2000). Reduction in the density and expression, but not G-protein coupling, of serotonin receptors (5HT1A) in 5-HT transporter knock-out mice: Gender and brain region differences. Journal of Neuroscience, 20, 7888–7895. Luria, A. R. (1980). Higher cortical functions in man. Moscow: Moscow University Press. Matsumoto, K., Ojima, K., & Watanabe, H. (1995). Noradrenergic denervation attenuates desipramine enhancement of aggressive behavior in isolated mice. Pharmacology, Biochemistry and Behavior, 50, 481–484. McCarthy, M., Auger, A., & Perrot-Sinal, T. (2002). Getting excited about GABA and sex differences in the brain. Trends in Neuroscience, 25, 307–312. Miczek, K. A., DeBold, J. F., Van Erp, A. M., & Tornatzky, W. (1997). Alcohol, GABA-abenzodiazepine receptor complex, and aggression. Recent Developments in Alcohol, 13, 139–171.
Neuromodulators in Aggression
279
Mohler, H., Fritschy, J. M., Benke, D., Benson, J., & Luscher, B. (1996). Genetics and function of GABAA-receptor subtypes. European Neuropsychopharmacology, 6, S4. Morgane, P. J., Austin-LaFrance, R., Bronzino, J. D., Tonkiss, J., Diaz-Cintra, S., & Cintra, L., et al. (1993). Prenatal malnutrition and development of the brain. Neuroscience and Biobehavioral Reviews, 17(1), 91–128. Nakao, H., Tashiro, N., Kono, R., & Araki, R. (1979). Effects of GABA and glycine on aggressive-defense reaction produced by electrical stimulation of the ventromedial hypothalamus in cats. In M. Itoh (Ed.), Integrative control functions of the brain (pp. 332–334). Amsterdam: Elsevier. Nelson, C. (1999). Neural plasticity and human development. Current Directions in Psychological Science, 8, 42–45. Nielsen, D. A., Virkkunen, M., Lappalainen, J., Eggert, M., Brown, G. L., Long, J. C., Goldman, D., & Linnoila, M. (1998). A tryptophan hydroxylase gene marker for suicidality and alcoholism. Archives of General Psychiatry, 55(7), 593–602. Nieoullon, A. (2002). Dopamine and the regulation of cognition and attention. Progress in Neurobiology, 571, 1–31. Parsey, R. V., Oquendo, M. A., Simpson, N. R., Ogden, R. T., Van Heertum, R., Arango, V., & Mann, J. J. (2002). Effects of sex, age, and aggressive traits in man on brain serotonin 5-HT1A receptor binding potential measured by PET using [C-11]WAY100635. Brain Research, 954, 173–182. Pihl, R. O., Peterson, J., & Finn, P. (1990). The inherited predisposition to alcoholism: Characteristics of sons of male alcoholics. Journal of Abnormal Psychology, 99, 291–301. Ronnekleiv, O., Fang, Y., Choi, W., & Chai, L. (1998). Changes in the midbrain-rostral forebrain dopamine circuitry in the cocaine-exposed primate fetal brain. Annals of the New York Academy of Sciences, 846, 165–181. Roy, A., Pickar, D., De Jong, J., & Karoum, F. (1989). Suicidal behavior in depression: Relationship to noradrenergic function. Biological Psychiatry, 25, 341–350. Sari, Y., Powrozek, T., & Zhou, F. C. (2001). Alcohol deters the outgrowth of serotonergic neurons at midgestation. Journal of Biomedical Science, 8, 119–125. Sasaki-Adams, D., & Kelley, A. (2001). Serotonin-dopamine interactions in the control of conditioned reinforcement and motor behavior. Neuropsychopharmacoloogy, 25, 440–452. Seeman, P., Bzowej, N., Guan, H., & Bergeron, C. (1987). Human brain D-sub-1 and Dsub-2 dopamine receptors in schizophrenia, Alzheimer’s, Parkinson’s, and Huntington’s diseases. Neuropsychopharmacology, 1, 5–15. Siegel, A., Roeling, T., Gregg, T., & Kruk, M. (1999). Neuropharmacology of brain-stimulation-evoked aggression. Neuroscience and Biobehavioral Reviews, 23, 359– 389. Simler, S., Puglisi-Allegra, S., & Maudel, P. (1982). Gamma-aminobutyric acid in brain areas of isolated aggressive or non-aggressive inbred strains of mice. Pharmacology, Biochemistry and Behavior, 16, 57–61. Sirvio, J., Jakala, P., Mazurkiewicz, M., Haapalinna, A., Riekkinen, P., Jr., & Riekkinen, P. J. (1993). Dose- and parameter-dependent effects of atipamezole, an alpha 2–antagonist, on the performance of rats in a five-choice serial reaction time task. Pharmacology, Biochemistry and Behavior, 45, 123–129. Spear, L. P. (2000). The adolescent brain and age-related behavioral manifestations. Neuroscience and Biobehavioral Reviews, 24, 417–463.
280
DETERMINANTS OF AGGRESSION
Spoont, M. R. (1992). Modulatory role of serotonin in neural information processing: Implications for human psychopathology. Psychological Bulletin, 112, 330–350. Stolk, J. M., Connor, R. L., Levine, S., & Barchas, J. D. (1974). Brain norepinephrine metabolism and shock-induced fighting behavior in rats. Journal of Pharmacology and Experimental Therapeutics, 190, 193–209. Tassin, J. P. (1992). NE/DA interactions in prefrontal cortex and their possible roles as neuromodulators in schizophrenia. Journal of Neural Transmission, 36, 135–162. Tremblay, R., Boulerice, B., Harden, P., McDuff, P., Pérusse, D., Pihl, R. O., & Zoccolillo, M. (1996). Do children in Canada become more aggressive as they approach adolescence? Growing up in Canada: The National Longitudinal Survey of Children (pp. 127–136). Ottawa: Statistics Canada. Van Erp, A. M. M., & Miczek, K. A. (2000). Aggressive behavior, increased accumbal dopamine, and decreased cortical serotonin in rats. Journal of Neuroscience, 20(24), 9320–9325. Virkkunen, M., Nuutila, A., Goodwin, F. K., & Linnoila, M. (1987). Cerebrospinal fluid monoamine metabolite levels in male arsonists. Archives of General Psychiatry, 44, 241–247. Volkow, N., Gur, R., Wang, G., Fowler, J., Moberg, P., Ding, Y., Hitzemann, R., Smith, G., & Logan, J. (1998). Association between decline in brain dopamine activity with age and cognitive and motor impairment in healthy individuals. American Journal of Psychiatry, 155, 344–349. Walker, E. (2002). Adolescent neurodevelopmental and psychopathology. Current Directions in Psychological Science, 11, 24–28. Williams, G., Srinivas, G., & Goldman-Rakic, P. (2002). The physiological role of 5HT2A receptors in working memory. Journal of Neuroscience, 22, 2843–2854. Xiao, L., & Becker, J. B. (1994). Quantitative microdialysis determination of extracellular striatal dopamine concentration in male and female rats: Effects of estrous cycle and gonadectomy. Neuroscience Letters, 180(2), 155–158. Zhou, F. C., Sari, Y., Zhang, J. K., Goodlett, C. R., & Li, T. K. (2001). Prenatal alcohol exposure retards the migration and development of serotonin neurons in fetal C57BL mice. Brain Research Developmental Brain Research, 126, 147–155.
DETERMINANTS Hormones and Aggression OF AGGRESSION
14 Hormones and the Developmental Origins of Aggression S TEPHANIE H. M. V AN G OOZEN
In considering biological influences on the development of behavior, in this case aggression, one should (and this cannot be emphasized enough) always see the behavior as the outcome of a complex interplay of individual, developmental, and social factors. Some of these factors might play a more or less important role as a function of the age of a child. For example, with respect to social factors, harsh parental discipline with cruel punishment plays a causal role in the development of antisocial behavior in childhood, whereas the absence of parental supervision is an important factor in antisocial behavior in late childhood or adolescence (Lahey, Waldman, & McBurnett, 1999). To give an example of a more individual factor, the early (perinatal) hormonal environment is assumed to play a role in shaping temperament and its development before the onset of important social influences such as peer interactions; at puberty, when pronounced hormonal changes occur, these individual factors may again become more relevant. When we consider young children, it is most likely that the origin of aggressive behavior lies in a combination of a toddler/child with a difficult temperament and a nonoptimal environment in which ineffective socialization plays a key role: A difficult child elicits harsh, inconsistent, and negative socialization behaviors, as a result of which a difficult temperament ultimately develops into aggressive or disruptive behavior. Although there are factors that contribute to antisocial behavior in childhood becoming chronic, not all antisocial children become antisocial adolescents, and not all antisocial adolescents become antisocial adults. A study of the individual factors in the development of aggression could help explain why these behavioral patterns change over time. 281
282
DETERMINANTS OF AGGRESSION
In this chapter the role of two hormones—testosterone and cortisol—in the development and maintenance of aggressive and/or antisocial behavior are discussed, namely the androgen testosterone and the stress hormone cortisol, the primary reason for their selection being that most research on human aggression has focused on these two hormones. Before considering the evidence of hormonal factors in aggression, it is important to keep in mind that aggression can occur in both normal and pathological contexts. In the first case we mean normal individual variations in aggressive behavior, whereas in the second case aggression occurs more frequently and in more serious forms and is usually more persistent. In the latter case we are dealing with evidence from psychopathological studies. Because there have been very few studies addressing the influence of hormones on the normal development of aggression, evidence in support of the influence of hormones on both normal variations in aggression and more serious and persistent aggression are discussed. Another complex issue concerns the behavior of interest, aggression, and, in particular, its multiple causations. Many aggressive, antagonistic, or hostile acts are elicited by events, for instance, frustrating or otherwise aversive events. Yet aggression results from the individual’s disposition to react aggressively to such events. Such a disposition can be conceptualized in various ways, and problems have always existed with the predictive validity of instruments that measure such traits. One reason for such problems may be that much aggressive behavior results from emotional responses and that individuals differ in their proneness to emotional responding of particular kinds. It is not difficult to find examples from everyday life showing that aggression and emotion are intimately connected phenomena. Aggression can result from fear, hate, or even love. Still, the most common emotion related to the occurrence of aggressive behavior is anger. Whether a certain situation leads to anger, and perhaps to aggressive behavior, depends to an important extent on individual differences, such as personality characteristics, and these can be partly influenced by hormonal (organizing or activating) factors. Moreover, such individual differences can be of a short-term (temporary mood changes in irritability and quick-temperedness) or long-term character. It is clear that items in most aggression questionnaires, whether the focus is on adults or children, measure the occurrence of overt aggressive acts without reference to relevant antecedents, such as emotions and situational cues. Although there is important research examining the relationship between aggression and anger (Berkowitz, 1989), little has been done to study possible relationships between aggression and individuals’ proneness to anger, and as far as I know, nobody has studied hormonal influences on anger proneness in young children. A final comment concerns the type of aggression under investigation. An emotional or reactive form of aggression (with high levels of emotional arousal) is commonly distinguished from a more instrumental, proactive, or calculated (low emotional arousal) form (Vitiello & Stoff, 1997). Within these categories aggression can be of a more direct or indirect nature, it can be phys-
Hormones and Aggression
283
ical or verbal, and other subcategories can also be distinguished. It is beyond the scope of this chapter to discuss these differences, but it should be clear that highly emotional and impulsive aggression and cold-blooded, calculated aggression have different neuropsychological and/or neurobiological underpinnings. For present purposes it is important to keep in mind that studies investigating the relationship between hormones and aggression have rarely addressed this issue, focusing usually on one type of aggressive behavior in particular.
TESTOSTERONE Why Is Testosterone an Interesting Hormone? Studies of aggressive behavior in animals indicate that the exchange of antagonistic and hostile acts is predominately a male attribute, exhibited between males. Criminological studies reveal a similar clear-cut sex difference: Men use far more physical violence and are more often the victims of aggression. Biological explanations have linked these findings to the social dominance of the male and to testosterone, the most important male sex hormone. Whereas many animal studies have demonstrated that the presence of this hormone is a prerequisite for the display of aggressive behavior, results of such studies in primates (Dixson, 1980; Eberhart, Keverne, & Meller, 1980) and humans are less clear-cut and at best suggestive of a positive association (Albert, Walsh, & Jonik, 1993; Archer, 1991; Archer, Graham-Kevan, & Davies, in press; Harris, 1999; Mazur & Booth, 1998). However, the interest in the aggression– testosterone relationship remains a strong one and can most probably be explained by the extensive literature on gender differences in aggressive behavior. If one investigates behaviors that clearly show differences between men and women, it is clear from a biopsychological perspective that the role of gonadal hormones is an important one on which to reflect. Indeed, gonadal hormones have been shown to be a relevant factor in the origin and expression of sex differences in behavior: The so-called “organizing” and “activating” effects have been well established in studies on neuroendocrine functioning, in anatomical studies, and in studies of behavior (Goy & McEwen, 1980). The development of internal and external genitalia, of sex differences in figure and appearance, and of various classes of sex-specific behaviors, such as the occurrence of sexual behavior or aggression (but by no means only these!) are to an important extent related to the presence or absence of testosterone during particular prenatal and postnatal developmental phases (Collaer & Hines, 1995). In adulthood both males and females are exposed to different classes and patterns of gonadal hormone stimulation, which, again, affect brain and behavior. However, understanding the origin of sex differences in behavior is complicated by the fact that such differences are not only manifestations of the organizing and activating effects of hormones; they are also the result of
284
DETERMINANTS OF AGGRESSION
interactions between these hormonal influences on behavior, on one hand, and social influences on the individual during various phases of development, on the other.
Organizing and Activating Hormonal Effects The long-term developmental and structural effects of pre- and postnatal exposure to sex hormones on sex-dimorphic behavior have been termed “organizing” effects. These developmental effects are distinguishable from “activating” effects of gonadal hormones in that activating effects are reversible, repeatable, and not limited to a critical phase of development. It is, however, generally assumed that organizing, and activating hormone effects interact, and the distinction is therefore not as easy to make as the definition of the terms suggests. Activating hormone effects presumably act in concert with perinatal hormonal and early developmental experiential factors to perpetuate sex-dependent and individual behavioral characteristics (Van de Poll & Van Goozen, 1992). Organizing effects are usually tested by manipulating the exposure to sex hormones during fetal development, whereas activating effects are investigated at the time of puberty or adulthood. For obvious reasons, experimental research on these effects has focused almost exclusively on animals, rodents in particular. Androgenic gonadal hormones have been shown to be involved in various aspects of fighting behavior, including “intermale” aggression (Moyer, 1976). Studies of female rodents also demonstrate that the female, when exposed to testosterone prenatally or in adulthood, may exhibit high levels of aggressive behavior equal to those of males (Van de Poll, Van Zanten, & DeJonge, 1986). Although animal studies can teach us a great deal about the role of testosterone in behavior, and in aggression in particular, extrapolating the results to humans and implicitly postulating similar mechanisms in humans may present us with an unduly simplified and misleading picture of human functioning. The important issue to be addressed next is whether individual differences in human aggression have anything to do with differences in hormone levels or hormonal functioning.
Organizing Effects: Sex Differences in Aggression Evidence concerning the development of aggression is obviously important for the study of sex differences in aggression. One can assume that socializing processes start early, so that if one wants to research gender differences, these have to be studied at an early age. Boys and girls differ in play styles and toy preference: Boys are more energetic and active and show more aggressive and destructive behavior, involving so-called rough-and-tumble play, a form of play-aggression that has also been observed to be sex-dependent in monkeys (Hines & Green, 1991). Some of these sex differences have been reported in
Hormones and Aggression
285
children as early as 12 months of age. Recently, Hines et al. (2002), in a large population-based study, found a relationship between testosterone levels, as measured in the blood samples of pregnant women, and gender role behavior in 3.5 year old girls (n = 337), such that as maternal testosterone levels increased, so did masculine-typical gender role behavior. No relationships between hormones and gender role behavior were found in boys (n = 342). The largest gender differences, however, are manifested in the years that follow, and this is a strong argument for the influence of socialization, modeling, and norms and values. A fair amount of evidence also exists for gender differences in aggressive feelings and behaviors during adolescence. Most of these associations involve the androgens and, as expected, larger effect sizes have been reported for boys than for girls. Several studies found evidence of a mediating role for impatience and impulse control in the association between hormonal levels and aggression (Susman et al., 1987; Warren & Brooks-Gunn, 1989). Olweus, Mattson, Schalling, and Low (1988) reported a causal effect of pubertal hormones on aggressive behavior via increased impatience and irritability. Hormonal influences on the development of the human brain and behavior can be investigated in individuals who have been exposed to unusual gonadal hormone environments, pre- or neonatally, due to genetic disorders or because their mothers were prescribed hormones during pregnancy. Reinisch (1981) concluded on the basis of her study that the development of aggressive behavior in humans is partly determined by prenatal androgens. Aggression proneness in children was measured with a response hierarchy test, consisting of a series of drawings picturing daily situations (e.g., a child takes away your toy) together with a number of possible reactions (hitting, shouting, go to parents, or crying); the number of aggressive reactions was the aggression score. Reinisch found a clear sex difference in (physical) aggression between boys and girls. Moreover, the administration of progestins to the mother during pregnancy (i.e., progestin being a synthetic sex hormone with androgenic action, prescribed to avoid the risk of possible miscarriage) had an aggressionenhancing influence on exposed boys and girls. Animal research on mice has demonstrated that sex differences in aggression can be counteracted by exposing female mice to testosterone or by castrating male mice postnatally. Later studies of similar effects in girls with congenital adrenal hyperplasia (CAH: an autosomal recessive disorder associated with an enzyme defect in the pathway of cortisol and aldosterone biosynthesis), who had been exposed prenatally to elevated levels of androgens, have not shown clear effects on aggressive behavior, although different aspects of gender role behavior were affected. These girls were more energetic, played with male peers and with boys’ toys more, and were identified by themselves and others as tomboys (Ehrhardt & MeyerBahlburg, 1981; Berenbaum & Resnick, 1997; Berenbaum, Duck, & Bryk, 2000). These studies of the prenatal effects of androgens on aggression used
286
DETERMINANTS OF AGGRESSION
small samples and different research methods. For example, Reinisch (1981) measured aggression proneness as a more or less stable way of reacting to a number of standard situations. In studies in which no effects were found, a more or less standardized interview was often conducted with the children, their parents, and sometimes also their teachers. The differences in findings may be due to the different research methods used. Gender differences in aggression seem to be found most often when the subjects are offered response alternatives, for instance, a choice between verbal versus physical aggression. Reviewing the results of these studies of humans exposed prenatally to excessive androgens, one can conclude that certain aspects of gender-related behavior are influenced by prenatal hormones. The influence of these hormones is primarily on “tomboyism” or over-energetic and playful behavior. In girls, interest in maternity may be affected, but many androgenized women who are able to have children do become mothers. Clear-cut differences in gender identity and sexual orientation have not been found (Ehrhardt & Meyer-Bahlburg, 1981; Ehrhardt et al., 1985; Gooren, Fliers, & Courtney, 1990), although androgenized females sometimes show a delay in exhibiting normative teenage sexual experience and motivation, possibly because these girls look “different” and are less likely than their normal counterparts to develop a typically female body shape and may be masculinized if not treated by estrogens. Thus, whereas clinical studies with humans suggest organizational effects of excessive androgens on gender role behavior (specifically, oppositesex relationships and activity level), they do not appear to have a clear influence on aggression or other gender-typed behavior such as sexual behavior or sexual orientation (Paikoff & Brooks-Gunn, 1990). In sum, there are not many studies that have addressed the issue of whether prenatal androgen levels influence human aggression. The limited support for hormonal effects on aggression comes from studies with good experimental designs, good selection of control groups, and standardized measures of aggression (Berenbaum & Resnick, 1997). The literature reviewed earlier in this section suggests that there is no good evidence of a link between androgens and aggression in children, apart from Reinisch (1981). The recent study by Hines et al. (2002) and the research on girls with CAH were cited as evidence of effects of androgens on energy levels, play activity, and gender role behavior, not aggression.
Activating Effects of Testosterone on Aggression in Adolescents and Adults Correlations in men between androgens and self-reports of aggressive behavior or scores on aggression inventories are generally low (Meyer-Bahlburg, Boon, Sharma, & Edwards, 1974). For example, in one study a correlation of 0.28 was found between total testosterone and scores on a physical aggression subscale; the correlation with free testosterone was even lower, at 0.14 (Gladue, 1991). A study relevant to the issue of whether pubertal increases in
Hormones and Aggression
287
testosterone levels produce an increase in physical aggression is that by Halpern, Udry, Campbell, and Suchindran (1994). They found no evidence of either an increase in aggression or an association between testosterone and aggression in a 3-year longitudinal study of 100 teenage boys (ages 12–13 at study entry) going through puberty. Recently, in a high-quality multitrait, multimethod, and multisampling study, Granger et al. (2003) also found no relationships between (saliva) testosterone and aggression, delinquency, or symptoms of disruptive behavior disorders in a relatively large sample of adolescent boys (n = 106). However, present or past aggressive behavior has sometimes been found to correlate positively with testosterone concentrations. For instance, Dabbs, Frady, Carr, and Besch (1987) measured saliva testosterone in 89 male prison inmates. They found that testosterone concentration was related to the type of crime: Men convicted for violent crimes (murder, rape, robbery) had higher testosterone concentrations than men convicted for theft or burglary. Most of the studies on activating effects have been correlational in design, thereby limiting the possibility of establishing a causal role for testosterone in eliciting aggression. A more compelling line of research is one that examines the effects of testosterone administration or depletion on aggression. Finkelstein et al. (1997) investigated the effects of testosterone administration on aggression in hormone-deficient (delayed puberty) adolescents. The data demonstrated significant testosterone effects, specifically on physical aggression and aggressive impulses in boys (and similar effects were found after estrogen administration in teenage girls). O’Connor, Archer, Hair, and Wu (2002) studied the effects of exogenous testosterone on aggression and mood in 30 eugonadal and 8 hypogonadal men. No significant increases in a range of aggression measures and no testosterone-related mood effects were observed in the eugonadal men. However, significant positive mood effects were found in the hypogonadal group. Another group of human subjects that is interesting from the standpoint of the study of activating effects of testosterone on aggression are transsexuals. In the process of their sex reassignment procedure, these patients are treated with high doses of so-called opposite sex hormones. In two studies it was found that when female-to-male transsexuals received high doses of androgens, their anger and aggression proneness increased, whereas the reverse happened in male-to-female transsexuals receiving anti-androgens and estrogens (Van Goozen, Frijda, Van de Poll, 1994; Van Goozen, Cohen-Kettenis, Gooren, Frijda, & Van de Poll, 1995). It is nevertheless clear that the evidence of a link between androgens and aggression is much weaker in humans than in animals. The results seem to be clearer when the relation is studied in younger age groups, as shown in studies of adolescents (Olweus et al., 1988; Schaal, Tremblay, Soussignan, & Sussman, 1996). This could be because social restrictions on aggression are still limited at a younger age and behavior has yet to be shaped. One therefore wonders what evidence there is for a relation between testosterone and aggression in prepubertal children.
288
DETERMINANTS OF AGGRESSION
Androgens and the Development of Aggression in Young, Prepubertal Children Testosterone Given that there has been little research on testosterone and aggression in older age groups, it does not come as a surprise that there is even less information on young children. Until very recently it was not easy to measure testosterone levels in children. This is because prepubertal plasma levels are very low,1 and if one wants to sample testosterone in saliva, which is the least complicated way, only the unbound fraction of testosterone is available. To my knowledge, only four studies have measured testosterone levels (two in plasma and two in saliva) in preschool or schoolaged children and related these levels to aggression. Constantino et al. (1993) measured testosterone in plasma of 18 highly aggressive boys and 11 controls, all children being between 4 and 10 years old, and found neither a difference in testosterone between the groups nor a relationship between testosterone and the children’s Child Behavior Checklist (CBCL; Achenbach, 1991) aggression scores. The authors concluded that these findings raise questions about inferences from adult studies suggesting that testosterone may play a causal role in the development of aggression. However, an explanation for the negative results may lie in the composition of the groups, because 10 out of 18 aggressive boys had a comorbid diagnosis of psychotic disorder. Psychosis has been considered an exclusion criterion (Kruesi et al., 1990), because aggression as a temporary symptom of a psychotic episode is clearly different from aggression as observed in children with a psychiatric diagnosis of oppositional–defiant disorder (ODD) (American Psychiatric Association, 1994) or conduct disorder (CD) (American Psychiatric Association, 1994), let alone in normal, healthy children. Moreover, the age range in the subgroups was large (between 4 and 10 years), and the investigators had to divide the groups into smaller subgroups (i.e., an adrenarchal subgroup consisting of 6 psychotic–aggressive, 4 nonpsychotic–aggressive, and 6 normal controls; and a preadrenarchal subgroup of 4 psychotic–aggressive, 4 nonpsychotic–aggressive, and 5 normal controls). Van Goozen, Matthys, Cohen-Kettenis, Thijssen, and Van Engeland (1998b) also found no difference in plasma testosterone between boys with CD (n = 15) and normal controls (n = 25; all participants were between 8 and 12 years old), and there were no significant relationships between testosterone and CBCL ratings of aggression and delinquency. A third study focused on normal, healthy 4-year-olds and measured testosterone in saliva (SánchezMartín et al., 2000).2 No difference in testosterone was observed between boys (n = 28) and girls (n = 20), but testosterone in boys was positively correlated with observations of more serious aggression in social interaction. No such relations were observed for more playful aggression, and no relations between testosterone and behavior were observed in girls. Finally, Scerbo and Kolko (1994) measured saliva testosterone in forty 7–14-year-old clinic-
Hormones and Aggression
289
referred disruptive children and found a significant positive relationship between testosterone and staff-rated aggression, after taking age and physical characteristics into account, but no differences in mean levels for children with or without CD, ODD, or ADHD.
DHEA/DHEAS From about age 6, children exhibit a gradual increase in androgens of adrenal origin, a period called the adrenarche, and it is not until puberty that gonadal androgens, such as testosterone, become more important. Research in prepubertal children should therefore not only focus on testosterone, but should also consider adrenal androgens, such as dehydroepiandrosterone (DHEA), its sulphate (DHEAS), and androstenedione. Of these, DHEA and DHEAS are particularly interesting because they are also endogenously synthesized by the brain and act as neurosteroids (Robel & Baulieu, 1995), they increase neuronal excitability, enhance neuronal plasticity, and have neuroprotective properties (Wolf & Kirschbaum, 1999) and, finally, they appear to have potent antagonistic effects on central gamma-aminobutyric acid (GABAA) mechanisms, which are involved in aggressiveness (Majewska, 1992). One would therefore expect to find elevated DHEA/DHEAS levels in individuals who are aggressive. Until now, very few studies have focused on the role of adrenal androgens in children. Interestingly, a pattern of DHEA hyposecretion has been found in children and adolescents with major depression (Goodyer et al., 1996). So far, three studies have examined DHEAS levels in prepubertal aggressive boys. The previously mentioned study by Constantino et al. (1993) found no relationship with aggression, but, as noted earlier, there are some grounds for treating this null finding with caution. Van Goozen et al. (1998b) found that DHEAS levels were significantly higher in boys with CD than in normal controls and were strongly related to aggression and delinquency scores; differences between the two groups in androstenedione and testosterone were, respectively, marginally significant and not significant. However, no data on developmental status were collected and no psychiatric control group was included. One could speculate that if higher adrenal androgen levels reflect adrenal response to stress, and if chronic stress stimulates adrenal development and secretion (Sapolsky, 1997), it is possible that elevated DHEAS levels would also be found in other psychiatric groups as a result of stress associated with the psychiatric disorders. The aim of a subsequent study (Van Goozen et al., 2000b) was to investigate DHEAS levels in a new group of prepubertal children with ODD and normal and psychiatric controls, including a group of children with attention-deficit/hyperactivity disorder (ADHD). DHEAS levels were examined in relation to developmental status, psychiatric diagnosis, and intensity of aggression and delinquency during the preceding 6 months as rated by parents/care workers. The results again showed that the ODD group
290
DETERMINANTS OF AGGRESSION
had significantly higher levels of DHEAS than both the normal and psychiatric control groups. At present it is not clear how best to explain these results. It may be that the DHEAS levels of the children with ODD were higher from early childhood onward; alternatively, it may be that these children had higher levels only since the beginning of adrenarche. However, by measuring the development of the secondary sex characteristics and finding that the different groups of children were similar with respect to physical development, Van Goozen and colleagues could rule out the possibility that the children with ODD were physically more mature than their counterparts in the two control groups. The positive relation observed between ODD and/or CD, on one hand, and plasma levels of DHEAS, on the other, is of interest. As mentioned earlier, higher endogeneous levels of this neuroexcitatory steroid may contribute to more aggression because of its GABAA antagonistic actions (Majewska, 1992). It is also possible that higher DHEAS levels contribute to a larger pool of endogeneous testosterone. However, in our own study measuring different androgenic hormones (i.e., testosterone, DHEAS, and androstenedione) in children with CD, we found that the relation between DHEAS and aggressive behavioral problems is stronger than that between testosterone and such problems, suggesting a more important role of CNS excitatory mechanisms in aggressive behavior of children (Van Goozen et al., 1998b). These results demonstrate that adrenal androgen functioning is an important subject for future research on the developmental origin of aggression in children.
Methodological Considerations A large part of human aggression research has been questionnaire-based, and it is worthwhile to reflect on the potential for measuring the effects of hormones on aggression by means of this methodology. If one accepts the evidence that hormonal effects on aggression are rather subtle, and admits that the assessment of the type of behavior under consideration is complicated, not only because there are many different ways of expression, but also because one is dealing with a sensitive (and rather negatively valued) issue, it follows that asking people to report on their own aggression or asking informants for their observations about other people’s aggression is fraught with difficulties. The evidence shows that the assessment of the frequency and intensity of children’s aggression is, to a large extent, dependent on whether one asks the parents, the teachers, or the children themselves, and that there are considerable cross-informant discrepancies (see, e.g., Loeber, Green, Lahey, & StouthamerLoeber, 1991). If the assessment of the target behavior is so variable, how can one expect to find a reliable relationship with hormonal parameters? It is safe to conclude that questionnaire studies of the relationship between testosterone levels in humans and aggressive disposition have generated a conflicting body of evidence and that, on the whole, directly observational research or studies that treat differences in hormonal condition as an inde-
Hormones and Aggression
291
pendent variable (e.g., Finkelstein et al., 1997; Reinisch, 1981; SánchezMartín et al., 2000; Van Goozen, Frijda, Wiegant, Endert, & Van de Poll, 1996) have been better able to reveal the rather subtle hormonal effects on overt angry and aggressive behavior.3 Given that there are very few data on children’s aggression, it is our belief that future studies on the development of aggression in children, in focusing on the involvement of hormonal parameters, should rely on such methods to explore this issue further.
CORTISOL A second hormonal factor contributing to individual differences in antisocial behavior is cortisol. Cortisol is related to the experience of stress, and it seems likely, on the basis of observing overt aggressive behavior, that the stress system is involved in displaying aggression. Specifically, one could hypothesize that people who engage in aggression are, relatively speaking, unafraid of the negative consequences of their actions or, stated differently, are less sensitive to punishment (Matthys, Van Goozen, De Vries, CohenKettenis, & Van Engeland, 1998; Van Goozen et al., 2004). One important biological stress system is the hypothalamic–pituitary–adrenal (HPA) axis, which plays an important role in the regulation of both physical and psychological stress. It has been proposed that a reduced activity of the HPA system is part of the biological predisposition to antisocial behavior (Van Goozen, Matthys, Cohen-Kettenis, Buitelaar, & Van Engeland, 2000a). Low activity or arousal has been linked to fearlessness, sensation seeking, and/or risk-taking behavior (Raine, 1996). Specifically, one can predict that individuals who are less sensitive to stress (i.e., stress hyporesponsive) are more likely to engage in aggression because they do not fear the negative consequences of their actions because of a lack of fear conditioning. A lack of fear of punishment would also reduce the effectiveness of social conditioning and could therefore explain why this pattern of behavior, especially when observed in disruptive and antisocial clinical cases, can be so persistent and hard to treat.
The Stress System There are clear indications that stress plays an important role in the explanation of individual differences in antisocial behavior, and in this connection it is useful to draw on concepts such as stress sensitivity, stress regulation, and coping with stress. The activity of the neuroendocrine HPA system is central to stress resistance and can be assessed by using measures of its end products, the glucocorticoids. The primary glucocorticoid in humans is cortisol, and this hormone can be relatively easily assessed in saliva (Dettling, Gunnar, & Donzella, 1999). A well-functioning HPA system is necessary for survival, but hyperactivity of this system may have deleterious effects on immune system
292
DETERMINANTS OF AGGRESSION
activity and brain areas involved in memory and attentional processes and may influence the expression of defensive behaviors (Gunnar, Tout, de Haan, Pierce, & Stansbury, 1997). Thus, healthy adaptation involves efficient regulation of HPA system activity. Generally speaking, the expression of an adrenocortical stress response is dependent on (1) whether the situation is perceived as potentially threatening, (2) whether the outcome is important to the individual, and (3) whether the individual expects to have the resources necessary to manage the threat (Gunnar et al., 1997). Factors such as control, sense of success and efficacy, and predictability are therefore very important in the activity of the HPA system. The starting point of research on the relationship between stress and antisocial or aggressive behavior is the hypothesis that aggressive individuals are less fearful or sensitive to stress. This can be deduced from the fact that these individuals are less inhibited and more prone to engage in risky, stressful, or dangerous situations than other people are. If this is true, there are two possible explanations for a relationship between a lower stress sensitivity and aggressive or antisocial behavior. One theory claims that antisocial or aggressive individuals have low levels of fear (Raine, 1996). A relative lack of fear or a less inhibited temperament would lead to aggression or antisocial behavior because the person is less sensitive to the negative consequences of his or her own or other people’s behavior in general, and the receipt of punishment in particular, and because the person overestimates his or her own coping strategies. If this is the case, the implication for the treatment of serious aggression or antisocial behavior is clear. Thus, aggressive individuals, including children, have problems in conditioning, so trying to point out the negative consequences of behavior, or punishing unacceptable behavior, is likely to have little or no effect. A different stress theory focuses on sensation seeking (Zuckerman, 1991). Here it is argued that a certain level of stress is needed in order to feel pleasant and that too little or too much stress is experienced as unpleasant. Aggressive individuals are supposed to have an elevated threshold for stress. They are easily bored and are not put off by situations that most people find too exciting, stressful, or dangerous.
Human Studies on Cortisol and Aggression What evidence is there that a dysfunctional stress system plays a role in aggressive behavior? Several studies have found that antisocial and aggressive adults have lower levels of cortisol. There are also indications of inverse relationships: The more aggressive the behavior, or the more serious the antisocial behavior disorder, the lower the biological stress levels (King, Jones, Scheurer, Curtis, & Zarcone, 1990; Virkkunen, 1985; Woodman, Hinton, & O’Neill, 1978). This type of research has also been conducted on children with antisocial behavior, albeit less frequently. Although the results are less consistent
Hormones and Aggression
293
than those from the adult studies, the predicted relations have been found. For example, Tennes and Kreye (1985) and Tennes, Kreye, Avitabe, and Wells (1986) found a relationship between lower cortisol levels in urine and the intensity of aggressive behavior as expressed toward peers or a teacher. In other studies researchers have found a negative relation between conduct disorder severity and cortisol (McBurnett, Lahey, Rathouz, & Loeber, 2000; Pajer, Gardner, Rubin, Perel, & Neal, 2001; Van Goozen et al., 1998a; Van Goozen et al., 2000a; Vanyukov et al., 1990). However, the findings for children and adolescents are not as unequivocal as those for antisocial adults, in that positive relationships between cortisol and antisocial behavior have also been found in some studies. In normal, healthy adolescents a positive relationship was found between aggression and cortisol response level during an experimentally induced aggression task (Gerra et al., 1997) and during an emotion-arousing and painful procedure (Susman, Dorn, Inoff-Germain, Nottelmann, & Chrousos, 1997). Moreover, McBurnett et al. (1991) reported high levels of cortisol in conduct-disordered (CD) children, but only when they had a comorbid anxiety disorder. Finally, some studies have found no relationship between cortisol and antisocial or aggressive behavior (Klimes-Dougan, Hastings, Granger, Usher, & Zahn-Waxler, 2001; Kruesi, Schmidt, Donnelly, Hibbs, & Hamburger, 1989; Scerbo & Kolko, 1994; Schulz, Halperin, Newcorn, Sharma, & Gabriel, 1997; Stoff et al., 1992; Susman et al., 1999). Apart from the direct relationship between stress and antisocial or aggressive behavior, cortisol also seems to play a part in the prediction of antisocial behavior in the long term (McBurnett et al., 2000). At present is not known whether a reverse relationship also exists, that is, that aggressive children with an elevated sensitivity to stress are more fearful and therefore have a lower risk of persisting in their antisocial behavior. Such data already exist for autonomic nervous system measures (Brennan et al., 1997). An important future line of research is therefore to establish whether the development of aggression over time is related to young children’s stress sensitivity or level of fearfulness, in the sense that children with elevated HPA axis activity levels are biologically protected against acting aggressively on a frequent and/or more serious basis.
Methodological Considerations The mixed findings in regard to cortisol and aggression in children and adolescents may be due to important methodological differences between these studies. First, the label “aggression” or “antisocial” has been used for behaviors as different as physical aggression, running away from home, stealing, and drug use (Coie & Dodge, 1998; Tremblay, 2000, 2003). Furthermore, studies that specifically assess physical aggression have generally not taken into account whether it is of a reactive or proactive type. Reactive
294
DETERMINANTS OF AGGRESSION
and proactive aggression have been observed in children, adolescents, and adults (Brendgen, Vitaro, Tremblay, & Lavoie, 2001; Dodge, Lochman, Harnish, Bates, & Pettit, 1997). Reactive aggression is impulsive, often accompanied by disinhibition and affective instability, but not necessarily by antisocial tendencies; it is characterized by high levels of bodily arousal. Proactive aggression, however, is nonimpulsive and controlled and occurs in the context of persistent antisocial behavior. Proactively aggressive individuals are less likely to have unstable affect, their aggression is goal directed, and the level of arousal is usually low (Vitiello & Stoff, 1997). It seems theoretically plausible that proactive aggression is likely to be associated with low levels of cortisol, whereas reactive aggression is accompanied by elevated levels of cortisol. Differences in assessments of cortisol may also explain the diversity of results. Some studies measured cortisol under resting conditions (Kruesi et al., 1989; McBurnett et al., 2000; Tennes & Kreye, 1985), and other studies measured cortisol before the occurrence of an anticipated stressful event (Dawes et al., 1999; Moss, Vanyukov, & Martin, 1995), during an aggression-provoking task (Gerra et al., 1997), or under highly stressful conditions (Van Goozen et al., 1998a, 2000a). Moreover, most studies on children’s cortisol involved measurements varying over the day, without controlling for the clear circadian rhythm in cortisol secretion or for the kinds of activity in which participants had been engaged before the start of the study (McBurnett et al., 2000); only a few studies measured cortisol at specific time points, having kept the conditions constant for all participants (Susman et al., 1997; Van Goozen et al., 1998a, 2000a). Finally, some studies used only clinical samples (Scerbo & Kolko, 1994; McBurnett et al., 2000), some compared clinical samples with nonclinical samples (Kruesi et al., 1989; Pajer et al., 2001; Stoff et al., 1992; Van Goozen et al., 1998a, 2000a), and others used community samples (Gerra et al., 1997; Klimes-Dougan et al., 2001). To summarize, research on aggressive child psychiatric patients (i.e., children with a psychiatric diagnosis of CD or ODD) demonstrates that these children are characterized by a lower sensitivity to stress, as reflected by a lower secretion of cortisol during experimental conditions involving stress exposure (Van Goozen et al., 2000a). These data support the fearlessness theory of aggression as described by Raine (1996). The situation is less clear with respect to the issue of whether persistently aggressive children also have lower resting levels of cortisol. Furthermore, normal variations in HPA system (re-) activity seem to play a clear role in temperamental differences between children, with higher activity being linked to a shy, inhibited, and anxious temperament and low activity with a more impulsive and aggressive temperament (Gunnar et al., 1997; Kagan, Resnick, & Snidman, 1988; Susman, Schmeelk, Ponirakis, & Gariepy, 2001). However, there has been hardly any research directly studying the relationship between cortisol and the occurrence and development of aggression. This is clearly an important issue for future research.
Hormones and Aggression
295
The Role of Early Experiences We do not currently know why aggressively acting individuals have lower HPA-axis activity levels, although it is, of course, possible that genetic factors play a role (Bakshi & Kalin, 2000). However, there is also evidence that stressful events—by which is meant serious stressors such as neglect, abuse, and trauma—in the first years of life play an important role in “programming” the HPA axis. The evidence comes mainly from animal studies, but it seems likely that the neurobiological consequences of the severe stress that can be manipulated in animal studies also occur in humans. Interactions between a person’s biological makeup, on one hand, and the environment in which that person is raised, on the other, determine his or her physical and psychological development. Physical and biological problems during important phases in pre- or postnatal development (e.g., birth complications, intense stress or illnesses during the mother’s pregnancy), together with aversive early psychosocial experiences (such as malnutrition, poverty, neglect, and abuse) contribute importantly to the development of personality and psychopathology (Gottlieb & Halpern, 2002). There is increasing evidence that such interactions between biological and environmental factors affect the developing brain both prenatally and postnatally, including the functioning of the HPA axis (Francis, Caldji, Champagne, Plotsky, & Meaney, 1999; Kaufman, Plotsky, Nemeroff, & Charney, 2000; Liu et al., 1997) as well as the serotonergic system (see Suomi, Chapter 4, this volume). Although preclinical studies offer important suggestions regarding the way in which early experiences can influence the development of specific biological systems that are involved in aggressive behavior, the conclusions are based on animal data and it is obvious that such experimental studies cannot be done on humans. There is, however, some preliminary evidence that similar processes can occur in humans. In a study conducted in an orphanage in Romania, Carlson and Earls (1996) measured the cortisol levels of children who had been severely neglected. They found that these children failed to exhibit a normal diurnal cortisol rhythm. In another phase of the study they exposed some of these children to an intervention. In this intervention the children lived in smaller groups, and they were positively stimulated by their caretakers and were given toys. After 13 months their cortisol levels were measured again in order to establish whether they exhibited a normal diurnal pattern. This was not the case. Such a finding is consistent with the view that a long and intense period of neglect had a permanent effect on cortisol functioning, although it is, of course, also possible that the intervention was too short or too weak to allow the children to recover normal functioning. Although there is presently not much solid evidence that processes similar to those observed in animals also occur in humans, it is known that abuse and neglect are serious psychosocial stressors, that such experiences can occur at a
296
DETERMINANTS OF AGGRESSION
young age, and that a high proportion of children with aggressive behavioral problems have had such experiences.
FUTURE RESEARCH ON ADRENAL STEROIDS Testosterone has been found, albeit inconsistently, to play a role in aggression, but levels are very low in prepubertal children and it is therefore difficult to assess its role in the development of this behavior from an early age. This chapter reviewed the evidence that children with aggressive or antisocial behavior have increased plasma levels of the adrenal androgen DHEAS and either a decreased basal level of another adrenal steroid, cortisol, or a decreased cortisol reactivity to stress. It is not unusual to observe these steroids changing in opposite directions. Adrenocorticotropic hormone (ACTH) stimulates the secretion of cortisol, DHEA, and DHEAS, but there are many instances of dissociation between the secretion of cortisol and adrenal androgens, indicating that they are governed by separate regulatory mechanisms and that other factors may modulate their biosynthesis. Moreover, although both are secreted by the adrenal cortex, they have different sites of origin. Clarke, Fearon, Cunningham, and McKenna (1996), in their research on the hypothalamic– pituitary–adrenal (HPA) axis, demonstrated that beta-endorphins (the levels of which rise at adrenarche at the same time as the secretion of DHEA/ DHEAS increases) and joining peptide stimulated adrenal androgen production but inhibited ACTH-stimulated cortisol production. One could infer from these results that a dissociation between cortisol and adrenal androgen secretion might be found in people with antisocial behavior, but it is also possible that other regulatory mechanisms will be uncovered in future research. Clearly, adrenal androgens in young children should be studied more extensively, and preferably in combination rather than separately. The observation of an association between increased DHEAS levels, decreased cortisol levels, and a current diagnosis of ODD or CD in children (Van Goozen et al., 1998a, 1998b, 2000a, 2000b) suggests that it would be worthwhile to study the mechanisms underlying adrenal steroids alterations in childhood. If future studies replicate the findings of an aberrant secretion of adrenal steroids (i.e., an excess of DHEAS and a deficit of cortisol reactivity), this would confirm the idea that such a hormonal profile contributes to aggressive or antisocial tendencies. Such a hormonal line of research on the developmental origins of aggression should focus on children as young as possible, preferably in the prenatal period, and should extend from the immediate postnatal period into early and late childhood. The implication is that hormonal samples would be taken frequently and controlled for various internal (e.g., diurnal rhythm) and external (e.g., life events) influencing factors. Not only is such a line of research timeconsuming and expensive, it is also difficult to carry out for practical reasons. Large sample sizes are required, and these have to be followed up over time.
Hormones and Aggression
297
Moreover, as far as the prenatal part of such a project is concerned, recent evidence from my own research group suggests that it is not steroid levels measured in the umbilical cord or maternal serum (assessed over the three trimesters of pregnancy) that provide the best source of information if one wants to assess androgen exposure during fetal development (Van de Beek, Thijssen, Cohen-Kettenis, Van Goozen, & Buitelaar, 2004). Rather, it is the analysis of the amniotic fluid that provides the most valid assessment of fetal androgen exposure during a very sensitive period of development (about 16–18 weeks into pregnancy). Specifically, it was found that the mother’s own androgen levels were unrelated to the androgen exposure of her child as reflected in the amniotic fluid. If one accepts these results, this poses severe restrictions on conducting this type of research, because it is not easy to collect amniotic fluid simply for scientific purposes. Nevertheless, if one wants to investigate hormonal influences on the very early development of temperament in general, and aggression in particular, this seems to be the way to go about it. The picture is somewhat different for cortisol. An adequate physiological stress response, characterized by a rapid increase and subsequent decrease in stress hormones, is essential in adapting to and coping with stress. As discussed earlier, variations in both basal activity and stress responses have been associated with the existence of problem behavior and the development of psychopathology. Long-term changes in stress hormonal functioning play a key role not only in the pathophysiology of aggressive behavior but also in depression (Goodyer et al., 1996). Adequate physiological stress responses are also important in infants. Neonates (4–7 days old) who were alert during a physical examination and well orientated had relatively higher baseline cortisol values and weaker cortisol responses (Spangler & Scheubeck, 1993). Insecure–avoidant and disorganized 1-year-old children showed a clear increase in cortisol during the Strange Situation Test (SST) as compared with securely attached children, who showed an adequate behavioral response and no increase in cortisol (Spangler & Grossmann, 1993). These examples illustrate the relationship between physiological and behavioral responses to stress. However, the precise relationship between behavioral and physiological stress responses is complex and dependent on age and developmental phase, type of stressor, and the state the child is in. Given that physiological stress responses play an important role in normal and deviant forms of adaptation, it is important to establish what determines individual differences in reactivity. Animal research has shown that prenatal and early postnatal stress results in long-term changes in the regulation of the offspring’s stress hormonal (HPA) axis (Francis, Diorio, Plotsky, & Meaney, 2002; Francis, Szegda, Campbell, Martin, & Insel, 2003). Thus, it can be assumed that individual differences in maternal hormones and emotions can have a significant impact on fetal neural development and, specifically, that the amount of prenatal stress the mother experiences has a programming effect on the reactivity of the child’s HPA axis (Anisman, Zaharia, Meaney, & Merali, 1998). In humans, prospective research examining the in-
298
DETERMINANTS OF AGGRESSION
fluence of prenatal stress on the HPA axis of mother and fetus, and later HPA axis functioning and development in the child, is virtually absent (but see Susman et al., 2001). At present it is therefore unclear how prenatal and early postnatal stress in the mother affect the infant’s HPA axis and temperament. Following birth, maternal hormone concentrations and emotions can still contribute to the development of the infant’s temperament as mediated by the quality, quantity, and regularity of maternal interactions. The postpartum period is for many women a time of intense emotions attributable to the dramatic psychological and role-requirement changes that accompany the transition into motherhood. New mothers have to assume responsibility for a helpless infant, often without adequate preparation. If these challenging roles and responsibilities occur in a family context without adequate social or economic support, negative effects on the mothers and their children may well occur. Postpartum maternal emotions are in many women characterized by dysphoria or other depressive symptoms, and the rapid decrease in gonadal and adrenal hormone concentrations in this period is likely to be a contributing factor. It has been found that children of affectively ill mothers have earlier and significantly more depressive and disruptive behavior problems, and it is likely that a suboptimal mother–child relationship plays an important role in this effect (Susman et al., 2001). On the basis of the substantial literature on animals, one can expect maternal stress and alterations in human prenatal hormones to affect fetal development and later infant behavior. For example, one can hypothesize that infants born from pregnancies with elevated levels of prenatal and/or postnatal stress will have a more reactive stress system and problems with coping and adaptation, resulting in behavioral problems such as aggression.
CONCLUSIONS To avoid any misunderstanding, it is worth emphasizing that biological contributions to human behavior should not be regarded as deterministic. Rather, the hormonal environment may affect behavior by biasing the neural system in such a way as to create behavioral dispositions. It is then that environment and culture add their influence by rewarding, punishing, and/or ignoring behaviors, thereby exaggerating, diminishing, distorting, or permitting the expression of biologically based behaviors (Reinisch & Sanders, 1992). Aggression is, of course, an optional behavioral strategy: As humans, we are exquisitely sensitive to our social environment and we are usually capable of inhibiting this form of behavior, although biological factors may play a role in how easy it is for us to do so. Any consideration and reflection of the biological—in this case the hormonal—influence on aggression is controversial. On one hand, it is obvious that individual differences in aggression originate, at least partly, in differences in the biological substrate and the central nervous system. On the other
Hormones and Aggression
299
hand, we find it hard to accept that these differences, and the way in which they are established, may also involve specific possibilities and restrictions. In light of this debate, it is important to remember that the evidence that hormones play a role in individual differences in aggression, certainly when we focus on normal aggression, is small. Nevertheless, research on the influence of biological factors on the origin and development of behavior in humans is important, not in order to demonstrate or reject the importance of biological factors, but rather because these biological factors are part and parcel of a process of growth and development that can be influenced and disturbed in numerous ways.
NOTES 1. The Constantino et al. (1993) study provides some information about pre-adrenarchal plasma testosterone values: in one study the mean T value for 3- to 5-year-olds was found to be 6.0 ± 2.0 ng/dl; the mean T value for 6- to 8-year-olds was 8.0 ± 2.8 ng/dl (data published by Forest, 1989); in another study by Lashansky, Saenger, & Fishman (1991) a mean T value for 1- to 5year-olds of 1.98 ± 0.85 ng/dl and a mean T value for 6- to 12-year-olds of 6.8 ± 6.0 ng/dl was observed. Data from our own study on testosterone in 8- to 12-year-old boys found a mean T value of 0.86 ± 1.5 nmol/l (range = 0.1–6.8), with a mean T value of 0.2 nmol/l in 8-year-old boys (Van Goozen, Matthys, Cohen-Kettenis, Thijssen, & Van Engeland, 1998b). With respect to saliva, Scerbo and Kolko (1994) found a mean T level of 0.037 ng/ml in those younger than 9 years old (n = 6), and a mean T level of 0.057 ng/ml in those between 9 and 14 years old (n = 34). Clearly, T values in prepubertal children are low, and reliable norms for serum and saliva androgen levels are needed. 2. Testosterone can be reliably detected in saliva using immunoassay in boys and girls ages 8 and older (Granger, Schwartz, Booth, & Arentz, 1999). It is therefore unclear how seriously one should regard the testosterone values and findings in the study published by Sánchez-Martín et al. (2000). 3. It is beyond the scope of this chapter to discuss in any detail the possibility that the influence of testosterone on aggression could be largely non-neural and indirect, and work primarily through its effects on the body (increased physical size, muscle mass, and appearance of secondary sex characteristics) and activity. For example, Mazur and Booth (1998) conclude that testosterone affects adolescent behavior mostly through indirect social responses elicited by maturation rather than through direct activation of target receptors in organs or the nervous system. Similarly, the activity differences between girls and boys seem to be associated with play and toy preferences that are at least partly influenced by testosterone (Hines et al., 2002), leading to sex segregation and to different socialization experiences in girls and boys.
REFERENCES Achenbach, T. M. (1991). Manual for the Child Behavior Checklist and 1991 Profile. Burlington, VT: University of Vermont Department of Psychiatry. Albert, D., Walsh, M., & Jonik, R. (1993). Aggression in humans: What is its biological foundation? Neuroscience and Biobehavioral Reviews, 17, 405–425. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Anisman, H., Zaharia, M. D., Meaney, M. J., & Merali, Z. (1998). Do early-life events
300
DETERMINANTS OF AGGRESSION
permanently alter behavioral and hormonal responses to stressors? International Journal of Developmental Neuroscience, 16, 149–164. Archer, J. (1991). The influence of testosterone on human aggression. British Journal of Psychology, 82, 1–28. Archer, J., Graham-Kevan, N., & Davies, M. (in press). Testosterone and aggression: A re-analysis of Book, Starzyk and Quinsey (2001). Aggression and Violent Behavior. Bakshi, V. P., & Kalin, N. H. (2000). Corticotropin-releasing hormone and animal models of anxiety: Gene–environment interactions. Biological Psychiatry, 48, 1175– 1198. Berenbaum, S. A., Duck, S. C., & Bryk, K. (2000). Behavioral effects of prenatal versus postnatal androgen excess in children with 21–hydroxylase-deficient congenital adrenal hyperplasia. Journal of Clinical Endocrinology and Metabolism, 85(2), 727– 733. Berenbaum, S. A., & Resnick, S. M. (1997). Early androgen effects on aggression in children and adults with congenital adrenal hyperplasia. Psychoneuroendocrinology, 22(7), 505–515. Berkowitz, L. (1989). Frustration–aggression hypothesis: Examination and reformulation. Psychological Bulletin, 106, 59–73. Brendgen, M., Vitaro, F., Tremblay, R. E., & Lavoie, F. (2001). Reactive and proactive aggression: Predictions to physical violence in different contexts and moderating effects of parental monitoring and caregiving behavior. Journal of Abnormal Child Psychology, 29, 293–304. Brennan, P. A., Raine, A., Schulsinger, F., Kirkegaard-Sorensen, L., Knop, J., Hutchings, B., et al. (1997). Psychophysiological protective factors for male subjects at high risk for criminal behavior. American Journal of Psychiatry, 154, 853–855. Carlson, M., & Earls, F. (1996). Psychological and neuroendocrinological sequelae of early social deprivation in institutionalized children in Romania. Annals of the New York Academy of Sciences, 807, 419–428. Clarke, D., Fearon, U., Cunningham, S. K., & McKenna, T. J. (1996). The steroidogenic effects of b- endorphin and joining peptide: A potential role in the modulation of adrenal androgen production. Journal of Endocrinology, 151, 301–307. Coie, J. D., & Dodge, K. A. (1998). Aggression and antisocial behavior. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology: Social, emotional, and personality development (pp. 779–862). New York: Wiley. Collaer, M. L., & Hines, M. (1995). Human behavioural sex differences: A role for gondal hormones during early development? Psychological Bulletin, 118, 55–107. Constantino, J. N., Grosz, D., Saenger, P., Chandler, D. W., Nandi, R., & Earls, F. J. (1993). Testosterone and aggression in children. Journal of the Academy of Child and Adolescent Psychiatry, 32, 1217–1222. Dabbs, J. M., Frady, R. L., Carr, T. S., & Besch, N. F. (1987). Saliva testosterone and criminal violence in young adult prison inmates. Psychosomatic Medicine, 49, 174–182. Dawes, M. A., Dorn, L. D., Moss, H. B., Yao, J. K., Kirisci, L., Ammerman, R. T., & Tarter, R. E. (1999). Hormonal and behavioral homeostasis in boys at risk for substance abuse. Drug and Alcohol Dependence, 55, 165–176. Dettling, A. C., Gunnar, M. R., & Donzella, R. (1999). Cortisol levels of young children in full-day childcare centers: Relations with age and temperament. Psychoneuroendocrinology, 24, 519–536. Dixson, A. F. (1980). Androgens and aggressive behavior in primates: A review. Aggressive Behavior, 6, 37–67.
Hormones and Aggression
301
Dodge, K. A., Lochman, J. E., Harnish, J. D., Bates, J. E., & Pettit, G. S. (1997). Reactive and proactive aggression in school children and psychiatrically impaired chronically assaultive youth. Journal of Abnormal Child Psychology, 106, 37–51. Eberhart, J. A., Keverne, E. B., & Meller, R. E. (1980). Social influences on plasma testosterone levels in male Talapoin monkeys. Hormones and Behavior, 14, 247–266. Ehrhardt, A. A., & Meyer-Bahlburg, H. F. L. (1981). Effects of prenatal sex hormones on gender-related behavior. Science, 211, 1312–1318. Ehrhardt, A. A., Meyer-Bahlburg, H. F. L., Rosen, L. R., Feldman, J. F., Veridiano, N. P., Zimmerman, I., & McEwen, B. S. (1985) Sexual orientation after prenatal exposure to exogeneous estrogen. Archives of Sexual Behavior, 14, 57–75. Finkelstein, J. W., Susman, E. J., Chinchilli, V. M., Kunselman, S. J., D’Arcangelo, M. R., Schwab, J., Demers, L. M., Liben, L. S., Lookingbill, G., & Kulin, H. E. (1997). Estrogen or testosterone increases self-reported aggressive behaviors in hypogonadal adolescents. Journal of Clinical Endocrinology and Metabolism, 82(8), 2433– 2438. Forest, M. G. (1989). Physiological changes in circulating androgens. Pediatric Adolescent Endocrinology, 19, 104–129. Francis, D. D., Caldji, C., Champagne, F., Plotsky, P. M., & Meaney, M. J. (1999). The role of corticotropin-releasing factor-norepinephrine systems in mediating the effects of early experience on the development of behavioral and endocrine responses to stress. Biological Psychiatry, 46, 1153–1166. Francis, D. D., Diorio, J. F., Plotsky, P. M., & Meaney, M. J. (2002). Environmental enrichment reverses the effects of maternal separation on stress reactivity. Journal of Neuroscience, 22(18), 7840–7843. Francis, D. D., Szegda, K., Campbell, G., Martin, D., & Insel, T. R. (2003). Epigenetic sources of behavioral differences in mice. Nature Neuroscience, 6(5), 445–446. Gerra, G., Zaimovic, A., Avanzini, P., Chittolini, B., Giucastro, G., Caccavari, R., Palladino, M., Maestri, D., Monica, C., Delsignore, R., & Brambilla, F. (1997). Neurotransmitter-neuroendocrine responses to experimentally induced aggression in humans: Influence of personality variable. Psychiatry Research, 66, 33–43. Gladue, B. A. (1991). Aggressive behavioral characteristics, hormones, and sexual orientation in men and women. Aggressive Behavior, 17, 313–326. Goodyer, I. M., Herbert, J., Altham, P. M. E., Pearson, J., Secher, S., & Shiers, S. (1996). Adrenal secretion during major depression in 8 to 16 year olds: I. Altered diurnal rhythms in salivary cortisol and dehydroepiandrosterone (DHEA) at presentation. Psychological Medicine, 26, 245–256. Gooren, L., Fliers, E., & Courtney, K. (1990). Biological determinants of sexual orientation. Annual Review of Sex Research, 1, 175–196. Gottlieb, G., & Halpern, C. T. (2002). A relational view of causality in normal and abnormal development. Development and Psychopathology, 14(3), 421–435. Goy, R. W., & McEwen, B. S. (1980). Sexual differentiation of the brain. Cambridge, MA: MIT Press. Granger, D. A., Schwartz, E. B., Booth, A., & Arentz, M. (1999). Salivary testosterone determination in studies of child health and development. Hormones and Behavior, 35, 18–27. Granger, D. A., Shirtcliff, E. A., Zahn-Waxler, C., Usher, B., Klimes-Dougan, B., & Hastings, P. (2003). Salivary testosterone diurnal variation and psychopathology in adolescent males and females: Individual differences and developmental effects. Development and Psychopathology, 15, 431–449.
302
DETERMINANTS OF AGGRESSION
Gunnar, M. R., Tout, K., de Haan, M., Pierce, S., & Stansbury, K. (1997). Temperament, social competence, and adrenocortical activity in preschoolers. Developmental Psychobiology, 31, 65–85. Halpern, C. T., Udry, R., Campbell, B., & Suchindran, C. (1994). Relationships between aggression and pubertal increases in testosterone: A panel analysis of adolescent males. Social Biology, 40, 8–24. Harris, J. A. (1999). Review and methodological considerations in research on testosterone and aggression. Aggression and Violent Behavior, 4(3), 273–291. Higley, J. D., Mehlman, P. T., Poland, R. E., Taub, D. M., Vickers, J., Suomi, S. J., & Linnoila, M. (1996). CSF testosterone and 5–HIAA correlate with different types of aggressive behaviors. Biological Psychiatry, 40, 1067–1082. Hines, M., Golombok, S., Rust, J., Johnston, K. J., Golding, J., & the Avon Longitudinal Study of Parents and Children Study Team. (2002). Testosterone during pregnancy and gender role behavior of preschool children: A longitudinal population study. Child Development, 73(6), 1678–1687. Hines, M., & Green, R. (1991). Human hormonal and neural correlates of sex-typed behaviors. Review of Psychiatry, 10, 536–555. Kagan, J., Reznick, J. S., & Snidman, N. (1988). Biological bases of childhood shyness. Science, 240, 167–173. Kaufman, J., Plotsky, P. M., Nemeroff, C. B., & Charney, D. S. (2000). Effects of early adverse experiences on brain structure and function: Clinical implications. Biological Psychiatry, 48, 778–790. King, R. J., Jones, J., Scheurer, J. W., Curtis, D., & Zarcone, V. P. (1990). Plasma-cortisol correlates of impulsivity and substance abuse. Personality and Individual Differences, 11, 287–291. Klimes-Dougan, B., Hastings, P. D., Granger, D. A., Usher, B. A., & Zahn-Waxler, C. (2001). Adrenocortical activity in at-risk and normally developing adolescents: Individual differences in salivary cortisol basal levels, diurnal variation, and responses to social challenges. Development and Psychopathology, 13, 695–719. Kruesi, M. J. P., Rapoport, J. L., Hamburger, S., Hibbs, E., Potter, W. Z., Lenane, M., & Brown, G. L. (1990). Cerebrospinal fluid monoamine metabolites, aggression, and impulsivity in disruptive behavior disorders of children and adolescents. Archives of General Psychiatry, 47, 419–426. Kruesi, M. J. P., Schmidt, M. E., Donnelly, M., Hibbs, E. D., & Hamburger, S. D. (1989). Urinary free cortisol output and disruptive behavior in children. Journal of the American Academy of Child and Adolescent Psychiatry, 28, 441–443. Lahey, B. B., Waldman, I. D., & McBurnett, K. (1999). Annotation: The development of antisocial behavior: An integrative causal model. Journal of Child Psychology and Psychiatry, 40, 669–682. Lashansky, G., Saenger, P., & Fishman, K. (1991). Normative data for adrenal steroidogenesis in a healthy pediatric population: Age and sex-related changes after ACTH stimulation. Journal of Endocrinology and Metabolism, 73, 674–686. Liu, D., Diorio, J., Tannenbaum, B., Caldji, C., Francis, D., Freedman, A., Sharma, S., Pearson, D., Plotsky, P. M., & Meany, M. J. (1997). Maternal care, hippocampal glucocorticoid receptors, and hypothalamic-pituitary-adrenal responses to stress. Science, 277, 1659–1662. Loeber, R., Green, S. M., Lahey, B. B., & Stouthamer-Loeber, M. (1991). Differences and similarities between children, mothers, and teachers as informants on disruptive child behavior. Journal of Abnormal Child Psychology, 19, 75–95.
Hormones and Aggression
303
Majewska, M. D. (1992). Neurosteroids: Endogeneous bimodal modulators of the GABAA receptor: Mechanisms of action and physiological significance. Progress in Neurobiology, 38, 379–395. Matthys, W., Van Goozen, S. H. M., De Vries, H., Cohen-Kettenis, P. T., & Van Engeland, H. (1998). The dominance of behavioral activation over behavioral inhibition in conduct disordered boys with or without attention deficit hyperactivity disorder. Journal of Child Psychology and Psychiatry, 39, 643–651. Mazur, A., & Booth, A. (1998). Testosterone and dominance in men. Behavioral and Brain Sciences, 21, 353–397. McBurnett, K., Lahey, B. B., Frick, P. J., Risch, C., Loeber, R., Hart, E. L., Christ, M. A. G., & Hanson, K. S. (1991). Anxiety, inhibition, and conduct disorder in children: II. Relation to salivary cortisol. Journal of the American Academy of Child and Adolescent Psychiatry, 30, 192–196. McBurnett, K., Lahey, B. B., Rathouz, P. J., & Loeber, R. (2000). Low salivary cortisol and persistent aggression in boys referred for disruptive behavior. Archives of General Psychiatry, 57, 38–43. Meyer-Bahlburg, H. F. L., Boon, D. A., Sharma, M., & Edwards, J. A. (1974). Aggressiveness and testosterone measures in man. Psychosomatic Medicine, 36, 269– 274. Moss, H. B., Vanyukov, M. M., & Martin, C. S. (1995). Salivary cortisol responses and the risk for substance abuse in prepubertal boys. Biological Psychiatry, 38, 547– 555. Moyer, K. E. (1976). The psychobiology of aggression. New York: Harper & Row. O’Connor, D. B., Archer, J., Hair, W. M., & Wu, F. C. W. (2002). Exogenous testosterone, aggression, and mood in eugonadal and hypogonadal men. Physiology and Behavior, 75, 557–566. Olweus, D., Mattsson, A., Schalling, D., & Low, H. (1988). Circulating testosterone levels and aggression in adolescent males: A causal analysis. Psychosomatic Medicine, 50, 261–272. Paikoff, R., & Brooks-Gunn, J. (1990). Physiological processes: What role do they play during the transition to adolescence? In P. Montemayor, G. Adams, & T. Gullota (Eds.), Advances in adolescent development: Transition from childhood to adolescence (Vol. 2, pp. 63–81). Newbury Park, CA: Sage. Pajer, K., Gardner, W., Rubin, R. T., Perel, J., & Neal, S. (2001). Decreased cortisol levels in adolescent girls with conduct disorder. Archives of General Psychiatry, 58, 297– 302. Raine, A. (1996). Autonomic nervous system activity and violence. In D. M. Stoff & R. B. Cairns (Eds.), Aggression and violence (pp. 145–168). Mahwah, NJ: Erlbaum, Reinisch, J. M. (1981). Prenatal exposure to synthetic progestins increases potential for aggression in humans. Science, 211, 561–562. Reinisch, J. M., & Sanders, S. A. (1992). Prenatal hormonal contributions to sex differences in human cognitive and personality development. In A. A. Gerall, H. Moltz, & I. L. Ward (Eds.), Handbook of behavioral neurobiology: Sexual differentiation (Vol. 11, pp. 221–243). New York: Plenum Press. Robel, P., & Baulieu, E. E. (1995). Dehydroepiandrosterone (DHEA) is a neuroactive neurosteroid. Annals of the New York Academy of Sciences, 774, 82–110. Sánchez-Martín, J., Fano, E., Ahedo, L., Cardas, J., Brain, P., & Azpíroz, A. (2000). Relating testosterone levels and free play social behavior in male and female preschool children. Psychoneuroendocrinology, 25, 773–783.
304
DETERMINANTS OF AGGRESSION
Sapolsky, R. M. (1997). The importance of a well-groomed child. Science, 277, 1620– 1621. Scerbo, A. S., & Kolko, D. J. (1994). Salivary testosterone and cortisol in disruptive children: Relationship to aggressive, hyperactive and internalizing behaviors. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 1174– 1184. Schaal, B., Tremblay, R., Soussignan, R., & Susman, E. (1996). Male testosterone linked to high social dominance but low physical aggression in early adolescence. Journal of the Academy of Child and Adolescent Psychiatry, 34, 1322–1330. Schulz, K. P., Halperin, J. M., Newcorn, J. H., Sharma, V., & Gabriel, S. (1997). Plasma cortisol and aggression in boys with ADHD. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 605–609. Spangler, G., & Grossmann, K. E. (1993). Biobehavioral organization in securely and insecurely attached infants. Child Development, 64(5), 1439–1450. Spangler, G., & Scheubeck, R. (1993). Behavioral organization in newborns and its relation to adrenocortical and cardiac activity. Child Development, 64(2), 622– 633. Stoff, D. M., Pasatiempo, A. P., Yeung, J., Cooper, T. B., Bridger, W. H., & Rabinovich, H. (1992). Neuroendocrine responses to challenge with dl-fenfluramine and aggression in disruptive behavior disorders of children and adolescents. Psychiatry Research, 43, 263–276. Susman, E. J., Dorn, L. D., Inoff-Germain, G., Nottelmann, E. D., & Chrousos, G. P. (1997). Cortisol reactivity, distress behavior, and behavioral and psychological problems in young adolescents: A longitudinal perspective. Journal of Research on Adolescence, 7, 81–105. Susman, E. J., Inoff-Germain, G., Nottelmann, E. D., Loriaux, D., Cutler, G., & Chrousos, G. P. (1987). Hormones, emotional dispositions, and aggressive attributes in young adolescents. Child Development, 58, 1114–1134. Susman, E. J., Schmeelk, K. H., Ponirakis, A., & Gariepy, J. L. (2001). Maternal prenatal, postpartum, and concurrent stressors and temperament in 3–year-olds: A person and variable analysis. Development and Psychopathology, 13, 629–652. Susman, E. J., Schmeelk, K. H., Worrall, B. K., Granger, D. A., Ponirakis, A., & Chrousos, G. P. (1999). Corticotropin-releasing hormone and cortisol: Longitudinal associations with depression and antisocial behavior in pregnant adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 460– 467. Tennes, K., & Kreye, M. (1985). Children’s adrenocortical responses to classroom activities and tests in elementary school. Psychosomatic Medicine, 47, 451–460. Tennes, K., Kreye, M., Avitable, N., & Wells, R. (1986). Behavioral correlates of excreted catecholamines and cortisol in second-grade children. Journal of the American Academy of Child and Adolescent Psychiatry, 25, 764–770. Tremblay, R. E. (2000). The development of aggressive behaviour during childhood: What have we learned in the past century? International Journal of Behavioral Development, 24, 129–141. Tremblay, R. E. (2003). Why socialization fails: The case of chronic physical aggression. In B. B. Lahey, T. E. Moffitt, & A. Caspi (Eds.), Causes of conduct disorder and juvenile delinquency (pp. 182–224). New York: Guilford Press. Van de Beek, C., Thijssen, J. H. H., Cohen-Kettenis, P. T., Van Goozen, S. H. M., & Buitelaar, J. K. (2004). Sex steroids in pregnant women: Relationships between hor-
Hormones and Aggression
305
mones assessed in amniotic fluid, and maternal and umbilical cord serum. Hormones and Behavior, 46(5), 663–669. Van de Poll, N. E., & Van Goozen, S. H. M. (1992). Hypothalamic involvement in sexuality and hostility: Comparative psychological aspects. Progress in Brain Research, 93, 343–361. Van de Poll, N. E., Van Zanten, S., & De Jonge, F. H. (1986) Effects of testosterone, estrogen and dihydrotestosterone upon aggressive and sexual behavior of female rats. Hormones and Behavior, 20, 418–431. Van Goozen, S. H. M., Cohen-Kettenis, P. T., Gooren, L. J. G., Frijda, N. H., & Van de Poll, N. E. (1995). Gender differences in behaviour: Activating effects of cross-sex hormones. Psychoneuroendocrinology, 20, 343–363. Van Goozen, S. H. M., Cohen-Kettenis, P. T., Snoek, H., Matthys W., Swaab-Barneveld, H., & Van Engeland, H. (2004). Executive functioning in children: A comparison of hospitalized ODD and ODD/ADHD children and normal controls. Journal of Child Psychology and Psychiatry, 45(2), 284–292. Van Goozen, S. H. M., Frijda, N. H., & Van de Poll, N. E. (1994). Anger and aggression in women: Influence of sports choice and testosterone administration. Aggressive Behavior, 20, 213–222. Van Goozen, S. H. M., Frijda, N. H., Wiegant, V. M., Endert E., & Van de Poll, N. E. (1996). The premenstrual phase and reactions to aversive events: A study of hormonal influences on emotionality. Psychoneuroendocrinology, 21, 479–497. Van Goozen, S. H. M., Matthys, W., Cohen-Kettenis, P. T., Buitelaar, J. K., & Van Engeland, H. (2000a). Hypothalamic–pituitary–adrenal axis and autonomic nervous system activity in disruptive children and matched controls. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 1438–1445. Van Goozen, S. H. M., Matthys, W., Cohen-Kettenis, P. T., Gispen-deWied, C., Wiegant, V. M., & Van Engeland, H. (1998a). Salivary cortisol and cardiovascular activity during stress in oppositional defiant disorder boys and normal controls. Biological Psychiatry, 43, 531–539. Van Goozen, S. H. M., Matthys, W., Cohen-Kettenis, P. T., Thijssen, J. H. H., & Van Engeland, H. (1998b). Adrenal androgens and aggression in prepubertal boys. Biological Psychiatry, 43, 156–158. Van Goozen, S. H. M., Van den Ban, E., Matthys, W., Cohen-Kettenis, P. T., Thijssen, J. H. H., & Van Engeland, H. (2000b). Increased adrenal androgen functioning in children with oppositional defiant disorder: A comparison with psychiatric and normal controls. Journal of the American Academy of Child and Adolescent Psychiatry, 39(11), 1446–1451. Vanyukov, M. M., Moss, H. B., Plail, J. A., Blackson, T., Mezzich, A. C., & Tarter, R. E. (1990). Antisocial symptoms in preadolescent boys and in their parents: Associations with cortisol. Psychiatry Research, 46, 9–17. Virkkunen, M. (1985). Urinary free cortisol secretion in habitually violent offenders. Acta Psychiatrica Scandinavica, 72, 40–44. Vitiello, B., & Stoff, D. M. (1997). Subtypes of aggression and their relevance to child psychiatry. Journal of the American Academy of Child and Adolescent Psychiatry, 36(3), 307–315. Warren, M. P., & Brooks-Gunn, J. (1989). Mood and behavior at adolescence: Evidence for hormonal factors. Journal of Clinical Endocrinology and Metabolism, 69, 77– 83. Wolf, O. T., & Kirschbaum, C. (1999). Actions of dehydroepiandrosterone and its sulfate
306
DETERMINANTS OF AGGRESSION
in the central nervous system: Effects on cognition and emotion in animals and humans. Brain Research Review, 30(3), 264–288. Woodman, D. D., Hinton, J. W., & O’Neill, M. T. (1978). Cortisol secretion and stress in maximum security hospital patients. Journal of Psychosomatic Research, 22, 133– 136. Zuckerman, M. (1991). Psychobiology of personality. Cambridge, MA: Cambridge University Press.
DETERMINANTS Executive Function OF in Aggression AGGRESSION
15 Executive Function in Early Physical Aggression J EAN R ICHARD S ÉGUIN and P HILIP D AVID Z ELAZO
Physical aggression is characteristic of early childhood but typically becomes less frequent with development (Tremblay et al., 1996). During this same period of their development, there are marked age-related increases in children’s executive function (Zelazo & Müller, 2002). Executive function generally refers to the self-regulation of thought, action, and emotion—processes that depend on the integrity of neural systems involving prefrontal cortex (e.g. Owen et al., 1999; Robbins, 1996; Stuss, 1992). We propose that this pattern of declining aggression and increasing executive function reflects the fact that, during development, children typically acquire the executive function skills needed to regulate physical aggression. Moreover, we propose that in those cases (about 4–6% of the population) where children’s levels of physical aggression remain high (Nagin & Tremblay, 1999; Tremblay et al., 2004), the development of executive function will be atypical. The literature is replete with studies that have examined the relation between global cognitive measures and global indices of antisocial or externalizing behavior problems. To date, however, very few studies have tested the specific hypothesis that executive function development is particularly relevant to the regulation of physical aggression. Executive function is a specific aspect of cognitive function that is rarely captured by global measures. Similarly, physical aggression is a specific aspect of aggression that may not be captured by externalizing behavior problems or global scales of aggression, disruptiveness, antisocial behavior, or delinquency. Thus, a test of the hypoth307
308
DETERMINANTS OF AGGRESSION
esis requires taking into account (1) that executive function is a specific subset of more global cognitive abilities and (2) that physical aggression is a specific behavior within externalizing, aggressive, disruptive, antisocial behavior problems or delinquency.
PHYSICAL AGGRESSION AMONG GLOBAL MEASURES OF BEHAVIOR PROBLEMS Physical aggression can be found in the form of hitting, kicking, biting, assaults, confrontation with a weapon, getting involved in fights, threatening to use physical force, rape, and homicide. Physical aggression is also studied indirectly, via its relation to various externalizing disorders, and we now know from retrospective and longitudinal studies that externalizing disorders begin a chronic course in early childhood (Tremblay et al., 1999). Physically aggressive behavior is found in conduct disorder (CD), and it is related to, but not a diagnostic feature of, attention-deficit/hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD). ADHD itself can be subdivided into a subtype in which attention deficits predominate and another subtype characterized by hyperactive and impulsive behavior (Milich, Balentine, & Lynam, 2001). Within ADHD, it is the hyperactive–impulsive subtype that is more strongly associated with ODD and CD (Willcutt, Pennington, Chhabildas, Friedman, & Alexander, 1999), and possibly with physical aggression, although early ADHD behavior problems are not typically related to later violence (Farrington & Loeber, 2000; Lahey, Loeber, Burke, Rathouz, & McBurnett, 2002; Loeber, Burke, & Lahey, 2002; Nagin & Tremblay, 1999; Stouthamer-Loeber, Loeber, Wei, Farrington, & Wikström, 2002). CD, ADHD, and ODD co-occur at a greater than chance level, and this has been a growing source of concern in understanding the development of externalizing disorders (Lilienfeld, 2003). In a longitudinal study of a community sample of more than 1,000 boys from low-socioeconomic-status (SES) schools (Séguin, Boulerice, Harden, Tremblay, & Pihl, 1999), close to 90% of adolescents who met criteria for ADHD or CD between the ages of 14 and 16 were found to have had a history of occasional or chronic physically aggressive behavior (based on teacher ratings of physical aggression between the ages of 6 and 12 years). This is in contrast to 61% of boys without such a diagnosis. It should be noted, however, that although a history of physical aggression during childhood is relatively common in most adolescents with ADHD or CD, only about a quarter of boys with a history of physical aggression meet the criteria for ADHD or CD (Séguin et al., 1999). Finally, besides externalizing disorders, physical aggression is also found among other global categories such as “disruptive disorders,” in children defined as “hard to manage,” among “antisocial disorders,” and in “delinquency” and “criminality.”
Executive Function in Aggression
309
Developmental Course of Physical Aggression Research on the developmental course of physical aggression per se suggests (Robins, 1991) that the frequency of physical aggression actually peaks at around 2 years of age (See Tremblay & Nagin, Chapter 5, this volume). Most children who begin with high levels of physical aggression gradually desist at various rates. However, there is a small but significant proportion of children (about 4–6%) for whom frequencies of physical aggression remain very high throughout childhood and adolescence (Broidy et al., 2003). We propose that these children will display early impairments in executive function.
EXECUTIVE FUNCTION AND ITS DEVELOPMENT Executive function is a popular topic in contemporary research, but definitions of executive function differ widely. A traditional approach is to treat executive function as a higher order cognitive mechanism or ability, as when Denckla and Reiss (1997) suggest that “executive function refers to a cognitive module consisting of effector output elements involving inhibition, working memory, and organizational strategies necessary to prepare a response” (p. 283, emphasis in original). Unfortunately, this approach (see also Baddeley & Della Sala, 1996; Norman & Shallice, 1986) essentially invokes a homunculus, or little man in the head, rather than explaining in detail how executive function, including inhibition, working memory, and strategy use, is accomplished (Parkin, 1998; Zelazo & Müller, 2002). An alternative and more common approach is more empirically driven. This approach involves devising comprehensive neuropsychological batteries and using factor analysis to reveal the underlying structure of executive function (see Zelazo & Müller, 2002, for a review). Such studies generally reveal three or four factors—a result that has been taken to suggest that there are dissociable dimensions of executive function, consistent with efforts to “fractionate” executive function based on lesion studies in nonhuman animals (Robbins, 1996). Although the results of factor-analytic studies are potentially misleading, because providing labels for factors may give the erroneous impression that researchers actually understand the cognitive processes underlying performance on various tasks, this approach does suggest that there are both common and diverse aspects of the measures that have been used to assess executive function (Miyake et al., 2000). Research on executive function has demonstrated that executive function first emerges early in development, probably around the end of the first year of life, and it develops across a wide range of ages, with important changes occurring between about 2 and 5 years of age. Adult-level performance on many standard tests of executive function is reached at about 12 years of age, and performance on some measures continues to change into adulthood (see
310
DETERMINANTS OF AGGRESSION
Zelazo & Müller, 2002, for review). One of the most widely studied examples of executive function in infants is performance on the A-not-B task. In this task, an infant watches as an object is hidden at one location (A). Following a brief delay (e.g., 3 seconds), the infant is permitted to search for the object. After a number of A trials, the object is then hidden conspicuously at a different location (B). Despite witnessing the hiding event at B, most 9-month-olds continue to search perseveratively at A on the first B trial, thereby committing the A-not-B error (for a recent meta-analysis, see Marcovitch & Zelazo, 1999). However, by about 12 months of age, most infants search flexibly on this task. The changes occurring between about 2 and 5 years have been carefully documented using a variety of measures of rule use. The ability to use a rule systematically to control behavior seems to be acquired at about 2.5 years of age. However, children at this age are highly susceptible to perseverative responding, and it is not until 3 years that most children are able to use two rules simultaneously (Luria, 1973; Zelazo, Frye, & Rapus, 1996a; Zelazo & Reznick, 1991). Then, although 3-year-olds can use a pair of rules, it not until about 5 years that most children can switch flexibly between two incompatible pairs of rules, or two different perspectives on a situation, as seen in tasks used in assessing causal reasoning (e.g., Frye, Zelazo, Brooks, & Samuels, 1996; Zelazo, Reznick, & Piñon, 1995), behavioral prediction (Zelazo, Helwig, & Lau, 1996b), card sorting (e.g., Zelazo et al., 1996a), and understanding false beliefs (see Wellman, Cross, & Watson, 2001, for a meta-analysis). The Wisconsin Card Sorting Test (WCST) has frequently been used to explore the development of executive function beyond about 5 years of age. In the WCST (Grant & Berg, 1948; Robinson, Heaton, Lehman, & Stilson, 1980), participants are presented with four target cards that differ on three dimensions (number, color, and shape) and asked to sort a series of test cards that match different target cards on different dimensions. Participants must discover the sorting rule by trial and error, and after a certain number of consecutive correct responses, the sorting rule is changed. In a representative study, Chelune and Baer (1986) documented a linear increase in performance on the WCST between the ages of 6 and 10 years, with 6-year-olds performing at the level of patients with prefrontal cortical damage (e.g., Milner, 1963) and 10-year-olds performing like healthy adults. This pattern was obtained for the three variables assessed: number of perseverative errors, number of categories achieved, and failures to maintain set.
THE RELATION OF EXECUTIVE FUNCTION TO EXTERNALIZING OR ANTISOCIAL DISORDERS Most research on the relation between executive function and physical aggression has relied on relatively global measures of both executive function and aggression. In most cases, the measures of executive function correspond to
Executive Function in Aggression
311
standard neuropsychological tests that are either taken directly from the adult literature or adapted from adult tasks. For example, several studies have used the WCST. The WCST taps numerous aspects of executive function, and, as a result, the origin of errors on this task is difficult to determine (e.g., see Delis, Squire, Bihrle, & Massman, 1992). To perform correctly on the WCST, one must first construct a representation of the problem space, which includes identifying the relevant dimensions. Then, one must choose a promising plan—for example, sorting according to shape. After selecting a plan, one must (1) keep the plan in mind long enough for it to guide one’s thought or action and (2) actually carry out the prescribed behavior. Keeping a plan in mind to control behavior is referred to as intending; translating a plan into action is rule use. Finally, after acting, one must evaluate one’s behavior, which includes both error detection and error correction. Inflexibility can occur at each phase, so there are several possible explanations of perseverative performance on the WCST—and on global executive function tasks more generally. For example, perseveration can occur after a rule change in the WCST either because a new plan was not formed (one type of representational inflexibility; Zelazo et al., 1995) or because the plan was formed but not carried out (an example of lack of response control; Zelazo et al., 1995). Nonetheless, research using the WCST has proven useful in documenting that executive function is diminished in groups of children who are likely high in physical aggression. Indeed, relatively high levels of perseveration on the WCST have been documented in individuals with antisocial behavior (Kandel & Freed, 1989), in children with externalizing behavior problems (Matson & Fisher, 1991), in children with ADHD (Pennington, 1997), and in adolescents with CD (Toupin, Déry, Pauzé, Mercier, & Fortin, 2000). Performance on the WCST was also related to childhood social abilities (Bonino & Cattelino, 1999). Perseveration on other card-playing tasks that also involve error detection and correction has also been shown in adult psychopathic populations (Newman & Wallace, 1993) and in our work on adolescents with a welldocumented history of physical aggression (Séguin, Arseneault, Boulerice, Harden, & Tremblay, 2002). Several studies along these lines have shown that poor executive function is involved in ADHD and antisocial disorders. These have recently been reviewed, respectively, by Pennington and Ozonoff (1996) and Morgan and Lilienfeld (2000). This latter review covered studies published between 1942 and 1997 for ages 13–40 years and revealed a medium-sized effect (mean standard difference) of about 0.6 in comparing antisocial and nonantisocial groups. Unfortunately, however, the review could not focus on violent antisociality because most studies lacked information about violence. This is indeed a serious problem in the field (Tremblay, 2000). Further, Morgan and Lilienfeld (2000) and Pennington and Ozonoff (1996) point out that the studies they reviewed have typically not taken into account the comorbidity between ADHD and antisocial disorders such as CD. A third recent review further examined problems associated with tests for the specificity of executive
312
DETERMINANTS OF AGGRESSION
function to the three externalizing disorders, and to Tourette’s syndrome and high-functioning autism (Sergeant, Geurts, & Oosterlaan, 2002). The problem of examining physical aggression per se (instead of antisocial disorders or CD/ODD) has been addressed, however, in several recent studies that have supported the hypothesis that executive function problems are related to physical aggression when observable physical aggression was entered into statistical models even after controlling for ADHD (see Giancola, Mezzich, & Tarter, 1998; Séguin et al., 1999; Toupin et al., 2000). The failure to take into account physical aggression and hyperactivity at the outset has typically resulted in a weaker assessment of one or the other behavior problem. In our earlier work we found that boys from a community sample with a history of physical aggression performed most poorly on tests of executive function even after controlling for verbal and spatial abilities as well as cerebral dominance (Séguin, Pihl, Harden, Tremblay, & Boulerice, 1995). In that study and others that followed, the core executive function tasks were derived from the neuropsychological literature and had been validated with brain imaging techniques. The Self-Ordered Pointing task and a number randomization task involved primarily rule use, and Conditional Association tasks also involved error detection and correction in addition to rule use (Petrides, Alivisatos, Evans, & Meyer, 1993a; Petrides, Alivisatos, Meyer, & Evans, 1993b). In addressing a question raised by Pennington and Ozonoff (1996) about that study, we further controlled for chronic ADHD. We also added a control for IQ and still found poor executive function in the physically aggressive boys (Séguin et al., 1999). Moreover, in an attempt to understand more fully the relation between executive function and physical aggression and hyperactivity, we selected study participants based on the histories of both behavior problems. We chose to use the narrow-band behavior of hyperactivity instead of the broader-band construct of ADHD because hyperactivity was found to attenuate the relation between physical aggression and card perseveration (Séguin et al., 2002), although ADHD, as a whole, was not related to executive function in our previous study (see Séguin et al., 1999). Using this methodology also allowed testing for additive and interactive effects of physical aggression and hyperactivity on cognitive function in general and executive function in particular (Waschbusch, 2002). We did not find an interaction, but did find clear additive effects of both behavior problems on executive function, although the effects were clearer for physical aggression than they were for hyperactivity (Séguin, Nagin, Assaad, & Tremblay, 2004).
THE RELATION OF PHYSICAL AGGRESSION AND EXECUTIVE FUNCTION IN EARLY CHILDHOOD Given that, for most children, the decline in physical aggression occurs between ages 24 and 60 months (see Tremblay & Nagin, Chapter 5, this volume) and that this period is marked by major changes in executive function, it
Executive Function in Aggression
313
seems especially important to examine the relation between aggression and executive function during the preschool years. Early identification of potential problems also has implications for the success of any interventions, because early childhood may be a period of relative plasticity and it may be easier to alter behavior before it becomes entrenched. Studies of this relation in preschoolers have addressed a broad range of relevant issues. Some studies focused more specifically on the relation between early hyperactivity and aspects of cognitive function relevant to, but not isomorphic with, executive function. For example, one community study found that three-quarters of the 2% of children who had hyperactivity problems and were “very hard to manage” in preschool still had problems at age 15 (McGee, Partridge, Williams, & Silva, 1991). They were worse off as adolescents on both behavioral and cognitive measures than preschoolers who were only “very hard to manage” (3% of that sample) and developmental control children (95% of that sample). However, it is not clear whether early poor cognitive abilities contributed to the maintenance of the behavior problems. Other studies of preschool hyperactivity have also failed to take into account concurrent conduct problems, let alone physical aggression. Such was the case in one study of hyperactive preschoolers from the community (vs. the clinic) who were found to have deficits in executive function (Mariani & Barkley, 1997).
Studies Using Global Measures of Cognition Only a handful or so of studies have examined more directly the relation between preschool cognitive function and behavior problems not restricted to hyperactivity. For example, poor intelligence assessed by the McCarthy Scales (McCarthy, 1970) was related to mother-rated Child Behavior Check List (CBCL; Achenbach, Edelbrock, & Howell, 1987) psychopathology indices in preschoolers recruited from pediatricians’ practices (Dietz, Lavigne, Arend, & Rosenbaum, 1997). However, because performance on the McCarthy Scales was related to internalizing as well as externalizing problems and because the McCarthy score was taken either as a whole or only broken down into verbal and performance IQ, the study does not provide information about the specificity of behavior problems in relation to cognitive function. Another longitudinal study with assessments at about 55 and 80 months of age found that preschool IQ was related to CBCL teacher ratings of externalizing problems at follow-up and that preschool expressive language abilities were related to CBCL maternal ratings of externalizing problems at follow-up (Heller, Baker, Henker, & Hinshaw, 1996). However, these associations were accounted for by maternal education and socioeconomic status. This again suggests that parental characteristics may influence both cognitive development and externalizing problems, but it is not clear that this poor cognitive development represents an incremental risk factor beyond parenting effects, because interactions with cognitive abilities were not tested. Interestingly, preschool ratings of
314
DETERMINANTS OF AGGRESSION
externalizing behavior problems did not attenuate first-grade IQ and language abilities over and above the variance explained by preschool cognitive abilities. Although preschool language abilities predicted stability and pervasiveness of externalizing disorder across assessments, this was not the case for IQ. Another prospective longitudinal study failed to support Moffitt’s (1993) hypothesis that adolescents with early-onset/persistent conduct problems would show early neuropsychological deficits (Aguilar, Sroufe, Egeland, & Carlson, 2000). Based on CBCL externalizing scores, the authors contrasted four groups (i.e., never antisocial, child-limited, adolescent-limited, and early onset/persistent) in a one-factor (group) design instead of a 2 × 2, child by adolescent, factorial design. Adolescent scores differentiated the adolescentonset and early-onset/persistent groups, as in Moffitt’s earlier work (see Moffitt, 1990; Moffitt & Henry, 1989; Moffitt & Silva, 1988b; Moffitt & Silva, 1988a), but the effect seems to be the result of having had behavior problems as a child, because those children who recovered from behavior problems (i.e. child-limited) appear to have performed as poorly as those that did not (i.e. early-onset/persistent). Nonetheless, early cognitive scores were associated with later ones, but stability of cognitive performance was not used to identify those children who performed poorly in childhood and adolescence. Despite the many merits of that study, its test of Moffitt’s hypothesis may be severely limited. First, most children in that sample were at high risk for behavior problems. Second, about a third of the sample was assigned to the early-onset/persistent category, when we would expect about 4% of chronic cases in normative samples (Broidy et al., 2003). Third, the study included a roughly equal number of males and females, yet trajectories of early-onset problems do not predict the same maladaptive outcomes in males and females (Broidy et al., 2003; see also Archer & Côté, Chapter 20, this volume). Fourth, Moffitt suggested that the presence of aggression and hyperactivity should be used as a criterion for early onset, and it is not clear that this is the case here. Finally, there were no tests of executive function. Raine and colleagues (Raine, Yaralian, Reynolds, Venables, & Mednick, 2002) have proposed a different cognitive interpretation of persistent antisocial behavior. Their early-starter spatial impairment model of antisocial behavior suggests that early visuospatial (right hemisphere) impairments can predispose a child to persistent antisocial behavior by interfering with early attachment via emotion recognition and regulation. The model is admittedly speculative, but is reasonably well based on a series of community studies, including a longitudinal study of children from ages 3–17 years. Unlike most studies reviewed so far, this study included stability of behavior as an important design feature. As in the Aguilar et al. (2000) study, four groups were formed, based on whether children were above or below the top 33rd percentile of an antisocial behavior problem index at ages 8 and 17 years: a comparison group, a child-limited group, an adolescence-onset group, and a persistent group. Results show that spatial abilities at age 3 years were poorest in the
Executive Function in Aggression
315
persistent group, and that this effect appeared to be stronger than the effects of groups on verbal deficits. This effect was robust to statistical control for level of hyperactivity and test motivation at age 3 years. However, by the time the children reached age 11 years, spatial and verbal deficits were equally impaired in this group and social adversity attenuated considerably the group effect on spatial deficits. There appears to be a loss of specificity of spatial deficits over time; the persistent children now seemed generally impaired. Although a spatial deficit hypothesis was formulated, it is difficult to argue that the tasks that formed that spatial index (which represents essentially performance IQ tasks) were not contaminated by several other abilities. The problem of task purity is found everywhere in this type of research, and, unfortunately, tasks that more explicitly and specifically involve executive function were not administered in that study. A close examination of the battery does not suggest that any tasks could have been construed in terms of executive function either. Although the Wisconsin Intelligence Scale for Children (WISC) Mazes subtest, which involves some degree of planning, was administered, individual subtest scores were not available. Further, physical aggression was not examined. Thus, the question of the specific relevance of preschool executive function to the concurrent or later regulation of physical aggression remains unanswered.
Studies Using Specific Measures of Executive Function in Community Samples Moving more closely toward the concept of executive function, “disruptiveness” in preschoolers, aged about 4½ years, was found to be negatively associated with verbal and visuospatial abilities (Cole, Usher, & Cargo, 1993). “Executive function” was related to another measure of behavioral control (i.e., following the rule not to play with a forbidden but attractive object) but not to disruptiveness. In that study verbal and visuospatial abilities were assessed with the McCarthy Scales (Kaufman & Kaufman, 1977; McCarthy, 1970), as well as with the Florida Kindergarten Screening Battery (Satz & Fletcher, 1982), which includes the Beery Visual-Motor integration test (Beery, 1982) and the Peabody Picture Vocabulary Test (Dunn & Dunn, 1981). The executive function composite was made up of a tapping test (Becker, Isaac, & Hynd, 1987), a rapid-alternating-stimulus-naming test (Wolf, 1986), a hand movement test (Kaufman, 1983), a block sort (based on the WCST; Heaton, 1981), and a visual search test (adapted from Plude & Doussard-Roosevelt, 1989). All five tests loaded onto one principal component accounting for 42% of the variance, but the sum of scale scores was used instead of the principal component score. Individual scores correlated significantly with one another between r = .27 and r = .64. Cognitive risk was defined as scores below one standard deviation from the mean. Behavior was assessed with four scales. The mothers completed the CBCL (Achenbach, 1999; Achenbach & Edelbrock, 1983) and the Eyberg Child Behavior Inventory (Eyberg & Robinson,
316
DETERMINANTS OF AGGRESSION
1983), and teachers completed the Teacher Report Form (TRF; Achenbach & Edelbrock, 1986) and the Preschool Behavior Questionnaire (Behar & Stringfield, 1974). In regard to behavior, three risk levels had been identified in this study on the basis of combined informants and scale scores (i.e., low < 1 SD below mean for all informants, high > 1 SD above mean according to at least one informant or T score > 70 on CBCL total score or > 85th percentile on the TRF, and moderate for the rest of the sample). We note that these criteria for behavior risk definition were very liberal and may have increased the chance for false positive identification of risk. Analyses contrasted a combination of the two high groups (moderate and high) with the low-risk group. Therefore, we do not know the relative positions of performance scores between the three groups. It was also clear that the “high” group was elevated not only in disruptive behavior but also in other behavior problems, which may limit further the specificity of the finding to disruptive behavior. Another study also referring explicitly to the concept of executive function found that “hard-to-manage” children performed more poorly than control children not only on executive function tasks, but also on theory of mind and emotion understanding tasks (Hughes, Dunn, & White, 1998). Interestingly, the relation between number of successful executive function tasks and behavior problems was partly accounted for by the introduction of socioeconomic status and verbal abilities to the model, although the executive function battery was explicitly designed to require minimal verbal abilities. In that study, executive function was measured with six tasks, and the authors were guided by a model of executive function put forward by Welsh, Pennington, and Groisser (1991). This model emphasized three components: working memory and planning (measured with the Tower of London task ([Shallice, 1982] and a “noisy book” auditory sequencing task [Hughes, 1998]), inhibition of maladaptive prepotent responses (measured with Luria’s hand game [Hughes, 1998] and a “detour-reaching box” [Hughes & Russell, 1993]), and self-monitoring/attentional flexibility (measured with a pattern reproduction task [Frith, 1971] and a card sort [Hughes, 1998] similar to the WCST). All test scores were converted to success/failure scores following Hughes (1998), and significant differences were observed on four of the six tasks, excluding the “noisy book” and card sort tasks. In the first of two follow-up studies with these same children, poor executive function and poor language ability were both related to violent fantasy (Dunn & Hughes, 2000). A second follow-up study of these “hard-to-manage” children involved playing with a teacher-nominated best friend in a room with an attractive toy (Hughes, White, Sharpen, & Dunn, 2000). Scores on the “detour-reaching box” and the Tower of London tasks, seen as the two tasks most sensitive to antisocial behavior by Hughes et al. (1998), were used to estimate executive function. Although the “hard-to-manage” children did not display more verbal aggression relative to the control group, they snatched toys, teased, bullied, broke rules or the other child’s toy, hurt the other child, and engaged in sex play (i.e., with a doll, themselves, or the other child in the room) more often. Verbal
Executive Function in Aggression
317
abilities were negatively related to hurting the other child (an observational index of physical aggression) within the “hard-to-manage” group. The executive function scores were associated with frequency of observed antisocial behavior in this group even after controlling for verbal abilities and maternal education, but this was not the case for theory of mind scores. This latter study suggests that executive function problems, as opposed to verbal abilities or theory of mind, may be specifically related to antisocial behavior in children observed in the laboratory. However, this study does not focus on physical aggression in the analyses involving executive function nor does it take into account the potential behavior problems of the teacher-nominated best friend. Thus, it is not clear whether antisocial peers, which are more likely to associate with one another, have contributed to the escalation of behavior problems in the “hard-to-manage” group. Peer dynamics is a different but relevant issue (see Boivin, Vitaro, & Poulin, Chapter 18, this volume).
Studies Using Specific Measures of Executive Function in Clinic Samples Most of the preschool studies reviewed so far have focused on community samples. One study examined clinic-referred preschool boys (ages 47–68 months) with ODD and ADHD, an ODD group without ADHD, and a matched nonclinic control group (Speltz, DeKlyen, Calderon, Greenberg, & Fisher, 1999). In addition to being referred and having a primary DSM-III-R (American Psychiatric Association, 1987) diagnosis of ODD, these children also had a T score > 65 on the CBCL aggression scale. Executive function was operationalized with two measures, one of motor planning (adapted from the Luria–Nebraska Battery for Adults) and the other of verbal fluency (taken from the McCarthy Scales). Verbal abilities were measured with the Expressive One Word Picture Vocabulary Test—Revised (Gardner, 1990), the Peabody Picture Vocabulary Test (Dunn & Dunn, 1981), the Wechsler Preschool and Primary Scales of Intelligence—Revised (WPPSI-R; Wechsler, 1989) arithmetic and comprehension subtests, a Test of Early Reading Ability, and a test of vocabulary for affective states. Visual–motor abilities were measured with the Beery Visual-Motor Integration Test (Beery, 1982) and the WPPSI-R block design and picture completion subtests. Although confirmatory factor analyses showed that this dimensional grouping of variables was appropriate, test scores were summed within dimension and reliabilities were ranging from poor to moderate, that is, from .41 for executive function, .52 for visual– motor, and .68 for verbal abilities. Clinic-referred boys combined did more poorly on all dimensions and on test behavior in univariate analyses. However, contrary to the authors’ hypothesis, the effect for executive function was accounted for by test behavior (even after removing test inattention and persistence) and by verbal abilities. However, when breaking down the clinicreferred group into its subtypes, univariate comparison of the groups with ODD-ADHD and ODD-only revealed poorer verbal (about 0.4 mean standard difference) and executive function abilities (about 0.6 mean standard dif-
318
DETERMINANTS OF AGGRESSION
ference, a moderate effect size) in the comorbid group. Although it was not tested for significance, the ODD-only and comparison groups differed by about 0.5 SD in verbal abilities and 0.4 SD on executive function. Further, the contrast for executive function was not controlled for performance in verbal abilities. In the absence of an ADHD-only group, support for a specific relation of ODD-ADHD to poor executive function remains elusive. In a second follow-up study, these same clinic-referred boys also generated more aggressive solutions to social information–processing vignettes than nonclinic boys, independently from ADHD (Coy, Speltz, DeKlyen, & Jones, 2001). However, verbal abilities at age 57 months, which were then worse in clinic-referred boys, were no longer associated with clinic status at age 82 months. Unfortunately, executive function at age 57 months was not used as a predictor of later cognitive abilities or behavior in this follow-up study. Interestingly, and surprisingly, several physically aggressive behaviors (e.g. occasional shoving or hitting, or hitting of parent or sibling during a tantrum) were explicitly discounted for establishing the diagnosis of CD in that study so that the diagnosis would emphasize deliberate confrontation. If most expressions of aggression are reactive instead of proactive, and poor executive function is more likely to be associated with reactive aggression (Giancola, Moss, Martin, Kirisci, & Tarter, 1996), then these studies may be missing important opportunities to understand the cognitive underpinnings of CD. (See Vitaro & Brendgen, Chapter 9, this volume, for a more detailed discussion of reactive and proactive aggression.) Further, despite the longitudinal design of the second study, we do not know anything about the stability of problems across the three measurement times. The longitudinal study did not examine whether the children with comorbid ODD-ADHD at time 1 were more likely to have an externalizing disorder at time 3. The effects of having been referred and presumably being under professional care are also unknown.
THE UTILITY OF A PROBLEM-SOLVING FRAMEWORK FOR UNDERSTANDING AGGRESSION We have already emphasized the need for greater specificity of behavior problems that are the object of study instead of broad-band categories such as delinquency, externalizing disorders, or even CD. We also suggest that measurements across time may be more powerful in determining children at highest risk for later violence. This concept is becoming the standard, as many of the studies reviewed thus far indicate. Similar methodological features need to be systematically implemented for the study of cognitive aspects involved in the behavior problems of interest. In their empirical review, Pennington and Ozonoff (1996) point to difficulties (1) in isolating components of executive function and (2) in considering also the many factors that may disrupt executive function performance on traditional tasks. Because executive function appears to be more strongly and specifically involved in physical aggression
Executive Function in Aggression
319
than does general cognitive development, indexed, for example, by IQ tests, at least in adolescence and young adulthood, we propose that executive function in preschoolers be examined from a functional perspective according to a problem-solving framework. This approach, which follows Luria (e.g., 1973), views executive function as a functional construct that makes reference to (but cannot be equated with) the psychological processes involved in self-regulated, goal-directed problem solving (Zelazo, Carter, Reznick, & Frye, 1997). Like all functions, executive function is defined solely in terms of what it accomplishes—not the mechanisms that accomplish the function. From this perspective, the various subfunctions of problem solving, from initially representing a problem, planning a strategy, executing the plan, to eventually evaluating the adequacy of an attempted solution, can all be seen as contributing to the function of goaldirected problem solving. For example, when searching for a hidden toy, children need to represent the problem, select a plan for action, actually execute the plan, and then evaluate the outcome. Treating executive function as a functional construct does not explain executive function (it remains something to be explained), but it does lay the groundwork for an explanation by facilitating the formulation of specific hypotheses regarding the role of basic cognitive processes (e.g., attention, memory, action monitoring) in different aspects of executive function. It also provides a framework for devising more precise measures of executive function (e.g., measures of planning that do not also require rule use). A major advantage of the problem-solving framework is that it integrates diverse executive function processes (and disparate lines of research) into a single, coherent model of components. The problem-solving framework allows one to understand executive function in a way that suggests specific executive function assessments (see Zelazo et al., 1997, for an exhaustive list) and clarifies the way in which different aspects of executive function work together to fulfill the higher-order function of problem solving while avoiding conceptualizing executive function as a homuncular ability (e.g., as a supervisory attentional system; Shallice, 1988). Further, by designing tasks in light of this framework, one can discover more precisely when performance breaks down in the process of problem solving, and this, in turn, bears on specific issues regarding the cognitive bases of unwarranted physical aggression. 1. Do some children engage in physical aggression because they fail to represent a problem adequately? For example, some children may be biased to represent certain situations as threatening because they have an underlying difficulty in attaining a more objective perspective and flexibly reinterpreting several situations more adaptively. 2. Alternatively, or in addition, some children may fail to plan or think ahead properly. They may fail to anticipate the negative consequences of a strategy that can be aggressive, for example. 3. In some cases, children may understand the rules that govern a social
320
DETERMINANTS OF AGGRESSION
situation (e.g., “I should not hit others” or “I should be quiet during class” or “I should do as I am asked by my teacher”) but fail to use these rules much in the same way that children fail to use rules that they know on tests of rule use (Zelazo & Frye, 1998). 4. Finally, difficulties with action evaluation may contribute to inflexibility and persistence of inappropriate responding despite relevant feedback (e.g., failing to learn from negative consequences of behavior).
THE UTILITY OF DISTINGUISHING BETWEEN HOT AND COLD EXECUTIVE FUNCTION In addition to the distinctions between different phases of problem solving, one might usefully differentiate between the relatively “hot” affective aspects of executive function associated more with ventral and medial regions of prefrontal cortex and the more purely cognitive “cool” aspects associated more with dorsolateral prefrontal cortex (Zelazo & Müller, 2002). Whereas cool executive function is more likely to be elicited by relatively abstract, decontextualized problems, hot executive function is required for problems that involve the regulation of affect and motivation (i.e., regulation of basic limbic system functions). This characterization of hot executive function in contradistinction to cool executive function is consistent with several recent proposals regarding the function of orbitofrontal cortex (e.g., Damasio, 1994; Rolls, 1999). For example, based on single-cell recordings of neurons in orbitofrontal cortex together with neuroimaging data and evidence that damage to orbitofrontal cortex impairs performance on simple tests of object reversal and extinction, Rolls (e.g., 1999) suggests that orbitofrontal cortex is required for the flexible representation of the reinforcement value of stimuli. Damasio proposed a rather different theory, the somatic marker theory (e.g., 1994). According to this theory, orbitofrontal cortex is required for processing learned associations between affective reactions and specific scenarios, and this processing plays a crucial but often overlooked role in decision making. Despite their differences, however, both approaches capture the fact that the control of thought and the control of action depend on different cortical systems, depending on whether or not such control occurs in motivationally significant contexts. Although a strict orbitofrontal cortex account of antisocial behavior, and particularly physical aggression, is questionable (Séguin, 2004), the relevance of hot executive function is not. Traditionally, research on executive function in human beings has focused almost exclusively on cool executive function, using measures such as the WCST. Recently, however, there has been growing interest in hot executive function. The hot aspect of executive function appears to be involved, for example, in theory of mind (Zelazo & Müller, 2002), delay of gratification (Mischel, Shoda, & Rodriguez, 1989), and affective decision making (Damasio, 1994). An experimental approach to hot executive function using tasks such as delay of gratification or the children’s gambling task (Kerr & Zelazo, 2004)
Executive Function in Aggression
321
might reveal that hot executive function may also play an important role in regulating physical aggression. However, an approach that integrates hot and cool executive function is likely to be most informative. As shown earlier, cool executive function was related to behavioral control—that is, following the rule not to play with a forbidden but attractive object (Cole et al., 1993). The toy’s attractiveness invokes hot executive function, but it seems likely that children regulate their behavior in this type of situation in part by translating the problem into a more cool, decontextualized problem (Mischel et al., 1989). We also found that boys with a history of physical aggression played more cards associated with rewards and punishments than nonaggressive boys even after controlling for cool executive function (in this case working memory; Séguin et al., 2002). Whereas one group of stable aggressive boys did so perhaps because of an underlying attentional problem, card perseveration was related to emotional regulation (in this case neuroticism) in a group of unstable aggressive boys. Individuals high in neuroticism are more likely to be responsive to the hot or motivationally significant aspects of task demands, which increases their likelihood of dysregulated behavior.
CONCLUDING COMMENTS Implicit in the problem-solving framework is the idea that deficits in physically aggressive preschoolers are not necessarily global, although they may be for a subset of children. Much research in childhood and adolescence has attempted to specify the nature of the cognitive deficit in physical aggression (e.g., Séguin et al., 1999; Séguin et al., 1995) and hyperactivity (e.g., Pennington, Bennetto, McAleer, & Roberts, 1996). As this review shows, only a handful of studies of preschoolers have attempted to break down cognition into subcomponents (e.g., Raine et al., 2002) and isolate executive function deficits from more global deficits (e.g., Cole et al., 1993; Dunn & Hughes, 2000; Hughes et al., 1998; Speltz et al., 1999). Still, many of the poor executive function abilities identified in physical aggression (i.e., working memory; Giancola, 1995; Séguin et al., 1995) and in hyperactivity (i.e., planning, inhibition; Pennington & Ozonoff, 1996) can be reformulated in terms of the problem-solving framework. For example, much research on child aggression (not specifically physical) has focused on the way in which children represent problems. This research emphasizes distortions, biases, beliefs, and schemas (Huesmann, 1997), such as those that are the object of cognitive therapies (Lochman, Magee, & Pardini, 2003) and that often fall under the label of social information processing (Pardini, Lochman, & Frick, 2003). In their formulation of an information-processing model, Crick and Dodge (1994) emphasized four stages: encoding, attribution, problem solving, and response evaluation. This model is strongly based on distortions and biases that predispose an individual to automatically construe problems in preconceived ways. Yet there are deficit models that emphasize more fundamental processing diffi-
322
DETERMINANTS OF AGGRESSION
culties (Séguin et al., 1995; Wallace & Newman, 1997), which could be more directly based in brain anatomy or physiology but still interact with individual motivational predispositions and which may be influenced greatly by early experience (Mezzacappa, Kindlon, & Earls, 2001). Thus the problem-solving framework suggests that a child’s inability to take another person’s perspective in a relatively cool task (underlying deficit) is unlikely to change a more emotional bias. Conversely, the child who can show flexibility on a relatively cool task (no underlying deficit) is more likely to be able to change perspective on a problem that is laden with affective tone and for which that child shows a maladaptive response bias. Several studies have also examined family characteristics and early experiences that influence not only preschoolers’ behavior problems but also their cognitive development (e.g., Heller et al., 1996; Mezzacappa et al., 2001; Morrell & Murray, 2003; Pierrehumbert, Ramstein, Karmaniola, & Halfon, 1996). As such studies refine their measurement of executive function and behavior problems, we will not only be able to see how family characteristics affect specific as opposed to global behavior problems but will also be able to identify family characteristics that may more specifically influence the subphases of the problem-solving framework. As they become embedded in longitudinal designs, a study of the components of executive function will better inform early intervention strategies that will target families or the child at risk. In conclusion, there is growing evidence that poor executive function is involved in externalizing problems as early as the preschool years. However, the findings lack consistency largely because there is considerable variability in study objectives, and in cognitive and behavioral assessments across studies. Further, very few studies have explicitly examined preschool executive function or focused on physical aggression. However, on the basis of studies spanning childhood to adulthood, we would expect that physical aggression alone, or in combination with other externalizing problems, becomes more specifically related to poor executive function. There is thus a need for studies that explicitly chart the joint development of physical aggression, its comorbid characteristics, and executive function. Those studies would also do well to consider predisposing and maintenance factors that affect this joint development in order to better inform prevention and intervention studies.
REFERENCES Achenbach, T. M. (1999). The Child Behavior Checklist and related instruments. In M.E.Maruish (Ed.), The use of psychological testing for treatment planning and outcomes assessment (2nd ed., pp. 429–466). Mahwah, NJ: Erlbaum. Achenbach, T. M., & Edelbrock, C. S. (1983). Manual for the Child Behavior Checklist and Revised Child Behavior Profile. Burlington, VT: University Associates in Psychiatry. Achenbach, T. M., & Edelbrock, C. S. (1986). Manual for the Teacher’s Report Form and
Executive Function in Aggression
323
the teacher version of the Child Behavior Profile. Burlington, VT: University Associates in Psychiatry. Achenbach, T. M., Edelbrock, C. S., & Howell, C. T. (1987). Empirically based assessment of the behavioral/emotional problems of 2- and 3-year-old children. Journal of Abnormal Child Psychology, 15, 629–650. Aguilar, B., Sroufe, L. A., Egeland, B., & Carlson, E. (2000). Distinguishing the early-onset/persistent and adolescence-onset antisocial behavior types: From birth to 16 years. Development and Psychopathology, 12, 109–132. American Psychiatric Association (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev.). Washington, DC: Author. Baddeley, A., & Della Sala, S. (1996). Working memory and executive control. Philosophical Transactions of the Royal Society of London, 351, 1397–1403. Becker, M. G., Isaac, W., & Hynd, G. W. (1987). Neuropsychological development of nonverbal behaviors attributed to “frontal lobe” functioning. Developmental Neuropsychology, 3, 275–298. Beery, K. E. (1982). Revised administration, scoring, and teaching manual for the developmental test of visual–motor integration. Cleveland, OH: Modern Curriculum Press. Behar, L., & Stringfield, S. (1974). A behavior rating scale for the preschool child. Developmental Psychology, 10, 601–640. Bonino, S., & Cattelino, E. (1999). The relationship between cognitive abilities and social abilities in childhood: A research on flexibility in thinking and co-operation with peers. International Journal of Behavioral Development, 23, 19–36. Broidy, L., Nagin, D. S., Tremblay, R. E., Bates, J. E., Brame, R., Dodge, K. A., et al. (2003). Developmental trajectories of childhood disruptive behaviors and adolescent delinquency: A six site, cross-national study. Developmental Psychology, 39, 222–245. Chelune, G. J., & Baer, R. A. (1986). Developmental norms for the Wisconsin Card Sorting Test. Journal of Clinical and Experimental Neuropsychology, 3, 219–228. Cole, P. M., Usher, B. A., & Cargo, A. P. (1993). Cognitive risk and its association with risk for disruptive behavior disorder in preschoolers. Journal of Clinical Child Psychology, 22, 154–164. Coy, K., Speltz, M. L., DeKlyen, M., & Jones, K. (2001). Social-cognitive processes in preschool boys with and without oppositional defiant disorder. Journal of Abnormal Child Psychology, 29, 107–119. Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social informationprocessing mechanisms in children’s social adjustment. Psychological Bulletin, 115, 74–101. Damasio, A. R. (1994). Descartes’ error. New York: Harper & Collins. Delis, D. C., Squire, L. R., Bihrle, A., & Massman, P. (1992). Componential analysis of problem-solving ability: Performance of patients with frontal lobe damage and amnesic patients on a new sorting test. Neuropsychologia, 30, 683–697. Denckla, M. B., & Reiss, A. L. (1997). Prefrontal–subcortical circuits in developmental disorders. In N. A. Krasnegor, G. R. Lyon, & P. Goldman-Rakic (Eds.), Development of the prefrontal cortex: Evolution, neurobiology, and behavior (pp. 283– 293). Baltimore: Paul H. Brookes. Dietz, K. R., Lavigne, J. V., Arend, R., & Rosenbaum, D. (1997). Relation between intelligence and psychopathology among preschoolers. Journal of Clinical Child Psychology, 26, 99–107.
324
DETERMINANTS OF AGGRESSION
Dunn, J., & Hughes, C. (2000). “I got some swords and you’re dead!” Violent fantasy, antisocial behavior, friendship and moral sensibility in young children. Child Development, 72, 491–505. Dunn, L. M., & Dunn, L. M. (1981). PPVT: Peabody Picture Vocabulary Test—Revised: Manual for forms L and M. Circle Pines, MN: American Guidance Services. Eyberg, S. M., & Robinson, E. A. (1983). Conduct problem behavior: Standardization of a behavioral rating scale. Journal of Clinical Child Psychology, 12, 347–354. Farrington, D. P., & Loeber, R. (2000). Epidemiology of juvenile violence. Child and Adolescent Psychiatric Clinics of North America, 9, 733. Frith, U. (1971). Studies in pattern detection in normal and autistic children. II. Reproduction and production of colour sequences. Journal of Experimental Child Psychology, 10, 120–135. Frye, D., Zelazo, P. D., Brooks, P. J., & Samuels, M. C. (1996). Inference and action in early causal reasoning. Developmental Psychology, 32, 120–131. Gardner, M. F. (1990). Expressive One Word Picture Vocabulary Test—Revised. Novato, CA: Academic Therapy. Giancola, P. R. (1995). Evidence for dorsolateral and orbital prefrontal cortical involvement in the expression of aggressive behavior. Aggressive Behavior, 21, 431–450. Giancola, P. R., Mezzich, A. C., & Tarter, R. E. (1998). Executive cognitive functioning, temperament, and antisocial behavior in conduct-disordered adolescent females. Journal of Abnormal Psychology, 107, 629–641. Giancola, P. R., Moss, H. B., Martin, C. S., Kirisci, L., & Tarter, R. E. (1996). Executive cognitive functioning predicts reactive aggression in boys at high risk for substance abuse: A prospective study. Alcoholism: Clinical and Experimental Research, 20, 740–744. Grant, D. A., & Berg, E. A. (1948). A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Weigl-type card-sorting problem. Journal of Experimental Psychology, 38, 404–411. Heaton, R. K. (1981). A manual for the Wisconsin Card Sorting Test. Odessa, FL: Psychological Assessment Resources. Heller, T. L., Baker, B. L., Henker, B., & Hinshaw, S. P. (1996). Externalizing behavior and cognitive functioning from preschool to first grade: Stability and predictors. Journal of Clinical Child Psychology, 25, 376–387. Huesmann, L. R. (1997). The role of social information processing and cognitive schema in the acquisition and maintenance of habitual aggressive behavior. In R.G.Geen & E. Donnerstein (Eds.), Human aggression: Theories, research, and implications for policy. New York: Academic Press. Hughes, C. (1998). Executive function in preschoolers: Links with theory of mind and verbal ability. British Journal of Developmental Psychology, 16, 233–253. Hughes, C., Dunn, J., & White, A. (1998). Trick or treat? Uneven understanding of mind and emotion and executive dysfunction in “hard-to-manage” preschoolers. Journal of Child Psychology and Psychiatry, 39, 981–994. Hughes, C., & Russell, J. (1993). Autistic children’s difficulty with mental disengagement from an object: Its implications for theories of autism. Developmental Psychology, 29, 498–510. Hughes, C., White, A., Sharpen, J., & Dunn, J. (2000). Antisocial, angry, and unsympathetic: “Hard to manage” preschoolers’ peer problems, and possible cognitive influences. Journal of Child Psychology and Psychiatry, 41, 169–179.
Executive Function in Aggression
325
Kandel, E., & Freed, D. (1989). Frontal-lobe dysfunction and antisocial behavior: A review. Journal of Clinical Psychology, 45, 404–413. Kaufman, A. S. (1983). Kaufman Assessment Battery for Children (K-ABC). Circle Pines, MN: American Guidance Services. Kaufman, A. S., & Kaufman, N. (1977). Clinical evaluation of young children with the McCarthy Scales. Boston: Allyn & Bacon. Kerr, A., & Zelazo, P. D. (2004). Development of “hot” executive function: The Children’s Gambling Task. Brain and Cognition, 55, 148–157. Lahey, B. B., Loeber, R., Burke, J. D., Rathouz, P. J., & McBurnett, K. (2002). Waxing and waning in concert: Dynamic comorbidity of conduct disorder with other disruptive and emotional problems over 17 years among clinic-referred boys. Journal of Abnormal Psychology, 111, 556–567. Lilienfeld, S. O. (2003). Comorbidity between and within childhood externalizing and internalizing disorders: Reflections and directions. Journal of Abnormal Child Psychology, 31, 285–291. Lochman, J. E., Magee, T. N., & Pardini, D. A. (2003). Cognitive behavioral interventions for children with conduct problems. In M. Reinecke & D. A. Clark (Eds.), Cognitive therapy over the lifespan: Theory, research and practice (pp. 441–476). Cambridge, UK: Cambridge University Press. Loeber, R., Burke, J. D., & Lahey, B. B. (2002). What are adolescent antecedents to antisocial personality disorder? Criminal Behaviour and Mental Health, 12, 24– 36. Luria, A. R. (1973). The working brain. New York: Basic Books. Marcovitch, S., & Zelazo, P. D. (1999). The A-not-B error: Results from a logistic metaanalysis. Child Development, 70, 1297–1313. Mariani, M. A., & Barkley, R. A. (1997). Neuropsychological and academic functioning in preschool boys with attention deficit hyperactivity disorder. Developmental Neuropsychology, 13, 111–129. Matson, D. E., & Fisher, M. (1991). A comparison of internalizers, externalizers, and normals using the WISC—R and Wisconsin Card Sorting Test. Journal of Psychoeducational Assessment, 9, 140–151. McCarthy, D. (1970). McCarthy Scales of children’s abilities. New York: Psychological Corporation. McGee, R., Partridge, F., Williams, S., & Silva, P. A. (1991). A twelve-year follow-up of preschool hyperactive children. Journal of the American Academy of Child and Adolescent Psychiatry, 30, 224–232. Mezzacappa, E., Kindlon, D., & Earls, F. J. (2001). Child abuse and performance task assessments of executive functions in boys. Journal of Child Psychology and Psychiatry, 42, 1041–1048. Milich, R., Balentine, A. C., & Lynam, D. R. (2001). ADHD combined type and ADHD predominantly inattentive type are distinct and unrelated disorders. Clinical Psychology Science and Practice, 8, 463–488. Milner, B. (1963). Effects of different brain lesions on card sorting. Archives of Neurology, 9, 90–100. Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay of gratification in children. Science, 244, 933–938. Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks. Cognitive Psychology, 41, 49–100.
326
DETERMINANTS OF AGGRESSION
Moffitt, T. E. (1990). Juvenile delinquency and attention deficit disorder: Boys’ developmental trajectories from age 3 to age 15. Child Development, 61, 893–910. Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674–701. Moffitt, T. E., & Henry, B. (1989). Neuropsychological assessment of executive functions in self-reported delinquents. Development and Psychopathology, 1, 105– 118. Moffitt, T. E., & Silva, P. A. (1988a). Neuropsychological deficit and self-reported delinquency in an unselected birth cohort. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 233–240. Moffitt, T. E., & Silva, P. A. (1988b). Self-reported delinquency, neuropsychological deficit, and history of attention deficit disorder. Journal of Abnormal Child Psychology, 16, 553–569. Morgan, A. B., & Lilienfeld, S. O. (2000). A meta-analytic review of the relation between antisocial behavior and neuropsychological measures of executive function. Clinical Psychology Review, 20, 113–136. Morrell, J., & Murray, L. (2003). Parenting and the development of conduct disorder and hyperactive symptoms in childhood: A prospective longitudinal study from 2 months to 8 years. Journal of Child Psychology and Psychiatry, 44, 489–508. Nagin, D. S., & Tremblay, R. E. (1999). Trajectories of boys’ physical aggression, opposition, and hyperactivity on the path to physically violent and nonviolent juvenile delinquency. Child Development, 70, 1181–1196. Newman, J. P., & Wallace, J. F. (1993). Diverse pathways to deficient self-regulation: Implications for disinhibitory psychopathology in children. Clinical Psychology Review, 13, 699–720. Norman, D. A., & Shallice, T. (1986). Attention to action: Willed and automatic control of behavior. In R.J.Davidson, G. E. Schwartz, & D. Shapiro (Eds.), Consciousness and self-regulation (pp. 4–18). New York: Plenum Press. Owen, A. M., Herrod, N. J., Clark, J. C., Downey, S. P., Carpenter, T. A., Minhas, P. S., et al. (1999). Redefining the functional organization of working memory processes within human lateral prefrontal cortex. European Journal of Neuroscience, 11, 567–574. Pardini, D. A., Lochman, J. E., & Frick, P. J. (2003). Callous/unemotional traits and social-cognitive processes in adjudicated youths. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 364–371. Parkin, A. J. (1998). The central executive does not exist. Journal of the International Neuropsychological Society, 4, 518–522. Pennington, B. F. (1997). Dimensions of executive functions in normal and abnormal development. In N.A.Krasnegor, G. R. Lyon, & P. Goldman-Rakic (Eds.), Development of the prefrontal cortex: Evolution, neurobiology, and behavior (pp. 265– 281). Toronto: Brookes. Pennington, B. F., Bennetto, L., McAleer, O., & Roberts, R. J. (1996). Executive functions and working memory: Theoretical and measurement issues. In G. R. Lyon & N. A. Krasnegor (Eds.), Attention, memory, and executive function (pp. 327–348). Baltimore: Paul H. Brookes. Pennington, B. F., & Ozonoff, S. (1996). Executive functions and developmental psychopathology. Journal of Child Psychology and Psychiatry, 37, 51–87. Petrides, M., Alivisatos, B., Evans, A. C., & Meyer, E. (1993a). Dissociation of human mid-dorsolateral from posterior dorsolateral frontal cortex in memory processing.
Executive Function in Aggression
327
Proceedings of the National Academy of Sciences of the United States of America, 90, 873–877. Petrides, M., Alivisatos, B., Meyer, E., & Evans, A. C. (1993b). Functional activation of the human frontal cortex during the performance of verbal working memory tasks. Proceedings of the National Academy of Sciences of the United States of America, 90, 878–882. Pierrehumbert, B., Ramstein, T., Karmaniola, A., & Halfon, O. (1996). Child care in the preschool years: Attachment, behaviour problems and cognitive development. European Journal of Psychology of Education, 11, 201–214. Plude, D. J., & Doussard-Roosevelt, J. A. (1989). Aging, selective attention, and feature integration. Psychology and Aging, 4, 98–105. Raine, A., Yaralian, P. S., Reynolds, C., Venables, P. H., & Mednick, S. A. (2002). Spatial but not verbal cognitive deficits at age 3 years in persistently antisocial individuals. Development and Psychopathology, 14, 25–44. Robbins, T. W. (1996). Dissociating executive functions of the prefrontal cortex. Philosophical Transactions of the Royal Society of London, 351, 1463–1470. Robins, L. N. (1991). Conduct disorder. Journal of Child Psychology and Psychiatry, 32, 193–212. Robinson, A. L., Heaton, R. K., Lehman, A. W., & Stilson, D. W. (1980). The utility of the Wisconsin Card Sorting Test in detecting and localizing frontal lobe lesions. Journal of Consulting and Clinical Psychology, 48, 605–614. Rolls, E. T. (1999). The brain and emotion. Oxford, UK: Oxford University Press. Satz, P., & Fletcher, J. M. (1982). Florida Kindergarten Screening Battery. Odessa, FL: Psychological Assessment Resources. Séguin, J. R. (2004). Neurocognitive elements of antisocial behavior: Relevance of an orbitofrontal cortex account. Brain and Cognition, 55, 185–197. Séguin, J. R., Arseneault, L., Boulerice, B., Harden, P. W., & Tremblay, R. E. (2002). Response perseveration in adolescent boys with stable and unstable histories of physical aggression: The role of underlying processes. Journal of Child Psychology and Psychiatry, 43, 481–494. Séguin, J. R., Boulerice, B., Harden, P., Tremblay, R. E., & Pihl, R. O. (1999). Executive functions and physical aggression after controlling for attention deficit hyperactivity disorder, general memory, and IQ. Journal of Child Psychology and Psychiatry, 40, 1197–1208. Séguin, J. R., Nagin, D. S., Assaad, J.-M., & Tremblay, R. E. (2004). Cognitiveneuropsychological function in chronic physical aggression and hyperactivity. Journal of Abnormal Psychology, 113, 603–613. Séguin, J. R., Pihl, R. O., Harden, P. W., Tremblay, R. E., & Boulerice, B. (1995). Cognitive and neuropsychological characteristics of physically aggressive boys. Journal of Abnormal Psychology, 104, 614–624. Sergeant, J. A., Geurts, H., & Oosterlaan, J. (2002). How specific is a deficit of executive functioning for attention-deficit/hyperactivity disorder? Behavioural Brain Research, 130, 3–28. Shallice, T. (1982). Specific impairments of planning. Philosophical Transactions of the Royal Society of London, B298, 199–209. Shallice, T. (1988). From neuropsychology to mental structure. New York: Cambridge University Press. Speltz, M. L., DeKlyen, M., Calderon, R., Greenberg, M. T., & Fisher, P. A. (1999).
328
DETERMINANTS OF AGGRESSION
Neuropsychological characteristics and test behaviors of boys with early onset conduct problems. Journal of Abnormal Psychology, 108, 315–325. Stouthamer-Loeber, M., Loeber, R., Wei, E., Farrington, D. P., & Wikström, P.-O. H. (2002). Risk and promotive effects in the explanation of persistent serious delinquency in boys. Journal of Consulting and Clinical Psychology, 70, 111–123. Stuss, D. T. (1992). Biological and psychological development of executive functions. Brain and Cognition, 20, 8–23. Toupin, J., Déry, M., Pauzé, R., Mercier, H., & Fortin, L. (2000). Cognitive and familial contributions to conduct disorder in children. Journal of Child Psychology and Psychiatry, 41, 333–344. Tremblay, R. E. (2000). The development of aggressive behaviour during childhood: What have we learned in the past century? International Journal of Behavioral Development, 24, 129–141. Tremblay, R. E., Boulerice, B., Harden, P. W., McDuff, P., Pérusse, D., Pihl, R. O., et al. (1996). Do children in Canada become more aggressive as they approach adolescence? In Growing up in Canada (pp. 127–137). Ottawa: Statistics Canada. Tremblay, R. E., Japel, C., Pérusse, D., Boivin, M., Zoccolillo, M., & Montplaisir, J. (1999). The search for the age of “onset” of physical aggression: Rousseau and Bandura revisited. Criminal Behavior and Mental Health, 9, 24–39. Tremblay, R. E., Nagin, D. S., Séguin, J. R., Zoccolillo, M., Zelazo, P. D., Boivin, M., et al. (2004). Physical aggression during early childhood: Trajectories and predictors. Pediatrics, 114, e43–e50. Wallace, J. F., & Newman, J. P. (1997). Neuroticism and the attentional mediation of dysregulatory psychopathology. Cognitive Therapy and Research, 21, 135–156. Waschbusch, D. A. (2002). A meta-analytic examination of comorbid hyperactive–impulsive–attention problems and conduct problems. Psychological Bulletin, 128, 118–150. Wechsler, D. (1989). Manual for the Wechsler Preschool and Primary Scale of Intelligence—Revised. Toronto: Psychological Corporation. Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory of mind development: The truth about false belief. Child Development, 72, 655–684. Welsh, M. C., Pennington, B. F., & Groisser, D. B. (1991). A normative-developmental study of executive function: A window on prefrontal function in children. Developmental Neuropsychology, 7, 131–149. Willcutt, E. G., Pennington, B. F., Chhabildas, N. A., Friedman, M. C., & Alexander, J. (1999). Psychiatric comorbidity associated with DSM-IV ADHD in a nonreferred sample of twins. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 1355–1362. Wolf, M. (1986). Rapid alternating stimulus naming in the developmental dyslexias. Brain and Language, 27, 360–379. Zelazo, P. D., Carter, A., Reznick, J. S., & Frye, D. (1997). Early development of executive function: A problem-solving framework. Review of General Psychology, 1, 198–226. Zelazo, P. D., & Frye, D. (1998). Cognitive complexity and control: II. The development of executive function in childhood. Current Directions in Psychological Science, 7, 121–126. Zelazo, P. D., Frye, D., & Rapus, T. (1996a). An age-related dissociation between knowing rules and using them. Cognitive Development, 11, 37–63.
Executive Function in Aggression
329
Zelazo, P. D., Helwig, C. C., & Lau, A. (1996b). Intention, act, and outcome in behavioral prediction and moral judgment. Child Development, 67, 2479–2492. Zelazo, P. D., & Müller, U. (2002). Executive function in typical and atypical development. In U. Goswami (Ed.), Handbook of childhood cognitive development (pp. 445–469). Oxford, UK: Blackwell. Zelazo, P. D., & Reznick, J. S. (1991). Age-related asynchrony of knowledge and action. Child Development, 62, 719–735. Zelazo, P. D., Reznick, J. S., & Piñon, D. E. (1995). Response control and the execution of verbal rules. Developmental Psychology, 31, 508–517.
DETERMINANTS Language Development OF AGGRESSION and Aggressive Behavior
16 Language Development and Aggressive Behavior G INETTE D IONNE
Aggression has been associated with low language proficiency as early as the second year of life and throughout the lifespan. Although language deficits are not sufficient or necessary correlates of aggressive behavior, they have received much attention as frequently co-occurring with behavior problems (for reviews, see Beitchman, Cohen, Konstantareas, & Tannock, 1996; Cantwell & Baker, 1977, 1987; Cohen, 2001; Donahue, Cole & Hartas, 1994; Gallagher, 1999; Goodyer, 2000; Howlin & Rutter, 1987; Silva, 1987; Stevenson, 1996). This has led to a consensus around the view that the link between language and behavior is functional, not spurious. Just what this functional link is, however, remains unclear. Previous reviews of the association between language and behavior have mainly focused on the variety of behavioral and emotional correlates of language deficits in childhood. One difficulty in attempting to write a chapter specifically on aggression and language is that aggression is rarely directly assessed in developmental psychopathology. Yet aggression, possibly leading to violence and criminality, has a huge social impact. As such, it is important that we understand how it comes about, and language should not be disregarded as an associated factor. This chapter examines empirical evidence and surveys theoretical explanations of the association between language deficits and aggression. The goal is to look at the numbers, then search for the story (or, possibly, stories) behind the numbers. For the sake of brevity, it does not attend to specific syndromes in which aggression and pervasive language deficits are severely incapacitating, as in general developmental delay, pervasive developmental disorders, autism, childhood-onset schizophrenia, psychosis, or structural brain injuries. Where aggression is not assessed per se, as is the case in most studies 330
Language Development and Aggressive Behavior
331
of early development, the review is extended to include conduct disorders. Early-onset conduct disorders have been shown to be particularly characterized by the use of physical aggression (Lahey et al., 1998). Finally, more broadly defined externalizing behavior problems are also included, but with an added caution in regard to making inferences about aggression. In line with the developmental focus of this book, this chapter takes a developmental approach to the review of prevalence data from clinical, epidemiological, and longitudinal studies. As the picture unfolds from descriptive data, theoretical models of the language–aggression association are outlined, with a focus on the empirical evidence available to support these models. As this empirical evidence still remains scarce, this chapter concludes with a look at some considerations that research designs could address to take on the current challenges in understanding the functional links between language and aggression.
THE HISTORICAL PERSPECTIVE The notion that language and behavior are intrinsically linked in humans is hardly new. Although it can be traced back to ancient philosophers, the functional role of language in moral action and social order became a favorite topic of late 17th- and 18th-century philosophers such as Locke and Rousseau. By the early 20th century, the renewed interest from an empirical perspective largely stemmed from observations in clinical settings. British language and literacy specialist Samuel T. Orton concluded, in his classic 1937 essay on dyslexia and language, that communication problems led to secondary behavior and social problems. A few years later, Stella Chess (1944) echoed this view in an early paper in which she presented language disabilities as a factor in “personality distortions” in children. On another front, early Russian psychologists were speculating on possible causal processes linking language to behavior regulation. Pavlov’s (1927) concept of language as a second signaling system, on which excitatory and inhibitory processes operate to direct and control behavior, pointed to the possible role of language in the inhibition of aggressive impulses. Both Luria (1961) and Vygotsky (1962) later formalized the notion that language, spoken or as an inner process, was a necessary means to behavior regulation. Although much of this earlier work was speculative in nature, it created the basis for the next four decades of theorizing and research. In the early 1970s, Montare and Boone (1973) formulated an explicit “language–aggression hypothesis”: They proposed that language was related to physical aggression in two ways—low levels of proficiency linked to higher levels of observable aggression, and high proficiency linked to lower levels of aggression. Their empirical work on a group of children and teenagers from diverse ethnic backgrounds provided mitigated support for the hypothesis (Boone & Montare, 1976), and subsequent investigations of a specific link be-
332
DETERMINANTS OF AGGRESSION
tween aggression and language proficiency have been few and far between (Cole, 2001; Dionne, Tremblay, Boivin, Laplante, & Pérusse, 2003; Mack & Warr-Leeper, 1992; Stattin & Klackenberg-Larsson, 1993; Piel, 1990). During this time frame, general intelligence became the focus of many studies as a correlate of delinquency and criminality, low intelligence more often characterizing violent offenders (see Moffitt, Gabrielli, Mednick, & Schulsinger, 1981). However, as Stattin and Klackenberg-Larsson (1993) point out, early language delays have been shown to be highly correlated with lower verbal and nonverbal intelligence. In this regard, they hypothesize that “the early differences between future delinquents and nondelinquents might be found in the area of language development” (p. 370). Is there support for this assertion? The next sections of this chapter look into empirical evidence of concurrent and longitudinal links between language development and aggression at different periods of development. But first, the terms and criteria used to define atypical development in both fields are addressed and the normative developmental course of language and socialized behavior is outlined to provide a background against which empirical data may be understood.
DEFINING LANGUAGE DEFICITS AND AGGRESSION To study the interface between language and behavior in psychopathology, one must be aware of the variety of terms and criteria used to define atypical development in both fields. In psycholinguistics, a deficit may permeate the whole language system or may affect specific aspects of language differentially. On one hand, general deficits are often referred to as specific language impairments (SLI)—although what is specific to SLI remains somewhat controversial—indicating language proficiency below that which is expected for an otherwise normally developing child’s IQ range. Speech and language difficulties (SLD), on the other hand, is a more general construct that refers to more or less severe nonoptimal functioning in many language-related skills. A variety of general and specific measures of language deficits are used in the prevalence studies reviewed. Basically, language skills may be affected at three different levels: production, content, and use of language. Production problems linked with speech processes such as stuttering or articulation have, for the most part, not been linked with a higher occurrence of conduct problems (Beitchman et al.,1999; Cantwell & Baker, 1987; Gualtieri, Koriath, Van Bourgondien, & Saleeby, 1983; Love & Thompson, 1988). The content and use of language however, have been linked with behavioral problems. The content of language refers to the mastery of a sufficient number of words (lexical development) and their meaning (semantic development) and of the rules that govern how they come together to form sentences and convey meaning. To understand and create these sentences, a child must master phoneme recognition, be able to segment the stream of speech to decipher basic units, and be able to memorize and use these units effectively. Deficits at this level, whether
Language Development and Aggressive Behavior
333
expressive or receptive, are more readily diagnosed because the resulting communication impairment is often obvious. In addition, they are assessed more systematically in screening procedures. Although emotional and social problems appear to be the most prevalent forms of psychopathology in children with deficits in the content of language (Cantwell & Baker, 1977), a few studies suggest that there may be long-term effects of early deficits in regard to violent behavior (Beitchman et al., 1999; Stattin & Klackenberg-Larson, 1993). Finally, use of language comprises the notions of turn taking in communication and use of contextually appropriate and socially efficient language. Deficits in the pragmatics of language, as they are called, are more readily missed because they do not audibly impair everyday communication. Yet, over time, they result in a breakdown in communication with others and unrewarding communication experiences. A few studies point to pragmatics as the prevailing language deficit in aggressive children (Cole, 2001; Mack & Warr-Leeper, 1992; Minutti, 1991). As pragmatics are often overlooked unless audible language presents some concern, it is not surprising to find that a majority of children referred to clinics for externalizing behavior problems have undiagnosed language deficits (Cohen, 2001). Aggression, on the other hand, is rarely assessed directly in most largescale studies dealing with the language–behavior interface. For one thing, aggression is not a psychiatric diagnosis, but rather a criterion associated with conduct or behavior disorder. Most large-scale studies reviewed have dealt with either psychiatric outpatient clinic populations or populations ascertained for language deficits. When behavior problems are considered, they may be labeled as conduct disorders or externalizing problems (including or excluding ADHD) in preschool and school-age children, antisocial personality, or delinquency or registered criminality in adolescents and adults. Any of these labels may or may not imply overt aggression. For instance, although 7 of 15 criteria for conduct disorder refer to overt aggression, a diagnosis of conduct disorder may be given in the absence of an aggression component. In addition, although aggression studies point to different developmental trajectories involving overt, covert, and verbal aggression as well as reactive versus proactive types of aggression, these distinctions have never been addressed in studies looking into comorbid language deficits. Inferences are thus limited with respect to specific forms of aggression. Nonetheless, the convergence or absence thereof, between studies of aggression and those looking at more general forms of externalizing behavior problems, may be of particular interest to decipher what, if any, developmental processes they may share.
PREVALENCE THROUGHOUT DEVELOPMENT How the Developmental Story Unfolds These conceptual limitations aside, the first question to address is the prevalence of these co-occurring problems throughout development. As with many
334
DETERMINANTS OF AGGRESSION
other considerations in developmental psychopathology, most of what we know about co-occurring language and behavior problems is documented for school-age children, and to a much lesser extent, adolescents and adults, often in follow-up studies. Although the exact proportions of overlap vary from study to study, it is estimated that more than 50% of children with behavior problems have language deficits, many of them undiagnosed (Cohen, 1998). Similar proportions of language-impaired children are reported to display behavior problems severe enough to warrant intervention. It is safe to assume that a significant proportion of these behavior problems would include some forms of aggression. Whether the initial focus is on behavior or language, the general outlook is the same: A child with one of these problems is up to three or four times more likely to have the other, than a child in the community at large (Goodyer, 2000; Stevenson, 1996). This childhood snapshot, however, says little about what leads to these compelling statistics and even less about when and how the story unfolds. Which came first? In any attempt to understand co-occurring problems, this question inevitably arises. And as with hens and eggs, the answer in this case may be equally problematic. Decades of research on aggression in humans still leaves the century-old question of the onset of physical aggression unresolved or, at best, with controversial answers (Tremblay, 2000). Psycholinguists, however, have less difficulty in agreeing that early infancy is the vessel of early language development. The first spoken words typically appear around the first birthday, but a child is able to comprehend words and link them to objects somewhat earlier. By the end of the second year most children experience a vocabulary spurt, which generally precedes the onset of word combinations into sentences. By age 3, simple grammatical forms are used to make complete sentences, albeit language remains somewhat telegraphic at this age. At any point in time, a child’s understanding of language, and thus his or her receptive abilities, precede the ability to use the same grammatical forms in expressive language. At ease with the rudiments of spoken language, most 3-year-olds are able to make demands, give information verbally, and have an active role in conversation. With the availability of normed instruments of early language development, we are able to diagnose early delays in the lexical and grammatical content of spoken language, as well as in a child’s ability to understand spoken language. Although most delays before age 3 appear transient (Dale, Price, Bishop, & Plomin, 2003), some 40% of slow talkers still show signs of delay by the end of the preschool years and beyond. Those with receptive delays, who show early difficulties in understanding language, tend to have a worse prognosis. Theoretical disputes aside, the developmental story of aggression shares some interesting milestones with language development. As Goodenough (1931) illustrated years ago, the frequency of aversive types of behavior, from fussing and crying in the first year, to kicking, biting, hitting, and tantrums when physical ability enables such behavior, shows a marked increase from 6
Language Development and Aggressive Behavior
335
months onward to reach a peak between the ages of 18 and 36 months. Goodenough, and others after her (Vygotsky, 1962; Luria, 1961), have argued that a child’s growing verbal ability is instrumental in the following decrease of physically disruptive behaviors after 36 months. This milestone does coincide with the onset of efficient communicative skills in most children, as stated earlier. In an attempt to prospectively track the onset of physical aggression in a representative birth cohort, Tremblay and colleagues (Tremblay et al., 1996) showed that by 17 months, 80% of children were reported to exhibit some form of physically aggressive behavior. From the third year onward, physical aggression was shown to decrease (Tremblay et al., 1996), with a second peak around midadolescence, mainly for boys. As with early language deficits, high levels of physical aggression in infancy tend to remain stable in a proportion of children (Hay, Castle, & Davies, 2000; Keenan & Shaw, 1993), placing them most at risk of following a high aggression trajectory leading to later violence and criminality. Some forms of physical aggression and rough-and-tumble play may be normative during the period fondly known to parents as the “terrible twos.” In most children, after this period, physical aggression is replaced by less overt forms of aggression, namely, verbal and indirect aggression (Bjorkvist, Osterman, & Kaukiainen, 1992). Although minimal language ability is required to use verbal aggression, it remains unclear whether the use of verbal aggression is highest among verbally proficient toddlers. If, as is the case in school-age children, rates of verbal and physical aggression remain associated (Pepler, Craig, & Roberts, 1998), it may be that some physical aggressors add verbal aggression to their repertoires and others move from physical to verbal forms of aggression. In time (it is hoped), language becomes for most children a social tool for increased prosocial interactions. Whatever the case may be, from both the behavior and language perspectives, the early toddler years serve as the center stage for emergent individual differences that may have a profound impact on later adjustment. From that point onward, language abilities continue to progress and physically aggressive behaviors diminish in frequency. At no other moment in life will the frequency of aversive-type behaviors reach the peak and subsequent sharp decline observed in the toddler years; similarly, at no other point in time will language abilities show a growth as staggering as that evidenced in the first 3 years of life. For some children, however, language has not progressed as anticipated at this milestone, and for some, physical aggression continues to attract increasing concern. The question is to what extent are these the same children? Studies offer different answers, using different methodologies at various periods of development. Two separate concerns are generally addressed in regard to prevalence: (1) concurrent behavior and language problems and (2) longitudinal outcomes of earlier language deficits. The following sections offer a review of prevalence studies throughout development.
336
DETERMINANTS OF AGGRESSION
Concurrent Prevalence in Infancy and the Preschool Years The youngest age at which epidemiological data are available on the language– behavior association comes from two recent twin studies addressing the genetic etiology of this association and a small community sample. We return to the etiological answers provided by the genetic studies later. A first question they attend to pertains to the age of onset of this association. How early is there evidence of an association between language deficits and aggression? In the Québec Newborn Twin Study (QNTS), a modest but significant correlation of –.20 was found between expressive vocabulary assessed via a parent checklist and parent reports of physical aggression at 19 months (Dionne et al., 2003). Although there were no sex differences in either vocabulary or aggression at this age, the correlation was slightly higher for boys than for girls. In contrast, when more general conduct problems were considered in relation to vocabulary in the Twin Early Development Study (TEDS), Plomin and colleagues (Plomin, Price, Eley, Dale & Stevenson, 2002) found correlations of –.05 at 24 months for both boys and girls. In addition, languagedelayed children (below the 5th and 10th percentiles) in this sample did not show a higher incidence of conduct problems. There was also no significant association between language and behavior in a small community sample comparing language-delayed (n = 11) and language normal (n = 53) 24month-olds on a CBCL measure of total externalizing behavior (Carson, Klee, Perry, Muskina, & Donaghy, 1998). Subsequent analyses of the Québec sample, however, show this not to be the case when physical aggression is considered. Indeed, posthoc analyses on this sample of 900 twins indicate a significantly higher incidence of physical aggression in 19-month-olds below the 15th percentile on vocabulary (p < .05). Although there appears to be an association, albeit very modest, of delayed language with aggression before age 2, this does not seem to be the case when more general conduct problems are considered. By age 3, prevalence reports markedly differ. In the TEDS sample, correlations at follow-up assessments at 36 and 48 months remain modest: –.13 for boys and –.09 for girls at 36 months and –.18 for boys and –.15 for girls by 48 months. Again, the incidence of conduct problems and total behavior problems is not higher in the delayed children at both ages. This is in sharp contrast with results from an earlier U.K. study by Stevenson and Richman (1978), in which half of the language-delayed children, from a sample of 828 3-year-olds, showed behavior problems. A majority of these problems pertained to general immaturity and overactive behavior. The most frequent language delay pertained to difficulty with language structure. In a later study, Silva, Williams, and McGee (1987) assessed a normative sample of 1,037 3-year-olds in New Zealand. In this study, a higher incidence of behavior problems, as rated by parents and teachers on the Rutter Scale, was found for
Language Development and Aggressive Behavior
337
children with general language delay and comprehension delay but not for those with expressive delay. The variety of language and behavior measures make comparisons and conclusions difficult before age 3, but the general trends in the U.K. and N.Z. studies of 3-year-olds appear sharper by age 4–5. In the late 1970s, Cantwell and Baker (1977) studied behavioral outcomes in incoming patients to a community speech and hearing clinic. Two-thirds of these children were of preschool age (median 4.8 years). The 600 children identified within a 3-year period were divided into three language attainment categories: pure speech disorder, speech and language disorder, and pure language disorder, based on a comprehensive speech and language assessment. Speech-disordered-only children showed fewer concurrent psychiatric disorders. Overt behavioral disorders were identified in 30% of the speech and language group, and in 47% of the pure language group. However, approximately two-thirds of the overt behavioral disorder were related to attention deficits, and the incidence of conduct disorder was only slightly more elevated in these groups than would be expected in the normal population. Cantwell and Baker (1977) cautioned that because a conduct disorder diagnosis requires a long pattern of behavior, it was rarely given for the younger children. This may account for the low incidence. However, as there is a high prevalence of physical aggression in attentiondeficit disorders, the co-occurrence of language deficits and aggression in the late preschool period may have been underestimated in this sample. A stronger association has been shown in one of the most extensive longitudinal investigations of psychopathology in language-delayed children. This study was initiated in a community-based sample of 5-year-olds in Ottawa, Canada, by Joseph Beitchman and his colleagues (Beitchman, Nair, Clegg, Ferguson, & Patel, 1985). At initial assessment, the boys (n = 90) among the 142 children in the language-delayed group showed a higher incidence (13.6% versus 3.4% in controls) of conduct disorder than controls, but not the girls. A recent community study of 4- and 5-year-old preschoolers also reports this gender difference (Ortiz, Stowe, & Arnold, 2001). Lower levels of expressive and receptive vocabulary skills in this sample (n = 56) were strongly associated with disruptive behavior as coded from video observations and from teacher ratings, but for boys only. Although none of these studies focused specifically on aggression, the gender-biased results are consistent with those of previous studies (Beitchman et al., 1985; Dionne et al., 2003). Thus, during infancy and the preschool years, some studies show evidence of a higher incidence of aggressive behaviors in preschoolers with lower language proficiency, particularly for boys. However, other studies report no such evidence. In the studies where there is a significant association in this age range, pure speech impairments are not the culprit. Some studies report a stronger association for receptive deficits, whereas other studies report associations for both receptive and expressive language deficits, mainly for vocabulary. Overall, however, these associations remain very modest.
338
DETERMINANTS OF AGGRESSION
Concurrent Prevalence in Childhood and Adolescence By the early school years, the specificity of language deficits linked with aggression emerges. In one of the earlier childhood studies, Camarata, Hughes, and Ruhl (1988) showed that the language performance of children identified as having mild to moderate behavior disorder was typically 2 or more standard deviations below the normative mean. These children displayed problems with expressive syntax while having a satisfactory knowledge of vocabulary. Results from Cohen and colleagues (Cohen, Davine, Horodezky, Lipsett, & Isaacson, 1993) point to similar language deficiencies linked with use of language more than content of language in aggressive children. They systematically screened 399 psychiatric outpatients ages 4–12 years and found that one-third of the referred children had an unsuspected language impairment. Overall, more than 50% of these children exhibited some form of language impairment, previously identified or not. Although semantics, syntax, and phonology were more affected in the group with previously identified language impairment, a higher proportion of children from the unsuspected language impairment group showed problems with receptive syntax, auditory memory, and fluency. Cohen (1998) later reported that although the prevailing behavior diagnosis for children with language impairments was ADHD, the children with unsuspected language impairment were rated as more aggressive and having more problems with delinquency. These two studies suggest that aggressive children may have more problems with the use and understanding of complex language structures than with lexical knowledge. The specificity of pragmatic difficulties associated with aggression or behavior problems is reported in other studies on smaller samples. In a study of 54 special education second- and third-grade children, Minutti (1991) found that among children in special classes, with either behavioral problems or learning problems, language deficiencies of more than 2 standard deviations below the normative mean identified those having more behavior problems, based on teacher ratings. The most important deficit in this subgroup was associated with difficulty in formulating coherent sentence structures to convey meaning. A more thorough investigation of pragmatic language development in association with physical aggression was conducted in two separate studies by Cole (2001) and Mack and Warr-Leeper (1992). In the study by Cole (2001), aggressive (n = 19) and nonaggressive (n = 26) 8–13-year-old boys were compared on standardized receptive and expressive language measures, as well as measures of syntactic and narrative complexity. Although the groups did not differ on the composite scores of expressive and receptive language, the aggressive boys did perform more poorly on both the syntactic and narrative complexity measures. Mack and Warr-Leeper (1992) had previously identified similar problems in chronic-behavior-disordered boys, even once IQ level was taken into account. In their sample, 16 of 20 boys scored below the average range on at least 4 of 20 language assessments. The main areas of language compromised were use of complex linguistic structures, use of abstract
Language Development and Aggressive Behavior
339
language, and use of concepts. More important, however, for half of these boys, language assessments were below expected levels based on IQ. This suggests that pragmatic and sentence structure difficulties may be associated with the more severe forms of behavior problems and aggression in childhood and early adolescence, over and beyond IQ level. Although these studies offer compelling results, inferences remain limited by the fact that they are based on clinical populations of behaviorally compromised children and that, for the most part, these are mainly boys (from 75– 100% of the samples). Do normative samples of boys and girls replicate the association between language deficits and conduct disorder or aggression in this age range? For the most part, they do. Although total prevalence is higher in boys, the co-occurence rates with language impairments appear to be similar for boys and girls (Stevenson, Richman, & Graham, 1985). In one of the few studies looking simultaneously at physical and verbal aggression, Piel (1990) added another consideration to gender differences. He studied language maturity in association with responses on an apperceptive test of aggressive behavior in 108 second- and third-grade children. Although children’s behaviors were not assessed directly, the best predictor of physically aggressive responses to vignettes was low language maturity, accounting for 20% of the variance in both boys and girls. Sex and social class were not associated with the selection of physically aggressive responses. Interestingly, girls did use verbally aggressive responses more often than boys, and language maturity was associated positively with this form of aggressive response in girls, albeit less strongly (4% of the variance). The more verbally proficient girls were more prone to offer verbally aggressive responses to the vignettes. Even though apperceptive tests have important methodological limitations in generalizing to behavior, these results suggest that the developmental shift from physical aggression to verbal aggression may be mediated by language maturity, yet differently in boys and girls. Tomblin and colleagues (Tomblin, Zhang, Buckwalter, & Catts, 2000) offered yet another consideration in looking at the co-occurence of behavior and language problems in school-age children. They studied language, reading, and behavior in a normative sample of second-grade boys and girls. The overall association between externalizing behavior problems assessed by teachers and parents and a composite language score (including expressive and receptive vocabulary and pragmatic complexity) was –.30. Both were negatively associated with reading problems. Subgroups of children were then identified as language impaired or reading impaired. When investigating the pattern of association between reading, language, and behavior problems at this age, log-linear model fitting indicated that behavior problems in the children with language impairments were mediated by reading problems whereas behavior problems in the reading-impaired group were not mediated by language problems. This suggests that by the early school years, reading difficulties may act as an additional risk factor for behavioral problems, either linked with early language impairment or as a novel liability. Written forms of lan-
340
DETERMINANTS OF AGGRESSION
guage may have specific links with aggression and externalizing behaviors that need to be addressed separately. In sum, by the early school years and adolescence, many studies have documented the link between language proficiency and behavior problems. Although both problems are more prevalent in boys at this age, their rate of co-occurrence appears generally similar across genders. Contrary to what was observed in the preschool years, studies specifically pertaining to aggression show patterns of co-occurences similar to those shown in studies of general behavior problems or conduct disorders. But although the co-occurence may be of the same magnitude, these studies suggest that low pragmatic language skills and difficulties with sentence structure may be more specifically associated with aggression during this period.
Longitudinal Data: Do Early Language Deficits Predict Later Aggression? Probably the most striking findings of the association between language and aggression are those reported in longitudinal lifespan studies. Although language has been a continuous focus in studies of early development, studies in adulthood shifted to low IQ as a concurrent correlate of physical aggression, delinquency, and criminality (Moffitt et al., 1981). Language proficiency may have appeared as an unlikely concurrent candidate by adulthood, as, for the most part, adults seem to have mastered the basic linguistic skills, even the less educated adults. Prospective longitudinal studies, however, tell a different story when early language skills are associated with adulthood outcomes. But first, we consider longitudinal results during childhood. In three studies described earlier, children assessed as preschoolers were followed into later childhood. In all of these studies, early behavior problems (ages 3–4) predicted later behavior problems (at age 8 in Stevenson et al., 1985, and Benasich, Curtiss, & Tallal, 1993; at ages 7, 9, and 11 in Silva et al., 1987). However, early language delays were limited in predicting later behavior problems. In Benasich and colleagues’ U.S. sample of language-impaired preschoolers and controls, neither degree of language impairment at the initial assessment nor the language improvement between assessments predicted behavioral outcomes at age 8. In Stevenson and colleagues’ U.K. sample, low language structure at age 3 did not predict antisocial behavior scores at age 8 but did predict neurotic problems. In the N.Z. sample of Silva and colleagues, however, the general language delay group at age 3 did score significantly higher for total behavior problems on the Rutter Teacher Scale at age 9 and on the Rutter Parent Scale at 11. Although behavior problems remained quite stable, the predictive value of early language impairment appears mitigated in these samples of boys and girls. The story, when looking at late adolescence and early adulthood followups, is quite different. Beitchman’s Canadian sample of language-impaired children initially assessed at age 5 were followed well into late adolescence. Boys and girls with initial language impairement had strikingly divergent
Language Development and Aggressive Behavior
341
trajectories by age 19 (Beitchman et al., 1999). Both speech- and languageimpaired girls showed a higher prevalence of psychiatric disorders than controls, predominantly affective and anxiety disorders. However, antisocial personality disorder was the most prevalent diagnosis for the boys in the language-impaired group. Boys in this group were up to four times more likely to have this diagnosis than controls. Similar results are shown for boys in relation to registered criminality in a longitudinal follow-up of a sample of 122 Swedish boys (Stattin & Klackenberg-Larsson, 1993). Infant measures of language development were conducted at 3, 6, 9, 12, 18, and 24 months. Of the 122 boys, 36 were sporadic (n = 22) or frequent (n = 14) offenders by age 17. Registered criminality was significantly but modestly correlated with 6-, 18-, and 24-month infant language measures (correlations ranging from –.16 to –.19). Moreover, language ratings by a psychologist were obtained at ages 3 and 5. Registered criminality was both related to boys language maturity at age 3 (r = –.16, p < .05) and age 5 (r = –.17, p < .05) and to the mother’s self-reported difficulty in understanding the speech of the child at ages 4 and 5. Intelligence measures were also obtained (ages 3, 5, 8, 11, 14, and 17 years), and all except one (age 8) predicted later registered criminality. However, once SES was controlled for, only the age 3 intelligence measure differentiated offenders from nonoffenders, but 18- and 24-month language ability, 3-year comprehension of language, and 5-year maturity of language use and comprehension of language remained significant predictors of criminality. Thus, early language development appeared as the early aspect of these boys’ lives that better distinguished later offenders from nonoffenders, even once SES and IQ were taken into account. The longitudinal picture these studies paint appears to vary according to two criteria: gender and the specificity of the behavior measures. Longitudinal outcomes for children with early language delays seem to differ considerably as a function of gender. The antisocial trajectory, possibly more directly associated with the use of aggression, is up to 4 times more likely in boys with language impairments than in control boys, but almost 10 times more likely than in girls with language impairments. Girls with language impairments at age 5 do not even reach the rate of antisocial disorder in control boys. The specific behavior measure also distinguishes the predictive value of early language deficits. Results in childhood studies assessing general behavior problems are more inconsistent. However, when criminality and antisocial disorders were targeted, early language impairments distinguished offenders from nonoffenders, beyond SES measures, whereas intelligence measures in childhood did not. Overall, the striking feature of the prevalence studies reviewed is not the strength of the association between language deficits and aggression. The coprevalence may be high in clinical populations, but the linear association remains modest, with correlations ranging at best from –.17 to –.33. Rather, it is the consistency with which results appear in many studies based on a variety of designs, populations, age ranges, and even construct definitions.
342
DETERMINANTS OF AGGRESSION
Is the Association between Language Proficiency and Aggression Linear? A major question arises from the consistency of the modest association between low language proficiency and aggression throughout development. Why are correlations so low when the prevalence of one problem can reach as much as 80% when children have the other? Low correlations, aside from indicating a modest association, could also stem from one of two other situations: Either good language skills are not associated with lower aggression scores—that is, the relationship is not linear—or heterogeneous subgroups, with respect to the association between language skills and aggression, create competing effects that outweigh each other in the assessment of a linear trend. Although studies have looked at the higher co-occurrences of the problematic aspects of language and aggression, few have considered looking at whether good language skills predicted lower aggression scores. This question is important, because if the trend is not linear in nature, the resulting correlations may have been underestimations of the true underlying relationship between language impairment and aggression. In addition, if better language skills are not associated with less aggression and more prosocial skills, then there is little to infer from these studies in terms of prevention. Post hoc analysis of the Québec sample offers some reassurances in that respect. At 19 months, the infants above the 85th percentile for vocabulary development did show a lower incidence of physically aggressive behaviors (p < .05), but more important, they also showed a higher incidence of prosocial behaviors (p < .05). This trend persists at 30 months and is true for both boys and girls. Thus, language proficiency may be playing a protective role in the process of early socialization of behavior. The other possibility for the modest correlations is that physical aggression occurs in some verbally proficient individuals, just as some languageimpaired individuals show no signs of using aggression at higher rates. In other words, language deficits, although they occur more frequently in aggressive individuals, may lead to other behavioral and emotional problems not involving aggression, or to no other specific problem. In our attempts to make sense of co-occurring milestones of language and socialized behavior in toddlerhood, we may have overlooked the possibility that distinct subgroups of aggressive individuals with differing levels of language proficiency may emerge in early development. In an examination of communicative strategies in preschool classroom disputes, Danby and Baker (2001) have shown how dominant boys in such group disputes use a variety of verbal strategies to assert their identity, construct group affiliations, and generate terror in targeted rejected children. Although it is unclear whether these boys are in fact more proficient verbally, it is clear that they are able to use language to attain goals involving some forms of aggression. If such different subgroups exist, the etiology of aggression accompanied by low language proficiency may differ from the etiology of aggression in verbally proficient individuals. Different statistical strategies are needed if we are to distinguish these nonlinear associations
Language Development and Aggressive Behavior
343
and, more important, identify individuals following at-risk developmental trajectories for aggression with or without language impairment. This possible missing piece to the puzzle suggests that we may only be starting to understand the complexity of the interface between language and aggression, and to some degree, between language and socialized behavior.
THEORETICAL MODELS OF THE LANGUAGE–AGGRESSION ASSOCIATION There have been numerous theoretical propositions put forth to explain the language–aggression association. By and large, these fall within two sets of developmental models. Models of the first type are those that assume language deficits and aggression share similar origins. The second type of models are those that contend that one problem leads to the other as a secondary consequence, either via distinct pathways or through a combination of multiple pathways acting in transaction, or that both problems feed on each other during development, creating a consolidation of the co-occurence in time.
Shared Etiology Pathways There are three types of shared etiology models that have been suggested to explain the language–aggression association. First, there are models that imply that both language deficits and aggression may stem from the same genetic and biologically based predispositions (Stevenson, 1996). The origins of both problems would then be organic in nature. Stevenson (1996) has demonstrated that ADHD and reading problems do share a common genetic etiology in school-age children. However, even though language and reading problems often co-occur, and aggression and ADHD often co-occur, there are no studies showing that this generalizes to language deficits and aggression. In fact, both the Québec Newborn Twin Study and the U.K. Twin Early Development Study have found no evidence of shared genetic influences—between physical aggression and delayed vocabulary in the Québec sample, and between general conduct problems and delayed vocabulary in the U.K. sample. The shared genetic liability pathway thus seems unlikely in early childhood. However, as genetic influences on most traits tend to increase with time, a possible shared genetic etiology may emerge only later. In addition, empirical tests of this model should take into account the possibility that the etiology of the association of behavior and language may vary, depending on the specific behavior and language deficits. Second, there are models that suggest at-risk environmental features can lead to both unsocialized behavior and poor cognitive-intellectual development, including or secondary to poor language skills. In this macro-social pathway, parenting styles and harsh economic conditions are viewed as environmental liabilities that steer children into trajectories of poorly socialized behavior and leave them poorly stimulated in language development. The link
344
DETERMINANTS OF AGGRESSION
between aggression and parenting has been documented in numerous studies, such as in work by Diana Baumrind (1993) and Gerald Patterson and his colleagues (Patterson, 1982; Patterson, Reid, & Dishion, 1992) at the Oregon Social Learning Center. These have generally demonstrated that inconsistant and harsh parenting, with poor child supervision, create a context within which a child is more susceptible to develop behavior problems. In a similar fashion, but within a completely different field, psycholinguists have demonstrated how parental input (moderated by such factors as socioeconomic status, parental education levels, and neighborhood safety, to name a few) has a profound effect on children’s language development (Storch & Whitehurst, 2001). Michael Tomasello (2003) offered compelling empirical evidence of the effect of input in tasks of non-word learning in helping a child learn new words to designate novel objects. Others have shown how parents’ reading to children (Lyytinen, Laakso, & Poikkeus, 1998; Sénéchal, LeFèvre, Thomas, & Daley, 1998), and parents’ rephrasing of early language mistakes (Bohannon & Stanowicz, 1989) also play a role in the development of language in children. The absence of these forms of early language stimulation is more likely to be accompanied by parental discipline lacking in attention to a child’s needs. Such parental discipline may be characterized by harshness and lack of appropriate supervision, often associated with behavior problems in children, and may account for the association of problems regarding both language and behavior. Finally, the co-occurence of both traits has been suggested to be linked to the same cognitive deficit, such that poor judgment and/or lack of ability to understand others’ perspectives may be accompanied by language difficulties but also heighten the risk for social exclusion and deviant behaviors. The next section of this chapter explores in more detail the theoretical propositions involving a causal contribution of language to the development of aggression.
Causal Pathways Causal models of the language–behavior association have received more attention than their shared etiology counterparts. The prevailing hypothesis in this respect suggests that low language proficiency makes one more liable to use aggression (Cohen, 2001). Two studies have conducted direct tests of this hypothesis, the Québec Newborn Twin Study (Dionne et al., 2003) and an intervention study conducted more than three decades ago (Slaby & Crowley, 1977). In the Québec study, the investigators compared models specifying shared etiological contributions and models testing the significance of unidirectional and reciprocal contribution paths from language to aggression at 19 months. These analyses were conducted using structural equation modeling and the additional statistical flexibility offered by genetically informative data. Results indicated no significant shared etiology component, whether genetic or linked to environments shared by twins at this age, except for a unique environmental component specific to twins within families. However,
Language Development and Aggressive Behavior
345
causal genetic models revealed that a causal path from vocabulary to physical aggression could account totally for the covariance between the two. Assumptions of the genetic methodology should, however, be taken into consideration in generalizing these results. In addition, the results remain specific to the time frame for which models were tested. The most compelling evidence of causal effects comes from experimental designs using and assessing interventions. Many preschool programs targeting at-risk children have an incorporated intervention focus on language and literacy, yet these rarely supply a proper control group to assess the efficacy of interventions. In a study conducted in the late 1970s (Slaby & Crowley, 1977), when the language–aggression hypothesis attracted attention, disruptive children were identified in classroom settings. Half of them received an intervention based on increasing their linguistic skills in social contexts, the other half did not. Teachers reported that the children having received the intervention had significantly decreased their use of disruptive behavior in class and aggression in the schoolyard. There was no significant change in the nonintervention group. The authors concluded that these results were indicative of a causal contribution of enhanced language proficiency toward more socialized behavior. Because there was not a separate group receiving intervention targeting the behavior problems, it is difficult to conclude that regulating behavior does not also contribute to better language learning. Although there were many methodological limitations to this small study, it serves to highlight the necessity of conducting experimental intervention studies to assess causal pathways. Another limitation of these two studies is that although they suggest that language proficiency can affect behavior, they do not address the mechanisms possibly involved. Three major views on how language deficits can lead to behavior problems, including aggression but not specific to it, have been outlined. They are referred to here as the micro-social pathway, the self-regulation pathway, and the social-cognition pathway. The first of these views assume that language plays a significant role in sustaining fulfilling social interactions. This micro-social pathway, along with others reviewed later, has been formalized and explored by Brinton, Fujiki, and colleagues (Brinton & Fujiki, 1993; Brinton, Fujiki & McKee, 1998). In one study, language-delayed school-age children were compared with controls in regard to their negotiation skills in a series of negotiating sequences with two other peers. As compared with controls, the language-delayed children contributed significantly less and developmentally lower-level negotiation strategies to their teams. Interestingly, the language-delayed children did not speak less often than controls. Rather, the maturity level of their verbal contributions to the team was what distinguished them from their peers. The authors caution that lack of efficient negotiation skills may place the languagedelayed children at risk for confrontations with peers and lead them to adopt more disruptive behaviors. Brinton and Fujiki have also documented the fact that language-delayed
346
DETERMINANTS OF AGGRESSION
children show more emotion regulation problems (Fujiki, Brinton, & Clarke, 2002). Cohen (2001) points out that language is an important means by which one controls one’s behavior and emotions. This notion goes back to early propositions by Luria (1961) and Vygotsky (1962) to the effect that inner language is a necessary tool to inhibit antisocial impulses, among other things. Emotion regulation and self-regulation are generally viewed as requiring complex linguistic tools such as the ability to analyze social situations, organize thoughts about one’s own emotions, and plan behavior according to social rules. Given this central role of language in emotion regulation, in the self-regulation pathway, children with language difficulties are expected to be at a disadvantage (Gallagher, 1999). There is some evidence showing that language-delayed children are more at risk of also having regulation problems. In a study of 82 school-age children, half of whom were language delayed, Fujiki and colleagues (Fujiki et al., 2002) reported lower scores overall for the language-delayed children on the Emotion Regulation Checklist, especially for boys. Although emotion regulation problems may stem from overregulation as well as underregulation (Cole, Michel, & O’Donnell-Teti, 1994), it may be expected that the underregulated children would be the ones at risk for aggression. Empirical evidence is, however, not yet available to support the assumption that the contribution of language delays to aggression is mediated by underregulation problems. Social cognition pathways offer yet other explanations for possible mediators of language deficits leading to aggressive behaviors. Among these, Dodge’s social information-processing model suggests that aggression develops as a consequence of deficits in social problem-solving abilities. The model assumes that there are six steps involved in the social problem-solving process. Once a child has (1) encoded cues, the child must (2) interpret those cues before deciding (3) what his or her goal is, (4) constructing his or her response, (5) choosing it, and (6) acting on it (Crick & Dodge, 1994). Although aggressive children tend to construct more aggressive responses and act on them, Lochman and Dodge (1994) have shown that what mainly distinguishes them from nonaggressive children is the tendency to perceive hostile intent in ambiguous situations. This presumably causes them to act on these perceived hostile intents by choosing aggressive responses. It is less clear, however, how language delay may lead a child into this attribution biases pathway. If, as Cohen (2001) points out, language-delayed children have had more negative social encounters, and have more difficulty in making complex analyses because of their linguistic limitations, they may in fact be more at risk of forming hostile attributions. However, this needs to be documented empirically. In addition, the attribution biases pathway seems an unlikely explanation for aggression that appears as early as late infancy. The same can be said about another approach to the social cognition pathway. Theory of mind has been receiving much attention in recent years as a possible central cognitive impairment, linked to autism in particular. Theory of mind involves understanding that individuals may have different views
Language Development and Aggressive Behavior
347
about a given situation, and that sometimes one may act on the basis of an assumption that appears true, which is in fact a false belief (Dunn, 1996). Children begin to be able to consider another person’s perspective, and understand that thoughts, beliefs, and emotions are determined by a person’s unique perspective, at about age 4. Autistic children do not master this task (Wimmer & Perner, 1983). Happé and Frith (1996) have shown that children with social/emotional problems have more problems mastering this task as well, and Cutting and Dunn (1999) have shown theory of mind mastery to be more difficult for language-delayed children. However, there is no evidence to support a theory of mind pathway by which the effects of language delay on socio-emotional problems and/or aggression are mediated. Finally, although it has received much less attention, the alternative causal pathway, whereby aggression leads to secondary language delays, has not been ruled out. For instance, aggressive individuals could be less sensitive to linguistic stimulation in their environment or too disrupted by their own behavior to benefit from this stimulation, leaving them with linguistic deficits they are unable to overcome. As Cantwell and Baker (1977) pointed out decades ago, the main impediment to this hypothesis has been that language delays have always appeared to precede the onset of aggressive behaviors in development. But because overt aggression has rarely been assessed around the time language appears, their conclusion may have been premature. Until proven otherwise, we should consider the possibility that aggression may play the causal role at one moment of development, and not at another, or both may influence each other reciprocally. For example, language-based formal schooling represents a specific time frame within which specific demands are placed on children, requiring both good language skills and social skills. If low language skills appear to drive aggressive behavior in early childhood, can we infer that low language skills drive aggressive behavior once schooling starts? Or does aggression lead to social interactions and choices of peers that have adverse consequences on schooling and limit the sophistication of language skills that comes with schooling? If this is the case, concurrent language skills may not be the best focus for intervention with adult delinquents or violent criminals, but may be targeted in early development to promote prosocial development. For now, we are unable to discern whether any or all of these alternatives are at play.
LOOKING FOR ANSWERS As this review illustrates, research designs of the past decades have been mainly concerned with documenting the prevalence of and, to a lesser extent, exploring the etiology of the language–behavior association. The initial clinical observations of Orton (1937) and Chess (1944) hold true across many samples, large and small, from the normative to the clinical, concurrently and longitudinally. Research on aggression, although more scarce, also suggests
348
DETERMINANTS OF AGGRESSION
language and aggression are associated. That much we know. Any answers to the question concerning how language and aggression become linked during development, however, are at best, tentative for the time being. For the most part, we are left to speculate, given incomplete correlational information. To be more efficient, research designs will need to incorporate measures of mediating variables, such as economic considerations, concurrent measures of IQ, literacy, and discipline-oriented parental practices, to assess their specific contributions in developmental pathways. In addition, aggression should be targeted specifically if designs seek to understand the specificity of the language–aggression association. Similarly, a survey of language skills should include the full array of linguistic abilities. To assess possible shared physiological contributions to both problems, lower-level cortical functioning, such as executive functions and time processing, could be assessed. Obviously, gender differences in both prevalence and developmental mechanisms need to considered. Finally, the true test of causality rests on experimental designs. To scan the full array of possibilities, such designs should divide groups to consider the etiology and course of aggression and language deficits separately, as well as in association, and study the impact of interventions on both aspects longitudinally.
REFERENCES Baumrind, D. (1993). The average expectable environment is not good enough: A response to Scarr. Child Development, 64, 1299–1317. Beitchman, J. H., Cohen, N. J., Konstantareas, M. M., & Tannock, R. (1996). Language, learning, and behavior disorders: Developmental, biological, and clinical perspectives. New York: Cambridge University Press. Beitchman, J. H., Douglas, L., Wilson, B., Johnson, C., Young, A., Atkinson, L., et al. (1999). Adolescent substance use disorders: Findings from a 14-year follow-up of speech and language impaired and control children. Journal of Clinical Child Psychology, 28, 312–321. Beitchman, J. H., Nair, R., Clegg, M., Ferguson, B., & Patel, P. G. (1985). Prevalence of psychiatric disorders in children with speech and language disorders. Journal of American Academy of Child Psychiatry, 25, 528–535. Benasich, A. A., Curtiss, S., & Tallal, P. (1993). Language, learning, and behavioral disturbances in childhood: A longitudinal perspective. Journal of the American Academy for Child and Adolescent Psychiatry, 32, 585–594. Bjorkqvist, K., Osterman, K., & Kaukiainen, A. (1992). The development of direct and indirect aggressive strategies in males and females. In K. Bjorkqvist & P. Niemela (Eds.), Of mice and women: Aspects of female aggression (pp. 51–64). San Diego: Academic Press. Boone, S. L., & Montare, A. (1976). Test of the language–aggression hypothesis. Psychological Reports, 39, 851–857. Brinton, B., & Fujiki, M. (1993). Clinical forum: Language and social skills in the schoolage population: Language, social skills, and socioemotional behavior. Language, Speech, and Hearing Services in Schools, 24, 194–198.
Language Development and Aggressive Behavior
349
Brinton, B., Fujiki, M., & McKee, L. (1998). Negotiation skills of children with specific language impairment. Journal of Speech, Language, and Hearing Research, 41, 927–940. Camarata, S., Hughes, C., & Ruhl, K. (1988). Mild/moderate behaviorally disordered students: A population at risk for language disorders. Language, Speech, and Hearing Services in Schools, 19, 191–200. Cantwell, D. P., & Baker, L. (1977). Psychiatric disorder in children with speech and language retardation: A critical review. Archives of General Psychiatry, 34, 583–591. Cantwell, D. P., & Baker, L. (1987). Clinical significance of children with communication disorders: Perspectives from a longitudinal study. Journal of Child Neurology, 2, 257–264. Carson, D. K., Klee, T., Perry, C. K., Muskina, G., & Donaghy, T. (1998). Comparison of children with delayed and normal language at 24 months of age on measures of behavioral difficulties, social and cognitive development. Infant Mental Health Journal, 19, 59–75. Chess, S. (1944). Developmental language disability as a factor in personality distortion in childhood. American Journal of Orthopsychiatry, 14, 483–490. Cohen, N. J. (1998). Language impairments and psychopathology [Letter to the editor]. Journal of the American Academy for Child and Adolescent Psychiatry, 37, 461–462. Cohen, N. J. (2001). Language development and psychopathology in infants, children and adolescents. Thousand Oaks, CA: Sage. Cohen, N. J., Davine, M., Horodezky, N., Lipsett, L., & Isaacson, L. (1993). Unsuspected language impairment in psychiatrically disturbed children: Prevalence and language and behavioral characteristics. Journal of American Academy for Child and Adolescent Psychiatry, 32, 595–603. Cole, D. (2001). Narrative development in aggressive boys. Behavioral Disorders, 26, 332–343. Cole, P. M., Michel, M. K., & O’Donnell-Teti, L. (1994). The development of emotion regulation and dysregulation: A clinical perspective. Monographs of the Society for Research in Child Development, 59(Serial No. 240). Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social informationprocessing mechanisms in children’s social adjustment. Psychological Bulletin, 115, 74–101. Cutting, A. L., & Dunn, J. (1999). Theory of mind, emotion understanding, language, and family background: Individual differences and interrelations. Child Development, 70, 853–865. Dale, P. S., Price, T. S., Bishop, D. V. M., & Plomin, R. (2003). Outcome of early language delay: I. Predicting persistent and transient delay at 3 and 4 years. Journal of Speech Language and Hearing Research, 46(3), 544–560. Danby, S., & Baker, C. D. (2001). Escalating terror: Communicative strategies in a preschool classroom dispute. Early Education and Development, 12, 343–358. Dionne, G., Tremblay, R., Boivin, M., Laplante, D., & Pérusse, D. (2003). Physical aggression and expressive vocabulary in 19–month-old twins. Developmental Psychology, 39, 261–273. Donahue, M., Cole, D., & Hartas, D. (1994). Links between language and emotional/behavioral disorders. Education and Treatment of Children, 17, 246–254. Dunn, J. (1996). Children’s relationships: Bridging the divide between cognitive and social development. Journal of Child Psychology and Psychiatry, 37, 507–518.
350
DETERMINANTS OF AGGRESSION
Fujiki, M., Brinton, B., & Clarke, D. (2002). Emotion regulation in children with specific language impairment. Language, Speech, and Hearing Services in Schools, 33, 102– 111. Gallagher, T. M. (1999). Interrelationships among children’s language, behavior, and emotional problems. Topics in Language Disorders, 19, 1–15. Goodenough, F. L. (1931). Anger in young children. Minneapolis: University of Minnesota Press. Goodyer, I. M. (2000). Language difficulties and psychopathology. In D. V. M. Bishop & L. B. Leonard (Eds.), Speech and language impairments in children: Causes, characteristics, intervention, and outcome (pp. 227–244). Philadelphia: Psychology Press. Gualtieri, T., Koriath, U., Van Bourgondien, M., & Saleeby, N. (1983). Language disorders in children referred for psychiatric services. Journal of the American Academy of Child and Adolescent Psychiatry, 22, 165–171. Happé, F., & Frith, L. (1996). Theory of mind and social impairment in children with conduct disorder. British Journal of Developmental Psychology, 14, 385–398. Hay, D. F., Castle J., & Davies L. (2000). Toddlers’ use of force against familiar peers: A precursor of serious aggression? Child Development, 71, 457–467. Howlin, P., & Rutter, M. (1987). The consequences of language delay for other aspects of development. In W. Yule & M. Rutter (Eds.), Language development and disorders (pp. 271–294). Oxford, UK: Blackwell. Keenan, K., & Shaw, D. S. (1993). The development of aggression in toddlers: A study of low income families. Journal of Abnormal Child Psychology, 22, 53–77. Lahey, B. B., Pelham, W. E., Stein, M. A., Loney, J., Trapani, C., Nugent, K., Kipp, H., Schmidt, E., Lee, S., Cale, M., Gold, E., Hartung, C. M., Willcutt, E., & Baumann, B. (1998). Validity of DSM-IV attention-deficit/hyperactivity disorder for younger children. Journal of the American Academy for Child and Adolescent Psychiatry, 37, 695–702. Lochman, J. E., & Dodge, K. A. (1994). Social-cognitive processes of severely violent, moderately aggressive, and nonaggressive boys. Journal of Consulting and Clinical Psychology, 62, 366–374. Love, A., & Thompson, M. G. G. (1988). Language disorders and attention deficit disorders in young children referred for psychiatric services. Journal of Orthopsychiatry, 58, 52–63. Luria, A. R. (1961). The role of speech in the regulation of normal and abnormal behavior. Oxford, UK: Pergamon Press. Lyytinen, P., Laasko, M.-L., & Poikkeus, A.-M. (1998) Parental contribution to child’s early language and interest in books. European Journal of Psychology of Education, 13(3), 297–308. Mack, A. E., & Warr-Leeper, G. A. (1992). Language abilities in boys with chronic behavior disorders. Language, Speech, and Hearing Services in Schools, 23, 214– 223. Miniutti, A. M. (1991). Language deficiencies in inner-city children with learning and behavioral problems. Language, Speech, and Hearing Services in Schools, 22, 31–38. Moffitt, T. E., Gabrielli, W. F., Mednick, S. A., & Schulsinger, F. (1981). Socioeconomic status, IQ, and delinquency. Journal of Abnormal Psychology, 90, 152–156. Montare, A., & Boone, S. L. (1973, February). Language and aggression, an exploratory study among black, Puerto Rican youth. Paper presented at the meeting of the Educational Research Association, New Orleans.
Language Development and Aggressive Behavior
351
Ortiz, C., Stowe, R., & Arnold, D. H. (2001). Parental influence on child interest in shared picture book reading. Early Childhood Research Quarterly, 16, 263–281. Orton, S. T. (1937). Reading, writing and speech problems in children. New York: Norton. Patterson, G. R. (1982). Coercive family processes. Eugene, OR: Castalia. Patterson, G. R., Reid, J. B., & Dishion, T. (1992). Antisocial boys. Eugene, OR: Castalia. Pavlov, I. P. (1927). Conditioned reflexes (G. Anrep, Trans.). London: Oxford University Press. Pepler, D. J., Craig, W. M., & Roberts, W. L. (1998). Observations of aggressive and nonaggressive children on the school playground. Merrill-Palmer Quarterly, 44, 55–76. Piel, J. A. (1990). Unmasking sex and social class differences in childhood aggression: The case for language maturity. Journal of Educational Research, 84, 100–106. Plomin, R., Price, T. S., Eley, T. C., Dale, P. S., & Stevenson, J. (2002). Associations between behaviour problems and verbal and nonverbal cognitive abilities and disabilities in early childhood. Journal of Child Psychology and Psychiatry, 43, 619–633. Sénéchal, M., LeFèvre, J.-A., Thomas, E., & Daley, K. (1998). Differential effects of home literacy experiences on the development of oral and written language. Reading Research Quarterly, 32, 96–116. Silva, P. A. (1987). Epidemiology, longitudinal course and some associated factors: An update. In W. Yule & M. Rutter (Eds.), Language development and disorders (pp. 1–15). Oxford, UK: Blackwell. Silva, P. A., Williams, S., & McGee, R. (1987). A longitudinal study of children with developmental language delay at age three: Later intelligence, reading and behaviour problems. Developmental Medicine and Child Neurology, 29, 630–640. Slaby, R. G., & Crowley, C. G. (1977). Modification of cooperation and aggression through teacher attention to children’s speech. Journal of Experimental Child Psychology, 23, 442–458. Stattin, H., & Klackenberg-Larsson, I. (1993). Early language and intelligence development and their relationship to future criminal behavior. Journal of Abnormal Psychology, 102, 369–378. Stevenson, J. (1996). Developmental changes in the mechanisms linking language disabilities and behavior disorders. In J. Beitchman, N. Cohen, M. Konstantareas, & R. Tannock (Eds.), Language, learning and behaviour disorders (pp. 78–100). Cambridge, UK: Cambridge University Press. Stevenson, J., & Richman, N. (1978). Behaviour, language and development in threeyear-old children. Journal of Autism and Childhood Schizophrenia, 8, 299–313. Stevenson, J., Richman, N., & Graham, P. (1985). Behaviour problems and language abilities at three years and behavioural deviance at eight years. Journal of Child Psychology and Psychiatry, 26, 215–230. Storch, S., & Whitehurst, G. J. (2001). The role of family and home in the literacy development of children from low-income backgrounds. In P. R. Britto & J. BrooksGunn (Eds.), The role of family literacy environments in promoting young children’s emerging literacy skills: New directions for child and adolescent development (pp. 53–77). San Francisco: Jossey-Bass. Tomasello, M. (2003). Constructing a language: A usage-based theory of language acquisition. Cambridge, MA: Harvard University Press. Tomblin, J. B., Zhang, X., Buckwalter, P., & Catts, H. (2000). The association of reading
352
DETERMINANTS OF AGGRESSION
disability, behavioral disorders, and language impairment among second-grade children. Journal of Child Psychology and Psychiatry, 41, 473–482. Tremblay, R. E. (2000). The development of aggressive behaviour during childhood: What have we learned in the past century? International Journal of Behavioural Development, 24, 129–141. Tremblay, R. E., Boulerice, B., Harden, P. W., McDuff, P., Pérusse, D., Pihl, R. O., & Zoccolillo, M. (1996). Do children in Canada become more aggressive as they approach adolescence? In Human Resources Development Canada and Statistics Canada (Eds.), Growing up in Canada: National Longitudinal Survey of Children and Youth (pp. 127–137). Ottawa: Statistics Canada. Vygotsky, L. S. (1962). Thought and language. New York: Wiley. Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and the constraining function of wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128.
DETERMINANTS Intergenerational Transmission OF AGGRESSION of Aggression
17 The Intergenerational Transmission of Aggression and Antisocial Behavior M ARK Z OCCOLILLO, E LISA R OMANO, D AVID J OUBERT, T ANIA M AZZARELLO, S YLVANA C ÔTÉ , M ICHEL B OIVIN , D ANIEL P ÉRUSSE , and R ICHARD E. T REMBLAY
In this chapter the intergenerational transmission of childhood-onset persistent aggression (CPA) is reviewed. The evidence for both genetic and environmental risk is presented. The association between parental antisocial behavior and environmental risk factors is examined, and some new data from a large longitudinal study are presented. The difficulties posed by the strong association between parental antisocial behavior and environmental risk factors and the lack of adequate control for parental antisocial behavior and gene–environment correlation in past studies are also discussed. The chapter concludes with a discussion of how future studies of risk factors for childhoodonset persistent aggression can be improved by incorporating assessments of parental antisocial behavior when examining environmental risk and by studies of the early environment of children with antisocial parents.
CHILDHOOD-ONSET PERSISTENT AGGRESSION AND THE PROBLEM OF DEFINING AGGRESSION A recent review of the development of aggressive behavior has noted that defining aggression in studies has been problematic: “The aggregation of different types of aggressive behaviours, and the aggregation of aggressive behaviours with different forms of antisocial behaviour clearly creates an important problem for a developmental science aiming to understand the origin and development of these behaviours” (Tremblay, 2000, p. 130). For example, the aggression scale for one of the few longitudinal and intergenerational studies 353
354
DETERMINANTS OF AGGRESSION
of aggressive behavior included the following items: disobeys teacher, gives dirty looks, makes up stories and lies, does things that bothers others, gets in trouble (Huesmann, Eron, Lefkowitz, & Walder, 1984). Another problem in regard to studying aggression has been requiring an intent to harm as part of the definition. This would preclude studies of aggressive behavior in infants where intent to harm cannot be assessed, and of aggression following anger or fear as “anger and fear [that] lead to reactions which are clearly not under the control of anyone’s will” (Tremblay, 2000, p. 131). Nonetheless, despite these problems with defining aggression, several longitudinal studies of physical aggression in children (rigorously defined as fighting, bullying, or threatening others) have all concluded that there is a small group of children who are persistently aggressive from preschool into adolescence or adulthood (Broidy et al., 2003). Furthermore, children in this high-aggression group are much more likely to engage in violent delinquency than nonaggressive children. It is this childhood-onset persistent aggression (CPA) that is the focus of this chapter. Studies of the intergenerational transmission of rigorously defined CPA are few. There are many more studies of more broadly defined childhoodonset and persistent antisocial behavior (CPASB), which includes both aggressive and nonaggressive antisocial behaviors (Zoccolillo, Price, Hwu, & Ji, 1998; Moffitt, Caspi, Rutter, & Silva, 2001; Robins, 1966). This group has been given different names depending on the domain of research, including the diagnoses of antisocial personality disorder (APD) (within the DSM-IV (American Psychiatric Association, 1994) diagnostic framework) and childhoodonset and life course persistent antisocial disorder (Moffitt et al., 2001). Individuals who are severely and persistently aggressive also engage in less serious and nonviolent antisocial behaviors (Zoccolillo et al., 1998; Robins, 1996; Moffitt et al., 2001; Loeber, Burke, Lahey, Winters, & Zera, 2000). Conversely, among both men and women who met full DSM-III diagnostic criteria for APD in a general population survey of the prevalence of psychiatric disorders in the United States, 85% of men and 83% of women with APD met the criterion for violent behavior defined as follows: hit or threw things at partner more than once, spanked a child hard enough to bruise, had more than one physical fight (other than with spouse) (Robins & Regier, 1991, p. 261). Children with rigorously defined CPA are also more likely to engage in nonviolent delinquency (Broidy et al., 2003). Studies that focus on the more broadly defined CPASB are also informative in regard to CPA and are included in this chapter. Studies of rigorously defined CPA are highlighted.
WHAT CAUSES CHILD-ONSET PERSISTENT ANTISOCIAL BEHAVIOR OR AGGRESSION? Two recent studies have focused specifically on risk factors for childhood persistent aggression. In one study, risk factors for high versus low trajectories of
Intergenerational Transmission of Aggression
355
physical aggression, defined using three items (fights with other children; kicks, bites, or hits other children; bullies or intimidates other children), were examined in boys of ages 6–15 from a high-risk sample (Nagin & Tremblay, 2001). Maternal low education status, teenage mother, and family not intact were environmental variables that were significantly associated with trajectories of high aggression. In a study of infants’ trajectories of aggression from 17 to 42 months of age in a random population sample, using the same items for aggression, three trajectories of aggression were found, including one group with stable high aggression (Tremblay et al., 2004). Risk factors for this trajectory of high stable aggression included young siblings, low family income, early motherhood, mother’s own history of conduct problems, maternal smoking during pregnancy, high family dysfunction, and maternal coercive parenting. Loeber and Farrington (2000) have recently reviewed the literature on young children who commit crime, and listed 40 factors across five domains (child, family, school, peers, and neighborhood). Hinshaw has reviewed issues related to process and mechanism in regard to externalizing behavior (Hinshaw, 2002). In commenting on the list of risk factors generated by Loeber and Farrington, he notes “The sheer size of this list, which must be considered only a partial consideration of potential risk factors, betrays the field’s lack of ability to synthesize or to tell a fully coherent story about the development and maintenance of externalizing behavior.” In regard to methodologic issues and promising practices for the field, he warns investigators of psychosocial influences: “Environmental factors may well account for independent variance in the field; but the typical means of inferring their effects (i.e., correlating a socialization variable with a child outcome) are sufficiently confounded and uninformative that considerably more sophisticated tests are in order.” As to promising avenues of investigation, he highlights the importance of genetically informative designs and conceptually driven investigations of interactions and transactions across multiple levels of influence.
CPASB/CPA IS A FAMILIAL DISORDER A major, and often neglected, aspect of CPASB/CPA is that it is a familial disorder. A recent meta-analysis has examined the association separately between maternal and paternal antisocial personality disorder and offspring externalizing disorders and found significant associations between both mother and father antisocial personality disorder and child externalizing disorder (Connell & Goodman, 2002). The weighted mean r for mothers was .17 and for fathers .16, or small effect sizes. A number of intergenerational studies have found that CPASB/CPA runs in families. As noted earlier, maternal history of conduct problems was a significant risk factor for persistent high levels of aggression in infants.
356
DETERMINANTS OF AGGRESSION
Why Does Antisocial Behavior and Aggression Run in Families? Familial transmission can be due to genes, environment, or both. Genetically informative designs can disentangle genetic and environmental effects through twin and adoption studies (Plomin, 1997; Lahey et al., 1998). Environmental risk (e.g., hostile parenting) in parents with antisocial disorder is likely to affect all children in a family. If shared environment effects are found, this suggests (but does not prove) that early-onset antisocial disorder runs in families, in part due to cultural transmission. A longitudinal twin study has attempted to address the specific contributions of genetic influences, shared environment influences, and nonshared environmental influences separately for aggressive and nonaggressive antisocial behavior (Eley, Lichtenstein, & Moffitt, 2003). This study is of particular interest because both aggression and nonaggressive antisocial behavior, child versus adolescent onset, and persistence were examined. Twin pairs (n = 1,232) from the population-based Swedish Twin Registry were assessed at two different ages (8–9 years and 13–14 years). Aggressive antisocial behavior was rated by a parent using the Aggression subscale of the CBCL, and nonaggressive antisocial behavior was rated using the Delinquency subscale of the CBCL. At both ages aggressive and nonaggressive antisocial behavior were highly correlated. At age 8–9 aggressive behavior was highly heritable, with little evidence for shared environment. Nonaggressive antisocial behavior was influenced by both genes and shared environment. At age 13–14 both aggressive and nonaggressive antisocial behavior were influenced by genes and shared environment. Continuity in aggression was largely mediated by genetic influences, and continuity in nonaggressive antisocial behavior was mediated by both genes and shared environment. A limitation of this study is that the measure of aggression included nonaggressive items such as arguing a lot, stubborn, and jealous. A meta-analysis of 51 twin and adoption studies was conducted recently (Rhee & Waldman, 2002). Antisocial behavior was operationalized four different ways: (1) diagnosis of antisocial personality disorder or conduct disorder by DSM criteria, (2) criminality or delinquency by official records, (3) behavioral aggression, and (4) antisocial behavior (antisocial behaviors but not operationalized as diagnoses). For diagnosis, aggression, and antisocial behavior, the models that included additive genetic effects, shared environmental effects, and nonshared environmental effects were the best-fitting models, with genetic effects explaining between 44 and 47% of the variance.
CPASB/CPA and Environmental (Nongenetic) Risk Given the evidence that there is a shared environment effect on CPASB and, possibly, aggression, what might the risk factors be? This is a difficult question to answer because of possible gene–environment correlations. In regard to environmental risk factors, three questions can be asked:
Intergenerational Transmission of Aggression
357
1. What are the potential environmental (nongenetic) mechanisms by which antisocial behavior may be transmitted within families and which occur early enough (prenatally to kindergarten) to precede or coincide with the first signs of early-onset antisocial disorder or aggression? 2. Are these risk factors confounded with genetic risk for early antisocial disorder or aggression, as measured by either parental antisocial behavior or aggression or through genetically informative designs? 3. Have researchers adequately controlled for gene–environment correlations in examining environmental risk factors for child antisocial behavior? Five specific environmental risk factors are discussed briefly in this chapter: having a young or teen mother, maternal cigarette smoking during pregnancy, maternal sensitivity and attachment, coercive parenting practices, and child maltreatment. A composite risk of multiple adversity is also discussed. These risk factors are chosen because they are believed to be the final pathway through which other factors operate, are supported by developmental theory, are potentially modifiable and therefore candidates for intervention studies, or, if causal, open up a new field of investigation (e.g., prenatal exposure to tobacco smoke). In addition, they also illustrate the difficulties of disentangling gene from environment effects in studying the intergenerational transmission of early-onset antisocial disorder.
Young or Teenage Parenting In two longitudinal studies of trajectories of physical aggression, teen motherhood was a significant risk factor for trajectories of high aggression. This was the case for aggression trajectories between ages 6 and 15 years (Nagin & Tremblay, 2001) and 7 to 42 months of age (Tremblay et al., 2004). For boys ages 6–15, having a teen mother also predicted which boys would be in a chronic high aggression path versus a high but declining path. Young maternal age at birth is associated with serious offspring antisocial behavior (Nagin, Pogarsky, & Farrington, 1999; Wakschlag et al., 2000; Martin & Burchinal, 1992; Jaffee, Caspi, Moffitt, Belsky, & Silva, 2001a; Hardy et al., 1997; Christ et al., 1990; Fergusson & Lynskey, 1993). This association is strongest for maternal age at first birth rather than maternal age per se (Wakschlag et al., 2000; Jaffee et al., 2001a).
Maternal Cigarette Smoking during Pregnancy Maternal smoking was a significant risk factor for stable high aggression trajectories in infancy (Tremblay et al., 2004). A number of research studies have suggested that maternal prenatal tobacco smoking is a risk factor for conduct disorder or related constructs (criminality, delinquency, and violent criminal-
358
DETERMINANTS OF AGGRESSION
ity) (Olds, 1997; Fergusson, Woodward, & Horwood, 1998; Rasanen et al., 1999; Weissman, Warner Wickramaratne, & Kandel, 1999; Brennan, Grekin, & Mednick, 1999; Wakschlag, Pickett, Cook, Benowitz, & Leventhal, 2002; Wakschlag et al., 1997). A recent study has found an interaction between maternal sensitivity and maternal smoking on the development of conduct disorder in offspring (Wakschlag & Hans, 2002).
Maternal Responsiveness and Attachment Maternal responsiveness to infant behaviors is broadly defined as action by the mother that is contingent on a cue from the infant and is appropriate in its type, timing, and intensity to the needs of the infant. Unlike coercive parenting behaviors, which are dependent on response to oppositional behaviors seen in toddlers and older children, maternal responsiveness can be assessed near birth and in the context of infant behaviors that are unrelated to aggression or oppositional behavior. It is also believed that maternal responsiveness provides the foundation for the development of behavioral regulation and social competence in children (Wakschlag & Hans, 1999). Maternal responsiveness during infancy (rated at 4, 12, and 24 months) has been shown in one study to predict offspring disruptive behavior (diagnoses of DSM-III-R conduct disorder or oppositional defiant disorder or a symptom count of criteria from these disorders) in middle childhood. The mothers were either opiate addicts or controls for low-income neighborhoods (Wakschlag & Hans, 1999). In a subsequent study on the same sample, maternal responsiveness was found to interact with exposure to cigarette smoke prenatally in predicting offspring conduct problems (Wakschlag & Hans, 2002). Although early research on the role of attachment in the development of aggression has highlighted the importance of the avoidant category (Renken, Egeland, Marvinney, Mangelsdorf, & Sroufe, 1989), more recent work has uncovered an association between attachment disorganization, specifically, and aggressive behaviors in children. In a sample of high-risk infants, clinically elevated levels of aggressive behavior at age 5 were found in 44% of disorganized children but in only 5% of the secure group (Lyons-Ruth, Alpern, & Repacholi,1993). The association between attachment disorganization and aggression was observed in the same sample 2 years later (age 7). Another study found that the majority (60%) of children classified as disorganized also showed clinically elevated levels of aggressive behavior, as compared with 17% of their peers classified as secure (Shaw, Owens, Vondra, & Keenan, 1996). Although attachment disorganization seems to be a significant contributor to the development of aggression in children, recent evidence indicates that it is better considered in interaction with other risk factors. In one study, attachment disorganization predicted child aggression only in the context of problems in information processing (“mild mental lag”) (Lyons-Ruth, Easter-
Intergenerational Transmission of Aggression
359
brooks, & Cibelli, 1997). Infants both classified as disorganized at age 1 and perceived by their mothers as having a difficult temperament at age 2 were particularly at risk for manifesting high levels of aggression at school age (Shaw et al., 1996).
Coercive Parenting Practices A test of one of the most comprehensive models in the field of child aggression (Patterson, 1982; Reid & Patterson, 1989, Eddy, Leve, & Fagot, 2001) provided evidence for a link between coercive maternal behavior and aggressive behavior for both boys and girls. Maternal coercive behavior and lack of affectionate behavior predicted both high and increasing levels of aggressive behavior, especially in boys (McFadyen-Ketchum, Bates, Dodge, & Pettit, 1996). In research on aggression in toddlers, maternal negative dominance in interaction with male gender and difficult (dysregulated) temperament was a predictor of aggression in the child (Rubin, Hastings, Chen, Stewart, & McNichol, 1998). Observed aggression and mother-reported externalizing problems were associated significantly with dysregulated temperament only for boys with mothers who demonstrated relatively high levels of negative dominance. In a study designed to clarify the influence of socioeconomic factors (operationalized as mother’s and father’s years of education and occupational status) in the development of conduct problems and aggressiveness in young children, several socialization indicators predicted externalizing and aggressive behavior (Dodge, Pettit, & Bates, 1994). Of this set, the strongest individual contribution to the prediction of aggressive behavior in the children came from the level of exposure to harsh discipline as experienced by the child. In a study of 46 parent-referred problem children, negative and directive maternal behaviors were predictive of maternal ratings of aggressive behavior in the child (Campbell, Breaux, Ewing, & Szumoski, 1986). The authors concluded that negative and coercive mother–child interactions are significant contributors to the long-term maintenance of early-onset aggressive behavior. Maternal coercive parenting at 5 months predicted a stable high trajectory of infant aggression from 17 to 42 months (Tremblay et al., 2004).
Child Maltreatment Child maltreatment has been associated with a number of problematic developmental outcomes, especially physical aggression and aggression-related behaviors. A number of researchers have found that physical abuse, in particular, is predictive of elevated levels of aggression in toddlers, preschoolers, and school-age children (Dodge, Bates, & Pettit, 1990; Dodge, Pettit, Bates, & Valente, 1995; Downey & Walker, 1992; Erickson, Egeland, & Pianta, 1989; Malinosky-Rummel & Hansen, 1993). Moreover, physical abuse during childhood has been linked to aggression-related behaviors, such as delinquency, antisocial behavior, and criminal activity, during adolescence and
360
DETERMINANTS OF AGGRESSION
adulthood (Herrenkohl, Egolf, & Herrenkohl, 1997; Muller & Diamond, 1999; Widom, 2001). A recent longitudinal birth cohort study suggests that child maltreatment may interact with a common gene on the X chromosome for monoamine oxidase A (MAOA) to produce CPA, at least in boys (Caspi et al., 2002). In that study, 36.8% of the males had the low-MAOA-activity form of the gene. By itself, the low form did not confer risk for antisocial behavior or for being maltreated, as compared with the high-MAOA-activity form. Maltreatment alone, however, did confer some increased risk. The low-MAOA gene, in interaction with maltreatment, had a significant and meaningful increased risk for antisocial behavior, defined four different ways: conduct disorder, conviction for violent offenses, a continuous score of disposition toward violence, and a continous score measure of antisocial personality disorder.
Multiple Adversity Some studies suggest that it is the cumulation of risk factors that is important in the development of early-onset antisocial disorder (Rutter, Tizard, Yule, Graham, & Whitmore, 1976; Shaw, Vondra, Hommerding, Keenan, & Dunn, 1994; Fergusson, Horwood, & Lynskey, 1994). The evidence is strong that the more adversity, the greater the probability of developing CPA/CPASB. The effect of multiple risks is above and beyond that of the additive effects of each risk, suggesting that it is multiple risk itself that is possibly causal.
The Association between Parental Antisocial Behavior and Risk Factors Do these risk factors occur more commonly in families in which parents exhibit antisocial behavior? The evidence for the association between the aforementioned risk factors and parental antisocial behavior comes from two types of studies—studies of adult outcomes of children with CPASB and studies that have assessed the relationship between antisocial disorder in parents and the risk factors. The adult outcomes of children with early-onset antisocial disorder include the very outcomes that are also considered risk factors for early-onset antisocial disorder. Girls and boys with antisocial behavior are more likely to become teen parents. In a recent adult outcome study of a birth cohort of children in Dunedin, New Zealand, the authors stated: “Research into the consequences of teenage parenthood should begin to address the implications of the strong link between antisocial behaviour and teenage parenthood. Most of the infants born to teenage parents have at least one antisocial parent, and many are born to two [antisocial] parents” (p. 197) (Moffitt et al., 2001). There are also strong associations between conduct disorder in girls and smoking (Boyle et al., 1993; Hawkins, Catalano, & Miller, 1992; Kandel et al., 1997). Multiple adversity in adult life appears to be particularly associated with childhood histories of conduct problems. Among a sample of adults who in
Intergenerational Transmission of Aggression
361
childhood were either in the care of child welfare authorities or from a poor neighborhood, multiple social adversity was found almost exclusively in those with a history of conduct disorder (Zoccolillo, Pickles, Quinton, & Rutter, 1992). The current diagnostic construct of antisocial personality disorder is based on the empirical observation that adults with pervasive dysfunction across several domains and who have antisocial behavior almost always have histories of child conduct problems (Robins, 1966). Turning to studies on parents, there is again good evidence for an association between antisocial behavior and the five risk factors examined here. Coercive parenting has been found to be associated with parental antisocial behavior (Verlaan & Schwartzman, 2002; Patterson, DeGarmo, & Knutson, 2000; Bosquet & Egeland, 2000; Johnson, Cohen, Kasen, Smailes, & Brook, 2001). Adults who report child neglect have higher rates of antisocial personality disorder (Robins & Regier, 1991, Table 11-1, p. 261), and infants and toddlers of women who were aggressive in childhood have higher rates of hospitalization for accidents (Serbin, Peters, & Schwartzman, 1996). Maternal sensitivity has been found in three separate studies to be associated with maternal conduct disorder (Cassidy, Zoccolillo, & Hughes, 1996), a history of maternal aggression (Serbin, Peters, McAffer, & Schwartzman, 1991), or maternal antisocial personality disorder (Hans, Bernstein, & Henson, 1999). Maternal smoking during pregnancy is associated with paternal antisocial personality disorder and parental criminality (Fergusson et al., 1998; Wakschlag et al., 1997).
Parental Antisocial Behavior and Risk Factors for CPASB/CPA in Québec Although many studies have found an association between parental antisocial behavior and/or disorder and risk factors for CPASB/CPA, there has been no population-based study of child development that has prospectively assessed both maternal and paternal antisocial behavior and risk factors. This has made it difficult to estimate the strength of the association between risk variables and parental antisocial behavior. To supplement these data, we present findings from the Longitudinal Study of Child Development in Québec (LSCDQ) (Jette, Desrosiers, Tremblay, & Thibault, 2000; Zoccolillo, 2000). The LSCDQ assessed maternal and paternal antisocial behavior in a sample of 2,223 infants who were 5 months old. These infants were selected from birth certificate records using a three-stage sampling design in order for them to be representative of all infants of the province of Québec, Canada. Maternal and paternal conduct problems, specified as behaviors occurring before the end of high school, were assessed by self-report. Maternal behaviors included stealing more than once (17.8%), starting more than one fight (3.3%), involvement with the police or youth protection because of misbehavior (4%), skipping school more than twice in one year (47.6%), and running away from home overnight (9.6%). Paternal behaviors included stealing more than once (27.1%), starting more than one fight (10%), involvement
362
DETERMINANTS OF AGGRESSION
with the police or youth protection because of misbehavior (8.7%), and suspension/expulsion from school (20%). The associations between the number of maternal conduct problems and a variety of potential risk factors for adverse child development are shown in Table 17.1. Note that maternal conduct problems predicted whether the biologic father was missing from the home. Table 17.2 shows the association between the number of paternal conduct problems for biological fathers living in the home and potential risk factors. Tables 17.1 and 17.2 illustrate very clearly, in this population-based sample, the association between a measure of parental antisocial behavior (number of conduct problems) and a variety of possible risk factors. The risk is most pronounced at the level of 3–5 conduct problems, which is similar to the DSM-IV cutoff for a diagnosis of conduct disorder. What proportion of infants with multiple risks come from homes with antisocial parents? An index of multiple risk was created by a sum score of
TABLE 17.1. Risk Factors by the Number of Maternal Conduct Symptoms Number of conduct symptoms 0 (n = 955)
1 (n = 759)
2 (n = 275)
3–5 (n = 129)
7.3
12.9
13.4*
24.7*
< .001
13.5
18.5
18.8
32.2
< .001
7.1
8.6
10.2*
19.3 *
< .05
24.5
24.0
28.4
41.7
< .01
13.0 *
< .001
p
Sociodemographic status Had first child when age 19 or younger No high school degree Biological father no longer in the home Insufficient household income (Statistics Canada definition) Substance use Used illegal drugs in the past 12 month
1.2**
Ever drunk (5 or more drinks on one occasion) in past 12 months
9.2
16.2
23.5
31.8
< .001
15.0
29.4
38.8
50.2
< .001
Smoked during pregnancy Used illegal drugs during pregnancy
0.2**
3.7
2.1**
4.2**
2.0**
5.9 **
Note. Data from Institut de la Statistique du Québec, ELDEQ 1998–2002. *Coefficient of variation between 15% and 25%; interpret with caution. **Coefficient of variation higher than 25%; imprecise estimate for descriptive purposes only.
< .001
Intergenerational Transmission of Aggression
363
TABLE 17.2. Risk Factors by the Number of Paternal Conduct Symptoms in Two-Parent Families Number of conduct symptoms 0 1 (n = 1,067) (n = 457)
2 (n = 208)
3–5 (n = 102)
p
No high school degree
13.2
17.6
24.6
25.6*
< .001
College degree
43.2
35.4
26.9
12.5**
< .001
Unemployed
11.6
11.4
18.0*
20.3*
< .05
Currently daily smoker
21.5
33.8
42.8
55.3
< .001
Used illegal drugs in the past 12 month
4.6*
7.6*
11.0*
20.6*
< .001
Mean number of times drunk (5 or more drinks on one occasion)
4.2
7.4
7.0
13.1
< .001
Maternal behaviors of partner Mother had 2 or more conduct symptoms
12.9 20.0 (n = 1,048) (n = 451)
30.8 (n = 204)
30.7* (n = 101)
< .001
Mother smoked during pregnancy
19.4 26.1 (n = 1,062) (n = 456)
31.3 (n = 208)
42.7 (n = 103)
< .001
Note. Data from Institut de la Statistique du Québec, ELDEQ 1998–2002. *Coefficient of variation between 15% and 25%; interpret with caution. **Coefficient of variation higher than 25%; imprecise estimate for descriptive purposes only.
four risk factors: (1) mother does not have high school diploma, (2) family in lowest income category, (3) mother smoked during pregnancy, (4) mother had first child before age 21. A dichotomous measure of parent antisocial disorder was created. Because one conduct problem was relatively common in the mothers and studies have suggested that it is the presence of multiple conduct problems that is particularly associated with pervasive and persistent antisocial disorder (Zoccolillo et al., 1992; Robins, 1966), infants were dichotomized into two groups where data on antisocial disorder was available for both parents—infants who had mothers and fathers who had no or one conduct symptom versus all other infants. Studies have suggested that biological fathers not living in the home are more antisocial than nonmissing fathers (Caspi et al., 2001; Jaffee, Caspi, Moffitt, Taylor, & Dickson, 2001b), and in our sample infants with the most antisocial mothers were also the most likely to have biological fathers not living in the home. Therefore, infants with biologic fathers not living in the home were also placed in the “all others” group. Finally, some mothers and fathers living in the home did not return the self-administered questionnaires that contained the questions on antisocial behavior, and a small subset of infants not participating in the longitudinal followup (103) were removed from the analyses, leaving a total of 1986 infants.
364
DETERMINANTS OF AGGRESSION
The association between the number of risks was cross-classified by the dichotomous parental antisocial disorder/father not in the home measure (Table 17.3). There was a strong and significant association between multiple risk and parental antisocial behavior/father not in the home. Most infants without risk do not have antisocial parents, and most infants with multiple risk have at least one parent with at least two conduct problems or have fathers not living in the home. The association between multiple risk and CPASB/CPA is confounded by parental antisocial behavior.
Controlling for Parental Antisocial Behavior When Examining Risk Factors for CPASB/CPA To summarize the findings: (1) Parental antisocial behavior and offspring early antisocial behavior are associated; (2) parenting practices, maternal sensitivity, child maltreatment, young motherhood, maternal smoking during pregnancy, and multiple adversity are all predictors of CPASB and or CPA; (3) these same risk factors are also associated with parental antisocial behavior; (4) genetically informative designs have consistently found evidence for both shared environment and genetic effects for antisocial behavior and aggression. Given these strong relationships between environmental risk and parental antisocial behavior, have previous studies adequately controlled for gene–environment correlations in examining environmental risk factors for child antisocial behavior? An important question is how can the confound of genetic association be
TABLE 17.3. Proportion of Infants with Nonantisocial Parents by Level of Adversity
Adversity indexa
Mother and father have no or one conduct problem and father is not missingb
None (n = 1,117)
75.1%
One (n = 453)
56.5%
Two (n = 179)
50.3%
Three (n = 107)
36.6%
Four (n = 40)
21.4%
Note. n = 1,896; χ2 = 121.34, 4 df, p < .000 (corrected for design effect of 1.3). aSum of 4 risks: Mother does not have high school degree; in lowest income category; mother smoked during pregnancy; mother had first child before age 21. bInfants classified into two parent groups: (1) Mother and father have no or one conduct problem and father is not missing; (2) all other combinations of maternal and paternal conduct problems and those with fathers not living in the home. Homes where mother or father was present but did not return valid questionnaire on antisocial behavior were excluded.
Intergenerational Transmission of Aggression
365
controlled for when testing a specific risk factor? A critical question is whether genetically informative designs (twin and adoption studies) are necessary to control for parental antisocial disorder when testing a specific risk factor. The adoption design is not feasible to examine the critical question of why children exposed to maltreatment and multiple adversities are more likely to grow up to be antisocial, because adopted children are generally not placed in adoptive homes where maltreatment and severe adversity will occur (Rutter, 1997). Another, related, problem with adoption studies is that the maternal prenatal environment is often not well assessed. Given the association between maternal and paternal antisocial disorder, maternal prenatal smoking, and offspring antisocial behavior, this could lead to an overestimate of genetic effects. Twin studies are limited because the families most of interest are those of twins with one or more antisocial parents and/or twins with severe adversity, maltreatment, or other risk factors. Furthermore, the ideal studies are prospective studies and studies based on observational data collected before the twins show aggressive or antisocial behavior. This requires recruiting from the pool of twin births. Because approximately 1 in 100 deliveries is a twin birth, recruiting 100 twin pairs requires screening 10,000 births (Vogel & Motulsky, 1996). However, if twins with an antisocial parent are required (assuming 10% of twins have an antisocial mother or father), then 100,000 births will have to be screened and 1,000 twin pairs recruited. Even this is limiting, as the end result will be approximately 30 monozygotic and 70 dizygotic twin pairs with an antisocial parent. A recently developed twin study of high-risk infants is the Environmental Risk (E-Risk) Longitudinal Twin Study (Jaffee, Moffitt, Caspi, & Taylor, 2003). This sample of twins was based on the Twins Early Development Study, a register of all twins born in England and Wales in 1994–1995. During this time 15,906 twin pairs were born, of whom 71% joined the register. From this pool, a subset of 1,116 twin pairs was drawn, of whom one-third were those of mothers whose first birth was less than 20 years of age and the rest were representative of all twins. Of note is that this twin sample was specifically designed to look at environmental risk and includes assessments of antisocial behavior in the parents. Few researchers, however, are able to access such large twin samples. Can the twin design be useful in identifying specific environmental effects? First, as noted earlier, it is important to study twins with antisocial parents so as to have the twins exposed to the type of risk seen in infants of antisocial parents. Apart from the large number of twins needed to find twins meeting this criterion, the twin design itself poses a particular problem for studying children of antisocial parents. Mothers of dizygotic twins are considerably older than mothers of monozygotic twins (Vogel & Motulsky, 1996). Furthermore, mothers who postpone childbearing to later in life, because of education and/or career, and who then require fertility treatment are likely have dizygotic twins as a result. Because antisocial mothers have children at a
366
DETERMINANTS OF AGGRESSION
younger age, this raises the possibility that identical twins will be more likely to have antisocial parents than dizygotic twins. Differences in environment between monozygotic and dizygotic twins must be properly examined and accounted for, as the key assumption of twin studies is that the environments of monozygotic and dizygotic twins are similar. Nonetheless, a twin high-risk design is a powerful tool for examining environmental effects. In the E-Risk study noted earlier, combining behavior genetic and epidemiologic analysis showed that “children who resided with antisocial fathers received a ‘double whammy’ of genetic and environmental risk for conduct problems” (Jaffee et al., 2003, p. 109). Because adoptive designs are not feasible, and twin designs suitable for addressing questions such as maltreatment or multiple risk are methodologically intensive and require large populations of twin births, is there another alternative? The answer is a qualified yes. The solution is to control for antisocial behavior in both biological parents when measuring the effects of the suspected risk factor. It is important to note that this method can be used to control for genetic transmission but cannot be used to address the question of proportion of variance of the outcome explained by genes or environment. It is also important to note that the “genotype” measured is very imperfectly measured through parental phenotype. If the wrong phenotype is measured, then it will not be an adequate control. An example of how this can occur is provided by the adoption and family studies examining the association between unexplained somatic complaints in women and antisocial behavior or disorders in their male relatives. The evidence is strong that multiple somatic complaints in women are genetically related to antisocial behavior in men, seen most dramatically in the excess of somatic complaints in the adopted daughters of violent alcoholic fathers (Bohman, Cloninger, von Knorring, & Sigvardsson, 1984). Controlling for overtly antisocial behaviors in the mother may still not adequately control for genotype. Before turning to the literature to see whether studies of suspected risk factors for offspring early-onset antisocial disorder have controlled for parental antisocial disorder, some issues of measurement need to be addressed: (1) classification of parents for use as a control variable, (2) assessing both parents, and (3) modeling both parents. These issues are discussed in turn in the following paragraphs. Classifying parents for use as a control variable is problematic because there is no clear boundary between “antisocial” and “not antisocial.” For example, even one child conduct-disorder symptom in childhood is associated with an increased risk for antisocial behavior in adult life (Robin, 1966). Because the reason for controlling for parental antisocial behavior is to remove any covariance between parental antisocial behavior and child antisocial behavior in order to test the independent effects of other variables, the measure of parental antisocial behavior should be broad enough to capture the full continuum of parental antisocial behavior.
Intergenerational Transmission of Aggression
367
Two common measures used as a control for parental antisocial behavior are a diagnosis of antisocial personality disorder and criminality (e.g., arrests, convictions). Both are problematic because they are dichotomous measures, and many parents without a diagnosis or without criminality are still likely to have antisocial behavior. With regard to DSM-IV antisocial personality disorder (APD), three child symptoms and three adult symptoms are required to meet the criteria (American Psychiatric Association, 1994). This means that those with three child symptoms and two adult symptoms, or two child symptoms and three adult symptoms, would be classified as “No disorder.” In previous work, it was shown that DSM-III (American Psychiatric Association, 1980) criteria for APD identified only about half of young adults with childhoodonset persistent and pervasive antisocial behavior (Zoccolillo et al., 1992). It is highly likely that many parents with some antisocial symptoms, but not enough to meet DSM criteria for APD, will have childhood-onset persistent and pervasive antisocial behavior. Classifying these parents as not antisocial when investigating risk factors for offspring CPASB/CPA does not adequately control for the possibility of confounding by parental antisocial behavior. Similarly, criminality is also a poor control because, in population samples of adults with DSM-III diagnoses of APD, almost half the men and most women with APD did not have major criminal convictions (Robins & Regier, 1991). Both biological parents need to be assessed for antisocial disorder. The heritability for antisocial behavior was not found to differ for men and women (Rhee & Waldman, 2002), and the effects of parental APD on offspring externalizing disorder were also the same (Connell & Goodman, 2002). Assessing both parents is particularly important because environmental risk factors may be associated with antisocial disorder in one or the other parent. For example, antisocial men may be more prone to mate with teenage females, independent of whether the female also had an antisocial disorder. Although there is good evidence for assortative mating, it is not complete (Krueger, Moffitt, Caspi, Bleske, & Silva, 1998). This means that there will be a large proportion of families in which only one parent has an antisocial disorder. Another major concern is missing biological fathers. Missing biological fathers are likely to be more antisocial than fathers who are present in the home (Jaffee et al., 2001b). Asking the biological mother about the missing biological father is essential and appears to give reliable information (Caspi et al., 2001). Alternatively, missing biological fathers can be incorporated in the theoretical model being tested as a separate category. Both parents also need to be included in the analyses with the child as the unit of analysis. For example, in the LSCDQ there was a strong association between maternal smoking during pregnancy and the number of paternal conduct symptoms (Table 17.2). Controlling only for maternal antisocial behavior when evaluating the effect of maternal prenatal smoking would not adquately control for the child’s genetic background. We have been unable to find any studies of the environmental risk factors
368
DETERMINANTS OF AGGRESSION
that meet all the aforementioned criteria, except for one recent study of infant aggression (Tremblay et al., 2004) and the Environmental Risk (E-Risk) Longitudinal Twin Study (Jaffee et al., 2003). Although some investigators have attempted to measure antisocial behavior in the parent, most used a diagnostic category of parental antisocial personality disorder versus no disorder (Johnson et al., 2001; Wakschlag et al., 1997), relied on parental criminality (sometimes specified only as familial criminality) (Moffitt et al., 2001; Patterson et al., 2000; Fergusson et al., 1998; Rutter et al., 1976), or used personality disorder scales from the Minnesota Multiphasic Personality Inventory (Patterson et al., 2000; Wakschlag et al., 1997). To summarize, none of the studies have demonstrated good control of parental antisocial disorder. It would be difficult to conclude that maternal smoking during pregnancy, teenage motherhood, maternal sensitivity, parenting practices, maltreatment, or multiple adversity is causal for the development of early-onset antisocial behavior, as these factors are confounded by parental antisocial disorder.
DIRECTIONS FOR FUTURE RESEARCH It is critical for studies that are attempting to identify the causes of antisocial behavior to take into account the familial transmission of antisocial disorder and that antisocial disorder in parents is strongly associated with a multiplicity of environmental risk factors for their offspring. At the population level, the association between parental antisocial disorder and adversity, particularly multiple adversity, is large. Researchers cannot ignore that fact that a significant proportion of children living in adversity he found to have an antisocial parent. There are several implications of these findings. First, it is critical to study families with antisocial parents. Although the difficulties of studying these families and the complexity of multiple risk factors may appear to be daunting, these families are not rare and account for a large proportion of antisocial youth, particularly the most severely affected children. Second, children with antisocial parents are exposed to multiple risk factors that start very early. Teasing out which risk factors are causal for antisocial disorder in the child is the challenge. Tests of hypotheses need to be well structured to control for confounds and to test for interactions. Of critical importance are studies of the developmental sequence of risk factors and behaviors. Priority should be given to studies that start prenatally or at birth. To test whether poor parenting practices are a causal risk factor for early antisocial disorder requires demonstrating that these parenting practices precede the emergence of early-onset antisocial disorder. This is not an easy task, inasmuch as coercive parenting practices may emerge only when the child develops normal opposition behaviors in the toddler years, making it difficult to determine the time order of effects. Controlling for parental antisocial disor-
Intergenerational Transmission of Aggression
369
der is critical in such design because the same genetic mechanism could potentially lead to an antisocial parent, poor parenting practices, and offspring antisocial disorder (Ge et al., 1996; O’Connor, Deater-Deckard, Fulker, Rutter, & Plomin, 1998). Third, as much attention needs to be paid to defining “antisocial behavior” in the parent as in the child. Future studies should avoid using criminality as the sole measure of parental antisocial disorder, as it is neither sensitive nor specific enough to be useful. The LSCDQ study has shown that parents in a population sample will answer questions on their own child and adult antisocial behaviors. An advantage of assessing a broad range of child and adult antisocial behaviors in parents is that the researcher can then use latent variable methods to cluster parents into homogenous groups or to create a latent continous variable. It also becomes possible to test whether there are distinct familial types of antisocial behavior—for example, early-onset versus adolescentonset. Ideally, the same dimensions of aggression and antisocial behavior assessed in the child should be assessed in the adult. Little is known about whether parental aggression is specifically associated with child aggression, for example. Fourth, both biological parents must be assessed. Controlling for one parent only is not adequate, as assortative mating is incomplete and antisocial disorder is more common in males than in females. Missing biological fathers need to be accounted for, either by proxy interview of other informants or by having a separate category of “missing father.” Analytic strategies for incorporating data by proxy interview on missing fathers with data from present fathers need to be refined. The analysis of data that incorporates measurements of antisocial behavior in both parents is more complex than treating maternal and paternal antisocial disorder as separate variables in a logistic or linear regression model. For example, fathers are more likely to be missing if the mother is antisocial and are therefore not missing at random which regard to many analyses of interest. This poses considerable problems for most statistical analyses in that maximum likelihood estimates for parameters often cannot be estimated (Little & Rubin, 1987). Analyses of interactions between parental antisocial disorder and other variables (as proxy measures for gene– environment interactions) are also complex. For example, a parent contributes only half of his or her genes to offspring, and path models of parent effects on offspring would need to take this into account (Vogel & Motulsky, 1996). Fifth, randomized controlled trials of interventions to prevent CPA/ CPASB should assess parental antisocial behavior. Little is known about whether promising prevention interventions are also effective in families with antisocial parents. Furthermore, if modifying a risk factor is effective in preventing CPA/CPASB even among children with parents with antisocial behavior, then this strongly suggests a causal role for that factor. In a recent review of studies of conduct disorder, the authors noted,
370
DETERMINANTS OF AGGRESSION
“Parsing out the contributions of genetic effects from environmental risk factors to familial antisocial behavior is a significant task that remains largely incomplete” (Burke, Loeber, & Birmaher, 2002, p. 1277). A major challenge of the next 10 years is to fill in the holes in our knowledge of genetic and environmental contributions to the development of early-onset antisocial behavior and aggression through new and better research designs.
REFERENCES American Psychiatric Association. (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Bohman, M., Cloninger, C. R., von Knorring, A. L., & Sigvardsson, S. (1984). An adoption study of somatoform disorders. III. Cross-fostering analysis and genetic relationship to alcoholism and criminality. Archives of General Psychiatry, 41, 872– 878. Bosquet, M., & Egeland, B. (2000). Predicting parenting behaviors from Antisocial Practices Content Scale scores of the MMPI-2 administered during pregnancy. Journal of Personality Assessment, 74, 146–162. Boyle, M. H., Offord, D. R., Racine, Y. A., Fleming, J. E., Szatmari, P., & Links, P. S. (1993). Predicting substance use in early adolescence based on parent and teacher assessments of childhood psychiatric disorder: Results from the Ontario Child Health Study follow-up. Journal of Child Psychology and Psychiatry and Allied Disciplines, 34, 535–544. Brennan, P. A., Grekin, E. R., & Mednick, S. A. (1999). Maternal smoking during pregnancy and adult male criminal outcomes [see comments]. Archives of General Psychiatry, 56, 215–219. Broidy, L. M., Nagin, D. S., Tremblay, R. E., Bates, J. E., Brame, B., Dodge, K. A., Fergusson, D., Horwoood, J. L., Loeber, R., Laird, R., Lynam, D. R., Moffitt, T. E., Petit, G. S., & Vitaro, F. (2003). Developmental trajectories of childhood disruptive behaviors and adolescent delinquency: A six-site, cross-national study. Developmental Psychology, 39, 222–245. Burke, J. D., Loeber, R., & Birmaher, B. (2002). Oppositional defiant disorder and conduct disorder: A review of the past 10 years. Part II. Journal of the American Academy of Child and Adolescent Psychiatry, 41, 1294–1305. Campbell, S. B., Breaux, A. M., Ewing, L. J., & Szumowski, E. K. (1986). Correlates and predictors of hyperactivity and aggression: A longitudinal study of parent-referred problem preschoolers. Journal of Abnormal Child Psychology, 14, 217–234. Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., Taylor, A., & Poulton, R. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297, 851–854. Caspi, A., Moffitt, T. E., Newman, D. L., & Silva, P. A. (1996). Behavioral observations at age 3 years predict adult psychiatric disorders: Longitudinal evidence from a birth cohort. Archives of General Psychiatry, 53, 1033–1039. Caspi, A., Taylor, A., Smart, M., Jackson, J., Tagami, S., & Moffitt, T. E. (2001). Can women provide reliable information about their children’s fathers? Cross-infor-
Intergenerational Transmission of Aggression
371
mant agreement about men’s lifetime antisocial behaviour. Journal of Child Psychology and Psychiatry and Allied Disciplines, 42, 915–920. Cassidy, B., Zoccolillo, M., & Hughes, S. (1996). Psychopathology in adolescent mothers and its effects on mother–infant interactions: A pilot study. Canadian Journal of Psychiatry [Revue Canadienne de Psychiatrie], 41, 379–384. Christ, M. A., Lahey, B. B., Frick, P. J., Russo, M. F., McBurnett, K., Loeber, R., Stouthamer-Loeber, M., & Green, S. (1990). Serious conduct problems in the children of adolescent mothers: Disentangling confounded correlations. Journal of Consulting and Clinical Psychology, 58, 840–844. Connell, A. M., & Goodman, S. H. (2002). The association between psychopathology in fathers versus mothers and children’s internalizing and externalizing behavior problems: A meta-analysis. Psychological Bulletin, 128, 746–773. Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990). Mechanisms in the cycle of violence. Science, 250, 1678–1683. Dodge, K. A., Pettit, G. S., & Bates, J. E. (1994). Socialization mediators of the relation between socioeconomic status and child conduct problems. Child Development, 65, 649–665. Dodge, K. A., Pettit, G. S., Bates, J. E., & Valente, E. (1995). Social information-processing patterns partially mediate the effect of early physical abuse on later conduct problems. Journal of Abnormal Psychology, 104, 632–643. Downey, G., & Walker, E. (1992). Distinguishing family-level and child-level influences on the development of depression and aggression in children at risk. Development and Psychopathology, 4, 81–95. Eddy, J. M., Leve, L. D., & Fagot, B. I. (2001). Coercive family processes: A replication and extension of Patterson’s Coercion Model. Aggressive Behavior, 27, 14–25. Eley, T. C., Lichtenstein, P., & Moffitt, T. E. (2003). A longitudinal behavioral genetic analysis of the etiology of aggressive and nonaggressive antisocial behavior. Development and Psychopathology, 15, 383–402. Erickson, M. F., Egeland, B., & Pianta, R. (1989). The effects of maltreatment on the development of young children. In D. Cicchetti & V. Carlson (Eds.), Child maltreatment: Theory and research on the causes and consequences of child abuse and neglect (pp. 647–684). New York: Cambridge University Press. Fergusson, D. M., Horwood, L. J., & Lynskey, M. (1994). The childhoods of multiple problem adolescents: A 15-year longitudinal study. Journal of Child Psychology and Psychiatry and Allied Disciplines, 35, 1123–1140. Fergusson, D. M., & Lynskey, M. T. (1993). Maternal age and cognitive and behavioural outcomes in middle childhood. Paediatric and Perinatal Epidemiology, 7, 77–91. Fergusson, D. M., Woodward, L. J., & Horwood, L. J. (1998). Maternal smoking during pregnancy and psychiatric adjustment in late adolescence. Archives of General Psychiatry, 55, 721–727. Ge, X., Conger, R., Cadoret, R., Neiderhiser, J., Yates, W., Troughton, E., & Stewart, M. A. (1996). The developmental interface between nature and nurture: A mutual influence model of child antisocial behavior and parent behaviors. Developmental Psychology, 32, 574–589. Hans, S. L., Bernstein V. J., & Henson L. G. (1999). The role of psychopathology in the parenting of drug-dependent women. Development and Psychopathology, 11, 957– 977. Hardy, J. B., Shapiro, S., Astone, N. M., Miller, T. L., Brooks-Gunn, J., & Hilton, S. C. (1997). Adolescent childbearing revisited: The age of inner-city mothers at delivery
372
DETERMINANTS OF AGGRESSION
is a determinant of their children’s self-sufficiency at age 27 to 33. Pediatrics, 100, 802–809. Hawkins, J. D., Catalano, R. F., & Miller, J. Y. (1992). Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance abuse prevention. Psychological Bulletin, 112, 64–105. Herrenkohl, R. C., Egolf, B. P., & Herrenkohl, E. C. (1997). Preschool antecedents of adolescent assaultive behaviour: A longitudinal study. American Journal of Orthopsychiatry, 67, 422–432. Hinshaw, S. P. (2002). Process, mechanism, and explanation related to externalizing behavior in developmental psychopathology. Journal of Abnormal Child Psychology, 30, 431–446. Huesmann, L. R., Eron, L. D., Lefkowitz, M. M., & Walder, L. O. (1984). Stability of aggression over time and generations. Developmental Psychology, 20, 1120–1134. Jaffee, S., Caspi, A., Moffitt, T. E., Belsky, J., & Silva, P. (2001a). Why are children born to teen mothers at risk for adverse outcomes in young adulthood? Results from a 20–year longitudinal study. Development and Psychopathology, 13, 377–397. Jaffee, S. R., Caspi, A., Moffitt, T. E., Taylor, A., & Dickson, N. (2001b). Predicting early fatherhood and whether young fathers live with their children: Prospective findings and policy reconsiderations. Journal of Child Psychology and Psychiatry and Allied Disciplines, 42, 803–815. Jaffee, S. R., Moffitt, T. E.,Caspi, A., & Taylor, A. (2003). Life with or without father: The benefits of living with two biological parents depend on the father’s antisocial behavior. Child Development, 74, 109–126. Jette, M., Desrosiers, H., Tremblay, R. E., & Thibault, J. (2000). Longitudinal study of child development in Quebec (ELDEQ 1998–2002) (Vol. 1). Québec: Institut de la Statistique du Québec. Johnson, J. G., Cohen, P., Kasen, S., Smailes, E., & Brook, J. S. (2001). Association of maladaptive parental behavior with psychiatric disorder among parents and their offspring. Archives of General Psychiatry, 58, 453–460. Kandel, D., Johnson, J., Bird, H., Canino, G., Goodman, S. H., Lahey, B. B., et al. (1997). Psychiatric disorders associated with substance use among children and adolescents: Findings from the Methods for the Epidemiology of Child and Adolescent Mental Disorders (MECA) Study. Journal of Abnormal Child Psychology, 25, 121– 132. Krueger, R. F., Moffitt, T. E., Caspi, A., Bleske, A., & Silva, P. A. (1998). Assortative mating for antisocial behavior: Developmental and methodological implications. Behavior Genetics, 28, 173–186. Lahey, B. B., Loeber, R., Quay, H. C., Applegate, B., Shaffer, D., Waldman, I., Hart, E. L., McBurnett, K., Frick, P. J., Jensen, P. S., Dulcan, M. K., Canino, G., & Bird, H. R. (1998). Validity of DSM-IV subtypes of conduct disorder based on age of onset. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 435–442. Little R. A., & Rubin, D. R. (1987). Statistical analysis with missing data. New York: Wiley. Loeber, R., Burke, J. D., Lahey, B. B., Winters, A., & Zera, M. (2000). Oppositional defiant and conduct disorder: A review of the past 10 years. Part I. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 1468–1484. Loeber, R., & Farrington, D. P. (2000). Young children who commit crime: Epidemiology, developmental origins, risk factors, early interventions, and policy implications. Development and Psychopathology, 12, 737–762.
Intergenerational Transmission of Aggression
373
Lyons-Ruth, K., Alpern, L., & Repacholi, B. (1993). Disorganized infant attachment classification and maternal psychosocial problems as predictors of hostile-aggressive behavior in the preschool classroom. Child Development, 64, 572–585. Lyons-Ruth, K., Easterbrooks, A., & Cibelli, C. (1997). Infant attachment strategies, infant mental lag, and maternal depressive symptoms : Predictors of internalizing and externalizing problems at age 7. Developmental Psychology, 33, 681–692. Malinosky-Rummel, R., & Hansen, D. J. (1993). Long-term consequences of childhood physical abuse. Psychological Bulletin, 114, 68–79. Martin, S. L., & Burchinal, M. R. (1992). Young women’s antisocial behavior and the later emotional and behavioral health of their children. American Journal of Public Health, 82, 1007–1010. McFadyen-Ketchum, S. A., Bates, J. E., Dodge, K. A., & Pettit, G. S. (1996). Patterns of change in early childhood aggressive-disruptive behavior: Gender differences in predictions from early coercive and affectionate mother–child interactions. Child Development, 67, 2417–2433. Moffitt, T. E., Caspi, A., Rutter, M., & Silva, P. (2001). Sex differences in antisocial behaviour. Cambridge, UK: Cambridge University Press. Muller, R. T., & Diamond, T. (1999). Father and mother physical abuse and child aggressive behaviour in two generations. Canadian Journal of Behavioural Science, 31, 221–228. Nagin, D. S., Pogarsky, G., & Farrington, D. P. (1999). Adolescent mothers and the criminal behavior of their children. Law and Society Review, 31, 137–162. Nagin, D. S., & Tremblay, R. E. (2001). Prenatal and early childhood predictors of persistent physical aggression in boys from kindergarten to high school. Archives of General Psychiatry, 58, 389–394. O’Connor, T. G., Deater-Deckard, K., Fulker, D., Rutter, M., & Plomin, R. (1998). Genotype–environment correlations in late childhood and early adolescence: Antisocial behavioral problems and coercive parenting. Developmental Psychology, 34, 970– 981. Olds, D. (1997). Tobacco exposure and impaired development: A review of the evidence. Mental Retardation and Developmental Disabilities Review, 3, 257–269. Patterson, D. R., DeGarmo, D. S., & Knutson, N. (2000). Hyperactive and antisocial behaviors: Comorbid or two points in the same process? Development and Psychopathology, 12, 91–106. Patterson, G. R. (1982). Coercive family processes. Eugene, OR: Castalia. Plomin, R. (1997). Behavioral genetics (3rd ed.). New York: Freeman. Rasanen, P., Hakko, H., Isohanni, M., Hodgins, S., Jarvelin, M. R., & Tiihonen, J. (1999). Maternal smoking during pregnancy and risk of criminal behavior among adult male offspring in the Northern Finland 1966 Birth Cohort. American Journal of Psychiatry, 156, 857–862. Reid, J., & Patterson, G. (1989). The development of antisocial behaviour patterns in childhood and adolescence. European Journal of Personality, 3, 107–119. Renken, B., Egeland, B., Marvinney, D., Mangelsdorf, S., & Sroufe, A. (1989). Early childhood antecedents of aggression and passive-withdrawal in early elementary school. Journal of Personality, 57, 257–281. Rhee, S. H., & Waldman, I. (2002). Genetic and environmental influences on antisocial behavior: A meta-analysis of twin and adoption studies. Psychological Bulletin, 128, 490–529. Robins, L. N. (1966). Deviant children grown up. Baltimore: Williams & Wilkins.
374
DETERMINANTS OF AGGRESSION
Robins, L. N. (1996). Deviant children grown up. European Child and Adolescent Psychiatry, 5(Suppl. 6). Robins, L. N., & Regier, D. A. (1991). Psychiatric disorders in America: The epidemiologic catchment area study. New York: Free Press. Rubin, K. H., Hastings, P., Chen, X., Stewart, S., & McNichol, K. (1998). Intrapersonal and maternal correlates of aggression, conflict, and externalizing problems in toddlers. Child Development, 69, 1614–1629. Rutter, M. (1997). Nature–nurture integration. The example of antisocial behavior. American Psychologist, 52, 390–398. Rutter, M., Tizard, J., Yule, W., Graham, P., & Whitmore, K. (1976). Research report: Isle of Wight studies. Psychological Medicine, 126, 493–509. Serbin, L., Peters, P., McAffer, V., & Schwartzman, A. (1991). Childhood aggression and withdrawal as predictors of adolescent pregnancy, early parenthood, and environmental risk for the next generation. Canadian Journal of Behavioural Science, 23, 318–331. Serbin, L. A., Peters, P. L., & Schwartzman, A. E. (1996). Longitudinal study of early childhood injuries and acute illnesses in the offspring of adolescent mothers who were aggressive, withdrawn, or aggressive-withdrawn in childhood. Journal of Abnormal Psychology, 105, 500–507. Shaw, D. S., Owens, E. B., Vondra, J. I., & Keenan, K. (1996). Early risk factors and pathways in the development of early disruptive behavior problems. Development and Psychopathology, 8, 679–699. Shaw, D. S., Vondra, J. I., Hommerding, K. D., Keenan, K., & Dunn, M. (1994). Chronic family adversity and early child behavior problems: A longitudinal study of low income families. Journal of Child Psychology and Psychiatry and Allied Disciplines, 35, 1109–1122. Tremblay, R. E. (2000). The development of aggressive behaviour during childhood: What have we learned in the past century? International Journal of Behavioral Development, 24, 129–141. Tremblay, R. E., Nagin, D. S., Seguin, J. R., Zoccolillo, M., Zelazo, P. D., Boivin, M., Pérusse, D., & Japel, C. (2004). Predictors of high level physical aggression from 17 to 42 months after birth. Pediatrics, 114, e43–e50. Verlaan, P., & Schwartzman, A. E. (2002). Mother’s and father’s parental adjustment: Links to externalizing behaviour problems in sons and daughters. International Journal of Behavioral Development, 26, 214–224. Vogel, F., & Motulsky, A. (1996). Human genetics: Problems and approaches (3rd ed.). Berlin: Springer. Wakschlag, L. S., Gordon, R., Lahey, B. B., Loeber, R., Green, S., & Leventhal, B. L. (2000). Maternal age at first birth and boys’ risk for conduct disorder. Journal of Research on Adolescence, 10, 417–441. Wakschlag, L. S., & Hans, S. L. (1999). Relation of maternal responsiveness during infancy to the development of behavior problems in high-risk youths. Developmental Psychology, 35, 569–579. Wakschlag, L. S., & Hans, S. L. (2002). Maternal smoking during pregnancy and conduct problems in high-risk youth: A developmental framework. Development and Psychopathology, 14, 351–369. Wakschlag, L. S., Lahey, B. B., Loeber, R., Green, S. M., Gordon, R. A., & Leventhal, B. L. (1997). Maternal smoking during pregnancy and the risk of conduct disorder in boys. Archives of General Psychiatry, 54, 670–676.
Intergenerational Transmission of Aggression
375
Wakschlag, L. S., Pickett, K. E., Cook, E., Jr., Benowitz, N. L., & Leventhal, B. L. (2002). Maternal smoking during pregnancy and severe antisocial behavior in offspring: A review. American Journal of Public Health, 92, 966–974. Weissman, M., Warner V., Wickramaratne, P., & Kandel, D. B. (1999). Maternal smoking during pregnancy and psychopathology in offspring followed to adulthood. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 892– 899. Widom, C. S. (2001). Child abuse and neglect. In S. O. White (Ed.), Handbook of youth and justice (pp. 31–47). New York: Kluwer Academic/Plenum Press. Zoccolillo, M. (2000). Parents’ health and social adjustment. Part II. Social adjustment. In M. Jette, H. Desrosiers, R. E. Tremblay, & J. Thibault (Eds.), Longitudinal study of child development in Québec (ELDEQ 1998–2002). Québec: Institut de la Statistique du Québec. Zoccolillo, M., Pickles, A., Quinton, D., & Rutter, M. (1992). The outcome of childhood conduct disorder: Implications for defining adult personality disorder and conduct disorder. Psychological Medicine, 22, 971–986. Zoccolillo, M., Price, R. K., Hwu, H. G., & Ji, T. (1998). Antisocial personality disorder: Comparisons of prevalence, symptoms, and correlates in four countries. In C. Slomkowski, P. Cohen, & L. N. Robins (Eds.), Where and when: Historical and geographic effects on psychopathology. Mahwah, NJ: Erlbaum.
DETERMINANTS Peer Relationships and OF Aggressive AGGRESSION Behavior
18 Peer Relationships and the Development of Aggressive Behavior in Early Childhood M ICHEL B OIVIN , F RANK V ITARO, and F RANÇOIS P OULIN
Children’s peer relationships have been the object of persistent attention by the developmental research community in the last decades. The prevalent view among members of this community is that children’s peer relationships serve important developmental functions (Asher & Coie, 1990; Hartup, 1983, 1996; Rubin, Bukowski, & Parker, 1998). They provide contexts where children may acquire new social skills, learn to self-regulate and solve interpersonal conflict, expand and validate their self-knowledge, and discover the social roles, norms, and processes involved in interpersonal relationships (Rubin et al., 1998). Unfortunately, peer relationships are not always beneficial to the child, as we discuss in this chapter; there are many ways in which specific features of peer relations may impede child development. The present review focuses on peer relationship problems as they relate to aggressive behaviors, inasmuch as aggression has been identified as the prime behavioral correlate of a variety of peer relationship problems in childhood (Rubin et al., 1998; Coie & Dodge, 1998). However, a comprehensive understanding of the association between aggressive behaviors and peer relation problems requires a differentiated view of childhood peer relationships.
THE MANY FACES OF PEER RELATIONSHIPS In examining children’s peer relationships, it is important to recognize that childhood peer relations are multifaceted. Children experience peer relations 376
Peer Relationships and Aggressive Behavior
377
through the group aggregate—for example, positive peer status, peer rejection, and peer victimization—through peer networks (Cairns, Leung, Buchanan, & Cairns, 1995), and through dyadic relationships, generally within friendship relations (Hartup, 1996). The distinction between these broadband classes of peer experiences is especially relevant when examining the association between aggressive behaviors and peer relationships. At least two main streams of research have extensively documented the link between aggressive behaviors and peer relationships in childhood. The first line of research is on peer rejection. Indeed, aggressive behavior is the most commonly cited behavioral correlate of peer rejection, an index of negative peer status derived from aggregated sociometric nominations (Rubin et al., 1998). Peer rejection is assumed to reflect difficulties in social integration within the peer group. The second stream of research focuses on the nature of the peer affiliations of aggressive children. According to this line of research, aggressive children, mostly boys, are not socially left out by their peers, but rather tend to associate with each other (Boivin & Vitaro, 1995; Cairns, Cairns, Neckerman, Gest, & Gariépy, 1988; Dishion, Andrews, & Crosby, 1995a). Thus, aggressive children may suffer from exclusion and ostracism by normative peers, as well as from problematic associations with deviant peers. More fundamentally, this distinction in peer experiences also pertains to putative processes of peer influence (i.e., peer exclusion processes), which are mostly derived from aggregate measures, versus deviant peer socialization processes, which mainly stem from dyadic relationship assessments. In other words, processes of peer influence are heterogeneous; that is, peers may marginalize, as well as provide support to, aggressive children. This review is organized to reflect this distinction in peer experience and process. Over the past 50 years there has been substantial research aimed at documenting, and then at understanding, the nature and meaning of the link between aggressive behaviors and peer relation problems (Hartup, 1983; Rubin et al., 1998). However, most of the more recent research effort has been centered on school-age children. To our knowledge, no one has specifically reviewed the evidence concerning the association between aggressive behaviors and peer relation problems during the preschool period. This lack exists despite evidence that peer relationships are already well established and quite differentiated in the preschool years (see the following section), and notwithstanding their potential influence on the early developmental course of aggressive behaviors. Furthermore, this question is especially timely and relevant to social policy issues because a growing number of children are exposed to peers early in their lives through daycare (NICHD Early Child Care Research Network, 2003). The purpose of this chapter is to provide an overview of the research bearing on the issue of peer relationships and aggressive behaviors in early childhood, that is, during the preschool years and in kindergarten. We start by
378
DETERMINANTS OF AGGRESSION
reviewing some of the basic milestones in the early development of peer relationships and aggressive behaviors.
THE FOUNDATIONS OF PEER INTERACTIONS AND PEER RELATIONSHIPS IN EARLY CHILDHOOD Early toddlerhood witnesses the development of many social skills underlying peer interactions. Social interest in peers is clearly present in the first year of life (Buhler, 1931; Vandell, Wilson, & Buchanan, 1980), but it is really during the second year that important gains in interactive skills can be observed. By the end of that year, toddlers have the ability to coordinate behavior in games with play partners. They can reciprocally imitate each other and progressively learn to alternate roles in play, thus indicating that they have acquired expectations about the behaviors of the other (Ross, Lollis, & Elliot, 1982; Strayer, 1989). With the onset of language, peer interactions become more refined and elaborate. Between age 3 and age 5, there is a dramatic increase in pretend play and in positive dyadic social behaviors, reflecting the augmented capacity of the child to adopt the perspective of the play partner (Rubin, Watson, & Jambor, 1978). These emerging social interactive skills are the foundation of early peer relationships. Signs of these relationships are seen in toddlers’ increasing behavioral preference for specific peers (Howes, 1987; Strayer, 1989), and those preferences culminate in stable preschool friendships. These early friendships are often, but not exclusively, embedded in extensive affiliative networks (Strayer & Santos, 1996) and progressively become sex-segregated (Lafrenière, Strayer, & Gauthier, 1984; Maccoby, 1998; Pellegrini, in press). By age 3–4 years, children initiate more play behavior and display more mutual activity with friends than with nonfriends. Further, behavioral homophily, that is, the tendency to befriend or affiliate with peers whose behavior is similar to their own, is already present by the end of preschool. Nevertheless, preschool friendships are not without disagreements, quarrels, and conflicts; there is indeed more conflict among preschool friends than among preschool nonfriends. However, these conflicts are resolved more equally and more often by the use of negotiations and disengagement among friends than among nonfriends (Hartup & Laursen, 1991). Preschoolers also progressively internalize their perceptions about their friends, as well as about other peers with whom they are in contact. At least by age 4, and when asked appropriately, preschoolers will readily and reliably identify best friends, as well as peers they like and peers they dislike. The aggregation of these perceptions reveals a coherent peer status structure within the larger group (Boivin & Bégin, 1986; Boivin, Tessier, & Strayer, 1985; Howes, 1987). Physically aggressive behaviors are also manifested very early in life (Caplan, Vespo, Pederson, & Hay, 1991; Goodenough, 1931; Tremblay,
Peer Relationships and Aggressive Behavior
379
2000). They become prevalent by the age of 2, peaking around age 3, before declining steadily with age (Strayer, 1989; Tremblay, 2000; see also Tremblay & Nagin, Chapter 5, and Archer & Coté, Chapter 20, this volume). The emergence of sex differences in aggressive behaviors seems to coincide with the time children start interacting more regularly in preschool groups (see Archer & Coté, Chapter 20, this volume). This transition is also characterized by a shift toward affiliation in the dyadic structure of social participation. Social dominance hierarchies based on agonistic behaviors are established very early, even among groups of infants. However, the ratio of initiated agonistic activity to initiated affiliation is much lower among older preschoolers (Strayer, 1989). Many factors may explain the decline in physical aggression, including language development (Dionne, Chapter 16, this volume), an augmented capacity for perspective taking and empathy (Selman, 1980; Zahn-Waxler, Radke-Yarrrow, & King, 1979), and an increasing awareness of dominance hierarchies (Strayer & Trudel, 1984). There is also the possible emergence of peer group norms and implicit rules with respect to aggressive behaviors. These norms and rules are likely to curb aggressive behaviors by providing a convergent feedback from group members to aggressors. All of this suggests that maturational changes and peer experiences compel a majority of young children to progressively develop a capacity to use other means than overt physical aggression to meet their goals. However, not all of them do. Physical aggression is one specific form of aggressive behavior, and we also need to attend to the other types and functions of aggressive behaviors to fully understand the nature of the developing association between these behaviors and peer relationships. Of special interest here is the distinction between direct and indirect or relational aggressive behaviors, as well as between reactive and proactive aggression, as they reflect changes in the maturational and social processes underlying aggression. For instance, Bjorkqvist, Lagerspetz, and Kaukiainen (1992) have argued that the prevalent form of aggressive behaviors is likely to change over the years, with physically aggressive behaviors giving way to more verbal forms of aggression, and then to more subtle and indirect forms (see Archer & Coté, Chapter 20, and Vaillancourt, Chapter 8, this volume). According to Bjorkqvist and colleagues (1992), these changes would be brought about by the parallel maturational changes in language and social cognitive skills. It is important to determine whether there is heretotypic stability in this normative change (i.e., the same children displaying these different forms of aggressive behaviors over the years) and to assess to what extent peer relationships play a role in the various developmental trajectories of aggression. Clearly, the social lives of preschoolers are elaborate and refined, and the development of peer relationships and aggressive behaviors are intertwined in the preschool years. It is thus important to document and understand the nature of the early association between peer relations and aggressive behaviors. Thus, in the following section of this chapter, we focus on experiences of seg-
380
DETERMINANTS OF AGGRESSION
regation by peers in early childhood, specifically peer rejection or ostracism, as they relate to aggressive behaviors. Then we turn our attention to the role of affiliating with aggressive peers in the development of aggressive behaviors during the same age period. In doing so, we adopt an ecological perspective; we consider different contexts of peer influence (e.g., friends and cliques, larger peer groups, daycare centers and schools) and examine the potential influence of specific nonpeer agents (parents, teachers) on children’s relationships with peers.
AGGRESSIVE BEHAVIORS AND PEER REJECTION AND OSTRACISM IN THE PRESCHOOL YEARS Between 5 and 10% of children experience chronic rejection by peers (Coie & Dodge, 1983). These children are at risk for a variety of future adjustment problems, including both internalizing and externalizing difficulties, dropping out of school, and delinquency (Rubin et al., 1998; McDougall, Hymel, Vaillancourt, & Mercer, 2001). Peer rejection could be a cause of these adjustment problems, but this predictive association could also result from an underlying factor, such as the child’s tendency to aggress or another related factor (see Parker & Asher, 1987). There has been much valuable research aimed at confirming the truly causal contribution of peer rejection to later maladjustment (see McDougall et al., 2001). However, evidence substantiating the causal model has been mixed, especially in studies examining the longitudinal associations between peer relationships and aggressive behaviors. This is why the nature of the association between peer rejection and aggressive behaviors has been, and is still, the focus of much interest. There has been abundant research linking aggressive behaviors and negative peer status. Some of this research has been conducted with preschool children (e.g., Boivin, Dorval, & Bégin, 1990; Hartup, Glazer, & Charlesworth, 1967; Masters & Furman, 1981). This research indicates that preschool negative peer status is not only associated with a negative aggressive behavior style, but is also linked to more negative peer experiences (Boivin et al., 1990; Masters & Furman, 1981). These negative peer experiences could reflect the give-and-take of aggressive exchanges. They could also antecede, or follow from, the child’s aggression. However, because behavior and status were measured concurrently in these studies, the causal nature of the relation was in doubt. The question of the direction of the relation between aggressive behaviors and peer rejection was tested more straightforwardly in a series of experimental play-group studies (Coie & Kupersmidt, 1983; Dodge, 1983). In these studies, unacquainted boys were typically assembled and observed in small groups, where their emerging peer status was also assessed. Verbal and physical aggressive behaviors were shown to antecede the emergence of peer rejection, which suggested that these behaviors could be considered as a proximal
Peer Relationships and Aggressive Behavior
381
determinant of these difficulties (Coie, 1990). However, because these studies were conducted with school-age children, it is not clear whether this could be the case in preschool. Dodge, Coie, Pettit, and Price (1990) came close to providing an answer to this question in a study that looked at the play-group behaviors of firstgrade children and those of third-grade children in order to investigate age differences in the behavioral predictors of peer status. They also looked at different forms of aggressive behaviors, including reactive aggression, instrumental aggression, bullying, and rough play. Both angry reactive aggression and instrumental aggression, but not rough play, were associated with negative peer status at the end of the sessions at both ages. Bullying was negatively related to status, but only in the third-grade groups. In the first-grade groups, the more popular boys displayed more bullying against aggressive boys, as if to establish their dominance, whereas for older boys, this goal could be attained by more sophisticated strategies. However, it was not clear whether this could be a finding specific to the social ecology of first-graders or whether it reflected a more general developmental trend. Further research in the preschool years is needed to clarify this issue. The studies described so far concur in indicating that aggressive behaviors could be a proximal cause of peer rejection, at least in middle childhood. However, it is actually difficult to form a definitive conclusion for the preschool period. Experimental play-group studies aimed at assessing different types of aggressive behaviors of unacquainted young children may thus be a good strategy to determine whether specific aggressive behaviors antecede peer rejection in the preschool years. The fact that aggressive behaviors may be a proximal cause of peer experiences does not mean that the experience of peer rejection could not also be a cause of aggressive behaviors. In other words, the relation between aggressive behaviors and peer rejection could very well be bidirectional. In fact, there is at least some indirect evidence suggesting that rejection by peers yields aggressive interactions. For instance, when the same behavior is manifested by a popular child and a rejected child, the rejected child received a less positive answer (Dodge, 1983). In familiar groups, rejected boys are more often the target of aggression than popular boys, whereas no difference is found in nonfamiliar groups (Coie & Kupersmidt, 1983). These observations are in line with the two studies of preschoolers presented earlier (Masters & Furman, 1981; Boivin et al., 1990). They suggest that status acquisition in peer group differs from status maintenance, in which reputation biases and scapegoat processes operate (Coie, 1990; Hymel, Wagner, & Butler, 1990). Thus, once established, peer rejection may be associated with negative peer experiences, and aggressive behaviors could also be adopted as a response to these experiences. Could peer rejection and negative peer experiences during the preschool years contribute to subsequent adjustment problems, including aggressive behaviors? The evidence with preschool children is scarce, but a series of studies
382
DETERMINANTS OF AGGRESSION
showed that cumulative peer rejection, starting in kindergarten, predicted growth in reactive aggression (rather than proactive aggression; Dodge et al., 2003). Interestingly, in one of these studies, peer rejection in kindergarten contributed uniquely to the prediction of aggressive behaviors in third grade, beyond peer rejection in first grade and peer rejection in second grade. Furthermore, early aggression also moderated the effect of rejection: Only children who were above the median in aggression were likely to show increases in aggression linked to their cumulative experience of peer rejection over the years. These results strongly suggest that the experience of rejection in kindergarten may exacerbate aggressive behavior development only among children initially disposed toward aggression. The fact that it affected reactive aggression, rather than proactive aggressive behaviors, also supports the view that peer rejection is a condition predisposing the child to aggressive behaviors. In sum, the empirical evidence favors a bidirectional model of the association between peer rejection and aggressive behaviors. Aggressive behaviors may lead to rejection by peers, and the experience of being rejected by peers in childhood may play an incremental role in the later development of aggressive behaviors. Not only can this evidence be traced back to the preschool years, but the experience of rejection in kindergarten seems critical in that prediction. It would be highly relevant at this point to substantiate further the potential contribution of peer rejection experiences in preschool to later forms of aggression. In particular, there are specific lines of inquiry that would require attention. The first line of inquiry is process oriented. We need to move beyond the simple description of the associations between aggressive behaviors and peer relations, however specific they may be, to better understand the “dynamics” of these specific behavior–relationship associations during early childhood. Specifically, intraindividual sociocognitive and emotional processes (e.g., social information processing deficits; Crick and Dodge, 1994) may underline, moderate, or mediate the links between peer relations and aggressive behaviors. The early experiences of rejection could influence the child’s sociocognitive processes, leading to social-cognitive deficits and biases, and result in increased aggressive behaviors. For example, retrospective reports of harsh parenting has been associated with the child’s aggressive behaviors in kindergarten, and specific social-cognitive deficiencies, including hostile biases in attributions, accounted for this relation (Dodge, Bates, & Pettit, 1990). These results suggest that social-cognitive processes may mediate the relation between specific aversive experiences and the early development of aggressive behaviors. Mediation through social-cognitive process was also found by Dodge et al. (2003): Negative peer status significantly predicted later aggression, and a significant part of that prediction was accounted for by socialcognitive processes. What is needed is a closer examination of these processes as a function of different forms and functions of aggressive behaviors in preschool. The second line of inquiry is also process oriented, but aims at describing
Peer Relationships and Aggressive Behavior
383
the nature of the peer experiences that are conducive to higher aggression. Research on peer victimization may be highly relevant to this objective in that it specifically targets children who truly experience negative treatments by peers. Indeed, peer rejection and peer victimization are related phenomena, but they should not be confounded. Peer rejection is a measure of disliking and reflects a negative attitude of the peer group toward a child, whereas peer victimization is a class of negative behaviors displayed toward a child (Boivin & Hymel, 1997; Boivin, Hymel, & Hodges, 2001). Peer victimization has mostly been documented in middle childhood, but there is evidence that such difficulties may exist in the preschool years (Alsaker & Valkanover, 2001; Kochenderfer & Ladd, 1996), even as early as age 3 for specific forms of indirect or relational victimization (Crick, Casas, & Ku, 1999). However, the main point here is that we need to document the nature of the peer interactions that are associated with actual peer rejection in preschool. Asher, Rose, and Gabriel (2001) have recently identified more than 30 types of rejecting behaviors, which they classified in six different categories (from controlling and dominating a child, to preventing a child from having access to a friend). Some of these ways of rejecting may be relevant for preschoolers. What is needed is a more complete and integrated view of how preschoolers reject another child. Third, social peer interaction processes may also underlie, mediate, or moderate the negative impact of aversive peer experiences, as well as the association between peer relations and aggressive behaviors. For example, group norms (Boivin, Dodge, & Coie, 1995; Stormshak, Bierman, Bruschi, Dodge, & Coie, 1999) and specific relationships (e.g., affiliation with aggressive children; Boivin & Vitaro, 1995; having a protective friend; Hodges, Boivin, Vitaro, & Bukowski, 1997) have been shown to moderate the links between peer experiences and aggressive behaviors among school-age children. We need to know whether and how these moderation effects operate in the preschool period. Two final considerations are warranted: First, it is important to note that some aggressive children may actually possess fairly high social status in the group (e.g., Prinstein & Cillessen, 2003; Rodkin, Farmer, Pearl, & Van Acker, 2000), especially if the groups norms are supportive or neutral in regard to aggressive behaviors (Boivin et al., 1995). This is more likely the case among young children, because specific forms of aggressive behaviors (i.e., the instrumental/ proactive type) are sometimes positively related to popularity. When this is the case, aggressive behaviors may be more likely to be valued and reinforced by, as well as generalized within, the peer group. Consequently, high-status aggressive preschoolers should be the focus of attention in future research. Finally, despite the growing awareness that girls and boys may express aggressive behavior differently (Bjorkqvist et al., 1992; Crick & Grotpeter, 1995; Underwood, 2003), most of the research on peer relations and aggression has been conducted with boys. Greater attention should be devoted to sex differences in the early study of the linkages between direct and indirect/ relational aggression and peer rejection in the preschool years.
384
DETERMINANTS OF AGGRESSION
AN ECOLOGICAL PERSPECTIVE OF PEER INFLUENCE ON AGGRESSIVE BEHAVIOR DURING EARLY CHILDHOOD A full understanding of the role played by the child’s peer affiliates must take into account the fact that these peer relationships are embedded in other, larger social contexts (Dishion, French, & Patterson, 1995b). Accordingly, different levels of the young child’s social world are discussed in the following sections of this chapter. The interactions between friends and clique members are examined first. Typically, however, these interactions are embedded in a larger peer group and take place in settings such as children’s homes, neighborhoods, daycare centers, and schools. These settings vary greatly in terms of the amount of structure, adult supervision, and type of peers encountered, all conditions that may have a bearing on the development of aggressive behaviors. Finally, in analyzing these different contextual “levels,” we examine the roles of parents and teachers in overseeing, supervising, and managing the child’s relationships with peers during the preschool period (Ladd & Le Sieur 1995; Parke & O’Neil, 1999).
Friends and Cliques Although they are rejected by conventional peers, many aggressive preadolescents and early adolescents have friends, and most of them participate in cliques (Boivin & Vitaro, 1995; Cairns et al., 1988). This seems to be true also for young children (Farver, 1996; Snyder, Horsch, & Childs, 1997). For instance, Farver (1996) showed that children at the center of teacher-identified peer networks of 4-year-olds were also the most aggressive. Only a few studies have examined whether homophily with respect to aggression (i.e., the tendency for aggressive children to affiliate with other aggressive children) exists among young children. For instance, members of teacher-identified cliques of 4-year-olds were found to be moderately similar in aggression, as assessed by independent observers (Farver, 1996). Four- to five-year-old children were more likely to select friends who were similar in respect to teacher-rated aggression; that is, 50% of aggressive children’s friends were aggressive, as compared with 12% of nonaggressive children’s friends (Snyder et al., 1997). A further study is worth considering, in spite of the fact it was conducted with grade school children; in this study, clique membership in first grade was associated with similar characteristics in respect to aggression (Estell, Cairns, Farmer, & Cairns, 2002). Thus, the few available studies suggest that aggression-related homophily exists in early childhood at the dyadic and the clique levels, conditions that extend onward, and perhaps increase, into middle childhood and adolescence. Many convergent processes may account for aggression homophily (Kandel, 1978). Aggressive children may associate because they are actively rejected by conventional peers and gradually cast out of the group, leaving them with no other choice than to affiliate with other aggressive children
Peer Relationships and Aggressive Behavior
385
(Patterson, Reid, & Dishion, 1992). It may be that aggressive children proactively select other aggressive children, partly because they share an interest for rough play, but also because their mutual selection allows them to dominate other peers and to have better access to resources in the peer group (Boivin & Vitaro, 1995; Poulin & Boivin, 2000). This latter view is in line with the confluence model of Dishion, Patterson, and Griesler (1994) that posits a gradual convergence of aggressive children into homogeneous groups of aggressive and risk-taking adolescents. The confluence model implies an increasingly active selection of aggressive peers who share similar attitudes toward aggression. Once established, these deviant associations may also accentuate the behavioral similarity by having aggressive children mutually influencing each other, through deviancy training (Dishion et al., 1995a; see the discussion of “mechanisms” below). To the extent that these mechanisms have cumulative effects over time, aggressive children could be progressively trapped in a social environment that is increasingly segregated from that of their better-adjusted peers. There is evidence that aggression homophily increases with age from middle childhood to late adolescence (Dion & Boivin, 2003; Neckerman, 1992). However, whether these processes are operative in the preschool years is not known.
Are (Deviant) Friends Significant for Later Adjustment in Early Childhood? Very few studies have examined this question among preschool children. Snyder et al. (1997) reported a short-term effect of interacting with aggressive peers. They found that the amount of time preschoolers spent with aggressive peers predicted an increase in observed and teacher-rated aggressive behavior over a 3-month interval. Conversely, children who spent minimal time (less than 15%) with aggressive peers showed a decrease in aggression over the subsequent 3-month period. The findings of Snyder et al. (1997) indicate that aggressive friends may have a negative influence during early childhood. However, this effect was measured only over a 3-month period. Whether a similar long-term effect exists at that age is still unknown.
What Are the Mechanisms That Could Account for the Possible Influence of Aggressive Friends? The answer to this question assumes that aggressive peers mutually influence each other through some form of socialization mechanisms and that this influence is not spurious, that is, attributable to a third factor such as a personal disposition or a family factor (see Vitaro, Tremblay, & Bukowski, 2001). Work by Dishion and colleagues on rule-breaking talk during dyadic interactions involving deviant adolescents indicates that positive verbal reinforcement by peers of past, actual, and future deviant behaviors becomes increasingly important in shaping social behavior (Dishion et al., 1995a). This process, labeled “deviancy training,” has received substantial empirical sup-
386
DETERMINANTS OF AGGRESSION
port. For example, the association between exposure to deviant peers by age 10 and growth in arrests, substance use, and sexual intercourse from ages 10 to 18 is mediated by deviancy training at age 14 (Patterson, Dishion, & Yoeger, 2000). However, very little is known about deviancy training in young children. A precise examination of this process with young children is needed to confirm its existence in early childhood. Direct training through conflicts and coercion among friends or clique members may also play a role in explaining how and why affiliation with aggressive friends might instigate or increase aggressive behaviors. In their classic observational study of preschoolers’ aggressive interactions, Patterson, Littman, and Bricker (1967) showed that the principles of operant conditioning maintain aggressive behavior in the peer context. They found that the response to an aggressive behavior conditioned the likelihood of reenacted aggression in the future. When a child who was the victim of aggression reacted by withdrawal and submissive behaviors, the aggressor was more likely to subsequently attack the same target than when the victim retaliated. Negative reinforcement in coercive peer interactions may thus shape young children’s behaviors in the same way it shapes this behavior during child–adult coercive interchanges. Similarly, the use of verbal aggression is followed by termination of a conflict episode for aggressive children, whereas a positive verbal interchange or the absence of a response terminates the conflict episode for nonaggressive children (Snyder & Brown, 1983). In other words, aggressive children are negatively reinforced for using coercive behaviors with their peers, whereas nonaggressive children are reinforced for using nonaversive strategies. Hence, peer interactions during preschool years are sometimes occasions for coercive interchanges among peers which may, under some conditions (e.g., child’s submissiveness, adult and peer tolerance of aggression), serve as learning opportunities for aggression. If these conditions are present, then preschool and early elementary school settings may provide training grounds for further shaping of aggressive behaviors in children who may already have learned them at home. It is not clear, however, whether this type of training is taking place within dyadic friendships or among members of a clique. Given the higher incidence of conflicts among preschool friendships involving at least one aggressive child, this could be the case (Snyder et al., 1997). However that may be, there are possible caveats to the long-term impact of deviancy training in preschool.
Caveats to Deviancy Training in Preschool Why would affiliating with aggressive friends have little long-term effect on the behavior trajectory of young children? One possible reason is that preschool children do not systematically and exclusively affiliate with other aggressive children as much as preadolescents or adolescents do. Indeed, in the three preschool studies reviewed earlier, aggression-related homophily was
Peer Relationships and Aggressive Behavior
387
moderate. Hence, behavioral homophily may not yet be firmly established and specialized during the preschool years. Young aggressive children may have access to a variety of peers and friends, and these experiences may mitigate the negative influence of deviant peers. For instance, to the extent that many aggressive children are not rejected by conventional peers during early childhood, they would be less likely to be deprived of positive socialization experiences. Another related possibility is that the associations among aggressive preschoolers may be short-lived. There is indeed evidence showing that childhood friendships are not always stable and that clique composition may change substantially over a period of weeks, especially when involving aggressive children (Cairns et al., 1995; Parker & Seal, 1996). However, Snyder et al. (1997) found no difference in the stability of strong mutual affiliations of aggressive and nonaggressive children, although aggressive children were less likely than nonaggressive children to establish strong mutual affiliations with classmates. Other researchers also reported that friendships during the preschool years can be relatively stable and intimate (Howes, 1987). Furthermore, young children’s friendships may not carry the same weight as those of older children. Indeed, these early friendships are based on concrete exchanges rather than shared values and interests, loyalty, and mutual support (Bigelow, 1977), qualities that may induce greater mutual influence on behavior. Finally, aggressive preschoolers may simply be more sensitive to adult rules than older aggressive children. Hence, parents and teachers may exert a moderating effect on aggressive friends’ influence by controlling children’s behavior more closely and effectively. Parental supervision may also play a similar role. In contrast to those of adolescence, the preschool peer environments are more closely structured and supervised by adults. However, even during preschool years the social settings may vary greatly in terms of adult structure and supervision, and this could make a difference with respect to exposure to deviant peers. For instance, the degree of parents’ monitoring of their kindergarten child has been associated with the child’s exposure to deviancy-producing experiences and aggressive peers in a variety of contexts (Kilgore, Snyder, & Lentz, 2000). Longitudinal studies that examine whether aggression-related homophily increases from the preschool years to late childhood are clearly needed. In addition, it would be important to determine whether homophily is driven by specific forms of aggressive behaviors (i.e., proactive aggression rather than reactive aggression), as this seems to be the case with older children (Poulin & Boivin, 2000). Moreover, it would be relevant to learn about the qualitative features of friendships during this period in order to understand why the early encounters with aggressive children may or may not leave traces in the long term.
388
DETERMINANTS OF AGGRESSION
The Larger Peer System By and large, our discussion of the peer influence on aggression in early childhood has been limited so far to the peers with whom the child forms affective bonds and spends more time interacting with, such as friends or clique members. However, the larger peer group to which the child is exposed may also have a unique impact on the level of aggression, especially in the case of stable peer groups, such as daycare groups or kindergarten classrooms. As already suggested, group norms with respect to aggression seem to be very salient during early childhood and may interact with initial dispositions for aggression. For example, it was found that a high level of aggression in a first-grade classroom uniquely contributed to an increase in aggression over a 5-year period for children who were already displaying high levels of aggression (Kellam, Ling, Merisca, Brown, and Ialongo, 1998). In other words, aggressive firstgraders may become more aggressive when they are placed in a classroom where the level of aggression is high. A recent observational study found evidence of a contagion effect in daycare: Children were significantly more likely to behave aggressively after an aggressive act had occurred than when no aggressive act had occurred (Goldstein, Arnold, Rosenberg, Stowe, & Ortiz, 2001). Interestingly, contagion was stronger for aggressive behaviors directed toward another child than for aggressive behaviors directed at an object. Aggressive behaviors are not isolated acts in the classroom, but rather seem to be part of a larger group system, characterized by chain reactions. A next step would be to examine whether such contagion effects are associated with changes in individual behaviors. These findings suggest that the ecology of peer influence during early childhood is complex and that all levels of the peer system should be considered.
The Behavior Settings As early as 2 years of age, children start to spend more time with peers than with adults (Ellis, Rogoff, & Cromer, 1981). During the early years, these peer interactions are mainly confined to the home, where parents are more likely to orchestrate and supervise social interactions with siblings and other children. Indeed, there are some indications that interactions with peers within the family may account for individual differences in early developmental trajectories of aggression. In particular, siblings appear to be a predominant source of agonistic behavior for preschoolers and may provide a training ground in aggressive behaviors for younger siblings (Dunn & Kendrick, 1982). For instance, Tremblay et al. (2004) found that having a sibling of close age (i.e., less then 2 years in age difference) was the best overall predictor of being described as highly physically aggressive between 17 and 42 months of age. These findings underscore the potential contributing role of sibling relationship for revealing, and perhaps fostering, aggressive behaviors very early in life.
Peer Relationships and Aggressive Behavior
389
However, parental monitoring of these aggressive episodes are key to mitigating this potential negative effect. There are different ways in which parents may influence their child’s peer interactions and relationships. For example, parents can provide opportunities for positive friendships while prohibiting deviant ones. Parents may provide direct opportunities for peer interactions by organizing social encounters between their child and selected other children. They can also provide indirect opportunities by coaching their child about peer interactions (Ladd & LeSieur, 1995). In doing so, parents may improve their child’s ability to access more competent peers (Parke & Ladd, 1992). Parental practices such as monitoring their children’s peer affiliations (Dishion & McMahon, 1998) and prohibiting their interactions with deviant peers (Mounts, 2001) have been extensively studied during the late childhood/ adolescence period. More research is needed on how parents manage young children’s affiliation with aggressive peers. As children get older, they interact with peers in a larger variety of settings (Sinclair, Pettit, Harrist, Dodge, & Bates, 1994). Other significant adults, such as classroom teachers, daycare staff, or after-school activity leaders, may then also contribute in shaping a child’s peer environments (Searcy & Meadows, 1994). These settings are likely to vary with respect to the degree of structure and supervision provided by adults, as well as of the characteristics of the peers (i.e., heterogeneity in age, level of peer aggression, etc.). Some settings may be more likely than others to provide opportunities for associations among aggressive children, as well as for deviancy training. Typically, these are settings where the demands placed on the child and the degree of adult supervision are low. Three behavioral settings are relevant to this discussion: (1) the neighborhood, (2) the childcare facility, and (3) the school.
The Neighborhood During the preschool years, children gradually begin to form informal peer groups in their neighborhoods (Ladd & Golter, 1988). These peer groups are typically mixed-aged groups (Ellis et al., 1981). The behavioral characteristics of a child’s friends in this setting are likely to vary according to the socioeconomic status of the families living in the neighborhood. In socially disadvantaged neighborhoods, such as public housing neighborhoods, the density of aggressive children is higher, potentially increasing the child’s exposure to deviant peers. In these neighborhoods, close parental supervision is crucial for limiting the child’s access to deviant peers (Pettit, Bates, Dodge, & Meece, 1999). Deviant friendships in a neighborhood have been studied mostly with adolescents (Dishion et al., 1995b), but this context is also relevant for younger children because their friendship networks are mostly composed of peers from the same neighborhood (Ladd & Golter, 1988). According to Sinclair et al. (1994), the neighborhood is where 4–5-year-old children, especially children from low SES families, spend the greatest amount of time with peers. This is also where they are most frequently exposed to aggressive peers. Devi-
390
DETERMINANTS OF AGGRESSION
ancy training in these settings is likely, given the lack of structure and the lack of adult supervision. However, it is not clear whether these experiences predict an increase in children’s aggressive behaviors over time, above and beyond personal dispositions, family characteristics, and social interactions at the dyadic level.
Daycare and After-School Care Programs In western societies, the amount of time young children spend in daycare has been steadily increasing in the last decades (NICHD Early Child Care Research Network, 2003). In typical daycare settings, children spend most of their time interacting with peers. Preliminary evidence seems to indicate that time spent in daycare may be associated with higher rates of aggression (Belsky, 2002; NICHD Early Child Care Research Network, 2003). Peer effects may partly be responsible for this (Fabes, Hanish, & Martin, 2003). At about 2–3 years of age, a clear social structure involving specific patterns of affiliations and cliques emerges (Strayer & Santos, 1996). Children of that age, especially boys (Hawley, 1999), often use aggressive means to reach high status in the social structure. Some of these cliques may thus contribute to the use of aggression by modeling or reinforcement processes. As already alluded to, the degree of adult supervision and structure may play a moderating role with respect to the possibility of a peer influence on aggression in daycare settings. Indeed, research on the impact of daycare on children’s adjustment constantly points to adult variables (e.g., child–staff ratio, staff qualifications) to explain the observed effects (Vandell & Shumow, 1999). For example, Goldstein et al. (2001) found that the immediate presence of the daycare teacher was associated with less contagion of aggression in 36–72-month-old children. With close supervision of peer interactions, adults may simply prevent aggressive interactions from occurring at all. Arrangements for child care are still necessary when children reach kindergarten and first grade. Indeed, because of the increasing rates of maternal employment, the 3:00–5:00 P.M. period represents a new opportunity for additional peer interactions (Vandell & Shumow, 1999). Some arrangements for this period of time involve a low level of peer contact, such as staying with a parent or with a sitter, whereas others involve more peer contact, such as after-school centers and out-of-school formal activities including lessons (music, dance), clubs (scouts), and organized sports. After-school centers vary considerably in respect to child–staff ratio, degree of structure, types of activities, and heterogeneity in children’s age (Vandell & Shumow, 1999). Some of these settings may be a fertile ground for peer influence on aggression. Pierce, Hamm, and Vandell (1999) found that negative peer interactions in afterschool care settings were associated with boys displaying high levels of externalizing behaviors at school, suggesting continuity in peer difficulties from one setting to another. Similar findings concerning after-school care were also reported (Vandell & Corasaniti, 1988).
Peer Relationships and Aggressive Behavior
391
Deviant peer processes are less likely to take place in out-of-school formal activities: First-graders involved in these activities tend to show fewer behavioral problems (Posner & Vandell, 1994). In these settings, aggressive behaviors are not tolerated because they disrupt the activities. The levels of structure and adult supervision are usually high, and participants also tend to come from higher-income families.
The School When they reach kindergarten, children spend several hours every day interacting with a stable group of same-age classmates. Most of the research on children’s friendship networks and aggression reviewed earlier has been conducted in such same-age classroom settings. However, the school structure may also influence the nature of the peer relations of aggressive children. As early as first grade, aggressive children may be assigned to a segregated setting, such as a special classroom that involves other students with behavior problems. These special classrooms create an environment where aggressive children are given opportunities for developing friendships with each other (Searcy & Meadows, 1994). In addition to increasing aggression through the mechanisms already described, these friendship groups may contribute to the creation of an in-group/out-group type of social division between aggressive children and their mainstream nonaggressive peers.
CONCLUSION Preschoolers’ peer relationships are complex. Preschool children may face a variety of peer experiences in a diversity of contexts, and it is important not to neglect the potential role of these experiences in the development of aggressive behaviors. A review of the research suggests that starting in preschool, there is a bidirectional and differentiated association between peer relationship problems and aggressive behaviors. Hostile/reactive aggressive behaviors, rather than instrumental/proactive forms of aggression, appear associated with, and perhaps augmented by, peer rejection. Homophily, in respect to aggression, is noticeable during the preschool years. This pattern of association should be of concern. Be that as it may, the empirical evidence linking peer relationships and aggressive behaviors in preschool is still thin. For instance, most of the research effort has been conducted with kindergarten children, and we still know very little about the association between early experiences with peers and aggressive behaviors in prekindergarten children. Nor is it clear whether these early preschool peer experiences have long-term consequences. Longitudinal studies, starting early in life, are necessary to evaluate the potential impact of peer experiences in the long term with respect to subtypes of aggressive behaviors. These future studies should measure the various
392
DETERMINANTS OF AGGRESSION
forms of peer relationships at different levels (e.g., dyads, cliques, groups) and in a variety of contexts (i.e., home, daycare, neighborhood). They should examine the microprocesses involved in interactions between aggressive children and their friends, in addition to the mediating and moderating processes linking aggressive behaviors and peer experiences. Specifically, gender differences, as well as the direct and indirect role of parents, other adults, and siblings, should be documented more systematically. Research may also take advantage of genetically informative and longitudinal designs to test specific models relative to the personal and family determinants of peer relationships, and to examine the specific role of peer relationships in later adjustment. Of special interest is the possibility that peer experiences may account for a significant part of nonshared environmental influences on children’s later aggressive behaviors. Such studies could also help unravel the developmental processes presiding over the development of children’s peer relationships. Specifically, there is a possibility that gene– environment correlation may partly explain the nature of young children’s peer experiences, for instance, rejection by peers of a child’s behavior, but also his or her selection of specific peers (see Deater-Deckard, 2001). Finally, future research may also examine the feasibility and the impact of experimental interventions aimed at exposing young aggressive children to nonaggressive peers and modifying the composition of their social networks. These interventions could serve as experimental manipulations in order to test more decisively the putative causal role of these peer-related variables in respect to aggressive behaviors.
ACKNOWLEDGMENTS Preparation of this chapter was supported in part by grants from the Canadian Institute of Health Research and the Canada Research Chair Program awarded to Michel Boivin.
REFERENCES Alsaker, F. D., & Valkanover, S. (2001). Early diagnosis and prevention of victimization in kindergarten. In J. Juvonen & S. Graham (Eds.), Peer harassment in school: The plight of the vulnerable and victimized (pp. 175–195). New York: Guilford Press. Asher, S. R., & Coie, J. D. (1990). Peer rejection in childhood. New York: Cambridge University Press. Asher, S. R., Rose, A. J., & Gabriel, S. W. (2001). Peer rejection in everyday life. In M. Leary (Ed.), Interpersonal rejection (pp. 105–142). New York: Oxford University Press. Belsky, J. (2002). Quantity counts: Amount of child care and children’s socioemotional development. Journal of Developmental and Behavioral Pediatrics, 23(3), 167– 170.
Peer Relationships and Aggressive Behavior
393
Bigelow, B. J. (1977). Children’s friendship expectations: A cognitive developmental study. Child Development, 48, 246–253. Bjorkqvist, K., Lagerspetz, K. M. J., & Kaukiainen, A. (1992). Do girls manipulate and boys fight? Developmental trends in regard to direct and indirect aggression. Aggressive Behavior, 18, 117–127. Boivin, M., & Bégin, G. (1986). Temporal reliability and validity of three sociometric status assessments with young children. Canadian Journal of Behavioral Sciences, 18, 167–172. Boivin, M., Dodge, K. A., & Coie, J. D. (1995). Individual–group behavioral similarity and peer status in experimental play groups: The social misfit revisited. Journal of Personality and Social Psychology, 69(2), 269–279. Boivin, M., Dorval, M., & Bégin, G. (1990). Statut auprès des pairs et comportement social à la maternelle: Profils individuels et dyadiques. Revue Canadienne des Sciences du Comportement, 22, 121–139. Boivin, M., & Hymel, S. (1997). Peer experiences and social self-perceptions: A twostage mediational model. Developmental Psychology, 33, 1, 135–145. Boivin, M., Hymel, S., & Hodges, E. V. E. (2001). Toward a process view of peer rejection and harassment. In J. Juvonen & S. Graham (Eds.), Peer harassment in school: The plight of the vulnerable and victimized (pp. 265–289). New York: Guilford Press. Boivin, M., Tessier, O., & Strayer, F. F. (1985). La cohérence des choix sociométriques et l’évaluation de l’amitié chez les enfants d’âge pré-scolaire. Enfance, 329–343. Boivin, M., & Vitaro, F. (1995). The impact of peer relationships on aggression in childhood: Inhibition through coercion or promotion through peer support. In J. McCord (Ed.), Coercion and punishment in long-term perspectives (pp. 183–197). Cambridge, UK: Cambridge University Press. Buhler, C. (1931). The social behavior of the child. In C. Murchison (Ed.), A handbook of child psychology (pp. 393–431). New York: Russell & Russell. Cairns, R. B., Cairns, B. D., Neckerman, H. J., Gest, S. D., & Gariépy, J.-L. (1988). Social networks and aggressive behavior: Peer support or peer rejection? Developmental Psychology, 24, 815–823. Cairns, R. B., Leung, M.-C., Buchanan, L., & Cairns, B. D. (1995). Friendships and social networks in childhood and adolescence: Fluidity, reliability, and interrelations. Child Development, 66, 1330–1345. Caplan, M., Vespo, J. E., Pedersen, J., & Hay, D. F. (1991). Conflict over resources in small groups of 1- and 2-year-olds. Child Development, 62, 1513–1524. Coie, J. D. (1990). Toward a theory of peer rejection. In S. R. Asher & J. D. Coie (Eds.), Peer rejection in childhood (pp. 365–402). New York: Cambridge University Press. Coie, J. D., & Dodge, K. A. (1983). Continuities and changes in children’s social status: A five-year longitudinal study. Merrill-Palmer Quarterly, 29, 261–281. Coie, J. D., & Dodge, K. A. (1998). Aggression and antisocial behaviour. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology: Social emotional and personality development (pp. 779–862). New York: Wiley. Coie, J. D., & Kupersmidt, J. B. (1983). A behavioral analysis of emerging social status in boys’ groups. Child Development, 54, 1400–1416. Crick, N. R., Casas, J. F., & Ku, H. C. (1999). Relational and physical forms of peer victimization in preschool. Developmental Psychology, 35(2), 376–385. Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social informationprocessing mechanisms in children’s social adjustment. Psychological Bulletin, 115(1), 74–101.
394
DETERMINANTS OF AGGRESSION
Crick, N. R., & Grotpeter, J. K. (1995). Relational aggression, gender, and social-psychological adjustment. Child Development, 66, 710–722. Deater-Deckard, K. (2001). Annotation: Recent research examining the role of peer relationships in the development of psychopathology. Journal of Child Psychology and Psychiatry, 42(5), 565–579. Dion, É., & Boivin, M. (2003). Band apart: A developmental perspective on aggressive boys’ peer affiliation. Manuscript submitted for publication. Dishion, T. J., Andrews, D. M., & Crosby, L. (1995a). Antisocial boys and their friends in early adolescence: Relationship characteristics, quality, and interactional process. Child Development, 66, 139–151. Dishion, T. J., French, D. C., & Patterson, G. R. (1995b). The development and ecology of antisocial behavior. In D. Cicchetti & D. Cohen (Eds.), Manual of developmental psychopathology (pp. 421–471). New York: Wiley. Dishion, T. J., & McMahon, R. J. (1998). Parental monitoring and the prevention of child and adolescent problem behavior: A conceptual and empirical formulation. Clinical Child and Family Psychology, 1, 61–75. Dishion, T. J., Patterson, G. R., & Griesler, P. C. (1994). Peer adaptations in the development of antisocial behavior: A confluence model. In L. R. Huesmann (Ed.), Aggressive behavior: Current perspectives (pp. 61–95). New York: Plenum Press. Dodge, K. A. (1983). Behavioral antecedents of peer social status. Child Development, 53, 1386–1399. Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990). Mechanisms in the cycle of violence. Science, 250, 1678–1683. Dodge, K. A., Coie, J. D., Pettit, G., & Price, J. (1990). Peer status and aggression in boys’ groups: Developmental and contextual analyses. Child Development, 61, 1289– 1309. Dodge, K. A., Lansford, J. E., Salzer Burks V., Bates, J. E., Pettit, G. S., Fontaine, R., & Price, J. M. (2003). Peer rejection and social information-processing factors in the development of aggressive behavior problems in children. Child Development, 74(2), 374–393. Dunn, J., & Kendrick, C. (1982). Siblings. Cambridge, MA: Harvard University Press. Ellis, S., Rogoff, B., & Cromer, C. C. (1981). Age segregation in children’s social interaction. Developmental Psychology, 29, 63–73. Estell, D. B, Cairns, R. B., Farmer, T. W., & Cairns, B. D. (2002). Aggression in inner-city early elementary classrooms: Individual and peer-group configurations. Merrill– Palmer Quarterly, 48, 52–76. Fabes, R. A., Hanish, L. D., & Martin, C. L. (2003). Children at play: The role of peers in understanding the effects of child care. Child Development, 74(4), 1039–1043. Farver, J. A. M. (1996). Aggressive behavior in preschoolers’ social networks: Do birds of a feather flock together? Early Childhood Research Quarterly, 11, 333–350. Goldstein, N. E., Arnold, D. H., Rosenberg, J. L., Stowe, R. M., & Ortiz, C. (2001). Contagion of aggression in day care classrooms as a function of peer and teacher responses. Journal of Educational Psychology, 4, 708–719. Goodenough, F. L. (1931). Anger in young children. Minneapolis: University of Minnesota. Hartup, W. W. (1983). Peer relations. In P. H. Mussen & E. M. Hetherington (Eds.), Handbook of child psychology: Vol. 4. Socialization, personality, and social development (4th ed., pp. 103–196). New York: Wiley.
Peer Relationships and Aggressive Behavior
395
Hartup, W. W. (1996). The company they keep: Friendships and their developmental significance. Child Development, 67, 1–13. Hartup, W. W., Glazer, J. A., & Charlesworth, R. (1967). Peer reinforcement and sociometric status. Child Development, 38, 1017–1024. Hartup, W. W., & Laursen, B. (1991). Relationships as developmental contexts. In R. Cohen, & A. W. Siegel (Eds.), Context and development (pp. 253–279). Hillsdale, NJ: Erlbaum. Hawley, P. H. (1999). The ontogenesis of social dominance: A strategy-based evolutionary perspective. Developmental Review, 19, 97–132. Hodges, V. E., Boivin, M., Vitaro, F., & Bukowski, B. (1999). The power of friendship: Protecting against an escalating cycle of peer victimization. Developmental Psychology, 35(1), 94–101. Howes, C. (1987). Peer interaction of young children. Monographs of the Society for Research in Child Development, 53(No. 217). Hymel, S., Wagner, E., & Butler, L. (1990). Reputation bias: View from the peer group. In S. R. Asher & J. D. Coie (Eds.), Peer rejection in childhood (pp. 156–186). New York: Cambridge University Press. Kandel, D. B. (1978). Homophily, selection, and socialization in adolescent friendships. American Journal of Sociology, 84, 427–436. Kellam, S. G., Ling, X., Merisca, R., Brown, C. H., & Ialongo, N. (1998). The effect of the level of aggression in the first grade classroom on the course and malleability of aggressive behavior into middle school. Development and Psychopathology, 10, 165–185. Kilgore, K., Snyder, J., & Lentz, C. (2000). The contribution of parental discipline, parental monitoring, and school-based risk to early-onset conduct problems in African-American boys and girls. Developmental Psychology, 36, 1–11. Kochenderfer, B. J., & Ladd, G. W. (1996). Peer victimization: Cause or consequence of school maladjustment? Child Development, 67, 1305–1317 Ladd, G. W., & Golter, B. S. (1988). Parents’ initiation and monitoring of children’s peer contacts : Predictive of children’s peer relations in nonschool and school settings? Developmental Psychology, 24, 109–117. Ladd, G. W., & Le Sieur, K. D. (1995). Parents’ and children’s peer relationships. In M. H. Bornstein (Ed.), Handbook of parenting (pp. 377–409). Hillsdale, NJ: LEA. Lafrenière, P. J., Strayer, F. F., & Gauthier, R. (1984). The emergence of same-sex affiliative preferences among preschool peers: A developmental/ethological perspective. Child Development, 55, 1958–1965. Maccoby, E. E. (1988). The two sexes: Growing up apart, coming together. Cambridge, MA: Harvard University Press. Martin, C. L., & Fabes, R. A. (2001). The stability and consequences of young children’s same-sex peer interactions. Developmental Psychology, 37, 431–446. Masters, J. C., & Furman, W. (1981). Popularity, individual friendship selection, and specific peer interaction among children. Developmental Psychology, 3, 344–350. McDougall, P., Hymel, S., Vaillancourt, T., & Mercer, L. (2001). The consequences of early peer rejection. In M. Leary (Ed.), Interpersonal rejection (pp. 213–247). New York: Oxford University Press. Mounts, N. S. (2001). Young adolescents’ perceptions of parental management of peer relationships. Journal of Early Adolescence, 21, 92–122. Neckerman, H. J. (1992). A longitudinal investigation of the stability and fluidity of so-
396
DETERMINANTS OF AGGRESSION
cial networks and peer relationships of children and adolescents. Dissertation Abstracts International, 53, 9–B. NICHD Early Child Care Research Network (2003). Does amount of time spent in child care predict socioemotional adjustment during the transition to kindergarten? Child Development, 74(4), 976–1005. Parke, R. D., & Ladd, G. W. (1992). Family–peer relationships: Modes of linkage. Hillsdale, NJ: Erlbaum. Parke, R. D., & O’Neil, R. (1999). Social relationships across contexts: Family–peer linkages. In W. A. Collins & B. Laursen (Eds.), Relationships as developmental contexts: Minnesota Symposia on Child Psychology (Vol. 30, 211–239). Hillsdale, NJ: LEA. Parker, J. G., & Asher, S. R. (1987). Peer acceptance and later personal adjustment: Are low-accepted children “at risk”? Psychological Bulletin, 102(3), 357–389. Parker, J. G., & Seal, J. (1996). Forming, losing, renewing, and replacing friendships: Applying temporal parameters to the assessment of children’s friendship experiences. Child Development, 67, 2248–2268. Patterson, G. R., Dishion, T. J., & Yoerger, K. (2000). Adolescent growth in new forms of problem behavior: Macro- and micro-peer dynamics. Prevention Science, 1, 3–13. Patterson, G. R., Littman, R. A., & Bricker, W. (1967). Assertive behavior in children: A step toward a theory of aggression. Monographs of the Society for Research in Child Development, 32(No. 5). Patterson, G. R., Reid, J. B., & Dishion, T. J. (1992). A social learning approach: Vol. 4. Antisocial boys. Eugene, OR: Castalia Press. Pellegrini, A. D. (in press). Sexual segregation in childhood: A review of evidence for two hypotheses. Animal Behaviour. Pettit, G. S., Bates, J. E., Dodge, K. A., & Meece, D. W. (1999). The impact of after-school peer contact on early adolescent externalizing problems is moderated by parental monitoring, perceived neighbourhood safety, and prior adjustment. Child Development, 70, 768–778. Pierce, K. M., Hamm, J. V., & Vandell, D. L. (1999). Experiences in after-school programs and children’s adjustment in first-grade classrooms. Child Development, 70, 756–767. Posner, J. K., & Vandell, D. L. (1994). Low-income children’s after-school care: Are there beneficial effects of after-school programs? Child Development, 65, 440–456. Poulin, F., & Boivin, M. (2000). The role of proactive aggression and reactive aggression in the formation and development of boys’ friendships. Developmental Psychology, 36, 233–240. Prinstein, M. J., & Cillessen, A. H. N. (2003). Forms and functions of adolescent peer aggression associated with high levels of peer status. Merrill–Palmer Quarterly, 49, 310–342. Rodkin, P. C., Farmer, T. W., Pearl, R., & Van Acker, R. (2000). Heterogeneity of popular boys: Antisocial and prosocial configurations. Developmental Psychology, 36, 14– 24. Ross, H. S., Lollis, S. P., & Elliot, C. (1982). Toddler–peer communication. In K. H. Rubin & H. S. Ross (Eds.), Peer relationships and social skills in childhood (pp. 73– 98). New York: Springer-Verlag. Rubin, K. H., Bukowski, W., & Parker, J. G. (1998). Peer interactions, relationships, and groups. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology: Social emotional and personality development (Vol. 3, pp. 619–700). New York: Wiley.
Peer Relationships and Aggressive Behavior
397
Rubin, K. H., Watson, K., & Jambor, T. (1978). Free play behaviors in preschool and kindergarten children. Child Development, 49, 534–536. Searcy, S., & Meadows, N. B. (1994). The impact of social structures on friendship development of children with behavior disorders. Education and Treatment of Children, 17, 255–266. Selman, R. L. (1980). The growth of interpersonal understanding. New York: Academic Press. Sinclair, J. J., Petit, G. S., Harrist, A. W., Dodge, K. A., & Bates, J. E. (1994). Encounters with aggressive peers in early childhood: Frequency, age differences, and correlates of risk for behaviour problems. International Journal of Behavioral Development, 17, 675–696. Snyder, J., & Brown, K. (1983). Oppositional behavior and noncompliance in preschool children. Behavioral Assessment, 5, 333–348. Snyder, J., Horsch, E., & Childs, J. (1997). Peer relationships of young children: Affiliative choices and the shaping of aggressive behavior. Journal of Clinical Child Psychology, 26, 145–156. Stormshak, E. A., Bierman, K. L., Bruschi, C., Dodge, K. A., & Coie, J. D. (1999). The relation between behavior problems and peer preference in different classroom contexts. Child Development, 70, 169–182. Strayer, F. F. (1989). Co-adaptation within the early peer group: A psychobiological study of social competence. In B. H. Schneider, G. Attili, J. Nadel, & R. Weissberg (Eds.), Social competence in developmental perspective (pp. 145–174). Boston: Kluwer Academic. Strayer, F. F., & Santos, A. J. (1996). Affiliative structures in preschool peer groups. Social Development, 5, 117–130. Strayer, F. F., & Trudel, M. (1984). Developmental changes in the nature and function of social dominance among young children. Ethology and Sociobiology, 5, 279–295. Tremblay, R. E. (2000). The development of aggressive behavior during childhood: What have we learned in the past century? International Journal of Behavioral Development, 24, 129–141. Tremblay, R. E., Nagin, D. S., Séguin, J. R., Zoccollillo, M., Zelazo, P., Boivin, M., Pérusse, D., & Japel, C. (2004). Physical aggression during early childhood: Trajectories and predictors. Pediatrics, 114(1), e43–e50. Underwood, M. K. (2003). Social aggression in girls. New York: Guilford Press. Vandell, D. L., & Corasaniti, M. A. (1988). The relations between third graders’ afterschool care and social, academic, and emotional functioning. Child Development, 59, 868–875. Vandell, D. L., & Shumow, L. (1999). After-school child care programs. The Future of Children, 9, 64–80. Vandell, D. L., Wilson, K. S., & Buchanan, N. R. (1980). Peer interaction in the first year of life: An examination of its structure, content, and sensitivity to toys. Child Development, 51, 481–488. Vitaro, F., Tremblay, R. E., & Bukowski, W. M. (2001). Friends, friendships, and conduct disorders. In J. Hill & B. Maughan (Eds.), Conduct disorder in childhood (pp. 346– 378). Cambridge, UK: Cambridge University Press. Zahn-Waxler, C., Radke-Yarrow, M., & King, R. A. (1979). Child-rearing and children’s prosocial initiations toward victims of distress. Child Development, 50, 319–330.
DETERMINANTS Social Capital and Physical OF AGGRESSION Violence
19 Social Capital and Physical Violence U BERTO G ATTI and R ICHARD E. T REMBLAY
Man is made fit for Society not by nature, but by training. —THOMAS HOBBES, On the Citizen (1647/1998, p. 25) God makes all things good; man meddles with them and they become evil. —ROUSSEAU, Émile (1762/1986, p. 5)
Although Rousseau and Hobbes disagreed on the nature of “natural” dispositions, they agreed on the fact that the social environment has an important impact on social development. It is generally accepted today, as it has been by philosophers since Plato and Aristotle, that the social development of children is influenced by the environment in which they live, and in particular by the relationships they have with their own families and communities. The assumption is that “good” social environments will contribute to good psychosocial development, whereas, on the contrary, “bad” social environments will facilitate social adjustment problems. The aim of this chapter is to review the literature on the effects of social environments on the development of physical aggression. We have chosen to consider the social environment in terms of a relatively recent unifying concept, social capital. In recent years, this concept has increasingly come to the fore. Initially formulated to explain such phenomena as social stratification and economic development, it has since been utilized to interpret various aspects of life in society. Social capital has been defined as interpersonal relationships that facilitate action (Bourdieu, 1980). Social capital therefore represents an aspect of social reality that individuals (or groups) utilize as a resource in order to further their own interests and to achieve goals that would otherwise be beyond 398
Social Capital and Physical Violence
399
their reach (Coleman, 1990); that is to say, it is the investment and the instrumental use of the resources inherent in social networks (Lin, 1999). Or, again, it has been seen as interpersonal trust (Fukuyama, 1995), or networks, norms, and trust that enable participants to act together more effectively to pursue shared objectives (Putnam, 1993). Social capital constitutes a resource that is often compared to physical capital (tools, machinery, etc.) and human capital (skills acquired by the individual through education and socialization). Physical capital is created through the transformation of raw materials; human capital through the transformation of persons; social capital through changes in the relationships between persons (Coleman, 1990). These succinct definitions hint at two different (though not entirely independent) interpretations of the concept of social capital: The first, which is chiefly microsocial, stresses the importance of relationships, social networks or links that the individual can utilize; the second, which is mainly macrosocial, attaches greater importance to the norms of reciprocity and civic engagement as the features of a given society. In general, however, it has to be admitted that the definitions proposed for social capital are quite numerous; indeed, debate is currently under way with a view to clarifying and defining the concept, the excessive extension of which may compromise its heuristic value (Portes, 1998). The notion that social capital is linked to the development of children has been argued especially by Coleman (1990), who draws a distinction between social capital within the family and outside the family. In reality, however, leaving aside this dichotomy, it can be claimed that many diverse forms of social capital exist and that these can be detected at different levels, ranging from the microsocial, which centers on the family, to the macrosocial, which takes the nation, region, or city as its unit of analysis; at an intermediate level, the district or neighborhood may be used as the unit of observation. Moreover, social capital may exert different effects in different phases of development. It therefore seems appropriate to reflect on how the different types of social capital might influence aggressive behavior at various ages. By setting the developmental phases against the greater or lesser breadth of the context in which the effect is felt, we created a classification for our analysis of the effects of social capital. Table 19.1 illustrates the hypothesis that, in general, social capital at the microsocial level acts primarily in the early years of life, and at the macrosocial level its effects are chiefly felt during adulthood.
THE CHILD WELFARE SYSTEM A form of social capital that may affect children’s adjustment, and therefore the likelihood of aggressive behavior, is constituted by social and financial assistance to parents (generally mothers) in difficulty. In modern societies, welfare systems have been developed that loosen the bonds between the citizen and the labor market and compensate for financial disadvantages, thereby
400
DETERMINANTS OF AGGRESSION
TABLE 19.1. Action of Social Capital at Different Ages and in Different Contexts 0–2 years 3–5 years 6–11 years 12–18 years > 18 years Child welfare system
+
+
+
Family
+
+
+
+
Daycare, preschool, school
+
+
+
+
Peers
+
+
+
+
Neighborhood
+
+
+
Town or region
+
+
Gang
+
+
Violent subculture and organized crime
+
+
making health, social, and educational services available to people who would otherwise be unable to afford them. The relationships between such investments and crime rates have generally been examined through synchronous data. Messner and Rosenfeld (1997), for example, showed that investments in welfare were able to reduce anomic pressure and, consequently, homicide rates in the various nations considered. In an investigation involving several countries, Pampel and Gartner (1995) examined the effect of demographic changes—particularly increases in the number of young people—on general homicide rates. These authors found that in countries where institutions for collective social protection were weak, an increase in the number of young people led to a rise in the homicide rate, whereas no such effect was recorded in countries where these institutions were strong. In comparing homicide rates in various countries, Savolaien (2000) found that the effect of economic inequality on this type of crime was nonexistent in countries with high rates of investment in welfare. DeFronzo conducted a series of studies on the relationship between the financial level of welfare aid and crime rates. In a study of 39 U.S. metropolitan areas, he found, after controlling for other social and economic factors, that the level of public assistance to poor families correlated negatively with the rates of various crimes, including homicide, sexual assault, and burglary (DeFronzo, 1983). In a subsequent study (DeFronzo,1997) he examined the effect of the mean level of social assistance, per person, in 141 U.S. cities, on homicide, while controlling for a whole range of socioeconomic variables, and found that the level of social assistance (Aid to Families with Dependent Children, AFDC) had a direct negative impact on homicide rates; moreover, it exerted an indirect negative effect on these rates through its association with family status (by reducing the number of families headed by a single mother). The high level of economic support therefore apparently reduced homicide rates through two different causal mechanisms. Note that all these studies concerned general crime rates, resulting from both juvenile and adult crime.
Social Capital and Physical Violence
401
These results fall within the framework of the broader question of the effects of welfare intervention that have been debated in recent decades. Indeed, such intervention is regarded by some as being counterproductive, fostering dependence and passivity, stifling effort, and promoting marginalization; others, however, regard it as an indispensable tool for reducing the deprivation of those who belong to the marginal sectors of society, and, in the final analysis, for improving their social adjustment. Awareness of the effects of welfare intervention acquired particular importance in the United States when, in 1996, a radical change in welfare policy took place. In that year the AFDC program was abolished and was replaced by the more restrictive Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA), which limited the period of eligibility for aid and imposed prescriptions in the field of work. As a result, the number of families on assistance, which had peaked at 5 million in 1994, had been halved by June 1999 (Ku, 2001) and further declined to about 2 million in 2001. According to The Economist (2001), the success of this reform went beyond the most optimistic forecasts, yet other commentators judged it to be a substantial failure (Cancian, 2001). Still other experts, however, viewed the results of the reform as partly positive and partly negative. Zaslow, Tout, Smith, and Moore (1998), for instance, saw the results as positive in those cases in which mothers actually succeeded in finding and keeping a job, thus increasing family income, but as potentially negative when mothers were unable to find or keep a job once the maximum permitted time on assistance had been reached. Morris (2002) analyzed the main evaluation studies and concluded that the new welfare structure may be advantageous for children but potentially harmful to adolescents; since the reform was implemented, smoking, alcohol use, small-time delinquency, and maladjustment at school appear to have increased among teenagers, who receive less supervision from their families. A particular negative effect of the reform has been highlighted by Paxon and Waldfogel (2001), who claim that as a result of reduced aid to families in difficulty, cases of child abuse have increased. The heated debate and the research efforts aimed at assessing the results of aid programs emphasize the importance of welfare systems in the politics of a nation, especially in regard to their repercussions on families and child development. From a criminological point of view, it has been suggested (Vila, 1994) that all of the factors that enhance child development, such as health care, education, and parenting, have a long-term effect and are able to reduce crime a generation later (the “nurturance” hypothesis). To test this hypothesis, Savage and Vila (1997) carried out analyses, with countries as units, aimed at ascertaining whether indicators of child welfare were able to predict crime levels 20 years later. The nurturance hypothesis was not supported for crime in general, but only for crimes of violence. In a further study, again carried out on a sample of countries, Savage and Vila (2002) examined the relationships between changes in child care services and changes in crime rates recorded in a later period. As covariates, the authors considered measures of economic
402
DETERMINANTS OF AGGRESSION
growth and inequality, factors that reflected the opportunities to commit crimes, and demographic factors. The results supported the nurturance hypothesis: (1) Improvements in primary school enrollments and in the number of hospital beds available per capita predicted declines in homicide rates, (2) improvements in infant mortality rates, life expectancy at birth, and the number of hospital beds predicted declines in the rates of violent crime, (3) and improvements in life expectancy at birth were associated with lower rates of theft. Similar results were obtained by Gatti, Schadee, and Tremblay (2002a), who tested the hypothesis that better maternal and infant health care will be reflected in reduced crime rates a generation later. They analyzed the relationship between infant mortality and the lagged rates of homicide (20 years later) in the 95 Italian provinces, controlling for gross domestic product. Another form of social capital is constituted by programs that provide support for high-risk families. The Elmira Home Visitation Study (Olds et al., 1997) is one of the best examples. This randomized clinical trial showed that home visitation by a nurse, starting during pregnancy and lasting until the child’s second birthday, significantly reduced child abuse by the parents, substance abuse and arrests among the mothers, and juvenile delinquency among the children. The relationships between the quality and intensity of welfare intervention, on one hand, and children’s adjustment, on the other, may be various when viewed from the standpoint of social capital. These interventions may help weave a network of relationships that goes beyond the nucleus of the family and facilitates access to different types of resources (e.g., educational and recreational). To some extent, proper welfare intervention should make up for the deficiencies in the family’s social capital, even though, as mentioned earlier, the effects of the welfare system are not completely clear and are still debatable.
FAMILY SOCIAL CAPITAL How social capital should be defined at the family level is controversial. On one hand, some, such as Wright, Cullen, and Miller (2001), claim that a very broad conception should be adopted—one that includes any type of relationship involving the child, both within and through the family, and which therefore takes into account the time and effort that parents devote to their children, the intense and lasting affective links that are forged, and the prosocial directives that are proposed. On the other hand, according to Sampson, Morenoff, and Earls (1999), many researchers adopt an excessively broad conception of social capital, including aspects of life within the family, such as parental monitoring and expectations, which should not be included; in this perspective, the notion of social capital should concern links with the local community. In this review we adopt an intermediate stance—one that considers those aspects of family life that imply a greater or lesser link with the com-
Social Capital and Physical Violence
403
munity. This is in line with the suggestions made by Coleman (1990), who stressed that some characteristics of the family, in particular single-parent status and the presence of a working mother, may weaken links with the community. In regard to the single-parent family, several studies revealed a correlation between this condition and antisocial behavior on the part of children (see, e.g., Rutter, Giller, & Hagell, 1998). Vaden-Kiernan, Ialongo, Pearson, and Kellam (1995) investigated the relationships between the family structures and teacher-rated aggression of children, within the framework of a longitudinal study of 1,197 children from 19 public elementary schools in Baltimore, Maryland. The relationships between fourth-grade family structure and sixthgrade aggression were examined separately for boys and girls, using logistic regression analyses comparing mother-alone families with the other families, and controlling for income level, urban area, and fourth-grade aggression. The boys from mother-alone families were more likely to be rated as more aggressive by their teachers, whereas this was not the case for the girls. Pagani, Boulerice, and Tremblay (1997) conducted a similar study with a random sample of the population of kindergarten children in the Canadian province of Québec (947 girls and 882 boys). They used logistic regressions to predict age12 fighting from family status, controlling for age-6 fighting and other family characteristics (education, age, poverty). A short-term increase in fighting for boys and girls could be attributed to family transitions in both directions, that is, divorce (becoming a single parent) and remarriage (ending single-parent status). With a large sample (n = 1,037) of kindergarten boys from low socioeconomic areas of Montreal, Nagin and Tremblay (2001) studied the predictive value of family break-up before school entry for trajectories of physical aggression from 6 to 15 years of age. After controlling for parent characteristics (age at birth of child, education, employment status) and child characteristics (IQ and behavior problems), they showed that family breakup before school entry increased the odds of being on a high physical aggression trajectory by close to 50%. In the Dunedin study, the fact that a 13-year-old was the child of a single parent predicted the likelihood of convictions for violent crimes at 18 years (Henry, Caspi, Moffitt, & Silva, 2004). In London, Farrington (1989) found that parent–child separation before the age of 10 years predicted self-reported violence in adolescence and adulthood, as well as the likelihood of convictions for violent crimes. In a more recent study, Tremblay et al. (2004) found that children born in Canada in 1997–1998 to single mothers were more at risk of being on a high-level physical aggression trajectory between 17 and 42 months after birth than other children. These results were obtained after controlling for a large number of parent and family characteristics. Another aspect of family life that has long been investigated concerns the relationship between the delinquency of parents and that of their children. This relationship has been variously interpreted in terms of genetic predisposition, psychological influence, or environmental factors. From the
404
DETERMINANTS OF AGGRESSION
social capital perspective, it may also be claimed that parental delinquency reduces family social capital, in that involvement in illicit activities might not only restrict children’s access to socially valid interpersonal and community relationships, but also facilitate contact with groups involved in crime. Several longitudinal studies have revealed the link between parental delinquency and violence on the part of children (Smith & Farrington, 2003), as recorded through self-reported surveys and official records of violent offenses committed either in adolescence or adulthood. Particularly significant in this field is the research conducted by Keenan and Shaw (1994), who measured the frequency and pervasiveness of aggression from videotapes of the behavior of 98 mother–child couples when the children (males and females) were 18 and 24 months old. This is one of the very few studies that have examined physical aggression at a very early age, and measured it not only through parental perception but also through direct observation. These authors found that parental delinquency was one of the strongest predictors of aggression for boys at 24 months (but not at 18 months). They therefore claimed that the link between familial delinquency and boys’ aggression was stronger when the children had reached a developmental stage at which the parents began to discipline their children. Using the U.S. National Youth Study, Wright et al. (2001) evaluated family social capital, using a composite index that included the time parents spent with their children, the strength of the bonds between various family members, and the parents’ critical attitude toward delinquency. They found an inverse relationship between social capital and both short- and long-term involvement in delinquency, evaluated through answers given by the youths about their involvement in a range of crimes, such as theft, assault, robbery, and drug dealing. McCord (2002) obtained similar results with data from the Cambridge–Somerville study. Social capital was measured by social workers’ observations of 232 families with children between 10 and 16 years of age at the start of the research. Outcome data, collected more than 40 years after the beginning of the study, indicated that the social capital of parents reduced the risk of their children’s being convicted of violent crimes. According to Coleman (1990), maternal employment also has an effect on social capital. He suggests that if a mother works, she will have less chance to participate in the life of the community, to develop relationships with neighbors, and to share educational problems, and that this will have a negative influence on children’s adjustment. This aspect of family social capital assumes particular relevance in modern societies, in which women’s employment outside the home has become the norm. Some have expressed concern regarding the well-being of children on the grounds that their development and adjustment may be hindered if their mothers work (Belsky, 2001). Vander Ven, Cullen, Carrozza, and Wright (2001) used data from the National Longitudinal Survey of Youth and found that the features of the mothers’ jobs had relatively little or no direct influence on the delinquency of their children, but did have a slight indirect effect through supervision. Lack of supervision had
Social Capital and Physical Violence
405
an effect regardless of whether early or current employment was considered. It was found that the effects of maternal employment might differ according to the population considered. For instance, Han, Waldfogel, and Brooks-Gunn (2001) found that, in the general population, maternal employment in the first year of the child’s life had an effect that could be perceived in the sphere of externalizing problems when the child was 7–8 years old, but among African American children this effect did not emerge. However, research on the effects of maternal employment must include research on the impact of the quality of care given to young children while their parents are working. We turn next to this topic.
DAYCARE AND EARLY CHILDHOOD DEVELOPMENT CENTERS In the last few decades, a heated debate has raged in regard to children’s daycare, to the extent that the term daycare war has been coined to describe the widespread and fierce clash of views on such institutions (Karen, 1994). Some argue that early child care services are an important social resource that mothers can exploit to improve their children’s socialization. Others argue that daily separation from the mother during the first few years of the child’s life may put the child’s development at risk. We obviously need good data to decide which resources constitute a sound capital for children’s development. But few good studies have been carried out to date, and fewer than a handful deal with the development of aggression. There appears to be a consensus that the effects will depend on the quality of the daycare environment rather than on the presence or absence of daycare (Tremblay, Barr, & Peters, 2003). Some early studies on children’s disruptive behavior concluded that those who spent several hours of the day away from their mothers during the first year of life were more disobedient and more aggressive toward their peers between 3 and 8 years of age (Haskins, 1985; Rubenstein & Howes, 1983). These studies were criticized because they often failed to take into account the quality of the daycare environment and selection effects, that is, family and child characteristics before the beginning of daycare. For example, a study of a representative sample of 3,431 Canadian children 2–3 years of age (Borge, Rutter, Côté, & Tremblay, 2004) found a higher level of aggression among children who were not sent to daycare groups, but were looked after by their parents. This relationship emerged only among children in high-risk families; among those in low-risk families, who constituted the majority (87%), no difference in the level of aggression was recorded. Moreover, within the subgroup of children attending group daycare, placement instability was associated with a higher level of aggression after controlling for a family risk index (occupational level, maternal education, number of siblings, and family functioning). The U.S. National Institute of Child Health and Human Development created the NICHD Early Child Care Research Network (2002, 2003) to tackle the shortcomings of previous studies. It regularly assessed 1,300 chil-
406
DETERMINANTS OF AGGRESSION
dren and their families starting with the children at 6 months of age. Children who spent more time in child care during the first years of life displayed higher levels of externalizing problems and problems with aggression at 24 months and 54 months, as well as in kindergarten and first grade. These results certainly fueled the public debate. However, we will need much more information before the debate is settled. First, the differences were statistically significant, but more time in daycare did not increase the risk of being considered in the clinical range; second, developmental trajectories of aggression were not studied to determine which type of children are more likely to be affected. Are the affected children those who tend to be highly aggressive? For example, Raine, Reynolds, Venables, Mednick, and Farrington (1998) reported that 3-year-olds who had a larger body size were more likely to use aggression at 11 years of age. One can imagine that being with peers for longer periods of time, and dominating them, makes it less likely that a child will learn not to use physical aggression. Alternately, could the affected children be those who normally tend to learn rapidly not to use physical aggression? These children may need to keep their level of aggression high enough to survive in a tough peer environment. Developmental trajectories during the preschool years, coupled with data from the elementary school years, should help in understanding who is affected and whether these effects are long lasting. A third caveat is that we are looking at a longitudinal study. We will need experiments to determine which types of preschool environments help children learn alternatives to aggression. Experimental designs are indeed needed to answer questions concerning cause and effect. A number of randomized experiments have attempted to test the effect of early preventive interventions, carried out in the preschool years. The Perry preschool project, also known as the High Scope project, is probably the best known of these experiments (Weikart, 1998; Weikart & Schweinhart, 1992). The main aim of this program was to prevent school failure among poor African American children by intervening when they were 3 and 4 years old. Children attended quality day centers and teachers visited mothers at home. The long-term results revealed not only that the treated subjects were more successful at school than the control subjects, but also that they were arrested less often both as adolescents and as adults (at ages 19 and 24 the treated subjects were five times less likely to have been arrested). They committed less serious crimes and had lower rates of recidivism. Significant reductions in antisocial behavior were also found in the Syracuse University Family Development Project, and a reduction in aggressive behavior was observed in both the Yale Child Welfare Project and the Houston Parent Child Development Center (for a review of the results, see Tremblay & Craig, 1995). Administered during the preschool years, all these programs increase social capital, in that they improve relationships between children and parents, and between parents and the community. Unfortunately, none of these studies appear to have specifically studied their subjects’ development of aggression.
Social Capital and Physical Violence
407
SCHOOL SOCIAL CAPITAL Hagan and McCarthy (1997) have stressed that school is an important component of the social institutions network within which social capital is generated. Indeed, in the light of the social capital paradigm, school is seen as a context in which relationships that are important for the pupil’s adjustment are developed—relationships that also involve teachers and parents. A ready indication of the relationships between school social capital and pupil adjustment can be deduced from those studies that have correlated the size of the school with the pupils’ behavior. Some studies revealed that students in larger schools experience a greater sense of alienation and frustration than those in small schools, and that the smaller schools are able to provide their students with a better social environment. Moreover, it has been shown that large schools report a far higher number of serious crimes and behavior problems (Ferris & West, 2002; National Center for Education Statistics, 2002) than small schools; it may be supposed that the social relationships within the smaller institutions are more closely knit, that the people know each other better and are more inclined to lend a hand in solving one another’s problems. Social capital at the school level particularly concerns the involvement of parents in school activities. By using different samples of adolescents (in residential treatment, ambulatory treatment, and untreated) Buysse (1997) found that a lack of social capital in the family and at school engendered a high risk of antisocial behavior on the part of the youth. Parcel and Dufur (2001) also analyzed the influence of capital at home and at school and considered both human and social capital. Using the National Longitudinal Survey of Youth (NLSY), to which they added indicators in the children’s schools (such as a measure of school-wide parental involvement in school activities and scales that reflect school environment), they observed that family social capital and child human capital had a large impact; but effects of school social capital could also be detected, and interactions between family and school social capital were important as well. In studying the relationships between school environment and the antisocial behavior of youth, Rutter, Maughan, Mortimore, Ouston, and Smitj (1979) identified a set of characteristics that apparently prevented deviant behavior on the part of pupils. The school environment thus created was one in which teachers provided an example of good conduct, monitored and stimulated their pupils, and implemented firm and consistent disciplinary measures. More recently, Leblanc, Swisher, Vitaro, and Tremblay (2004) measured the social climate (discipline and academic pressure, in regard to teacher–pupil relationships, and autonomy and job satisfaction, concerning relationships within the teaching staff) of 217 high schools attended by 1,233 pupils who had been followed since kindergarten. Students attending high schools where teachers perceived fewer classroom behavior problems reported less physically aggressive antisocial behaviors. The analyses controlled for the students’ de-
408
DETERMINANTS OF AGGRESSION
velopmental trajectory of aggression during elementary school, as well as for the adverse family environment of the pupil and the sociodemographic factors of the school, including school size. Prevention experiments are, of course, better tools than correlational studies to test the causal effects of school social capital. Four experiments— the Seattle Social Development Project (Hawkins, Catalano, Kosterman, Abbott, & Hill, 1999), the Baltimore Prevention Study (Kellam, Rebok, Ialongo, & Mayer, 1994), the Montreal Longitudinal-Experimental Study (Tremblay, Kurtz, Mâsse, Vitaro, & Pihl, 1995), and the Fast Track prevention trial (Bierman et al., 2002)—measured the development of antisocial behavior after attempting to improve relationships among pupils, parents and teachers, and the school environment in general. These four prevention experiments started in the first 2 years of elementary school and lasted between 2 and 6 years. Both the Baltimore and Montreal experiments lasted 2 years and showed a significant long-term reduction (4–8 years after the intervention) of aggression in boys with the highest levels of physical aggression when they started school (Kellam et al., 1994; Lacourse, Nagin, Tremblay, Vitaro, & Claes, 2003). The Seattle experiment also showed a significant reduction of violent delinquency during adolescence for boys and girls who participated in the program from first to sixth grade, but not for those who participated only during fifth and sixth grade. Fast Track is the latest of these four preventive trials. Four years after the start of the intervention, the authors reported that children in the control group, as compared with those in the intervention group, displayed less aggressive behavior, according to teacher and parent ratings (Bierman et al., 2002; Conduct Problems Prevention Research Group, 2002). The results from these four experimental programs suggest that trajectories of physical aggression can be deflected by interventions that improve relationships between the pupil, his or her parents, peers, and teachers. They indicate, however, that the younger the children when the intervention starts, the more impact it will probably have.
PEER SOCIAL CAPITAL Well-socialized peer groups are an important aspect of social capital—a resource that is available to children and may influence socialization in a positive manner. As the child grows up, the peer group apparently becomes increasingly important, while the impact of the family environment declines. In reference to the social capital paradigm, and therefore to resources in terms of social relationships that can be utilized, we concentrate here on the effect exerted by the presence of prosocial peers, or the lack of this resource resulting from the phenomenon of peer rejection. Various studies have reported that nondeviant friends are able to reduce involvement in antisocial behavior (Brown, Lohr, & McClenahan, 1986) and that a prosocial group seems to lighten the impact of other risk factors
Social Capital and Physical Violence
409
(Quinton, Pickles, Maughan, & Rutter, 1993). It has also been observed that peer disapproval of delinquency reduces the likelihood that the individual will subsequently commit violent crimes in general (Elliott, 1994), and sexual crimes in particular (Ageton, 1983). With a sample of 868 boys selected from kindergarten classrooms in poor areas of Montreal, Vitaro, Brendgen, and Tremblay (2002) observed that, by early adolescence, those who were most influenced by the aggressive–disruptive characteristics of their friends were those who already had moderate behavior problems. These problems tended to become milder if their friends were not aggressive–disruptive, but if their friends were aggressive–disruptive, then their own problems increased. It is interesting to note that boys who had serious problems, and boys who had no problems, were not influenced by the characteristics of their friends. In a similar study of 585 families in which there was a 5-year-old boy or girl, Criss, Pettit, Bates, Dodge, and Lapp (2002) found that peer acceptance and friendship acted as a moderator between the degree of family adversity and the child’s externalizing behavioral problems. In contrast, peer rejection can cause a child to miss out on a fundamental experience in the socialization process (Vitaro, Tremblay, & Bukowski, 2001b) and can facilitate the emergence of adjustment problems and deviant behavior. In a review of the literature on this subject, Parker and Asher (1987) concluded that there was a relationship between peer rejection and subsequent problems of social adjustment and delinquency. However, they also raised the question of whether this rejection exerted an independent effect on the subsequent behavioral disorders, or whether peer rejection was linked to the subject’s aggressive behavior from the outset, and the subsequent disorders were chiefly linked to this early aggressiveness. Subsequent investigations enabled this issue to be tackled and yielded some preliminary answers. On the basis of behavior observation, teacher and peer ratings, and peer interviews, Bierman, Smoot, and Aumiller (1993) subdivided a sample of 95 children examined at three successive times (when they were 6–8, 8–10, and 10–12 years old) into four subgroups: aggressive and rejected, aggressive and nonrejected, rejected and nonaggressive, and neither aggressive nor rejected. Aggressive-rejected children exhibited more diverse and severe conduct problems than aggressive children, in that the former displayed high levels of aggressiveness and disruptive behavior and a low level of prosocial activity, whereas the latter displayed only a higher level of aggressive behavior. Moreover, 2 years later, the aggressive-rejected children continued to manifest a higher degree of aggressive and disruptive behavior and a greater deficiency in prosocial activities than the children initially classified as aggressive-nonrejected. In a longitudinal study, which involved subjects from third grade to adolescence, Coie, Lochman, Terry, and Hyman (1992) observed that both aggressiveness and rejection, as evaluated in the third grade, independently predicted externalizing problems in the sixth grade. The combination of aggressiveness and rejection predicted (only in males) acts of criminal assault and, in general, forms of antisocial behavior that were more serious than those predicted by aggressiveness alone.
410
DETERMINANTS OF AGGRESSION
Reactive aggression appears to be associated with peer rejection, whereas active aggression is enacted by children who are accepted by their peers (Dodge, Lochman, Harnish, Bates, & Pettit, 1997). The effects of peer relationships on aggressive behavior have been also described in a relatively new field of research, concerning the consequences of mutual antipathies (i.e., relationships characterized by reciprocal disliking) in childhood and adolescence. In a study on 2,348 school-age children and 2,768 adolescents, Abecassis et al. (2002) found that, controlling for peer rejection, same-sex antipathies were associated with aggressive behavior (perceived by peers) and bullying others (self-reported), both in childhood and in adolescence, and to emotionality and lack of friendship support among adolescents; however, mixed-sex antipathies were related to antisocial and bullying behavior in boys but not in girls. Finally, it should be noted that the influence of the peer group in reducing and controlling aggressive behavior can also be evaluated through experimental interventions. Vitaro, Brendgen, Pagani, Tremblay, and McDuff (1999), for instance, found that association with less deviant peers partly mediated the effect on conduct disorders, assessed at the age of 13 years, of a multicomponent prevention program that targeted disruptive boys when they were ages 7–9. In a subsequent study with the same sample, Vitaro, Brendgen, and Tremblay (2001a) showed that reduced disruptiveness and increased parental supervision by age 11, as well as association with nondeviant peers by age 12, were part of a chain of events that mediated the effect of the program on the initial level of delinquency at 13 years. Results of this kind have therefore confirmed the notion that relationships with prosocial companions provide a form of social capital that exerts a positive influence on the socialization process and on the progressive control over disruptive behavior, including aggression.
NEIGHBORHOOD SOCIAL CAPITAL At the beginning of the 1980s, Sampson and Groves (1989) used data from 238 localities in Great Britain with a view to ascertaining whether some parameters of community life, drawn from a national survey of 10,905 residents, were able to predict crime rates. They demonstrated that some features of the community, which today would be identified as social capital (friendship networks, participation in community activities, and supervision and control of adolescent groups), were associated with lower crime rates, as measured both through self-reports and through victimization surveys. Sampson, Raudenbush, and Earls (1997) subsequently utilized the concept of “collective efficacy” to interpret differences in the occurrence of violence in Chicago neighborhoods. Collective efficacy refers to social cohesion in a neighborhood, combined with the willingness of residents to act in favor of common interests. It was measured by a scale of social control and a scale of social cohesion and trust. Controlling for race, age, and homicide rates in the previous years,
Social Capital and Physical Violence
411
collective efficacy at the neighborhood level correlated inversely with homicide, violence as perceived by citizens, and violence assessed through victimization surveys. Using the same indexes employed by Sampson et al. (1997), Browning (2002) found that collective efficacy was negatively associated with both intimate homicide rates and nonlethal partner violence and that, furthermore, in neighborhoods displaying higher levels of collective efficacy, women were more likely to disclose relationship conflicts to potential sources of support. In order to understand the origins of collective efficacy, Sampson et al. (1999) conducted a subsequent study in which they tried to identify, again at the local community level, three aspects of social organization that influence children’s lives, development, and adjustment: intergenerational closure for children (when adults know the parents of their children’s friends, watch out for the safety of children, etc.), reciprocal exchanges (the degree of interaction among families regarding their children’s education and the swapping of advice, material possessions, and information on childcare), and the degree of informal control combined with reciprocal support (expectations of action within a community). They used the data from investigations conducted in 1995 on a sample of 8,782 residents in 342 different Chicago neighborhoods. Residential stability and wealth were good predictors of the bond between generations and of reciprocal exchanges. However, concentrated disadvantage correlated strongly with low expectations of shared social control over children. Spatial aspects proved to be very important, in that proximity to an area with strong bonds between generations, frequent reciprocal exchanges, and good informal control had a favorable impact above and beyond the structural features of the neighborhood. Sampson et al. (1999) introduced these indicators on the grounds that some aspects of social capital, such as the existence of groups, associations, and networks, might not, on their own, be sufficient to favor child development and might even have a negative effect on social life if they operated in such a way as to exclude others. They therefore claimed that the objectives of local associations should be examined before deciding whether such groupings favor children’s collective needs. Bellair (1997) used an interesting and innovative research method to investigate the effect of various types of social interaction on crime across 60 urban neighborhoods. He found that getting together once a year or more with neighbors had a strong negative association with burglary, motor vehicle theft, and robbery and that this form of interaction mediated a significant proportion of the effect of ecological characteristics on crime. Bellair suggested that even infrequent interaction and knowing one another could prompt neighbors to engage in supervision and guardianship. This infrequent interaction corresponds to what Granovetter (1973) defined as “weak ties,” which can strengthen the community by creating linkages across networks. The preventive effect of social bonds may not, however, be homogeneous. For instance, Warner and Rountree (1997) showed that strong social ties exerted a preventive effect in regard to assault rates in predominantly white neighbor-
412
DETERMINANTS OF AGGRESSION
hoods, whereas in predominantly minority and mixed-race neighborhoods such ties were largely ineffective. A particular form of violent behavior, child maltreatment, has been seen to be associated, at the neighborhood level, with the lack of social resources, when income levels are equal (Vinson, Baldry, & Hargreaves, 1996). Dubet (1987), through ethnographic studies, identified another relationship between violent behavior and scant social bonds. He ascribed the outbreaks of youth violence seen in recent years in the suburbs of some French cities to the disappearance of the traditional working-class culture. In workingclass neighborhoods there used to be a whole range of meeting places and spare-time activities and groups, which provided the young with role models and common objectives. In traditional working-class neighborhoods the factory provided a nucleus for aggregation and socialization; children and adolescents therefore grew up within a network of relationships and expectations that fostered social and occupational development and provided models that were endorsed by the whole community. Although this setting also featured social conflicts, the community was somewhat close-knit and rich in social relationships. De-industrialization and suburban transformation have demolished this social fabric; commerce is rare, meeting places nonexistent, buildings dilapidated, and transport inadequate. What has emerged is a sort of no-man’s-land, where the young try to survive, drug pushing abounds, crime and violence are endemic, and a deep sense of insecurity prevails. In the light of the social capital theories, what has disappeared from these neighborhoods is a set of social networks, shared expectations, and community commitment. These are so essential to the social life and socialization of the young that their disappearance has facilitated the outbreaks of violence that have afflicted many urban districts in Europe.
TOWN AND REGION The relationship between regional social capital and violent crimes has recently been examined in a series of studies in the United States. Using data from the 50 U.S. states, Kennedy, Kawachi, Prothrow-Stith, Lochne, and Gupta (1998) measured social capital by membership of groups and associations and social trust, two items from the U.S. General Social Survey. They found a correlation between a low level of social capital and marked inequality on one hand, and violent crimes on the other. In a cross-sectional analysis that also included the 50 U.S. states, Putnam (2000) found that an index of social capital was inversely correlated with homicide rates and that this correlation was very high (r = –.80) for the period 1980–1995. Controlling for the variable “fear of crime,” the (partial) correlation between social capital and homicide remained high (r = –.53). Putnam concluded that the direction of the relationship could not possibly be from homicide to social capital, because any such relationship would have been mediated by the fear of crime. To ob-
Social Capital and Physical Violence
413
tain a clearer picture of the relationship that had emerged, he introduced a range of social and economic factors as independent variables in a multiple regression. He found that the model best able to predict homicide rates in the various states included four variables: social capital, mean poverty level, ethnic makeup of the population (percentage of nonwhites), and the proportion of the population living in cities. In a similar study, Hemenway, Kennedy, Kawachi, and Putnam (2001) measured, at the state level, the degree of interpersonal trust (through the replies provided to the U.S. General Social Survey) and social commitment (assessed through participation in voluntary activities, religious observance, etc.). After controlling for the level of urbanization, poverty, and the percentage of nonwhite residents, they concluded that these indicators of social capital were inversely associated with the percentages of homicides and suicides committed with firearms. Focusing their analysis on a more restricted territorial area, and thus increasing the sample size, Rosenfeld, Messner, and Baumer (2001) investigated the relationships between social capital and homicide in 99 geographical areas of the United States. They measured two aspects of social capital, civic commitment and social trust: The former was measured through voter turnout and membership in voluntary associations, and the latter through respondents’ replies to the U.S. General Social Surveys of 1993, 1994, and 1996. A low level of social capital was found to be associated with a high homicide rate, after controlling for a range of socioeconomic variables. Social capital mediated the effect of the area’s being located in the southern states, but it did not mediate the effect of economic deprivation. The statistical relationship between social capital and homicide was not due to the influence of homicide on social capital. Subsequently, Messner, Baumer, and Rosenfeld (2002) broadened this focus to include robbery and assault, demonstrating that these crimes were also inversely related to social capital. In one of the rare studies on social capital and violence outside the United States, we used Italian data and calculated an index of “civicness” (Putnam, 1993) for each of the 20 administrative regions (Gatti, Tremblay, & Larocque, 2003) and 95 provinces (Gatti, Tremblay, & Schadee, 2002b). The civicness index included three indicators: (1) voter turnouts in referendums or elections, (2) the percentage of citizens over the age of 13 years who read a newspaper every day, and (3) the number of recreational, sporting, and cultural associations per 100,000 inhabitants. The results revealed a negative correlation between civicness and the reported number of some types of serious crimes of violence, as recorded in the 1970s, 1980s, and 1990s, at the regional level. Data from the vast victimization survey conducted by the Istituto Nazionale di Statistica/National Institute of Statistics (ISTAT) confirmed the negative association between civic sense and violent crimes. At the provincial level, multiple regression analyses were carried out, using as control variables unemployment, family breakdown, and urbanization as predictors of homicide, in addition to civic sense. Notwithstanding the introduction of these control variables, civic sense continued to be negatively associated with homicide;
414
DETERMINANTS OF AGGRESSION
however, the interactions among variables revealed that the protective effect of civic sense was exerted only in the southern provinces, where homicide rates are higher. The criminogenic effect of urbanization and unemployment was also observed only in the southern provinces, whereas family breakdown was associated with homicide in the whole of Italy. Finally, the preventive effect of civicness in regard to robbery was examined. This analysis revealed that civic sense has a preventive effect on this type of crime, but only in the more urbanized provinces, where robbery rates are higher.
NEGATIVE SOCIAL CAPITAL Although the majority of authors stress the positive aspects of social capital, some also underline its possible negative consequences (e.g., Portes, 1998). Indeed, social networks and the norms of reciprocal cooperation may be used to exclude those who do not belong to the group from access to resources (see Vaillancourt, Chapter 8, this volume). Community bonds may constitute a pervasive social control tool, which hampers autonomy and privacy, and members of marginal minority groups may be tied down by bonds of solidarity that make it difficult for them to accept the dominant social norms and to gain access to the world of traditional work. Finally, strong social bonds may be used for patently illegal purposes within the sphere of organized crime or youth gangs (Portes, 1998). In regard to children’s development, the fact that youth gangs may constitute an environment in which aggressive behavior is facilitated has been well documented, especially for males. Numerous studies have revealed that youths who belong to a gang commit more crimes than those who do not, and that this difference is greater with respect to serious crimes and crimes of violence. Longitudinal studies are especially suited to test three models, which Thornberry, Krohn, Lizotte, Smith, and Tobin (2002) have defined as the selection model, the socialization model, and the enhancement, or mixed, model. According to the selection model, the high level of delinquency and violence among gang members does not depend on the influence of the gang, but rather is due to the fact that gangs attract and select those youths who already commit the most crimes. In contrast, the social facilitation model holds that it is the gang that facilitates and triggers deviant behavior in youths who, before joining the gang, are no different from other young people. The enhancement model combines these two views: Youths who join gangs already display a higher level of delinquency, and joining the gang exacerbates their deviant behavior. Using data from the Rochester Youth Development Study, Thornberry, Krohn, Lizotte, and Chard-Wierschem (1993) showed that, before joining a gang, members did not display higher rates of delinquency or drug use than nonmembers. However, such behaviors increased upon entry into the gang, and delinquency declined once an individual left the gang (Thornberry et al.,
Social Capital and Physical Violence
415
1993). In a subsequent study, covering a longer period of time, Thornberry et al. (2002) found that, in addition to a strong facilitation effect, a weaker selection effect was also present. Strong evidence for the social facilitation model and partial support for selection was provided by Zhang, Welte, and Wieczorek (1999) through their analysis of the data from the Buffalo Longitudinal Survey of Young Men. The enhancement model was confirmed by the Denver Youth Survey, a longitudinal study of families in high-risk neighborhoods (Esbensen & Huizinga, 1993), by a longitudinal study of schools in six American cities (Esbensen, Peterson, Freng, & Taylor, 2002), and by the Montreal Longitudinal Experimental Study (Lacourse et al., 2003; Vitaro, Tremblay, Kerr, Pagani, & Bukowski, 1997). The latter study also revealed that the criminogenic influence of the gang was independent of other risk factors, such as previous delinquency, disruptiveness, and lack of parental supervision; gang membership emerged as a significant predictor of the frequency of aggressive behavior, above and beyond the effects of having delinquent friends (Gatti, Tremblay, Vitaro, & McDuff, in press). Data from the Seattle Social Development Project revealed that respondents who were gang members always had the highest frequency of delinquent behavior and drug use, and that comparing gang members, nongang youths who had delinquent friends, and nongang youths who did not have delinquent friends, gang membership predicted selfreported and officially recorded delinquency beyond the effects of having delinquent friends and prior delinquency (Battin-Pearson, Thornberry, Hawkins, & Krohn, 1998). The results from these studies demonstrate that the criminogenic effect is not due simply to contact with other delinquent youths, but that it is also linked to the influence of the gang itself. Some ethnographic studies have revealed that the gang may constitute an important form of social capital for youths—a resource that they can exploit in order to achieve their own objectives. Jankowsky (1991) obtained participant observations over a 10-year period among 37 gangs in the metropolitan areas of Los Angeles, New York, and Boston. He concluded that in the American suburbs, gangs are not only an expression of social disorganization, but they are also an organizational response aimed at enhancing the competitiveness of their members in acquiring scarce resources. Indeed, gang members enjoy certain advantages in terms of money, status, and power, and the gang accepts youths as members if they are able to contribute to its prestige and efficiency. Criminal organizations and subcultures may also be regarded as a form of social capital that facilitates violence. In particular, violent subcultures, which were first investigated by Wolfgang and Ferracuti (1967) and subsequently analyzed by numerous criminologists, may, like the army or a boxing coach, train the individual in using physical aggression. According to Wolfgang and Ferracuti, a subculture is “a normative system of some sub-group smaller than the whole society,” which implies a process of learning shared values. There are subcultures that value the use of physical violence, and certain members (e.g., adult males) are required to use violence in certain circumstances. In such situations, the individual is expected to act violently, and
416
DETERMINANTS OF AGGRESSION
nonviolence may lead to ostracism or condemnation; in contrast, violence is rewarded rather than condemned and does not arouse guilt feelings in the individual. There are a few published biographies that shed light on the relationships between criminal organizations, subcultures, and aggressive behavior. A biography of Raffaele Cutolo (Marrazzo, 1984), a Neapolitan crime boss who reigned over the Campania region for many years, and who is now in prison, provides us with the following illuminating episode. Raffaele’s father was an agricultural worker, who for years tilled a field as a sharecropper to support his family. One day, when Raffaele was still a child, the owner of the land told his father that the following year the field would be used for a different purpose, and that his services would therefore no longer be required. In desperation, Raffaele’s father turned to the local Mafia boss, whose word was law in the village. The Mafia boss invited the Cutolo family to his home, and after Mass one Sunday, dressed in their best clothes, both parents and children respectfully presented themselves at the boss’s door. Raffaele’s father was ushered into another room by the boss, while the lady of the house offered cakes to Mrs. Cutolo and her children. When the two men reemerged, Mrs. Cutolo anxiously asked her husband if there was any news. She was told that everything had been settled. The following day, the landowner informed Mr. Cutolo that he had changed his mind and that his contract would be renewed. It is easy to imagine the psychological impact that this event must have had on all the members of the family, and especially on the young Raffaele, who had seen this family crisis solved by appealing to a local institution that was both powerful and efficient, although illegal. A few years later, however, when Raffaele was an adolescent, he was summoned by the Mafia boss, who instructed him to deliver a pistol secretly to a mafioso in another village. Other requests followed, and finally Raffaele Cutolo found himself involved in serious crimes. From there, he went on to become the head of a powerful organized crime syndicate. It is clear from this account that the criminal fraternity constituted a form of social capital—a resource to be exploited in order to solve problems and achieve otherwise unreachable goals. But this social capital was also permeated by norms of reciprocity and expectations centered on the use of violence, which ended up conditioning the social development of those living within this culture. Some systematic studies have also evaluated the importance of violent subcultures. Felson, Liska, South, and McNulty (1994), for example, studied 2,123 boys in 87 public high schools. Respondents were asked a series of questions designed to measure whether they approved of nonaggressive responses to some types of provocation. The questions consisted of a series of personal values, and respondents were asked whether certain acts were “good thing[s] for people to do.” Disagreement with these values was deemed to indicate a high score on the subculture of violence scale. The authors reached the conclusion that group-level values predicted violent behavior, regardless of individual-level values. Furthermore, this research demonstrated that subcul-
Social Capital and Physical Violence
417
ture differences in values were more frequently encountered within small groups involving interactions between their members, rather than within broad aggregations based on race, social class, or religion. Such examples illustrate the fact that social capital may also exist in illegal forms and that, in some cases, crime and violence constitute a shared resource that can be exploited in order to achieve ambitions or, in certain circumstances, simply to survive.
CONCLUSIONS We have seen that social capital can be defined in various ways and may operate in more or less broad contexts and at different ages from early childhood to adulthood. From the developmental standpoint, several empirical studies have demonstrated that both familial and extrafamilial social capital have an effect, even in the long term, on children’s adjustment, on academic results and socioeconomic success (Furstenberg & Hughes, 1995), and on health and well-being (Morrow, 1999). In general, the social capital perspective has proven useful in helping us to understand aggressive behavior and violence, though much more research is needed. Not all forms of social capital act in the same way, and several studies have found that social capital seems to exert a preventive effect on aggressive behavior only in certain circumstances and only in regard to some types of persons. We have seen that the development of children and young people can be influenced by the quality of relationships within the family, at school, and among peers, and by the characteristics of the communities and regional environments in which they live. For the most part, each of these levels of influence has been studied separately, though some researchers have begun to consider different aspects of social capital simultaneously. This appears to be an important advancement, in view of the multidimensional nature of social capital and its potential to link “the overly narrow purview of psychology and the overly broad purview of sociology” (Furstenberg and Hughes, 1995). In addition to its multidimensional character, the developmental aspect of social capital should also be considered, by examining how the loss or acquisition of a given form of social capital during the course of the individual’s life might influence engagement in aggressive behavior. The various types of social capital have a differential impact according to the age of the individual. Family social capital and the relationship resources offered by the child welfare system chiefly act during early childhood to middle childhood, when learning to regulate physical aggression appears to be at its peak. The influence of peers has been shown mainly during late childhood and adolescence; however, it is most likely present very early at home, through siblings, and in daycare. Violent subcultures and organized crime are more likely to exert a direct effect on adolescents and adults. Broadly speaking, it seems that social capital at the
418
DETERMINANTS OF AGGRESSION
microlevel plays a stronger role during childhood, whereas the macrolevel acts especially during adolescence and adult life. Thus, the influence of social capital on the regulation of aggressive behavior should be studied from a developmental perspective. It can be claimed that the process of progressively waning physical aggression from early childhood to adulthood (Sampson & Laub, 2003; Tremblay & Nagin, Chapter 5, this volume) is facilitated by the majority of the social relationships we have defined as social capital. All cultures, including nonhuman cultures, benefit from helping their young ones to control their ability to physically aggress others. This does not mean that they completely inhibit the power to physically aggress. Armies are, or should, always be ready for war. Individuals and subgroups are at an advantage if they can defend themselves from an aggression. Thus, social capital that is advantageous for one group can be disadvantageous for another. This is why gangs and organized crime are considered threats to the mainstream, but constitute a resource for individuals who, for different reasons, are not or have chosen not to be part of this mainstream. In general, attempts should be made to improve social support (Colvin, Cullen, & Vander Ven, 2002) at all levels—through the family, welfare, and academic institutions, formal and informal groups, neighborhoods and cities— to strengthen social networks, to increase community participation and social commitment, and to reduce individualism in favor of a community spirit. These objectives should be pursued both through specific intervention, for example, by providing support for mothers in difficulty, and through a broader cultural evolution toward greater cooperation and solidarity. Research on the effects of these interventions should help us to understand which are the most effective strategies for a given culture, while testing our theories concerning the origins and development of aggressive behavior. We will, of course, need to study to what extent different types of social capital have different effects on different types of aggressive behavior, at different periods of the life course.
REFERENCES Abecassis, M., Hartup, W. W., Haselager, G. J. T., Ron, H. J., Scholte, R. H. J., & Van Lieshout, C. F. M. (2002). Mutual antipathies and their significance in middle childhood and adolescence. Child Development, 73(5), 1543–1556. Ageton, S. S. (1983). Sexual assault among adolescents. Lexington, MA: Lexington Books. Battin-Pearson, S. R., Thornberry, T. P., Hawkins, J. D., & Krohn, M. D. (1998). Gang membership, delinquent peers, and delinquent behavior. Bulletin No. NCJ 171119. Youth Gang Series. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention. Bellair, P. E. (1997). Social interaction and community crime: Examining the importance of neighbor networks. Criminology, 35(4), 677–703. Belsky, J. (2001). Developmental risks (still) associated with early child care. Journal of Child Psychology and Psychiatry, 42, 845–860.
Social Capital and Physical Violence
419
Bierman, K. L., Coie, J. D., Dodge, K. A., Greenberg, M. T., Lochman, J. E., McMahon, R. J., & Pinderhughes, E. E. (2002). Using the Fast Track randomized prevention trial to test the early-starter model of the development of serious conduct problems. Development and Psychopathology, 14(4), 925–943. Bierman, K. L., Smoot, D. L., & Aumiller, K. (1993). Characteristics of aggressive-rejected, aggressive (non-rejected) and rejected (non-aggressive) boys. Child Development, 64, 139–151. Borge, A., Rutter, M., Côté, S., & Tremblay, R. E. (2004). Early childcare and physical aggression: Differentiating social selection and social causation. Journal of Child Psychology and Psychiatry, 45(2), 367–376. Bourdieu, P. (1980). Le capital social: Notes provisoires, Actes de la Recherche en Sciences Sociales, 31, 2–3. Brown, B. B., Lohr, M. J., & McClenahan, E. L. (1986). Early adolescents’ perceptions of peer pressure. Journal of Early Adolescence, 6, 139–154. Browning, C. R. (2002). The span of collective efficacy: Extending social disorganization theory to partner violence. Journal of Marriage and Family, 64(4), 833–850. Buysse, W. H. (1997). Behaviour problems and relationship with family and peers during adolescence. Journal of Adolescence, 20(6), 645–659. Cancian, M. (2001.) Rhetoric and reality of work-based welfare reform. Social Work, 46(4), 309–314. Coie, J. D., Lochman, J. E., Terry, R., & Hyman, C. (1992). Predicting early adolescent disorder from childhood aggression and peer rejection. Journal of Consulting and Clinical Psychology, 64, 53–63. Coleman, J. S. (1990). Foundation of social theory. Cambridge, MA: Harvard University Press. Colvin, M., Cullen, F. T., & Vander Ven, T. (2002). Coercion, social support, and crime: An emerging theoretical consensus. Criminology, 40(1), 19–42. Conduct Problems Prevention Research Group. (2002). Evaluation of the first 3 years of the Fast Track prevention trial with children at high risk for adolescent conduct problems. Journal of Abnormal Child Psychology, 30, 19–35. Criss, M. M., Pettit, G. S., Bates, J. E., Dodge, K. A., & Lapp, A. L. (2002). Family adversity, positive peer relationship, and children’s externalizing behavior: A longitudinal perspective on risk and resilience. Child Development, 73(4), 1220–1237. DeFronzo, J. (1983). Economic assistance to impoverished Americans. Relationship to incidence of crime. Criminology, 21(1), 119–136. DeFronzo, J. (1997). Welfare and homicide. Journal of Research on Crime and Delinquency, 34(3), 395–406. Dodge, K. A., Lochman, J. E., Harnish, J. D., Bates, J. E., & Pettit, G. S. (1997). Reactive and proactive aggression in school children and psychiatrically impaired chronically assaultive youth. Journal of Abnormal Psychology, 106, 37–51. Dubet, F. (1987). La galère: Les jeunes en survie. Paris: Fayard. The Economist. (2001, August 25). Welfare reform: America’s great achievement. The Economist, 25–27. Elliott, D. S. (1994). Serious violent offenders: Onset, developmental course, and termination—The American Society of Criminology 1993 presidential address. Criminology, 32, 1–21. Esbensen, F., & Huizinga, D. (1993). Gangs, drugs, and delinquency in a survey of urban youth. Criminology, 31, 565–589. Esbensen, F., Peterson, D., Freng, A., & Taylor, T. J. (2002). Initiation of drug use, drug
420
DETERMINANTS OF AGGRESSION
sales, and violent offending among a sample of gang and non-gang youth. In C.R. Huff (Ed.), Gangs in America (3rd ed., pp. 37–50). Beverly Hills, CA: Sage. Farrington, D. P. (1989). Early predictors of adolescent aggression and adult violence. Violence and Victims, 4, 79–100. Felson, R. B., Liska, A. E., South, S. J., & McNulty, T. L. (1994). The subculture of violence and delinquency: Individual vs. school context effects. Social Forces, 73(1), 155–173. Ferris, J. S., & West, E. G. (2002). Economies of scale, school violence, and the optimal size of schools. Carleton Economic Papers, No. 02–01. Ottawa, QC: Carleton University. Fukuyama, F. (1995). Trust: The social virtues and the creation of prosperity. New York: Free Press. Furstenberg, F. F. Jr., & Hughes, M. E. (1995). Social capital and successful development among at-risk youth. Journal of Marriage and the Family, 57(3), 580–592. Gatti, U., Schadee, M. A. H., & Tremblay, R. E. (2002a, September). Lagged effects of nurturance on crime in Italian provinces. Paper presented at the Second Conference of the European Society of Criminology, Toledo, Spain. Gatti, U., Tremblay, R. E., & Larocque, D. (2003). Civic community and juvenile delinquency: A study of the regions of Italy. British Journal of Criminology, 43, 22–40. Gatti, U., Tremblay, R. E., & Schadee, H. (2002b, July). Social capital and violent crime in Italian provinces. Paper presented at the 15th World Meeting of the International Society for Research on Aggression, Montreal, Canada. Gatti, U., Tremblay, R. E., Vitaro, F., & McDuff, P. (in press). Youth gangs, delinquency and drug use: A test of the selection, facilitation, and enhancement hypotheses. Journal of Child Psychology and Psychiatry. Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1370–1380. Hagan, J., & McCarthy, B. (1997). Mean streets: Youth crime and homelessness. Cambridge, MA: Cambridge University Press. Han, W., Waldfogel, J., & Brooks-Gunn, J. (2001) The effects of early maternal employment on later cognitive and behavioral outcomes. Journal of Marriage and Family, 63(2), 336–354. Haskins, R. (1985). Public school aggression among children with varying day-care experience. Child Development, 56, 689–703. Hawkins, J. D., Catalano, R. F., Kosterman, R., Abbott, R., & Hill, K. G. (1999). Preventing adolescent health-risk behaviors by strengthening protection during childhood. Archives of Pediatrics and Adolescent Medicine, 153, 226–234. Hemenway, D., Kennedy, B. P., Kawachi I., & Putnam, R. D. (2001). Firearm prevalence and social capital. Annals of Epidemiology, 11, 484–490. Henry, B., Caspi, A., Moffit, T. E., & Silva, P. A. (1996). Temperamental and familial predictors of violent and nonviolent criminal convictions: Age 3 to age 18. Developmental Psychology, 32, 614–623. Hobbes, T. (1647/1998). De Cive. New York: Cambridge University Press. Jankowsky, M. S. (1991). Islands in the street: Gangs in American urban society. Berkeley, CA: University of California Press. Karen, R. (1994). Becoming attached. New York: Warner Books. Keenan, K., & Shaw, D. S. (1994). The development of aggression in toddlers: A study of low-income families. Journal of Abnormal Child Psychology, 22(1), 53–77. Kellam, S. G., Rebok, G. W., Ialongo, N., & Mayer, L. S. (1994). The course and mallea-
Social Capital and Physical Violence
421
bility of aggressive behavior from early first grade into middle school: Results of a developmental epidemiologically-based preventive trial. Journal of Child Psychology and Psychiatry, 35, 259–281. Kennedy, B. P., Kawachi, I., Prothrow-Stith, D., Lochne, K., & Gupta, V. (1998). Social capital, income inequality, and firearm violent crime. Social Science and Medicine, 1, 7–17. Ku, I. (2001). The effect of welfare on children’s education. Social Service Review, 75(2), 245–270. Lacourse, E., Nagin, D. S., Tremblay, R. E., Vitaro, F., & Claes, M. (2003). Developmental trajectories of boys’ delinquent group membership and facilitation of violent behaviors during adolescence. Development and Psychopathology, 15, 183–197. LeBlanc, L., Swisher, R., Vitaro, F., & Tremblay, R. E. (2004). Antisocial behavior and high school social climate: A 10 year longitudinal and multilevel study. Manuscript submitted for publication. Lin, N. (1999). Social networks and status attainment. Annual Review of Sociology, 25, 467–487. Marrazzo, G. (1984). Il camorrista. Napoli: Tullio Pironti. McCord, J. (2002, July). Social capital, family socialization, and aggressive behavior. Paper presented at the 15th World Meeting of the International Society for Research on Aggression, Montreal, Canada. Messner, S. F., Baumer, E., & Rosenfeld, R. (2002, July). Social capital and violent crime rates across U. S. communities. Paper presented at the 15th World Meeting of the International Society for Research on Aggression, Montreal, Canada. Messner, S. F., & Rosenfeld, R. (1997). Political restraint of the market and levels of criminal homicide: A cross-national application of institutional-anomie theory. Social Forces, 75(4), 1393–1416. Morris, P. (2002). The effect of welfare reform policies on children. Social Policy Report, 16(1), 4–19. Morrow, V. (1999). Conceptualizing social capital in relation to the well-being of children and young people: A critical review. The Sociological Review, 47(4), 744– 765. Nagin, D., & Tremblay, R. E. (2001). Parental and early childhood predictors of persistent physical aggression in boys from kindergarten to high school. Archives of General Psychiatry, 58, 389–394. National Center for Education Statistics. (2002). Digest of education statistics. Washington DC: U.S. Department of Education. NICHD Early Child Care Research Network. (2002). Child-care structure → process → outcome: Direct and indirect effects of child-care quality on young children’s development. Psychological Science,13(3), 199–206. NICHD Early Child Care Research Network. (2003). Does amount of time in child care predict socioemotional adjustment? Child Development, 74(4), 976–1005. Olds, D. L., Eckenrode, J., Henderson, C. R., Jr., Kitzman, H., Powers, J., Cole, R., Sidora, K., Morris, P., Pettitt, L. M., & Luckey, D. (1997). Long-term effects of home visitation on maternal life course, and child abuse and neglect. Journal of the American Medical Association, 278, 637–643. Pagani, L., Boulerice, B., & Tremblay, R. E. (1997). The influence of poverty upon children’s classroom placement and behavior problems during elementary school: A change model approach. In G. Duncan & J. Brooks-Gunn (Eds.), Consequences of growing up poor (pp. 311–339). New York: Sage.
422
DETERMINANTS OF AGGRESSION
Pampel, F., & Gartner, R. (1995). Age-structure, socio-political institutions, and national homicide rates. European Sociological Review, 11, 243–260. Parcel, T. L., & Dufur, M. J. (2001). Capital at home and at school: Effects on child social adjustment. Journal of Marriage and Family, 63(1), 32–47. Parker, J. G., & Asher, S. R. (1987). Peer relations and later personal adjustment: Are low-accepted children at risk? Psychological Bulletin, 102, 357–389. Paxon, C., & Waldfogel, J. (2001, November). Welfare reforms, family resources and child maltreatment. Paper presented at the 23rd annual research conference of the Association for Public Policy Analysis and Management, Washington, DC. Portes, A. (1998). Social capital: Its origins and applications in modern society. Annual Review of Sociology, 24, 1–24. Putnam, R. D. (1993). Making democracy work: Civic tradition in modern Italy. Princeton, NJ: Princeton University Press. Putnam, R. D. (2000). Bowling alone. The collapse and revival of American community. New York: Simon & Schuster. Quinton, D., Pickles, A., Maughan, B., & Rutter, M. (1993). Partners, peers and pathways: Assortative pairing and continuities in conduct disorder. Development and Psychopathology, 5, 763–783. Raine, A., Reynolds, C., Venables, P. H., Mednick, S. A., & Farrington, D. P. (1998). Fearlessness, stimulation-seeking, and large body size at age 3 years as early predispositions to childhood aggression at age 11 years. Archives of General Psychiatry, 55(8), 745–751. Rosenfeld, R., Messner, S. F., & Baumer, E. P. (2001). Social capital and homicide. Social Forces, 80(1), 283–310. Rousseau, J. J. (1762/1986). Émile (B. Foxley, Trans.). London: Dent. Rubenstein, J., & Howes, C. (1983). Social-emotional development of toddlers in day care: The role of peers and individual differences. In S. Kilmer (Ed.), Advances in early education and day care (Vol. 3, pp. 13–45). Greenwich, CT: JAI Press. Rutter, M., Giller, H., & Hagell, A. (1998). Antisocial behavior by young people. New York: Cambridge University Press. Rutter, M., Maughan, B., Mortimore, P., Ouston, J., & Smitj, A. (1979). Fifteen thousand hours: Secondary schools and their effects on children. Cambridge, MA: Harvard University Press. Sampson, R. J., & Groves, B. W. (1989). Community structure and crime: Testing socialdisorganization theory. American Journal of Sociology, 94, 774–802. Sampson, R. J., & Laub, J. H. (2003). Life-course desisters? Trajectories of crime among delinquent boys followed to age 70. Criminology, 41(3), 555–592. Sampson, R. J., Morenoff, J. D., & Earls, F. (1999). Beyond social capital: Spatial dynamics of collective efficacy for children. American Sociological Review, 54(5), 633– 660. Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhood and violent crime: A multilevel study of collective efficacy. Science, 277, 918–924. Savage, J., & Vila, B. (1997). Lagged effects of nurturance on crime: A cross-national comparison. Studies on Crime and Crime Prevention, 6(1), 101–120. Savage, J., & Vila, B. (2002). Changes in child welfare and subsequent crime rate trends: A cross-national test of the lagged nurturance hypothesis. Applied Developmental Psychology, 23, 51–82. Savolaien, J. (2000). Inequality, welfare state, and homicide: Further support for the institutional anomie theory. Criminology, 38(4), 1021–1039.
Social Capital and Physical Violence
423
Smith, C. A., & Farrington, D. P. (2003). Continuities in antisocial behavior and parenting across three generations. Journal of Child Psychology and Psychiatry, 44, 1–18. Thornberry, T. P., Krohn, M. D., Lizotte, A. J., & Chard-Wierschem, D. (1993). The role of juvenile gangs in facilitating delinquent behavior. Journal of Research in Crime and Delinquency, 30, 55–87. Thornberry, T. P., Krohn, M. D., Lizotte, A. D., Smith, C. A., & Tobin, K. (2002). Gangs and delinquency in developmental perspective. Cambridge, UK: Cambridge University Press. Tremblay, R. E., Barr, R. G., Peters, R. D. (Eds.). (2003). Encyclopedia on Early Childhood Development [Online]. Montreal, QC: Centre of Excellence for Early Childhood Development. Available at http://www.excellence-earlychildhood.ca/documents/TremblayANGxp.pdf. Tremblay, R. E., & Craig, W. (1995). Developmental crime prevention. In M. Tonry & D. Farrington (Eds.), Building a safer society: Strategic approach to crime prevention. Chicago: The University of Chicago Press. Tremblay, R. E., Kurtz, L., Mâsse, L. C., Vitaro, F., & Pihl, R. O. (1995). A bimodal preventive intervention for disruptive kindergarten boys: Its impact through mid-adolescence. Journal of Consulting and Clinical Psychology, 63(4), 560–568. Tremblay, R. E., Nagin, D. S., Séguin, J. R., Zoccolillo, M., Zelazo, P., Boivin, M., Pérusse, D., & Japel, C. (2004). Physical aggression during early childhood: Trajectories and predictors. Pediatrics, 114, e43–e50. Vaden-Kiernan, N., Ialongo, N. S., Pearson, J., & Kellam, S. (1995). Household family structure and children’s aggressive behavior: A longitudinal study of urban elementary school children. Journal of Abnormal Child Psychology, 23(5), 553–568. Vander Ven, T. M., Cullen, F. T., Carrozza, M. A., & Wright, J. P. (2001). Home alone: The impact of maternal employment on delinquency. Social Problems, 48(2), 236– 257. Vila, B. (1994). A general paradigm for understanding criminal behavior: Extending evolutionary ecological theory. Criminology, 32, 501–549. Vinson, T., Baldry, B., & Hargreaves, J. (1996). Neighbourhoods, networks and child abuse. British Journal of Social Work, 26, 523–542. Vitaro, F., Brendgen, M., Pagani, L., Tremblay, R. E., & McDuff, P. (1999). Disruptive behavior, peer association, and conduct disorder: Testing the developmental links through early intervention. Development and Psychopathology, 11, 287– 304. Vitaro, F., Brendgen, M., & Tremblay, R. E (2001a). Preventive intervention: Assessing its effects on the trajectories of delinquency and testing for mediational processes. Applied Developmental Science, 5(4), 201–213. Vitaro, F., Brendgen, M., & Tremblay, R. E (2002). Reactively and proactively aggressive children: Antecedent and subsequent characteristics. Journal of Child Psychology and Psychiatry, 43(4), 495–505. Vitaro, F., Tremblay, R. E., & Bukowski, W. M. (2001b). Friends, friendship and conduct disorders. In J. Hill & B. Maughan (Eds.), Conduct disorders in childhood and adolescence (pp. 346–378). Cambridge, UK: Cambridge University Press. Vitaro, F., Tremblay, R. E., Kerr, M., Pagani, L., & Bukowski, W. M. (1997). Disruptiveness, friends’ characteristics, and delinquency in early adolescence: A test of two competing models of development. Child Development, 68, 676–689. Warner, B., & Rountree, P. W. (1997). Local social ties in a community and crime model:
424
DETERMINANTS OF AGGRESSION
Questioning the systemic nature of informal social control. Social Problems, 44, 520–536. Weikart, D. P. (1998). Changing early childhood development through educational intervention. Preventive Medicine, 27(2), 233–237. Weikart, D. P., & Schweinhart, L. J. (1992). High/Scope preschool program outcomes. In J. McCord & R. E. Tremblay (Eds.), Preventing antisocial behavior: Interventions from birth to adolescence (pp. 67–86). New York: Guilford Press. Wolfgang, M. E., & Ferracuti, F. (1967). The subculture of violence: Towards an integrated theory in criminology. London: Tavistock Publications. Wright, J. P., Cullen, F. T., & Miller, J. T. (2001). Family social capital and delinquent involvement. Journal of Criminal Justice, 29, 1–9. Zaslow, M., Tout, K., Smith, S., & Moore, K. (1998). Implications of the 1996 welfare legislation for children: A research perspective. Social Policy Report. Society for Research in Child Development, 12(3), 1–35. Zhang, L., Welte, J. W., & Wieczorek, W. F. (1999). Youth gangs, drug use, and delinquency. Journal of Criminal Justice, 27, 101–109.
DETERMINANTS Sex Differences in Aggressive OF AGGRESSION Behavior
20 Sex Differences in Aggressive Behavior A Developmental and Evolutionary Perspective J OHN A RCHER and S YLVANA C ÔTÉ
INTRODUCTION: EVOLUTIONARY AND DEVELOPMENTAL EXPLANATIONS Explanations of Behavior Tinbergen (1963) set out four types of explanation of behavior. Two concerned the evolution of behavior, its historical antecedents and its survival value, and two concerned causes within the individual, its development and immediate causation. A full explanation of any form of behavior that humans share with animals involves all four explanations. Aggression is such a form of behavior, having its origins in evolutionary history. Psychologists studying human aggression typically concentrate on individual-level explanations, immediate causation and development, referred to as proximal explanations. Evolutionary psychologists and ethologists are also concerned with the two evolutionary-based explanations, referred to as distal explanations. In this chapter we consider the development of sex differences in human aggression in relation to evolutionary explanations.
Evolutionary History The evolutionary history of aggression can be appreciated by noting its widespread occurrence throughout the animal kingdom, from animals with very simple nervous systems, such as sea anemones, to humans (Archer, 1988). It is therefore not something that can realistically be viewed only as a consequence of social influences on human development, although there are social environments that greatly increase an individual’s level of aggression, and individuals 425
426
DETERMINANTS OF AGGRESSION
who are more aggression-prone than others whatever their social environment.
Evolutionary Function Fighting is widespread throughout the Animal kingdom. It occurs in two main contexts, in warding off danger and as a consequence of competition for resources (Archer, 1988). Competition is a wider term than physical aggression and results from a simple ecological principle. Organisms typically produce more offspring than their environment can sustain; resources are therefore scarce, and some will obtain more than others. Occasionally, this is the result of chance, but animals are typically selected for features that enable them to outcompete and outbreed others of their kind. This process often occurs over a period of time and is not immediately obvious. For example, one animal’s digestive system may be more efficient than another’s, or one animal may be better at finding food than another. Aggressive behavior is a form of competition that has evolved where organisms have the ability to displace one another quickly, and there are resources worth fighting over. Thus, animals that can remove others by force will seek to do so. Resources that are located in one place and can be obtained by actively displacing a rival will be fought over, whereas those that are evenly and abundantly distributed over a wide area will not. Thus, hyenas fight over a carcass whereas cows do not fight over grass. Broadly speaking, the occurrence of aggression in animals can be understood in terms of the costs and benefits of fighting (Archer, 1988), and this principle can be extended to human aggression.
Sexual Selection A more specific application of the cost–benefit analysis is incorporated into the principle of sexual selection (Darwin, 1871), which involves intermale competition and female choice. With a few exceptions, males have more to gain than females have through fighting one another (or more to lose through not fighting). Trivers (1972) explained why this is the case. The specialization of germ cells leads to unequal parental investment. Females show greater initial investment in terms of the energy and time costs of producing egg cells, and therefore a greater cost is sustained if they choose a low-fitness mate. Males show less initial investment, and therefore sustain a lower cost through choosing a low-fitness mate. These factors result in females being more selective in their mate choice. They also lead to males being in competition with other males for both access to females and the resources that make them more acceptable to females. These considerations enable us to understand why greater physical aggression by males than females is observed in many animal species and is com-
Sex Differences in Aggressive Behavior
427
mon throughout the mammals. Trivers’s (1972) theory of parental investment predicts that sex differences in aggression (and in associated features such as size and strength) are most pronounced when male parental investment is absent and competition between males is at its most intense. This is observed in species such as the elephant seal (Le Boeuf, 1974), when one male defends a group of as many as 40 females against other males. In species in which the male contributes considerably to subsequent parental investment, sex differences are minimal. In those relatively unusual cases in which male parental investment exceeds that of females, it is the females who are larger and more aggressive.
Human Sex Differences in Evolutionary Perspective Comparison of sex differences in body size with those in other primates suggests that there has, in evolutionary history, been a tendency toward mild polygyny in humans (Plavcan & van Schaik, 1997). A range of studies suggests that the sexual natures of men and women, and sex differences in aggression, fit this pattern (Archer & Lloyd, 2002), although there is also considerable variation between males (Archer & Mehdikhani, 2003). A range of evidence on direct physical aggression shows that it is greater among males than females—at least toward adults of the same sex. There are large sex differences in measures of physical aggression (Archer, 2004), and in recorded violent crime (Courtwright, 1996; Daly & Wilson, 1990), for same-sex encounters.
Development from an Evolutionary Perspective Given this evolutionary background, we can understand why human aggression occurs and why there are sex differences in direct forms of human aggression. However, how aggression develops and when the sex differences first appear in development is another question, not readily answered by these evolutionary considerations alone. To appreciate some of the complexities that are involved when we consider development from an evolutionary perspective, it is worth noting a number of general principles that have arisen from research on development from an evolutionary perspective (Archer, 1992a; Bjorklund & Pellegrini, 2002; Stamps, 2003). In each case, we indicate briefly how the principle might apply to the development of aggression and sex differences in aggression. The first principle is that behavior serving an evolutionary function may arise from different developmental processes in different animals. Thus, greater aggressiveness by males is achieved in different ways in different species. In some birds and mammals, testosterone directly enhances male aggressiveness, but as we show in this chapter, it is likely that different processes operate to produce this same end point in humans.
428
DETERMINANTS OF AGGRESSION
Second, some types of behavior that develop during childhood serve the function of preparing the individual for a particular role in adulthood. Sex differences in aggression and rough-and-tumble play are likely examples of this in humans (Bjorklund & Pellegrini, 2002). In other cases, behavior may serve an adaptive function during childhood itself. We show that physical aggression develops early in childhood, and that the objective of obtaining and retaining resources may be one reason why this occurs. Third, children are born not with a tabula rasa, but with a program for development that has evolved in human ancestral environments to serve an adaptive outcome. However, the potential range of outcomes, or phenotypes, that can be produced from this program is wide. This is termed the norm of reaction, and it is a particularly important principle for considering development in modern human environments, which are very different from those of our prehistoric ancestors. Thus, the modern school environment may lead to once-adaptive behavior, such as high levels of aggression and rough-and-tumble play, being viewed as maladaptive. Fourth, some evolved forms of behavior are little affected by variability in the environments in which the individuals develop, whereas others are much affected by it. The onset of physical aggression early in life may not be readily susceptible to environmental modification, whereas its subsequent development is likely to be influenced to a much greater extent by the type of social and physical environment in which the individual develops. Fifth, individuals select and modify the environments in which they develop. For example, as a consequence of their own preferences, boys and girls form sex-segregated groups early in their social development, and these last through to the teenage years (Maccoby, 1998). Such groups constitute different social environments, in which boys and girls come to develop different patterns of behavior including aggression, boys showing more direct physical forms, and girls more indirect forms involving manipulation of relationships. A sixth principle is that selection may favor phenotypic diversity rather than a single adaptive type. There are various reasons for this, such as that different phenotypes are adapted for different environments, or that a rare phenotype has an advantage simply because it is unusual. Males of many animal species show what are termed alternative reproductive strategies, for example, specialized either for fighting other males or for sneaking up on females when other males are distracted. In humans, it has been suggested that males vary from the more aggressive risk-prone male who favors a short-term promiscuous reproductive strategy to the less aggressive, more risk-averse male who favors a longer-term parental reproductive strategy (Archer & Mehdikhani, 2003). It follows that a variety of interlinked characteristics, or trait syndromes, are likely to develop together in the same individual. Thus, an aggression-prone individual may show a number of other high-risk and impulsive types of behavior that in our society are typically labeled as antisocial or criminal.
Sex Differences in Aggressive Behavior
429
THE DEVELOPMENT OF PHYSICAL AGGRESSION The social learning perspective has provided an influential account of the development of social behavior, including aggression, in terms of principles originally derived from learning theory (Bandura, 1973; Loeber & StouthamerLoeber, 1998; Reiss & Roth, 1993). Aggressive and violent behaviors are viewed as being learned during childhood and adolescence through exposure to social influences such as violent television, aggressive role models, or deviant peers (Johnson, Cohen, Smailes, Kasen, & Brook, 2002; Patterson, Dishion, & Bank, 1984; Thornberry, 1998). From this perspective alone, we would expect children to become more aggressive and violent as they become older. The social learning perspective provides a stark contrast with a view of development informed by evolutionary principles. First, there is no recognition of an evolved program for the development of aggression to serve the immediate function of resource competition during childhood or as a preparation for conflicts during adulthood. Evolutionary approaches to development recognize that social learning is an important influence in behavior (Stamps, 2003). However, it is one among a wider range of genetic and environmental influences and is likely to extend beyond the facilitation of aggression by social learning that has been the focus of most investigations from this perspective. Recent longitudinal studies indicate that children become less, not more, physically aggressive and violent over time (Broidy et al., 2003; Cairns, Cairns, Neckerman, Fergusson, & Gariepy, 1989; Nagin & Tremblay, 1999; Tremblay et al., 1999). Studies have found that most children followed declining trajectories of physical aggression between kindergarten and grade 6 (Broidy et al., 2003; Nagin & Tremblay, 1999). If the development of aggression were dependent solely on the cumulative effects of social influences tending to promote aggressive behavior, we would expect a significant group of children to begin their use of physical aggression during the elementary school years. Such a group has not been not identified (Kingston & Prior, 1995; Loeber & Stouthamer-Loeber, 1998; Tremblay et al., 1991). The majority of those at risk for later violent behavior were already on high trajectories of physical aggression in kindergarten (Broidy et al., 2003; Nagin & Tremblay, 1999). The possibility that aggression in middle childhood simply reflects the continuation of a behavior pattern already present during the first years of life is supported by research on the development of disruptive behavior disorders during the preschool years. Physically aggressive behavior is common among infants and toddlers (Goodenough, 1931; Hay, Castle, & Davies, 2000; Tremblay et al., 1999). However, most studies have examined the development of aggression during limited periods of development. An exception is the National Longitudinal Study of Canadian Youth (NLSCY), in which the developmental trajec-
430
DETERMINANTS OF AGGRESSION
tories of physical aggression between infancy and preadolescence were estimated, with longitudinal data collected on a population sample (n = 8,208). The developmental patterns identified with mother reports indicated that aggression was part of most (65.5%) preschoolers’ behavioral repertoire, but that the majority of children had successfully ceased to use physical aggression by the end of the elementary school years. As shown in Figure 20.1, four distinct trajectory groups were identified in the NLSCY. The first comprised children with stable low levels of aggression. The second group—the rapid-desister group—included children with moderate levels of physical aggression in toddlerhood, who desisted rapidly and maintained low levels from school entry to preadolescence. The third group was labeled the “moderate desisters”: it included children with slow but constant declines in physical aggression between 2 and 11 years. Finally, children
FIGURE 20.1. Trajectories of physical aggression for boys and girls (n = 8,208).
Sex Differences in Aggressive Behavior
431
in the high physical aggression group were also desisting. However, their declines were of such small magnitude that in preadolescence (age 11) they still exhibited notably higher levels of physical aggression than did their peers. It is important to note that no significant group of children with an onset of physical aggression during this period (between 2 and 11 years) was identified. According to a retrospective study with younger children, most infants start to use physical aggression between the first and the second years after birth (Tremblay et al., 1999). Its onset is therefore likely to be earlier and to be relatively stable, rather than being dependent on specific environmental influences that may lead to variable onset (see Tremblay & Nagin, Chapter 5, this volume). This suggests that physical aggression is an evolved pattern of behavior that develops early in life, irrespective of the sort of social influences highlighted by social learning theory. From the evidence reviewed in this section, it is clear that human aggression appears early in life and is part of social interactions right from their beginning. Social learning theory has perhaps laid too much emphasis on the learning of specific forms of aggressive behavior as a general mechanism for explaining the developmental progression of aggression. Although such specific aggressive actions can undoubtedly be learned from peers, parents, and media representations, they do not account for broad changes with age. However, social factors such as coercive parenting, violent models, or violent media may be related to the failure, for a minority of children, to inhibit and regulate their aggression. Why do high levels of physical aggression appear at such young ages? We have already highlighted the importance of competition for resources in functional explanations of aggression. Among young children, we would predict that aggression arises when more than one child claims the same resources. In modern affluent societies, children do not have to fight for their lives, as occurs occasionally in other species (Frank, 1997; Stinson, 1979), or over resources necessary for survival, such as food, as occurs in the young of many birds and mammals (e.g., Bekoff, 1981; Geist, 1978). They may, however, fight over other resources, such as a preferred food, a toy, or a play location. Campbell (1999) argued that most conflicts at preschool ages involve competing for limited resources, such as guarding a toy or a territory.
THE DEVELOPMENT OF SEX DIFFERENCES IN PHYSICAL AGGRESSION Sex differences have been found from the onset of physical aggression itself, in children as young as 17 months, according to mothers’ reports of specific acts of physical aggression (Baillargeon, 2002). In this study, boys were twice as likely as girls to hit another child frequently. Observational studies show differences of fairly substantial magnitudes at around 3 years of age. Hay et al. (2000) found an effect size of d = 0.64, for a (relatively large) sample of 66 children between 18 and 30 months of age. Campbell, Shirley, and Caygill
432
DETERMINANTS OF AGGRESSION
(2002) observed considerably more negative interactions with peers by boys than girls in children who were 27 months old. Negative interactions included a child’s grabbing another child’s toy (d = 1.42) and a child’s resisting another’s attempt to grab his or her toy (d = 0.86).1 Observational studies of children 4 years of age have also yielded large effect sizes (e.g., Sears, Rau, & Alpert, 1965: d = 1.27; McGrew, 1972: d = 1.29). The conclusion that sex differences in physical aggression are present early in life is counter to earlier accounts based on the assumption that they arise as a result of the differential socialization of boys and girls (e.g., Tieger, 1980). In a review of sex differences in several types of childhood behavior problems, Keenan and Shaw (1997) put forward a similar view. They reviewed findings from five observational studies (n = 20–90) concerned with the development of aggression during the preschool years. They reported that boys were more aggressive than girls in only one of these studies (Gunnar, Senior, & Hartup, 1984) and concluded that there is no sex difference in aggression for toddlers, although there is an increasing divergence in girls’ and boys’ behavior during the preschool period. This was based on two longitudinal studies indicating that girls exhibited less aggression by age 5. Keenan and Shaw’s conclusion that there are no sex differences for toddlers is not supported by closer examination of the evidence they cite. Their own study of children ages 18 and 24 months interacting with their mothers showed sex differences in global aggression at both ages (18 months: d = 0.49; confidence interval [CI] = 0.07–0.91; p = .02; 24 months: d = 0.39; CI = 0.02– 0.80; p = .06), those at the younger age being greater (figures calculated from Table 2 in Shaw, Keenan, & Vondra, 1994). Jacklin and Maccoby (1978) observed dyads of 33-month-old children: Comparing aggression among male dyads with that among female dyads (i.e., taking each dyad as a unit) produced a small effect size in the male direction (d = 0.18). Cummings, Iannotti, and Zahn-Waxler (1989) observed children playing with their best available friends at 2 and at 5 years. For the two ages combined (the only calculation possible), effect sizes for sex differences in aggression were d = 0.26 in the male direction. Caplan, Vespo, Pedersen, and Hay (1991) observed 1- and 2year-olds in mixed-sex groups of three for 25 minutes. The main comparison was between the level of conflict in groups with one male and groups with one female. A footnote indicated that there was no significant sex difference at an individual level, but no direction was reported (for this sample size, the d value could have been anywhere between -0.57 and 0.57). The remaining study (Gunnar et al., 1984) found significantly longer duration of conflicts among male than female dyads (d = 1.04). A meta-analysis of the four studies that enabled effect sizes to be calculated produced a weighted mean of d = 0.44 (CI = 0.15–0.74; p = .003), showing that there was a moderate sex difference in the male direction. This contrasts with Keenan and Shaw’s conclusion, based on a narrative review, that there are no sex differences at young ages. It supports the conclusion from the
Sex Differences in Aggressive Behavior
433
other studies described above, that the sex difference occurs as early as it is possible to measure children’s aggressive behavior. Keenan and Shaw’s (1997) review also raised the question of how sex differences in physical aggression change throughout childhood, as they claimed that it increased from about 2 to 5 years of age, based on two longitudinal studies. A meta-analysis of 40 observational studies carried out between 1965 and 2000 (Archer, 2004) showed no significant difference in the effect size for the sex difference in aggression (overall d = 0.53) between three age categories, under 6 years, 6–11 years, and over 11 years. However, this involved cross-sectional comparisons of different studies, and there are limitations to the inferences that can be drawn from such evidence. Large-scale representative samples provide a way of overcoming the limitations, although at younger ages they involve less direct assessments of behavior, inasmuch as they generally rely on mothers’ reports. In the NLSCY (see the preceding section), sex differences in physical aggression were examined by comparing the proportion of boys and girls assigned to the different trajectory groups. The results showed that boys were more likely to be on the highest physical aggression trajectories (odds ratio [OR] = 1.56; 95% CI = 1.41–1.71) and less likely to be on the lowest trajectories (OR = 0.85; 95% CI = 0.82–0.89). This suggests that boys were already more physically aggressive than girls at 2 years of age, a difference that persisted throughout childhood. These sex differences are consistent with the findings from the individual observational studies presented earlier. The NLSCY study also found that the initial sex difference became more pronounced during the course of childhood, as boys were less likely to be assigned to the rapid-desister trajectory group (OR= 0.88; 95% CI = 0.82– 0.94). These findings support Keenan and Shaw’s (1997) conclusion that sex differences in physical aggression increase during the preschool years and that this is attributable to girls’ faster rate of decline between 2 and 5 years. Thus, girls may learn more quickly than boys to inhibit or control (physically) aggressive behaviors (Kochanska, Murray, & Harlan, 2000). Finally, there were no significant sex differences for the moderate desister group (OR = 1.05; 95% CI = 1.00–1.09). One further point about the sex differences found in the NLSCY study is that they are far from absolute. Many boys are found in the low groups and the rapid-desister groups, and many girls are found in the high-trajectory group. Such variation may be worth investigating in relation to the evolutionary hypothesis of alternative reproductive strategies, which would predict consistent ways of responding, with an early developmental onset and a consistent outcome. A meta-analysis of sex differences in self-reports of physical aggression (Archer, 2004) showed similar effect sizes for the age categories 6–11 years, 12–3 years, and 14–17 years. The lack of any further widening in the sex difference at puberty provides indirect evidence that pubertal testosterone does not increase male aggressiveness, as it does in some other mammals (e.g.,
434
DETERMINANTS OF AGGRESSION
Beeman, 1947; Edwards, 1969). More direct support for this conclusion comes from a longitudinal study of boys going through puberty (Halpern, Udry, Campbell, & Suchindran, 1994), which found no influence of rising testosterone levels on physical aggression. The meta-analysis of self-reported physical aggression (Archer, 2004) did find larger sex differences in the age categories 18–21 and 22–30 years than at younger (14–17 years) or older ages (31–55 years). Studies of violent crime and of homicides show that there is a pronounced peak in males between about 20 and 29 years of age. This was found for homicides in data sets from four locations with different homicide rates, England and Wales, Canada, Chicago, and Detroit (Daly & Wilson, 1990), and for violent crimes in the modern United States, 19th-century England and Wales, and early 19th-century Europe (Courtwright, 1996; Quetelet, 1833/1984). Although these studies differ from those we have considered so far in that they involve much rarer violent acts, defined by their damaging consequences rather than the type of behavior, they show a striking parallel in their pattern of age and sex distribution. The evidence reviewed in this section indicates that there is a large sex difference in physical aggression, beginning early in life, suggesting that it is not the product of a gradual process such as socialization (Tremblay et al., 1999). However, there is also evidence from a longitudinal study that the sex difference in physical aggression, such as hitting, biting, and attacking others, may become larger between toddlerhood and kindergarten. There is no evidence that the sex difference widens at puberty as a consequence of the action of the male hormone testosterone. However, studies of self-reported physical aggression and of violent acts converge to suggest a widening of the sex difference between 18 and 30 years of age. These are peak years for the development of a constellation of characteristics that enable a young man to successfully compete with other males.
THE DEVELOPMENT OF SEX DIFFERENCES IN PHYSICAL AGGRESSION: AN EVOLUTIONARY VIEW We noted that resource competition provides an evolutionary reason for the appearance of physical aggression early in life. However, it is less clear why there would be sex differences in this aggression early in life. We might expect, in the absence of reproductive competition, boys and girls to behave similarly, inasmuch as they would be competing for similar resources. Campbell (1999) articulated this view, arguing that, as toddlers, children’s conflicts are mainly related to competing for limited resources, such as guarding a toy or a territory. She also argued that there are no sex differences in this type of resource-oriented aggression among toddlers, whereas there is a sex difference in status-oriented aggression by 4 years of age. The first conclusion was based on a narrative review of a very small number of studies that sought to
Sex Differences in Aggressive Behavior
435
distinguish these different categories of aggression. It is inconsistent with other findings (e.g., Gunnar et al., 1984; Jacklin & Maccoby, 1978), and with a later study by Campbell et al. (2002), which reported very large sex differences at 27 months in negative interactions (including a child’s grabbing another child’s toy and a child’s resisting another’s attempt to grab his or her toy). Campbell is more likely to be correct in her identification of status disputes as an important area in which the sexes diverge in the early years. It is clear that as children become older there is increasing sex segregation and different types of social relations in boys’ and girls’ groups, based initially on self-selection of playmates (Archer, 1992b; Maccoby, 1998). Boys are more concerned with status and dominance—for example, who is tough and who is the boss (Weisfeld, 1994)—and they operate in larger play groups. Girls form more intense relations with one or two close friends (Archer, 1992b). These contrasting patterns have wide-ranging implications for later sex differences in several areas of social behavior, including other forms of aggression (Maccoby, 1998). Because of the greater emphasis on physically based dominance in childhood, boys’ exchanges are more likely to develop into physical aggression, whereas girls develop strategies of indirect aggression that become prominent in the teenage years. Nevertheless, as we noted earlier, the expectation that there will be a greater contrast between the frequencies of physical aggression in boys’ and girls’ groups as they become older has, as yet, only limited empirical support. Sex differences in physical aggression prior to the ages of reproductive competition are found in the young of many mammals, together with sex differences in rough-and-tumble play. In addition to the advantages during childhood of gaining tangible rewards, such as food and other valued resources, success in physical fights, whether over status or resources, may have two other longer-term consequences. The first involves the learning of techniques of fighting that may prove valuable in young adulthood, and the learning of cues associated with when it is preferable not to fight (assessment strategies; Archer, 1994). The second involves the longer-term advantage, conferred by a physically based dominance acquired during childhood, for dominance in young adulthood. This may have been considerable in the ancestral environment, and may still be so where physical prowess forms the basis of dominance orders in adulthood, as occurs in many masculine subcultures. If the sex difference in physical aggression found in childhood arises mainly from the concern of boys with dominance relations and the learning of ways of fighting in preparation for adulthood, we would expect it to be largely manifest in relation to same-sex disputes. If instead it arises from resource competition (as suggested for early ages), we would expect the aggression to be directed to any child who has a preferred toy, irrespective of his or her sex. There are relatively few studies that have separated the sex of the opponent when considering children’s aggression. Where they have, the evidence supports the view that sex differences in physical aggression concern aggres-
436
DETERMINANTS OF AGGRESSION
sion to same-sex opponents. A different pattern (in the female direction) is found whenever the (less frequent) between-sex physical aggression is studied. For example, an observational study by Barrett (1979) of 6- to 7-year-old children found a substantial sex difference in physical aggression in the male direction for same-sex opponents (d = 1.05)2, and a difference in the female direction for opposite-sex opponents (d = –0.53). Pellegrini and Long (2002) followed up a sample of children at 6-monthly intervals from 12.8 years of age and found at all ages except the youngest one, a sex difference in the male direction for same-sex aggression and in the female direction for opposite-sex aggression. Effect sizes were in the moderate range, from 0.29 to 0.46. Cairns and Cairns (1994) reported similar findings using questionnaires at 11 and 14 years of age. A questionnaire study involving an older age (16.8 years) found a similar pattern, with d = 1.03 for same-sex aggression and –0.95 for oppositesex encounters (Hilton, Harris, & Rice, 2000). Gergen (1990) and Harris (1992) found similar results using self-reports at 19 years of age.
SEX DIFFERENCES IN VERBAL AGGRESSION The evolutionary origins of verbal aggression are seen in the vocal threat displays of animals. Some aspects of human aggression are directly comparable, such as shouting in a loud and intimidating way at someone, accompanied by threatening gestures. These forms are found as early as children show social interactions, and accompany physical aggression. Other types of verbal aggression are uniquely human, for example, making verbal threats that specify intended violent actions, such as “I’ll smash your head in.” Insults are also uniquely human and generally involve remarks or accusations that challenge important aspects of a person’s identity, denigrating the person in some way. For example, a teenage boy may be called names that imply that he is cowardly, such as “chicken” or “yellow,” whereas a teenage girl may be called names that imply that she is sexually promiscuous, such as “slag” or “slut” (Campbell, 1995). These insults challenge characteristics that are especially valued in the particular sex. Traditionally, insults were viewed as attacks on a person’s honor or reputation and, in circumstances where there was no effective rule of law, had to be avenged to prevent loss of face and status for both the person and his or her family (Nisbett & Cohen, 1996; Ruff, 2001). Thus, insults such as “son of a bitch” or “bastard,” which have largely lost their potency in modern society, were once regarded as denigrating statements about a person’s whole family. More subtle forms of verbal aggression develop later in childhood, for example, making someone the butt of remarks that are denigrating to that person, but humorous to others. Being able to use language in this way is particularly important in groups of teenage boys and young men (Benson & Archer, 2002). There is therefore considerable diversity within the category “verbal aggression.” Yet verbal aggression by both sexes has generally been considered as a
Sex Differences in Aggressive Behavior
437
single overall category, so that the use of particular forms of verbal aggression by the two sexes has seldom been studied. Meta-analytic summaries of sex differences in aggression can only provide a broad view of changes with age. The sex difference for self-reported verbal aggression is in the same direction as that for physical aggression, but is generally smaller in size (Archer, 2004). The difference was smaller at 6–11 years (d = 0.19) than at 11 to 17 years (d = 0.36). Studies involving observations showed little difference across age categories for verbal aggression. However, there are individual studies showing sex differences in verbal aggression at young ages. Fagot and Hagan (1985) found differences in the male direction of d = 0.31 between 2 and 3 years of life. McGrew (1972) and Sears et al. (1965) found larger differences (d = 0.53 and 1.05) at 4 years of age. As in the case of physical aggression, these sex differences are restricted to same-sex encounters, when the sex of the opponent is considered (Barrett, 1979; Tucker, 1989; Hilton et al., 2000).
THE DEVELOPMENT OF SEX DIFFERENCES IN INDIRECT AGGRESSION Indirect aggression refers to covert forms of aggression, as opposed to direct or face-to-face forms. It may be physical, such as in destroying someone’s property, or verbal, through deliberately attacking a person’s social standing or via malicious gossip or social ostracism (Archer, 2001). Most studies have concentrated on verbal forms, such as getting others to dislike a person, becoming friends with another as a form of revenge, telling a person’s secrets to another, saying bad things about people behind their backs, and telling others to avoid someone. Lagerspetz, Björkqvist, and Peltonen (1988) were the first to carry out a systematic study of indirect aggression using peer ratings, which are particularly suited to a study of this form of covert aggression. They documented large sex differences in the female direction, which were found in later studies (Björkqvist, Lagerspetz, & Kaukiainen, 1992a). Sex differences have also been found in a very similar form of aggression, termed relational (Crick & Grotpeter, 1995). It is a matter of dispute whether relational aggression is the same as indirect aggression (Björkqvist, 2001) or closely overlaps with it (Archer, 2001). Björkqvist, Osterman, and Kaukiainen (1992b) proposed that the expression of aggression changes over the life course. Normatively, as children mature, aggression changes from mainly physical to verbal, and then from verbal to indirect. In their cross-sectional analysis, levels of particular forms of indirect aggression, such as gossip, were much lower in both sexes at 8 years of age, but were higher at 11, 15, and 18 years, and sex differences were particularly pronounced at these three older ages. A meta-analysis of studies using peer reports to investigate indirect aggression (Archer, 2004) showed that effect sizes increased with age over the range of 6–17 years, so that by the teenage years they were substantial. This coincides with the larger sex differences
438
DETERMINANTS OF AGGRESSION
found from 11 years onward in Björkqvist’s analysis, and with evidence from the NLSCY, presented in the following discussion. Vaillancourt (Chapter 8, this volume) presents in detail theories of the development of indirect aggression, as well as developmental trajectories (for ages 4 and 11 years) estimated from the NLSCY (see also Vaillancourt et al., 2004). Briefly, the results indicate that the development of indirect aggression follows a pattern opposite to that for physical aggression in the same sample: While physical aggression decreases during childhood, indirect aggression increases. The development of physical and indirect aggression found in the NLSCY support the notion that boys and girls use different types of aggressive behavior over the course of childhood. Indeed, the results show not only a stronger tendency for girls to cease using physical aggression, they also show a stronger tendency for girls to increasingly use indirect aggression over time (i.e., girls are more likely than boys to follow rising trajectories of indirect aggression). It is noteworthy that the NLSCY shows that both sexes use both forms of aggression, physical and indirect. Early in childhood (from age 2), girls already use less physical aggression and more indirect (or relational) aggression than boys do. These sex differences become even more pronounced over time, with boys being less likely to exhibit declining levels of physical aggression prior to kindergarten, and girls being more likely to follow rising trajectories of indirect and relational aggression during middle childhood. Even so, boys and girls were represented on all types of trajectories, so that we again have to caution that these generalizations hide considerable individual variation. For instance, a substantial proportion of girls (33.1%) followed a high and mostly stable trajectory of physical aggression, and a substantial proportion of boys (40.3%) followed a high and rising trajectory of indirect aggression. Thus, the sex differences were related to the proportion of boys and girls on trajectories reflecting high levels or changing levels of physical and indirect aggression. These longitudinal data broadly agree with the cross-sectional analysis of Björkqvist and his colleagues, and with the meta-analysis of age differences. An analysis of the same data found distinct factors for physical and indirect aggression, which were stable over time (Vaillancourt, Brendgen, Boivin, & Tremblay, 2003). Sex differences at younger ages (4 to 7 years) were in the female direction, but small in size, across four cohorts. Generally, they became larger from 8 to 11 years, –0.26 and –0.28 being the largest values.3
THE EVOLUTIONARY SIGNIFICANCE OF SEX DIFFERENCES IN INDIRECT AGGRESSION Campbell (1999) proposed that females tend to use forms of aggression that carry low risks for physical injury, such as indirect aggression (see also Björkqvist, 1994). She related this, in evolutionary terms, to the greater value
Sex Differences in Aggressive Behavior
439
attached by women to protecting their own lives, because the survival of their infants is more dependent on the mothers’ care and defense than the fathers’. In this context, the cost of direct and physical aggression, the forms that represent risk for one’s physical safety, is greater for women than men. The greater use of indirect and relational aggression among girls, and of physical aggression among boys, is also relevant for competition with peers in the social domain. In an earlier section, we referred to the importance of same-sex peer groups during childhood, and the values adopted in these groups, for explaining sex differences in physical aggression. This can be extended to an understanding of indirect aggression. Specifically, Maccoby (1998) and Crick and Grotpeter (1995) suggested that the structure of female friendships, which tend to be based on intimacy and cohesiveness, facilitates the use of indirect and relational forms of aggression among girls. A cohesive social network enables this form of covert aggression to be used more effectively (Green, Richardson, & Lago, 1996).
CONCLUSIONS In this chapter we have sought to place the development of human sex differences in aggression into an evolutionary context. The evolution of aggression can be understood in terms of its origins in the natural world, and its benefits and costs in particular situations. Sex differences in aggression can be understood in similar terms, by considering the imbalance in parental investment in the two sexes. Aggression begins early in postnatal life in many animals, and data from small-scale observational studies, and from a large-scale longitudinal study, indicate that this is the case for humans. Such aggression consists largely of physical forms, with accompanying vocalizations. Higher levels are found in boys than girls from an early age, and these encounters arise from both property-related and dominance disputes. Following the trajectories of aggression in a large representative sample showed a gradual decline in physical aggression overall, suggesting that rather than having to be socialized into being aggressive, children gradually learn to inhibit their direct forms of aggression. More boys than girls were found on the developmental trajectories containing the highest rates of physical aggression, and more girls than boys were found on the lowest trajectories. Indirect or relational aggression appears later in development, and girls were more likely than boys to follow rising trajectories in this form of aggression, coinciding with findings that the sex difference in indirect aggression is particularly pronounced in the teenage years. These contrasting developmental pathways in girls’ and boys’ aggression are not absolute differences, but show an overlap between the sexes. Nevertheless, they can be understood in terms of the different forms of social relations that are typical of boys’ and girls’ social groups, as well as in terms of the different selection pressures on males and females resulting from sexual selection.
440
DETERMINANTS OF AGGRESSION
NOTES 1. All these effect sizes were calculated from the published data using DSTAT (Johnson, 1989). 2. Effect sizes were calculated by the first author, using D-STAT (Johnson, 1989). 3. Values calculated from the authors’ data.
REFERENCES Archer, J. (1988). The behavioural biology of aggression. Cambridge, UK: Cambridge University Press. Archer, J. (1992a). Ethology and human development. Savage, MD: Barnes & Noble. Archer, J. (1992b). Childhood gender roles: Social context and organization. In H. McGurk (Ed.), Childhood social development: Contemporary perspectives (pp. 31–61). Hove, UK, & Hillsdale, NJ: Erlbaum. Archer, J. (1994). Violence between men. In J.Archer (Ed.), Male violence (pp. 121–140). London and New York: Routledge. Archer, J. (2001). A strategic approach to aggression. Social Development, 10, 267–271. Archer, J. (2004). Sex differences in aggression in real-world settings: A meta-analytic review. Review of General Psychology, 8, 291–322. Archer, J., & Lloyd, B. B. (2002). Sex and gender (2nd ed.). Cambridge, UK: Cambridge University Press. Archer, J., & Mehdikhani, M. (2003). Variability among males in sexually-selected attributes. Review of General Psychology, 7, 219–236. Baillargeon, R. (2002). Gender differences in physical aggression at 17 months of age: Results from the longitudinal study of child development in Quebec. In S. Côté (Chair), Sex differences and similarities in aggression: Correlates and consequences from infancy to mid-life. Symposium conducted at the 15th world meeting of the International Society for Research on Aggression, Montreal, Canada. Bandura, A. (1973). Aggression: A social learning analysis. Englewood Cliffs, NJ: Prentice-Hall. Barrett, D. E. (1979). A naturalistic study of sex differences in children’s aggression. Merrill–Palmer Quarterly, 25, 193–203. Beeman, A. E. (1947). The effect of the male hormone on aggressive behavior in mice. Physiological Zoology, 20, 373–405. Bekoff, M. (1981). Development of agonistic behaviour: Ethological and ecological aspects. In P. F. Brain & D. Benton (Eds.), Multidisciplinary approaches to aggression research (pp. 161–178). Amsterdam: Elsevier/North Holland. Benson, D. A., & Archer, J. (2002). An ethnographic study of sources of conflict between young men in the context of the night out. Psychology, Evolution and Gender, 4, 3– 30. Bjorklund, D. J., & Pellegrini, A. D. (2002). The origins of human nature: Evolutionary developmental psychology. Washington, DC: APA. Björkqvist, K. (1994). Sex differences in physical, verbal and indirect aggression: A review of recent research. Sex Roles, 30, 177–188. Björkqvist, K. (2001). Different names, same issue. Social Development, 10, 272–275. Björkqvist, K., Lagerspetz, K. M. J., & Kaukiainen, A. (1992a). Do girls manipulate and boys fight? Developmental trends in regard to direct and indirect aggression. Aggressive Behavior, 18, 117–127.
Sex Differences in Aggressive Behavior
441
Björkqvist, K., Osterman, K., & Kaukiainen, A. (1992b). The development of direct and indirect aggressive strategies in males and females. In K. Björkqvist & P. Niemela (Eds.), Of mice and women: Aspects of female aggression (pp. 51–64). San Diego, CA: Academic Press. Broidy, L. M., Nagin, D. S., Tremblay, R. E., Bates, J. E., Brame, B., Dodge, K., Fergusson, D., Horwood, J., Loeber, R., Laird, R., Lynam, D., Moffitt, T. E., Pettit, G. S., & Vitaro, F. (2003). Developmental trajectories of childhood disruptive behaviors and adolescent delinquency: A six site, cross-national study. Developmental Psychology, 39, 222–245. Cairns, R. B., & Cairns, B. D. (1994). Lifelines and risks: Pathways of youth in our time. New York and London: Harvester Wheatsheaf. Cairns, R. B., Cairns, B. D., Neckerman, H. J., Ferguson, L. L, & Gariepy, J. L. (1989). Growth and aggression: I. Childhood to early adolescence. Developmental Psychology, 25, 320–330. Campbell, A. (1995). A few good men: Evolutionary psychology and female adolescent aggression. Ethology and Sociobiology, 16, 99–123. Campbell, A. (1999). Staying alive: Evolution, culture, and women’s intrasexual aggression. Behavioral and Brain Sciences, 2, 203–252. Campbell, A., Shirley, L., & Caygill, L. (2002). Sex-typed preferences in three domains: Do two-year-olds need cognitive variables? British Journal of Psychology, 93, 203– 217. Caplan, M., Vespo, J., Pedersen, J., & Hay, D. F. (1991). Conflict and its resolution in small groups of one- and two-year-olds. Child Development, 62, 1513– 1524. Courtwright, D. T. (1996). Violent land: Single men and social disorder from the frontier to the inner city. Cambridge, MA: Harvard University Press. Crick, N. R., & Grotpeter, J. K. (1995) Relational aggression, gender, and social-psychological adjustment. Child Development, 66, 710–722. Cummings, E. M., Iannotti, R. J., & Zahn-Waxler, C. (1989). Aggression between peers in early childhood: Individual continuity and developmental change. Child Development, 60, 887–895. Daly, M., & Wilson, M. (1990). Killing the competition: Female/female and male/male homicide. Human Nature, 1, 81–107. Darwin, C. (1871). The descent of man and selection in relation to sex. London: Murray. Edwards, D. A. (1969). Early androgen stimulation and aggressive behavior in male and female mice. Physiology and Behavior, 40, 333–338. Fagot, B. I., & Hagan, R. (1985). Aggression in toddlers: Responses to the assertive acts of boys and girls. Sex Roles, 12, 341–351. Frank, L. G. (1997). Evolution of genital masculinization: Why do female hyaenas have such a large “penis”? Trends in Ecology and Evolution, 12, 58–62. Geist, V. (1978). On weapons, combat and ecology. In L. Krames, P. Pliner, & T. Alloway (Eds.), Advances in the study of communication and affect: Vol. 4. Aggression, dominance and individual spacing (pp. 1–30). New York: Plenum Press. Gergen, M. (1990). Beyond the evil empire: Horseplay and aggression. Aggressive Behavior 16, 381–398. Goodenough, F. L. (1931). Anger in young children. Westport, CT: Greenwood Press. Green, L. R., Richardson, D. R., & Lago, T. (1996). How do friendship, indirect, and direct aggression relate? Aggressive Behavior, 22, 81–86. Gunnar, M. R., Senior, K., & Hartup, W. W. (1984). Peer presence and the exploratory
442
DETERMINANTS OF AGGRESSION
behavior of eighteen- and thirty-month-old children. Child Development, 55, 1103–1109. Halpern, C. T., Udry, J. R., Campbell, B., and Suchindran, C. (1994). Relationships between aggression and pubertal increases in testosterone: A panel analysis of adolescent males. Social Biology, 40, 8–24. Harris, M. B. (1992). Sex and ethnic differences in past aggressive behaviors. Journal of Family Violence, 7, 85–102. Hay, D. F., Castle, J., & Davies, L. (2000). Toddlers’ use of force against familiar peers: A precursor of serious aggression? Child Development, 71, 457–467. Hilton, Z., Harris, G. T., & Rice, M. E. (2000). The functions of aggression by male teenagers. Journal of Personality and Social Psychology, 79, 988–994. Jacklin, C. N., & Maccoby, E. E. (1978). Social behavior at 33 months in same-sex and mixed-sex dyads. Child Development, 49, 557–569. Johnson, B. T. (1989). Software for the meta-analytic review of research literatures. Hillsdale, NJ: Erlbaum. Johnson, J. G., Cohen, P., Smailes, E. M., Kasen, S., & Brook, J. S. (2002) Television viewing and aggressive behavior during adolescence and adulthood. Science, 295, 2468–2471. Keenan, K., & Shaw, D. (1997). Developmental and social influences on young girls’ early problem behavior. Psychological Bulletin, 121, 95–113. Kingston, L., & Prior, M. (1995). The development of patterns of stable, transient, and school-age onset aggressive behavior in young children. Journal of American Academy of Child and Adolescent Psychiatry, 34, 348–358. Kochanska, G., Murray, K. T., & Harlan, E. T. (2000). Effortful control in early childhood: Continuity and change, antecedents, and implications for social development. Development and Psychopathology, 36, 220–232. Lagerspetz, K. M. J., Björkqvist, K., Peltonen, T. (1988). Is indirect aggression typical of females? Gender differences in aggressiveness in 11 to 12–year-old children. Aggressive Behavior, 14, 403–414. Le Boeuf, B. J. (1974). Male-male competition and reproductive success in elephant seals. American Zoologist, 14, 163–176. Loeber, R., & Stouthamer-Loeber, M. (1998). Development of juvenile aggression and violence. Some common misconceptions and controversies. American Psychologist, 53, 242–259. Maccoby E. E. (1998). The two sexes: Growing up apart, coming together. Cambridge, MA: Belknap Press/Harvard University Press. McGrew, W. C. (1972). An ethological study of children’s behavior. New York: Academic Press. Nagin, D. S., & Tremblay, R. E. (1999). Trajectories of boys’ physical aggression, opposition, and hyperactivity on the path to physically violent and non violent juvenile delinquency. Child Development, 70, 1181–1196. Nisbett, R. E., & Cohen, D. (1996). Culture of honor: The psychology of violence in the South. Boulder, CO: Westview Press. Patterson, G. R., Dishion, T. J., & Bank, L. (1984). Family interaction: A process model of deviancy training. Aggressive Behavior, 10, 253–267. Pellegrini, A. D., & Long, J. D. (2002). A longitudinal study of bullying, dominance, and victimization during the transition from primary school through secondary school. British Journal of Developmental Psychology, 20, 259–280. Plavcan, J. M., & Van Schhaik, C. P. (1997). Intrasexual competition and body weight di-
Sex Differences in Aggressive Behavior
443
morphism in anthropoid primates. American Journal of Physical Anthropology, 103, 27–68. Quetelet, A. (1984). Research on the propensity for crime at different ages. (S. F. Sylvester, Trans.). Cincinnati, OH: Anderson. (Original work published 1833) Reiss, A. J., & Roth, J. A. (1993). Understanding and preventing violence. Washington, DC: National Academy Press. Ruff, J. R. (2001). Violence in early modern Europe 1500–1800. Cambridge, UK: Cambridge University Press. Sears, R. R., Rau, L., & Alpert, R. (1965). Identification and childrearing. Stanford, CA: Stanford University Press. Shaw, D. S., Keenan, K., & Vondra, J. I. (1994). Developmental precursors of externalizing behavior: Ages 1 to 3. Developmental Psychology, 30, 355–364. Stamps, J. (2003). Behavioural processes affecting development: Tinbergen’s fourth question comes of age. Animal Behaviour, 66, 1–13. Stinson, C. H. (1979). On the selective advantage of fratricide in raptors. Evolution, 33, 1219–1225. Thornberry, T. P. (1998). Membership in youth gangs and involvement in serious and violent offending. In R. Loeber & D. P. Farrington (Eds.), Serious and violent juvenile offenders: Risk factors and successful interventions (pp. 147–166). Thousand Oaks, CA: Sage. Tieger, T. (1980). On the biological basis of sex differences. Child Development, 51, 943– 963. Tinbergen, N. (1963). On the aims and methods of ethology. Zeitschrift fur Tierpsychologie, 20, 410–413. Tremblay, R. E., Japel, C., Perusse, D., Boivin, M., Zoccolillo, M., Montplaisir, J., & McDuff, P. (1999). The search for the age of “onset” of physical aggression: Rousseau and Bandura revisted. Criminal Behavior and Mental Health, 9, 8–23. Tremblay, R. E., Loeber, R., Gagnon, C., Charlebois, P., Larivée, S., & LeBlanc, M. (1991). Disruptive boys with stable and unstable high fighting behavior patterns during junior elementary school. Journal of Abnormal Child Psychology, 19, 285– 300. Trivers, R. (1972). Parental investment and sexual selection. In B. B. Campbell (Ed.), Sexual selection and the descent of man (pp. 136–179). Chicago: Aldine. Tucker, M. L. (1989). Anger and aggression in early adolescence: The influence of hormones and social context. Dissertation Abstracts International, 50(2), 761B. (UMI No. 8821665) Vaillancourt, T., Brendgen, M., Boivin, M., & Tremblay, R. E. (2003). Longitudinal confirmatory factor analysis of indirect and physical aggression: Evidence of two factors over time? Child Development, 74, 1628–1638. Vaillancourt, T., Côté, S., Farhat, A., Leblanc, J. C., Boivin, M., & Tremblay, R. E. (2004). The development of indirect aggression among Canadian children. Manuscript submitted for publication. Weisfeld, G. (1994). Aggression and dominance in the social world of boys and girls. In J. Archer (Ed.), Male violence (pp. 42–69). London and New York: Routledge.
This page intentionally left blank
Part IV CHALLENGES FOR THE FUTURE
This page intentionally left blank
CHALLENGES Where Are We Going? FOR THE FUTURE
21 The Developmental Origins of Aggression Where Are We Going? R ICHARD E. T REMBLAY and S YLVANA C ÔTÉ
The aim of this book was to review the state of knowledge on the developmental origins of aggression. Contributions were divided in two categories: Chapters 2–10 describe different types of aggressive behavior and the changes that occur as individuals increase in age; Chapters 11–20 review factors often considered to be proximal and distal determinants of aggressive behavior. Like all attempts to carve nature at its joints, separating the development of aggressive behavior from putative determinants is a perilous enterprise, especially when nature has not yet been precisely described. The aim of this final chapter is to summarize where we come from, where we stand, and where we appear to be going with reference to selected key issues. As we planned this book, it became clear that we were following a path traced by another book published 30 years earlier (De Wit & Hartup, 1974). In Chapter 1 of this volume, Willard W. Hartup, an editor of the previous book, highlights three important changes that have occurred over these past three decades. First, investigators shifted their attention from the aggressive act to the aggressive individual. Second, the research became truly developmental; that is, longitudinal data were collected, enabling investigators to look at changes over time within individuals. Third, investigators increasingly aggregated the concepts of aggression and antisocial behavior. In this chapter, we add four more changes that we believe are important to future research on aggression: More studies are focused on aggression during the first 3 years after birth; there is an explosion of studies on the biological bases of human aggres447
448
CHALLENGES FOR THE FUTURE
sion; we have started to systematically study gene–environment interactions; and preventive interventions are being used as experiments to test causal factors. Six of these seven changes should mark the next few decades of research, because they are essential for the unraveling of the developmental origins of aggression. However, one of the seven changes (aggregating aggression and antisocial behavior) may be a serious handicap in achieving this task.
CHANGES IN AGGRESSIVE ACTS OVER TIME WITHIN INDIVIDUALS Two of the three changes Hartup highlights in Chapter 1 go hand in hand. If we are studying aggressive behavior, the aggressive act must obviously remain the basic phenomenon of study; however, understanding development implies a focus on changes over time within individuals. Here are examples of developmental questions that target the aggressive acts of individuals over time: When, during an individual’s life course, do different types of aggressive acts start? Does the frequency of these acts change over the life course? Do the victims of these acts change during the life course? Do the determinants of these acts change during the life course? Do the consequences of these acts change during the life course? How and when, during the life course, can we change these developmental trajectories? The traditional approach to answering these questions was a crosssectional design, that is, comparing samples of individuals of different ages. In the De Wit and Hartup (1974) book there are two very good examples of this (Hapkiewicz, 1974; Hoving, LaForme, & Wallace, 1974). Note that Hapkiewicz’s chapter has a perfect developmental title: “Developmental Patterns of Aggression.” But there are many more recent examples, such as the following chapters in other books: “The Development of Direct and Indirect Aggressive Strategies in Males and Females” (Björkqvist, Österman, & Kaukiainen, 1992) and “Do Children in Canada Become More Aggressive as They Approach Adolescence?” (Tremblay et al., 1996). We now have longitudinal data to study intra-individual changes in aggression, and these studies are reviewed in the first 10 chapters of this volume. However, many, if not most of the studies on the determinants of aggression (Chapters 11–20) provide little developmental information on aggression, because the aggressive acts (the dependent variable) to be explained by one or many determinants (independent variables) tend to be assessed at only one point in time. Many longitudinal studies are used to conduct cross-sectional analyses, or predictive analyses, that is, assessing the strength of the correlation between independent variables and aggression either at the same point or at two different points in time. For example, see Chapter 11 on genetics and aggression (Pérusse & Gendreau), and count how many of the cited studies used the within-individual development of aggressive behavior as the phenotype to be explained by genes and environment. Studies on intra-individual
Where Are We Going?
449
change in playful aggression (Chapter 7, by Peterson & Flanders), indirect aggression (Chapter 8, by Vaillancourt) and proactive–reactive aggression (Chapter 9, by Vitaro & Brendgen) are either nonexistent or extremely rare. Most of the studies attempt to find causes of a phenotype measured at one point in time, and hence not the development of a behavior over time. The maintenance of a cross-sectional approach is most evident in the data available for the meta-analyses of sex differences in aggression presented in Chapter 20 by Archer and Côté. It is tempting to conclude that studying the development of a phenomenon does not come naturally to humans. This is best exemplified by the behavior of many of the young, and not so young, investigators who come to work on our longitudinal data sets. They say they are attracted by the richness of the longitudinal data that we have collected. They spend weeks, and often months, reviewing the literature, examining the data that was collected, and come up with questions such as, “Do obstetrical complications predict violent behavior in late adolescence?” “Does behavior in kindergarten predict violent behavior in early adulthood?” “Do different types of aggression in early adolescence predict equally well spouse aggression?” Each of these questions has some theoretical and practical value. Each necessitates longitudinal data that span decades. But they are not asking one of the developmental questions asked at the start of this section. The preoccupation is with prediction from time X to time Y, rather than describing the development of a phenotype and identifying the mechanisms that unfold over time. Humans are in a hurry to know—“can’t wait for 20 years to confirm a theory.” From a developmental science perspective, showing that characteristics at time X predict characteristics at time Y is as simplistic as showing that one, or an array of genes predict behavior at a given point in time. Selecting two points in time from a data bank that includes repeated assessments over long periods of time is similar to dichotomizing a continuous measurement scale; there is a huge loss of information. By the same analogy, the important developmental questions are concerned with the pattern of the distribution of the scores, not by the creation of two groups that are somewhat different. Taking a developmental perspective means, first, to understand the development (change over time) of the phenomenon we are trying to predict, and second, to describe the causal chain of events over time that explain the development. However, a premium is generally given to studies that aim at testing causal theories (that is: testing the strength of the correlation between a “determinant” and an “outcome”). Although there is a need for studies with such objectives, the need for studies that describe the development of a phenomenon over time is obviously a prerequisite for understanding the cause of the development. Those who tend to value theory testing more than descriptions should look back to the revolutions created by the simple descriptions of Copernicus, Galileo, and Mendelyev (see also Tinbergen, 1963). Galileo certainly did not find it easy to convince his colleagues that Co-
450
CHALLENGES FOR THE FUTURE
pernicus’s description of the solar planet trajectories was closer to the truth than the trajectories their theories led them to imagine, notwithstanding who or what caused the phenomenon. As the longitudinal data become available and the next generation of investigators come into their prime, the traditions should change. The research efforts aimed at explaining behavior development should be accompanied by efforts aimed at describing that development, ideally over the life course. Unfortunately, the best designed and longest running longitudinal studies with birth cohorts apparently did not obtain repeated assessments of aggressive behaviors between birth and school entry (e.g., Fergusson & Horwood, 1995; Moffitt, Caspi, Dickson, Silva, & Stanton, 1996; Power, Manor, & Fox, 1991; Raine et al., 2001). Thus, to describe the life course development of different forms of aggressive behavior, we will have to count on new birth cohorts. This means that complete life course data will become available in not less than 90 years! Someone should start soon, concentrate on collecting as much data as possible over the early years, be persistent, and be prepared to train many generations of investigators. This is the price we have to pay to make serious claims of conducting “life course” research on human development.
CLASSIFICATION: THE AGGRESSIVE AND ANTISOCIAL BEHAVIOR CASE Chapter 2 by Gendreau and Archer shows the complexity of creating a classification of aggressive behaviors, the basic building block of a science. This is not a new problem (see e.g., De Wit & Hartup, 1974), and it is a problem that all sciences have to handle with care (Appel, 1987; Burkhardt & Smith, 1988). From that perspective, we strongly believe that the tendency over the past 30 years to aggregate aggressive and antisocial behaviors (see Hartup, Chapter 1, this volume) is a serious mistake. First, from an evolutionary perspective, aggression is an adaptive behavior. It is a tool that individuals learn to control during their development. For example, playful aggression is a natural behavior of nonhuman (e.g., juvenile rats, cats, dogs, monkeys) and human animals (see Pellis, Pellis, & Foroud, Chapter 3, and Petersen & Flanders, Chapter 7, this volume). Adult humans continue to play fight through physically aggressive games such as boxing, wrestling, rugby, American football, and ice hockey. Humans also produce and consume a huge amount of fictive violence in books, plays, electronic games, and movies. Many humans learn to master different complex forms of physical aggression for defensive purposes, for example, by training in martial arts, by training in the art of boxing, or by learning to use weapons such as guns and rifles. The practice of aggregating aggression and antisocial behavior that appeared in the research over the past 30 years (see Hartup, Chapter 1, this volume; Coie & Dodge, 1998; Tremblay, 2000) probably reflects the ambivalent relationship humans have with aggression. For example, there have been strong movements in the past 30 years to prevent children from playing
Where Are We Going?
451
aggressive games (e.g., parents should not buy toy guns) and to protect children from aggressive contents in books, movies, and video games. It is argued that these activities train them to become “antisocial” (e.g., Anderson et al., 2003; Jones, 2002). Meanwhile, violent movies and violent video games have the biggest commercial success, games such as hockey are said to have become more violent than ever, thousands of young men are trained for preemptive wars, and we see, every day on television, the thousands of people violently killed during these wars. These are certainly not the most noble of human behaviors; however, they should not be aggregated with “antisocial behaviors” that are defined as law-breaking behaviors by criminologists and as mental illnesses by psychiatrists. Similarly, indirect aggressions (see Vaillancourt, Chapter 8, this volume) manifested by elementary school girls or university professors are aggressions, but they should not be confused with the “antisocial behaviors” that are of interest to criminologists, psychiatrists, and clinical psychologists. Second, if not all aggressions are antisocial, it is also true that not all antisocial behaviors are aggressions. For example, in their review of the state of knowledge on aggression, Coie and Dodge (1998) include behaviors such as substance abuse and risky sexual behavior in their definition of antisocial behavior. It seems a far stretch to label these “aggressive.” Many of the chapters in this book had to rely on studies of the general “antisocial behavior” concept, because there were not enough studies that specifically targeted aggressive behavior. One would expect that age of onset and determinants of verbal and indirect aggression trajectories would be different from onset and determinants of antisocial behaviors described as smoking marijuana (substance abuse) and not wearing a condom during sexual intercourse (a risky sexual behavior) (Coie & Dodge, 1998). We believe that instead of aggregating aggressive behaviors with antisocial behaviors, we need to go in the opposite direction if we are to understand the developmental origins of aggression. We need to disaggregate aggressive behaviors. Many studies over the past few years have targeted subcategories of aggression, such as physical aggression (see Tremblay & Nagin, Chapter 5, this volume), indirect aggression (see Vaillancourt, Chapter 8, this volume), playful aggression (see Pellis, Pellis, & Foroud, Chapter 3, and Petersen & Flanders, Chapter 7, this volume), and proactive and reactive aggression (see Vitaro & Brendgen, Chapter 9, this volume). We predict that in the long run, the “splitters,” those who disaggregate the components of aggressive behaviors, will make a more important contribution to understanding the developmental origins of aggression, than the “lumpers,” those who aggregate many forms of aggression and a large variety of antisocial behavior. The success of the physical and biological sciences has been strongly linked to the “splitters” approach. For instance, from measuring the circumference of the skull in the nineteenth century (Lombroso, 1896), investigators moved on to identifying the large components of the brain (MacLean, 1949) in the middle of the 20th century. We are now able to precisely describe the lifespan development of
452
CHALLENGES FOR THE FUTURE
every part of an individual’s brain without using a scalpel (see Paus, Chapter 12, this volume), and we are discovering the importance of prions (Si et al., 2003). The work of Eric Kandel, the senior collaborator of the last cited article, is probably one of the best examples of the scientific value of “reductionism” (Kandel, 2000). The young psychoanalyst decided that the human and nonhuman primate brains were too complex to understand. He chose to study how a sea slug’s brain (Aplysia) learns. His reductionist work over four decades revealed how memory is created in the brain of a mollusk, a rat, and a human. Without doubt, Sigmund Freud, the young biologist, would be proud of the accomplishment of his Vienna-born colleague. The development of different forms of aggressive behavior appears much less complex than the development of the human brain. We simply need more investigators who will patiently observe and describe this development.
AGGRESSION DURING EARLY CHILDHOOD Aristotle starts the second chapter of Book I of Politics with the following sentence: “He who considers things in their first growth and origin, whether a state or anything else, will obtain the clearest view of them” (Aristotle, 1943, p. 248). How obvious! It has been repeated many times since then, once by Hamburg and van Lawick-Goodall in a chapter in De Wit and Hartup (1974): “There is a great need for direct studies, utilizing systematic observational techniques, of aggressive behavior and its precursors in human infancy and early childhood” (p. 72). Out of 45 chapters in that book, only three provided some information on preschool children. There were a number of important studies on preschool children in the 1920s and 1930s (see Hay, Chapter 6, and Tremblay & Nagin, Chapter 5, this volume). There was also a wave of these studies in the 1970s and 1980s, apparently stimulated by the ethological work with primates (e.g., Blurton Jones, 1972; McGrew, 1972; Restoin et al., 1985; Smith, 1974; Strayer & Trudel, 1984), but they were cross-sectional. There are still occasional observational studies on aggression in early childhood (see Hay, Chapter 6, this volume). But the vast majority of studies on the development of aggression in humans were, and still are carried out with individuals in either elementary or secondary schools. One reason why few longitudinal preschool studies have been conducted is certainly because they are exceptionally difficult to do. The tradition has been to use direct observations—methods of data collection that are time-consuming both at the data collection and the data analysis levels, even for very small samples. Yet, there are good methodological reasons for choosing to study schoolage children: (1) They are easy to find, because they all have to go to school, (2) representative samples of populations can be easily created, and (3) reports of aggressive behaviors can be obtained from different sources—teachers, parents, peers, self-reports. However, the methodological challenges of studying preschoolers should not be a sufficient cause for their paucity. Recent epidemiologi-
Where Are We Going?
453
cal studies with very large representative cohorts of newborns have shown that the methods used with older children (i.e., parent and educator reports on children’s behavior) can provide useful information on the developmental origins of aggression during early childhood (e.g., Côté, Vaillancourt, LeBlanc, Nagin, & Tremblay, 2004; Dionne, Tremblay, Boivin, Laplante, & Pérusse, 2003; Kingston & Prior, 1995; Tremblay et al., 2004b). A second reason for neglecting early childhood aggression over the past 30 years may have been the popularity of the social learning theory (Bandura, 1973). It is probably not an exaggeration to say that most of the research on the development of human aggression in the last few decades of the 20th century was strongly inspired by the social learning paradigm. Although Bandura had studied aggressive behavior in nursery schoolchildren, most of the longitudinal studies initiated with a social learning perspective were attempting to find how schoolchildren learn to aggress from their environment. Parents, peers, movies, and television were the obvious models (e.g., Cairns & Cairns, 1994; Farrington, 2003; Huesmann & Miller, 1994; Lefkowitz, Eron, Walder, & Huesmann, 1977; Loeber, Farrington, Stouthamer-Loeber, & Van Kammen, 1998; Patterson, 1982; Thornberry, 1998; Tremblay, Vitaro, Nagin, Pagani, & Séguin, 2003). A third reason for neglecting early childhood aggression is possibly related to funding. Research priorities have been and will remain dictated by policy issues. Each time an adolescent commits a homicide, the pressure increases to understand the causes of violence during adolescence and to find ways of helping or restraining at-risk adolescents. It is not easy to convince agencies and peers that a study of infants’ aggression is the answer to understanding, and preventing, violence during adolescence. Thus, most funding of research on violence is channeled toward studies of individuals who are big enough to hurt others seriously when they aggress. The increasing evidence that control over physically violent behavior is mostly learned by humans during early childhood (see Tremblay & Nagin, Chapter 5, this volume) should help convince policymakers and funding agencies that research on early childhood aggression could help our understanding of the developmental origins of aggression and eventually prevent the violent behaviors of adolescents who scare the population. However, if there were a sudden increase in funds to study aggression during early childhood, there would be a receptor problem, inasmuch as there are very few investigators with the necessary training. Fortunately, there are examples indicating that investigators working on violence during adolescence or adulthood can become interested in studying aggression in the early years. In fact, the investigators of aggression needed for the future are individuals who focus on the development of aggression throughout the lifespan. The development of human behavior cannot be understood if it is cut into investigator-convenient age slices. The problem with direct observation, discussed earlier, will probably be solved in the near future by multidisciplinary teams that will include this timeconsuming but important approach within a global strategy. For example,
454
CHALLENGES FOR THE FUTURE
large population cohorts of children can be studied using parent ratings, while subsamples are strategically selected for intensive direct observations of social interactions, paired with key biological data collections such as saliva samples and brain scans (see Paus, Chapter 12, Pihl & Benkelfat, Chapter 13, and Van Goozen, Chapter 14, this volume).
BIOPSYCHOSOCIAL STUDIES OF AGGRESSION The great divide between social and biological approaches to human behavior is probably nowhere more evident than in the research on aggression over the past three decades. The De Wit & Hartup (1974) book showed a preoccupation with considering both sides of the question. But each article was concerned with either one or the other perspective. Numerous books on aggression and violence during the past decades omitted biological issues, and books on the biological bases of aggression did not include psychosocial factors. Most investigators preoccupied with the social “causes” find it difficult to learn the biological jargon. In fact, many resist considering the hypothesis that violence could in some ways be influenced by biological factors. For example, in the not so distant past, during the planning phase of one of the most expensive longitudinal studies of child development and antisocial behavior in the United States, plans for biological assessment apparently had to go underground (for the original plan, see Chapter 15 and Appendix II in Tonry et al., 1991). The National Consortium on Violence Research, another very large enterprise of the 1990s, funded by the U.S. National Science Foundation, was essentially focused on the social factors. Similarly, those who focused on the biological aspects of aggression generally did not take into account social factors. Considering these parallel trajectories, the changes that are occurring are impressive. The 1996 NATO Advanced Study Institute on the Biosocial Bases of Violence represented an important step in that direction (Raine, Farrington, Brennan, & Mednick, 1997). Compare the number of biological studies on humans reviewed in this book (especially in Chapter 12, by Paus; Chapter 11, by Pérusse & Gendreau; Chapter 13 by Pihl & Benkelfat; and Chapter 14, by Van Goozen) with those in the De Wit and Hartup (1974) book. The ideas were present in 1974. An example that strongly influenced our work comes from Hamburg and van Lawick-Goodall (1974): “We are inclined to believe that the hormonal changes of puberty, especially in males, may facilitate the learning of aggressive behavior” (p. 76) (see Van Goozen, Chapter 14, for tests of this hypothesis). The increasing number of biopsychosocial studies appears largely due to the development of biological measurement techniques that are relatively easy to use with humans. This is the case with molecular genetics, brain scans, and radio immunoassays of saliva. This methodological progress will continue, but biopsychosocial studies are still a small part of the numerous studies on human aggressive-antisocial behaviors.
Where Are We Going?
455
The best incentive for behavioral and social scientists to include the biological components is probably the increasingly clear demonstration that the environment, both physical and social, has an important impact on the developing organism and that, in turn, these biological consequences have a longterm impact on social behavior (Francis, Diorio, Plotsky, & Meaney, 2002; Keating & Hertzman, 1999; Meaney, 2001; Weaver et al., 2004). Smoking during pregnancy is a very good example. Children of mothers who smoked during pregnancy are at higher risk of numerous problems, including low birth weight, hypertension, hyperactivity, inattentiveness, impulsiveness, and aggression (Tremblay, Barr, & Peters, 2004a; Pausova, Paus, Sedova, & Bérubé, 2003; Tremblay et al., 2004b; Wakschlag, Pickett, Cook, Benowitz, & Leventhal, 2002). At first sight, this is the domain of biologists, however, a second look shows that women who smoke during pregnancy tend to have low education, to be poor, to have a history of problem behaviors, to have a mate who has similar problems, or to be a single parent already during pregnancy. To what extent are the effects of smoking during pregnancy on aggression explained by these social characteristics or, vice versa, the effects of the social characteristics explained by the effect of the nicotine on the developing central nervous system? There is no way of knowing if the social scientist conducts a study without measuring smoking while the biologist is doing a parallel study and does not measure the history of antisocial behavior of the parents. The best alternative for the advancement of knowledge is for the biologist and the social scientist to use the same subjects and to do analyses that take into account the biological and social variables. This is actually happening, and the teams that take these multidisciplinary approaches are flourishing. The next step is to realize that if smoking during pregnancy is injuring brain development, a biological phenomenon, the preventive solutions are mostly in the domain of the psychosocial specialists. How do we get women with a history of problem behaviors to stop smoking during pregnancy? This leads back to our comments earlier concerning the importance of early childhood in understanding the development of aggressive behavior. If early in development the social environment has an important impact on biological development, and this biological change has a long-term impact on social behavior development, it becomes still more obvious that we need to follow Aristotle’s advice: Start at the beginning if you want to understand the developmental origins of aggression.
GENE–ENVIRONMENT INTERACTIONS The gene–environment interactions (G×E) issue is but a sub-section of the previous discussion. In the De Wit & Hartup (1974) book there were two chapters on genetics. One was a summary of an impressive master’s degree thesis on the criminality of adopted children in Denmark (Hutchings, 1974). The
456
CHALLENGES FOR THE FUTURE
second was a comparative study of agonistic behavior in different strains of mice (Stewart, 1974). There were no twin studies and no molecular genetic studies. The review by Pérusse and Gendreau in Chapter 11 of this volume shows the impact of the genetic revolution over the past two decades. A new generation of social scientists can hear the word genes associated with human behavior without becoming inflamed and referring to eugenics. Many “psychosocial” investigators are recognizing that genetic information can be used as a control variable to study more adequately environmental effects. During the 1980s and 1990s, David Rowe made almost yearly presentations at the annual meeting of the American Society of Criminology, aimed at convincing criminologists that controlling for genetic factors through twin, sibling, and adoption designs was important for identifying environment effects, especially family effects, on antisocial behavior (see Chapter 15 by Rowe in Tonry et al., 1991). An important number of social scientists have moved much beyond what Rowe would have expected before his untimely death. It is obvious that most investigators with rich longitudinal data sets on human development need to collect genetic information on their subjects to add to the huge amount of data they have collected on behavior and the environment. This will lead to an important increase in G×E papers. It should eventually become next to impossible to publish a paper on human development in a top scientific journal without having a wide range of information on genes and environment. In the same way that we routinely control for socioeconomic status (SES), controlling for specific genetic polymorphism will most probably become standard practice. This will obviously not simplify our task. But it will lead to sounder knowledge, once we have learned to deal with the idea of G×E in our hypotheses and our analyses. Concerning the developmental origins of aggression, we must repeat that the first problem we need to solve is not the measurement of genes, nor the measurement of the environment, but the measurement of the phenotype. It is unlikely that there is a simple G×E leading to any expression of aggression at any point in time, or even multiple points in time, during the life course. The studies reviewed by Pérusse and Gendreau in Chapter 11 of this volume clearly used a hotchpotch of assessments at various ages, with various populations, to identify the “aggression” phenotype. We are far from a G×E science of the development of aggressive behavior in humans, but we are much closer than we were 10 years ago. Fortunately, we can count on animal studies, as can be seen in Chapter 4 of this volume, by Suomi.
TOWARD PSYCHOSOCIAL EPIGENOMICS One of the important changes of the past three decades has been the increasing awareness of the need for experimental preventive interventions (e.g., Dishion, McCord, & Poulin, 1999; Farrington, Ohlin, & Wilson, 1986;
Where Are We Going?
457
McCord, & Tremblay, 1992; Tremblay et al., 1999a; Tremblay, 2003). De Wit and Hartup (1974) did not include any experimental, or even quasi-experimental, intervention studies. Two studies made use of juvenile delinquents as subjects (Parke, 1974; Leyens & Camino, 1974), and Patterson (1974) presented an analysis of a boy referred to treatment for a high rate of “noxious” behavior. However, in their concluding chapter, De Wit and Hartup (1974) made an interesting set of comments on how children learn to regulate aggression. They suggested that two processes are involved: “a) moment-to-moment control of the ongoing actions of the child, and b) training the child for long-term selfcontrol.” These processes can be linked to research domains covered in this volume: Chapter 17 on family factors (Zoccolillo et al.) and Chapters 12 (Paus), 16 (Dionne), and 15 (Séguin & Zelazo), respectively, on brain development, language development, and development of executive functions. Research on parent training and children’s social skills training blossomed during the past three decades; however, judging from the reviews in this book, these efforts did not generate clear knowledge on the developmental origins of aggression. There appear to be four main reasons for this situation. First, children who were the targets of the interventions were generally school-age children, well past the “origins” phase. The study by Gross et al. (2003) is a notable exception. Second, rather than targeting aggression, intervention programs targeted the general concepts of aversive, externalizing, or obnoxious behavior (e.g., Patterson, 1982; Wahler, 1987; Webster-Stratton, 1998). Third, interventions showing long-term effects (more than a few years) were multimodal, precluding the attribution of the effect to a specific component (e.g., Lacourse et al., 2002; Olds et al., 1998). Fourth, the vast majority of these studies were not designed to test developmental hypotheses concerning aggression and did not control for most of the other potential causal effects. Chapter 17, by Zoccolillo et al., in this book describes the many competing explanations for a trajectory of chronic physical aggression: for example, genetic factors, parents’ history of problem behaviors, parents’ education, smoking during pregnancy, parental responsiveness to the child, coercive parenting, abuse and neglect. Whether parenting skills are determinant still needs to be demonstrated by controlling for alternative hypotheses, including genetic and perinatal effects (see Harris, 1998; Raine, 2002; Rowe, 1994). In regard to the gene–environment interaction, Caspi et al. (2002) reported a significant G×E for violent behavior, using assessments of child maltreatment and a MAOA genetic polymorphism. This is one good indication that studies that do not take into account both genetic and environmental factors are unlikely to provide adequate information on causal factors. However, we often forget that G×E studies, even with longitudinal data, are simply correlational studies if there is no manipulation of genes or environments. Correlational studies cannot prove causal effects. Hence, G×E studies of the developmental origins of aggression with large-population birth cohorts will simply give indications of possible causal factors. The ultimate proofs of cau-
458
CHALLENGES FOR THE FUTURE
sality must come from experiments in which potential causal factors are manipulated. Most probably, we are far from genetic manipulations studies to test their effects on the developmental trajectories of human aggression. Manipulations of environments, however, are the daily business of thousands of people paid to help families and individuals with behavior problems or who are at high risk of having these problems. Governments are investing billions in these social programs, without much evidence that they are having the intended effects (e.g., McCord, 2003; Timmins, 2001; Tremblay, LeMarquand, & Vitaro, 1999b). The next decade should see a substantial increase in experimental manipulations of environments to test for G×E. This may sound far-fetched to some, but it is not much different from genotyping subjects in a longitudinal study. One of the new fields riding on the crest of the genetic revolution is pharmacogenomics. Chemists create medications that compensate for genetic deficits or take into account genetic differences. These chemical products act at the protein level, but they are environmental products and thus constitute an environmental manipulation. So is nutrition; special diets are given to children born with a multitude of genetic defects (e.g., Scriver & Kaufman, 2001). The aim of these diets is either to give the children proteins they are lacking because of the genetic defect or to prevent them from eating nutrients that could harm their development because of a genetic deficit. Such a diet may be needed throughout the individual’s life, or for only a limited time during a critical period for the growth of a physiological structure. We can generalize this approach by thinking about parenting as a “nutrient.” The results of the longitudinal birth cohort study in New Zealand reported by Caspi et al. (2002; see also, Pérusse & Gendreau, Chapter 11, this volume) indicates that boys maltreated by their parents (inadequate nutrient) are more or less at risk of becoming chronically aggressive, depending on which of two MAOA genotypes (high or low activity) they inherited. A practical implication of these results is that, because maltreated boys with the highactivity MAOA polymorphism appear “genetically” protected from chronic physical aggression, preventive interventions should specifically target boys without that protective factor. The parents of these at-risk boys should receive parent training, as early as possible, to prevent maltreatment. The Elmira study (Olds et al., 1998) has shown that a nurse home visitation program could prevent both child abuse by parents and delinquency when the children become adolescents. Such effects should be observed mainly for the males without the MAOA protective polymorphism. Thus, an experimental preventive intervention that targets quality of parenting, for boys who lack the protective genetic factor, would not only demonstrate the cost-effectiveness of parenting interventions based on genetic screening, but would also provide an experimental demonstration of the causal nature of the statistical interaction shown in correlational evidence. With psychosocial interventions, we can con-
Where Are We Going?
459
duct psycho-socio-genomic experiments, but the most exciting challenge is psycho-socio-epigenomics. There is increasing evidence that gene expression is turned on and off by environmental factors (Szyf, 2003). For example, a recent study of maternal behavior with rat pups shows that increased licking and grooming, as well as arched-back nursing altered rat pups’ DNA methylation (Weaver et at, 2004). These results suggest that environment manipulation not only can compensate for genetic defects but can also turn on or off gene expression at different points in time during the developmental process. In the not too distant future we should see numerous preventive experiments that will collect genetic information to test the cost-effectiveness of the intervention and, at the same time, test G×E hypotheses. However, we can immediately go back to subjects in successful interventions, such as the Elmira study, and genotype them to verify whether indeed the subjects who most benefited from that intervention had the low-activity MAOA genotype or any other genetic risk. Unfortunately, such analyses would probably not provide useful information on the developmental origin of early aggression, inasmuch as the published work from the Elmira study indicates that early aggression data were not collected. This leads us to underscore the fact that prevention experiments with a G×E design need to specifically target the period from pregnancy to school entry, if we are to understand the developmental origins of aggression. These early studies will also be extremely useful to test the biological embedding hypothesis (Keating & Hertzman, 1999) by studying the mediating role of endophenotypes such as brain development (see Paus, Chapter 12, this volume) and HPA activity (see Van Goozen, Chapter 14, this volume). Similar studies can also be undertaken to test other putative environmental causes of chronic aggression, such as smoking during pregnancy (see Wakschlag et al., 2002), influences of peers (see Boivin, Vitaro, & Poulin, Chapter 18, this volume), and television (see Johnson, Cohen, Smailes, Kasen, & Brook, 2002). In fact, we should be testing within the same studies the differential impact of parents, peers, and fictive aggression (on television, in movies, and in video games).
CONCLUSION Planning what is needed to take us from where we are to where we would like to be forces us, first, to take stock of what has been done; second, to identify what has not been done; third, to figure out what else needs to be done; and finally, to deal with the whirl of new ideas stimulated by the planning process. As we glance at this feast of new research within our reach, we can imagine how excited a clone of Charley Darwin would be to contribute to the next 50 years of research on the developmental origins of aggression.
460
CHALLENGES FOR THE FUTURE 12 August 1849 To M. A. T. Whitby My dear Madam, I cannot express too strongly my thanks for the extraordinary trouble which you have taken in the interesting experiment, of which you send me the results.—I had given up all hopes of knowing whether peculiarities in the caterpillar state were hereditary, but now the point is amply proved: there is indeed a wide difference between a probability, however high & such an experiment as you have made.— . . . I dare not do more than hint my curiosity to know whether the Frales would prove hereditary,—i.e., whether it would be possible to make a breed with cocoons destitute of silk.—In the eyes of all silk-growers, this assuredly would appear the most useless of experiments ever tryed.— (Darwin, 1832/1985, p. 248)
REFERENCES Anderson, C. A., Berkowitz, L., Donnerstein, E., Huesmann, L. R., Johnson, J. D., Linz, D., Malamuth, N. M., & Wartella, E. (2003). The influence of media violence on youth. Psychological Science in the Public Interest, 4(3), 81–110. Appel, T. (1987). The Cuvier–Geoffroy debate: French biology in the decades before Darwin. New York: Oxford University Press. Aristotle. (1943). On man and the universe. New York: Walter J. Black-Roslyn. Bandura, A. (1973). Aggression: A social learning analysis. New York: Holt. Björkqvist, K., Österman, K., & Kaukiainen, A. (1992). The development of direct and indirect aggressive strategies in males and females. In K. Björkqvist & P. Niemelä (Eds.), Of mice and women: Aspects of female aggression (pp. 51–64). Toronto: Academic Press. Blurton Jones, N. (1972). Characteristics of ethological studies of human behaviour. In N. Blurton Jones (Ed.), Ethological studies of child behaviour (pp. 3–33). New York: Cambridge University Press. Cairns, R. B., & Cairns, B. D. (1994). Life lines and risks: Pathways of youth in our time. New York: Cambridge University Press. Caspi, A., McClay, J., Moffitt, T., Mill, J., Martin, J., Craig, I. W., Taylor, A., & Poulton, R. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297, 851–854. Coie, J. D., & Dodge, K. A. (1998). Aggression and antisocial behavior. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology: Social, emotional, and personality development. (Vol. 3, pp. 779–862). Toronto: Wiley. Côté, S., Vaillancourt, T., LeBlanc, J., Nagin, D. S., & Tremblay, R. E. (2004). The development of physical aggression during childhood: A nation wide longitudinal study of Canadian children. Manuscript submitted for publication. Darwin, C. (1832/1985). The correspondence of Charles Darwin, Vol. 4. London: Cambridge University Press. De Wit, J., & Hartup, W. W. (Eds.). (1974). Determinants and origins of aggressive behavior. The Hague, Netherlands: Mouton.
Where Are We Going?
461
Dionne, G., Tremblay, R. E., Boivin, M., Laplante, D., & Pérusse, D. (2003). Physical aggression and expressive vocabulary in 19-month-old twins. Developmental Psychology, 39(2), 261–273. Dishion, T. J., McCord, J., & Poulin, F. (1999). Iatrogenic effects in early adolescent interventions that aggregate peers. American Psychologist, 54(9), 755–764. Farrington, D. (2003). Key results from the first forty years of the Cambridge study in delinquent development. In T. P. Thornberry & M. D. Krohn (Eds.), Taking stock of delinquency: An overview of findings from contemporary longitudinal studies (pp. 137–183). New York: Kluwer Academic/Plenum Press. Farrington, D. P., Ohlin, L. E., & Wilson, J. Q. (1986). Understanding and controlling crime: Towards a new research strategy. New York: Springer-Verlag. Fergusson, D. M., & Horwood, L. J. (1995). Early disruptive behavior, IQ, and later school achievement and delinquent behavior. Journal of Abnormal Child Psychology, 23(2), 183–199. Francis, D. D., Diorio, J., Plotsky, P. M., & Meaney, M. J. (2002). Environmental enrichment reverses the effects of maternal separation on stress reactivity. Journal of Neuroscience, 22, 7840–7843. Gross, D., Fogg, L., Webster-Stratton, C., Garvey, C., Julion, W., & Grady, J. (2003). Parent training of toddlers in day care in low-income urban communities. Journal of Consulting and Clinical Psychology, 71(2), 261–278. Hamburg, D. A., & van Lawick-Goodall, J. (1974). Factors facilitating development of aggressive behavior in chimpanzees and humans. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 59–86). The Hague, Netherlands: Mouton. Hapkiewicz, W. G. (1974). Developmental patterns of aggression. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 201–207). The Hague, Netherlands: Mouton. Harris, J. R. (1998). The nurture assumption: Why children turn out the way they do. New York: Free Press. Hoving, K. L., LaForme, G. L., & Wallace, J. R. (1974). The development of children’s interpersonal aggression in a competitive task. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 193–200). The Hague, Netherlands: Mouton. Huesmann, L. R., & Miller, L. S. (1994). Long-term effects of repeated exposure to media violence in childhood. In L. R. Huesmann (Ed.), Aggressive behavior: Current perspectives (pp. 153–186). New York: Plenum Press. Hutchings, B. (1974). Genetic factors in criminality. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 255–266). The Hague, Netherlands: Mouton. Johnson, J. G., Cohen, P., Smailes, E. M., Kasen, S., & Brook, J. S. (2002). Television viewing and aggressive behavior during adolescence and adulthood. Science, 295, 2468–2471. Jones, G. (2002). Killing monsters: Why children need fantasy, super heroes, and makebelieve violence. New York: Basic Books. Kandel, E. R. (2000, December 8). The molecular biology of memory storage: A dialog between genes and synapses. Nobel lecture. Stockholm: Karolinska Instituet. Keating, D. P., & Hertzman, C. (1999). Developmental health and the wealth of nations. New York: Guilford Press. Kingston, L., & Prior, M. (1995). The development of patterns of stable, transient, and
462
CHALLENGES FOR THE FUTURE
school-age onset aggressive behavior in young children. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 348–358. Lacourse, E., Côté, S., Nagin, D. S., Vitaro, F., Brendgen, M., & Tremblay, R. E. (2002). A longitudinal-experimental approach to testing theories of antisocial behavior development. Development and Psychopathology, 14, 909–924. Lefkowitz, M. M., Eron, L. D., Walder, L. O., & Huesmann, L. R. (1977). Growing up to be violent. A longitudinal study of the development of aggression. New York: Pergamon Press. Leyens, J., & Camino, L. (1974). The effects of repeated exposure to film violence on aggression and social structures. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 509–516). The Hague, Netherlands: Mouton. Loeber, R., Farrington, D. P., Stouthamer-Loeber, M., & Van Kammen, W. B. (1998). Antisocial behavior and mental health problems: Explanatory factors in childhood and adolescence. Mahwah, NJ: Erlbaum. Lombroso, C. (1896). L’umo delinquente (5th ed., Vols. 1, 2, & 3). Torino: Broca. MacLean, P. (1949). Psychosomatic disease and the “visceral brain”: Recent developments bearing on the Papez theory of emotion. Psychosomatic Medicine, 11, 338– 353. McCord, J. (2003). Cures that harm: Unanticipated outcomes of crime prevention programs. Annals of the American Academy of Political and Social Science, 587, 16–30. McCord, J., & Tremblay, R. E. (Eds.). (1992). Preventing antisocial behavior from birth through adolescence. Experimental approaches. New York: Guilford Press. McGrew, W. C. (1972). An ethological study of children’s behavior. New York: Academic Press. Meaney, M. J. (2001). Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Annual Review of Neuroscience, 24, 1161–1192. Moffitt, T. E., Caspi, A., Dickson, N., Silva, P. S., & Stanton, W. (1996). Childhood-onset versus adolescent-onset antisocial conduct problems in males: Natural history from ages 3 to 18 years. Development and Psychopathology, 8(2), 399–424. Olds, D., Henderson, C. R., Cole, R., Eckenrode, J., Kitzman, H., Luckey, D., Pettitt, L., Sidora, K., Morris, P., & Powers, J. (1998). Long-term effects of nurse home visitation on children’s criminal and antisocial behavior: Fifteen-year follow-up of a randomized controlled trial. Journal of the American Medical Association, 280(14), 1238–1244. Parke, R. D. (1974). Some methodological problems and some future trends. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 499– 508). The Hague, Netherlands: Mouton. Patterson, G. R. (1974). Stimulus control in natural settings. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 391–400). The Hague, Netherlands: Mouton. Patterson, G. R. (1982). A social learning approach to family intervention: III. Coercive family process. Eugene, OR: Castalia. Pausova, Z., Paus, T., Sedova, L., & Bérubé, J. (2003). Prenatal exposure to nicotine modifies kidney weight and blood pressure in genetically susceptible rats: A case of gene–environment interaction. Kidney International, 64(3), 829–835. Power, C., Manor, O., & Fox, J. (1991). Health and class: The early years. London: Chapman & Hall.
Where Are We Going?
463
Raine, A. (2002). Annotation: The role of prefrontal deficits, low autonomic arousal, and early health factors in the development of antisocial and aggressive behavior in children. Journal of Child Psychology and Psychiatry, 43, 417–434. Raine, A., Farrington, D. P., Brennan, P., & Mednick, S. A. (Eds.). (1997). Biosocial bases of violence. New York: Plenum Press. Raine, A., Venables, P. H., Dalais, C., Mellingen, K., Reynolds, C., & Mednick, S. A. (2001). Early educational and health enrichment at age 3–5 years is associated with increased autonomic and central nervous system arousal and orienting at age 11 years: Evidence from the Mauritius Child Health Project. Psychophysiology, 38(2), 254–266. Restoin, A., Montagner, H., Rodriguez, D., Girardot, J. J., Laurent, D., Kontar, F., Ullmann, V., Casagrande, C., & Talpain, B. (1985). Chronologie des comportements de communication et profils de comportement chez le jeune enfant. In R. E. Tremblay, M. A. Provost, & F. F. Strayer (Eds.), Ethologie et développement de l’enfant (pp. 93–130). Paris: Editions Stock/Laurence Pernoud. Rowe, D. C. (1994). The limits of family influence. Genes, experience, and behavior. New York: Guilford Press. Scriver, C. R., & Kaufman, S. (2001). Hyperphenylalaninemia: Phenylalanine hydroxylase deficiency. In C. R. Scriver, A. L. Beaudet, W. S. Sly, & D. Valle (Eds.), The metabolic and molecular bases of inherited disease (pp. 1667–1724). New York: McGraw-Hill. Si, K., Giustetto, M., Etkin, A., Hsu, R., Janisiewicz, A. M., Miniaci, M. C., Kim, J. H., Zhu, H., & Kandel, E. R. (2003). A neuronal isoform of CPEB regulates local protein synthesis and stabilizes synapse-specific long-term facilitation in aplysia. Cell, 115(7), 893–904. Smith, P. K. (1974). Aggression in a preschool playgroup: Effects of varying physical resources. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 97–105). The Hague, Netherlands: Mouton. Stewart, J. M. (1974). Genetic and ontogenetic determinants of agonistic behavior in animal societies: Prototypic experiments with mice. In J. De Wit & W. W. Hartup (Eds.), Determinants and origins of aggressive behavior (pp. 267–276). The Hague, Netherlands: Mouton. Strayer, F. F., & Trudel, M. (1984). Developmental changes in the nature and function of social dominance among young children. Ethology and Sociobiology, 5, 279–295. Szyf, M. (2003). DNA methylation: Enzymology.In D. Cooper (Ed.), Nature encyclopedia of the human genome (pp. 1–8). New York: Nature Publishing Group. Thornberry, T. (1998). Membership in youth gangs and involvement in serious and violent offending. In R. Loeber & D. P. Farrington (Eds.), Serious and violent juvenile offenders: Risk factors and successful interventions (pp. 147–166). Thousand Oaks, CA: Sage. Timmins, N. (2001, August 27). Fact, not fashion in testing time for social policy: Governments are taking an interest in a pioneering think-tank’s “what works” approach to policy making. Financial Times, Comment and Analysis. Tinbergen, N. (1963). On aims and methods of ethology. Zeitschrift fur Tierpsychologie, 20, 410–433. Tonry, M., Ohlin, L. E., Farrington, D. P., Adams, K., Earls, F., Rowe, D. C., Sampson, R. J., & Tremblay, R. E. (1991). Human development and criminal behavior: New ways of advancing knowledge. New York: Springer-Verlag. Tremblay, R. E. (2000). The development of aggressive behaviour during childhood:
464
CHALLENGES FOR THE FUTURE
What have we learned in the past century? International Journal of Behavioral Development, 24(2), 129–141. Tremblay, R. E. (2003). Why socialization fails: The case of chronic physical aggression. In B. B. Lahey, T. E. Moffitt, & A. Caspi (Eds.), Causes of conduct disorder and juvenile delinquency (pp. 182–224). New York: Guilford Press. Tremblay, R. E., Barr, R., & Peters, R. de V. (Eds.). (2004a). Tobacco and pregnancy. Encyclopedia on early childhood development. Centre of Excellence for Early Childhood Development. Retrieved October 2003, from http://www.excellence-earlychildhood. ca/theme.asp?id=1&lang=EN Tremblay, R. E., Boulerice, B., Harden, P. W., McDuff, P., Pérusse, D., Pihl, R. O., & Zoccolillo, M. (1996). Do children in Canada become more aggressive as they approach adolescence? In Human Resources Development Canada and Statistics Canada (Eds.), Growing up in Canada: National Longitudinal Survey of Children and Youth (pp. 127–137). Ottawa: Statistics Canada. Tremblay, R. E., Japel, C., Pérusse, D., McDuff, P., Boivin, M., Zoccolillo, M., & Montplaisir, J. (1999a). The search for the age of “onset” of physical aggression: Rousseau and Bandura revisited. Criminal Behavior and Mental Health, 9, 8–23. Tremblay, R. E., LeMarquand, D., & Vitaro, F. (1999b). The prevention of ODD and CD. In H. C. Quay & A. E. Hogan (Eds.), Handbook of disruptive behavior disorders (pp. 525–555). New York: Kluwer Academic/Plenum Press. Tremblay, R. E., Nagin, D. S., Séguin, J. R., Zoccolillo, M., Zelazo, P., Boivin, M., Pérusse, D., & Japel, C. (2004b). Physical aggression during early childhood: Trajectories and predictors. Pediatrics 114(1), e43–e50. Tremblay, R. E., Vitaro, F., Nagin, D. S., Pagani, L., & Séguin, J. R. (2003). The Montreal Longitudinal and Experimental study: Rediscovering the power of descriptions. In T. Thornberry (Ed.), Taking stock of delinquency: An overview of findings from contemporary longitudinal studies (pp. 205–254). New York: Kluwer Academic/ Plenum Press. Wahler, R. G. (1987). Contingency management with oppositional children: Some critical teaching issues for parents. In J. Wilson & G. C. Loury (Eds.), From children to citizens: Vol. 3. Families, schools, and delinquency prevention (pp. 112–131). New York: Springer-Verlag. Wakschlag, L., Pickett, K. E., Cook, E., Benowitz, N. L., & Leventhal, B. (2002). Maternal smoking during pregnancy and severe antisocial behavior in offspring: A review. American Journal of Public Health, 92(6), 966–974. Weaver, I. C. G., Cervoni, N., Champagne, F. A., D’Alessio, A. C., Sharma, S., Shecke, J. R., Dymov, S., Szyf, M., & Meaney, M. J. (2004). Epigenetic programming by maternal behavior. Nature Neuroscience, doi: 10.1038/nn1276. Webster-Stratton, C. (1998). Preventing conduct problems in Head Start children. Journal of Consulting and Clinical Psychology, 66(5), 715–730.
Index
Index
“f” following a page number indicates a figure; “t” following a page number indicates a table Abstract play, 142–151 Abuse. See Maltreatment Academic achievement predictors of violence and, 210 proactive and reactive aggression and, 187 Active defense, 111–112 Adjustment problems peer rejection and, 380 proactive and reactive aggression and, 186–188 schools and, 407 Adolescence. See also Development age–violent crime curve and, 85–87t aggression prior to, 87–90, 89f gender differences and, 285 homicide and, 209 indirect aggression and, 160–161, 163 language development and, 338–340, 340–341 neuromodulators and, 262–263, 271– 273 overview, 98–101, 99f proactive and reactive aggression and, 185 subtyping aggression and, 38 testosterone and, 286–287 violence during, 202 Adoption studies, 231–232 Adrenocorticotropic (ACTH) hormone, 296. See also Hormones Adrogenic gonadal hormones, 284. See also Testosterone Adrogens, 288–290. See also Hormones
Adulthood, play fighting in rats during, 50–51 Affective aggression, 37–38 Affective disorders. See Mood disorders After-school care, 390–391. See also Daycare Age differences, 160–161 Age-graded theory, 209 Age–violent crime curve aggression prior to adolescence and, 87– 90, 89f overview, 85–87f Aggression aggressive competence, 125 aggressive performance, 125 anger as a cause of, 94 antisocial behavior vs. prosocial aggression, 5–6 association with language, 343–347 brain mechanisms influences on, 36–38 child characteristics and causes of, 14, 93 during childhood, 87–90, 89f, 93–98, 97f, 430f–431, 452–454 childhood-onset persistent, 353–370, 362t, 363t, 364t cognitive antecedent of, 31–33 consequences of, 31–33 defensive, 38, 111–112 defining, 5, 108–110, 353–354 development of, influences on, 38–39 development of prior to age 6, 90–93f, 92f developmental trajectory of, 7–8
465
466
Index
Aggression (continued) direct, 34 environment and causes of, 95, 96 evolution and causes of, 95 expression of, 30f, 33–34, 38–39 friends as determinant of, 15–16, 17–18 frustration–aggression model, 178–179 game, 38 genetic–developmental model and, 224– 227, 228–234, 230f genetics and causes of, 12–13, 96–97, 100 heritability of, 355–368, 362t, 363t, 364t hostile, 31–33, 109, 183–184 impulsive, 264 indirect, 34, 158–171, 383, 437–439, 451 individual differences and causes of, 96 during infancy, 107, 108–110, 124–126 instrumental, 31–33, 38, 109, 184, 381 integrative model of, 39–41f intermale, 37–38 language–aggression hypothesis, 331– 332 language development and causes of, 96, 340–341 motivational system and causes of, 94 neocortex, influences on, 37 neural circuitries, influences on, 37 nonviolent, 5–6 offensive, 37 parental, 28 parenting as determinant of, 14 peer relationships and causes of, 15, 96 personally motivated, 32–33 physical, 83–85, 308–309 pleasure-motivated, 32 predicting, 449–450 as a predictor for antisocial behavior, 5 prefrontal cortex, influences on, 37 pregnancy and causes of, 96 in preschool age group, 90–93f, 92f, 430f–431, 452–454 proactive, 40–41f, 179–197 prosocial, 5–6, 33 punishment as, 33 rage-like, 37 reactive, 36, 40–41f, 178–197, 381 regulation of, 142–143 relational, 6, 34, 159–160, 383, 437 relief-motivated, 32 risk factors for early development of, 121–124
serotonin metabolism and causes of, 96 siblings and causes of, 97 smoking and causes of, 97 social aggression, 34, 159–160 social learning hypothesis and causes of, 94 social systems and causes of, 100 socially motivated, 32–33 studies of, 250–255, 252f, 254f, 448– 450 subtypes of, 26–29, 30f, 31–42, 41f, 450–452 taxonomies of, 27–29, 39–41f teasing, 38 testosterone and causes of, 96 during toddler years, 124–126, 430f– 431, 452–454 verbal, 108, 436–437 war as, 33, 85 weaning, 28 Aggressive competence, 125 Aggressive performance, 125 Alcohol use. See also Substance use molecular genetics and, 232, 233 neuromodulators and, 262 during pregnancy, 210 in rhesus monkeys, 76 Allelic variations, 13 Alternative reproductive strategies, 428 Androgens, 285–286 Anger. See also Reactive aggression as a cause of aggression, 94 defining aggression and, 354 during infancy, 110 overview, 179, 282 peer relationships and, 381 proactive and reactive aggression and, 188–189, 195 in young children, 91–93f, 92f Antecedents of aggression integrative model and, 40 polythetic analysis and, 30f of proactive and reactive aggression, 194–196 subtyping aggression based on, 34– 39 Anthropomorphism, 26–27 Antidepressants, 263 Antisocial behavior. See also Homicide; Violence adoption studies on, 231–232 aggression as a predictor for, 5
Index cognitive skills and, 314–315 cortisol and, 292–293 daycare and, 406 developmental pathway model of, 205f– 206 evolutionary perspective of, 450–452 executive function and, 310–312 as a familial disorder, 356 friendships and, 16 future research on, 368–370 intergenerational transmission of, 355– 368, 362t, 363t, 364t language development and, 333, 346 maltreatment and, 359–360 in parents, 360–368, 362t, 363t, 364t peer relationships and, 409 prefrontal damage and, 148–149 vs. prosocial aggression, 5–6 school environment and, 407 Antisocial personality disorder, 37 Anxiety indirect aggression and, 167 language development and, 341 molecular genetics and, 233 during pregnancy, 123 proactive and reactive aggression and, 187, 188 Appearance, 167 Arousal, 268 Attachment system childhood-onset persistent aggression (CPA) and, 358–359 as a risk factor, 123–124 Attention-deficit/hyperactivity disorder (ADHD) DHEA/DHEAS and, 289–290 executive function and, 311–312, 317– 318 genetic factors and, 122 language development and, 343–344 neuromodulators and, 270 physical aggression and, 308 proactive and reactive aggression and, 187 testosterone and, 289 Attentional control, 273–275 Attribution biases pathway, 346 Authority conflict pathway, 205f–206. See also Developmental pathway Autism, abstract play and, 148 Autonomy, during infancy, 111–112 Avoidance of conflict, 116–117. See also Conflict
467
B Behavior changes in, 12–19, 35–36 evolutionary perspective of, 427– 428 executive function and, 310–312, 320– 321, 322 explanations of, 425 language development and, 333, 336– 337, 339–340 physical aggression and, 308–309 relationship to brain mechanisms, 242– 243 Behavioral control, 261 Bilateral view of socialization, 13–18. See also Socialization Biological indicators. See also Hormones contextualism and, 18–19 overview, 13, 298–299, 454–455 proactive and reactive aggression and, 188–190 Biopsychosocial research, 454–455 Birth complications, 96. See also Pregnancy Boxing stance, across species, 26–27 Boys. See also Gender differences developmental trajectory of aggression in, 7–8 in play fighting in rats, 49–50 proactive and reactive aggression and, 196 use of indirect aggression by, 161– 162 Brain mapping. See also Brain mechanisms overview, 242–243, 255–256 principles of, 243–246, 244f, 245f, 246f studies of aggression using, 250–255, 252f, 254f Brain mechanisms. See also Brain mapping; Prefrontal cortex abstract play and, 144–146 influences of on aggression, 36–38 MRI studies of brain development and, 246–250 neuromodulators and, 262–263 relationship to behavior, 242–243 smoking during pregnancy and, 455 Bullying history of, 84 peer relationships and, 381 relationships and, 17 as a research construct, 5–6
468
Index
C Cats, gamma-aminobutyric acid (GABA) and, 269 Cause of aggression during early childhood, 93–98, 97f language–aggression association and, 344–347 of proactive and reactive aggression, 194–196 Cerebral cortex, 250–255, 252f, 254f. See also Brain mechanisms Child care contagion effect in, 388 peer relationships and, 377, 390–391 as a risk factor, 124 social capital and, 405–406 Child characteristics causes of aggression and, 93 as a determinant of aggression, 14 Child development perspectives, 10–12. See also Development Childhood. See also Development; Preschool age group aggression during, 87–90, 89f, 430f– 431, 452–454 causes of aggression during, 93–98, 97f empathy during, 136 executive function and, 309–310, 312– 318 gender differences and, 432–434 hormones and, 288–290 language development and, 334–335, 338–340 morality in, 117–118 overview, 98–101, 99f risk factors and, 121–124 violence during, 202 Childhood-onset persistent aggression (CPA). See also Childhood causes of, 354–355 as a familial disorder, 355–368, 362t, 363t, 364t future research on, 368–370 overview, 353 Chimpanzees, 135–137. See also Monkeys Cingulate cortex, 251, 252. See also Brain mechanisms Cingulate motor areas, 251. See also Brain mechanisms Classification of aggression. See Subtypes of aggression Cliques, 384–387. See also Popularity Coercive control, 25
Coercive parenting, 359. See also Parenting Cognitive antecedent of aggression, 31–33 Cognitive development, 96. See also Cognitive skills; Development Cognitive skills global measures of, 313–315 during infancy, 112 neuromodulators and, 273–275 peer relationships and, 379 play fighting and, 55–57, 58 Cold executive function, 320–321, 322. See also Executive function Collective efficacy, 410–411 Competition abstract play and, 143 reproductive objectives and, 167 for resources, 435 Complex social play, 144. See also Play Conduct disorder cortisol and, 189, 293, 294 DHEA/DHEAS and, 290 executive function and, 311–312 hormones and, 296 language development and, 333, 337 lithium and, 275 physical aggression and, 308 smoking and, 357–358, 360 subtyping aggression and, 39 testosterone and, 288–289 Conflict infancy and, 108–109 normative course of, 110–118, 114f, 115f during the toddler years, 116–117 Consequences of aggression polythetic analysis and, 30f subtyping aggression based on, 31–33 Constructs in aggression literature, 4–6 Contextualism, 18–19 Cooperation lack of, 149–150 rough-and-tumble play and, 138– 142 Corpus collosum. See also Brain mechanisms MRI studies of brain development and, 247–249 overview, 251 Cortisol. See also Hormones; Hypothalamic–pituitary–adrenal (HPA) axis overview, 282–283, 291–296 proactive and reactive aggression and, 189–190
Index Covert pathway, 205f–206. See also Developmental pathway Criminal behavior age–violent crime curve and, 85–87f childhood-onset persistent aggression (CPA) and, 355 daycare and, 406 developmental trajectory of, 206–207 genetic studies on, 228–234, 230f language development and, 333, 341 maltreatment and, 359–360 neighborhoods and, 411–412 peer rejection and, 409–410 proactive and reactive aggression and, 187–188 social capital and, 413 testosterone and, 287 welfare system and, 400–402 CSF 5–HIAA interaction with environment, 74–76f parenting and, 74 in rhesus monkeys, 67–69, 69–70, 70– 72, 71f Culture proactive and reactive aggression and, 196 role of, 133–134
D Daycare contagion effect in, 388 peer relationships and, 377, 390–391 as a risk factor, 124 social capital and, 405–406 Defensive aggression development and, 38 during infancy, 111–112 Defining aggression during infancy, 108–110 problem of, 353–354 for research purposes, 5 Delinquency childhood-onset persistent aggression (CPA) and, 354 developmental aspects of, 207–209 developmental trajectory of, 206–207 friendships and, 16 genetic studies on, 228–234, 230f language development and, 333 maltreatment and, 359–360 in the parents, 403–404 peer relationships and, 409 predictors of, 209–211
469
proactive and reactive aggression and, 187–188, 195 social capital and, 414–415 Depression indirect aggression and, 167 maternal, 123–124, 140 serotonin and, 264–265 Development. See also Adolescence; Brain mechanisms; Childhood; Infancy; Language development; Preschool age group; Toddler years abstract play and, 149 aggression prior to age six, 90–93f, 92f brain, 246–250 causes of aggression and, 93–98, 97f child development vs. lifespan perspectives of, 10–12 contextualism and, 18–19 daycare and, 405–406 determinants of change in, 12–19 evolutionary perspective of, 427–428, 429–431, 430f of executive function, 309–310 friendships and, 16 gender differences and, 439 genetic factors and, 223–224, 456 of homicide, 204–207, 205f, 207–209 influences of on aggression, 38–39 neuromodulators and, 270–273 normative change and, 6–10f, 7f, 9f overview, 98–101, 99f of physical aggression, 83–85, 309 of rhesus monkeys, 65, 67–69, 68f, 77– 78f social capital and, 399, 400t, 414, 417 of social understanding, 138 studies of aggression using, 448–450 testosterone and, 284–287 welfare system and, 401–402 Developmental model. See also Development overview, 448–450 of proactive and reactive aggression, 190–193 Developmental pathway. See also Development of homicide, 204–206, 205f prevention and, 214 Developmental trajectories. See also Development of homicide, 206–207 overview, 7–8 sociometric status and, 10 violence and, 211–214t, 212f, 213f
470 Deviancy training overview, 385–386 in preschool children, 386–387 DHEA/DHEAS. See also Hormones in childhood, 289–290 future research on, 296–298 Direct aggression, 34 Dominance, 138–142 Dominance hierarchy, 135–136 Dopamine. See also Neuromodulators development and, 270–273 executive control and, 273–275 inhibition and, 261–262, 276 overview, 265–268 play fighting in rats and, 57 Dorsal premotor cortex, 251. See also Brain mechanisms Dorsolateral prefrontal cortex. See also Brain mechanisms executive control and, 273 functions of, 144–145 Drug use dopamine and, 270–271 gang involvement and, 414–415
E Ecological perspective, 384–391 Elephant seals, 236 Emotion regulation, 346 Emotional antecedents, 34–36 Emotions, 188–190, 316 Empathy, 133–138 Empiricism, vs. nativism, 12–13 Employment, maternal, 404–405. See also Daycare; Socioeconomic status Enemies, 16–17. See also Relationships Engagement, 268 Enhancement model, 414–415 Environment. See also Social learning hypothesis biological components and, 455 causes of aggression and, 95, 96 childhood-onset persistent aggression (CPA) and, 355, 356, 356–360 contextualism and, 18–19 evolutionary perspective and, 100, 428 genetic factors and, 228–234, 230f, 455–456 hormones and, 295–296 interaction with genetics, 74–76f language–aggression association and, 343–344 maltreatment and, 457–459
Index neuromodulators and, 262 norepinephrine and, 268 predictors of violence and, 210–211 proactive and reactive aggression and, 191 as a risk factor, 123–124 role of, 398 rough-and-tumble play and, 139 school, 407–408 Escalation theory overview, 208–209 prevention and, 214 Ethnicity, 196 Etiology pathways, 343–344 Evolution of aggression causes of aggression and, 95 gender differences and, 434–436, 437– 439 indirect aggression, 158–159, 165–170 overview, 84–85, 425–428, 429–431, 430f subtyping aggression and, 450–451 Executive control, 273–275 Executive function development of, 309–310 in early childhood, 312–318 hot and cold, 320–321 overview, 307–308, 321–322 problem-solving framework and, 318– 320 relationship with externalizing or antisocial disorders, 310–312 Explanatory models, 32 Exploratory activity, 139 Expression of aggression development and, 38–39 polythetic analysis and, 30f subtyping aggression based on, 33–34
F Family. See also Family relationships; Parenting childhood-onset persistent aggression (CPA) and, 355–368, 362t, 363t, 364t, 368–370 social capital and, 402–405 Family history, 185 Family relationships. See also Relationships child development vs. lifespan perspectives and, 11 developmental trajectory of aggression and, 8 as a risk factor, 123–124
Index Fantasy, 138 Fast Track prevention trial, 408. See also Prevention Fear responses to, 35–36 rough-and-tumble play and, 139 Fear system, 37–38 Females. See also Gender differences aggressive competence and, 125 developmental trajectory of aggression in, 8 in play fighting in rats, 49–50 proactive and reactive aggression and, 196 use of indirect aggression by, 161–162, 163–170 50:50 rule. See also Play fighting adult play fighting in rats and, 51 juvenile play fighting in rats and, 55– 56 overview, 48, 58 5–HIAA concentration levels interaction with environment, 74–76f parenting and, 74 in rhesus monkeys, 67–69, 69–70, 70– 72, 71f 5–HT. See also Neuromodulators dopamine and, 265, 267 gamma-aminobutyric acid (GABA) and, 266–267 inhibition and, 261–262, 274–275 overview, 262, 263–265 5–HTT interaction with environment, 74–76f interaction with rearing conditions, 13 fMRI, 245f, 246f, 255–256. See also Brain mapping Force during infancy, 109, 112–115f, 114f toddlers’ use of, 125 Friends. See also Peer relationships; Relationships as a determinant of aggression, 15–16, 17–18 preschool age children and, 384–387 Frontal cortex adolescence and, 272–273 brain mapping studies on, 250–255, 252f, 254f Frontal lobes, 250–255, 252f, 254f. See also Brain mechanisms Frontopolar cortex, 251. See also Brain mechanisms Fruit flies, boxing stance in, 26–27
471
Frustration–aggression model, 178–179. See also Reactive aggression Functional MRI, 245f, 246f, 255–256. See also Brain mapping Funding of research, 453
G Game aggression, 38 Game theory, 51 Gamma-aminobutyric acid (GABA). See also Neuromodulators inhibition and, 261–262, 274–275, 276 overview, 262, 269–270 serotonin and, 266–267 Gang involvement, 414–416 Gender differences aggressive competence and, 125 childhood-onset persistent aggression (CPA) and, 356 development of, 431–436 dopamine and, 267 evolutionary perspective of, 427 interaction with environment, 74–76f language development and, 339, 341, 348 neuromodulators and, 262 normative change and, 8 overview, 5, 439 peer relationships and, 379 in play fighting in rats, 49–50 preschool age children and, 383 in proactive and reactive aggression, 196 serotonin, 264–265 sexual selection and, 426–427 single-parent families and, 403 testosterone and, 283–284, 284–286 in toddlers’ use of force, 118–119, 120, 120–121 in use of indirect aggression, 161–162, 163–165, 170–171, 437–439 in verbal aggression, 436–437 Genetic–developmental model. See also Genetics of aggression in animals, 224–227 of aggression in humans, 228–234, 230f overview, 234–236 Genetics causes of aggression and, 96–97, 100 childhood-onset persistent aggression (CPA) and, 355–368, 356, 362t, 363t, 364t as a determinant of aggression, 12–13 environment and, 455–456
472
Index
Genetics (continued) hormones and, 295–296 human studies and, 228–234, 230f language–aggression association and, 343–344, 344–345 maltreatment and, 457–459 mouse studies on, 224–227 neuromodulators and, 262 overview, 223–224, 234–236 rhesus monkeys and, 69–70 as a risk factor, 122 Genotype driven approaches, 223–224 Girls. See also Gender differences aggressive competence and, 125 developmental trajectory of aggression in, 8 in play fighting in rats, 49–50 proactive and reactive aggression and, 196 use of indirect aggression by, 161–162, 163–170 Grey matter, 252–254. See also Brain mechanisms
H Harm, subtyping aggression based on, 31 Health care, 402 Heritability of aggression. See also Genetics childhood-onset persistent aggression (CPA) and, 355–368, 362t, 363t, 364t as a risk factor, 122 High Scope project, 406. See also Daycare Hippocampus, 144. See also Brain mechanisms Homicide. See also Antisocial behavior; Violence developmental aspects of, 207–209 heterogeneity of, 203–207, 205f overview, 202–203 predictors of, 209–211 prevention of, 214–215 social capital and, 413 violence trajectories and, 211–214t, 212f, 213f welfare system and, 400 Hormones. See also Cortisol; Testosterone as an explanation for violence, 86–87t future research on, 296–298 genetic factors and, 295–296 overview, 281–283, 298–299, 454–455
Hostile aggression during infancy, 109 overview, 31–33, 183–184 Hot executive function, 320–321. See also Executive function Hunger, 139 Hunting, 28 Hyenas, empathy in, 136 Hypothalamic–pituitary–adrenal (HPA) axis. See also Brain mechanisms cortisol and, 291–296 future research on, 296–298 genetic factors and, 295–296 proactive and reactive aggression and, 189–190 rough-and-tumble play and, 141
I Impulsive aggression, 264 Indirect aggression. See also Relational aggression; Social aggression age differences in, 160–161 correlates of, 162–163 gender differences in, 161–162, 437– 439 overview, 34, 158–159, 159–160, 170– 171, 451 in preschool children, 383 use of by females, 163–170 Individual differences causes of aggression and, 96 changes in aggression and, 448–450 compared to normative change, 6–10f, 7f, 9f 5–HIAA concentration levels in rhesus monkeys and, 68–69 genetic factors and, 235 homicide and, 209 play fighting in rats and, 57 in rhesus monkeys, 66–67 in toddlers’ use of force, 118–121 Infancy. See also Development aggression during, 107, 124–126 defining aggression during, 108– 110 empathy during, 136 executive function and, 310 individual differences and, 118–121 innate features of, 137–138 language development and, 334–335, 336–337 normative course of conflict and aggression in, 110–118, 114f, 115f
Index proactive and reactive aggression and, 191–192 risk factors and, 121–124 Inhibition lack of, 149 neuromodulators and, 262–263, 273– 274 overview, 261, 276 proactive and reactive aggression and, 188 Inhibition theories, 142–143 Initiative, 111–112 Injury, subtyping aggression based on, 31 Innate features of aggression, 133–138 Instrumental aggression. See also Proactive aggression development and, 38 during infancy, 109 overview, 31–33, 184 peer relationships and, 381 Instrumental force, 109 Insults. See Verbal aggression Integrative model of aggression, 39–41f Intelligence executive function and, 312, 313–314 language development and, 332, 338– 339, 341, 348 neuromodulators and, 273 Intent, subtyping aggression based on, 31– 33 Intergenerational transmission of CPA future research on, 368–370 overview, 353, 355–368, 362t, 363t, 364t Intermale aggression, 37–38 Interventions child development vs. lifespan perspectives and, 11 childhood-onset persistent aggression (CPA) and, 369 daycare and, 406 language–aggression association and, 345 need for, 456–459 peer relationships and, 392 proactive and reactive aggression and, 195–196 schools and, 408 social, 11 state dependence theory and, 209 Isolation, 52–53, 56
473
L Labeling, 209 Language–aggression hypothesis, 331–332. See also Language development Language development. See also Development causes of aggression and, 96 history of, 331–332 overview, 330–331, 347–348 peer relationships and, 378, 379 prediction of future aggression and, 340–341 in the preschool age group, 314 prevalence throughout, 333–343 theoretical models of association to aggression, 343–347 Learning, 95 Lexical development, 332–333. See also Language development Lifespan perspectives, 10–12 Limbic structures, 146 Linnoila, Markku, 63. See also Rhesus monkeys Lithium, 275
M Macaques, 56–57. See also Monkeys Macro-social pathway, 343–344 Males. See also Gender differences developmental trajectory of aggression in, 7–8 in play fighting in rats, 49–50 proactive and reactive aggression and, 196 use of indirect aggression by, 161–162 Maltreatment childhood-onset persistent aggression (CPA) and, 359–360 genetic factors and, 457–459 hormones and, 295–296 monoamine oxidase and, 13 neighborhoods and, 412 play and, 150 proactive and reactive aggression and, 186 Maternal employment, 404–405. See also Daycare Maternal responsiveness, 358–359 Maturation status, 169 Mediofrontal cortex, 252. See also Brain mechanisms Memory. See also Executive function neuromodulators and, 273–275 overview, 321
474
Index
Mesial frontal cortex, 251. See also Brain mechanisms Micro-social pathway, 345–346 Mid-dorsolateral frontal cortex, 251, 252. See also Brain mechanisms Mid-ventrolateral frontal cortex, 251, 252. See also Brain mechanisms Molecular genetics, 232–234. See also Genetics Monkeys. See also Rhesus monkeys indirect aggression among, 167 neuromodulators and, 275 overview, 135–137 rough-and-tumble play in, 56–57, 284 Monoamine oxidase genetic factors and, 232–234 maltreatment and, 360, 458–459 parental maltreatment and, 13 Mood, testosterone and, 287 Mood disorders language development and, 341 serotonin, 264–265 Moralistic aggression, 28 Morality in early childhood, 117–118 play and, 143 rough-and-tumble play and, 141–142 Motivational system abstract play and, 144–146 brain mechanism research on, 37 causes of aggression and, 94 overview, 137, 151–152 predation and, 28 subtyping aggression based on, 31–33 Mouse models. See also Rodents dopamine and, 270 gamma-aminobutyric acid (GABA) and, 270 gender differences in, 285 genetic–developmental model, 224–227 norepinephrine and, 269 MRI. See also Brain mapping functional, 245f, 246f overview, 255–256 structural, 243–244f studies of aggression using, 250–255, 252f, 254f studies of brain development, 246–250
N Name calling. See Verbal aggression Nativism, vs. empiricism, 12–13 Natural selection, 166
Neglect. See Maltreatment Neighborhoods peer relationships and, 389–390 social capital and, 410–412 Neocortex influences of on aggression, 37 integrative model and, 40 Neural circuitries abstract play and, 144–146 influences of on aggression, 37 Neuroendocrinological status, 37 Neuroimaging. See Brain mapping Neuromodulators development and, 270–273 executive control and, 273–275 inhibition and, 261–262 overview, 262–263, 276 Neurotransmitters. See also Neuromodulators dopamine, 265–268 gamma-aminobutyric acid (GABA), 269– 270 inhibition and, 261–262 norepinephrine, 268–269 serotonin, 263–265 Nonviolent aggression, 5–6 Noradrenaline, 268 Norepinephrine. See also Neuromodulators inhibition and, 261–262 overview, 268–269 Norm of reaction, 428 Normative change, 6–10f, 7f, 9f
O Object-oriented nature of aggression, 38 Offensive aggression, 37. See also Proactive aggression Ontogenetic organization comparing different levels of, 26–29 overview, 39 Oppositional–defiant disorder (ODD) cortisol and, 294 DHEA/DHEAS and, 289–290 executive function and, 312, 317– 318 hormones and, 296 physical aggression and, 308 testosterone and, 288–289 Orbitofrontal cortex, 251, 252. See also Brain mechanisms Orbitofrontal–prefrontal cortex, 144 Organized crime, 416, 417–418. See also Gang involvement
Index Ostracism, 380–383 Overt pathway, 205f–206. See also Developmental pathway
P Pain, 35–36 Parallel development model, 191 Parent–child relationship, 14–15. See also Family relationships; Parenting; Relationships Parental aggression, 28 Parenting childhood-onset persistent aggression (CPA) and, 357, 358–360, 368–369 cognitive skills and, 313–314 as a determinant of aggression, 14 gene–environment interactions and, 75– 76f hormones and, 298 interventions and, 457–459 language–aggression association and, 343–344 peer relationships and, 389 predictors of violence and, 210 proactive and reactive aggression and, 185, 191 in rhesus monkeys, 73–74 as a risk factor, 123–124 rough-and-tumble play and, 140 serotonin transporter gene (5–HTT)’s interaction with, 13 social-cognitive processes and, 382 socialization and, 14–15 use of force during infancy and, 112– 115f, 114f Parents antisocial behavior of, 360–368, 362t, 363t, 364t social capital and, 403–404 Peer rejection. See also Peer relationships during the preschool years, 380–383 social capital and, 409–410 Peer relationships. See also Friends; Relationships causes of aggression and, 96 as a determinant of aggression, 15 in early childhood, 378–380, 384–391 gene–environment interactions and, 75– 76f indirect aggression and, 162–163 during infancy, 110, 112 overview, 376–378 predictors of violence and, 210
475
during the preschool years, 380–383, 391–392 proactive and reactive aggression and, 185–186, 188, 191–192 rejection from, 380–383, 409–410 in rhesus monkeys, 73–74 as a risk factor, 124 social capital and, 408–410 during the toddler years, 116–117 use of force during infancy and, 113 Performance theory, 14. See also Socialization Perry preschool project, 406. See also Daycare Person-oriented nature of aggression, 38 Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA), 401. See also Welfare system Personality disorders language development and, 331 molecular genetics and, 232–233 proactive and reactive aggression and, 187 Personally directed force during infancy, 109 toddlers’ use of, 119 Personally motivated aggression, 32–33 Phenotype driven approaches, 224, 235– 236 Phylogenetic organization, 26–29, 40 Play. See also Play fighting; Rough-andtumble play abstract, 142–151 overview, 152, 451 social understanding and, 138 Play fighting. See also Rough-and-tumble play in adult rats, 50–51 in juvenile rats, 52–55, 54f, 55–57 overview, 47–48, 58, 138–142 in rats, 48–50, 49f Pleasure-motivated aggression, 32 Polythetic analysis, 29–30f. See also Subtypes of aggression Popularity, 163, 165, 169 Population heterogeneity theories overview, 208 prevention and, 214–215 Postpartum period, 298 Predatory aggression. See Proactive aggression Predicting aggression, 449–450
476
Index
Prefrontal cortex. See also Brain mechanisms abstract play and, 145, 146–147 brain mapping studies on, 250–255, 252f, 254f damage to, 148–149 executive control and, 273 influences of on aggression, 37 inhibition and, 276 integrative model and, 40 overview, 251 Pregnancy causes of aggression and, 96 childhood-onset persistent aggression (CPA) and, 357–358 dopamine and, 270–271 hormones and, 285–286, 297–298 predictors of violence and, 210 risk factors and, 123 smoking during, 123, 210, 455 Premotor cortex, 251. See also Brain mechanisms Prenatal malnutrition, 271. See also Pregnancy Preschool age group. See also Childhood; Development aggression during, 90–93f, 92f, 430f– 431, 452–454 executive function and, 310, 312–318 gender differences and, 433 language development and, 334–335, 336–337 peer rejection and ostracism during, 380–383 peer relationships and, 378–380, 384– 391, 391–392 Prevention. See also Interventions of homicide and violence, 214–215 need for, 456–459 schools and, 408 state dependence theory and, 209 Primary motor cortex, 251. See also Brain mechanisms Primates. See also Rhesus monkeys indirect aggression among, 167 neuromodulators and, 275 overview, 135–137 rough-and-tumble play in, 56–57, 284 Proactive aggression. See also Instrumental aggression causality of, 194–196 correlates of, 184–190 developmental model of, 190–193 gender, ethnicity, and culture and, 196
integrative model and, 40–41f measurement and validity issues regarding, 180–184 overview, 179–180, 197 Proactive use of force during infancy, 109 during the toddler years, 119 Problem solving, 195 Problem-solving framework executive function and, 318–320 overview, 321–322 Procedural morality, 141–142. See also Morality Prosocial aggression vs. antisocial aggression, 5–6 overview, 33 Protest, 111 Proximal antecedents integrative model and, 40 overview, 34–36 Psychoanalytic view, 134–135 Punishment as aggression, 33 childhood-onset persistent aggression (CPA) and, 359 physical, 84
Q Quantitative genetics, 228–232, 230f. See also Genetics
R Rage-like aggression, 37 Rage system, 37–38 Random effect growth modeling, 206– 207 Rats. See also Rodents boxing stance in, 26–27 brain mechanism research on, 37 dopamine and, 267, 270–271 empathy in, 136 neuromodulators and, 271, 275 norepinephrine and, 268 play fighting in, 48–50, 49f, 55–57, 58, 139, 142 play fighting in adults, 50–51 play fighting in juveniles, 52–55, 54f Reactive aggression causality of, 194–196 correlates of, 184–190 developmental model of, 190–193 gender, ethnicity, and culture and, 196 integrative model and, 40–41f
Index measurement and validity issues regarding, 180–184 overview, 36, 178–180, 197 peer relationships and, 381 Reactive use of force, 109 Regions, 412–414 Regulation of aggression, 142–143 Rejection, peer. See also Peer relationships during the preschool years, 380–383 social capital and, 409–410 Relational aggression. See also Indirect aggression compared to indirect aggression, 159– 160 gender differences in, 6, 437 overview, 34 in preschool children, 383 Relationships. See also Family relationships; Peer relationships in rhesus monkeys, 64–66 socialization and, 13–18 violence in, 188 Relief-motivated aggression, 32 Reproductive objectives gender differences in, 428, 435 indirect aggression and, 166–170 Repulsion hypothesis, 17 Resistance, 111 Resolution of conflict, 116–117. See also Conflict Retaliation, 111 Revenge, 150 Rhesus monkeys. See also Monkeys development of, 67–69, 68f effects of early social experiences on, 72–74 empathy in, 136 field studies of, 70–72, 71f gene–environment interactions and, 74– 76f genetic factors and, 69–70 individual differences and, 66–67 overview of research with, 77–78f rough-and-tumble play among, 139–140 social aspects of, 64–66 Risk factors childhood-onset persistent aggression (CPA) and, 355, 355–368, 362t, 363t, 364t early development of aggression, 121– 124 homicide and, 209–211 population heterogeneity theory and, 208
477
Rodents boxing stance in, 26–27 brain mechanism research on, 37 dopamine and, 267, 270, 270–271 empathy in, 136 gamma-aminobutyric acid (GABA) and, 270 gender differences in, 285 genetic–developmental model, 224–227 neuromodulators and, 271, 275 norepinephrine and, 268, 269 play fighting in, 48–50, 49f, 55–57, 58 play fighting in adults, 50–51 play fighting in juveniles, 52–55, 54f rough-and-tumble play among, 139, 142 testosterone and, 284 Role play, 138 Rough-and-tumble play. See also Play fighting gender differences in, 284, 428, 435 overview, 138–142, 152 peer relationships and, 381 in rhesus monkeys, 65–66 social understanding and, 138
S Schizophrenia, 274 School achievement predictors of violence and, 210 proactive and reactive aggression and, 187 School departure, 167 Schools environment of, 428 peer relationships and, 391 social capital and, 407–408 Selection model, 414–415 Selective breeding, 225–227. See also Genetics Self-esteem, 167 Self-regulation language–aggression association and, 346 overview, 307–308 Self-selection model, 209 Semantic development, 332–333. See also Language development Serotonin. See also Neuromodulators dopamine and, 265, 267 gamma-aminobutyric acid (GABA) and, 266–267 inhibition and, 261–262, 274–275 overview, 262, 263–265
478 Serotonin metabolism causes of aggression and, 96 in rhesus monkeys, 66–67, 69–70 Serotonin receptors, 232–234 Serotonin transporter gene (5–HTT) interaction with environment, 74–76f interaction with rearing conditions, 13 Sex differences. See Gender differences Sex segregation, 435. See also Gender differences Sexual encounters overview, 451 play fighting and, 48, 49–50, 52, 55 Sexual selection, 166, 426–427 Shared etiology pathways, 343–344 Shared genetic liability pathway, 343–344 Siblings. See also Family relationships causes of aggression and, 97 conflicts with, 117 preschool age children and, 388–389 Single-parent family, 403. See also Family Smoking. See also Substance use causes of aggression and, 97 childhood-onset persistent aggression (CPA) and, 357–358, 362t–363t conduct disorder and, 360 during pregnancy, 123, 210, 455 Social aggression. See also Indirect aggression; Relational aggression compared to indirect aggression, 159– 160 overview, 34 Social bonding, 49 Social capital daycare and, 405–406 at the family level, 402–405 negative, 414–417 neighborhoods and, 410–412 overview, 398–399, 417–418 peer relationships and, 408–410 schools and, 407–408 town and region and, 412–414 Social cognitions language–aggression association and, 346 parenting and, 382 proactive and reactive aggression and, 183 Social development, in rhesus monkeys, 64–66. See also Rhesus monkeys Social–development perspective, 38–39 Social experience, in rhesus monkeys, 72– 74 Social facilitation model, 415
Index Social forces, 137 Social-interactionist perspective, 32 Social interventions, 11. See also Interventions Social learning hypothesis. See also Environment causes of aggression and, 94 overview, 125, 429, 453 proactive aggression and, 179 Social organization innateness of, 133–138 in rhesus monkeys, 64–66 Social skills abstract play and, 147–149 indirect aggression and, 162–163 language development and, 342 peer relationships and, 378, 379 play fighting and, 50, 52–53, 55–57, 58 proactive and reactive aggression and, 195 in rhesus monkeys, 73–74 training in, 457 Social status, 162–163 Social systems causes of aggression and, 100 contextualism and, 18–19 Social testing, 49 Socialization gender differences and, 432 language–aggression association and, 343–344 model of, 414–415 overview, 95 responses to, 151–152 in rhesus monkeys, 64–66 from unilateral to bilateral views of, 13– 18 Socially motivated aggression, 32–33 Socioeconomic status. See also Welfare system age–violent crime curve and, 89 cognitive skills and, 313–314 indirect aggression and, 169–170 language–aggression association and, 343–344 language development and, 341 peer relationships and, 389–390 physical aggression and, 308 predictors of violence and, 210–211 proactive and reactive aggression and, 187 social capital and, 413 Sociometric status, 10
Index Specific language impairments (SLI), 332. See also Language development Speech and language difficulties (SLD), 332. See also Language development Splenium, 249. See also Brain mechanisms State dependence theory overview, 208–209 prevention and, 214 Strain differences, 69–70 Stress system cortisol and, 291–292 hormones and, 297–298 Structural MRI, 243–244f. See also Brain mapping; MRI Subcortical structures, 37 Substance use. See also Alcohol use; Smoking lack of behavioral control and, 261 overview, 451 proactive and reactive aggression and, 187, 195 serotonin and, 264 Subtypes of aggression. See also Taxonomies of aggression based on its antecedents, 34–39 based on its expression, 33–34 based on the consequences, 31–33 integrative model and, 39–41f multilevel analysis of, 29–30f overview, 41–42, 450–452 phylogenetic and ontogenetic organization of, 26–29 Suicide indirect aggression and, 167 molecular genetics and, 233 serotonin and, 264 Supervision, 404–405 Supplementary motor area, 251. See also Brain mechanisms
T Tantrums, 100 Taxonomies of aggression. See also Subtypes of aggression early, 27–29 integrative model and, 39–41f overview, 39 Teasing aggression, 38. See also Verbal aggression Temperament proactive and reactive aggression and, 184–185, 191 as a risk factor, 122
479
Terrorism, 33 Testosterone. See also Hormones causes of aggression and, 96 as an explanation for violence, 86–87t future research on, 296–298 overview, 282–283, 283–291 Theory of mind abstract play and, 149–150 executive function and, 316, 320–321 language–aggression association and, 346–347 Threat, 35–36. See also Reactive aggression Toddler years. See also Development; Infancy aggression during, 124–126, 430f–431, 452–454 anger during, 110 executive function and, 310 gender differences and, 432 individual differences and, 118–121 language development and, 334–335 peer relationships and, 378–380 risk factors and, 121–124 use of force during, 114–115f Towns, 412–414 Transsexuals, 287 Tryptophan hydoxylase genetic factors and, 233 overview, 264 Type A personality, 178–179
U Unilateral view of socialization, 13–18. See also Socialization
V Ventral premotor cortex. See also Brain mechanisms abstract play and, 145 overview, 251 Ventrolateral–medial hypothalamus, 37–38 Ventromedial cortex, 251. See also Brain mechanisms Verbal abilities, 315 Verbal aggression gender differences in, 436–437 overview, 108 Verbal organization, 273–275 Victimization in preschool children, 383 relationships and, 17 as a research construct, 5–6
480 Video games, 150 Vigilance, 268 Violence. See also Antisocial behavior; Homicide development of, 83–85, 207–209 genetic studies on, 228–234, 230f language development and, 333 neighborhoods and, 412 vs. nonviolent aggression, 5–6 overview, 202–203 play and, 149, 150 predictors of, 209–211 prevention of, 214–215 proactive and reactive aggression and, 187–188, 195 serotonin and, 264 testosterone and, 287 trajectories of, 211–214t, 212f, 213f Visuospatial abilities, 315 Vocabulary. See Language development
Index W War as aggression, 33 evolution of aggression and, 85 Weaning, 52 Weaning aggression, 28 Welfare system. See also Socioeconomic status overview, 399–402 social capital and, 418 White matter, 248–250. See also Brain mechanisms Wisconsin Card Sorting Test (WCST), 310, 311. See also Executive function World wars, 85
Y Young male syndrome, 169
Z Zoomorphism, 26–27