aggression, socialization, societies, human aggression, conceptual framework, Global Perspective, Arnold P. Goldstein Marshall H. Segall, anthropological literature, internal conflict, polygyny, Pergamon, Marshall H. Segall, Arnold P. Goldstein, Toronto Sydney Paris Frankfurt Pergamon Press, Pergamon Press, aggressive behavior, Syracuse University, cross-cultural study, Whiting & Whiting, cultural differences, Mexican mothers, internal warfare, Melvin Ember, ethnographic research
Pergamon Titles of Related Interest Alexander/Gleason BEHAVIORAL AND QUANTITATIVE PERSPECTIVES ON TERRORISM Center for Research on Aggression PREVENTION AND CONTROL OF AGGRESSION Goldstein/Carr/Davidson/Wehr IN RESPONSE TO AGGRESSION Miron/Goldstein HOSTAGE Rapoport/Alexander THE MORALITY OF TERRORISM Related Journals* INTERNATIONAL JOURNAL OF INTERCULTURAL RELATIONS INTERNATIONAL JOURNAL OF LAW AND PSYCHIATRY JOURNAL OF CRIMINAL JUSTICE *Free specimen copies available upon request.
PERGAMON GENERAL PSYCHOLOGY SERIES EDITORS Arnold P. Goldstein, Syracuse University
Leonard Krasner, SUNY at Stony Brook Aggression in Global Perspective edited by Arnold P. Goldstein Marshall H. Segall Center for Research on Aggression Syracuse University PERGAMON PRESS New York Oxford Toronto Sydney Paris Frankfurt
Pergamon Press Offices:
Pergamon Press Inc., Maxwell House, Fairview Park, Elmsford, New York 10523, U.S.A.
Pergamon Press Ltd., Headington Hill Hall, Oxford 0X3 OBW, England
Pergamon Press Canada Ltd., Suite 104, 150 Consumers Road, Willowdale, Ontario M2J 1P9, Canada
Pergamon Press (Aust.) Pty. Ltd., P.O. Box 544, Potts Point, NSW 2011, Australia
Pergamon Press SARL, 24 rue des Ecoles, 75240 Paris, Cedex 05, France
FEDERAL REPUBLIC OF GERMANY
Pergamon Press GmbH, Hammerweg 6, D-6242 Kronberg-Taunus, Federal Republic of Germany
Copyright © 1983 Pergamon Press Inc.
Library of Congress Cataloging in Publication
Data Main entry under title:
Aggression in global perspective.
(Pergamon general psychology series ; 115)
1. Aggressiveness (Psychology)--Cross-cultural
studies. I. Goldstein, Arnold P. II. Segall,
Marshall H. III. Series.
All Rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic tape, mechanical, photocopying, recording or otherwise, without permission in writing from the publishers. Printed in the United States of America
Preface Aggression in its diverse individuals and collective forms has long been, and remains, a worldwide problem of the first magnitude. When viewed in global perspective, contemporary aggression takes many guises--violence and vandalism by juveniles in schools and in their communities; child and spouse abuse and other forms of domestic or familial violence; assaults, muggings, and homicides; rape and other sex-related crimes; politically-motivated terrorism; racially- or economically-motivated mob violence; and aggression in many forms directly or indirectly initiated by the state. As the chapters which follow will make clear, this is far from an exhaustive list. We could add athletic mayhem, clan blood feuds, ritual torture, police brutality, organized warfare, and much, much more. The variety, intensity, frequency, and overall prevalence of overt aggressive behavior throughout the world is starkly and appallingly high. This book was initiated, organized, and written in direct response to this state of affairs, to this unremittingly high level of global aggression. Our specific purposes are several. The chapters which follow describe contemporary manifestations of aggression in a large number of nations representing almost the entire world. These descriptions are placed in a cultural context, thus helping us understand why, for the given country or region, aggression currently assumes particular forms, rates, and intensities. Such contextual information is also utilized in most of the ensuing chapter to aid in understanding how aggression "fits in" or is conceptualized in each nation's stream of daily living. Each chapter author was also charged by us to deal similarly with his or her society's efforts to control aggression, and to promote prosocial alternatives to aggression. How is it done? When and why does it succeed or fail? What additional controls and alternatives might lie ahead for each nation or region? Topics such as these, addressed by most of this volume's contributors, enabled us to devote our introductory chapter to comparative cross-cultural efforts. Here we seek to draw upon these individual cultural perspectives
on aggression, aggression control, and aggression alternatives to offer a more unified, global perspective. We compare, contrast, distill differences and similarities, put forth what we feel we may optimally learn from one another, and suggest specific directions for future research and applied efforts at better understanding of aggression, more effectively controlling it, and more successfully enhancing its prosocial counterparts. A prefatory comment seems appropriate from us regarding our bases for selecting the nations represented in this book, as well as the particular social or behavioral scientists who wrote its chapters. Our goal in country selection, as this book's title makes clear, was truly global representation . All major areas of the world were to be included. Our efforts toward this end were especially energetic and persistent and, while we were largely successful, two important areas are not represented in this book. We feel this absence keenly, as our goal was truly total and comprehensive representation, and our diverse efforts toward this end were protracted and vii
very considerable. It is, perhaps, a significant cross-cultural statement itself that, in spite of these diverse efforts, no Soviet or Arab social or behavioral scientist felt able or willing to join us. We are both saddened and chagrined at this state of affairs, and can only hope that the future will bring more open international scientific communication, especially about such volatile topics as aggression, its control, and its alternatives. The chapter authors are noted scientists and authors on the topic of aggression. They were recommended to us via their writings, their involvement in such organizations as the International Society for Research on Aggression, and their high standing in the global community of scientists. Collectively, they constitute a distinguished, expert panel of knowledgeable scientists. We are delighted with their participation. Our goal has been to organize and offer a book which presents both an elucidating and a utilitarian picture of aggression in global perspective: elucidating, in that it serves to help deepen our understanding of the meaning and nature of aggression throughout the world; utilitarian, in that its companion focus on aggression controls and alternatives in global perspective actually functions to aid the constructive, prosocial, anti-aggression efforts which do exist, or might exist, to more readily and more fully succeed.
Arnold P. Goldstein Marshall H. Segall
1 Aggression in Global Perspective: A Research Strategy Marshall H. Segall INTRODUCTION: THE RESEARCH OBJECTIVE AND STRATEGY The thrust of the research to be described in this volume is to understand why humans behave aggressively and violently. That they do is beyond doubt. That they do so to different degrees, in different manners, and for a variety of different reasons is also clear, as all of the contributions to this volume illustrate. At the same time, there are doubtless some universal features of human aggression, as these same contributions also show. The pancultural similarities and the cross-cultural differences in aggressive behavior together comprise the puzzle on which we and our international collaborators are working. The chapters which they have contributed, plus our own account of aggression in the United States, are, in a sense, the puzzle pieces. Assembling them into a more or less coherent picture is the ultimate objective of the research effort in which we are all engaged. The picture we are aiming for in this book cannot possibly be the definitive picture of human aggression in global perspective. That, obviously, must await more information than is presently available to us. But, as the eighteen chapters which follow this one show, there is more than enough to prompt us to make a first pass at the picture. We do so with the hope that the sketch which results will reflect reality at least in broad outline. As the reader will discover, it is not easy to digest and merge all the information contained in the accounts of aggression in the 18 different settings covered in this volume. While each is a case that is interesting in its own right, case accounts are raw data in the context of our real objective. They need to be compared and contrasted and examined as a collectivity in an effort to glean whatever generalizations, however tentative, they might support. They must also be compared with the works of those earlier students of aggression who share our ultimate goal and who have assembled their own tentative pictures. Especially noteworthy, in this regard, are the social scientists, including anthropologists, political scientists
, and cross-cultural psychologists whose research orientation, like ours, is comparative. What they have learned must be set down here next to what we have learned and all of it must be viewed as a whole. Their pieces and our pieces are all part of the same puzzle. We will, therefore , gather in as many pieces that seem to be a part of the puzzle before we attempt our tentative solutions. 1
Aggression in Global Perspective
A Methodological Preamble
But first, a word about our general strategy for a continuing research program in which the present volume comprises a first step. What we will try to do is identify as many as possible of the various factors in the natural and man-made environments of humans which might influence, in reliable ways, their behaviors relating to conflict, aggression, and violence. These ecocultural factors we consider to comprise the context in which such behaviors occur. We recognize that these factors are distributed in a dazzling variety of ways across human habitats and that they include features of the habitats themselves (e.g., climate and terrain) plus many other factors (e.g., economic systems, political structures, and socialization practices), some correlated with environmental features and some which cut across them. In our overall research strategy, this large set of ecocultural, contextual factors, singly, in combination, and very likely in interaction with each other, comprise our Independent Variable
s. But which ones of the seemingly endless list of independent variables deserve our attention? How do we choose them; how do we even commence the search? We find compelling, and we accept, the advice of Robert LeVine, a psychological anthropologist.
Research into causal relations can begin with dependent or independent variables. . . . Starting with independent variables entails the risk that their effects will turn out to be trivial, irrelevant, or otherwise uninteresting according to the criteria by which the scientist evaluates his research. ... . Beginning with dependent variables guarantees the scientist that his efforts will be directed toward explaining a phenomenon he considers important or interesting. This could be thought of as the etiological approach, in which a well documented variation of definite interest to the investigator and perhaps broader social significance poses an explanatory problem for research, that is, a search of causes. [LeVine, 1970, pp. 565-566]
The enterprise we are engaged in is an etiological approach, a search for causes of aggression. But to make that search worth commencing, we must have differences in aggressive behavior that require explanation. Unless we have differences in the ways human beings who happen to be scattered around the world react when confronted by a situation that might make some frustrated, some angry, or some pugnacious, then we don't have a problem in the sense that LeVine, and we following him, understand problem. By the same token, whenever such a difference has been documented, our work has just begun; then must commence the search for the operative independent variables. To begin, then, we need differences. Yet, the real goal of the enterprise is to emerge with universals. There is no paradox here, since the differences we refer to are behavioral differences, while the universals we seek are principles, or laws, that explain that behavior whenever it occurs. In other words, we are not seeking behavioral differences in aggression merely to demonstrate them, nor to proclaim gleefully their existence because we prize human variety. In the case of aggression, there is, as we shall argue below, some good reason to call attention to the simple fact of behavioral differences, but the essence of our work is to seek those differences in order then to determine the variables which produce them, anywhere and everywhere. It is, of course, epistemologically obvious that the establishment of universal generalizations about behavior requires variation in both that behavior and in the numerous factors which co-vary with it. This is true of all science; it is certainly true of cross-cultural psychology, the research strategy that characterizes our present effort to understand human aggression. To say that we are engaged in cross-cultural psychology is not to say, however, that we shall use culture to explain aggression. Culture is, undoubtedly, conceptually unclear (Jahoda, 1980). It just won't do as an independent variable, not for aggression, not for any behavior
A Research Strategy
that interests us. Whatever culture is (and we settle for the appropriately broad and loose definition of Moore and Lewis (1952): " . . . anything one person can learn from all other persons") , it is too gross, too diffuse, too multifactored to qualify as a variable. It is a valuable concept because the term, especially as Moore and Lewis and we use it, calls attention to the simple fact that the stuff of culture is learnable and indeed learned; and that, via learning, it is transmitted over time, usually through interpersonal contacts. So, the process of cultural transmission concerns us; nonetheless, we resist the temptation to explain differences in aggression as "caused by culture." Such a statement is empty. Following the suggestions of the anthropologist/psychologist team, Robert L. and Ruth Munroe (Munroe & Munroe, 1980) we think of culture as dissected into numerous, separable (albeit often correlated) contextual factors, including basic institutions, subsistence patterns, social and political organizations, languages, social rules governing interpersonal relations, divisions of labor by sex, age or other dimensions, population density, dwelling styles, and more, so many more that an exhaustive list of such potential independent variables is impossible. All those anthropologists and cross-cultural psychologists who pursue hologeistic research in the tradition of Murdock (1957, 1967) and Naroll (1973), among others, employing either coded ethnographies like the Standard Cross-Cultural Sample or the Human Relations Area Files or systematically coordinated multiple-site field work in order to test universal hypotheses about human behavior, employ variables like those enumerated above. In all such work, behavioral differences across cultural groups are related to one or more (and it is almost always more) variables as presence or absence of, or degree of presence of, some specific contextual factor. And so it is in our work on aggression.
A Conceptual Framework Assuming that we are confronted with facts about aggression that reveal reliable behavioral differences between human groups and that our objective is to relate them to the kinds or variables listed here, it is useful also to have an overarching conceptual framework about how human behavior changes over time and, by implication, comes to differ over space. We find that framework most cogently expressed in Donald Campbell's evolutionary epistemology (1977). Campbell's view of behavioral development at the individual level (and of change at the social level as well) is a model based on (but not identical with) contemporary theories of biological evolution. It is a model which emphasizes random variation and selective retention, tending toward adaptiveness in a circuitous manner and only in the long run. The behavioral variations are nonprescient; environmental selection for adaptiveness occurs only after the fact of behavioral change. Analogous to physiological natural selection
, which operates effectively at the species level, is trial and error learning at the individual level (and through social influence of one kind or another, at various group levels). Thus, the learning of each organism and the dynamics of all cultures tend toward adaptiveness, again in a nonprescient way, with environmental response to the trial variations determining their viability. In a virtual paraphrase of the Campbellian position, and in equally cogent language, B. F. Skinner (1981) asserted, "Selection by consequences is a causal mode found only in living things or in machines made by living things. It was first recognized in natural selection, but it also accounts for the shaping and maintenance of the behavior of the individual and the evolution of cultures [p. 501]." It should, of course, be obvious that neither Campbell nor Skinner, by referring to natural selection, were suggesting genetic determination for aggression, or for any other complex human action, or for cultural change. Rather, they were simply pointing out that biological evolution and human learning (and hence cultural change, since human learning is what culture is
Aggression in Global Perspective
all about) are parallel processes both tending toward the same end--survival--through nonprescient selection. So, if we have reason to believe that some behavior related to aggression is culturally influenced, we still have to ask, "In what way might this particular behavior, which we find in this/ these particular place(s) be adaptive?" In this formal respect, we must think as biologists and ask, "What do we know about the ecocultural contexts of the groups for which behavior A prevails and those of the groups for which behavior A -prime prevails that makes each of those behaviors fit the contexts in which they are found?" This conceptual framework which we have adopted is in tune with a growing ecocultural emphasis in cross-cultural psychology, as is found, for example, in Berry's (1975) perspective wherein most complex human behaviors and the cultural recipes for them are cumulatively selected for their adaptive character and are transmitted intergenerationally by social process
es, including socialization and enculturation, rather than by genetic adaptation. The Nature-Nurture Controversy
Because aggression is our topic, it may be necessary to be very explicit about where we stand on the question of genetic determination, particularly since we have just emphasized our interest in processes that parallel natural selection. As is well known, human aggression is widely thought to have a genetic basis. It may well be that this is the single most popular view. It certainly is among laypersons. And over the past few decades, many scientists have written works which have restored respectability to it. However, in accord with most students of aggression, we do not find the aggression-instinct arguments, in any of the several versions which have appeared in recent years (e.g., Ardrey, 1966; Lorenz, 1963; Lumsden & Wilson, 1981; Morris, 1967; Storr, 1968; Wilson, 1975), at all compelling. While we share with Lumsden and Wilson their notion that social learning can stimulate Mendelian transmission, we reject their notion of "epigenetic rules" whereby genetically-based neurobiological constraints channel behavior in particular directions. We do not believe that complex social behaviors are determined by natural selection; we know of no evidence for the existence of a heritable genetic bias in the development of complex social behavior. As for the Lorenzian insistence on an aggressive instinct, the compelling reasons for rejecting it are almost too familiar to bear restatement. Berkowitz (1969) and Kim (1976), among others, have documented many of the errors of fact and reasoning which mark Lorenz's work. As Kim so aptly put it,
the Lorenzian theory of aggression and war is seriously flawed on conceptual, methodological, and substantive grounds. The cavalier use of operationally ill-defined terms and concepts, the almost exclusive reliance on causal anecdotes, the disregard of empirical studies contradicting the monocausal paradigm, the inductive/extrapolative leaps to solve the level-of-analysis problem, the cross-species generalizations based on the recurrent tendency to advance argument in finalistic terms with little supporting evidence--all these weaknesses warrant scientific disapproval of the Lorenzian theory, [pp. 270-271]
The mere fact of differences across societies in degree, form, and concomitants of aggressive behavior is, in our view, an overwhelming problem for instinct theorists and sociobiologists. We doubt that they are equipped to solve it by remaining within their theoretical frameworks. And we will not try to do it for them. The behavioral differences with which we are here confronted are, of course, also a problem for us, who work within our theoretical framework--social learning theory--buttressed by Campbellian evolutionary epistemology and Skinnerian selection-by-consequences. But,
A Research Strategy
as we have tried to make clear in the foregoing discussion, differences are precisely what our theoretical predilections have primed us to expect. It may not be easy to make sense out of them, but we welcome them, indeed require them, as the raw material to be entered into our search for explanations of human aggression. For us, the differences are a problem only in the happiest sense of that term, a scientific puzzle, one for which we believe the solution resides in a social learning approach. And, one more preliminary word on the nature vs. nurture controversy as regards human aggression. As we have noted, genetic hypotheses and learning hypotheses, as competing alternatives, both must ultimately be assessed in the light of their competence in pointing to systemic fit. While we find learning more plausible than biological evolution as the mechanism that provides that fit for aggression, we nonetheless acknowledge that only when the adaptiveness of the aggressive behavior is accounted for can that behavior be said to be understood . So, in the end, our criterion for success is the same as the sociobiologist's or the instinct theorist's. Like them, we have to explain aggression wherever it occurs, in whatever form. We have to show how it fits its setting. But, unlike some instinct theorists, we do not ever have to claim that aggression is "good." Functional, yes, in the sense that a given degree of aggression fits other features of a particular ecocultural setting, but good for the aggressors or further victims? Absolutely not! Only instinct theories or other frameworks which employ hydraulic system metaphors force their adherents into that morally repugnant and scientifically indefensible stance.
ANTHROPOLOGICAL LESSONS ON AGGRESSION The anthropological literature on aggression and violence is both ethnographic and nomothetic. The ethnographic portion is descriptive of particular societies, intensively studied, one at a time, usually by a particular anthropologist who specializes in that society. (The case studies of the large-scale contemporary societies included in this volume are like ethnographies in most of these respects.) The nomothetic portion of the anthropological literature is composed of what are often called hologeistic studies (see Schaefer, 1977, for a history and contemporary description of the field), which are comparative, across many societies, extensively examined, usually via secondary analyses of previously published ethnographies. Scholars engaged in hologeistic research seek to test cross-cultural, universal hypotheses about human behavior, usually employing quantitative techniques that reveal patterns of associations. The long-range objectives of the research program of which the present book marks a start are clearly akin to those of the hologeistic scholars. We believe that many non-anthropologist students of aggression are unfamiliar with the anthropological literature on aggression, particularly several recent works of the hologeistic variety. We have consulted many of these works and will describe or summarize those which we believe offer intriguing hypotheses, with supporting data, about ecocultural and socialization correlates of aggression. Our review of the literature will not be exhaustive; it is too large and varied. And, given the objectives of this chapter, we shall pay less attention to the purely ethnographic works that deal with aggression, which are very numerous indeed, than we shall to the less numerous comparative studies. The Ethnographic Evidence First, however, we shall say a few words about the ethnographic literature, for it offers students of aggression some important preliminary lessons. To begin with, that literature shows clearly that there are societies in which adults, in their dealings with each other, with their chil-
Aggression in Global Perspective
dren, and sometimes even with outsiders, seldom behave aggressively. As Ashley Montagu (1978) stated, in his introduction to a collection of ethnographic reports dealing with several widely dispersed, non-literate societies:
Many human societies cannot be characterized as aggressive. . . . There are societies in which both inter- and intragroup aggression is low, as among the Toda of Southern Indi and there are societies in which both inter- and intragroup aggression are nonexistent, as among the Tasady of Mindanao, * in the Philippines. . . . [Other] societies that are notable for their unaggressiveness . . . are the Punan of Borneo, the Hadza of Tanzania, the Birhor of Southern India, the Veddahs of Ceylon , the Arapesh of New Guinea, the Australian aborigines, the Yamis of Orchid Island off Taiwan, the Semai of Malaya, the Tikopia of the Western Pacific, the Land Dayaks of Sarawak, the Lepchas of Sikkim, the Papago Indians, the Hopi, the Zuni, and the Pueblo peoples generally, the Tahitians, and the Ifaluk of the Pacific, [pp. 3-5].
This long, recently compiled list is not exhaustive. Far from that. Ethnographers have documented other cases of societies, mostly small, and mostly nonliterate ones, where the frequent acts of aggression and violence which seem so "natural" in large, complex, usually technologically advanced societies, like nearly all of those discussed in later chapters of the present volume, are, rather than frequent, virtually absent. However, a review of the ethnographic literature would also show that some forms of aggression occur in nearly every society, including most of these societies in Montagu's list. Moreover, there are good reasons to be skeptical about assertions that aggression is "nonexistent" in any society, if only because no observer can ever be sure to have detected every behavior that takes place nor to have appropriately and fully interpreted every act that was observed. So, this first lesson from ethnography regarding societal differences in aggression is that such differences exist and do indeed span a wide range of magnitudes but that aggression is at the same time quite likely a universal p h e n o m e n o n . Its universality coupled with its wideranging variation is what makes it such an intriguing scientific puzzle. The second lesson concerns the level of complexity that must be expected in any satisfactory solution to that puzzle. The ethnographic literature, in addition to pointing to many cases of small, nonliterate societies that display very low levels of aggression, contains cases of similar societies, also small and nonliterate, that are highly aggressive, such as the somewhat notorious Y a n o m a m o of Southern Venezuela, whose frequent warfare has been described as ferocious, an adjective that seems appropriate, given the ethnographic details provided by Chagnon (1968). So, the puzzle cannot be solved simply by appealing to societal size, simplicity, or level of technological development. The variability in h u m a n aggression, the range and diversity of aggressive behaviors, and their distribution over numerous other variables--geographic, cultural, economic, politicalare what convince us that research that focuses on ecocultural, social, and experiential antecedents of aggression is both necessary and promising. That aggression's antecedents are many and complex and probably intertwined, interactive, and confounded as well, further convinces us that research that has any hope at all of understanding aggression must be multidisciplinary, multimethod, and very open-minded. There is a third lesson to be gleaned from ethnographic research, since that research suggests that aggression's antecedents are many and complex, there must be some sets of condi-
*Ed. Note. The Tasaday are a special case. Comprising only 26 individuals, 13 of them children, when discovered in 1966, they have not been well studied. As noted below in this chapter, skepticism regarding the "nonexistence" of aggression in such societies is clearly in order.
A Research Strategy
tions that only are found to exist in some societies, but also are encouraged to take root and flourish in other societies, if human aggression is to be controlled and reduced. An optimistic hope, then, which can be derived from what the ethnographies tell us about the nonstereotypy of human aggression, is that competent research can reveal potentially pragmatic approaches to the inculcation and enculturation of nonaggressive behavioral alternatives to situations where aggression is presently the norm. If aggressiveness varies in some as yet undetermined but systematic fashion, and if the conditions which contribute to lesser aggressiveness, wherever they prevail, can confidently be ascertained, then those conditions might be encouraged anywhere. However, to realize these promises--both the scientific one of understanding and the pragmatic one of controlling human aggression--we must go beyond ethnography per se. Digesting available ethnographic descriptions or compiling additional ones would serve merely to underscore the lessons just reviewed and to reiterate the promises just enunciated .The ethnographic facts alone do not provide the generalizations we seek, either to understand human aggression or to intervene in efforts to control, reduce, or minimize it. For that task to be undertaken properly, we must adopt the kind of hypothesis-testing stance that is characteristic of hologeistic research.
Hologeistic Studies of Aggression
Using ethnographies as data sources, hologeistic research aims for generalizations about human behavior. Exploiting the many kinds of variations that exist across the many hundreds of societies in the world, hologeistic researchers seek regularities in association and covariations in diverse characteristics that might serve as the basis for plausible inferences regarding causality. The causal statements about human behavior that emerge from hologeistic studies are, of course, no more than inferences since the forms of analysis available in such studies are no better than correlational. The hologeistic research must take the ethnographic facts as they are (or, more properly stated, as they have been described in the first instance by others) and examine their covariates. But whatever limitations regarding the establishment of causality that this fact implies, and however numerous and serious the threats to validity inherent in having to employ secondary data sources of varying quality (for discussion of these problems and ways to solve them, see Naroll, 1962, 1973; and Naroll, Michik, & Naroll, 1980), hologeistic research is an unparalleled research strategy for discovering what goes with what on the broadest possible canvass--the many and diverse habitats in which the human animal survives. To employ this method, one samples from the many hundreds of human societies for which ethnographic descriptions have already been compiled. These exist in several convenient sets and in highly usable formats, with very large numbers of characteristics extracted, coded, stored in accessible form, and very much subject to quantitative analysis. The best known and most convenient of these sets are the Human Relations Area Files (see Barry, 1980 for a thorough description), the Standard Cross-Cultural Sample (Murdock & White, 1969) and the Ethnographic Atlas (Murdock, 1967). Sets of coded variables from one or another of these data set
s can conveniently be found in Barry and Schlegel (1980) and Textor (1967). Examples of hypothesis testing employing such coded variables can also be found in those two sources, among many others. And the topics covered are wide ranging, many of which, of course, have little to do with aggression. However, among the coded variables available in these and similar sources are many that do relate to aggression and many more that describe social structures, economic systems, socialization practices, and other societal characteristics that plausibly might be expected to re-
Aggression in Global Perspective
late to aggression. Not surprisingly, then, considerable research that seeks cross-cultural generalizations about human aggression has already been accomplished. Equally not surprisingly, the goal of such research--truly to understand why aggression varies the way it does--remains elusive. In an effort to determine how far hologeistic research has carried us toward that goal, we turn now to some selected examples of studies that at least have asked some of the questions that need to be asked. We shall see that many of the tentative answers raise still other questions. We shall also see that many such questions remain to be raised. And we shall see, as was suggested in our cursory review of the relevant ethnographic literature, that the story of the antecedents of aggression is indeed a complex one. We shall begin our digest of that story by considering some cross-cultural studies relating to socialization emphases.
Inculcation of Aggressiveness During Childhood
Considerable cross-cultural evidence exists to support the general proposition that many traits which are functional for adult activities in particular societies are, in part at least, inculcated during childhood. Thus, Barry, Child, and Bacon (1959) found relations between adult subsistence economic activities and various socialization emphases during childhood, in such domains as training to be nurturant, to be independent, to be obedient, and other such traits. If aggressiveness may be thought of as a trait, it is reasonable to ask (1) whether societies differ in the degree to which it is inculcated during childhood and (2) whether such differences relate to other trait-inculcations or other aspects of culture, especially economic activities. To pursue this question, Barry, Josephson, Lauer, and Marshall (1976, 1980) scored nearly 150 societies drawn from the Standard Cross-Cultural Sample (Murdock & White, 1969, 1980) on inculcation of aggressiveness among children beween the ages of approximately four and twelve years of age (both early and later during that age span), for both boys and girls, and along with a number of other variables. They found considerable variation across societies in inculcation of aggressiveness, defined as:
aggressive behavior toward people (including peers) or animals, which may be implicitly inculcated or condoned by adults, e.g., parental urging to stand up for oneself or retaliate against aggression. Exhortations or frequent retelling of heroic myths may also instill aggressiveness; overt and covert inculcation are both included. [Barry et al., 1976, 1980, p. 215]
Using a 5-point scale (subdivided with plus and minus for finer discrimination) with 3 indicating "moderately strong inculcation," 5 "extremely strong inculcation," and 1 "no inculcation of the trait or strong inculcation of its opposite," they were able to make confident judgments for 148 societies on, for example, aggressiveness among boys during later childhood. Regarding this trait, the judgments were distributed as follows:
· 92 societies(62 %) were judged to employ moderately strong inculcation of aggressiveness, i.e., were scored from 3 ~ to 3+ · 24 (16%) were scored 4 " or 4 · 8 (5%) were scored 4+ or 5 · 24 (16%) were scored from 2" to2 + · No societies scored lower than 2"
As this distribution of scores shows, the modal tendency in this sample, to employ moderately strong inculcation of aggressiveness in boys, was characteristic of nearly two thirds of the soci-
A Research Strategy
eties, which indicates considerable worldwide pervasiveness of aggression-inculcation. In addition, another fifth of the societies appeared to inculcate aggressiveness to an even higher than modal degree. Just under a fifth scored low. Societies scoring both high and low were found in all regions of the world; regional differences across the high and low scoring societies were minor, with a slightly disproportionate number of high aggression-inculcation societies in Africa and a slightly disproportionate number of low aggression-inculcation societies in East Asia. So, the first finding in this study is that there are cross-cultural differences in the degree to which aggression is inculcated in boys during later childhood. Similar differences were found also for boys during early childhood, so similar that there is no need here to describe those differences in any detail. For girls, at both stages of childhood, there were again similar patterns of differences among the societies. However, a sex difference was found; in most societies, and on the average over all societies, there was more inculcation of aggressiveness for boys than for girls, with the sex difference more marked during the later period than during early childhood. Also, for both boys and girls, the mean rating of degree of aggression-inculcation was higher in the later childhood stage than in the earlier one (Barry et al., 1976, 1980, p. 219). Despite these mean differences involving gender and stage, plus the interaction between gender and stage, the following generalization clearly holds: Societies vary in aggression-inculcation during all of childhood and for both sexes, with the variation across societies following the pattern described above in detail for boys during later childhood. With what does this cross-cultural variation in aggression-inculcation correlate? It correlates with other socialization variables. For example, it correlates positively with inculcation of "fortitude" (training to suppress pain and fear reactions) and with inculcation of "competitiveness" (encouragement to achieve superiority over others). The correlations among all three of these inculcation variables was high enough (Pearson J?s between .38 and .45 for all possible pairs) that Barry and his colleagues could justifiably combine them into a single cluster which they labeled "toughness." This cluster was independent of four other inculcation clusters included in this study (maturity, sociability, dutifulness, and submission). However, aggression-inculcation by itself was found to be negatively correlated with two measures of the inculcation of sociability cluster, namely inculcation of honesty and inculcation of trust. So, training to be aggressive is stressed where training to be honest and trusting of others is not stressed. And, training to be aggressive tends to be a part of a general toughness-inculcation syndrome. The toughness-inculcation syndrome (which, as we have just seen, includes aggression-inculcation) is in turn related to certain structural characteristics of the societies. One of these, which involves marriage systems, lineage, and residence patterns, can be characterized as "availability of contact with another during infancy." High contact with mother during infancy was found to be associated with high inculcation of aggressiveness during later childhood, for both boys and girls, but more strongly for boys. In addition, one part of the toughness-inculcation syndrome, inculcation of competitiveness, was associated with importance of animal husbandry, a subsistence economy variable. (However, aggression-inculcation per se was not, nor was the over all toughness-inculcation syndrome.) So, the best, albeit tentative generalization to be gleaned from the Barry, Josephson, Lauer, and Marshall (1976, 1980) study isthat, despite considerable worldwide inculcation of aggressiveness during childhood (with the mode characterizable as moderately high), there are variations in degree, with more of it likely in societies where infants have much contact with mothers, and with one of its correlates, inculcation of competitiveness, more likely to occur in societies where animal husbandry is an important feature of the economy.
Aggression in Global Perspective
With regard to this last point, it may be instructive to consider the Masai, the society which in this study earned the highest aggression-inculcation score. The Masai are well-known East African pastoralists who, during infancy, rely exclusively on female caretakers. Later in childhood, following genital maturation (and perhaps as late as 18 years of age), both boys and girls are subjected to relatively severe initiation ceremonies. These involve genital operations (circumcision and clitorodectemy) and are judged anthropologically (e.g., by Schlegel & Barry, 1979, 1980) to stress same-sex bonding and, especially for males, valor (rather than wisdom or responsibility, which are the foci of many other initiation ceremonies). Thus, the Masai serve well to illustrate the kinds of societies which inculcate aggression more than most, and to underscore the kinds of cultural variables which reinforce that socialization emphasis. For contrast, consider the case of the Tahitians. Not included in the Barry et al. (1976, 1980) study, they have been described recently by Levy (1978) as gentle, affectionate, tender, and living in harmony in an easy, nonfrustrating environment. Relevant to our present concern with inculcation of aggressiveness and its correlates, it is clear from Levy's account that the Tahitians score low in this regard. Moreover, Levy reports, "Young infants are the center of attention, cherished, fussed over, gratified, protected and a sense of basic trust is instilled [p. 227]." "Also communicated, are the notions that masculinity, striving and aggression are dangerous, but cooperation with nature will be successful [p. 228]." "And childcare, from infancy onward, is the responsibility of many, not only mothers, with more than half the homes in the community having adopted children [p. 228]." As the reader of this volume encounters the various descriptions of aggression in 18 contemporary nation-states, he might profitably read them with an eye toward hints regarding the magnitude of aggression inculcation present therein. Do economic variables and degree of contact with mothers operate in such societies as they do in the smaller, nonliterate societies covered by the Barry et al. (1976, 1980) study?
The Socialization of Aggressive Behavior by Children
When children behave aggressively, caretakers and others react. Those reactions, whatever they may be, presumably have consequences for strengthening or weakening of the behaviors. This notion, widely shared by most social learning theorists, was a feature of a classic cross-cultural study of mothers and children in six selected societies (Minturn & Lambert, 1964; Whiting, 1963; Whiting & Whiting, 1975). While not a hologeistic study, strictly speaking, since it involved only six societies, it is certainly in the hypothesis-testing tradition of such studies and, in certain respects, possesses some methodological advantages over them, e.g., the use of standard observations for all cases in the sample. The six-cultures-study ranged widely over a series of questions, but a portion of the study dealt with aggression and related matters. It is that portion to which we now turn. Observations were made by anthropologist/psychologist teams of mothers and of children of both sexes in two age groups (3-6 years and 7-11 years) in Okinawa, India, the Philippines, Mexico, Kenya, and New England in the United States. Among the children, two forms of aggressive behavior were observed, termed by the authors "opportunity aggression" and "retaliatory aggression" (roughly speaking, unprovoked and provoked). Interviews with mothers provided information regarding their ways of coping with such behavior when displayed by their children. Numerous cultural differences were found in this study. Among those most relevant to our present concerns are the following. Mothers in these six societies varied in their reactions to aggression directed to themselves by their children. The African sample (Nyansongo, a Gusii community in Kenya) was the
A Research Strategy
most punitive in this regard, the Indian mothers (the Rajputs of Khalapur) least. An interesting correlate was the degree to which mothers make independent economic contributions (maximal in Nyansongo, minimal in Khalapur). Regarding aggression directed by children against other children, Mexican mothers (Mixtecans of Juxtlahuaca) were most punitive and mothers in the United States ("Orchard Town" in New England) sample were least punitive, even encouraging retaliation by their children to aggression stemming from their peers. The most interesting correlate here was the degree of structural relationship with neighbors. The less close the relationship, the less concern mothers showed over the peer-directed aggression of their children. As one Mexican mother put it, ". . . if our children don't get along, then my brother and I may come to a parting of the ways [Lambert, 1971, p. 51]." What impact did these and other socialization tendencies have on the behavior of the children? No simple answer is possible. The facts regarding aggressive behavior by children cannot be understood out of context. And that context included numerous other kinds of behaviors. However, these were reduced to 12 behavioral categories: seeks help, seeks attention, seeks dominance, suggests responsibility, offers support, offers help, acts sociably, touches, reprimands, "assaults" sociably, assaults, and symbolically aggresses, plus a separately treated category of compliance. These behavioral categories were multidimensionally scaled and two dimensions emerged, one defined as "dependent and dominant versus nurturant and responsible" and the other, "sociable versus generally aggressive." The first was found to be related primarily to cultural complexity, with less complexity paired with nurturance and responsibility. The second related mostly to household structure (most notably whether families were nuclear or not), with children in nonnuclear families more aggressive. But, more important, the two behavioral dimensions and the two societal characteristics were interrelated; both sets of characteristics have to be taken into account to understand the different patterns of the two behavioral dimensions. In other words, socioeconomic complexity and household structure combine to influence the combination of behaviors characteristic of each society. Consider two concrete examples. The children from the New England town, whose setting involved a nuclear household structure and a complex socioeconomic system, were alone in scoring both on the nonnurturantnonresponsible side of the first behavioral dimension and on the sociable side of the second behavioral dimension. The children in the Kenyan village, whose setting involved a nonnuclear household structure and a relatively simple socioeconomic system, were alone in scoring on both the nurturant-responsible side of the first dimension and toward the authoritarianaggressive side of the second one. To complicate the picture further, mothers' workload was implicated in the cross-cultural differences. But, if one were, in spite of these complexities, to venture a "simple" hypothesis regarding both (a) the socialization of aggression and (b) manifest aggression, it would be as follows. It comes in several parts.
· Aggression against caretakers is punished most where children's obedience is necessarily valued by mothers who contribute economically and need help from their children, especially in households where other adults are present. · Aggression against other children is punished most where economics demand good relations with close neighbors. · Whatever the sanctions against aggression, it is most likely to occur in cultures with patrilineal extended families, combined with simple economic systems.
In short, the socialization of aggression, whether severe or lax, does not by itself predict the
Aggression in Global Perspective
manifestation of aggression. Severe socialization of aggression does not reduce it. Indeed, it may enhance it. Or, it may coexist with it because high manifestation of aggression by children may elicit stronger sanctions against it from their caretakers. So, while we have simplified the picture regarding aggression and its socialization that emerges from the six-cultures study, it is a picture that is not, in an absolute sense, simple at all. But we do have a sense of what variables deserve further research: socioeconomic complexity, household structure, and the workload of mothers. Recent work by members of the six-cultures-research team (e.g., Lambert & Tan, 1979) on three different aggressive styles of children (self-instigated, retaliatory, and "surprise" aggression) showed again that such variables are implicated both in mothers' behavior and that of their children. From another large-scale, on-going hologeistic research project, one that focuses on correlates of parental acceptance and rejection (Rohner, 1975; 1980; Rohner & Nielson, 1978; Rohner & Rohner, 1979), evidence has consistently accumulated to support the view that rejected children throughout the world are more hostile or passively aggressive than are accepted children, particularly when caretakers behave hostilely toward them, thereby providing aggressive models to emulate (Rohner, 1980, p. 6). And, of course, it is usually the case that a given generation's hostile children grow up to be the next generation's hostile adults (as shown in Rohner, 1975, with a sample of 101 societies), so that conditions for intergenerational transmission of a syndrome of hostility, mistrust, and rejection seem easily established and very difficult to remove. In both the Rohner project and the work done by Lambert and his associates within the framework of the six cultures study, there is concern with maternal warmth as a variable influencing tendencies to aggress in children. Warm mothers are, in Rohner's terminology, accepting mothers. Lambert's (1971) analysis of maternal warmth (showing little hostility in dealings with children, rarely using physical punishment) showed that it is more likely to be the characteristic mode of behavior of mothers when there are other adults in the household who can share tasks (especially child-rearing), when there are fewer siblings around to function as competitors, and when fewer "courtyard cousins" are present to serve as potential targets for mothers to respond with severe sanctions. To summarize, the several studies that deal with socialization of aggression in children show that caretakers' (particularly mothers') reactions to children impact in nonsurprising ways on children's tendencies to aggress. In addition, the relevant behaviors of the caretakers are a complex product of socioeconomic and household structural variables. These notions now appear to be panculturally valid. Do they apply, then, to technologically developed societies such as those described in this volume?
Sex Differences in Aggression
A recent review of 130 studies done on sex differences in North America from the 1930s through the mid-1970s (Rohner, 1976) confirmed the well-known fact that American males, both men and boys, are more aggressive than females. Rarely did a study show the reverse; most supported the generalization of more male aggressiveness, while a few studies showed no significant sex difference. The sex differences were not as large among adults as among children, but the proportion of studies that found males more aggressive was almost as high for adults as for children. In all these respects, the cross-cultural evidence leads to the same conclusion. Using the same sample of 101 societies in the parental acceptance and rejection study, Rohner (1976) found that, while aggression by males and females was positively correlated (in societies in which males were more aggressive than in other societies, so were fe-
A Research Strategy
males), males tended to be more aggressive than females in most societies. And while that sex difference was less marked during adulthood, it does not totally disappear, even as women become more aggressive and men less aggressive. Thus, the North American findings that males from early in life are more aggressive than females can now, in Rohner's words, ube raised to the level of a panspecies generalization [p. 69]." This generalization, Rohner noted, is subject to diverse explanations. One would attribute the sex difference to differential socialization pressures; but, while evidence for such differential socialization can indeed be adduced, the best available cross-cultural evidence on inculcation of aggression (Barry et al., 1976) showed the sex difference in this regard to be small, albeit consistent. Indeed, in many societies in that study, the pressures on boys and girls to be aggressive are about the same. So, this simple hypothesis seems wanting. Another attributes the sex difference to the sexual division of labor, which in turn leads to different socialization pressures on boys and girls because the adult roles they will have to play call for different adult traits, with the male roles requiring more aggressiveness. Again, support for this view can be found; and, though Rohner rejects it, we find it rather compelling. Rohner's objection is based on the fact that differentiation between the sexes during childhood, as found, for example, by Barry, Bacon, and Child (1957), is so consistent across cultures that one must suspect that the differentiation reflects a "phylogenetically acquired, species predisposition" [Rohner, 1976, p. 70]. Rohner suggests that, even if males and females are assigned different tasks during adulthood and it is preparation for these tasks that calls for differential socialization, one still must explain, for example, "why boys around the world are assigned tasks that lead to greater aggression than girls [p. 70]." But, in our view, Barry et al. (1957) satisfactorily answered that question by noting that the sexual division of labor known to most subsistence-level societies is one that derives primarily from the simple fact that females are child-bearers and that close-to-home activities being assigned to females would be an adaptive response, at least in many kinds of ecologies, as has also been argued persuasively by Van Leeuwen (1978). And, in one of the reports to emerge from the six cultures study (Whiting & Edwards, 1973), there is at least one important finding that supports this argument. In two of the societies--in Kenya where some child-care and other domestic tasks are assigned to boys, and in New England where tasks assigned to girls are not stereotypically feminine ones--sex differences in behavior (including responding aggressively) were smaller or less frequent than in the other societies. Furthermore, the Whiting and Edwards paper has become a center of controversy involving Maccoby and Jacklin (1980) and Tieger (1980). The former two psychologists had argued earlier (1966) that the cross-cultural evidence, particularly that from the six cultures study, supported their contention that males are biologically predisposed toward aggressive behavior. Tieger, on the other hand, found their argument wanting, and emphasized the same finding from Whiting and Edwards (1973) that we cite here, namely that, as Tieger (1980) puts it, "the greatest variation . . . is found in societies where children's work roles differ from traditional patterns most strongly [p. 945]." In their 1980 rejoinder, Maccoby and Jacklin emphasized a measure, based on data pooled over societies, which does support an early sex difference . It is impossible to resolve the dispute here. Accordingly, some attention must be given to Rohner's third and preferred explanation for sex differences in aggression, which posits a genetic predisposition which is subject to interaction with enculturative pressures. He thereby places himself in the Maccoby and Jacklin camp. Whatever the explanation, sex differences in aggressiveness do seem to be very wide spread and a continuing search for the contributing factors is likely to be a feature of cross-cultural research on aggression. Perhaps in the chapters solicited for this volume, some such factors may be discerned.
Aggression in Global Perspective
Organized Violent Conflict A kind of aggression that is well known in the more industrialized societies described in detail in this volume is intergroup hostility, one form of which, warfare, is both a feature of their recent histories and a continuing possibility for which many such nations remain vigilant. Organized feuding and warfare are not unknown among the kinds of societies that anthropologists study, either. The anthropological literature on organized, sanctioned violence deals with both feuding and war. As noted earlier, the anthropological literature (in this respect among others) is both ethnographic (idiographic) and hologeistic (nomothetic), and it is a very large literature. As such, it can not exhaustively be reviewed here. But we will present a summary of some of it, especially recent works of the hologeistic variety, since these offer some intriguing hypotheses about why groups engage in feuding and warfare. Since the early 1960s, a number of studies have been completed which illustrate the complexity of influences that seem to impinge on various forms of organized combativeness, including feuding and both internal and external warfare. These three, while they share certain characteristics, need to be distinguished from each other, as will become clear when we review the findings from a few studies. The distinctions among them relate to the anthropological concepts of political community and cultural unit, with the former defined as "a group of people whose membership is defined in terms of occupancy of a common territory and who have an official with the specific function of announcing group decisions [Naroll, 1964, p. 286]" and with cultural units composed of contiguous political communities that speak a common language and in other respects are culturally similar. The three forms of organized combativeness relate to these social group distinctions as follows: Feuding occurs within a political community, internal warfare occurs between political communities that are parts of the same cultural unit, and external warfare occurs between cultural units.
Feuding. Feuding is usually characterized (e.g., Otterbein, 1968) as an ambush by an organized band on a member of their own political community, usually to avenge an act of homicide. The victim is often a relative of the earlier killer, and he is usually killed in turn. At least two studies have revealed that the frequency of feuding, so defined, relates to the presence or absence of fraternal interest groups--localized groups of related males, a form of social organization which is found in societies that have patrilocal residence patterns. The first study (Van Velzen & Van Wetering, 1960) compared patrilocal with matrilocal societies on five different indices of intrasocietal conflict and found the patrilocal societies to be less peaceful than the matrilocal ones on all five. The second study (Otterbein & Otterbein, 1965) concentrated on feuding per se and indexed the importance of fraternal interest groups by two factors--patrilocal residence and polygyny. In this study, it was shown that feuding was more likely to occur in societies with fraternal interest groups.
Internal Warfare. Keith Otterbein later reasoned (1968) that the existence of fraternal interest should "also produce warfare between political communities [p. 281]." But, in this study, only one of two indices of fraternal interest groups (i.e., polygyny) was significantly related to internal warfare (X2 = 3.93, N = 42 societies, p < .05); patrilocal residence was not. A complication, then, that needs to be kept in mind is that polygyny and patrilocality, while both indices of fraternal interest groups, are different variables in their own right. In one set of 50 societies examined by Otterbein, of 33 patrilocal ones, only 18 were polygynous; among the 17 nonpatrilocal ones, 3 were polygynous. Thus, the notion that fraternal interest groups "produce" internal warfare might better read "polygynous societies are groups in which internal warfare is more likely." We shall return to this point later.
A Research Strategy
Putting aside for the moment the questionable relationship of fraternal interest groups and internal warfare, we can extract from Otterbein's (1968) study another interesting (but essentially unconfirmed) hypothesis. It concerns the degree of political centralization that characterizes a society. Centralized political systems have authority patterns that usually limit the initiation of warfare to one designated individual or agency (as in modern industrialized states and in many traditional societies as well); in uncentralized political systems, warfare can usually be initiated by anyone. This fact, demonstrated empirically for 42 societies (16 centralized, 26 not) in Otterbein's own study, led him to predict that centralized societies would be more likely to engage in frequent internal warfare. In fact, they are not. There was no significant relationship found and the trend was even in the opposite direction. This forced him, and us, to push the analysis a step further. When Otterbein broke down his sample into polygynous and nonpolygynous societies, he found that it was primarily in the former that political centralization was related to internal warfare; in patrilocal societies, the politically centralized ones were those most likely to inhibit internal warfare, while in nonpatrilocal societies, no such relationship prevailed. Similarly, political centralization and war inhibition were related in polygynous societies, but not in nonpolygynous societies. So, it is tempting once again to suggest that fraternal interest groups (presence or absence) is a relevant variable, since, as noted above, patrilocality and polygyny are both indices of that variable. Accordingly, Otterbein reasoned as follows: Since anyone can initiate warfare in uncentralized political systems, especially if patrilocal and/or polygynous, then in such societies, internal war should be more frequent. He did find the expected significant interaction: in uncentralized political systems, those with fraternal interest groups were more prone to internal war, while in centralized political systems, those without fraternal interest groups were more prone. It was, however, not possible to determine whether the interest group variable or the ability to initiate war variable was the better predictor of internal warfare proneness. Despite this ambiguity, Otterbein (1968) asserted that "societies with fraternal interest groups are more likely to have both feuding and internal war than societies without fraternal interest groups [p. 287]." In our opinion, this often cited assertion must be qualified by stating that fraternal interest groups possibly increase the frequency of internal warfare only in uncentralized political systems, where it can be initiated by nearly anyone.
External War as an Independent Variable. In both the study of feuding (Otterbein & Otterbein, 1965) and the study of internal warfare (Otterbein, 1968), the frequency of external war was employed as an independent variable. Arguing, as many students of aggression might, that intergroup conflict, such as was between different cultural units, creates cohesion within such units, it can be predicted that external war would be negatively correlated with both feuding and internal warfare. However, in the case of feuding, the centralized/uncentralized political system variable intervened to affect that predicted relationship; in centralized societies, feuding and war were negatively correlated but, in uncentralized societies, they were actually positively related. On the other hand, no such interaction prevailed with regard to internal warfare. So, to summarize the picture to this point: External war seems not to reduce internal warfare in either kind of society (centralized or not). And warfare per se, whether internal or external, is accompanied by diminished feuding only in centralized societies where, presumably, leaders can suppress feuding when necessary.
External War as a Dependent Variable. We have just considered some evidence that external warfare has a dampening effect on feuding, but only in centralized societies, and no such effect on internal warfare, regardless of degree of political centralization. We saw earlier that feuding was more common in patrilocal and polygynous societies (which, as we have
Aggression in Global Perspective
noted more than once may or may not best be explained in terms of the presence of fraternal interest groups) and that internal warfare was more common in polygynous societies but not in patrilocal ones. Polygyny thus related to both feuding and internal warfare. Does it relate to external war, too? Is the presence of polygyny a factor in the frequency of external war? Can some of the variation in the frequency of external war be attributed to this widespread form of marriage? And, since polygyny is one index of fraternal interest groups, is the existence of such groups also implicated in the external war story? To approach answers to these questions, we must once again weave through a complex set of hologeistic studies. The first in this set, interestingly, did not ask whether polygyny contributes to war, but whether war contributes to polygyny. Melvin Ember (1974) sought to explain the existence of polygyny, noting the availability of alternative, but not necessarily competing, hypotheses: (1) Long post-partum sex taboos, shown by John Whiting (1964) to be more frequent in protein-poor subsistence level societies, led men to acquire multiple wives as sex objects; and (2) Sex-ratios favoring women would set the stage for polygyny. Ember's (1975) study found more support for the sex-ratio hypothesis. Moreover, and importantly for our present concerns, he found an explanation for gender imbalances favoring women in the fact of warfare mortality. When Ember controlled statistically for male mortality in warfare, Whiting's (1964) relationship between a long post-partum sex taboo and polygyny disappeared. In more warlike societies that suffered high male mortality, polygyny was more common. Ember interpreted this to mean that polygyny was an adaptive response to warfare mortality, since polygyny served as a device for maximation of reproductive rates that could lead to a replenishing of a regularly depleted pool of potential warriors. Keeping in mind that the relationship found by Ember, like those in all hologeistic studies, is correlational, its directionality is a matter of conjecture. One could propose that polygyny sets the stage for warfare, rather than vice versa. Plausibly, one could argue that in polygynous societies, where males are shared by females, males are relatively dispensible and hence war and the male mortality which results from it would be relatively affordable. What Ember (1974) found was that warfare, mortality, sex ratios, and polygyny were interrelated. His suggested causal chain ran from warfare mortality to polygyny; we are offering, as a plausible alternative, one running from polygyny to warfare. Our hypothesis would be challenged by a well-known paper by Divale and Harris (1976) which argued that warfare is fought in order to remedy shortages of women, a notion that contradicts ours, and one that requires that warfare be more frequent in societies with sex ratios favoring males. While Divale and Harris presented some data showing that to be the case on a cross-cultural sample, a recent reanalysis of their data by Ember (1981) cast serious doubt on their conclusion. And when Ember in his 1981 paper compared 7 high warfare frequency societies with 8 low frequency ones as to their sex ratios, he clearly found female surplus significantly more often in the high warfare frequency societies. So, the best available evidence at present is that polygyny and warfare are positively related. Nevertheless, Ember (1981) does not point to polygyny as a "cause" of warfare; he merely used the sex ratio finding as a basis for rejecting the Divale and Harris (1976) hypothesis that fighting for women is a "cause" of warfare. If shortage of women is not a cause of war, struggle over other resources that might be in short supply could well be. This notion was recently challenged by Sillitoe (1977) who alleged to have demonstrated that among 26 small societies in New Guinea there was no relationship between warfare and population pressures. But a reanalysis by Ember (1981) of Sillitoe's own data showed that warfare was indeed more common in parts of New Guinea where population pressures were higher. This led Ember to argue that "people . . . go to war to try to gain access to scarce resources [p. 2]." This he dubbed an ecological explanation of warfare. Using
A Research Strategy
another ecological variable--famine or severe food shortage--Ember found this also related to frequency of warfare in New Guinea. So, land and food pressures may be contributing factors. Another popular line of argument regarding warfare has been that proneness to war is rooted in childhood socialization, with excessive punishment during childhood leading to an abundance of warlike personalities. Several studies support this line, including Russell (1972), Eckhardt (1973), and Prescott (1975). Some of the correlations reported in Textor (1967) also point in this direction. But, as Ember (1981) notes, if one employs societies for which Rohner (1975) provides ratings of parental acceptance/rejection and for which Ember has warfare frequency scores (18 societies), warfare is found to be more common among the "accepting" societies. Thus, this particular psychological explanation of warfare remains on shaky ground. Some support for it, however, is found in another recent cross-cultural study; but, again, its findings are very complex and the one concerning socialization, on which we are presently focusing, has to be viewed in context. Marc Ross (1980) contrasted two broad explanations for external warfare (and also for internal conflict). One of these was structural in nature, dealing with such variables as the existence of cross-cutting ties created by local exogamy (preference for marriage outside local communities), and by matrilocality (which disperses males following marriage), and patrilocality and polygyny which imply a lack of cross-cutting ties. The other was dispositional in nature and deals with such variables as severity of socialization, fostering of affection and warmth, and closeness/distance of father-child ties. He investigated these two classes of variables (structural and dispositional) in a sample of 90 societies from the Standard Cross-Cultural Sample (Murdock& White, 1969, 1980). A general finding was that dispositional variables were more important (accounted for more than structural variables). More specifically, lack of affection during childhood and harshness of socialization were both related to overt conflict, and independently of structure. Lack of affection was associated more with external conflict, and harshness of socialization (which includes aggression inculcation) was associated more with internal conflict, but these two socialization variables contributed to both kinds of conflict in both politically centralized and uncentralized societies. On the other hand, some structural variables also related to overt conflict, but in more complex ways. In uncentralized societies, polygyny and animal husbandry combined with the dispositional variables to account for internal conflict; but, in such societies, there was no effect of these variables on external warfare. In centralized societies, lack of checks on political authority and socioeconomic complexity combine with dispositional variables to account for internal conflict; for external conflict, polygyny and socioeconomic complexity are important. These findings can be better comprehended, perhaps, if reordered as follows: When Ross (1980) divided his sample into centralized and uncentralized societies, he found that in uncentralized societies,
1. child-rearing patterns that de-emphasize affection and warmth and training styles which are severe and aggressive correlate with conflict, both internal and external, and 2. the existence of cross-cutting ties limit internal conflict.
In centralized societies, he found that
1. the differences between internal and external conflict are greater, 2. internal conflict relates to severe socialization and to unchecked political authority, and 3. external warfare is associated with an early end to childhood and weak father-child ties, to greater socioeconomic complexity, and to polygyny.
Aggression in Global Perspective
For our present purposes, the most important fact to be gleaned from this very thorough study is that "societies using harsher and less affectionate socialization practices have higher levels of both internal and external conflict [p. 13]." This finding derived from an analysis in which both types of societies, centralized and uncentralized, were merged. In a subsequent analysis, the two kinds of societies were treated separately, and the harshness of the socialization variable was broken down into its several component variables (degree of corporal punishment, degree of pain infliction on infants, and degree to which fortitude and aggression are inculcated during childhood). For uncentralized societies, aggression inculcation (which we learn from Barry, Josephson, Lauer, and Marshall, 1976, 1980 relates to mother-child contact during infancy, degree of father absence, and importance of animal husbandry) is a very good predictor of external warfare. Along with lack of affection toward children, aggression inculcation accounted for nearly all of the variance in external warfare among the uncentralized societies in Ross's sample. Fortitude inculcation (but not aggression inculcation per se) was an important predictor of internal conflict. Combined once again with lack of affection toward children, it predicted nearly all of the variance in internal conflict among uncentralized societies. In centralized societies, as we have already noted, these socialization variables did better in accounting for internal conflict than for external warfare. But other socialization variables were, as already noted, associated with external warfare in centralized societies (e.g., an early end to childhood and weak father-child ties). Since none of the structural hypotheses mentioned earlier found consistent support across the two types of societies and across the two forms of conflict, while the dispositional hypotheses did, a tentative conclusion derivable from Ross's efforts is that future research on feuding and warfare which focuses on socialization emphases and consequent individual dispositions may well be fruitful. While structural variables surely can't be ignored either, they may best be treated as setting the stage for certain socialization emphases. For example, we know from Berry, Josephson, Lauer, and Marshall (1976, 1980) that importance of animal husbandry (a structural variable which Ross found to relate to internal conflict in uncentralized societies) predicts, and may set the stage for competitiveness, a component in the Barry et al. "toughness training," a variable which overlaps considerably with Ross's "aggression and toughness in socialization." So it does appear that a useful model is one in which structural variables are treated as ecocultural settings affecting the probability that certain socialization practices, emphases, and styles will prevail; and that these, in turn, affect the dispositions of adults to engage in overt conflict. Of course, one cannot preclude the possibility of direct effects of certain structural variables, such as polygyny, for example, which we have seen above relates to warfare frequency. But even where such relationships exist, it might be reasonable to expect complementary socialization effects. At least, Ross's work underscores the need to look for them.
Universally, some forms of aggression are defined as criminal. At least under certain conditions, assaults on persons or transgressions against their property are subject to negative sanctions. Industrialized societies, such as those described in this book, have elaborate criminal justice systems that are ostensibly designed to control crime. Debates continue in many of these nations regarding these systems and their effectiveness, or lack thereof. The attention paid to crime by our colleagues who prepared the chapters in this volume probably reflects the attention paid to it by the media and the public in all of the countries represented here. In smaller, subsistence-level and near subsistence-level societies, crimes are also recognized and dealt with, equally ineffectively, by various means. Clearly, then, those aggressive
A Research Strategy
acts defined as crimes occur universally, to a lesser or greater extent. It matters little that the nature of the acts defined as criminal may vary somewhat from society to society. Most striking is that, everywhere, some aggressive acts are perceived as unacceptable and dealt with by some institutionalized practices meant to keep them at some minimal level of occurrence. Thus, not only do sociologists in industrialized societies devote considerable attention to criminal behavior, so do anthropologists. While the anthropological literature on crime is not very large, it contains considerable material, both empirical findings and theoretical notions, that students of aggression in global perspective can ill afford to ignore. We shall here consider two studies, one employing ethnographic data from the Human Relations Area Files, the other a part of the six cultures study. From these we shall see (1) criminal behavior is mostly a male phenomenon, (2) it relates to both structural and socialization variables, and (3) sex-role identity may provide an explanation for criminal aggression. Despite some recent reports, mostly journalistic, of a narrowing of the gap between male and female crime statistics, it remains strikingly true that most crimes are committed by males. This sex difference is well documented in industrialized societies (in the United States, for example, in the 1970s, more than nine out of every ten crimes were committed by males), and anthropologists and cross-cultural psychologists who have studied crime hologeistically (e.g., Bacon, Child, & Barry, 1963) "have no doubt that this sex difference characterizes most societies [p. 292]." So, cross-cultural research that seeks childhood antecedents of criminal behavior has tended to focus on the childhood experiences of boys. This was so for Bacon, Child, and Barry (1963) who found 48 societies among a sample of 110 in the Human Relations Area Files for which comparative ratings on criminal behavior, socialization practices, and family structure could confidently be made. Their study yielded several findings of considerable interest. For example, and of primary importance, the 48 societies varied in frequency of crime of two types, theft and personal crime. These two types correlated with each other (R = .46) but, when examined separately, had different antecedent correlates, which we shall examine a bit later. For the moment, it should be stressed that the kinds of aggressive behavior which qualify as criminal, both stealing property and harming persons, occur to varying degrees in various nonliterate societies. Secondly, variation relates systematically to a feature of family structure and household composition in a way that suggests that an important antecedent to criminal behavior among males is a lack of opportunity for young boys to form an identification with the father. In monogomous nuclear family societies, which provide maximal presence of fathers during infancy and childhood, frequency of theft and of personal crime is relatively low. In polygynous mother-child household societies, characterized by maximal father absence, the frequencies are relatively high. Indeed, the frequency of both types of crime increase steadily over four forms of household characterized by decreasing degree of father presence, of which the two mentioned above are the extremes. This finding fits with data collected some years ago in the United States on juvenile delinquents (e.g., Glueck & Glueck, 1950), on lower-class culture (Miller, 1958) and with cross-cultural data on male initiation rites (Whiting, Kluckhohn, & Anthony, 1958) all of which point to an overcompensatory effort on the part of males who lacked early opportunity to identify as such to assert their masculinity in an aggressive way. Thus, while this finding provided by Bacon and his colleagues (1963) concerns a structural variable (household structure), a plausible psychological hypothesis (which might be dubbed compensatory machoism) can be adduced to account for it. The third finding to be noted from the Bacon et al. (1963) study is that, while both forms of crime relate to father absence, as was just detailed above, theft and personal crime have different patterns of socialization correlates. Theft is negatively correlated with childhood indul-
Aggression in Global Perspective
gence and positively correlated with socialization anxiety (or the degree to which punishment is employed and anxiety provoked during socialization). This finding may also be related to the U.S. data on delinquents (Glueck & Glueck, 1950) with both sets of findings suggesting that parenting styles that do not instill feelings of being loved lead to impulses to steal. As Bacon and his colleagues note, this hypothesis is similar to the Psychoanalytic View
of the motivation underlying kleptomania. Of more direct relevance to our concerns is a fourth finding, this one concerning personal crime, which involves attempts to injure or kill persons, assaults, rapes, and other forms of aggression and violence. The pattern of child-training factors which correlate positively with personal crime include dependence socialization anxiety (which is composed of abrupt transition to independence training, severity and frequency of punishment for dependency, and evidence of emotional disturbance in children surrounding independence training) plus prolonged and exclusive mother-child sleeping arrangements. Bacon et al. (1963) suggest that these experiences produce persistent attitudes of rivalry, distrust, and hostility (p. 298), a suggestion which finds support in another finding from their study that frequency of personal crime is negatively correlated with a measure of general trustfulness and positively correlated with hostility in folk tales, an index of adult suspiciousness about the social environment. This complex of findings from the Bacon et al. (1963) cross-cultural study provides the basis for an hypothesis about the sociopsychological conditions that might predispose males in any society to acts of personal aggression. The key conditions include any childhood experiences that set the stage for compensatory machismo, doubts concerning one's ability to function independently, and a view of the social environment as threatening, hostile, and not to be trusted. In the kinds of small societies that were included in the Bacon et al. study, the childhood experiences were perhaps not the sort that one finds readily in the kinds of societies featured in this book (e.g., exclusive mother-child sleeping arrangements), but there may well be some functionally identical or at least similar ones that exist, to varying degrees, in industrialized societies. For example, father absence, independence training, teachings regarding the nature of the social environment, and the instilling of trust are all variables that might be expected anywhere to produce the psychological states, including motives and attitudes, that Bacon and his colleagues have suggested contribute to aggression. Following on the Bacon, Child and Barry (1963) study, Beatrice Whiting (1965) pursued the masculine protest (or, as we have termed it, compensatory machismo) notion, adding to it the concept of status envy (identifying with adult males who are perceived as controlling valued resources). Reasoning that males reared during early childhood primarily by females will be most susceptible to this kind of status envy, she examined field notes collected in the six cultures study and found that the two societies in which physical assault and homicide were at high levels (Khalapur in India and Nyansongo in Kenya) were characterized by lower father salience and higher adult male prestige. More specifically, in both of these societies, husbands and wives do not regularly eat, sleep, work, or relax together. Infants seldom see their fathers. Fathers and all adult males enjoy many prerogatives. Both societies also have a tradition of extolling warriors and cattle-raiders. Note, for example, the following ethnographic observation about the Kenyan society:
The initiation rites that every adolescent boy in Nyansongo experiences stress the value of manliness--especially toughness and bravery. In the old days after initiation, a boy joined other warriors in the cattle villages and spent his time defending the herds and raiding other groups, stealing their cattle and sexually assaulting their women. [Whiting, 1965, p. 131] And, she concludes, it would seem as if there were a never-ending cycle. The separation of the sexes leads to a conflict of identity in the boy children, to unconscious fear of being feminine, which
A Research Strategy
leads to "protest masculinity," exaggeration of the difference between men and women, antagonism against and fear of women, male solidarity and hence to isolation of women and very young children [Whiting, 1965, p. 137]
This brief consideration of Whiting's findings marks the end of our effort in this chapter to extract some generalizations about human aggression from the anthropological literature. We have by no means finished the job. We have, instead, simply brought it to a temporary halt. We have halted in this abrupt fashion because we cannot complete this very substantial task in space available to us, nor is it our purpose in this chapter to do so. We are willing to stop here because we believe we have gone far enough to suggest the major aggression-relevant ideas which we believe reside in the anthropological literature. The reader will by now recognize that we believe that literature teaches us to take a functional, ecocultural perspective on aggression, to focus on both structural variables and socialization variables, to expect that parenting matters mightily, and to very much be concerned with relations between the sexes, if we expect ever to understand why human aggression is, on the one hand, as pervasive as it is and, on the other, so varied across human societies. Rather than overwhelm the reader with further details, we turn now to a presentation of a conceptual framework which we believe is compatible with the lessons from the anthropological research, one that can guide us and our colleagues in the necessarily continuing effort we must all make to study aggression in global perspective.
TOWARD CONCEPTUAL CLARITY; SOME DEFINITIONS The literature on aggression employs terms whose meaning is often imprecise. Communication among students of aggression suffers as a consequence. If we are to attempt a conceptual framework for future research, an effort to clarify the terminology used in that framework is useful. Following Strauss (1979), we suggest that some firm distinctions need to be made among such terms as conflict, hostility, aggression, and violence. While we cannot expect that our own definitions will erase the conceptual confusion that is present in the literature, or supersede the diverse, often competing conceptual schemes of others (e.g., Coser, 1956, and Dahrendorf, 1959, who interpret the same term--conflict--in very different ways), we hope at least to make clear what we mean by the various aggression-related terms that we use. And, by so doing, some theoretical implications of our framework may become more obvious. Further, it may assist the reader as he moves through the 18 case studies which comprise the largest part of this book. Conflict. We shall use conflict in the Dahrendorf (1959) and Strauss (1979) sense, i.e., to refer to conflict of interest, a bona fide disagreement between two or more persons or groups about allocation of resources of whatever kind. Thus, conflict is a situational matter, a state of affairs, a characteristic of an interaction involving persons and some feature(s) of their environment. Conflict in this sense is similar to a competition; the term says nothing about how the persons involved in it will behave. It merely notes the existence of a situation in which two or more persons want different outcomes, or want the same outcome when not all can have it. Any situation, then, in which A's anticipated gain is potentially at B's expense, or in which one's winning presupposes the other's losing, is what we mean by conflict. The only behavioral component in this concept is a cognitive one; we assume that the participants in a conflict will be aware, to at least some minimal degree, of its existence and their involvement in it. That
Aggression in Global Perspective
awareness, of course, can only be assumed or inferred from some behavior that appears to be a consequence of the conflict, especially some apparent effort to resolve it.
Conflict Resolution. Whenever a conflict exists, we assume the parties involved in it will attempt to resolve it in one way or another. Often, perhaps most often, they will each try to resolve it in their own favor, at least in the early stages of the attempted resolution. Whatever means are employed, these are, in our opinion, best thought of as strategies, tactics, and management procedures. They are behaviors, acts by the participants in the conflict that make the situation dynamic, moving it toward one kind or another of resolution. These acts are, for all participants, probably multiply determined; the determinants include such variables as the nature of the conflict, cultural or subcultural norms for conflict resolution, personality variables of the participants, and their short-term "states" including emotional levels. One such state could be, of course, anger. But more of that, later.
Frustration. By frustration, we prefer to indicate a feature of environmental states of affairs, rather than a state of the organism. However, it is that feature of situations, including and especially conflicts, which impinge directly on any person involved in them, which comprises a block to need or want satisfaction. It can lead to anger or to hostility and, usually whenever it does, it makes likely forms of behavior we call aggression.
Anger. By anger, we refer to a hypothetical construct, in the sense that it is a presumed-toexist state of the organism, inferred from some observed behavior, that denotes a feeling or an emotion often experienced by persons involved in conflicts or in other circumstances in which they find their needs or wants frustrated. We reserve the term for a relatively short-lived feeling, usually provoked by the behavior of other actors in the conflict, who serve as the source of frustration. Anger is usually directed toward those persons, although it may sometimes be transferred (displaced) to others.
Hostility. We prefer to employ hostility for another presumed-to-exist state of the organism that resembles anger in all respects except that hostility is longer lasting. If anger is an acute state, hostility tends to be chronic, existing long after its original instigation, and often directed rather diffusely, as in the case (for example) of enduring ethnic hostility. We conceive it as usually less intense than anger and often not as readily detected.
Aggression. Obviously, aggression is the key term in our conceptual framework; it is also the one that presents the most difficulties. In contrast with much popular use of the word, we use it to refer to any behavior by one person that inflicts harm on another. We would stretch that definition somewhat to allow aggression to include behavior which, by a consensus of observers, appears to have harm infliction as its intent. Thus, we would score as an aggressive act, pulling the trigger of a gun pointed at another person, even when the gun misfires and no visible harm results. But we are uneasy about the subjectivity of intent judgments and, for this reason, do not define aggression as behavior that stems from the intent to harm, as some students do. Better, we think, to define aggression as behavior that has harm infliction as its consequence, but allow consensual inference of harm infliction as intent. In any case, it is highly desirable to emphasize that aggression is a class label for behaviors, thereby avoiding any necessary connotations of underlying motives, intentions, or feelings, even though all such states of the organism performing the behaviors might be reasonable inferences. In our conceptual scheme, we will theorize about such states, but prefer to keep aggression as a relatively pure behavioral term, albeit one defined in terms of its primary consequence, harm infliction. By so
A Research Strategy
doing, we avoid some unnecessary problems: for example, Aronson (1972) distinguished between instrumental aggression (responding in a harm-inflicting way in order to get what one wants without wanting to hurt anyone) and intentional aggression (responses for which harm infliction is the desired end); Aronson urged reserving the label aggression for the latter. As we use the term, aggression covers both. While we see merits in Aronson's suggested distinction, we see it relating to matters of substance rather than of definition. We do, however, offer our own distinction centering around the concept of aggression and it involves our own term, protoaggression.
Protoaggression. By protoaggression, we refer to any behavior which has the following characteristics: intense, gross, diffuse efforts to satisfy needs, with those efforts sometimes directed against others. The feature of aggression which it lacks is any intent whatsoever to harm. Thus, it is different from Aronson's instrumental aggression, where, although harm infliction is not the desired end, it is nonetheless part of the presumed intentional state of the aggressor. Protoaggression includes no such intent. We will use the term primarily in reference to the behavior of human infants, who, we shall argue, often behave in"protoaggressive" ways, struggling diffusely and intently to get what they want without meaning to inflict harm, even in instances in which they (inadvertently) do. This concept will be discussed at some length later.
Violence. Like aggression, violence is behavior that inflicts harm. We distinguish it from aggression in only a few ways. We think of it as often (but not always) more intense than aggression, often having multiple perpetrators and victims. The violent actors often act in concert, sometimes in sanctioned, authorized, or agitator-led ways. Examples of violence include gang attacks, riots, large-scale vandalism, and, as a special case having many characteristics peculiar to it, warfare.
A CONCEPTUAL FRAMEWORK FOR THE STUDY OF AGGRESSION IN GLOBAL PERSPECTIVE The few terms defined above--conflict, conflict resolution, frustration, anger, hostility, aggression, protoaggression, and violence--are the cornerstones of our conceptual framework, which will now be set out. A few additional terms will be introduced when appropriate, but those (e.g., culture, socialization) are general concepts that pertain to much more than aggression . But to reveal our framework in the most meaningful fashion, we must begin with one of those more general concepts, culture. The Ecocultural Context Segall (1981) recently asserted that the overarching problem of cross-cultural psychology is to identify the various factors in the natural and man-made environments of humans which influence their behaviors in reliable ways. These factors comprise the context in which human behaviors occur. And that context includes some of the distal causes of those behaviors. The term culture (itself too gross, diffuse, and variously defined to be useful except heuristically) can be and is regularly dissected into various contextual factors, including basic institutions, social organizations, subsistence patterns, languages, rules governing interpersonal relations, educational systems, division of labor schemes, and the like. These sorts of variables comprise the man-made part of any human environment. In addition, there is the physical environ-
Aggression in Global Perspective
ment, and certain other man-made phenomena, like habitats (houses, tents, houseboats) and certain demographic features (e.g., population density) that are rather directly contingent on features of the physical environment. Together, all such characteristics comprise the ecocultural context, and all these characteristics constitute, potentially, the independent variables for cross-cultural psychology. Any one or any combination of them might, we assert at the outset, be relevant to our attempt to understand aggression. Nevertheless, we single out a few ecocultural contextual features as most pertinent for the study of aggression.
Probability of Conflicts Over Resources. Simply put, some environments are rich, some are poor. But within both rich and poor, allocation of resources may be more or less equitable, setting the stage for less or more conflict over their distribution.Thus, in the South Fore highlands of New Guinea, in the not-too-distant past, when the Fore people were hunters and gatherers, we are told by Sorenson (1978) that land was abundant, population sparse, resources sufficient for all, and, "aggression and conflict within communities . . . unusual and the subject of considerable comment when it occurred [p. 15]." By the same source, we are told, that later, with increasing population pressure and competition for agricultural land, adult fighting emerged [p. 16]. This example, which could be duplicated many times over, is the basis for our asserting that contexts vary, over both time and space, in the likelihood of realistic, intragroup conflict.
Probability of Frustration. Many features of a culture can determine the probability that numbers of individuals will experience frustration. Distribution of resources, mentioned above, is only one. Certain features of child-rearing practices comprise others. So do prevalent attitudes governing acquisition and consumption of valued goods. For example, in many large-scale technological societies, characterized by consumptionist values, reinforced by advertising and other marketing practices, unrealistically high levels of aspiration (for goods) may set the stage for high levels of personal frustration among large portions of a population. Conversely, communal ownership in a setting that lacks all but the most essential of commodities may produce lower levels of frustration. Together, the probability of conflict over resources and the probability of individual frustration, and whatever variables determine those probabilities, either make likely or prevent the occurrence of conflict and frustration. In this manner, the interpersonal situations in response to which aggression and violence may occur are culturally conditioned. Thus, the first set of linked concepts in our framework is as shown in figure 1.1.
Cultural Norms for Conflict-Resolution. Every society has traditions, institutions, and rules that pertain to the resolution of conflicts. These include systems for adjudicating disputes, when the prime actors in them are unable to reach a settlement. There are likely also to be widely shared expectations about the behaviors of persons engaged in such disputes. These expectations may be clearer in some societies than in others, and their actual content may differ. Various subgroups within societies may have different rules, as would be the case with various religious groups that might stress nonviolence or various regional groupings that might encourage vigilante-like efforts to resolve conflicts. Closely allied to these features are sanctions for or against aggression, which clearly vary across societies. For example, from data on maternal behavior collected in the six cultures study (Minturn & Lambert, 1964), marked differences could be found across six comparable communities in punishment for peer-directed aggression among children; mothers in an American New England town were very unlikely to do so. Numerous other examples could be adduced from the cross-cultural literature that sustains the proposition that societies vary in the nature of sanctions pertaining to
A Research Strategy
Ecocultural Context 1. Probability of conflicts over resources land availability, population density, etc. 2. Probability of frustration consumption aspirations distribution of goods
Interpersonal Situation Conflict competition over resources Frustration blockage of needs or wants, impinging especially on weaker actors in conflicts, the relatively poor
Fig. 1.1. Conceptual framework, Part One: Cultural roots of conflict and frustration aggressive behavior and in other norms that relate to strategies for conflict resolution. These would also include, as we saw earlier in this chapter, variations in aggression-inculcation. This proposition is illustrated in figure 1.2. Child-Rearing Emphases. The literature on differences in content, timing, targets, and agents of socialization across cultures is vast. Very little about socialization is universal beyond the bald fact that children everywhere are socialized. There is also evidence of considerable consistency across cultures in differentiation between the two sexes in child rearing emphases; e .g., /a society places more emphases on nurturance training for one sex than for the other, it is nearly always, if not always, females who receive more nurturance training than males (Barry, Bacon, & Child, 1957). But, even in this regard it must be noted that some societies differentiate between the sexes much more than do other societies (Barry, Child, & Bacon, 1959). So, it is clearly safe to state that one of the most striking differences across cultures is in child-rearing practices. In a variety of rather obvious ways, such differences have implications for aggression, by affecting the experiences individuals are likely to have growing up. A culture and personality framework, whether of a Freudian variety or of a more contemporary learning theory sort, provides linkages that run from socialization emphases to individual experiences to prevailing behavioral dispositions. We shall illustrate such linkages in our own framework, but not before introducing one more cultural context variable. Availability of Models. In any society, children learn not only from deliberate teaching, as occurs in socialization, but in other ways as well. Among those other ways surely is emulation of models, including adults, older siblings, and some peers under certain circumstances. That
AP Goldstein, MH Segall