HUMAN VIOLENCE AND AGGRESSION:
AN ANNOTATED BIBLIOGRAPHY FROM A PSYCHOLOGICAL PERSPECTIVE
Prepared by Clark McCauley, Co-Director, Solomon Asch Center
Violence between individuals
Impulsive versus instrumental aggression
Individual differences in aggression
Cultural differences in aggression
Cultural regularities in aggression
Frustration-aggression theory in group conflict
Iniquity theory in group conflict
Patriotism, nationalism, and ethnic identification
Group centrism in public opinion
Social identity: the minimal group effect
Perceptions of ethnic and national groups
The contact hypothesis
Men in combat
Violence in the system: dominance and hegemony
Some social psychology of extreme behavior
Other analyses of extreme behavior
Escalation as a social trap
Understanding human aggression and violence requires an initial and crucial distinction between individual violence and group violence. In general terms, it is the least socialized who are disproportionately involved in individual violence, whereas it is often the best socialized who are involved in intergroup violence. This is not to say that well socialized members of a culture are never involved in violence between individuals, nor that poorly socialized members of a culture never appear in group conflict. Nor is it to say that social norms and group dynamics are irrelevant to individual violence. Still, the distinction is important because the origins of individual violence are mostly to be found in individual differences that tell us little about group conflict. Instead the origins of group conflict are to found in social perceptions of groups (stereotypes), social constructions of events, group dynamics and cultural norms. The first section of this bibliography deals with individual violence and aggression, the second section with intergroup violence and aggression.
Violence is not only in individuals and groups, however; it is often in the system--the rules of the game, the whole culture. Relativists may say that no culture is better than another, that the only test of culture is survival. This bibliography takes a different view: that some cultures dominate and dehumanize in ways that are worth fighting against. The third section of the bibliography deals with this kind of violence in an effort to understand how a system or culture does violence in controlling the very definition of reality.
This bibliography was prepared by a social psychologist who has tried to learn some of what other social scientists have to teach. The result is inevitably one person’s vision, and the author apologizes in advance to all whose work has been inadequately represented.
VIOLENCE BETWEEN INDIVIDUALS
Caution: This literature depends almost entirely on research with North American and European subjects.
This is American psychology’s most popular contribution to understanding aggression. Battered and revised as a theory of individual aggression, it remains popular in various theories of intergroup conflict (see below).
Dollard, J., Doob, L, Miller, N., Mowrer, O. & Sears, R. (1939). Frustration and Aggression, New Haven: Yale University Press.
The original statement of the theory was that frustration (interference with goal response) is necessary and sufficient for aggression (behavior aimed at harming another member of same species). Much of the original statement of the theory was a translation of Freudian concepts into the animal learning theory of the day.
Miller early recognized that the theory was too strong. Aggression could occur without prior frustration (frustration not necessary); frustration could be followed by something other than aggression (e.g. problem solving) so that frustration is not sufficient for occurrence of aggression.
The next two papers are classic studies of the F-A hypothesis, still often and inaccurately cited in secondary literature.
Often cited as showing frustration leading to hostility, this study actually showed frustration leading to less positive attitudes toward outgroups, with no change in negative attitudes. Boys at a summer camp were promised a trip to town for a movie, then denied the trip at the last minute. Attitudes toward various outgroups were measured by having boys check personality traits to describe these groups; in a pretest-posttest design, fewer positive (but not more negative) traits were checked following the frustration.
Often cited as showing a correlation in U.S. southern states between lynching blacks and low price of cotton. Conceptually, it is difficult to take low cotton price as a frustration; what is the blocked goal response?
This update of the previous study uses better statistical techniques to show that the relation between economic measures and lynching is weak and that this weak relation holds for white as well as black victims of lynching.
This review aims to salvage the F-A hypothesis by distinguishing two kinds of aggression: instrumental and impulsive. Instrumental aggression is used as a means to some other goal; impulsive or angry aggression is aimed at hurting someone (the reinforcer is the perception of distress). This important distinction is of considerable current importance (see below).
Berkowitz revised F-A theory to make explicit that it applies only to impulsive aggression. The theory is further revised to predict that frustration is necessary and sufficient for the instigation to aggression, but whether this instigation (anger) is expressed in aggressive behavior depends upon the reinforcement history of the individual and upon situational cues previously associated with aggression.
This was the first of many studies to note that frustration usually has to be seen as arbitrary--unexpected and unjustified--for aggression to follow. For instance, the bus that spashes by you without stopping is an instigation to aggression if it has your destination on its signboard but not if the signboard says "Out of Service."
Zillmann, D. (1979). Hostility and aggression. Hillsdale NJ: Erlbaum.
A masterful review and critique of F-A theory.
Berkowitz once again tries to save F-A theory by reformulating to say that any kind of noxious experience produces an instigation to impulsive aggression (anger); frustration is only one type of noxious experience but some kind of noxious experience is now necessary for impulsive aggression. Whether impulsive aggression actually occurs depends on the cues and contingencies in the situation.
This formulation can deal with evidence that violent crime increases in hot/muggy weather. Still trying to maintain behaviorist purity over mentalist fuzziness, Berkowitz tries to deal with the role of arbitrariness by saying that unexpected frustration is a greater frustration than expected frustration, and unjustified frustrations are greater only to the extent that they are unexpected. This line avoids having to deal with anger as a moral emotion, but suffers some implausibilities in relation to examples such as Pastore offers. For instance, it is not clear that having your bus splash on by you is more uncommon (less expected) than having an out-of-service bus do the same.
In The Rhetoric, Book 2, Aristotle argued that the source of human anger and aggression is the perception of insult. This theory differs from F-A theory in emphasizing that anger follows a moral appraisal that is sensitive to social norms: the definition of insult differs enormously across cultures and centuries. For Aristotle, the moral judgment is the cause of anger and motivation to hurt; for F-A theory, the moral judgment is an epiphenomenon or rationalization of a blind reaction to noxious stimulation. Aristotle’s theory remains at the core of some modern psychological analyses of anger and aggression.
Insult-anger theory can deal easily with results such as those of Pastore that indicate that anger follows not just any but especially unjustified frustration. The out-of-service bus splashing by you is not an insult the way your bus splashing by you is. F-A theory says that we feel anger and instigation to aggression no matter which bus goes by, but that we suppress our anger (another frustration!) at the out-of-service bus because we have no rationalization available. Insult-anger theory recognizes that we may be swallowing some anger even at the out-of-service bus because it is to a degree still insulting. It reminds us of our low status: we are dependent on the bus while others drive.
Similarly, encountering a red light when we are in a hurry is a reminder of low status and lack of control: important people are swept through red lights by police escorts. Physical infirmity, heat, cold, and hunger remind us of the weakness of the body: our bodies make us animals. Noxious events small and large insult us with the reminder that we are subject to what we cannot control, including that final uncontrolled loss that is death.
Insult-anger theory and F-A theory are agreed that impulsive aggression--the experience of anger and the desire to hurt--are linked to noxious experience and the perception of transgression. The issue, unresolved in the literature, is whether the moral appraisal is causal or only rationalization.
Sabini, John. (1995). Social Psychology, 2nd Edition. New York: Norton. Chapter 13, "Aggression," provides a succinct and readable comparison of F-A theory and insult-anger theory.
Averill asked people about specific incidents in which they had been angered in the past. About 80% of those reporting being angered by some particular instigator also reported reacting with some form of aggression. The goal of the aggression, and of other behaviors following instigation, was more often to restore self-image than to change the instigator’s behavior. According to Averill (1982, p.177), the most commonly reported focus of anger is "to reassert...authority or independence, and to improve (one’s) image."
IMPULSIVE vs INSTRUMENTAL AGGRESSION
The distinction between impulsive and instrumental aggression, which came out of research on the F-A hypothesis, is now being put to work to further understanding of sociopathy, childhood aggression, and violent crime
Irritable/impulsive but not instrumental aggression is related to low levels of serotonin in the brain
Psychopathy is related to use of instrumental rather than impulsive aggression. Psychopaths use aggression coldly, as a means of controlling others
Reactive and proactive aggression correspond to impulsive and instrumental aggression.
Juveniles with histories of impulsive violence are less likely recidivists than juveniles with histories of instrumental (cold) violence.
INDIVIDUAL DIFFERENCES IN AGGRESSION
Individual differences in aggression, assessed by peer ratings or teacher ratings, are almost as stable as IQ scores; from gradeschool to teen years, retest correlations are about 0.50.
Across cultures, males are more physically aggressive than females. Recently, a different form of aggression, "relational aggression" , has been recognized. Relational aggression is aimed at hurting another’s relations with others, by mean words, gossip, and social exclusion. The class of behaviors called aggression is defined by intention to hurt others, and there is little doubt that humans can aim to hurt one another with words as well as blows. Relational aggression is more common in females than in males, such that total of relational and physical aggression may be little different for females than for males. Relational aggression, like physical aggression, may be either impulsive or instrumental.
The authors review a great deal of evidence that low IQ is a strong predictor of the likelihood of perpetrating violent crime (or at least being convicted of perpetrating violent crime). The cognitive contributions to aggressive behavior are explored in the next two papers.
This is a dual-process theory of the linkage between aggressiveness and low IQ. The first process operates early in life when the low IQ child is frustrated in everyday problem-solving, especially in school; frustration then leads to aggression and the learning of mostly aggressive strategies or "scripts" for responding to frustration. Once learned, these aggressive strategies tend to interfere with the child’s relationships with peers and teachers. In this second process, aggression leads to continued poor performance in school and continued failure to take advantage of intellectual opportunities. Thus low IQ leads to aggression and then aggression interferes with intellectual development. Although plausible, this theory mostly depends on the results of a single longitudinal study that found aggressiveness at age 8 negatively correlated with IQ at age 8 and, more surprisingly, that age 8 aggressiveness was better than age 8 IQ as a predictor of intellectual functioning at age 30.
Dodge and his colleagues have found that more aggressive children have a hostile-attribution bias that leads them to see others as hostile and malevolent; their aggression is then a reaction to this attribution. This hostile-attribution bias is associated with greater sensitivity to cues of hostile intent, greater likelihood of interpreting ambiguous behaviors as hostile, easier access to aggressive scripts, and more practice enacting aggressive behaviors. Dodge is currently testing his theory with a multi-million dollar grant to help highly aggressive children with a kind of cognitive therapy that will change their attributional style.
Katz, J. (1988). Seductions of crime. New York: Basic Books.
A useful contrast to the emphasis on the cognitive side of aggression is Katz’s account of the emotional side of crime, especially violent crime. Of course there is a cognitive side to aggression--cues, attributions and scripts--but it is easy to loose sight of the emotional side of aggression in the pull of the cognitive wave that has overtaken psychology in general and social psychology in particular. Katz provides a counterweight to the cognitive emphasis of Huesman and Dodge; he deconstructs the idea that career criminals use violence coldly--in instrumental aggression. From interviews with violent criminals, Katz highlights the powerful moral emotions that are involved in their violence. "Doing stickup", in particular, brings the glory of respect on the street and high status in the criminal world, brings the thrill of risk taking from the sizeable minority of victims who offer some resistance, and brings, least for a while, feelings of dominance and control to cover feelings of humiliation and shame.
CULTURAL DIFFERENCES IN AGGRESSION
These authors brought together crime statistics from 110 nations and 44 major cities, for 1900 to 1970. Cross national comparisons of rates of homicide, rape, and criminal assault indicated that the U.S. is relatively high compared with other industrialized nations; among these nations, the U.S. is highest in homicide rates. The high homicide rate in the U.S. probably has something to do with the easy availability of firearms, although questionnaire studies indicate that U.S. respondents are also more likely than respondents of other nations to suggest aggressive solutions to interpersonal conflicts. Archer and Gartner find that more violent crime is associated with waging and winning wars: nations participating in WWI and WWII, especially nations on the winning side in these wars, show increases in homicide after the war is over. Their interpretation is that social norms about use of violence are liberalized when use of violence is rewarded.
Nisbett examines crime statistics and stages laboratory experiments to show that white non-Hispanic males in the Southern U.S. states are more likely than other Americans to respond to threats and insults with violence. In a celebrated experimental assessment, Nisbett arranges for students at the University of Michigan to be bumped and called an "asshole" by an experimental accomplice. Students from southern states react more aggressively than other students. Nisbett’s interpretation is that southern students participate in a culture of violence originating in the need of pig farmers to deter rustling their stock. Another possible interpretation is that southern norms of politeness are higher than norms in other states, so that the same provocation is a greater norm violation for southerners.
CULTURAL REGULARITIES IN AGGRESSION
Daly, M., & Wilson, M. (1988). Homicide. New York: A. De Gruyter.
An evolutionary approach to aggression leads to the expectation that there will be some powerful cross-cultural consistencies in violence and aggression. Where genes are shared, violence should be less likely; where genetic interests are in conflict, conflict should be more likely. There is some evidence consistent with these predictions.
Sociologists and criminologists have long noted the high incidence of homicide among family members. Of course there is more opportunity for conflict where interaction is higher, but the pattern of violence within families is telling. Preschool children living with at least one parent genetically unrelated (foster parent, step-parent) are 70-100 times more likely to be abused than children living with both natural parents. Risk of homicide for unrelated people living together is 10-11 times higher than for genetically related people sharing a residence.
Males in the U.S. consistently commit about 80% of all homicides. A similar sex difference in homicide is found in most nations studied, and in the same nation over decades and even centuries. This sex difference is consistent with genetic interests of the two sexes: males can father hundreds and even thousands whereas females can give birth to dozens at most. It follows that males should compete more for status (linked to reproduction) than females. Indeed male violence is particularly high among males who are young, unmarried, and poor. Male-to-male violence is most often a response to humiliation (see above F-A theory vs. Insult-injury theory). Male to female violence is predominantly associated with male sexual jealousy. These statistical tendencies seem to be quite reliable cross-culturally.
Of course evolutionary theory does not assert that reproductive interests are the conscious reasons for violence. The link between reproductive interest and the psychology of the perpetrator is usually left unaddressed in the evolutionary approach. This approach says only that prediction of violence can be improved by noting that people act as if they were aware of their reproductive interests.
There are many forms of intergroup violence, ranging from street gangs and soccer hooligans to modern warfare between nations in arms. Theories of intergroup conflict usually aspire to span this entire range. My view is that the larger the groups involved in conflict, the more organizational and leadership psychology is implicated; that is, the larger the groups in conflict, the greater the difference in the interests of leaders and led. In small groups, leaders are usually running the same risks as followers or even greater risks; in larger groups, especially at the level of the nation state, leaders risk loss of status and power but not their lives. Nevertheless, this review will largely ignore the leadership psychology and organizational psychology in assuming that the dynamics of intergroup conflict are substantially the same for gangs as for nation states.
FRUSTRATION-AGGRESSION THEORY IN GROUP CONFLICT
Battered and revised as a theory of individual aggression, F-A theory remains popular in various theories of intergroup conflict.
In his classic experiment, Sherif assigned boys to two groups matched for school and playground skills and took them separately to Robbers Cave State Park. For the first week each group hiked, camped, and played without knowledge of the other group. Then one group was allowed to see that another group had been playing on "our" ballfield. Sheriff introduced both groups to a weeklong tournament, with jacknives as prizes for the winners. By the end of the tournament, intergroup hostility and negative stereotypes were strong. The tournament losers decided that the winners had not played fair, and burnt the winners’ banner. Hostilities escalated, and Sherif had to intervene to prevent physical violence. In the third phase of the experiment Sherif tested ways of reducing hostility. Bringing the groups together for a party was a failure; the group arriving first took all the best goodies and the other group reacted with more hostility. Sherif succeeded finally by staging several "emergencies" (supply truck stuck, water supply cut off) in which boys from both groups had to cooperate for common goals. Hostility was reduced and the boys of both groups voted to ride back to the city on the same bus.
Sherif’s interprets his results as supporting "realistic group conflict theory", which says that competition over resources (pen knives) leads to intergroup hostility and aggression. At bottom, this is a frustration-aggression theory: competition means a zero-sum game in which one side’s gain is the other side’s frustration. Cooperation for common goals is the antidote.
Gurr, T. (1970). Why men rebel. Princeton, NJ: Princeton University Press.
Relative deprivation theory is more psychological than realistic group conflict theory; it says that the source of hostility toward another group is the perception that the status or welfare or one’s own group is less than it should be in relation to the outgroup. The deprivation may be in relation either to the level expected or the level desired. Again, this is a frustration-aggression theory: relative deprivation is understood as a form of frustration. Relative deprivation theory is prominent in political science, where it can explain why rebellions occur, not when economic stress is highest, but when life is improving (e.g. French revolution). If the improvement for a group is slower than they expect, their experience of frustration leads to aggression. In Sherif’s experiment, relative deprivation suggests that group conflict might have emerged if there were no prizes (penknives), nothing at stake except the status of "winner."
This study shows that violence in Northern Ireland is not associated with economic ups and downs; rather violence is associated with re-negotiations of political ?rules of the game?. Thus the results are inconsistent with realistic group conflict theory but consistent with relative deprivation theory if Protestant vs. Catholic political arrangements are seen as expressions of relative group status.
INIQUITY THEORY IN GROUP CONFLICT
Iniquity theory is still more psychological than relative deprivation theory in holding that it is the social construction of moral violation that is the immediate cause of intergroup hostility and conflict. Iniquity theory is an extension of Aristotle’s insult-anger theory to intergroup conflict. Iniquity suggests that a crucial origin of hostility in the Sherif experiment was the judgment by the losing group that the winners had not played fair. Knauft, an anthropologist, provides another kind of example. Gebusi villagers of New Guinea experience many incursions of others’ pigs in their garden; only when the incursion is interpreted (or constructed) as willful and insulting do the villagers attack the village from which the pigs came.
Similarly, the theory notes that, in persuading America and its allies to support violence against Saddam Hussein, President Bush spoke little of oil (realistic group conflict theory) and little of disrespect from a nation the U.S. had helped against Iran (relative deprivation theory). Rather President Bush spoke of moral violation: of a new Holocaust led by a new Hitler. He spoke in particular of a violation of community. According to Fiske, community-sharing is one of four possible models of human relationships (along with authority-ranking, equality-matching, and market-pricing). It seems possible that Fiske’s four models are also a taxonomy of moral violation.
Welch, a political scientist, has proposed that "most decisions by great powers to go to war in the last 150 years have been driven not by security and power goals but by a concern for ?justice.? To activate the justice motive, one must convince oneself that the other side threatens something--territory, resources, status--to which one is entitled. The resulting reaction is not cold, rational, and calculating, but rather emotional, self-righteous, moralistic, and simplistic. Outrage triggered by perceived threats to entitlements provides the psychological momentum for dehumanizing adversaries, deactivating the normative constraints on killilng them, and taking big risks to achieve ambitious objectives." (Tetlock, 1998, p. 884).
Whether or not President Bush was personally outraged by the invasion of Kuwait, it is clear that moral outrage was his means of bringing the U.S. public and the U.S. Congress behind him in the Gulf War. This observation is a challenge to realistic group conflict theory and relative deprivation theory, for which moral outrage is epiphenomenal. Just as F-A theory at the individual level maintains that insult and anger are only rationalizations, so realistic group conflict theory and relative deprivation theory maintain that moral outrage is only rationalization of group conflicts originating in the blind linkage of frustration with aggression. As Tetlock points out, the long-time controversy over whether ideas can be reduced to interests is difficult to resolve; the line between moral entitlement and material interest is often psychologically blurry. Nevertheless, this review leans toward seeing a clear causal role for moral outrage in President Bush’s argument for the Gulf War.
Becker, Ernest. (1973). The denial of death. New York: Free Press.
Iniquity theory asserts that moral appraisal is the proximate cause of anger and hostility between groups. This moral basis makes sense of the fact that individuals will sacrifice themselves for their ethnic and national groups, and makes sense as well of the important role of religious differences in ethnopolitical conflicts. Religion is that part of culture that offers the most explicit answer to the question, What does it mean that I am going to die? So far as we know, man is the only animal to face this knowledge, and Becker has argued that it is our mortality, rather than our sexual or aggressive impulses, that we most need to repress. Our mortality is linked to animality; all animals die. Not just religion but all of culture distinguishes man from other animals; thus all of our culture functions to defend us against our own mortality. So far as we know, no sane human has ever had the idea that he or she alone was destined for some form of immortality; rather it is participation in and contribution to a group--family, religion, nation, cause--that offers some form of immortality.
Bloch, an anthropologist, describes the link between religion and violence in terms of "rebounding violence". He describes rituals in which the spirit world first attacks and kills the pig in us--the selfish animal self--then invigorates us to conquer the animal world with the power of the spirit world. The expression of this spiritual power in the ritual group is often an attack on nearby groups. It is not difficult to extend Bloch’s model to the rituals of major religions and the violence of one religion against another. The power of the spirit world, its immortal right order, is made manifest and persuasive in conquering those who do not share in it.
The social psychological version of this insight is represented in about three dozens studies by Greenberg and his collaborators. They show that individuals who are asked to write about their own death (versus writing about TV) are more hostile toward those who violate cultural norms and more positive toward those who uphold the norms. For instance, American students given the death-salience manipulation give higher evaluations to a pro-American essay and its writer, and lower evaluations to an anti-American essay and its writer. Similar results are obtained with death reminders presented subliminally. This is the beginning of understanding the power of patriotism and nationalism in the same way as we understand the power of religion: death in the service of the group that offers immortality is preferable to life without meaning, that is, life as just another animal.
PATRIOTISM, NATIONALISM, AND ETHNIC IDENTIFICATION
It must be acknowledged from the outset that social psychology has had little to say about the origins and consequences of attachment to national and ethnic groups. "Nationalism" and "patriotism" do not appear in the index of the 1998 Handbook of Social Psychology.
Feshbach is one of a number of scholars who have sought to distinguish patriotism, a good thing (e.g. "I love my country") from nationalism, a bad thing (e.g. "Other countries should try to make their government as much like ours as possible). Feshbach devised a questionnaire with separate scales for patriotism and nationalism and reported a low correlation between them. In his 1991 report, however, he found a correlation of .61 between the two scales. The distinction between loving one’s country and thinking one’s country superior is not well supported.
A limitation of the questionnaire approach is that it focuses on individual differences in attachment to state or nation. The phenomenon of interest is rather how the whole of a public is mobilized to an increased level of patriotism, that is, how the distribution of patriotism scores can be moved up the scale. On this point, there is little research.
In larger perspective, patriotism and nationalism are part of a larger phenomenon of group identification. The paradigm case of group identification is caring for one’s family. Identification with family is generally seen as unremarkable; after all, we share directly in the good and bad things that happen to our family. It is not so easy to say why we easily come to care about the successes and failures of sports teams, ethnic groups, and national or state groups. In these larger groups, the outcomes of the group are not immediately and obviously linked to individual outcomes. It seems that humans are prepared, perhaps biologically prepared, ready and even eager to pour themselves into others. Economic man, maximizing short-run self-interest, is not easy to find in our relation to groups.
GROUP CENTRISM IN PUBLIC OPINION
The importance of group identification has emerged in public opinion research, where it has been determined that the key to predicting voting behavior is not the individual interest of the voter but the voter’s attachment to various political groups and causes. Here is Kinder’s overview (citations omitted in this and subsequent quotations from Kinder). "The unemployed do not line up behiind policies designed to alleviate economic distress. The medically indigent are no more likely to favor government health insurance than are the full insured. Parents of children enrolled in public schools are generally no more supportive of government aid to education than are other citizens. Americans who are subject to th draft are not especially opposed to military intervention or to the escalation of conflicts already under way....On such diverse matters as racial busing for the purpose of school desegretation, antidrinking ordinances, mandatoryh college examinations, housing policy, bilingual education, compliance wiht laws, satisfaction with the resolution of legal disputes, gun control, and more, self-interest turns out to be quite unimportant."
Instead political opinions are predicted by group identification: by feelings for and against politically salient groups. Here is Kinder again. "Support among whites for affimative action or school desegregation reflects sympathy for the plight of blacks. Opposition to social welfare programs derives from hostility toward the poor. Support for a tough foreign policy reflects fear of communists."
Kinder’s summary: "American public opinion turns substantially on the beliefs and feelings that Americans invest in social groups--membership groupos they belong to and perhaps identify with, as well as reference groups that elicit their tympathy or resentment....In matters of public opinion, citizens seem to be asking themselves not ?What’s in it for me?? but rather ?What’s in it for my group?" (As well as "What’s in it for other groups?"). "The power and persistance of group-centrism in public opinion amounts to a dramatic disconfirmation of the expectation, issued so confidently not so long ago, that categories of race and ethnicity and religion and the like were about to become obsolete. Thanks haven’t turned out that way--not in the United States, and not around the world, where conflict organized around social and cultural difference has become a murderous commonplace."
SOCIAL IDENTITY: THE MINIMAL GROUP EFFECT
Tajfel and his collaborators (see Brown) have shown in many studies that arbitrary division of individuals into two groups, even by flipping a coin, is enough to generate ingroup preference. Individuals prefer the products of their own group, and favor members of their own group when distributing rewards and costs. The interpretation offered by Tajfel and his collaborators is that a purely cognitive and perceptual manipulation has motivational effects in producing in-group bias. The theory assumes that our self esteem depends to great extent on the status and success of the groups we belong to; if we belong to a group, however arbitrary, it must be above-average.
This interpretation must be qualified in two ways. First, the ingroup bias measured in minimal groups experiments never goes as far as hostility toward the outgroup; the ingroup is seen as more positive but the outgroup remains positive. Second, studies 2-5 in the above list show that the minimal group manipulation works because the the groups created by the experimenter believe that the experimenter will be treating these groups differently. That is, members of the same group believe that they share "common fate". Thus it is perceived interdependence, the motivational source of cohesion recognized by Kurt Lewin and his followers in group dynamics research, that is at the bottom of minimal group effects. In effect, the minimal groups procedure is an operational but not a theoretical minimum for group identification.
PERCEPTIONS OF ETHNIC AND NATIONAL GROUPS (STEREOTYPES)
Lippmann, W. Public Opinion. New York: Harcourt, Brace, 1922.
At least as far back as Plato, theorists have recognized that we react not to the world as it is but to the world as we perceive it. Lippmann introduced the word "stereotype" to refer to the over-simplified "pictures in our heads" that guide our behavior in the social world.
Gilbert, G. M. (1951). Stereotype persistence and change among college students. Journal of Abnormal and Social Psychology, 46, 245-254.
These three classic studies are typical of the first thirty years of stereotype research in asking subjects to check the five traits "most typical" of the stereotyped group. Results indicated that stereotype strength (agreement across respondents in trait choices) was as strong for unfamiliar groups (Turks) as for familiar groups (English); the implication is that stereotypes are largely socially constructed, rather than learned from personal experience with members of the stereotyped group. Stereotypes of Japan and Germany became more negative in association with WWII, but recovered to generally positive by the 1960s. For U.S. respondents, there was a long-term trend toward avoiding or denying negative stereotypes of ethnic or national groups. The three studies are also typical in decrying stereotypes as inaccurate, based on hearsay, and irrationally resistant to information about individuals who contradict stereotypic expectations.
These essays opened a new approach to stereotypes as probabilistic predictions about group differences. Lippmann’s metaphor of stereotypes as pictures was too simple: almost no one has 100% all-or-none beliefs about group differences. Once recognized as probabilistic, stereotypes are not obviously incorrect and, statistically, should not change in response to encounters with small and non-random samples of individuals from the stereotyped population.
Recent research indicates that, counter to the "kernel of truth hypothesis", stereotypes are not generally exaggerations of real group differences. Social perception does not seem to show automatic sharpening and leveling at category boundaries the way physical perception does. However, mutual stereotyping of groups in conflict may exaggerate real group differences.
Stereotypes are not the same as attitudes; two individuals can agree about group differences (e.g. between black and white Americans) but have different feelings toward the two groups associated with different theories about the causes (genetic, cultural, economic) of the differences. Stereotypes may include theories of the origins of group differences.
Current research is beginning to explore the cognitive developmental origins of human conceptions of race and ethnicity.
THE CONTACT HYPOTHESIS
This hypothesis has been social psychology’s perennial favorite for reducing negative stereotyping and intergroup hostility. It can be seen at work in programs to improve relations between black and white in America, between Jews and Arabs in Israel, and between Protestant and Catholic in Northern Ireland.
Allport, G. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley.
The classic formulation of the contact hypothesis: intergroup contact that is interpersonal, equal-status, cooperative, and supported by authority will lead to reduction of intergroup hostility and negative stereotyping.
Mutual stereotyping of Greek and American undergraduates was assessed before and after Americans spent a semester at the University of Athens; stereotyping increased after contact.
Contact between white and black students in desegregated schools did not lead to reduction of stereotyping or more positive attitudes. In retrospect, it seems clear that the conditions of the contact hypothesis were not met: contact was not equal-status (black students not equal in school performance) and occasions of cooperation for common goals were weak.
Various failures of contact predictions have led to a multiplication of the conditions required for the hypothesis to apply; Stephan now lists approximately a dozen such conditions. Taken together these conditions amount to saying that the contact hypothesis applies only to intergroup contact in which, on average, the representatives of the two groups do not differ in any way that either group cares about.
This classic experiment is often cited as supporting the contact hypothesis: hostilities between two groups of boys was reduced by arranging contact (responding to staged emergencies) that was interpersonal, equal-status, and cooperative. Generalizability of the result is limited by the fact that the two groups were equated by random assignment at the beginning of the experiment; that is, the two groups began as similar as the experimenter could make them. Thus groups that do not differ can be brought by competition to conflict and negative stereotyping, and contact under the right conditions can reduce conflict and stereotyping when the two groups do not in fact differ.
A magisterial overview of the history and vicissitudes of the contact hypothesis.
One of the most realistic studies of prejudice ever conducted. Applicants for a blue-collar job generally preferred to work with someone of a different race who shared their values than with someone of the same race who did not share their values. Rokeach argues that racism is not an irrational reaction to skin color or physiognomy but a reaction to cultural difference associated with race.
This review distinguishes between a strong and a weak form of Rokeach’s theory of prejudice. The strong form says that racial prejudice is entirely a matter of cultural difference; the weak form says that at least a substantial part of racial prejudice is perception or expectation of cultural difference. The review concludes that there is good evidence for the weak form, but not for the strong form.
Schelling, T. C. (1978). Micromotives and macrobehavior. New York: Norton.
Chapter 5, "Sorting and Mixing: Race and Sex," offers an important reminder of how small a preference for similarity can result in strongly segregated groups. Schelling describes a game with checkers on a checkerboard. Arrange black and red checkers randomly on the board. Consider each checker to be the center of a 3x3 neighborhood with thus 8 neighbors. Assume each checker wants at least half of his neighbors to be the same color as himself. Remove perhaps 5 checkers randomly to create some holes on the board. Move checkers one at a time to satisfy the the desire not to be a minority. The surprising and usual result is strong segregation.
This demonstration is important as a counter to the assumption that strong segregation means strong preferences for similarity and against difference. A very moderate preference for similarity can produce striking levels of group segregation.
This research contributes to understanding how normal and even idealistic individuals can become capable of atrocities against noncombatants, including women and children. The logic of terrorist groups starts with the French revolution: the people are the state, enemies of the people are targeted by class membership, and the army is the nation in arms.
Early research on terrorism sought explanations in individual psychology and psychopathology; this reivew argues that terrorism is instead a phenomenon of intense but normal group dynamics.
Describes the trajectory of political groups that seek first political reform, then overthrow of an unresponsive political system, and finally a radical political vision to be realized through terrorism against anyone who does not join their assault on the state.
Summarizes research on terrorism for the attention of policy makers. A terrorist group is described as the activist apex of a pyramid of more passive supporters and sympathizers; the relations between the terrorist group and its sympathizers is key to its strategies and its survival.
MEN IN COMBAT
van Crevald, M. (1982). Fighting power. Westport, CN: Greenwood Press.
Holmes, R. (1985). Acts of war: The behavior of men in battle. New York: Free Press.
How do men advance and fight, when self interest seems to dictate avoiding combat and running away? Three kinds of answer have been given serious attention: because it is safer to fight than to run, because men don’t want to let down their buddies, and because men believe in a cause worth dying for. The third of these--identification with a cause--tends to get the most attention from civilians, whereas professional military men tend to focus only on the first two.
Keegan, and to a lesser extent Marshall, emphasizes that men fight because they fear to run. They fear being killed by an enemy undistracted by opposing fire. They fear prison and shame for disobeying orders. They fear being shot for desertion in the face of the enemy. (Only one U.S. deserter was shot in WWII but the German army of WWII shot hundreds on the record and probably many more off the record. The Red Army of WWII is in a different league; "punishment battalions" filled with deserters, anti-Communists, and trouble makers of every sort were given carbines with a few rounds and sent against German machine guns with Russian machine guns at their backs. In air combat, punishment battalion men were chained to their seats as tail-gunners.)
In short, the power of the state and the power of the enemy are both powerful reasons to fight; once on the battlefield, soldiers will usually find it safer to fight than to flee or cower. One of Marshall’s observations to the contrary got him in a lot of trouble. He reported that, even in major engagements and even when hard pressed by the enemy, only about a third of U.S. soldiers fired their weapons. He suggested that they were frozen by a combination of fear and civilized inhibitions against violence. Other observers note that not firing is the right answer if you can’t see a target, and a modern battlefield is a lonely place in which combatants are mostly down and out of sight and most casualties come from artillery rather than individual weapons. Others also note that firing off in all directions without a target is typical of raw and undisciplined troops, and dangerous in exposing the position of the one firing. Finally, other reports from the battlefield (Holmes) have not supported Marshall’s initial observation of a reluctance to shoot.
Today, the dominant explanation of why men fight is that they do not want to let their buddies down. There is considerable evidence for this view. A sizeable fraction of U.S. Medal of Honor winners got the award for something akin to throwing themselves on a grenade to save their buddies. When queried, these men say they did not make a decision, they acted automatically to do what had to be done to save the group.
William Manchester (quoted by Holmes, p.300) was wounded in WWII but went AWOL from a field hospital to rejoin his comrades: "It was an act of love. Those men on the line were my family, my home. They were closer to me than I can say, closer than any friends had been or ever would be. They had never let me down and I couldn’t do it to them...men, I now know, do not fight for flag or country, for the Marine Corps or glory or any other abstraction. They fight for one another. Any man in combat who lacks comrades who will die for him, or for whom he is willing to die, is not a man at all. He is truly damned."
More systematically, Stouffer reports interviews with WWII soldiers fresh from combat. Asked what was helpful when the going was tough, about 60% of enlisted mentioned "not letting my buddies down". Prayer was even more often cited, but promotion, decorations, and finishing the job in order to get home again were much less often cited. For officers, the thought that they could not let the other men down was most often mentioned (about 80%). Hatred of the enemy was mentioned by only a small minority; evidently WWII was more a matter of instrumental than impulsive aggression--at least for U.S. soldiers.
The power of the primary group in combat is also evident in the relation between group cohesion and combat effectiveness. There is no doubt that the German army was, man for man, the best in WWII. Battle results (van Crevald) indicate that, to get an even chance of winning against a German force, the U.S. or British force had to be almost twice the size of the German force. The Germans seen some of their armed forces mutiny at the end of WWI, and they took care to maximize group cohesion in WWII. In particular, their replacement policy gave primacy to development of group cohesion. Even when hard pressed in the late stages of WWII, an entire German division would be taken from the line for a period of months to rest, add replacements, and retrain with the new replacements. At a minimum, German replacements trained in a replacement battalion commanded by officers and NCOs rotated out of the division the replacements were joining. In contrast, U.S. policy in WWII was to feed replacements piecemeal into units still in combat, so that the newcomers were often casualties before they had become effective members of their units.
In the Korean war (Little), group cohesion suffered from a rotation policy in which an individual left his unit after putting in a certain number of months of combat. Each soldier in a unit had his own date of departure, a self-interest in conflict with group cohesion. In WWII, cohesion was enhanced by "common fate"; every soldier in a combat unit stayed with that unit until he was killed, wounded, or the war was won.
In the Vietnam war (Moskos), a variation of the Korean War rotation policy was put in place. In addition to individual rotation dates, the new policy harmed group cohesion by introducing a distinction between officers and men. Officers rotated out of combat units in six months, enlisted men rotated out in 12 months; the goal was to allow more officers the career benefit of combat experience.
The third explanation of why men fight is now in disrepute. U.S. military doctrine used to emphasize the importance of morale and fighting spirit based on the common soldiers’ understanding of and commitment to the goals of the war. From this perspective, it was a big surprise to find that over half of Chinese prisoners captured by U.N. forces in the Korean War were anti-Communist; they mobilized themselves into anti-Communist factions in their POW camps. This despite the fact that the Chinese troops fought very well against the U.N., seldom surrendering except when their situation was objectively hopeless. Bradbury, Meyers and Biderman resolved this paradox by showing how the Chinese army used "brainwashing" or "thought-reform" techniques to control the behavior of soldiers who had no commitment to the cause they were fighting for. As described below under Violence in the System, these techniques depend crucially on the manipulation of small group dynamics.
Still, there is something to be said for the power of ideology. German SS divisions, sworn to special devotion to National Socialism, were particularly feared opponents on WWII battlefields. Perhaps more important than its impact on the battlefield, ideology may be important in determining who is on the battlefield. In the Vietnam War, for instance, most of the best and the brightest of U.S. draft-age males found it expedient to be somewhere else: graduate school, teaching, or even Canada. In contrast, the brother of Israel’s political leader, Netanyahu, was the only casualty of the Israeli anti-terrorist action to free hostages at Entebbe. Two earlier leaders of Israel, Menachem Begin and Yitzak Shamir, put themselves at risk as youthful leaders of anti-Arab and anti-British terrorist groups. Examples such as these suggest that ideology is not to be ignored in understanding combat performance; without ideological commitment an army will not get its best young men to the battlefield. The importance of getting the best in the field is indicated by the success of the Israeli Defence Force against larger Arab forces; Herzog’s history of these successes indicates that they owe more to the intelligence, initiative, and risk taking of low-level officers and NCOs than to superior weapons or genius staff work.
VIOLENCE IN THE SYSTEM:
THE PSYCHOLOGY OF DOMINANCE AND HEGEMONY
Both individual and group violence have been reviewed above with a focus mostly on physical violence (with the exception of recognizing that females may often use verbal or "relationship aggression"). But there are less obvious forms of violence in which an entire social system operates to control not just the behavior but the social reality of those in the system. Perhaps the most notorious example of such a system is the collection of social control techniques embodied in "brainwashing" or "thought reform."
Brainwashing came to public attention when some U.S. soldiers, capured by the Chinese in the Korean War, refused repatriation at the end of the war. How could American boys not want to come home? Brainwashing was the answer: a system of small-group study of Communist-party materials, joined with sessions of criticism and self-criticism for failures to live up to "the people’s viewpoint."
This technology originated with Lenin as a means of maintaining Party discipline across the underground of Bolshevik Party cells (Whyte). The Soviets exported this technology to China in founding the Chinese Communist Party, and Mao extended it from Party members to the whole Red Army. It disappeared in the U.S.S.R. after Stalin denounced the "fetishism of the study circles," but it flourished in China. As the Red Army rolled across China after WWII, Mao extended thought-reform groups into schools, factories, rural work teams, labor camps, and even residential neighborhoods.
Pasqualini, a Chinese national whose father was French, has described the horrific quality of thought reform procedures applied to Chinese sentenced to "reform by labor." The camps grind men down with heavy work and starvation, as food is contingent on political correctness.
Foreign prisoners do not ordinarily show up in the labor camps, presumably because authorities expect that they will one day leave China. A more genteel version of thought reform is applied to foreigners, a version that does not involve labor or starvation. Lifton, a psychiatrist, has famously described this application.
Lifton interviewed 25 Western civilians who had stayed in China after the Communists took over, spent years in prison, confessed to "spying" against the Chinese people, and finally been deported from China. Lifton put his interviewees into three groups: 2 who seemed 100% converted to "the people’s viewpoint," 8 apparent resistors, and 15 "confused and searching." Every prisoner had been persuaded to some degree, however. Even the resistors expressed grudging admiration for the intensity and consistency of their captors, felt that they had learned something about themselves from the experience, and thought that the West needed to learn from the Communists in order to succeed against Communism.
Lifton’s explanation of the power of thought reform was cast mostly at the individual level. He acknowledged the power of threat and torture to get the first confession, but to explain the impact on prisoners’ private opinions he pointed to individual differences in need for black-and-white thinking, to captors’ control of all information entering the cell, to mystical manipulation in which Party control was wrapped in the "people’s" cause, and so forth. Lifton gave special emphasis to guilt associated with confessing selfish behaviors at odds with prisoner’s professed values: a physician who said he came to help the Chinese people had to recognize the significance of depending on a Chinese cook, a Chinese housekeeper, a Chinese gardner, and Chinese pedicab drivers.
A more social psychological analysis is called for, however, because very much the same kinds manipulation (including torture) were used by the N. Vietnamese on downed U.S. fliers imprisoned in the Hanoi Hilton. The one big difference in manipulation was that the fliers were in solitary confinement and there was no use group dynamics against the prisoners (Stockdale). A parallel big difference in result was obtained: there was little impact on the private opinions of prisoners in the Hanoi Hilton.
SOME SOCIAL PSYCHOLOGY OF EXTREME BEHAVIOR
A social psychological analysis begins by noting that Lifton’s prisoners seldom met with the judge/prosecutor to whom their confessions were addressed. Every waking hour of every day in the cell was devoted to group discussion, with group members assisting in one another’s re-education with criticism and self-criticism. An important insight of social psychology is that the power of the group over the individual depends upon the number of groups to which the individual belongs. Most of us belong to many groups, including family, friends, co-workers, church, teams, and neighborhoods. When the individual’s social world is reduced to a single group, the norms of that group have enormous power.
Three lines of research indicate the power that can be wielded even by temporary groups arranged by an experimenter.
Asch, S. (1952). Social psychology. New York: Prentice Hall.
A classic experiment by Asch showed that 75% of subjects in a "study of perception" would deny the evidence of their own eyes when three friends (experimental accomplices) lined up to give a wrong judgment of line length. Inquiry revealed that subjects thus conforming mostly did so to avoid being laughed at (public compliance without private acceptance). But about a third of the conforming occured because subjects were persuaded that their friends must be able to see the line better than they could (private acceptance of majority judgment). That is, the unanimous majority could affect the definition of reality for the lone minority. The power of three friends in an hour’s experiment is small compared with the power of a cell group--all unanimous in expressing the people’s viewpoint--that is a prisoner’s whole world for years at a time.
Brown, R. (1989). Social psychology: the second edition, Chapter 6, pp. 200-248. New York: Free Press.
The phenomenon of group extremity shift can also illuminate what happens in thought reform. Group discussion tends to make group opinions and judgments more extreme in the same direction favored by most group members to begin with. The mechanisms of this phenomenon are two.
First, opinions become more extreme because most of the arguments offered in discussion are biased in the same direction as initial group opinion. The imbalance of arguments in discussion moves opinions further in the favored direction. The second mechanism depends on the fact that individuals above-average in their support of the group favored direction are seen as better people; thus group discussion is the occasion for individuals to compete for status by the extremity of their support for the group favored direction.
The arguments discussed in thought reform groups are clearly biased toward the people’s viewpoint, and competition for the status of "reformed" and ready for discharge from the cell is obvious in Lifton’s account of prison experiences.
Milgram, S. (1974). Obedience to authority. New York: Harper & Row.
Sabini, John. (1995). Social Psychology, 2nd Edition. New York: Norton. Chapter 2, "Social Influence."
A third social psychological contribution to understanding thought reform comes out of Milgram’s famous experiments on obedience. Milgram brought subjects into a laboratory for a supposed study of learning, in which the teacher (the subject) was asked to shock the learner (an experimental accomplice) every time the learner made a mistake. The experimenter asked the teacher to raise the shock one level at each mistake, with shock switches starting at 15v and going up by 15v increments to 450v labeled "XXX DANGER STRONG SHOCK." Milgram originally expected his subjects to refuse at a relatively low level, but found instead that about 60% of subjects would go all the way to 450v.
Some experimental variations showed that full obedience (giving 450v) declined if the learner was closer to the teacher but increased if the learner and his programmed complaints and screams were unseen and unheard. If the experimenter left the room, obedience decreased. These results parallel common experience that it is easier to hurt others if they can be made distant or dehumanized, and that authority is stronger when it is immediate and supervisory.
One of Milgram’s variations merits special attention. Here the teaching job was divided between the real subject giving the shocks and another subject who read the questions and determined which responses were mistakes. The experimenter left before saying anything about shock levels, and the the other subject (actually another accomplice of experimenter) came up with the idea of raising the shock level with each mistake. Still 20% of subjects went to 450v. Of course 20% is less than 60%, but what explains the 20% when the authority of the experimenter was never associated with raising the shock level?
The answer seems to be that the fine-graded shock levels are a slippery slope. No reason not to give the first one: 15v is nothing. If the first is nothing, how bad can the second be at 30v? And so on. Thus the best reason for giving the next shock is that the subject has already given the last shock. To stop at any point means recognizing that there was something at least somewhat wrong about having given the last shock. In this way, the subject’s past behavior becomes not just a predictor but a cause of his future behavior. This is an important part of the banality of evil; it comes not as a single clear choice but as a slow downward spiral into defining as acceptable what once would have been unthinkable.
In thought reform groups, this mechanism is at work in the slowly escalating demands for participation in criticism and self-criticism, including especially the quality of the confession. Lifton’s interviewees differed in how long they held out against making their first confession; some confessed within hours, others after weeks of interrogation, lack of sleep, chains, and beatings. But all came to the same initial strategem: a wild confession of imaginary crimes against the Chinese people. This confession was rewarded with improved conditions, but thereafter the confession and group participation had to be ever more realistic and biographical. Sooner or later the prisoner had to implicate real persons as collaborators in crime, knowing that these persons would be imprisoned in their turn. On this slippery slope, each additional detail of biography and confession is a reason to give the next one.
OTHER ANALYSES OF EXTREME BEHAVIOR
Recruitment to religious cults is often described as brainwashing. Barker’s study shows that most of the persuasion is based on normal small-group dynamics, as above. Legally and morally, the big difference between cult recruiting and brainwashing is that cult recruiting begins with promise and reward rather than threat and punishment.
Research has found that psychopathology is no more common among terrorists than among non-terrorists from the same background. Rather it is normal social psychology, intensified in the terrorist cell by being cut off from all other kinds of social group, that explains how terrorists can commit extreme violence against even non-combatants. Very much the same account can be given of how some of America’s best young men were brought to drop fire bombs on Dresden, Hamburg, and Tokyo, and atomic bombs on Hiroshima and Nagasaki.
This book adds up the numbers killed by governments in the 20th century, up to 1987. Deaths in war and civil war add up to 38 million; deaths by government massacres of helpless people add up to 151 million. Rummel finds some hope in the tendency for democracies to kill less than totalitarian states.
The banality of evil is explored in a detailed examination of the ordinary middle-aged Germans who followed orders and even competed in executing Jews in Poland. Browning rejects explanations in terms of psychopathology in favor of much the same normal social psychology cited in relation to thought reform.
Staub, E. (1989). The roots of evil. New York: Cambridge University Press.
This book aims to provide an integrated account of the psychological, social, and cultural origins of mass murder and genocide. If its ambition leaves both author and reader with many questions, it succeeds at least in leaving us with interesting questions. It focuses particularly on the Holocaust and, to lesser extent, on the genocide of Armenians in Turkey, the genocide of Cambodians by Cambodians, and the "disappearances" in Argentina.
Smith points to structural/cultural factors such as land competition and govenrment propaganda, and to individual factors such as authoritarian personality development. He has almost nothing to say, however, about the dynamics of the groups involved in the killing.
This paper focuses chiefly on structural/cultural factors such as history, theology, nationality, and inequality. It suggests that psychological approaches have only a modest role to play, although Tajfel’s social identity theory is seen as promising. In this theory, people fight for their group because part of their self-esteem is bound up in group security and group status. More is needed here about how one group identity is mobilized from among the many that may contribute to an individual’s self esteem.
This even-handed paper, written by an Israeli-Palestinian and an Israeli, explains the conflict on the basis of widely-held beliefs in which both sides see themselves as victims of an illegitimate enemy. The Israelis see themselves surrounded by a sea of Arabs, threatened by Arab terrorism. The Palestinians see themselves as a minority in Israel, the victims of Israeli confiscation of their land and water. There is an important idea here: intractable ethnic conflicts tend to be "double-minority" conflicts, where both sides can see their way of life as threatened. Northern Ireland and Sri Lanka have conflicts of this sort: Protestants see themselves as the threatened minority in a united Ireland, Catholics see themselves as a threatened minority in Northern Ireland; in Sri Lanka, Tamils see themselves as theatened minority amid Sinhalese and Sinhalese see themselves as threatened minority amid South Indian Tamils.
This paper identifies three strands of interpretation of the conflict. One strand deconstructs the "primordial" ethnic identities often cited, one applies early twentieth-century theorizing about crowd behavior, and one explores the meaning of pain and suffering for victims of ethnic violence. This case seems to offer considerable opportunity for more modern treatment of conflict origins.
Even the most thorough-going system of dominance tends to wear and fail. The Soviet Union is no longer with us, and China is no longer in the grip of Communist ideology even if it remains in the hands of Communists. Scott offers an unusual look at the means by which men in their weakness resist state power in its hubris. Slaves, serfs, prisoners, and taxpayers open safe spaces in language and inaction; in these spaces dissident subculture grows.
ESCALATION AS A SOCIAL TRAP
Axelrod, R. M. (1984). The evolution of cooperation. New York: Basic Books.
Teger, A. (1980). Too much invested to quit. Elmsford, NY: Pergamon.
For political scientists, escalation tends to be something of a black box. One state does something inimical to the interests of another, the other responds, and both are into the box. Conflict escalates.
From another perspective, it is not at all obvious why escalation should occur. The tit-for-tat strategy, in which one begins with cooperation and then does exactly what the opponent does, has been the clear winner in Axelrod’s computer tournaments of prisoner’s-dilemma strategies. Unfortunately, the world is not a computer tournament. Striking back with exactly the same amount of harm one has received is difficult, especially in a world where the recipient of harm is likely to estimate it greater than the perpetrator does.
Teger studied escalation in a more natural context: the dollar game. In the dollar game, an experimenter auctions off a dollar to the highest bidder in a classroom. The rules of the game are that both the highest bidder and the previous highest bidder have to pay their bids. Typically as the bids approached $1.00, bidders dropped out until only two were left. These two always went over $1.00 and sometimes went to $20. Business students went particularly high. The emotional experience of the bidders changed as the bids went past $1.00; the excitement of winning was replaced with anxiety about losing. Teger’s laboratory variations and analyses indicate that the definition of the game is changed as bids go past $1.00; the attractions of reward are replaced with fear of three kinds of loss: loss of status in giving up first, loss of the money value of the last bid, and relative loss in losing more money than the winner.
Teger finds some ways people can get out of the game. If one person bids $1 and the other follows with a bid of $2, the first person is likely to quit. In this situation, both bidders lose one dollar, and many pairs that stumbled onto this sequence did quit. Teger also has some hints about recognizing and avoiding this kind of social trap, which can be found in arms races, industrial bargaining, and divorce proceedings.
One interesting aspect of this game is that the escalation is usually graded by slow increments in bidding, a variant of the slippery slope identified in Milgram’s obedience experiments. Once again past behavior becomes a cause of more extreme behavior, and the power in the situation is the power of increasing commitments. More abstractly, the power in the situation is the power of a social system--the rules of the auction--that can move reasonable individuals to unreasonable behavior. It is difficult to avoid reflecting that "there ought to be a law" against games like this one.