Searching for the truth (Khmer version) July & August, Number 31 & 32, 2002

A Magazine of the Documentation Center of Cambodia

AGENTS OF DEATH: EXPLAINING THE CAMBODIAN GENOCIDE

IN TERMS OF PSYCHOSOCIAL DISSONANCE

Alex Hinton

Department of Anthropology

Rutgers University

ahinton@andromeda.rutgers.edu

Many people have asked me: how could the Cambodian peasant, whom we had always regarded as gentle and charming and smiling and civilized turn into the kind of tough and grim and even brutal revolutionary who entered Phnom Penh on April 17? I have no easy answers. [Schanberg 1975:29]

The existence of dissonance, being psychologically uncomfortable, will motivate the person to try to reduce the dissonance and achieve consonance. [Festinger 1957:3]

            People often used to characterize Cambodia as a "gentle land" inhabited by nonviolent Buddhists who were always courteous, friendly, and ready with a smile. Beginning in the late 1960's, however, the country was rocked by socioeconomic unrest, civil war, intensive U.S. bombing, and, finally, social revolution. While around six hundred thousand of Cambodia's eight million inhabitants perished during these years, up to a million and a half people later died from disease, starvation, overwork, and execution during Democratic Kampuchea (1975-1979).[i] Survivor accounts are replete with stories of how the Khmer Rouge shot, bludgeoned, stabbed, and tortured legions of their own country people. This type of violence demands the attention of scholars. How could the seemingly "gentle" Cambodians come to commit such genocidal acts?

            While other disciplines have addressed this challenging issue, anthropology has been remarkably silent on the topic of large-scale genocide, an omission that is particularly striking because anthropologists have demonstrated the ability to productively explain the roots of violence in other, non-genocidal contexts.[ii] To help to redress this deficiency, this essay will provide a psychosocial explanation of how people come to commit acts of genocide. Drawing on the psychological concept of cognitive dissonance (Festinger 1957), modified and enhanced by anthropological notions of the self, cultural models, and emotion, I will analyze the events that took place in Cambodia in terms of what I call "psychosocial dissonance" (PSD). By doing so, I intend to provide a starting point for anthropological debate on large-scale genocide, to delineate an explanation for how people come to commit genocidal atrocities, to develop a general explanatory model that can be applied to other genocides, and to explain the psychosocial origins of the Cambodian genocide.

            The first section of this essay briefly outlines the theory of cognitive dissonance and then illustrates how it can be developed by anthropological concepts into a model of "psychosocial dissonance." I then describe two salient cultural models that existed prior to Democratic Kampuchea, the "gentle ethic" and the "violent ethic." The third section shows how these models came into conflict when the Khmer Rouge attempted to radically transform Cambodian society, thus creating a situation of PSD. I then illustrate the steps the DK regime took to help reduce the PSD of its "agents of death." The fourth section illustrates how, in part by drawing on this "state level response," Khmer Rouge cadre made a series of cognitive moves to reduce their PSD -- i.e., the "individual level response." I conclude by discussing the implications of PSD for the anthropological study of large-scale genocide.

A Psychosocial Model of Cognitive Dissonance

Cognitive Dissonance

            Leon Fstinger's (1957) original formulation of "cognitive dissonance" (CD) theory asserts that if a person holds two conflicting cognitions, she or he will be motivated to reduce the resulting state of psychological discomfort in a manner similar to drive reduction. Upon hearing a report that cigarette smoking is bad for their health, for example, many smokers will likely be motivated to reduce the resulting psychological discomfort/dissonance by: changing cognitions to make them more compatible (e.g., dismissing the research out of hand); circumspectly adding new cognitions that bridge the gap between the cognitive elements (e.g., finding information that indicates smoking is less dangerous than driving a car); or changing her or his behavior (e.g., stopping smoking). The stronger the "magnitude" of dissonance, the more a person will be motivated to reduce it.

            This theory has generated a great deal of research, much of which indicates that cognitive dissonance is greatest when an individual has a behavioral commitment to one or both of the conflicting cognitions (Aronson 1992). Such experimental findings and a growing dissatisfaction with the vagueness of the original formulation of dissonance theory have led Elliot Aronson to assert that cognitive dissonance "is clearest and greatest when it involves not just any two cognitions but, rather, a cognition about the self and a piece of our behavior that violates that self-concept" (1992:305). Dissonance therefore arises in situations in which a person is confronted with behavioral expectations that conflict with this concept of the self.

            While the theory of cognitive dissonance and its later reformulations provide a great deal of insight into human thought and behavior, this concept is nevertheless predicated on some of the biases of Western psychology. From an anthropological perspective, this cognitive dissonance research is problematical in its: 1) treatment of the relationship between culture and individual variation; 2) lack of an adequate theory of motivation; 3) insufficient attention to contextual complexity; 4) transcendent conceptualization of the self; and 5) disregard of emotion. As opposed to simply dismissing the theory, however, we can revise it in accordance with theoretical developments about cultural models, the self, and emotion.

Anthropological Contributions: Cultural Models and the Self

            Problem #1 - Culture: While Festinger (1957) included "cultural mores" as a potential source of dissonance in his original formulation of CD theory, culture was never treated in a very sophisticated manner in later research. Recent developments in cognitive anthropology, however, have made it possible to bridge the gap between culture and cognition through the concept of cognitive schema, or knowledge structures through which people interpret stimuli and determine appropriate behavioral responses (D'Andrade 1995).

            The key move cognitive anthropologists have made is to illustrate that a large number of these cognitive schemas are culturally shared. While no exact correspondence exists between such cultural models and the way they are internalized by different social actors, individual variability is somewhat constrained by language and shared social experience (D'Andrade and Strauss 1992). When going to a restaurant, for example, the expectations and behavior of most members of our society are largely mediated by a "restaurant script" (e.g., entering, ordering, eating, and exiting; see Shank and Abelson 1977) that they have learned through childhood observations, personal experience, dramatizations on TV, books, etc. One of the strengths of cultural models theory lies in its ability to illustrate the relationship between such personal knowledge structures and their social analogs (Shore 1996). By connecting cognition and culture, the concept of cultural models thus provides a means for culture to be actively incorporated into CD theory.

            Problem #2 - Motivation: Dissonance theory lacks an adequate concept of motivation (Kunda 1990). Festinger originally asserted that people would be motivated to decrease dissonance in a manner similar to the way they pursue drive reduction. While Aronson's introduction of the self-concept adds greater specificity to this "black box," the processes underlying CD motivation remain vague. The concept of motivation delineated in cultural models research can therefore help bring much needed specificity to CD theory.

            Cognitive anthropologists have argued that, as opposed to just being a response to an internal drive stimulus, motivation can be explained more complexly in terms of schemas since they often establish goals that instigate action (D'Andrade and Strauss 1992). Thus, within the restaurant script, dissatisfaction with service may motivate a person to complain to the manager. It is important to recognize the hierarchical arrangement of such schemas. In this case, the "unsatisfactory service" schema is one component of the higher-level restaurant script. While this lower-level schema may have links to other models (e.g., to a "shopping" or an "auto-repair" schema), interpretations tend to be passed on to "the topmost level of interpretation which is typically linked to the actions by which the organism operates in its environment. That is, the top-level schemas tend to be goals" (D'Andrade 1992:30). Since many of these high-level schemas are cultural models, culture sets goals and therefore has motivational force.

            Cultural models are also internalized in disparate ways by different actors (Shore 1996). For some people, a given model may be regarded with relative indifference, while for others it may be highly salient. In each case, the motivational significance of the model will vary according to the degree to which the model is internalized and the degree of affective force it carries (D'Andrade 1992). To determine the motivational salience of any cultural model for a given individual, it is necessary to understand something about the extent to which the model has acquired emotional force during her or his life history. However, some extremely high-level models will seem more "natural" and thus tend to motivate the behavior of most people in a given society (e.g., "individual rights" and "equality" in the United States).

            Problem #3 - Contextual Variation: While CD theory has always acknowledged the importance of taking situational factors into account, it has lacked criteria by which to determine which cognitions are salient within a given context (Schlenker 1992:342). Cultural models theory can provide this guidance. In addition to varying between individuals, the significance of schemas also differs across social contexts (Holland and Quinn 1987). Thus, while a restaurant script mediates behavior when people dine out, few follow this event sequence when eating at home. Such alternative schemas may even be inconsistent. To determine which cognitive schema are germane in such situations, a researcher must specify the contextual background that frames an interactional sequence using discourse analysis, life-histories, behavioral observations, and/or the study of socialization practices (D'Andrade and Strauss 1992; Holland and Quinn 1987).

            Problem #4 - The Self: CD researchers have found that dissonance is strongest in situations which necessitate behavior that violates a person's self-concept. Aronson states that this psychological discomfort occurs because people strive to "preserve a consistent, stable, predictable . . . competent . . . morally good sense of self" (1992:305). Anthropologists have pointed out, however, that this more extreme view of the egocentric self ignores the fact that the self-concept may vary across social contexts.

            Katherine Ewing (1990), for example, has proposed a "model of shifting selves" in which people are conceptualized as having multiple, often inconsistent self-concepts that, while experienced as "whole" at a given moment in time, are contextually defined. To illustrate her point, Ewing describes how one of her Pakistani informants, Shamim, held disparate representations of herself (i.e., as a "good daughter" who would obediently follow her parents' wishes and become a good wife and as a "politician" who employed various strategies to achieve her personal educational/work goals) in different contexts. We can use this anthropological notion of the contextually-defined self-concept to refine CD theory. In particular, as opposed to arising when a transcendent concept of the self is violated, dissonance can be said to occur when a person must act in ways that violate a contextually-defined self-representation.

            Problem #5 - Emotion: Festinger's model of cognitive dissonance is predicated on the Western Judeo-Christian emphasis on conceptual consistency. Anthropologists have shown, however, that in other cultures dissonance often arises only in situations in which discordant cognitions are emotionally salient to actors (DeVos 1975; Nuckolls 1993). The notion that dissonance occurs only in situations which necessitate behavior that violates a person's contextually-defined self-representation dovetails nicely with the anthropological view that emotions are "embodied thoughts, thoughts steeped in the apprehension that 'I' am involved" (Rosaldo 1984:143). In other words, situations that implicate the self are emotionally significant.

            George DeVos has pointed out that, in most societies, people are socialized by taboos to avoid directing aggression toward those with whom they share a collective identity and social bonds: "We cannot tolerate the emotional conflict, or as I term it the 'affective dissonance' which results from killing within the social group" (1975:82). As we shall see, one way of getting individuals to kill in such a situation is to redefine who belongs to the group so that the victims are excluded from the ingroup and dehumanized as the "enemy." What is important to note, however, is that when a person must act in a manner that violates a contextually-defined self-concept (e.g., to kill a person with whom she or he shares a social identity), she or he will likely experience an emotionally charged form of cognitive dissonance. This insight forms the basis for a model of psychosocial dissonance.

Anthropological Contributions: A Model of Psychosocial Dissonance

            Drawing on the aforementioned anthropological insights, we can now reformulate Aronson's theory of cognitive dissonance in the following manner. Cognitive dissonance arises when an (often culturally informed) emotionally salient cognition about the (culturally informed and context-dependent) self comes into conflict with another (often culturally informed) emotionally salient cognition that motivates behavior which violates that context-dependent self-concept. Psychosocial dissonance is reserved for those cases in which an emotionally salient cultural model about the context-dependent self comes into conflict with another emotionally salient cultural model that violates that context-dependent self-concept. Thus, Shamim experienced PSD when her parent's demands that she marry (i.e., the "good daughter" model that shaped her self-concept in familial interactions) began to conflict with her desire to achieve personal occupational goals (i.e., the "clever politician" model that informed her self-representations in the workplace). Whereas previously these contradictory models had been salient in different contexts, they began to overlap and caused PSD. PSD is thus a subset of CD, since all cultural models are cognitions, but not vice versa.

            The degree of dissonance is a function of the emotional salience of a cognition. If being a "good daughter" was not particularly important to Shamim, she would only have experienced a small degree of PSD when her parents raised the subject of marriage. Because both the "good daughter" and the "clever politician" models were extremely important to her, however, Shamim experienced a great deal of PSD. People are motivated to reduce such dissonance because the cognitions in question establish goals that necessitate contradictory courses of action. In Shamim's case, her goal of getting married established a behavioral sequence that directly conflicted with her goal of advancing her career. I will examine the precise ways in which people attempt to reduce such dissonance in a later section.

            First, I will illustrate how this model of Psychosocial Dissonance can be applied to genocidal acts, using Cambodia as an example. As DeVos notes, in almost every society, powerful injunctions exist against killing other members of a social community. Within some contexts, however, the destruction of "others" is often culturally and/or legally legitimated (e.g., military, penal code, political enemies). Because cultural norms against killing tend to be so strong, some of the individuals who kill will experience PSD. In the case of large-scale genocide, PSD has the potential to be much greater because these acts are often being perpetrated against people who were formerly members of the same society. In other words, two extremely salient models fostering extremely contradictory behavioral goals -- "kill" versus "don't kill" -- begin to overlap within a context, thus potentially causing a great deal of PSD for the "agents of death." If a regime is to accomplish its genocidal goals, it must make changes in the environment that will potentially decrease the degree of PSD its minions experience. These perpetrators of genocide will draw both on this "state level" response and on their own "individual level" dissonance reduction strategies to achieve consonance with the idea of killing. This is exactly what transpired in Cambodia.

Psychosocial Dissonance Applied: the Cambodian Genocide

            Many people have been struck by both the friendly demeanor of Cambodians and their ostensible lack of conflict in daily interactions. Given the harmonious atmosphere that is so prominent in everyday life, it is easy to be somewhat taken aback by the political violence that has characterized Cambodian history. Such violence illustrates the fact that while everyday communal life (i.e., relations with fellow members of a family, village, or organization) is frequently mediated by a high-level cultural model that fosters prosocial behavior, larger sociopolitical interactions (i.e., relations with an "enemy" in military activity, law enforcement, or national politics) are often informed by an extremely salient, yet potentially contradictory cultural model that promotes aggression. These two models, which I will hereafter respectively refer to as the "gentle ethic" and the "violent ethic," were significant in different interactional contexts and thus rarely came into conflict in pre-DK Cambodia.[iii] The conditions for PSD arose, however, when the violent ethic was legitimated in everyday communal interactions during DK. The unfortunate result was a situation in which acts of extraordinary violence took place.

The Gentle Ethic

            While disputes sometimes occurred, the few ethnographies that were conducted in pre-DK Cambodia are striking in their description of the overall harmony of village life (Delvert 1961; Ebihara 1968; Martel 1975). Over the course of a year, for example, May Ebihara observed only a few significant intracommunal quarrels. This apparent lack of discord was due to several factors. Members of a community often shared a strong sense of social solidarity that was developed through years of association, cooperative labor exchanges, mutual aid, overlapping friendship and kinship networks, communal activities, and a sense of identification with and loyalty to the group (Ebihara 1968).

            Cambodian socialization practices emphasized that, when interacting within such a known community, a person should attempt to "have friendly relations with others" (roap an knea, reak teak). In addition to avoiding conflict and potentially making patronage connections, Cambodians who were friendly and polite were respected by others and thus gained honor (Hinton n.d.). Individuals who transgressed social norms, in contrast, were subject to gossip, avoidance, and public censure. In a culture in which "face" (mukh) is highly valued, the threat of such potentially shameful (khmas ke) consequences represented an effective control on individual behavior.

            Buddhism also promoted prosocial behavior among villagers. The temple, for example, reinforced social bonds by serving as a center both for communal and religious activity and for the moral education for children (Ebihara 1968; Martel 1975). While people did not necessarily know the intricacies of Buddhist doctrine, all were familiar with the five moral precepts which told them not to lie, steal, have immoral sexual relations, drink, or kill living creatures. Such prohibitions were coupled with Buddhist notions of karma and merit that encouraged villagers to maintain harmonious relations (Ebihara 1968).

            Although domestic violence existed, intrafamilial relations, like communal life, appear to have been generally harmonious. Family members were tied together by economic production, daily social interaction, sharing, and joint participation in important ceremonies (Ebihara 1968; Martel 1975). Harmony was also promoted by rules of proper etiquette (Ebihara 1968; Ledgerwood 1990). Like Cambodian society in general, the family constituted a mini-hierarchy in which people were accorded different degrees of respect depending on age and sex. Folk-tales, didactic poems, and terms of linguistic etiquette reinforced such patterns of appropriate behavior that could both regulate interactions and diffuse conflict. The threat that village and/or familial spirits would cause an innocent member of the family to become ill when siblings quarreled provided a more proximate mechanism for curbing improper behavior (Ebihara 1968; Martel 1975). Such values and practices were part of a high-level cultural model that fostered a disposition toward prosocial behavior (i.e., the "gentle ethic") in the context of everyday interactions within a known community.  

The Violent Ethic

            Cambodia's history of violence against those defined as sociopolitical enemies began long before DK. During the Angkorian period, "god-kings" led their troops into battles against foreign states; both these external foes and any internal opposition were dealt with in a brutal fashion (Chandler 1992). Political violence was rampant during the nineteenth century and continued during French rule. The Khmer Issarak, in particular, were known for a style of ruthless violence against enemies that foreshadowed later practices of the Khmer Rouge (Bun 1973). After independence, Prince Sihanouk reestablished the royal tradition of absolute authority (Becker 1986; Chandler 1991): opposition to his rule was considered treasonous and dealt with accordingly. The actions of Lon Nol's troops were not much different. Such evidence illustrates that while a person might have been "gentle" within a known community, a different ethic often held sway in the context of larger sociopolitical interactions.

            The origins of this tradition of sociopolitical violence can be traced back to Brahmanistic notions of status and function (Chandler 1992; Ponchaud 1977). The "naturalness" of given social roles was embodied in the concept of dharma, the "cosmic doctrine of duty in which each sort of being in the universe . . . has by virtue of its sort, one ethic to fulfill and a nature to express - the two things being the same" (Geertz 1983:196). This duty could vary across place and time. In one context, a warrior had to be respectful to his social superiors, while, in another situation, his dharma would require him to "crush the enemy" (kamtech khmang) without hesitation.

            While the arrival of Buddhism significantly altered many aspects of Cambodian society, this "warrior heritage" (Bit 1991) was retained and even reinforced. By asserting that the king was the defender of dharma, for example, Buddhism legitimated the use of force against enemies who threatened social order (Thion 1990). This norm was reinforced by Cambodian proverbs, didactic poems, and folk-tales that encouraged people to act in accordance with their station (Chandler 1982; Ledgerwood 1990). From a young age, children learned about the virtues of "warriors" (neak tâsou) who gained honor by distinguishing themselves through bravery, fulfilling their duty, and heroically fighting the enemy. This type of "Cambodian machismo" was premised upon an honor code which held that those who dared to kill a sociopolitical enemy in battle gained face, while those who did not were shamed.

            One of the most popular stories was the Reamke, the Cambodian version of the Ramayana. Set in a Brahmanical world of violence and duty, this epic also reflects Cambodian ideals about virtuous behavior within a known community and against a sociopolitical enemy (Chandler 1992; Jacobs 1986). Thus, Komphâkar displays proper respect and obedience toward his elder brother and king, Reap, while at court. On the battlefield, however, he bravely fights his adversaries to the death in accordance with his duty as a warrior. In each domain, a different ethic predominates. Like the followers of Sihanouk, Lon Nol, and Pol Pot, Komphâkar was disposed to engage in acts of ruthless aggression against those defined as sociopolitical enemies.

Psychosocial Dissonance in the DK Killing Fields

            Festinger asserts a person can reduce cognitive dissonance by 1) changing given cognitive elements; 2) changing her or his behavior; 3) circumspectly adding new cognitions that bridge the gap between the dissonant cognitions; or 4) changing the situation in which given cognitions are salient. These dissonance reduction strategies can also, by extension, be used to reduce psychosocial dissonance. For example, a person experiencing PSD might 1) change one of her or his cultural models; 2) change the behavior that one of the cultural models entails; 3) add new, lower level schemas that bring the dissonant cultural models into consonance; and/or 4) alter the context in which the given cultural models are salient. This latter strategy is often difficult, since an individual rarely has the power to single-handedly change her or his environment. Totalitarian states, however, often do. While PSD reduction ultimately takes place on the individual level, a totalitarian state help transform people into "agents of death" both by: 1) promoting an ideology that modifies existing cultural models; and 2) by changing the context in which the given models are salient.

Psychosocial Dissonance I: The State Level Response

            When the Khmer Rouge victoriously entered Phnom Penh on April 17, 1975, their first order of business was to evacuate Phnom Penh and the provincial capitals (Becker 1986; Chandler 1991). This dispersal of the urban population was designed to control the citizenry, level class distinctions, create a strong labor base for the new agrarian, communist society, and weed out opposition. Leading military and civilian officials from the old government were rounded up and often executed. There was also a campaign to identify other potential traitors (e.g., teachers, students, bureaucrats, technical workers, and professionals). While some of these "class enemies" were killed, others were sent to be reeducated in special camps or through rural peasant life. At least one to two hundred thousand people died in this first wave of DK killing (Thion 1993). Having dealt with these potential sources of opposition, the Khmer Rouge instituted a number of social and ideological reforms that served to facilitate genocide by altering the environment in which "agents of death" perpetrated their deeds.

            DK Social Transformations: The DK regime introduced a number of radical changes which undermined the "gentle ethic" that had previously characterized communal interactions. Whereas Cambodian life had formerly revolved around the village, cooperatives became the fundamental socioeconomic unit in DK (Ebihara 1990; Jackson 1989a). Economic and ecological conditions that had previously necessitated cooperation were rendered irrelevant. In contrast to the polite and friendly relations Cambodians had developed through kin/friendship networks and years of communal association, interactions between "old people", "new people," soldiers, and local DK cadre in the cooperatives were often characterized by fear and suspicion (May 1986; Yathay 1987). While people had previously observed patterns of etiquette that both regulated and diffused conflict, they were now told that everyone was equal and that obedience was due only to "the Organization" (Ângkar).

            Intergroup harmony was further eroded by the destruction of Buddhism. Many of Cambodia's leading monks were executed immediately after the revolution, and the rest of the religious order was eventually forced to resume a secular life (Becker 1986; Ponchaud 1989). Temples were often physically destroyed or desecrated, sacred texts were burned, and statuary was defaced. If a Cambodian child had previously received her or his earliest lessons on morality at the temple, she or he was now indoctrinated into an ideology that glorified revolutionary violence and blood sacrifice (Jackson 1989a). Communism replaced Buddhism as the new "religion."

            Whereas the family had previously constituted the primary social unit in Cambodian life, its bonds represented a threat to the DK regime (Ebihara 1990; Ponchaud 1989). Consequently, the Party attempted to diminish the importance of the family by eliminating its social and economic functions. Family members were systematically separated by housing restrictions, relocation, communal meetings, and long work hours in sexually segregated work teams. Such separation was part of a larger movement to redirect familial attachment to the state (Ebihara 1990; Ponchaud 1989). In accordance with their official policy of egalitarianism and with their high valuation of children as the future of the revolution, the Party subverted patterns of etiquette that had traditionally governed interactions between family members. Indoctrination sessions informed children that they no longer had to act deferentially toward their parents. Mothers, father, children, and neighbors were all "comrades" (mitt) now.

            DK Ideology: As the gentle ethic was being undermined, the violent ethic was ideologically legitimated at the local level and began to inform everyday communal relations. While the leaders of the old regime had been eliminated and socioeconomic transformation begun, the DK regime was determined to bring the "spirit of combative struggle" to the cooperatives. Khmer Rouge ideology frequently employed the word tâsou ("to fight/struggle bravely") to reference the warrior spirit. Everyone was expected to enlist in the revolutionary fight to "build and defend" (kâsang neung karpear) the country. The first battleground was the work site. Daily activity was reorganized along military lines (Carney 1977; Ponchaud 1977). "Squads," "platoons," "companies," "battalions," and "divisions" of workers were sent to plant and harvest crops, to clear land, and to dig irrigation dams and canals. Like the military, this economic army was subject to strict discipline, harsh living conditions, and long work hours.

            National security constituted the second battlefield (Chandler 1991; Jackson 1989a). In addition to feeling threatened by external enemies like Vietnam, the DK regime feared internal foes. Everyone was enjoined to seek out "traitors" (kbat) who could potentially sabotage and/or co-opt the revolution. Initially, this command entailed mounting an offensive against "class enemies" and secret agents of foreign countries who were likened to a "sickness" that needed to be "treated" or "cut out" (Ponchaud 1977). Following economic failures, a possible coup attempt, increased hostilities with Vietnam, and internal fights within the Party, a second wave of killings began at the end of 1976. In the brutal purges that ensued, high ranking DK officials were tortured at the infamous interrogation center at Tuol Sleng; their families and subordinates, guilty by association, often followed. In the East Zone alone, over one hundred thousand people having "Khmer bodies and Vietnamese minds" were killed in 1978. Such terror was glorified "in the name of the revolution. Violence became a virtue. Waging war became prestigious. So did smashing the enemies of the party" (Chandler 1991:241-2).

            If the battle to "build and defend" the country was to succeed, all Cambodians had to adopt a proper revolutionary spirit. This new mentality required both complete obedience to the DK leadership and the renunciation of material goods, reactionary attitudes, and previous loyalties (Chandler 1991; Jackson 1989a). Since the Party represented the people, any sign of disobedience was tantamount to treason. As in war, such enemies of the state were subject to summary execution. Because local officials were commanded to root out these internal "microbes" without criteria for how to do so, a great deal of local level variation in the pattern of violence ensued (Vickery 1984). In some areas, hard-line cadre would execute suspected traitors without hesitation. In other locales, officials were relatively moderate in their actions. No doubt many of these individuals would have experienced a great deal of PSD when given orders to kill that came into conflict with the "gentle ethic" (i.e., an emotionally charged cultural model which prohibited them from harming fellow members of their community) that had been such an integral part of their social identity. We will now examine how these "agents of death" dealt with their PSD.

Psychosocial Dissonance II: The Individual Level Response

            How do people become genocidal killers? As we have seen, one factor in this conversion process comes from a "state level response." In the case of Cambodia, the DK regime helped to reduce PSD by altering the environment (i.e., by undermining the gentle ethic and bringing the violent ethic to the local level) and by providing an ideology that could be used to modify these two cultural models (e.g., redefining the "enemy," ordering the execution of "traitors," promoting revolutionary violence). Ultimately, however, psychosocial dissonance occurs and is reduced on the individual level.

            This "individual level response" will vary for each person. Based on her or his life history, an individual will need to take certain steps to become an "agent of death." Some people may just require a suitable environment to enact potentialities that they have already actualized. Others may have to undergo one or a series of transformations to become a killer. Still others may refuse to participate in a genocidal regime. Most genocidal killers probably fall into the middle category. This section will thus be concerned with delineating several of the cognitive "moves" such individuals may make to reduce their PSD to a point at which they are transformed into "agents of death." In particular, I will argue that the "genocidal self" emerges in situations in which an actor is able to: 1) dehumanize victims; 2) employ euphemisms to mask her or his deeds; 3) undergo moral restructuring; 4) become acclimated to killing; and/or 5) deny responsibility for her or his actions. I will deal with each of these dissonance reduction strategies in turn.

            Dehumanization: A group of Stanford social psychologists once initiated a mock prison experiment in which college student subjects were randomly assigned roles as "prisoners" and "guards" (Zimbardo et al. 1974). Six days later, the planned two-week study had to be halted after the guards exhibited increasingly abusive behavior toward the prisoners. This experiment illustrates two interrelated aspects of the dehumanization process that facilitates genocidal killing -- exclusion and devaluation. "Exclusion" refers to the process by which people lose their personal identity and are viewed in terms of a group category that is differentiated from the larger social community (Kelman and Hamilton 1989). "Devaluation" refers to the way in which such groups of people are increasingly marginalized from humanity. Exclusion and devaluation both contributed to the extreme dehumanization that took place in DK.

            In an attempt to erase hierarchical and class distinctions, the Khmer Rouge set out to divest the populace of "individualistic" qualities associated with a "capitalist" mentality (Becker 1986; Ngor 1987). Personal property was abolished; work and eating were communized. Everyone was required to wear identical black garb, to cut their hair short, to adopt stereotypical patterns of "appropriate" speech and behavior, and to divest themselves of individualistic traits that precluded a proper revolutionary "consciousness." The ostensible goal was to create a homogeneous society in which the individual was subsumed by the state.

            In reality, this homogeneous mass was divided along several lines (Chandler 1992; Stuart-Fox 1985). First, a clear distinction was made between the "true" Khmer who were a part of Ângkar and those who were its "enemies." Since Ângkar represented the people, any opposition to it was treasonous. Local level cadre were ordered to root out these "class enemies" who were attempting to subvert the Revolution. The first people to come under suspicion were "new" people: the urbanites and rural refugees who had been expelled from the cities and were suspect for having (in)directly supported the Lon Nol forces which the Khmer Rouge had defeated. Their very exposure to foreign influence and imperialism suggested that new people were not "real Khmer" and thus enemies who should be treated in accordance with the violent ethic. This group was sharply distinguished from both Khmer Rouge cadre and soldiers and the "old" people who had lived under the Khmer Rouge during the difficult war years. From the very beginning, the relocated "new" people were "outsiders" who were treated more harshly.

            In addition to being excluded from normal communal life, "new" people and other suspected enemies were subjected to dehumanizing practices. "New" people spoke of being crammed into trucks for many hours during later relocations (May 1986; Moyer 1991). Often they had to defecate or urinate where they stood; the trucks didn't stop, even if someone died of suffocation. "We were being treated worse than cattle, the victims of methodical, institutionalized contempt . . . we [were] no longer human beings" (Yathay 1987:73). People were also required to work like animals. Not only were they expected to labor obediently for extremely long hours on starvation rations, but they did so under the watchful gaze of armed soldiers and/or supervisors who had the power to have them executed. Many people have recounted the miserable living conditions in DK. "We were hungry, too tired to wash or clean our clothes, and we lost all sense of hygiene. We didn't care what we ate . . . where we had a shit, or who saw us. Disease spread through the village - cholera, malaria, dysentery, diarrhoea and skin infections" (May 1986:165).

            These dehumanizing practices were mirrored by Khmer Rouge ideology. People were instructed to be like oxen -- "Comrade Ox never refused to work. Comrade Ox was obedient. Comrade Ox did not complain. Comrade Ox did not object when his family was killed" (Yathay 1987:171). A soldier told one "new" person that it was better that her mother had died "than a cow . . . [cows] help us a lot and do not eat rice. They are much better than you pigs" (Moyer 1991:123). Part of this extreme devaluation stemmed from the fact that new people were often regarded as "war slaves" (Criddle and Mam 1987; Ngor 1987). "Many times we heard soldiers shout, 'Prisoners of war! You are pigs. We have suffered much. Now you are our prisoners and you must suffer'" (Moyer 1991:81). While sometimes tolerated, such "enemies" were expected to work hard and to be obedient. If they committed offenses, their execution would be no loss to DK. When explaining why his commune leader, Comrade Chev, both killed and ordered the execution of so many people, Haing Ngor notes: "We weren't quite people. We were lower forms of life, because we were enemies. Killing us was like swatting flies, a way to get rid of undesirables" (Ngor 1987:230).

            Such Khmer Rouge were indoctrinated into an ideology which instructed them to have no feeling for the enemy. As one cadre told me, "We were brainwashed to cut of our heart from the enemy, to be willing to kill those who had betrayed the revolution, even if the person was a parent, sibling, friend, or relative. Everything we did was supposed to be for the Party." This ideology of cutting off one's sentiment toward a now excluded and dehumanized "enemy" helped many Khmer Rouge reduce PSD both by redefining who was to be included in the new communist society and by creating a target group onto which they could project any anxiety-producing feelings. Because the revolutionary struggle continued in the cooperatives, Khmer Rouge cadre had little problem invoking the "violent ethic" to execute these hated enemies who were threats to the revolution, not "true" Khmer, and less than human.          

            Euphemism: By using euphemisms, perpetrators of violence are able to mask the true nature of their actions with expressions that make them seem benign or even respectable (Bandura 1986). In Nazi Germany, for example, Jews were referred to in terms of a medicalized vocabulary that made their elimination seem like a public health decision. Jews arriving at Auschwitz were prepared for "special treatment" in which "disinfection squads" would pour cyanide into their shower/gas chambers (Lanzmann 1985:155). Such linguistic maneuvers provide genocidal perpetrators with a sanitizing "discourse in which killing [is] no longer killing; and need not be experienced, or even perceived, as killing" (Lifton 1986:445).

            As one might expect in a culture in which indirect speech is extremely common, a similar euphemistic discourse was prevalent in DK. As in Nazi Germany, much of the violence was described in a medicalized manner. The elimination of enemies was justified as a necessary "cleansing" of "diseased elements" (Criddle and Mam 1987:164). This purification process continued throughout DK. In late 1976, Pol Pot stated: "there is a sickness in the Party . . . We cannot locate it precisely. The illness must emerge to be examined . . . If we wait any longer, the microbes can do real damage" (Chandler, Kiernan, and Boua 1988:183). The race to eliminate this "infection" (Picq 1989:100) led to increasingly violent purges both on the local level and within the upper echelons of the party itself. In the East Zone bordering Vietnam, for example, many individuals -- particularly the new people -- were executed because they were suspected of being "infected" or "contaminated" by a "pro-Vietnamese virus" (Stuart-Fox 1985).

            Local level DK cadres used a variety of other euphemisms to mask their genocidal deeds. People were sometimes threatened with being taken "into the forest," "to the field behind the village," or to "work on the mountain" (May 1986; Szymusiak 1986; Yathay 1987). In Cambodian culture, these places were associated with the "wild," the feared non-civilized domain in which violence took place (Chandler 1982). Before being led off to be tortured, imprisoned, and/or executed, people were often told that they were being taken "to learn" or "to see Angka" (e.g., Ngor 1987; Stuart-Fox 1985). Such palliative expressions served to disguise and legitimate the acts of violence that the Khmer Rouge cadre were committing.

            Moral Justification: Both dehumanization and the use of euphemisms make it easier for killers to morally justify their actions.  In the Stanford mock prison experiment, for example, "prisoners" constituted a category that it was legitimate to treat in a harsh manner. Similarly, since Jews were likened to a "disease" that threatened the larger German community, Nazis could regard their eradication as a necessary public health decision. In each case, the group of persecuted individuals was marked in such a way that normal moral standards no longer applied to them.

            In DK, this type of moral restructuring was facilitated by the local-level legitimation of the violent ethic and the ideological glorification of violence in general. As noted earlier, individuals suspected of being "enemies" of Ângkar were strongly dehumanized and subject to execution. Khmer Rouge cadre had been taught not only to hate such people but to destroy them without hesitation or pity. For a person like Comrade Chev, killing "enemies" was both a "political necessity" and a moral imperative: "If he purged enough enemies, he satisfied his conscience. He had done his duty to Angka" (Ngor 1987:229). As "traitors," those who criticized Ângkar's policies, stole food, attempted to escape, conspired against the state, or did not display the "proper" revolutionary spirit could be legitimately punished and/or executed (Picq 1989; Szymusiak 1986). Such violence against enemies was glorified by Khmer Rouge ideology, as illustrated by the frequent references to blood in the national anthem, political speeches, and revolutionary songs (Becker 1986; Jackson 1989b).

            This blood imagery also invoked another theme in Cambodian culture -- kum, or a grudge that often leads to disproportionate revenge. Kum is "a long-standing grudge leading to revenge much more damaging than the original injury. If I hit you with my fist and you wait five years and then shoot me in the back one dark night, that is kum . . . Cambodians know all about kum. It is the infection that grows on our national soul" (Ngor 1987:9). Because one of the most common situations in which Cambodians developed this type of grudge was when another person made them lose face or suffer, many peasants harbored resentment toward the rich and powerful who looked down upon them and enjoyed a much easier lifestyle. Khmer Rouge ideology played upon this resentment by explaining to its supporters that the rich had traditionally oppressed the poor and that the situation could only be changed through class struggle (another concept that invoked the warrior tradition). When they took power, many Khmer Rouge thus had a class grudge against "capitalists" who had often treated them with a lack of respect, were responsible for their impoverishment and suffering, and who had supported the Lon Nol forces that had killed many of their compatriots.

            A prescient awareness that the Khmer Rouge might attempt to take revenge upon the urbanites after obtaining power is evident in comments made by several survivors (Criddle and Mam 1987; May 1986; Stuart-Fox 1985). One person reported how "Khmer Rouge speakers publicly admitted, they had been fired by 'uncontrollable hatred' for members of the 'old society'. 'We were so angry when we came out of the forest', one speaker allegedly said, 'that we didn't want to spare even a baby in its cradle' (Chandler, Kiernan, and Muy 1976:9). Ngor also relates an anecdote about how, during a propaganda dance, costumed cadre would pound their chests with clenched fists and shout over and over again at the top of their lungs: "'BLOOD AVENGES BLOOD!' . . . Blood avenges Blood. You kill us, we kill you. We 'new' people had been on the other side of the Khmer Rouge in the civil war . . . Symbolically, the Khmer Rouge had just announced that they were going to take revenge" (Ngor 1987:140-1). Khmer Rouge could morally justify their killing as an act of revenge against people who had been responsible for traditional class inequalities and the wartime deaths of numerous comrades.

            Two other forms of moral justification for killing were also operative during DK: the use of palliative comparisons and of torture to extract confessions. First, the suffering and death of the people was sometimes legitimized by negatively comparing such privations to those the Khmer Rouge endured during the war. A common Khmer Rouge response to questions about overwork, disease, starvation, living conditions, and/or random execution was that "the revolutionaries suffered ten times worse than you during the war" (Ponchaud 1977:58). Through the use of such comparisons, Khmer Rouge cadre were able to minimize the harmful effects of the death and privations that the populace was enduring. Second, the Khmer Rouge often sent suspected "enemies" to prison centers where they were tortured until they confessed to their "crimes." The conditions in these "reeducation" centers were appalling: prisoners were often chained together in extremely hot and cramped quarters, severely beaten or disfigured, left lying in their excrement and urine, tortured in horrific ways, and/or randomly executed (May 1986; Ngor 1987; Stuart-Fox 1985). Such dehumanization made it easier to torture victims. Torture produced their confessions. Confession morally justified the entire process.

            Desensitization: Upon completing a post-war study, the U.S. armed forces were stunned to discover than only fifteen percent of trained combat riflemen reported having fired their weapons in World War II (Dyer 1985). On the basis of this information, basic training was revamped so that recruits would be desensitized to killing. Indoctrination techniques were geared to getting soldiers used to the idea of "wasting" an enemy. As one marine reported: "'Kill, kill, kill, kill,' It was drilled into your mind so much that it seemed like when it actually came down to it, it didn't bother you" (Dyer 1985:121).

            Khmer Rouge cadre seem to have undergone an analogous desensitization process. Recruits were put through an intensive indoctrination program that "filled their hearts and minds with a seething, unquenchable hatred for the [enemy]" (Sihanouk 1980:27). Propaganda meetings, ideological training, self-criticism, and membership in various associations were all geared toward producing this proper "revolutionary spirit" (Carney 1977). Extremely young people from the poorest segments of society were favored for the army, since they were, in Maoist terms, like "a blank page on which we can write anything" (Picq 1989:60).

            The Khmer Rouge had also become acclimated to violence during years of guerrilla warfare, U.S. bombing, and civil war. Many survivors have commented on how tough and battle-seasoned Khmer Rouge soldiers appeared when they victoriously entered Phnom Penh. Their attitude toward violence was often similar to that of Comrade Chev, for whom "the act of killing other human beings was routine. Just part of the job. Not even worth a second thought" (Ngor 1987:229). One informant explained to me that at first Khmer Rouge like Chev "wouldn't have the heart to kill people and might have even pitied their victims. After executing a few people, however, killing became normal to them, a way of proving their bravery." The banality of death was reinforced by DK ideology, particularly the local-level implementation of the violent ethic and the general glorification of blood sacrifice. Like the Nazi Doctors (Lifton 1986), some of the initiates to this world of violence reportedly had to get drunk before they could execute their victims. Eventually, however, many of these "soldiers were able to kill without being intoxicated. Some even learned to relish it and bragged about it afterward" (Criddle and Mam 1987:98-9). For such "agents of death," repeated exposure to violence gradually blunted their sensitivity to killing.

            Obedience to Authority: If many people were skeptical about Arendt's (1964) description of Eichmann as a normal bureaucrat who was epitomized the "banality of evil," these doubts were largely erased by Stanley Milgram's (1974) research on "obedience to authority." Milgram found that a large majority of his subjects would obey the experimenter's commands to continue shocking a "learner," even when they could clearly see that the voltage designation read: "Danger - Severe Shock." Afterwards, many subjects explained that they were only "doing as I was told." Such research illustrates that people will more readily commit acts of violence in situations in which they are given orders which "must be obeyed" and can displace blame onto authority figures.

            Strong social precursors for this type of obedience existed in Cambodian culture. Like other Southeast Asian countries, Cambodian political interactions often took place on the basis of patron-client relationships (bâks puok, khsae royeah). In return for protection and assistance in matters in which only someone with more power could be effective, a client incurred a moral debt which had to be repaid through loyalty, gifts, and obedience to her or his patron, whose own influence was thus increased incrementally. During DK, Ângkar became the new political authority to whom such loyalty and absolute obedience were due. People were told not to think -- "Angkar thinks for you" (Szymusiak 1986:151). Obedience was enforced by intimidation, spying, criticism sessions, indoctrination, and terror.

            In addition to the expectation that the populace would serve Ângkar, soldiers were trained to unquestioningly obey Ângkar's orders. Recruits returning from indoctrination training were reported to have been converted into well-disciplined "fanatics who would not deviate from [a] prescribed course of action" (Quinn 1976:24). Given that the violent ethic had been legitimated on the local level, this military ethos of strict discipline and obedience remained in force throughout DK. "In the civil war [such soldiers] had been trained to kill Lon Nol forces. When they were ordered to kill 'new' people on the front lines they obeyed automatically" (Ngor 1987:229). One Khmer Rouge executioner told me that he killed because "I had to obey the orders of my superior. If they ordered me to do something, I would do it.  If we didn't obey, we would have been killed."

            Such obedience was reinforced by group norms. Cambodians place great importance upon maintaining face because honor is gained through the respect and obedience of others (Hinton n.d.). In social interactions, Cambodians remain extremely concerned about how others evaluate the way they perform the duties expected of someone of their status. This desire to gain honor through the positive evaluations of others contributed to group conformity. In DK, Khmer Rouge soldiers belonged both to Ângkar and to smaller units whose survival often depended on one another. In addition to following military discipline and being loyal to their leaders, individuals would also have felt pressured to obey orders for fear of losing face in front of their comrades. One person explained that if a soldier didn't obey an order to kill, "it meant that she or he was a coward, the most inferior person in the group, the one who had lost to the others." Moreover, those who distinguished themselves in the performance of their duty gained honor and might be promoted. The result was that "when the order came from Angkar to kill, they obeyed" (Stuart-Fox 1985:145).

            The lack of contrition evinced by some of these Khmer Rouge was in part due to their ability to deflect responsibility for the acts they were perpetrating. Orders came from Ângkar and thus were not the responsibility of the individuals actually committing acts of violence. Relocation orders could be explained by saying: "This is Angka's rule, not my rule" (Ngor 1987:195). Torturers would tell their victims to "please tell Angka the truth" (Ngor 1987:242). Since Ângkar represented "everyone," accountability for such actions could be successfully diffused.

            The Conversion Process: How are people like Comrade Chev converted into genocidal killers? As noted previously, the specific inputs required to turn someone into an "agent of death" will vary depending on that individual's life history. One, several, or all of the PSD reduction strategies discussed above may be pivotal in creating people who can commit acts of evil. These people will draw upon the "state level response" as they make their own "individual level response" to the PSD that arises when they kill. Such cognitive restructuring involves a dialectic in which complex processes interact to push the individual along the "continuum of destructiveness" (Staub 1989; see also Darely 1992). The exclusion and devaluation of a group of individuals sets them outside of a given community. Dehumanization morally justifies the harm of these people. By using euphemisms and deflecting responsibility onto authority figures, any remaining culpability can be diffused. As people are harmed, the perpetrators become acclimated to violence. Desensitization makes the dehumanization of victims seem more normal.

            Many Khmer Rouge would have experienced PSD when they were asked to kill people who had previously been members of their community (i.e., when the violent and gentle ethics came into conflict). These "agents of death" reduced their PSD in a number of ways. First, they changed these cultural models (i.e., they were able to do this because the DK regime effectively glorified violence and undermined the gentle ethic in the creation of the "killing fields"). Second, these actors changed their behavior (i.e., once they had killed, such violence became more routine). Third, Khmer Rouge cadre acted in an environment that had been dramatically altered (i.e., DK ideology not only legitimated but actually glorified the violent ethic in local level interactions). Finally, these individuals were able to add new, lower level schemas that made the larger cultural models more consonant (i.e., through dehumanization, the use of euphemism, moral justification, the deflection of responsibility). The result of this process of cognitive restructuring was the creation of "agents of deaths" like Comrade Chev who could commit genocidal atrocities.

Conclusion: Anthropological Thoughts on Genocide

There are blows in Life so violent - I can't answer!

Blows as if from the hatred of God; as if before them,

the deep waters of everything lived through

were backed up in the soul . . . I can't answer!

Not many; but they exist . . . They open dark ravines

in the most ferocious face and in the most bull-like back.

Perhaps they are the horses of that heathen Atilla,

or the black riders sent to us by Death.

-- Cesar Vallejo, "The Black Riders"

            While I agree with Cesar Vallejo that it is difficult to comprehend how people can come to kill each other so easily, I do think we can to understand events like large-scale genocide. When confronted with the horror of genocide, scholars have all too often turned away from exploring the origins of such violence. Anthropologists have been particularly negligent in this respect. Perhaps the relativistic ethos that guards "our" cultures of study has prevented us from making stronger statements about the morality of genocidal events. As is evident from this essay, I do not think that this is a responsible stance. It is precisely because anthropologists can make insightful analyses of the conditions that facilitate large-scale genocide that we must begin to examine this topic.

            This essay has been an attempt to initiate such a dialogue by drawing on both psychology and anthropology. I would argue that, before discarding western psychological concepts, we must first determine if they can be productively applied cross-culturally -- albeit in a form modified in terms of anthropological understanding. By doing so, we can enter into a interdisciplinary debate about genocide. Psychosocial dissonance represents one concept that can be fruitfully forged from a synthesis of psychological and anthropological insights and can help us understand the genocidal events that took place in Cambodian and, potentially, elsewhere.

            This essay has not only shown the ways in which perpetrators of violence attempt to deal with PSD, but also how a genocidal regime deliberately manipulates cultural models to transform the consciousness of these "agents of death." Because it can account for both microlevel and macrolevel factors, the model of PSD I propose is able to provide a more complete explanation than can either a "top-down" or "bottom-up"  perspective alone. For example, a macrolevel approach which assumes that genocidal killers are homogeneous automatons who blindly follow state ideology cannot explain variation in the "individual-level response" of disparate actors. Likewise, a microlevel framework that focuses primarily on psychodynamic processes has difficulty accounting for how the "state-level response" shapes the contexts in which genocidal violence takes place. Because a model of PSD is able to draw insights from both the micro and macro levels of analysis in a culturally-sensitive manner, it represents a distinct anthropological contribution to our understanding of the origins of large-scale genocide. Given their work on violence and other complex topics, anthropologists can develop many more insightful analyses that increase our understanding of large-scale genocide even further. I urge them to begin doing so.

            One of the most disturbing conclusions that has emerged from this analysis is that genocide does not seem to be something only "sadistic killers" are capable of performing. Long ago, Arendt correctly pointed out that such killing can become banal, something almost anyone can come to do with appropriate training and/or cognitive restructuring. If this is true, how can we prevent genocide? First, as indicated above, we can develop analytical constructs that provide insight into how genocide arises and, once it is occurring, how it can be stopped. For example, the notion of PSD helps us understand how people can be converted into "agents of death" through a combination of ideological, sociocultural, and cognitive changes. To recognize genocide in the making, we can therefore monitor a given society's "state level response" to see if it is setting up the preconditions for genocidal PSD reduction (e.g., by undermining cultural models that foster nonviolence, promoting violence against devalued groups through ideology, producing a euphemistic discourse of destruction). Similarly, once genocidal violence is taking place, we can attempt to interfere with, or counteract, the "individual level response" that converts a person into an "agent of death." Since dehumanization is so crucial to the conversion process, we can take steps to "rehumanize" victims. Obviously, each case must be examined separately to determine exactly how to prevent/stop genocidal killers. While remaining sympathetic to the spirit of Vallejo's words, we must try to produce such sensitive analyses of genocide. We can answer.

Notes

Acknowledgments: I would like to gratefully acknowledge the funding sources that have supported my research: AAS SEAC Small Grant, FLAS, ISWP Dissertation Fellowship, NIH NRSA Fellowship, NSF Graduate Fellowship, NSF Doctoral Dissertation Improvement Grant, and an SSRC International Doctoral Research Fellowship. In addition to expressing my appreciation to Barbara and Dennis Tedlock, Howard Stein, David Griffith, and two other anonymous reviewers, I want to thank Nicole Cooley, David Chandler, Bruce Knauft, Fredrik Barth, May Ebihara, and Bradd Shore for their helpful comments on the manuscript.

References Cited

Arendt, Hanna

   1964  Eichmann in Jerusalem. New York: Viking Press.

Aronson, Elliot

  1992  The Return of the Repressed: Dissonance Theory Makes a Comeback. Psychological Inquiry 3:303-352.

Bandura, Albert

   1986  Social Foundations of Thought and Action. Englewood Cliffs, NJ: Prentice-Hall.

Becker, Elizabeth

   1986  When the War as Over. New York: Touchstone.

Bit, Seanglim

   1991  The Warrior Heritage. El Cerrito, CA: Seanglim Bit.

Bun Chân Mol

   1973  Châret Khmaer. Phnom Penh.

Carney, Timothy Michael

   1977  Communist Party Power in Kampuchea (Cambodia). Data Paper No. 106.

   Ithaca: Cornell University Southeast Asia Program.

Chandler, David P.

   1982  Songs at the Edge of the Forest: Perception and Order in Three Cambodian Texts. In Moral Order and the    

   Question of Change. David K. Wyatt and Alexander Woodside, eds. Pp. 53-77. New Haven: Yale University        

   Southeast Asia Studies.

   1991  The Tragedy of Cambodian History. New Haven: Yale University Press.

   1992  A History of Cambodia. Boulder: Westview Press.

Chandler, David P., Ben Kiernan, and Chanthou Boua

   1988  Pol Pot Plans the Future. Monograph Series 33. New Haven: Yale University Southeast Asia Studies.

Chandler, David P., Kiernan, Ben, and Muy Hong Lim

   1976  The Early Phases of Liberation in Northwestern Cambodia: Conversations with Peang Sophi. Melbourne:          

   Monash University Centre of Southeast Asian Studies Working Papers.

Connor, John W.

   1989  From Ghost Dance to Death Camps: Nazi Germany as a Crisis Cult. Ethos 17(3):259-288.

Criddle, Joan D., and Teeda Butt Mam

   1987  To Destroy You is No Loss. New York: Anchor.

D'Andrade, Roy

   1992  Schemas and Motivation. In Human Motives and Cultural Models. Roy D'Andrade and Claudia Strauss, eds.

   Pp. 23-44. New York: Cambridge University Press.

   1995  The Development of Cognitive Anthropology. New York: Cambridge University Press.

D'Andrade, Roy and Claudia Strauss, eds.

  1992  Human Motives and Cultural Models. New York: Cambridge University Press.

Darley, John M.

   1992  Social Organization for the Production of Evil. Psychological Inquiry 3(2):199-218.

Delvert, Jean

   1961  Le Paysan Cambodgien. Paris: Mouton.

DeVos, George A.

   1975  The Dangers of Pure Theory in Social Anthropology. Ethos 3(1):77-91.

Dyer, Gwynne

   1985  War. New York: Crown Publishers.

Ebihara, May

   1968  Svay, A Khmer Village in Cambodia. Ph.D. dissertation, Anthropology Department, Columbia University.

   1990  Revolution and Reformulation in Kampuchean Village Culture. In The Cambodian Agony. David A. Ablin and

   Marlowe Hood, eds. Pp. 16-61.              

    Armonk, NY: M. E. Sharpe.

Ewing, Katherine P.

  1990  The Illusion of Wholeness: Culture, Self, and the Experience of Inconsistency. Ethos 18(3):251-279.

Festinger, Leon

  1957  A Theory of Cognitive Dissonance. Evanston, IL: Row, Peterson.

Geertz, Clifford

   1983  Local Knowledge. New York: Basic Books.

Goffman, Erving

   1967  Interaction Ritual: Essays on Face-to-face Behavior. New York: Pantheon.

Hinton, Alexander L.

  1994  Cultural Models and the Study of Aggression: An Examination of the Psychosocial Origins of Cambodian

   Violence. Southern Anthropologist 21(2):11-25.

   n.d.  Why Did You Kill? Face, Honor, and the Struggle for Power Spit in the Cambodian Genocide. Unpublished

   manuscript.

Holland, Dorothy and Naomi Quinn, eds.

  1987  Cultural Models in Language and Thought. New York: Cambridge University Press.

Jackson, Karl D.

   1989a  Cambodia 1975-1978. Princeton: Princeton University Press.

   1989b  The Ideology of Total Power. In Cambodia: 1975-1978. Karl Jackson, ed. Pp. 37-78. Princeton: Princeton

   University Press.

Jacobs, Judith M.

   1986  Reamker (Ramakerti): the Cambodian Version of the Ramayana. London: The Royal Asiatic Society.

Kelman, Herbert C., and V. Lee Hamilton

   1989  Crimes of Obedience. New Haven: Yale University Press.

Kunda, Ziva

   1990  The Case for Motivated Reasoning. Psychological Bulletin 108:480-498.

Kuper, Leo

   1981  Genocide. New York: Penguin.

Lanzmann, Claude

   1985  Shoah: An Oral History of the Holocaust. New York: Pantheon.

Ledgerwood, Judy

   1990  Changing Khmer Conceptions of Gender: Women, Stories, and the Social Order. Ph.D. Dissertation,             

   Anthropology Department, Cornell University.

Lewin, Carroll McC.

  1992  The Holocaust: Anthropological Possibilities and the Dilemma of Representation. American Anthropologist

  94:161-166.

  1993  Negotiated Selves in the Holocaust. Ethos 21:295-318.

Lifton, Robert Jay

   1986  The Nazi Doctors. New York: Basic Books.

Martel, Gabrielle

   1975  Lovea, Village des Environs d'Angkor. Paris: Publications d'École Française d'Extrême-Orient.

May, Someth

   1986  Cambodian Witness. New York: Random House.

Milgram, Stanley

   1974  Obedience to Authority. New York: Harper and Row.

Moyer, Nancy

   1991  Escape from the Killing Fields. Grand Rapids, MI: Zonervan Publishing House.

Nagengast, Carole

   1994  Violence, Terror, and the Crisis of the State. Annual Review of Anthropology 23:109-36.

Ngor, Haing

   1987  A Cambodian Odyssey. New York: Warner.

Nuckolls, Charles W.

   1993  The Anthropology of Explanation. Anthropological Quarterly 66(1):1-21.

Picq, Laurence

   1989  Beyond the Horizon: Five Years with the Khmer Rouge. New York: St. Martin's. Ponchaud, François

   1977  Cambodia Year Zero. New York: Holt, Rinehart and Winston.

   1989  Social Change in the Vortex of Revolution. In Cambodia 1975-1978. Karl Jackson, ed. Pp. 151-177. Princeton:

   Princeton University Press.

Quinn, Kenneth M.

   1976  Political Change in Wartime: The Khmer Krahom Revolution in Southern Cambodia, 1970-1974. Naval War

   College Review 28(Spring):3-31.

Rosaldo, Michelle Z.

   1984  Toward an Anthropology of Self and Feeling. In Culture Theory: Essays on Mind, Self, and Emotion. Richard

   A. Shweder and Robert A. LeVine, eds. Pp. 137-157. New York: Cambridge University Press.

Schanberg, Sidney H.

   1975  The Enigma of Khmer Rouge Purpose. Saturday Review August 23:29-34.

Schlenker, Barry R.

   1992  Of Shape Shifters and Theories. Psychological Inquiry 3(4): 342-344

Shank, Roger C. and Robert P. Abelson

  1977  Scripts, Plans, Goals, and Understanding. Hillsdale, NJ: Lawrence Erlbaum.

Shiloh, Ailon

   1975  Psychological Anthropology: A Case Study in Culture Blindness? Current Anthropology 16(4):618-620.

Shore, Bradd

   1996  Culture in Mind: Cognition, Culture, and the Problem of Meaning. New York: Oxford University Press.

Sihanouk, Prince Norodom

   1980  War & Hope. New York: Pantheon.

Staub, Ervin

   1989  The Roots of Evil. New York: Cambridge University Press.

Stein, Howard F.

   1993  The Holocaust, the Self, and the Question of Wholeness: A Response to Lewin. Ethos 21(4):485-512.

Stuart-Fox, Martin

   1985  The Murderous Revolution. Chippendale, Australia: Alternative Publishing Cooperative Limited.

Szymusiak, Molyda

   1986  The Stones Cry Out. New York: Hill and Wang.

Thion, Serge

   1990  The Pattern of Cambodian Politics. In The Cambodian Agony. David A.          

   Ablin and Marlowe Hood, eds. Pp. 149-164. Armonk, NY: M. E. Sharpe.

   1993  Genocide as a Political Commodity. In Genocide and Democracy in Cambodia.        

   Ben Kiernan, ed. Pp.163-190. Monograph Series 41. New Haven: Yale University Southeast Asia Studies.

Vickery, Michael

   1984  Cambodia: 1975-1982. Boston: South End Press.

Yathay, Pin

   1987  Stay Alive, My Son. New York: Touchstone.

Zimbardo, Philip G., Craig Haney, W. Curtis Banks, and David Jaffe

   1974  The Psychology of Imprisonment: Privation, Power, and Pathology. In Doing Unto Others. Zick Rubin, ed. Pp.

   61-73. Englewood Cliffs, NJ: Prentice-Hall.