powerful organize how we think and thereby what we think.
Herbert Kelman (1973) theorizes institutional sway over cognition in a morally extreme action context—that of massacring innocents. Kelman first refutes individual explanations. He notes that psychologically normal people are seen to perpetrate this kind of violence. Nor, in his view, do emotions such as frustration and hostility motivate such violence, though they often accompany it. For Kelman anger and sadism are not “major motivating forces in their own right” (p. 36) but at best facilitators and even outcomes of the actual causal factors, which are conditions that weaken erstwhile moral restraints against violence. Such weakening is the result of three social processes—authorization, routinization, and dehumanization: “Through processes of authorization, the situation becomes so defined that standard moral principles do not apply and the individual is absolved of responsibility to make personal moral choices. Through processes of routinization, the action becomes so organized that there is no opportunity for raising moral questions and making moral decisions. Through processes of dehumanization, the actor’s attitudes toward the target and toward himself become so structured that it is neither necessary nor possible for him to view the relationship in moral terms” (p. 38). Although Kelman scarcely refers to cognition per se, the three morality-attenuating processes that he identifies operate cognitively. Authorization and dehumanization lead us to invest in particular definitions of the situation, self and target, for example, a view of compulsion by (or paramount loyalty toward) authorities, and the idea that the victim is not part of the human community.3 Authorization can also work more subtly and more pervasively when authoritarian regimes cast doubt on facts and the media that disseminate them; in this way the government delegitimizes potentially contradictory ideas. Routinization, like bureaucratization, is a material process, “transforming the action into routine, mechanical, highly programmed operations” (Kelman 1973, p. 46). Kelman’s theory echoes Arendt’s (1963) “banality of evil thesis” according to which the Nazi regime transformed people carrying out the genocide into “functionaries and mere cogs” (p. 289). Routinization operates via ideas in the sense that it helps actors avoid contemplating what they are doing or the outcomes they are contributing to: “Routinization fulfills two functions. First, it reduces the necessity of making decisions, thus minimizing occasions in which moral questions may arise. Second, it makes it easier to avoid the implications of the action since the actor focuses on the details of his job rather than on its meaning” (Kelman 1973, p. 46). Kelman’s moral actor is situated to think certain things and, more important, not to think others. Not thinking about the ethics of one’s actions—or inactions—is highly consequential and wholly socialized. Thus, too, Bandura (1999) refers to “decisional arrangements of foggy nonresponsibility” where “authorities act in ways that keep themselves intentionally uninformed” (p. 197). These arrangements insulate actors from sanction in the event that harm is publicized, as well as self-sanction, for these actors “also have to live with themselves” (p. 197).
Not thinking about what one is doing brings us to habit and Bourdieu’s (1977) concept of habitus, “a system of lasting, transposable dispositions which, integrating past experiences, functions at every moment as a matrix of perceptions, appreciations, and actions” (pp. 82–83; emphasis in original). Perception, appreciation, and action are enmeshed in this perspective. Habitus is embodied, something on the order of playbook moves, not necessarily thought of yet reciprocally related to ideas as well as other structures. Discursive forms are among the structures that inculcate and reflect habitus (see Fleetwood 2016).
DISCURSIVE FORMS
Ideas and mechanisms that manipulate ideas and thinking are made concrete and disseminated through semiotic processes. Messages of harm’s legitimacy are embodied in particular statements. For example, boys’ routine violence lends itself to a popular figure of speech called an epanelepsis: “Boys will be boys.” Is it significant that the normalization of boys’ violence is captured in discourse? In other words, does discursivity have a unique impact?
The answer from many quarters is yes. Philosophers and psychologists have advanced a variety of theses to the effect that language constitutes thought. Carruthers (2002) observes that the weakest of these, that “language makes some cognitive difference” (p. 659), enjoys broad support. Yet, he gathers evidence for the stronger claim that language is a necessary medium of thought across domains. Language both integrates and communicates ideas, whereas other “modules” such as the visual do not. That is, language “has both input and output functions” (p. 666). We make the world meaningful to ourselves and to others through language.
But texts do not merely clarify; they also make things happen (Austin 1962). Or, as Barthes (1957) says of myth, “It makes us understand something and it imposes it on us” (p. 117). Texts establish positions and institutions. Even where the ideas that inspired them are not consciously received or even accepted, discourses govern through the hierarchies they construct. Thus, van Dijk (1992) takes note of the discursive strategies with which racialized hierarchies are maintained. Where overtly racist expressions are inconsistent with prevailing norms, whites use strategies like reversal, where “anti-racists tend to be represented as the ones who are intolerant” (van Dijk 1992, p. 94). Van Dijk (1993) situates the role of language this way: “A discourse analytical approach does not imply that we reduce the problem of racism to a language or communication problem. Obviously, racism also manifests itself in many non-discursive practices and structures, such as discrimination in employment, housing, health care, and social services, or in physical aggression. Our major claim and interest, then, are twofold: (1) Racism also manifests itself in discourse and communication, often in relation with other social practices of oppression and exclusion, and (2) the social cognitions that underlie these practices are largely shaped through discursive communication within the dominant white group” (p. 13). Here, discourse both reflects and molds thought. The idea of that reciprocal relation is compatible with Foucault’s (2000) perspective on the political impact of ideas as discourses, which flow from particular “regimes of truth.” Among discourses I distinguish between what I call wording and narrative.
Wording
By wording I mean to demarcate linguistic processes that are more or less contained in a few words and whose use is relatively flexible across statements. They may amount to systems, but it is useful to focus attention on distinct devices—words or groups of words.4 For example, we find across harms the negative labeling of targets: “The use of labels helps to deprive the victims of identity and community. Terms like ‘gook’ help to define them as subhuman, despicable, and certainly incapable of evoking empathy. Terms like ‘Communist’ allow their total identity to be absorbed by a single category, and one that is identified by the perpetrators of the massacre as totally evil” (Kelman 1973, p. 50).
Not just labels, but labeling schemes—classifications, or “differentiations” (Lévi-Strauss 1966)—are a vital aspect of the enculturation of harm. A raft of studies has turned up classificatory systems that are essential to war, colonialism, slavery, prostitution, and more, distinguishing, for example, virgin and whore, ruler and subject, Occidental and Oriental, and master and slave. Abortion is opposed by reference to “killing babies” (Lakoff 2002). The sharp distinction between victims and offenders, however those roles are cast, is essential to punishment schemes. Yet, harm-promoting classifications are not necessarily binary, illustrated by Lombroso’s (1876) typology of criminals (e.g., the born criminal, criminaloids, the criminal by passion, and others) and South Africa’s rule of apartheid (black, white, Coloured, and Indian).
The harm-inducing effects of labels that pertain to deviance and criminality are central to the labeling perspective in sociology. The labeling perspective applies to mass harm to the extent that it concerns itself with groups and definitions of crime that criminalize them, following conflict theory (see, e.g., Quinney 1970). The labeling of target groups facilitates their mistreatment. Numerous studies have inventoried figurative expressions for victims of mass harm. In particular, much has been written about the dehumanizing terms that
