Veteran Dutch sociologist Abram De Swaan’s new study is about “mass annihilation”: killings on a very large scale of human beings who lack means to defend themselves, with the support of the dominant regime, primarily in the nineteenth and twentieth centuries. De Swaan opts not to ground his sociological analysis in legal terms like “genocide,” which necessitate detailed legal discussions of governments’ intentions to destroy entire groups. Instead, his focus is on the rank-and-file perpetrators who are in close contact with their victims. Their lack of restraint is framed in Eliasian terms: for centuries violence had been moving out of public view, and such destructive impulses had become increasingly regulated. How, then, did such perpetrators become so ubiquitous in the nineteenth and twentieth centuries?

De Swaan aims to distinguish himself from two positions. On the one hand, he does not resurrect the long-discredited view that “genocidaires” are pathological. But on the other hand, he criticizes what he sees as the dominant “situationist” position, according to which immediate social situations can explain murderous behaviors, regardless of long-term dispositions. De Swaan sees these dispositions as a crucial part of the answer, since they explain differences in the way people react to such extreme situations. I shall return below to some of his intriguing critiques of the situationist position.

The problem of “decivilization,” or, more specifically, the breakdown of long-term trends towards the restraint of public violence, has already been tackled by Elias himself in The Germans [1996]. But whereas Elias analyzed one historical case in detail (Germany, with its lack of an influential “civilizing” court society, and the resulting preservation of a militaristic ethos), De Swaan’s strategy is to cover a vast range of cases in order to formulate generalizations. These include the Armenian Genocide, the Holocaust, Cambodia and Rwanda; Stalin’s terror and famines, together with those of Mao; atrocities in the Belgian Congo, Partition in India, mass killings in Indonesia, and several other violent episodes.

The author’s response to the problem he identifies comes in two parts. First, he describes conditions through which uncivilized violent impulses could be given full expression, despite long-term civilizational trends. The historical creation of these conditions is named “compartmentalization,” and it is initially defined as a multi-level process: psychological dis-identification with “different” others, lack of social interaction with them, exclusion from social institutions, and finally deporting them to separate spaces of extermination [11]. Second, he posits several social and psychological mechanisms through which perpetrators accustom themselves to this compartmentalization: conqueror’s frenzy (soldiers’ rage resulting from brutalization in battle), rule by terror (violence carried out by trained specialists), losers’ triumph/turning passive into active (a situation where those who are about to lose a war alleviate their anxiety by victimizing defenseless others), and megapogroms—“outbursts of improvised, local mob violence” [161] occurring simultaneously in dozens of locations. These are all defined as “modes of mass annihilation.”

While the book covers a very wide range of cases, it is clear that the underlying paradigmatic example is the Holocaust (described in the Preface as a “formative event” for the author). And indeed, the description of compartmentalization mirrors conventional interpretations of the history of German Jews under Nazism—from their continuous demonization by the regime, through their gradual isolation from their fellow citizens, until their eventual deportation to concentration camps in the East. However, as De Swaan surveys his other cases, it becomes apparent that “compartmentalization” does not apply to them as effectively as to this key case. Often, the violence was not aimed at a clearly defined “other”: under Stalin, targeted “individuals [were] not that easily classified, [and] new distinctions had to be applied forever” [155-156]; in Mao’s China “the target groups were fuzzy and might change at any moment” (160); in Indonesia, a purge of “communists” was a cue to local bands to kill “whomever they wanted,” including “Chinese shopkeepers and traders” as well as “personal enemies” [165-166]. Similarly, there was frequently no spatial compartmentalization: in India, as well as in Rwanda, victims were attacked by their neighbors within their homes [111; 198]; during the Armenian Genocide, many of “the massacres were hardly hidden from view; anyone who wanted could witness them” [175].

The second part of the theory—the “modes of mass annihilation”—fares no better empirically. The Belgian King Leopold II’s entire rule in Congo, from 1887 to 1908, which led to the death of an estimated 10 million inhabitants, is unconvincingly attributed to an initial “conquerors’ frenzy” (stemming from the brutalization of troops in battle). The Cultural Revolution is presented under the rubric of “rule by terror,” executed by specialists, despite its being “a free-for-all, especially for young students […] [I]n this instance the actors were not professionals” [161]. The Khmer Rouge in Cambodia are said to have “turned passive into active,” victimized helpless civilians in the face of imminent defeat—but they engaged in genocide from 1975 to 1979, whereas the Vietnamese invasion that toppled the regime lasted a mere two weeks, at the end of this period. Indian Partition is initially explained as a “megapogrom” of improvised mob violence, yet this is later refuted by citing “recent research, [which] has established a pattern of large-scale and effective organization behind the ‘spontaneous’ mob attacks” [202]. It turns out that these attacks were initiated and coordinated by local politicians [201]. Of course, there are some cases which do fit the theses better (Stalin’s secret police are more convincingly described as experts), but the number of exceptions is striking.

As the book progresses, De Swaan seems to retreat from his initial hypotheses. He admits, for instance, that in Turkey, “the genocidal episodes were not thoroughly compartmentalized” [175]. Instead, there is an increasing reliance on hypothesized psychological mechanisms. This leads to a shift of emphasis towards so-called “mental compartmentalization,” where the genocidaires’ original (moral) self can “disavow the other self, thus enabling it to commit whatever violent act without compromising the original self” [248]. Genocidaires are also said to have a low sense of agency, as well as reduced empathy [247]. This new set of hypotheses appears very late, in the final chapter, and although they are plausible (“few perpetrators ever uttered a word of commiseration with their victims”—228), there is not enough attempt to connect these psychological dispositions to the wider social processes that could have created them (the discussion of Germans being brutalized by combat during World War I on p. 222 is too brief). In addition, almost all the evidence in this chapter is drawn from research on the best-known perpetrators, the Nazis, and De Swaan can only speculate on its relevance to the other cases.

A well-known essay on ethnic and nationalist violence by Brubaker and Laitin concludes [2004: 115]: “[t]here is no reason to believe that these large-scale components of large-scale ethnic violence can be understood or explained through a single theoretical lens. Rather than aspire to construct ‘a theory’ of ethnic and nationalist violence—a theory that would be vitiated by its lack of meaningful explanandum—we should seek to identify, analyze, and explain the heterogeneous processes and mechanisms involved […] This can be accomplished only through a research strategy firmly committed to disaggregation in both data collection and theory-building.” This advice seems to apply to the no less problematic category of “mass annihilation.” While they all ended in massive death, the Holocaust, Belgian atrocities in the Congo and the riots of Partition in India may be too dissimilar to be analyzed within a single theoretical framework. I was not convinced that there really are parallel mechanisms underlying these highly heterogeneous cases.

However, De Swaan does provide some hints that point in other, more promising, directions. Early in the book, in one of its most provocative chapters, he reviews and critiques key works in what he sees as the “situationist” tradition. One of these is Milgram’s famous obedience experiments, commonly cited as proof that most ordinary people will commit atrocities if ordered to do so. As readers will recall, authoritative experimenters managed to pressure participants to administer what these participants thought were a series of increasingly strong electric shocks to a group of “learners,” after the latter failed to memorize sequences of words (in reality, the situation was simulated, and “learners” only pretended be harmed as part of the experiment). But as De Swaan astutely points out, “if the situation had been ‘real’, the compliant subjects would have risked electrocution themselves, since the penalty for murder in the state of Connecticut was still the electric chair” [29]. What explains the participants’ behavior was not simply their inclination to obey authority, but “significant ambiguities in the presentation of the experiment to the subjects”: the experimenter continuously assured them that there was “no danger” involved, but the buttons they pressed were explicitly labeled “danger, severe shock.” This ambiguity generated what De Swaan calls a “gamelike” aspect to the situation [29], which better explains the obedience effect.

Clearly, in the actual cases described in the book, violence was real, not simulated. Nevertheless, many of the empirical descriptions (certainly not all) make references to this ambiguous definition of the victims and of the killings: in Rwanda, local Tutsi civilians were presented as a dangerous “fifth column,” potential military allies of an external Tutsi-led front that was to eventually topple the regime [107]; the 1932-1934 famine in the Ukraine was a result of grain being requisitioned, but under Stalin severe food shortages had been a recurrent result of forced collectivization [156]—another possible source of ambiguity; in Indonesia, attacks on “Communists” led to the death of shopkeepers and traders who happened to be Chinese [165]; riots in India seemed like spontaneous religious outbursts, but were engineered by political bosses to rid themselves of local competitors, and to acquire the latter’s property [201]. The use of euphemisms such as the “final solution” to cloak killings may also be relevant here.

My point is not that as researchers we need to accept these situations’ ambiguity at face value, as a justification for the atrocities that occurred. Rather, I argue that if we are interested precisely in De Swaan’s problem—how to explain the participation of the perpetrators—we would be well advised to follow the production and the performance of the violence, alongside the way in which its representation was produced and reproduced in other spaces by the immediate perpetrators and by less visible faraway authorities. This should help us understand the conditions for the aforementioned “ambiguity” and its effect on the perpetrators’ actions and self-justifications. Such an approach may apply to at least some of the cases described above, in which classifications of victims became dynamic and fuzzy. In addition, the way in which the violence was interpreted likely differed based on the previous dispositions of particular groups: this could be a way of incorporating those dispositions into the analysis, in order to avoid an extreme “situationism.”

Specific case studies, like the ones De Swaan cites, have already uncovered several of the mechanisms through which massive violence is performed and represented. In my own work, I have described how coalitions of American settlers managed to officially define their indiscriminate killing of Native Americans as “war,” and how such definitions were sometimes overturned by competing coalitions of other settlers [Pessah, Reference Pessah2014]. There may, however, be new insights yet to be drawn from systematically comparing particular manifestations of these mechanisms, if we remember not to lump them together too hastily. Read in this light, The Killing Compartments does provide us with some thought-provoking observations and materials that could stimulate further research.