Chapter 5. Ritual Cannibalism: A Case Study of Socially-Sanctioned Group Violence

5.1 Introduction

This chapter argues that ritual or customary cannibalism shares a number of features with war, which make it a relevant precedent for the practice and potential abolition of war. Like war, ritual cannibalism:

• was relatively widespread (among hunter-gatherer societies) for centuries or millennia;
• was institutionalized as a socially-sanctioned and, often, socially-required practice, in which various members of society played well-defined roles;
• involved a violation of the human body which, in the abstract, was perceived as abhorrent and hateful; and
• was undertaken despite its abhorrent nature because it was believed to be justified by an important social value.

Finally, as I argue is likely to be the case for the abolition of war, when key features of the culture changed, the culturally-determined reasoning that had justified customary cannibalism faded away; and once the justifying reasons disappeared, the practice not only stopped, but became an unthinkable horror.

The Meaning of Ritual Cannibalism

In everyday usage, the word ‘cannibalism’ conjures up the image of a tribe boiling or roasting human flesh for dinner. This is a caricature of actual cannibal practices, an ancient myth perpetuated, largely inadvertently, by modern explorers and conquerors. The evidence available today suggests that regularly practiced, socially-sanctioned ‘food cannibalism’ may have occurred in one or two cultures, but that if it ever did exist, it was extraordinarily rare.

In contrast, starvation cannibalism—individuals painfully violating an internalized taboo against consuming human flesh—has occurred in times of famine in all parts of the world. Some cases of starvation cannibalism have involved murdered victims, others have involved flesh taken from dead bodies; but there is no evidence that the practice was ever socially sanctioned as customary behavior. Cannibalism as an aberrant (not sanctioned) practice has also occurred as a product of sociopathic behavior and mental illness.

Virtually all socially-sanctioned cannibalism—and nearly all of the practices which anthropologists study—are very different from food cannibalism. These practices typically involve the ritual or customary ingestion of some part of a dead member of one’s own band or tribe (‘endocannibalism’) or of a stranger or a member of an enemy tribe (‘exocannibalism’) in the belief that this will have a particular impact on the life or the after-life of the consumer or the consumed, or both. In acts of ritual or customary cannibalism, specified members of a society consume all or part of certain body parts, which are believed to contain and convey attributes of the deceased. Often, the organ to be consumed is the heart, the brain, or the liver, and the attribute to be preserved (endocannibalism) or appropriated (exocannibalism) is the soul stuff, the spirit, the courage, or the fecundity of the deceased.

In many cases of ritual or customary cannibalism, the part of the human body that is ingested does not take the form of food or resemble food; in some cases, it is obscured by food. For example, the ash of the burnt bones of a deceased relation may be dissolved in a drink; or some of his or her blood may be mixed in with some form of food or drink. Even when ritual cannibalism does involve the consumption of flesh in more substantial quantities and gruesome forms, it is still utterly unlike food cannibalism. Ritual or customary cannibalism occurs only under certain circumstances, and then at a special time set aside for the purpose. The ingestion of human substance is usually just one step in a much longer and more elaborate ritual. The activity is generally considered to be of the utmost importance as a social obligation or sacred duty.

The Debate over the Existence of Ritual Cannibalism

To the modern mind, cannibalism in any form is so abhorrent that it is hard to credit. We find it difficult to comprehend or even imagine cannibal acts undertaken as a freely-chosen activity by ordinary, sane, non-sadistic, non-sociopathic human beings. The gulf between the modern sensibility and that in societies with ritual cannibal practices is so wide that it undermines confidence in our ability to separate fact from fiction, or to avoid ethnocentric projections of feared, horrible ‘otherness’ onto peoples whose cultures are vastly different from our own. Moreover, there were well-known tendencies among the explorers, traders, and conquistadors of earlier centuries to exaggerate the adventures and encounters they experienced during their travels, sometimes bringing to life the monsters of myth, fairy tale, and nightmare, and to plagiarize fantastic stories from one another.

Thus, it is not altogether surprising that some years ago an anthropologist challenged the accepted view that various forms of socially-sanctioned cannibalism were once relatively widespread. The 1979 book by William Arens, The Man-Eating Myth: Anthropology and Anthropophagy, sent a shock wave through the field of anthropology with the claim that all or virtually all reports of socially-sanctioned cannibal practices are myths originally invented by ethnocentric or interested parties and later credulously repeated, with little investigation and insufficient evidence, by sloppy, ethnocentric social scientists with a vested interest in sensationalizing the subjects of their study.

Though rejected by virtually all reviewers, Arens’ claim left lingering doubts among anthropologists who have not studied the topic. In light of the controversy, it might have seemed wise to drop ritual cannibalism as a case study in socially-sanctioned group violence. Instead, I decided to retain the study for several reasons.

First, as an instance of socially-sanctioned group violence, the alleged practice of cannibalism was too central to ignore. As noted in Chapter 3, many socially-sanctioned forms of violence have appeared in one culture or another over the course of human history, but only five forms of institutionalized, socially-sanctioned group violence have been widely practiced in all parts of the world for millennia: war; punishment (often corporal punishment) for breaking the law, or violating law-like customs; slavery; human sacrifice; and, allegedly, cannibalism. With cannibalism included in the set, a strong case can be made that all forms of socially-sanctioned group violence are culturally determined and susceptible to change, or abolition, as a function of changes in culture and in culturally-determined moral ideas. If customary cannibalism did not exist, however, it might be argued that widespread socially-sanctioned group violence comprises two forms (slavery and human sacrifice) which were practiced for a time and then ended, and two other forms (war and punishment) which have been practiced throughout human history. In other words, if cannibalism did not exist as a widespread, socially-sanctioned group practice, the evidence for systematic, culturally-determined variation in all forms of socially-sanctioned violence and related moral beliefs would be considerably weaker.

Second, the controversy over the existence of cannibalism concerns the very issue which is at the heart of this study: Can behavior (like cannibalism) which is unthinkable in one period of history be institutionalized and even routine in another? And if so, can behavior (like war) which is institutionalized in a given culture become unthinkable as culture changes? While granting that ethnocentric projection could play a role in the attribution of cannibal practices to simple societies, this study underscores the potential for bias in the opposite direction. Throughout the twentieth century, anthropologists reacting against the ethnocentrism and naivete of the field’s nineteenth-century founding fathers have, quite reasonably, eschewed developmental concepts and language which suggest that simple, non-literate tribal societies were the unfortunate precursors to our own ‘advanced’ society. The effort not to be judgmental or derogatory about simple cultures has been admirable and effective; but it can create a counterproductive form of political correctness. If cannibalism did exist, then the ostensibly politically-correct critics of the literature on cannibalism, such as Arens, would be the parties guilty of ethnocentrism, for having mistakenly labeled some customs of simple societies as too terrible to be human; by the same token, the anthropologists who accepted the evidence of ritual cannibalism would be the genuinely non-judgmental scholars.

Thus, the dispute between the two views of ritual cannibalism could not be more pertinent to the argument advanced in this essay: if ritual cannibalism did exist, that tends to confirm my hypotheses, first, that people are capable of rationalizing any action in the framework of a justifying world view and set of moral beliefs; and, second, that forms of violence believed to be vital to society at some times in history can become completely irrelevant to the preservation or loss of important social values other times.

Finally, a careful review of the Arens-related claims and counter-claims, reported in the Appendix, along with the evidence concerning the incidence of customary cannibalism given in the next section of this chapter, persuaded me that ritual forms of cannibalism not merely existed, but were widespread, being practiced by at least 50 percent of simple societies in most parts of the world, and more in some areas.

The Organization of the Chapter

The chapter begins with a detailed discussion of the evidence concerning the global incidence and cultural correlates of ritual cannibalism. Because the prevalence and even the existence of ritual cannibalism are contentious issues even among professional anthropologists who have devoted time to this issue, I have gone to some length to identify and assess evidence concerning the global incidence of the practice.

Next, I discuss the purposes and forms of ritual cannibalism, and the morality and affect associated with it in the practicing societies. These sections highlight some remarkable similarities in the context, justification, and emotional coloring of ritual cannibalism and war.

Finally, I tum to the question of how practices of ritual cannibalism ended and the inferences we might draw from that historical precedent for the potential abolition of war.

5.2 The Global Incidence and Correlates of Cannibalism

Two anthropologists have attempted in very different ways to provide a global overview of the incidence, forms, and meaning of cannibalism: Evald Volhard in the 1939 study Kannibalismus (Cannibalism), and Peggy Reeves Sanday in the 1986 book Divine Hunger. Volhard, an anthropologist at the Institute for Cultural Studies at Gothe University in Frankfurt, exhaustively compiles descriptions of every reported cannibal custom, current or past, anywhere in the world, that the author could find in any source. In the first half of his study, which is organized geographically by continent, region, and tribe, Volhard gives a thumbnail sketch of the time period, the sources, the specific actions, and the purposes involved in each reported custom or practice. This text is supplemented by numbered lists of the tribes or cultures (or language groups) in each region with reported cannibal practices, and by detailed maps that use the same numbers to show the approximate location of each tribe or culture by continent and region. In the second half of the study, Volhard regroups the earlier material, giving a cross-cultural analysis of cannibal customs grouped by their main purpose or meaning. (The typologies of cannibalism given by Volhard and others are discussed in the next section.)

Drawing on primary and secondary material in English, French, and German, as well as some primary material originally published in other languages (particularly Italian, Spanish, Dutch, and Portuguese) and later translated into English, French, or German, Volhard uses roughly 800 sources to identify 914 tribes or cultures which have been reported to have practiced some form of cannibalism, in nearly all cases customary or ritual cannibalism. Of the cultures identified by Volhard as having cannibal practices, 93 percent (857 cultures, many if not most no longer extant) were located in Central or South America, Africa, or Oceania (Australia, New Zealand, the Pacific islands, and Indonesia), while a scant 7 percent (57 cultures) were located in North America and the entire Eurasian continent.

Peggy Reeves Sanday, a cultural anthropologist at the University of Pennsylvania, used the Human Relations Area Files and Murdock and White’s Standard Cross-Cultural Sample of societies133 to estimate the incidence of cannibalism by geographic region, and to assess the impact on its occurrence of several likely factors.134 Out of a total cross-cultural sample of 186 cultures, Sanday found enough information to assess the presence or absence of cannibalism in only 109 cultures. Of the latter, 37 (one-third) had evidence of cannibal practices.

The geographic distribution of the cases in Sanday’s sample study loosely resembles that in Volhard’s global study. Most of Sanday’s cases were located in Africa, Oceania, and Latin America, and only 5 percent were in located in Eurasia. (See Table 5-1 for the detailed data supporting this and the next several paragraphs.) For North America, however, Volhard’s and Sanday’s findings appear to diverge considerably. North American accounted for only 2.5 percent of Volhard’s cases (23 out of 914) but 30 percent of Sanday’s much smaller sample (11 out of 37). In Sanday’s sample, the rate of cannibalism in North America was about the same as that in Latin America, Oceania, and Africa—close to 50 percent of the cultures surveyed in each area (Table 5-1, col. 5). A comparable figure cannot be given for Volhard, for whom there is no basis for judging the relative intensity of cannibal practice by region. But Volhard’s low absolute figure for North America—23 cases, compared with 400 cases for Africa and 351 cases for Oceania—implies either that the universe of cultures in North America, compared with that in other regions, is, for unknown reasons, far smaller in Volhard than in Sanday; or, alternatively, that contrary to Sanday’s findings, the intensity of cannibalism in North America is comparatively much lower than that in other regions; or both.

Table 5-1. Quantitative results of Volhard’s and Sanday’s global surveys of reported cannibal practices

Another measure of the overall similarities and differences between Sanday’s and Volhard’s results is the ratio of Sanday’s cultures with cannibal practices to Volhard’s, by region and for the world as a whole (Table 5-1, col. 9). Globally, Sanday’s cases represent 4 percent of Volhard’s, and for Africa, Oceania, Central and South America, and Eurasia, the ratios lie in the plausible surrounding range of 2–6 percent. In other words, in relative counts by continent, the two sources diverge from their relative global totals by not more than 50 percent. In the case of North America, however, they differ by a factor of ten: in North America, instead of representing 4 percent of Volhard’s cases, Sanday’s cases represent 48 percent.

This discrepancy raises questions about possible sampling bias in Volhard (North American cultures under-represented) or Sanday (North American cultures over-represented). In addition, for both studies an explanation is needed to account for the very low incidence of reported cannibal practices in Eurasia compared with that in Africa, Oceania, and Central and South America. The remainder of this section addresses these two points.

In an effort to resolve the apparent discrepancy between Volhard and Sanday for North America and to clarify the reasons for the low incidence of reported cannibal practices in Eurasia, I began by considering Sanday’s statistical analysis of factors that might be expected to affect the incidence of cannibalism. Sanday assessed three such factors: (1) the degree of social or political complexity, defined as levels of superordinate authority above the local band or village; (2) the regular presence of hunger, famine, or protein deficiency; and (3) the presence of a lengthy (greater than six months) post-partum sexual taboo. The expectations in the three cases (for which Sanday gives no explanation) are that cannibalism will be more frequent in simpler societies, in societies where hunger or famine is common and where the post-partum sexual taboo is long.135

Sanday found strong positive correlations for all three factors (see Table 5-2). Of the three, social complexity was the dominant factor: In simple societies (no political structure or integration above the local village), the incidence of cannibal practices was about 45 percent; among modestly complex societies (one or two levels of authority above the local village), it was 30 percent; and among the most complex (three or more levels above the local) it dropped to 15 percent.

Table 5-2. Results of Sanday’s analysis of factors affecting the occurrence of cannibalism

Food stress (famine) and lengthy post-partum sexual taboos were also associated with a higher incidence of cannibalism, particularly among the more complex societies. Among simple societies, some food stress increased the rate of cannibalism slightly (from 44 to 49 percent); ample food decreased it noticeably (to 18 percent). In contrast, in simple societies, the presence of a lengthy post-partum sexual taboo had no significant impact (incidence increased from 44 to 47 percent). Among societies with jural levels above the local village, 23 percent had cannibal practices; in the smaller sample of such societies with food stress the rate was 30 percent, and in the even smaller sample with lengthy post-partum sexual taboo, it was 47 percent.

The frequency of cannibal practices in simple and complex cultures: Sanday’s statistical analysis of contributing factors did not shed any light on the differences between Volhard and Sanday regarding North America, but it did suggest a possible reason for the findings in both of an exceptionally low incidence of cannibal practices in Eurasia compared with that in other continents: the earlier development of large, complex societies on the Eurasian continent resulted in an incidence of simple societies (for the period and sources covered by Volhard and Sanday) much lower than that for other continents. Sanday does not identify the 109 cultures included in her survey, nor does she give cross tabulations of factors by region. It was possible, however, to check the complexity of cultures by region in Murdock and White’s original cross-cultural sample of 186 cultures—and to review the quality of the sample itself with respect to coverage of continents.

Table 5-3. The global distribution of cultures, and proportions of simple cultures, in the Volhard, Murdock and White, and Sanday samples

Table 5-3 shows four Murdock and White measures of cultures by continent:

Col. 2. The number of cultures described and coded in their original Ethnographic Atlas.

Col. 3. The Murdock and White estimate by continent of all ‘adequately described’ cultures in the world, some 200 of which were not yet in the Atlas.

Col. 4. The number of ‘independent clusters’ of cultures in the Atlas, which could be considered distinct from one another for the purpose of cross-cultural comparison.

Col. 5. Murdock’s estimate of the total number of independent clusters that would appear if all adequately described cultures had been included.

Out of Murdock and White’s estimate of 433 independent cultures (col. 5), 129 or one-third are in Eurasia; and of the 186 cultures in the Standard Cross-cultural Sample, 62 or one-third are in Eurasia. Of these, 30 (one-half) are stateless or have only one jural level above the local. In contrast, for the world as a whole, the share of cultures with no or one jural levels above the local is 69 percent: it is 61 percent in sub-Saharan Africa, 77 percent in Oceania, 88 percent in Central and South America, and 91 percent in North America.

Thus, the distribution of simple and complex cultures by continent appears explain part of the exceptionally low rate of cannibalism in Eurasia in both Volhard and Sanday—but it does not account for all or even most of the difference between that region and others. For example, in Sub-Saharan Africa, where simpler cultures represent just over 60 percent of Sanday’s sample, nearly half of the sample have cannibal practices, but in Eurasia, where nearly 50 percent of the cultures are stateless or have minimal states, only 6 percent of the sampled cultures had such practices. (See Table 5-1.)

Although it was not possible to check directly for a relevant bias in the 109 cultures in Sanday’s subset of the cross-cultural sample which had adequate evidence to judge the presence or absence of cannibal practices, the distribution of cultures by region and the overall incidence of cultures that were stateless or had minimal political organization (68 percent) were both close to those in the larger, 186-society sample (Table 5-3, compare columns 7 and 8 with columns 10 and 12).

In a final effort to clarify the source of the low incidence for Eurasia in both Sanday and Volhard, I checked the specific cultures included in the Murdock and White sample. I was stunned to find that except for some very small, outlying groups (in Europe, the Lapps, the Irish Celts, and the Basques), the Murdock and White 186-society sample includes for Europe only one society (Rome); and for the entire span from Western Russia to the East coast of China one modem Russian village and one modern Chinese village. Most of the 62 cultures in the “Eurasia” regional sample are found in North Africa, the Middle East, and South Asia. The stress on these areas is undoubtedly appropriate for some sampling purposes; and the paucity of historical and pre-historical cultural samplings from Europe, Russia, and China, alone, does not account for the lower than usual of rate cannibal practices among the simpler cultures of this large region. Perhaps the best explanation for the low incidence of ritual cannibalism in Eurasia is that the customs of the simpler societies in this vast region may have been influenced by the proximity and frequent intrusions of more complex societies, which must have been far more extensive and frequent than was the case for simple societies in the Americas, Oceania, and Sub-Saharan Africa.

In the case of North America, the full Murdock and White cross-cultural sample shows this region to have the highest proportion of simple cultures of any continent. This only heightens the perplexing contrast between Volhard’s finding of few North American cultures with cannibal practices and Sanday’s finding that 50 percent of the cultures sampled had such practices. Moreover, the low overall numbers of distinct cannibal ‘cultures’ or ‘tribes’ identified in the Volhard study cannot be reconciled with the numbers of cultures by continent listed in Murdock’s World Ethnographic Atlas (see Table 5-3, cols. 1 and 3). Clearly, there were discrepancies in what Volhard, on the one hand, and Murdock and White, on the other, counted as a tribe or culture—and such definitional issues could affect even broad, general inferences about the global, historical incidence of cannibalism.

Reconciliation of the Volhard and Sanday findings for North America: In order to identify the sources of the sampling differences between Volhard and Sanday, I reviewed a number of checklists of tribes, cultures, and languages prepared by anthropologists who specialize in cultural geography. Initially, I had hoped to place the tribes listed in Volhard in the context of a universal list of all tribes, so as to derive estimates of cannibal practices by region to compare with those of Sanday. I found, however, that no existing checklist was sufficiently comprehensive in scope, time period, and detail to permit complete identification of the tribes listed by Volhard. Moreover, various experts disagreed on the degree of difference in language or custom that merited a distinction between one tribe and another.

Given these formidable obstacles to reconciling Volhard and Sanday globally, I decided to look more closely at the North American subset of the Volhard and Sanday studies. The purpose of this excursion was to determine the source of the statistical gap for this region; to illustrate the methodological problems that are bound to arise in any comprehensive cross-cultural study of a widespread, long-lived cultural practice; and to obtain a better sense of whether cannibalism in a region occupied mainly by simple societies may have occurred in as many as half of all cultures, as suggested by Sanday’s results, or was rare and exotic, as suggested by Volhard’s findings for North America.

My foray into North American Indian ethnography and linguistics, whose results are presented in Table 5-4, met these goals satisfactorily. The brief answer to the Volhard-Sanday discrepancy is that Volhard counted as single tribes all of the tribes that spoke either of two languages, Iroquois and Dakota, and either of two groups of languages, Athapaskan and Algonquian; and each of these four clusters represents hundreds of tribes, as tribes counted in the Human Relations Area Files. If we count all the tribes which use a language or language group used by any tribe with a cannibal practice—shown in Table 5-4 in parts 1, 2, and 3—then well over half of all distinctly-named and -located, independent tribes in North America may have had cannibal practices. More specifically, in terms of Sanday’s sample, of the 31 North American tribes in Murdock and White’s cross cultural sample (an unspecified 23 of which were coded for the presence or absence of cannibalism by Sanday), 8 match specific cases identified by Volhard (Table 5-4, part 1); another 4 concern tribes located in the same region and with the same language as a tribe cited by Volhard (Table 5-4, part 2); and a further 4 involve tribes with the same language as a tribe with cannibalism and located in a region for which Volhard has other reports of cannibalism (Table 5-4, part 3). These 16 cases make up over half of Murdock and White’s original 31-tribe sample for North America. (For good measure, another six tribes in the sample belong to tribes with the same language family as tribes with cannibal reports in Volhard, and appear in regions for which there are cannibal reports for other language families.) In other words, the comparatively small number of North American tribes identified by Volhard as having reported cannibal practices involves several very large aggregates of tribes; and when the counting system is rebased to the tribal level which Volhard uses elsewhere (and which Murdock and White use everywhere), the result is an incidence of cannibalism among simple cultures similar to that on other continents: about 50 percent.

Table 5-4. The North American tribes coded for cannibalism by Volhard and Sanday compared with all North American tribes, grouped by language and region.

An independent review of the situation in North America is provided in Kannibalismus bei den nordamerikanischen Indianern und Eskimo (Cannibalism among the North American Indians and Eskimos), a review of evidence of cannibal practices among native North Americans conducted by German anthropologist Herman Schöppl von Sonnwalden (1992), specifically in response to the Arens’ allegation that there are no first-hand witness reports of cannibal customs anywhere. Schöppl von Sonnwalden reviews the claims of cannibalism by language group, citing original sources to distinguish between some areas where there are credible eye-witness reports, and others where the evidence appears to be nothing more than sensationalistic hear-say. He sums up his conclusions in a passage worth citing in full (p 91, Forsberg translation):

Cannibalism among North American Indians and Eskimos did not assume the extensive form which it took in, for example, Melanesia, Central and West Africa, or Central and South America; generally speaking, it was not practiced. It did, however, occur in a region that stretched from Labrador across the eastern woodlands, through Florida and to the Gulf of Mexico. [This huge region is represented in Volhard by just two tribes, which are actually language groups encompassing a large fraction of North America: Algonquian and Iroquois. RF] Then, with an interruption for the non-cannibal areas, across Northeast New Mexico into Central America and the Caribbean. A second small area included parts of the Northwest Pacific Coast. Among the Eskimos, there was a tradition of eating dead relatives and tribe members in cases of extreme hunger. The eating of dead tribe members was nowhere practiced with the exception of the Pacific Northwest Coast. (In any case, as practiced by the Kwakiutl, it did not involve ‘patrophagie’—killing one’s seniors to eat them—but rather the use of mummified corpses for ritual purposes.)

Schöppl von Sonnwalden supplements Volhard’s sources with another 125 sources available (in the original or in translation) in English or German. After reviewing these, Schöppl von Sonnwalden comes to conclusions that match extremely closely both Volhard’s original account of incidence by language group and Sanday’s account of incidence by subgroup or tribe (as identified by me from within the larger White and Murdock sample).

While this does not fully account for the Volhard-Sanday differences in counts for North America compared with those for other regions, it increases confidence in the main conclusions suggested by both scholars:

First, ritual or customary cannibalism has been far more frequent in simple societies than in societies with complex, multilevel political organization.

Second, taking into account a sampling bias (simple cultures in the regions of Europe, Russia, and China lie so far back in prehistory that they are missing from the sample), we can assume that the high incidence of cannibal practices in the Third World and the low incidence in Eurasia are products of the length of time that certain societies in each region have had complex political systems with widespread cultural influence.

Third, given the extraordinarily good convergence between Volhard and Sanday (after North America is reconciled)—two fully independent studies—it is reasonable to conclude that the incidence of ritual or customary cannibalism among simple societies is likely to have been as high as 50 percent or more.

5.3 The Purposes and Forms of Cannibalism

Several authors who survey cannibalism in various cultures offer schematic representations of its purposes and forms. It is useful to start with that provided by Volhard, who divided cannibal practices into four main forms:

• Profane cannibalism (treating human flesh as food, whether by preference to other food or as a means of preventing starvation)
• Legal or punishment cannibalism
• Magical cannibalism
• Ritual cannibalism, including:
  • Worship of gods
  • Mortuary (death and burial) rituals Ancestor preservation
  • Celebration of victory in war
  • Puberty rites
  • Fertility rites136

In Cannibals! (1994), a book for the general public, Hans Askenasy makes a useful distinction among various forms of what Volhard calls ‘profane’ cannibalism and what others call ‘food’ or ‘gastronomic’ cannibalism. Askenasy divides the eating of flesh as food into five sub-categories, each with a slightly different connotation from the others. First, he treats starvation cannibalism (cannibalism by necessity) as a special case of food cannibalism. Then he distinguishes among cases of starvation cannibalism according to the cause of the starvation:

• Natural famines
• Man-made famines, such as the siege of a city in wartime
• Travel accidents that leave people stranded in remote locations, such as at sea, while traversing mountains, or in plane accidents

He then adds to these forms of starvation cannibalism two forms of ‘food cannibalism’:

• Cannibalism among sadists and the mad
• Cannibalism by tribes which chose human flesh by preference over other available food.

My review of cases of cannibalism among sadists and the mad suggests that the causation here is probably much closer to that in cases of ritual or customary cannibalism than to that in food cannibalism: generally, the cannibal actions sadists and the mentally ill are driven by a construction of feelings and motives in which the ingestion of part of another human does not serve as food, but has an important symbolic role. There also appears to be a special related sub-category for a few butchers who seem otherwise completely sane but who secretly killed people on a regular basis during times of famine or near famine in order to process their flesh into meat for sale, from which they made a profit. This kind of behavior is sociopathic: deadly violence which seems to involve little if any insanity except in the utter lack of conscience and guilt. (In some cases of human butchery, however, the butchers killed themselves before they could be brought to trial, indicating their expectation and fear of unbearable shame.)

Eli Sagan, who does not cite Volhard or the large body of German and French literature on cannibalism, offers an independent yet largely overlapping scheme of categorization.137 The examples of cannibalism on which Sagan draws, all taken from English-language publications, include anthropological ethnographies, popular books, and the accounts of travelers. Since many of Sagan’s sources are the same as those of Volhard, it is not surprising that Sagan’s list of ritual forms of cannibalism overlaps to some extent with that of Volhard. Both discuss funeral customs, rites to aid fertility, and rites to mark great occasions. Sagan adds to Volhard’s list several other special purposes which are probably more typical of rituals of human sacrifice with a cannibal component than of cannibal rituals more narrowly defined: the passing on the powers of the king, and rites to expiate sin, to enhance the general welfare, or to insure success in warfare.

A war-related category not listed prominently by either Volhard or Sagan but extremely common in ethnographic descriptions of exo-cannibalism is the consumption of the blood, or heart, and/or in some cases others parts of enemies slain in battle and of war captives tortured and killed after the return to the victor’s village.138 The foremost objective seems to have been to deprive the opponent’s tribe of the soul stuff or courage or fighting strength of this warrior, and to transfer one or more of those qualities to one’s own tribe. The consumption of enemies was also intended to convey disrespect of the enemy tribe and demoralize them.

Not surprisingly, the main objective of endo-cannibalism, particularly in tribes which practiced both, was often a mirror image of the purpose of exo-cannibalism: the goal was to retain the soul stuff, courage, strength, and fertility in the tribe, for the good of the group. In other cases, such as the South American Waari (which practiced both endo- and exo-cannibalism), described by Conklin (1995), the purpose of endo-cannibal customs was to insure the successful passage of the dead into the afterworld, to make sure that the soul did not linger in some kind of limbo, and in some cases, to make sure that the soul would proceed to a place or state from which a cycle of reincarnation would later be completed, with the individual’s returning either in another human body or else in a prescribed animal intermediary form. In the case of the Waari, being buried, rather than consumed by relatives, was expected to result in the soul getting stuck in the mucky ground forever and not being able to participate in an unending cycle of reincarnation. For this reason, in the 1950s, after Europeans insisted that the Waari abandon cannibalism and bury their dead, the last person to be consumed after death (secretly) was an old woman who on her death-bed begged her relatives to do her a favor and violate the ban so that she would not be lost forever.

The two main purposes and forms of customary cannibalism in China—which was a complex rather than simple society by the time of the first reports—are substantially different from those most common elsewhere, though not entirely unrelated. First, there are many reported cases in which emperors and warlords used cannibalism as form of punishment for treason, disloyalty, or opposition. In some cases the victims were tortured, killed, and then eaten as a form of disrespect worse than death. In other cases, punishment involved the victims being forced to consume parts of their own bodies or of the bodies of relatives (children or parents).

Second, the offering of flesh to an ill parent or a superior was recognized as an expression of filial piety or loyalty. In the most frequent example of this form of cannibalism, children cut a piece of flesh out of their thigh or arm, and turned it into broth in an effort to save the life of a parent suffering from a life-threatening illness. There is also substantial evidence of medicinal forms of cannibalism—that is, the use of powdered, dried, or otherwise reduced forms of human body parts in medical prescriptions—in 18th and 19th century Europe and North America.139 In the conclusion of the only major study on cannibalism in China, Chong (1990) provides a useful statistical summary of the specific, historically documented cases of cannibalism which the book describes. It is quite remarkable (the more so because apparently not noticed by Chong) that out of 653 documented cases of cannibalism for reasons of filial piety, only six involve the cutting out of flesh by males; all of the others involve younger, subordinate females making the precious donation (see Table 5-5).

Table 5-5. Chong’s analysis of reported incidents of cannibalism in China by purpose and form

5.4 Morality and Affect in Customary Cannibalism

The reason that cannibalism, human sacrifice, and slavery represent such appropriate and important parallels to war is that in these cases, just as in the case of war, a destructive practice is sanctioned and enabled by a powerful sense of moral righteousness, which outweighs less powerful senses of moral wrong-doing, repugnance, horror, shame, and sadness.

Given the complex mixes of moral belief and affect associated with each practice, and the differences in those mixes among different individuals, we should expect the process of identifying and characterizing the feelings involved to be difficult. This is all the more true because inherent in the social sanction is having been taught that one will be respected and applauded for doing a given action, and condemned and despised for failing to do it. The feelings of the practitioner about the expected social punishments or rewards thus become intertwined with feelings about the action itself in a manner that is hard to extricate. The sense of doing something wrong and of being hurt by doing it, or feelings of sadness for the victim or remorse for the act—feelings which I hypothesize exist in most people—are likely to be concealed and repressed beneath an emotional barrier created by the social acceptability of and requirement for the action, and the moral reasoning by which it is justified. This means that we are likely to find evidence of a sense of wrong-doing and hurt in indirect forms of expression more than in explicit, verbal statements.

In the case of ritual and customary cannibalism, there are a variety of ways in which repugnance and revulsion are associated with the actions in different cultures. Among the Kwakiutl of British Columbia, a designated cannibal goes through a series of ritual processes which result in small bits of flesh being consumed whole and then vomited up and rigorously accounted for. In this case and many others, the designated consumer becomes contaminated and must undergo cleansing rituals which last for weeks, months, or even a year. During the period of purification various forms of separation are required: for example, the cannibal may be forbidden to speak to or live with his family, and sexual activities are forbidden. In some societies, where many people are expected to partake of small symbolic bits of flesh, it is recognized that some of the less hardy individuals will secretly spit theirs out. In other cases, rites of passage in puberty are required to harden the individual sufficiently to be able to consume flesh.

In cases of endo-cannibalism where large parts of the body are consumed by relatives, it is common that the deceased will not have the funeral ritual until all of the appropriate relatives have gathered, which can take one or two days. The degree of decomposition of the body then makes the consumption of flesh particularly difficult, even after it is cooked. The test of love for the deceased is how much of the flesh the relations manage to consume rather than cremate. In some cultures only blood relatives consume the dead, while in others only in-laws do so: in either case, the task is a duty (not a pleasure) in which the consumers are supported (though not aided) by the non-consumers.

The Chinese forms of cannibalism reflect parallel feelings about exo- and endo-cannibalism: Used as a punishment, the important qualities of cannibalism were that because it violated the dignity and integrity of a person, it was revolting and it showed disrespect. Used as an expression of filial piety, the important quality was that of self-sacrifice—such a great sacrifice that only the lowest person on the family totem pole was expected to make it.

5.5 The Demise of Customary Cannibalism

My hypotheses about the about the causes and variability of socially-sanctioned group violence do not pertain narrowly to any single case: that is, the implications for the abolition of war of the rise and demise of ritual or customary cannibalism, discussed in this chapter, do not differ substantially from the implications for war of the rise and demise of the practices discussed in Chapter 4. It may be useful, nonetheless, to pause here and consider the more general points briefly in light of the material on cannibalism.

Customary cannibalism adds an important case to the universe of socially-sanctioned forms of group violence. It is an important case in part because any form of cannibalism is so alien to the modem mind and in part because the practice seems to have been relatively common among simple societies—quite possibly more common than war has been in recent centuries among complex societies.

As a form of socially-sanctioned group violence, what features does cannibalism have in common with war which might help us understand the tenacious roots of war and the possibility of its abolition? The key features are that, like war, customary or ritual cannibalism was relatively widespread; was institutionalized as a socially-sanctioned routine in which various members of society played well-defined roles; involved forms of behavior which, in the abstract, were perceived as abhorrent and hateful; and was undertaken in spite of its abhorrent aspects because it was believed to be extremely important to the survival or thriving of society.

Finally, the practice and subsequent demise of cannibalism is relevant to the potential abolition of war because, as is likely to be the case for war, cannibalism involved justifying reasons which were embedded in a larger cultural framework and which became irrelevant or invalid when key features of the culture changed. As the main form of socially-sanctioned violence in simple societies, cannibalism addressed the most serious issues in the world view of hunter-gatherers: death and rebirth, immortality, and soul stuff as a finite resource whose retention, revitalization, possession, or loss was all-important to the survival of the group.

When economic structure, scale, and political organization changed from hunter-gatherer, roving and small, and simple to agricultural, settled and large, and complex, the central preoccupations of the group changed from soul stuff, reincarnation, and the roles of the hunter and the hunted, to crops and herds; and the means to survival changed from holding together body and soul, maximizing soul stuff, and keeping up courage, fighting strength and fertility to having good weather and fertile soil. This externalization of the means of survival, combined with the growth of the social unit, led to the decline of cannibalism as a means of surviving and thriving, and the rise of pantheons of gods, who controlled weather and disease, and who, like humans, had to be fed to be happy. Thus cannibalism, like human sacrifice, was a justifiable evil which was centrally associated with survival in the societies in which it was practiced.