The aim of this paper is to review a sample of the evidence for hominin cannibalism within the prehistoric archaeological record. The review aims to ascertain whether testable motivations and social contexts for prehistoric cannibalism can presently be offered through current interpretative techniques. This paper will also attempt to identify any discernable patterns to the act of cannibalism within the prehistoric record, whether these patterns are consistent across time and space, and will discuss the interpretive behavioural implications for any patterns found. This will be achieved through a comparative examination of six possible prehistoric cannibal sites from different temporal and geographical zones and hominin species, examined through a particular categorising system of signatures in order to identify the specific type of cannibalism conducted. An examination of cannibalistic practices within the natural world and the genetic claims for prehistoric cannibalism are also discussed here in order to place hominin cannibalistic practices in the wider contexts of time and nature to allow for a more objective review of the hominin evidence.
Keywords: cannibalism, signatures, patterns, genetics, nature
Cannibalism amongst modern humans and their ancestors has always been a taboo topic which modern society finds to be both controversial and uncomfortable. By definition, a cannibal is “a person or animal that eats any type of tissue of another individual of its own kind” (Andrews and Fernandez-Jalvo 2003: 59). References to cannibal practices have been suggested all over the world in both prehistoric and historic periods. Despite the controversy cannibalism invokes, it has always been a topic of morbid interest and heated debate, with many ethnographic and anthropological studies reporting instances of cannibalism (Du Chaillu 1861; Schweinfurth 1873; Sanday 1986; Conklin 1995; Goldman 1999 and Vilaca 2000). Many authors have disputed the existence of cannibalism as a human behavioural norm, the most infamous being Arens’ 1979 publication “The Man-Eating Myth: Anthropology and Anthropophagy ” which claimed that there were no completely satisfactory ethnographic or historical evidence for cannibalism being a consistent part of the human social structure. As a result, this debate has been especially strong in the study of human remains from a wide variety of archaeological sites (Hillson 2000). Many of the early claims for cannibalism at such sites as Krapina ( Croatia c/f 130ka) and Monte Cicero ( Italy c/f 50 ka) have recently been re-examined and cannibalism subsequently rejected (Trinkaus 1985; Russell 1987; White and Toth 1991).
Yet, despite the many negative responses and oppositions to the idea of cannibalism in prehistory, the current view of prehistoric cannibalism, when faced with the overwhelming biological, anthropological and archaeological evidence, would be that cannibalism was a common practice amongst our ancestors. Having escaped the antiquarian view that cannibalism was a barbarous act of primitive societies, modern interpretations only extend to identifying the type of cannibalism practiced, without venturing possible hypotheses as to why the act of cannibalism occurred. The fact that cutmarks on hominin bones have been observed on different continents and come from a range of different periods prompted an observation by Stringer that, “the fossil record is only a tiny sample of people who lived in the past. If we are picking up butchery in this very sparse sample of humans and human behaviour in the past, then it cannot have been a very rare event” (Korn et al 2001: 34).
Thus, in order to try to understand the reasons behind the act of cannibalism it is important to grasp the abstract and physical implications of the word. ‘Cannibalism’ describes a huge range of behaviours and motivations across a number of species. For example, in humans, cannibalism has been related to any number of combinations of the following (after Villa et al 1986; Fernandez-Jalvo et al 1999; Taylor 2002):
- Nutritional necessity (survival cannibalism)
- Psychological imbalance (psychotic or criminal cannibalism)
- Aggression (hunting enemies and eating them)
- Affection (consuming friends or relatives)
- Spiritual (eating the dead as a funeral practice)
- Gastronomic or dietary cannibalism
- Medicinal cannibalism (health concerns)
All of which can be applied to inter-group cannibalism (exocannibalism) and intra-group cannibalism (endocannibalism).Thus, when looking at the possible archaeological evidence for cannibalism, it is important to specify what type of cannibalism is being discussed (by using the system of signatures given below), and not just label the act with the general term of, ‘cannibalism’. It is also important to note at this stage that the evidence for cannibalism within the prehistoric record from different geographic locations and time periods is highly unlikely to represent isolated incidents of psychopathic or survival cannibalism (Taylor 2002).
At this point, it may be prudent to look at the actual likelihood of prehistoric cannibalism having occurred. In order to do this, I shall review the evidence for neurodegenerative diseases within antiquity, which I believe provides a completely independent check on the data from osteology and archaeology. In addition, this will be accompanied by a brief synopsis of cannibalism within the natural world
Kuru and Genetic evidence for Archaic Cannibalism
Recent studies into transmissible spongiform encephalopathies (TSEs) or prion diseases have revealed important results bearing on the behavioural practices of our hominin ancestors. According to the National Institute of Neurological Disorders and Stroke (NINDS), TSEs are a group of rare neurodegenerative brain diseases which incorporate kuru and Creutzfeldt-Jakob disease (CJD), fatal familial insomnia (FFI), and Gerstmann-Straussler-Scheinker disease (GSS) in humans; scrapie and bovine spongiform encephalopathy (BSE) are the most common forms found in animals (Mead et al 2003; Roach 2003; NINDS). Prion diseases are thought to be caused by the malformation of a type of protein called a prion ( proteinaceous infectious particle) which, in turn facilitates the further malformation of normal prions, encouraging accumulation and clustering on the surface of brain tissue, ultimately resulting in the formation of small cavities in the brain (Mead et al 2003; Roach 2003; NINDS).
TSEs occur in three ways: “sporadically, as hereditary diseases or through transmission from infected individuals” (NINDS). Humans that carry two identical copies (homozygosity) of the prion protein would appear to be more susceptible to prion diseases, whilst those that carry two unmatched copies of the gene (heterozygosity, M129K or E219K polymorphisms - as exemplified by Mead et al 2003) appear to be protected against the prion diseases as heterozygosity “(inhibits) homologous protein-protein contact” (Mead et al 2003: 641). In order to test the distribution and occurrence of heterozygosity in the world’s populations, Mead et al (2003) “sequenced and genotyped” over 2000 chromosomes selected to represent worldwide genetic assortment. The results showed that one of the two polymorphisms (indicating heterozygosity) were present within all the populations investigated (Mead et al 2003: 641).
The global patterns of polymorphisms would certainly suggest that prehistoric TSEs were certainly a part of prehistoric hominin life. The cause of these TSEs is harder to determine, one possible explanation could be the spread of a prehistoric animal prion disease that managed to transverse the “transmission barrier to carnivorous humans” (Mead et al 2003: 643). However, given the increasing evidence for cannibalism in prehistory, Mead et al (2003) proposed that a more likely scenario explaining the polymorphic changes would be as a direct response to the reprocessing of prions within human populations through this act. Repeated exposure to the effects of TSEs, as a result of cannibalistic activities, may have necessitated the polymorphisms as a “natural selective advantage” of ancient populations (Mead et al 2003: 643). This research into TSEs and their causes has provided strong genetic evidence for a natural selective advantage occurring as a direct reaction to the common cannibalistic behavioural practices of our hominin ancestors.
Cannibalism within Nature
Cannibalism is found in many spheres of the natural world where genetically speaking, the promotion of one’s own genes at the expense of others, has tremendous survival benefits which offset the risks of cannibalising your own kith and kin (Taylor 2002). Cannibalism has been observed in over 1, 300 species of animal in the wild, ranging from protozoa (Hyman 1940), snails (Manzi 1970), centipedes (Eason 1964), fish (Larkin 1956, Poulson 1963) birds and 75 species of mammal (Southwick 1955, Yom-Tov 1974, Taylor 2002). These numbers are increasing all the time as zoologists start to take more of a keen interest in observing and recording instances of cannibalism. Along this vein, there would appear to be two extremes in examples of cannibalism from the natural world: Base levels of cannibalism (such as those episodes seen in spawning fish), where an indefinitely large number of young are produced, so the consumption of some of these young would not affect the population dynamics of the spawning species. Diametrically opposed to this form of cannibalism is the type of cannibalism seen in higher mammals (such as those episodes seen in chimpanzees) where there is a high investment in the survival of the few numbers of offspring produced and a higher level of social co-operation would involve more complex issues of the social group dynamic. As the higher mammals, such as chimpanzees, are closest to our selves biologically and socially a closer examination of the cannibalism encountered in these natural populations follows below.
As a direct result of excellent work conducted by many primatologists (Bartlett, Fossey, Goodall, Hamai, and Hrdy to name but a few), a complex picture of the myriads of primate behaviour has emerged. The wild chimpanzee’s (Pan troglodytes) of the Gombe National Park, Tanzania are perhaps the most infamous for their very human like behaviour sets displaying complex social relations intermixed with episodes of bereavement, joy, valour, compassion, infanticide and cannibalism (Goodall 1977, 1986; Bartlett et al 1993; Taylor 2002). In 1970, the Gombe chimpanzee society divided itself into two groups, the Kasakela and the Kahama. It is within the Kasakela community that the notorious duo Passion and Pom conduct their habitual cannibalistic infanticide. Passion and Pom have been directly reported to have killed and eaten at least three infants and are inconclusively linked to the disappearance of at least seven other infants, all belonging to other mothers within their community (Goodall 1986; Taylor 2002). Kasakela males and other chimpanzees from the M Group of the Mahale Mountains, Tanzania have also been reported to have committed acts of infanticide and endocannibalism (Goodall 1986; Hamai et al 1992).
The picture of chimpanzee cannibalism painted so far could reflect a form of sexual selection related to infanticide (Hrdy 1974, 1977) where females who lose their offspring will return to fertility more rapidly than those who maintain their offspring. It has been observed that infant killings seem to be more a result of male-male aggressive episodes rather than the focus of attack (Bartlett et al 1993). However, the almost exclusive focusing on infants by Passion and Pom could suggest a form of cannibalistic infanticide motivated towards sexual selection. Passion and Pom seem to have planned their attacks with an almost military like efficiency. The pair (either individually or together) would attack the mother and infant, chase the mother away, seize the infant and bite “into his forehead and face” (Goodall 1977: 269). The form and occurrence of this infanticide execution seems particularly brutal to be linked to a form of sexual selection. Head biting infanticide has also been observed among langurs (Hrdy 1999), which may imply that this form of infanticide is more common than its brutal nature would suggest. Cannibalism has also been reported amongst other nonhuman primate groups such as the red-tailed monkey (Cercopithecus ascanius), Japanese monkey (Macaca fuscata), chamba baboon (Papio ursinus) and Mountain Gorilla (Gorilla gorilla beringei) (Itani 1982; Fossey 1984).
Motivational issues behind primate cannibalism are however, not a major proponent of this discussion. We cannot begin to presume that the reasons for a chimpanzee cannibalism attack would correspond to those of our hominin ancestors nor indeed our anatomically modern selves. What is important here is more the fact that the instances of cannibalism occur throughout nature from the most base level to the most complicated (both biologically and socially). Therefore, the fact that cannibalism is evident within our evolutionary past should not be surprising nor disturbing as it would appear to be a natural, and in some cases an integral part of the world in which we live.
Within this paper, in order to examine the evidence for ‘cannibalism’ in the prehistoric record, an overview of six sites (from differing time periods and hominin species) will be compared with each other with the aim of establishing: 1) clear evidence for the act of cannibalism, and the specific type of cannibalism; and 2) whether there are any patterns within these data sets to identify which individuals were being cannibalised and why.
Archaeologists are expected to infer behaviour from mute remains. Any claim that cannibalism took place has to be based on tell tale signatures. The list of signatures below is generated using a combination of signatures compiled from a number of previous studies on prehistoric cannibalism (Villa et al 1986; White 1992; Turner 1993; Fernandez-Jalvo et al 1999; Hillson 2000; Andrews and Fernandez-Jalvo 2003; Taylor 2002).
- Lack of a cranial base (to get to the brain) on otherwise complete or near complete skeletons.
- The virtual absence of vertebrae (due to crushing or boiling to get at marrow and grease).
- Cut and chopmarks: result of striking the bone surface with a sharp stone tool leaving a deep, wide V-shaped scar (related to cutting strong muscle attachments or dismembering).
- Cutmark arrangement: position, number and placement on element (whether on muscle/ligament attachments etc).
- Long bone breakage (to get at the bone marrow).
- Anvil abrasions left where a bone has rested on a stone anvil whilst it is broken with a hammer stone.
- Comparable butchering techniques in human and animal remains: incidence, position and type of cut and chopmarks on human and animal/ food bones should be directly comparable (although allowance should be made for anatomical differences between humans and animals).
- Post processing discard: have the bones been treated in the same way as the faunal remains? Indicating that hominins and animals were treated the same way in terms of food preparation and discard.
- Evidence of cooking in the form of burnt bone. If present this evidence would seem to suggest the comparable treatment of human and animal remains
- Peeling: a roughened bone surface with parallel grooves or fibrous texture is produced when fresh bone is fractured and peeled apart.
- Percussion pits: the point of impact where a stone or any solid matter struck the bone cortex and scarred the surface.
- Scraping marks: the result of periosteal and muscle removal by scraping the bone surface and identified by a concentrated series of parallel and superficial striations on a broad area of bone.
It is evident that it is insufficient to assign the general term ‘cannibalism’ to a single ‘signature’ left on a hominin skeletal assemblage. Rather, by combining the signatures it is possible to identify, in certain circumstances, the specific type of cannibalism practiced. The sample sites studied below have been examined in accordance with the amalgamated list of signatures given above and through the original authors’ criteria to create Table 1. The amalgamated list of signatures given above not only expands the criteria of evidence for the act of cannibalism into a single comprehensive list, but the use of the amalgamated list allows for a more refined interpretation concerning the specific type of cannibalism practised from the physical remains of the act. By applying the amalgamated list of signatures to the sample sites studied here, it allows for a testing of the original authors’ methodology and vice versa, a test for the methodology laid out in this paper. For example, the differentiation between ritual and nutritional cannibalism would, primarily be based on the comparison of hominin and faunal remains from within the same archaeological context. If hominin and faunal remains are found within synonymous contexts with different patterns of exploitation and distribution, then ritual or other interpretations of cannibalism should be considered for the hominin remains (Villa et al 1986; White 1992; Fernandez-Jalvo et al 1999). From the signatures given above, it becomes evident that marrow extraction along with brain extraction and a discard pattern similar to those of the faunal remains would all be indications for gastronomic cannibalism. Ritual cannibalism could cautiously be inferred if the bones display evidence of defleshing through cut and scrape marks without marrow extraction taking place, the only feasible reason for the extraction of bone marrow would be for nutritional consumption. However, the ability to distinguish between nutritional and ritual cannibalism cannot solely be made on the comparison between animal and hominin butchery practices, because they are going to be the same or similar (allowing for anatomical differences) regardless of intent. A similarity in butchery practices only establishes the intent to deliberately dismember a hominin body, the post-depositional discard patterns must be closely considered in this light when inferences of cannibalism are to be made within the sites’ interpretation.
This theory also holds true for the treatment of different parts of the skeleton. For example, if the skull is treated differently to the postcranial skeleton as at (Gough’s Cave Andrews and Fernandez-Jalvo 2003), by being kept intact, whilst the rest of the skeleton showed signs of processing in a way that indicated consumption. This could imply that the skull may have held a ritualistic or deeper meaning for the hominins involved, while the remainder of the body was processed for food. Thus, it becomes clear that when dealing with the act of cannibalism, it is necessary to look at all aspects of the skeleton. This kind of analysis also has the potential to provide insights into the way some hominins may have viewed different parts of the body (to be discussed further below).
At this juncture, the differences between the evidences for cannibalism and excarnation (defleshing) may seem unclear. In order to clarify this I turn to a synopsis of a study conducted on the Neanderthal remains from the Hušnjakovo rock shelter, Krapina, Croatia. Krapina has yielded an astonishing number of Neanderthal remains, forty-three adults, adolescents, and several juveniles less than fourteen years of age (Trinkaus 1985 cited in Russell 1987). There has been much controversy over whether Krapina was a cannibalistic site or a site of secondary/excarnated burial (secondary burial being the act of burying the cleaned/defleshed remains of the hominins). Russell (1987) conducted a comparative study of the site, comparing the cutmarks on the Neanderthal remains to a Mousterian reindeer butchery site, Combe Grenal (Dordogne, France), and the secondary burial/excarnation site of Juntunen (Michigan, U.S.A). The comparative basis of the study was as follows. If a bone were cleaned for secondary burial (as at Juntunen), there would be significantly more cutmarks on the bone as the flesh is cut right off the bone in an attempt to remove all the soft tissue. Conversely, in butchery (as at Combe Grenal) the butcher would want to extract the meat and soft tissue without striking the bone too often and subsequently dulling the edge of the cutting implement. The results showed that the Krapina bone assemblage had more affiliation with the Juntunen bone assemblage in frequency, location and orientation of the cutmarks present (Russell 1987). The post-processing discard patterns support this. An unusually large number of relatively fragile skeletal elements such as scapulae and juvenile claviculae were preserved at Krapina, suggesting that these bones were buried rather than left on the cave floor surface (Trinkaus 1985). Recent studies have claimed that Krapina was in fact a cannibalism site (White 2001), however, Russell’s study has methodological credibility and sound argument. The controversial site of Krapina would thus seem to exemplify the cannibalism debate perfectly.
Russell’s study clearly highlights the differences between cannibalistic cutmark remains and those that have undergone excarnation (in terms of frequency, angle and depth of cutmark). Further evidence that may help in telling the difference between excarnated and cannibalised hominin remains may be seen within the post-processing discard patterns. Excarnated remains are unlikely to be discarded randomly within the faunal remains from the hominin occupation site. If a hominin has taken the trouble to deflesh their contemporary, it is assumed that this would be for the express purpose of treating those hominin remains in a manner different to those of the faunal remains. By contrast, cannibalised remains can expect no such differential treatment (unless there is evidence for ritual cannibalism) and would therefore be discarded within the faunal remains. The next question to be answered would be how to tell the difference between ritualised and excarnated hominin remains? The answer lies within Russell’s study where the position and type of cutmark change radically between the two types of mortuary practice. For a cannibalised skeleton, the cutmarks are likely to be placed/concentrated more on the muscle/ligament attachments with minimal amounts of scrape/cutmarks found in between the two extremities of the bone. Whereas excarnation activities are more likely to have cut and scrape marks located along the entire length of the bone, with no apparent concentration zones for cutmarks and an increased number of scrape marks. Furthermore, scarcity of bone breakage for marrow extraction and manner of bone deposition sets excarnation apart from cannibalistic activities (Villa et al 1986).
Thus, by using the amalgamated list of signatures above, and with an awareness of the different post-processing patterns left from excarnation and cannibalised remains, it is possible to apply this methodology to archaeological sites and arrive at a feasible interpretation as to the validity of a cannibalism claim and to specifically identify the type of cannibalism that was practiced. Motivation for the act however remains illusive and shall be addressed in further detail below.
Overview of the Sites under Review
Table 1 below shows the evidence for cannibalism from six sites encompassing differences in hominin species, geographic location and period. Subsequent tables provide more specific data from individual sites.
The famous Bodo cranium was found in an outcrop of Middle Pleistocene sediments from the site of Bodo in Ethiopia’s Middle Awash Valley (Conroy et al 1978) along with Acheulean technocomplex (Kalb et al 1982, 1984) or with developed Oldowan stone tools (Clark et al 1984). The Bodo cranium represents one of the earliest examples for the intentional defleshing of a hominin (White 1986). White clearly illustrates this point in his examination of the cranium that “intentional post-mortem defleshing of some kind occurred” (White 1986, 508). However, White also points out that there is not enough evidence to be able to infer the act of cannibalism per se. The lack of the cranial base means that brain removal through the foramen magnum cannot be clearly established. This serves to illustrate the earlier point that unless a majority of the signatures can be satisfied, then an inference of cannibalism cannot be made. There may be many other hominin behaviours that could explain the cutmarks on the cranium, ranging from curation to decoration (White 1986). This well-known example is included to demonstrate the importance of fulfilling the majority of the criteria laid out above before cannibalism can inferred.
Atapuerca, Tranchera Dolina (TD) layer 6 Aurora Stratum
The site of Gran Dolina is located in the southern part of the Sierra de Atapuerca, near Burgos, Spain and has revealed the remains of six individuals belonging to Homo antecessor, two adults, two adolescents and two infants (see Tables 1 and 2) from the Aurora stratum (TD6) found mixed together with stone tools (pre-Acheulian Mode 1) and non-hominin faunal remains (Carbonell et al, 1995, Fernandez-Jalvo et al 1999). Both the hominin and faunal remains display evidence of hominin modification ( Fernandez-Jalvo et al 1999). The hominin and faunal remains were examined according to numbers 3, 10, 11 and 12 of the signatures given above.
The examination prompted observations that the butchery techniques for the two taxa were very similar seen in the analogous methods of periosteum removal, marrow extraction, filleting, dismembering and skinning (Fernandez-Jalvo et al 1999) (see Table 2). As with any direct anatomical comparison, allowances for anatomical differences should be made, for example, hominin faces have strong muscle attachments that make them likely to have more cutmarks and modifications than other fauna (Fernandez-Jalvo et al 1999). Fernandez-Jalvo et al (1999) further observed that similar breakage patterns (to extract marrow), percussion and conchoidal fractures were found on large, medium and small-sized animals and hominins (see Table 2). Identical patterns of post-processing discard of hominin and animal bones were observed where the remains of hominin and non-hominin animals were randomly dispersed with no distinctive depositional arrangement for any of the taxa (Fernandez-Jalvo et al 1999).
From the signatures given in the Methodology above, the butchering techniques observed in the Aurora Stratum would indicate meat and marrow extraction from both hominin and non-hominin animal remains. From this evidence, no ritual treatment of the hominin remains is apparent. Thus, nutrition would seem to be the purpose of the type of cannibalism practised at this site. Nutritional cannibalism can be divided by definition into “(a) survival, where cannibalism is incidental or a short-term measure, and (b) dietary or gastronomic cannibalism, which is associated with long periods during which hominins are feeding on other hominins as part of their regular diet” (Fernandez-Jalvo et al 1999: 620). To determine which type of nutritional cannibalism was practised, a number of other factors have to be examined. Pollen analysis seems to indicate that the environment around Gran Dolina at the time of occupation was temperate, and the large species diversity in the faunal record indicates a plentiful and ready food supply (Fernandez-Jalvo et al 1999). Although dietary stress can occur in warmer climatic environments, through drought, disease and animal migration, the contemporary mixing of the hominin remains with the animal remains indicates a ready and plentiful food supply within the Gran Dolina locality. It can be suggested that the hominins occupying this site did not suffer from starvation. Thus, in agreement with the methodology laid out for this paper, the type of cannibalism encountered here should be considered gastronomic cannibalism, showing that hominins were a regular part of the diet of other hominins (Fernandez-Jalvo et al 1999).
The site of Herto Bouri located in the Middle Awash area of Ethiopia’s Afar Depression, has yielded the remains of two adults and one infant from Homo sapiens idaltu, radioisotopically dated to between 160, 000 and 154, 000 years ago (Clark et al 2003; White et al 2003). The remains were found in correlation with archaeological assemblages containing elements of Acheulean and Middle Stone Age technocomplexes (Clark et al 2003; White et al 2003). The crania remains display clear evidence of post-mortem modification in the form of cut and scrape marks presumably to remove the skin, muscles, nerves and blood vessels (Lemonick and Dorfman 2003; White et al 2003). However, as with the Bodo cranium, the evidence is not strong enough to infer cannibalism, although the apparent special treatment of the infant skull, with its smooth polished areas not caused by depositional erosion (Lemonick and Dorfman 2003) may have some interesting implications concerning the recognition of the individual (discussed below).
The cave site of Moula Guercy is located in southeastern France on the west bank of the Rhone River, Ardèche. The Neanderthal remains extracted from level XV of the Moula-Guercy site show a minimum number of six individuals, two adults, two adolescents and two infants (see Table 1) found in conjunction with a Ferrassie Mousterian lithic assemblage and non-hominin faunal remains dated to between 100, 000 and 120, 000 years ago (Defleur et al 1999). From the remains, “only one identifiable Cervus specimen showed carnivore modification (whilst) none of the hominin remains (did)” (Defleur et al 1999: 129). Contrary to this, both hominin and deer remains displayed explicit evidence for deliberate post-mortem modification. These modifications were analysed and quantified according to numbers 3, 5, 6, 10 and 11 of the signatures given above. Interestingly, in some instances, the cut and percussion marks indicate successive strokes of the same implement, whilst evidence through similar post-discard polish on the hominin and non-hominin assemblages, may indicate that the cave continued to be occupied after the butchery incident/s occurred (Defleur et al 1999).
As has been emphasised before, one of the key criteria in assessing a possible incidence of prehistoric cannibalism is the direct comparison and demonstration that the faunal and hominin remains were exposed to similar post-processing discard patterns. At the site of Moula-Guercy, as at Gran Dolina, there is clear archaeological evidence to this effect. Table 3 provides data on the representation and modifications of the deer and Neanderthal bones. As can be seen from Table 3, there is a direct correlation between the cutmark and fracture pattern for the hominin remains and the faunal remains, except where biological and functional differences between the taxa influence the modifications made to the bones. This in turn seems to indicate that the hominin and deer carcasses may have been butchered in similar ways, where the apparent objective was the removal of soft tissues and marrow (Defleur et al 1999). According to the signature criteria laid out above (and in agreement with the original site authors), an inference of cannibalism for Moula-Guercy level XV would be reasonable, as no evidence was found that the human modification to the hominin and faunal remains represented any recognisable form of mortuary ritual (Defleur et al 1999). The authors (Defleur et al 1999), although concluding that cannibalism was practised at Moula-Guercy, did not specify what type of cannibalism was practiced. Therefore, based on the methodology specified for this paper, and due to the species diversity found within the same layer as the hominin remains (indicating a possible ready and plentiful food source), and the similar post-processing discard of animal and hominin remains, it is suggested that the type of cannibalism practiced at Moula-Guercy, was gastronomic.
The site of Gough’s Cave located in Cheddar Gorge, Somerset, England has yielded the remains of five individuals belonging to Homo sapiens, two adults, two adolescents and one infant found with abundant faunal remains and lithic artefacts (Creswellian technocomplex) from Oxygen Isotope Stage 2 deposits (Stringer 2000; Andrews and Fernandez-Yalvo 2003). The human induced bone modifications were recorded according to numbers 3, 6, and 12 of the signatures given above. As with the Aurora Stratum, and Moula-Guercy, the analysis of the Gough’s Cave assemblage showed that the butchering techniques observed on human and nonhuman skeletons were similar, apart from differences arising from anatomy (see Table 4).
The one major difference in the similar butchery and post discard pattern between the hominin and faunal remains is the difference in skull completeness. Although the human skulls and faces at Gough’s Cave have a higher intensity of cutmarks than the faunal remains, two of the human skulls are almost complete (Andrews and Fernandez-Yalvo 2003). At the other sites where nutritional cannibalism has been identified (Aurora Stratum TD6, Fernandez-Jalvo et al 1999; Moula Guercy, Defleur et al 1999) and from more modern cannibalism sites from the United States (for example 800 years ago at Mancos 5MTUMR-2346, White 1992; or the American southwest in general, Turner and Turner 1999), skull completeness has tended to vary greatly. The deliberate damage to the hominin skulls at all of these sites is extensive and many of the authors have interpreted this as a consequence of accessing the brain. The completeness of the two human skulls at Gough’s Cave (despite their intensive modifications) is an exception to the general patterning of the Gough’s Cave assemblage. The authors interpreted the Gough’s Cave assemblage as possible “ritual skull treatment’ following Vila et al 1986 interpretation of the skulls found at Fontbregoua (see below) (Andrews and Fernandez-Jalvo 2003).
When looking at the radiocarbon dates for human bones and humanly modified bone, antler and ivory (see Table 5) it is clear that there are two distinct periods to the humanly modified bones. The first period from around 12,940 14cyr BP until 12,300 14cyr BP shows almost exclusively animal remains being modified by humans, whereas from 12,300 14cyr BP to 11,480 14cyr BP there is a shift to almost exclusively human bones being modified. This could reflect two distinct phases in the occupation of Gough’s Cave, a hunting phase for the animal modifications, changing to a ritual phase for the human bone modifications (tying in with the ‘ritual’ suggested by the complete skulls). It must be noted however, that not too much emphasis should be placed on the radiocarbon dates due to their large time ranges in accuracy. The apparent shift in usage in Gough’s Cave (Table 5), similarities in butchery techniques between human and animal remains (Table 4) and the large representation of faunal remains evident within the Gough’s Cave assemblage suggests that the type of cannibalism practiced here was gastronomic or nutritional overall, with a possible ritual aspect involving the human skulls (Andrews and Fernandez-Yalvo 2003).
The site of Fontbregoua (southeastern France) has revealed the remains of three adults, two children and one individual of indeterminate age belonging to Homo sapiens sapiens dating from 6, 000 to 7, 000 years before present found in association with pottery, stone tools, domestic and wild animal remains (Villa et al 1986). The human induced modifications on both human and animal bones were examined according to numbers 3, 4, 5, 8 and 9 of the signatures given above. The results are given in Table 6.
The results show very similar methods of processing and discard for the human and faunal remains and evidence to suggest cooking is lacking in both assemblages. Evidence for the fracturing of long bones to get to the marrow for both human and faunal remains is however, very clear, which suggests nutritional cannibalism. The primary difference is that the human crania show a much larger amount of defleshing marks (Villa et al 1986), whilst still being maintained relatively intact (as at Gough’s Cave). This special treatment has been interpreted by Villa et al as a possible ritual treatment or trophy procurement as the special treatment of the skulls appears to be “consistent with an interpretation of exocannibalism” (Villa et al 1986: 436). The interpretation of exocannibalism may not necessarily be the only explanation for the maintenance of the skulls (as discussed below), and such an inference should only be made with caution.
However, based on the similar butchery patterns, long bone breakage to extract the marrow and post-processing discard patterns, cannibalism would appear to have been practised at Fontbregoua (Villa et al 1986). However the authors, although laying down certain criteria pertaining to dietary cannibalism did not explicitly state what type of cannibalism was practised at this site. Given the evidence as discussed above, I would conclude that the type of cannibalism practised at Fontebregoua, as at Gough’s Cave, was nutritional cannibalism with a possible ritual aspect involving the skulls.
What follows in this section is a discussion on a number of patterns within the data followed by a set of hypothesised interpretations of what those patterns may represent in a motivational and social context. The motivational and social contexts behind the act of cannibalism are often complex and integrated within abstract ideals or notions. Such notions as the varying philosophies of humanness and cannibalism depending on the predator/prey, relative/stranger relationship seen in the Amerindian tribe the Wari’(Vilaca 2000) or the head hunting life force ideologies, seen in many head hunting communities of Borneo (Izikowitz 1951, Needham 1976), all exemplify the fact that the motivations and social contexts surrounding cannibalism are never constant nor synonymous. Therefore, due to the nature of the archaeological data given above, it has become evident that a testable supposition for the motivation and social contexts behind the cannibalistic act cannot be reached with any degree of certainty or satisfaction.
From Table 1 it is clear that from four of the six sites shown, the age ranges of the individuals consumed illustrate a consistent pattern of indiscriminate choice in the ages of the individuals being cannibalised. Whether this also implies that any individual was cannibalised regardless of social standing (assuming that the older the individual, the more social status they may hold), is impossible to prove from the evidence. However, if the assumption just stated is accepted then the data could imply that every individual within a hominin group, regardless of age (certainly) and social status (possibly), could potentially be eaten. Table 1 also illustrates the fact that cannibalism within the sites studied was predominately conducted for nutritional reasons (stated by the respective authors, and supported by this author using the system of signatures and given above).
Table 1 also highlights that from the data set under examination, nutritional cannibalism would appear to be the dominant form with the possible element of ritual cannibalism becoming apparent with the advent of Homo sapiens (although it should be made clear here that the sample size is too small for this to be empirically tested or quantified). The sites of Gough’s Cave and Fontebregoua illustrate this where the skulls are almost intact despite evidence of extensive cut and percussion marks. This paradox of extensive damage yet completeness could imply (as has been suggested by the original authors) that the skulls held some special purpose or meaning for the groups involved (Andrews and Fernandez-Jalvo 2003; Villa et al 1986). Other sites not related to Homo sapiens do not display any such careful maintenance of any skeletal component; all are intermixed within the same contexts as other non-hominin animal remains. Indeed, within the Homo sapiens’ sites the skulls are the only skeletal elements that show such careful preservation. The rest of the Homo sapiens’ skeletal bones are all intermixed with the other animal bones. Naturally, the suggestion of possible ritualised elements to cannibalism, occurring with the advent of Homo sapiens does not necessarily apply to all Homo sapiens cannibalism sites. The oldest human remains from Klasies River ( South Africa) dated to 110, 000 years are fragmentary and mostly consisting of skull parts (Deacon 2001) (completely opposite to the Homo sapiens sites in Europe studied for this paper). The bone remains from Klasies have been submitted to clear post-mortem modification through cut and tear marks, percussion impacts and burning (Deacon 2001; by the signatures of cannibalism laid out above, burning combined with cutmarks etc can be taken as evidence for cannibalism). The remains only occur in specific horizons within the site and not from conventional burials, although it is impossible to state whether their occurrence in particular horizons indicates some ritual of secondary burial practice or interpersonal violence (Deacon 2001). Thus, as is shown by the evidence from the Klasies River site, a ritual aspect to Homo sapiens cannibalism sites must not be taken as a universal factor, only when there is sufficient evidence should a hypothesis of ritual behaviour be proposed.
This does not mean to imply that other hominin species did not practice or have conceptual awareness of a ritual act, as this is clearly not the case. For example, the Neanderthal secondary burials at Krapina (Russell 1987) could imply that the Neanderthals at least had some higher cognitive ability, sense of individuality and ritual. Furthermore, the Herto site (White et al 2003) may circumstantially provide further evidence for pre-behaviourally modern Homo sapiens’ conceptual awareness. An infant’s skull displays severe cutmarks, but portions of the skull are smoothed and polished. The polished areas were not caused by post-depositional processes because the polish patterning transverses the breaks between the recovered pieces, prompting Tim White to suggest that it looked “as though it (had) been fondled repeatedly” (Lemonick and Dorfman 2003: 78). Whether this repeated fondling was ‘ritual’ or, more personal, is impossible to say from the evidence so far known about the site. This repeated fondling of the skull may well imply that the hominins from Herto may have had a sense of personal individuality, and fondled the skull in order to remind themselves of the child lost. However, inferences concerning intentionality from assemblages like Herto remain intrinsically difficult to prove and any number of far more pragmatic interpretations could explain this curious phenomenon.
One possible explanation for the careful maintenance of the Herto juvenile skull, and the intact Homo sapiens skulls from Gough’s Cave and Fontebregoua, maybe found when viewing the skull as the source of what defines someone as a ‘human’. Taking into account the diverse views of what constitutes a human in the modern world, it is plausible that it is the face that could epitomize an individual’s humanity. The face expresses an individual’s emotive state. By looking into someone’s face we (as humans) gauge how someone will react to a situation, tell if they are lying, see if they are happy or sad, angry or thoughtful. By this reasoning, it is conceivable that the face could exemplify an individual’s humanness with the skull being a physical representation of the face perhaps not dissimilar to a memento.
From the data in Table 1, and other sites mentioned in this paper, it becomes evident that cannibalism was almost certainly practised throughout prehistory, and it is reasonable to hypothesise on what type of cannibalism was practised in the majority of the examples used for this data set. The author recognises that in some cases, due to lack of evidence or bad preservation of some sites, it may not always be possible to identify what type of cannibalism was practised in accordance to the ‘Methodology’ given above. In these instances, cannibalism should only be inferred with the utmost caution. It has also been shown that there are at least some recognisable patterns of cannibalism evident from the archaeological record.
It has been shown from archaeological and genetic evidence that nutritional or dietary cannibalism has been a part of Homo’s heritage and way of life from the earliest times (Villa et al 1986, White 1992, Defleur et al 1999, Fernandez-Jalvo et al 1999, White 2001, Andrews and Fernandez-Jalvo 2003). Although testable motivations and social contexts for prehistoric cannibalism cannot satisfactorily be offered, the combination of the signatures stated in the methodology creates a comprehensive list by which the evidence for prehistoric cannibalism can be scientifically analysed. The list was generated from criteria suggested by other researchers (see text for references). I am advocating that cannibalism only be recognised as such when the data matches all or the majority of these criteria, not just a small percentage. What I have tried to do is present a sample of the evidence illustrating this point clearly. I have also attempted to show that not only is cannibalism an important part of hominin behaviour, but a wholly natural act in the animal world, of which humans are still very much a part. It has also been shown that within archaeological research it is not sufficient to simply state that the act of ‘cannibalism’ was practised at a given site. Rather, it is important that the act of ‘cannibalism’ be further refined and categorised into the type of cannibalism practised using the signatures given in the methodology. Through the application of these signatures, this paper has shown that this type of specific classification is indeed possible and vital to the correct interpretation of an identified cannibalistic site. By identifying some clear patterns within the archaeological record, which transcend both time and species. I have established a testable proposition, namely that there does not appear to be any discrimination in terms of age, as to who was consumed. Whether a correlation between age and social standing can ever be established has yet to be seen. From the data under examination, no such inference can be definitively made at present. Similarly, the theory that ‘ritualistic cannibalism’ is only present in Homo sapiens (from Gough’s Cave and Fontebregoua) remains to be tested by further analysis.
I would like to thank Dr J. McNabb for the continued advice and invaluable guidance throughout the writing of this paper. I would also like to thank Professor C. Stringer and Dr R. Jacobi for their help in the beginning stages of my BA dissertation, from which this paper ensued, written at the University of Southampton .
Andrews, P. and Fernandez-Jalvo, Y. 2003. Cannibalism in Britain: Taphonomy of the Creswellian (Pleistocene) faunal and human remains from Gough’s Cave. Bulletin of the NaturalHistoryMuseum. Geology Series 58: 59-81
Arens, W. 1979. The Man-Eating Myth: Anthropology and Anthropophagy. New York: Oxford University Press.
Bartlett, T.Q., Sussman, R.W., Cheverud, J.M. 1993. Infant Killings in Primates: A Review of Observed Cases with Specific Reference to the Sexual Selection Hypothesis. American Anthropologist 95: 958-990.
Bowman, S.G.E., Ambers, J.C. and Leese, M.N. 1990. Re-evaluation of British museum radiocarbon dates issued between 1980 and 1984, Radiocarbon 32(1): 59-79.
Carbonell, E., Bermudez de Castro, J. M., Arsuaga, J. L., Diez, J. C., Rosas, A., Cuenca-Bescos, G., Sala, R., Mosquera, M. and Rodriguez, X. P. 1995. Lower Pleistocene hominins and artefacts from Atapuerca-TD6 ( Spain). Science 269: 826-830.
Clark, J.D., Asfaw, B., Assefa, G., Harris, J.W.K., Kurashina, H., Walter, R.C., White, T.D. and Williams, M.A.J. 1984. Palaeoanthropological discoveries in the Middle Awash Valley, Ethiopia. Nature 307: 423-428.
Clark, J.D., Beyene, Y., WoldeGabriel, G., Hart, W.K., Renne, P.R., Gilbert, H., Defleur, A., Suwa, G., Katoh, S., Ludwig, K.R., Boisserie, J-R., Asfaw, B. and White, T. 2003. Stratigraphic, chronological and behavioural contexts of Pleistocene Homo sapiens from Middle Awash, Ethiopia. Nature 423: 747-752.
Conklin, B.A. 1995. “Thus are our bodies, thus was our custom”: mortuary cannibalism in an Amazonian society. American Ethnologist 22 (1): 75 – 101.
Conroy, G.C., Jolly, C.J., Cramer, D. and Kalb, J.E. 1978. Newly discovered fossil hominin skull from the Afar Depression, Ethiopia. Nature 275:67-70.
Deacon, H.J. 2001. Guide to KlasiesRiver. Unpublished Typescript http://www.sun.ac.za/internet/academic/arts/archaeology/KRguide2001.PDF (Accessed 20.04.04)
Defleur, A., White, T., Valensi, P., Slimak, L., Cregut-Bonnoure, E. 1999. Neanderthal Cannibalism at Moula-Guercy, Ardèche, France. Science 286: 128-131.
Du Chaillu, P.B. 1861. Explorations & adventures in equatorial Africa: with accounts of the manners and customs of the people, and of the chace of gorilla, crocodile, leopard, elephant, hippopotamus, and other animals. London: Murray.
Eason, E.H. 1964. Centipedes of the British Isles. London: Warne.
Fernandez-Jalvo, Y., Diez, J.C., Caceres, I. and Rosell, J. 1999. Human cannibalism in the Early Pleistocene of Europe (Gran Dolina, Sierra de Atapuerca, Burgos, Spain). Journal of Human Evolution 37: 591-622.
Fossey, D. 1984. Infanticide in Mountain Gorillas (Gorrilla gorilla beringei) with Comparative Notes on Chimpanzees. In G. Hausfater and S.B. Hrdy (eds). Infanticide: Comparative and Evolutionary Perspectives. Pp 217 – 235. New York: Aldine Publishing.
Gillespie, R., Gowlett, J.A.J., Hall, E.T., Hedges, R.E.M. and Perry, C. 1985. Radiocarbon dates from the AMS system: Archaeometry Datelist 2. Archaeometry 27 (2): 237-246.
Goldman, L.R. (ed). 1999. The Anthropology of Cannibalism. Westport: Bergin & Garvey.
Goodall, J. 1977. Infant Killings and Cannibalism in Free-Living Chimpanzees. Folia Primatologica 28: 259-282.
Goodall, J. 1986. The Chimpanzees of Gombe: patterns of behaviour. Cambridge MA: Belknap Press of Harvard University Press.
Gowlett, J.A.J., Hall, E.T., Hedges, R.E.M. and Perry, C. 1986. Radiocarbon dates from the Oxford AMS system: Archaeometry Datelist 4. Archaeometry 28: 206-221.
Hamai, M., Nishida, T., Takasaki, H. and Turner, L.A. 1992. New records of Within Group Infanticide and Cannibalism in Wild Chimpanzees. Primates 33: 151-162.
Hedges, R.E.M., Housley, R.A., Law, I.A., Perry, C. and Gowlett, J.A.J. 1987. Radiocarbon dates from the Oxford AMS system: Archaeometry Datelist 6. Archaeometry 29: 289-306.
Hedges, R.E.M., Housley, R.A., Law, I.A. and Bronk, C.R. 1990. Radiocarbon dates from the Oxford AMS system: Archaeometry Datelist 10. Archaeometry 32: 101-108.
Hedges, R.E.M., Housley, R.A., Bronk, C.R. and van Klinken, G.J. 1991. Radiocarbon dates from the Oxford AMS system: Archaeometry Datelist 12. Archaeometry 33: 121-34.
Hedges, R.E.M., Housley, R.A., Bronk-Ramsey, C. and van Klinken, G.J. 1994. Radiocarbon dates from the Oxford AMS system: Archaeometry Datelist 18. Archaeometry 36: 337-374.
Hillson, S. 2000. Editorial: Cannibalism and Violence. International Journal of Osteoarchaeology 10: 1-3.
Hrdy, S.B. 1974. Male-Male Competition and Infanticide among Langurs (Presbytis entellus) of Abu, Rajasthan. Folia Primatologica 22: 19-58.
Hrdy, S.B. 1977. Infanticide as a Primate Reproductive Strategy. American Scientist 65: 40-49.
Hrdy, S.B. 1999. Mother Nature: A History of Mothers, Infants, and Natural Selection. New York: Pantheon
Hyman, L.H. 1940. The Invertebrates: Protozoa through Ctenophora. New York: McGraw
Itani, J. 1982. Intraspecific killing among Non-Human Primates. Journal of Social and Biological Structures 5 (4): 361-368.
Izikowitz, K.G. 1951. Lamet: hill peasants in French Indo-China (Etnol. Stud. 17). Goteborg: Etnografiska Museet.
Kalb, J.E., Jolly, C.J., Mebrate, A., Tebedge, S., Smart, C., Oswald, E.B., Cramer, D., Whitehead, P., Wood, C.B., Conroy, G.C., Adefris, T., Sperling, L. and Kana, B. 1982. Fossil mammals and artefacts from the Middle Awash Valley, Ethiopia. Nature 298: 25-29.
Kalb, J.E., Jolly, C.J., Oswald, E.B. and Whitehead, P. 1984. Early hominin habitation in Ethiopia. American Science 72: 168-178.
Korn, D., Radice, M. and Hawes, C. 2001. Cannibal: a history of the People-Eaters. London: Channel 4/Macmillan.
Larkin, P.A. 1956. Interspecific competition and population control in freshwater fish. Journal of the Fisheries Research Board of Canada 13: 326-342.
Lemonick, M.D. and Dorfman, A. 2003. The 160,000-Year-Old-Man. Time. Jun 23, 2003.
Manzi, J.J. 1970. Combined effects of salinity and temperature on the feeding, reproductive, and survival rates of Eupleura caudata (Say) and Urosalpinx cinerea (Say) (Prosobranchia: Muricidae). Biological Bulletin 138: 35-46.
Mead, S., Stumpf, M. P. H., Whitfield, J., Beck, J. A., Poulter, M., Campbell, T., Uphill, J., Goldstein, D., Alpers, M., Fischer, E. M. C. and Collinge, J. 2003. Balancing Selection at the Prion Protein Gene Consistent with Prehistoric Kuru like Epidemics. Science 300: 640-643.
Needham, R. 1976. Skulls and Causality. Man. New Series. 11 (1): 71-88.
NINDS – National Institute of Neurological Disorders and Stroke. Transmissible spongiform encephalopathies information page: http://www.ninds.nih.gov/disorders/tse/tse.htm (Accessed 22.10.05)
Poulson, T.L. 1963. Cave adaptation in amblyopsid fishes. American Midland Naturalist 70: 257-290.
Roach, J. 2003. Cannibalism Normal for Early Humans? National Geographic News. April 10 2003.
Russell, M.D. 1987. Mortuary Practices at the Krapina Neanderthal Site. American Journal of Physical Anthropology 72: 381-397.
Sanday, P. 1986. Divine Hunger: Cannibalism as a Cultural System. New York: Cambridge University Press.
Stringer, C. 2000. The Gough’s Cave human fossils: an introduction. Bulletin of the NaturalHistoryMuseum. Geology Series 56: 135-139.
Southwick, C.H. 1955. Regulatory mechanisms of house mouse populations: social behaviour affecting litter survival. Ecology 36: 627-634.
Scheinfurth, G. 1873. The heart of Africa. Three years’ travels and adventures in the unexplored regions of central Africa. From 1868 to 1871. 2 vols. London: Sampson Low, Marston, Low & Searle.
Taylor, T. 2002. The Buried Soul: How Humans Invented Death. London: Fourth Estate.
Trinkaus, E. 1985. Cannibalism and Burial at Krapina. Journal of Human Evolution. 14: 203-216.
Turner II, C.G. 1993. Cannibalism in Chaco Canyon: the charnel pit excavated in 1926 at Small House Ruin by Frank H.H. Roberts, Jr. American Journal of Physical Anthropology 91: 421-439.
Turner, C.G.II. and Turner, J.A. 1999. Man Corn: Cannibalism and Violence in the Prehistoric American Southwest. University of Utah Press. Salt Lake City.
Vilaca, A. 2000. Relations between Funerary Cannibalism and Warfare Cannibalism: The Question of Predation. Ethnos 65: 84-106.
Villa, P., Bouville, C., Courtin, J., Helmer, D., Mahieu, E., Shipman, P., Belluomini, G. and Branca, M. 1986. Cannibalism in the Neolithic. Science 233: 431-437.
White, T.D. 1986. Cutmarks on the Bodo Cranium: A Case of Prehistoric Defleshing. American Journal of Physical Anthropology 69: 503-509.
White, T.D. 1992. Prehistoric Cannibalism at Mancos 5MTUMR-2346. Princeton: University Press. Princeton.
White, T.D. 2001. Once were Cannibals. Scientific American (August): 86-93
White, T.D. and Toth, N. 1991. The Question of Ritual Cannibalism at Grotta Guattari. Current Anthropology 32 (2): 118-138.
White, T.D., Asfaw, B., Degusta, D., Gilbert, H., Richards, G.D., Suwa, G. and Howell, F.C. 2003. Pleistocene Homo sapiens from Middle Awash, Ethiopia . Nature 423: 742-747.
Yom-Tov, Y. 1974. The effect of food and predation on breeding density and success, clutch size and laying date of the crow (Corvus corone L.). Journal of Animal Ecology 43: 479-498.
About the author
James Cole’s main research interest is the interpretation of cutmarks on hominin skeletal remains in the Palaeolithic. He is currently working on constructing a nutritional template of the human body to be used in cut mark comparison to identify whether specific ‘nutritional zones’ are targeted by hominins when butchering their contemporaries remains. He also aims to use the model to help develop a solid methodology for distinguishing the differences between butchery and excarnation, which would in turn help archaeologists to better interpret deliberately modified hominin remains.