As Sahlins notes, cannibalism is always symbolic even when it is real. Other than those situations of disaster and starvation, cannibalism is rarely just gastronomic. The case of the Miyanmin of Papua New Guinea may be one exception, where blood feuding is thought to have turned into a quest for human meat. We can dismiss as unsubstantiated Levi-Strauss's claim that cannibalism formed part of a symbolic gastronomic opposition in which kin were boiled and outsiders were roasted. Cultural ecological research in the 1960s and 1970s focused on the nutritional value, and adaptive potential, of eating human fiesh. Aztec nobles' eating of limbs from human sacrifices, observed by the conquistadores, has been interpreted by Harner and Harris as an adaptive means of acquiring protein in an environmentally exhausted ecosystem even though most sacrifices were carried out at the time of the harvest when food was plentiful. For Harner and Harris, the symbolic motivations of rebirth for the eaten and the divine nature of human flesh were simply rationalizations of a latent ecological materialism - symbolism and culture 'covering up' for meat protein hunger a position rejected as patently absurd and ethnocentric by others.

Like ethnography, archaeology is one of those fields in which the otherness of the savage and primitive is reified through claims of cannibalism, thereby serving to reinforce selfidentities of civilized superiority. The popular imagination, as reflected by newspaper interest, is always ready to devour stories on archaeological discoveries of cannibalism.

There has been something of a recent revival in cannibalistic interpretations from archaeological evidence. At the cave site of Fontbregoua in south-eastern France, human bones among Neolithic deposits of butchered animal bones exhibit the same cut marks, while long bones are broken as if to extract marrow, all in the absence of carnivore tooth marks. Modifications of human bones found in deposits of the Pueblo II/III period (AD 1000-1300) at Chaco Canyon and Mancos - sites belonging to the Anasazi culture of the south-western USA - have been considered as revealing probable cases of violent death followed by cannibalism. Criteria that archaeologists have used in identifying cannibalism from the study of human remains include:

No single instance is judged sufficient to provide evidence of cannibalism and all criteria have to befflet. For example, detailed taphonomic studies of Neanderthal bones previously thought to be indicative of cannibalism, especially the burnt and split bones from Krapina and the damaged skull from Guattari Cave on Monte Circeo, do not support the cannibalism hypothesis. Those archaeological cases that do qualify all appear to involve aggressive cannibalism, inflicted on defeated outsiders, which is perhaps riot surprising since compassionate cannibalism, of the sort practiced by kin during funerary rites, may be considered less likely to result in such extensive bone modification.

Cannibalism is, therefore, extremely difficult to demonstrate archaeologically beyond doubt; absolute proof requires the identification of human tissue or bone in human coprolites. The difficulty of proving the existence of cannibalism from archaeological remains is, on one level, methodological. But there is also an ideological problem, outlined above, in that the cultural baggage of chauvinist and racist attitudes to the 'primitive' is so strong in popular perceptions that archaeologists are often at pains to refute such claims because they are deeply reluctant to bolster unquestioned attitudes of discrimination which, in the west, are centuries old. Until cannibalism can be understood as simply one means among many of transforming the dead corpse, rather than the monstrous bogey so repellent to the civilized mind, such sensationalizing stereotyping will continue to be reinforced.