The identification of prehistoric ethnic groups suffers from troubling epistemological difficulties. The assumption that archaeological cultures correspond perfectly with biological populations, languages, and ethnic groups is faulty because the amount of overlap between the elements of the 'race-language-culture' triad must be determined empirically. The use of such overlap to identify prehistoric cultural groups is therefore highly suspect. Moreover, the study of ethnicity in sociocultural anthropology has demonstrated that it is a matter of self-identification at various social scales that does not necessarily correspond with any particular feature of material culture and is manipulated by individuals for social purposes. No cross-cultural regularities have been discovered that would enable the identification of ethnic groups in prehistory. Therefore, the study of ethnicity is impossible for most prehistoric contexts. However, in protohistoric and late prehistoric contexts, such as the Iroquoian Late Woodland period, ethnic attributions based on ethnohistoric and linguistic data may be possible by correlating historically known toponyms and ethnonyms with archaeologically known sites.