A cross-cultural assessment of similarities in interior Salish myth content

Descriptive statistical techniques were applied to the cross-cultural study of Interior Salish myth content. The use of an electronic computer has facilitated the analysis of a large body of myth data. This work is a departure from traditional historically oriented cross-cultural studies in that the entire mythologies, rather than motifs or tale-types, of peoples are compared. A pragmatic approach necessitated by the size of the corpora analyzed requires that the number of cultures be relatively small (eight were included in the analysis) and that the units of analysis be rudimentary (the word is the unit employed). The computer was used to determine the frequencies of all words in each corpus. Frequential analysis of the myth data was limited to three semantic areas: religion, social organization, and technology. Two correlation coefficients were employed. Spearman's Rho (a statistic which measures similarity of rank orderings) was tested and rejected in favor of a modified form of Robinson's Index of Agreement (a statistic which measures differences in percentage frequencies). Both the comparison of relative frequencies of individual words and the overall assessment of similarities in frequency through submission of correlation matrices to Smallest Space Analysis result in the recognition of two cultural groups: a Canadian Plateau group consisting of the Lillooet, Lower Thompson, and Upper Thompson corpora and an American Plateau group consisting of the Okanagon, Colville, Sanpoil, Kalispel, and Coeur d'Alene corpora. This division of the Interior Salish myth corpora into two geographically distinct groups coincides with previously observed cultural and linguistic classifications.Arts, Faculty ofAnthropology, Department ofGraduat