This item is not viewable because it is not yet digitized or has special access permissions. Please log in or request access.
Numerical Taxonomy and the Classification of Salish Indian Languages
Academic Work
In 1950, Morris Swadesh presented a classification of Salishan languages; his classification was based on percentage similarities in 'selected basic vocabulary' of all pairs of thirty Salishan languages and dialects (see matrix of similarities, Table 1). In 1962, Dyen proposed a general procedure for classifying languages into family trees based on the same kind of data as Swadesh used; Dyen in fact eexamplified the method with the use of Swadesh's matrix of similarities for the thirty Salishan languages. In 1969, Jorgensen using a different procedure - one based on the work of numerical taxonomists in biology - provided a third classification of the same thirty languages. Swadesh, Dyen, and Jorgensen all realized that the data used - the matrix of similarities - reflected for the most part two distinct kinds of historical processes: genetic drift and borrowing. While Swadesh and Jorgensen did not distinguish between the effects of drift and borrowing in their classifications, rather letting them reflect the effects of both, Dyen did attempt to eliminate particular similarities values which were judged to be 'inflated' by borrowing. It is our purpose in this paper to provide yet a fourth classification of the same matrix of similarities of Salishan languages, one which we think an improvement on the earliest classifications, especially in distinguishing the effects of borrowing from the effects of drift. Because of doubts about the adequacy of the Salish data, we are not so much attempting to definitvely classify the Salishan languages (see Dyen 1962: 256, footnote 7 on Swadesh's doubts about the data) as to present a method. We will outline the problem first, by briefly discussing the classifications and classification methods used by Swadesh, Dyen, and Jorgensen. Second, we will describe our method and the assumptions on which it is based. Third, we will describe the major differences in classifying Salishan languages that our method yields as compared with the results of the earlier methods; we will give our reasons for thinking our classification preferable to the earlier ones. Finally, we will indicate what further work needs to be done in the way of improving on the method and testing it. Throughout our discussion we will be making use of special and novel assumptions about the nature of the data and the workings of our method. Some of these assumptions are certainly subject to debate. However, insofar as they place strong constraints on the data, and insofar as these strong constraints are met, we infer some support for the assumptions.
AW.00127
1974
Language and Linguistics
UBC Working Papers in Linguistics Press
International Conference on Salish and Neighbouring Languages, 9, 119-162
International Conference on Salish and Neighbouring Languages, 9, 119-162
Working Papers
Do you have a comment, story, or something you would like us to know related to this item?