Identifying Linguistic Structure in Aggregate Comparison

Dialectometric techniques for analyzing variation in the aggregate are maturing rapidly, but there is still little agreement on how to extract linguistic structure from aggregate comparison. The present paper explores one means of comparing aggregate analyses in order to determine linguistically concise restrictions, essentially the use of factor analysis. Using the Southern states data which Guy Lowman collected as part of LAMSAS, we apply factor analysis to the vowels involved in aggregate analyses in order to determine which alternations in pronunciation tend most to