ABSTRACT. Holmquist presents results for phonological conditioning of the vocalic variables: overall, the high and raised-mid variants of the front- and back-vowel variables are not the most frequent in the data. Kenstowicz argues for a more direct analysis of phonological similarities between morphologically related words in terms of a family of constraints requiring phonological identity between words that stand in certain morphological relationships. Albright observes that the uniform exponence account formalizes the intuition that paradigm leveling is due to a pressure for non- alternating paradigms. (pp. 145–149)