Asymmetric Distributions of Glides In General American English

The purpose of this paper is to provide an explicit formal account for asymmetric distributions of glides in General American English. The data, which we attempt to analyze in the present study, can be summarized as follows. First, the labial glide [w] and the palatal glide [j] cannot coocur with preceding labial and coronal consonants, respectively. Second, only [w] cannot cooccur with preceding sonorants. Finally, unlike [w] which can occur before any vowels, [j] can occur only before [ul.Most previous analyses employ somewhat complicated and arbitrary mechanisms. For instance, in Davis and Hammond (1995), different syllable positions, onset and nucleus, are assigned to [wl and [j], respectively. Also, several intermediate representations are posited during the derivation. In the present study, we provide a constraint-based parallel analysis within the framework of Optimality Theory. Markedness constraint ocr is adopted to account for the lack of homorganic CG clusters such as *[pwl and *[tj]. Another Markedness constraint, which prohibits a sequence of sonorant consonants, is adopted to account for the lack of sonorant-[wl sequences like *[nw]. Finally, to account for the fact that [j] can occur only before [ul, we propose a faithfulness constraint for a certain sequence, Le. MAX(i-u) which requires that an underlying sequence /iu/ must have correspondents with identical precedence relations at the surface. We have shown that the interaction of these constraints produce all the asymmetric patterns of glides in General American English