Coronal Specification and Licensing in Place Assimilation

Language, Information and Computation(PACLIC 11), 1996, 489-500 Coronal Specification and Licensing in Place Assimilation' Seung-Hoon Shin Indiana University sshin@indiana. edu Abstract Based on unmarkedness of coronal cross-linguistically, coronal has been considered to be underspecified. Also, place assimilation where coronal tends to assimilate to the adjacent labial or velar by a feature- filling rule has been support the underspecification of coronals. I instead argue that coronal should be specified in Optimality Theory despite its unmarkedness, and place assimilation no longer provides evidence of underspecification of coronal. In addition, I maintain that place assimilation is a way of respecting licensing requirements and that a licensing hierarchy with respect to place feature of a coda empirically resolve current issues regarding coronal underspecification and place assimilation. 1. Introduction The theory of Underspecification stipulates that either all unmarked (predicted) features (Kiparsky 1982, Archangeli 1988, Avery and Rice 1989, 1989, Pulleyblank 1988 among others) or redundant features (Steriade 1987, Archangeli 1988, Mester and Ito 1989) be not specified in underlying representation. Hence, given that coronal is cross-linguistically the most unmarked consonant, under Radical Underspecification theory, it was considered as underspecified (Kiparsky 1985, Goldsmith 1989, Avery and Rice 1989, Paradis and Prunet 1989 among others) unless the coronal consonant has a dependant feature, which depends on the language. For instance, as Paradis and Prunet (1989) observe, in some languages, vowel or tone spreading occurs through a transparent coronal. In addition to its transparent property in some phonological processes, coronals in many languages are most likely to assimilated to the following labials or velars (Kiparsky 1985, Avery and Rice 1989, Steriade 1995, Crimson 1989). Based on such evidence, it has been argued that coronals do not have either a Place node (Paradis and Prunet 1989) which presumably is not opaque but is transparent to vowels and tones, or a Coronal node (Avery and Rice 1989), which forces it to be assimilated. Although the underspecified thing is different, both hypotheses agree in that coronals should be underspecified in some sense. Nevertheless, it is still believed that underspecification of a feature hinges on the phonological property of a language. Also, there have been a number of phonologists who claim that coronal should be specified despite its unmarkedness (Clements 1987, Steriade 1987, Kim H.S. 1994, Ito and Mester 1989). They contend that all unmarked features should be fully specified and only a redundant 489 feature is underspecified; otherwise all features, although they are unmarked and predictable, are specified. The goal of this paper is to solve problems raised in the traditional analyses regarding underspecification of coronal within the framework Optimality Theory (Prince and Smolensky 1993, McCarthy and Prince 1993). Hence, in this paper, I will argue based on place assimilation examples that unmarked coronal should be fully specified underlyingly, and that ranking of licensing constraints rather than coronal underspecification explains unmarkedness of coronal and typology of place assimilation. In the following sections, we will review place assimilation of some languages which has been used to establish coronal underspecification. 2. Place Assimilation …

Content truncated to 500 words