Coronal Unmarkedness and Clusters in Correspondence Theory

Language, Information and Computation(PACLIC 11), 1996, 415-424 Coronal Unmarkedness and Clusters in Correspondence Theory Shinsook Lee and Mi-Hui Cho Hoseo University and Pukyong National University leess@dogsuri.hoseo.ac.kr and mhcho@dolphin.pknu.ac.kr Abstract This paper examines paradoxes involving coronal specification. Although several phonological phenomena such as place assimilation, insertion and deletion argue for coronal underspecification, other facts like morpheme structure conditions demand coronal specification. This paper shows that this problem can be solved with the markedness hierarchy which ranks coronals low under Correspondence Theory. The hierarchy also accounts for the distributional bias of coronals in word clusters without appealing to Yip's Cluster Condition and coronal underspecification. 1. Introduction Until recently, phonological theories crucially depended on the assumptions of underspecification and feature-filling application of rules in a derivational mode. In particular, Paradis and Prunet (1991) and Cho (1991), among others argue that coronals are underspecified due to their asymmetrical behavior in several phonological phenomena such as place assimilation, deletion, and epenthesis. However, there has been some evidence against coronal underspecification. For example, morpheme structure conditions in both English and Korean need to refer to the unmarked coronals to rule out certain sequences of coronal consonants (McCarthy and Taub 1992). English plural and past tense suffixes should also refer to coronals in order to account for vowel insertion between "two like (or similar) coronals". Then how can we account for the paradoxes involving coronal specification? That is, on the one hand, coronals should be underspecified because of the asymmetrical behavior, but on the other hand, coronals should be specified due to their reference in the description of some other phonological phenomena. In this paper we argue that coronal paradoxes can be resolved in Correspondence Theory which assumes constraint interactions in a parallelistic mode (McCarthy and Prince 1995). In specific, we contend that with the low ranking of Coronal, coronal unmarkedness is naturally derived, without appealing to coronal underspecification. We also argue that Yip's Cluster Condition (1991) which crucially depends on coronal underspecification cannot be maintained. Rather we show that the fact that coronals frequently show up in clusters can receive a unified account within Correspondence Theory. The organization of this paper is as follows: Section 2 examines the analysis of coronal unmarkedness in terms of underspecification theory. Section 3 provides evidence against coronal underspecification from several phonological phenomena, in languages such as English and Korean. Section 4 gives a unified account of coronal unmarkedness and cluster facts within Correspondence Theory, solving the problem of coronal paradoxes. Section 5 summarizes our findings. 2. Coronal Unmarkedness in Underspecification Theory It has been assumed in the literature that coronals are underspecified because of their unmarked status in several phonological phenomena, as reviewed in Paradis and Prunet (1991). First, coronals are potential targets of 415 rules like place assimilation in languages like English and Korean. Some relevant data are given in (1) and (2) (Gimson 1962, Jun 1995, Cho 1991): (1) Place assimilation in English a. coronal --> labial that pen [8wp pen]� ten pounds [tern pawndz] b. coronal --> …

Content truncated to 500 words