Compensatory Lengthening and Cluster Reduction in First Language Acquisition: A Comparison of Different Analyses

In this paper we will discuss phonological structure in the light of first language acquisition data obtained in a case study. These data exhibit cluster reduction (henceforth: CR) and compensatory lengthening (henceforth: CL). The main problem here is that for the description and explanation of these different phonological processes, different models seem to be more adequate than others. Working within the framework of Universal Grammar (UG), with its generalizations and economy of representation, this should raise some eyebrows. Do we believe that a child is endowed with different models for the same phonological domain? Obviously, in UG there is only one way that leads to Rome; of all possible grammars, one should meet the UG-criteria best. This paper is organized as follows. In section 2 we introduce a constituent syllable model (Cairns and Feinstein 1982) and a model that incorporates weight in the syllable representation (Hayes 1989). In section 3 we will show that the former model seems to be better equipped to account for CR-data and in section 4 we will discuss the claim that CL is best accounted for by the latter. In this section we will also discuss some fundamental flaws of this model. We will propose an obvious synthesis of the different models in section 5. However, this synthesis will prove to suffer from the same demerits as the original models on which it is based. Section 6 deals with a recently proposed foot-based account of CL by Schane (1994) and we will see how our data fit in. Schane’s account can be converted to the recently proposed framework of Optimality Theory (Prince and Smolensky 1993) (henceforth: OT) (section 7). Because OT also enables us to deal with CR, it will prove to provide us with the most general account of the different data discussed in this paper.