The representation of sC-clusters: A Strict CV analysis

Abstract

sC-clusters have always posed a challenge for theories of phonological representation because they behave like coda-onset clusters, but they can also occur word-initially (where their behaviour typically differs from that of branching onsets comprising an obstruent followed by a sonorant). [s] in such initial clusters has been analysed as an appendix, as a coda, or as part of a complex segment. Out of these possibilities, the coda analysis (where initial [s] is assumed to be preceded by an empty nucleus) is shown to account for the data in the most satisfactory way. However, it faces the problem of how to ensure that this empty nucleus remains silent (called Magic Licensing). Instead, I propose a Strict CV representation of initial sC-clusters, where the melody of [s] is shared by a neighbouring V position, by a language specific choice: in Italian and Portuguese [s] occupies the preceding V position, while in English the following one. As the initial nucleus does not remain empty in this analysis, no special licensing is necessary. I also show that not only is the branching representation of [s] identical to that of syllabic sonorants in this model, their distribution is also parallel. In English, syllabic sonorants branch on the left, while syllabic [s] branches on the right, and the differences in their behaviour follow directly from the difference in the direction of branching. Finally, the marked structure of syllabic consonants is only permitted if without their branching the representation is ill-formed. This means that they must always be flanked by a consonant at least on one side.