Title: What is the sound of a rhyme branching?
Abstract: In this talk, I will illustrate the cross-linguistic validity of the claim that “syllable rhymes seek to branch”. In Government Phonology (Kaye et al. 1985, 1990), there are two ways in which a rhyme may achieve branching: either by acquiring a coda through “coda licensing” (Kaye 1990) - the case in non-final closed syllables - or by having a long nucleus. In simplex vocalic expressions like [i] and [u], long nuclei are achieved by spreading of lexical vowels to two timing slots, or “points”. Such spreading, in turn, is only possible under licensing from a following contentful nucleus (i.e. only in open syllables) or through parametric licensing of the final point. This leaves one syllabic position unaccounted for, namely final closed syllables. Since final consonants are never codas, the rhyme may branch; but since the second timing slot is not licensed by a following vowel, it may not do so through spreading.
The talk shows two cases where this situation is resolved through the insertion of an additional element A, lowering final high vowels in exactly this situation (all data have been collected by me). This is the case in Jaffa and Jerusalem Arabic, where unstressed [i] and [u] are lowered only in final closed syllables. These vowels are always longer than non-final closed syllable, be they stressed or not, attesting to the effective branching of the rhyme. In Nazareth Arabic, this also happens in final open syllables, which is expected if the licensing of final time slots in indeed parametric and disables in this dialect. In Qaraqosh Neo-Aramaic, extra length is added to this syllable only under stress. Accordingly, the same phenomenon is attested, though only under stress (and only in my data, not in those reported by the only grammar, Khan (2002)). I conclude that in these two languages, the sound of a non-licensed rhyme branching is A, or lowness. Time permitting, I will present the case of Québec French, where unlicensed branching is expressed in laxness.