Vowel deletion before sibilant-stop clusters in Latin: issues of syllabification, lexicon and diachrony

Abstract

AbstractSyncope of a short vowel before a clustersT(R)(T: stop;R:lorr) is attested in Latin and Sabellic: Latinfēstra(besidefenestra),sēstertius,iuxtā, Oscanvezkeí, minstreis, Umbrianetuřstamu,mersto. This phenomenon raises important questions both for the historical phonology of the Italic languages and for the typological study ofsT(R)-clusters. In Latin and Sabellic, syncope normally took place only in open syllables. Three competing strategies are possible in order to explain this paradox. (i) It has been argued that the clustersT(R)was an onset, which would imply that the vowel standing before the clustersT(R)was in an open syllableat the time of the syncope. (ii) It has been proposed that the sequencesTbehaves as a single consonant. (iii) It could be assumed that the syncope did not take place in a closed syllableexceptif the closing consonant wass(or its allophonez). Furthermore, a careful study of the relevant material shows that in some Latin words a vowel standing beforesT(R)may have been deleted by a phonological process distinct from the syncopestricto sensu(haplology,noundinum-rule).