12h00
en ligne
Parallel Spirantisation but Cyclic Syllabification in Syriac? No way!
Dominique Stephan Bobeck, Universität Leipzig, dom.bobeck [at] gmail.com ()style="color:rgb(108, 42, 15);cursor:pointer;text-decoration:none;" rel="nofollow noopener noreferrer noreferrer nofollow noopener noreferrer" target="_blank"
Atelier de phonologie
Syriac (Semitic, Aramaic) is characterised by post-vocalic spirantisation and the deletion of unstressed vowels (Nöldeke 1880, Edzard 2001). Note that vowel deletion counterbleeds spirantisation with the consequence of overapplication. The past tense paradigm seems to comprise three distinct bases, i.e. kθav-, keθb-, and kaθv-. It is proposed that all these forms have the underlying representation /katab/ ‘write’. Examples are: kθav (write.3sg/pl.m) ‘he wrote’, keθb-aθ (write-3sg.f) ‘she wrote’, and kaθv-uː-x (write-3pl.m-2sg.m) ‘they wrote to you’. These data show that the original vowels of the base katab- reappear if it is concatenated with V(ː)C-shaped suffixes within a single phonological cycle, cf. /katab+uː+x/ → [kaθvuːx], not *[keθbuːx]; cf. kθav ‘he/they wrote’. However, how can keθbaθ (3sg.f) and keθbeθ (1sg.c) be explained? Should not /katab+aθ/ yield *[kaθvaθ]? Why can [a] surface in kaθvuːx but not in keθbaθ? Why does spirantisation apply in one case but not in the other one? Although we get the impression of parallel phonology for spirantisation, this presentation shows that only cyclic syllabification can predict all forms of the Syriac verbal paradigm. Apparently parallel phonology, however, is explained via more complex exponents, i.e. -aθ (2sg.f) and -t (1sg.c) plus floating [−continuant], which reconvert derived fricatives into stops.
