Session double de l'Atelier de phonologie

11
oct.
2017.
10h00
12h00
Brandão de Carvalho - Faust : Session double de l'Atelier de phonologie

Salle 159

Why is tl a bad onset? / Joaquim Brandão de Carvalho

Languages having (i) phonemic laterals and rhotics, and (ii) TR onsets typically show /pr, tr, kr/ but only /pl, kl/ and not /tl/ (except in the case of a bogus cluster). The problem raised by tl-onsets has been accounted for hitherto in terms of perceptive inadequacy (Flemming, Hallé & Best, Hallé). Here I will focus on the possible representational basis of this perceptive defect. To understand why /tl/ is a bad onset, we need to explain (i) why /tr/ is a good one, and find out the difference between the internal structure of the two liquids, and (ii) the structure of tautosyllabic sequences of the muta cum liquida type. I will make two assumptions on these points, from which it follows that *tl simply results from the well-known Obligatory Contour Principle (OCP).

 
Support for the root hypothesis: allomorphy in Mehri imperfective plurals / Noam Faust
 
In Semitic, words are traditionally analysed as morphologically complex. A word like Modern Hebrew [χaʃav] 'he thought' is decomposed into a "root" √χʃv and a template CaCaC, essentially because 1) there are many words with related meaning that share with [χaʃav] only the elements attributed to the root, and 2) CaCaC and other templates reccur with many roots. Starting with Bat-El (1994 et passim), a trend has emerged in studies of Semitic morphology which argues against the morphemic status of "roots" in Semitic. According to Bat-El, a word like Modern Hebrew [χaʃav] is stored as a stem. While the template is a morpheme, the root is "only a residue". In this short talk, I will show a case of allomorphy from Mehri, in which the consonantal allomorph of a plural suffix is abandoned because of incompatibility with a vowel of the base; but it is not sensitive to a consonant of the base, which is linearily closer to it than the vowel. I conclude that morpho-syntactically, roots must appear on a different, more embedded level than the template they occupy, supporting morpho-syntactic decomposition of such words into root and template.