Laryngeal Contrast and Phonetic Voicing: A Cross-dialectal Laboratory Phonology Approach for Arabic

The thesis investigates the relationship between VOT and the phonological representation of laryngeal contrasts in Arabic, in a series of three studies, based on (i) plain and emphatic coronal plosives in eight Arabic dialects in data from the IVAr corpus; (ii) experimental data investigating the effect of speech rate on VOT in the full range of plosives in the Najdi and Hijazi dialects of Saudi, and the Tunisian dialect; and (iii) a parallel speech rate experiment on VOT in plain and emphatic coronal plosives in Modern Standard Arabic (MSA) produced by the same speakers (Najdi / Hijazi / Tunisian). Recorded data of real/nonsense words yielding emphatic/plain minimal pairs from 88 speakers in total across eight Arabic dialects (Moroccan, Tunisian, Egyptian, Syrian, Jordanian, Iraqi, Kuwaiti, Omani) were used from the IVAr corpus. The findings reveal a continuum of variation (rather than a clear-cut dichotomy) between dialects with three VOT categories (plain voiceless ~ emphatic voiceless ~ voiced) versus two VOT categories (voiceless ~ voiced). This finding adds to the classification of dialects as having either two or three VOT categories based on the literature on VOT and laryngeal contrasts (Bellem, 2007, 2014). Data for the speech rate studies were from 64 speakers (18 Najdi / 18 Hijazi / 28 Tunisian) in both registers (dialect speech and MSA). Evidence from speech rate effects in dialect and MSA speech indicate a complex set of relationships between the number of VOT categories and the number of active phonological features, resulting in over-specification at some stages of the putative sound change from one end of the continuum to the other. The implications for Arabic phonological representations, and for the modelling of a theory-driven typological analysis of laryngeal features in Arabic dialects and of phonetic uniformity in the context of sound change, are discussed.