Retour vers liste

Détail de la contribution

Auteur: Sam HELLMUTH

Co-Auteur(s): Rana Alhussein ALMBARK, University of York

A typology of word stress in Arabic dialects: phonetic correlates and rhythmic variation

Abstract/Résumé: This paper seeks to establish the phonetic correlates of word stress in a sample of spoken Arabic dialects and to explore whether any observed dialectal variation co-varies with other prosodic properties (speech rhythm, syllabic structure, prosodic timing). The study is part of a wider investigation of the intonational phonology of Arabic dialects ( Spoken Arabic dialects are well known for variation in the position of word stress, but there has been little work on the phonetic realization of word stress in Arabic. Investigation of the phonetic correlates of stress in Arabic has been limited by the absence in most dialects of minimal pairs distinguished by stress alone, as exploited in Beckman’s (1986) study of English and Japanese. Two studies which used stress near-minimal pairs suggest variation in the correlates of word stress across dialects: in Jordanian Arabic F0, duration and intensity are reported (Zuraiq 2005) whereas in Tunisian Arabic it is F0, spectral balance and F1 lowering (Bouchhioua 2008). Secondly, investigation of the correlates of word stress presupposes reliable identification of different levels of prosodic prominence (Beckman & Edwards 1990), which is problematic in dialects whose intonational phonology is not yet fully understood. We present here a comparative survey of the phonetic correlates of word stress in three regional dialects (Egypt, Gulf and Jordan), based on acoustic analysis of f0, duration, intensity, F1/F2 and spectral balance in stress near-minimal pairs (following Bouchhioua 2008). With a working analysis of the intonational phonology of each dialect in hand, it will be possible to disambiguate levels of phrasing, and we expect to find variation in the phonetic correlates of word stress across the three dialects. We also explore possible correlations between any observed variation and other prosodic properties of the dialects under investigation, including speech rhythm, syllabic structure (degree of vowel reduction, incidence of consonant clusters) and prosodic timing (accentual and/or pre-boundary lengthening). Rhythmic variation across Arabic dialects, in a continuum from more to less ‘stress-timed’, has already been documented (Ghazali et al. 2002). We will present rhythmic analysis of read speech sentences using a range of rhythm metrics (Wiget et al. 2010). We will also present the results of preliminary analysis of accentual lengthening and pre-boundary lengthening in each dialect, which are argued to contribute to variation in speech rhythm as much, if not more than, differences in vowel reduction and the incidence of consonant clusters (Prieto et al. 2012).