Abstract
In his eighth-century treatise al-Kitab the Baghdad-based Persian scholar Sibawayh divided the Arabic consonants into two classes, majhur and mahmus. The focus of this study is his inclusion of ta, qaf and hamza among the majhur consonants. 2 Their classification has stimulated much discussion because while all other majhur consonants are voiced, ta' and qaf are voiceless in Modern Standard Arabic (McCarus 2012), and there has been a general reluctance to describe hamza as voiced. We demonstrate that modern instrumental techniques of phonetic investigation, applied to Modern Standard Arabic (MSA) fusha style speech, can help us to make sense of Sibawayh's classification without having to assume either that he was mistaken about their voicing or that ta and qaf were voiced at that time. Aerometric data from three speakers show that volume-velocity airflow at the release of majhur stop consonants is much lower than for mahmus stop consonants, and laryngographic and spectrographic analysis of the speech of eight speakers provides evidence that for ta' and qaf airflow is reduced by laryngeal adjustment while hamza is shown to be voiced, albeit typically with irregular glottal cycles. We relate our results to glottal states as described in Esling and Harris (2005), and conclude that the defining feature of majhur consonants is low airflow regulated by glottal constrictions of various kinds and possibly also by reduced initiator power in the pulmonic airstream mechanism.