Intonational Variation on Saudi Dialects: a cross-dialectal corpus-based approach

Intonational patterns have been shown to vary across Arabic dialects on a regional level. This thesis presents one of the first investigations of inter-dialectal intonational variation within a single Arabic-speaking country (Saudi Arabia). The aim of the thesis is to describe the prosodic realisation of yes/no questions (ynqs) and focus utterances in three urban Saudi dialects (Najdi, Hijazi, and Jizani) in a speech production corpus stratified by age and gender. Qualitative analysis of prosodic contours in ynqs from two speech styles showed two main findings: a dialectal difference in which ynqs are typically realised with one of two contours: a rise (Hijazi and Najdi) versus a rise-fall (Jizani), and a potential indication of dialectal change in which young Jizani speakers use more rises than older Jizani speakers. Quantitative analysis of the acoustic cues employed in prosodic marking of focus from two elicitation tasks found similar off-focus cues in both pre-focus and post-focus positions across dialects, but small dialectal differences in the type and degree of acoustic cues observed in on- focus positions; the Jizani speakers displayed less focus enhancement in subjects. Finally, qualitative analysis of non-prosodic focus-marking strategies (such as ellipsis) in data excluded from quantitative analysis due to disfluency, with comparison to data from unscripted portions of the speech corpus, revealed dialectal difference in use of syntactic focus marking strategies, which are used more by older Jizani speakers. From this first parallel description of two key features of three urban Saudi dialects, the thesis argues that the three patterns found in the results are not independent of each other, but due to a trade-off in the allocation available of prosodic strategies for marking of ynqs and marking of focus.