Abstract:
The super-heavy syllable structure (CVCC) of Modern Standard Arabic (MSA) has been claimed to be the best syllable to study syllabification and syllable structures in all Arabic varieties (Kenstowicz, 1986). Tautomorphemic lexical items that have the CVCC syllable of MSA do not have underlying vowels in their canonical syllable structure and exhibit all possible manifestations within sonority sequencing (Al Tamimi & Al Shboul, 2013). Previous research of coda clusters on Arabic dialects has generally documented clusters of the CVCC syllable in one of two ways: Either the last two consonants in the cluster observe the sonority hierarchy and the clusters do not undergo any change, or the sonority hierarchy is violated and anaptyxis of the cluster occurs (e.g., Abboud, 1979; Prochazka, 1988; Ingham, 1994; Al-Mozainy, 1981). Najdi Arabic (NA) and Hijazi Arabic (HA) have been documented to adhere to this generalization, but there are also documented exceptions (e.g., Al Qahtani, 2014; Al-Mohanna, 1998). This dissertation reports on three experiments designed to ascertain how coda clusters of the CVCC syllable surface in NA and HA. The first experiment is a production experiment that is designed to elicit coda clusters with different sonority profiles by NA and HA speakers. The results of the production experiment show variation of clusters that undergo and forego anaptyxis within sonority rises, but indicate sonority falls and plateaus as acceptable clusters by speakers. Results of this experiment also find sonority distance to play a predictive role when anaptyxis occurs. The second and third experiments are perception experiments that are designed to complement the first experiment and minimize methodological issues associated with the production experiment. Results of these two experiments show that variation exists mostly within sonority rises, and less so within falls and plateaus; hence, the results of the perception experiments indicate that categorical use of sonority-based explanations is not sufficient to explain the variation within each coda type. Similar to the results of the first experiment, results of the perception experiments indicate that sonority distance imposes restrictions when anaptyxis occurs or does not occur. Results of the perception experiments also indicate that that other non-sonority-based factors, such as the order of the place of articulation of the coda sequences and Obligatory Contour Principle-NASAL can have an effect on when clusters undergo anaptyxis. The results of this dissertation suggest that speakers of NA and HA do not treat coda clusters in an all-or-nothing binary fashion.
