Dynamic specification of vowels in Hijazi Arabic.

Research on various languages shows that dynamic approaches to vowel acoustics - in particular Vowel-Inherent Spectral Change (VISC) - can play a vital role in characterising and classifying monophthongal vowels compared with a static model. This study's aim was to investigate whether dynamic cues also allow for better description and classification of the Hijazi Arabic (HA) vowel system, a phonological system based on both temporal and spectral distinctions. Along with static and dynamic F1 and F2 patterns, we evaluated the extent to which vowel duration, F0, and F3 contribute to increased/decreased discriminability among vowels. Data were collected from 20 native HA speakers (10 females and 10 males) producing eight HA monophthongal vowels in a word list with varied consonantal contexts. Results showed that dynamic cues provide further insights regarding HA vowels that are not normally gleaned from static measures alone. Using discriminant analysis, the dynamic cues (particularly the seven-point model) had relatively higher classification rates, and vowel duration was found to play a significant role as an additional cue. Our results are in line with dynamic approaches and highlight the importance of looking beyond static cues and beyond the first two formants for further insights into the description and classification of vowel systems.

Static and dynamic cues in vowel production in Hijazi Arabic.

Static cues such as formant measurements obtained at the vowel midpoint are usually taken as the main correlate for vowel identification. However, dynamic cues such as vowel-inherent spectral change have been shown to yield better classification of vowels using discriminant analysis. The aim of this study is to evaluate the role of static versus dynamic cues in Hijazi Arabic (HA) vowel classification, in addition to vowel duration and F3, which are not usually looked at. Data from 12 male HA speakers producing eight HA vowels in /hVd/ syllables were obtained, and classification accuracy was evaluated using discriminant analysis. Dynamic cues, particularly the three-point model, had higher classification rates (average 95.5%) than the remaining models (static model: 93.5%; other dynamic models: between 65.75% and 94.25%). Vowel duration had a significant role in classification accuracy (average +8%). These results are in line with dynamic approaches to vowel classification and highlight the relative importance of cues such as vowel duration across languages, particularly where it is prominent in the phonology.