Second-language (L2) vowel production depends on how learners relate non-native vowel categories to their first-language (L1) phonetic system, yet little is known about how vowel-specific difficulty and improvement can be explained within an L2 speech learning model. This study examined English monophthong production by Hijazi Arabic-speaking learners through Flege’s Speech Learning Model and compared mobile-based Interactive Phonetic Training (IPT) with Print-Based Phonetic Instruction (PBI). Forty Hijazi Arabic-speaking English majors completed pre-test and post-test production tasks involving eleven English monophthongs. Productions were analysed through vowel-specific confusion matrices and modelled using mixed-effects logistic regression. The results showed that learners had the greatest difficulty with vowels lacking direct Hijazi Arabic equivalents, especially /ɛ/, /ɜː/, /ɔː/, /ɒ/, and /ʌ/. The IPT group showed significantly greater post-test improvement than the PBI group, particularly for these difficult vowels. The findings suggest that interactive phonetic training can reduce L1-based substitutions and support more accurate L2 vowel production.
Wael ALMURASHI (Wed,) studied this question.