Cross-Linguistic Phonotactics: A Comparative Study of Arabic and English Stress Patterns

Thesis Abstract

This study addresses a gap in phonotactic research by examining how stress interacts with cross-linguistic phonotactics in Arabic and English, with particular attention to how stress placement influences permissible syllable structures, prosodic boundaries, and phoneme sequencing in word formation and loanword adaptation. The aim is to elucidate cross-language variation and universals in stress-driven phonotactics, and to determine how Arabic and English listeners and speakers perceive and produce stressed sequences, contributing to models of stress assignment and syllable structure constraints. Specific objectives are to (i) compare stress placement rules and their interaction with phonotactic filters in both languages; (ii) identify language-specific and cross-linguistic tendencies in permissible consonant clusters, vowel quality, and syllable structure under stress conditions; (iii) assess perceptual salience and misperception of stressed syllables using reaction-time paradigms; and (iv) test the applicability of existing theoretical frameworks to explain cross-language stress-phonotactics interactions. A mixed-methods design is employed to triangulate corpus-based analysis, perception experiments, and production data. The population comprises native Arabic speakers (N=60) and native English speakers (N=60) residing in multilingual urban centers, with balanced representation of gender and age ranges (18–45). A stratified random sampling approach yields two samples of 60 participants each for clarity in cross-language comparisons. Data collection instruments include a stress-evoking wordlist and nonword paradigm, a targeted corpus of 5,000 tokens drawn from contemporary Arabic dialects and British and American English corpora, and perception tasks consisting of forced-choice identification and speeded discrimination with reaction-time measurement. Production data are captured via reading and repetition tasks administered to language-malance cohorts, while perceptual data are collected through computerized experiments implemented with E-Prime. Analytical methods combine quantitative and qualitative techniques. Corpus-derived phonotactic probabilities are estimated using Bayesian surprisal metrics and grapheme-to-phoneme alignment to quantify stress-conditioned phonotactic constraints. Production data are analyzed with mixed-effects logistic regression to model the probability of acceptable stress-phonotactic sequences, alongside linear mixed-effects models for articulatory duration and intensity. Perception data utilize ANOVA and generalized linear models to compare accuracy and reaction times across language groups, with post hoc Tukey tests to identify cross-language differences. A factorial design examines the interaction between stress position (initial, medial, final) and syllable complexity (CVC, CVCC, CVCCC) on acceptability judgments. Theoretical interpretation is guided by the Metrical Theory of Stress and the Articulatory Phonology framework, with supplementary insights from Optimality Theory’s constraints on phonotactics and a usage-based perspective on cross-linguistic phonotactic variability. Data reliability is ensured through test-retest procedures for perception tasks, inter-rater agreement for coding of production data, and bootstrapping to validate model estimates. Key expected findings include language-specific stress-phonotactic patterns Arabic is anticipated to exhibit stronger consonant cluster constraints and vowel quality variation in stressed syllables due to prosodic-phonotactic coupling, while English is expected to show greater tolerance for certain coda clusters under stress and more consistent stress-sensitive syllable weight adjustments. Perceptual results are expected to reveal quicker and more accurate identification of stress-bearing sequences in the native language, with cross-language transfer effects detectable in bilingual or multilingual participants. Production data are anticipated to reveal measurable differences in articulation duration, voicing, and intensity correlated with stress position and syllable complexity. The study is expected to contribute to knowledge by clarifying how stress interacts with language-specific phonotactic constraints and by proposing an integrative model that accounts for cross-linguistic similarities and differences in stress-phonotactics across Arabic and English. The research will advance theoretical understanding of phonotactics by integrating metrical and articulatory perspectives within a cross-linguistic framework, and it will have practical implications for second language teaching, speech synthesis, and automatic speech recognition by informing language-specific stress modelling and phonotactic constraint encoding. The main conclusion is anticipated to be that stress acts as a predictive cue modulating phonotactic permissibility in language-specific ways, while revealing universal tendencies in stress-driven phonotactic reorganization. Recommendations include developing language-aware pronunciation training tools that adapt to Arabic and English stress-phonotactic profiles, and extending the model to include additional SLG (second-language grammar) interactions and dialectal variation to enhance generalizability.