| Although intelligibility has been widely studied in European languages, empirical research on Indonesian Non-Native English-Speaking (INNES) lecturers remains limited. This study investigates English as a Foreign Language (EFL) learners’ comprehension of INNES lecturers’ pronunciation, analyzing the influence of segmental and suprasegmental features, familiarity, syntactic cues, and lexical frequency. Quantitative analysis was utilized using the Mann-Whitney test. The data were derived from segmental and suprasegmental features and collected from test results obtained through two tasks, with the lecturers’ voice recordings serving as stimuli. In the first task, two female lecturers (LA and LB) who recorded 38 pairs of near-homophone target words embedded in both nonsense and make-sense sentences. Each lecturer was assigned twelve listeners, comprising six listeners familiar with their voices and six unfamiliar, yielding a total of twenty-four listeners. Using these recordings, listeners were instructed to complete the blanks in the sentences with the appropriate target words. Findings revealed a higher prevalence of unintelligible target words (57%) than intelligible ones (43%). The articulation of vowels and consonants emerged as a key factor influencing reduced intelligibility, with problematic vowels and consonants that affected the intelligibility of the lecturers’ pronunciation features. Additionally, the pronunciation of target words and their near-homophone pairs affected the intelligibility of the lecturers’ speech. Familiarity with the lecturers’ voices enhanced performance than those without, with only 42% of in-group listeners finding the words unintelligible, compared to outgroup at 58%. While syntactic cues provided some benefit to the listeners, those unfamiliar with the lecturers’ voices struggled to complete the task. Word frequency also affected the lecturers' pronunciation features. However, some highfrequency words were not perceived, while some low-frequency words were. The Mann-Whitney test results indicated that the lecturers’ in-group listeners performed significantly better outcomes in achieving target words (R1) and reducing non-achieved words (R3), although no significant differences were found between the groups in near-homophones achieved (R2). In the second task, lexical frequency and syntactic cues were found to affect the lecturers’ pronunciation features, whereas familiarity showed no effect. Two male lecturers (LC and LD) recorded homographic words functioning as nouns or verbs within forty sentences. Twenty-four listeners identified word stress placement to distinguish word classes. The results indicated no statistically significant difference between in-group (twelve listeners) and out-group (twelve listeners), as confirmed by the Mann- Whitney tests (p-values = 0.855 and 0.405, greater than 0.05). Both groups performed similarly, despite a slightly higher mean rank for LD’s in-group (7.42 vs. 5.58). Lexical frequency of words affected task performance. Listeners could perceive words that have a higher frequency of usage in the corpora NOW, though some of them could not. Syntactic structure of the verb provided as a cue to identify word stress placement, distinguishing verbs from nouns. The implications suggest that while both two features contribute to clarity and fluency, segmental features play a more prominent role. While native-like pronunciation persists in EFL contexts and pronunciation teaching, effective communication prioritizes intelligibility over accent perfection. Therefore, it is implied that pronunciation features can improve communication with intelligibility as the goal. |