International intelligibility of English spoken by college students in the Bashu dialect area of China

The Bashu (Sichuan) dialect, spoken in Sichuan, Chongqing, and adjacent regions, differs from Standard Mandarin in syllable structure, tone, lexicon, and grammar, and its unique pronunciation features can negatively transfer to English production, potentially hindering international intelligibility. As English functions as a global lingua franca in trade, culture, and education, understanding intelligibility among Sichuan-Chongqing speakers is important for cross-cultural communication and educational practice. Grounded in World Englishes/ELF perspectives, the study seeks to assess intelligibility for diverse international listeners and to inform teaching models that address phonetic barriers in L2/EFL contexts. Research questions: 1) What is the level of international intelligibility of English among Bashu dialect-speaking college students with intermediate proficiency? 2) How does intelligibility differ among Bashu speakers with minor regional variation (Sichuan vs. Chongqing)? 3) Which factors influence their international intelligibility? 4) What strategies can effectively enhance intelligibility for improved international communication?

Foundational work by Catford (1950) framed intelligibility as communicative success distinct from mere accuracy. Smith and Nelson (1985, 1992) distinguished intelligibility, comprehensibility, and interpretability, emphasizing interactional processes between speaker and listener and proposing a multi-level taxonomy. Four major paradigms in intelligibility research are outlined: psycholinguistic (speech perception/processing), sociolinguistic (social and regional influences; World Englishes), communicative competence (ELF and core phonetic norms among L2/EFL speakers), and educational linguistics (pedagogical applications of intelligibility constructs). Levis (2005) proposed a Speaker–Listener Intelligibility Matrix to map interactions across Inner, Outer, and Expanding Circle users, reflecting a shift from Inner Circle listener dominance toward a pluralistic, global perspective as L2/EFL users outnumber L1 speakers. Studies on Chinese English report mixed results: Munro & Derwing (1995) and later Zhang (2012) and Zhou et al. (2019) indicate that strong accent does not necessarily decrease intelligibility. Reported intelligibility rates vary (e.g., Zhang, 2015: 64.4% “mostly intelligible”; Kirkpatrick et al., 2008: 81% “high intelligibility”), and listener background matters (Crowther et al., 2015; Orikasa, 2016). A gap is noted: prior work often treats a unitary “Chinese accent,” underexploring dialectal variation (e.g., Bashu), motivating the present study.

Design: Mixed-method phonetic experimental study with questionnaires and listening tests. Participants: 80 total—40 speakers (20 Sichuan, 20 Chongqing; undergraduates who had not lived outside the region >3 months; balanced gender 20F/20M) and 40 educated listeners from Inner Circle (e.g., UK, US, Australia, Canada, New Zealand), Outer Circle (e.g., Philippines, India, Malaysia, Singapore), and Expanding Circle (e.g., China, South Korea, Japan), approximately 13–14 per circle. Speakers’ English proficiency: Intermediate (CEFR B1), determined via Oxford Quick Placement Test (OQPT) and corroborated by exam scores. Materials: Test 1—74 target words from Cambridge Pronunciation Dictionary, embedded in the carrier sentence “Please say again.” Word list included minimal pairs and challenging contrasts (e.g., ship/sheep; beer/bear; vest/west; think/thank etc.). Test 2—passage “The Boy Who Cried Wolf” (216 words); for the cloze, 30 content words were omitted to test contextual intelligibility. Recording: Zoom H6 recorder in quiet classrooms/labs, 44.1 kHz sampling rate. Procedure for listeners: (1) Orthographic transcription of the 74 isolated test words; (2) Cloze task to fill 30 omitted words while listening to the passage; (3) Overall intelligibility rating on Likert-9 scale; (4) Qualitative evaluation of pronunciation features affecting recognition/comprehension and suggestions for improvement. Data analysis: SPSS 26.0. Computed descriptive statistics for ratings and test scores. Independent/paired sample t-tests to compare Sichuan vs Chongqing means (two-tailed; df reported as 19 for paired regional comparisons). Pearson correlation between word and passage intelligibility scores. Thematic analysis of listener qualitative feedback to identify segmental and suprasegmental issues and improvement suggestions.

- Overall intelligibility: Mean listener rating 6.2/9 (≈69/100), indicating generally intelligible speech despite noticeable accents (Table 2). Sichuan mean 6.0 (Table 3), Chongqing mean 6.4 (Table 4). - Word intelligibility (74 items): Max correct 61, min 29, mean 48.6 correct (SD 9.42), i.e., 65.7% accuracy (Table 5). Sichuan mean 47 (SD 9.94; Table 6), Chongqing mean 50.3 (SD 8.82; Table 7). Difference not statistically significant (paired t = -0.924, p > 0.05, df = 19). - Passage (cloze, 30 omissions): Max 30, min 16, mean 26.2 (SD 2.94), i.e., 87.3% accuracy (Table 8). Sichuan mean 25.8 (SD 2.69; Table 9), Chongqing mean 26.6 (SD 3.2; Table 10). Difference not statistically significant (paired t = -0.758, p > 0.05, df = 19). - Correlation between word and passage intelligibility: Pearson r = 0.105, p > 0.05 (Table 11), indicating no significant association; context substantially aids comprehension independent of isolated word recognition. - Listener feedback highlighted persistent pronunciation issues affecting intelligibility: vowel length distinctions (/i:/ vs /ɪ/, /ɔ:/ vs /ɒ/, /u:/ vs /ʊ/), vowel quality confusions (/ɔ:/–/ɒ/, /ʊ/–/u:/), voicing contrasts (voiced vs voiceless), nasal codas (/m/ vs /n/, /ŋ/ confusions), liquids and alveolars (/l/–/r/, /l/–/n/), /w/–/v/ confusion, substitution of dental fricatives /θ, ð/ with /s, z/ or /t, d/, consonant cluster deletion/insertion, stress/rhythm, and intonation issues. - Thematic analysis of feedback underscored that accent, lack of context, and prosody can hinder recognition of minimal pairs (e.g., ship/sheep, beer/bear, wet/wait), while clear loudness and contextual cues improve comprehension. - Suggestions from listeners: emphasize long–short vowel contrasts with exaggeration drills and word lists; focus on specific consonant clusters and tongue/lip positioning (e.g., ch/sh, th vs s); develop fluency and intonation with strategic pausing; exposure to and mimicry of L1 models via media; use practical tools (e.g., smartphone TTS/translators) and targeted practice lists. Overall, despite accent, average listeners reported comprehension remained generally adequate.

The study demonstrates that intermediate-level Bashu dialect speakers are generally intelligible to international listeners, addressing RQ1. Slightly higher scores for Chongqing than Sichuan speakers were observed across overall, word, and passage tests, but none reached statistical significance, addressing RQ2 and suggesting minor regional variation within a shared Bashu-influenced accent profile. The strong improvement from isolated word accuracy (65.7%) to contextual passage accuracy (87.3%) highlights the role of semantic/pragmatic context in supporting comprehension, aligning with prior findings on contextual facilitation. Listener qualitative analyses identified specific segmental (vowel length/quality contrasts, fricatives, liquids, nasals, /w/–/v/, voicing, clusters) and suprasegmental (stress, rhythm, intonation) factors that influence intelligibility (RQ3). These results situate within World Englishes/ELF literature showing that accented speech can remain intelligible and that listener background shapes comprehension. The findings underscore pedagogical value in focusing on high-impact phonetic contrasts and prosody rather than native-like accent targets. Recommendations derived from international listener feedback directly address RQ4, offering practical strategies to enhance intelligibility and communicative confidence for Bashu speakers in global interactions.

Using questionnaires and controlled phonetic experiments with 40 intermediate Bashu-dialect speakers and 40 international listeners, the study shows that English produced by Sichuan-Chongqing college students is generally intelligible (mean 6.2/9), with no statistically significant differences between Sichuan and Chongqing groups across word and passage tasks. Context significantly boosts comprehension (87.3% in passage cloze vs 65.7% isolated words) and is uncorrelated with isolated word performance. Key intelligibility challenges include specific vowel length/quality contrasts, dental fricatives, liquid and nasal distinctions, voicing, consonant clusters, and prosody. Practitioner-oriented strategies include prioritizing high-yield phonemic contrasts, explicit articulatory practice (e.g., th vs s; ch/sh), exaggerated long–short vowel training, prosody (stress/rhythm/intonation) development, and increased exposure to L1 models, while focusing on intelligibility over native-like accent. The findings inform targeted pronunciation instruction for dialect-influenced learners and are potentially transferable to other locales where local L1 phonology impacts L2 English. Future research should broaden participant diversity (regions, ages, proficiency levels, educational backgrounds) and expand sample sizes to enhance generalizability and refine pedagogical recommendations.

The sample is relatively small and homogeneous: 40 undergraduate speakers (all intermediate-level, B1) from Sichuan and Chongqing only, and 40 listeners. This limits generalizability across proficiency levels, ages, educational backgrounds, dialects within China, and international listener profiles. Results are based on read speech in controlled tasks (word lists, a single passage), which may not fully represent spontaneous conversational intelligibility. Future work should include larger, more diverse participant pools, additional dialect regions, varied proficiency levels, broader listening panels, and multiple speech genres (e.g., spontaneous dialog).