Investigating the Comprehensibility Paradigm: Key Fluency Predictors in L1 English Speakers’ Evaluations of Japanese Speakers’ L2 English

Abstract

The present study investigates the relationship between first language (L1) and second language (L2) oral fluency and their association with comprehensibility in lower-intermediate Japanese learners of English, examining which L2 fluency features are most closely associated with L1 English listeners’ comprehensibility ratings after accounting for individual differences in L1 fluency. Thirty participants completed parallel speaking tasks in their L1 (Japanese) and L2 (English), and speech samples were analysed for eight established utterance fluency measures encompassing speed, breakdown, and repair fluency. Inexperienced L1 English listeners rated comprehensibility, the perceived ease or difficulty of understanding a speech sample, on a 1,000-point continuous scale. Bootstrapped Least Absolute Shrinkage and Selection Operator regression and dominance analysis identified filled pause ratio, silent pauses within clauses, and articulation rate as the strongest L2 predictors of comprehensibility, together explaining nearly half the variance in listener ratings. Correlational analyses revealed moderate cross-linguistic links for pausing ( r = .46–.51) and repair behaviours ( r = .63–.80), suggesting that certain fluency traits, particularly repair behaviours, show cross-linguistic consistency. However, when both L1 and L2 fluency measures were included in regression models, only L2 features remained significant predictors, suggesting that listener judgements are shaped primarily by L2-specific speech behaviours rather than L1 speaking styles. These findings clarify the relationship between language-general and language-specific fluency traits, indicating that comprehensibility is most strongly linked to how fluently learners speak in the L2 itself, and illustrate the value of regularised regression techniques for identifying core fluency predictors in small-sample research.

The authors' abstract, as published at the source. Language and Speech, 2026 · DOI ↗