Abstract Bilingualism researchers have intensively studied how learning and using multiple languages affects all levels of linguistic structure. In this strand, examining diversity in the bilingual experience and the extent to which variables like language dominance regulate crosslinguistic interaction has been of special interest. However, most studies sample small groups of bilinguals from a single research site, creating a twofold generalizability problem. First, with small samples it is unlikely that researchers will be able to fully capture and quantify the range of variables known to affect findings. Second, when bilinguals are recruited from a single site, it is impossible to determine if findings are site-specific or apply to bilinguals more broadly. To address these issues, we propose a large(r)-scale, multisite approach to bilingualism research. We believe that such an approach, when informed by open science practices, has the potential to significantly advance the state of the art.

Abstract Orthography has well-documented effects on L2 pronunciation accuracy. This study investigates the effects of orthographic input on the devoicing of L2 German learners’ production of voiced stops /b, d, g/ in final position. Additionally, we investigate the interaction of cognate status and orthographic input. Thirty L2 German students completed read-aloud, picture-naming, and delayed-repetition tasks using target words classified as identical, near, and noncognates. Analysis of voicing using stop closure duration and the amount of glottal pulse during stop closure showed orthography interfered with pronunciation accuracy. Cognate status resulted in more target-like pronunciation only for noncognates in the read-aloud and delayed-repetition tasks. The results confirm earlier findings that orthographic input in L1 English-L2 German matching leads to less accurate pronunciation and that identical cognates and near cognates suffer most from orthographic effects.

Abstract In this paper, we discuss a set of 60 listening quizzes, created to support adult learners of Hul’q’umi’num’ (Coast Salish) in fine-tuning their listening and speaking skills. Hul’q’umi’num’ has a rich consonant inventory, including many sounds not found in learners’ L1 (English). The goal of the quizzes was twofold: provide learners with opportunities to practice hearing these sounds and, at the same time, inform us about the features of Hul’q’umi’num’ L2 speech perception. Findings showed which sounds were particularly easy or challenging, laying the foundation for creating more targeted resources to better aid sound acquisition among Hul’q’umi’num’ learners. Evidence of improvement in perceptual ability after taking the quizzes was also found. This work contributes to diversifying scientific approaches to second language acquisition by showing how speech perception research and pedagogy can be combined in an Indigenous language revitalization context.

Abstract This study investigates the intelligibility and comprehensibility of French-accented speech in an academic context. L1 French and L1 English listeners heard speech samples in three different accent conditions: a marked French accent, an unmarked French accent and a Southern British English (SBE) accent. They were asked to perform two word recognition tasks, a speech comprehension task and provided subjective ratings of certainty, comprehensibility, cognitive load and accentedness. Results showed that for English listeners sharing the same first language (L1) had a facilitating effect, whereas varying the levels of French-accentedness had a detrimental effect. French listeners, however, did not find French-accented speech significantly more intelligible and comprehensible than SBE-accented speech. These findings deepen our knowledge of the relationship between intelligibility, comprehensibility, accentedness and cognitive load.

Abstract This study examines the validity of WER as a proxy for pronunciation quality in EFL contexts. Human ratings of comprehensibility and accentedness were compared with WER and automated pronunciation scores from six ASR systems — Kaldi, wav2vec 2.0, HuBERT, Whisper (Base and Large-v3), and Microsoft Azure — using 190 read-aloud recordings by Korean elementary learners. With respect to pronunciation scoring, Azure’s phoneme-level accuracy scores demonstrated moderate correlations with human judgments, while Kaldi’s GOP scores showed no meaningful association. Analysis of WER revealed a critical trade-off between ASR accuracy and perceptual sensitivity: high-performing systems such as Whisper Large-v3 and Azure produced near-zero WERs but weakly correlated with human ratings. In contrast, mid-performing systems such as Whisper Base and HuBERT showed stronger correlations, indicating that moderate WER values may better reflect pronunciation variation. These results underscore the limitations of WER in advanced ASR systems and the need for perceptually grounded, interpretable metrics.