Kotaro Takizawa

To contribute to the ongoing discussion on the role of pitch processing in grammar learning, this study examines the extent to which pitch statistical learning (pitch statistical learning (SL); the ability to detect and internalize pitch patterns in auditory input) affects second language (L2) morphosyntax learning outcomes. In the context of 93 Japanese learners of English, piece-wise regression analyses were conducted to compare the roles of their pitch SL abilities and pitch acuity in L2 morphosyntactic knowledge. The results revealed a weak but significant positive correlation between pitch SL and morphosyntactic knowledge, whereas pitch acuity showed no significant association. Further analysis identified a threshold effect: pitch SL has a strong association with morphosyntactic knowledge for learners with lower pitch SL abilities, but not for those with higher pitch SL abilities. These findings suggest that the lack of pitch SL ability could substantially slow down L2 morphosyntactic learning.

Given the critical role of phonological vocabulary in second language (L2) listening proficiency, scholars have increasingly focused on measuring not only the form–meaning mapping aspects of vocabulary knowledge but also the extent to which learners can access this knowledge accurately and promptly in real-time contexts. To assess the automatized dimension of vocabulary knowledge, we built upon prior efforts to develop the lexicosemantic judgment task (LJT). Study 1 successfully replicated previous findings, showing that LJT outcomes could be distinguished from measures of declarative vocabulary knowledge (meaning recognition) and that LJT exhibited relatively strong associations with general L2 listening proficiency, even when controlling for other influencing factors. To further refine the construct validity of LJT as a measure of automatized L2 knowledge, Studies 2 and 3 developed and tested a timed version of the LJT. The results of the timed LJT, compared to the untimed version in Study 1, revealed similar links with listening proficiency, suggesting that the effect of time pressure on performance may be relatively minor. The test materials are provided in various formats to facilitate future research and pedagogical applications.

The present study explored anxiety and its fluctuation during an artificial intelligence (AI)-based English conversation test as a test validation study. Applying an idiodynamic approach, 13 Japanese university students learning English as a foreign language (EFL) took the test and participated in pre- and post-test interviews, which were analyzed by conducting reflexive thematic analysis. Individual differences in the pattern of anxiety experience fluctuations were observed, while factors affecting their anxiety experiences were also explored. These were classified into either authentic factors observed in both testing and non-testing contexts (e.g., perceived English proficiency, perceived communication skills, a two-way relationship between cognition and anxiety, and features of conversation) or factors unique to the test (evaluative situation and AI’s unnatural behavior). From the perspective of test validation, this study emphasized the importance of paying attention to construct-irrelevant variance, such as anxiety unnecessarily arising from test-specific factors.

This study investigates the nature of co‐construction in roleplays conducted with human versus AI interlocutors for assessing interactional competence (IC) in L2 English. Seventy‐five university students in Japan completed roleplay tasks with both human tutors and an AI agent. The AI agent is a multimodal dialog system integrated with a large language model (LLM), designed to allow synchronous interaction with the participant through autonomous turn‐taking. Using conversation analysis, 24 interactions were analyzed to investigate how participants managed preference organization, sequence expansion, and turn‐taking. The analysis revealed that the AI‐delivered roleplays elicited some IC‐relevant practices and that participants treated the roleplay as a co‐constructed interaction, responding contingently to the AI's contributions. While the data suggested both human and AI interlocutors maintained mutual understanding, striking differences in turn‐taking practices were observed, including more frequent overlaps and inter‐turn gaps in the AI‐delivered condition. The study concludes that LLM‐integrated multimodal dialog systems, by producing recognizable verbal actions and multimodal signals, have the potential to effectively elicit co‐constructed interactional performances relevant to IC assessment.

Prior studies developed a lexicosemantic judgment task (LJT) to assess automatized phonological vocabulary knowledge, which reflects the accurate, prompt, and stable access to L2 phonological vocabulary knowledge in contexts. Automatized vocabulary knowledge has been shown to strongly predict general listening ability. Shifting the focus on automaticity in speech production, the current study explored the role of automatized vocabulary knowledge as a measure of L2 cognitive fluency predicting L2 utterance fluency (UF) in spontaneous speech. A total of 210 university students took a multiple-choice vocabulary test and the LJT to assess the declarative and automatized aspects of phonological vocabulary knowledge, respectively. UF was measured as articulation rate and mid/end-clause silent pause ratio, elicited through picture narrative and personal opinion tasks, each presenting differential lexical demands. Memory-based cognitive aptitude was also considered. Mixed-effects regression analyses revealed that automatized vocabulary knowledge, rather than the declarative counterpart, underlies fluent speech performance free of undue pauses. Task effects were identified, indicating that automatized vocabulary knowledge is sensitive to the lexical demands of speech tasks.

Despite the preponderance of research on extensive reading (ER), the research field still suffers from a lack of common understanding of what ER is and how it should be implemented. From the perspective of Day and Bamford’s (2002) seminal ten principles for teaching ER, the present methodological synthesis aimed to systematically review and critically appraise the reporting practice of how previous studies implemented ER. A total of 72 primary studies were retrieved and reviewed. The findings suggested that many studies did not explicitly report ER characteristics in their articles, making it difficult to understand how ER was operationalized and practiced. Suggestions for future studies are discussed.

This study explored how second language (L2) speakers’ use of multiword sequences in speech predicted perceived fluency ratings while controlling for their utterance fluency. A total of 102 Japanese speakers of English delivered an argumentative speech, which was analyzed for bigram and trigram measures (frequency, proportion, and mutual information) and utterance fluency measures capturing three subdimensions: speed, breakdown, and repair fluency (Tavakoli & Skehan, 2005). Perceived fluency was assessed by 10 experienced L2 raters. Mixed-effects regression analyses revealed that after establishing the parsimonious model solely by UF predictors (marginal R² = .61), a frequency-based n-gram predictor––bigram proportion––slightly but significantly explained the remaining variance of fluency rating scores (0.8%). The results indicated that multiword sequences in speech had a small but systematic impact on perceived fluency, even controlling for the effects of utterance fluency. This finding contributes to the discussion concerning the role of multiword sequences in fluent speech production.