Fluency is an essential aspect of second language (L2) oral proficiency. Recent studies have demonstrated that L1 individual speaking style is connected to L2 fluency, suggesting that L2 speech fluency does not solely represent L2-specific skills. Furthermore, task mode (monologue vs. dialogue) has been shown to influence fluency. The present study examines the extent to which these two factors (L1 speaking style and task mode) can predict L2 speech fluency, and how such connections are modified by the learners’ L2 proficiency level. The data consist of monologic and dialogic speech samples from 50 advanced students of English in their L1 (Finnish) and L2 (English). The samples were analyzed for speed, breakdown, repair, and composite fluency. The results of multiple linear regressions demonstrated high predictive power for speed, breakdown, and composite fluency dimensions, while the model for repair fluency showed weak predictive power. The results have implications for L2 fluency research.

This study synthesized 65 (quasi-)experimental studies published between 2010 and 2024 that examined the use of mobile applications to develop language learners’ vocabulary learning. Bayesian meta-analysis was adopted to assess (1) overall effect size; (2) subgroup analyses (i.e. education level, vocabulary knowledge, aspects of vocabulary learning, learning environment, sample size, mobile application type, gender, and cultural background); and (3) publication bias. A large effect size of 1.28 was found for the overall effectiveness of using mobile applications for vocabulary learning when we restricted the studies to long-term treatment duration of 10 weeks or above. Each moderator was analyzed and discussed, and implications for language teaching and research were provided.

This essay explores the Danish concept of hygge, commonly glossed as “coziness,” as a structure of feeling attuned to particular qualities of light. It draws from an ethnographic study of Copenhagen Municipality’s Climate Plan to build the world’s first carbon-neutral capital. Homing in on one of the Climate Plan’s inaugural initiatives—the LED (light-emitting diode) conversion of street lighting—it tracks how ambient intensities of hygge are swept up with both changing lightscapes and changing national demographics. Via a semiotics of social difference, I examine how changing qualities of artificial light are experienced as eroding culturally configured sensory comforts, and how this erosion is grafted onto a fear of the city’s potentially diminishing “Danishness.” This semiotic process is evidenced in the lamination of racialized anxieties about “non-Western immigrants” onto discomforts derived from energy-efficient lighting technologies, and the apparent intrusion of both into habit worlds of hygge. In Copenhagen, I show how a semiotic account of atmosphere illuminates the fault lines of the Danish racial imagination.

Monolingual children tend to assume that a word labels only one object, and this mutual exclusivity supports referent selection and retention of novel words. Bilingual children accept two labels for an object (lexical overlap) for referent selection more than monolingual children, but in these previous studies, information about speakers’ language backgrounds was minimal. We investigated monolingual and bilingual 4-year-old children’s ability to apply mutual exclusivity and lexical overlap flexibly when objects were labelled either by one or two speakers with the same or different language backgrounds. We tested referent selection and retention of word–object mappings. Both language groups performed similarly for mutual exclusivity, were more likely to accept lexical overlap in the two-language than one-language condition, and performance was similar for referent selection and later retention. Monolingual and bilingual children can adapt their word-learning strategies to cope with the demands of different linguistic contexts.

We explored the relationships between L2 utterance fluency and cognitive fluency in monologic and dialogic tasks. The study involved 136 Chinese university-level English learners. Utterance fluency was measured through speed, breakdown, and repair fluency aspects. Cognitive fluency was indicated by L2 lexical and syntactic processing efficiency measures. Stepwise regression models, including metrics of L2-specific cognitive fluency, L2 knowledge, and L1 utterance fluency as predictors, targeted L2 utterance fluency as the dependent variable. We found that L2 cognitive fluency predicted limited variance in utterance fluency, with its influence more evident in monologues. L2 lexical processing efficiency paralleled syntactic processing efficiency’s importance in the monologic task but surpassed it in dialogues. Moreover, L2 processing speed had a more significant impact on utterance fluency than processing stability across both contexts. We suggest that cognitive fluency is not the sole determinant of utterance fluency; L2 knowledge and L1 utterance fluency play non-negligible roles.

From the early use of TF-IDF to the high-dimensional outputs of deep learning, vector space embeddings of text, at a scale ranging from token to document, are at the heart of all machine analysis and generation of text. In this article, we present the first large-scale comparison of a sampling of such techniques on a range of classification tasks on a large corpus of current literature drawn from the well-known Books3 data set. Specifically, we compare TF-IDF, Doc2vec and several Transformer-based embeddings on a variety of text-specific tasks. Using industry-standard BISAC codes as a proxy for genre, we compare embeddings in their ability to preserve information about genre. We further compare these embeddings in their ability to encode inter- and intra-book similarity. All of these comparisons take place at the book “chunk” (1,024 tokens) level. We find Transformer-based (“neural”) embeddings to be best, in the sense of their ability to respect genre and authorship, although almost all embedding techniques produce sensible constructions of a “literary landscape” as embodied by the Books3 corpus. These experiments suggest the possibility of using deep learning embeddings not only for advances in generative AI, but also a potential tool for book discovery and as an aid to various forms of more traditional comparative textual analysis.

This article introduces a strategy for the large-scale corpus analysis of music audio recordings, aimed at identifying long-term trends and testing hypotheses regarding the repertoire represented in a given corpus. Our approach centers on computing evolution curves (ECs), which map style-relevant features, such as musical complexity, onto historical timelines. Unlike traditional approaches that rely on sheet music, we use audio recordings, leveraging their widespread availability and the performance nuances they capture. We also emphasize the benefits of pitch-class features based on deep learning, which improve the robustness and accuracy of tonal complexity measures compared to traditional signal processing methods. Addressing the frequent lack of exact work dates (year of composition) in historical corpora, we propose a heuristic method that aligns works with timelines using composers’ life dates. This method effectively preserves historical trends with minimal deviation compared to using actual work dates, as validated against available metadata from the Carus Audio Corpus, which spans 450 years of choral and sacred music and contains 5,729 tracks with detailed metadata. We demonstrate the utility of our strategy through case studies of this corpus, showing how ECs provide insights into stylistic developments that confirm expectations from musicology, thus highlighting the potential of computational studies in this field. For example, we observe a steady increase in tonal complexity from the Renaissance through the Baroque period, stable complexity levels in the 19th and 20th centuries, and consistently higher complexity in minor-key works compared to major-key works. Our visualizations also reveal that vocal music was more complex than instrumental music in the 18th century, but less complex in the 20th century. Finally, we conduct comparative analyses of individual composers, exploring how historical and biographical contexts may have influenced their works. Our findings highlight the potential of this strategy for computational corpus studies in musicological research.

Despite comments in the ELT literature on the importance of word-stress for comprehensibility in English, there are many places where native speakers of English appear to pay it little attention, showing systematic variation as well as errors. At the very least, there is a paradox here, in that learners are told to get a feature right that native speakers feel free to ignore. More detailed consideration, though, shows that matters are not as simple as this implies. In this paper, several types of stress variation in English are exemplified, and it is also shown that in everyday usage native English speakers are flexible in what they will accept where stress is concerned. This raises questions about the best model for teaching stress in English as a second or foreign language. A simple right/wrong dichotomy is unlikely to reflect native usage.

To better understand language teacher turnover, this study closely replicates and extends McInerney et al.’s (2015) research, which found that teacher commitment predicted turnover intentions to schools (44.2%) and the profession (45.2%) among Hong Kong schoolteachers (N = 1,060). Given the relatively stable employment conditions in that context, the generalizability of these findings to more mobile populations, such as expatriate native English-speaking teachers (NESTs), remains uncertain. In this replication, (1) the population was changed to NESTs in East Asia, and (2) subgroup comparisons were extended to reflect distinctions relevant to the replication sample. Additionally, results were directly compared to the original. A total of 215 NESTs participated. Results showed similar directional patterns but stronger effects: commitment explained 51.8% of variance in turnover intentions to schools and 59.7% to the profession. Affective commitment was the strongest predictor, though NESTs reported lower commitment and higher turnover intentions than in the original study.