Our study examines chunking ability in the processing of auditory multi-word expressions (MWEs) in first (L1) and second languages (L2) using a single- and dual-task paradigm. The findings reveal that divided attention caused by dual tasks does not impair L1 speakers’ ability to bind individual words into a complete MWE (i.e., a unitary chunk). In contrast, L2 learners struggle to form complete MWEs under dual tasks, representing them as smaller, multiple chunks in memory. Divided attention also reduces the overall number of chunks recalled for both groups. Additionally, increased language proficiency and repeated practice through training are positively correlated with the formation of larger chunks, whereas greater working memory capacity is associated with the recall of a greater number of chunks. These findings underscore the challenges L2 learners face in acquiring relatively large chunks and suggest that L2 learning improves through gradually binding smaller units into larger chunks over time.

Accurate self-assessment is notoriously difficult for many second language (L2) speakers as they struggle to align self-evaluations of their performance with external assessments by raters or examiners. We investigated whether a brief peer-assessment activity helps L2 speakers align their self-assessment of comprehensibility with the evaluations by external raters. We also explored how speakers’ metacognitive knowledge contributes to their self-assessments. We recorded 40 L2 English-speaking international students completing an academic oral summary task and self-assessing their speech for comprehensibility. Half of the students then performed a brief peer-assessment activity, whereas the other half engaged in a filler task before all students self-assessed their initial performance again. The speech of all students was subsequently evaluated for comprehensibility by 30 external listeners, allowing us to estimate the extent to which the students’ and the external raters’ assessments converged. Whereas engaging in peer-assessment was generally associated for L2 speakers with greater alignment between their self-ratings and external listeners’ evaluations, peer-assessment appeared to mainly benefit L2 speakers with initially good self-assessment skills. Metacognitive knowledge was not associated with greater alignment between self- and other-assessments. We discuss whether and how brief peer- and self-assessment awareness-raising activities can help L2 speakers calibrate self- and other-assessments.

Adults who stutter (AWS) frequently engage in language monitoring to anticipate and manage stuttering. This linguistic monitoring may reallocate cognitive resources, with potential consequences for language production and memory. We investigated whether AWS’ increased monitoring during production imposes dual-task costs that limit encoding benefits, or whether it enhances memory through deeper conceptual engagement. Thirty-two AWS and sixty-four adults who do not stutter (AWNS) completed a referential communication task in which they described or identified pictures with an experimenter. To simulate AWS’ linguistic monitoring, half of the AWNS performed a simultaneous sound avoidance task (AWNS-SA), prohibiting certain word-initial phonemes. After the communication task, participants completed a recognition memory test for past referents. Results showed that AWS performed more similarly to AWNS than to AWNS-SA in both language production and memory, although AWS’ memory declined on a trial-by-trial basis when stuttering occurred. These findings suggest that linguistic monitoring in AWS does not impose substantial dual-task costs overall, but that stuttering moments can transiently disrupt memory encoding. Together, these results highlight the adaptive nature of linguistic monitoring in AWS and contribute to a broader understanding of how it supports language production and memory across AWS and AWNS.

Over the last few decades, linguistic gender-fair forms have become increasingly used by individuals and official institutions. In the French-speaking sphere, this has led to heated discussions among politicians and other stakeholders, some of whom claim that these forms render texts illegible and inaccessible to the general public. However, the processing of gender-fair forms in reading has been the topic of a few empirical studies. In the present paper, we add to this small body of research by reporting results from a pre-registered eye-tracking study where 58 native French-speakers read short texts which included a masculine form (voisins), complete double form (voisines et voisins), or contracted double form (voisin·es). Consistent with previous findings, the complete double forms were not more costly to process. In contrast, contracted double forms led to increased processing costs in intermediate and late stages of processing, but had no effect on the early stages of processing. Our data also indicate that the processing of contracted double forms becomes easier over time, and that it is facilitated by positive attitudes towards gender-fair language. These findings provide important insights that enlighten the current debate and should therefore be considered in the elaboration of official guidelines regarding gender-fair language.

This study investigates how individual differences in perceptual-cognitive aptitude and L2 learning experience shape the acquisition of English word stress in foreign language (FL) classroom settings, addressing gaps in prior research that has primarily focused on L2 segmental learning. A total of 119 Japanese English-as-a-foreign-language students completed a word stress identification task alongside measures of auditory processing, working memory, and L2 learning experience. Their performance was assessed in terms of both accuracy and processing fluency (reaction time and response stability). Results showed that learners with more precise auditory processing, higher working memory capacity, and more frequent L2 use outside the classroom demonstrated greater accuracy in perceiving English word stress. In contrast, classroom-based experience and the length or onset of FL learning had limited impact. Processing fluency was only weakly related to these factors, showing substantial individual variability. The findings suggest that in input-limited FL contexts, successful L2 suprasegmental learning depends heavily on perceptual-cognitive aptitude and opportunities for L2 use beyond the classroom.

This study investigates incremental processing of subject and object relatives in Italian by comparing two experimental paradigms. Previous research has shown that object relatives are more difficult to process than subject relatives, but the source and timing of this object disadvantage remain debated. Two experiments were conducted using the same linguistic stimuli. Experiment 1 employed a novel mixed design: a lexical maze task for the relative clause combined with a self-paced reading task for the matrix clause. Experiment 2 used a self-paced reading paradigm. Results confirmed a disadvantage for object relatives across both tasks. Critically, the modified maze task revealed that this difficulty emerges immediately after the complementizer and persists in later regions, a finding that was not evident in the self-paced reading task or in previous maze task studies. This shows the greater temporal resolution of the maze paradigm for identifying the locus of syntactic difficulty in incremental processing. Although frequency-based surprisal and memory-based integration accounts explain portions of data, we consider two approaches that can offer a unified account for both early and late effects: an extension of featural relativized minimality and the cue-based retrieval account. We conclude by indicating how future research can disentangle the two accounts.

We examined how first (L1) and second language (L2) English speakers interpret thematic roles when semantic cues conflict with syntactic cues in discourse contexts. We tested three groups: monolingual English speakers, Taiwanese Mandarin-speaking and Japanese-speaking L2 learners of English. Participants read written vignettes containing target sentences with by-phrases (e.g., Jerry was robbed by the thief). In these sentences, the embedded by-phrase noun was either semantically congruent or incongruent with the previous and subsequent discourse context. An offline comprehension task measured interpretations of the agent. Following predictions from the Competition Model and the Shallow Structure Hypothesis, we tested whether L2 speakers would rely more heavily on semantic cues as a general rule, or whether differences in how L1–L2 structures differ within and between participants would influence comprehension of the agents. Results from Bayesian multilevel regression models indicated that L2 speakers were more likely to choose agents based on semantic cues when compared to L1 speakers. No strong effects of L1–L2 cue similarity were obtained. Our results lend new support to the view that L2 speakers rely more heavily on discourse-level semantic cues than L1 speakers, highlighting the need for further research on cue competition in discourse contexts.

Whether listening to different talkers improves or impedes word identification has important implications for theory and practice. Yet, past research on children with hearing devices shows discrepant findings. This study tested 22 children with typical hearing (mean 5;0) and 20 with hearing devices (mean 4;11) on a remote, online 4-alternative forced-choice task (with a 4-picture display) delivered on iPads, with blocks containing 1 vs. 6 different talkers. All words were familiar to young children and were minimal pairs contrasting in voicing and place of articulation in the word-initial and word-final positions. Word identification was worse for place contrasts occurring word finally when listening to different talkers, but no effect was found for voicing contrasts. A consistent position effect was also found, where word identification was poorer across all word-final contrasts. However, no group differences were detected. These results suggest that even when listening to familiar words in good listening environments, the word-final position remains vulnerable to word misidentification, which can be further impeded by listening to different talkers. These effects impact children with and without hearing devices to a similar degree.

While prior research has established how people read in non-interactive media, little is known about the reading process in interactive multimedia such as video games. In this exploratory eye-tracking study, two levels of reading demand (high vs. low) were created during gameplay. Ninety-eight participants were randomly assigned to play a video game with either an English (L1; low reading demand) or an unintelligible foreign-language (FL; high reading demand) soundtrack. At the subtitle level, the FL group (vs. the L1 group) had higher dwell time percentages, more fixations, higher regression rates, and longer mean fixation durations. No group differences were found in saccade lengths. At the word level, the FL group skipped fewer words. An interaction was found between reading demand and word frequency, where the magnitude of skipping lower-frequency words (vs. higher-frequency words) was smaller in the FL group. Gaze duration showed a significant word frequency effect only. The FL group had longer total fixation time on words, and no interaction was found. These results showed that the FL group (vs. the L1 group) experienced greater processing efforts, as reflected in increased fixation-based measures and regressions. The current study provided empirical insights into task-driven reading in interactive multimedia.

The current study examines the effect of talker identity and linguistic experience on the perception of novel speech patterns by English speakers, focusing on vowel insertion in Korean-accented English. Experiment 1 shows that English speakers with no experience of living in Korea identify English words with vowel insertion as valid words more frequently throughout the experiment only when the talker is described as Korean, but not when the talker is American or Mexican. In Experiment 2, we find similar results with English speakers living in Korea, who provide more word responses to vowel-inserted English words in the Korean talker condition but not the American talker condition. Comparing Experiment 1 and Experiment 2, participants living in Korea show a greater preference for the inserted vowel that is similar to the one found in Korean-accented English ([ʊ]) over the control vowel ([ɪ]), as well as faster adaptation to this type. These results suggest that both talker identity and previous exposure to an accent influence how listeners perceive and adapt to foreign-accented speech, consistent with exemplar models of speech perception.

In typical development, conventional metaphors are supposed to be stored as related senses within a single lexical entry, unlike homonyms, whose meanings are represented in separate entries. Autistic individuals often face challenges in understanding metaphors, raising the possibility that they process conventional metaphors more like homonyms—as unrelated meanings. In this study, we tested this hypothesis by comparing autistic and non-autistic adults on a lexical decision task involving both homonyms and conventional metaphors. We predicted that autistic participants would show inhibition effects (slower access) for both subordinate homonym meanings and metaphorical senses, while non-autistic participants would show inhibition only for homonyms. Our results partially confirmed these predictions. Non-autistic participants exhibited inhibition for both homonyms and conventional metaphors, suggesting that accessing metaphorical senses is more effortful than previously assumed. In autistic participants, metaphorical senses were even more difficult to access than subordinate homonym meanings and more difficult than for non-autistic participants. These findings indicate that autistic individuals experience particularly strong inhibition from the literal meaning when processing conventional metaphors, suggesting that these metaphorical senses may not be fully integrated as related senses in their mental lexicon.

Although children with cochlear implants (CIs) have limited access to pitch information due to the suboptimal device transmission, durational cues are relatively well preserved, allowing for the acquisition of prosodic cues needed for communication. Recent findings show that Mandarin-speaking preschoolers with CIs can produce prosodic cues (e.g., duration and pitch) to disambiguate noun-noun compounds (e.g., xiong-mao “panda”) and lists (e.g., xiong, mao “bear, cat”), with those implanted early (before age 2) demonstrating production patterns similar to their typical hearing (TH) peers. This then raises questions about these children’s ability to perceive prosodic cues, and if early implantation again enhances their performance. These questions were investigated using a two-alternative forced-choice task with 57 Mandarin-speaking preschoolers with CIs and 66 TH peers. The results show that all preschoolers can perceive the prosodic cues needed to identify compounds but not lists, suggesting that, like English, the mapping between prosodic cues and postlexical meaning is also acquired late in children learning a tonal language. In terms of the effect of CIs, those implanted before age 2 performed as well as their TH peers. These findings suggest that preschoolers may rely more on other linguistic information rather than prosodic cues when comprehending compounds and lists, offering cross-linguistic evidence for this tendency. Furthermore, interventions for preschoolers with CIs should support the mapping of prosodic cues to discourse functions rather than just vocabulary training, improving daily communicative abilities.

This study explores whether the ability to process grammatical evidentiality is compromised in older adults speaking Turkish and Korean, two languages that grammatically encode evidentiality. Building on previous research that suggests cognitive demands associated with language structures may reduce processing capacity in older adults, we conducted self-paced reading experiments using sentence contexts involving grammatical evidentials. We tested adult groups of young (N = 44, ages 19–27) and older (N = 37, ages 48–70) speakers of Korean and young (N = 31, ages 18–31) and older (N = 42, ages 50–85) speakers of Turkish. The results indicate that both language groups rated mismatched evidential verb forms as unacceptable, with Turkish speakers more likely to interpret mismatches as acceptable than Korean speakers. Notably, older Korean adults exhibited longer reading times (RTs) for direct evidential mismatches, while older Turkish adults showed longer RTs for indirect evidential verbs, suggesting age-related disruptions in processing. The findings only partially support the hypothesis that predicts grammatical processing differences in older compared to younger adults.

The current study investigated two heuristic processing strategies, the agent-first strategy and an animacy-based strategy, in visual world eye-tracking data as well as sentence final interpretations of wh-questions in adolescent L1 German learners of English in both their L1 and their L2. We observed differences between online and offline measures, as well as L1-L2 differences, both in the selection and the time course of application of the heuristics. In L1 German, heuristics were visible only in online data, and the dominant heuristic was animacy-based. In L2 English, the animacy-based heuristic was applied later and to a lesser degree than the agent-first heuristic. The results speak against a direct transfer of heuristic strategies from the L1 to the L2. Instead, we suggest that low-proficiency learners may not have the capacity to use several heuristics at once, and may thus prioritize the agent-first strategy due to its broad domain of application.

This study investigates how lexical, phrasal, and contrastive stress are acoustically realized in American English, focusing on whether men and women differ in how they use pitch, amplitude, and duration to convey stress. Thirty-six native speakers completed minimal-pair stress production tasks online. We analyzed the resulting speech using prosodic contour measures, Bayesian ANOVAs, mixed-effects regression, Random Forest Classification, and human coder judgments. Results show greater acoustic overlap between lexical and contrastive stress than between either of those and phrasal stress. Duration was the primary cue for phrasal stress, while lexical and contrastive stress relied more evenly on multiple cues. Gender-based differences were especially evident in contrastive stress, which, to our knowledge, has not previously been studied in relation to gender: women relied more on pitch, while men emphasized amplitude and duration. These findings highlight the multidimensional acoustic nature of stress realization and demonstrate the value of combining computational and perceptual approaches in prosody research.

Many people experience workplace-related biases because of how they speak, which stems from listener processes including social categorization (placing speakers into groups), stereotyping (forming simplified beliefs about speakers), and processing (dis)fluency (struggling to understand speakers). However, it is unclear how these processes account for evaluations of speakers with intersecting, voice-cued identities across different job contexts. We recruited 192 listeners to assess the employability of men whose speech marks them as first (L1) or second (L2) language speakers and as gay- or straight-sounding men. The speakers were presented as applicants for jobs considered gay- or straight-typed and involving high or low communication demands. Besides employability, listeners evaluated speakers’ sexual orientation and ease of understanding (comprehensibility as processing fluency). Straight-sounding L1 speakers received the highest employability ratings, followed by straight-sounding L2 speakers and gay-sounding L1 speakers, with gay-sounding L2 speakers ranked lowest. Processing fluency mediated the effect of language status, with L1 speakers rated as more employable partly because they were more comprehensible. Job communication demands (but not job stereotypicality) interacted with speaker effects, where L1 and straight-sounding speakers were perceived as more employable in low-communication jobs. We discuss how speaker identity, job context, and listener experience shape evaluations.