Language and Speech最新文献

Outcomes of early phonological acquisition in multi-accent contexts can be especially wide-ranging, raising the question of whether children exposed to multiple accents in one community are building the same linguistic systems. This present study investigates the English coda clear laterals in the spontaneous, mother-directed speech of English-Malay early bilingual preschoolers raised in multi-accent Singapore. Previous work has shown that these children were exposed to highly variable input involving three different English coda /l/ variants within and outside of their ethnic community. To elucidate the complex nature of language acquisition in such diverse settings, we examine both individual differences and group behaviors. Our findings reveal that despite the considerable between- and within-child variation, production patterns are generally systematic. Malay children with close Chinese peers, however, exhibited greater variability and unpredictability in their production, revealing word-specific inconsistencies that suggest a restructuring of or instability in their phonological representations. This study underscores the complexity of phonological development in multi-accent contexts and highlights the challenges in predicting the contributors of these variable outcomes.

Word complexity indices, such as the Index of Phonetic Complexity (IPC) and the Word Complexity Measure (WCM), code a word in terms of featural and structural properties that pose difficulty in phonological development. Studies have investigated the influence of complexity indices on vocabulary development; however, few have examined their influence on consonant accuracy. Furthermore, these indices were developed for English-speaking children and have not been widely applied to other languages. This study investigates whether a word's phonological complexity influences how accurately it is produced in French-speaking children. Four databases consisting of the productions of children (n = 74), aged 1;11 to 4;9, were analyzed. Words were coded in terms of the IPC, WCM, and parameters that add complexity during phonological development. Using mixed-effects logistic regression, we examined whether phonological complexity as determined by the IPC, WCM, or by alternative indices better accounts for the influence of complexity on production. We also investigated whether the accuracy of a target sound/structure was influenced by a word's complexity. Results indicated that complexity based on the IPC or WCM significantly influenced consonant accuracy; however, indices tapping fewer features provided superior model fit. At younger ages, the presence of fricatives/liquids and, at all ages, the presence of alveopalatal fricatives, codas, and clusters significantly influenced accuracy. Findings were inconclusive as to whether whole word complexity influenced the accuracy of a target sound/structure. Results suggest that current complexity indices provide only approximate indications of how featural and structural properties of words influence production.

This tutorial paper introduces two approaches to modeling tongue contour data obtained with DeepLabCut using multivariate generalized additive models (MGAMs) and multivariate functional principal component analysis (MFPCA). For each method, we present a fully commented analysis of two illustrative data sets: VC coarticulation in Italian and Polish, and consonant emphaticness in Lebanese Arabic. All the materials (inlcuding data and code) are available in the research compendium of the tutorial at https://github.com/stefanocoretta/mv_uti. We conclude by discussing advantages and disadvantages of the two methods (MGAM and MFPCA) and we recommend researchers to prefer MFPCA over MGAM as an initial step for modeling tongue contours.

The perception of voice onset time (VOT) in word-initial stops has been widely studied in the second language, although it remains underexplored from a multilingual perspective. Previous limited research on multilingual VOT perception offers inconsistent results, with studies suggesting either cross-linguistic interactions or language-specific VOT patterns. This study aims to consolidate these findings by investigating VOT perception in trilingual speakers of L1 Polish, L2 English, and L3 Norwegian, as well as examining whether perceptual boundary locations between voiced and voiceless stops in L3 learners differ from those of native speakers and are influenced by L1, L2, and L3 interactions. To this end, VOT continua were created for three places of articulation (labial, coronal, velar) and three languages (Polish, English, Norwegian). Participants completed a two-alternative forced-choice task separately for each language. The results indicate that trilingual speakers experience cross-linguistic interactions, particularly based on the place of articulation, highlighting the complex nature of VOT perception in multilingual acquisition.

Tone perception has drawn much attention in recent years. However, previous studies have mainly focused on tones with different contours, while systematic investigation of tones with the same contour-particularly falling tones-remains scarce. Given that different tone contours may elicit distinct perceptual patterns, examining tones with the same contour is crucial for understanding how listeners distinguish closely related tonal categories. Taking the four falling tones in Bai (Meiba variety) as an example, this study first demonstrates the acoustic foundation of these tones using acoustic analysis. On this basis, the perceptual experiment reveals that identification of these four falling tones shows gradient boundaries and no sharp discrimination peaks, a pattern often described as continuous perception in previous literature. Moreover, the phonation types of tones also affect the identification score and boundary position. Based on the systematic relationship between phonation types and perceptual measures (i.e., identification score and boundary position), this study proposes an operational criterion to identify whether the phonation types of speech sounds are the same.

This study investigated Japanese, Cantonese, English, and French listeners' ability to perceive non-native quantity contrasts (e.g., short vs. long). In these languages, duration is used to mark phonemic quantity contrasts to different degrees. We had native listeners of these four languages listen to resynthesized pseudo-Japanese and pseudo-Estonian stimuli in AXB and identification tasks. The stimuli contrasted in consonant and vowel quantity. The results showed that while Japanese listeners, who have systematic phonemic quantity contrasts in their L1, generally outperformed other listeners in identification and in discrimination, their identification accuracy for overlong Estonian vowels and consonants was not as high as that for long Estonian vowels and consonants. Meanwhile, French listeners, who have no quantity contrasts in their L1 phonology, did not perform worse than the other groups as predicted. Our findings show that the role of duration in L1 phonology alone is not enough to predict the perception accuracy of non-native quantity contrasts when duration is the only acoustic cue.

This study investigated how social-pragmatic cues modulate causal inference during Chinese discourse comprehension, bridging social cognition and language processing. Two self-paced reading experiments manipulated causal strength (Strong vs. Weak) and respect consistency (Respect "nin" vs. Disrespect "ni") across coherent five-sentence discourses. Experiment 1, using an explicit causal judgment task, found that weak causality elicited longer reading times than strong causality. Notably, this causal strength effect was amplified under respectful address but reduced under disrespectful address. We interpret this finding as suggesting that the cognitive cost of processing a sociopragmatic violation-driven by mechanisms such as attentional diversion, emotional arousal, or pragmatic reanalysis-interferes with causal inference. Experiment 2, using an implicit reading paradigm, showed a persistent causal strength effect without modulation by respectfulness, and no spillover to subsequent sentences. This pattern indicates that sociopragmatic influences on inference may depend on task context and processing goals. Overall, the results suggest that while causal inference can occur automatically, its interaction with social cues appears constrained by cognitive resources and task demands. These findings offer new insights for how social-pragmatic norms interact with cognitive mechanisms in language comprehension, with implications for cross-cultural pragmatics and models of discourse coherence.

The study aims to contribute to the understanding of the role of the putamen-claustrum-insular region (region of interest, ROI) in speech motor control processes, considering functional magnetic resonance imaging (fMRI) and acoustic analysis. Forty-three neurotypical participants performed visually presented motor planning and execution tasks (two highly frequent non-word CV-syllable sequences), which differ in gestural complexity. First, it can be shown that the ROI generally plays a role during motor planning and motor execution. Second, comparing blood oxygen level-dependent (BOLD) responses across the complexity conditions revealed that syllables with greater gestural complexity elicited increased activation in the bilateral ROI. This region may help to coordinate articulatory movements in articulatory subsystems. The intricacy of these glottal and supraglottal articulatory movements defines the gestural complexity of the respective speech sound. Finally, our results suggest that these differences in BOLD response in the ROI reflect the differential contribution of articulatory mechanisms that are required to produce phonologically distinct speech sounds.

Previous research has shown that tonal second language (L2) experience can enhance the third language (L3) tone perception of native speakers of non-tone languages. However, tone is a multidimensional concept, and phonation type also serves as an important cue in the tone perception of some tone languages. In this study, 38 native English speakers of different Mandarin proficiency levels, 15 native English and 15 native Mandarin speakers, all of whom were naïve to the Wenzhou Wu dialect, participated in an AX discrimination task. The Wenzhou dialect was used for the naïve L3 stimuli because of its breathy voice feature in the low-register tones. Mandarin Tone 3 (T3) is often realized with creaky voice, while in English, creaky voice is a prosodic and sociophonetic marker. We asked whether the exposure to one phonation type (creaky voice) at a different linguistic level (indexical in English, allophonic in Mandarin) could lead to better performance on another phonation type (breathy voice). Our results showed that in addition to its effect on pitch perception, Mandarin-learning experience was associated with higher accuracy in phonation perception for native English speakers, and high-level L2 learners could even outperform native Mandarin speakers. Longer length of Mandarin-learning and Mandarin-immersion experience had a facilitative effect on naïve L3 tone perception, resulting in higher accuracy in the perception of both pitch and phonation to varying degrees. Moreover, the study demonstrated that acoustic similarity significantly affects pitch perception in the initial stages of L3 processing, with acoustically similar tone pairs posing greater perceptual difficulty than dissimilar pairs.

The tip-of-the-tongue (TOT) phenomenon is traditionally considered a retrieval failure; however, it may also reflect cognitive advantages in semantic memory. Semantic memory refers to the repository of general world knowledge, including word meanings, objects, people, and their relationships, and is underpinned by a network of multiple brain regions. Prior studies on the relationship between semantic memory and TOT occurrences have yielded inconsistent results, warranting further investigation. Here, we examined this relationship by applying graph theory to analyze the topological properties of the semantic memory network, using diffusion magnetic resonance imaging (dMRI) data from the Cambridge Centre for Ageing and Neuroscience (Cam-CAN) repository (N = 576; age: 18-87). Our results demonstrated that global metrics of the semantic memory structural network-global efficiency, local efficiency, mean degree centrality, and mean clustering coefficient-positively predicted TOT rates. At the nodal level, increased metrics in the right posterior lateral temporal cortex (pLTC) and left dorsomedial prefrontal cortex (dmPFC) were associated with higher TOT rates. These findings suggest that a more extensive semantic memory, characterized by a richer network of general knowledge, may hinder word retrieval.