Journal of Phonetics最新文献

Previous studies have examined whether speakers initiate longer utterances with higher F0. Evidence for such effects is mixed and is mostly based on point estimates of F0 at the beginning of the utterance. Moreover, it is unknown whether utterance length can influence F0 control solely at utterance onset or also during the utterance. We conducted a sentence production task to investigate how control of pitch register – F0 ceiling, floor, and span – is influenced by utterance length. Specifically, we test whether speakers adjust register both in relation to an initially planned utterance length – proactive F0 control – and in response to changes in utterance length that occur after response onset – reactive F0 control. Target sentences in the experiment had one, two, or three subject noun phrases, which were cued with visual stimuli. An experimental manipulation was tested in which some visual stimuli were delayed until after participants initiated the utterance. Evidence for both proactive and reactive control of register was observed. Participants adopted a higher register ceiling and broader span in longer utterances. Furthermore, they decreased the amount of ceiling compression upon encountering delayed stimuli. The findings suggest the presence of a mechanism in which speakers continuously estimate the remaining length of the utterance and use that information to adjust pitch register.

Recent research showed selectivity of perceptual learning in speech to linguistic variables and non-linguistic variables. With regard to the latter Keetels et al. (2016) reported that perceptual learning for one spatial location does not fully generalize to another. This spatial selectivity has been suggested to indicate that learning may target non-linguistic representations. We test whether spatial selectivity is a general property of perceptual learning or whether it is related to specific design choices, such as using a single nonword throughout the study. Therefore, we aimed to replicate spatial selectivity with a paradigm that makes use of a larger set of word and nonword stimuli. However, in three experiments, one in-person and two web-based, no effect of spatial selectivity was observed. A Bayesian analysis suggests that the null hypothesis is better supported by the data than the alternative hypothesis based on the previously reported effect size. Repercussions for the debate about pre-lexical representations in speech processing are discussed.

Relatively little research has been conducted on the effect of hearing an unfamiliar native English accent. This paper tests listeners with varying levels of familiarity with the Glaswegian linguistic environment, presenting them with naturalistic minimal pairs such as hut/hurt – produced by speakers raised in Glasgow – in two-alternative-forced-choice tasks. The results of Experiment 1 show a benefit of long-term familiarity in discriminating minimal pairs with derhoticised /r/, a phonetically eroded form of postvocalic /r/ in working class Glaswegian. Native Glaswegian listeners displayed high sensitivity to difference (d’), and low response bias (c) towards hearing either rhotic or non-rhotic words, indicating accurate perception. Unfamiliar listeners were less sensitive to stimulus difference, and were biased towards hearing plain vowels, demonstrating their unfamiliarity with Glaswegian /r/. Non-rhotic English listeners with a moderate level of experience with Glaswegian showed an effect of ‘perceptual hypercorrection’, i.e. over-reporting /r/ presence. Experiment 2 found that, following a short period of exposure, English listeners with very little experience with Glaswegian also started to show hypercorrection, suggesting rapid adaptation to novel phonetic detail. These results may be explained by some general principles underlying exemplar and hybrid theories, and contribute to the ongoing research into the complex nature of Scottish /r/.

This study investigates cross-linguistic influence of a second language (L2) on the first language (L1) perception in two distinct bilingual populations: Korean heritage speakers (n = 30) and long-term immigrants (n = 26) in the USA, compared to Korean-immersed speakers in South Korea (n = 30). By leveraging the differences in language-specific cue primacy in stop consonants between Korean and English, the present study examined L2 (English) influence in participants’ perceptual cue weighting and discrimination of Korean laryngeal categories through a three-alternative forced-choice identification task and a speeded AX discrimination paradigm. The results indicated divergent patterns of cross-linguistic influence for the two bilingual groups. While heritage speakers showed a decreased reliance on the onset f0 cue, suggesting an assimilatory effect of English, long-term immigrants relied more heavily on this cue than Korean-immersed speakers, suggesting dissimilation with English. Furthermore, in discriminating Korean stops based on f0 differences, heritage speakers demonstrated decreased accuracy while long-term immigrants outperformed Korean-immersed speakers. In addition, individual weighting of f0 in the identification task was found to be predictive of discriminatory performance. Overall, the findings underscore the importance of L1 input and experience in determining the nature of cross-linguistic influence in L1 speech perception.

Phonetic reduction arises in the course of typical language production, when language users produce a less clearly articulated form of a word. An important factor that affects phonetic reduction is the predictability of the information conveyed: predictable information is reduced. This can be observed in everyday use of reference in spoken language. Specifically, first mention of a referent is more carefully articulated than subsequent mentions of the same referents, which are often phonetically reduced. Here we ask whether phonetic reduction for predictable information exists in a young sign language, and, in particular, how phonetic reduction is realized in visual languages that exploit various articulators of the body: the hands, the head, and the torso. The only natural languages that we can observe as they emerge in real time are young sign languages, and we focus on one of these in the current study: Israeli Sign Language (ISL). We use 3D motion-capture technology to measure phonetic reduction in signers of ISL by comparing the production of referring expressions synchronically, at different points during a narrative (e.g., Introduction, Reintroduction, Maintenance). Our findings show: (a) that phonetic reduction is present in a young sign language; and specifically (b) that the actions of different articulators involved in discourse are reduced, based on predictability. We consider the importance of these findings in understanding predictability in language more generally.

The present study tests a prediction from the prevalent Autosegmental-Metrical (AM) model of American English intonation: the existence of distinct phonological contrasts among nuclear tunes composed of a pitch accent (here H*, L+H*, L*+H), phrase accent (H-, L-) and boundary tone (H%, L%), which in combination yield an inventory of 12 tonally distinct nuclear tunes. Using an imitative speech production paradigm and AX discrimination task with L1 speakers of Mainstream American English (MAE) we test the extent to which each of 12 predicted tunes is distinct from the others in the production and perception of intonation. We tackle this question with a series of analytical methods. We use GAMM modeling of time-series F0 trajectories to test for differences among all of the twelve nuclear tunes, and compare these results to a method that does not rely on pre-defined tune categories, k-means clustering for time-series data, to discover emergent classes of tunes in a “bottom-up” fashion. We complement these time-series analyses with an analysis of the temporal tonal center of gravity (TCoG) over the F0 trajectories of nuclear tunes to assess tonal timing distinctions and their relation to top-down tune classes (defined by the AM model) and bottom-up classes (emergent from clustering). Production results are further compared to perceptual discrimination responses, which together point to a hierarchy of distinctions among nuclear tunes: a set of primary tune distinctions are emergent in clustering and always distinct in perception. Other tune distinctions, although evident in top-down analyses of (labeled) F0 trajectories, are lost in emergent clusters, limited in magnitude and scope, and often confused in perception. Results are discussed in terms of implications for a theory of intonational phonology.

Listeners often make use of suprasegmental features to compute a prosodic structure and thereby infer an information structure. In this study, we ask whether listeners also use segmental details as a cue to the prosodic structure (and thus also the information structure) of an utterance. To this end, we examined the effects of segmental variation of German auxiliary haben (‘to have’)—i.e., hyperarticulated [habən], moderately reduced [habm], and strongly reduced [ham]. Three remotely accessed online mouse-tracking experiments were carried out by adapting the lab-based experimental paradigms used in Roettger and Franke (2019). They showed effects of pitch accent on the auxiliary haben, leading to the interpretation of an affirmative answer to a preceding question, thus anticipating an upcoming referent noun to be the same as the one given in the question (i.e., the verum focus effect). Experiment 1 adapted the design Roettger and Franke (2019) to an online setting. In Experiment 2, listeners were indeed found to make use of the segmental detail of the auxiliary haben, even in the absence of f0 (pitch accent) information—i.e., the hyperarticulated (full) form showed an effect similar to the pitch accented form, albeit smaller. In Experiment 3, we confirmed that the observed segmental effects were not simply due to learning that might have taken place during the experiment. Our results thus imply that the analysis of prosodic structure, which is often assumed to occur in parallel with the segmental analysis, must integrate segmental details that help to signal the prosodic structure.

This study explores the ways in which marginal contrasts are clearly realized by merged speakers as a result of exposure to a distinct speaker. The alveolar-retroflex sibilants in Taiwan Mandarin (TM) were chosen as a test case as the distinction is variably implemented, ranging from a complete merger to clear contrasts. In a spontaneous phonetic imitation task, merged and distinct TM speakers imitated the speech of a distinct model talker of Mainland Mandarin. The results show that merged speakers substantially increased the spectral distance between the two sibilants during imitation, essentially reversing the merger. Specific patterns of the merger reversal were further enriched by phonetic and social factors. The categories were unmerged by making reference to each individual's phonetic space; speakers with higher baseline spectral frequencies restored the underlying retroflex category, and those with lower baselines retrieved the alveolars through dentalization. Furthermore, the merger reversal showed preferential convergence conditioned by gender—female speakers were less willing to accommodate socially undesirable strong retroflexion. Taken together, phonetic imitation reflected speakers’ abstract phonological knowledge, but specific patterns were also shaped by a careful calibration of phonetic space and the desired phonetic norms of the speech community.

Context cues are useful for listeners to normalize speech variability and achieve perceptual constancy. It remains unknown whether this normalization strategy is language-independent and can be generalized directly from the perception of first language (L1) to second language (L2). To answer this question, Experiment 1 in the present study asked Mandarin learners of Cantonese to perceive ambiguous Cantonese tones with context cues. The results revealed a significant Cantonese-tone normalization process in Mandarin learners, but the effect size was smaller than native speakers, suggesting that speech normalization required language-specific knowledge and thus it was refined gradually during L2 acquisition. The results also showed that even with effective context cues, Mandarin learners tended to give more high level tone responses, a tone also in Mandarin, implying that L1 phonological system interacts with immediate L2 context during L2 speech normalization. Experiment 2 revealed that L2 immersion but not overall L2 proficiency or L2 phonological proficiency facilitated L2 normalization process, indicating that L2 speech normalization improved with perceptual practice and needed more high-level L2 knowledge than L2 phonology.

Previous studies on whether perceptual learning generalizes across multiple speakers have produced inconsistent results between generalization and speaker-specificity. A prior proposal is that the critical phonemes produced by two different speakers need to be phonetically similar for perceptual learning to generalize. To test this account, we investigated the perceptual generalization of sibilants across two pairs of speakers. In both cases, sibilants of the same male speaker were manipulated to induce either an /s/-favoring perceptual bias or an /ʃ/-favoring one in a training phase. We then examined whether the perceptual biases would generalize in a test phase to the /s/-/ʃ/ continua of two different female speakers, one resembling the training speaker and the other differing from the training speaker in the spectral frequency distributions of their sibilants. We found that generalization of perceptual learning occurred in both /s/-favoring and /ʃ/-favoring conditions between the speaker pair with similar sibilant productions. For the speaker pair with different sibilants, we found perceptual generalization in the /s/-favoring training condition but not in the /ʃ/-favoring condition, which is not predicted by the phonetic similarity account. To explain these unexpected results, we offer a novel phonetics–phonology mismatch account as a refinement of our understanding of when and why perceptual generalization might be blocked. The results shed light on the constant influence of the mapping between phonetics and phonology during the learning and generalization of phonetic variability.