Search

This paper explores the minimal knowledge a listener needs tocompensate for phonological assimilation, one kind of phono-logical process responsible for variation in speech. We usedstandard automatic speech recognition models to represent En-glish and French listeners. We found that, first, some typesof models show language-specific assimilation patterns com-parable to those shown by human listeners. Like English lis-teners, when trained on English, the models compensate morefor place assimilation than for voicing assimilation, and likeFrench listeners, the models show the opposite pattern whentrained on French. Second, the models which best predict thehuman pattern use contextually-sensitive acoustic models andlanguage models, which capture allophony and phonotactics,but do not make use of higher-level knowledge of a lexiconor word boundaries. Finally, some models overcompensate forassimilation, showing a (super-human) ability to recover theunderlying form even in the absence of the triggering phono-logical context, pointing to an incomplete neutralization notexploited by human listeners.

Introduction: Learning to understand and speak a new language can be challenging and discouraging for adults. One potential tool for improving learning is transcutaneous auricular vagus nerve stimulation (taVNS), which modulates perception, memory, and attention systems. It has recently been reported that taVNS can improve English speakers' ability to perceive unfamiliar Mandarin tones. The current project explored the potential benefits of taVNS for language learning beyond tone perception. Methods: We studied adults' ability to perceive and produce unfamiliar speech sounds as well as any potential change in language learningmotivation frompre- to post-training. Forty-five native English speakers were divided into three groups and were trained to perceive German sounds: one group received stimulation during easier-to-learn sounds (vowels), one group received stimulation during harder-to-learn sounds (fricatives), and a control group received no stimulation. Results and discussion: We did not find evidence that taVNS improved perception or production of the German sounds, but there was evidence that it did improve some aspects of motivation. Specifically, the group that received taVNS during easier sounds showed a significant decrease in feelings of tension/pressure about language learning, while the other groups did not. Overall, the present study does not find that taVNS holds benefits for the acquisition of new speech sounds; however, the field is nascent, and so the potential applications of taVNS for language learning remain to be clarified. [ABSTRACT FROM AUTHOR]

Purpose: Previous work has found that both young and older adults exhibit a lexical bias in categorizing speech stimuli. In young adults, this has been argued to be an automatic influence of the lexicon on perceptual category boundaries. Older adults exhibit more top-down biases than younger adults, including an increased lexical bias. We investigated the nature of the increased lexical bias using a sensorimotor adaptation task designed to evaluate whether automatic processes drive this bias in older adults. Method: A group of older adults (n = 27) and younger adults (n = 35) participated in an altered auditory feedback production task. Participants produced target words and nonwords under altered feedback that affected the 1st formant of the vowel. There were 2 feedback conditions that affected the lexical status of the target, such that target words were shifted to sound more like nonwords (e.g., less-liss) and target nonwords to sound more like words (e.g., "kess-kiss"). Results: A mixed-effects linear regression was used to investigate the magnitude of compensation to altered auditory feedback between age groups and lexical conditions. Over the course of the experiment, older adults compensated (by shifting their production of 1st formant) more to altered auditory feedback when producing words that were shifted toward nonwords (less-liss) than when producing nonwords that were shifted toward words (kess-kiss). This is in contrast to younger adults who compensated more to nonwords that were shifted toward words compared to words that were shifted toward nonwords. Conclusion: We found no evidence that the increased lexical bias previously observed in older adults is driven by a greater sensitivity to top-down lexical influence on perceptual category boundaries. We suggest the increased lexical bias in older adults is driven by postperceptual processes that arise as a result of age-related cognitive and sensory changes.

Purpose: Heterogeneous child speech was force-aligned to investigate whether (a) manipulating specific parameters could improve alignment accuracy and (b) forced alignment could be used to replicate published results on acoustic characteristics of /s/ production by children. Method: In Part 1, child speech from 2 corpora was force-aligned with a trainable aligner (Prosodylab-Aligner) under different conditions that systematically manipulated input training data and the type of transcription used. Alignment accuracy was determined by comparing hand and automatic alignments as to how often they overlapped (%-Match) and absolute differences in duration and boundary placements. Using mixed-effects regression, accuracy was modeled as a function of alignment conditions, as well as segment and child age. In Part 2, forced alignments derived from a subset of the alignment conditions in Part 1 were used to extract spectral center of gravity of /s/ productions from young children. These findings were compared to published results that used manual alignments of the same data. Results: Overall, the results of Part 1 demonstrated that using training data more similar to the data to be aligned as well as phonetic transcription led to improvements in alignment accuracy. Speech from older children was aligned more accurately than younger children. In Part 2, /s/ center of gravity extracted from force-aligned segments was found to diverge in the speech of male and female children, replicating the pattern found in previous work using manually aligned segments. This was true even for the least accurate forced alignment method. Conclusions: Alignment accuracy of child speech can be improved by using more specific training and transcription. However, poor alignment accuracy was not found to impede acoustic analysis of /s/ produced by even very young children. Thus, forced alignment presents a useful tool for the analysis of child speech.

Adults differ considerably in their perception of both native and nonnative phonemes. For instance, when presented with continua of native phonemes on two-alternative forced choice (2AFC) or visual analog scaling (VAS) tasks, some people show sudden changes in responses (i.e., steep identification slopes) and others show gradual changes (i.e., shallow identification slopes). Moreover, some adults are more successful than others at learning unfamiliar phonemes. The predictors of these individual differences and the relationships between them are poorly understood. It also remains unclear to what extent different tasks (2AFC vs. VAS) may reflect distinct individual differences in perception. In two experiments, we addressed these questions by examining the relationships between individual differences in performance on native and nonnative phonetic perception tasks. We found that shallow 2AFC identification slopes were not related to shallow VAS identification slopes but were related to inconsistent VAS responses. Additionally, our results suggest that consistent native perception may play a role in promoting successful nonnative perception. These findings help characterize the nature of individual differences in phonetic perception and contribute to our understanding of how to measure such differences. This work also has implications for encouraging successful acquisition of new languages in adulthood. Public Significance Statement: Successfully perceiving speech sounds is a crucial skill for spoken communication; yet individuals show differences in how they perceive both native and nonnative speech sounds. We studied the relationships between performance on different native and nonnative speech perception tasks, finding that (a) different tasks measure different subtleties and (b) people with consistent perception of native speech sounds tend to be better at accurately perceiving nonnative sounds. These findings have implications for understanding the nature of individual differences in speech perception and for helping adults to learn new languages successfully. [ABSTRACT FROM AUTHOR]

Purpose: This study examined whether older adults remain perceptually flexible when presented with ambiguities in speech in the absence of lexically disambiguating information. We expected older adults to show less perceptual learning when top-down information was not available. We also investigated whether individual differences in executive function predicted perceptual learning in older and younger adults. Method: Younger (n = 31) and older adults (n = 27) completed 2 perceptual learning tasks composed of a pretest, exposure, and posttest phase. Both learning tasks exposed participants to clear and ambiguous speech tokens, but crucially, the lexically guided learning task provided disambiguating lexical information whereas the distributional learning task did not. Participants also performed several cognitive tasks to investigate individual differences in working memory, vocabulary, and attentionswitching control. Results: We found that perceptual learning is maintained in older adults, but that learning may be stronger in contexts where top-down information is available. Receptive vocabulary scores predicted learning across both age groups and in both learning tasks. Conclusions Implicit learning is maintained with age across different learning conditions but remains stronger when lexically biasing information is available. We find that receptive vocabulary is relevant for learning in both types of learning tasks, suggesting the importance of vocabulary knowledge for adapting to ambiguities in speech.

Purpose: The developmental trajectory of English /s/ was investigated to determine the extent to which children's speech productions are acoustically fine-grained. Given the hypothesis that young children have adultlike phonetic knowledge of /s/, the following were examined: (a) whether this knowledge manifests itself in acoustic spectra that match the gender-specific patterns of adults, (b) whether vowel context affects the spectra of /s/ in adults and children similarly, and (c) whether children adopt compensatory production strategies to match adult acoustic targets. Method: Several acoustic variables were measured from word-initial /s/ (and /t/) and the following vowel in the productions of children aged 2 to 5 years and adult controls using 2 sets of corpora from the Paidologos database. Results: Gender-specific patterns in the spectral distribution of /s/ were found. Acoustically, more canonical /s/ was produced before vowels with higher F[subscript 1] (i.e., lower vowels) in children, a context where lingual articulation is challenging. Measures of breathiness and vowel intrinsic F[subscript 0] provide evidence that children use a compensatory aerodynamic mechanism to achieve their acoustic targets in articulatorily challenging contexts. Conclusion: Together, these results provide evidence that children's phonetic knowledge is acoustically detailed and gender specified and that speech production goals are acoustically oriented at early stages of speech development.

Purpose: Previous work has found that both young and older adults exhibit a lexical bias in categorizing speech stimuli. In young adults, this has been argued to be an automatic influence of the lexicon on perceptual category boundaries. Older adults exhibit more top-down biases than younger adults, including an increased lexical bias. We investigated the nature of the increased lexical bias using a sensorimotor adaptation task designed to evaluate whether automatic processes drive this bias in older adults. Method: A group of older adults (n = 27) and younger adults (n = 35) participated in an altered auditory feedback production task. Participants produced target words and nonwords under altered feedback that affected the 1st formant of the vowel. There were 2 feedback conditions that affected the lexical status of the target, such that target words were shifted to sound more like nonwords (e.g., less-liss) and target nonwords to sound more like words (e.g., kess-kiss). Results: A mixed-effects linear regression was used to investigate the magnitude of compensation to altered auditory feedback between age groups and lexical conditions. Over the course of the experiment, older adults compensated (by shifting their production of 1st formant) more to altered auditory feedback when producing words that were shifted toward nonwords (less-liss) than when producing nonwords that were shifted toward words (kess-kiss). This is in contrast to younger adults who compensated more to nonwords that were shifted toward words compared to words that were shifted toward nonwords. Conclusion: We found no evidence that the increased lexical bias previously observed in older adults is driven by a greater sensitivity to top-down lexical influence on perceptual category boundaries. We suggest the increased lexical bias in older adults is driven by postperceptual processes that arise as a result of age-related cognitive and sensory changes., The already complex task of speech perception becomes more difficult with age, as sensory and cognitive declines cause listening performance to decrease. Older adults are susceptible to a variety of [...]

Purpose: Heterogeneous child speech was force-aligned to investigate whether (a) manipulating specific parameters could improve alignment accuracy and (b) forced alignment could be used to replicate published results on acoustic characteristics of/s/ production by children.Method: In Part 1, child speech from 2 corpora was force-aligned with a trainable aligner (Prosodylab-Aligner) under different conditions that systematically manipulated input training data and the type of transcription used. Alignment accuracy was determined by comparing hand and automatic alignments as to how often they overlapped (%-Match) and absolute differences in duration and boundary placements. Using mixed-effects regression, accuracy was modeled as a function of alignment conditions, as well as segment and child age. In Part 2, forced alignments derived from a subset of the alignment conditions in Part 1 were used to extract spectral center of gravity of/s/ productions from young children. These findings were compared to published results that used manual alignments of the same data.Results: Overall, the results of Part 1 demonstrated that using training data more similar to the data to be aligned as well as phonetic transcription led to improvements in alignment accuracy. Speech from older children was aligned more accurately than younger children. In Part 2, /s/ center of gravity extracted from force-aligned segments was found to diverge in the speech of male and female children, replicating the pattern found in previous work using manually aligned segments. This was true even for the least accurate forced alignment method.Conclusions: Alignment accuracy of child speech can be improved by using more specific training and transcription. However, poor alignment accuracy was not found to impede acoustic analysis of/s/ produced by even very young children. Thus, forced alignment presents a useful tool for the analysis of child speech.Supplemental Material: https://doi.org/10.23641/asha. 7070105, Acoustic analysis of speech has traditionally required labor-intensive hand annotation of segment boundaries or acoustic events. The time-consuming nature of the process has limited the scale of these studies. There [...]

Purpose: The developmental trajectory of English /s/ was investigated to determine the extent to which children's speech productions are acoustically fine-grained. Given the hypothesis that young children have adultlike phonetic knowledge of /s/, the following were examined: (a) whether this knowledge manifests itself in acoustic spectra that match the gender-specific patterns of adults, (b) whether vowel context affects the spectra of /s/ in adults and children similarly, and (c) whether children adopt compensatory production strategies to match adult acoustic targets. Method: Several acoustic variables were measured from word-initial /s/ (and /t/) and the following vowel in the productions of children aged 2 to 5 years and adult controls using 2 sets of corpora from the Paidologos database. Results: Gender-specific patterns in the spectral distribution of /s/ were found. Acoustically, more canonical /s/ was produced before vowels with higher [F.sub.1] (i.e., lower vowels) in children, a context where lingual articulation is challenging. Measures of breathiness and vowel intrinsic [F.sub.0] provide evidence that children use a compensatory aerodynamic mechanism to achieve their acoustic targets in articulatorily challenging contexts. Conclusion: Together, these results provide evidence that children's phonetic knowledge is acoustically detailed and gender specified and that speech production goals are acoustically oriented at early stages of speech development., Along tradition of developmental studies assume (implicitly or explicitly) that what is heard by adult transcribers accurately reflects children's production targets; in other words, what adults transcribe represents what children [...]

English-speakers' learning of a French vowel contrast (/schwa/-/slashed o/) was examined under six different stimulus conditions in which contrastive and noncontrastive stimulus dimensions were varied orthogonally to each other. The distribution of contrastive cues was varied across training conditions to create single prototype, variable far (from the category boundary), and variable close (to the boundary) conditions, each in a single talker or a multiple talker version. The control condition involved identification of gender appropriate grammatical elements. Pre- and posttraining measures of vowel perception and production were obtained from each participant. When assessing pre- to posttraining changes in the slope of the identification functions, statistically significant training effects were observed in the multiple voice far and multiple voice close conditions.

Listeners are exquisitely sensitive to fine-grained acoustic detail within phonetic categories for sounds and words. Here we show that this sensitivity is optimal given the probabilistic nature of speech cues. We manipulated the probability distribution of one probabilistic cue, voice onset time (VOT), which differentiates word initial labial stops in English (e.g., "beach" and "peach"). Participants categorized words from distributions of VOT with wide or narrow variances. Uncertainty about word identity was measured by four-alternative forced-choice judgments and by the probability of looks to pictures. Both measures closely reflected the posterior probability of the word given the likelihood distributions of VOT, suggesting that listeners are sensitive to these distributions. (Contains 1 table and 4 figures.)

Purpose: This study examined whether older adults remain perceptually flexible when presented with ambiguities in speech in the absence of lexically disambiguating information. We expected older adults to show less perceptual learning when top-down information was not available. We also investigated whether individual differences in executive function predicted perceptual learning in older and younger adults. Method: Younger (n = 31) and older adults (n = 27) completed 2 perceptual learning tasks composed of a pretest, exposure, and posttest phase. Both learning tasks exposed participants to clear and ambiguous speech tokens, but crucially, the lexically guided learning task provided disambiguating lexical information whereas the distributional learning task did not. Participants also performed several cognitive tasks to investigate individual differences in working memory, vocabulary, and attention-switching control. Results: We found that perceptual learning is maintained in older adults, but that learning may be stronger in contexts where top-down information is available. Receptive vocabulary scores predicted learning across both age groups and in both learning tasks. Conclusions: Implicit learning is maintained with age across different learning conditions but remains stronger when lexically biasing information is available. We find that receptive vocabulary is relevant for learning in both types of learning tasks, suggesting the importance of vocabulary knowledge for adapting to ambiguities in speech., Listeners regularly encounter unexpected variation in speech, to which they must adapt in order to successfully perceive the signal. Various causes for these idiosyncrasies include accent and dialect differences, speech [...]