Segmentation of speech in a foreign language

The subjects' ability to segment foreign speech was examined. Naturalness judgments regarding three syntactically defined pauses [between constituents (noun and verb phrases), words, or syllables] were obtained using a paired-presentation, forced-choice paradigm. It was hypothesized that segmentation skill developed through exposure to lexical and syntactic markers. Teh existence and effect of such markers was investigated by assigning subjects to various exposure conditions.Results indicated that lexical and syntactic markers exist and can be utilized by subjects in segmenting speech. Contrary to previous research, however, exposure did not facilitate performance. All groups discriminated constituents from either words or syllables, and words from syllables. Results were interpreted as reflecting the interdependence of syntax and suprasegmental phonology. Results challenged the credibility of traditional associationist accounts of language acquisition and speech perception. Results were discussed in the context of Martin's theory of the rhythmic structure of speech.     

Infants' early ability to segment the conversational speech signal predicts later language development: a retrospective analysis

Two studies examined relationships between infants' early speech processing performance and later language and cognitive outcomes. Study 1 found that performance on speech segmentation tasks before 12 months of age related to expressive vocabulary at 24 months. However, performance on other tasks was not related to 2-year vocabulary. Study 2 assessed linguistic and cognitive skills at 4-6 years of age for children who had participated in segmentation studies as infants. Children who had been able to segment words from fluent speech scored higher on language measures, but not general IQ, as preschoolers. Results suggest that speech segmentation ability is an important prerequisite for successful language development, and they offer potential for developing measures to detect language impairment at an earlier age.     

Phonetic variability and prosodic structure in mothers

Spontaneous speech acoustics are highly variable. Such variability may be problematic for infants relying on phonological form to solve the segmentation problem. In the present study, acoustic measures of vowel duration and a computer model of speech segmentation were used to evaluate the problem of phonetic variability for a rhythm-based speech segmentation strategy. The specific questions under study were (1) whether or not mothers realized disyllabic vowel duration patterns consistently in spontaneous infant-directed speech, and (2) whether or not these patterns were distinctive enough in the context of an utterance to provide a useful cue for speech segmentation. Data from four English-speaking mothers indicated that the trochaic-like duration pattern may interact with phrase-position and with grammatical category, but when the resulting patterns are consistent, they provide useful segmentation cues for spontaneous infant-directed speech.     

An Acoustic Analysis of Rate of Speech Effects on Vowel Production in Aphasia

The present investigation examined the production of tense and lax vowel duration differences at two speaking rates in the speech of 10 nonfluent aphasics, 8 fluent aphasics, and 10 normal control subjects. Subjects produced four repetitions of each of the vowels [i e æ o u I ε υ Λ] at each speaking rate. Acoustic analyses revealed that subjects in all three groups were able to manipulate overall rate of speech. In addition, normal controls and fluent aphasic subjects produced vowels under the fast rate condition which were significantly shorter than those under the slow rate condition. Despite a change in overall speaking rate, the nonfluent aphasics did not exhibit a significant difference in vowel duration at the two rates of speech, suggesting a deficit in the implementation of this temporal parameter. Both normal controls and fluent aphasic patients produced nonoverlapping distributions of tense and lax vowels at both speaking rates. In contrast, the nonfluent aphasics demonstrated a great deal of overlap in the distribution of tense and lax vowel durations at the fast rate. Results are discussed in relation to the nature of the speech production deficits in nonfluent and fluent aphasic patients.     

Speech segmentation by statistical learning is supported by domain-general processes within working memory

The purpose of this study was to examine the extent to which working memory resources are recruited during statistical learning (SL). Participants were asked to identify novel words in an artificial speech stream where the transitional probabilities between syllables provided the only segmentation cue. Experiments 1 and 2 demonstrated that segmentation performance improved when the speech rate was slowed down, suggesting that SL is supported by some form of active processing or maintenance mechanism that operates more effectively under slower presentation rates. In Experiment 3 we investigated the nature of this mechanism by asking participants to perform a two-back task while listening to the speech stream. Half of the participants performed a two-back rhyme task designed to engage phonological processing, whereas the other half performed a comparable two-back task on un-nameable visual shapes. It was hypothesized that if SL is dependent only upon domain-specific processes (i.e., phonological rehearsal), the rhyme task should impair speech segmentation performance more than the shape task. However, the two loads were equally disruptive to learning, as they both eradicated the benefit provided by the slow rate. These results suggest that SL is supported by working-memory processes that rely on domain-general resources.     

The contribution of apraxic speech to working memory deficits in Parkinson's disease

This article links two formerly separate areas of research associated with Parkinson's disease (PD): speech and memory. It is proposed that speech deficits occur in PD not merely at the level of muscular control, as is commonly termed dysarthria, but also at the level of speech planning and programming, more aptly described as a form of apraxia. It is further argued that PD patient groups exhibit small deficits in verbal span, and the link between apraxic speech and verbal span is elucidated via Baddeley's (1986) model of working memory. An experiment is described in which aspects of speech of 36 PD and 43 healthy control subjects were rated and classified, and measures of span and articulation rate for words of different syllable lengths were taken. Twenty-three PD subjects had dysarthric speech, while 14 of them had apraxic speech, which was associated with lower memory span scores for longer words. It is concluded that apraxic speech can be a source of reduced memory span in PD. In addition to implications for rehabilitation and therapeutic work with PD sufferers, these findings advance our theoretical understanding of the Parkinsonian syndrome.     

Word rate and intelligibility of alternated speech

A continuous speech message alternated between the left and right ears retains generally good intelligibility, except at certain critical rates of alternation of about 3–4 switching cycles/sec. In the present experiment, subjects heard speech alternated between the two ears at eight different switching frequencies, and at four different speech rates. Results support an earlier contention that the critical intelligibility parameter in alternated speech is average speech content per ear segment, rather than absolute time per ear. Implications are discussed both in terms of critical speech segments in auditory analysis and in neural processing of binaural auditory information.     

The monolingual nature of speech segmentation by bilinguals.

Monolingual French speakers employ a syllable-based procedure in speech segmentation; monolingual English speakers use a stress-based segmentation procedure and do not use the syllable-based procedure. In the present study French-English bilinguals participated in segmentation experiments with English and French materials. Their results as a group did not simply mimic the performance of English monolinguals with English language materials and of French monolinguals with French language materials. Instead, the bilinguals formed two groups, defined by forced choice of a dominant language. Only the French-dominant groups showed syllabic segmentation and only with French language materials. The English-dominant group showed no syllabic segmentation in either language. However, the English-dominant group showed stress-based segmentation with English language materials; the French-dominant group did not. We argue that rhythmically based segmentation procedures are mutually exclusive, as a consequence of which speech segmentation by bilinguals is, in one respect at least, functionally monolingual.     

Rapid speech processing and divided attention: processing rate versus processing resources as an explanation of age effects.

The authors conducted a dual-task study to examine age differences in speech processing under varying loads. Younger and older adults listened to and immediately recalled spoken passages presented at various speech rates (140-280 words per min). This task was performed alone as well as in a divided-attention condition in which subjects concurrently performed a picture recognition task. Consistent with the slowing hypothesis, older adults' immediate memory performance was differentially depressed when speech rates were very fast. The Age x Speech Rate interaction, however, was not exacerbated in the divided-attention condition. This suggests that aging may reduce the rate at which the processing operations underlying memory for speech are completed, but this is conceptually distinct from an age-related reduction in attentional capacity.     

Segmentation of coarticulated speech in perception

The research investigates how listeners segment the acoustic speech signal into phonetic segments and explores implications that the segmentation strategy may have for their perception of the (apparently) context-sensitive allophones of a phoneme. Two manners of segmentation are contrasted. In one, listeners segment the signal into temporally discrete, context-sensitive segments. In the other, which may be consistent with the talker’s production of the segments, they partition the signal into separate, but overlapping, segments freed of their contextual influences. Two complementary predictions of the second hypothesis are tested. First, listeners will use anticipatory coarticulatory information for a segment as information for the forthcoming segment. Second, subjects will not hear anticipatory coarticulatory information as part of the phonetic segment with which it co-occurs in time. The first hypothesis is supported by findings on a choice reaction time procedure; the second is supported by findings on a 4IAX discrimination test. Implications of the findings for theories of speech production, perception, and of the relation between the two are considered.     

口语加工中的词语切分线索

词是语言的基本结构单位, 对词语进行切分是语言加工的重要步骤。口语语流中的切分线索来自于语音、语义和语法三个方面。语音线索包括概率信息、音位配列规则和韵律信息, 韵律信息中还包括词重音、时长和音高等内容, 这些线索的使用在接触语言的早期阶段就逐渐被个体所掌握, 而且在不同的语言背景下有一定的特异性。语法和语义线索属于较高级的线索机制, 主要作用于词语切分过程的后期。后续研究应从语言的毕生发展和语言的特异性两个方面考察口语语言加工中的词语切分线索。     

The effect of fast speech rate on stuttering frequency during delayed auditory feedback

Delayed auditory feedback (DAF) has been documented to improve fluency in those who stutter. The increased fluency has been attributed to the slowed speech rate induced by DAF, but recent experiments have suggested that increasing the speech rate may also decrease stuttering under DAF. This investigation described the effect of combining a fast speech rate and DAF on the fluency of four people who stutter. Fluency of the two mildly dysfluent subjects was the same for both no DAF and DAF conditions at normal and at fast oral reading rates. In contrast, the two severely dysfluent subjects improved in fluency from the no DAF to the DAF conditions. They were found to be dysfluent at both normal and fast oral reading rates without DAF. The results of the study point to the need for further research on the relationship between speech rate and stuttering frequency under conditions of DAF and no DAF. EDUCATIONAL OBJECTIVES: Readers will learn about and be able to describe how the frequency of stuttering is affected by: (1) speech rates; (2) DAF; and (3) how stuttering severity influences such effects.     

Infants' sensitivity to allophonic cues for word segmentation

A series of four experiments was conducted to determine whether English-learning infants can use allophonic cues to word boundaries to segment words from fluent speech. Infants were familiarized with a pair of two-syllable items, such as nitrates and night rates and then were tested on their ability to detect these same words in fluent speech passages. The presence of allophonic cues to word boundaries did not help 9-month-olds to distinguish one of the familiarized words from an acoustically similar foil. Infants familiarized with nitrates were just as likely to listen to a passage about night rates as they were to listen to one about nitrates. Nevertheless, when the passages contained distributional cues that favored the extraction of the familiarized targets, 9-month-olds were able to segment these items from fluent speech. By the age of 10.5 months, infants were able to rely solely on allophonic cues to locate the familiarized target words in passages. We consider what implications these findings have for understanding how word segmentation skills develop.     

Speech segmentation is facilitated by visual cues

Evidence from infant studies indicates that language learning can be facilitated by multimodal cues. We extended this observation to adult language learning by studying the effects of simultaneous visual cues (nonassociated object images) on speech segmentation performance. Our results indicate that segmentation of new words from a continuous speech stream is facilitated by simultaneous visual input that it is presented at or near syllables that exhibit the low transitional probability indicative of word boundaries. This indicates that temporal audio-visual contiguity helps in directing attention to word boundaries at the earliest stages of language learning. Off-boundary or arrhythmic picture sequences did not affect segmentation performance, suggesting that the language learning system can effectively disregard noninformative visual information. Detection of temporal contiguity between multimodal stimuli may be useful in both infants and second-language learners not only for facilitating speech segmentation, but also for detecting word–object relationships in natural environments.     

Speech timing in Parkinson''s and Huntington''s disease

The purpose was to determine the effects of two different basal ganglia diseases on speech planning, initiation, and production. Patients with Parkinson's Disease (PD) and Huntington's Disease (HD) were compared with normal subjects on speech timing tasks. Speech reaction time was unimpaired, while changes in duration at different speech rates were affected differently in the two patient groups. In HD, changes in the duration of syllables, of pauses between phrases, and of sentences were all reduced, while in PD only the control of sentence duration was impaired. The HD patients had reduced syllable repetition rates, but both patient groups maintained their repetition rates over 5 sec. The ratios of word to phrase time, and of phrase to sentence time, remained constant across regular and fast speaking rates and did not differ from normal in either patient group. The results suggest that PD and HD patients are not impaired in speech planning or initiation, but have poor control over the duration of speech events.     

Temporal properties of memory for speech in preschool children

The ability of four-year-old children to remember syllables was investigated, both to determine the temporal properties of memory for speech in subjects who do not use sophisticated retention strategies, and to determine what speech memory is available in the late preschool period. In Experiments 1 and 2, children were to remember pairs of nonsense syllables for 5, 10, 15, or 20 sec filled with a silent, manual task. In both experiments there was an apparent decay of one type of speech memory across 10-15 sec. In Experiment 2, strong interference was obtained when there was a final speech suffix item. An additional phase of this experiment verified that children cannot rehearse during the postlist period. However, in Experiments 3 and 4, suffix effects were not obtained with adult subjects in similar tasks. The results suggest that memory for speech persists for at least 10 sec even in preschool children, but that adults, unlike preschool children, engage in an additional type of speech coding that increases the resistance of memory to speech interference.     

Effects of rate and word boundary ambiguity on recall by normal and impaired listeners

Twenty normal-hearing and 20 hearing-impaired persons heard experimental sentences that included two syllables with an ambiguous word boundary and control sentences that did not include an ambiguous word boundary. The sentences were produced at two rates. Difficulty in generating a meaning for the ambiguous word boundary that made sense within the sentence appeared to reduce recall. For the normal-hearing persons, the difficulty with the ambiguous word boundary reduced recall of experimental sentence words when the speech was fast but not when it was slow. For the hearing-impaired persons, the difficulty reduced the recall of the experimental sentence words when the speech was slow but not when it was fast. Two weeks later, the same listeners heard a set of two syllables having ambiguous word boundaries that were isolated from their sentences. Varying speed of speaking affected the segmentation of the two syllables by the normal-hearing more than it affected the segmentation by the hearing-impaired.This work was supported by NINDS Grants No. NS10468 and 05362.     

The rate of internal speech in paced rehearsal

Hollnagel, E. The rate of internal speech in paced rehearsal. Scand. J. Psychol., 1973, 14, 241–243.-The purpose of the experiment was to measure the rate of internal speech in rehearsal. The method used was paced rehearsal, i. e. the subjects should synchronize their internal speech with an external signal. The known rate of the external signal, the pace rate, was used as an ndication of the rate of internal speech. For each subject the maximum pace rate was found, and the corresponding maximum rate of internal speech was calculated. The mean of the maximum rates of internal speech for 17 subjects was 126 ms/syllable. This is considerably faster than the results obtained from experimental studies up to now.     

Disfluencies and rate of speech in young adult nonstutterers

The purpose of this study was to identify disfluency levels and rates of speech in young adults while they were being interviewed and during reading. Fifty undergraduate students (25 males and 25 females) served as subjects. Each was recorded individually while being interviewed and while reading a standard clinical passage. Percentage of disfluency, using eight disfluency categories, and rates of speech were obtained for all conditions. Results provide information that allows for comparison of post-therapy stutterers with nonstutterers during a clinical interview.     

Language learning impairment: Integrating research and remediation

Timing cues present in the acoustic waveform of speech provide critical information for the recognition and segmentation of the ongoing speech signal. Research has demonstrated that deficient temporal perception rates, that have been shown to specifically disrupt acoustic processing of speech, are related to specific language-based learning impairments (LLI). Temporal processing deficits correlate highly with the phonological discrimination and processing deficits of these children. Electrophysiological single cell mapping studies of sensory cortex in brains of primates have shown that neural circuitry can be remapped after specific, temporally cohesive training regimens, demonstrating the dynamic plasticity of the brain. Recently, we combined these two lines of research in a series of studies that addressed whether the temporal processing deficits seen in LLIs can be significantly modified through adaptive training aimed at reducing temporal integration thresholds. Simultaneously, we developed a computer algorithm that expanded and enhanced the brief, rapidly changing acoustic segments within ongoing speech and used this to provide intensive speech and language training exercises to these children. Results to date from two independent laboratory experiments, as well as a large national clinical efficacy trial, demonstrate that dramatic improvements in temporal integration thresholds, together with speech and language comprehension abilities of LLI children, results from training with these new computer-based training procedures.