Toshisada Deguchi

Listener adaptive characteristics of vowel devoicing in Japanese dialogue

Listener adapative characteristics of Japanese vowel devoicing were investigated by analyzing (i) dialogues between professional teachers and hearing-impaired (HI) or normal-hearing (NH) children, and (ii) speech samples read by the teachers as fast and clearly as possible (RD). The teachers reduced the devoicing rate and lengthened the moras more in the HI vs NH samples, and even more in the HI vs RD samples. A logistic regression analysis of the devoicing rate suggests that the speech rate dependency of the devoicing rate is different among the HI, NH, and RD samples. When moras are lengthened, the predicted devoicing rate decreases more for the HI vs NH samples, and even more in the HI vs RD samples, suggesting that not only rate-dependent adjustments but also rate-independent adjustments significantly affect devoicing reduction. These results suggest the following: (1) Professional teachers of hearing-impaired children reduce their devoicing not only by lengthening the interval of successive voicing/devoicing gestures, but also by resizing component gestures to some extent, probably to improve the listeners comprehension; (2) vowel devoicing should be represented in terms of parameters of speech motor control; (3) it may be possible to develop an optimized communication method for hearing-impaired children by simulating such listener-adaptive adjustments in speech production.

Effects of language experience on speech perception: American and Japanese infants’ perception of /ra/ and /la/

Listening to language during the first year of life has a dramatic effect on infants’ perception of speech. With increasing exposure to a particular language, infants begin to ignore phonetic variations that are irrelevant in their native language. To examine these effects, 72 American and Japanese infants were tested at two ages, 6–8 months and 10–12 months, with synthetic versions of the American English /r/ and /l/ consonants. The /r–l/ contrast is not phonemic in Japanese. In both countries, the same experimenters, technique (head‐turn conditioning), and stimuli were used. The results revealed two significant effects. The first shows the impact of language experience on speech perception. At 6–8 months of age, American and Japanese infants did not differ. Both groups performed significantly above chance (American M=63.7%; Japanese M=64.7%). By 10–12 months of age, American infants demonstrated significant improvement relative to performance at 6–8 months (M=73.8%), while Japanese infants declined (M=5...

Children's Imitations of Movements are Goal-Directed and Context-Specific

Previous research indicates that imitation of gestures in preschool children is goal-directed. A goal may be a salient feature from a presented movement; that goal may be imitated correctly, but other features were ignored, resulting in observable errors. Objects (e.g., a dot on the table) can become the most salient features and presence or absence of objects influences imitation responses. Imitation responses were examined under conditions in which objects could not be used directly as the most salient feature. 60 children (M age = 5:6) were assigned to Gestural, Dot, No-dot, and Un-dot conditions, and they were asked to imitate 20 movements. The type of presented movement and the occurrence of correct, mirror, and error responses were examined. Responses in the Un-dot condition were similar to those in the Dot condition. Error responses in the Un-dot condition were related to age. Children may extract a more abstract feature in a context without visible objects. This ability is associated with a cognitive mechanism developed in preschool years.

Imitation of Modeled Manipulation of Objects by Children and Adults

Prior studies have investigated imitation by extracting a hierarchy of goals from the key elements of action models. The theoretical model is that all ages practice a method of imitation in which goal or target elements are more easily imitated correctly, while nontarget elements are not. The present study compared error responses among 32 children and 32 adults when imitating an action model for manipulating concrete objects constructed of five elements. The results indicated that the elements for which error responses were easily produced and those for which error responses were more difficult to produce were approximately the same in children and adults. It showed that the imitation mechanisms were similar in children and adults. In addition, children had higher omissions and error responses than adults. This result suggests that imitation and differences in working memory capacity may be related.

Effects of L1 phonotactic constraints on L2 speech perception and production

Effects of L1 phonotactic constraints on L2 speech perception and production were analyzed during the course of audio‐visual perceptual training for Japanese adult learners of English by observing articulation of words containing /l/, /r/, and /w/. The speech identification score drastically improved during the training. The improvement in non‐native word‐initial rl distinction was clearly associated with the changes in the perceptual, articulatory, and neuronal spaces, which represent dissimilarities between the non‐native and native phonemes in each domain assessed through perceptual, palatographic, and neuromagnetic measurements. Significant difficulty in identification of consonant‐rl clusters, however, remained for some trainees even after one year of training. Analyses of palato‐lingual contact patterns during word articulation suggested that more articulatory errors tend to occur in consonant‐rl clusters than in other phonotactic contexts in such a way that both /l/ and /r/ are substituted with Japanese /r/ sometimes associated with a vowel inserted. Results suggested that the L2 phonological system can be acquired through a multimodal training, although it is strongly interfered with by L1 phonotactic constraints.

Physical imitation and verbal description of modeled movements engage different encoding processes.

The present study analyzed the effects of performing a specified sequence of movements in imitation of a model, and examined the characteristics of the encoding involved in the process. Thirty-two college students were presented with a movement model that consisted of five elements and performed imitation tasks in which they either reproduced the movement sequence physically or described it verbally. During the period from the onset of the movement model to the imitation response, participants performed articulation suppression, movement suppression, and spatial suppression tasks, and their effects on the imitation response were analyzed. The results showed that the encoding processes for conversion to movement and to language were different, and that movements associated with articulations or planning of articulations and spatial processing were involved in converting the model to movement. In addition, imitation-specific goal selection was partially supported in movement conversion, but not in language conversion.

Perception of prosodic units in native and non‐native languages by 4‐ to 10‐month‐old Japanese infants

The present study investigated whether Japanese infants are sensitive to prosodic cues to clausal units in native (Japanese) and non‐native (English) languages. Two types of speech samples (child‐directed speech) of each language were constructed for use as stimulus materials: Coincident samples were created by inserting 1‐s pauses at all clause boundary locations, and noncoincident samples were created by inserting the same number of pauses between words within clauses. Groups of 6‐ and 10‐month‐old infants were tested using Japanese samples, and groups of 4‐ and 10‐month‐old infants were tested using English samples on the headturn preference procedure. The results showed that only the 10‐month‐old infants tested on Japanese speech samples demonstrated a significant preference for the noncoincident samples over the coincident samples. These results indicate that very young Japanese infants are not sensitive to clausal units in either language, but they come to be sensitive to clausal units only in their...

Children's Imitation is not Always Goal-Directed: Evidence from Goal Clarification Task

The most basic unit of imitation is the ability to translate observed movements into our own movements and reproduce them as our own actions. Goal-directed theories assert that imitation is affected by goals extracted from the observed movements, whereas generalist hypothesis asserts that imitation is affected by the salience of the components rather than the goals. This study developed a new imitation task in order to clarify the role of goals in imitating movement to examine the relative applicability of these two theories to children’s imitation. Young children (N = 55; mean age = 5:1 years, age range = 4 yr. 1 mo. - 6 yr. 0 mo.) were asked to imitate the movements of an experimenter with a toy train. While the children observed the experimenter moving the toy train, they also heard a commentary describing the train’s departure and the trip until the train arrives at the station. The commentary emphasized that the train’s movements had a start and a goal. We analyzed the children’s imitation errors, including whether the color salience of the goal influenced the accuracy of their imitation. The results supported the generalist hypothesis, because the part that was relatively accentuated was more accurately imitated.

Word recognition by Japanese infants

Building a lexicon is a necessary step in the process of language acquisition. The emergence and development of vocabulary have usually been observed in naturalistic settings through their external manifestations: the first attempts at producing words, and the various signs showing that an infant comprehends words. Naturalistic approaches, however, may underestimate the productive lexicon on one hand, and overestimate the receptive lexicon on the other hand. An experimental approach, using the headturn preference procedure, has been used to show that 11-to-12-month-old French infants can recognize familiar words without specific training (Halle and Boysson-Bardies, 1994). The spontaneous preference for familiar words, interpreted as word recognition, revealed the formation of an early receptive lexicon comprising a significant part of the familiar words that were used. At 12 months, recognition seemed to be firmly established, while it seemed to be just emerging at 11 months. Using the same procedure, the present study examined familiar word recognition in Japanese infants: 12-month-olds, but not 10-month-olds, showed a preference for familiar words, similar in intensity to that shown by 11-month-old French infants. These results are again interpreted as revealing the formation of a nascent receptive lexicon in Japanese infants by 12 months of age. Commonalties and differences between the two language groups are discussed.

Vowel devoicing in Japanese dialogue between teachers and hearing‐impaired or normal‐hearing children: Listener adaptive characteristics of dialogue speech production

Listener adaptive characteristics of Japanese vowel devoicing in dialogue were investigated by analyzing dialogue between teachers and hearing‐impaired (HI) or normal‐hearing (NH) children. Dialogue was recorded during a game through which seven teachers assessed the speech perception of seven hearing‐impaired and seven normal‐hearing children. The following results were obtained. (1) Devoicing rate was significantly lower in dialogue directed to HI than to NH. (2) For vowel /i/ in the second mora of /CVkita/, where C was one of /h,s,sh,ts/ and V /i,u/, unvoiced segments of not‐fully‐devoiced syllables that were mainly directed to HI tended to be shorter than those of fully devoiced syllables that were mainly directed to NH. This may indicate that the teachers avoided devoicing by shortening unvoiced duration and lengthening voiced duration for HI. (3) Monosyllables and pauses in dialogue directed to HI were longer than those directed to NH. These results suggest that the teachers controlled their speakin...