Francisco Lacerda

Voice Onset Time in Speech to Infants and Adults

Investigations of the phonetics of infant-directed speech (IDS) have mainly focused on prosodic aspects, showing extreme values or overspecifications in comparison to adult-directed speech (ADS). On the other hand, studies of segmental properties, especially consonantal ones, are very rare. The lack of knowledge of phonetic characteristics of the infant’s speech input is problematic, considering that the infant’s perceptual development seems to be attuned towards properties in the ambient language, as suggested by infant perception experiments. In the present study voice onset time (VOT) was measured in Swedish mothers’ IDS and in their ADS. VOT was significantly shorter in IDS than in ADS in both voiced and voiceless stops. The impact of stress was very clear in both IDS and ADS, showing significantly longer VOT in stressed positions as compared to the unstressed. The results are discussed in a framework assuming a differential weighting system of different phonetic aspects such as prosodic, consonantal and vocalic, as a function of infant development.

Does babbling sound native? Listener responses to vocalizations produced by Swedish and American 12- and 18-month-olds.

Previous studies of infants’ babbling have reported contradictory results as to the extent and timing of discernible phonetic influences of the ambient language. In the present experiment, five experienced phoneticians were asked to listen for ambient language effects on vocalizations produced by American and Swedish 12- and 18-month-olds (with 8 children in each language and age group), and to motivate their decisions in terms of word or phonetic cue perception. Group results indicated that listeners did not perceive effects of ambient language on pure babbles for either of the two age groups, whereas a clear effect appeared in both age groups given a more liberal definition of babbling. This is taken to suggest that results of ambient language listening tests may depend crucially on judgments of vocalizations’ word status. As compared to the group trends, listener responses to individual children’s vocalizations did not indicate that a majority of either 12- or 18-monthers were sufficiently native-sounding to be reliably identified on the basis of ambient language. A closer analysis of listeners’ use of phonetic cues indicated that one single phonetic property, the grave tonal word accent, was discerned by most listeners in vocalizations produced by the Swedish 18-monthers; this property was also discerned by one listener in vocalizations produced by Swedish 12-monthers. This result is consistent with the generally held belief in the primacy of tonal features in phonetic acquisition, and with experimental evidence indicating that Swedish mothers tend to enhance word accent contours in baby talk. In the final section of the paper, the results are discussed with a view to reconcile competing theories of babbling development, notably the ‘babbling drift’ and the ‘independence’ hypotheses.

Speech discrimination and phonological working memory in children with ADHD

This study examined phonological working memory and speech discrimination among children with attention‐deficit–hyperactivity disorder (ADHD) with and without motor problems. Forty‐one children were assigned to three groups; children with ADHD (N=9), children with ADHD plus developmental coordination disorder (ADHD+, N=13), and age‐matched control children (N=19). The subjects’ability to classify stimulus pairs was examined in two experiments. The first experiment required subjects to discriminate pairs of monosyllabic stimuli with contrasting consonants to test speech discrimination without using a working‐memory load. In the second protocol, subjects were exposed to two‐ to five‐syllabic non‐word pairs with contrasting vowels in order to test speech discrimination with a working‐memory load. The subjects classified the pairs as being either the same or different in both experiments. No significant differences were found between the subject groups in the discrimination task with monosyllables. When exposed to the two‐ to five‐syllabic stimuli, the ADHD+ group scored significantly lower than both other groups. This was attributed to a higher sensitivity to working‐memory load. Some possible explanations of this effect are discussed.

Perception of lexical tone across languages: evidence for a linguistic mode of processing

Pairs of Thai tones were presented for perceptual discrimination in three linguistic contexts [normal speech, low-pass filtered speech, and as musical (violin) sounds] to tonal language speakers, Thai and Cantonese, and non-tonal (English) language speakers. English speakers discriminated the tonal contrasts significantly better in the musical context than in filtered speech, and in filtered speech better than in full speech. On the other hand, both Thai and Cantonese speakers perceived the tonal contrasts equally well in all three contexts. Thus, developmental absence of exposure to lexical tone results in a linguistic mode of processing which involves the attenuation of a basic psychoacoustic ability, pitch discrimination.

Universality and language-specific experience in the perception of lexical tone and pitch

Two experiments focus on Thai tone perception by native speakers of tone languages (Thai, Cantonese, and Mandarin), a pitch–accent (Swedish), and a nontonal (English) language. In Experiment 1, there was better auditory-only and auditory–visual discrimination by tone and pitch–accent language speakers than by nontone language speakers. Conversely and counterintuitively, there was better visual-only discrimination by nontone language speakers than tone and pitch–accent language speakers. Nevertheless, visual augmentation of auditory tone perception in noise was evident for all five language groups. In Experiment 2, involving discrimination in three fundamental frequency equivalent auditory contexts, tone and pitch–accent language participants showed equivalent discrimination for normal Thai speech, filtered speech, and violin sounds. In contrast, nontone language listeners had significantly better discrimination for violin sounds than filtered speech and in turn speech. Together the results show that tone perception is determined by both auditory and visual information, by acoustic and linguistic contexts, and by universal and experiential factors.

Phonology: An emergent consequence of memory constraints and sensory input

This paper presents a theoretical model thatattempts to account for the early stages oflanguage acquisition in terms of interactionbetween biological constraints and inputcharacteristics. The model uses theimplications of stochastic representations ofthe sensory input in a volatile and limitedmemory. It is argued that phonologicalstructure is a consequence of limited memoryresources under the pressure of ecologicallyrelevant multi-sensory information.

Temporal Resolution of Auditory Perception in Relation to Perception, Memory, and Language Skills in Typical Children

This study examined temporal resolution of auditory perception (TRAP) in relation to speech perception and phonological working memory in one experiment and in relation to measures of rapid naming, understanding of grammatical contrasts, and phonological awareness in a second experiment. In the TRAP test, 6- to 9-year-old children discriminated four pairs of two sine tones, 878 Hz and 1350 Hz, presented in blocks with stepwise decreasing interstimulus intervals from 256 ms to 8 ms. In Experiment 1, weak but significant correlations between TRAP and speech perception and phonological working memory were found, as well as a significantly lower TRAP performance by girls. Experiment 2 indicated a significant relation between TRAP and the phonological awareness task. The results give qualified support to the notion of a weak relationship between TRAP and speech perception and language skills among typical children. However, the gender differences in TRAP performance, the limited variance explained by the regression models, and the finding that only one of the language parameters correlated with TRAP do not support the notion of a causal relationship between language skills and TRAP. The possibility that TRAP is connected to general neurological maturation rather than specifically to language abilities is discussed.

Multimodal Language Acquisition Based on Motor Learning and Interaction

This work presents a developmental and ecological approach to language acquisition in robots, which has its roots in the interaction between infants and their caregivers. We show that the signal directed to infants by their caregivers include several hints that can facilitate the language acquisition and reduce the need for preprogrammed linguistic structure. Moreover, infants also produce sounds, which enables for richer types of interactions such as imitation games, and for the use of motor learning. By using a humanoid robot with embodied models of the infant’s ears, eyes, vocal tract, and memory functions, we can mimic the adult-infant interaction and take advantage of the inherent structure in the signal. Two experiments are shown, where the robot learn a number of word-object associations and the articulatory target positions for a number of vowels.

An Exemplar-Based Account of Emergent Phonetic Categories

This paper describes how an exemplar‐based model can be used to study the early stages of language acquisition and phonetic category formation. The paper discusses in particular how a (conceptually) very simple model, inspired by Edelmans notion of Neural Darwinism, may account for the emergence of phonetic categories arising from the interaction between ambient language input and limited memory‐representation resources. It presents an explicit attempt to demonstrate how phonetic categorization may emerge ‘‘spontaneously’’ from exposure to natural infant‐directed speech data. The results suggest that phonetic categories may emerge if the auditory input is statistically correlated with one (or several) other sensory inputs. The perceptual‐magnet effect is also discussed in the light of the present model. According to the model, the perceptual‐magnet effect can be seen as a consequence of distributed memory representations. [Work supported by The Bank of Sweden Tercentenary Foundation, grant 94‐0435.]

The asymmetric structure of the infant’s perceptual vowel space

This paper reports a study of the infant’s perceptual sensitivity to F1 and F2 changes occurring in the central region of an adult male’s vowel space. The data were collected from a group of 16 Swedish infants whose ages varied from 6 to 12 months. The infants were tested with the headturn procedure in their ability to discriminate variants of a reference schwa vowel. Four variants were generated by introducing a fixed increment or decrement (in Bark) along either F1 or F2. The results are compatible with previously reported data obtained for vowel contrasts in the low/back region of the vowel space [F. Lacerda, J. Acoust. Soc. Am. 93, 2372(A) (1993)]. Thus the general notion that there is a perceptual asymmetry favoring F1 differences during the early stages of vowel perception is supported by the present results. The paper attempts to argue that the early structure of the infant’s vowel space might be influenced by the combined effects of both the perceptual dominance of F1 contrasts and the infant’s la...