Perceptual bases of early lexical acquisition: an integrated developmental program – PBELA

The acquisition of a lexicon requires several abilities that can be schematically grouped into three different sets: ability to represent the sound pattern of a given word; ability to build the concept corresponding to that given word; and ability to link a sound pattern to its appropriate concept (to form a 'word'). The present project concentrates on the first (speech perception) and third (effects of the onset of the acquisition of words per se on lexical-related speech perception abilities) sets of abilities. More specifically, we will investigate in infants acquiring French (a) the emergence of speech segmentation procedures, and (b) the format of sound pattern representations before and after they start being attached to a referent. (a) emergence of speech segmentation Many behavioral studies have explored the emergence in English of speech segmentation, a prerequisite to lexical acquisition corresponding to the speech processing step that allows the determination of the sequence of word forms making up the heard utterances. These studies have shown the emergence of this ability by 8 months of age. At that age, these infants rely on prosodic cues to word boundaries, but also, within the prosodically-defined units, on the distributional properties of syllable order. Later on, by 10-12 months, they start using other word boundary cues defined at the lexical level: allophonic and phonotactic cues. The present project will specify the emergence of segmentation abilities in French; crosslinguistic differences are expected given that some of cues used are language-specific. We will continue evaluating, at the behavioral level, the role of syllable-based prosodic segmentation in French, and then test the use of distributional cues. Using the method of evoked-response potentials (ERPs), we will also specify the electrophysiological correlates (latency, scalp distribution) of speech segmentation abilities, to determine whether different segmentation cues are processed by different neural networks. Based on this knowledge, we will explore the development of segmentation abilities in bilingual infants, and infants with atypical development (cochlear-implanted deaf infants and infants with Williams syndrome), in order to test hypotheses about the impact of specific deficits on speech segmentation abilities. (b) format of sound pattern representations Numerous studies have looked at the phonetic specificity of sound pattern representations at different processing levels: segmented word forms, known words, process of learning new words. However, no study has explored the sensitivity to the same contrasts at all these different levels, and most studies have focused on consonantal contrasts in word-initial positions and embedded in stressed syllables. The ambiguous results regarding the specificity at these different levels and at different developmental points, raises the issue of the generalizability of the results obtained so far, and call for the systematic exploration of the same contrasts at all these levels. This will be done first using behavioral methods, and then will using ERPs. Then, as before, the studies will be extended to bilingual infants and infants with atypical development. Taken together, these two lines of research will contribute to a better specification of the pattern of typical lexical acquisition in French, language for which there is little data, and will underline the influence of the phonological properties of this language on the acquisition pattern observed. These studies will help understand the relationship between phonological and lexical development, and bring new data to evaluate the hypothesis of a continuity between both developments. Some of the experiments will also address the issue of the interaction between and integration of information from different modalities and different levels: audiovisual integration when looking at the effect of audiovisual presentation on segmentation (see section on cochlear-implanted infants); link between phonological development and cognitive development when looking at the effect of word learning on phonetic specificity of sound pattern representations. Finally, the present project is concerned with bridging a gap between what is observed at the behavioral level (using the following classical techniques in the field of development: Headturn Preference Procedure, Intermodal Preferential Looking, object manipulation) and the cerebral bases that support this behavior (using the non-invasive brain imaging technique of high-density Evoked Response Potentials).