Adaptation to language: Evidence from babbling and first words in four languages

ADAPTATION TO LANGUAGE: EVIDENCE FROM BABBLING AND FIRST WORDS IN FOUR LANGUAGES Bénédicte de Boysson-BardiesMarilyn May Vihman C.N.R.S.-E.H.E.S.S.Rutgers University Differences among languages offer a way of studying the process of infant adaptation from broad initial capacities to language-specific phonetic production. We designed analyses of the distribution of consonantal place and manner categories in French, English, Japanese, and Swedish to determine (1) whether systematic differences can be found in the babbling and first words of infants from different language backgrounds, and, if so, (2) whether these differences are related to the phonetic structure of the language spoken in the environment. Five infants from each linguistic environment were recorded under similar conditions from babbling only to the production of25 words in a session. Although all of the infants generally made greater use of labials, dentals, and stops than of other classes of sounds, a clear phonetic selection could already be discerned in babbling, leading to statistically significant differences among the groups. This selection can be seen to arise from phonetic patterns of the ambient language. Comparison of the babbling and infant word repertoires reveals differences reflecting the motoric consequences of sequencing constraints.* Introduction 1. Comparison of different languages makes it possible to investigate the epigenetic processes involved in the attainment of what may be the most complex of human action patterns: the production of speech. It can be assumed that there exists a capacity for language which is common to members of all speech communities. But different languages require that this capacity be shaped into different patterns—patterns that may be quite precisely defined by analysis of the adult language in question. Crosslinguistic study of the course of evolution toward these patterns can yield considerable insight into the nature of species-wide phonetic capacities, both by establishing commonalities in the progress of normally-developing infants and by determining the modifications and the time course involved in achieving particular adult-language goals. Work in the area of speech acquisition over the past several years has led to a profound change in our views about the nature of infants' early vocal action capacities. The earlier view, based on the theories of Jakobson (1941/ 68), held that prespeech infants produce all of the sounds of the world's languages . Jakobson maintained, however, that with the advent of word production children revert to a highly restricted sound-production capacity, dedicated to the unfolding of a universal sequence of segmental contrasts. This view has now been replaced by one that emphasizes the constraints on prespeech ca- * We gratefully acknowledge support from NSF grant BNS 85-20048 to Charles A. Ferguson, with the participation of Bénédicte de Boysson-Bardies, Björn Lindblom, and Marilyn May Vihman . For their considerable assistance in the collection, transcription, and analysis ofthe children's data we thank Marlys A. Macken, Ruth Miller, and Hazel Simmons for English, Liselotte RougHellichius and Ingrid Landberg for Swedish, Catherine Durand for French, and Fumiko Arao for Japanese. We also thank three anonymous referees for their helpful comments on an earlier draft of the paper. 297 298LANGUAGE, VOLUME 67, NUMBER 2 (1991) pacities, as well as their essential relevance to the form ofthe first words (Locke 1983, Vihman et al. 1985). This more recent view has sometimes been taken to mean that production of the child's first words requires no language-specific motor capacities (Locke 1983). Earlier work attempting to show language-specific influences within the first year through adult-listener judgments has been criticized on methodological grounds (Thevenin et al. 1985). However, at least two recent studies have demonstrated language-specific production patterns in vowels in the course of the first year, before the appearance of the first words. These studies are based on statistical analysis of the relative distribution of vowel sounds in the adult target language and in the vocalization patterns of infants, as defined by formant frequencies (de Boysson-Bardies et al. 1986, de Boysson-Bardies et al. 1989). These two studies suggest that vowel production begins to approximate favored values in the adult language even before production of the first words. Can similar effects be shown for...