The Emergence of Faithfulness

Abstract

Any theory of language development, whether emergentist or nativist, must address the child's phonological development. Children's pronunciations of words are often quite different from those of adults. The child pronunciations may diverge notably from the target adult phonology, but in ways that make sense from a cross-linguistic perspective on the phonological systems of adult languages. In this chapter, we provide an overview of the phenomena of phonological development, and sketch how they can be accounted for within the Optimality Theory variant of Bernhardt and Stemberger (1998). We argue that the main mechanisms within this approach (constraints) are almost certainly emergentist in nature. This paper focuses on speech production during the acquisition of meaningful speech rather than on speech perception or the babbling period. We assume that, at least for children with normal hearing, the child's perception of the adult pronunciation is by-and-large accurate by the time the child begins to produce words (Ingram, 1989; Werker & Lelonde, 1988). Furthermore, we believe that, during the babbling period, the child learns the basics about the mapping from acoustics to articulation; thus, we do not discuss issues regarding development of the mapping in this paper (see Plaut and Kello, this volume, for some ideas about how this process might work). 1.0 What needs to be accounted for in phonological development Pronunciations of words in early child language development, especially before 30 months, but as late as 5 or 6 years of age for particular aspects of the pronunciation, are frequently very different from the pronunciations of adult speakers of the target language. We would like to stress the great range of variation across individuals. There are few generalizations about phonological development that hold true for all young children learning English, especially generalizations about particular sounds or sequences of sounds. The following examples illustrate just a few possible pronunciations of a few words, based on the patterns reported in the literature for individual children; some of the pronunciations for the final word are hypothetical.