Introduction

Despite decades of research on speech development in children, three fundamental questions remain unanswered:

1. Just how universal are patterns of phonological development?

2. How do basic biological and environmental factors interact to impinge on phonological learning?

3. What mechanisms account for high individual variability in the rate and route of phonological learning?

These questions have been raised repeatedly in recent years (e.g. Ingram, 1999; Macken, 1995; Stoel-Gammon, 1992) yet they remain largely unanswered. Recent work has provided new and intriguing possible answers to these questions. First, there is now widespread evidence that patterns of phonological development are universal, insofar as all development follows an implicational hierarchy of increasing feature complexity (Jakobson, 1968; Dinnsen et al., 1990; Macken, 1995; Stokes et al., in press). Second, recent studies of Chinese phonological development clearly indicate a computational interaction of ambient language (input) and articulatory (motor) constraints on consonant and vowel learning (Stokes & To, 2002; Stokes & I.-M. Wong, 2002; Stokes, Lau & Ciocca, 2002; Stokes & C. T.-Y. Wong, 2004; Stokes & Surendran, 2005). Third, Plaut and Kello (1999) provide a convincing demonstration of phonological learning by a connectionist model that reflects possible cognitive processes of phonological learning in the child. This chapter describes how recent research addresses learnability issues in phonological development, with a focus on Chinese.

Before describing this research, a digression is required to briefly outline the phonological systems of Cantonese and Putonghua.