Consonant Acquisition in Taiwanese: A Functional Optimality-Theoretic Account

This is a case study of a child acquiring Taiwanese as her first language, attempting to model the phonological error patterns as well as output variation based on an optimality-theoretic (hereafter OT) analysis for her initial consonant acquisition. Driven by the fact that children usually do not produce one single output form that is identical to adults’, and the rule-based approach would further complicate child phonology, we, following Liu and Tsay (2000), approached the variation phenomenon in children’s phonology acquisition based on the OT framework with the Floating Constraints (Nagy and Reynolds 1997). While different from Liu and Tsay (2000), we also considered the functional constraint *GESTURE [tongue: distal] (revised based on Boersma’s Production Grammar Constraint, 1998) in our OT model in order to better capture the phenomena of the velarization pattern, i.e., substitution with the velar stop [k], a phenomena not found in Liu and Tsay’s study and rarely reported in the relevant literature. We argue that the error pattern found here is biologically motivated by minimization of effort on the same ground as the motoric account Meier (2002) proposed for the sign language acquisition. The finding and the approach adopted in the current study support Myers and Tsay’s (2003) suggestion that a formal model with functional factors considered can better describe phonological patterns since the functionalist-based OT model can not only explicitly describe how constraints are selected, ranked, and evaluated, but it also can reflect the extra-grammatical factors, such as speech physiology and auditory perception. The organization of the remainder is as follows: Section 2 reviews the literature regarding Optimality Theory about language acquisition, language variation, and Liu and Tsay (2000). Section 3 reports methodology to the current case study and the general observation. Section 4 presents the analysis of the data based on a revised OT model. We will conclude our findings in the last section.