Chain shift scenarios spontaneously arise and then subside in the developing phonological systems of child language learners. This paper argues that these emergent patterns are natural intermediate stages driven by specific faithfulness constraints referencing perceptually-prominent input feature combinations. As has been proposed elsewhere (e.g., Hayes 2004, Prince and Tesar 2004, Smith 2000, Tessier 2007), these specific faithfulness constraints are biased toward high ranking and stand in a stringency relationship with their general counterparts. Given this framework, reranking of constraints on the basis of positive target-language evidence is all that is necessary for developmental chain shifts to both initially emerge and eventually dissipate. No other mechanisms are required. Furthermore, this approach provides insight into apparent exceptions to the chain shift pattern, defines the related developmental paths that language acquirers can follow, and makes strong predictions about the typology of possible child chain shifts. These claims are illustrated with data from the puzzle-puddle-pickle pattern of Amahl (Smith 1973) and the s-theta-f pattern of R.H. (Dinnsen and Barlow 1998).
Proceedings of the 2nd Conference on Generative Approaches to Language Acquisition North America (GALANA)
edited by Alyona Belikova, Luisa Meroni, and Mari Umeda
Table of contents