An Optimality Theoretic Approach to Child Language Acquisition

1 AN OPTIMALITY THEORETIC APPROACH TO CHILD LANGUAGE ACQUISITION 2 One goal of linguistics is to develop models for how children acquire language. Boersma, Escudero, and Hayes (2003) provide a model using an Optimality Theoretic (OT) approach to show how children might learn to group language specific F1 frequencies into phonetic categories. The model utilizes initially high-ranked discriminatory constraints and then learns to categorize the sounds into appropriate phonetic categories. This presents a possible problem: Why does the learner begin categorizing speech sounds as meaningful when the initial state discourages this? In this exploratory paper, I have modified the previous model to avoid perceiving incoming speech sounds as non-speech sounds. By replacing the initially high-ranked discriminatory constraints with high-ranked PERCEIVE constraints I have created an alternative and functional model of L1 acquisition. This approach acknowledges previous research on L1 acquisition and the new model better accounts for frequencies not present in the original training data. * Acknowledgements: I would like to thank Ania Lubowicz and Randy Fletcher for providing the literature references that led me to choosing this topic. I also want to say thank you to Timothy Hunter for advising this project and helping me through every step of the process. 3 1.0 INTRODUCTION AND BACKGROUND In this paper I have set out to modify an existing Optimality Theoretic model of child language acquisition. In Boersma, Escudero, & Hayes (2003), the authors created a model to show how children learn their native language by warping their perceptual space to account for the sounds necessary in their native language. While the previous model works well, I have attempted to create a different perspective that takes into account other types of research in the field. Optimality Theory (OT) is a linguistic theory that provides a means of analyzing language in terms of ranked constraints instead of rule ordering. Some of the basic motivations behind OT are to create a model for language analysis that is fairly generalizable and accounts for aspects of Universal Grammar (McCarthy, 2007). Instead of a rigid set of ordered, language-specific rules to account for language, OT utilizes a hierarchy of inherently violable constraints that can be rearranged. By having inherently violable constraints, the same hierarchy can account for " exceptions to rules " by showing that the optimal output, or optimal candidate, can still violate aspects of the grammar that are not highly ranked. Another advantage of …