Course Description
Linguistic theory seeks to specify the range of grammars permitted by the human language faculty, and thereby to specify the child’s “hypothesis space” during language acquisition. Hence, theories of grammatical variation have immediate implications for the study of child language. Surprisingly, the acquisitional predictions of those theories are seldom tested, presumably because the optimal research methods are not well known. This course shows, step by step, how the longitudinal record of child language acquisition can serve as a major testing ground for theories of grammatical variation, whether in syntax, semantics, phonology, or morphology. \\ Participants will work (either individually or in groups) on lab exercises and small projects using data from the Child Language Data Exchange System (CHILDES), which includes longitudinal corpora of child speech for a considerable number of languages. Statistical methods will be presented for hypothesis-testing with longitudinal data. The primary issues will be testing for concurrent acquisition of two points of grammar, and testing for an ordering effect between two points of grammar. Students will learn how to use Wilcoxon tests and correlation analysis (including partial correlation) on data from groups of ten or more children, and special-purpose non-distributional methods that are suitable for single-child case-studies.
Area Tags: Acquisition, Learning
(Sessions 1 & 2) Tuesday/Friday 3:00-4:20
Location: ILC S415
Instructor: William Snyder
William Snyder (Ph.D. MIT, 1995) is the author of Child Language: The Parametric Approach (OUP, 2007); he served as Co-Editor of the journal Language Acquisition from 2003 to 2012, and co-edited The Oxford Handbook of Developmental Linguistics (OUP, 2015). A Professor of Linguistics at the University of Connecticut, he is a faculty affiliate of both the Institute for Brain and Cognitive Sciences and the Graduate Certificate Program in the Neurobiology of Language.
