A corpus-based study to evaluate the generativist explanation of children’s error patterns in questions

Abstract

This study explores whether the generativist account, specifically the integration theory, could explain children’s percentage of errors in questions in general and whether it also applies to yes-no and non-subject wh-question. The current study adopts a corpus-based method to compare 2-to-3-year-old children’s percentages of errors in questions (and in yes-no and wh-question separately) including auxiliary DO and auxiliary HAVE. The results show that children’s rate of errors in questions including auxiliary DO is higher than that including auxiliary HAVE, which is also applicable to yes-no and non-subject wh-questions. The findings indicate that the generativist theory of child language acquisition could successfully explain children’s patterns of errors in questions. This study also emphasises the impact of the question type which should be carefully considered when constructing and improving the generativist theory of child question formation. The study provides empirical evidence for improving and refining the generativist account of child language acquisition generally and language question acquisition specifically.