Comparability of lexical corpora: Word frequency in phonological generalization

Statistical regularities in language have been examined for new insight to the language acquisition process. This line of study has aided theory advancement, but it also has raised methodological concerns about the applicability of corpora data to child populations. One issue is whether it is appropriate to extend the regularities observed in the speech of adults to developing linguistic systems. The purpose of this paper is to establish the comparability of lexical corpora in accounting for behavioural effects of word frequency on children's phonological generalization. Four word frequency corpora were evaluated in comparison of child/adult and written/spoken sources. These were applied post-hoc to generalization data previously reported for two preschool children. Results showed that the interpretation of phonological generalization was the same within and across children, regardless of the corpus being used. Phonological gains were more evident in low than high frequency words. The findings have implications for the design of probabilistic studies of language acquisition and clinical treatment programmes.National Institutes of Health DC00433, RR7031K, DC00076, DC001694 (PI: Gierut)This is an Accepted Manuscript of an article published by Taylor & Francis in Clinical Linguistics & Phonetics on June 2007, available online: http://wwww.tandfonline.com/10.1080/02699200701299891