Children resolving spoken mispronunciations: the role of sentence context

Purpose: Theories of set for variability identify a role for sentence context within the process of resolving mispronunciations during reading (Share, 1995; Venezky, 1999), yet prior work has focused on mispronunciations of isolated spoken words. The primary aim of the current study was to examine whether sentence context boosted correction accuracy, compared to isolated mispronunciations, and to what extent this benefit derived from syntactic and semantic support. A secondary aim was to better understand the mispronunciation correction process by examining predictors of correction success. Finally, correlations explored the relationship between the mispronunciation correction tasks and standardised reading and letter-sound measures. Method: Sixty-two Year 1 children completed three oral mispronunciation correction conditions: isolated mispronunciations; mispronunciations that followed syntactically supportive sentences; and mispronunciations that followed syntactically and semantically supportive sentences. Child-, word- and child-by-word measures were also administered. Results: A logistic linear mixed effects model predicting correction accuracy revealed better performance when context was present than when it was absent, and a small but significant benefit of greater semantic support from context. Accuracy was predicted by features at the child- (word reading), word- (transparency rating; frequency), and importantly, at the child-by-word levels (item-specific reading, lexical phonology and semantic knowledge). Compared to the isolated and syntactic conditions, syntactically and semantically supportive contexts exhibited a weaker relationship with word reading. Conclusion: Context assists with correcting mispronunciations, pointing especially to an enhancing role of semantic support, and supporting the existence of a direct relationship between item-specific word knowledge and mispronunciation correction success.