Lexical revisions and filled pauses : associations with vocabulary knowledge in bilingual children

Purpose: The current study examined lexical revisions and filled pauses as related to vocabulary knowledge and language experience in bilingual children with and without language impairment. Method: Participants included Spanish-English speaking bilingual children (n=30) aged 84-103 months. Children were designated into one of three language groups based on language ability (typically-developing, language-impaired, and at-risk typically-developing). Narratives from the Test of Narrative Language were transcribed and coded in Spanish and English for lexical revisions and filled pauses. Lexical revisions and filled pauses in each language were then correlated with measures of vocabulary knowledge provided through the narrative sample (MLU and NDW) and standardized testing (EOWPVT and BESAME Semantics scores). Results: The current study found that the typically-developing group exhibited significantly more lexical revisions in Spanish and significantly more filled pauses in English and Spanish compared to the at-risk typically-developing and language-impaired groups. NDW significantly correlated with total maze use, lexical revisions, and filled pauses in both English and Spanish. No significant correlations were observed between language dominance or language exposure to either lexical revisions or filled pauses. Lexical revisions and filled pauses were significantly correlated to each other within each language, and significantly correlated across the languages. Conclusions: Results suggest that lexical revisions and filled pauses are more closely related to vocabulary knowledge than to language dominance or exposure. Lexical revisions and filled pauses were produced more in children with higher levels of vocabulary knowledge as measured by NDW as opposed to MLU or standardized test measures. Lexical revisions and filled pauses demonstrated correlations across and within languages, indicating that language experience does not seem to play a role in their use.