The sensitivity of children with SLI to phonotactic probabilities during lexical access

Abstract

The procedural deficit hypothesis (Ullman & Pierpont, 2005) has been proposed to account for the combination of linguistic and nonlinguistic deficits observed in specific language impairment (SLI). According to this proposal, SLI results from a deficit in procedural memory that prevents children from developing sensitivity to probabilistic sequences, amongst other deficits. We tested the ability of children with SLI to rely on a specific type of probabilities characterizing sequences that occur in a given language: phonotactic probabilities. Twenty French-speaking children with SLI (M=10;1), 20 typically developing children matched for chronological age (M=10;0) and 20 typically developing children matched for receptive vocabulary (M=7;4) performed an auditory lexical decision task. Pseudoword stimuli were built with combinations of either frequently associated phonemes (high phonotactic probability) or infrequently associated phonemes (low phonotactic probability). Phonotactic probabilities had a significant impact on the accuracy and speed of pseudoword rejection in children with SLI, but not in the two control groups. SLI children's greater reliance on phonotactic probabilities relative to typically developing children appears to contradict the PD hypothesis. Phonotactic probabilities may help them to partially overcome their difficulties in developing and accessing the phonological lexicon during spoken word recognition.

Learning outcomes: After reading this article, readers will understand the importance of sensitivity to phonotactic probabilities in language processing. They will also learn that such sensitivity is preserved in children with SLI. Finally, readers will understand that children with SLI are more prone to use phonotactic information when accessing their lexicon than typically-developing children.

Keywords: Auditory lexical decision; Phonological representations; Phonotactic probabilities; Procedural deficit hypothesis; SLI; Statistical learning.