Spoken word recognition by children with and without specific language impairment : an examination of lexical characteristics

The current study had three aims: 1) To examine spoken word recognition by children with and without specific language impairment (SLI), 2) To investigate the effects of two lexical characteristics, age of acquisition and neighbourhood density, on spoken word recognition, and 3) To explore possible links between spoken word recognition and phonological short-term memory or semantic knowledge. Eight school-aged children with SLI and nine typically developing children matched for comprehension vocabulary completed three experimental tasks. First, they listened to 28 common nouns presented in a gating paradigm. The 28 words varied along the dimensions of age of acquisition (early- versus late-acquired) and neighbourhood density (words from sparse versus dense phonological neighbourhoods) so that there were four types of words: early/sparse, early/dense, late/sparse and late/dense. Participants were asked to identify the word being presented. Second, participants completed a word-picture matching task (from a choice of four pictures) that measured the depth of their semantic knowledge of the same 28 words. Third, children repeated nonsense words that varied from one to four syllables in length as a measure of their phonological short-term memory. Results from the gating task found no significant difference between groups. For both groups, children recognized early-acquired words from sparse neighbourhoods more quickly than early-acquired words from dense neighbourhoods and late-acquired words from both sparse and dense neighbourhoods. A significant correlation was found between the gating task and the nonsense word repetition task for all participants. The longer the phonological short-term memory span, the more quickly words were recognized for all children. A strong and marginally significant correlation was found between the gating task and the word-picture matching task for the children with SLI only: the higher the number of words they could correctly match to a picture the more quickly words were recognized in the gating task. These results suggest that children with SLI do not have a deficit relative to vocabulary-matched peers in the recognition of single, spoken, words. Furthermore, the current findings provide some degree of support for a relationship between spoken word recognition and both phonological short-term memory and depth of semantic knowledge.Medicine, Faculty ofAudiology and Speech Sciences, School ofGraduat