Beverly Anne Collisson, Bernard Grela, Tammie Spaulding, Jay G. Rueckl and James S. Magnuson
In a classic shape-bias paradigm, three and four-year-old children with typical language (TL) or Specific Language Impairment (SLI) saw an example object and then chose another from a set of three in Similarity Classification (‘See this? Which one goes with this one?’) and Novel Name Extension (‘See this? This is a [dax]', Find another [dax]') conditions. Children with TL replicated the classic pattern (random choices in Similarity Classification, large bias for shape in Name Extension), but those with SLI showed no apparent difference. This suggests that children with SLI fail to detect coherent covariation between linguistic and nonlinguistic properties that accelerates object name learning among their peers with TL. In a paired visual association task, children with TL showed significant learning over 4 days, but children with SLI did not. Performance in this task predicted individuals' shape bias better than any assessment, suggesting impairments of nonlinguistic learning abilities contributes to linguistic difficulties in SLI.