Characterizing language impairment in children

This study compares expressive language samples from normal children ( n = 20) and language impaired children ( n = 9), matched for age and intellectual capacity, across a range of grammatical and lexical dimensions. The aims of the inquiry are to determine if it is possible to characterize language impairment (so far as it pertains to the carefully selected group of children in this study) using such dimensions, and in particular to see if the identification can be made on the basis of a subset of the large set of categories used in the original descriptive framework. An initial comparison of incidence scores based on samples of 200 utterances from each child, for 65 grammatical and lexical categories, indicates that only a third of the categories are relevant to the characterization of impairment. A discrimin ant function analysis using a subset of these caregories identified two variables, which, taken together, were reasonably successful in discriminat ing the two groups. The categories identified relate to verb premodification and inflection, and lexical verb type frequency. The implications of these findings are discussed.