Sentence Imitation with Masked Morphemes in Czech: Memory, Morpheme Frequency, and Morphological Richness

Purpose: This study examined two markers of language impairment (LI) in a single experiment, testing sentence imitation and grammatical morphology production using an imitation task with masked morphemes. One goal was to test predictions of the morphological richness account of LI in Czech. We also tested the independent contributions of language and memory skills to sentence imitation performance. Method: Seventeen children with LI (5;1-7;6 [years;months]) and 17 vocabulary-matched typically developing (TD) children (3;8-4;11) were administered a sentence imitation task where each sentence had one noun or verb ending replaced by a coughing sound. In addition, a receptive vocabulary and the digit span (backward and forward) tasks were administered. Results: Children with LI were significantly less accurate than TD children in sentence imitation task. Both vocabulary and digit span had unique effects on sentence imitation scores. Children with LI were less successful in imitating the target words, especially verbs. However, if they succeeded, their completions of the masked morphemes were no less accurate than in TD children. The accuracy of completions was affected by the morpheme frequency and homophony, but these effects were similar in TD and affected children. Conclusions: Sentence imitation is a measure of language skills and verbal memory. Results on morpheme completions are consistent with processing models of LI, but some predictions of the morphological richness model were not confirmed. The results suggest that children with LI might have a deficit in organizing morphosyntactic relations in sentences, rather than in morphological processing proper.