Investigating sentence processing and language segmentation in explaining children's performance on a sentence‐span task

Abstract

Children with poor language abilities tend to perform poorly on verbal working memory tasks. This result has been interpreted as evidence that limitations in working memory capacity may interfere with the development of a mature linguistic system. However, it is possible that language abilities, such as the efficiency of sentence processing and the ability to segment language, directly influence performance on common working memory tasks. This study investigated the possible roles of sentence-processing efficiency and the ability to segment language in children's performance on a verbal working memory task. Participants were 37 children (aged 6;2-13;7) with varying oral language and reading abilities. Children completed a sentence-span task to assess working memory: the Competing Language Processing Task (CLPT). In the CLPT, children determine the truth value of sentences while maintaining sentence final words in memory. Sentence-processing efficiency was evaluated by measuring response latencies to CLPT sentence veracity judgments outside the context of the CLPT. The Elision and Blending Words subtests of the Comprehensive Test of Phonological Processing were used to measure the ability to segment and combine units of language. Children's ability to segment words into smaller units (Elision) was a significant predictor of word recall on the CLPT, which is commonly assumed to reflect working memory capacity. Sentence processing latency did not reach significance as a unique predictor. Individual differences on the sentence-span task are partly explained by the ability to segment language into smaller units. Future studies should further consider the metalinguistic and metacognitive demands of tasks used to measure working memory. If metalinguistic abilities directly impact performance on working memory tasks, the utility of working memory theories in explaining individual differences in language abilities is reduced. Inferences that link 'working memory capacity' to language ability become circular. One clinical implication of such a result would be that appropriate intervention strategies may not involve focusing on working memory capacity, but rather on building language skills, including metalinguistic knowledge, which in turn should directly improve children's processing capacities.