Contributions of Phonological and Semantic Short-Term Memory to Sentence Processing: Evidence from Two Cases of Closed Head Injury in Children Gerri Hanten and Randi C. Martin Rice University This paper reports the short-term memory and sentence processing performance of two head- injured children who were approximately 10 years of age at the time of test. The two children were found to have similarly reduced memory span on a variety of short-term memory tasks relative to age-matched control children, but differing patterns of performance. One child showed reduced (or absent) effects of phonological variables, and the other showed reduced effects of semantic variables. A dissociation in their performance on sentence processing tasks was also observed, with one child showing preserved comprehension but impaired repetition, and the other child showing the reverse. The results are consistent with previous findings from brain-damaged adults supporting a multiple- components model of short-term memory in which phonological and semantic components play different roles in sentence processing. Contrary to our initial hypothesis, the phonological component did not appear to play a more important role in sentence comprehension for children than for adults. © 2000 Academic Press Key Words: short-term memory; sentence comprehension; sentence repetition; multiple compo- nents of STM; language development. Evidence from normal and brain-damaged adults suggests that there are multiple compo- nents of short-term memory and that these com- ponents play different roles in sentence compre- hension (Martin & Romani, 1994; Martin et al., 1994). This multi-component approach has not been previously applied to studies of short-term memory and sentence processing deficits in children. This paper reports the short-term memory performance of two head-injured chil- dren who were approximately 10 years of age at the time of testing. The issue was whether a multicomponent approach could account for their short-term memory patterns and whether the relations between short-term memory and sentence processing were the same for these children as for adults. Some might question whether 10 year old children would be predicted to behave any differently than adults, given the widespread assumption that much of language development is complete at a very early age. For example, Lenneberg (1967) stated that the bulk of language acquisition was complete by age four. However, more recent research indi- cates that language development continues into adolescence (ages 13 through 19) in a variety of language areas, including syntactic and lexical acquisition (Nippold, 1988). Although the changes occurring in older children may be more subtle than those observed in preschool children, they have been well documented (e.g., Scott, 1988; Holcomb et al., 1992). In addition, general increases in speed of processing across cognitive domains have been documented through late adolescence (Kail & Salthouse, 1994), with large increases occurring between ages 10 and 18. Thus, because 10-year-old chil- dren may have less efficient language process- ing mechanisms or generally slower speed of processing than adults, it is possible that com- This work was supported in part by National Institutes of Health grants to Dr. Harvey Levin at Baylor College of Medicine (NS21889) and to Dr. Randi Martin at Rice Uni- versity (DC 00218), and by the Maurin Fund for research in congitive psychology to the Department of Psychology at Rice University. Address correspondence and reprint requests to Randi C. Martin, Dept. of Psychology, Rice University, P.O. Box 1892, Houston, Texas 77251. E-mail: rmartin@rice.edu. 335 0749-596X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved. Journal of Memory and Language 43, 335–361 (2000) doi:10.1006/jmla.2000.2731, available online at http://www.idealibrary.com on