This study compared school-age outcomes for 57 children born prematurely, at ≤32 weeks or <1500g, with outcomes for 57 children born at full term with no reported perinatal complications. The two groups were matched for age, sex, race, and parental education. Data came from the Western Reserve Reading Project and included discourse-level language samples collected at three points in time, each a year apart. In addition, standardized test results were obtained for IQ, digit span, and global language ability. The language samples were analyzed to yield a number of semantic and syntactic measures which were condensed via factor analysis to a semantic score and a syntactic score. Regression models showed statistically significant differences between the two groups for standardized test results, with more ambiguous results for the discourse-level language measures. The control group outperformed the premature group on both semantic and syntactic measures, but those differences never reached statistical significance and narrowed markedly at the third-year assessment point. These findings suggest that in the absence of frank neurological impairment, sophisticated semantic and syntactic skills may be relatively intact in the conversational and narrative language of children born prematurely. The decrements observed on standardized assessments of language and cognition may arise from deficits in domains such as attention or executive function, rather than reflecting significant impairment in their ability to learn language.