Purpose: In this study, the authors examined the role of syllabic structure in nonword repetition performance in typically developing (TD) children and children with specific language impairment (SLI). Method: Eighteen children with SLI (5;7--6;7 [years;months]) and 18 TD children matched for chronological age were tested on their ability to repeat phonemes in different positions within syllable structure (onset, nucleus, coda). The test involved 2 separate nonword repetition tasks differing in lexicality (high vs. low). High-lexicality nonwords contained subparts that are morphemes of the language (i.e., subparts were "lexical"), whereas nonlexical nonwords did not. Results: Repetition performance across the 3 syllabic positions showed a significant effect for both populations and on both tasks. However, although on the high-lexicality task the direction of the effect revealed the onset as the most error-prone constituent (onset greater than coda greater than nucleus), on the low-lexicality task, it was the coda that attracted most errors (coda greater than onset greater than nucleus). Conclusions: Results suggest that the procedures involved in computing syllabic structure are qualitatively similar in the 2 populations. We take these results to support the view that different syllabic positions involve different levels of phonological complexity and that tests that control for lexicality are crucial in illuminating these differences. (Contains 4 tables and 4 figures.)