Children often make early commitments to sentence structure but fail to revise after late-arriving conflicting evidence (Trueswell et al.,1999). While immature cognitive control is thought to contribute to parsing errors, correlations between syntactic and Stroop-like tasks remain inconclusive (Novick et al., 2005; Woodard et al., 2016; Huang & Hollister, 2019). In adults, real-time cognitive-control engagement improves revision of sentences like “Put the frog on the napkin into the box" (Hsu & Novick, 2016). Yet for children, these paradigms lead to less accurate revision (Huang et al., 2016). One explanation for this discrepancy is that children’s immature cognitive control is easily depleted, leaving fewer resources for revising sentences. Alternatively, cognitive-control engagement may be shifting which parsing cues children rely on. "Put” requires a location, but relying on this causes mis-predictions for these particular sentences. I'll present some evidence that cognitive control engagement doesn't always lead to resource depletion, and present a case where it helps children ignore an unreliable parsing heuristic and attend more to role information from verbs.