Children performed well on tasks involving pictures regardless of whether the pictures were named. Performance on atypical pictures was better when category labels were provided than when labels were not provided. A control study demonstrated that children ignored labels when they named a transient, as opposed to a stable, category. (RH)