Two-year-olds' difficulty with rule execution is robust and pervasive. In Experiment 1, 120 32-month-olds received 1 of 6 tasks: a task assessing knowledge about a series of items; a deductive card sort requiring children to use their knowledge to sort the items by rules; and 4 modifications of the card sort that provided various types of task support. Children performed better on the knowledge task than the other tasks, which did not differ. Errors increased over trials and were typically perseverative. Experiment 2 replicated the finding that children failed to use rules systematically even when items were labeled by the rules' antecedents. Improvements in rule use seem to reflect emerging control over actions rather than increased representational flexibility.