Two experiments were conducted to evaluate symbolic-deficit andmemory-deficit hypotheses to account for the cognitive problems seenin children with autism. Experiment 1 tested imitation, in immediateand deferred conditions, of familiar actions with different sets ofobjects representing the developmental progression from functional tosymbolic play. The results showed that the autism group and both theirreceptive language and nonverbal IQ-matched controls imitated familiaractions with realistic objects (evidence for functional play) andplaceholder objects (evidence for symbolic play) after delays rangingfrom 24 hr to 3 weeks. Experiment 2 tested familiar three-step eventsequences in which a placeholder object was substituted for the secondstep in half the events. The results showed that the autism groupremembered as many of the actions with the placeholder objects astheir language-matched controls and as many correctly orderedsequences, a finding that supports a symbolic-delay (rather thandeficit) hypothesis. These results were obtained in highly structuredtest situations and sharply contrast with the impairments seen inchildren with autism who are observed in naturalistic settings. Twointerpretations of these findings are offered. First, structured testsettings minimize distractions that typically occur in naturalisticsettings that may interfere or disrupt symbolic play in children withautism. Second, the results are consistent with an executive functiondeficit in that the autistic group demonstrated more knowledge in thetest settings than they demonstrate spontaneously in naturalisticones.