Studies claiming to have demonstrated facial imitation in the neonatal period are often cited as evidence that rudimentary representational abilities are present at birth. A critical examination of nine studies with infants ranging in average age from 2 days to 7 weeks demonstrates that the claims of early imitative abilities are not well founded. All nine studies investigated the imitation of tongue protrusion (TP) and one other gesture: head movement (HM) in one study and mouth opening (MO) in eight studies. Two studies acknowledged that there was a modeling effect for TP but not for MO. The remaining studies claimed to have obtained effects not just for TP but also for the other gesture investigated (MO or HM). But a close examination of these studies shows that in contrast to the TP effect, which was clear and consistent, the MO and HM effects were weak and inconsistent and in some cases derived from the TP effect. Because the matching behavior found is restricted to a single gesture, it is best explained as a specific, directly elicited response, rather than imitation.