Grasping is a fundamental skill that enables people to interface with and explore objects around them. The emergence of precision (thumb‐to‐forefinger) grasping during infancy represents a developmental shift in this skill and has been linked to more advanced action perception, particularly in detecting action incongruencies. In this study, ERPs known to be elicited in response to action were studied in 9‐ and 11.5‐month‐old infants as they watched whole‐hand and precision grasping actions congruent or incongruent with a target object. Components related to attentional (Nc, P400) and semantic (N400) processes were examined to determine whether infants' perception of grasp is based on attention and recognition, on higher‐level representations of action, or a mix of these two levels of processing. Effects of congruence were found for the P400 and the N400. The P400 effect was greater among the older age group. Infants' ability to produce a precision grip did not significantly affect their ERPs in response to actors' incongruent versus congruent grasps, which would have been expected if recognition of incongruous grasping actions were based on motor experience. Results indicate that infant ERPs differ between grasps that are congruent or incongruent with the form of a target object via multiple cognitive processes.