Eye gaze is an important communicative signal, both as mutual eye contact and as referential gaze to objects. To examine whether attention to speech versus nonspeech stimuli in 4- to 5-month-olds (n = 15) varies as a function of eye gaze, event-related brain potentials were used. Faces with mutual or averted gaze were presented in combination with forward- or backward-spoken words. Infants rapidly processed gaze and spoken words in combination. A late Slow Wave suggests an interaction of the 2 factors, separating backward-spoken word + direct gaze from all other conditions. An additional experiment (n = 15) extended the results to referential gaze. The current findings suggest that interactions between visual and auditory cues are present early in infancy.