Eye tracking was used to show that 18-month-old infants are sensitive to social context as a sign that others’ actions are bound together as a collaborative sequence based on a joint goal. Infants observed five identical demonstrations in which Actor 1 moved a block to one location and Actor 2 moved the same block to a new location, creating a sequence of actions that could be considered either individual actions or collaboration. In the test phase, Actor 1 was alone and sitting so that she could reach both locations. The question was whether she would place a new block in the location she had previously (individual goal) or in the location that could be considered the goal of collaboration (joint goal). Importantly, in the Social condition, the actors were socially engaged with each other before and during the demonstration, while in the Non-Social condition, they were not. Results revealed that infants in the Social condition spontaneously anticipated Actor 1 placing her block in the joint goal location more often than those in the Non-Social condition. Thus, the social context seems to allow infants to bind actions into a collaborative sequence and anticipate joint rather than individual goals, giving insight into how actions are perceived using top-down processing early in life.