We automatically represent others actions in terms of their goalfor actions with concrete external goals (e.g. reaching for an object; Woodward, 1998). Is this true for actions without concrete external goals? We hypothesized that movement itself could be considered a goal, and that this inference forms part of concepts like dance, exercise and ritual. In a between-subject experiment, participants saw either (a)movements with no external goal (animated character moving in an empty room); or (b)identical movements, with an external goal (manipulating objects). We first determined what goal participants inferred, if any. Then, we asked participants if the actions could be dance, exercise, or ritual. Nearly half of participants in the no-external-goal condition inferred that the movements themselves were the goal. Participants who inferred movement-as-goal later rated the actions as more likely to be dance, exercise, or ritual, even compared to other participants who saw the exact same stimuli.