When seeing someone’s arm reach for a hot cup of coffee we can predict that their hand will make contact with the cup by simply extrapolating the arm’s current path of motion into the immediate future. This simple process can be employed to anticipate the movement of many different types of object (e.g. cars, football).
However, human movement is governed not only by physical causes and constraints, but also by the goals and intentions of the actor, which can radically change our interpretation of their action and thus our prediction of how it will end. These mental states are not readily observable but must be inferred from contextual cues. For example, if our friend is looking at the hot cup of coffee whilst reaching for it, we can anticipate that he will pick it up and bring it to his mouth. However, if he is not looking at it, we can infer that he does not know it is there and that he will most likely spill it.
Therefore, although the actions themselves are the same in both situations, the actor’s gaze direction provides a cue regarding the goals and intentions of the actor and will affect how we predict the outcome of the action. Additional cues such as their emotional expression, the type of object, the shape of their hand, or any tools they may be holding may also provide an observer with information regarding what the other person intends to do. This is important as more accurate and efficient predictions enables us to quickly modify our behaviour accordingly, by either passing the cup of coffee to their friend, or moving the cup of coffee out of the way.