Some of the beings we encounter in our environment are inanimate. These things may be pushed and pulled, they may collapse or disintegrate. In each of these cases, the entities are fundamentally passive--if they move or change, one suspects that these movements and changes will be exhaustively explained by appealing to mechanical forces within or without.

But there is another kind of entity in our environment. These beings seem to be fundamentally goal-directed. To appearances, they are spontaneous initiators of their own actions. These are animate beings, or agents. The movements of these entities seem to be best explained not by appeal to mechanical causes of their activity, but to the goals that they are striving towards, the beliefs they have about the world, and their desires.

Giving a precise definition of what animacy or agency consists of is no easy task for the philosopher, but nonetheless we appear to have no difficulty at all recognizing animate motion and distinguishing it from the motion of inanimate objects. Even human infants, it seems, can detect animate motion and differentiate it from inanimate motion in point-light displays, even when occlusions are present.

But why should this be the case? Why would it be so difficult to give a theory of intentional action, and yet so easy to detect it? I will set out one suggestion. We do not, as has been proposed, infer intentions, beliefs, and desires as a part of a theory for explaining or predicting behavior. Instead, intentional action is behavior with certain distinctive, overt characteristics, which our perceptual systems have evolved to directly perceive. Goal-directed behaviors, I will suggest, are a very real kind of behavior out there in the world with distinctive characteristics. Furthermore, it is important that animals perceive and understand this particular kind of behavior, and sure enough, they are able to do so with astonishing acuity.

What are we saying when we explain the activities of another person (or of a non-human animal) by appealing to their intentions? Here I will draw on an analysis from Dennis Walsh: “A teleological explanation is one that explains the nature or activities of an entity, or the occurrence of an event, by citing the goal it subserves. A system has goal, E, just in case it exhibits goal-directed behavior toward E. Goal-directed behavior is a gross property of a system as a whole.” (p. 177)

What this amounts to is not an account of the intrinsic causal etiology of the agent’s behavior. Instead, we are locating that behavior in a chain of events that show a certain distinctive pattern. If an agent is trying to do X, then its behavior will flexibly reconfigure itself in the service of that goal. When a dropped object encounters the ground, it will stop. When an agent’s initial attempts to pursue some goal are thwarted, that agent will spontaneously and flexibly reconfigure its behavior so as to continue to pursue its goal. A human need not stop at the ground--they can retrieve a shovel, and perhaps a jackhammer or a drill if called for (if they really want to!) This is to say, when an agent is engaging in goal-directed activity, its behavior is robust against perturbations and obstacles in a way characteristically not present in inanimate objects.

If there are such systems in nature--systems that will reliably produce effects by marshaling their intrinsic causal capacities in the service of goals--then clearly the perceiving animal would be at an advantage if they could detect them when present! The challenge, from the perspective of an animal’s perceptual system, then, is to detect or pick up the information which specifies what the goals of other agents in their acting are. While this might sound like quite a feat, again, this is something we all seem to be very good at.

If agency amounts to the capacity for intentional action, and the preceding account of goal-directed behavior is sound, what basis might there be to deny that such a thing exists as a real phenomenon in nature, and a real attribute of natural beings?