Recently, the blogosphere has had a number of posts asserting that we have no free will, and claiming to prove that. One is tempted to ask whether the posters make this claim of their own free will (which they deny having), and if it is not made of their own free will one might wonder whether to take them seriously. In one such post, Jerry Coyne asked “So why don’t computers have free will?” In a comment to that post, I suggested that it was because computer systems are based on the “rational agent” model of behavior. Here, I shall expand on that point.
A rational agent is an agent who acts rationally. According to the accounts of rationality given by philosophers, a rational agent has desires (or goals), has beliefs (or knowledge), and receives input from the world. The rational agent then applies logic to input and beliefs in order to determine the behavior (or outputs) most likely to achieve those desires. Some versions, such as the Minimal Rationality of Cherniak, admit that the logic required for rationality might be intractably complex, and allow the use of heuristic shortcuts to make the most of the limited computational ability of the agent.
Under a rational agent model, decisions are either based on truth or on a random decision (when the rational model calls for a random choice). So, either way, it would seem that the decisions of the rational agent are either determined by inputs and knowledge, except when they are random. If randomness is handled by a deterministic pseudo-random number generator, then the decisions would be always determined by the inputs. Whether or not we live in a deterministic universe does not matter. The behavior of the rational agent is still determined.
The compatibilist view is that we choose to act rationally, and thus we have free will on the basis of our choices. That our behavior is deterministic is a result of our choice to act rationally. The incompatibilists claim that we have no free will.
For myself, I don’t see us as rational agents. Psychologists report that people are not rational, but that is not my concern. Rather, the rational agency model seems to misdescribe how we fit into our world. It seems implausible that we could have evolved as rational agents. If we were really rational agents, our species should have gone extinct long ago. To a first approximation, rational agency depends on living in a stable world where you can carry out rational planning for the future. But our world is far too dynamic for that to be a plausible model of behavior.
I now want to suggest a different model of behavior, one that I believe better describes how we fit into the world. The model is that of what I shall call an opportunistic agent. It is apparently due to J.J. Gibson, who is best known for his theory of direct perception.
According to this model we do not receive inputs. Rather, we seek and retrieve information from the world. That is, we are in charge of what information we receive, and are not merely helpless recipients of information. As an example, when we pay attention to something in the enviroment, we are concentrating our information gathering on what we are attending to.
In seeking information, what we are particularly seeking is information on opportunities that are available to us. Gibson called these “affordances”, for they afforded us opportunities. Then we select between the available opportunities on the basis of pragmatic judgment (what will work best for us).
We can compare the two models, based on the choices made. In the rational agency model, choices are made on the basis of truth. And since truth is normally seen as person independent, these choices are not really choices but are dictated by the requirement of rationality. In the opportunist agency model, the choices are more abundant because we are seeking an array of opportunities. And we make our choices pragmatically. So the choices are real, since pragmatic judgment is normally seen as person relative, so not the same for everybody.
It seems to me that our notion of having free will fits pretty well with the opportunistic agency model. Some people want to define “free will” as the ability to violate the laws of physics, and of course we do not have free will in that sense. But I think we do have it in the sense of pragmatically choosing opportunities from among those that we perceive as available to us.
The Heretical Philosopher 