I love making up new words. So here's a neologism for you. If a preamble is something that "walks before," then why not have a word for that which walks behind? So this post is a postamble to Andrew Monroe's post on the
Free will illusion illusion.
As I struggled with the concept of free will in several posts (e.g.), I found myself taking an ever hardening determinist point of view - and I stand by it. It is clear to me, and Andrew seems to agree, that the metaphysical notion of free will (i.e., a will that is itself uncaused) must be false. What remains is the issue of what to do with the concept of moral (and legal) responsibility. Andrew's work shows that ordinary people don't worry much about metaphysical arguments. They understand and use the idea of free will in a functional way. They use it, in other words, because "it works."
The key to Andrew's argument, as I understand it, is the notion of intentionality. Philosopher Dan Dennett coined the phrase "intentional stance." Humans have it and they have a hard time setting it aside. Andrew notes that the intentional stance develops with enthusiasm during early childhood. This invites the inference that evolution has selected it for its functional value. A mental module that performs a particular function ideally does so only when there is an actual need or cause. In a complex world, few mental modules perform perfectly in that sense. The engagement of a module is a decision under uncertainty, which entails two types of error: A false positive occurs when the module responds when it should not; a miss occurs when the module fails to respond when it should respond. The relative probability of these two errors tells us something about their utility.
Applied to the intentional stance, one view is that the intention-detection module is overactive. Humans see a lot of intent where there's none. We readily, instinctively, and enthusiastically conclude that dogs, gods, and gadgets do what they do because they want to. The decision-theoretic interpretation is that these errors can be tolerated because it would be so much worse if the intention-detection module were under-active, i.e., if we missed the intentionality of actions that were in fact intentional. This analysis, interesting as it is, begs the question of what intentionality is and who has it. Is it really an error to assume that dogs or apes act intentionally? If they do act intentionally, what does it mean and how do we know? Conversely, if we deny dogs or other interesting non-human animals (e.g., dolphins) intentionality, we commit ourselves to human exceptionalism. Intentionality emerges as a true (perhaps the true) distinguishing feature of humans. If so, how do we know that we have intentionality beyond just claiming that we do because that's what it feels like? If all attributions of intentionality are false, wouldn't it be the most fundamental of all errors to assume that someone acted intentionally?
If evolution saw to it (pardon the intentional-stance-like phrasing) that we humans see intentionality in many places, why did it? What are the practical or pragmatic advantages? The idea that it is a good thing to be able to "explain" behavior doesn't cut much ice. One might explain behavior without reference to intentionality. More interesting is the notion that attributions of intentionality are adaptive because they pick up on important invariance (as all good predictions do). Predictions work when they capitalize on invariances that were discovered in the past and that generalize into the future. The irony is that the more these predictions succeed, they more they undermine the metaphysical concept of free will, which, by definition, is unpredictable.
Some, e.g., Dan Wegner (see his book The illusion of conscious will), feel they have proven that intentionality doesn't do anything. This view implies that the decision-theoretic analysis of intentionality breaks down. All inferences about intentionality are false positives. I am not ready to go there (yet) and I look forward to the progress made in research programs such as Andrew's.
