A Naturalist Free Will?
I have been wondering about the goals, advantages, and limits of naturalism. Although for the most part philosophers seem to have dropped crude physical reductionism, naturalism in one form or another continues to occupy a dominant place. I suppose that the major appeal of naturalism is that by sticking to it, one avoids looking like a lunatic; after all, what could be more obvious than that the sciences give us hard truths? These may, of course, be revisable, but it seems ridiculous to suggest that we can ignore empirical evidence and come up with “facts” about reality from the comfort of an armchair (to be honest, I find couches much more comfortable, but those seem to have been reserved by psychoanalysts or, at least, their patients; philosophers in similar predicaments are supposed to use fly-bottles, and let me tell you, those are just not soft enough).
On the other hand, however, naturalism can threaten to reduce philosophy to bad science—that is, if philosophers start to focus primarily on formulating all philosophical theses in ways that open them to possible empirical verification, then strictly speaking there is no more philosophy; there is only a set of hypotheses which, presumably, science will one day be able to check out in order to resolve the problem once and for all. I am not saying that this shouldn’t be one of the things it may be worthwhile for philosophers to pursue, but any thoroughgoing naturalism seems to me to run into the standard Kantian problem: If science alone is the arbiter of truth, then what arbitrates the claim that science is the arbiter of truth? There are different ways of answering this question, but for obvious conceptual reasons, science simply cannot be self-legitimating. Thus, to take science as the ultimate—and sole—arbiter of truth is really to miss something about the facts as we know them, specifically that science is a human enterprise, linked to human projects and modes of reasoning that at some level cannot themselves be open to complete scientific analysis without begging the question.
If rejecting naturalism seems a bit ridiculous, accepting it in certain philosophical contexts can have far stranger consequences. Let’s take as an example the way philosophers commonly deal with free will, which strikes me as a non-naturalistic concept par excellence. One strong trend has been to do everything conceivable to avoid reaching that most horrible and utterly unscientific conclusion of noumenal causation, proposed by Kant. Even those who, like Davidson, want to say that not all meaningful discourse is reducible to physical description, still want to avoid a stance that veers too far from naturalism of some sort. Davidson’s anomalous monism is designed only to suggest how we might meaningfully talk about psychological causality without reducing it to physical causality. This is an important and interesting approach in its own right, but I am not convinced that it helps if one wants to develop a theory of libertarian freedom. Kant, after all, believed that psychological causality, even if it is not reducible to physical causality, still belongs to the domain of determinism because psychological states are phenomena; they are empirical entities, and are subject to the same law-like behavior as all other empirical entities (for Kant, this is the condition stipulated by the Second Analogy).
Of course it is possible that Kant was wrong in his insistence that all phenomena are subject to causal law, and that is in fact the position most commonly taken by incompatibilists: if we have libertarian freedom, then determinism must be false as van Inwagen argued. To accommodate naturalism, then, libertarians have gone to great lengths to devise a concept of agency that fits into a view of the rest of the world as we know it through science. For example, there is the theory of agent causation, i.e., that agents somehow are the causes of their own action. Or there is the attempt to argue that decisions are completely uncaused by prior events, and this is accompanied by a desperate appeal to quantum indeterminacy, which will hopefully somehow match up to our deliberative processes. Perhaps the most desperate, while also the most brilliant and innovative attempt is to hold on to some version of strong determinism but insist that it allows for backward causation—all physical events are caused by previous physical events, but our choices are capable of affecting the entire causal chain, future and past.
This desperation brings out what I think is one of the great advantages of naturalism in philosophy: by trying to squeeze themselves into a bottle, philosophers are forced to come up with increasingly ingenious approaches. But is it wise to be limited to this bottle in the first place? Are these attempts to explain free will really so much more reasonable than the proposal of noumenal causation? Why can’t we postulate a second discourse, irreducible to scientific discourse and yet still meaningful by other criteria? To be sure, it is probably wise to avoid claims that clearly oppose scientific data, and that very loose criterion of naturalism seems perfectly sound. But a properly formulated non-naturalist discourse would be a sort of discourse that could not, in principle, conflict with scientific conclusions. The attempt to formulate such a discourse was the point of Kant’s transcendental idealism, and it was carried over into the various projects of phenomenology. Why, then, should we dismiss the notion of a discourse that was, in a sense, tailor made to account both for free will and for the possibility of scientific cognition?
On the other hand, however, naturalism can threaten to reduce philosophy to bad science—that is, if philosophers start to focus primarily on formulating all philosophical theses in ways that open them to possible empirical verification, then strictly speaking there is no more philosophy; there is only a set of hypotheses which, presumably, science will one day be able to check out in order to resolve the problem once and for all. I am not saying that this shouldn’t be one of the things it may be worthwhile for philosophers to pursue, but any thoroughgoing naturalism seems to me to run into the standard Kantian problem: If science alone is the arbiter of truth, then what arbitrates the claim that science is the arbiter of truth? There are different ways of answering this question, but for obvious conceptual reasons, science simply cannot be self-legitimating. Thus, to take science as the ultimate—and sole—arbiter of truth is really to miss something about the facts as we know them, specifically that science is a human enterprise, linked to human projects and modes of reasoning that at some level cannot themselves be open to complete scientific analysis without begging the question.
If rejecting naturalism seems a bit ridiculous, accepting it in certain philosophical contexts can have far stranger consequences. Let’s take as an example the way philosophers commonly deal with free will, which strikes me as a non-naturalistic concept par excellence. One strong trend has been to do everything conceivable to avoid reaching that most horrible and utterly unscientific conclusion of noumenal causation, proposed by Kant. Even those who, like Davidson, want to say that not all meaningful discourse is reducible to physical description, still want to avoid a stance that veers too far from naturalism of some sort. Davidson’s anomalous monism is designed only to suggest how we might meaningfully talk about psychological causality without reducing it to physical causality. This is an important and interesting approach in its own right, but I am not convinced that it helps if one wants to develop a theory of libertarian freedom. Kant, after all, believed that psychological causality, even if it is not reducible to physical causality, still belongs to the domain of determinism because psychological states are phenomena; they are empirical entities, and are subject to the same law-like behavior as all other empirical entities (for Kant, this is the condition stipulated by the Second Analogy).
Of course it is possible that Kant was wrong in his insistence that all phenomena are subject to causal law, and that is in fact the position most commonly taken by incompatibilists: if we have libertarian freedom, then determinism must be false as van Inwagen argued. To accommodate naturalism, then, libertarians have gone to great lengths to devise a concept of agency that fits into a view of the rest of the world as we know it through science. For example, there is the theory of agent causation, i.e., that agents somehow are the causes of their own action. Or there is the attempt to argue that decisions are completely uncaused by prior events, and this is accompanied by a desperate appeal to quantum indeterminacy, which will hopefully somehow match up to our deliberative processes. Perhaps the most desperate, while also the most brilliant and innovative attempt is to hold on to some version of strong determinism but insist that it allows for backward causation—all physical events are caused by previous physical events, but our choices are capable of affecting the entire causal chain, future and past.
This desperation brings out what I think is one of the great advantages of naturalism in philosophy: by trying to squeeze themselves into a bottle, philosophers are forced to come up with increasingly ingenious approaches. But is it wise to be limited to this bottle in the first place? Are these attempts to explain free will really so much more reasonable than the proposal of noumenal causation? Why can’t we postulate a second discourse, irreducible to scientific discourse and yet still meaningful by other criteria? To be sure, it is probably wise to avoid claims that clearly oppose scientific data, and that very loose criterion of naturalism seems perfectly sound. But a properly formulated non-naturalist discourse would be a sort of discourse that could not, in principle, conflict with scientific conclusions. The attempt to formulate such a discourse was the point of Kant’s transcendental idealism, and it was carried over into the various projects of phenomenology. Why, then, should we dismiss the notion of a discourse that was, in a sense, tailor made to account both for free will and for the possibility of scientific cognition?
No comments:
Post a Comment