Friday, December 17, 2004

A Quasi-Defense of Flew's Quasi-Theism

Somewhat surprisingly, Flew's 'conversion' has occupied a lot of bloggers' attention recently. I think Flew's comments have been rather obscure; but there have been some odd things said in response to them. Julian Sanchez, for instance, says:

What's befuddling is why any of these considerations are supposed to provide any support whatever for the God hypothesis. To think that they do seems to rely on a kind of ignotum per ignotius: We have no satisfying account of complex phenomenon X, so we explain it in terms of, even more complex phenomenon Y, a mind capable of consciously producing X. Why is this supposed to be satisfying? Why, in the absence of a culture in which religion is pervasive, would anyone resort to this kind of explanation? Indeed, why would anyone count it as an explanation at all?


Pressed a bit by Joe Carter in the comments, he says:

Yes, obviously given the assumption of some complex intelligence Y, the account you give of how it might produce X will be more straightforward than a bottom-up emergent account of X. (Though, come to think of it, one doesn't see much speculation on the mechanism by which God does his creating.) But equally obviously, if you're not precomitted to believing in Y, it does no fundamental explanatory work in terms of net complexity. It just shifts the complexity "problem" from X to Y. Sure, if you want to explain complexity in one part of a system--a watch in the desert, a strange machine in space--complex intelligence somewhere else is a viable hypothesis. But, again, if you're not precomitted to coming up with "God" as your answer, it should be fairly transparent that this won't work all the way down, as an account of all complexity.


This seems to assume that Flew's claim is intended as "an account of all complexity," but as Stuart Buck notes in response in the comments, there is no reason to think this is the case, and particularly no reason to think that its not being so somehow makes it unfitting as an explanation (as the original passage seems to imply). One could, perhaps, argue, that for any explanandum and its explanans, what is appealed to in the explanans must be simpler than what is appealed to in the explanandum, but this is obscurantist. For instance, some explanations explain traits of organisms by appealing to populations of organisms + selection pressures; which explanans is necessarily appealing to something more complex than the explanandum. Nor can it be that the explanans itself has to be more simple than the explanandum; for all the features of the explanandum to be explained by the explanans, the explanans has to be at least equal in complexity to the explanandum, because for the explanans to explain all features of the explanandum it must in explaining it tie up to all the features of the explanandum. I really have no notion what else Sanchez could possibly mean. Arguments can be dealt with without making up nonsensical requirements for what counts as explanation.

I do agree with Sanchez that Flew is being very obscure about why he finds his explanans satisfying (if he does; it might just be that he thinks it better than the competition, and nothing else); and it is very obscure, in particular, as to why these considerations suddenly have won Flew over given all the pro-atheism considerations Flew has made a career bringing up. But there seems to be no real problem with "why anyone would count it as an explanation at all".

Although I'm not really interested in Flew himself, never having considered any of his arguments very impressive (but then, despite much posturing on the part of certain sorts of atheists, I've never found any arguments for atheism I considered impressive), I am interested in Flew's new position to the extent that so far it sounds broadly Humean. It is commonly thought that Hume's Dialogues Concerning Natural Religion rejects design inferences; this is an egregious misreading of the text. First, there is no passage in Hume's works in which Hume clearly rejects the design inference. Second, there are several passages in which he seems clearly to accept it. Third, on the Humean analysis of analogical reasoning in the Treatise, which doesn't seem to be rejected in the later work, it is unreasonable to reject analogies. Analogies are irrefutable; it is a waste of time to try to reject them. What one can do is tweak their conclusions by introducing other things that have to be considered. On Humean principles it is, and will always, be the case, the design theist is right that the order of the world will suggest some ordering cause (and what is more, it would follow from Humean principles that this suggestion will get stronger as the sciences advance); the only question is what other things have to be considered in talking about this ordering cause. Fourth, when we recognize this, two supposed 'mysteries' in Dialogues interpretation completely vanish: the confounding of Philo in Part III is seen because he starts out by trying to refute the analogy, and gets trounced by Cleanthes because of it. Demea, however, opens the argument again by criticizing what Cleanthes thinks is the result of the inference rather than the inference itself, and Philo trounces Cleanthes by taking Demea's lead here. And then the alleged 'reversal' of Philo in Part XII turns out not to be a reversal at all. And so forth.

Now, this Humean admission of the value of the design argument is, like Flew's very weak and vague; and this is a result of the weakness of the inference. Lindsay at Majikthise describes Flew's position (as stated so far) as claiming the existence of a "supernatural non-conscious intelligent design force". And this seems fair enough. It would be a good description of Hume's position, too, although he prefers just to call it "invisible intelligent power" or "first intelligent Author". Hume avoids all contradiction, however, because of the nature of the inference with which he is working. We don't really know whether Flew's argument has similar results, because, as Lindsay says, he's being very coy. All we know is that it is some sort of design argument having to do with complexity. There is nothing in this that implies that it is what has recently become "Intelligent Design theory"; it might be, but there are dozens of different forms a design inference can take. There is nothing in it that implies that is being put forward as a scientific hypothesis, rather than a philosophical argument starting from a set of scientific facts (whatever a few naturalists might want to think, their own position is of this general sort). Contrary to some, there is nothing in this that implies that there is an appeal to an "omnipotent Creator" going on, unless one uses the terms extremely loosely (likewise, it doesn't seem to be the case, contrary to some, that Flew has accepted the fine-tuning argument, since he contrasts that argument with whatever one he holds). And so forth. It's really quite surprising to me how many people are willing to pronounce Flew's argument bad when they confessedly don't know what it is. And a lot of people, who have nothing better to say on the subject, throw out empty rhetorical clichés or dismiss Flew, without evidence, as doddering, or senile, or 'a fearful old man'. But the commentary hasn't been totally along these lines, fortunately. I think Yglesias's point is perceptive and interesting, for instance; it suggests yet another way in which Flew's position should be considered broadly Humean, because Hume makes the same point (in Dialogues XII). If Flew is proposing something more or less along these lines, a broadly Humean position on this point is a respectable philosophical position, despite some occasional odd attempts to run a smear campaign against it without bothering to figure out what it is. It's a position that I'm certain fails, due to a rather hefty number of things that can be said against it; but it's a respectable position, with a number of things to be said for it. [But of course, to make the point again, we really don't know much about what Flew's inference is supposed to be, just its starting-point and its conclusion. -ed.]