It is popularly believed that if a theory has theological implications, then the theory is somehow "unscientific." A post (NOTE: MovableType won't let me link directly to this post because the URL contains an unescaped ' contrary to the HTTP spec so the above link goes to the daily archive) at the Florida Student Philosophy Blog challenges this claim. I think the article is unnecessarily long and involved, but I'm quite impressed with the insight. The argument is a reductio that works more or less like this:
That's rather a beautiful proof, isn't it? Of course, the proponent of the strict separation of science and theology is free to respond by further refining his notion of 'implication' to mean something narrower than logical implication, but I've never heard any such account.
Update (10:20 PM PDT): As Richard (of Philosophy, etc.) points out below, the comments to the linked post are worth reading and discuss precisely the point I just mentioned: the fact that the best response for the opponent to make is to simply state that he doesn't mean logical implication (the modal stuff is not needed - if God exists, then any proposition implies that God exists in the strict logical sense; I should have seen this immediately, and I think it makes the argument slightly less impressive, though the discussion is still of interest). He should set up something else as what he really means by implication. My own suggestion (similar to some of those made in the comments) is that he might use something like a Bayesian theory of evidence: that is, if P(S|H&B) > P(S|B) (where S is some supernatural claim, H is some hypothesis, and B is our background knowledge) then H is not a scientific hypothesis. Of course, it would be much more reasonable for the critic to say that even if P(S|H&B) > P(S|B), to infer S from H is not to make a scientific inference, because we have left the subject matter of science. If this was the claim, then the ID proponents would be exactly right to call the claim that the evidence supports some kind of designer a scientific claim, but discussion of the identity of the designer unscientific - even though the inference is obvious it is, according to this idea, an unscientific inference. But the previous claim, that H is unscientific because it provides evidence for a supernatural claim, seems to me to be the worst kind of dogmatism - it is to specify in advance as methodology what science can and can't find. Besides, it's impossible to uphold: for instance, the fact that natural laws hold consistently and tend to be mathematically elegant tends to provide evidence for the God of traditional monotheism over and against Zeus and friends, a theological claim. Furthermore, these people tend to think that scientific evidence against God is ok. So what we should really say is that as long as our hypothesis itself is entirely natural, it can be a scientific hypothesis (though it is not automatically a good one), but conclusions about the supernatural may still be able to be drawn (by philosopher-theologians) from the hypothesis.Posted by Kenny at June 26, 2007 12:52 PM
Return to blog.kennypearce.net