Thursday, June 03, 2004

Sandefur on Satisfactory Explanations

In a recent post, Sandefur demonstrates his open-mindedness on the subject of the supernatural:
Magical explanations aren’t really explanations at all; they simply move the Mystery to a higher shelf. * * *
Richard Dawkins: “If you’re allowed just to postulate something complicated enough to design a universe intelligently…[y]ou’ve simply allowed yourself to assume the existence of exactly the thing which we’re trying to explain…. You’re simply not providing any kind of explanation at all.”
This argument is wrong. An example: John Smith is dead of a head wound in a cabin in the woods. Someone posits that some unknown person must have killed him. The response: "But that isn't an explanation at all, because now we just have the mystery of who the unknown person is and why that person would kill Smith. Therefore, an unknown person couldn't have been involved."

Wrong. You can explain one thing on one level, and still be left with a mystery on another level. If John Smith really was killed by some unknown person, then that is the explanation for his death. We might now need an explanation for the unknown person, but that is an entirely separate question.

Let's generalize: Suppose that we see the phenomenon Y happening. Someone says that X must be the cause. It is absolutely, totally, and completely beside the point to say, "But we don't have an explanation for X itself." So what? That's no basis for ignoring X. If X is the cause of Y, that's just the way things are. It would be nice to have an explanation for X too, but we can know that X caused Y even if we don't know (yet) that W caused X.

The same is true here. Assume that God really did heal someone. If that happened, then it might be a mystery why God exists, and why God would heal her and not someone else. But that doesn't change the fact that -- by the very assumption of this hypothetical-- God's healing is the proper explanation for why the person was healed. The fact that a mystery remains doesn't undermine the fact that there IS an explanation for her healing.

The source of the confusion is that Sandefur and Dawkins seem to rely on a principle akin to this: "Nothing can count as an explanation unless it explains everything, without pointing to anything that is currently unexplained." But there is no reason to think that this principle is true. Indeed, they wouldn't accept this principle if the explanations involved were naturalistic. No one would say, "You can't rely on the existence of dark matter to explain the movements of stars, because that just pushes the mystery to another level." No: If dark matter causes the motion of stars, then that's the way things are, whether or not we currently understand all there is to know about dark matter.

Or perhaps the confusion arises because the unstated principle is this: "Nothing outside the scope of my knowledge or understanding can possibly exist. Thus, if something is offered as an explanation but points to something I can't fully understand, it must be false." Again, there's no reason to think that this principle is true. No one is guaranteed that they will personally understand every explanation of everything that ever happens.

Or another variant: "No explanation is true unless I find it satisfying. I find it very unsatisfying to be told that something -- anything -- might be due to some supernatural cause that I might not be able to understand." Again, not true. The world has not constructed itself around our desires, and there is no reason whatsoever to think that all truths must be satisfying.

Now, one can think of many possible supernatural explanations that are absurd and untrue. Those whose mind runs to the absurd will respond with something like this: "Ah, but are you really saying that I can 'explain' the rain by saying that it was caused by invisible fairies?" No. Neither I nor anyone else has ever implied that every phenomenon should be slapped with a supernatural explanation willy-nilly. As it happens, the fairies-causing-rain explanation doesn't work. But why? Not because fairies themselves haven't been explained satisfactorily. The reason this explanation doesn't work because there's no evidence that there are fairies or that they caused the rain in the first place. There and there alone is where a supernatural explanation might fall short.

And indeed, most or all explanations might happen to fall short there. But if there were evidence of fairies, no matter how slight, it would be silly to dismiss that evidence on the grounds that fairies can't cause anything unless their existence has been fully explained.

UPDATE: Sandefur has responded, with his main idea being that a true explanation shouldn't be adaptable enough to cover all conceivable possibilities. Here's a sampling:
But that answer could be applied to Y and not-Y equally. What we want to know is why Y happened instead of, say, Z, which did not happen, but which, had it happened, could be explained with equal plausibility by appealing to God’s will. Statement Q is worthless unless it is at all possible to say not-Q.

* * * What I’ve said -- and so clearly that I can’t really think Buck misses it by accident — is that nothing counts as an explanation if it explains everything. Buck’s “It’s Magic!” answer would explain every conceivable alternative state of facts. It is therefore worthless as an explanation.
So how about natural selection (and variants such as sexual selection, kin selection, and pleiotropy)? It could equally be charged with explaining "every conceivable alternative state of facts." Example: Why do humans have eyebrows? Because a) natural selection selected for organisms who kept sweat out of their eyes; or b) because sexual partners liked eyebrows for some reason; c) because there was some other preferable gene that happened to correlate with having eyebrows; or d) because having eyebrows somehow helped the rest of the group to survive.

But what if we didn't have eyebrows? The answer would be nearly the same: Because a) natural selection selected for organisms who are more aerodynamically streamlined; or b) because sexual partners liked smooth foreheads for some reason; c) because some other gene that eliminated eyebrows must have been selected; or d) because not having eyebrows somehow helped the rest of the group to survive. You could repeat this process of hypothetical explanations for every conceivable quality: We have large heads because big brains were an advantage. But if we had small heads, it would be because smaller heads were easier on the birth canal, or the like. No matter what happens, some variant of natural selection can be brought in to explain it.

Now does that mean that natural selection doesn't count as an explanation? Of course not. If the actual fact is that humans who had eyebrows were selected -- for whatever the reason -- then by definition that is what happened. That is the immediate cause, no matter whether natural selection could be brought in to explain the opposite as well. An explanation might be extraordinarily flexible -- but true nonetheless.

Same goes for the supernatural. I know it is awfully hard to use the imagination, but imagine that God really does exist and that he really did heal someone. If that actually happened, then that is the only true explanation. Period. It is absolutely irrelevant that someone else could proffer God's inaction as "explaining" someone else's continuing sickness. That's what I continue to say that the only legitimate objections here are that 1) God doesn't exist, or 2) God didn't act in some particular instance.

FURTHER UPDATE: Read Jeremy Pierce's analysis.


Blogger Pepper Jack said...

Like many people of course, I've also encountered arguments like Sandefur's. Thanks for the simple and well-thought-out rebuttal.

11:25 AM  
Blogger Kimber said...

Interesting and well thought out argument until the end there. You ran yourself into a hole with the fairy thing because there is no evidence of god's existence either. So under your theory, saying fairies cause rain is just as plausable as god healing someone.

Sorry, better luck next time.

2:46 PM  
Blogger Pepper Jack said...

Kimber, I believe you missed the point of the post. I don't think Stuart was trying to show that the existence of God is provable, only that the "shifting the mystery" type of argument isn't valid.

5:12 PM  
Blogger Jeremy Pierce said...

I decided I had more to say on this than I thought would be good for a comment, so I've posted my thoughts here.

10:18 AM  
Anonymous Anonymous said...


There is plenty of evidence for the existence of God. For example, there are claims (by people) that he became flesh, dwelt amonst us, and told the people around him of that. There's lots more, too (in different traditions).

Like the evidence for most things, it consists of testimony, and perhaps you're discounting testimony as evidence. That's a tenable position, but please do realize that it puts you completely apart from the great bulk of humanity, who believe in things like science and history and geography and train schedules.

3:26 PM  
Blogger David Nieporent said...

The flaw in your argument is this: when science asks about a phenomenon, it is not asking, "Who did this?" It is asking "How did this happen?" If we see John Smith's dead body mysteriously lying on the ground and your answer is "someone did it," that doesn't answer anything at all. If your answer is, "His head was bashed in," then we have made progress, and we can discuss who did it.

It's not that we should reject "X explains Y" because we don't also have an explanation for X, but rather that X isn't an explanation. It's a word. You might as well say, "Flgwak5gt3vf is responsible." It provides just as much informational content, and is just as verifiable.

To respond to your update, "natural selection" isn't an explanation; it's just a restatement of the question. "X provides an evolutionary advantage because of such-and-such" is an explanation. It's testable (at least in theory) and it explains how. Whereas "God did it," or "magic did it" or "Flgwak5gt3vf did it" ends the investigation without giving us anywhere to go; we know no more than we did before the answer was given.

6:07 PM  
Anonymous Anonymous said...

Natural selection would indeed not count as an explanation if it only provides just-so stories that could be made to be consistent with anything. That was Sandefur's point, that if a theory can explain anything it is not an explanation at all. An explanation should be verifiable through supporting evidence that its postulates would predict exist. God can be invoked to explain anything and is an unfalsifiable postulate.

1:31 AM  
Anonymous Anonymous said...

Interesting article. There is another guy who wrote about something similar on his site. I liked his analysis and I think he's a scientist.

11:23 AM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home