Sunday, November 17, 2013

Coyne: No Faith In Science

I disagree with Jerry Coyne about this.

I'm not sure that faith in science is on par with the kind of faith that most people seem to have in mind when they assert that science and religion are on par--they seem to be talking about "blind" faith. (Of course, many religious folks treat faith as if it were magical--as if it conferred all the advantages of knowledge onto their preferred beliefs, without requiring any of that pesky proof...)

For some time now I've been pretty interested in Richard Smyth's interpretation of Peirce with respect to the justification of the logic of science. In short, that does involve a kind of appeal to Kantian rational faith. The argument begins with the recognition that we (currently) have no non-circular justification of the types of inference requried by science (and, since circular justifications don't work, we can just leave out the 'non-circular' in that sentence; it's superfluous...) Most notoriously, we don't  know how to justify induction. Coyne makes an appeal to Dawkins--always a mistake in these matters--who writes:
There is a very, very important difference between feeling strongly, even passionately, about something because we have thought about and examined the evidence for it on the one hand, and feeling strongly about something because it has been internally revealed to us, or internally revealed to somebody else in history and subsequently hallowed by tradition. There's all the difference in the world between a belief that one is prepared to defend by quoting evidence and logic and a belief that is supported by nothing more than tradition, authority, or revelation.
But this won't work. The problem at the end of this trail is that we don't know how to justify the principles we use in our inferences. Sure, we can look at the evidence, but if we don't know how to justify induction--for example, we don't know how to show that it's more rational to reason inductively than to reason counter-inductively (though you don't have to make the argument that way)--then we seem to have no rational grounds for doing what we do with that evidence instead of something else.

Coyne writes:
What about faith in reason? Wrong again. Reason—the habit of being critical, logical, and of learning from experience—is not an a priori assumption but a tool that’s been shown to work. It’s what produced antibiotics, computers, and our ability to sequence DNA. We don’t have faith in reason; we use reason because, unlike revelation, it produces results and understanding. Even discussing why we should use reason employs reason!
 This, however, doesn't seem to work either, for well-known reasons. Think again about induction. The claim that we are justified in reasoning inductively because it works is patently circular, even on the best reconstruction of it--all we know is that induction has worked well so far. To conclude that we should reason inductively tomorrow is (apparently) to reason inductively: induction has worked in the past, so it will work in the future. Circularity, again. (Note: I don't think the past-to-future characterization of induction is right...I'm just employing it here for brevity.)

Coyne also writes:
Finally, isn’t science at least based on the faith that it’s good to know the truth? Hardly. The notion that knowledge is better than ignorance is not a quasi-religious faith, but a preference: We prefer to know what’s right because what’s wrong usually doesn’t work. We don’t describe plumbing or auto mechanics as resting on the faith that it’s better to have your pipes and cars in working order, yet people in these professions also depend on finding truth.
But a mere preference does not constitute a justification. So, though it's right to say that this appeal can't prove the conclusion that science is based on faith, that's a hollow victory. This response says that science is based on something exactly as irrational as faith. But, of course, Coyne tries to slip in proof to bolster preference at the end of the argument: we just prefer it...but we prefer it because it works...circularity and other nastiness threaten again...

Honestly, I don't see why glib, unsound defenses of science are much better than glib, disingenuous attacks on science.


Anonymous Anonymous said...

It seem Coyne wavers a bit as to how normative he takes science to be. At first it seems like he might be relying on a traditional fact/value distinction, but then he ends up sneaking in preferences towards the end. So, while he's admitted that science is normative through and through, I suppose the question is how different we want to make "faith" and "preference." Obviously how he defines them makes them fairly distinct (insofar as his version of faith is fairly cartoonish); but I'm not so sure. Most, if not all the things I have faith in, e.g. that society might be more just in the future than it is today, are built on my preferences. But the reason why I said Coyne “sneaks in” preferences is that it seems like he doesn't appreciate the extent to which faith can and is involved in them. He glibly says he has his preferences, e.g. to prefer knowledge to ignorance, because they work. But that seems to beg the question somewhat, because the further question would be, "work for what?" They work to further a certain trajectory in scientific inquiry, but it doesn't seem like you can have a faith-free preference in terms of where something that oriented to the long-run is going. In other words, it seems like there will have to be a certain faith (or simply a much more loaded "preference" than he allows), as regards the notion that where inquiry is taking us in the long-run is better than where we'd end up otherwise. Now, I'm certainly not one to block the road of inquiry, but it just seems like it's not so obvious that that road just is lined with roses.

10:20 AM  
Anonymous Anonymous said...

Now, Winston, as to your musings about justifying our inferences, I wonder whether Peirce would worry about something like "justification" to the extent that you are. It seems like your worries still stem from the deductivist camp that he's trying to distance himself from. Now I could be wrong about that, i.e. that that's not what you're trying to do, but it seems like there is still this desire, one that wishes induction could be justified deductively, in order to get something like "certainty" up and running. For from within the deductivist framework, none of what Perice says makes much sense. For one, abduction would be a fallacy, insofar as it's essentially just affirming the consequent. But nonetheless, as Perice would say, and I agree, that's where inquiry starts; you can call it what you want. (Similarly, the idea that induction is a form of inference that can be gauged in terms of validity will sound equally foreign, etc.)

Thus, while I'm no crude instrumentalist about pragmatism, I don't think we have to shy away from saying that abduction and induction are justified because they work. Wouldn't he say something like, "Look how far we've come in terms of being able to string together certain explanatory theories. In an infinite universe, where there are hypothetically infinite hypotheses to choose from, the idea that a solution to a problem is available is crazy enough, but the idea that that solution could then help us figure out a further problem, is, statistically speaking, borderline impossible. So, what's going on? Well, it seems we have a predisposition towards truth, that are minds are attuned with the universe, and that progress in inquiry has occurred through the application of hypothetical inferences, and that, were there not something to them, we wouldn't have ever solved a single problem, etc."

I suppose I'm just trying to say that I think your worry is a bit overblown, insofar as it still somehow seems to react to Hume, who seemed all too glib about showing that you can't justify induction deductively, to which Peirce would say, of course you can't, that was a misguided idea in the first place; no inquiry in human history has started with deduction.

The most interesting philosophical question I know of is: "where did you get your premises?" And, so far as I can tell, the answers is always: from what was originally a hypothetical inference. That’s what we’ve got, and it seems the burden of proof is on whoever thinks they have something better to offer.

10:21 AM  
Blogger Winston Smith said...

Whoa, thanks for the interesting responses, anonymi...

At a glance, I think I'm inclined to agree with first Anonymous, and disagree with second Anonymous.

Quickly, wrt the latter: well, I'm really interested in Smyth's non-standard interpretation of Peirce, so that may be a source of our disagreement. "Where did you get your premises"--or, more precisely: "how do you justify them?" is one interesting question. Another interesting question is "how do you justify the principles you use to move from your premises to your conclusions?" I don't see how you can ignore either question, and, IMO, neither can Peirce.

Indeed, abduction is a deductive fallacy, but that doesn't mean that you can't justify abduction deductively. That's Hume's mistake, right? There's a huge difference between (with respect to a non-deductive inference from P to C) (a) showing that C follows from C deductively and (b) providing a deduction that shows that P follows from C non-deductively. It's like a kind of scope error, no?

Anyway, gotta go run before it starts to cool off around here... I'll try to say something less rushed later...

12:37 PM  
Blogger Pete Mack said...

I don't think "faith" vs. "reason" is the issue here: it's more a matter of judgment. Can you reasonably agree with the following postulates? Then we will deduce what follows... This is the basis of mathematics (from a very few axioms/postulates, to spectacular, subtle results in number theory) and physics (from a very few basic observations like the periods of Jupiter's satellites to spectacular theories like Newtonian and Quantum Mechanics.)

There's a much better reference here:

The whole faith/reason kind issue for science doesn't seem any more fruitful than trying to decide whether philosophy (or math) is based on faith or reason.

7:24 PM  
Blogger Winston Smith said...

Well, these issues aren't really amenable to being addressed in a short forum like this one... All I'm really doing is pointing out that Coyne doesn't make a successful case.

But, briefly: your suggestion is that it's about judgment not faith. Again, I'm sympathetic with Perice who thinks that speaking of judgment is a confused way of speaking about inference...which means that appeals to judgment can't make questions about the justification of the principles of inference go away. You put it like this:"Can you reasonably agree with the following postulates?" That just turns the problem into that of determining how to distinguish reasonable from unreasonable initial propositions... So, again, the problem doesn't go away.

None of this means that science must be based on faith--much less blind faith--but Peirce has an interesting and complicated story to tell about something like rational faith and rational hope where proof and disproof are both unavailable.

I'm just gesturing at something here, not claiming to have done anything more.

7:20 AM  
Anonymous Anonymous said...

I guess perhaps I just don't understand what you mean by "justification," e.g. where you say "appeals to judgment can't make questions about the justification of the principles of inference go away." I agree with you on that specific point; but I ultimately think you're after more justification than we can get, making said point somewhat moot. Take the following claim from Peirce in "The First Rule of Logic" (on pg. 49 of vol. 2 of E.P): "Now nothing justifies a retroductive inference except its affording an explanation of the facts." This is, I take, a sort of restatement of Emerson's claim in the introduction to Nature (the book): "Whenever a true theory appears, it will be its own evidence. Its test is, that it will explain all phenomena." Now, from the vantage-point of certain deductivist logics, both of those claims seem to be question-begging, i.e. don't count as the sort of non-circular justification you claim you're after. That said, begging the question is only a fallacy from within that deductivist framework from which Peirce is trying to move away from, making that criticism seem somewhat moot as well.

5:54 PM  
Blogger Winston Smith said...


You write:

"I guess perhaps I just don't understand what you mean by "justification," e.g. where you say "appeals to judgment can't make questions about the justification of the principles of inference go away." I agree with you on that specific point; but I ultimately think you're after more justification than we can get..."

Let me start by saying that I think that a lot of extremely non-Peircean "pragmatists" of the Quine-Goodman stripe commonly make argument that depend on a suppressed premise to this effect:

The best justification we can get must be good enough.

I'm not sure what you think about that premise, but you seem to flirt with adopting in above.

I don't think that the best we can get must be good enough. That is, I think that skepticism/epistemic nihilism is a real threat, that must really be defeated or capitulated to. And so does Peirce, so far as I can tell.

Second, I totally agree that Emerson is important to Peirce, but I don't think he (Peirce) thinks that anything is ever "it's own evidence." The rejection of Cartesianism/foundationalism seems to me to be an absolutely fundamental plank in his program.

His story about the justification of non-deductive inference is long and complicated and incomplete (and, of course, not obviously successful). But I don't see him at any point accepting circular justifications. And I personally don't see that circularity is only a problem for deduction... Is this something well-known that I'm ignorant of?

7:58 AM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home