Friday, June 09, 2006

Gimme some truth

Many people nowadays, including or even especially philosophers, seem to think it important that we acknowledge the value of truth. I was reminded of this by the recent release of Ophelia Benson's and Jeremy Stangroom's Why Truth Matters, but this is a familiar theme not only in the works of the usual pomo-phobic culture warriors, but also those of philosophers such as Michael Lynch and Simon Blackburn. In his 2004 book, itself called True to Life: Why Truth Matters, Lynch compares philosophic indifference or hostility to truth with complacent acceptance of the Bush administration's alleged blithe rationalization of insufficiently justified claims of Iraqi WMD as instrumentally valuable (e.g. for fostering consensus). As George Will would say (and it is not often that I quote the man, so listen up):
Should we indeed make a point of valuing truth? Not surprisingly, that depends on what we mean. Sometimes "truth" is used (usually – but not only, alas – by non-philosophers) as a name for a kind of commodity or stuff, like knowledge is, rather than for a property that a proposition has, or not, in virtue of the relation between its content (meaning) and the way the world is. Although this usage can be innocent (as in the pragmatist motto "seek truth, avoid error"), it can also blur the very real conceptual difference between truth and knowledge, thus making the following points more controversial than they should be. (It can also lead to confusing talk about "kinds of truth," which would be better thought of as domains of inquiry or discourse.)

First, perhaps it is too obvious to mention, but it can't be that what is to be valued is that there is such a thing as truth, i.e., that there are truths. All that would mean is that we are speaking a language; but while I suppose I'm glad we're discursive creatures, that can't be what truth-valuers value. What use is it to me, for example, that "Eto ne moy plotok" correctly denotes a certain state of affairs if I don't know which state of affairs that is? In order for truths to be valuable, or for truth to be a goal, I need to have access to them somehow, and it is that that must be the locus of value, not the truths themselves; so maybe it is the value of knowledge that we are meant to affirm.

However, this doesn't seem right either. The value of knowledge, while more easily debated, is a fairly straightforward matter, philosophically speaking. Naturally we can think of plenty of things we are better off knowing than not knowing (or being deceived about). But clearly not every particular thing is worth knowing; indeed, the vast majority of truths are entirely worthless as knowledge. It is presumably either true or false (pace verificationists) that the word "marble" was used at least once on October 12, 1977 within a six-meter radius centered at a particular spot in the Fulton Fish Market. But who cares? (Of course we could make up a story in which we might need to know this. But so what.) Some knowledge is even downright dangerous. On the other hand, we can hardly make sense of a choice between knowledge and ignorance in general. Believing only falsehoods – having no knowledge at all – isn't even conceptually possible. Even Thomas Anderson (Neo pre-red pill in The Matrix) has plenty of true beliefs (but let's not go there today).

If the value of knowledge in this (uninteresting) sense is the value of belief (that P) given the truth of P, maybe we can do better switching it around: what is the value of the truth of P given belief that P? This keeps the focus on the value of truth, while preserving the connection to belief without which, as we saw up top, the question makes no sense.

So, do we really want our beliefs to be true? This too is ambiguous. In one sense, it's not an issue about truth at all, but about our desires. Given that you believe the world to be a certain way, do you really prefer (what you see as) the actual state of affairs to, say, this other one? I believe that the hurricane was devastating, but I would be happy to be wrong about that; I would prefer that my belief be false (but it isn't, alas). In another sense, this reverts to the question about knowledge – the value of true belief. If my beliefs are true, then, just as when truths are believed, the epistemic connection to the world is made; but we've already seen that this may or may not be valuable.

A third sense must be what is intended. As the aletheiaphiles present the matter, the issue is one of the threat of instrumentalism, or skepticism, or nihilism, or all at once. Richard Rorty famously denies that justification-transcendent truth is a goal of inquiry, or even that there is any such thing in the first place; and this will never do. The question is this: given that you believe something (i.e., because it pays to do so, and has established its value already, say in terms of prediction and control), do you place an additional value on its being true as well as useful? Rorty claims that he cannot see what point there could possibly be to this. But if you do not go beyond valuing utility, we are encouraged to complain, you are (if I may borrow a phrase from elsewhere) an epistemic "free rider": belief without ontological commitment, that is, has all the advantages of doxastic theft over honest ontological toil. But does this accusation, and the accompanying apotheosis of truth, get at our real interest here?

Now I agree with the truth-lovers in a certain limited sense: we should indeed reject instrumentalism, skepticism, and nihilism. This may not seem to be a limited sense, but as we shall see, it amounts to somewhat less than they suggest. Again, our claim cannot be that there is no such thing as a useful fiction (or even a pious fraud); clearly there is. Rather, it must be that we should not elevate this fact into a general attitude toward truth, like that what matters in inquiry is the consequences of believing P, rather than that P is true. But this – rightly claimed as a "truism" by Lynch, but brandished as if it were a substantive result – is simply a condition of having beliefs at all. (In fact, Lynch's "truism" is that "truth is a worthy goal of inquiry," which is confused – it's like saying that "arriving at the solution is a worthy goal of doing a puzzle." If you're not trying to get the solution, you're not "doing the puzzle" at all.) It makes no sense to deny the "value of truth" in inquiry; but by the same token it makes no sense to affirm it either (i.e., as something that transcends what is normally available to us as the result of inquiry). One inquires, and forms beliefs, or one does not; and one cannot not believe.

The anti-skeptical point, as shown for example by Myles Burnyeat some time ago re: ancient skepticism, is that it is incoherent to take a third-person point of view on your own beliefs (although Burnyeat himself did not press the matter any farther). If you believe something, inquiry is over: the issue is settled. This doesn't mean you're stuck with your beliefs, as you may always reopen the matter. But to do so is to remove the proposition in question from your beliefs and put it into doubt; and doubt is only intelligible against a background of settled belief. So (at least on the pragmatist cum Davidsonian view I recommend) we never say: this belief of mine might be false; for to say of something that it might be false is to regard the issue of its truth as no longer settled – and thus no longer a belief at all.

Making a similar (but not identical) point (and ironically providing Rorty, who is otherwise not a big fan, with pragmatist cover), Peirce claimed that the goal of inquiry is not truth but belief: you may think you want truth, but once you believe, you find that that is sufficient (I can look up the quote if you want). The thought is right, but the conclusion is not – because once you believe, you thereby ipso facto believe that you know the truth; so of course fixation of belief is sufficient. What would be superfluous would be a purported philosophical demonstration that our beliefs are true, either individually or collectively. If you really do believe, you feel no need for such a "demonstration"; and if you don't believe, then such a demonstration wouldn't apply (as it applies only to beliefs).

As may be evident by now, an important confusion (or systematic conflation) that can cause trouble here is between "belief" as a) a proposition qua believed (never mind by whom), and b) a proposition actually believed by the agent in question. These must be kept distinct. Again: take a belief of mine, any belief. If I cease to believe it, now I see it as possibly false. So the belief is possibly false? No, the proposition is possibly false as far as I'm concerned: I'm in doubt. Once I give it up, it's not a "belief" at all. Maybe someone else believes it; then it's a belief of theirs.

Once the anti-skeptical point is in place, the anti-instrumental and anti-nihilist points follow quickly (or are superfluous; on the other hand, this is not to say that related metaphysical and methodological unconfusing – of realist and anti-realist alike – isn't necessary to keep the anti-skeptical point in place, as skepticism naturally reactivates in the presence of dogmatism). So, to sum up: the value of truth has to mean the value of true belief, lest it mean the value of there being truths at all; but it turns out to be impossible to isolate the truth component of true belief as the object of value – the value of true belief ends up being either 1) the value of the world's being a certain way, or 2) of our knowing that the world is the way it is (either for the informative value of true belief or for the instrumental value of lacking false belief), or 3) of our believing what we believe given that we believe it (or something). The first two aren't what we were after, and the third doesn't make any sense. Our concern is really with belief and inquiry properly construed; but that noble goal is remarkably poorly advanced with such slogans as "truth matters."

I was going to run through Lynch's WMD example, but there are so many conflations and confusions in there that we better save it. Next time.


Charles Johnson (Rad Geek) said...

Duck: "So (at least on the pragmatist cum Davidsonian view I recommend) we never say: this belief of mine might be false; for to say of something that it might be false is to regard the issue of its truth as no longer settled – and thus no longer a belief at all."

Come on. One often says that a belief of hers might be wrong. For example:

L.W.: Why are the detective magazines so bloody expensive this month?

N.M.: I believe that there's a paper shortage, but I might be mistaken.

What Moore reminded us of is that one never says that a (current) belief of hers is actually false. But one often admits the possibility that it might be. That's not abandoning the belief; it's just expressing some epistemic humility about the beliefs that you have.

In fact, it's precisely in such moments of epistemic humility that we're most likely to use the phrase "I believe that --," instead of simply asserting the content of the belief without qualification.

I have no firm idea as to whether this grammatical point bears on anything of substance in your post, though. I believe it might, but you never know...

Clayton Littlejohn said...

I gotta side with duck on this one...

Often we say 'I believe that p' as a way of indicating that we are not fully committed to the truth of p. When we omit the 'I believe' and directly express the belief, it sounds like a contradiction to say:
(1) Custer died at Little Big Horn but he might not have.

Charles Johnson (Rad Geek) said...


Careful with the modalities there. "Custer died at Little Big Horn but he might not have" is actually a sentence of a sort that we often say. For example, he might not have if he'd been a better tactician, or less of an asshole. It's just that here "might not have" raises a counterfactual possibility for Custer, rather than a salient error-possibility for the belief about what actually happened, which is the sort of possibility that Duck's point was about.

That quibble aside, I'm a bit confused by your reply. Are you siding with the view that you can't raise the possibility that a belief of yours might be false, without thereby treating it as something other than a belief? Or with the view that you can? Because Duck's view is apparently the former, but what you say about "I believe that Custer died at Little Big Horn" as vs. "Custer died at Little Big Horn" would seem to count in favor of the latter view, not the former.

It's true that asserting that Custer died at Little Big Horn expresses a belief to the effect that Custer died at Little Big Horn. And it's true that directly asserting that rules out raising an error-possibility for that belief. But it's precisely when you not only express a belief but also say that you believe so-and-so that you can raise error-possibilities for the belief in question. And I certainly find it hard to believe that you treat P less as one of your beliefs when you say "I believe that P" than when you say simply "P." You're treating the belief as something you're more ready to give up or revise; and you're treating its truth as less than certain; but you're not thereby actually giving it up, in whole or in part, or treating it as something other than a belief of yours.

Duck said...

RG and Clayton: thanks for your comments! A real discussion at our humble blog - I can't believe it.

RG, as you might imagine, I have heard this objection before. I hadn't intended to get into my decidedly unorthodox (but not original) epistemological views, but then the issue had to go and come up anyway. I had thought that my disambiguation of two senses of "belief" might forestall the objection, but you are right that there is another ambiguity here (related but not the same).

If you want to distance yourself from some claim – as Clayton says, to indicate that we are not "fully committed" to it – you can indeed do so by prefixing it with "I believe that ...". I did not intend to (seem to) deny this, nor did I intend to (seem to) appeal to ordinary usage. By my lights, ordinary usage is ambiguous in this respect. If I had my druthers, people would resurrect the perfectly good word "opinion" for use in this context, to mean something which falls short of full belief. I would never suggest that we give up the very idea of partial commitment, as it's very useful in just the cases you mention. (We also say: I think so, but I might be mistaken. That would be fine too.)

My use of "belief" is specific (in the way philosophers feel entitled to be for their own purposes). This makes "belief" the opposite of "doubt". You will grant that it is nonsense to say: "the issue of the truth of P is settled, but still I have my doubts" if this means: "it's settled, but it isn't settled." In my usage, this is the same as "I believe it, but it might be false."

As I mentioned in the post, that doesn't mean that I can't reopen the case. It is in the fact that even full belief is (as my sect puts it) "corrigible" that one's epistemic humility is manifested. It is true that most epistemologists are (not corrigibilists but) "fallibilists", sacrificing full belief on the altar of epistemic humility. That's why fallibilism is essentially skeptical (but don't tell anyone).

Still, people do speak in the way you suggest. I might even be prepared to concede that the specific locution "I believe that ..." is a special case of some kind, such that it might make sense to regard the speaker not as retreating into doubt but preparing to, or something. (It still sounds odd to me. Do you believe it or not? Make up your mind!) The main point is that seeing the issue as settled is incompatible with regarding P as possibly false anyway, and the "I believe that P (but I don't)" formula brings this out the most directly. (There are also different senses of "possibility" here, which I'd rather not get into.)

Clayton, I share RG's confusion here. Are you really siding with me? Your first sentence sounds like what he says more than what I said (at least at first, but I have now conceded the specific claim concerning usage). I think I agree with RG's first paragraph re: Custer (except it sounds a bit like something a Brit would say more than a Yank).

Clayton Littlejohn said...

Hey DR and RG,

I thought I was agreeing with you, DR. In particular, I was using the incoherence of 'p but it might be that ~p' speaks in favor of the point Burnyeat seems to be making.

Suppose I ask, 'What happened to Custer?' and you say 'He's dead, I see his body right here'. 'Oh, I say, so it didn't go well for him'. If you say 'But he might not be dead', it seems you've just taken back what you told me over the phone.

Maybe in the last post, I wasn't clear to distinguish this remark from the similar and coherent remark that 'He might have lived if only he had not been so reckless' but this is a 'might have been' statement and not a 'might be/maybe its true' statement. I take it that because we take the 'p but it might be that ~p' as a kind of Moorean absurdity, we should regard recognizing that it might (actually) be that ~p as just a way of not believing p or not having settled deliberation on the matter.

Anonymous said...

In fairness to Peirce, a belief is a habit of mind that directs action, a habit which we may or may not be aware of. As (for Peirce) we have no priviledged access to our beliefs, we make judgments about which beliefs we have based on our actions, but of course we might be wrong about them. We may judge we believe something while, in fact, believe exactly the opposite. We also might judge something to be p, but in fact (unbeknownst to us) believe it to be not-p. Since they are habits, it's clearer why the first criteria for their evaluation wouldn't be their truth and, obviously, since Peirce uses "belief" in this technical sense, the normal intuitions about the usage of the word don't apply.