Expressivism about the value of truth

As readers of this blog well know, there is a growing amount of interest in epistemic value. In my view, understanding epistemic value requires understanding its meta-normative status. A good place to start getting a hold on this status is a question which Jon raised on this blog over a year ago: Can we be expressivists about epistemic value?

Clearly, a chief – probably the chief – epistemic value is truth. So we might just as well ask: can we be expressivists about the value of truth? In order to answer that question, we need to get clear on at least two things: what it means to talk about “the value of truth” and what it would mean to be an expressivist about that value. My answers to both questions can be found in “Expressivism and the Value of Truth” a rough draft of which can be found over on Epistemic Value (Comments obviously welcome).

Here’s the really short version:

By “the value of truth” we can mean two nominally distinct things. First, we could mean: the value of (having) true beliefs. This is what we are on about when we talk about truth being the proper end of inquiry – we think it is a good state of affairs to believe what is true. Second, we could mean: the value of a belief having the property of truth. This is what we are on about when we talk about truth being a norm of belief: that a belief is good/right/correct when true.

Suppose we limit ourselves to sense (a). If so, then, at the very least, being an expressivist about the value of truth amounts to claiming that we are not describing the world when we say, e.g. truth is what inquiry should aim at, or that it is good to believe all and only what is true. Rather, we are expressing a positive attitude or sentiment.

Is it true, for the expressivist, that it is, e.g. good to believe what is true? An epistemic expressivist in the mold of old-fashioned moral expressivists might have said no: evaluations/expressions are precisely distinct from descriptions in that the former can’t be true or false while the latter can. But an epistemic expressivist of a more contemporary frame of mind would be apt to answer as Blackburn or Timmons would in the more familiar moral case. According to Blackburn/Timmons and other contemporary expressivists, from what Timmons calls the morally engaged perspective, ethical claims are true or false. That is, they are disquotationally true or false. But from the morally disengaged perspective they are not capable of being “objectively true” or capable of being TRUE, as Timmons might say.

Perhaps not surprisingly, this is just how Hartry Field’s own epistemic expressivism is articulated. From what we might call the epistemically engaged perspective, Field says it is true that is good to believe what is true and not what is false. This is because, just as Timmons and Blackburn say in the parallel ethical arena, from a normatively engaged stance, the truth predicate operates disquotationally. Yet from the disengaged perspective, the expressivist will say that it is not true (or as Field says, not “objectively true”) that any epistemic goal is better than any other.

This is an interesting view: it is consistent with naturalism about epistemic value; nonreductive (it doesn’t reduce the value to something else); and yet it allows us to speak with the vulgar (we get to say that it is true that truth is the proper end of inquiry). Yet I think it is implausible, for I am not convinced we can make the necessary distinction, in the epistemic case, between the engaged and disengaged perspective. For in order to make sense of such a distinction, we need to be able to make sense of someone having something other than truth as a goal of epistemic inquiry. But, or so I argue, that requires making sense of someone believing – that is, having real beliefs — without their doxastic practice being regulated by the norm of truth. I don’t think we can make sense of that. So I don’t think we can make sense of the distinction between the epistemically engaged and disengaged standpoints; and without that distinction, the epistemic expressivist can’t say what is (clearly) true: namely, that it is true that it good to believe what is true and not what is false.


Expressivism about the value of truth — 10 Comments

  1. Mike, a quick question here about the strategy, one that will show that I haven’t read your paper yet! I take your strategy of undermining expressivism to involve trying to show that it is metaphysically impossible to form and hold beliefs without, in some sense, aiming at having true beliefs. If such an argument works, that would be interesting, but it strikes me as taking a difficult path. I would have thought that the problem would lie elsewhere. What would be needed to characterize an epistemically disengaged perspective would be to characterize a perspective where truth doesn’t matter. Not that it doesn’t matter to those taking that perspective, but that it doesn’t matter. Period. And then refusing to allow the response that X matters only if X matters to someone.

    As I read your quick-and-dirty summary, I think you’d allow this last response, and try to show that there is no perspective an individual can take in which truth doesn’t matter to them. Is that right?

  2. Hi Jon and thanks: In general, I would be glad to allow the last response, and moreover, I am inclined to agree — and indeed have argued elsewhere– that there is no perspective an individual can take in which truth doesn’t matter to them (whether they are aware of that fact or not).

    But you are also right that here I am trying to take a harder path so to speak. One reason for this is that the expressivist will charge that the above strategy is unfair. For their point is that from the disengaged perspective, for any epistemic goal, it neither matters nor fails to matter. Mattering-talk, so to speak, is non-factual from the disengaged perspective. And so they may well claim that insisting it does fails to appreciate their view. Hence I am hoping to construct an argument that appreciates what is distinctive about their view, but nonetheless claims that — in the epistemic case, at least — the very distinction that allows for that distinctiveness can’t be made coherent.

    The move about belief I try to make in the paper is roughly this(this is not how I put it there however): in order to make sense of the disengaged perspective, we’ve got to be able to make sense of the possibility that someone’s engaging in inquiry without their practices being governed by the goal of believing what is true and not believing what is false. But (I claim) this can’t be done, for to engage in inquiry involves figuring out what to believe, and one whose practices are not governed by the truth-goal isn’t figuring out what to believe. And THAT is because (insert longer story) it is a constititutive fact about belief that its basic norm of correctness is truth.

    I’m afraid that this is hardly transparent; but perhaps it gives the general idea.

    ( looking back, I see this is a very longwinded way of saying “yes, but…” to your question!

  3. OK, I haven’t read the paper either, but why wouldn’t the following be a good response:

    Suppose deflationism. Then truth is not valuable, since there is no such thing as truth. When we affirm that “it is good to believe what is true” we mean simply, “it is good to believe [some sentences]”. We might or might not be expressivists about that.

    I wonder, because *IF* there is anything to that line of thought (i.e. if deflationism entails that truth is not valuable), it follows that: if truth *is* valuable, then deflationary theories are false. Which would be interesting.

  4. Heath, I don’t think many contemporary deflationists would agree that their view entails that there is no such thing as truth — even though that might be how a robust realist about the truth property might describe the matter. They would say that there is such a thing as truth — but it is a merely logical property in the way “being a conjunction” is a logical property. In any event, many deflationists, including Field and Horwich, do wish to affirm the thought that truth is valuable — presumably in both senses that I distinguish above.

    You are right, of course, that it is a good question as to whether they are entitled to affirm the value of truth. I don’t think they are so entitled, in fact. (For details see my “Minimalism and the Value of Truth” in The Philosophical Quarterly, Oct. 2004 and True to Life (MIT, 2005). And I think, like you, that this is bad news for deflationism.

    In this paper, however, I am remaining neutral about deflationism. I could be wrong, of course, but I don’t think any of this paper’s arguments hinge on the falsity of deflationism — in any of its many varieties!

  5. Like Michael, I also believe that the truth norm is constitutive of belief. Specifically, I think that it is a conceptual truth that a belief that p is correct iff p. But I don’t think that this can be used to undermine all expressivist views about the normativity of truth, since one can be an expressivist about the norm embedded in the concept of belief. This would mean, of course, that one would have to be an expressivist about belief ascriptions. But so be it! I sketch an expressivist view about belief-ascriptions in ‘Doxastic Deliberation,’ (co-authored with David Velleman), available here:

    If I understand l correctly what Michael wrote above, he is claiming that there is no perspective from which I can deny the norm of truth, since if I were to do so, this would itself be an expression of what I take to be a belief, and thus would commit me to accepting the value of truth. Furthermore, to espouse any expressivist view about anything involves asserting certain claims, which involves expressing a set of beliefs (or what the espouser takes to be beliefs), which in turn involves accepting the norm of truth for those beliefs.

    If we accept these claims this does not mean that expressivism about the normativity of truth is false, but it may mean that there is a significant cost for expressivism. If making sense of people’s claims that it is true that truth is valuable requires that the expressivist be able to state his view about the normativity of truth from a perspective that does not involve accepting the norm of truth , then the expressivist may not be able to make sense of such claims. But all theories on offer have costs, so it is unclear that this cost should make us abandon expressivism.

  6. Hi Nishi, thanks for the links and comments. More on that in a moment.

    First, let me try to clear up a misunderstanding, no doubt due to my lack of clarity in the original post. In the present paper, I am concerned not with expressivism about the norm of truth, properly so-called, but with the value of the truth-goal. By the “norm” I mean I think what you mean in your papers on these subjects: that any given belief is correct iff it is true. By the truth-goal, I mean the idea that it is prima facie/pro tanto good to have and only true beliefs.

    My question in this paper is whether we can be expressivists (or more simply, non-factualists) about the goodness of the truth-goal (of having true beliefs). In another paper, (a copy of which is floating around on Epistemic Value), I also consider expressivism about the truth-norm, and essentially make (among other arguments) the argument you make above. (Although I haven’t worked that out to my satisfaction. Reading your latest paper will no doubt help — I’ve found your previous work extremely valuable).

    But let’s return for a moment to the truth-norm. As I see it, to be an expressivist about that norm is to be an expressivist about the “correctness” of true beliefs. There is one worry I have about any such position that is much more basic than the argument you mention. It is this:

    Suppose we think that

    (TN): The belief that p is correct iff it is true.

    And yet hold that

    (C): The belief that p is correct

    fails to state a FACT, or have TRUTH-conditions, where by that, I mean that it fails to have the type of positive semantic status propositions about the physical world presumably have for the semantic expressivist. Given (TN), then it follows that

    (D): The belief that p is true.

    Must also fail to state a FACT or have a TRUTH-condition, because even on the weakest reading of the biconditional (TN), (TN)’s truth requires that its halves have the same truth-value. But of course

    (T): The belief that p is true iff p.

    You see where this is going: by parity of reasoning “p” can’t state a FACT or have a TRUTH-condition. And that seems bad.

    So it seems that the expressivist about (TN) better not put their view in any way that implies that (C) is non-FACTUAL. So how can they put it? One suggestion is to say the following (and I think perhaps Nishi might go this way, I dunno): Statements about whether a belief is correct or not are not strictly true or false but only true or false relative to epistemic goals (such as the goal of believing all and only what is true). But which epistemic goals are true or not is not a matter of TRUTH or FALSITY. In other words, the strategy is: relativize norms to goals, and then go non-factualist about the goals. This would be more Gibbardian. It is the way Field goes. And it is this latter strategy that I aim at in the present paper.


  7. I’m not quite seeing the problem. An expressivist about good might also be a hedonist, thinking that

    (H) A state of affairs is good iff it is pleasant.

    The claim that a state of affairs is good does not state a fact, according to this expressivist, but the claim that a state of affairs is pleasant does state a fact. I take it that in order to say these things the expressivist must deny that strictly speaking the acceptance of (H) involves belief in a truth-apt content.

    How are things any different with the truth norm? Well, I did say that (TN) is a conceptual truth. But what I meant is that anyone competent with the concept of belief accepts (TN), and this does not commit me to the claim that (TN) is true.

  8. You are certainly right that an expressivist about good might also be a hedonist, and I think that the initial best reply to the argument I raised is precisely to make just that point.

    But what that point reveals, I think, is that much of the issue hangs on how we interpret (TN). For like any biconditional, (TN) can be read differently. One way to read it is to say that it expresses the thought (or is underwritten by the thought) that

    (SN): A belief is correct in virtue of its being true.

    Read in this way, there is no problem thinking that one can be an expressivist and endorse (TN). But it does raise a further explanatory burden, namely explaining the “in virtue of” relation here that holds between truth and correctness. What is it about truth that explains why a belief that has that property is correct?

    But there is another natural — and more minimal — reading of (TN):

    (MN): A belief’s correctness is identical with its truth.

    The plausibility of (MN) trades on the thought that there is nothing more or less to saying a belief is correct than saying it is true and vice versa. Note too that with (MN) we have no further worry about how correctness supervenes on truth. Correctness for beliefs simply is truth. But of course interpreted this way, the original issue remains.

    As a further note, I suspect that deflationists must opt for reading (MN). And if the original problem remains when (TN) is read as (MN), then deflationists can’t be expressivists about the truth norm.

    In any event, you also raise the further point as to whether in accepting (TN), however read, as a conceptual truth I must also believe it. I am inclined to say yes, but that is because I think any account of acceptability presupposes belief. But here we part ways no doubt.

    By the by, the argument and distinctions above appear in a different paper, “The Value of Truth and the Truth of Values”, which appears on Epistemic Value (under paper posts). Comments welcome on that too!

  9. Michael,

    You say that much of the issue hangs on how we interpret the biconditional (TN), and you hint that the expressivist will not be able to give an adequate intrepretation of it. But surely this is a general issue about how to read biconditonals with normative predicates on one side and non-normative predicates on the other. Thus, if the expressivist can give an adequate interpretation of (H), he should be able to give an adequate interpretation of (TN). I thus don’t see that there is anything special about (TN) that presents an obstacle for expressivism.

    As to ‘acceptance’, I am using this as a technical term to refer to whatever state of mind is expressed by a normative judgment. Thus one cannot argue that because there is an ordinary notion of acceptance that entails belief that ‘acceptance’ of (TN) entails believing (TN).

Leave a Reply

Your email address will not be published. Required fields are marked *