Al, Betty, Carla, Dan, and Ed accept the claim, “Utilitarianism is true.” They accept it because they have been told it by their revered professor, Dr. Black. In fact, their reverence for Dr. Black is so great that they would revise any other claim they accept before revising this one. However, there is a twist. In addition to accepting that utilitarianism is true, each student also has certain other beliefs as well:
- Al believes, correctly, that utilitarianism is the moral theory that tells you to do what will bring about the greatest happiness for the greatest number.
- Betty believes, correctly, that utilitarianism is a moral theory, but she does not know which theory it is.
- Carla has no idea what utilitarianism is.
- Dan believes, incorrectly, that ‘utilitarianism’ is the name of a mathematical thesis, though he has no opinions about which mathematical thesis it names.
- Ed believes, incorrectly, that utilitarianism is the claim that the square of the hypotenuse of a right triangle is equal in area to the sum of the squares of the other two sides (i.e., he has utilitarianism confused with the Pythagorean theorem).
We would ordinarily say that all the students have at least one belief (or more generally, one mental state) in common, namely the belief that utilitarianism is true. And we would say that when each asserts, “Utilitarianism is true” they all mean the same thing. But can an expressivist make sense of these ordinary intuitions?
According to the expressivist, the state of accepting “Utilitarianism is true” is either a cognitive state (an ordinary belief) or a non-cognitive state. (If you don’t like that terminology, substitute your own.) The difference is supposed to be reflected in the conceptual role of the state. Since “Utilitarianism is true” looks like a good example of a moral belief, let’s begin by supposing that the mental state the five students have in common is a non-cognitive one. But why think that Ed’s acceptance of this claim is in any way non-cognitive? It plays the same role that “The Pythagorean theorem is true” would for a normal mathematically educated person. Perhaps, then, the state is a cognitive one. But why think this about Al, for whom the claim functions exactly as it did for, say, Jeremy Bentham? Perhaps, finally, Al’s mental state of accepting “Utilitarianism is true” differs from Ed’s. But it is hard to see any grounds for this position, especially since the acceptance is equally deeply entrenched for both of them.
I’ve not even brought up Betty, Carla, and Dan. What should the expressivist say about Carla? If you think a certain named theory is true, but have no idea what kind of theory it is, is your state cognitive or non-cognitive? What grounds could there be for saying one or the other?
Here’s a diagnosis of the expressivist’s difficulty. The usual thing to say is that the meaning of “Utilitarianism is true,” and consequently the content of the mental state of accepting this statement, is “wide,” i.e. does not depend solely on the psychological states of those who accept it. It depends on history, social usage, etc. The expressivist explains the meaning of any sentence in terms of the mental states it expresses, so any wideness in the meaning of a statement has to be a wideness in the content of the state expressed. But since part of what’s wide about “Utilitarianism is true” is whether ‘Utilitarianism’ names a moral or a mathematical theory, it follows that whether or not a mental state is cognitive or non-cognitive is a wide property of the state. But that just seems like too much for the expressivist to swallow.
Heath,
Whatever else one should say about your example, an expressivist should not claim that the judgment “utilitarianism is true” is a moral judgment, whether “utilitarianism” is thought to be a moral theory or not. Saying any theory is true (or false) is ordinarily a descriptive or cognitive saying. On the other hand, the judgment “Utilitarianism is abhorrent” or “Utlitiarianism is a good thing for us to believe” would indeed be a normative judgment about utilitarianism.
Heath,
I’ll grant that all five students accept the sentence ‘utilitarianism is true’ (i.e. they believe that this is a true sentence), but since Carla, for instance, has no idea what utilitarianism is, she has no idea what proposition the sentence expresses; so I don’t think she can be said to believe that utilitarianism is true. Similarly, if I hear my reliable French friend utter “Paris est une femme,” I (who, let’s imagine, know no French) might as a result form the belief that ‘Paris est une femme’ expresses a true sentence, but I don’t think that I would thereby believe that Paris is a woman.
Why would we think that they have one belief in common? Say that the professor convinces them that ‘Kennedy was the greatest politician of all times’. Take 4 students of which 3 have different Kennedy in mind and the 4th knows of no politician named Kennedy but accepts the claim anyway. Do they have a belief in common? My intuition is that they do not have a shared belief.
Robert,
Should the expressivist say that
(i) “one ought to be nice to children” and
(ii) “it is true that one ought to be nice to children”
express different (kinds of) mental states–the first cognitive and the second non-cognitive? That seems very strange to me, though I suppose it might solve the problem I raise. But why is everyone in the state expressed by (i) also in the state expressed by (ii)?
Troy,
How do we tell when someone “believes” something? This could get complicated, but it seems to me that if I tell you something, and you accept what I say, then you believe it, even if you understand it incompletely. (Most people are like this with E=mc^2.) I don’t think the French example is a good one, since my example doesn’t suggest that Carla or anyone else believes a translation of what they are told, but just: exactly what they are told.
Robert,
If I say, ‘It’s true that utilitarianism is abhorrent,’ is that normative or descriptive?
Heath & Troy,
Adding ‘it’s true that’ to a normative claim, according to expressivists nowadays, adds, if anything, only more normative oomph of some sort. But adding “it’s true that” to a non-normative claim, for instance, a theory of right and wrong, would not add something normative.
I supppose you may hold that utilitarianism is just a first order normative claim, such as ‘Torture is wrong’. In that case, saying “it’s true that all and only right actions maximize happiness” is both a normative claim and adds something normative to an embedded normative claim. I don’t see the theoretical claim that way, but I suppose one might.
Heath, I think Troy is right about this one, though you might be able to close the loophole. You say (in reply to Troy),
Doesn’t that confuse propositions with sentences? Sentences, and not propositions, have translations; but propositions, and not sentences, are believed; so Carla cannot believe a translation of anything.
Jamie and Troy,
Troy sought to avoid the problem I raised for expressivism, or part of it, by distinguishing between “accepting a sentence” and “believing a proposition.” I grant that there is such a distinction, it just seems to me that Carla (and in fact, A-E in my example) believe a proposition, namely that utilitarianism is true. But this hinges on what it takes to believe a proposition. Troy compared Carla, who does not know what utilitarianism is, accepting “utilitarianism is true,” to himself, who [in the example] knows no French, accepting, “Paris est une femme.” The idea was that, when you don’t know what a certain term means, claims that it is true are for you like hearing a foreign language. I don’t think this is a good comparison, but I suppose I need to say why more clearly, by way of saying what I think it takes to believe a proposition (rather than simply accept a sentence, as one would if one heard a foreign language).
It seems to me that you can believe propositions even if you only partially understand them—there’s plenty of work by e.g. Putnam and Burge to make this case. But, how little can you understand, and still be a believer? Imagine Fred, who has reliably been told, and accepts, that
– utilitarianism is true
– utilitarianism is a version of consequentialism
– consequentialism is incompatible with Kantianism
– John Stuart Mill endorsed utilitarianism
Other than this he knows nothing substantive about any of utilitarianism, consequentialism, Kantianism, or Mill. On this basis, however, he can draw a number of inferences: that a version of consequentialism is true, that Kantianism is false, that Mill was a consequentialist. This degree of inferential ability seems like enough to count as a believer in a proposition, rather than simply an accepter of a sentence.
But there is no sharp line between Fred and Carla: Fred just has more sentences in which terms are relatively undefined than Carla does. Moreover, I think you would be hard-pressed to draw a sharp line between Fred and Al: Al just has a good deal more of the same kind of thing that Fred does. We could say that the distinction between believing a proposition and accepting a sentence is vague, but that seems very unpromising.
I think the only important distinction is between someone who can parse a sentence and get the logical form out of it, and someone who cannot. This is what allows one to make inferences with the contents of the sentence, it is what separates Al, Carla, and Fred from the fictional Troy, and it is the only interesting criterion separating “accepting a sentence” from “believing a proposition.” After you can parse a sentence, understanding is a matter of degree, and there is no reason to think you have to “fully” (whatever that would mean) understand a sentence in order to believe the proposition it expresses.
So I think A-E (and Fred) all believe the proposition “utilitarianism is true,” even if they understand it only partially or even misunderstand it. Anyone who disagrees with this needs to provide some criterion for “believing a proposition” that accounts for ordinary cases, is not vague, and somehow distinguishes Carla from everyone else. I don’t think there is one.
Heath,
I think you’re right that there’s no sharp line between Fred and Carla, nor any between Fred and Al. But I doubt that there is a sharp line between believing a proposition and failing to believe it.
That seems very implausible. Suppose PEA Soup adopts a special terminology: we use ‘δ’ as a name for Dave Sobel and ‘ν’ as a predicate for being no fun at all. Robert, though, hasn’t been paying attention, so when I write “δ is ν” in the comments section, Robert assumes I am speaking with authority about some arcane bit of set theory, and he repeats it assertively later. Does Robert then believe that Dave is no fun at all?
It looks to me like there are two problems at stake in Heath’s original formulation. First, there is a problem about how expressivists are to treat truth-ascriptions. How to develop an expressivist brand of minimalism gets interesting precisely when we consider sentences like “Utilitarianism is true”, yet it is in precisely those cases that it is unclear what sort of mental state they should express.
This problem has nothing in particular to do with Heath’s choice of “utilitarianism”. Consider the sentence, “What Jon said is true”. What sort of attitude does it express? Does it depend on what Jon actually said (which the speaker may be unaware of)? Or on what the speaker thinks that Jon said (which may not be what he said at all)? Or both? If you start with the idea that “it is true that murder is wrong” should express a noncognitive attitude and “it is true that grass is green” should express a cognitive one, then no matter what you say about “what Jon said is true”, you’re going to run into funny problems.
For example, if it depends on what Jon actually said, then it would seem that whether the state the speaker is in when she sincerely asserts, “What Jon said is true” is cognitive or noncognitive is externally individuated. But that’s bizarre, and hardly fits well with the idea that cognitive and noncognitive attitudes have different motivational properties. On the other hand, if it depends on what the speaker _thinks_ Jon said, then that can come apart from what Jon actually said, leading to funny problems about speakers’ commitments, and the thought that what Jon said is true turns out to be affected holistically in a way that other thoughts are not.
All of this seems like a real issue, but it has specifically to do with how truth-ascriptions are handled, not with expressivism in general.
Now distinguish a second issue. Consider the sentence, “murder is shprong”. If “shprong” is a normative predicate, then this should express a noncognitive attitude, but if “shprong” is a descriptive predicate, then this sentence should express a cognitive attitude. The question is: can speakers use a word like “shprong” successfully and have thoughts about what is shprong, without realizing whether it falls into the normative or the descriptive category?
If so, then someone can assert that murder is shprong without realizing whether it is supposed to express a cognitive or noncognitive attitude. Similarly, she may be able to _think_ that murder is shprong without realizing whether this involves having a cognitive or noncognitive attitude. The former seems weird, but how weird it is depends on what you think is involved in the expression relation. The latter seems super-weird. How does your psychology know whether to get itself into a belief-like state or a desire-like state as a matter of thinking that murder is shprong, if nothing about your linguistic competence with the term clues you in about which it is supposed to be?
This problem looks like it is independent of the issues about truth-ascriptions. Of course, Heath was presumably thinking that it is more uncontroversial that two people can both think that utilitarianism is true without knowing what utilitarianism is, than that two people can both think that murder is shprong without knowing whether “shprong” (or any of its translations) is a normative or a descriptive predicate. Still, it looks to me like the two separate issues should be evaluated separately, and like Troy’s worry, for example, is based on Heath’s formulating them as a single issue.
Having hung around Dave Sobel some, let me say: ~(δ is ν).
Doug, don’t you speak English?
“δ is not ν.”
Here’s a first pass. Start from Mark’s claim ‘what Jon said is true’. If I don’t remember this wrong, the minimalists think that ‘true’ here works as a logical device to enable samesaying more easily. So, the way to unpack what is meant by the term is:
‘(Jon said that p & p) or (Jon said that q & q) or (Jon said that r & r), and so on.
Saying that sentence would be a mouthful. Would actually take infinitely long. So, we use rather ‘what Jon said’ as a demonstrative to refer to Jon’s statement and ‘is true’ of the disquotation schema to samesay or rather sameassert that statement. Very handy.
Now return to Clara’s statement ‘utilitarianism is true’. She can intend to use the name ‘utilitarianism’ to refer to the same theory as her professor. This is how we use names often. We inherit the reference from others. She then uses the ‘is true’ part as a logical device to affirm these claims, whatever they are, of that view on the basis of her trust for her professor. The emotivist can say that by affirming the view, she expresses a commitment to whatever commitments, belief-like or desire-like, the professor’s view commits anyone to.
Very neat, Jussi.
Notice that this commits the expressivist to the view that adding “is true” to any proposition is a way of expressing one’s acceptance of it, and no doubt of inviting others to accept it too, etc. In which case uses of “is true” look like they ought to be classified as normative by expressivists. In which case any expressivist can get across the supposed is-ought gap by this simple argument:
(Premiss) p;
(Conclusion) p is true.
The premiss is (let us grant) non-normative; the conclusion (as we’ve just seen) is, by your expressivist’s lights, normative. Voila, a valid is-ought inference.
Of course, I mention this consequence of a certain expressivist view about truth in a spirit of anything but helpfulness to expressivism… it’s something I’ve just argued out in a book I’m writing, as an objection (yet another objection) to expressivism, so I’d be pleased to hear any reactions.
Hi, Tim –
I think it follows from Jussi’s suggestion that different ‘true’ sentences will express different attitudes. ‘it is true that murder is wrong’ will express a noncognitive attitude, but ‘it is true that grass is green’ will express a cognitive one, and ‘what Jon said is true’ will apparently express whatever sort of attitude is expressed by infinite disjunctions with both descriptive and normative disjuncts. We presumably have to solve the traditional Frege-Geach problem for disjunctive sentences before we know just what that is. But in no case would your argument seem to be an is-ought inference, on this view, so far as I can tell.
As for the Jussi’s suggestion itself, this sort of thing always puzzles me. On the face of it, ‘what Jon said is true’ and ‘what Jon touched is gold’ have the same structure. So unless English works in some sort of magical way, the hypothesis that the apparent quantificational structure of the former is illusory, and really just an abbreviation for an infinite disjunction, would seem to require, for consistency, the hypothesis that the quantificational structure of the latter is illusory, too.
And similarly for every other sort of quantifier – ‘everything Jon said is true’ will be some crazy disjunction of conjunctions, and so, apparently, will ‘everything Jon touched is gold’. But if this way of developing minimalism about truth requires the hypothesis that there aren’t really any quantifiers in English after all, then it looks like the kind of thing we can have linguistic evidence against, just by considering sentences without the word ‘true’ in them at all.
Hi Mark:
I’m not sure I know what a cognitive attitude is anyway– the uses of attitude that go with “propositional attitude” and with “attitudinism” seem so different that it looks like an equivocation to treat them together.
That aside, why should different true sentences express different attitudes? My point was that one way to understand the truth-predicate– a natural one for certain sorts of expressivist– is to say that “it is true that p” expresses a normative attitude to p, for any content of p. Hence this expressivist understanding of the truth-predicate makes it look as if “it is true that p” is normative even where “p” is non-normative. But that’s obviously a crazy conclusion. You could draw various morals from the craziness of this conclusion; the one I want to draw is that the difference between assertion and expression does not give us a good model of the difference between the non-normative and the normative.
Mark;
Why is the upshot that there aren’t any quantifiers in English, rather than that a good way to explain how quantifiers work is by thinking of them as devices of infinite conjunction (disjunction)?
Tim;
I thought all expressivists agreed that the difference between assertion and expression isn’t a good model of the difference between the non-normative and the normative. When I assert “Snow is white”, I express my belief that snow is white, but I neither say nor do anything normative. All expressivists agree abut that, right?
Here’s the part of your argument that I’m not getting. Why is an expressivist supposed to think that calling p true is expressing a normative attitude toward p? When I say that it is true that snow is white, which normative attitude am I supposed to be expressing toward the proposition that snow is white?
Hi Jamie…
1) We may be at cross purposes here. I thought expressivists said that the difference between assertion and expression WAS the difference between non-normative and normative utterance. Don’t they say that?
2) An expressivist isn’t supposed to think that calling p true is expressing a normative attitude toward p, except on certain other assumptions. For instance, if an expressivist thinks that what we use the truth-predicate for is (basically) just to mark a sentence as to-be-believed, then it looks like he thinks that calling p true is basically) expressing a normative attitude toward p. Again, if an expressivist agrees, as I think he should, that truth brings with it normativity, then given that the expressivist wants to cash out normative assertions of as expressions of pro-attitudes, it looks like he will have treat the normativity of truth, too, as a matter of expressed attitudes.
Dave is no fun at all.
Or is that “Not (Dave is any fun at all)”?
Tim;
Well, whenever we assert that p, we express the belief that p, according to expressivists (and I thought this was supposed to be common ground, in any case). That’s so no matter whether p is normative or purely descriptive. So there are plenty of descriptive assertions that are also expressions of a state of mind, and those aren’t normative at all.
Hm. I suppose that depends on which sort of minimalism your expressivist adds on. I don’t exactly understand the idea of marking something as ‘to be believed’. I would have thought that insofar as it applied, it again should be common ground that asserting something is marking it ‘to be believed’ (truth predicate or no). Insofar as the marking is supposed to be anything beyond that, I’m unclear on what it’s supposed to be.
Suppose an expressivist adopts Paul Horwich’s view of truth. Then the meaning, the inferential capacity, of calling something true is delivered by the collection of instances of the T schema. So, for instance, when I straight out say of the proposition that snow is white, that it is true, we use it is true that snow is white iff snow is white to see that what I said implies that snow is white. No normativity there.
I share Jamie’s bafflement at Tim’s argument. It’s surely false that “p is true” either means or entails “p is to-be-believed.” It certainly isn’t analytic that if p is true anyone ought to believe it. In fact it’s surely false. There is a massive conjunctive truth which reports correctly the national insurance numbers of everyone in the United States but it is true of no one that they ought to believe that conjunction – it would be a pointless expensive waste of mental storage space. Or, as Gibbard observes somewhere I think, I might quite reasonably think it better if I refrain from forming true beliefs believing which would lead me to become morally corrupt (e.g. about how frightfully well off I might become were I to yield to the temptation to accept bribes.)
It could be replied that at least a pro tanto ought is implicit in saying p is true. But care seems appropriate. A *hypothetical* such ought is arguably in the offing when I assert that p is true IF I endose the norm widely held constitutive of belief that belief should aim at truth. But as the same is no less plausibly true, prior to any semantic ascent, when I assert simply that p, no”is”-“ought” gaps are getting crossed.
Tim,
about this:
“Notice that this commits the expressivist to the view that adding “is true” to any proposition is a way of expressing one’s acceptance of it, and no doubt of inviting others to accept it too, etc. In which case uses of “is true” look like they ought to be classified as normative by expressivists. In which case any expressivist can get across the supposed is-ought gap by this simple argument:
(Premiss) p;
(Conclusion) p is true.
The premiss is (let us grant) non-normative; the conclusion (as we’ve just seen) is, by your expressivist’s lights, normative. Voila, a valid is-ought inference.”
I think you are making a use-mention confusion. The introduction use for the logical truth-operator is rather:
Premise: p
Conclusion: ‘p’ is true.
You don’t add ‘is true’ to a proposition but rather a name of a proposition. What you get as a result of is asserting or thinking or accepting the original proposition – whatever it is. If the original proposition, p, is normative, then so is the saying the proposition ‘the proposition ‘p’ is true’ is it is the very same proposition. If not, then so is not the conclusion either. Just a matter of empty semantic ascent. And in the case we discuss ‘whatever Jon said’ is a demonstrative used to name some said proposition like ‘utilitarianism’ is a name for certain ethical theory and its claims.
Jamie, Jimmy, Jussi:
We’re moving away from where we started… not that matters.
The important point, I think, is about the normativity of truth. I think truth is a normative notion: not in the simple way that Jimmy rightly rejects (if p, then you ought to believe that p), but in some subtler way (if p, then if you believe anything on the question whether p, ceteris paribus you should believe p… or something like that). I don’t think (1) “p is true” MEANS (2) “there is (some as-yet-to-be-explained sort of) reason to believe p”– and, though it is tempting sometimes, I’m not outright attributing that view to anyone else. But I do think (2) can reasonably be inferred from (2) (NB I didn’t just say that (1) *entails* (2)). And I do find the denial of this view– which is what I mean by saying that truth is normative– puzzling. The question then is what expressivists who don’t deny the normativity of truth are to say about what that normativity consists in. As I’m writing on this at the moment, I’d be interested to hear more about your views, if you’ve time.
Doh. “I do think (2) can reasonably be inferred from (2)” should, of course, have read “I do think (2) can reasonably be inferred from (1)”.
Hi, Jamie.
Either way, it looks like a general hypothesis about quantifiers, even in sentences that don’t contain the word ‘true’ – and on the face of it, we were just supposed to be getting in for a minimalist account of ‘true’. And like I was saying, such general hypotheses look like they ought to be testable in their own right. If they fail, then so does this way of elaborating on minimalism in an expressivist-friendly way.
Also, standardly interpreted, quantifiers have more expressive power than abbreviations for long disjunctions, so long as we don’t have names for everything in our domain – don’t they? Now if you’re only worried about dealing with sentences like ‘what Jon said is true’, then presumably it is safe to assume that Jon couldn’t say it unless there is a way of saying it. But things get harder with sentences like ‘something beyond human comprehension is true’, and the assumption about having language covering the domain goes out the window, if we’re using this treatment for quantifiers everywhere, as in ‘we know relatively little about anything more than one billion light years from the Sun’. Clearly there aren’t sentences to even _be_ the disjuncts in the infinite disjunction that this would approximate.
And finally, on a more telling but different note, how are we supposed to understand ‘most things Jon said are true’ or ‘many things Jon said are true’ on this model? Maybe Jussi can help me out with this one, but binary quantifiers like these have more structure and more expressive power than standard Fregean quantifiers, so what is going to allow us to treat these sentences without any sort of quantifier? My point is just that the move Jussi alluded to is far from a gimme.
Tim –
I don’t see why your argument doesn’t cross-cut expressivism. Suppose that we agree with you that truth is ‘normative’. But suppose that we’re not expressivists. Still, the following argument will take us from a non-normative premise to a normative conclusion:
1) P.
2) It is true that P.
Now, you yourself bought into these assumptions. So you yourself get an ‘ought’ from an ‘is’. Now – what was the problem supposed to be? Was the problem supposed to be that _ad hominem_, expressivists don’t want to get an ‘ought’ from an ‘is’, but you actually don’t mind it? I’d want that idea to be sharpened a lot further. In precisely what sense don’t expressivists want this? Maybe expressivists will deny your ‘normativity of truth’ premise and be off the hook. Maybe they’ll distinguish between kinds of ‘is’-‘ought’ inferences that are okay and kinds which bother them. As it stands, I’m not quite getting the force of why the problem applies specifically to expressivists, though.
Tim,
going further away from the topic, the views that would support your claim about reasons and truth are anti-realist ones. If you think how things in the world are is independent of our beliefs and the justification for them, then the inference is not valid. One problem with the inference as it is stated is that of course there are yet to be explained reasons to believe propositions that are false. We often get good new evidence that actually gets us further away from the truth. So, from merely the existence of evidence we don’t have we cannot infer the truth of any proposition. BTW, talking about inferring from the existence of reasons we don’t know anything about sounds funny. There is a temptation to say that making an inference from something requires knowing what that is.
It’s also not that deflationists deny that truth is normative. Horwich is good on this. The just think that the normativity of truth as a mere logical property is derived from something else. We have reasons to believe true propositions but it’s not their *truth* that gives reason to believe them. What gives reason to believe that snow is white is that *snow is white*. The talk about the *truth* of that proposition that snow is white does not add anything to that reason there already is. In fact it is saying the same thing.
Mark,
I wasn’t saying that it is a solve it all solution but something worth trying. I thought the view was that ‘whatever Jon said’ is understood as a demonstrative that directly refers to Jon’s utterance (an event in the world). In that way, we *avoid* having to use the disjunctive that would have as its domain all possible statement Jon could say. I agree that the idea of all possible statements as a domain is problematic especially in the case of the propositions beyond human understanding. So, my suggestion was supposed to enable us to avoid all of that.
That we refer to the statement is just the reason why we need ‘is true’ – to help us get from a reference to the statement to using it.
The same goes for many and most things Jon said. I could say that there we are demonstratively referring to all Jon’s statements and saying that we are willing to accept many or most (vaguely) statements of that domain.
“I don’t think (1) “p is true” MEANS (2) “there is (some as-yet-to-be-explained sort of) reason to believe p”– and, though it is tempting sometimes, I’m not outright attributing that view to anyone else. But I do think (2) can reasonably be inferred from (1)”
WHat can be conceded is simply that from (1) p is true we may infer (2′) p.
SO, *given* (and only given) that we accept certaion episetmic norms directing us to aim at truth in belief, you ought (pro tanto), if you believe (1) p is true to believe (2′) p. But of course it is equally the case that, given the same acceptance, if you believe (2′) p you ought to believe (2′) p. So your conclusion still isn’t normative in a way your premise isn’t already normative.
If we accept a norm of truth (which isn’t the same as accepting that truth is a normative concept) then in some weak sense maybe: if p is true we ought to believe p. If you mean no more than that by saying if p is true there is some reason to believe p, fine. But we may with euqal propriety say, if p there is some reason to believe p so again premise and conclusion come out on all fours for “normativity”. (I assume you don’t mean something stronger like, If p is true then we have some evidence for it or some available warrant for it – that would commit you to a sort of anti-realism I doubt you want to buy.
Indeed if you are going to understand any sentence as normative if its assertion carries certain inferential commitment, Hume’s principle is trivialized as every sentence becomes normative. But it is surely neither necessary nor sensible to understand normativity like that.
Mark,
I’m not following, exactly — are you listing problems for minimalism, or more specifically for combining minimalism with expressivism? There are, certainly, problems that minimalists have to solve and I don’t personally know how to solve them, but I’ll let Hartry Field and Paul Horwich deal with them.
Tim,
I almost see your point, I think. Maybe it actually has nothing to do with the truth predicate at all? You think that from p, we can validly infer that one ought, ceteris paribus, to believe that p. This is an analytic ought from is. Is that right? This would be independent of one’s view about the truth predicate, and independent of one’s views about expressivism.
I seem to remember that Gideon Rosen has written something about that issue. I’m just about certain that Gibbard has something on it, too.
Hi, Jamie.
I meant to be elaborating on one of the problems I took to be posed by Heath’s original post. I took Jussi to have articulated one kind of answer to it (although now it looks to me like he has two), and I just meant to be articulating further potential costs of going that way. I’m not expert enough to assess whether it is a general problem for minimalism about truth, or only for versions that are compatible with expressivism. Some kinds of minimalism about truth are happy to quantify over propositions, which seems like it ought to help solve some problems, but expressivists apparently can’t go in for that variety.
Could we get back to the question of whether there is any fun at all that I am?
Mark,
sorry about that. I might have mislead. Usually how the two solutions is presented by the minimalists is that there are supposed to be two logically equivalent claims (i) one that lists all possible utterances and saying each one of them in disjunction [‘(Jon said that ‘p’ and p) or (Jon said that ‘q’ and q), and so on’] and (ii) one that refers to a particular utterance and then uses the disquotational feature of the truth-predicate to remove the quotation marks so to speak. The latter is supposed to just be the actual utterance ‘whatever Jon said is true’. We have pragmatic reasons to say that instead of the infinite disjunction that is supposed to be co-extensional. This is supposed the be the kind of purpose we have the minimalist truth-predicate in our language in the first place. But, you might have given good reasons to claim that (i) is not even in the offering.
If expressivists accept Gibbard’s fact/plan propositions, then they too could quantify over propositions in a way.
Take it from me, David, you’re lots of fun.