There’s been a great discussion on an earlier thread about expressivism.  I don’t want to kill it off, but I do want to raise a diferent worry about expressivism.  It appears below the fold.

 I’m going to assume in what
follows that Mark Shroeder is right about what the expressivist ought to say in
response to my earlier worries. I’m still
worried, but I want to push forward a

On behalf of the expressivist, we’ve distinguished between
BELIEF (the cognitive state) and — let’s call it "Elief" — the
state that I am in when I sincerely say that I believe something is wrong. 

According to the expressivist,

I do not BELIEVE that torturing cats is wrong, but I do
elieve it. Because expressivist
semantics is true, we can truly say I believe that torturing cats is
wrong, because (perhaps) in ordinary english "belief" means "BELIEF or elief." 

Now for the worry.

Here’s something everyone (except perhaps expressivists)
accepts: that what one believes can also be what one desires or fears or hopes,
etc. That is, anything that I believe I can fear, hope, or direct
some other propositional attitude towards.

An example: I believe that Tommy’s torturing the kitten is morally
wrong.  I
don’t like Tommy, and believe that God will punish people who do morally wrong
things. So I am glad that Tommy’s
torturing the kitten is morally wrong. 

The cognitivist says: there is a proposition — the proposition that Timmy’s torturing the cat is morally wrong — and this propsosition is the content of two attitudes, a belief and a hope.

What does the expressivist want to say?  Here are some possibilities.

Option 1:  There is a proposition — the proposition that Timmy’s torturing the cat is morally wrong — and it is the object of two attitudes, a hope and an elief (not a BELIEF.)

Question: If this proposition can be the object of a hope why can’t it be the object of a BELIEF?  If this proposition can be the object of a hope, then certainly it can be the object of a fear, a desire, a contemplation, etc. (Actually, what does the expressivism want to say about "contemplation" or "entertaining"?) Why then not a BELIEF? But if this proposition can be the object of a BELIEF, what’s left of expressivism?  Isn’t the expressivist committed to saying that I never BELIEVE propositions of this sort?

Option 2:  Just as we distinguished elief and BELIEF, we must also distinguish ear and FEAR, ope and HOPE, etc. etc.  In this case, I do not have a BELIEF and a HOPE,  I have an elief and an ope.  (The content of elief and ope is the same, but presumably this content is not a full proposition: instead, the content is something less — I am guessing that the best thing to say here is that the object of my ope and elief is simply Timmy’s act of torturing the cat.) 

Question: Isn’t this too much? Bad enough we need to distinguish two mental states (BELIEF and elief) where previously we thought we had only one — but now we must say that each genuinely propositional attitude has a deformed twin brother.   If expressivism requires this, shouldn’t we give up expressivism? 

Option 3: The case I’ve described can’t really happen.  I can’t hope that Timmy’s act is morally wrong.  Whatever is going on, that can’t be it.

Question: Ok, this isn’t really a question, but this is obviously not acceptable, right?

Option 4:  Someone smarter than me should help out here!

28 Replies to “Expressivist Psychology

  1. Kris,
    I wonder if the expressivists will want to be more presice on what the moral propositions are. Of course, they are going to accept that there is a proposition, in a non-committal, minimal sense, that ‘torture is wrong’. But, in this sense, to say that there is a proposition that such and such is the case is just a logical device to make a noun-phrase out of a declarative sentence.
    However, the PROPOSITION that is the content of the expression “Timmy’s torturing the cat is morally wrong” is the PROPOSITION [that Timmy is torturing a cat]. It is toward this content that the utterer of this claim has the elief attitude. And, of course, towards this content, ‘that Timmy is torturing a cat’, we can adopt attitudes belief, hope, and so on attitudes.
    Of course this still leaves the question of what is going on when we hope that such and such is wrong. I think the expressivist could accept some sort of second-order attitude account of this. So, saying that would be expressing favourable attitudes towards having pro or con attitudes towards certain act or acts. Of course, one can have such second-order attitudes with or without having the first-order attitudes – maybe it would make sense to say that ‘I wish kindness was wrong’ if I wished that I had con-attitudes towards kindness even if I actually thought well of kindness.
    The expressivist can also say that one can have a BELIEF about having con-attitudes towards certain acts, but we probably wouldn’t express this by saying that “I believe that torture is wrong’ but rather by saying that for instance that ‘I know that I disapprove of torturing’.

  2. I’m not sure I find option 3 unacceptable. I’m not at all sure you can hope that 2+2=4, for example. Hoping that a moral claim is true strikes me as even stranger (unless it’s a disguised way of hoping that some non-moral facts obtain, which may be the case in your example).

  3. Hi Aaron,
    I have to disagree with you pretty strongly about how strange it is to hope that a moral claim is true. Not only isn’t it strange — it’s incredibly common. If expressivism can’t make sense of our doing this, expressivism is in big trouble.
    Suppose I am wrestling with a decision. I believe that there are several moral reasons against doing the act in question. There are also several moral reasons in favor of doing the action. I don’t know how to weigh these reasons against each other. I’ve got to decide what to do. I perform the action, hoping all along that I am doing the right thing. Surely you’ve been in that situation? When you are in that situation, you certainly are not merely hoping that some non-moral fact obtain. You are hoping that you are doing the right thing.
    Hi Jussi,
    I’m not sure I understand your response. You wrote:
    “I think the expressivist could accept some sort of second-order attitude account of this. So, saying that would be expressing favourable attitudes towards having pro or con attitudes towards certain act or acts. Of course, one can have such second-order attitudes with or without having the first-order attitudes – maybe it would make sense to say that ‘I wish kindness was wrong’ if I wished that I had con-attitudes towards kindness even if I actually thought well of kindness.”
    Here’s my concern.
    Bob is hoping that Timmy’s act of torturing the cat is wrong. But he isn’t hoping that he (Bob) disapprove of Timmy’s torturing the cat. In fact, he is certain that he doesn’t disaprove of tortuing the cat — Bob is himself a sicko, and delights in the fact that Timmy is inflicting pain. And Bob doesn’t want to change. So he certainly is not hoping that he has (or will develop) a con attitude of some sort towards Timmy’s torturing the cat. He’s a sicko, and he wants to remain a sicko.
    Just to be explicit: He is pleased that Timmy is torturing the cat. He is pleased that he is pleased that Timmy is torturing the cat. He has no con-attitude towards Timmy torturing the cat. He has no pro-attitude towards himself having a con-attitude towards Timmy torturing the cat.
    But Bob does hope that Timmy’s torturing the cat is wrong. He hopes this because he believes that God will punish Timmy for doing something wrong, and he wants Timmy to be punished. (He’s a sicko remember.) So Bob’s hope is an “instrumental” or “means-end” hope, but it is a hope nonetheless. He hopes that Timmy’s act is wrong because if it is, certain effects will occur.
    It’s a sick, messed up world. There are probably actual people like Bob. Surely there could be, even if there aren’t.
    In any event, I’ve made the story dramatic just so it’s vivid. I’m confident I could tell a much more ordinary (but less interesting) story to make the same point.
    So I don’t see how appealing to second-order attitudes helps in this situation, but maybe I’m just not getting what your response is.

  4. I’m with Kris. This is one of the problems for expressivists in the philosophy of mind that I raised in my graduate seminar this semester, and discuss in a draft of a chapter for a book I plan to write. I think it is an _enormous_ problem for expressivists, and it is part of the evidence that I’ve been alluding to that it is far from trivial that expressivists will be able to give a semantics which gets everything to fall out right, just because Simon Blackburn has stipulated that that is what Quasi-Realists aspire to do.
    I don’t think, however, that Kris was right to characterize the toy expressivist semantics for ‘believes that’ that I offered as being disjunctive. It was a univocal account that appealed to the notion of expression and from which what Kris is calling BELIEF and elief fell out as special cases. The metaphysics was disjunctive, but not the semantics.
    Notice that this is not a problem just about attitudes. It is a problem for every complement-taking or infinitive-taking verb. ‘Jon illustrated that murder is wrong’, ‘Jon wondered whether murder is wrong’, ‘Jon plans for murder to be wrong’ and so on. The problem is just one small piece of the embedding problem for expressivism – expressivists must give an account for what these sentences even _mean_, because on standard accounts, the embedded clauses contribute propositions.
    As I see it, to solve this problem, an expressivist would have to 1) come up, for each such construction, with a distinct metaphysics for what goes on in the normative cases; 2) come up with a novel semantics; 3) explain why the traditional metaphysics falls out as a special case when descriptive sentences are embedded; 4) explain why the distinct metaphysics falls out as a special case when normative sentences are embedded; and 5) explain why the distinct metaphysics yields all of the predictions that we would ordinarily expect, and is not error-theoretic in its own right.
    For example, in the case of belief, Horgan and Timmons think the state Kris calls ‘elief’ is something they call ‘ought-commitment’. They are committed to explaining why ought-commitment has the same sorts of properties as what Kris calls ‘BELIEF’ and they call ‘is-commitment’ – for example, same sort of phenomenology, etc. Then they have a semantics for ‘believes that’ from which each state is supposed to fall out as special cases. This requires a lot of work just for belief. They will have to do it all over again for wondering whether – new psychological state, new semantics, new explanation of why hope and ope have similar phenomenological characteristics, new account of why they both fall out as special cases of the same semantics. If this sounds unpromising, then you understand why saying what they aspire to do is a long ways, for expressivists, from doing it. I’m incredibly pessimistic on their behalf on this score.

  5. Kris,
    you are probably right that the higher-order account won’t work in all cases. Here’s a worry and/or one possible solution. You write that:
    ‘But Bob does hope that Timmy’s torturing the cat is wrong. He hopes this because he believes that God will punish Timmy for doing something wrong, and he wants Timmy to be punished.’
    I know that many expressivists say that we don’t always use moral terms to express attitudes. Sometimes we for instance use them to describe what is forbidden in the society – the that’s ‘bad’ use. So, in this case it is enough for Bob to hope that God considers torturing the cat to be wrong. In this case then Bob’s utterance ‘I hope burning cats is wrong’ can be seen to say that I hope God disapproves of burning cats.
    I actually think that the expressivists could look at the uses of the embedded claims and by piece-meal investigation explain what attitudes they are used to express. For instance, I cannot imagine any uses for ‘Jon illustrated that murder is wrong’ or ‘Jon plans for murder to be wrong’. I can’t think of situations where these could be used or where they would be true or false. They are grammatical sentences but I’m not sure I understand their meaning if they have such. If you give a story where they would be appropriately used, it would be easier to start looking for semantic accounts of them – expressivist or not. But, ‘Jon wonders whether murder is wrong’ looks already easier. We can imagine where we could say this and even where Jon would say something like this himself. In the situation I have in mind Jon uses the utterance to express the uncertainness of his attitudes towards murder. We can use the third-personal claim to describe Jon’s confusion. Anyway, I’m not sure the expressivist needs to accept that the moral part in the embedded claims makes always the same contribution but she can look at the situation and give an explanation of what contribution the moral bit makes in the given embedded claims. It is a big project but I’m not too pessimistic that it could be carried out.

  6. I just looked at my original post. I switched from “glad” to “hope” within two seconds. Weird. I’m glad that didn’t throw anyone off though.
    If the best expressivism can do is take each case as it comes and try to tell a story about that particular case, then I would think that expressivism is sunk. But let me push forward still.
    Suppose Bob hopes that Timmy’s act is wrong. (Same case as before — he wants Timmy to get punished.)
    Suppose Fred hopes that Timmy’s act is wrong. But not because he wants to Timmy to be punished, but because Fred holds a moral theory that will be counterexampled if Timmy’s act is not wrong. No philosopher likes that to happen!
    You seem to be committed to saying that the content of Bob’s hope is not the content of Fred’s hope. But it’s the same content, isn’t it? (What’s true of course is that Bob hopes P because he thinks P leads to Q, whereas Fred hopes P because he thinks P leads to R. But they both hope Q.)
    I also want to respond to a minor claim of yours:
    “For instance, I cannot imagine any uses for ‘Jon illustrated that murder is wrong’ or ‘Jon plans for murder to be wrong’. I can’t think of situations where these could be used or where they would be true or false.”
    I’m not sure what’s meant by “illustrated” — “produce a visual image”, “demonstrated that”, etc., so I’ll focus on planning.
    Consider the following case:
    Jon holds two crazy views. First, he holds that God will punish the wrongdoer. Many people believe this. Second, he holds a kind of relativism: if enough people say that something is wrong, then it becomes wrong. Many people seem to believe this as well. So you shouldn’t have a hard time conceiving someone who believes both at the same time. Many freshmen do.
    Jon wants Sally to be punished by God. But Sally is sweetness and light. As things stand, she does nothing wrong. She spends a lot of time working at Habitat for Humanity. Her character is incorruptible.
    So Jon decides that the only way God will punish Sally is by changing what’s wrong. He plans to bring this about: specifically, he plans for working for habitat for humanity to be wrong. So he starts an insidious advertising campaign, which is designed to bring about widespread belief in the claim that helping the homeless is wrong. (Maybe he mails out copies of The Fountainhead.) People are convinced. Working for habitat for humanity is widely viewed as wrong.
    Jon is delighted. By his own lights, his plan has succeeded. That is, he planned for working for habitat for humanity to be wrong — and given what he believes, it is now wrong. And he is delighted, because now Sally will be punished by God.
    Now this story is admittedly goofy. But although cheesy relativism and religious fundamentalism make for strange bedfellows, many of our students have had a threesome with them. All my story requires is that one of these students cleverly think through a few steps.
    The moral I draw is that we should be very cautious about assuming that semantic compositionality fails in these cases. Maybe it does. But the expressivist should not get her hopes up, or base a defense of the view simply on it being initially somewhat difficult to see how these weird sentences could make sense.

  7. Kris,
    I don’t think that it is implausible to say that people use terms like ‘wrong’ occasionally in different senses and that we need to look at the cases to determine what is going on. The central case the expressivist is concerned is the use that is at least in part capture by the motivational internalism thesis. The idea being that when we use ‘wrong’ sans phrase this implies that we are motivated at least to some extent (or practically irrational). All the stories you have given seem to be ones where ‘wrong’ is used in some other, deviant way – either to denote what is forbidden by God or what is forbidden by the society. In these cases the internalism requirement is already failing so the terms are not used in the basic moral way the expressivist wants to account for.
    The case of Bob and Fred seems to illustrate this well. Bob is a scoundrel who does not care if something is wrong. Expressivists, as internalists, would that he isn’t then making actual wrongness judgments but rather ‘wrongness’ judgments. If Fred on the other hand uses wrong in the main sense, he is making wrongness judgments. This would make clear the difference in the contents of their hopes.
    Take the utterance ‘Jon plans helping the homeless to be wrong’ and the story you give where it is used. It’s clear that in that sentence you don’t understand wrong in the same way as you usually do. The story illustrates how Jon goes about making helping homeless *”wrong”*, i.e., disapproved by the society. But, my intuition, and I hope yours too, is that he hasn’t succeeded in making helping homeless *really* wrong by doing this. Even if helping homeless was disapproved by the society, it would be an open question whether it was wrong. In fact, it wouldn’t be. I don’t think that anyone would want to say that in this case wrong is used in the normal sense that makes the standard semantic contribution to the meaning of the complex claims.

  8. Hi Jussi,
    Thanks for continuing the discussion. It’s been really stimulating.
    You wrote:
    “Take the utterance ‘Jon plans helping the homeless to be wrong’ and the story you give where it is used. It’s clear that in that sentence you don’t understand wrong in the same way as you usually do.”
    I disagree — when stating the story, I used the word “wrong” in the way I ordinarily use it. In the story, Jon has a theory about what makes actions wrong in the ordinary sense. The theory happens to be false. That can happen. This is a story in which it does happen. I didn’t tell a story in which I used “wrong” in an unusual sense.
    It would be weird if I can’t tell a story in this way. Why not? Don’t people sometimes have false beliefs (or false eliefs, if you prefer) about what’s wrong? Doesn’t the expressivist want to say that:
    (1) It is false that what’s wrong is determined by what people believe.
    (Recall our ealier discussion — if the expressivist denies this, his view is obviously false, since this is a truism, a platitude, etc.)
    So doesn’t the expressivist want to say that :
    (2) Jon’s relativism is false, since it implies that what’s wrong is determined by what people believe.
    So Jon has a belief (maybe its just an elief) about what makes these things wrong, wrong in the ordinary sense. And it’s a stupid elief to have. But many people have it, and Jon has it. These people are making a mistake about what makes acts wrong, but they are intending to talk using the same sense of “wrong” as everyone else.
    (Here’s another example, to illustrate the same point. Suppose you are a utilitarian, and you dislike it when people do action F. You can plan to make doing actions of that sort wrong, and it’s pretty easy to see how.
    Just devise a plan to ensure that the consequences of doing F always make it the case that doing F is suboptimal. Then it will be wrong to do F. Surely someone could have this sort of plan?)
    Secondly, in response to: “But, my intuition, and I hope yours too, is that he hasn’t succeeded in making helping homeless *really* wrong by doing this.”
    I agree that he hasn’t succeeded in making helping the homeless really wrong. But that’s not relevant. One can plan that P and fail. Bob had a plan – a plan to bring about that helping the homeless is wrong — and of course it failed. The only reason Bob thinks it suceeds is that he holds a false normative theory. This is why Bob believes that his plan suceeded.
    So Bob definitely had a plan — a plan that failed miserably. (He should run for the US presidency.)
    Mary plans to get God to love her. So she helps the homeless, because she believes that God loves people who help the homeless.
    That’s a coherent story, and it’s coherent even if there is no God. When we describe what she is planning, we don’t need to assume that her plan even has a chance of succeeding. That Mary’s plan failed implied that she had a plan.
    The point of the example I used earlier is that it makes sense to say of Bob that he has a plan, a plan to make it the case that murder is wrong. There was a worry originally that the no one could ever have a plan like this — but what I’ve done is shown that someone could. It’s irrelevant that no one could have a *successful* plan like this — I didn’t try to show that, and I don’t think it’s possible. (I’m no relativist.)
    And the cautionary moral I drew from all this is that it is not that hard to come up with reasonably sensible examples of someone having other attitudes besides eleif to moral propositions. Cognitivists have zero difficulty explaining why this is. Moral propositions are propositions, differing from other propositions only in that their subject matter is different. It’s not surprising that one can hope and believe biological propositions, propositions about gym class, propositions about love, etc.
    Moreover, it looks like the cognitivist semantics can be much simpler: it could even be compositional.
    Expressivism, on the other hand, doesn’t seem able to make sense of these examples except in a piecemeal fashion. It looks like compositionality will fail in a radical and surprising way.
    (One final point: Since we both agree that Bob’s plan failed, we both agree that Bob had a plan for helping the homeless to be wrong. But Bob did succeed in bringing about that people disapprove of helping the homeless.
    So whatever the content of Bob’s plan is, it’s final end is not simply that public opinion go a certain way. That was a vital component of Bob’s plan, that component suceeded.)

  9. Kris,
    I’m still not convinced that you used ‘wrong’ in the ordinary sense in the story. If we accept even a weak judgment-internalism, then when you say that something is wrong in the ordinary sense you are making judgments that imply that you are motivated to avoid the relevant actions. I take it that in the story when you mention wrong, you did not refer to that which you are motivated not to do. And, clearly Jon is not making judgments about that as he pulled to act in the case he thinks something is wrong. Notice that if he had eliefs about wrongness as being forbidden by the society, then he would be motivated not to do wrong acts too. But, he is not.
    Of course, you might want to deny the internalism thesis, but then you are cutting expressivism off already at an earlier stage.

  10. Jussi –
    Let me point out just how weird something you said is. According to judgment internalism, someone who _judges that_ being friendly is wrong will be motivated, qualifications aside, not to be friendly.
    No one has ever claimed, that I know of, that someone who _hopes_ that being friendly is wrong will be motivated, qualifications aside, not to be friendly. After all, they might hope that it is wrong without judging that it is. And the same goes for every other attitude verb. No one has ever defended the view that anyone who wonders whether being friendly is wrong must be motivated not to be friendly – and rightly so.
    Judgment internalism is a test on _judgments_. That is, it is a test on _beliefs_, in the ordinary, everyday sense in which people agree that there are moral beliefs. It has never been, and could never plausibly be, a test on concept deployment even in other kinds of thoughts. The test you are suggesting can rule these other thoughts as not deploying the concept, ‘wrong’, is not the test of judgment internalism.
    Kris –
    I think you’re wrong about the failure of compositionality. That doesn’t directly follow. As I see it, and as I have been saying all along, it is the ambition of an expressivist view to give a genuinely compositional, univocal semantics for all of these constructions. I don’t, myself, think that this project is going to succeed, as I have also been trying to make clear. But the toy semantics I gave for ‘believes that’ was supposed to illustrate how a univocal compositional expressivist semantics for this particular construction could work.
    According to expressivists, sentences express mental states. So I think of an expressivist compositional semantics as operating on those mental states. ‘Jon believes that P’, my toy theory went, expresses the belief that Jon is in M, where M is the state expressed by ‘P’. That’s compositional in the same way as expressivist accouns of ‘not’ are compositional – they operate on the mental states expressed by the embedded sentences, to yield the mental state expressed by the whole.
    Of course, ”P’ expresses M’ can’t just mean, ‘To believe that P is to be in state M’. But I think expressivists typically have an independent idea of what expression is, and I myself have offered a view about how it should be understood, in order to make the most of expressivism (in ‘Expression for Expressivists’). I’m not saying that this semantics is going to work – I think it is open to some clear objections, and in the semantic framework I develop for Expressivists in my book manuscript, ‘Being For’, I can show how to do a little bit (though not much) better. The point is that it’s not beyond the ambitions of expressivists to think that they can make an account like this work, and do it in a compositional way. I’m with you on seeing no grounds to think that it can be done; this is a quibble: I just think you overstate what is obvious when you put things that way.

  11. Mark,
    thanks for that. I’m not sure I said anywhere that hoping at al are supposed to satisfy the internalism requirement. The thought was more that if I, for instance, hope that kindness was wrong, then I need to be using ‘wrongness’ in a way that I can make judgments that satisfy internalism about it. So, if I came to believe that my hope was satisfied in some odd scenarion and kindness became wrong, then I should be motivated to avoid it. This condition was not satisfied in the stories Kris gave. The person who hoped that kindness would become wrong didn’t care at all about what he referred to with kindness and would not start doing so. My hypothesis was that when he was hoping that such-and-such would be wrong, he was therefore not talking about wrongness in the ordinary sense.

  12. Jussi –
    Okay, then I interpreted you uncharitably. Apologies. I still don’t see why Kris’s stories turn on anything like that. I didn’t see him build it in, and I don’t see why you do, either. In fact, I don’t even see why we need Kris’s stories. If hopes puzzle you, try ones with negated contents. For example, does ‘Jon hopes that masturbation is not wrong’ sound like a meaningful sentence to you?
    Moreover, these sentences don’t even have to ever be _true_, for Kris’s objection to work. Take his example of planning. You don’t think he has come up with a real example of a plan to make something wrong. That’s because you think that in his example, the characters involved _don’t_ plan to make something wrong.
    But if that’s what you really think, then you clearly think that these sentences are meaningful – you think they are false, and their negations are true. Kris’s challenge is to say what their meanings are, and to do so in a systematic way.

  13. Mark,
    I guess my intuition just is that the expressivists might want to be able to say different things about the different cases. Some of the cases will turn on in which sense the terms are used – whether wrongness or ‘wrongness’ judgments are made.
    Here’s a first pass on what I would like to say about the masturbation example and it’s something I said earlier. Jon has a range of acts that he thinks are wrong – the ones towards which he has the suitable con-attitudes. I think it makes for him to hope that masturbation does not fall under that category irrespective of whether he disapproves masturbation, i.e., thinks it to be wrong. That case does not seem to problematic.
    I’m not sure that in the planning case I think the claim is false but rather non-sensical. I thought that it would help me undestand what is meant by ‘to plan that something is wrong’ if I could see cases where such plans could be attributed to someone. Kris came up with a case that he thought was a suitable one but my suspicion was that in this case wrong was used to mean something deviant. Similarly, I’m not sure what counts as planning to draw a square circle so I’m not sure what is meant by the claim that ‘Jon plans to draw a square circle’. If we think that all such claims are false and nothing counts as success, then I’m not sure meaning has been given for the expressions.
    I guess I’m wondering whether these are meaningful sentences in the first place despite of looking like they are. If they are not, then the expressivist should not be required to give meaning to them.

  14. Jussi –
    Something _does_ count as success at drawing a square circle. You have to draw a circle. And it has to be square. It is just that since necessarily, nothing is both a circle and square, it is impossible to succeed. But the only reason we _know_ that it would be impossible to succeed is that we know what it _would_ take to succeed, and _that_ is impossible.
    Compare: it is impossible to construct a trisection of an arbitrary angle using a straightedge and a compass. But some very smart people planned, over a very long period of time – roughly 2200 years – to do exactly this. From what I’ve heard, even Gauss tried to do it early in his career, and he ranks pretty highly on history’s list of great mathematicians. Yet by your standards, since it is impossible to do it, nothing would count as success, and so not only did they not plan to do it, it isn’t even possible to meaningfully say that they did. If I say, ‘they planned to trisect an arbitrary angle using only a straightedge and a compass’, I must be using ‘trisect’ or ‘arbitrary’ or ‘angle’ or ‘straightedge’ or ‘compass’ in some nonstandard, inverted-commas, sort of way. If that’s your view, I’m happy to part ways with you.
    Moreoever, the claim that ‘Jon plans for murder to be wrong’, where ‘wrong’ means what it usually does, is literally meaningless, does look like a violation of compositionality, of the kind Kris was complaining about.

  15. Mark,
    maybe you are right about the mathematics cases. I’m not sure. No matter how much I try to come with a plan to draw a square circle I can’t do it. I really don’t know what I would be aiming at. Or, what anyone would be aiming at. But, maybe this is just the limits of my imagination.
    I grant you the trisection case. In that case you are right that we know what would constitute a success – we know what using a straightedge and a compass is and we know the wanted end-result a trisection of an arbirtary angle is. This much must be true even if it is actually impossible.
    I’m not too bothered if it’s not true that all grammatically correct complex sentences get a definitive meaning compositionally. I’m one of those people who think that ‘colourless green ideas sleep furiously’ does not make any sense. Galen Strawson has often tried to convince me to think otherwise. Maybe he knows what the sentence means but I don’t.

  16. Hi Mark,
    Thanks for your clarificatory remarks. I am on board with what you say.
    Just out of curiousity, what do you want to say about this case, which is a variant on the paradox of the preface:
    I believe that I am a good guy. I think about all that I’ve done, and I approve. That is, I think of each individual action, and I view that action favorably. But I am aware that most people are incapable of morally perfect behavior, and after reflecting a bit, I conclude that I am no different from anyone else. I conclude that I probably have done *something* wrong, but for each action you pick, I’m very confident that this action is not wrong. (And of course, I *hope* that none of my actions is wrong.)
    According to the cognitivist, there is a puzzle here — but it is no more puzzling than the paradox of the preface. What do you want to say about this case?
    And in general, what does the expressivist want to say about sentences like:
    Some actions are morally wrong.
    Thanks for the great discussion, folks!

  17. Jussi, maybe what’s confusing you about the square circle case is that you can’t imagine what steps to take toward that goal – in that sense, you can’t plan it. But that’s compatible with what Mark said: we know what would count as success in that project, namely producing something that is both circular and square.
    I’ve had the same worry as Kris and Mark for a while (though they’ve articulated it better), and nothing I’ve seen so far seems like a satisfying response. A second-order attitude view would have the wrong object – when I hope that taking a pen from the departmental secretary’s desk isn’t wrong, I do not have an attitude toward my attitudes toward pen-pinching, but, precisely, an attitude toward its wrongness. Of course I can hope that I will morally approve of pen-pinching, but that’s a different thing altogether. I can also hope that it doesn’t have features F which I morally disapprove of (as you suggested in Jon’s case), but that again is different from hoping that features F are not to be disapproved (i.e. wrong), which is the original challenge.
    If you find Mark’s scenario difficult to imagine, picture Jon as subscribing to a divine command view according to which acts are wrong iff God disapproves of them. For one reason or another, he hopes, then, that masturbation isn’t among the things that God disapproves. He may even pray to that effect.(Here a certain kind of expressivist might say: Jon hopes that the sort of moral sensibility he approves of doesn’t result in disapproval of actions with features F. Ditto for fearing etc. I’m not sure how far this tactic would work, and how expressivist in spirit it would be.)

  18. Hi, Kris:
    Horgan and Timmons have an account of existentially quantified sentences like yours in their paper in their own [2006] collection, ‘Metaethics after Moore’. I won’t try to replicate it here; my own view is that their account doesn’t explain anything, and is more than a little bit unparsimonious. Neither Blackburn nor Gibbard says anything explicitly about quantifiers, that I know of.
    I actually show how expressivists _can_ treat quantifiers in ‘Being For’, though. I just posted the revised draft on my research page. My solution works perfectly for a language with only normative terms and only the expressive power of the predicate calculus, and reasonably well, given certain assumptions, for a language with both normative and descriptive terms and the same expressive power, although it raises a problem about how complex descriptive sentences get assigned the right truth-conditions. (I offer two solutions to this problem, neither of which is perfect.)
    The treatment I give is also motivated by what I think is the only promising expressivist approach to the negation problem – in fact, it falls directly out of this approach. The basic ideas are pretty simple, but still too complicated for a blog posting.
    The approach, if I’m right that it is the only promising way for expressivists to treat negation and quantifiers, also sets constraints on how to think about an expressivist semantics for the other kinds of things we’ve been discussing: attitude attributions, counterfactual conditionals, modals, tense, and so on.

  19. Kris,
    I share Aaron’s scepticism about hopes with moral contents.
    Following his example, suppose a student, after taking a maths test, says ‘I hope Pi is irrational’, because that’s an answer she gave and she’s unsure whether it’s correct. That sounds quite odd to me. I’m inclined to say, she doesn’t really hope that Pi is irrational. Rather, she hopes she got the right answer, and she’s chosen an odd way to express this hope.
    Similarly, when you say, ‘I hope that Timmy’s torturing the kitten is morally wrong’, you’ve expressed, in an odd way, the hope that Timmy is punished by God.
    One more example. Suppose a doctor is uncertain whether her patient has condition X or condition Y. If it’s X, only administering a certain drug will save the patient’s life. If it’s Y, only performing a certain operation will save the patient’s life. She can’t both administer the drug and perform the operation. After some careful deliberation, she decides that X is more likely and administers the drug. ‘I hope I’ve done the right thing’, she says. However, I’m reluctant to say she’s thereby expressed (or reported) a hope with a moral content. She’s quite sure that morality requires that she do what she can to save the patient’s life, but she’s unsure whether she’s done that. The hope she expresses is the hope that she has done what she can to save the patient. And that’s a non-moral content.

  20. Further to my previous comment:
    Consider the case where Bob, who wants Timmy to be punished and believes God punishes those who do wrong, says ‘I hope Timmy’s torturing cats is wrong.’ Now, suppose Bob discovers, or comes to believe, that torturing cats is not wrong, but that God has punished Timmy anyway (perhaps because Timmy did something else wrong). In that case, would we expect Bob to say ‘Damn! I didn’t get what I hoped for’? No. He’s got exactly what he hoped for: Timmy’s being punished.

  21. Here are two general strategies for metaethical expressivists. One strategy is to show that it is *possible* to give a big picture semantic story that explains how we could understand moral language and thought if moral language and thought work the way expressivists say they work, and then go some way towards giving the big picture semantic story. This is Gibbard’s strategy, about which he is explicit. But this strategy leaves a promissory note to show how expressivism can go all the way in showing how moral language and thought work. This strategy thereby leaves ample room for objections, like, “What about this case? What about that case? Etc.” and puts all the burden on expressivists to provide such a big picture semantic story. Another general strategy is find persuasive examples of nonmoral language and thought that work the way expressivists think moral language and thought work. This strategy, if it can be pulled off, is more effective than the former strategy. If successful, there will be persuasive examples of nonmoral language and thought that work the way expressivists say moral language and thought work, and if so, then it looks like there *must* be a semantic story that explains how we understand the nonmoral language and thought. It would be nice if expressivists could provide the details of such a big picture semantic story, but they would carry no special burden to do so, and thus need not leave any promissory note to do so.
    In a previous thread, I gave some examples–pejorative predicates–for which a persuasive case could be made that they work sufficiently like many complex expressivists think moral terms work. So, complex expressivists might, at this point, reply to Kris: “Kris, tell me how to understand the psychological state of hoping/desiring/fearing that Carrie is a Jesus-freak, and I’ll tell you how to understand the psychological state of hoping/desiring/fearing that Timothy’s torturing of the cat is wrong.”
    Can pure/simple expressivists find some persuasive examples of nonmoral language and thought that work the way they say moral language and thought work? I’ve been having a difficult time coming up with some. The best I can do is to find some examples that *I* think work the way pure expressivists think moral language and thought work, but I’m not sure how persuasive they will be to others. Most of the examples are the cringe-producing swear words, but let me try the most benign one I can find right now. Consider the sentence ‘Helping others sucks!’–especially taken as a sentence in the language of today’s teenagers and early twenty-somethings–and the psychology of one who sincerely utters this sentence, or a more complex sentence that embeds this sentence. To me, what an expressivist would say about the semantics of ‘Helping others sucks!’ is much more plausible than any descriptivist/representational/truth-conditional semantic story about it. But ‘Helping others sucks!’ is embeddable everywhere an indicative sentence is embeddable, including within the scope of attitude attribution, complement-taking, or infinitive-taking verb. So, at this point, I can imagine some pure expressivist (legitimately) saying to Kris, “Look, tell me how you understand the psychological state of hoping/desiring/fearing that helping others sucks, and I’ll tell you how to understand the psychological state of hoping/desiring/fearing that Timothy’s torturing of cats is wrong,” and saying to Mark, “Look, tell me your semantic story about how to understand ‘John illustrated that helping others sucks’ and I’ll give you my semantic story about how to understand, ‘John illustrated that murder is wrong’.”
    Does anyone else think that ‘sucks’ is an example of a purely expressive predicate? Can anyone give more persuasive examples of predicates that work the way pure expressivists think moral predicates work? If so, then there must be a semantic story that expressivists might be able to use. If not, then I think pure expressivists are restricted to the first strategy and, in that case, are in for a haul that will last as long as it takes to provide a big picture semantic theory–which is to say that it will be a very long haul.

  22. Thanks everyone for the comments. There’s so much to think about. Here’s something towards which my intuition is leading me. I guess i might want to say that part of the meaning of the terms is inferential connections they are in. Understanding some of these connections is part of grasping the meaning of the term. So, understanding the ‘term’ square in part is contituted of knowing that it is not a circle. Therefore, if you grasp the term, you know that when you would start planning to draw a square you are starting to plan not drawing a circle. Thus, you can’t start to draw a square circle and neither can anyone else who understands the terms.
    This is the difference to trisected angle case. Understanding trisection and angle, grasping those concepts, does not consist of knowing that you cannot construct a trisection of an angle by a straightedge and a compass.
    The next step would be to argue that mastering moral terms involves understaning that acts cannot be made to be right or wrong, good or bad, but rather they are right or wrong independent of our actions. This would explain why planning to make something wrong does not make sense. This is just a rough thought though.
    The hope case is the most interesting because it does sound so natural. Antti, when you give the divine command case that does sound like the second account you reject. That is, in that case I hope that the act falls under the extention of ‘wrong acts’ that is co-extensive with what is forbidden by the God. I think that that is still a good way to account for many uses of the relevant claims.
    But, I think that also more can be said for the first account. In the quasi-realist picture, our disapprovals are supposed to be projected to the world. We see the world coloured by the moral properties that are unbeknownst to us our projected attitudes. If that were right, then it might be somewhat plausible to say that you would hope that x had some moral property, the same moral colour, than some other actions appear to have for you. In effect this is saying that you hope that your attitudes would colour the world in a different way but you don’t know this. In the moralising perspective you are not supposed to be aware of the projection in any case.
    the case you give is an interesting one and of course paradoxical. I do think that the claim ‘I have done something wrong’ expresses a belief that some of the actions I’ve done have features which I disapprove of or maybe even disapproval of unspecified action of my own (for the sake of retaining the moral equivalence with others). There is something odd about the latter choice even if it would fit the idea that I don’t find that any of my specific actions had the general features I disapprove of. The first account also seems plausible. In the case I go through all my specific actions, I would of course have to conclude that my belief is false or that I’m fine with accepting a contradiction.

  23. Hi, Dan.
    I think there are limited versions of your kind of strategy that can be brought to bear by pure expressivists. Not that commands are uuncontroversially expressive of desire-like attitudes, but Hare and Smart both focused on complex commands. ‘Don’t shut the door’ conflicts with ‘shut the door’, and ‘shut the door and open the window’ conflicts with both ‘don’t shut the door’ and ‘don’t open the window’, just as ordinary conjunctions are contradicted by what contradicts each conjunct. Taking up Kris’s question about quantifiers, we can go further: ‘shut every door’ conflicts with ‘don’t shut this door’, for any choice of this door. So that seems to work like we would expect a quantifier to work. These kinds of things made Hare and Smart optimistic that explaining the logical properties of complex commands was everyone’s problem, just as for you, pejoratives are everyone’s problem. You might think, however, that their grounds for optimism don’t extend to embedding in, say, the past tense. So that strategy might only get them so far.

  24. Jussi,
    I guess i might want to say that part of the meaning of the terms is inferential connections they are in. Understanding some of these connections is part of grasping the meaning of the term. So, understanding the ‘term’ square in part is contituted of knowing that it is not a circle. Therefore, if you grasp the term, you know that when you would start planning to draw a square you are starting to plan not drawing a circle. Thus, you can’t start to draw a square circle and neither can anyone else who understands the terms.
    To make views like this work you are presumably going to have to equate knowledge of the terms with lots of different inferential connections. And then there are going to be some people who fail to make all of them. Presumably you don’t want to say that failure to reason in accord with each and every one is sufficient for not having thoughts with the relevant content.
    Now maybe you want to reply that some are more central than others and even so central as to be essential. And you might think treating circles as not square is central. OK. But it isn’t obvious to me that I can’t make some action wrong (say by making an action have bad results as someone has already suggested). But I’ll bet that much of the rest of the inferences I make about wrongness will still be in order by your lights. Can I have thoughts whose contents are well-characterized as involving wrongness?

  25. Mark,
    that’s a good question. Obviously you can make some action wrong by making it the sort of action that belongs to the extension of ‘wrong acts’. So, I can make moving my arm wrong by moving my arm near someone’s face so that I’m hitting that person. But, in the sense here we are fiddling with the action and not so much with wrongness. Take then definitive action of which we know its consequences, motives, and so on. Say that it is the particular action of cutting the grass that I did and which didn’t have any bad consequences. Could that act be made wrong? What would that be like? For me it sounds like painting the moral wrongness of an act blue. Anyone competent with wrongness would know that it is a property that cannot be painted. It might be that a certain act with all its consequences and motives are not in similar way made to be wrong either. I wouldn’t be surprised that this is part of the competence with the term.
    Several people have suggested an example where wrongness is equated by what is forbidden, say, by the community. In this case you could make some particular act wrong by getting the community to forbid that action. But, in an important sense you are still not making the act wrong. You are just making the act forbidden by the community – which is wrong no matter what you do on this view.

  26. Hi, Mark. I had forgotten about those examples. Yes, they do seem to constitute a kind of ‘it’s everyone’s problem’-type reply to objectors.
    Kris, I forgot to mention earlier that hopes and fears seem to be the paradigm of (logically) complex attitudes consisting of a belief and a desire broadly construed. For example, hoping that my parents come for Thanksgiving seems to be a complex state consisting of a belief that my parents might come for Thanksgiving and a desire that they do so; fearing that my parents might come for Thanksgiving seems to be a complex state consisting of a belief that my parents might come for Thanksgiving and a desire that they not do so.

  27. Let me try one more variation; maybe thinking in epistemic and temporal terms would help make the issue clearer.
    Case 1: I know that wrong actions have properties F. Then, I hope that my action doesn’t have properties F. This is the sort of case of “hoping that my action isn’t wrong” that Jussi seems to be talking about, and it’s not the sort of case that is meant to cause problems for expressivism.
    Case 2: I know my action has properties F’. Then, I hope that actions that have properties F’ are not wrong. This is the sort of case that prima facie raises a problem for expressivists.

  28. Antti,
    that’s helpful. I started to have something like that in the mind. The more I think the more ‘prima facie’ the problem must be though. The argument cannot start from the premise that there are moral properties in the world and we can hope that they would be on some other natural properties. That would beg the question. So, the expressivist must be allowed to the premise that the moral properties we think there are in the world are projections of our pro and con attitudes. This, however, is not supposed to be transparent for us in the internal perspective (which is odd to say but let it pass).
    If that’s true, then it makes sense to hope that acts with some other natural properties were the ones with the moral properties. Inter alia, this means that the hope is that our attitudes would be projecting the moral properties on some other natural features of the world. But, this external claim is supposed to be opaque to us in the moralising perspective in which we are hoping things to be different morally speaking. Again, we might want to take an issue with this but it’s a different objection.
    So, I’m starting to think that the main problem for the expressivist will be the more usual problem of ‘psychological chauvinism’. It’s not that the expressivist won’t have a story to tell what is going on but rather that the story is one which attributes attitudes to us that we have hard time identifying with.

Comments are closed.