No papers blog today because of an administrative snafu. (One of many administrative snafus around TAR headquarters it turns out.)

The spam attack only fully stopped when I found out how to IP block the addresses being used. Thanks to Will in comments for suggesting this. No feedback from, who are hosting the spammers.

When things settle down here, I mean to write something substantive about the paper on moral relativism Jesse Prinz did here on Tuesday. So for now I’ll say two insubstantive things. First, it’s a little disturbing how many hidden indexicals there are meant to be in language these days. (Jesse is really a moral indexicalist, like Jamie Dreier, not a full-blown relativist.) Second, it’s amazing how much work the open question argument (and variants on it) has been thought to do over the last 100 years. It’s like Bow down before me, for I am the Open Question Argument of Doom, and I can be used to derive Every anti-realist conclusion Ever Conceived. That wasn’t quite how Jesse put it, indeed he didn’t describe his argument as a variant on the open question argument, though don’t know if he’d disagree with the claim that he was appealing to (inter alia) something lke the open question argument in his argument for relativism. Six months ago I thought of starting a serious study of Moore in order to have a better grip on what was going on with all these arguments, but it never really happened.

As I said, actual serious substantive comments to follow, because it was a very interesting paper and there’s lots to say.

20 Replies to “Stuff”

  1. Brian: Would you mind explaining the distinction between the moral indexicalist and the full-blown moral relativist.

  2. It’s a completely tendentious use of words on my part. But here’s what I mean. The indexicalist thinks that which property is expressed by a use of ‘good’ depends on the speaker’s context/culture/values/emotions. The relativist thinks that the truth value of a sentence containing ‘good’ depends on the evaluators context/culture/values/emotions, whoever the utterer is. So when someone from a pro killing Americans culture says “Killing Americans is good” the indexicalist thinks he expressed a true proposition, while the relativist thinks he expressed a proposition that is false (i.e. false in our culture) even though it is true in his culture. (And all the worse for his culture, the kind of relativist I’m imagining could go on to say.)

    The coherence of the relativist position is a long-running issue (to say the least) but I think it’s closer to what people meant by ‘moral relativism’.

  3. I should say that I’m not a moral relativist, but I do have opinions on the relative plausibility of different relativist views, and I think even us anti-relativists should be interested in finding the most plausible relativist view to get a clear sense on what is at issue.

  4. One thing that seems odd about a hidden indexical view like Jesse’s is that there is not a strong conceptual connection between ‘wrong’ and the character that Jesse assigns to ‘wrong’. Jesse’s view, if I remember correctly, is that the character of ‘wrong’ is something like “that which causes disapprobation in me”. But, it seems that I can conceive of worlds where I have no emotional responses to an act yet I think the act is wrong. I am like Spock in this world. There is just a failure of duty, so it is wrong, but I do not have disapprobation towards that act. It does little good, I think, to argue that social scientists have data to suggest that in order for us to grasp the concept of wrong we must have the right kind of emotional responses developmently. Character reveals a necessary conceptual connection between an expression and its content, not just a conceptual connection at the actual world. But if anything seems like a conceptual necessity for ‘wrong’ it is that acts can be wrong whether or not I think (where ‘think’ includes having emotional responses) they are.

  5. Brian,

    Does your claim: “The relativist thinks that the truth value of a sentence containing ‘good’ depends on the evaluators context/culture/values/emotions, whoever the utterer is” imply that the truth-value of the utterance actually changes depending on who evaluates it?

    If so, that seems like a nonstandard version of relativism to me, and it seems to have the (perhaps) unwanted consequence that a single utterance is both true and false. Now, if the point is rather that in different ethical contexts one evaluator might think that the utterance is true while another might think that it is false, that seems not to be a version of relativism but the view that people disagree about ethical matters in a systematic way.

    I would have thought that relativists offer an account of the truth-conditions of a moral sentence-token, according to which they are implicitly relative to the context/culture/values/emotions of the person who utters it rather than the person who evaluates it. Isn’t that roughly the common thread of relativism shared by Westermark, Harman, Wong, Dreier, etc.?

  6. Hi Brian,

    Thanks for the post. A few remarks…

    (1) Relativism vs. Indexicalism? With Matthew Chrisman, I am strongly inclined to call indexicalism a species of relativism. If X is relative to Y then claims about things in the X domain cannot be asigned truth values without reference to something in the Y domain. This could amoount to the view that a single claim about an X can be both true and false. That approach courts incoherence. The alternative is to say that a claim in the X domain is incomplete (is not a candidate for a truth value) without filling in a context variable. X predicates need to be both saturated and grounded to express propositions, where grounding refers to the contextual filling in. The claim that X predicates need to be grounded amounts to a kind of indexicalism. With Matthew, I think this is what many relativists have had in mind. It may also be what physicists have in mind when, for example, they say that momentum claims must be evaluated relative to the reference frame of an observer.

    (2) Moore. Open question arguments generally go like this:

    P1. Assume, F = G
    P2. I can imagine and a that is F and not G
    P3. If P2, then: F not= G
    C. F not= G

    Open question arguments are designed to prove negative claims, and, in most domains they fail, because in most domains P3 is false. The argument I advanced in the talk is somewhat Moorean, but it was explicitly designed to overcome this worry. I said that moral properties are neither Millian properties nor Kantian properties. To arrive at this conclusion, I first (with the aid of the Moral Mary’s Room argument) showed that moral concepts are not Kantian or Millian (e.g., we do not think of the good as that which maximized utility). If I concluded from this alone that the Good is not that which maximizes utility, then my argument would be Moorean. But Moral Mary was designed to show something about moral concepts not about moral ontology. My claims about moral ontology required a further argument. I claimed that, if moral concepts refer to Kantian or Millian properties, they must do so by bearing a content conferring causal relation to those properties (because they cannot refer to those properties via description). I seriously doubt that any of the content conferring causal relations advanced by contemporary semanticists will do the trick. (a) GOOD probably doesn’t trace back to a baptism in which a Kantian or Millian property was instantiated (and if it did, there would be a serious qua problem); (b) tokens of GOOD are not reliably caused by Kantian or Millian properies; (c) it’s unlikely that GOOD was selected for this purpose (it wasn’t naturally selected for anything, and it’s cultural teleofunction probably has more to do with preserving extant social norms). If moral concepts do not refer in one of these ways, then the refer in a way that is more heavily determined by conceptually constituent features. Since moral concepts have emotions as conceptually constituent features, I suggested that moral concepts refer to the property of inducing those emotions (under idealized conditions). Going back to Moore, P3 is usually false, but it’s plausible in those cases where meaning is heavily constrained by our concepts. Such heavy constraints arise when causal relations are unavailable.

    (3) Nick on Spock’s morality. I don’t think Spock has (non-deferential) moral concepts. How then, can we imagine that something would be wrong even if everyone became indifferent? I think the answer is that we can refer rigidly to wrongness:

    “Killing would be wrong if no one was bothered by it” = “Killing would has the propery of [causing disappobation in me as I am currently constituted] in worlds where neither I nor anyone else am bothered by killing.”

    Indeed, in the accessible worlds in which I was never born, killing would have the prorty of causing disapprobabtion in me here in this world.

    Now suppose we undermine this reading by making the hidden indexical explicit:

    “Killing is (morally) wrong to those who do not have emotional disappropobation towards it.”

    No intuitions strongly support this claim. To insist that Spock finds killing morally wrong begs the question. I think a world of Spocks is a world of cool rational rules with no need for morality. Spock could reason that a certain course of action would lead to contradictions when universalized and still wonder whether that is the property we humans are referring to when we use the word “wrong.” He uncertainty would be well placed, because we almost certainly aren’t referring to properties of that kind.

    There’s so much more to say, but I have to sign off for now.

  7. Jesse,

    The clarification of what Moral Mary is meant to be doing is helpful. I suppose I still don’t think that she shows analytic consequentialism is false – she might just be missing something when she reads the books, or she might conclude Mill s right – but I’m independently sympathetic to that claim. So while I probably disagree with (b) or something like it, I don’t really have much to say about that part. (But maybe on the weekend I’ll say more.)

    I agree too that indexicalism is a form of relativism. I just don’t think it’s the form that most naturally captures relativist intuitions. The kind of relativism I was describing has the result that a single proposition can be true relative to some contexts and false relative to others. It can’t be true and false simpliciter, because that would mean it’s both true and false in our context, which is incoherent.

    In general I think these claims of incoherence go by pretty quick, and it’s far from obvious they work. Everyone thinks that propositions can be true in one world, false in another. Many people think propositions can be true at one time, false at another. If both of those positions are coherent, then the view that propositions take their truth values relative to, say, centred worlds looks prima facie coherent. It might be false – all the proposed cases of relativism are controversial – but it looks at least coherent.

  8. Brian,

    Second, it’s amazing how much work the open question argument (and variants on it) has been thought to do over the last 100 years. It’s like Bow down before me, for I am the Open Question Argument of Doom, and I can be used to derive Every anti-realist conclusion Ever Conceived.

    Has the OQA (and variants) really been used to derive anti-realist conclusions? I can’t think of any examples.

    Here’s what I think. I think the OQA (and kin) can be used to eliminate analytic descriptivist theories. One way to argue for anti-realism is to eliminate realist contenders, so some version of the OQA would be an element of an anti-realist argument.

    If that’s right, then it isn’t amazing that the OQA has been thought to do that work. How would you try to eliminate an analytic reductionist view other than by setting out some version of an OQA? In fact, doesn’t any attempt to refute an analytic reductionist view thereby count as ‘some version’ of an OQA?

    What Jesse said seems mainly right to me — let me just say what I don’t agree with. Well, I don’t see what the first premise, P1, is doing in the argument, but presumably that’s a typo due to a midstream switch in argument type. But mainly, I don’t think thin moral predicates have any descriptive sense to begin with, so there is no need to rigidify any description to avoid the difficulties that might be presented by some counterfactuals. Instead, moral predicates refer directly to their properties. Moral attitudes and emotions show up in the character, not in the sense.
    (The truth is, I don’t know how seriously I believe this stuff now, but I’m willing to defend it for a while and I do have strong opinions about which sort of relativism is best.)

  9. Has the OQA (and variants) really been used to derive anti-realist conclusions? I can’t think of any examples.
    According to the lore that they teach me, it was used to launch Ayer’s noncognitivism in the 1930s. When people accepted that ‘good’ couldn’t refer to any natural property, and they stopped believing in non-natural properties, they decided that there wasn’t any real property corresponding to the term, and noncognitivism became a promising option. (I’ve always been surprised that people turned to noncognitivism before error theory, though — error theory seems to me the more obvious antirealist option.)

  10. Brian,

    I agree with the basic point that arguments for incoherence are overstated. Moral sentences might express propostions like this: The person at the center of the world regards phi-ing with disapprobation. This would allow that P and ~P (where P is a moral claim) can both be true when uttered in the same world about the same event. But notice that P and ~P still can’t both be true in the same world of evaluation. Because they are essentially indexical, they can be evaluated only in centered worlds. The only cases where both are true in the same context of evaluation are cases where a speaker has an incoherent morality.


    Greetings, and thanks for chiming in. You are right about P1. I was going to set the argument up as a reductio, but it’s not really needed. On descriptive content, I would desagree however. By descriptive content, I just mean that there are features encoded in our moral concepts that place necessary conditions on the application of those concepts (as opposed to merely serving as reference fixers). I think the features in question are emotions or sentiments. Roughly, something must be disposed to cause negative sentiments if it is morally right or wrong. Moral emotions also place further constraints on moral concepts and moral content by introducing appropriateness conditions. Each emotion has an acceptable range of application (e.g., moral anger is restricted to the range of harms).

  11. Thanks for the comments Jesse. You write that no intuitions support the following claim:

    “Killing is (morally) wrong to those who do not have emotional disappropobation towards it.”

    Say that I am a Hannibal Lecter. Consequently, I have no emotional dispprobation towards killing. However, it seems that I can believe truly that killing is wrong. I think that no intuitions support the following claim. “I know that killing is wrong, yet I have no emotional responses to killing”. However, I do think intuitions support the following claim “I believe truly that killing is wrong even though I have no emotional responses to killing”.
    Likewise, with Moral Mary (if I am remembering the example correctly). She can believe truly that x is wrong, even though she cannot know that x is wrong. But, this just shows that x has the property of wrongness without Mary’s disapprobation. Or does it?

  12. hiNick,

    Thanks for pursuing the point. I think Lecter uses an inverted comma sense of “wrong.” He knows that others find it wrong. Suppose there are no such others. Suppose there is a society of emotionally deficient psychopaths and they have had no contact with any moralizers. Can they have thoughts about what counts are right and wrong? I don’t think there are clear intuitions on this.

  13. I should just add, as part of my claim that there was lots interesting in Jesse’s talk, that I thought the arguments for the inverted commas interpretation of Lecter’s moral language were very good, and moved me from hoping that was the right interpretation to believing it was.

  14. Jesse,

    By descriptive content, I just mean that there are features encoded in our moral concepts that place necessary conditions on the application of those concepts (as opposed to merely serving as reference fixers).

    Ok. But those could be in the character, surely. (Cf.: ‘yesterday’ has encoded in its concept the feature that it denotes the day before the context of use, but that feature is not a part of its content.)

  15. Hi. I wanted to chime in on the question of what kinds of accounts of the meaning of moral concepts are available given that the OQA undermines specifying the meaning by description.

    Above Jesse writes, “I claimed that, if moral concepts refer to Kantian or Millian properties, they must do so by bearing a content conferring causal relation to those properties (because they cannot refer to those properties via description). I seriously doubt that any of the content conferring causal relations advanced by contemporary semanticists will do the trick.”

    And I find plausible the suggestion that the standard causal accounts fail to link GOOD to Kantian or Millian properties. However, isn’t there a third option in contemporary semantics between (i) accounts that give meanings by description and (ii) accounts that give meanings by some causal story—namely, (iii) accounts that give meanings by some sort of communally instituted inferential role?

    Now, I don’t know how exactly an inferential-role account of GOOD would look. (It’d probably be more plausible to claim that Kant is giving an inferential role account of DUTY rather than GOOD in the first two books of the GW.) In any case, however, Jesse seemed to be running an argument by elimination against all accounts of the meaning of ethical concepts that don’t refer to emotional responses, yet in my view he didn’t eliminate one of the contenders.

  16. Matthew,

    Good point. I must admit that there is something very appealing about the idea that moral concepts get fixed by social inferential roles. If the roles were social in a truly distributed sense (no moral experts), then something of the unversalist perspective might emerge. Each person would have a stake in being regarded as an object of moral concern; each person would say its bad to harm me; and, thus, the GOOD would encompass rights for all members of the linguisitc community. (Imagine a marriage of Brandom and Gewirth.)

    Still, I have some doubts. First, for moral concepts to have a social semantics, users of moral terms would have to defer. I’m not sure how much deference there is. We don’t tend to defer when someone tells us we are incorrect to call something good or bad. Second, I’m not sure that social inferential roles would lead us to the view that moral concepts designate Kantian and Millian properties. It’s fair to say that the majority of scoieties show global disregard for the well-being of other groups, even when they recognize others as persons (see, e.g., Reed’s 1955 moral ethnography of the Gahuku Gama). No society operates under a purely Kantian or Millian morality. Relatedly, Gil Harman gives reason to think that social semantic would lead away from objectivity. Harman is both a wide functional role semanticist and a moral relativist. If we defer to others in deciding the extenion of “good,” we will end up arriving at ethics by negotatiations. Other society values will exert an influence, and those with greater power will be in a better position to dictate moral rules.

  17. Brian and Jesse,

    I’m not sure I see how the inverted comma stuff works. Suppose Lecter wasn’t always so bad: as a lad he thought that killing was wrong and he had emotional responses to it. Surely he could say the following:

    (1) I used to think that killing was wrong and I even had emotional responses to killing. I still think as I used to, only now I don’t respond emotionally.

    How does this statement work on the inverted comma view? If he’s using inverted commas now, can he report truly about what he used to think when he wasn’t using inverted commas? If so, then he should be able to use the eliptical “think as I used to” to talk about how he still thinks killing is wrong (without inverted commas). If he can’t talk about what he used to think, then not only is this a pretty weird result, but then we might wonder how inverted-comma people report about others who aren’t using such talk. So, for example, if Mary is a pretty morally respectable person, it seems Lecter could say (2):

    (2) Mary thinks killing is wrong and so do I.

    If he’s using inverted commas with “wrong,” his report about Mary if false; if he’s not using inverted commas, then he can use the “so do I” to attribute “wrong” (non-comma invertedly) to himself.

  18. Hi Meg,

    Thanks for your post and welcome on board. I think moral curruption is sort of like losing your sight. A person, call him Smedley, who has gone blind might say:

    “I used to think that cherries were red, and I still do”

    But now suppose we ask Smedly if he can recall what red is like. He won’t be able to. And ask if he could tell you whether the object presented before him is red. He’d have to ask. Now, finally, ask him what redness consists in, or what he means when he says of something that it’s red. He might tell you red is a color. He might say that strawberries and roses are red, and that red and yellow paint mix to make orange paint. But, besides recalling these facts about the extension of “red” and the behavior of “red,” he’d be forced to admit that he doesn’t really recall what red is, as such. There has been a major discontinuity in his thought. In addition to being blind to red, however, Smedly is somewhat blind to his deficit. He cannot say, thinking of red used to be like THIS (a phenomenal demonstrative) and now feels different. He cannot form the phenomenal demonstrative and make the comparison. This may make it hard for him to reflect on the change. He might even fail to realize that he used to think about red in a different way.

    Likewise for Lecte. He is doubly blind. He cannot recall what wrongness is, as such, and he is blind to his own disability. He can defer to his childhood self and to upright Mary, but he couldn’t, for example, determine whether an entirely novel case was good or evil, unless it was sufficiently close to familiar exemplars. Is it morally worse to stomp a person’s head in or to kill them with poison? My guess is that upright Mary would say the former and mature Lecter would say they are equally “evil.” And his inability to discern is precisely what warrant the inverted commas. His morality is cooly, and unprojectably nestled between the strokes.


  19. This move sounds fishy as a semantic move. If I go to China and experience it first hand but then forget everything I saw, it doesn’t seem to follow that I only remember going to ‘China’ as opposed to China. Perhaps its useful to make such distinctions when talking about modes of presentation under which a I can refer/think about China but it seems fairly hopeless to say that I lose the capacity to think about China as a I forget stuff I learnt. (Replace ‘china’ with a predicate that China falls under if you don’t like examples involving names).

    Are tricks using quotation marks intended to mark a real semantic difference, or just to mark a difference between typical modes of presentation and others? The answer to this question should bear on the effectiveness of Meg’s argument.

Comments are closed.