Moral Fictionalism

83 Replies to “Moral Fictionalism”

1. Josh,
   I’m not sure I buy MF either. Richard Boyd also defends it. But I don’t feel I understand the force of your criticism. It sounds right to say that first-order moral judgments are binary in the way you propose. So why couldn’t MF handle them by proposing that for any such judgment there is
   (1) a factual truth condition (that, say, executing murderers imposes extreme psychological hardship on them), and
   (2) a moral truth condition (that imposing extreme psychological hardship on a person is/is not wrong),
   and that for each such judgment, component (2) is false, since there is no moral fact to which the claim can refer?
   I suppose that this move does render component (1), the factual component, “irrelevant in determining the truth-value of moral claims,” as you say. But I gather that MF’s proponents would say that this shouldn’t bother us anymore than it should bother us when, in asserting (P&Q), our assertion is false whenever Q is. Q is just always false, according to MF.
   My uneasiness with MF has more to do with familiar worries about having-and-eating-cake strategies generally, including quasi-realism. In particular, the example
   (K) “Kryptonite is poisonous to Superman”
   is suggestive. This is true in the superhero narrative, but if it were false in that narrative, we would continue to think of that narrative as the Superman narrative. In contrast, isn’t part of moral discourse that moral claims are necessary truths, and if one thought of them as contingent, one would be talking about some other domain besides morality? Does MF compel us to think of moral claims in this way? If so, then we cannot think of morality as “fictional” in this sense, and MF would require us to adopt a certain double consciousness toward morality, attributing modal properties to statements that in fact lack them.

2. Josh,
   I agree that when it comes to truth and falsity, MF implies that all (positive) moral judgments are in the same boat — they’re all false. But MFers still have a way to separate the “acceptable” moral judgments from the “unacceptable” ones. The acceptable ones are the ones that are true according to the fiction of morality (which, I suppose, consists of the claims of “common sense morality” or whatever). So their theory can predict and explain why we who labor under the illusion of moral realism are inclined to think ‘torturing for fun is wrong’ is true while ‘selfishness is a virtue’ is false.
   Do you think this is good enough?
   It’s analogous to how we would think of people who labor under the illusion of some false religion. All their religiously-loaded claims are false. But they’re not all in the same boat with respect to a different standard — the standard of truth according to the fiction of their religion.
   Or am I missing your objection?
   (Don’t read ‘MFers’ the wrong way.)

3. Chris and Michael,
   Thanks for the comments. And, no, I would never read “MFers” the wrong way–I’d hope PEA Soup is a place where one can be proud to be an MFer.
   Chris, right, MF can distinguish between ‘true’ and ‘false’ moral claims within the fiction (just like we can say that it’s false that Spiderman is allergic to Kryptonite). But all moral claims (as you say) are for MF false in the literal sense. And I don’t think that’s enough. They need to account for truth-aptness in the literal sense, since literal-truth-apt claims are what ordinary discourse purports to make. At least, this is the task set in the article by Nolan, Restall, and West (their answer being that MF does recognize literal-truth-aptness, since all moral claims are literally false [if some are ‘true’ within the fiction]).
   Michael, I think I may have stated the worry a bit ambiguously. I’m only focused on your (2), not your (1). To my mind, whether imposing extreme psychological hardship on a person is wrong is to be determined (in ordinary discourse) by seeing whether imposing such hardship has wrongness, as opposed to permissiveness, attached to it. Certainly, if there is no wrongness, (2)-type claims will all be false. But then (I’m suggesting) we can’t make sense of the truth of “imposing hardship is wrong” in terms of whether imposing hardship has wrongness attached to it. Indeed, hardship-imposition will be totally irrelevant to determining the truth value of “imposing hardship is wrong.” But (I’m suggesting) the nature of imposing hardship is what determines the truth that it is wrong as opposed to permissible.
   I’m not sure what to think about whether MF causes us to revise the putative necessity of moral ‘truths’, but I think the MFer might say that moral ‘truths’ are necessary within the fiction. But I don’t know if that move works.

4. Josh,
   Is this your argument against MF?:
   1. If MF is true, then all moral judgments are literally false.
   2. But obviously some are literally true.
   3. So MF is not true.
   Although I personally accept 2, it sure seems question-begging against the MFer. Moreover, they would say they have pretty good reasons to reject 2 (all the reasons they see to reject naturalist realist theories, non-naturalist realist theories, and quasi-realism [if quasi-realism entails 2]).
   But maybe I’m not getting your objection.

5. Chris,
   That’s more ambitious than what I want to argue for (and I agree that (2) would be question-begging). I now see that in my initial reply to you and Michael, I occasionally used stronger language than I should have, language which might suggest your reconstructed argument. E.g., I said “the nature of imposing hardship is what determines the truth that it is wrong as opposed to permissible”, whereas I should have said “according to ordinary moral discourse, the nature of imposing hardship is what determines the truth that it is wrong as opposed to permissible.” What I want to claim is not that MF is not true, nor even that MF is weaker overall than other forms of non-realism. All I want to suggest is that, contra Nolan, Restall, and West, MF does not seem more advantageous than emotivism or quasi-realism with respect to (specifically) capturing the truth-aptness suggested in ordinary moral discourse.
   So I guess my argument could be put as follows:

   4. If MF is superior to other forms of non-realism with respect to accounting for the truth-aptness (supposedly) suggested by ordinary moral claims, then MF can adequately account for the truth-aptness (supposedly) suggested by ordinary moral terms.
   5. MF cannot adequately account for the truth-aptness (supposedly) suggested by ordinary moral terms.
   6. MF is not superior to other forms of non-realism with respect to accounting for the truth-aptness (supposedly) suggested by ordinary moral claims.

   So it’s a pretty modest conclusion. That said, depending on how one evaluates the strengths and weaknesses of other forms of non-realism, I guess their parity with respect to truth-aptness might knock MF out of contention.

6. I’m actually quite puzzled about the way Nolan, Restall, and West phrase MF, and suspect that some of the potential problems derive from their error-theoretic version of MF. There is a big contrast between the NRW MF (I love this) and the version of MF you get in Mark Eli Kalderon’s excellent recent book Moral Fictionalism (OUP 2005). I think Kalderon comes much closer to a view that takes the ‘fictionalism’ seriously in moral fictionalism and where fictions are understood in a way that is more truthful to our understanding of fiction in say art.
   The biggest difference there is between these two versions of MF is that for Kalderon Fictionalism is a non-cognitivist, non-descriptivist, and non-assertivist view. So for him moral thoughts and utterances have sui generis moral content, i.e., there are moral propositions. This corresponds to the way unicorn thoughts have sui generis content involving really unicorns. This of course is something expressivists would have to deny. But, then he denies that in moral thought we *believe* these contents or *assert* them while using the utterances. This corresponds to the way I do not believe that Sherlock Holmes was a friend of Watson’s or present that as true when I say it. So, in moral thought and discourse we are not committed to a systematic error like error theorists would have it. But, we could think that these propositions are in many ways good and useful to entertain, and so we do.
   Of course you can attach something like a deflationist account of truth on the top of the Kalderon type MF. I think that on that account the nature of torture for instance would have a role to play on whether it is true that torture is wrong.

7. Jussi, thanks for that. I haven’t read Kalderon’s book yet, but between that book and the new Kalderon-edited book Fictionalism in Metaphysics, which apparently includes a paper by Lewis arguing that quasi-realism is best understood as a kind of fictionalism, it seems like there is a lot of new material out there for anyone toying with the idea of becoming a MFer. [Yes, I’m enjoying this too…I’m tempted to adopt the position just so I can call myself a professional MFer].
   To be fair to NRW, they distinguish between descriptive and prescriptive MF. I was working with what I think they would consider prescriptive MF, while maybe on their map Kalderon would be a descriptive MF (though I can’t be sure, because I find their construal of the distinction slightly less than totally transparent).
   So the descriptive MFer can say that we already do think of moral claims as claims made within the fiction of morality. That is, ordinary discourse already treats all moral claims as false (or as non-assertive), because we already realize that those claims presuppose moral properties that are mythical. This would entail that the nature of ordinary moral claims’ truth conditions is not binary, as I have suggested. I suppose at this point, I just have a clashing intuition: it sure doesn’t seem to me as though ordinary people ordinarily think of moral claims as fictional, or even that they would were they to undergo some philosophical reflection that could (in principle) prompt them to revise their explicitly held beliefs. Maybe this would be a good one for examination by experimental philosophy. Or maybe my intuition will change when I read Kalderon.

8. Josh,
   A good point but you need to be careful here. I think that there is a way in which moral discourse and thought can be fictional (as seen from a metaethical point of view) without the participants themselves thinking or realising that what they are concern with is a myth or a fiction that strictly speaking is false. Think of the emotional involment with fiction in art. I pretty much cry everytime I see ‘Dancer in the Dark’. I would not if I thought it was all just a film and actors going through the motions on the screen for the sake of it (I wouldn’t think this even though someone philosophically pointed it out to me that that’s all it really is). So, at some level I have to think that what is going on is real and happening to people. But still I would deny that I’m entertaining beliefs about actual people that are false. I think there is something to this picture even though I do have too many cognitivist symphaties when it comes to morality.

9. It sounds like MF is an old idea with a new name. From Josh’s post, it sounds like the difference between MF and Error Theory (ET) is supposed to be that MF says moral discourse is useful and we should keep it. But ETers don’t deny that. Mackie doesn’t say that we should stop using moral discourse, nor does Richard Joyce. And if Mackie’s not an ETer, I don’t know who is.
   I thought MF was going to be different from ET by allowing some moral claims to be true. Consider two views one might have of the statement
   (K) Kryptonite hurts Superman.
   (1) K is true, roughly because the author of the Superman story consistently represents K as being the case in the story.
   An MFer might then say that positive moral claims are often true, and what makes them true is that the authors of the moral story (presumably, this means society) consistently represent them as being the case. Then you have the problems of cultural relativism. Whereas it is true that
   (K1) If the authors of the Superman story consistently represent Superman as being allergic to Kryptonite, then Superman is allergic to Kryptonite.
   it is not true that
   (M1) If society consistently represents meat-eating as permissible, then it is permissible to eat meat.
   (2) K is false, because neither Kryptonite nor Superman exist.
   An MFer would then have to say that all positive moral claims are false. This just has the obvious problems of ET. Notably,
   (P1) It is better to be lying on the beach in Hawaii sipping a Mai Tai, than to be stretched on the rack in a medieval torture chamber.
   seems more plausible (to say the least) than any of the following premises that might be used to support ET:
   (P2) All knowledge is empirical.
   (P3) All properties are natural.
   (P4) No property can be intrinsically motivating.
   (P5) Moral properties exist only if they are intrinsically motivating.
   By the way, what does MFer stand for–Moral Fictionalister? Moral Fictioner?
   Also by the way, what’s the status of the claims “Moral discourse is useful” and “We should continue to use moral discourse”? If all positive moral claims are false, I would think that the same arguments that show that would also show that all claims of the form “x is useful” and “We should X” are false.

10. Allow me to modify and then second Huemer’s last point. I would have said that MFers can allow that there are facts about what is useful. They should be able to say, I would have thought, that if doing X leads to Y, then X is useful for bringing about Y. So MFers could say that continuing to use moral language roughly as we have is useful to making the world contain more happiness. But then it seems to me the issue is how to get from any such premise to the conclusion that we ought to do anything. If the ought is a moral ought the problem is right there on its face. But if the ought is some other kind of ought (pragmatic, instrumentalist, etc.) then it is not clear how the premise combines with pragmatic or instrumentalist norms to yeild the conclusion.

11. The Nolan, Restall, West version of moral fictionalism is also an error theory — they acknowledge it. And I think Richard Joyce endorses a lot of what they say in the paper, for that matter — he thinks pretending is a very good idea. On the other hand, as Jussi points out, Mark Kalderon’s version does not include an error theory: falsehood doesn’t entail error unless the speech act aims at truth, to put it telegraphically.
    I know that Joyce thinks that moral value and moral oughts are special. Pragmatic oughts and instrumental goods can be straightforwardly true (not just true-in-the-fiction). Like Mackie, he thinks that ordinary moral practices are good and ought to be continued, in those non-moral senses of ‘good’ and ‘ought’. I can’t remember what Nolan and Restall and West think about that, but I presume it’s similar.

12. David says:

    
    If the ought is a moral ought the problem is right there on its face.
    

    I’d like to hear more about this. Couldn’t it be true in the moral fiction that it’s good for society to perpetuate a moral fiction? (Reminds me of the movie Scream 2, in which some characters in a film class discuss the merits of sequels. Or here’s another example: suppose it were part of the Santa fiction that children who don’t believe in Santa don’t get any presents.) If defenders of MF claim that we morally ought to perpetuate a moral fiction, they are, by their own lights, speaking falsely. But that is not to concede that MF is false. (As I’m thinking of it, the claim that we ought to keep up the pretense is not itself part of MF; it’s an optional extra. MF is a metaethical view.)

13. I once thought Fictionalism appealing. In addition to reasons like those discussed by Mike, I was (am) inclined to think MF has other problems.
    First, suppose the fictional operator selects something like a folk moral theory, some theory the folk represent. Then the true moral claims will likely include “homosexuality is wrong”. That should come out false, though. So, we need to idealise somewhat. The fictional operator should be “According to ideal folk moral theory…” But then, it seems we need a normative concept to distiguish ideal from non-ideal folk theories, and there goes the project.
    The phenemenology gets messed up too. I know what it is like to engage in fiction, to pretend that someone is a malicious person. But, when it seems to me that someone real is malicious, I know that I am not pretending. The phenomenology of discovery is very different from imagination and MF (some versions) implies that we imagine moral claims being true, rather than discover them as true, and that’s wrong.

14. See, this is what happens when you go to sleep on the other side of the world: everyone tees off on the MFers.
    Let me just add a few points to the robust set of points already added.
    a) As I follow their parsing, NRW set up prescriptive MF as an alternative to error theory. While both views believe in the error of moral talk, ET instructs us to abandon moral talk (307) and MF is (putatively) distinctive in instructing us to retain it (308). (As NRW set it up, the conclusion to draw from Mackie is a dilemma that MF can escape: “either embrace an alien ontology…or become eliminativists about morality…. As a consequence we ought to abandon false talk of rightness and wrongness…”) I also suspected that this was an overread of ET, but since I don’t know the orthodox interpretation of Mackie on this issue of what to do once you realize the error, I figured it best to remain silent. By contrast, descriptive MF (as I understand NRW’s parsing) holds that our practice already approaches morality as a fiction, which, I gather, standard ET cannot claim, since there goes the error of believing in queer prescriptive stuff. While I’m also not sure abuot this, I believe that Joyce is standardly read as an MFer, not an ETer, at least by those sympathetic with MF–NRW ally themselves with Joyce (313), and take themselves to be supplying more reasons, beyond those cited by Joyce, to believe in the usefulness of moral discourse (again, unless I’m reading them wrong).
    b) I like Michael and David’s suggestion that (prescriptive) MF has a hard time saying that it is literally true that we should keep moral discourse, for this non-moral should seems no less problematic than the moral should. It sounds similar to Shafer-Landau’s rejection of the queerness argument on grounds that that argument requires inferential norms. In the case of NRW, their grounds for MF making a recommendation (to conserve moral discourse) are pragmatic, and they also set it up such that MF utilizes norms of inference. But Jamie suggests that, at least in Joyce, moral normativity is uniquely strange.
    c) I’m not sure that MF needs to idealize (or, if it did need to, that that would threaten it). For they could still say that within the fiction, some moral claims that we think are (fictionally) true, are not. It’s a complicated fiction, and just like we can misunderstand what happens in a novel, we could misunderstand what happens in the moral fiction (because we misapply principles, etc.).
    d) On Michael’s point about how the MFer would treat (K), if by “true” and “false” we mean literally true and false, then the MFer would adopt (2). But, given that prescriptive MF (at least as set out by NRW) holds on to ideas of goodness (hence the pragmatism motivating their conservation claim) they can agree that (P1) is true. They just can’t agree that any prescriptive claim follows from (P1), such as (PC1) we ought to bring people Mai Tais rather than torture them. I agree that denying this prescriptive claim is odd, but I also share Chris’s worry that simply stating that there are obviously some moral (prescriptive) truths is, as an argument against MF, question-begging. However, it seems as though arguing that MF cannot account for part of ordinary moral discourse (namely truth-aptness) is not question-begging if MF argues that it can account for that part of ordinary moral discourse.

15. Returning to the ET vs. MF issue: If their disagreement is not about the literal truth of moral claims but about whether we ought keep treating such claims as literally true or false, then the concern raised by David and Campbell seems important: In what sense ought we continue to so treat them? It can’t, without threat of contradiction, be a moral ought. The most natural route for MF to follow (I would think) is to say that moral claims, if thought to be literally true, would be more likely to be accepted and acted upon, and that widespread adherence to such claims is better (in terms of people’s advancing their interests, creating the conditions for peaceful and mutually beneficial interaction, etc.) than the alternative. In other words, go Hobbesian. This is what several irrealists have done (Mackie and Joyce come to mind): argue that treating moral claims are being categorical rather than pragmatically or instrumentally justified, even though they are so justified, is pragmatically or instrumentally justified. Those more familiar with all these MF-ers: Is that a strategy they adopt?
    If so, then I wonder whether there’s more than a hair’s breadth of difference between MF and, say, a constructivist position that says that moral claims are made true by practical reasons looked at from some comprehensive, impartial, or universal point of view.
    Also, Josh – I appreciate your suggestion in connection with my point about the necessity of moral truths. But does saying that moral claims are necessarily true in the moral fiiction really help? There have to be some identity conditions for the discourse called “morality,” such that some accounts of this discourse would run far enough afoul of its essential features to be an account of some other (rival?) discourse. But morality does seem to have a tighter set of identity conditions than an ordinary fictional discourse (including that its true claims are necessarily true). While we might countenance that, say, “Romeo and Juliet” would still be “Romeo and Juliet” if Juliet hadn’t died at the end, that (assuming it’s true at all) torturing newborns for fun *might have been false* seems to me to violate our sense of moral claims as fixed by the non-moral facts on which they presumably supervene. I don’t have a theory of the identity of a fiction, but it strikes me that MF has much less latitude in, e.g., trying to convince us that morality is akin to a fiction than does the person who tries to convince us that the Abominable Snowman is a fiction.

16. Clarification: Why did I say “x is useful” is false if MF holds?
    “x is useful for y” might still be true, as a purely descriptive claim. But I think “x is useful,” full stop, is normative. You wouldn’t say, “Balls of lint are useful”, even though there surely exists a y such that balls of lint are useful-for-y. Being useful (generically) seems to require being useful for something that itself is valuable.
    Also, I think that saying “All we meant was that x promotes happiness” or something like that is a dodge. One is really presupposing a value system according to which happiness is what matters, while pretending not to assume any values. We don’t say “x is useful” or “we should preserve x” on the grounds that x tends to cause misery, or that x tends to create lint balls, simply because we don’t value misery or lint balls.
    Even if you say that moral discourse tends to help people satisfy their desires, I think that doesn’t explain why the MFer would say “we should preserve moral discourse,” unless the MFer is assuming some sort of quasi-preference-utilitarian ethic. Imagine someone who thought preference-satisfaction was neutral or even bad: this person obviously wouldn’t accept the inference “x helps people satisfy their desires; so we should preserve x.”
    Notice, btw, that when NRW say we should preserve moral discourse, they probably are not making the unphilosophical remark that “Preserving moral discourse helps satisfy Nolan’s, Restall’s, and West’s desires.” They are probably claiming that preserving moral discourse helps satisfy most people’s desires.

17. More on prudential and other “oughts”: If there is a prudential “ought” distinct from the moral “ought”, I think the prudential “ought” would be open to the same arguments that are supposed to show that moral oughts are a fiction. E.g., prudential oughts require non-empirical knowledge, non-natural properties, and intrinsically motivating properties. I say this because I think the concept of a person’s interests is irreducibly evaluative.
    For example, is it in a person’s interests to believe a comforting lie (that has no further negative consequences that they’ll ever be aware of)? I think one’s answer to that turns on a kind of evaluation. The question can’t be answered by observation or scientific experiments. Also notice that when you get an answer to that, it would presumably motivate you to behave in certain ways (well, maybe not, but I think this is as plausible as the claim that an answer to a moral question would necessarily motivate you).

18. Josh,
    I like your objection to MF. If I get it right, one way to formulate it might be as follows. Consider these 2 propositions:
    (K1) Kryptonite hurts Superman.
    (K2) In the fiction, Kryptonite hurts Superman.
    Standard fictionalism contains a semantic claim and a pragmatic claim. The semantic claim is that K2 is true in virtue of X (insert right truth conditions). The pragmatic claim is that in certain circumstances C, K1 is an acceptable shorthand for K2. (For instance, if you’re in a Superman book club, you’re probably OK skipping the “in the fiction” every time you open your mouth.) Fictionaists like to say that given these two claims, there is a sense in which some utterances of K2 can be said to be true (perhaps those made when both X and C hold), and that’s how you get some type of truth conditions for things like K1.
    If I understand it correctly, one way to frame your objection, Josh, is to say that the pragmatics of moral discourse are very different from those of Superman discourse. When I say “murder is wrong,” what makes my utterance true is not that certain circumstances hold – e.g., I’m in a murder-disliking club. It is not that there are conditions that make it OK for me to skip “in the fiction” when I talk about morality, in the way it works for Superman discourse.
    Actually, as I read what I just wrote, it doesn’t seem to be exactly what you’re up to. Well let’s say it’s Josh-inspired rather than a Josh interpretation…

19. Michael C., yes, MF does hold that we should keep moral claims around because they are useful–the claim “we ought to conserve moral discourse” is by their lights a pragmatic ought. The important difference (or, anyway, one important difference) between MF and constructivism is that for MF this does not make moral claims true.
    On the question of carving out a distinctive space for morality, I think an MFer would say that the imaginative resistance you speak of might be part of uniquely moral terrain (or perhaps morality plus a couple of other domains). On that score, different fictions can have different characteristics.
    Michael H., right, NRW are definitely saying that moral discourse is useful full-stop, or useful for the community, not that it is useful for NRW. But I think that there is a distinction to be drawn between moral and pragmatic normativity. It might be hard to say just how we should draw the line, but, anyway, the idea that there is such a line seems plausible. So I’m not confident that we can accuse the MFer of presupposing a moral value system, as long as they’re claiming only to be making a pragmatic claim.
    (BTW and for the record, I’ve been at times associating NRW with MF. As far as I recall, this is an overstatement: they’re simply trying to lay out the advantages and disadvantages of MF. I don’t remember them actually saying “We are moral fictionalists.”… But then again maybe it’s hard to publicly identify as a MFer 🙂 .)
    Uriah, yeah, I think that is a somewhat different objection. Your point, if I follow, is that in ordinary moral discourse, it is not true that the truth of moral claims–moral analogues to your (K1)–depends on something (some circumstances) about the claim-maker. I think that’s true (in fact, I think that’s a variation on the above-mentioned objection to quasi-realism’s account of moral claims’ truth-aptness). But can’t the MFer agree? That is, I think the MFer (at least as portrayed in the NRW article) would say that the truth-makers for first-order moral claims are not geared towards the speaker, but towards the world: the reason there are no moral truths (in the literal sense) is that there are no prescriptive things (or properties or whatever).
    Or maybe instead you were more specifically saying that moral claims’ truth values are not dependent on the speaker having stepped into the fiction–this sounds related to Christian’s point above that, phenomenologically, finding that, say, slaveholding is wrong, feels like discovery, not pretending.
    I might have misread your point, though. Just to be clear, my objection to MF was that the truth of moral claims is (in ordinary moral discourse) meant to be in part a function of the nature of the action to which rightness or wrongness purportedly attaches.

20. I think that Jussi and Jamie’s comments above give some useful insight into the sort of fictionalism that Mark Kalderon proposes. It is too late for me to say anything very coherent here, but it might be useful to say that among the kinds of fictionalism that are Mark’s paradigms for fictionalism in ethics are van Fraasen’s (I hope I spelled that correctly) constructive empiricism and fictionalism about possible worlds.
    As I recall it is consistent with such views that people can put forth the claims of the theory while expressing different attitudes towards them. For vF people can assert them as true or merely put them forward as empirically adequate. I think something similar is supposed to be true on Mark K’s story. Some people can use moral utterances to express their beliefs whereas others can merely be expressing their sense that these beliefs are useful for various purposes, or their beliefs that it is somehow as if this or that action has this or that property of the sort that the theory of morality proposes.
    Whatever the merits of such views when applied to science or morals, I’m pretty sure that I am expressing an attitude other than belief when I’m using possible worlds talk. But not all users of such talk were doing that. So it might be helpful to think about that as a model.
    I think this means that you could be a fictionalist and think that some moral/normative claims were true whereas others would be merely such that is were in some way as if they were true. Any given speaker would not even have to have made up her mind as to which are which with any determinacy. You’d need to look at the reasons for thinking that moral judgements were merely fictions to think about which parts of what we say would be mere fictions and which might in fact be straight forwardly true, if that were the line you wanted to take. (I say this last bit by way of oblique reply to the claims above that fictionalists seem to be refuted by their other seemingly normative commitments.)

21. Mark,
    I think that folks above were thinking that the rationale for saying that MF is refuted by its own normative commitments is that its own argument against realism is that moral norms purport to refer to queerly prescriptive stuff. But if pragmatic or inferential norms are also prescriptive, then that suggests that it is incoherent for MFers to appeal to pragmatic and inferential norms. That doesn’t presuppose that there might not be some sort of domain-specific reason to be a moral anti-realist, which might not be applicable to other domains of prescriptivity.
    Maybe Kalderon proposes such a domain-specific reason [again, I haven’t read his book yet], but as far as I can recall, NRW motivate non-realism for MF only on the grounds of the queerness argument–their task then being to see how MF squares against other kinds of non-realism.
    Or, (as seems to be emerging here in discussion for those of us who haven’t read the book) maybe he’s a descriptive fictionalist. If so, then he need not account for any prescriptive claim that we should retain the fiction. His claim instead would be that we simply do retain the fiction, knowingly (in some sense). But even then, presumably he’s got some arguments for his position, in which case he needs to account for prescriptivity in the inferential norms presupposed by those arguments. (This is more Shafer-Landau-esque than the comments above from Michael H. and David, because now the target is inferential, rather than pragmatic, norms, but I believe the gist is the same.)

22. Just a short comment on Kalderon’s position. I think in this sense he is a descriptive MF. It is for him a contingent matter whether our mental states in sincerely accepting moral claims are beliefs, sets of pro-attitudes or entertained fictions. He ends up with the last as a result of some kind argument from exclusion. Our moral commitments cannot be beliefs because our moral discourse lacks certain epistemic obligations in certain disagreement cases and such obligations would have been a necessary requirement for having a cognitive inquire in which the participants commit themselves to mental states aiming at truth. The argument against these mental states being sets of pro-attitudes is the classic Frege-Geach from the surface structure of our moral discourse. Thus what we end up with are moral fictions. I think he is open to the idea that it is an open question where we go from here as a moral community. We could sharpen up the epistemic norms to commit ourselves to beliefs or change the surface structure to allow us to express pro and con attitudes.
    I also know that there is undergoing work on the seeming incoherence of being an ET in ethics but accept realism about other areas of normativity. I know that there is a global error theory under construction, but not sure I can tell more details at this point (it’s not me).

23. Thanks, Jussi, very helpful. Except now I can’t stop wondering about global ET!

24. Cambo,
    I was thinking that the MFers were offering us a reason to keep moral discourse around, not just telling us that if we play a certain game which we have no reason to play, we will, by the rules of that game, have reason to play the game.
    Also, I guess there might be many motivations for error theory about morality, but some of the biggies seem to me to not carry over to cause trouble for prudential or instrumental norms. For example, one might think that morality asserts of itself that it provides everyone a reason to comply. One could deny that this is so while maintaining that everyone has a reason to satisfy their desires. Roughly, moral norms need not motivate rational people even after they know the facts, but there are conceptions of normativity that necessarily would, such as broadly Williamsian subjectivism.

25. I think David’s comments about different reasons to be a MF is relevant and important. I’ve only read drafts of various parts of Kalderon’s book and don’t know for sure what of what I read was in the final version, but one of the main arguments in draft had to do with the irresolvability of certain sorts of moral dispute. Those arguments may not carry over into prudential and instrumental reasoning as David suggests some reasons won’t.
    And (now this is me speculating and not attributing this to Mark Kalderon or anyone else) it also suggests that you could treat some parts of the “morality story” as fictitious without treating all moral claims as so. So you could think the idea that moral claims are true simpliciter (as opposed to true relative to this or that moral outlook) was a fictitious part of the theory which had value insofar as it keeps people talking and trying to persuade one another, even if it is strictly speaking false.
    You might try something analogous for the claim David suggests, that categorical reason-givingness might be false.

26. The qestion has been raised here, in what sense of ‘should’ do Nolan, Restall, and West claim that we should continue to use moral language? By way of an answer, let me quote from their paper:

    We believe that moral fictionalism is more attractive as a prescriptive doctrine: not that we do, but that we should, employ moral fiction. This “should” is not a moral should; rather, it is some sort of pragmatic “should”: given the interest we have in co-ordinating our actions, inculcating and supporting our deeply-held preferences, and perhaps other desires we have, we will be well served to morally fictionalise.

    The paper, which I highly recommend, is available here:
    http://www.arts.usyd.edu.au/departs/philos/staff/documents/mf.pdf

27. This brings us nicely to a problem which I’ve started to think is a biggie for NRW’s and Joyce’s MF. I’m sure they probably address the problem somewhere but it’s still worth of it to bring this up. Michael already mentioned a part of this. It seems that there is a temptation for the prescriptivist MFers to go Hobbesian, and say that there is a pragmatic ground, based on our personal preferences, for employing the moral fiction (as Campbell’s NRW quote says). I remember having hard time figuring out what distinguishes Joyce from say Gauthier (still not sure if there is a difference).
    Now, the classic problem with the Gauthier type contractarianism is that there is an important moral phenomenon which the view has deep problems coping with. I think the same goes for the prescriptivist MF. There are situations in which our personal preferences would be better served by directly pursuing them rather than by following the moral fiction/agreement. Some of these situations are even such that the burden squarely falls on someone else, the sucker. Your ability to beneficially co-operate later on is not threatened as a result nor is the general co-operative atmosphere of the moral community which benefits you. Sometimes one would just get away with things and others end up being suckers (There is a nice paper about this in Reading Gauthier by someone). Maybe for these purposes we may need to act as if we were in an inescapable grasp of the fiction.
    At this point the problem is that if the fictional moral ought really is constituted of the pragmatic ought, it just vanishes. It’s not easy to see how the MF view would be in a position to say that there is something objectionable in disregarding the fiction. The reasons for having the fiction in general just do not apply in these cases and so without the pragmatic ought the fictional ought looses its force in these cases (I’m very uncomfortable with these different oughts). But, I take it that this result goes against our deepest considered moral judgments with which we evaluate these views. We really ought not to do these things. How would a prescriptivist MF view based on the pragmatic justification of the moral fiction deal with this issue escapes me.

28. I have a very simple argument against moral fictionalism. It goes like this:
    1. Where 2 propositions p and q are both things that we have some reason to believe, but q iff not-p, we should believe whichever of p and q we have more reason to believe.
    2. We have very strong reason to believe the proposition that torture and rape are seriously bad.
    3. We have some reason to believe the proposition that moral fictionalism is true.
    4. If moral fictionalism is true, then it is false that torture and rape are seriously bad.
    5. “Very strong reason” is more reason than “some reason”.
    6. Therefore, since we cannot believe both the proposition that torture and rape are seriously bad and the proposition that moral fictionalism is true, we should believe that torture and rape are seriously bad and deny moral fictionalism.
    There’s something like this in Michael Huemer’s new book on intuitionism. Indeed, there may be something like this on this blog already; sorry if I’ve scrolled past it!
    Responses to this argument may well target premiss (2), and claim that this question-beggingly presupposes that ordinary belief in the badness of rape and torture is realist belief. To which I reply that there’s nothing question-begging about that presupposition. The onus is on the fictionalist to explain why we should move away from it.

29. I think that Tim Chappell’s principle 1 can’t be quite right as stated, since some reason to believe something might not be enough reason to rationalize believing it.
    Relatedly, I think it also matters whether the other alternatives to the thing we have less reason to believe are or are not also incompatible with the truth of the thing we have more reason to believe. If the reason we have less reason to believe fictionalism is that more standard non-cognitivism might instead be the right account, or that outright skepticism might be right, and these sorts of views are also incompatible with the truth of the proposition we think we have more reason to believe, then it looks like we maybe should not just accept the first order belief based only on our having more reason to believe it than fictionalism.
    This isn’t to say I believe fictionalism.

30. I also worry that a fictionalist could say that, from his perspective, the argument equivocates between different senses of belief and true/false. After all a fictionalist will make a distinction between believe(in a fiction) and believe (actually), and true (in a fiction) and true (*really*). So while (2) would speak about reasons for believing (in a fiction) that it is true (in a fiction) that torture is bad, (4) will just refer to the idea that it is (only?) false (*really*) that torture is bad. From this only follows that we have reasons to believe (actually) that it is false (*really*) that torture is bad. This means that the crucial conflict in the argument does not rise. I guess the point is that the fictionalist can use the same normative ethics/metaethics machinery as expressivists to avoid this kind of problems. I’m not saying that this is completely convincing but it looks like what they would say.

31. I agree with Tim’s argument (not surprisingly). In response to Mark, (1) could be rephrased:
    1′. If p and q are incompatible, and p is more justified than q, then we should not believe q.
    Take the case where you have more reason to believe realism than MF, but only because MF is competing with non-cognitivism and other anti-realist views. Say the probabilities are like this:
    P(realism) = .4
    P(MF) = .3
    P(Non-cognitivism) = .3
    P(anti-realism) = .6
    In this case, we clearly should not believe MF, since it’s less than 50% likely to be true, and that of course must be the case as long as we allow some incompatible view to be more justified than MF.
    However, note also that non-cognitivism is typically not supposed to be incompatible with first-order moral ‘beliefs’ — the non-cognitivist is supposed to be able to say, “Torture is bad”. It’s really only the ETer or MFer who denies that torture is bad.
    In response to Jussi: I think (2) should be read as saying that we have strong reason for genuinely believing “Torture is bad”, not just pretending to believe it (is that what “believing in a fiction” means?).
    Btw, this sort of argument is G.E. Moore’s response to skepticism (e.g., he thinks it’s more obvious that “I know this is a pencil” than that any of the skeptic’s premises are true). And I think we should view MF and ET in metaethics the same way we view skepticism in epistemology.

32. If the argument is analogical to Moore’s argument against scepticism, then it, rather than equivocating, just begs the question in premise (2). Just as the sceptic says to Moore that his premise ‘I know this is a pencil’ assumes that there really is a pencil to be known about (the conclusion Moore wanted to have), the error theorist/fictionalist is going to point out about (2) that having a reason to believe that torture is bad assumes realism (that there are instances of torture with the property of badness that provide us with reasons to believe in the existence of them). Thus, given that already the second premise of the argument assumes the truth of the conclusion, the argument cannot transfer more warrant to the conclusion.
    By believing (in a fiction) I tried to indicate a belief one may have while immersed with the fiction. When I believe that Sherlock was killed by Moriarty I’m certainly not pretending to believe anything.

33. Returning to the question of the kind of ought involved when MFers say we ought to accept moral fiction. One version of this would be merely to publically mouth the moral convictions that permit us to be on good terms with our neighbors. A stronger version would maintain that together with paying lip service to morality for prudential reasons, we also ought to play like we believe in the truth of moral claims by being moved by them, etc. Some, eg. Gauthier, claim that we can only accomplish persuasively being seen to care about morality if we in fact care about morality. This, if true, would connect the two thoughts, but it is a difficult argument to make with any robustness.
    Do MFers mean to recommend to us moral concern or merely lip service? If they also recommend the former on “pragmatic” grounds, do they buy into the Gauthier strategy or have some other strategy?
    If MFers do not recommend genuine moral concern, then I would have said that it is not clear that such a view counts as recommending that we accept the fiction that moral claims are true–for it seems to me plausible that accepting this fiction should strongly co-vary with moral concern.

34. Chappell and Huemer,
    It would help me if you would explain why an argument of this style is not analogous to the argument Chappell offers above.
    1) we have very strong reason to believe that this chair is solid and not made up mainly of empty space.
    2) we have some reason to believe that this chair is made up of very small particles with mostly space between them
    3) therefore we should reject modern science about chairs.
    I guess the reason the argument strikes me as at least minimally similar is that the first premise is a very confident intuition that we have pre-reflectively. The argument seems to show cases where confident intuitions we have should be revised on the basis of good argument and research. One might claim that 2 above, properly understood, really does not conflict with but vindicates 1 above. But I suppose the MFer might try to claim that they too mean ultimately to, in a sense, vindicate ordinary moral intuitions. Further, one might try claiming that 2 above is something that we really have more reason to believe than 1 above, but how to make this out without threatening Chappell’s argument?

35. “If they also recommend the former on “pragmatic” grounds, do they buy into the Gauthier strategy or have some other strategy?”
    I don’t think it can be the Gauthier strategy. Gauthier manages to provide (really, at best) a basis in rational self-interest for cooperating with constrained maximinzers in a very limited range of cases. It is a pretty thin morality that follows: an ultra-liberarianism. I’m guessing the MFers (‘MF-ists’ might be kinder) have in mind a broader morality and so might need more assumptions than Gauthier offers.

36. I do think Mike’s restatement of Tim’s argument avoids much of my worry about the formulation of the first premise. I’m not sure though that I see how the fictionalist needs to deny the first order claim about the wrongness of torture/slavery etc. more than a non-cognitivist does. In some sense it seems like both the non-cognitivist and the fictionalist deny the truth of the first order claims construed as realists would like. But I think that Jussi is right that the attitude of the fictionalist towards those truths does not seem very different from that of a sophisticated non-cognitivist who is careful never to deny the truth of what she is morally committed to.
    And it may even be that the fictionalist can believe the first order claims while treating morality as a whole as a fiction.
    Most fictions contain truths. Suppose you think what makes the whole moral story false is that it is committed to categorical imperatives, but that there cannot be any genuinely universal reason-giving properties. Yet you also believe that it is not always obvious that everyone is outside of the range of decent arguments for accepting (that is acting on) moral claims like the claim that slavery is wrong.
    You can believe that you therefore have reason to continue on talking and acting as if the moral fiction that moral properties give categorical reasons for action is true. You may fully believe that slavery and torture come as close as anything does to offering reasons for all people to oppose them. Furthermore, you may think that you have such reasons and have a commitment to oppose them as you can. As a result it seems to make sense to engage in the practice of speaking as though the fiction is true.
    Does this mean you don’t believe that slavery and torture are wrong? I’m not sure what to say. You think that in the fiction, something has the property of being wrong only if it offers a categorical reason to oppose it. And you don’t believe anything does offer such reason even while given your commitment to opposing slavery it may make sense to pretend that you do.
    When I believe a story, I know that some of what is in the story is false. But that does not mean that I disbelieve every claim in the story. I believe London is a city even though in many fictions London is a city. In some story, it may be true that some fact presupposes another. Yet it is not obvious that every such fact which in the story presupposes another is doubted by me whenever the presupposition is true only in the fiction but not in reality. Perhaps in some story, London is a city only if Metropolis is a city. But that does not mean I doubt that London is a city even when I treat that story as a fiction. So why should the fictionalist’s doubts about the categorical nature of morality force us to say he or she doubts that slavery and lying are wrong?

37. Because, Mark, the fictionalist’s view says this: “That slavery and lying are wrong is true-in-the-moral-fiction, and that slavery and lying are wrong is false”. Just detach the second conjunct, and there you are.
    Isn’t that enough to show why the fictionalist has to be committed to the falsehood of all first-order moral claims? Very simply, “true-in-the-moral-fiction” doesn’t conflict with “false”: that’s the whole point of “t-i-t-m-f”.
    Thanks, Mark, for pointing out that my first premiss should have been stated as Mike stated it, or as this:
    Where 2 propositions p and q are both things that we have some reason to believe, but q iff not-p, we should believe whichever of p and q we have more reason to believe (if we have enough reason to believe either).
    And thanks more generally, folks, for all your comments on my Moorean/ Huemerian argument, which are full of helpful things that I’m still thinking about.

38. Incidentally, I like David Sobel’s argument about chairs. I too worry that the Huemer/ Moore argument strategy leaves the door open to this sort of riposte.
    I guess my answer, in the particular case that David’s picked, is that there’s no conflict between what modern science says about the empty space in chairs, and what common sense says about their solidity: in its ordinary-language use, “solid” just isn’t inconsistent with the scientific facts about the spaces between and inside atoms.
    However, that may be a feature just of this example. In other cases, the inconsistency between science and well-rooted folk-belief will be clearer. And there will of course also be the question “Whose folk-belief: ours, or the Azande’s?”

39. Here are a few quick bloggish thoughts; I wish I had something more solid to say.
    First and incidentally, I got myself a little confused reading the last couple of posters call Michael Huemer ‘Mike’ just after Michael Almeida signed his comment ‘Mike’, but I’ve got it now (I mention this lest anyone else feel similarly confused).
    Second, I wish there were someone lurking who’d defend moral fictionalism. It’s a bit surprising that of all the metaethicists chiming in here not one seems to have any sympathy for fictionalism at all (including me). I can’t help worrying that we’re going to miss whatever interesting defense fictionalism could present.
    Finally and substantially, my gut feeling (this is the “I wish I had something more worked out to say” part) is that Tim Chappell’s argument is missing its mark. In speaking of Sherlock Holmes, we almost never preface our sentences with “In the fiction”. When we ask ourselves whether Sherlock lived in Baker Street, we are extremely confident in answering “yes”. The premise “We have very strong reason to believe that Sherlock lived in Baker Street” strikes us as obviously true. This does not refute fictionalism about fictional discourse. I doubt, therefore, that our similar intuitions about moral sentences and propositions can refute fictionalism about moral discourse.
    How exactly is TC’s argument blocked? This is where I admit to feeling shaky.
    

    Because, Mark, the fictionalist’s view says this: “That slavery and lying are wrong is true-in-the-moral-fiction, and that slavery and lying are wrong is false”. Just detach the second conjunct, and there you are.

    Detachment of the conjunct might be a lot slipperier than it looks. Roughly, when we consider the conjunction the explicit ‘in-the-fiction’ operator makes it clear that the sentence as a whole is fully, literally, …, asserted. When the second conjunct is inferred, it is left bare, without any signal that it is a straight assertion. The fictionalist thinks that such sentences in their typical use do not make assertions, but a different kind of speech act, and in that typical use they are perfectly acceptable and unobjectionable, which is why they strike us as so intuitively right, but they are not strictly true.
    I think probably other accounts of fictional discourse (say, with tacit ‘in the fiction’ operators) can provide a similar defense.

40. Sobel asks:

    
    Do MFers mean to recommend to us moral concern or merely lip service?
    

    I’m not sure what’s at issue here. Do you think that this question arises only for defenders of MF? Would adopting some other metaethical view make the question go away? I don’t see why it would. If we were moral realists, for example, we would hold that some moral sentences are literally true. But that a sentence is literally true is not a (pragmatic) reason either to “pay lip service” to it or to have concerns that correspond to it. So even realists must face up to the questions you ask, and I don’t see why answers will come any more easily for them than for moral fictionalists.

41. Thanks for those comments, Jamie (and Hi! by the way, since we’ve communicated before).
    I’m sticking to my guns, I think. The key thing about moral fictionalism, as I understand it, is its commitment to the falsehood of first-order moral utterance. For a fictionalist such utterances may be useful, serve a non-assertoric purpose, etc., just as for an expressivist. What then is the difference between the expressivist and the fictionalist? Surely just this: that the expressivist says these utterances are neither true or false, whereas the fictionalist says that they’re false. So I don’t see all that much to slip up on, around my proposed detachment of the second conjunct.
    As for Sherlock Holmes, I think I must come from a different “we” from you, because I quite often preface my sentences about him with (some ordinary language equivalent of) “In the fiction”. And when I ask myself whether Sherlock lived in Baker Street, I am not at all confident in answering “yes” without qualification– though I am extremely confident in answering “yes, in the book he did”. So, as you say, “the premise ‘We have very strong reason to believe that Sherlock lived in Baker Street’ may strike me as obviously true; yet not in the way that “Madam Tussaud’s is in Baker Street” strikes me as obviously true. And in explaining to some lost tourist from Rhode Island (say) what kind of mistake he was making, if I found him looking for Sherlock’s house in Baker St, I think it very likely that some ordinary language equivalent of the phrase “In the fiction” would cross my lips.

42. Sorry, the penultimate sentence in my last post should have appeared like this:
    So, as you say, the premise “We have very strong reason to believe that Sherlock lived in Baker Street” may strike me as obviously true; yet not in the way that “We have very strong reason to believe that Madam Tussaud’s is in Baker Street” strikes me as obviously true.

43. On the threat of repeating myself, I need to quote this from Tim to get started:
    ‘I’m sticking to my guns, I think. The key thing about moral fictionalism, as I understand it, is its commitment to the falsehood of first-order moral utterance. For a fictionalist such utterances may be useful, serve a non-assertoric purpose, etc., just as for an expressivist. What then is the difference between the expressivist and the fictionalist? Surely just this: that the expressivist says these utterances are neither true or false, whereas the fictionalist says that they’re false. So I don’t see all that much to slip up on, around my proposed detachment of the second conjunct.’
    I much doubt that with the current popular minimalist/deflationist views about truth around we can make the notion of truth do this much work in distinguishing between any philosophical positions – let alone metaethical ones. I bet every realist, expressivist, fictionalist and whatsoeverist (excluding error theorists) are bound to say that of course it is OK (and even required to) say:
    ‘Torture is bad’,
    ‘It is true that torture is bad’,
    ‘It is an objective, mind-independent fact that torture is bad’,
    ‘It is written in stone that torture is bad, full stop’, and so on.
    For all parties concerned these just come out meaning the same. Truth just is by default too cheap for grounding such distinctions unless you are prepared to really defend robust correspondance views of truth. That is heavy going. Simon Blackburn is (again) really pounding this point home in his new paper in Jamie’s new collection.
    If we are going to tell these metaethical views apart I think there are two more promising ways of trying to do this. First, you can start from the content of all the utterances above and the related speech-acts. Do they assert something about the world (realists (both naturalists and non-naturalists)), quasi-assert or characterise the fictional realm (fictionalists), or express pro/con -attitudes (expressivists)? The second way would be to pay attention to the psychological states in moral judgments and their functional role. Are they beliefs (cognitivists), sets of pro/con -attitudes (conativists), or entertained fictions (fictionalists)? Tackling these issues just requires more substantial philosophical work than arguments like the one above.

44. What then is the difference between the expressivist and the fictionalist? Surely just this: that the expressivist says these utterances are neither true or false, whereas the fictionalist says that they’re false. So I don’t see all that much to slip up on, around my proposed detachment of the second conjunct.

    Hm, I don’t get it. Of course, the fictionalist says that the second conjunct is false. But he thinks that when standing alone it should seem perfectly acceptable. That’s because he thinks that (typically) when standing alone it does not serve to assert the moral proposition in question. It is not used to assert anything. The fishiness of the detachment is to be found in the pragmatics, not the semantics.
    I boldly conjecture that over 90% of intelligent non-philosophers will answer Yes to the question, “Did Sherlock Holmes live in Baker Street?” Do they believe it? I don’t know. This is a hard question for the theory of fiction.
    As for the two ways of striking you as obviously true, I just have to admit that I am unable to discern two clearly distinct strikings.
    Nonetheless, I don’t for a moment believe moral fictionalism.

45. I have a question about the Chappell/Huemer argument, in particular this premise:

    2. We have very strong reason to believe the proposition that torture and rape are seriously bad.

    What is this “very strong reason” supposed to be? Perhaps it is that we often hear people say “torture and rape are seriously bad”, or that when we think about torture and rape we feel strong aversion or repulsion. But these are exactly the things we’d expect to hear and feel if there were a pervasive moral fiction according to which torture and rape are seriously bad. So, at best, they are very strong reasons to believe a disjunctive proposition: either torture and rape are seriously bad, or there is a pervasive moral fiction according to which they are seriously bad.

46. Jussi, your reference to the debate between deflationism and a substantive theory of truth is interesting and helpful. For my own part I’d be happy to go deflationist and rewrite the debate in the terms you suggest. However, the terms of the debate were set by the fictionalists, not by me. Notice, for instance, that way up at the top of this blog Joshua Glasgow, reporting Restall et alii’s report of the fictionalists’ views, characterises moral fictionalism in terms of its commitment to the view that moral claims are literally false.
    Jamie, I’m slightly puzzled by your comment that “Of course the fictionalist says that the second conjunct is false.” For the second conjunct was “that slavery and lying are wrong is false” [outside the moral fiction]– and I thought the point about the fictionalist was that he took this claim to be *true*.
    Campbell, the question ‘What is this “very strong reason” supposed to be?’ is one that Jimmy Lenman’s also been pressing on me. My answer to it is that I think it’s a ground-level intuition that murder etc. are seriously wrong. Compare the ground-level intuition that I have hands– for from there, as noted, the argument proceeds in Moorean fashion.
    I should say, incidentally, that since I’m not sold on the Moorean style of argument in epistemology, I’m not sold on it in metaethics either; but I do think it’s interesting to give it a run for its money.

47. Sorry, I got tied up in the falses. I meant that the fictionalist will agree with the second conjunct that the moral proposition in question is false.

48. In defense of MMR (Moorean Moral Realism):
    (1) Begging the question?
    If Moore’s anti-skeptical argument ‘begs the question,’ then I think it doesn’t matter whether one begs the question or not. It seems clear to me that Moore’s premise, “I know this is a pencil,” was more plausible than any of the skeptic’s premises (let alone the conjunction of all of them), and that’s what matters to deciding which position to adopt. It seems clear that Moore’s argument was therefore the more rational (at least, I don’t know of any reason why it wouldn’t be). For example, “No one is ever directly aware of a physical object” is less plausible than “I know this is a pencil” (in the appropriate context). So it would make more sense to argue that people can be directly aware of physical objects on the grounds that this is the best way to account for the fact that I know this is a pencil, than to argue that I don’t know this is a pencil on the grounds that (among other things) no one can ever be directly aware of a physical object. I can’t see any way that the latter argument could come out as more rational.
    Similarly, we have to ask ourselves whether the premises of ET are more or less plausible than the propositions they’re used to ‘refute’, like the claim that having bamboo shoots stuck under one’s fingernails is bad. It seems to me that that (and not “who’s begging the question”) is the relevant question in deciding whether to be an MFer or a MRist.
    (2) David’s Chair argument
    I think there are three issues with the chair argument. First, I don’t think that “x is solid” is incompatible with “x is made of atoms with relatively large spaces between them.” I think atomic theory actually explains the properties characteristic of solidity, properties such as having a relatively fixed shape, resisting shear forces (unlike liquids), resisting compression (unlike gases), and so on. In contrast, though, I take it that MF is explicitly inconsistent with ordinary moral claims: MF (as we’ve been discussing it) says that all positive moral claims are false.
    Second, I don’t think that we have strong reason to believe that the chair is not made up mostly of empty space. I don’t see what reason we have to believe that. We can’t observe spaces of the size that atomic theory says exist, so we have no empirical reason to believe that; and I don’t think this is the sort of thing we could expect to know a priori.
    Third, I think that the second premise is an understatement: we have very strong reason to believe atomic theory. In contrast, though, I don’t think there is any argument for MF or ET that is remotely as powerful as the scientific case for atomic theory. (If there were, the community of experts would probably agree on MF about as much as scientists agree on atomic theory, instead of its being the case that it’s hard to find someone to defend the view.)
    (3) “We have very strong reason to believe that Sherlock lived in Baker Street”.
    As Jamie notes, that seems obviously correct. However, I think it is elliptical for something like: “We have very strong reason to believe that, according to the Sherlock Holmes stories, Sherlock lived on Baker Street.” If instead you say: “We have very strong reason to believe that Sherlock Holmes really lived on Baker Street,” that’s obviously false.
    In contrast, in the ethical case, we have very strong reason to believe that being subjected to the oral pear [a medieval torture device I learned about from Ryan Wasserman] would really be bad. I don’t think we merely have strong reason to think that according to people who do morality-talk, it would be bad.
    (4) Slavery:
    By the way, it strikes me as wishful thinking to suppose that almost everyone has prudential or other non-moral reasons for opposing slavery. The prudential reasons why slave-owners supported the institution are pretty obvious. I’m inclined to view any argument that it wasn’t in their interests as sophistry. So I don’t think you’re going to make a case that nearly everyone should have opposed slavery, without relying on the moral facts.
    (5) Our reason for thinking torture is bad
    Well, I think our strong justification for thinking that torture is bad consists in the (strong, widespread, unwavering) intuition that torture is bad (which is also consistent with every seriously proposed moral theory).
    I don’t see MF as accounting for this at all. I don’t have anything like the attitude toward morality that I have towards real fictions (like Sherlock Holmes). I’m not inclined to insist that Sherlock really lived on Baker St., nor do I plan my life around supposed facts about him.
    (6) Did Sherlock Holmes live on Baker Street?
    In most contexts, if you ask this, most people (familiar with the stories) will say “Yes.” However, in some contexts–where the context somehow makes clear that you’re asking the question literally, as when you don’t realize that Sherlock didn’t really exist–people will clearly answer “no.” And I think the effect of the “really” operator is instructive: “Did he really live on Baker St.?” will presumably elicit a “No.” I suggest that this is because in fact, no such person ever lived on Baker St. “Really” just functions to remove the implicit “According to the story” operator that the listener might be charitably inserting.

49. Thanks for the clarification, Jamie. Seems to me that, on your view, a claim like this–
    (1) That Madam Tussaud’s is in Baker St, and that Sherlock’s home is in Baker St, are both truths
    — should be utterable without any sense of weirdness: so to speak, you should be able to say it without “seeing the join” between the two halves of the sentence. But I doubt that it is so utterable. (1) sounds like a joke to me, precisely because I can “see the join” in it: it’s a zeugma, like “she left in tears and a taxi”.
    Overnight I’ve been thinking a bit more about whether Campbell’s question leaves the Huemer/ Moore argument in the soup (*sigh*… sorry). Comparison with the epistemic analogues of the argument does suggest that premiss (2) ought to be the Achilles’ heel of the argument. And so, perhaps, it is: in one email to me yesterday, Jimmy Lenman called it a charter for intellectual complacency. I see the force of that protest, but I still want to try and make more philosophical sense out of the fact that there are claims (in ethics and elsewhere) that we find irresistibly plausible without further argument or defence. “2 + 2 = 4”; “there is a sun”; “murder is bad”; that sort of thing. We find these claims to be obviously true, and do so without the need of argument (which is not to say that arguments can’t be provided for them). “Intuitions” is as good a name for such claims as any.
    It seems to me that there is *some* evidential value in the mere fact that we find these intuitions convincing. I can see why one might worry about this claim; but take particular cases. I find it very very convincing to say that there’s a computer screen in front of me now. That looks to me like evidence that there is a computer screen in front of me right now. Similarly with the claim about murder.
    I doubt the argument I originally posed can get off the ground without some such intuitionist backdrop as this. And then the weakness of the argument, and in particular of premiss (2), comes down to this: my very convincing intuition that p may warrant me, subjectively, in believing that p, but it only justifies my belief that p, objectively, if I am a reliable observer in p-related areas. So the upshot of the argument is that intuitions can give anyone who takes herself to be a reliable observer subjective warrant for her beliefs, but can only give objective justification for her beliefs to someone who is in fact a reliable observer.
    If that’s right, it explains what the Moorean/ Huemerian style of argument can achieve; and also what it can’t achieve, and in particular, why it doesn’t refute scepticism.
    As you’ll have noticed, the explanation is not specific to the case of moral-fictionalist scepticism.

50. I see Mike H and I posted at the same time… the point up to which we agree should be fairly clear from my last post. I (now) think intuitionism only leads to subjective warrant, and not to objective justification, and hence doesn’t answer scepticism; and that’s the problem with the argument I originally posed. Presumably Mike will disagree.
    Off topic, by the way, Mike, Phil Stratton-Lake has agreed to review your book for the PQ– but he won’t be getting the PQ’s review copy till I’ve finished reading it! I think it’s terrific.
    And back on topic: I think moral fictionalism faces problems about motivation. Schematically, the problem is going to take this form. Just as “p is true” states a reason for believing p and “p is true-in-the-Sherlock-fiction” does not, so likewise “that x is right” states a reason for doing x, “that x is right-in-the-moral-fiction” does not. That’s no more than the outline of an objection– but I think it’s very likely to prove cogent.

51. hmh. I thought that in philosophy we wanted valid and sound arguments that transmit warrant from premises for our views in a way that we can persuade our rational objectors to our side. Thus, saying that it doesn’t matter whether we beg the question or not seems odd. By saying just that your views are more plausible than someone else’s we won’t get anywhere. And, I still have my doubts about the begging the question issue in Moore’s argument. I thought his argument was:
    1. If this is a pen, then there are physical objects and I’m not a BIV.
    2. I know that this is a pen.
    3. Therefore, there are physical objects and I’m not a BIV.
    Similarly, one can argue that:
    1*. If torture is bad, then some first-order moral claims are true and fictionalism is false.
    2*. I know that torture is bad.
    3*. Therefore, some first-order moral claims are true and is false.
    Now take the perspective of sceptic or an error theorist. Are they given any new reasons here for abandoning their views? I doubt it. And the rest of us already believe the conclusions.
    They are going to contest your warrant for claims 2 and 2*. Knowledge requires reliability – that you generally hold the given opinion only in circumstances where it is true and abandon where it is false. Now, the sceptic is going to say that your opinions ‘this is a pen’ and ‘torture is bad’ fail this criterion. You would have the same opinion even if you were a BIV or engaged with a moral fiction. So, on this challenge you lack knowledge in the crucial premises.
    To argue against the sceptic and the fictionalist for the premise that you do know after all that this is a pen and torture is bad, and that you would not hold these opinions in the wrong circumstances, you then need to argue that you are not now a BIV or in moral fiction and you wouldn’t think so if you were. But, the point is that you cannot now appeal in this further argument for the premise that you know that this is a pen or that torture is bad without the argument becoming circular and begging the question – after all we are now trying to argue just for the premise that you know that this is a pen and torture is bad. Thus, if you think that there is a sceptical and fictionalist challenge then you will need some external warrant for 3 and 3* that at the same time are warrant for 2 and 2*. Whatever warrant you have for 2 and 2* are thus the same warrant that there is for 3 and 3*. On the other hand, if you don’t think that there are such challenges in the first place, then we don’t really need to bother with the ‘argument’. There is a big literature on this following Dretske, Nozick, Davies and Wright that we need to take seriously if we are using Moore’s argument in ethics. I know for instance that Alison Hills is doing serious work on this topic which I’m excited about.
    Tim,
    could you try to explain what you mean by subjective warrant?

52. Tim,
    I’m still not buying the argument. Wouldn’t the existence of a pervasive moral fiction according to which torture is bad explain why people have the strong intuition that torture is bad? Sadly, many of my undergraduate students find the theory of evolution highly counterintuitive. They have a strong intuition that humans did not evolve from monkeys. How can we explain this? Here’s one explanation: they’ve been indoctrinated with a religious fiction according to which it’s true.
    On the motivation issue, I suggest you read the paper by Nolan, Restall, and West. They argue that MF can accommodate internalism.

53. This is a great thread.

54. Cambo,
    I was asking above whether MFers recommended on pragmatic grounds moral motivation or merely lip service to morality. You replied:
    Do you think that this question arises only for defenders of MF? Would adopting some other metaethical view make the question go away? I don’t see why it would.
    Me neither. All I was asking was how much of the package of morality they mean to be pragmatically vindicating. Surely that is a legit question to ask people who say they recommend on pragmatic grounds pretending that moral claims are true. I see the question as asking them what we are to pretend when we pretend morality is true. Are they recommending that we pretend that it gives us reasons such that we live as if it did (even when no one is watching)? That seems hard to pragmatically vindicate (unless one tries some fancy Gauthier type move).
    Some are attracted to judgement internalism about moral belief, but I am assuming that internalism about moral pretence is much less attractive. So there is room to think that agents that pretend that moral claims are true would behave differently from agents that believe that moral claims are true. I was also wondering if MFers accept that there might well be this difference.
    Chappell and Huemer,
    I think your responses effectively take care of my example. I do still wonder, with Tim, if examples of that kind could be constructed that would be more of a problem. But your help in responding to the chair case has helped me see that one would need to say much more than I was thinking of to mess with the Huemer-style argument.

55. Thanks, all, for the great comments. I’ve been away from the Internet, which has been very frustrating! But this has all been very educational. Let me just add a couple of points.
    1) Yes, the fictionalist says that all first-order moral claims are false.
    2) Let’s grant that intuition provides some reason for belief. But surely intuition can be defeated, right? Couldn’t a defeater be the queerness argument, for example? All I mean to say is that the fictionalist does give some reason to deny intuitions about moral reality, and then the discussion shifts to the merits of MF vis-a-vis other non-realisms.
    3) On motivation: presumably the MF reasons for doing x is not that x is right, but that we gain other (pragmatic) benefits from doing the “right” things, of which x is one.
    4) On what to believe if we really should believe the fiction: perhaps this is unwarranted, but I get the impression that prescriptive MF treats the moral code as a “package”, so that you’re supposed to believe the whole deal. Maybe if we let in free riders, this would weaken the systemic benefits of having the fictional code? Also, by “believe the fiction,” I think they have in mind “Fully internalize the MF, i.e., believe it as much as any story.”

56. The simple fictionalist view that seems to be the main target in this debate appears to consist of two thesis:
    1. All (positive) first-order moral beliefs are false.
    2. We still ought to continue believing them for say prudential reasons.
    I have tried to resist this formulation of fictionalism for variety of reasons and to push for a more sophisticated fictionalism that would have at least some chance. But, it may be that this is Joyce’s and NRW’s fictionalism (certainly isn’t Kalderon’s) and thus the proposal on the table.
    I’m starting to think that there is another simple, non-question begging argument against this very view (I think the more plausible versions of fictionalism escape it). It’s just too close to violating ‘ought implies can’ by being psychologically way too demanding. Say that I’m a good fictionalist, conscientious moral agent. I am thoroughly convinced of the truth of 1 and 2. I know that all my moral beliefs are false and that I ought to still keep them for good reasons. Can I? It’s hard to see how I could. I often try to make myself have useful make-beliefs but rarely succeed. How nice would it be to believe that one was a happy, nice and succesful person.
    Once one really knows that something is false, it just seems too hard to believe it. This much is almost built into what beliefs are – states that aim at truth. Deliberate self-deception is an art few are capable of at least on this scale – we seem to lack a free will in adopting beliefs. As John Heil and David Owens put it ‘We are largely at the mercy of our belief-forming equipment’ and ‘In the end, it is the world that determines what I believe, not me… the evidence takes over and I loose control’ (see Watson 2004, ch.5). But it is just this kind of mass-self-deception the fictionalist view on the table is composed of. If what it requires is something even a willing agent cannot deliver, then so much worse for its requirements. The consequence seems to be either empty howling on us about what we ought to do or collapse into error theory.

57. Sobel,
    Nolan, Restall, and West argue that judgement internalism is consistent with moral fictionalism. We can describe a moral fiction F such that, if the assertability of moral claims were governed by F, then a moral claim would be assertable only by someone who had the corresponding motivation. (See p. 26 of their paper.) Perhaps, then, your question is whether moral fictionalists recommend that we pretend that a specifically internalist moral fiction is true. As far as I can tell, Nolan, Restall, and West don’t address this question explicitly. But it might be the case that an internalist fiction would do a better job than an externalist one at furthering the kinds of interests they discuss — e.g. co-ordination.

58. Jussi-
    “Once one really knows that something is false, it just seems too hard to believe it. This much is almost built into what beliefs are – states that aim at truth. Deliberate self-deception is an art few are capable of at least on this scale – we seem to lack a free will in adopting beliefs”
    Again, I doubt I’m ultimately sympathetic to MF, but on this point, why insist that the self-deception be deliberate or require an exercise of free will? Patients with monothematic delusions (at least to my understanding) often hold delusional beliefs that they cannot be convinced out of and which appear to exert significant influence on their behavior and yet they sincerely affirm that those beliefs are bizarre or delusional. I’m not sure if this is categorizable as self-deception exactly, but it’s an example of acknowledged but unreconciled doxastic contradiction that doesn’t result from anything like deliberation or an exercise of will. (It is also consistent with judgment internalism in a funny way.) So perhaps the MF retort: we’re all deluded about the epistemic or metaphysical status of our moral judgments, we know we are, and we can’t do a thing about it!

59. Jussi,
    On my reading, moral fictionalism as proposed by Nolan/Restall/West does not include your thesis 2. They suggest that moral discourse may be worth retaining despite its being false. But that’s not to say that we should believe moral claims, only that we should talk as if do.

60. Related to several people’s comments about both the pragmatic reasons to retain the fiction and the issue of internalism I think the following is worth considering:
    A person may have commitments to the content of a certain moral view and the reason they have to continue acting as if they believe the fiction can be that by doing so they may get others to go along with them.
    A lot here turns on what makes the fiction false. If it is that not everyone has reason to do what morality suggests/demands, that does not mean that parties to the fictitious use of the story don’t have such reasons. But they need not be instrumental to something outside of morality. They can stem from their commitments to what in the fiction is the content of morality.
    Again, I have no idea if this fits what N/R/G say, but it fits with at least the motivations for Kalderon’s sort of fictionalism where it is the alleged fact that moral disagreements are irresolvable that leads to the idea that the moral story is false.

61. Michael,
    there might be something to that. But see the situation is that the fictionalist *believes* apparently that she has sufficient grounds for regarding all moral beliefs as false. And she knows that she would need them for prudential reasons. So she does seem to need some way for retaining the beliefs against the overwhelming contrary evidence she explicitly thinks she has. You are right that she could let this happen undeliberately. But I take it that in other contexts that happens quite rarely, and so we can ask why should we expect it to be more common with regards to these beliefs. Usually when we have enough contrary evidence our beliefs disappear. It seems a far shot that when it comes to morality we could all actually suffer from monothematic delusions. If chance would not be on our side in this issue, then we would need to knowingly get ourselves to have deluded ‘beliefs’ and that’s the hard bit. Another issue is that if it turns out that our moral opinions are not generally responsive to evidence, then this just proves that they are not beliefs which by definition are.
    Campbell,
    That’s the puzzle. NRW first say that ‘moral fictionalism earns its cognitivists credentials in two ways: first, it retains the idea that moral claims are truth-apt (it is just that all positive ones are false); and second, it takes moral claims to express genuine, full-blooded beliefs (NRW 2005, 320).’ But then they go on to note that the plausible form of fictionalism says that ‘we should employ a moral fiction and continue to assert apparently morally committed claims while withholding belief from them (NRW 2005, 322).’ These just seem directly contradictory. If we withhold belief from our moral claims as we should, then they do not express genuine, full-blooded beliefs and fictionalism does not earn its cognitivists credentials. You cannot both have your cake and eat it too. I in that argument, like Huemer and Chappell above (their argument was wholly formulated in terms of beliefs), was targeting the first view. But you are right, if the second claim is the account then both our arguments are missing the target. But that account seems to make many things mysterious. For one, on that view it would have to make sense to say that ‘Torture is bad but I don’t believe it’, and that just seems paradoxical in Moore’s way.
    Mark,
    Kalderon doesn’t state that moral story is false. That’s because he doesn’t think that it is a story *asserted* at all *about* our world, which it would fail then to be true about. And, his ground for going fictionalist is not so much the existence irresolvable moral disagreements but rather that in disagreements about moral reasons where the participants are just as informed and coherent one doesn’t have epistemic obligations to inquire further into one’s grounds of acceptance of the relevant claims. He argues that such obligations are present in all cognitive discourses. So, it’s not so much the disagreements (that can be found in all inquiries) but rather how we as a moral community epistemically react to disagreement. I think this is much better argument than Mackies’ old one from relativism but I still think the argument fails.
    But I think you and Campbell are right about internalism not being a problem for fictionalists. Surely the moral fiction would include claims about moral reasons. When looking from the perspective being midst the fiction the (fictitious) moral properties would be (fictitiously) reason-providing. And, surely the same mechanism that is used to explain how our judgments about reasons motive us can be used to explain how our judgments about reasons motive us when these judgments are about fictitious reasons.

62. Tim C,
    Your distinction between subjective & objective justification is interesting. (I’d call them internalist and externalist justification, alluding to the dispute in epistemology.) I think we agree on this: Something’s seeming to S to be the case means that S is prima facie (internally) justified in believing it. It does not follow that S has external justification. To get the latter, S also has to actually be reliable.
    Now, does Moore refute skepticism? I still think the answer is yes. For, first, he provides internal justification for denying skepticism. And second, you’ve agreed that, provided that Moore was reliable in his belief formation, he also provided external justification for denying skepticism. But surely Moore was reliable in his method of belief-formation. I haven’t heard anything about Moore having bad eyesight or being subject to delusions, so presumably he was in fact reliable when it came to identifying hands and pencils.
    Btw, thanks for the nice words about my book. I hope Stratton-Lake likes it.
    I am so happy that I have succeeded in turning PEA Soup into an epistemology blog.

63. Jussi,
    The aim of philosophy:
    I don’t think the goal of philosophical discourse is to change the minds of your opponents. If it is, then philosophical discourse is clearly practically irrational, since the evidence that it does not in fact accomplish that goal is overwhelming. Anyone who thinks that either I am going to convince Richard Joyce to be an intuitionist or he’s going to turn me into a moral skeptic, would have to be either naive or delusional.
    Then what is the purpose of philosophical discourse? One purpose is to better understand our own and each other’s positions. Another is to influence people who haven’t already made up their minds.
    Of course you’re not going to convince a die-hard proponent of the diametrically opposite position, who has already spent years thinking about how to defend his position and has already rejected all the objections to it that have appeared in the last 100 years’ literature. The fact that that guy wouldn’t be convinced doesn’t mean your arguments aren’t reasonable.
    Moore’s argument:
    Moore has two arguments, one first-order and one second-order. The first-order argument was something like this:
    1. I know this is a pencil.
    2. If [Hume’s premises], then no one knows any physical fact.
    3. Therefore, ~[Hume’s premises].
    where [Hume’s premises] is a conjunction of claims that Hume uses to support skepticism (as I recall, they’re roughly these: “No one is ever directly aware of a physical fact”; “One is indirectly aware of something only if one is directly aware of something that one knows is a sign of it”; “One knows that A is a sign of B only if one has been directly aware of things of each type previously”).
    The second-order argument is epistemological/methodological, something like this (with a lot of reconstruction):
    1. Upon discovering an inconsistency between P and Q, it is irrational to reject the more plausible of the two while retaining the less.
    2. Common sense propositions like “I know this is a pencil” are more initially plausible than skeptical premises like [Hume’s premises].
    3. So it is irrational to reject common sense propositions on the basis of skeptical premises.
    This argument might not convince Peter Unger; again, that doesn’t mean it’s not good. It also won’t change the mind of anyone who already believes (3). That doesn’t make it a bad argument either. Anyone who already believes (3) probably believes it, at least implicitly, because of this argument. Anyone who disbelieves (3) probably disbelieves it because they’ve already rejected this argument. So that’s why those two classes of people would be unconvinced. But that doesn’t vitiate the argument — the fact that some of your audience has already either accepted or rejected an argument doesn’t make the argument bad.
    Reliability:
    Reliabilism and externalism are standardly taken as anti-skeptical positions. Skeptics do not say that our belief-forming methods are unreliable. Of course, if you were a brain in a vat, you’d be unreliable, but skeptics don’t think that you are a brain in a vat! Rather, they say that you fail some internalist justification condition for knowing claims about the external world (very roughly, because you don’t have a noncircular argument that you’re not a BIV or that your senses are reliable, etc.).
    The tracking account of knowledge is likewise an anti-skeptical weapon. No one thinks we fail the tracking condition for things like “This is a pencil” (even though we fail it for “I’m not a BIV”). If you want to apply tracking to moral claims, an analysis analogous to what Nozick says would be: I know that torture is wrong (in part) because in the nearest world in which it isn’t wrong, I don’t believe it is; this would be a world, say, in which most people had strong desires to be tortured, in which it was necessary for mental development, etc. Since he denies Closure, Nozick would probably say we don’t need to know MF is false to know that torture is wrong.
    Is there a skeptical/fictionalist challenge?
    You suggest that if there is a legitimate challenge, then the Moorean approach doesn’t answer it; while if there isn’t a legitimate challenge, then we don’t need to say anything. There’s another possibility: that the Moorean argument explains why there is no legitimate challenge.
    My own view is that skepticism and fictionalism are totally unreasonable positions. An argument should not be judged by what someone with an unreasonable position would say, so the rationality of Moore’s argument doesn’t depend on what an MFer would say. But it also isn’t merely unnecessary to give the Moorean argument, since it explains to people who might not already know this why I think MF is unreasonable.
    The Moore-Hume symmetry
    It might be worth stressing that, logically, Moore and Hume are in symmetrical positions with respect to each other. Moore says “I know this is a pencil; therefore, ~[Hume’s premises]”. Hume says, “[Hume’s premises]; therefore, you don’t know that is a pencil.” Schematically:
    Moore: A. Therefore, ~B.
    Hume: B. Therefore, ~A.
    If Moore is “begging the question”, then Hume is doing at least as bad (worse, since he’s also saying something absurd).

64. Josh,
    On one point: could Mackie’s queerness argument defeat our intuitive justification for moral beliefs? No, because the queerness argument is no good. Briefly, “x is weird” does not seem to count as evidence that x doesn’t exist. It’s not clear that “x is weird” is even a cognitive remark.
    The Moorean argument of course doesn’t show that moral beliefs are indefeasible. It’s just supposed to show that the MF/ETer would need more powerful arguments than the ones they have so far given. I actually think the Moorean point is trivial.

65. Michael,
    a big thank you for some very helpful comments indeed.
    I agree that the aim of philosophy is not to actually change anyone’s minds. But I do think that arguments that go towards this are more interesting than ones that do not. So I do think finding internal problems from others’ views and arguments for our own views that are based on premises that can be accepted as widely as possible and which are less closer to being reformulations of our own views are to be prefered. I think the ability to defend one’s views, as rational support for them, is on par in importance with self-understanding.
    I might still have one question. I still wonder about the symmetry thesis at least in the moral case (without going into Hume exegesis). I take it that many moral fictionalists are not eliminativists or revisionists. What they are trying to do is to account for all features of our moral practices in a way that makes fewer ontological assumptions. If you can explain all the same things then Occam’s razor seems like a good way to go.
    It looks like the gist of the argument above is that moral knowledge is an ineliminable part of our practice that they couldn’t account for. If moral fictionalists regard all first-order moral claims as false, then we cannot have positive moral knowledge (I know that torture is wrong). Well, many of us above seem to think that this misrepresents the fictionalist view. I can see no reason to think that the fictionalist view implies that I could not know that torture is bad. I can give plenty of reasons for why torture is bad, usually my badness judgments fit well those of others around me, and so on, – all of this seems to be compatible with the idea that I’m engaged with a fiction. We can use the assertibility norms internal to our practice to account for a minimalist truth notion within our moral fiction, and give for instance a reliabilist account of knowledge.
    At this point it looks like to me that, if the fictionalist can accept that one can know that torture is bad, i.e., denial of this does not follow from her premises, then the symmetry you refer to seems to be broken. If we can give a fictionalist metaethical account of moral knowledge in our practice, then it seems to be the task of the realist to either explain why the fictionalist reconstruction of moral knowledge is inadequate or what else there is in our practice that cannot be accounted for without positing more extensive ontology with REAL MORAL FACTS AND BELIEFS. At least for me metaphysically sparser views have an edge by default.
    Maybe the result of this is that if one is fictionalist the Kalderon line seems much more promising than the NRW line.

66. Call me naive, but I do think the point of writing philosophy is to change people’s minds. My own views in philosophy have changed hugely in (e.g.) the last 10 years, for two main reasons: (1) my own thinking and(2) other people’s writings have shown me holes in positions I previously thought correct, and fruitful philosophical possibilities that I hadn’t even noticed or explored. Given that I’m still engaging closely with Mike H’s book, this is happening to me right now!
    Now my obstinacy is legendary among those who know me. So if such change happens to me, a fortiori it happens to other philosophers.
    I have to admit, mind you, that on the deepest issues, e.g. theism and moral realism, I haven’t changed my mind; I’ve just found more sophisticated (or less unsophisticated?) ways of being a theist and a realist.
    But enough of me talking about myself all the time… Why don’t some of you other folks talk about me for a bit? 😉

67. Hi all,
    It was previously hoped that a lurking fictionalist would speak up. I think I might fit that description.
    By way of introduction, I’ll just say that I agree with Jussi Suikkanen’s suggestions on how MFers would be wise to progress. For reasons close to those set forth by Nadeem Hussain in a recent article, I just don’t think it’s the job of any metaethicist to tell people their first-order moral claims are false. It is the job of the metaethicist to say something about what those first-order moral claims are doing, what makes them true if T-apt, etc. On this view, hermeneutic MF is the only MF metaethic, and I agree.
    I find hermeneutic MF attractive in part because it seems possible to trace some moral idea’s movement from fiction to supposed real property (I have rights in mind, but sins, duties, and rules might also be instances).
    But I’m not totally on board just yet. My primary concern right now is that morality, unlike possible worlds talk or math, lacks any structure for which to guide debate within the fiction. I have the Horgan/Timmons Moral Twin Earth intuitions: I accept that anyone is competent with the term—our term— “morally good” if they are guided merely by the non-substantial platitudes of our use. So if someone can be competent with the term “morally good” when they claim that torturing babies for fun is morally good, then the moral fiction would never have been of any use. No thin moral term would, probably. And if the fiction wasn’t of any use, then it wouldn’t have lasted this long. So the argument:
    1. A fiction (of the sort morality would have to be) must be useful (e.g., for guiding action, ordering groups, limiting discussion, etc.)
    2. For a fiction to be useful, rules governing the use of language based on it must be more rule-governed than moral language is.
    3. So morality isn’t fictional.
    Any comments on this argument would be appreciated by a humble grad student. Right now, I wonder if the MF answer to this challenge might be that morality is a fictional determinable, and that the morally thick concepts pick out moral determinates of (as opposed to species of) the fiction of morality—and these determinates do the heavy lifting of guiding behavior. If so, then maybe the Moral Twin Earth intuition could be accommodated for the same reasons that some feel there could be more determinate colors, while at the same time allowing that the determinate colors of our experience govern our color-talk. (this is inchoate, obviously)

68. Hi Dan,
    great to get you taking part on this blog. No need to be a *humble* grad student. To let you in on a secret I’m one too (there goes my credibility if there was any). I think we need more grad students taking part in these discussions – it’s great fun and you learn a lot. So, any lurking grad students out there feel encouraged to take part. I hope the bloggers of this board share my feelings here and not get annoyed by us.
    To your question. First, I’m not sure we get those intuitions from the Timmons-Horgan thought-experiments. Sure we get the intuition that a competent user of ‘morally good’ does not have to be guided by norms that are so substantial that they fix the term to a particular natural kind. But, I don’t think that it follows from that all that is required from them is familiarity with purely non-substantial platitudes (ah – the strange Australian animals…).
    I mean look at it this way. Say you went to a Twin-Earth on which the natives would not speak English and so you needed to start radical translation process. Eventually you find a network of concepts from their language which inferential connections would correspond our thin moral terminology, and this network was related to actions and blaming practices in the same way as ours. But, where ‘morally good’ is on Earth paradigmatically attached to actions like helping the elderly, keeping promises even when burdensome, and so on, their corresponging term in the network would be attached to torture, boiling and eating babies, and so on. And, you can imagine the same shift for all the other thin notions. Now, the question is would you be pushed to translate that terminology to our moral language. My intuition is that you wouldn’t. I think Philippa Foot was right that there really are substantial limits on what counts as use of *moral* concepts. Anyway, that’s my humble grad student reply.

69. Jussi: I disagree. I would translate their word as “good”—that is, I would say that the Twin-Earthers think torturing babies and so forth are good.
    Isn’t that really our intuition, Star Trek watchers? The universal translator translates Klingons as saying that dying in battle is good, Harojan hunters that killing persons in hunts is good, and so forth. I am not aware of any Star Trek viewers who balk at this use.
    If members of these races felt committed to do the acts they felt to be “good”, if they allowed that “goodness” supervenes on the descriptive facts, etc. then it looks like they are expressing our concept, GOOD.
    Of course we may disagree with them about what falls under the concept, but we wouldn’t say these races are incompetent with the term.
    These considerations motivate my premise (2). Deconstruction aside, you cannot say Hamlet is an unreflective fellow. The play _Hamlet_ limits what we can say about Hamlet. The axioms of mathematics limit what can be truly said in the mathematical fiction (if it’s a fiction). But morality? The thin moral terms just don’t limit moral discourse enough for them to ground a fun game. It’s a little like you shoot your brother with a death-ray, and he says that he survived because he had his death-ray scrambler on, and you respond that you were actually using your next-generation scrambler-blocker, and so on. That game isn’t fun or useful for guiding our behavior. If morality is like that, then it couldn’t be fictional (absent something more being said—I floated one possible strategy for a reply).

70. Jussi,
    I agree with you that, if the MF allows moral statements to be true, or at least non-false, then the Moorean objection doesn’t apply.
    However, I think if the MF lets moral claims be true in the way we might think that “Superman was vulnerable to Kryptonite” is true, then MF is really (just) a variant of subjectivism/relativism, and then it’s going to be vulnerable to the standard objections to that, e.g.,
    “For any x, if society says x is right, then x is ipso facto right.”
    That’s false, but relativism implies that it should come out true, and I think the same holds for an MF that allows moral statements to be true. The MF will be saying that moral statements are true as long as they are true according to the fiction, which means they’re true, basically, when and only when society says they are.

71. Dan,
    Good point. I’m still not sure though, but maybe it is a case of differing intuitions. For one you switched from talking about ‘moral good’ to plain ‘good’. And, most Trekkies think that teletransportation is a way of travelling too so I wouldn’t rely on their intuitions.
    Michael,
    I’m probably not as worried about MF collapsing into subjectivism or relativism even though I’m very opposed to those views. Infallibility is one difference. In those views it is must be the case that such and such is wrong if I or the community disapproves of it. But, I or the community can be wrong about what is wrong according to the fiction they are engaged with.

72. Michael H., you say, “could Mackie’s queerness argument defeat our intuitive justification for moral beliefs? No, because the queerness argument is no good. Briefly, “x is weird” does not seem to count as evidence that x doesn’t exist. It’s not clear that “x is weird” is even a cognitive remark.”
    Okay, let’s grant that the QA doesn’t work. That doesn’t mean it couldn’t work, or couldn’t have worked, right? (Or am I confused about the modal status of arguments?) Nor does it mean, of course, that some other anti-realist argument doesn’t work. This is relevant because of Tim’s premise (2) above (here’s the address of his original comment way up there: http://peasoup.typepad.com/peasoup/2006/01/moral_fictional.html#comment-13563024):
    “2. We have very strong reason to believe the proposition that torture and rape are seriously bad.”
    Unless I mistracked the dialectic (not unlikely!), I thought that the evidence for this is intuition. And thus the argument becomes one of burden-shifting. As Tim wrote: “Responses to this argument may well target premiss (2), and claim that this question-beggingly presupposes that ordinary belief in the badness of rape and torture is realist belief. To which I reply that there’s nothing question-begging about that presupposition. The onus is on the fictionalist to explain why we should move away from it.”
    All I meant to suggest was that the fictionalist does explain why we should move away from it. The example I used was that NRW lean on the the QA, but if memory serves, folks above have suggested that Kalderon offers a different anti-realist argument. So, it’s not quite right to say that the onus is on MF. Rather, the onus is on realist anti-MF to explain why MF’s anti-realism is misplaced. Until then, Tim’s premise (2) isn’t a given.

73. Mike writes:
    “For any x, if society says x is right, then x is ipso facto right.”
    That’s false, but relativism implies that it should come out true.
    Mike, that seems to quick.
    Other than the relativists in intro ethics books (often included so as to refute relativism) I don’t know any relativists who actually believe this.
    Most philosophical relativists are not society relativists, since the same arguments that would support relativism at that level would also support relativism of a more fine-grained sort. And most relativists that I know of have done quite a lot to work into their proposals ways that a given person or group’s moral judgements can be wrong. For example, they say that what is morally required for X is what X would decide to do under conditions of full information and coherence (perhaps starting from his or her current set of moral commitments). By building in some idealization requirement, they make it the case that actual judgements are wrong. But so long as they think that the idealization process doesn’t eliminate all disagreement, the theory will generate relativity to persons (or groups of persons) of some sort.
    I’m pretty sure that in a similar fashion fictionalists can think that not just any interpretation of how the fiction goes corresponds to a truth about the fiction or even a truth for that person about the fiction.

74. Good points, Mark. For the sophisticated relativist you describe — I’d prefer to call him an ‘ideal observer theorist’ — I want to pose a dilemma. Either the ideal deliberation requirement lets the person keep their current (basic) desires and emotions, or it does not.
    If it does, then I claim you’re still going to get undesirable consequences in the neighborhood of undergraduate-relativism. This is because the outcome of one’s deliberation is going to be informed by whatever irrational prejudices a person has (apart from false beliefs). For instance, if someone has negative emotional reactions to homosexuals–they make him feel bad, regardless of his beliefs–then that person’s ‘correct’ moral code is going to include anti-homosexual norms. If a racist had racist feelings ingrained in him through childhood training, such that they’re now just part of his basic emotions, then his ‘correct’ moral code will include racist norms. And so on.
    And there’s really no reason to assume that more non-moral information or better logic would change much of people’s moral atitudes; no reason to think it would have made slave-holders in the 19th century convert to abolitionism, etc.
    On the other hand, if ideal deliberation does not let you keep your current basic desires and emotions, then you end up with nihilism, at least according to the theory, since desires and emotions are the only ultimate source of evaluations.
    Or suppose you let people have some basic desires and not others. Then you’re in the business of stipulating that some desires are legitimate sources of value and others are not, and that’s going to seem arbitrary (for a non-realist). Then I’m basically going to argue that you can’t motivate any restrictions on basic desires without appealing to some objective values.
    Suppose you try to restrict the class of desires and emotions by reference to how they were acquired — e.g., desires caused by ‘brainwashing’ are not going to count towards what is morally right for you. I’d argue that that’s just arbitrary. From a broadly Humean point of view, desires instilled by ‘brainwashing’ (where these desires do not require current false beliefs) are no less legitimate motivators than any other source of desire. Brainwashing is no less good a way of acquiring a basic desire (on an anti-realist view) than having certain genes, or certain chemical reactions in your brain, or however other desires are usually acquired.

75. Mike,
    I’m not really the one to carry on the defense of relativism here, since my views are only relativist around the edges if at all. But a relativist theory need not say that desires are the only sources of value. It need only say that what is right for you is determined by what rational revision of your beliefs and commitments would commit you too. And then whatever principles of rational belief and commitment revision the theorist accepts has to be one that will have different outputs for depending on differing starting points, at least for some different starting points. In principle, it seems to me, that a cognitivist about morals with a theory on which belief revision was always a matter of coherence could be this sort of relativist, though actual relativists seem also to include non-cognitive commitments as among the inputs and outputs.
    But again, I’m not the person to defend relativism here.

76. Thanks, Mark. Maybe some relativist or subjectivist will chime in. Anyway, here would be my argument about the source of value. I think the states that might give rise to evaluations are:
    a. Non-cognitive states. I think the salient ones would be desires, emotions, and pleasant and painful sensations.
    b. Cognitive states. These could be classified according to the objects of the cognition.
    b1. Cognitions of psychological facts about the agent.
    b2. Cognitions of psychological facts about others.
    b3. Cognitions of physical facts.
    b4. Cognitions of non-natural facts.
    I think this exhausts the relevant alternatives (but someone please tell me if I’m overlooking something). I don’t want to get into details now, but basically I think b2 and b3 are implausible, and anti-realists will reject b4. So only a and b1 are left. Under b1, I think it will only be plausible that cognitions of psychological facts that entail something about the non-cognitive states of the agent will lead to evaluations. (Roughly, if the occurrence of other kinds of states doesn’t itself support some evaluation, then second-order awareness of facts about those states probably doesn’t support an evaluation either.)

77. Mike,
    Can you tell me what a source of value is? I can think of several different interpretations, but I don’t know whether any of them are the intended one. And why are we limiting the candidates to states of persons, cognitive and non, as opposed to say states of the world? I keep thinking that someone like Moore might say that there are many several different sources of value, including beauty which he seems to think retains its value whether or not people know of it. Since the list is supposed to include implausible as well as plausible views about sources of value, I’m wondering why this view does not seem to be up there.
    But maybe this is a topic for a new thread . . .

78. Sorry, I meant sources of evaluations (or valuings)–things that might lead the ideal observer to make an evaluation.
    I think Moore’s view is perfectly plausible. But beauty as such still couldn’t be a source of evaluation, because no one would evaluate something as good (even if it were in fact beautiful) unless they had the appropriate mental state (e.g., an aesthetic experience or some other awareness of beauty).
    I don’t mean to deny, though, that external facts can cause appropriate mental states (e.g., aesthetic experiences), and thereby indirectly cause evaluations. But I just want to stipulatively call the mental state that’s involved the ‘source of the evaluation’, rather than any external state or fact.

79. I’d like to thank Joshua for the discussion of our paper – there’s a lot of interesting things in the comments, but I realised if I wait to write a comprehensive comment I won’t ever get it written. So I’ll just comment on one or two things from the original post.
    Joshua says:
    I suggest that MF similarly accounts for the truth-aptness of moral claims in the wrong kind of way (though not because the truth conditions it posits are subjective). When the ordinary person ordinarily claims ‘Executing murderers is wrong,’ the truth of this claim is not supposed to simply be a matter of whether wrongness is real. It is supposed to be a matter of whether wrongness attaches to the execution of murderers. That is, whether or not executing murderers is wrong is supposed to depend on something about both executing murderers and wrongness—does the nature of wrongness line up with the nature of executing murderers?
    This ‘binary’ nature of the truth conditions in ordinary moral claims is not capturable by MF, because MF renders the truth condition for moral claims ‘singular’—something to be settled solely by the existence of prescriptive entities. Once it has been determined that there are no such entities, whether executing murderers has the properties upon which wrongness supervenes is irrelevant to the truth of whether executing murderers is wrong. That is, MF renders the non-moral facts irrelevant in determining the truth-value of moral claims; it makes moral claims truth-apt, but in the wrong kind of way. Like quasi-realism, the truth-aptness it accounts for seems to abandon a part of ordinary moral discourse.
    I think the truth-value of the claim “executing murders is wrong” depends on whether there is such a thing as wrongness and whether it attaches to the execution of murderers. I think the MFer should agree. It’s just that if there is no wrongness, it follows that it does not attach to anything, including the execution of murderers. (Let me assume that “murderer” is merely descriptive here – if you’re unhappy with that, change the example to a case where what is being claimed is wrong is described entirely descriptively.)
    Obviously the mere existence of “wrongness” (or a positive extension to the predicate, to be less ontological) wouldn’t be enough – it would need to attach to the executions. So the MFer agrees that the truth-condition is “binary” – it’s just that, at least when we take the claim literally, the first component needs to be satisfied for the second to be. Since the first isn’t, neither is.
    This helps us see what the MFer might say about Joshua’s
    “it is curious in light of the supposition that ordinary moral discourse suggests that not all moral claims have the same truth value (and this is a front on which quasi-realism does not come up short). Or maybe there’s a way around this objection for MF?”
    It’ll depend on the MFer – my favoured variant is one who admits that a lot of people running around are realists of some sort, and so they don’t treat the positive claims as all having the same truth value because they’ve made a mistake. For the MFer that thinks we are already, by and large, treating moral talk as fictional, he also has a response: according to the fiction, the claims don’t all have the same truth value, so we talk as if they don’t. That’s why ordinary moral discourse suggests that not all of these claims have the same truth-value: it’s broadly the same kind of reason that Lord of the Rings discourse suggests that “Bilbo is a hobbit” and “Aragorn is a hobbit” don’t have the same truth value.

80. I’ve removed Robert’s earlier comment here–it has been relocated in the What Is A Categorical Imperative discussion

81. Daniel, thanks for the comment. I like your account of how the truth conditions for moral claims are binary for MF, so let me back away from that diagnosis of why MF fails to accommodate the truth-aptness in moral claims.
    That said, I still suspect that MF has problems accounting for truth-aptness (however we diagnose the problems). With error theoretic revolutionary/prescriptive fictionalism (with which I was originally concerned), it appears that MF just inherits the problems error theory faces on this front, by, say, being unable to explain how a moral proposition and its opposite (i.e., p and ~p) can have opposite (literal) truth values. With respect to descriptive/hermeneutic fictionalism, if we don’t literally believe in and assert moral claims, we couldn’t have moral beliefs and assertions that are truth-apt. (This is not to say that we couldn’t have beliefs and assertions about what happens in a story about morality, like we do about Bilbo).
    If that’s right, I still want to find a diagnosis, though, again, perhaps the binary-truth-condition approach is the wrong way to go. Note, also, that this isn’t to say that MF can’t say some things to explain and perhaps debunk the platitude that first-order moral claims are truth apt (e.g., by saying we don’t really believe and assert them). It’s just to say that MF can’t accommodate that platitude.

82. Joshua,
    I guess I’m still not getting why moral fictionalists don’t accommodate a platitude associated with truth-aptness. As we describe moral fictionalists, at any rate, they do (or can) believe exactly one of p and ~p is true – we draw a distinction between “positive” and “negative” moral claims. Drawing that in some cases might be tricky, but e.g. we might say that a negative moral claim is compatible with the claim that there are no moral obligations, and also the claim that nothing has moral value. (That’s probably not quite enough, but I hope you get the idea.) After all, if “killing the innocent is wrong” is false, it doesn’t seem to require much more for “it is not wrong to kill the innocent” to be true. Of course the prescriptive moral fictionalist probably does disagree to a great extent with ordinary moral commonsense about which moral claims are true and which are false, but that doesn’t seem to me to be a problem connected with allowing they are truth-apt.
    As for the “hermeneutic” MFer, I don’t see why they can’t, or wouldn’t allow that we literally have moral beliefs. At least a reflective hermeneutic fictionalist will probably believe most of the negative moral claims, and disbelieve the positive ones. Whether the ordinary folk are very opinionated about the literal truth-values of moral claims could go either way, I suppose, but I wouldn’t be surprised if descriptive moral fictionalists would want to hold that we implicitly disbelieve the positive ones and believe the negative ones.
    Finally, what should the descriptive MFer say about whether we have truth-apt moral assertions in everyday talk? Here I at least think it’s going to depend on the details of what the best theory is of this sort of “fictional assertion”. According to some people this sort of talk is only “pretend assertion”, and not an assertion at all. According to others it is an assertion which, in ordinary contexts, has an implicit “according to the fiction” operator (e.g. David Lewis). According to still others it is an assertion with a truth-condition quite different from the truth-conditions of such an assertion produced literally (see e.g. Yablo). A lot of these views say that the things people say when they’re engaging in that assertion-like talk are assertions.

83. Daniel,
    Yeah, I find the positive/negative claim distinction a little tricky. I believe, though, that it’s similar to what I’m thinking of as the first-order/higher-order moral claims distinction. So, e.g., “It’s not the case that killing innocents is wrong” is higher-order. Also, claims like, “There are no moral properties” are non-normative, and so outside the scope of the truth-aptness platitude I was thinking of.
    If that’s similar to what you have in mind, my claim could be recast as saying that MF can’t preserve the truth-aptness of positive (or first-order) normative moral claims. Here’s where you and I might have been construing p & ~p differently. If “Killing is not wrong” is the ~p to the p of “Killing is wrong”, we can read the former in two ways. It could mean, “It is not the case that killing is wrong,” in which case I would agree with you that this statement is true for MF, but it’s also not a first-order moral claim, so it lies outside the scope of the question about whether MF can account for the truth-aptness of first-order moral claims. If, instead, ~p is “Killing is permissible,” then this is a first-order moral claim, and so it is within the scope of the question, but it is false for MF. And it seems like MF must say that all first-order moral claims are in this way false (or at least not true).
    On the belief front, I was in my last comment (i.e., now that I’ve read his book, as opposed to before, when I hadn’t!) taking Kalderon as sort of authoritative here, which might not be fair. But, in any case, it’s hard to see how the descriptive fictionalist can say that we believe first-order/positive moral claims. Just as I when I utter “Superman is allergic to kryptonite,” I don’t believe it literally (because I don’t believe in Superman), so for MF I don’t literally believe “Killing innocents is wrong.” But, here again, if we’re talking instead about higher-order/negative moral claims, I agree with you. I’m taking the relevant platitude about truth-aptness to only be about first-order moral claims. (Not to deny that there’s also another platitude about higher-order moral claims being truth-apt.)
    Something similar strikes me as true about assertion (and, again, I’ve perhaps unfairly been under the spell of Kalderon lately). While I’m kind of partial to Joyce’s argument against Lewis’ tacit prefix view and for pretend assertion (possibly consistent with Lewis’ preface, rather than prefix, view), and I haven’t read Yablo on this, it still seems to me that all of the views you mention must deny that, for MF, when we utter first-order, normative moral claims, we are asserting them as such. The first view says we are pretend (i.e., not) asserting them; the prefix view says we are asserting something about a fiction, not about morality, and so we’re not really uttering normative claims at all; and, while again I’m not directly familiar with Yablo’s view here, it seems like fictionalism has to say something different than that we are straight-up asserting first-order, normative moral claims, or else it’s hard to see (for me, anyway) how MF is allied with more ordinary forms of fictionalism (e.g., with respect to discourse about Superman). But I clearly need to get caught up on Yablo’s view, so maybe I’m going astray here.