How can you say that if you’re an error theorist?!

Now and then, when I’m involved in discussion of some question of normative ethics or the like, I’ll get a response along the lines of, “How can you say that when you’re an error theorist?!”

Note that large assumptions are being made here. One is that I am, in fact, an error theorist as that is understood in contemporary metaethics. In fact, I tend to use formulations such as that I think moral error theory “has a point”, or that it’s the standard metaethical position that I think is “closest to the truth”, or that I am “attracted” to moral error theory, etc. What I try to avoid doing, though I don’t say I’ve always succeeded (since it’s often necessary to take conversational shortcuts), is to say, outright, “I am a moral error theorist.”

That’s partly because moral error theory has come to mean something quite specific that is not necessarily what J.L. Mackie advocated in the first place, and I don’t necessarily buy a theory quite that specific even though I agree with 90 per cent of what I read in Mackie’s Ethics: Inventing Right and Wrong. I actually prefer to call myself a moral sceptic (or “skeptic” if you prefer), which is a vaguer term that can cover a range of positions.

The difficulty here is that moral error theory has come to mean the claim that all of our first-order moral judgments (or perhaps just a very large sub-set of our standard kinds of first-order moral judgments) are truth-apt but false. The usual way, though by no means the only way, that this result is derived is to begin with the claim that there are no objectively binding behavioural standards or objectively prescriptive moral properties. This is then combined with a semantic claim that first-order moral judgments purport to refer to such standards or properties. For example, it might be that “Torturing babies is morally wrong” means something like “Torturing babies is prohibited by an objectively binding behavioural standard.” Since no such objectively binding behavioural standards exist, “Torturing babies is morally wrong” turns out to be false – in much the same way that “Samantha is a (real) witch” will always turn out to be false because there are no (real) witches in the requisite sense (i.e., no women with supernatural powers, involvement with the devil, etc.).

Even if we think that there are no objectively binding behavioural standards, in the relevant sense, or objectively prescriptive moral properties, in the relevant sense, it does not follow that moral error theory is true. It would only follow that moral error theory is true if we accepted a moral semantics in which first-order moral judgments purport to refer to such non-existent standards, properties, etc. Perhaps we should accept such a moral semantics, but it might get complicated. And of course there are notoriously analyses of moral language that do not require any such semantics – non-cognitivist analyses, moral naturalist analyses, relativist analyses of various kinds, and doubtless others.

It’s also likely, I think, that our moral language is not monolithic and is not simple even in particular cases. For example, some of our moral language, but not all of it, might best be analysed along non-cognitivist lines. Some of it might be best analysed along moral naturalist lines – for example, if I say, “Torturing babies is cruel” I might be saying something that is quite true, and yet this is a moral judgment. Perhaps it combines a factual statement about the painful consequences of torturing babies with an expression of repugnance at the practice and/or a prescription that others avoid it. Moral judgments, particularly “thick” ones, but perhaps not only those, might have mixed content of some kind.

The point that I want to suggest at this stage is that scepticism about objectively binding behavioural standards, objectively prescriptive moral properties, and the like, need not cash out in the belief that first-order moral judgments are simply false, or that all of them are.

This can actually get very messy, and I don’t claim to have got to the bottom of it all. For what it’s worth, I do tend to think that at least some of our first-order moral judgments are, strictly speaking, false, for the sorts of reasons typically advanced by moral error theorists. But that is a long way from accepting moral error theory of the the simplistic kind that is usually portrayed in undergraduate philosophy courses or even in philosophy text books.

But let’s assume for the sake of arguments that all first-order moral judgments actually are false. Perhaps so! Does it follow that we should give up making such judgments? Not obviously. Take the judgment that torturing babies is morally wrong. If this means that torturing babies is forbidden by an objectively binding behavioural standard, and assuming there are no such standards, then, strictly speaking, the sentence is false. But there may well be – I’m sure there are – true statements in the vicinity.

For example, it might still be true that: “Torturing babies is forbidden by standards that everyone in this conversation accepts.” And/or it might still be true that: “Torturing babies is forbidden by standards that it would be prudent for me to follow as a package, to promote my own long-term self-interest.” Or it might still be true that “Torturing babies is forbidden by standards that it would be prudent for everyone involved in this conversation and everyone else in their societies to follow as a package, in order to produce mutual advantage.” Or it might still be true that “Torturing babies is forbidden by standards that I try to follow and invite you to follow.” And so on.

I’m not suggesting that “Torturing babies is morally wrong” means any of the things in the previous paragraph, though we could probably find theorists who would defend one or other of these meanings. Nonetheless, there may be a causal story as to why we make the moral judgments that we do, involving the truth of some of these and related propositions, even though the statements that we make when we make moral judgments are, strictly speaking, false. We actually do, for example, have moral standards, these are largely shared, and they are not entirely arbitrary. They may not be objectively binding on us, but they may well have personal (for long-term self-interest) and social benefit.

Moral error theorists don’t have to deny any of this. In which case, it’s not obvious that moral error theorists should advocate abolishing language such as “Torturing babies is morally wrong.”

Even if I were a full-blown textbook moral error theorist, with no misgivings about the theory at all, I might think that there is utility in continuing to employ this kind of language, even in my own self-talk, thereby buying into a useful fiction that torturing babies is forbidden by an objectively binding standard (not merely a personally or socially beneficial one).

Or I might think that there is benefit in going on using such language while having in mind something more like “Torturing babies is wrong by a standard that I accept and invite you to accept, and which I think you probably have good reasons to accept given your own values.” If she is open, in appropriate contexts, that this is what she has in mind, a moral error theorist might think that the meaning of such sentences will ultimately be revised – people generally, or at least those she is likely to be talking to, will eventually come to use the language in this revisionary way. After all, she might think, the real point (in some sense) of first-order moral language is to make judgments based on standards that are personally and socially beneficial, and it is not strictly necessary for us to rationalise these standards as also being objectively binding.

All that said, some moral error theorists – moral abolitionists – actually do think it is more beneficial to give up on making moral judgments, once we see through them, as it were. There might also be a partial abolitionist position that suggests that we stop making some kinds of moral judgments but not others – the complexities of moral semantics and our social situations might support some nuanced approach along these lines.

In the upshot, scepticism about such things as objectively binding moral standards (a scepticism that I definitely share) goes only part of the way toward moral abolitionism: advocacy of the total abolition of moral judgments. In my own case, I am certainly aware of these issues when writing about how we should behave, what dispositions of character are virtuous or vicious, etc., and the language that I use is, indeed, moulded to an extent by my tentative views about the issues I’ve discussed in this post. I engage, I suppose, in a mix of revisionism and partial abolitionism.

The point is simply that even a moral sceptic – indeed, even a textbook moral error theorist – can have plenty of reasons not to abandon moral talk entirely. To assume otherwise is to skate over a host of complex and controversial issues.

Leave a comment ?


  1. What I find frustrating about the discussion above (and all your previous posts on this topic) is that you always leave out performatives and declaratives. I think this is a peculiar omission. Any plausible form of moral veridicalism will exploit the fact that the reflective folk will use truth-talk to discuss statements about social constructs (like, say, money, or the state).

    I mean, why should anyone have any allegiance to the view that ‘truth’ is used to mean something like correct description or designation? The only attraction of the correspondence view, so far as I can see, is that it would satisfy the preferred idiom of the hard-headed realist. But it only makes the realist’s heart swell by ignoring the way informed and reflective people actually use words. And ignoring evidence undermines the scientific status of realism as a theory of language.

    Correct me if I’m wrong, but at the moment, the major clog appears to be over realist tendencies in the analytic philosophy of language.

    Reference has a central place in philosophical semantics. The idea that a semantic theory is based on a theory of reference is tenacious, and it has been supported by theories that are both sophisticated and insightful. And if we talked about it, it would involve digression into a high-tech discussion of truth-conditions in contemporary semantics.

    But I invite that discussion. It’s about time we figured out that the reference-fixation in the philosophy is language now supports an intellectual programme that is dogmatic, entrenched, and ultimately unscientific. Its only real virtue is that, as far as mature semantic theories go, it is the only game in town.

    And here is the alternative. The veridicalist will point out that the focal semantic point in all these cases is deferential, and not referential. And so, “Torturing babies is wrong” is best translated as, “Torturing babies is forbidden (by x)”, where x is a elliptical placeholder for an authority which may or may not require being filled in. Any further questions — e.g., “What authority?” — require no answer from the analyst. Inquiry into the nature of authority is part of our meta-semantics, not part of a semantic theory; the suggestion of deference, by itself, is part of the literal contents of what the sentence means.

    There are other things you can say, if one has the time, discipline, and inclination to pursue this project. Indeed, from what I can tell, something like this view is now being cashed out explicitly and formally in the journals. And some avenues by which we’d formally incorporate deference into a philosophical semantic theory are lying latent in the classics. (e.g., authority was central to Sidgwick’s conception of meta-ethics; deference to the initial baptism was part of Kripke’s fixing conditions in his causal-historical theory of naming; deference is involved in early-Putnam’s stuff on expertise and scientific authority; etc.)

    Whatever — I wish we’d just get on with it.

  2. The problem with the possible triumph of the Moral Error theorist in civilization would be that the loss of the artifice of moral authority would reveal the limits of the state to maintain even the pretense of an orderly society. The truth (and I believe this is a falsifiable, and therefore testable, statement) is the vast majority of the order of our society is derived from people believing there are objective standards to which we ought to be bound and by which they will live even when they don’t believe they are being monitored.

    So perhaps the dark-humorous corollary to Moral Error theory would be that, since Moral Error theory may be true, but morality is a useful fiction it would not be immoral to kill all the Moral Error theorists.

  3. Part of the problem identifying yourself as a supporter of one or another extreme position (and here I am using the word extreme to be indicative of what general commonsense people would say about moral and ethical values) is that by definition you force yourself into a corner.

    Truth is like a many faceted diamond. When I philosopher attempts to restrict themselves to only certain schools of thought, then the ability to describe certain facets may be enhanced, but unlikely the entire kernel of truth.

    If philosophy is the method of explaining commonsense and the arts of life, then to make it (philosophy) complex via language, is to make it irrelevant to all except the academics.

    Morality, ethics, right and wrong, courage, bravery, fidelity, etc., are all understood by the common man, therefore their existence is de facto.

  4. Claire Creffield

    I haven’t got any useful comments, but I just wanted to say I really enjoyed reading this lucid and thought-provoking piece. Re the comment directly above, to forbid complexity would be to forbid philosophy, I think.

  5. My own argument for moral error theory would concentrate more on the fact that moral claims are not rooted in epistemic success, and consequently there is no evidence for their truth (apart from the fact that people believe them).

    Our claims about the physical world are clearly rooted in epistemic success, particularly in science. They help us to predict, explain and control reality. The same is true for our mathematical claims (though we can extend mathematics in ways that need make no further contact with reality). Even pragmatic, non-moral “oughts”, which tell us how best to achieve our goals, help us to control reality (which is closely related to predicting how reality will respond to our actions).

    But nothing like this is true for moral claims. Moral claims are not in the business of doing any of these things. They are only in the business of motivating human actions. And for this they have no need of truth or epistemic success. All they need is for people to believe them. (Our tendency to believe them is most likely rooted in an adaptively evolved moral faculty. But, unlike our general cognitive faculties, our moral faculty didn’t evolve for truth-conduciveness, because our moral beliefs don’t need to be true.)

  6. OK, my previous comment didn’t address the main subject of the OP: how can a moral error theorist make moral claims? I’m tempted to answer: “by using his vocal chords, just like anyone else”.

    But really the point is that the moral error theorist may still find himself wanting to make moral claims. Speaking for myself, at an intellectual level I’m pretty convinced of moral error theory. But that doesn’t mean my moral faculty has stopped working. I still have moral values and attitudes at a non-verbal level. And even at a verbal level I might find moral judgements coming to mind. And in particularly stressful situations I might blurt them out without prior reflection. Besides that, I recognise the value of moral claims in influencing behaviour, often in ways I approve of, so I’m reluctant to give them up, and may still use them reflectively too. Yes, in that case I may be saying something I believe (at an intellectual level) to be false. It’s an uncomfortable position to be in. But truth isn’t always the most important consideration.

    Richard Joyce has argued that society could maintain an effective moral discourse even if moral error theory was generally accepted. He calls this view “moral fictionalism”. I think the idea is that we can be moral error theorists in our most critically aware moments, while the rest of the time having a moral realist frame of mind, allowing us to engage in moral discourse without experiencing a sense of a contradiction. Maybe, but speaking for myself I haven’t achieved such a comfortable reconciliation.

    Frankly, I feel a bit guilty about arguing for moral error theory, and potentially putting more people in this uncomfortable position. On the other hand, I assume philosophers (and the philosophically-inclined) want to know the truth, whatever the consequences. And I think an acceptance of moral error theory opens the way to progress in other areas of philosophy too. So, read my comments at your own peril. 😉

  7. Richard wrote:

    My own argument for moral error theory would concentrate more on the fact that moral claims are not rooted in epistemic success, and consequently there is no evidence for their truth (apart from the fact that people believe them).

    I like this passage, because I think it articulates a point that the veridicalist can cleanly deny. I am inclined to say that it is most certainly not a fact that moral claims are not rooted in epistemic success. The problem is with the error theorist’s mistaken stance about what counts as epistemic success.

  8. @BLS

    I see no mention of any evidence. How do you recognise epistemic success?

  9. P.S. To be more precise, what is the epistemic success in which you think moral claims are rooted? And how can we see it?

  10. Consider conventional performatives. The evidence in such cases usually has to deal with whether or not some act has actually been performed. e.g., the captain breaks the bottle against the ship’s hull and utters a name, and the ship is baptized; if asked for evidence that this happened, we look to the broken glass, and we ask the witnesses what they heard. The priest says to the man and woman, “I now declare you man and wife”; if we want evidence that this was done, we ask the witnesses. So we say, “It is true that Bob and Susie are now married”, or “It is true this ship’s name is ‘Bertha'”.

    Suppose, just for the sake of argument, that moral claims are performatives of a special kind, behaviors or actions with a certain special aura. From this point of view, any error theory that thinks that the truth-conditions of a performative sentence are based on descriptions of What’s Out There Independent of the Mind, is an error theory that is in error. (It is a reasonable error, given that folk conceptions of morality can be complete metaphysical trash — but it’s an error all the same.) This account may not be obviously true, but it is a rival account, and it deserves to be attended to.

    Even a moral veridicalist might think that there is something fishy about this aura which surrounds moral performatives. So long as this extra aura has not been explained, and it is not immediately obvious what distinguishes morality from mere conventions — then morality must collapse into relativism. So if you think that this extra aura is troubling hocus-pocus, then you can certainly be a skeptic about the objectivity of moral claims. But if you take the principle of charity seriously, then you can’t be a skeptic about the truth of moral claims. In a worst case scenario, moral claims end up being conventional performatives with extra pretentions.

    As a matter of fact, I do think that this aura can be explained, and that the epistemic objectivity of claims can be secured. I invite that discussion, and think you’re right to press the point. But I don’t think that discussion can be had so long as people insist that error theory is even a valid option in meta-ethics.

  11. Moral language (e.g. a good car!?!)* is a useful shorthand for us to enable conversation is anything like a normal fashion. But the Moral Error theorist, or moral sceptic, must always be prepared to stop the conversation and make the point clearly that there are no moral truths when the conversation gets to the point where there are moral differences and the shorthand obfuscates the discussion rather than illuminating it.

    FWIW moral truths exist, but they only exist at a given moment for a given person, and they may not even be accessible to that person… Neuroscience may show this more clearly in time.

    * The new Ferrari is not a ‘good’ car, but a comfortable car with better handling than most, enviable equipment, better acceleration, braking and top speed than most and, given the above, is relatively fuel efficient.

  12. Keddaw wrote:

    But the Moral Error theorist, or moral sceptic, must always be prepared to stop the conversation and make the point clearly that there are no moral truths when the conversation gets to the point where there are moral differences and the shorthand obfuscates the discussion rather than illuminating it.

    There are two points you’re making there, one about the objectivity of the claims, and one about the truth of the claims. But they are not the same. You can believe that claims about marriage or the law can be true, even while also believing these claims are relative and subjective to a society, or even to a small subset of powerful people within the society.

    e.g., if I pointed out to you that the law of the state is essentially and hotly contested, and you replied that the law is not strictly speaking true because it’s just a useful shorthand to help us speak in a normal fashion, then while that denial of legal truth would be motivated with the best of philosophical intentions, you’d be doing an injustice to the internal reason-guiding aspect of law (at best), and be in contempt of the law (at worst).

  13. BLS, what I was getting at is the fact that for the moral sceptic any moral claims are a tightly packaged ball of truth claims that the user may have no idea of what it contains. Each of these truth claims may be True, or they may be true to them, or they may be true on an agreed upon scale (e.g. does a Prius get ‘good’ mileage?). We can use this packaged claim to enable debate and allow various statements to go unchallenged or unqualified in that particular discussion.

    What I am saying is that should the general agreement on morals become iffy, or should the individual components of the moral claim be in dispute, that is when the moral sceptic should stop the conversation and focus/tease out what each of these truth claims are and see if any of them hold up to scrutiny.

    I don’t accept your ‘law’ example, not least because there’s a huge difference between saying the law doesn’t exist and morality doesn’t exist. Huge claims are placed on morality (objective, independent of humans) that are what causes scepticism, not that it’s simply an agreement among people – that’s just moral relativity with stuff written down.

  14. Also, the law has other differences:

    There is a statute book one can look up to find out what’s illegal;
    There is no ‘ought’/internal compulsion to obey the law, simply punishments for breaking it;
    There is no way to guess what is legal and illegal, whereas with the supposed moral law there should be some commonality;
    The law, in most instances, has some kind of reason/logic, however twisted, behind it – and one could disagree with that logic whereas morality is indivisible.

  15. Keddaw, I have no problem with teasing out particular problematic claims. But if that’s all you’re saying, then your advice is not for moral skeptics; it’s for everybody. If you’re giving advice to a moral skeptic in particular, then you have to be saying something a bit more forceful or general than that.

    The point of bringing up law to you was to point out that there is a sense in which claims about truth can be epistemically subjective, in the sense of that they are relative and hotly contested by rational people. The discovery that some form of discourse is relative is not enough to conclude that it is strictly speaking false. If the skeptic thinks this is a valid basis for their skepticism, then they need to justify that.

    One way to justify it would be to say, as you have, that moral claims purport to be epistemically objective (objective claims resolvable by rational dispute) and ontologically objective (independent of the mind, independent of humans). So when you question their truth, you are questioning those two features.

    But as a piece of sociology, the assumption that moral talk involves talk about ontological objectivity is just not right. While there are some people out there who believe in the divine right of kings, and in divine command theory, they do not have any ownership over the moral vocabulary. They might have during the age of Templars, but this is the 21st century.

    I do think that moral claims are epistemically objective, and so I would be inclined to challenge any denial of that. But for the sake of argument, I’ll grant it to you. Suppose that moral talk ends up just being prejudice and convention, even though it pretends to have an extra-conventional element. If you are charitable, do you have sufficient grounds to question whether or not the claim is true? The example from law and of ordinary performatives demonstrates that, no, it is not sufficient grounds: so long as the truth of the claim is tied to descriptions of human behavior, and serves an action-guiding purpose, it can be said to be true. So the skeptic, if they are a skeptic, has to abandon charity to remain in doubt.

  16. @BLS

    You’ve pointed out (rightly) that some of our other (non-moral) discourses are relative. But you’ve given us no reason to think that moral discourse is of that sort. You seem only to have made an argument of the following character: we must save moral discourse; we can only save moral discourse by adopting some new interpretation of moral language; therefore we must adopt a new interpretation of moral language. Saving a common discourse has some merit. But if we care about the truth we must not save a discourse at all costs. In any case, saving a discourse at the cost of the kind of redefinition you propose is not genuinely saving it. It’s only switching to a new discourse disguised as the old one.

    To take an extreme analogy, we might try to save flat-earth discourse by redefining “flat” to mean “inhabitable”. In this new discourse the claim “the earth is flat” would be true. But it would no longer be a genuine flat-earth discourse. Flat-earthers will not thank us for “saving” their discourse in this way. This new discourse will violate their ordinary sense (their sprachgefuhl) for the meaning of the word “flat”, and the new discourse will be of no interest to them. It will not serve their purpose.

    Similarly, most people’s sprachgefuhl tells them that their moral claims are ontologically objective (to use your term). They will not thank you for redefining their moral discourse as a relative one. They will reject your proposed discourse as not fit for purpose. It stops them from making the ontologically objective moral claims that they want to make.

    Of course the case of moral discourse is much less clear than my flat-earth case. The meanings of moral terms do not strike us in the strong way that the meaning of “flat” does. The word “flat” easily evokes mental images, but we have no mental images of moral properties. So redefinitions of moral language don’t seem so obviously to be redefinitions. What’s more, moral claims matter so much to people that their desire to save moral discourse can override their sprachgefuhl.

    Speaking for myself, before accepting moral error theory I briefly settled on non-cognitivism, because I couldn’t make sense of error theory. But I was never comfortable with non-cognitivism, which seemed inconsistent with both my sprachgefuhl and the evidence from moral discourse. I still remember the discomfort (the cognitive dissonance if you like) of holding a position that contradicted my sprachgefuhl. I remember the opposing thoughts going round and round in my head, “But it can’t be. But it must be. But it can’t be.” It was quite a relief when (after about 3 months) I was finally able to make sense of moral error theory and switch to that position.

    Neither you nor I have made any argument here for our interpretations of moral language. I think the burden is more on you, because your interpretation goes against the common sprachgefuhl. Even if you don’t agree that you have an epistemic burden, you have (like it or not) a practical burden: you will have to work harder to persuade people of your counterintuitive relativist interpretation.

    Anyway, my brief argument above attempted to obviate the need for arguing over the meaning of moral language, by considering the function of moral language instead. But it’s not designed to talk someone out of a revisionist definition that they’ve already adopted.

    A couple of additional points:
    1. I agree with Russell that moral language is not simple or monolithic, and there is probably some relative element to it. But I think the ontologically objective element is predominant.
    2. There are probably many folk who naively say such things as “morality is relative”. I don’t think they should be taken as giving a relativist interpretation of moral language. They seem to be saying something about morality itself, not moral language. I doubt that many naive moral relativists say, “moral language is relative.” Among the folk (and even some philosophers) I think the term “moral relativism” gets used in a very broad sense to refer to any rejection of the traditional objective picture of morality.

  17. The problem with moral error theory, as I see it, is that they are expecting brute facts and failing to recognize the social ontology of morality (I’m using Searle’s terminology because I can’t think of a better way to describe it). It is like someone who comes to the realization that money is only worth something because we believe it is worth something and then decides that that means money isn’t real. Money is real, even if it is only our collective belief that makes it real. Morality is also real, but it is social reality. It can be discussed objectively, the same way politics and marriage are discussed objectively.

  18. Hi Richard,

    Things might get tricky at this point, because the arguments I’ve been making is inextricably embedded in a conversation, and has to be understood in the context of that give-and-take. So I’d like to clarify a few things.

    My argument to keddaw was that their characterization of the subjective seems to imply that relativism counts as a form of it. And so my point to them was just to say: “Even if that were so, that won’t be enough to be skeptical of its truth”. After all, law, performatives, etc., are all culturally embedded, and relative, and obviously true. So it’s an immanent critique, which accepts keddaw’s premises and tries to show some different conclusions.

    That said, I myself don’t believe that all forms of relativism entail epistemic subjectivity (of the “that’s a nice car” type, or the “this counts as a move in chess” type). Speaking for myself, I do happen to believe that the law is epistemically objective. But I recognize that this is an additional step, and it will take some table-setting to say why; and I do not think it would be appropriate for this thread to go into the reasons, since that argument would end up being almost a complete departure from Russell’s OP. Long story short: I do not mean to argue that moral discourse is as a matter of fact merely conventional or relative.

    What I am saying is as follows. I think that, if you’re a skeptic who follows the principle of charity, in the absence of further information, you have an obligation to read people as relativists or conventionalists. The idea of charity is that you maximize the attribution of true beliefs to your interlocutor based on what they say; and if someone says something that seems literally false, it will depend on figuring out the most relevant beliefs which have the closest fit to what they say (given what you know about their other beliefs, etc.). And in a secular age, what’s a closer fit to “Thou shalt not kill”: is it “There is a property of non-killingness in the world”, or “God decrees that you ought not kill”? Or is it, “According to the law of the land, you should not kill”?

    If forced to choose between these two options, I think the latter is the more psychologically and sociologically plausible one. It may be ultimately the wrong way of thinking about morality — indeed, I do think it is misleading as an interpretation of morality. But so far as charity goes, it’s the superior choice. It is not the superior choice because I assume, a priori, that moral talk is worth keeping around (though as it happens, I do, but that’s neither here nor there). It’s the superior choice because it’s a better explanation of how the average person in a secular society conceives of their own vocabulary, and that it would be recognized as such by any genuine experts of the culture.

    Your flat earth analogy is extreme, and I think most people would balk at it. (Indeed, you balk at it.) The distance between concepts of the “flat earth” and “inhabitability” is several orders of magnitude larger than that between the concepts of “morality” and (a kind of) “performance”. However, this might not have always been the case. In the 12th century, I suppose it really would have been just as unthinkable to hold moral convictions as performatives; if I lived in the 12th century, then perhaps I ought to have been an error theorist about the dominant mode of speech. But then there’s an obvious question: why are we basing our contemporary semantic analysis on a lexicon that is a millenium out of date, and/or by deferring to a people who have little to no perspective on the causes of social behavior?

    When you say, “most people’s sprachgefuhl tells them that their moral claims are ontologically objective”, I intuitively doubt it. I think the people who I defer to as a linguistic community (broadly speaking, Canadians) probably wouldn’t care about that stuff — instead, they would settle for epistemic objectivity. Anyway, even if they did cling on to metaphysics, they’re not necessarily in charge of meaning; instead, semantics belongs to the experts, and the people I defer to as experts of social behavior and cognition (social scientists), who I really don’t think would interpret any necessary commitments to metaphysics.

    But, ultimately it’s an empirical question, what people are willing to do with words. And given the uselessness of our divided intuitions, the burden of proof is on both of us equally. It is increasingly impossible to make offhanded attributions to the folk without evidence. So this would make for an interesting project in experimental philosophy.

  19. I haven’t used (as opposed to ‘mentioned’) moral language whatsoever for around five years with virtually no issues. I rather just say what I mean – i.e. that I want such and such, or that I would feel uncomfortable seeing such and such etc.

  20. @BLS

    You seem to lean very heavily on “charity” in your defence of moral discourse. I don’t think it’s helpful to talk about “charity” here. We’re not trying to understand an interlocutor’s views, but to get to the truth of a philosophical question about moral discourse in general. In any case, some people expressly state that their discourse is not relative. It can hardly be called charitable to attribute to those people a view which they expressly deny.

    What I can reasonably interpret you as doing is appealing to the native intelligence of the folk: surely, you might say, the folk cannot be so confused in their moral discourse as to make claims which cannot be true.

    Yes, this consideration is a pertinent one, but it should be only one factor in our overall inference to the best explanation. How much weight should we give to it? Here are several factors that count against its weight:
    1. The folk get a lot of things wrong. It would be wise to start from a position of some skepticism about folk beliefs generally.
    2. The folk are very divided, both in their moral claims and in the sorts of justifications they give for their claims. Even if the folk are sometimes right about morality, they seem to be wrong much of the time. It’s not such a stretch to think that they’re wrong all the time.
    3. Those folk who say something about the meaning of their discourse (e.g. that it’s objective or relative) are at best divided and at worst mostly wrong (given your premise that morality is relative and my assertion that most people think it isn’t). More evidence that the folk are confused.
    4. Philosophers are very divided about the meaning of moral language. If philosophers are so confused, it can hardly be considered an insult to the intelligence of the folk to think that they are confused too.
    5. If we accept an evolutionary account of the human moral faculty, then (as I argued briefly above) we shouldn’t expect it to be truth-conducive.

    This isn’t the whole of the case for moral error theory. Just reasons not to be overly committed to saving moral discourse.

  21. BLS, people can happily talk about ‘moral’ issues, but that’s not what they’re discussing, they’re discussing tightly balled up, often contradictory, concepts usually without knowing it.

    When people use moral language they are using confused concepts that, in spite of your claims, do not allow communication on the important issues (contraception, abortion, women’s rights, food, alcohol, gambling etc.) because some people think it is just wrong, as in built in to the fabric of the universe wrong, and even those who are less objective will still say that torturing a baby for fun is wrong, not why it is, just that it is.

    Which isn’t to say there aren’t many situations where the bundling might be a useful shorthand. e.g. Saying freedom or democracy is good when the people taking part in the discussion have values that align to make them think these are ‘good’ can save time and effort, but it becomes worse than useless if a theocrat joins the discussion.

    Heck, it might even be objectively true, for them, or even their culture, but that simply renders it about as useful as a form of slang (something being “wicked” does not necessarily mean it is intentionally causing suffering and something being “cool” does not mean it has less heat energy than its surroundings) which is simply a barrier to communication with outsiders and leads to misunderstandings and arguments.

  22. Richard, if we’re not “trying to understand an interlocutor’s views”, then your earlier comments about the ‘sprachgefuhl’ have no force. It seems to me that this is a new development.

    For those who claim belief in moral absolutes even after having been prompted by relevant counter-examples, it is entirely right to say that their claims are false. We do it all the time. But [the absolutist] mode of speaking is not commonly recognized, at least not in my experience. For that reason, I do not see that it can be used as leverage by the moral skeptic — at least, so long as they have experiences like mine. (Of course, if you live in something more like a 12th century society, and the only social science you know about is from the 12th century, then perhaps skepticism about the truth of moral discourse is the right thing to go on.)

    It is fine to point out that we can speak either about the normative justification behind the use of ‘true’ and ‘false’ in moral discourse, or about the descriptive sense of how the folk make use of their discourse. As you say — quite nicely — inference to the best explanation is what we end up having to do when we’re trying to arrive at the normative point. But that’s why I proposed that a mature social science tells us what folk expressions of ‘true’ and ‘false’ really mean when it comes to moral discourse, regardless of what the folk intend to say. Should you erase Neitzsche, Durkheim, Freud, the Frankfurt School, Austin, Searle, etc., from history, and you will have secured a case for skepticism about moral discourse.

    Suppose you grant me those arguments, and decided to shift gears to your fifth point, which is that “if we accept an evolutionary account of the human moral faculty, then we shouldn’t expect it to be truth-conducive.” But why should we think that, unless we have an outmoded theory of truth?

    My inclination is to say that true propositions are representations of causes. (So, e.g., imperatives are neither true nor false. If I command you to do something, then the point of a command to you is to cause you to do an action, and not to meditate at all upon the representation of the relationship between my speech and your compliance. Imperatives can be assertable, but we do not say they are true. “Pass me the salt”: this is either executed or not executed, but it is neither true nor false.) So I think an evolutionary account of true propositions (in whatever domain) must be formulated in such a way that representations of the world are representations of causes. And, as I think you’re suggesting, since the moral faculty doesn’t seem to deal with representations of causes in the sense that moral language doesn’t map onto stuff in the world, this is prime facie reason to think that the moral faculty does not fit into the evolutionary picture.

    Unfortunately, that argument tacitly trades on a correspondence theory of truth, which I reject. Once you realize that performances are causally effective representations, you come to appreciate that they are truth-conducive. Performatives represent a social state of affairs as true, and genuinely cause the representation to become true, simply by virtue of others’ trust that the thing has been represented accurately. Unlike imperatives, they represent causes. In order to develop an essential sense of trust with others and with oneself, performatives are a necessary survival tool.

  23. Keddaw: I think a lot of my comments to Richard might do double-duty as a reply to you.

    To the extent it matters what the folk think, I don’t believe that divine command theory is commonly recognized as the constitution of morality. Secular humanism might not be the dominant paradigm, but it is not that uncommon.

    You seem to be suggesting, in a sense, that the majority of people might be (effectively) moral bullshitters. In that case, if you think that we ought to defer to the folk when it comes to their meta-ethics, then I guess you really could have some solid grounds to be a moral skeptic. I can make my peace with that kind of skeptic when it comes to the issue of whether or not folk use of moral talk is epistemically objective (e.g., computer science) or subjective (like aesthetics). But the bullshit of the folk is not sufficient grounds for doubting that moral discourse is capable of truth; after all, even a bullshitter can get things right, and even get things right as a matter of course. And more importantly for this thread, the error theorist and their fellow-travellers most decidedly do not get off the hook.

    That said, you evidently don’t want to just defer to the folk. You seem to think that experts are better suited to evaluate the real cognitive contents of what the folk are saying. In that event, you have to make a case based on the best philosophy and social science at hand, by digging deep into the antecedents of moral discourse.

    I haven’t yet seen any appeals to mature social science, apart from Richard’s allusion to evolution. As far as I can tell, Richard there claimed that moral discourse relies upon cognitive faculties that are not truth-responsive, and hence, the moral discourse inherits the non-cognitive contents. (This is, as a matter of fact, a huge inferential leap from the uses of language to the phylogenesis of minds — and that is something quite contestable. But for the sake of argument, I shall accept the premise that the truth of language uses track the truth of the underlying faculties.) And then I say: well, okay, but the underlying faculties *are* truth-conducive, because the correspondence theory of truth is only half-true. The capacity to represent and communicate causation is the key feature of the mental engine which drives moral discourse (or, at least, the truth of performatives), and which also drives ordinary representation of boring things in the world.

  24. @BLS

    I don’t think your last comment was responsive to mine. Perhaps I didn’t make clear the point of my comment, so let me recap.

    You’ve pointed out that our discourses can be relative or performative, but you haven’t given any evidence to support the claim that moral discourse is of one of those types. Indeed, as far as I can see, these are two different interpretations and you haven’t argued for one or the other. As I see it, your only argument has been that “charity” requires us to find some interpretation of moral discourse that will make it truth-capable. My previous two comments have been primarily responses to that argument.

    Philosophical arguments frequently founder on questions of what words mean. (I like Wittgenstein’s aphorism that philosophy is a battle against the bewitchment of our intellect by our language.) But the primary source of evidence of our own meaning is our sprachgefuhl. We can only hope to get the right answers if we adopt a skeptical attitude towards our intuitive beliefs and pay careful attention to our sprachgefuhl. Your stance on this question is the opposite: to subordinate your judgement about meaning to the need to save moral discourse. This is perfectly understandable, and common among philosophers. Our deepest moral intuitions are very powerful, and it’s very difficult for most people to seriously entertain the possibility that they might all be untrue.

    I’m going to finish there. Thanks for the discussion. It’s been interesting.

  25. Mike has made a post ( where he seems very sure what is moral and what is immoral. It seems hard, given this discussion, to be remotely charitable to his definitions – and I’m not.

  26. “Ethics so far as it springs from the desire to say something about the ultimate meaning of life, the absolute good, the absolute valuable, can be no science. What it says does not add to our knowledge in any sense. But it is a document of a tendency in the human mind which I personally cannot help respecting deeply and I would not for my life ridicule it.”
    Ludwig Wittgenstein

  27. Hi Richard,

    While I tried my level best to make my post above responsive to yours, I acknowledge that I have not done enough signposting. I apologize for any confusion this may have caused.

    Correct me if I’m wrong, but I took your reference to the sprachgefühl of the folk as reason to believe that you think the folk’s meta-ethical convictions make a difference to the contents of moral sentences. And I replied, perhaps that is so; but even if it were, this is no evidence of the absolutist nature of moral sentences, since we do not live in the 12th century. So if you want to take this line of argument, then charity demands that the context-invariant meaning of the sentences cannot be tied to any dotty metaphysics.

    But — stress, emphasis, bolded! — this is not my only argument. It is my argument in response to one of your worries. I’m not committed to charitable deference to the folk’s meta-ethics, one way or another. I’m letting you pick your battles.

    Here is an example of an alternative line of argument. I am delighted to agree that your points (1-3) are good reasons to doubt that the truth or falsity of the folk’s utterances can depend effectively on folk meta-ethics. My response to that is: fine, so let’s pay attention to the role of performative truth as understood by a mature social science. I barely made an argument about what constitutes a mature social science — I only did some name-dropping. But you cannot say I have not been responsive.

    So it is not the case that I take a “save moral judgment at all costs” angle, or that we should “subordinate your judgement about meaning to the need to save moral discourse”. Indeed, I would advocate the elimination and refutation of the truth in moral discourse under certain conditions, i.e., if all the folk were absolutists, and if there were no such thing as a mature social science, or if the mature social sciences came out in favor of error theory.

    I ignored (4) because in any philosophical conversation it’s redundant to point out that philosophers disagree. I don’t agree with the Wittgenstein quoted there, for example; and I don’t believe that morality is just another language-game. [Frankly, I think the metaphor of a ‘language-game’ is one-dimensional, for most purposes. “Linguistic tool” is better; and “linguistic institution” even better than that.] But this goes far afield.

    The only area where I am genuinely not confident that I’m not on the same page is when you make an appeal to evolution (your point 5). You haven’t said quite enough for me to grapple with. I had to take a stab at it, though if I missed the point I wouldn’t be surprised.

    Hi Keddaw,

    It is very important, I think, to make sure that we’re making the same conceptual distinctions. So in their reference to Mike’s latest post, you seem to be worried about the morally real. But Mike’s use of moral discourse involves talking about moral truth. They are not the same — there are an enormous number of true (and banal) claims that involve the representation of causes, but which don’t map or correspond onto mind-independent facts in the world.

  28. @BLS

    Just to clarify. The purpose of my replies to you has not been to argue directly for the absolutist nature of moral discourse, but to counter your argument against absolutism, namely your argument from charity. Of course undermining your argument against absolutism would indirectly support absolutism. But you seem to have taken me to be making direct arguments for absolutism, and because I didn’t make any and you therefore couldn’t find any, you’ve attempted to read between the lines, and attributed to me arguments that I haven’t made. I’m sorry I didn’t make my intentions clearer.

    Perhaps you will say you’ve made other positive arguments, beside the argument from charity. I can only say that I failed to see any.

  29. Richard,

    I have not argued anywhere above for or against moral absolutism.

    I have argued that if you think theocratic absolutist folk meta-ethics (say, beliefs about the foundations of morality in mind-independent facts about the will of God) is part of the meaning of morally charged sentences, then you’re not being charitable, because as a matter of fact, secular humanism has had an impact upon the conventions which govern mainstream moral discourse. I think that if you are being charitable, the diversity of meta-ethical positions adopted by morally competent people should indicate that it is just a mistake to think that absolutism or relativism are a part of the meaning of morally charged sentences at all.

    Also, I have not attributed to you any direct argument for moral absolutism. I understood your argument in more or less the way you presented it: that some people expressly state that their discourse is “not relative” (absolute), and hence that it would not be charitable to attribute to them a relativist view. I agree that you need not interpret them as relativistic if they insist on calling themselves absolutists; indeed, I don’t think their position on those issues makes any difference to the meaning of what they’re literally saying. Charity only requires you to interpret them as asserting “conventions with extra pretentions” in a “worst case scenario”, e.g., when they’re deeply and genuinely confused about their meta-ethics but generally on an even keel in their morality. (Nov. 18)

    The argument from charity is not itself what most people would call a positive argument. It is a negative argument, in the sense that it originates as a criticism. The positive argument is summed up in a slogan: “Performatives are true. Moral claims are performatives. Moral claims are true.”

  30. Personally, I’m a trial and error theorist.

  31. It always struck me as odd Mackie went that way.

    If you agree moral predicates are thought to denote but in fact don’t then a more natural view, I think, is that declarative sentences like “Eating snails with butter is wrong” in moral use are not false but meaningless and do not express moral propositions, considerations, or reasons (of which there are none), morality itself being a mechanism of social control depending on mystification, bluff, and coercion.

    Just a thought.

  32. The Error Theorist’s Dilemma | The Skeptical Stingray - pingback on January 23, 2015 at 10:38 pm

Leave a Comment

NOTE - You can use these HTML tags and attributes:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

Trackbacks and Pingbacks: