As we all know saving the differences is difficult.
Distinguishing between realists and anti-realists is coming increasingly
difficult. For me, this is a fascinating topic. I’m reading our fellow pea
souper’s, Michael Huemer’s, book Ethical
. He doesn’t explicitly investigate the topic but still seems
to take a stand on it in classifying different metaethical views. I wonder how
efficient his way of describing the differences is. Much of what I have in mind
here is nothing original but I still would like to ask Michael if he had more
in mind that I’m picking up.

Huemer seems to think that we can use the concepts of
‘objective’ and ‘subjective’ for making the metaethical difference. He defines
subjectivity in this way:


( Subj) F-ness is subjective = Whether something is F
constitutively depends on at least in part on the psychological attitude or
response that observers have or would have towards that thing.


I take it that he thinks that F-ness is objective if and
only if it is not subjective, i.e., F-ness does not depend even in part on the
attitudes or responses of the observers. The crucial question he poses is whether goodness is
objective. If you are a moral realist, then your answer to this question must
be yes and if you are an anti-realist
your question must be no. So far, so


Huemer then lists three anti-realist theories: subjectivism,
non-cognitivism and nihilism a.k.a. the error theory. I agree with him that
these theories are anti-realist views but I cannot see how his criteria can
pick them out as such.


Start with error theory. Would an error theorist accept or
reject (Subj)? I take it that he would reject it and say that goodness does not
depend on the attitudes of the subjects. In this sense they would reply to
Huemer’s question yes – goodness is
an objective quality, just nothing happens to be good.


The fate of non-cognitivism is trickier. It depends whether
(Subj) says more than this:


(Subj*) If we did not have certain attitudes towards the
object, then the object would not be good.


If it doesn’t, then the non-cognitivist denies subjectivity. Blackburn is famous about this. Any claims of
the sort (Subj) or (Subj*) are part of morally suspicious first-order
moralizing to which the non-cognitivist does not want to be committed. As
Blackburn puts it nicely ‘The criminality of the Iraq war is
dead-innocent-Iraqi-dependent, not Republican-sentiment-dependent’.


So, it looks like both error theorists and non-cognitivists
answer Huemer’s question yes. Goodness
is objective. But, if Huemer’s question classified realists and anti-realists,
this would make error theorists and non-cognitivists realists. That cannot be
right. So, the question cannot distinguish between realists and anti-realists.


Huemer has another classification that does better with the
error theorists. This test distinguishes the kinds of facts you believe in. The
error theorists does not accept moral facts. But, of course, again the
sophisticated expressivists is happy to accept the talk of mind-independent
irreducible moral facts.

16 Replies to “Huemer on Objectivity and Realism

  1. Jussi,
    What’s bad about an error theory being labeled as a realist theory? Is it just that a realist says there are objective goods and an error theorist says there are not objective goods?
    If that is it, then shouldn’t we just read Mike’s account of an objective good as existence-entailing. This, then, is what it would look like:
    (OBJ) F-ness is objective = There is an F and whether something is an F does not constitutively depend on the psychological attitude or response that observers have or would have towards that thing.
    A realist believes there are goods that are also objective. Now, however, an error theorist is “not” a realist because she thinks there are no objective goods. Does that not solve the worry?

  2. Christian,
    Thanks. I think most people have the intuition that error theorists are not realists. But, many people’s classifications do not imply this (like Dummett’s). Your way of phrasing the question deals with the error theorists well but still leaves the non-cognitivists who would accept (OBJ) on the realist side.

  3. My classification is very different from Mike Huemer’s, but I also put Mackie on the realist side. I’d agree with you though, Jussi, that if both Mackie and Blackburn get counted as realists, the classification scheme has at least got the labels wrong.
    I don’t like Christian’s way of formulating the distinction, by the way, because it makes the property of being a gold mountain come out non-objective. Of course, taxonomists can use ‘objective’ and ‘realist’ as technical terms and let the chips fall where they may, but my view about gold mountains doesn’t have much philosophically interesting in common with Blackburn’s view about moral wrongness and Dummett’s view about large cardinals. (Whoa, that Albert Pujols is unreal!)

  4. Jamie,
    That is a nice counterexample. I want to add something too, take the state of being in love, or a state of knowledge, or a state of truly believing the world is good.
    These depend on subjects, attitudes, or more generally, something “subjective” for their existence. However, these goods are objective.
    Mike might say these properties do not “constitutively depend” upon attitudes, but just depend upon them in some other way. I don’t know what that other way is. Nonetheless, drawing the subjective/objective distinction is tough, and I, for one, haven’t seen it drawn to my satifaction.

  5. As long as people are trying to get clear on this, could anyone address where in any of these taxonomies sensibility theories fall?
    I ask because as an outsider to metaethics trying to get up to speed, there was a gap in time between when I first heard the position articulated and when I later heard people refering to it as “response-dependent realism”. Am I right in thinking that under Huemer’s classification this label is oxymoronic?

  6. Andrew,
    that’s an excellent question. The problem is that there are so many versions responce-dependence theories and so many ways of carving the metaethical landscape to realist and anti-realist theories. For what it’s worth, here’s at least the Wrightian way of how I understand the question. I take it that response-depence accounts are based on bi-conditionals like this one:
    X is F iff in circumstances C agents of the type A would have the response R (experience of Fness usually).
    This conditional is then understood as neutral between realist and anti-realist views. The realism question then becomes the question of what direction should we read the bi-conditional, i.e., what is the direction of explanation for it.
    The anti-realist response-dependence theorists will take the right-hand side to be the starting-point and say that the relevant responses determine the extension of F. F is projected to the world as a result of the responses of the ideal agents. This also implies that given the idealised responses it could not be that X was not F.
    The realist response-dependece theorists will take the left-hand side as a starting point and say that ideal responses only track the relevant property which is *humanly speaking* an objective feature of the world. The idea is something like that it just takes certain kind of human sensitivies that result from the correct education to be in touch with the properties that would be non-existent from ‘the point of view of the universe’. The properties, unlike other realist properties, might remain undetectable for beings with other kinds of sensibilities. But, as the properties are still ‘out-there’ there is a possibility that even the ideal responses would fail to respond to them and for this reason we earn the right to the realist title.
    I hope that helps somewhat. It is a complicated issue that would be well worth investigating further.

  7. Pardoning the pun, but it seems like people who accept the biconditional form are all trying to “gild or stain” this thing called “direction of fit” onto an otherwise neutral formulation…
    But my understanding of McDowell at least, and possibly Wiggins, is that the account of response-dependence explicitly denies that there even is such such a thing as a direction of explanation to be found; it’s supposed to be a (non-viciously) circular account. If this is a coherent view, then it seems that none of the taxonomies discussed so far will capture it, since it’s projectivists like Blackburn who end up on the anti-realist side and realist objectivists who pick the opposite direction of fit on the other (could anyone point me towards someone who maintains this view? I’d be interested to see how “response-dependence” figures in it at all except as a purely epistemic marker).
    So, the Direction of Explanation taxonomy won’t cleanly place this sort of sensibility theory anywhere, but I think that Heumer’s taxonomy would still have it coming out on the subjective, anti-realist side of the divide, since even a circular characterization still “constitutively depends” on the emotive response half of the circle.

  8. Andrew,
    you are right. One further condition in Wright’s way of carving the metaethical space is that if in specifying the circumstances C or the type of agents A you need to refer to any particularly moral properties, then you are putting forward a realist response-dependence view. This is because you need to assume the existence of some out-there moral properties to specify the responses that count with the respect to the property you are giving an account of. You are right that McDowell and Wiggins openly do so and for this reason they correctly count as realists in this classification. I think that Huemer’s classification gets them wrong. Not many people have tried to formulate the anti-realist versions of response-dependence views. The problem is that if you are allowed only to use naturalistic vocubulary in specifying the circumstances and agents you are probably ending up with a relativistic account of the response-dependent property.

  9. Christian’s suggestion is roughly in line with what I intended. Except that I wouldn’t modify the definition of an objective property; I would just modify the definition of “realism”. I meant moral realism to be the view that some things are objectively good (or right, etc.), that some things have objective evaluative properties, or that some (positive, 1st-order) evaluative statements are objectively true.
    As I see it, the nihilist and the non-cognitivist would deny that the property of goodness even exists. However, you might say a Platonist nihilist (if there are any) would say that goodness still “exists”, even though it is uninstantiated. Still, they would deny that anything has an objective value property.
    Now, here are some other properties mentioned in the discussion above:
    1. Being a golden mountain: This is objective, per my original definition–

    (Subj) F-ness is subjective = Whether something is F constitutively depends on at least in part on the psychological attitude or response that observers have or would have towards that thing.

    –although I am not a realist about golden mountains.
    2. Being in love: This is also objective, per the stated definition. Note that the def. does not say F-ness is subjective provided that F-ness constitutively depends on subjects or attitudes. It says that F-ness is subjective provided that x’s being F constitutively depends on the attitudes/reactions of observers towards x. x’s being in love does not constitutively depend on the attitudes or psychological reactions of observers toward x.
    3. Knowing that p, truthfully believing the world is good: Both objective. Same point as above.
    Lastly, what about the “sophisticated” expressivists? Jussi says:

    But, of course, again the sophisticated expressivists is happy to accept the talk of mind-independent irreducible moral facts.

    I wouldn’t say “of course” here. Timmons would be happy to say that, but I think very few other expressivists would. And my view is that in saying “there are mind-independent, irreducible moral facts,” the expressivist is either contradicting his central (expressivist) position, or simply changing the meanings of words like “mind-independent”, “fact”, or even “there are”. If you use those words in their ordinary senses, then (I would say of course) expressivists deny the existence of mind-independent moral facts.
    p.s. I’ve been gone from the blogosphere for a while. I have a very busy semester. I’ll try to stop in occasionally.

  10. Michael,
    thanks for the clarifications. About the last point, maybe there is a difference in the kind of expressivist we have met. I’m not sure I’ve ever met an expressivist who would be unhappy about the fact talk.
    As illustration, here is what Blackburn says in Ruling Passions: ‘Minimalism seems to let us end up saying, for instance, that ‘kindness is good’ represents facts’, ‘The ethical proposition is what it is and not another thing; its truth means that it represents the ethical facts or the ethical properties of things. We can throw in mention of reality: ethical propositions are really true’, ‘we might even find ourselves saying that we know moral propositions to be true. Or, really true, or really factually true, or really in accord with the eternal harmonies and verities that govern the universe’, and ‘We can add flowers without end:’it is good to be kind to children’ conforms to the eternal normative structure of the world’.
    Gibbard says similar things in THoL: ‘Anyone who reasons what to do, I argue, is committed to something very much like *facts* of what to do’, ‘There is indeed a property that constitutes being good’, ‘In many ways I end up sounding like a non-naturalist, and in some ways, like certain kind of naturalilists’, and so on.
    I guess I find it harder to find an expressivist who would want to revise our moral discourse in a way that would abandon all the talk about facts, truths, and so on.

  11. Yeah, I was thinking of the older, more forthright non-cognitivists (like Ayer and Stevenson). What Blackburn and Gibbard are saying in the quotations you produced is pretty obviously something they are not entitled to say–it directly and obviously contradicts their whole anti-realist meta-ethics.
    Of course, they think they can say those things because they think they can (re)interpret “fact”, “true”, and all related expressions in anti-realist terms. I think they’re just trying to (radically) change the meanings of those words.
    Now, you might think that even so, our classification of metaethical positions should take into account what these new expressivists would *say*. But if philosophers are allowed to change the meanings of words, then coming up with a classification scheme that expresses the distinction between realism and anti-realism in words that all philosophers would accept, is an impossible task.

  12. Michael,
    I think I see things very differently. First, I don’t see how Blackburn’s or Gibbard’s fact claims could contradict their metaethics when the fact talk is supposed to be metaphysically non-committal for them. I also wonder whether the charge of changing the subject is accurate. It would be if our moral truth and fact talk was inflated and came with a metaphysical cost.
    I’m not sure it is though. Is there more to be learned about truth and facts in general than the relevant instantiations of the T-schema and the corresponding biconditional for facts? And, especially do these terms amount to more in the moral discourse? At least for me, it is not so obvious that in the ordinary discourse these notions are used with a metaphysical weight. If they are and the expressivists are changing the subject, then you need at least a strong argument to that conclusion.
    I think that’s one of the better places to draw the realism vs. anti-realism contrast in any case. Btw, it’s slightly unfair to take Ayer and Stevenson as representatives of expressivism and target the general arguments against expressivism against their views. It’s a bit like their side would take Moore to be the representative of realists and dismiss realism by arguing against him.

  13. Jussi,
    If fact-talk were metaphysically non-committal, then assertions about moral facts would be compatible with anti-realist metaethics. But fact talk isn’t metaphysically non-committal–how could it be (other than by changing the meanings of words)? Assertions like “there are facts of type F” commit one to *the existence of facts of type F*. If that’s not a metaphysical commitment, I don’t know what is.
    About truth: we could have a separate discussion about truth, and deflationary views thereof (in which case, we should probably start a new thread). Very briefly, I think trying to ‘deflate’ the notion of truth by appealing to the T-sentence schema fails, because (1) you can’t thereby explain certain uses of “true” in which it is not stated exactly what was true (as in “do you swear to tell the truth?”)–this is a lot like the Frege-Geach problem, btw. (2) In order to explain what the T-schema really means, you have to use use some other semantic notion (like “meaning”). The T-schema,
    “P” is true iff P
    really means something like this: For any proposition, x, and sentence, y, if y means that (asserts that, expresses, etc.) x, then y is true iff x (is true).

  14. Michael,
    that’s exactly the point. I don’t know what the metaphysical commitment is when one accepts the existence of the facts of the type F. I do know what the metaphysical commitment is when I accept the existence of chairs, cars, electrons and protons, but I have no idea what facts add to the world. And, I don’t know for what that addition of facts to our ontology would be needed. Strawson’s line on facts has always seemed appealing. Of course, true sentences and facts match. This is because we introduce the whole fact talk in order to have something that makes our claims true. As he eloquently put it: “Of course, statements and facts fit. They were made for each other. If you prise the statements off the world you prise the facts off it too; but the world would be none the poorer.” (Strawson 1950b, 197). If there is more to facts, I really would like to know what.
    You are right that one cannot be deflationist about all semantic notions across the board, or at least it is difficult. But, I don’t see how this is an argument against deflationism about truth. One can accept inflated notion about meaning, say a use theory, and be a deflationist about truth just as Horwich is. You can even stipulate a logical dummy property to correspond to the uses of truth like the one you mention to get around the Frege-Geach problem.

  15. Just a quick point, on Michael Huemer’s objection (2):
    I don’t think anybody has to ‘explain what the T-schema really means’. It doesn’t mean anything, in the colloquial sense of meaning something. It’s a schema! All of its instances had better mean something. But each of its instances seems to have a straightforward meaning that doesn’t involve any semantic notions other than truth. The truth predicate, then, is implicitly defined by the collection of instances of the schema.

  16. So suppose that your theory of truth is (summarized in) your (Jamie’s) last sentence: “truth” is implicitly defined by the collection of instances of the schema.
    I then ask: exactly what collection are you talking about? You give me two or three examples. I say, “I’m sorry, I don’t know what set you’re gesturing at here. There are infinitely many sets of biconditionals that contain those three examples. You need to tell me what all the members of the set you’re pointing at have in common.”
    What do you then say? I think you say something like this: What they have in common is that they are all of the form “x is true iff y”, where y is replaced by something that expresses a proposition, and x is replaced by something that names a sentence that expresses that same proposition.
    What’s supposed to be interesting about this: well, if you need a (non-‘deflated’) semantic notion anyway, you might as well define “truth” in terms of that semantic notion, in which case I don’t see in what sense the account of truth is “deflationary”. (I really don’t know what “deflationary” means. But I assume that if x is defined in terms of y, and one’s account of y is ‘inflationary’–whatever that means–then one’s account of x is also inflationary.)

Comments are closed.