Expressivism, Subjectivism and Reasons

This post will be difficult to write as I’ll have to reign in my frustrations (I was thinking of calling this ‘Must Do Better’ or ‘All Souls Night, Part II’…). I’ve been reading Katarzyna de Lazari-Radek’s and Peter Singer’s The Point of View of the Universe – Sidgwick & Contemporary Ethics. Chapter 2 (sections 2–4) of this book contains a ‘Parfit-inspired’ criticism of expressivism and Blackburn, which I find quite upsetting. It makes me feel sad for Simon Blackburn (he really must get tired of objections like this) and it also makes me wonder about OUP’s editorial processes (I don’t see how this section could pass a peer-reviewed journal). In any case, here I want to start from a very basic distinction and then go through quickly some of the fairly outrageous claims de Lazari-Radek and Singer make about expressivism.

There are two different views in this context that are both inspired by David Hume. One of these is a first-order normative view about what reasons a given agent has. This view is sometimes called ‘existence internalism’ and it has been perhaps most clearly formulated by Bernard Williams. Roughly, his version of the view claims that a consideration is an reason for an agent to phi if and only if there is a sound deliberative route for the agent from her pre-existing motivational set to phying via that consideration. On this view then, it is true that what reasons an agent has at a given moment is a function of her desires (plus other items in the motivational set, plus sound deliberation, plus imagination…). Of course, since Williams introduced the modern version of this view, there has been huge amount written on whether existence internalism is true.

The second view is expressivism. Expressivism is a semantic (or better still a metasemantic) view about the meaning of normative sentences/utterances and the content of normative judgments. This means that it tells a story of what normative sentences mean on the basis of the desire-like, world-to-mind direction of fit mental states, which these judgments express. For example, one expressivist view about reasons claims and thoughts would be that to think that ‘The fact that P is a reason for S to phi in C’ is to be for taking the fact that P into account in phi-ing friendly way in C (to borrow something like Antti Kauppinen’s suggestion) and when you utter the sentence you conventionally express this thought.

The important point is there is no connection whatsoever between existence internalism and expressivism (other than that they are both Hume-inspired). As a semantic view of the meaning of reasons sentence and thoughts, expressivism is completely neutral about what reasons agents have, whether these reasons are facts, propositions, desire-belief pairs, and especially whether the having of a reason by a given agent depends on that agent’s current desires. Expressivism merely is a view about what you think when you think that existence internalism is true: when you say that Charlie’s reasons depend on his current desires you are expressing your attitude of being for Charlie only using those considerations which have the power to move him into account in action-friendly way in deliberation. Expressivism is equally also a view about what it is to think that existence externalism is true: when you say that Charlie’s reasons do not depend on his desires you are expressing your attitude of being for Charlie also using other considerations (that don’t move him given his current motivational set) into account in action-friendly way in deliberation. Expressivism doesn’t decide which one of these first-order normative views is true, full stop. You can be an expressivist and endorse either.

With this distinction in mind, we can turn to de Lazari-Radek and Singer’s treatment of expressivism. They call existence internalism subjectivism. Here’s what they write ‘Simon Blackburn is an expressivist, and therefore a subjectivist…’ (49). As explained above, ‘therefore’ here is just confused. Accepting expressivism does not commit you to the first-order normative view that what reasons an agent has depends on her desires. This means that by criticising subjectivism you do not criticise expressivism. For this reason, the objections de Lazari-Radek and Singer make to expressivism just do not hit the target.

To see this, let’s start from de Lazari-Radek and Singer’s objections to subjectivism. This is based on Parfit’s Future Tuesday Indifference case. There’s a person, call him Jack, who ‘cares about his pleasures and pains in just the same way as we do, with one important exception – he doesn’t care about them if they happen on any future Tuesday’. The idea then is that Jack has an objective fact-given reason to care about his suffering on Future Tuesdays and it is rational for him to do so, whereas existence internalism (‘subjectivism’) entails that Jack has no such reason and that it is therefore not irrational for him not to care about Tuesday suffering. Let’s accept that this charge against subjectivism is true (I have my reservations about this). This still is in no way a problem for expressivism. All expressivism says is that when Parfit, de Lazari-Radek, Singer, me, you or anyone else thinks that Jack has an objective object-given reason to avoid Future Tuesday suffering we are all for Jack taking his Future Tuesday suffering into account in deliberation in avoidance-friendly way. And, given that we deeply care about Jack and all other human beings and their well-being at any time, presumably we are for this. Nothing in expressivism says that we wouldn’t or shouldn’t be.

This means that the Future Tuesday Indifference is not a problem for expressivism in any way whatsoever (and so are not objective object-given reasons). De Lazari-Radek and Singer do make other objections too. The main one is that expressivists cannot explain how our moral attitudes are significantly different from our allegiances to sporting teams. I do think that this is a good objection to expressivism. It is called the ‘Moral Attitude Problem’. This is a problem Ayer already consciously struggled with and it has since been extensively discussed (there’s a nice overview in Miller’s intro to metaethics). I was surprised not to see any acknowledgement of this fascinating debate.

However, De Lazari-Radek and Singer tie this objection to the previous objection about reasons and this is again where things go wrong. They think that what distinguishes moral judgments from personal preferences is that we can give reasons for our moral judgments whereas for personal preferences we can give only causal explanations. The claim then is that because expressivists do not ground moral judgments on reasons they fail to explain the difference. This must be a point Blackburn is tired of hearing. I can only say that of course expressivists do ground moral judgments on reasons just as much as we all do. I can quote Blackburn: “Cruelty is bad … because it exhibits the intention to cause pain” (Ruling Passions, 307–308). Equally, “the truth … that you have an obligation to your children is child dependent, and the truth that you should not kick friendly dogs for fun is dog dependent” (Must We Weep for Sentimentalism?). This means that, as should already be clear, expressivists like Blackburn can give exactly the same objective object-given reasons for moral judgments as Parfit, de Lazari-Radek and Singer. Their metaethical view merely gives an account of what it is to think that there are such reasons for our moral attitudes. This is why it is inappropriate to say that ‘expressivists cannot defend the claim that moral issues really matter’ (50). To repeat, expressivists can defend the claim that moral issues really matter by giving exactly the same reasons why they matter than everyone else. 

De Lazari-Radek and Singer’s discussion of expressivism also contains a number of other problems that are due to failure to fully understand quasi-realism. Here’s just one illustration:

“To accept Blackburn’s view, however, brings us close to some kind of mental schizophrenia. If I utter the sentence “It is wrong that people are starving to death in Somalia and we are doing nothing to help them” and at the same time think that in saying this I am just expressing my attitudes, rather than stating anything that is true, then there is no way in which my judgment can express the idea that what I am saying is important independently of my present attitude toward it”.

There are a number of problems here. There is nothing “mere” or “just” about our attitudes or expressing them – these are the deepest concerns that make us who we are; according to Blackburn you are stating something true when you say that sentence, and your judgment (or his or mine) in no way is important because of your present attitudes (but rather for good reasons). I could go on and on, but I think I have to end here as I am sure you have the picture.

34 Replies to “Expressivism, Subjectivism and Reasons

  1. Amen. There is a substantive issue about whether expressivism is subjective in any interesting sense, of course, but this turns on rather subtle issues about how to understand claims like ”If I didn’t disapprove of kicking dogs, it wouldn’t be wrong to kick dogs”. Are they first-order claims which we should endorse? This would be bad, obviously, but luckily there is no reason for a (reasonable and moral!) expressivist to endorse them. Are they metasemantic claims about moral judgments? More plausible, but controversial. And it is not nearly as obvious why we should care much about “subjectivism” in our metasemantic views.
    These issues have been EXTENSIVELY discussed by Blackburn, Dworkin, Peacocke, here on Peasoup if I remember correctly, and so on. These issues are serious and interesting and not to be brushed under the rug in either direction, it seems to me.

  2. It is depressing and embarrassing how little progress we are able to make. I assign Brand Blanshard’s paper on “The New Subjectivism in Ethics”, where Blanshard is guilty of the conflation you’ve pointed out here, between expressivism and existence internalism, to undergraduates. I tell them that at least one of them is going to make the same mistake in a paper, and that they will do so even though I’m warning them.
    Blanshard’s paper is from 1949, so he had an excuse. Since then the mistake has been made so many times I doubt anyone knows about all of them. Mark Schroeder rooted out the error in several of its appearances and gave a kind of recipe for explaining what’s wrong with each, in ‘Does Expressivism Have Subjectivist Consequences?’, but I guess that didn’t do the trick. Of course, Mark hasn’t published the paper, but it can be found on his research page.
    As I’m sure you know, “Must we weep for sentimentalism?” also contains an explanation of the conflation. I guess Singer and de Lazari-Radek didn’t read that.
    Anyway, good job. Someone has to clean up messes, so thanks for doing your share, Jussi.

  3. Hi Jussi,
    you’re surely right that expressivism and existence internalism are independent of each other. However, as I recall, part of what motivates Gibbard and Blackburn to develop their expressivist theories is the idea that claims about attitude-independent reasons seemingly make sense. It is a nice feature of expressivism, they argue, that it can give an account of the meaning of claims about attitude-independent reasons. (Remember that in contrast, Williams said, in one of his main papers on the topic, that he thinks that we cannot make any sense of claims about “external” reasons.) So, though you’re surely right that expressivism has no substantive implications, or at least that it does not imply existence internalism, there are connections between the two having to do with the motives behind the development of these theories: some of the top expressivists wish to explain how it can make sense to make normative claims that conflict with what existence internalism allows for.

  4. Hi Sven
    completely agree. I was just reading All Souls Night again and Blackburn makes your point (and most of my original post) wonderfully eloquently. The more I read that paper the more I love it.

  5. Hi Sven and Jussi (or anyone else),
    Can you point me to substantive expressivist discussions of Williams’ internalism? I am aware of Gibbard’s paper “Reasons, Thick and Thin,” but would love to know of any other explicit expressivist defenses of externalism.
    Thanks,
    Brad

  6. Hi Brad,
    I am away from my Gibbard- and Blackburn-books at the moment. But just quickly from memory: I think Gibbard also talks about existence-internalism in Wise Choices, Apt Feelings. And Blackburn talks about these issues in Ruling Passions, saying, among other things, that normative and motivating reasons can come apart. (The idea being, I seem to remember, that though motivating reasons may need to always involve the agent’s desires, there is no reason, as Blackburn sees things, why normative reasons would need to.) Blackburn also talks briefly about these things in a piece of his on Sharon Street’s view (which at some point was posted on his website). Sorry for not being able to give more helpful and precise references at the moment, but like I said, I am on a trip and don’t have access to my books at the moment.

  7. Hi Brad
    for a substantive discussion see pages 264-266 in Ruling Passions and especially Blackburn’s “Majesty of Reason” paper (in the Practical Tortoise Raising collection). The discussion of Parfit in All Souls Night paper (still on Blackburn’s website) is also instructive. I recall various other people writing and speaking about this but I’m not sure where to find the papers from.

  8. It really is baffling and frustrating that this mistake can still be made so baldly after all this time. I’d like to read All Souls Night but can’t find it online. Can someone less incompetent in such matters link to it please? Thanks.

  9. Brad,
    As I remember it, Gibbard talks about Brandt’s subjectivist-style view early on in Wise-Choices.

  10. Non-philosopher here, so I apologise if I get this completely wrong, but it seems to me like you’re splitting a hair to no great effect. I recognise the conceptual difference between expressivism and internalism, as you label them, but to say, “there is no connection whatsoever between existence internalism and expressivism” is just absurd. There is a very obvious connection between them, and seeing as the connection is precisely what Parfit writes about, it seems relevant to this conversation.
    I think the premises for my argument are something like this:
    1) Either there are objective, non-internal moral reasons, or there aren’t.
    2) There are two proper types of motivations for action: internal desires and objective morals.
    3) Moral language is used in discussions of motivations.
    Now, if in fact there are no objective moral reasons, two things follow: a) the real referent of moral language is desires (expressivism); b) the only proper motivation for action is desires (internalism).
    Obviously you’re correct that there is no causal relationship between E and I, but they are both (I think necessary) consequences or corollaries of the claim that there are no such things as objective, external moral reasons. Therefore an argument against either one of them is at the very least *relevant* to the other.

  11. Hi everyone, just to let you know that we have read the discussion. Thank you for it [even though it’s hard to say we are happy with it :)]. We are going to answer shortly!

  12. Jussi,

    I don’t disagree with your main point, but I am puzzled by some things you say.

    First, you say:

    Expressivism is a semantic (or better still a metasemantic) view about the meaning of normative sentences/utterances and the content of normative judgments.

    Later you say:

    Expressivism merely is a view about what you think when you think that existence internalism is true …. Expressivism is equally also a view about what it is to think that existence externalism is true …

    But I am not sure that these are normative thoughts (or normative judgements). Internalism and Externalism are views about the concept of a reason, I would have thought, not substantive claims about what reasons we have. You describe them as "first-order normative views", but that seems wrong to me.

  13. Hi Campbell
    I hope you are well. Yes and no. Consider the two following formulations of Internalism:
    (Internalism 1) The fact that P is a reason for A to phi in C iff there is a sound deliberative route from A’s motivational set to phi-ing in C via the fact that P.
    This biconditional is merely a claim about when and what reasons an agent has. In this sense, it is a first-order normative claim (and metaethically neutral given that expressivists, non-naturalists and different forms of naturalism can accept this claim about extension of reasons).
    It’s true that some internalists (and I think Williams is amongst them) accepted the following stronger version of internalism:
    (Internalism 2) For the fact that P to be a reason for A to phi in C is for there to be a sound deliberative route from A’s motivational set to phi-ing in C via the fact that P.
    This is a conceptual claim (or a metaphysical reduction) as you are probably thinking of internalism. This thesis reductively analyses reasonhood in terms of motivation. Internalism 2 then is not primarily a normative claim but rather a reductivist naturalist metaethical claim (which perhaps has normative implications). But no expressivist worth their salt would accept this reduction.

  14. I’m not convinced that Internalism 1 is a (first-order) normative claim. Here’s one reason to think that it is not. Internalism 2 is not a normative claim. But I2 implies I1 (I assume). And it is natural to think that the normative claims are closed under the inverse of implication; i.e., if X implies Y, and Y is normative, then X is normative. (Another view inspired by Hume!)

  15. Hi Campbell
    that cannot be right. If it were, there not be any normative claims whatsoever. Here’s the argument:
    Take any apparently normative claim of the form ‘It is N to do X’ where N is a normative predicate (reasons, oughts, right, wrong whatever normative predicate we want to look at – even if for someone of these predicates we would need to change the structure of the claim).
    Then take a metaethical view that reductively analyses being N in terms of doing X. This view would say that: to be N is to be an instance of doing X. This metaethical view is not a first-order normative claim itself as it says something about the metaphysical structure of Nhood. It does, however, entail the first claim: It is N to do X. So, by your criterion the first claim would not be a normative claim.
    This shows that as long as there are suitable candidates for reductive analyses of normative claims available there are no first-order normative claims. That can’t be right.
    To repeat, internalism 1 is just a claim about what reasons an agent has. It also has implications for what an agent ought to do and this claim can be accepted by the major metaethical views. So, it is a normative claim.

  16. Jussi: But not all normative claims have that form. Consider, e.g., “Some acts of lying are wrong”, or “Tom’s lying to his friend was wrong”. I doubt that these are implied by any reductive analyses of normative properties. The second example implies that Tom lied to his friend. But I don’t think any analysis would imply this.
    Jamie: Yes, I know. But there is a better version of Hume’s Law than the one refuted by Prior.

  17. For an act to be wrong is for it to be an instance of lying implies both of those claims.

  18. “For an act to be wrong is for it to be an instance of lying”. But this doesn’t imply that Tom lied.

  19. For an act to be wrong is for the act to an instance of lying and Tom lied implies that claim. That conjunction is not not a normative claim and it entails the claim that you accept is normative. We can do this all day. You give a normative claim and I can always generate reductions and factual claims that entail it.

  20. “For an act to be wrong is for the act to an instance of lying and Tom lied”. Why isn’t this normative?

  21. You said above that reductions like Internalism 2 are not normative claims. All we have added here to a reduction like that is that Tom lied. A conjunction of two non-normative claims isn’t normative. Also, if you think that this is a normative claim then surely Internalism 1is.

  22. “A conjunction of two non-normative claims isn’t normative.” I reject this. (I have a paper on this stuff if you’re interested.)
    “Also, if you think that this is a normative claim then surely Internalism 1is.” Why? I don’t see this.
    Anyway, I’m afraid I’ve derailed this thread. Apologies for that.

  23. Campbell,
    Indeed, but you appear to have relied on the bad version: “if X implies Y, and Y is normative, then X is normative.” The Prior disjunction is normative, right?

  24. Jamie,
    No, the Prior disjunction is not normative, according to me.
    Prior’s argument refutes this version of HL: If X1, X2, … Xn jointly imply Y, and Y is normative, then at least one of X1, X2, … Xn is normative. But this is the bad version, not the one I was relying on.

  25. “…sometimes we have no idea whether a proposition is normative..even when we know exactly what it means.”
    This is quite plausible. Your friend comes off the phone to your mutual enemy saying “well, at least the last thing he said was true.” Whether what your friend said is normative plausibly varies with whether what the enemy said is normative. But you can know that what your friend meant, and even that what she said is true, without knowing what your enemy said.

  26. Good point.
    That seems different, since I know my friend has endorsed some proposition without knowing which proposition it is; his endorsement carries the force of that proposition. Whereas the Kiwi classification leaves the propositions in question completely transparent.
    In some moods I feel like it’s okay — when the non-normative disjunct is true, then the plain old non-normative world is enough to make the disjunction true, so that’s not really normative (the speaker accrues no normative commitment). But then sometimes it just seems crazy. If someone has merely described the world, how could a change in sipping habits ten thousand miles away make her assertion normative?
    The Kiwi-Normative also fails to close under conjunction, which seems pretty odd.

  27. Thanks Jussi!
    Brad,
    Another nice paper on mistakes like this is Horgan and Timmons, “Expressivism, Yes!, Relativism, No!” in the Oxford Studies in Metaethics Vol 1. I don’t recall them discussing reasons in particular, and so do not focus on the conflation of existence internalism for reasons and expressivism about reason discourse. But they do a great job with the genus of mistake of which that is a species, diagnosing it, helping people avoid it, and so on.

  28. Hi Matt,
    Thanks very much. I have never had a chance to read that, had forgotten about it, and am sure it must be good.
    If you, or anyone else who happens on this comment, knows of any good discussions of Gibbard on objectivity, I would love to hear about that too!

Comments are closed.