I am grateful that my post  on Ethics Etc. finally reached David Sobel, and that he has taken the trouble to respond to it . On the face of it, his response looks pretty devastating – I am sure it must seem so to him. But I offer the following rejoinder:
I suppose I should have anticipated this reply, given the way Parfit presents what he calls ‘the agony argument’ and the way in which Sobel in his article exploits the notions of liking and disliking. Nor can it be my aim to speak for Parfit or defend the letter of his texts (he is surely more able to do that for himself than I could possibly be on his behalf). My aim, however, was to criticize the strategy Sobel uses in defending subjectivism about reasons against Parfit’s objections. In a short post, it was impossible for me to do this in a way that avoids misunderstanding. For this reason, the present note is quite a bit longer, because it will take longer to explain what I believe to be at issue. I agree that Sobel’s reply seems quite apt in relation to the letter of what Parfit says. However, this reply also perpetuates the fatal confusion on which I believe the argument of Sobel’s article rests.
It is quite possible to distinguish between “the way a sensation feels” or its “intrinsic qualitative features” (in one sense of the term), and whether it is pleasant or painful. That is what Parfit is doing in the passage Sobel quotes. There is a distinction between “the way a sensation feels” (or its qualitative properties), in this sense, and whether a particular subject finds it pleasant or unpleasant. For example: Some people find certain textures of cloth against their skin pleasant, while others find these same sensations annoying. Parfit apparently finds the taste of milk chocolate unpleasant (disgusting), whereas many people find it pleasant. (Honey meringues seem to be even worse.) I think that is the sense in which Parfit means that the intrinsic qualitative features of sensations give us no reasons to like or dislike them. However, the fact that we like or dislike a sensation (the feel of such-and-such fabric, the taste of milk chocolate) can also be regarded as (in another sense of the phrase) part of the way it feels to us. We have no reason to like or dislike the sensation in this sense, but the fact that we do like or dislike it could give us an object-based reason for seeking or shunning the sensation. That is the sense in which the way pain feels gives us a reason to avoid it or prevent it, if we can, and also to take certain negative attitudes toward it.
The notions of liking and disliking, however, lie on the cusp between (a) mere brute facts about the way things are experienced — which facts might give us reasons for conative states or attitudes — and (b) conative states or attitudes themselves, which might have, and sometimes even require, reasons for us to be in those states or to take those attitudes. I would like to say that the terms ‘like’ and ‘dislike’ and their cognates, when used in sense (a), are being used in their ‘merely factual’ sense, or their ‘basic’ sense. But such terms can also be used in sense (b), which I will call their ‘conative’ sense.
The fact that Parfit dislikes the taste of milk chocolate, in the factual sense (a) might give him a reason for being in (b) the conative state of dislike or aversion toward eating milk chocolate, and then it might also give him an object-based reason for preferring or desiring not to eat milk chocolate. ‘Dislike’ when it refers to this conative state, is not being used in its basic or merely factual sense, it is being used in a different sense. Dislikes in the basic or merely factual sense can sometimes provide reasons for conative attitudes. I think Parfit takes our dislike of pain to be such an object-given reason for shunning sensations we find painful, because we dislike them in the basic or factual sense. But some likes and dislikes in this basic, factual sense merely motivate likes and dislikes in the conative sense, without providing reasons for them (or at any rate, without providing good reasons).
There is one single unique work of great literature which always contains somewhere within the rich texture of its drama and dialogue an apt illustration for almost any philosophical point, no matter what the point may be. No, it is not a work by Shakespeare, or Sophocles, or Goethe. It is The Importance of Being Earnest, by Oscar Wilde. In this case, the illustration is this:
In the final scene, Lady Bracknell finds out that ‘Ernest’ was the name of her deceased brother-in-law, the father of Algernon and (as we have just discovered) also of the erstwhile ‘Jack Worthing’ – the brother-in-law Lady Bracknell describes as a general who was “essentially a man of peace, except in his domestic life”. Her reaction to this information is: “Yes, the General’s name was Ernest. I knew I had a particular reason for disliking the name.” I submit that if we take her statement literally, she cannot be using ‘dislike’ in the basic or merely factual sense; she must be using it in the conative sense. For there can be no reason for disliking anything in the merely factual sense. At most there might be a causal explanation for this (e.g. the association of the name ‘Ernest’ with her late brother-in-law and his quarrelsome behavior). There might also be a causal explanation of why she dislikes the name in the conative sense. But she could have no reason – no good reason — for disliking the name in this sense. But of course Lady Bracknell is typical of her social class: she is arrogant and unreasonable enough to think that her likes and dislikes in the merely factual sense always give her legitimate reasons for any conative states or attitudes they might happen psychologically to motivate.
One trouble with subjectivism is that it might tend to license such an uncritical attitude toward one’s merely factual likes and dislikes, perhaps confusing them with the conative states they might (irrationally) motivate. Then these unreasonable conative states might be taken to provide reasons for further conative or attitudinal states (e.g. Lady Bracknell’s unwelcoming attitude toward her newly revealed nephew Ernest as the imminent marriage partner for her daughter Gwendolen). But – and here’s the point — conative states never provide reasons for the actions or attitudes they might motivate except insofar as we have independent, object-based reasons to be in these conative states in the first place. This is the decisive argument against the sort of subjectivism about reasons that Sobel is trying to defend. It is one of those arguments for which Lady Bracknell also expresses dislike: “I dislike arguments of any kind. They are always vulgar and often convincing.”
With apologies to Oscar Wilde, let’s now switch examples, descending from the heights of dramatic sublimity to the tawdry lives we academics lead in our petty everyday dealings. Suppose Sobel and I were colleagues in the same philosophy department; at the APA our department is interviewing Thomas Twit, a job candidate. After the interview, the Department takes a vote on whether to invite Twit for a campus visit and I vote no. When asked why I did so, I say: “I dislike him”. This might be merely a report of dislike in a factual or basic sense, and it might have a causal (psychological) explanation – for instance, Twit has a superficial mannerism that reminds me of someone I used to know who mistreated me. At this point, Sobel, who favors for inviting Twit for an on-campus interview, might legitimately ask me: “What reason do you have for disliking Twit?” If all I can say is what I have just reported, then Sobel could rightly reply: “That’s no reason at all for disliking him!” What he might mean by this is: that you dislike him in the merely factual sense gives you no reason for disliking him in the conative sense. And Sobel would be 100% right. I have no reason for voting against Twit. I shouldn’t have done so merely because I dislike him merely in the basic or factual sense. That gives me no reason for disliking him in the conative sense, and no reason for voting against inviting him to campus. If a subjectivist theory of reasons yields a different result (because it holds that my disliking him counts as a reason for the way I voted) then that theory has to be wrong.
But now consider a variant on this story. Suppose I reply: “The way he answered Professor Poynter’s objections was dogmatic, evasive, and intellectually dishonest.” If what I said was true, then that would provide me with a good (object-based) reason for disliking Twit in the conative sense. If Sobel decides on reflection that what I have said was true, then he might (and even should) agree with me that I have a good reason for disliking Twit and voting against inviting him. Indeed, if Sobel decides that what I said with true, that should give Sobel himself a good reason for changing his vote. It is the nature of genuine (objective, object-given) reasons that they are capable of doing this, in the way that subjective reasons never can — except perhaps contingently, based on the fact that people happen to share their (objectively groundless) likes and dislikes (in the conative sense).
In his criticisms of Parfit, I think Sobel is trading on the ambiguity between the merely factual or basic and the conative senses of ‘like’, ‘dislike’ and their cognates. When Parfit says: “Whether we like, dislike, or are indifferent to these various sensations we are not responding or failing to respond to any reasons,” he is (or at any rate, he ought to be) using these terms in the basic or merely factual sense. It is in this sense that disliking pain (along with, but in addition to, what Parfit calls the ‘intrinsic qualitative features’ of the sensation) is part of “the way pain feels.” That was the sense in which I used that phrase ‘the way pain feels’ in my earlier post. “The way pain feels” in this sense is a mere fact about it, which involves basic or factual dislike, and does provide us with object-based reasons for avoiding pain. It gives us such reasons because it also gives us object-based reasons considering pain bad — and also reasons for disliking pain in the conative sense. For disliking, in the conative sense, is very close to, if not equivalent to, a state of desire or aversion – in this case, the state of aversion to pain. But disliking pain in the basic or factual sense is not a state of desire or aversion. It is a fact about the way pain feels (in the broad sense I have just explained), that might give us an object-based reason for disliking pain in the conative sense.
Sobel writes: “In my paper I discuss this case Parfit makes for this difference and explain why I think the difference Parfit points to between likings and desires is unconvincing.”
It would indeed be unconvincing for Parfit to try to draw a distinction between likings and desires if ‘likings’ is used in the conative sense, since a liking in this sense is very close to a desire. But it is not in the least unconvincing for him to distinguish liking from desire, or disliking from aversion, if ‘liking’ and ‘disliking’ are used in the basic or merely factual sense. I think Sobel’s argument trades on this ambiguity. In my post, I avoided using the terms ‘like’, ‘dislike’ etc. because they are dangerously ambiguous in the way I have just been trying to explain.
When Parfit speaks of ‘disliking’ pain, using the term (as I would read him) in the merely factual or basic sense, it looks to me as if Sobel takes him to be countenancing the thought that likes and dislikes (in the conative sense) can provide reasons. But for Parfit to have done that would, in my view, have been a mistake. Parfit should have kept distinct the following two things:
1. The true (objectivist) claim that certain facts (e.g. about the way pain feels, including our dislike of it in the basic or merely factual sense) give us reasons for certain desires
2. The false (subjectivist) claim that conative states, which we might be in for no reason at all, could provide us with reasons for doing things, or for being in further conative or attitudinal states
For instance, he should keep my dislike of Thomas Twit (caused by the way his superficial mannerisms remind me of someone else) distinct from my dislike of Twit based on the dogmatic, evasive and dishonest answers he gave to Professor Poynter’s objections. The former is a merely factual state, which may or may not give me reasons for the latter conative state – in this case, they give no reasons, or at most bad reasons for it. But the latter dislike is itself a conative state, which is based on certain (possibly quite good) reasons. The fact that a philosopher is dogmatic, evasive and dishonest in discussions might be a good reason (an object-based reason) for not liking him (in the conative sense) and also a good reason for not wanting to consider him further as a job candidate.
Summary: Subjectivism about reasons is false because no conative or attitudinal state (including likes and dislikes, when these terms are used in the conative sense) ever provides a reason for acting or for anything else unless we have a reason to be in it. But there are facts about our sensory states (which include likes and dislikes in the basic or factual sense of those terms) which sometimes provide reasons for us to be in certain conative or attitudinal states (that can also be described as likes or dislikes). However, these factual likes or dislikes sometimes merely motivate the respective conative likes or dislikes, providing no reasons (or only bad reasons) for them, and therefore also no reasons (or only bad reasons) for further attitudes or actions motivated by those states. Sobel’s critique of Parfit trades on a fatal ambiguity in terms such as ‘like’ and ‘dislike’, making subjectivism about reasons seem plausible when in fact it is nothing of the kind.