My first reaction to a recent exchange between Brad DeLong and Will Wilkinson was simple befuddlement. In response to Wilkinson’s (I thought unexceptionable) assertion that people value things other than—and often more highly than—happiness, DeLong objected, not just that Will had said something substantively wrong, but made some kind of semantic error, asserting a tautological falsehood (what we used to call a “falsism” in debate):

The response–against which Wilkinson has no defense except to issue squidlike clouds of obfuscating ink–would be that Wilkinson believes that if he were to sacrifice his freedom for his happiness, that if he were to do so he would then look back on the choices he made and look ahead to his future life, and that he would be unhappy. If Wilkinson says otherwise–that he would look back on the choices he made and look ahead to his future life and be happy, but that he would still regret what he had done and wish he had done otherwise–Wilkinson is simply saying, “Baa baa buff.” He would be demonstrating that he does not understand the rules of conversation using the English language.

Now, as Will responded, this just seemed strange. I think we’d find it perfectly intelligible for someone to say (borrowing an example from Derek Parfit): “I’m making myself unhappy in many ways slaving over this novel I’m trying to write. And indeed, if I went and did something else instead, I can predict I wouldn’t feel myself torn by pangs of regret. I would, in short, be happier. But I think that this is such worthwhile thing to do, I’m willing to accept being less happy for it.”

I’m thinking that what’s going on here is that DeLong is using a theoretically freighted sense of “happiness” here that’s tied up with his conflation of Bentham’s hedonistic utilitarianism (pushpin is as good as poetry and all that) with later preference-based versions. Let me attempt a very rough oversimplified version of the theoretical history here. You start with the view that what’s morally good is to maximize “happiness” in the sense of subjective good feelings (orgasms, love, the feeling of a job well done, the joy of exercising one’s capacities, etc.). There are some obviously appealing features of this view. First, “happiness” seems like the right kind of candidate to bridge that infamous fact/value gap: Utilitarianism urges the promotion of states of affairs (facts) that are also subjective states we all seek (values). That happiness is good and pain is bad has the comforting feel of tautology that avoids tangled puzzles raised by comparable assertions about, say, violating people’s rights. Second, it has a fairly straightforward maximizing structure that allows competing claims to be weighed in the same balance. Actually doing the weighing may still be quite tricky, but we know in principle what we’re trying to get at, whereas other theories may run into puzzling conflicts (or at least apparent conflicts) between various rights and interests.

The trouble is that the appeal of utilitarianism as a normative theory rested on a psychological premise that turned out to be pretty dubious: That what everyone really values most is their own happiness, and that all our purposeful behavior is really directed toward this goal. This required some rather dodgy moves: If I give a few bucks to a homeless guy, on this theory, I’m really just seeking the psychic satisfaction I’ll get from it. But, of course, that leaves the question of why I’ll feel that satisfaction: Isn’t it because I think it’s an independently worthwhile thing to do? In short, it became apparent to many people that we don’t just care about how we feel subjectively. (That’s the point of Nozick’s Experience Machine, of course.) Indeed, we may have strong desires that things go a certain way even after we’re dead, and not around to experience anything.

No problem, say the utilitarian theorists (and economists), we’ll switch to preference utilitarianism, wherein “utility” or “happiness” are defined in terms of the satisfaction of preferences. This has the virtue of realigning the object of maximization with what people subjectively value, which makes for a sturdier fact/value bridge. It has the disadvantage of making it much less clear whether that appealingly simple maximizing structure is still a good fit for the task. Even rendering the very different forms of satisfaction and dissatisfaction people are capable of feeling comparable seemed a bit of a stretch. (How many of the bon vivant‘s wild nights on the town does it take to equal the same amount of “utility” in the monk’s serene satisfaction in a day of contemplation?) Preferences over states of the world that need not be experience by the subject who prefers them make things a much bigger tangle. And should we think it’s better for people to have more (and more intense) preferences, so that more of them can be satisfied? But I digress.

DeLong’s problem seems to be that he’s using this second, preference-based sense of “happiness.” But this isn’t the colloquial sense of the word. In effect, DeLong is like a Ptolemaic astronomer who, confronted with the proposition that the earth revolves around the sun, exclaims that this is equivalent to the absurd proposition that the earth doesn’t exist, since “the earth” is defined as the thing at the center of the universe.

Now, if we’re defining happiness in terms of preferences, then DeLong’s semantic objection isn’t as wacky as it initially seemed: The claim that freedom might be more important than happiness amounts to the claim that one prefers freedom to the satisfaction of one’s preferences.

But DeLong then goes on to argue in terms of “regret” and “being happy” when contemplating one’s future life, which certainly sounds as though he’s conflating the preference-based and experiential senses of happiness. Note also that DeLong’s shifting the focus from how one feels in one’s future life to how one feels now about that future life. Now, imagine I’m a soldier deciding whether to jump in front of a grenade that will surely kill my less-well-armored fellows. I suddenly have a vision of my rough future life if I do it, and am maimed in the explosion. I simulatneously have a vision of my future if I don’t do it: Perhaps I feel some initial guilt, but it doesn’t torment me. By any ordinary measure, I am far happier (both in experiential terms and in terms of the extent to which I can satisfy my future preferences) in the second scenario.

Actually, reflecting on this vision of a future life introduces a certain amount of unnecessary confusion: Someone might well be willing to leap on the grenade even if it killed him—no future life. So the discussion further blurs the line between a technical, preference-based sense of happiness in which DeLong’s claim sounds right, with happiness as a sort of feeling about how things will go for you. But if we want to stick with the preference-based version of “happiness”, then (and this certainly runs against the grain of ordinary English) it no longer makes much sense to talk about “being happy”, since “happiness” might be maximized by the choice that leads to my being killed instantly, or by events transpiring after my death (one of my great life projects is carried on and comes to fruition after I die—or doesn’t).

Now, I haven’t read Layard’s book, but I find it pretty implausible that this very techical sense of “happiness” is the one he intends. The book is described as making use of a series of psychological studies about how satisfied people are. And it’s only the experiential sense of “happiness” that’s going to turn up in studies like that: First, because if you ask people how “happy” they are, then everybody but Brad DeLong is going to interpret that question in the colloquial way. Second, because asking people about their happiness can only yield information about happiness in the subjective sense. Imagine two guys, Adam and Bob, who both believe their wives, Alice and Bea, to be loving and faithful. Alice actually is, but Bea secretly holds her husband in contempt. Now, experientially, these guys might be perfectly identical, and they’ll report comparable levels of happiness. Yet on a preference-satisfaction definition of “happiness”, they’re night and day: Though he doesn’t know it, the world doesn’t match up very closely at all with the way Bob wants it to. So this can’t be the sense of happiness that Layard is talking about.