Is Meta Always Bettah?

In that New York Times piece on the new paternalism that I linked a couple of weeks ago – you know, the one that cites me – the author relies on the notion of meta-preferences, or preferences about preferences, as reason to favor your farsighted self over your shortsighted self. This paragraph appears just after the one in which I’m cited:

If the goal is to promote freedom, though, there is an interesting argument favoring the long-run self. A distinctive quality of humans, as the third earl of Shaftesbury observed three centuries ago, is that we do not simply have desires; we also have feelings about our desires. Take the unhappy heroin addict: he gives himself an injection because he desires the drug, but he also has a desire to be rid of this desire. The philosopher Harry Frankfurt has given such "second order" desires a central role in his analysis of free will: we act freely, he submits, when we act on a desire that we actually desire to have, one that we endorse as our own. Beings that do not reflect on the desirability of their desires -- like animals and infants and, perhaps, our short-run selves -- are what Frankfurt calls "wantons."

Frankfurt is right to say the capacity to have preferences about our preferences is a key characteristic of being human (though I’m less sanguine about the whole “free will” thing, inasmuch as I’m not sure it exists). But it’s a non sequitur to conclude that meta-preferences are somehow superior to regular (first-order) preferences. The key question is why we have such meta-preferences. I think the most likely answer is that meta-preferences reflect the opportunity costs or forgone pleasures of alternative choices we could have made. If you prefer A to B, and therefore you choose A, that doesn’t mean B ceases to be desirable. You might experience ongoing dissatisfaction from the absence of B, and that dissatisfaction could manifest itself as a meta-preference. In no way does that mean you really (or really really) prefer B.

Meta-preferences do reflect the existence of within-person conflict, and such conflict itself can be a source of displeasure (a point often made by economist David George). But the conflict can be resolved in more than one way. Your regular preference could change to match your the meta-preference. Or your meta-preference could change to match your regular preference. Either route would eliminate the internal conflict. The analytical error here is closely akin to that of new paternalists who observe the internal conflict represented by differing rates of time preference (the phenomenon of hyperbolic discounting) and then, without justification, assume that the conflict should be resolved in favor of the more distant rate of time preference, the one that places more weight on the future. The mere existence of a internal conflict cannot tell us how it ought to be resolved.

To drive home the point that there’s nothing necessarily special or laudable about meta-preferences, it’s worth noting that meta-preferences often take the form of socially approved preferences. Philosopher Marilyn Friedman gives the example of a woman who prefers careerism over homemaking and therefore chooses careerism, but then expresses the wish that she preferred homemaking over careerism. But a recent Dan Savage column provides an even better example: the gay man who doesn’t want to be gay. The example was offered in the rather different context of a guy who keeps saying he wants be cuckolded but then says he doesn’t want it after all. Both examples are instructive:

No one begs to be f***ed quite as sincerely, graphically, or desperately as some frat boy who hasn’t yet reconciled himself to being gay.

But, oh, the moment a closet case gets what he came for – the moment he comes – his tone changes dramatically. Not only does he stop begging to be f***ed, he will deny he ever wanted to be f***ed in the first place. ...

Like those “straight” frat boys I f***ed back at the University of Illinois, your boyfriend wants it. He wants you to f*** around with another guy, preferably in front of him. But he doesn't wanna want it and wishes it would go away.

So if a closeted gay man has a meta-preference to be straight, should we conclude the meta-preference is somehow better? Should he take one of those de-queer-ifying courses advertised by the Christian Right in order to change his regular preference? Or should he just reconcile himself to being gay, thereby changing his meta-preference? To me, the latter seems like the better option. The meta-preference is probably just an internalized version of a discriminatory social norm, and he’s better off without it.

I don’t mean to imply that every meta-preference ought to be ditched. Maybe the guy with the cuckolding fetish really ought to get over it, for the sake of his unwilling girlfriend. Then again, maybe he should find a different girlfriend who shares his fetish. Sometimes the regular preference should prevail, sometimes the meta-preference. Simply pointing out our capacity to form meta-preferences is not sufficient to make a general statement about their superiority.