Suppose someone mistakenly thinks he can fly by jumping off a bridge. He wants to fly, so he has a preference to jump off the bridge. Should we respect that preference even though it is not “informed”?

I think we should understand preferences as desires – specifically as desires that are compared in strength to other desires, so that a preference is the stronger desire. We can tell which desires are stronger, because people choose to satisfy them in preference to weaker desires.

So let’s return to the bridge. Suppose our potential birdman wants to fly south for the winter. In that case, his preference to jump off the bridge would disappear as soon as we managed to persuade him that that will not be the outcome of his jumping. We might also notice that he seems to have a preference not to die (yet). In that situation, we would surely be obliged to go to considerable lengths to prevent him jumping off the bridge. Our obligation stems from respecting his preference for flying south for the winter rather than dying immediately, and his preference to avoid death rather than dying immediately. Our remonstrating with him would be justified by our respect for other preferences. If we decided to forcefully overrule his preference for jumping off the bridge, it would be out of respect for even stronger desires that he already has – stronger than his desire to jump off the bridge. In other words, it would be out of respect for his preferences.

Preferences aren’t true or false. They’re not themselves informed or misinformed, but instead belong to an agent whose beliefs are true or false, so that the agent himself is informed or misinformed. Like Humphrey Bogart’s character Rick in Casablanca, he can be “misinformed” about the availability of waters in the desert, by having false beliefs about where waters are found, but his preference for such waters cannot be.

If an agent’s beliefs are false, he is less likely to satisfy his preferences. So if we respect those preferences, we are obliged to inform him, if we can. But sometimes we too are misinformed, and cannot inform him any better than he can inform himself. We should take all of these possibilities into account when remonstrating with someone, and we should avoid overriding his judgement just because we think our own judgement is better than his, which often amounts to an assumption of our own infallibility.

We might think of the birdman’s preference as “misguided” rather than false. But then we must ask what is misguided about it. I would argue that inasmuch as it’s an “intermediary preference” to be satisfied as a condition of satisfying something more important to the agent, it’s simply weak. We can tell it’s weak, because the agent’s own choices reveal that he would happily choose “alternative routes to his preferred destination” if they were available.

That should be familiar to us all: what we really want is to go to the movies, say, so we form the intermediate goal of taking the 7.20 bus. If we miss the 7.20 bus, we just take 7.30 bus, or a taxi – an annoyance rather than a thwarting of our main preference to see the movie. The preference to take the 7.20 bus is weak because it’s ephemeral.

We might think of a preference as “misguided” because it’s aimed at an impossibility. But that doesn’t strike me as a reason to think it’s misguided. We all have a preference for things that are statistically impossible, such as never to lose another hair on our heads, or never to be sick again. The preference to live indefinitely might be understood as the preference never to die. Inasmuch as these preferences are strong, they deserve respect regardless of the impossibility of their being satisfied.