Monday, July 11, 2011

Are Pleasant Placeholder Beliefs Acceptable?

I'm interested in getting reactions to the idea, or the advisability of, limited and correctable unverified beliefs. I think most of us in the community would reject the idea of holding non-provisional beliefs which we know we don't have the ability to falsify. (I'll give a concrete example in a second.) However there's an assumption buried in there that may admit of exceptions to this.

Here's the concrete example: say you're out running. You're really thirsty. You can see your turnaround point coming up a mile or so away. You see what might be a water fountain, but until you get to it you won't be able to tell if it actually is one. You (okay, I) think the following: whether or not that is a water fountain, it's not going to change my decision to keep running, and my life and health are not in danger. However, if I believe that it's NOT a water fountain, and it's not, I will be miserable for the rest of the run, starting right now. If on the other hand I believe that it IS a water fountain, and it's not, I will at least have been miserable for five fewer minutes (the time between right now and when I get to what turns out to be a non-water-fountain.) So in my silly head I'm going to decide that it definitely is a water fountain, and then update my non-decision-or-health-affecting belief once I get to it. This seemingly trivial example can make some people cringe: I'm adopting a belief, that I have no way of verifying, AS TRUE - just because it makes me feel good.

The reason I'm spending any time on this seemingly trivial example is because it exposes a buried assumption about why beliefs matter and what they do. As a working definition: "Beliefs matter because they affect the world beyond our central nervous system." If the belief is not affecting your decisions, and it's not changing anything else in the world outside the experience of pain or pleasure happening inside your own skull, AND it's subject to eventual updating anyway, what's the problem, morally or psychologically? (To be clear, not only is it not changing my decision-making, I'm definitely not about to try to convince anybody else that there's a water fountain there until I actually see it.)

I've asked people this before and they point out that we could consider the water fountain situation a form of hypothesis testing. That said, people in skeptical communities are less uncomfortable with true hypotheses, which you might argue differ in purpose from my water fountain belief. Why? Because of the attitude we take toward the not-yet-evaluated belief, and the experience produced by that attitude. Hypotheses (in science and elsewhere) are explicitly provisional, adopted for maximum clarity of decision-making. For one thing, with hypotheses there's no thought (supposedly) given to which one is the most pleasant to believe, versus the example above where it seems to me that beliefs in those limited, non-action-affecting circumstances can be adopted entirely for the purpose of the internal experience they produce. It's also possibly important that because the hypothesis is explicitly provisional (you KNOW you don't know whether it's false the whole time), the experience of holding the proposition is different than one you think is actually a reflection of reality. The state of the water fountain belief, while I'm still running toward what might be a water fountain, is stronger than just "I prefer this hypothesis to be true", because that wouldn't do the trick. Of course the degree to which we have control over believing something is relevant here but it's a separate question (Dennett has written a lot on this).

In closing, an easier example: I once took two baby birds that fell out of their nest in front of my house and made a little artificial nest for them (full story here). I saw one of them fly away, but as night fell the other was still in its new nest. In the morning it was gone. I've since decided it flew away, because this makes me happier than the alternative. I will almost certainly never know whether it really did fly away, or something came out of the canyon behind my house and ate it. For this reason (its unfalsifiability), I think this example is even less objectionable than the previous one. Why? Should this situation present itself again my previous experience and my unverifiable belief will not inform my decision. And if some sadist shows up at my door with night-scope film from that evening of a coyote getting in there and chowing down, I'll have to say fine, it got eaten. Still, questions perhaps better addressed by cognitive science take on greater relevance. Is it really possible for us to know that a provisional belief (even a silly for-temporary-happienss one) isn't affecting us? Is it even possible to really hold a belief provisionally (if not, we're kidding ourselves in science when we say "we don't [think we] know" yet because we haven't done the experiment)?

See, I bet you'd like to believe that the one on the right flew happily away too. And since you'll never know and aren't about to apply the belief to rescuing other baby birds, why not?

On Lesswrong they call ideas or theories that it's bad to repeat (thereby confusing or sucking in other people) "basilisks". Maybe these seemingly more harmless ones would be "salamanders"?

1 comment:

Anonymous said...

Or you could stop tricking yourself into having to have assumptions about water fountains and either:
A. plan your route better and know where the fountains are.
B. carry enough water with you such that you don't need fountains.
C. realize that running equals abject misery in the first place! (Happy while're a madman!)