Friday, May 30, 2008

What If You Are Special?

OK, so there's this way of thinking that goes like this: if you don't know whether you're "typical" or special with respect to some unknown thing, then you should assume you're typical.

Since, you know, odds are that you are typical.

So, for instance, when people talk about crazy shit like the end of the world, they use this reasoning to predict "with 95% "confidence", that humanity will disappear within 9120 years." That's from the Wikipedia entry on Doomsday arguments. The idea is that you and I probably aren't among the very first people to live, and we're probably not among the very last people to live. Probably. So, probably, we're somewhere in the middle. Which if you add in a few
assumptions, leads to the 9120 number.

Then also there's that book I wrote about before on C and C, Stumbling on Happiness. There, Dan Gilbert says that since most people are really bad at predicting how happy or unhappy something will make you, you should inform yourself by asking how happy or unhappy that same thing made other people.

So, if you're not sure how upsetting it will be when your team loses, or how happy you'll be when you get a raise, you should just check with someone else whose team lost and who got a raise.

This assumes you're pretty typical, which, probabilistically is probably a pretty safe assumption. But what if you're not? If you're not typical, there's no way to reason yourself out of a clearly false conclusion.

So maybe you know that, statistically, almost everyone is happier being married. It seems to follow that in the absence of knowledge of your own specialness, you too should assume you will be happier married. But if you're in the minority of people who just isn't going to be happier being married, well, there's no way you could factor that in. There's no way not to come to the wrong conclusions.

It almost makes you think there's something fishy about the whole thing. I mean, if I don't know anything about how typical I am, how can I assign a probability to my own typicality? Maybe I'm so underinformed about my own typicality that I should just assume ignorance.

In that case, coming up against questions like whether to get married, you'd have to guess, you know, or punt. Kind of wing it. At least then there would be the off chance that if you're atypical, you'll make an atypical decision. And maybe live happily ever after, who knows?

Unless, of course, the world comes to an end.

No comments: