From: Rafal Smigrodzki (rafal@smigrodzki.org)
Date: Tue Apr 29 2003 - 14:31:02 MDT
I'd like to repost my objection to the Doomsday argument from wta:
George Dvorsky wrote:
>
> Bostrom uses a great analogy, that of the ball dispensing machine(1).
> You know you have two machines: one that dispenses 10 balls, and one
> that dispenses 1 million balls, but you don't know which one is
> which. Each ball is numbered. You pull a ball from one of them, and
> you get the number 4. Any reasonable person would therefore assume
> that they pulled the ball from the 10-ball machine.
>
> This same reasoning is what drives the doomsday argument. Thus, as
> rational people, we should *expect* that we, as the 106th billion
> persons, are ball #4; it's far less likely that we're 106th billion
> out of 106 trillion than 106th billion out of 200 billion.
>
> But by the same token we should not take that to be an empirical fact.
>
### Again, you make some big assumptions here, regarding the numbers of each
type of machine or civilization, or, more precisely, their measures. If you
take the equal measure of ball machines, or civilizations, as your Bayesian
prior (i.e., you *know* that exactly 50% of them dispense 10 balls, or 10
billion people, and 50% dispense 1 million balls, or 10 million billion
people), you can easily plug the numbers into the Bayes' equation and
calculate the odds of a ball (or human) coming from the 10 ball machine (or
civilization) if the ball's (or human's) number is 4. No problem here.
In fact, if you really knew that 50% of civilizations die right after the
10-billionth person is born, you could treat the prediction of impending
doom as a "fact" (=the result of Bayesian reasoning with very high odds of
being true), because you would be more than 99.9% sure that you are going to
die.
However, it has no bearing on the issue of doom and humanity - we do *not*
know the measure of civilizations producing a total of 10 billion people vs.
the measure of civilizations producing 10 million billion people. If the
latter are a million times more common than the former, the odds of us being
in either one are roughly equal. Unless you use some quite complex reasoning
and a lot of empirical data, you cannot even begin to derive this prior.
We know little about civilizations in general. The Doomsday argument tries
to wring too much knowledge out of too little data, and is for the time
being still completely useless.
Rafal
This archive was generated by hypermail 2.1.5 : Tue Apr 29 2003 - 11:42:28 MDT