On Sat, 8 Jan 2000, Fred Chen wrote:
> Jacques Mallah wrote:
> > That's not true. The population size distribution is determined
> > by the effective probability distribution for the various histories that
> > could happen. We don't know what that distribution is.
>
> I may have been unclear here. Rather than population size at a given
> moment, the population size should be taken to be the number of people
> since the dawn of time.
And until the end. It was understood.
> True, we don't really know what the distribution is, but the natural
> expectation is that since events that can wipe out mankind at any given
> instant are still considered rather unlikely, the probability
> distribution, whatever it is, is expected to favor longer-lasting
> populations.
Here you are invoking additional information, so you are really
_not_ addressing the 'doomsday argument' which explicitly says, in the
absence of additional information, ...
Besides, it is not so clear that such events are so
unlikely. Disease, nuclear war, asteroids, etc. could all do the
trick. But I do agree that with the information we have it is reasonable
to modify the prior distribution so as to expect larger populations than
one would predict just from birth order.
> > You are not likely to be the
> > last person ever born! You are probably around the middle of the pack,
> > which is the point.
>
> Okay, this depends on the population growth pattern, and for some reason I was
> envisioning a population that grows linearly or faster up to a certain point
> (N=10^14, or 10^11, or whatever), and terminates. Then the
> instantaneous probability of finding yourself in a particular order
> becomes more likely the later you appear.
Actually it doesn't matter. The point of maximum probability
density is of no significance, what matters is that the region of birth
order between, say, 10% < you < 90% is quite likely.
> But this is indeed very artificial and in fact does not reflect the real world's
> behavior either, as I'll mention below.
>
> > > (Now, the current world population growth rate is ~5 people every 2 seconds.
> > > So each moment we find ourselves in a less and less likely position in the
> > > history of the human race.)
> >
> > Again, the above make no sense.
> > In general, in a period of exponential growth, would would expect
> > that growth to reverse within a time constant or so. Of course, we do
> > have some additional information that modifies that in this case.
> >
>
> Well, the population is indeed still growing, but the rate may not be increasing as
> much as it used to (which is probably very good for the planet). It's 6 billion now.
> In 1960, the population was 3 billion. In 1800, it was 1 billion, most of whom died
> before the 1960 count. So there were at least ~4 billion to account for between 1800
> and 1960. If the world ended in 2000, we will find ourselves 'exceptionally' late
> (past the median) in the group between 1800 and 2000. And if you include those before
> 1800, that only makes us appear even later. (Source:
> http://ssl.zpg.org/popframe.htm). For us to be in the middle, the world has to keep
> on going for maybe at least another century. But then what will our children
> conclude?
Of course our children *should* predict a later "doomsday",
because the information they would have on birth order is, by definition,
different from the information we have.
For example, suppose population was constant for 10^6 years. We
would then expect it to remain so for about another 10^6, betting that
sometime around then doomsday would occur. But if we were born around
2*10^6 years, our expected timescale would double. It would keep
increasing until, one day, some of our ancestors would be proved
right. This end would come as a surprise to those alive at the end, but
this betting proceedure would maximize the fraction of people whose bets
were in the right ballpark over all time.
- - - - - - -
Jacques Mallah (jqm1584.domain.name.hidden)
Physicist / Many Worlder / Devil's Advocate
"I know what no one else knows" - 'Runaway Train', Soul Asylum
My URL:
http://pages.nyu.edu/~jqm1584/
Received on Tue Jan 11 2000 - 12:42:16 PST