From: gts (gts_2000@yahoo.com)
Date: Fri Apr 25 2003 - 22:54:52 MDT
Ramez Naam wrote:
> From: gts [mailto:gts_2000@yahoo.com]
>> You must, at least according to DA theory, accept that
>> you, Ramez Naam of 2003, are a single random sample taken
>> from the population of all humans ever to live past or present
>> [or future].
>
> Why should I accept that I'm a random sample taken from the
> population of all humans ever to live past or present? I see no
> justification for that.
I don't want to get too firmly entrenched here because like you I have some
doubts about the DA, at least at an intuitive level. The deeper
understanding of this problem is something I've put on the back-burner until
I can find more time to devote brain cells to it. :) For that reason I am
hardly an expert on the subject compared to others here have who have
probably spent more time than me puzzling over it (see some other threads
here on the subject, past and present).
However the assumption you question is called the SSA (Self-Sampling
Assumption). The SSA does not itself seem unreasonable to me. It is defined
thusly: "Every observer should reason as if she were a random sample from
the set of all observers in their reference class."
Put simply, why should you assume that your observation of any particular
phenomenon should be different from anyone else's? For example when you do a
statistical test on a coin to determine if the coin is fair, you assume that
your observation of the coin's flips is a random sample of coin flip
observations taken from the population of all possible coin flip
observations of that coin that might be taken by all possible humans; i.e.,
you assume that your observations are a random sample from the set of all
observers in your reference class with regard to that coin, where your
reference class is "humans" or "intelligent beings." If you don't make that
assumption then you cannot say that your analysis of the coin is meaningful
to other humans.
Now, you argue that people in the distant past who accepted the DA would
have been wrong. But consider for a moment that if the DA is true, the
number of people who would have been wrong would have represented a
relatively small % of humans ever to have lived. This is so because the
population of homo sapiens has been growing exponentially through time, or
roughly so.
In other words, if 60 billion people have so far lived, (a reasonable
estimate), and if doomsday is in fact imminent, the chances of any random
person taken from those 60 billion persons making a wrong prediction
regarding doomsday in their time is rather small, because most of those
observers live or lived in modern times rather than in ancient or
prehistoric times. If the argument is true then it is much more likely that
you will randomly select a correct DA predictor from the history of all
humans.
> Let me put it more formally. A good hypothesis is
> falsifiable. So how would we falsify the Doomsday Argument?
> Wait some arbitrary period of time to see if the world really
> ends? How long would we have to wait before deciding that
> it's moot? 1 day? 1 year? 1 billion years?
I don't think it can be falsified through experiment. It's a mathematical
argument open to falsification only through reason alone.
As I interpret it, (and again I am no expert), all the DA really can (try
to) say is that whatever your estimate about the probability of doomsday was
*before* you considered the argument, it must now be revised toward a more
pessimistic view *after* you have considered the argument.
-gts
This archive was generated by hypermail 2.1.5 : Fri Apr 25 2003 - 23:03:26 MDT