From: Nick Bostrom (nick@nickbostrom.com)
Date: Fri Oct 27 2000 - 23:00:40 MDT
Robert Bradbury wrote:
>I think I'm responding to Nick Bostrom, following some discussion
>by Jason Thompson & Hal Finney, but I'm not sure given the
>forwarding this went through.
>
> >But finding a second source of intelligent life, in our relative vicinity,
> >would dramatically boost the support for those theories which say that
> >intelligent life is common. (My PhD thesis at
> >http://www.anthropic-principle.com/phd has more on this and many other
> >related topics.)
>
>I think you may need to refine the term "intelligent life" a bit.
Yes and no. No in the sense that for this particular application I don't
think it makes much difference how you define an "observer", namely if you
think that the real difficulty was getting to complex organisms (like
vertebrae) and that the step from there to human-level intelligence was
*relatively* easy. But yes in the sense that there are many other cases
where how you define "observer" will make a big difference to what results
you get when trying to account for observational selection effects. The
more general problem here is how to define the reference class. (Remember,
the Self-Sampling Assumption says you should think of yourself as a random
sample from all observers in your reference class, so you need to know what
that class is.) The precise solution to this is not currently known (though
I have some ideas that I haven't yet articulated that possibly might solve
this.)
> >For example, we can compare
> >the time it actually took for intelligent life to evolve and compare that
> >to the time that was available (an upper bound on which is given by the
> >time when the sun engulfs the earth in about five billion years).
>
>Hsssss... The sun engulfing the planet is a very poor limit unless
>you are taking a very anthropic view of what "life" must look like.
>The habitable water zones shrink as the stars get smaller (K & M
>class), but they are much more numerous and live much longer. Free-formed
>or wandering planets (flung out of solar systems by gravitational sling-shots)
>could still evolve life based on energy from gravitational contraction
>or radioactive decay if they formed in regions significantly enriched
>by SN explosions with radioactive elements. While the matter density
>of gas clouds is low, we know they are seedded with the building
>blocks of life, have energy sources in nearby stars and have very
>large (>> solar) masses (which may offset their low density).
I'm talking about the cut-off for the interval during which our earth could
evolve life. Presumably you agree that that is impossible after earth has
been evaporated and been absorbed by a red giant, don't you agree? (If
posthumans were to construct some kind of hardy life form that could
continue to evolve after that point, this obviously doesn't count, since
we're talking about the first point when intelligent life evolved, not how
long is can survive.)
As for the other life forms you mention, although they would be irrelevant
to the point I'm making, I also happen to think that they are even more
extremely improbable. If evolution of intelligent life is so hard even on a
nice planet like earth, then I can see no reason to think that it is a
frequent event in less hospitable environments. But that's a sidetrack.
>Going back as far as Shklovsky and Sagan, there are concepts
>of civilizations advanced only slightly beyond ours, to reengineer
>their star to extend its longevity. There are also concepts for moving
>the planet. So, any limits on the maximum longevity of technological
>civilizations are very questionable.
Sure, but that's irrelevant for the point I was making (see above). To
clarify, the argument is this: Trace back to the starting point of our line
of intelligent life. It happened apprx 4.5 billion years after the Earth
became habitable. Also, we know it couldn't have happened later than after
apprx 10 billion years. Now, if there were many critical steps in the
evolution of intelligent life, then one should have expected it to have
appeared closer to the cut-off than it did. (This can be shown
mathematically - see e.g. Robin's paper http://hanson.gmu.edu/hardstep.ps.)
Conclusion: probably, there were not very many critical steps. (If we
assume that the real cut-off was earlier than 10 billion years, which it
might be because of climate change, then there could be more critical
steps. You need to assume either the number of critical steps or the true
cut-off time to be able to estimate the other.)
>record of? On other planets, these could have taken faster
>(or slower) approaches to the development of complex multi-cellular
>forms.
Yes. That's a separate issue.
> >..., except that all the evidence we
> >have is perfectly consistent with the hypothesis that it is extremely
> >difficult for intelligent life to evolve. This is also consistent with (and
> >in my opinion is the most probable explanation of) the absence of any signs
> >of extraterrestrial intelligence (the Fermi paradox).
>
>There is also a reasonable amount of evidence (IMO) to suggest that
>there is no Fermi Paradox. We are almost completely lacking in the
>ability to detect "life" at a distance, especially life at the limits
>of physical laws.
I agree that there is no Fermi "Paradox". A paradox is supposed to be where
there are compelling arguments for each of two incompatible claims. But
there is no compelling argument why life should be common in the universe.
I think there are some semi-compelling arguments for thinking that if
space-travelling life were common, it would have been here by now, but
that's been debated here a lot quite recently so let's not go into that at
this point.
>There are a lot more PhD thesis that need to be written about
>the possible end-points for technological civilizations and how
>we might recognize them.
That could be interesting. Unfortunately, it's not a PhD topic that many
departments would encourage.
Nick Bostrom
Department of Philosophy
Yale University
Homepage: http://www.nickbostrom.com
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:31:46 MST