Re: poly: eschatology

From: Nick Bostrom <>
Date: Tue Jan 13 1998 - 15:59:46 PST

> From: Hal Finney <>

> In "classical" general relativity the three alternatives are as I listed
> them. They are referred to as Friedmann cosmologies, and assume that
> the universe is homogeneous and that there is no cosmological constant.
> If you introduce a CC then you can make a spatially closed universe
> which expands forever (just throw in an expansion term) or a spatially
> open universe which eventually recontracts (need a compression term).

I see. So what do the latest data indicate about the cosmological

> The introduction of inflationary theories adds some new twists. As I
> understand it, an inflationary model is technically not homogeneous.
> The universe forms as an initially small, rapidly expanding bubble off of
> a larger manifold. Near the center of the bubble the observed properties
> may be different from at the edges. (I don't understand inflation very
> well so take this with a grain of salt.)

And at least according to some versions of inflation theory (e.g. the
one by Andrei Linde, I think) it is also the case that there are
innumerous other spacetime bubbles beside the one we are living in.
I'm not sure whether there are supposed to be infinitely many of
these alternative universes. Even if the theory implies this,
one can of course raise skeptical doubts as to whether we could
actually have strong reason to believe in an hypothesis that
goes so far beyond what we can observe.

> > It makes a big difference to my world view whether the universe is
> > only big or whether it is literaly spatially infinite at the present
> > time. If it's infinite, then I can be sure that there are infinitely
> > many exact copies of myself in existence right now, and none of my
> > actions will make any difference to the total quantity of anything in
> > the universe that is not local: for example, I can't in the slightest
> > increase of decrease the total amount of pleasure in the universe.
> > This seems interesting enough that it is surprising that not
> > attracted more attention.
> This is an interesting philosophical speculation. You have a similar
> result in the many-worlds interpretation of QM.

Yes. There are different versions of the MW interpretation, though.
The usual way it is presented, it seems possible that at each point
of time there are only finitely many worlds (if the splitting process
started a finite time ago and there initially were only a finite
number of worlds). But it seems to me, among others, that the MW
interpretation doesn't get the probabilities right unless it
postulates an *uncountable* number of world-instances corresponding
to each eigenstate. (The reason is basically this: if the world
splits into two, how comes we usually find ourselves in the branch
which has the greatest amplitude?)

> Larry Niven had a story where people made contact with parallel universes,
> and it produced widespread feelings of nihilism and frustration, leading
> to a rash of suicides.

It would seem better to take Prozac. Philosophical causes for despair
or exhaltation are powerless compared to pharmacological ones.

> The way I look at this case is to define my consciousness, my sense of
> self, to encompass all those instances throughout the universe which are
> in the same state (to the limits of perception). As time goes on, there
> is divergence, as fluctuations or differences in the various instances
> begin to make their presence felt. This is analogous to the splitting
> of universes in the MW interpretation.
> In this case you could still hope to make a difference in the universe,
> as your decisions affect more than the local instance of you, but all
> the infinite numbers which share your mental state.

Interesting, but I'm not sure it makes any difference: Even if there
are a countable infininity of Hal:s, and you can control all
of them, you could only affect a countable change in the number of
pleasure-unites the universe contains. But is there are other people
also, with similar abilities, then the total quantity of pleasure and
pain (and other non-local quantities) would all seem to be countably
infinite in any case, with or without your contribution.

Come to think of it, it is not unproblematic to measure these things.
For example, if suffering has negative value, and pleasure has
positive value, and there is infinitely much of both pleasure and
suffering, what is the total value of the system? Addition for
infinite cardinals is not generally defined. One can't subtract
aleph-null from aleph-null.

Nick Bostrom
Received on Tue Jan 13 23:54:14 1998

This archive was generated by hypermail 2.1.8 : Tue Mar 07 2006 - 14:45:29 PST