Re: Many worlds theory of immortality

From: Hal Finney <hal.domain.name.hidden>
Date: Mon, 18 Apr 2005 12:25:33 -0700 (PDT)

Jesse Mazer writes:
> >From: hal.domain.name.hidden ("Hal Finney")
> >Generally, I don't think the same logic applies to copying a mind in a
> >single universe than to splitting of worlds in the MWI. Copying a mind
> >will double its measure, while splitting one leaves it alone. That is a
> >significant practical and philosophical difference.
>
> Doubles its measure relative to who? If I am copied while my friend is not,
> perhaps it makes sense that my measure is doubled relative to his. But what
> if our entire planet, or entire local region of the universe, was copied?
> The relative measure of any two people would not be changed, it seems.

Copying this much would eliminate most of the practical problems, in
terms of how many wives or how much money you would end up with. But it
is important to understand that this is an extremely IMpractical thought
experiment. To the extent that we are focusing on practical issues
we ought to try to stick to at least somewhat plausible experiments.
Copying a person will perhaps be feasible someday, especially if he is an
AI or uploaded person who runs as a computer program. Such people will
have to deal with the practical as well as philosophical considerations
around potential duplication every day. Copying an entire galaxy seems
physically infeasible and is not something that our descendants are
likely to have to deal with.

> Perhaps you could say that the measure of observer-moments that take place
> after the the copying is higher than the measure of observer-moments that
> take place before it, but I'm not sure that'd be true either, it really
> depends on what your theory is about how measure should be assigned to
> different observer-moments.

Yes, I would say this. It is a standard prediction of the MWI (to the
extent that the MWI is standard!). Measure has a certain definition
in this flavor of QM, such that when a universe splits its measure is
reduced in each of the branches. Activities which take place within a
universe (neglecting irrelevant splits) do not get their measured reduced.

We use a similar concept of measure in the AUH (all universe hypothesis).
Schmidhuber defines the inverse exponential of the length of the
computer program to generate a universe as its measure. If we think
of a universe splitting a la the MWI within the framework of the AUH,
the new universe(s) require more information to specify them, namely
the outcome of the coin flip or whatever it was that caused the split.
This additional information reduces the measure of the universes by
making their information description longer. If there is no split and
simply a duplication of some subset, this could happen in a deterministic
universe and there would be no change in the measure. The result is that
universe splitting reduces measure while subset duplication does not.

> Part of the problem is you seem to be assuming
> measure can somehow be derived from the number of physical copies in a
> single universe, whereas I lean more towards the view that a TOE would
> ultimately be stated simply in terms of observer-moments and the measure on
> each, with the appearance of a "physical universe" just being a consequence
> of the particular types of observer-moments that have higher measure. So it
> seems that it partly depends whether one believes the third-person
> perspective or the first-person perspective is more fundamental. (Although
> even if you take the first-person perspective as more basic, you'd need more
> of a fleshed-out theory of how the appearance of an objective physical
> universe comes about to say for sure whether copying a mind in a single
> universe is the same or different from many-worlds splitting.)

I would say that the first person view is consistent with Schmidhuber's
approach as well as the MWI. Both of these models define a measure
over observer moments that can in principle be calculated precisely.
This then determines what we are likely to see and experience.

Hal Finney
Received on Mon Apr 18 2005 - 16:23:58 PDT