From: Bruno Marchal <[EMAIL PROTECTED]>
Subject: Re: Fw: Quantum accident survivor
Date: Sat, 08 Nov 2003 15:56:31 +0100

At 14:36 07/11/03 -0800, Hal Finney wrote:


Well, I do believe in continuity of consciousness, modulo the issues
of measure.  That is, I think some continuations would be more likely to
be experienced than others.  For example, if you started up 9 computers
each running one copy of me (all running the same program so they stay
in sync), and one computer running a different copy of me, my current
theory is that I would expect to experience the first version with 90%

Almost OK, but perhaps false if you put *the measure* on the (infinite)
computations going through those states. I mean, if the 9 computers
running one copy of you just stop (in some absolute way I ask you to conceive for
the benefit of the argument), and if the one computer running the
different copy, instead of stopping, is multiplied eventually into many
self-distinguishable copies of you, then putting the measure on the histories should
make you expect to experience (and memorized) the second version more probably.

It is the idea I like to summarize in the following diagram:

\        /                     |      |
  \    /                       |      |
    \/               =        |      |
     |                         |      |
     |                         |      |

That is, it is like a "future" bifurcation enhances your present measure.
It is why I think comp confirms Deutsch idea that QM branching is really
QM differentiation. What do you think? I mean, do you conceive that the
measure could be put only on the "maximal" possible computations?


This is an important point which I think people often miss about the QTI. It is sometimes spoken of as if the QTI only goes into effect at the moment you are about to die (and thus have no successor observer-moment), which would often require some fantastically improbable escape, like quantum tunneling away from a nearby nuclear explosion. But if later bifurcations can effect the first-person probability of earlier ones, this need not be the case.

Consider this thought experiment. Two presidential candidates, let's say Wesley Clark and George W. Bush, are going to be running against each other in the presidential election. Two months before the election, I step into a machine that destructively scans me and recreates two copies in different locations--one copy will appear in a room with a portrait of George W. on the wall, the other copy will appear in a room with a portrait of Wesley Clark. The usual interpretation of first-person probabilities is that, all other things being equal, as the scanner begins to activate I should expect a 50% chance that the next thing I see will be the portrait of George W. appearing before me, and a 50% chance that it will be Wesley Clark.

But suppose all other things are *not* equal--an additional part of the plan, which I have agreed to, is that following the election, the copy who appeared in the room with the winning candidate will be duplicated 999 times, while the copy who appeared in the room with the losing candidate will not experience any further duplications. Thus, at any time after the election, 999 out of 1000 versions of me who are "descended" from the original who first stepped into the duplication machine two months before the election will remember appearing in the room with the candidate who ended up winning, while only 1 out of 1000 will remember appearing in the room with the losing candidate.

The "last minute" theory of quantum immortality is based on the idea that first-person probabilities are based solely on the observer-moments that qualify as immediate successors to my current observer-moment, and this idea suggests that as I step into the duplication machine two months before the election, I should expect a 50% chance of appearing in the room with the portrait of the candidate who goes on to win the election. But as Bruno suggests, an alternate theory is that later bifurcations should be taken to influence the first-person probabilities of earlier bifurcations--under this "anticipatory" theory, I should expect only a 1 out of 1000 chance that I will appear in the room with the portrait of the losing candidate. This would lead to a weird sort of "first-person precognition", where after the duplication but before the election, I'd have good reason to believe (from a first-person point of view) that I could predict the outcome with a high probability of being right. But this kind of prediction would be useless from a third-person point of view, since all outside observers would see two symmetrical copies who both seem equally certain that their candidate will be the winner. Of course this is not much stranger than the basic quantum immortality idea that if I am in some dangerous accident, most third-person observers will see me end up dead, while I have a close to 100% chance of surviving from a first-person POV.

Applied to quantum immortality, this "anticipatory" idea suggests it would not be as if the universe is allowing events to go any which way right up until something is about to kill me, and then it steps in with some miraculous coincidence which saves me; instead, it would be more like the universe would constantly be nudging the my first-person probabilities in favor of branches where I don't face any dangerous accidents which require "miracles" in the first place. Of course since this would just be a probabilistic effect, I might still occasionally face accidents where I had to be very lucky to survive, but the lower the probability there is of surviving a particular type of accident, the less likely I am to experience events leading up to such an accident.

Jesse Mazer

Great deals on high-speed Internet access as low as $26.95. (Prices may vary by service area.)

Reply via email to