[extropy-chat] Probability of identity

Russell Wallace russell.wallace at gmail.com
Wed Oct 11 00:28:31 UTC 2006

On 10/10/06, Jef Allbright <jef at jefallbright.net> wrote:

> Here's where the confusion becomes apparent.  It's the usual confusion
> between objective and subjective points of view, reinforced by culture,
> language, and personal experience.  It's Descartes'' error in not so subtle
> disguise, the assumption within "cogito, ergo sum" that *I* am experiencing,
> rather than the more defensible assertion that *some process* seems to be
> reporting on an experience.

That's how I would be inclined to come at it too, but the problem is that we
implicitly use this stuff to make predictions.

In an infinite universe (at any of the Tegmark levels), there are infinitely
many identical instances of you, and there is no fact of the matter as to
which one of those instances you are; you are all of them.

Because there is a finite (albeit very small) probability that a flock of
giant albino penguins will materialize in your living room in five minutes,
the number of instances that experience such an event is the same as the
number that does not: infinity (of the same cardinality) in each case.

In practice, we predict a giant albino penguinless future - and this
prediction always comes true. Why? The best explanation is that in some
obvious sense (cardinality notwithstanding) there are far _fewer_
penguin-experiencing instances than penguinless ones, so one's subjective
probability of the latter experience is almost 1. Empirically this works:
our subjective experiences are in fact those we would expect from the
probability argument.

But this explanation appears to break down here, because a count of
instances gives 999:1, yet it seems that the actual subjective experience
should be 50:50. In other words, this paradox appears to break the one
explanation we have as to how the world manages to be a semi-predictable
place. That's why I'm still chewing on it.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20061011/65a8e0c5/attachment.html>

More information about the extropy-chat mailing list