--- GSLevy.domain.name.hidden wrote:
> jackmallah.domain.name.hidden writes:
> > OK. The advantages of my approach are
> > 1) It does not require any special definition
> of identity, while your
> > approach seems to give one a fundamental role.
> >
> > 2) In principle, it allows all kinds of
> (statistical) retrodictions.
> > Your approach seems to only allow a few kinds.
> >
> > 3) It is unclear how you deal with time based
> on your previous and current statements. In my
> approach it is clear that each observer-moment has
> > some measure, and a set of related observer
> moments can be used for a 'person'
> > if desired, so that the person's total measure is
> the sum of that of his constituent observer-
> moments. This clearly rules out QTI.
> >
> > 4) In the MWI of QM the measure of an observer
> in a 'branch' is defined.
> > This is proportional to
> > dM(total)/ d tau [where dtau ~ dt * clock speed]
> > in my approach.
> > In your approach it must be divided by an
> identity-dependent factor
> > resulting in possible near-zombies.
> >
> 1) True. I do require definition of identity or more
> precisely of the self.
> However this requirement is absolutely trivial.
Then state the precise definition.
> a) First, no matter how the self is defined, (i.e.,
> what boundaries you draw
> around it) the normalized measure for the self is
> always unity.
> Normalized Measure = M(Self) / M(Self)
> Which is nice and very egalitarian. We all have the
> same measure.
Then why don't people similar to us, but who see
white rabbits, have the same measure?
> b) Second, the assumption of the self is absolutely
> essential no matter what
> method (yours or mine) is used. We MUST start with
> the assumption "I think."
> Otherwise it's not even worth thinking about nature
> or anything.
Observer-moments exist. At least one I can vouch
for; the rest I must extropolate.
> 2) I don't understand. Maybe you should come up with
> an example.
Ok. Suppose there are two possible kinds of
people: those that are (at least as far as they can
tell) physically the product of a Darwinian evolution
process, and those that just came into being due to
random fluctuations. In my approach, given the laws
of (AUH or whatever) I can in principle predict that
the former is more 'probable' than the latter. But
for you, everyone has the same measure, so you can't.
> 3)
I already explained this to you: cardinality isn't
the issue. Just take a limit in the right way.
> 4) I don't understand. Are you trying to calculate
> the output measure in a
> branch, given the total input measure? What are tau
> and t? I can't comment on zombies.
No, just the contribution of each branch to the
total measure of some defined set of observer moments.
t is time, and tau is brain-computer proper time.
If you simulate a brain on a Pentium, it gives rise to
more measure per unit time than the same algorithm on
a 8086.
> With my method, you must distinguish between first
> person and third person
> observations. Third person observations match
> classical physics. First person
> observations do not. White rabbits do appear
> especially if they are essential
> in maintaining the existence of the observer. In our
> case, for example, the
> Big Bang which is definitely a first person event.
Sounds crazy.
With my method, there is no such distinction.
There's a measure distribution. White rabbits appear
only with very small effective probability.
"Maintaining an observer" is undefined until
"observer" is defined.
=====
- - - - - - -
Jacques Mallah (jackmallah.domain.name.hidden)
Physicist / Many Worlder / Devil's Advocate
"I know what no one else knows" - 'Runaway Train', Soul Asylum
My URL:
http://hammer.prohosting.com/~mathmind/
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
Received on Fri Apr 07 2000 - 14:53:39 PDT