Le 27-nov.-05, à 02:25, Saibal Mitra a écrit :
The answer must be a) because (and here I disagree with Jesse), all
that
exists is an ensemble of isolated observer moments. The future, the
past,
alternative histories, etc. they all exist in a symmetrical way. It
don't
see how some states can be more ''real'' than other states.
But then how could we ever explain why some states seem to be more
*near*, or more probable than others from our point of view?
Is the choice between Papaioannou's "a", "b" reflecting(*) the ASSA
and RSSA difference?
Recall: ASSA = absolute self-sampling assumption. RSSA = relative
self-sampling assumption.
(*) Stathis Papaioannou writes:
If on the basis of a coin toss the world splits, and in one branch I am
instantaneously killed while in the other I continue living, there are
several possible ways this might be interpreted from the 1st person
viewpoint:
(a) Pr(I live) = Pr(I die) = 0.5
(b) Pr(I live) = 1, Pr(I die) = 0
(c) Pr(I live) = 0, Pr(I die) = 1
Even on this list, there are people who might say (a) above is the case
rather than (b) or (c).
Saibal:
So, you must think of yourself at any time as being randomly sampled
from
the set of all possible observer moments.
This could make sense in a pure third person perspective, but then it
is no more a perspective. And, indeed, to predict the result of
anything I decide to test, I need to take into account relations
between observer-moments. Let me throw a dice. Are you saying to us
that to predict the result I need to take into account all
observer-moments and sample on them in some "uniform" way. Why should
people buy lotto-tickets? They could make the big win by their OM being
sampled on all OMs.
I'm not saying you are false, but your absolute sample does not
correspond tour first person experience (including physics) which we
want to explain. It seems to me.
To get to answer b) you have to
redefine your identity so that experiencing having done the experiment
becomes a necessary part of your identity.
Not some absolute identity, but memories are part of our relative,
mundane, identity.
But this is cheating because you
wouldn't say that if ''death'' were replaced by a partial memory
erasure
such that the experience of having done the experiment were wiped out
form
your memory.
OK, but that is why the experiment is proposed with (absolute) death
(if that exists) and not with memory erasure. This could change the
probabilities a lot, and this can admit many different protocol for
verifying the probability distributions. It is another experiment.
Perhaps I miss your point.
Bruno
http://iridia.ulb.ac.be/~marchal/
