At 05:27 21/04/04 -0400, Kory Heath wrote:
>Bruno Marchal wrote a 10-point argument about determining whether or not
>we are simulated by some massive computer. Here is point 9 from that post:
>
>>9) Now, from computer science and logic, startlingly enough perhaps,
>>we can isolate a measure on the 1-person comp histories, and this
>>give us the laws of physics (this is too technical to be put here, and
>>actually I have derived only the logical structure of the probability
>>one, but I got something non trivial and very close to a quantum logic
>>(which can be seen as the logic of the quantum probability one, as
>>von Neumann has shown).
>
>Bruno, I agree with everything in points 1-7 of the post I'm referring to.
Cool.
(Kory is referring to my post
http://www.escribe.com/science/theory/m5384.html )
>I think I understand the main thrust of points 8 and 10, but they both
>depend on the truth of the quoted point 9.
Not at all. The point 8 follows only from the points 1-7. Actually the points
1 to 8 sum up the main result of my thesis, that is the reversal
physics/machine's psychology (once we assume comp).
This made the point 9 just a mathematical confirmation.
I will try to explain, but actually it is me who does not understand why
you don't see that the point 8 follows from the point 1-7.
The proof really stop at 8.
9 is a mathematical confirmation with
observer modelized (or even identify) with "equivalence class" of
states of sound universal Turing machines.
>I find this point [9] highly dubious, and I'm interested to hear more
>about why you believe it (or whether I'm actually interpreting it correctly).
OK. I promise I will try to make it clear, but that will be possible
only if you realize that 8 follows from 1-7. Keeping we assume comp all
through.
I will first continue to comment your post, as a sort of preparation.
>Is it your contention that if we start only with the idea of Platonia
>(which contains all possible computational structures and "events"), and
>we analyze the structure of all possible 1st-person histories weaving
>their way through this Platonia, we would see that a majority of them find
>themselves in universes like the one we currently inhabit, with stars,
>galaxies, relativity, quantum physics, etc.?
No. Such a contention is not mine. What you say here
does not follow from neither 1-8, nor 9.
Only the laws of physics follows, but the existence of the moon
or of galaxies are not necessarily "laws of physics", but are
quite possibly "geographical laws" (like being in Moscow
instead of Washington). Now the stability of
the appearance of matter is related to the laws of physics,
and so, if (to talk like you) we analyze the structure of all
possible 1st-person histories going trough your actual
Kory-Heath state (including your memory of stars and
galaxies) then the measure of your computational
continuations in Platonia without stars and galaxies
will (or should) be negligible. The disappearance of galaxies
could be seen as a "talking white rabbit" event.
So, only worlds with entities very similar to Kory Heath
should contains galaxies. If this was false (in Platonia)
then comp would be false (and refutable).
>I find that extremely hard to believe; there are an infinite number of
>multi-dimensional cellular-automata worlds that contain nothing like
>stars, galaxies, or the familiar physics we know,
But then they don't contain entities sufficiently
close to Kory Heath, (they don't go through Kory-Heath comp
states) and those world/stories don't belong to your
actual first person indeterminacy domain.
>and I don't see any reason to believe that all the 1st-person histories
>weaving their way through these computational structures (which surely
>exist in Platonia) have less measure than those that weave through the
>structures more resembling our own universe.
You are right. No doubt there are worlds without stars
and galaxies, but having srats and gilixas instead, and
some Burno arguing to some Kyro about the geographical
nature of those objects. Nothing in 1-8 prevents this.
>I find it much easier to believe - in fact, it seems almost necessarily
>true - that, when we analyze the structure of all the 1st-person histories
>weaving their way through Platonia, we see that virtually all observers
>observe indeterminacy. This in itself seems like a non-trivial finding. It
>would be even more non-trivial if one could analyze the structure of the
>indeterminacy that virtually all 1st-person observers must observe, and
>find that it is very much like "quantum logic". This seems to be one of
>your claims, and it certainly seems possible (though I don't know enough
>yet to judge the issue for myself). However, it's a much further leap to
>conclude that we can go on to derive "the laws of physics" from the
>analysis of 1st-person histories weaving their way through Platonia. In
>the quote above, you acknowledge that you haven't actually done this -
>"actually I have derived only the logical structure of the probability
>one, but I got something non trivial and very close to a quantum logic" -
>but you seem to believe that it can be done. Why?
All right, you miss something. As I said, I feel *I* miss
something because it seems to me that the point 8 follows
directly from 1-7. (I suggest people print the post
http://www.escribe.com/science/theory/m5384.html
for having the 1-9 points in front of the eyes). Point 10
is not necessary because it concerns the particular
question by Brett Hall. First let me copy the point 8,
and let me rewrite it more carefully (if not less
diplomatically!).
Point 8: (original version from post 5384)
8) So, if we are being simulated by a massive computer somewhere
in a real universe, then to make any first person verifiable prediction
we must take into account all the possible computations going through
our actual state. So "the laws of physics", which we can temporarily
define as the laws of whatever we can predict consistently, should
be given by a measure on those 1-person computational histories.
Point 8bis: (Attempt toward a clearer ameliorated version. I have
also make some emphases):
8bis) So, if we are being simulated by a massive computer
somewhere in a real universe, then to make ANY first person
verifiable prediction, we must take as domain of 1-person
indeterminacy all the possible maximal consistent computational
computations GOING THROUGH, AND FROM, OUR ACTUAL STATE,
and existing in Platonia (more precisely in UD*, which is the complete
infinite execution of the Universal Dovetailer).
Let us (re)defined the laws of physics as the laws we can
always predict and verify consistently (if any!).
Now, having accepted the 1-7 points, the occurrence of such
laws must have a measure 1, so the laws of physics must be
derivable from what has measure 1 relatively to the measure
on the computational histories. OK?
Let me illustrate a simple experiment in physics. Let us
decide to let a chalk fallen. And let us compare some theories.
A) The naive common sense theory: As far as I remember
chalk always falls on the ground, so I infer that this will
happen again.
B) The first physicist: giving the laws F=ma, and Newton's
gravitation law, I can predict with reasonable accuracy
that the chalk will fall on the ground about here and about
that time.
C) The second physicist: the chalk is a many body
quantum object in interaction with many bodies, so let us
apply some gravitationnal form of Feynman summation on
all quantum computational histories. After some phase
randomization I can predict the chalk will fall here and at
some time with more than 60 correct decimals.
(He has been quick! I guess he used a quantum pocket
calculator!)
D) The computationalist: first let me bet on a comp level
of substitution (which exists *hopefully*), and let me bet
on my actual comp states, that is the one which describes
my brain/body state in front of the chalk which I will let
fallen. Now giving 1-7 (through comp) I "know" that
my future first person experience is giving by a sum (average)
on all computations (existing in Platonia) going from and
through that state, and RELATIVE to that state.
Obviously to compute that you need to run the complete
UD in finite time and that's impossible, also you would be
unable to recognize your own comp histories when described
at other levels, so the computationalist will probably just
answer "uncomputable".
But "platonistically" it remains that if comp is true
the actual physical invariant must emerge as an average
on ALL the maximal consistent extensions relative to our
actual states (worlds, observer-moments, whatever ...).
Although that can be proved useless for actually predicting the
behavior of the chalk, it is enough for deriving physics.
And that derivation goes in the reverse sense, that is, the
more easy part will be the abstract logical structures, but
that is enough for giving the geometry of the multiple histories.
Well I must go now, and if I hurry myself to much I will
begin to be unclear ;-) Tell me if you have understand that
keeping faith in comp through 1-7 leads necessarily to 8.
Don't hesitate to give me an objection.
The only difficulty in "8", I think, is to remember that
we *assume* comp.
Then all what is needed for 9 (the mathematical confirmation)
is to accept to "modelize" or
better "identify" a platonistic observer by a sound modest
(lobian) universal church-turing-post-markov-fortran-lisp-
java-whatever machine (including quantum one).
And to interview it about those relative consistent
extensions and its inferable platonistic geometries
and what is stable in their discourses. That has been
made possible by the discovery of incompleteness
(Godel 31) and its "axiomatizability" (Solovay 76).
Bruno
http://iridia.ulb.ac.be/~marchal/
Received on Thu Apr 22 2004 - 10:24:44 PDT