On 04 Jul 2015, at 02:27, Bruce Kellett wrote:
meekerdb wrote:
On 7/3/2015 5:57 AM, Bruce Kellett wrote:
Bruno Marchal wrote:
No. This does not work. Everything cannot be computable, once we
are turing emulable.
I have never understood why you say this. Given that the world is
explicable in terms of regular physical laws, then it is
computable. Unitary evolution of the wave function is a prime
example of this. The only problem might be that that laws of
physics do not necessarily give the boundary values. But
multiverse models eliminate even this difficulty.
I think he assumes (as do many proponents of MWI) that "everything
happens" in some sense, which is not computable in the limit even
though approximations to it are computable.
Yes. I was thinking in the terms in which the number of possible
histories in any O-region is finite. I.e., given Uncertainty
Principle limits on measurement, only a finite number of histories
are physically distinguishable.
Since he wants to explain all experience, including physics, from
computation, he doesn't want to assume unitary evolution or any
other physics. He just assumes all possible computation (the UD).
But this leaves the "white rabbit"
problem.
The computations of the UD include the computation which gives the
physical universe, complete in every detail (at least down the the
UP limits),
If that exists.
All we can "see/feel" is our own consciousness, number relation, some
strongly felt as belief in other people and many other universal being.
Somehow, from a computer science perspective, a digital physicalist is
someone who believe that one special universal numbers makes it all.
That is not yet completely logically excluded from (classical)
computationalism, but the chance are that we might need more than one
fundamental universal numbers, to tame the infinities of computation
below the substitution level.
Anyway, if that unique "physical" universal number exists, to soleve
the mind-body problem, we have to derive it from number's theories, if
only to distinguish what is true on the machine, from what the machine
can prove, constructively, or not.
so every physically distinguishable history is automatically
included. We don't need to worry about summing over computations
that go through our particular conscious state, or which have
inconsistent extensions (white rabbits). We have only to see
ourselves as instantiated in the calculation that instantiates our
universe (and us in it).
Well UDA is an explanation why this cannot work, or if you prefer, can
only work by a use of a god-of-the-gap argument.
You replace the wrong explanation "God made the universe" by the non-
explanation "There is a universe/or special universal number".
Unless you equate your brain with the entire physical universe/number,
you need to soleve the FPI problem. In fact you must still show that
this solve the measure problem.
No need to sum over anything. As I have said earlier, this
instantiates all of Tegmark's level I and II multiverse, and
possible levels III and IV as well.
We must come back to step 8. But I am not yet sure you get entirely
step seven.
Reminding us that a theory that explains everything fails to
explain at all.
Exactly. We see the universe around us as it is because that happens
to be the universe we are in. Nothing is actually explained, and
that is the problem with any TOE.
I disagree. Explaining is looking for a morphism or a representation
between the thing we try to understand into something on which we
already agree.
Even if rough and false, comp gives a nice example of (possible)
explanation, as it derives a first person consciousness flux and
stabilizing (hopefully, if S4Grz1, Z1*, X1* continue to work well!)
predictable realities, and semantical fixed points. It explains why
god and consciousness, and truth, are not definable, yet that we can
meta-reason about through hypothesis and axiomatic definitions.
Consciousness supervenes on the physical brain, or physical
computer if required. In either case, it obeys regular laws, so is
computable.
I don't think that follows. First, if you allow randomness then
that's not computable. Second, if you avoid randomness by assuming
everything possible gets computed then that's subject to Cantor
diagonalization so Turing computation can't compute everything; you
need a hypercomputer. And third, our best theories, QM and GR,
both assume real numbers which are not necessarily computable.
Some real numbers are necessarily not computable -- possible
computer programs are countable but the reals are not, so there are
some reals for which no program will reproduce the infinite decimal
expansion. But given UP limits in the digital universe, this problem
goes away for distinguishable universes.
See above. You still need to justify that u, or it looks like a pseudo
religion. You forget that we want to solve the mind-body problem
(called nowadays "the hard problem of consciousness", but with comp
the cpnsciousness part is easy, it is the matter part which becomes
complex and interesting, if only by suggesting where the physical laws
come from, and why it hurts.
Randomness is not really a problem. You might not be able to predict
a result, but you can calculate probabilities, and that is
sufficiently law like for our purposes. Of course, MWI simply
computes all outcomes, so there is no randomness.
OK. Just 1p-randomness. I agree. Same with computationalism. There is
only numbers + the additive-multiplicative structure, the rest are
indexical modal internal views of this.
I use the standard self-reference theory, valid in the ideal case of
self-referentially correct machine. There, incompleteness forces the
machine to take into account the "philosophers" nuance between
justifying, knowing, observing, feeling. + the cross nuance of true
and justifiable, which interferes will all other nuances.
Then I model comp itself, in arithmetic, by the restriction to the
sigma_1 arithmetical sentences. They have the property that when they
are true they are provable: p -> []p.
So the measure one will be defined by []p & <>t, with p sigma_1 (and
both []p & p and []p & Dt & p are variants, in fact).
Then indeed this leads to quantizations and quantum logics, and ...
open problems. One has been solved by Vandenbussche years ago.
Bruno
If you interpretation of the UDA leads to non-computability, then
that itself is a strong argument against comp because it would
imply that some behaviour in the universe is not law like.
The universe is then understood in terms of computations, but
these are a consequence, secondary and not primary.
That seems self-contradictory to me. If computations are
secondary, why explain the universe in term of them?
Any physical model is secondary, yet we routinely explain things
in terms of such models.
This is what I see as a defeater of the MGA. There's another
possible reason that the movie graph is not conscious; namely that
it does not function in a context. Bruno dismisses this by saying
that the context can be included and then the argument still goes
through. But it doesn't because if the context has to include a
whole world it is no longer 'absurd' that the movie graph is
conscious in that context. It may still be unintuitive, but not
absurd.
Brent
Computation is a purely mathematical, even arithmetical notion.
Without giving a theory which would be able to just give a
physical definition of computation (not using the arithmetical
one) I can not make sense of your proposition.
I don't know why you think that a separate physical definition
would be necessary. Mathematics is derived from physics, and so is
computation.
Bruce
--
You received this message because you are subscribed to the Google
Groups "Everything List" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
http://iridia.ulb.ac.be/~marchal/
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
