On 7/12/2014 2:16 AM, Bruno Marchal wrote:
On 11 Jul 2014, at 22:26, meekerdb wrote:
On 7/11/2014 11:21 AM, Bruno Marchal wrote:
On 11 Jul 2014, at 09:41, David Nyman wrote:
On 11 July 2014 00:54, meekerdb <meeke...@verizon.net> wrote:
As I understand the MGA it assumes physicalism and then purports to show
that computation still exists with minimal or zero physical activity - it
evaucates the physics and keeps the computation.
For heaven's sake, Brent! This is what you originally said to Liz.
What you're referring to is Maudlin's argument. It's the *opposite* of
my understanding of the MGA, which seeks to show how physical action
can be preserved unchanged even in cases where the original
"computational relations" have been completely disrupted. I spent
several paragraphs describing this with additional examples. You then
commented this with "I agree with all you wrote", which led to some
further discussion based (as I thought) on this understanding.
Your comment above now leaves me hopelessly confused. I would be
grateful if you would review our recent discussion and clarify what
you do or do not agree with in my analysis.
I think that it will help to define perhaps more precisely what is a
computation.
I will reread the thread (many posts) when I have more time, and make only one
comment.
We have a computation when a universal machine compute something. We have an
intensional Post-Church thesis 'which follows from the usual Post-Church-Turing
thesis), which makes possible to translate "universal machine compute something" in
term of numbers addition and multiplication + one existential quantifier.
Now, when are two computations the same? If we fix a base phi_i, we might define a
computations by sequences of step of the universal base computing some phi_k, that is
the nth steps phi_k(j)^n of the computation by the base of the program k on the input
j, with n = 0, 1, 2, 3, etc.
But now that very computation will recure infinitely often, and not always in
(algorithmically) recognizable way.
You can conceive it might not be obvious that the evolution of a game of life pattern
(GOL is Turing-universal) is simulating a Fortran interpreter simulating a Lisp
program computing the ph(j)^n above.
And is it not the case that there will exist a mapping to a different base such that
this same evolution of the GOL is simulating a Python interpreter computing some
different phi. This why I have trouble with the concept to two computations being "in
the same state". ISTM that "same state" is relative to the enumerated basis functions
and the functions cannot be recognized from any finite sequence of states.
That is exactly why our computations, with and without (and in between) their
environment (with and without oracles) recurre infinitely often in the sigma_1 truth
(UD*).
So two computations can be the same at some level of description, and yet occurs in
quite different "places" in the UD*.
Is there a canonical level of description at which they are the same,
Yes. When the level of description is chosen correct, it can be the same from your or
God (arithmetic truth) view. But it is not a constructive or intuitionist notion (like
"not-halting", computer science is full of such type truth).
or are you just saying there exists some mapping which makes them the same over a
finite number of steps?
No. It is for a possibly infinite number of steps. the number of steps is not relevant.
It is misleading to define a computation only by a sequence of steps. It is a sequence
of step + a universal machine or number bringing those computational steps. That makes a
computation well defined.
But this depends on knowing the universal number; which cannot be inferred from a finite
piece of the computation. Yet that seems to be what Quentin requires in order to say to
instances of the MG compute the same function. Knowing the universal number or knowing
the function is like the problem of knowing all the correct counterfactuals.
Comp says that there is a level of description of myself such that those computation
*at the correct level" "carries my consciousness".
There's where I agree with JKC. You keep fudging what "comp" means. The above is
*not* the same as betting that the doctor can give you a physical brain prosthesis that
maintains your consciousness.
I don't see this. Please explain.
I think the level description would have to include not only you but your "world". So I
could say "yes" to the doctor even though I don't think the computational brain he
installs in me is sufficient, by itself, to instantiate my consciousness.
But Brent, and Peter Jones, adds that the computation have to be done by a "real
thing".
This is a bit like either choosing some particular universal number pr, and called it
"physical reality", and add the axioms that only the phi_pr computations counts: the
phi_pr (j)^n.
I think Peter, like me, questions the existence of numbers as any more than elements fo
language.
This is conventionalism. I consider that this view is refuted by number theory
implicitly, and by mathematical logic explicitly. The existence of not of infinitely
many prime number twins is everythi,g but conventional. With comp, the existence of your
dreams in arithmetic, and their relative proportions, are not conventional.
So it is not like choosing a universal number, it's saying that some things exist and
some don't.
Define "exist". If you say "exists physically" then you beg the question, and I will ask
you to define "physics".
Define "exists".
Well, this would just select (without argument)
It's based on observation not axiomatic inference.
That is explain in the comp theory. Observation is an internal modality of the
arithmetical truth.
You are using some "real existence" fuzzy notion to make a reasoning invalid, in the
same way that a creationist can say that "Evolution Theory" needs a God-of-the-gap.
a special sub-universal dovetailing among (any) universal dovetailing. The only
"force" here is that somehow the quantum Everet wave, seen as such a phi_pr do solve
the measure problem (accepting Gleason theorem does its job).
But just choosing that phi_pr does not solve the mind-body problem, only the body
problem in a superficial way (losing the non justifiable parts notably).
Or they make that physical reality non computable (as comp needs, but they conjecture
that it differs from the non (entirely) computable physics that we can extract from
arithmetic (with comp). But then it is just a statement like "your plane will not
fly". Let us make the test, and up to now it works.
Yes, I'm willing to accept your argument as an hypothesis.
Comp is the hypothesis. the argument is not.
But it seems to me that it proves that consciousness and physics necessarily complement
one another.
It is more than that. It makes physics the analog of a surface of what is real
independent of me (the mind of the universal machine) which is more like a volume having
that physical surface as a border.
Starting from arithmetic you must solve both the mind problem and the body problem at
the same time. I don't see that you've made psychology more fundamental than physics.
You've made arithmetic more fundamental.
ARITHMETIC ==> NUMBER's PSYCHOLOGY ==> CONSCIOUSNESS ==> MATTER APPEARANCE ===>
PHYSICS.
I agree with Brent, and I think everybody agree, when he says that reducing does not
eliminate. But we can't use that to compare consciousness/neurons to
temperature/molecules-kinetic. In that later case we reduce a 3p high level to a 3p
lower level. And indeed, this does not eliminate temperature. But in the case of
consciousness, we have consciousness which is 1p, and neurons which are 3p. Here, the
whole 3p, be it the arithmetical or physical reality fails (when taken as a complete
explanation). The higher level 1p notions are not just higher 3p description, it is
the intimate non justifiable (and infinite) part of a person, which wonderfully enough
provably becomes a non-machine, and a non nameable entity, when we apply the
definition of Theaetetus definition to the machine.
But what does it mean to "apply a definition to a machine". And why should we apply
*that* definition, which is far from axiomatic.
The accepted axiomatic is T or S4, that is []p -> p (with or without []p ->
[][]p).
machine's have their "[]" well defined, and to apply the Theatetus' definition consist
in define knowledge of p by []p & p.
But you equivocate on []. Sometimes is means "provable (from some axioms...Peano?)" and
sometimes it means "believes".
This is then a definition of knowledge which is applicable to (correct) machine, which
can then be defined in some axiomatic. But you are right, although we get an axiomatic
at the meta-level (S4Grz), it describes an entity which is not capturable by an
axiomatic (that's the vaccine against reductionism).
Interesting! We are at the crux of the crux! I see that Gerson(*) follows Socrates,
and take the Theaetetus definition ([]p & p) as a "description" of knowledge, but the
universal machine can understand that this is not true when applied on machine
(ironically enough). The modal "[]p & p" can define knowledge without providing any
description or code. "Worst" (but this is why this strategy works!), not only "[]p &
p" definition does not provide a description of the knower, but it is constructively
immune against all descriptions. The apparently little inner god, is a god from his
own first person view, that here, he share with the outer god. (if machine and
self-referentially correct).
This seems to be your love of mysticism turning wishes into proofs. You think it is
right *because* it fails description.
It is enough to make Gerson invalid. (Despite it get close to the point and he concludes
like me that the modern should read and borrow some part of Plotinus). Gerson is just
not aware how far the machines makes it correct, except for his somehow naive view of
Theaetetus' definition. he becomes correct on Plotinus. He is just not correct on the
machine's Bp & p notion.
You seem to be the one invoking wishes of primary matter *and* comp, but the argument
shows that it simply not work.
I don't invoke "primary" matter, just necessary matter.
Brent
--
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 everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.