Le 22-juil.-12, à 21:10, Stephen P. King a écrit :
On 7/22/2012 7:24 AM, Bruno Marchal wrote:
Le 21-juil.-12, à 20:04, Stephen P. King a écrit :
On 7/21/2012 7:57 AM, Bruno Marchal wrote:
Hi Stephen,
I appreciate very much Louis Kauffman, including that paper. But I
don't see your point. Nothing there seems to cast any problem for
comp or its consequences.
Why not read the MGA threads directly, and address the points
specifically?
I already did. My contention is that computational universality
is NOT the separation of computations from physical systems, it is
the independence of a given computation from any one particular
physical systems.
Computational universality is an arithmetic notion. You don't need
UDA to separate it from physics, you need only a good intro to
computer science. This critics is wrong at the very start.
Dear Bruno,
Could you be more specific on what they are wrong about and how?
You might wait I come back on this on the FOAR list. I have already
xplain this here, but you can also read any textbook on theoretical
computer science, or any paper: the notion of universality can be
defined in arithmetic and has nothing to do with physics.
The independence of a given computation from any particular physical
system is obviously part of the comp assumption, and should not be
confused with the impossibility of any physical system to capture or
produce consciousness, which is related to the mathematical and the
theological by comp, and that is the consequence of UDA including
MGA.
By addressing MGA, I meant you to quote that text, and that text
only, and tell me were you disagree, and for what reason.
Let me quote some previous arguments by some others that are
making the same case:
http://old.nabble.com/Re%3A-Movie-Graph-Argument-p32993663.html
You should better quote the answer I gave to them. You should try to
make your own mind so as to be able to ask a specific point you feel
mistaken, without any interpretation and still less philosophical
beliefs, except those needed (like comp) for the sake of the argument.
Re: Movie Graph Argument <inconnu.png>
<inconnu.png><inconnu.png><inconnu.png>
by pedro7 Dec 17, 2011; 07:08am
On Dec 17, 4:39 pm, Russell Standish <li...@...> wrote:
> On Fri, Dec 16, 2011 at 08:26:21PM -0800, Pierz wrote:
>
> ...snip...
>
> > The problem is even deeper than this, however. How does the system
> > ‘know’ when two locations should be bilocated? This works OK for a
> > single copy of Klara, since she is a static system. But if she
must
> > physically interact with all the previous editions of herself
further
> > back in the calculation chain, then she will be forced to ‘build’
> > pipes on the go, a ridiculously contrived procedure that totally
> > vitiates the idea of a mindlessly proceeding, inert system. And
how
> > does Klara (or rather, Olympia) remember which path she has
followed
> > in order to know which trough to drain? New mechanisms must be
devised
> > which effectively mean retaining the activity of previous Klaras
in
> > the chain and are no different from a form of backtracking.
>
> My understanding is that to construct Olympia, we take n copies of
> Klara, and run each copy to step i of the program, where i=1..n-1.
The
> construct the sequence of water troughs such that they are equal to
> that of K_i at step i. We also connect K_i to Olympia at that point,
> ready to take over in the event of a counterfactual being true.
>
> I don't think the issue of pipes is a problem - we can assume each
> trough in state i is connected to the troughs of states i-1 such
> that when the armature moves through to state i, it closes a valve
> connecting the troughs to the previous state's troughs.
>
> It may seem complex, but it is mere complication, not complexity, if
> you understand the difference.
>
> > If Maudlin’s argument is a foundation of the UDA, then it seems to
me
> > the UDA is on shaky ground, though I have yet to investigate the
MGA
> > in depth. People talk about the Movie Graph Argument, but the
links
> > provided refer to Alice and a distant supernova with lucky rays
that
> > substitute for functional neurons. I don’t see a connection to the
> > idea of a recording or a filmed graph. Can someone enlighten me?
>
> Maudlin's argument has been compared with the MGA, which is step 8
of
> the UDA. The previous steps are independent of Maudlin.
>
> Olympia can be compared with a recording of the computation. That is
> the "filmed graph" (aka movie graph).
>
> --
>
>
-----------------------------------------------------------------------
---- -
> Prof Russell Standish Phone 0425 253119 (mobile)
> Principal, High Performance Coders
> Visiting Professor of Mathematics hpco...@...
> University of New South Wales http://www.hpcoders.com.au
>
-----------------------------------------------------------------------
---- -
... [show rest of quote]
> Maudlin's argument has been compared with the MGA, which is step 8
of
> the UDA. The previous steps are independent of Maudlin.
I understand that, but all the steps are necessary to support the
argument. If consciousness supervenes only on physical computation,
then one requires a physical instantiation of the UD, not a purely
arithmetical one.
But the step 8 point is precisely that consciousness cannot supervene
on the physical computation only.
> My understanding is that to construct Olympia, we take n copies of
> Klara, and run each copy to step i of the program, where i=1..n-1.
The
> construct the sequence of water troughs such that they are equal to
> that of K_i at step i. We also connect K_i to Olympia at that point,
> ready to take over in the event of a counterfactual being true.
Invalid because of the infinite regress problem. How can we run the
program on the individual Klaras without connecting them to the
Olympia in the first place?
?
That is supposed to have been done.
The Klaras cannot calculate anything
without the counterfactual mechanism of all the other Klaras ensuring
they don't go wrong.
The Klaras are build to keep the conunterfacualness correct, but
without any change in the physical activity of Olympia,
so if you accept the 323 principle, you can already abandon the
physical supervenience at this stage.
If all the Klaras have already been run somehow
so the troughs prior to the branch onto the active Klara contain the
calculated values then there is no need to run Oylmpia at all. The
state of the last Klara already contains the output of the calculation
and we can discard Olympia and just say that we already calculated the
value in the past. This makes a mockery of the entire elaborate
mechanism Maudlin postulates and the business about inert parts and so
on is irrelevant. I don't think that saying that a live calculation
can always be replaced by one that was completed in the past solves
anything. Certainly consciousness (or a computer) may draw on the
results of completed calculations in order to speed up its work (a
computer doesn't need to recalculate the value of pi every time it
needs that constant), but it cannot solve every problem that way,
obviously! A computer game may pre-render an explosion made by
computing hundreds of thousands of particles, as a shortcut, but it
cannot pre-render every possible game and just branch into the
relevant branch of that movie as required. Unless you grant it
infinite calculation resources in the past and none in the present, an
abject sophistry.
The result of the computations are not relevant, as the physical
supervenience has to associate
consciousness to a process, not to a result of a computation.
I can't find anything in Maudlin's paper that suggests the method you
propose - pre-running every copy of Klara as if it had dealt with all
prior counterfactuals.
Maudlin builds the Klara from a pre-running of Olympia.
Each copy is merely another dumb Klara ready to
wrong the next instant. That is both essential to the argument, and
its fatal flaw.
***
I don't see any flaw here. If someone else understand, please provide
some help to Stephen to convey it.
Here is another:
http://old.nabble.com/Re%3A-Movie-Graph-Argument-p32978306.html
Re: Movie Graph Argument
<inconnu.png> <inconnu.png><inconnu.png><inconnu.png>
by Joseph Knight Dec 14, 2011; 05:09pm ::
On Wed, Dec 14, 2011 at 1:51 PM, meekerdb <meekerdb@...> wrote:
On 12/14/2011 10:40 AM, Joseph Knight wrote:
On Tue, Dec 13, 2011 at 11:32 PM, Kim Jones <kimjones@...> wrote:
Any chance someone might précis for me/us dummies out here in maybe
3 sentences what Tim Maudlin's argument is? Nothing too heavy -
just a quick refresher.
I'll try, but with a few more than 3 sentences. Suppose the
consciousness of a machine can be said to supervene on the running
of some program X. We can have a machine run the program but only
running a constant program Y that gives the same output as X for one
given input. In other words, it cannot "handle" counterfactual
inputs because it is just a constant program that does the same
thing no matter what. Surely such a machine is not conscious. It
would be like, if I decided "I will answer A B D B D D C A C..." in
response to the Chemistry test I am about to run off and take, and
happened to get them all correct, I wouldn't really know Chemistry,
right?
But I think Russell has reasonably questioned this. You say X
wouldn't know chemistry. But that's a matter of intelligence, not
necessarily consciousness. We already know that computers can be
intelligent, and there's nothing mysterious about intelligence
"supervening" on machines. Intelligence includes returning
appropriate outputs for many different inputs. But does
consciousness?
I was really just using my Chemistry test as an imperfect analogy to
the machine running Y being conscious (or not), so it doesn't affect
the rest of the argument. But I see your point. Would you argue that
a constant program (giving the same output no matter the input) can be
conscious in principle? Maudlin assumes that such a program cannot be
conscious, in his words, "it would make a mockery of the computational
theory of mind." I am agnostic. In my opinion the
Filmed Graph argument is more convincing than Maudlin, because
with Maudlin one can still fall back to the position "consciousness
can in principle supervene on a constant program".
But this would contradict comp, and finish the reductio ad absurdum, as
a program is supposed to make a sophistiocate computation, and I would
say no to any doctor who would use the argument above to replace my
brain for a constant machine.
(For those interested, here is the article itself)
Brent
So consciousness doesn't supervene on Y. But Maudlin (basically)
shows that you can just add some additional parts to the machine
that handle the counterfactuals as needed. These extra parts don't
actually do anything, but their "presence" means the machine now
could exactly emulate program X, i.e., is conscious. So a
computationalist is forced to assert that the machine's
consciousness supervenes on the presence of these extra parts, which
in fact perform no computations at all.
I think what Russell said about this earlier, i.e., in a multiverse
the extra parts are doing things, so consciousness then appears at
the scale of the multiverse -- is fascinating. But I am out of time.
Hope this helped. I would recommend reading the original paper for
the details.
--
The idea is that we can obtain a fully explanatory model of
interaction between multiple minds that allows us to reconstruct
physics *and* that this explanation need not have anything to do with
physicality at all. It "all is just the dreams of numbers". It is this
thesis that I am claiming is wrong, for it ignores the necessity of
relatively persistent media on and in which representations of the
models can occur. If you cannot write a description of a theory or
model such that some other entity can read it, how do you assume that
it is real? Reality requires incontrovertibility, the lack of
contradiction for all involved.
Thus there has to be a means by which a plurality of observers,
each autonomous in thought and action (at some level), can met in a
common medium. That common medium is, I claim, our physical world and
such cannot be abstracted out of existence.
Good, but the last claim would be a problem with comp only if you mean
"primitive physical world". If not you are restating the conclusion of
UDA. We have to derive that physical world by the coherence conditions
on machine's dream in arithmetic. We do not abtyract it out of
existence: we justify its need from arithmetic (and comp).
Keep in mind that UDA reformulate the mind body problem, and show that
if comp is true, then the solution cannot be coherent with physicalism.
The former Seperation is categorical in that one has seperate
categories with no connection between them whatsoever. The latter is
a duality between a pair of categories in that for the class of
equivalent computations there is at least one physical system that
can implement it and for a class of equivalent physical systems
there is at least one computation that can simulate it. (Equivalent
between physical systems is defined mathematically in terms of
homologies such as diffeomorphisms) This idea was first pointed out
by Leibniz and known as Leibniz equivalence. See
http://plato.stanford.edu/entries/spacetime-holearg/
Leibniz_Equivalence.html
This makes sense in Aristotelian physics, which, that is the point,
does not make sense if comp is true. Unless you can find a flaw. May
be you have a problem with what a proof consists in. Proofs does not
depend on the interpretation of the terms and formula occurring in
it. A proof in math, and in applied math, is always complete in
itself. Keep in mind that comp does not presuppose any theory of
physics. It assumes only that the physical reality is Turing complete
at least. If not, asking for an artificial digital brain would not
make sense.
The flaw is the assumption of separation = independence.
Then you have to work harder on what is wrong with step 8.
You are assuming that Platonism is a immaterial monism theory, it is
actually a weak dualism theory as it requires a plurality of entities
"to whom" the meanings of the Forms obtains. Yes, there must be some
ontological level where all of the "seperate" entities are united and
singular. The materialist monist would claim that the singularity is
in the physical world and thus nothing else exists but the physical
world. The immaterialist and ideal monist claim that all is in the
mind of God. I claim that the Mind and the Body of God are degenerate
at the ultimate singuloarity level, they are indistinguishable, thus
the ground is neutral and singular from which all emerges.
I do the same, and it is already done (not a project), but the neutral
ontology (neutral in the philosophy of mind sense) is given by any
Turing complete theory, like arithmetic. This makes comp testable, etc.
You are to vague with your "neutral monism" to see if we are close on
this. You seems unable to be put it into a theory, which makes
impossible to progress, and make your "invalidity" argument not
communicable. Mere existence is not a theory.
Bruno
http://iridia.ulb.ac.be/~marchal/
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/everything-list?hl=en.
