On 08 Mar 2010, at 06:46, Jack Mallah wrote:
--- On Tue, 3/2/10, David Nyman <[email protected]> wrote:
computationalist theory of mind would amount to the claim that
consciousness supervenes only on realisations capable of
instantiating this complete range of underlying physical activity
(i.e. factual + counterfactual) in virtue of relevant physical laws.
Right (assuming physicalism). Of course, implementing only part of
the range of a computation that leads to consciousness might lead to
the same consciousness, if it is the right part.
What do you mean by right part?
In the case of a mechanism with the appropriate arrangements for
counterfactuals - i.e. one that in principle at least could be "re-
run" in such a way as to elicit the counterfactual activity - the
question of whether the relevant "physical law" is causal, or
merely inferred, would appear to be incidental.
Causality is needed to define implementation of a computation
because otherwise we only have correlations. Correlations could be
coincidental or due to a common cause (such as the running of a
movie).
Is is physical causality? Or computational causality. Which needs only
a universal mathematical base (like arithmetic, combinators, ...). I
try to understand your "physical or platonic".
--- On Fri, 3/5/10, Stathis Papaioannou <[email protected]> wrote:
If the inputs to the remaining brain tissue are the same as they
would have been normally then effectively you have replaced the
missing parts with a magical processor, and I would say that the
thought experiment shows that the consciousness must be replicated
in this magical processor.
No, that's wrong. Having the right inputs could be due to luck
(which is conceptually the cleanest way), or it could be due to pre-
recording data from a previous simulation. The only consciousness
present is the partial one in the remaining brain.
I have no clue by what you mean by "partial consciousness". Suppose a
brain is implementing a computation to which some pain can be
associated to the person owning that brain. Suppose that neuron B is
never used, and remain inactive, during that computation. Eliminating
the neuron B does not change the physical activity, but would change
the counterfactuals. Would such an elimination of inactive neuron
alleviate the pain? But then the person will change its behavior
(taking less powerful pain killer for example), and this despite the
brain implements the same computation? Continuing in that direction,
we could build a partial zombie. Partial consciousness does not make
sense for me.
"Computationalism" doesn't necessarily mean only digital
computations, and it can include super-Turing machines that perform
infinite steps in finite time. The main characteristic of
computationalism is its identification of consciousness with systems
that causally solve initial-value math problems given the right
mapping from system to formal states.
That is weird. Can you give a reference?
I should also note that if you _can't_ make a partial quantum brain,
you probably don't have to worry about the things my argument is
designed to attack, either, such as substituting _part_ of the brain
with a movie (with no change in the rest) and invoking the 'fading
qualia' argument.
Like in the movie graph? Look at MGA3 thread of last year.
Bruno, do you have the link? I searched the list archive but the
only references to fading qualia I could find are to the argument I
mentioned, in which a brain is progressively substituted for by a
movie, as Bishop does to attack computationalism. It _is_ different
than Chalmers, who substitutes components that _do_ have the right
counterfactuals - Chalmers' argument is a defense of
computationalism (albeit from a dualist point of view), not an
attack on it.
Search the thread MG11, MGA2 and MGA3. I don't use the expression
"fading qualia". I ask if consciousness disappear or not. It is is an
old argument already in my 1988 papers and earlier talks.
All of the 'fading qualia' arguments fail, for the reason I
discussed in my PB paper: consciousness could be partial, not
faded. I am sure that yours is no different in that regard.
You are the one saying there is something wrong, you are the one who
should be sure about this, and cite the passage you have refuted. What
do you mean by partial consciousness? In what sense this would deter
the movie graph. You don't give any indices.
If consciousness supervenes on the physical realization of a
computation, including the inactive part, it means you attach
consciousness on an unknown physical phenomenon. It is a magical
move which blurs the difficulty.
There is no new physics or magic involved in taking laws and
counterfactuals into account, obviously. So you seem to be just
talking nonsense.
If consciousness supervene of laws, which laws? The (physical)
supervenience thesis is that consciousness supervene on particular
instantiation of laws, and if physically inactive parts part play a
role (to get the counterfactual) then the physical supervenience is
already false. Then by saying "physical or platonic", you are just
agreeing with I try to explain, except that the label "physical" does
loose its explanative power both for consciousness, and physical
appearance (by uda-step seven). If you don't agree with this, it means
you do give a computational power to non active part for a particular
instantiation of consciousness, but then I am not allowed to say "yes"
to a doctor which propose to me an emulator of the computation of my
brain which may work differently from my current brain.
The only charitable interpretation of what you are saying that I can
think of is that, like Jesse Mazer, you don't think that details of
situations that don't occur could have any effect on consciousness.
Did you follow the 'Factual Implications Conjecture' (FIC)? I do
find it basically plausible, and it's no problem for physicalism.
For example, suppose we have a pair of black boxes, A and B. The
external functioning of each box is simple: it takes a single bit as
input, and as output it gives a single bit which has the same value
as the input bit. So they are trivial gates. We can insert them
into our computer with no problem. Suppose that in the actual run,
A comes into play, while B does not.
The thing about these boxes is, while their input-output relations
are simple, inside are very complex Rube Goldberg devices. If you
study schematics of these devices, it would be very hard to predict
their functioning without actually doing the experiments.
Now, if box A were to function differently, the physical activity in
our computer would have been different. But there is a chain of
causality that makes it work. If you reject the idea that such a
system could play a role in consciousness, I would characterize that
as a variant of the well-known Chinese Room argument. I don't agree
that it's a problem.
It's harder to believe that the way in which box B functions could
matter. Since it didn't come into play, perhaps no one knows what
it would have done. That's why I agree that the FIC is plausible.
However, in principle, there would be no 'magic' involved even if
the functioning of B did matter. It's a part of the overall system,
and the overall system implements the computation.
I don't think so. The overall systems implements all possible
computation making the subject reacting in possibly different
situation. For any particular computation, a part of the system may
implement the relevant computation needed for consciousness. Or,
again, you attach consciousness to the laws in general, and digitality
entails that consciousness will supervene on the platonic computation,
not on any particular universal machine (be it physical) running it.
You say you find FIC plausible, but you argue like it was not. What
happens in A depend on the level of substitution.
If consciousness supervenes, in "real time and place" to a physical
activity realizing a computation, and this "qua computatio" then
consciousness supervenes on the movie (MGA2).
We already agreed that it _doesn't_ supervene on the activity
alone. It requires the counterfactuals too.
But how could the neurons or the person knows if those counterfactual
neuron are present or not given that they have no activity at all.
Neither physical, nor computational. It looks like pure magic to me.
But if those definitions leads to a Turing emulable process, you
just lift the difficulty on another level.
As far as we know, physics is computable (although analog). So
what? That brings in no difficulty.
It brings the fact that if it is turing emulable, then it occurs an
infinity of time in the universal dovetailing, and that whatever I can
observe in my current state is emerge from a statistic on the
computation going through that state.
Do you agree with the notion of first person indeterminacy?
Eventually you talk like if we knew which universal system supports
us.
No, I don't and don't need to.
Why do you refer all the time to physical system, then, and causality.
Causality is not a notion of computer science, or it is defined in
term of higher level logical construction, with some modal logic, or
by referring to a fixed universal system.
Chalmers' dualist beliefs are not relevant to what we have been
discussing - I only mentioned him because he wrote the most widely
cited paper on implementation of computations - but I don't think he
would agree that the person would have any telepathic connection
between the places even epihenomenally. It seems more likely he
meant something else, like that he would define them both to be the
same person even though they each feel like they're the full original.
Well, he explicitly refuted this. He would have accepted the local
first person indeterminacy. But his later paper shows that he may have
see the point since.
That is especially likely given that he does favor the MWI, which
already involves similar splits and obviously no telepathy between
branches.
This is not clear at all. Given that the MWI has been invented
(discovered) to restore monism.
Didn't you see the quotes around 'physical'? As should have been
obvious, I use it as short for "physical or Platonic". Really
that's not important here.
As far as you pretend to have found something wrong in the derivation,
I would say it is the key point here.
If it is platonic, and thus arithmetic, then the mind body problem is
reduced to the problem of justifying the laws of physics from a
statistics on computations.
The point is that MGA fails because it doesn't take the needs for
the laws into account.
Which laws, the one for the arithmetical platonia (elementary
arithmetic) or the physical laws?
OK. I see you have not get the point. MGA does not need the notion
of counterfactuals. It just show that IF neurons, or basic entity
have no prescience, then the movie has the same physical activity,
as far as realizing a computation in real time, than the boolean
graph. So, to *avoid* prescience, we have to make consciousness
supervening on the counterfactuals. But those related to the
computation are mathematical, immaterial, defined only in computer
science. The physical counterfactuals have to be non relevant, or
Turing emulable.
The above paragraph doesn't make any sense. Why do the physical
counterfactuals have to be non-relevant?
Because MGA shows that a universal computing machine cannot
distinguish a "real" computation from a platonic computation. The
notion of computation is a purely mathematical one. Not a physical
one. It happens that it looks like we are embedded in a (quantum)
computational system, but below our level of substitution all
computations statistically interfere.
That would be the case if the work did not pass the academical test.
Sadly, some crackpots do get their work published and get people who
should know better to agree with them. The argument from authority
holds no water with me. Penrose is a crackpot, as is John Cramer,
and none more so than Joy Christian (who claims to have disproved
Bell's theorem and got himself a nice position at the Perimeter
Institute).
Penrose is not crackpot. he made an error, and that's all. All
logician can see the error. But you have not yet find an error, I
guess you would have shown it to me since. you refer to another
theory, which is very fuzzy, seems to rely on physics, admit partial
conscious being, etc. You cannot invoke a personal theory to show an
error in a reasoning, you have to show what is non valid.
Who are these people? You are unable to explain yourself,
The subject is difficult. But all people to which I explain
understand, soon or later.
so maybe it would be better for me to correspond with someone more
articulate. Who would you say is best equipped to explain the
"prescience" aspect of MGA among these other people? I guess it
should be one of the physicists, they are more likely to talk in
clear language.
You may read the detail report by Gochet (in french). Physicist knows
nothing about mathematical logic and theoretical computer science. Or
just tell me what do you don't understand. You are the one saying that
a piece of a machine having no role in a computation has still a role
for the consciousness attached to that computation.
It makes the computer science notion of emulation senseless. Actually
it makes computer science a branch of physics, like Landauer defends,
but this leads to physicalism and to the idea that to simulate a
brain, I have to simulate much more than what is necessary for the
computations (including the counterfactual).
Is there someone else who can explain this?
I would also like to know who these people are. What media has
discussed your work? Also you have left me (a physicist) out; I
have been saying it for years here and to your virtual face.
I got the price for the best thesis in 1998 by the journal Le MONDE,
then this has only extend moral harassment from Brussels to Paris (and
elsewhere). If you insist to have name I can give you them out of
line. None have published anything in this field. They told me that
the concept of "consciousness" is crackpot (in the seventies), but
also quantum computation, etc. I think you just have prejudice on me,
not based on any reading of my work. You never mention AUDA and my use
of the self-reference logic, which provides an utterly clear
arithmetical interpretation of what I say in UDA (but needs
mathematical logic). I do derive the embryo of physical laws in the
manner made obligatory by the UDA argument. In your paper you mention
"all computation" without mentioning the universal dovetailer. you
don't mention the first person indeterminacy. Most on this list are
familiar now with those notions.
I don't. There is no motivation given for such a passage.
Counterfactuals are needed but they could be Platonic or physical.
The whole point is there. Especially if you address to physicists. Of
course they are not to well placed to appreciate, given that it made
physics emerging from the non physical (the mathematical, or the
arithmetical).
What you may consider insults (and I can only guess) is nothing you
have not brought on yourself.
You are the one repeating term like non sense or crackpot instead of
saying things like I don't understand this or that, here or there in
this or that paper or posts. Asking for more precision you refer to
your on theory, which is not the way science work. You may use your
theory to say that that an argument is invalid, but then you have to
do it.
To say that consciousness supervene on laws and counterfactual, and
then to blur the platonic/physical distinction does not help to
clarify your saying.
Do you have a problem with the seven first step of the uda? I can only
guess a problem with the 8th step (MGA), based on the fact that you
pretend that an inactive piece of material is needed for the
particular consciousness (and not the particular computation).
You don't seem aware that computation is a notion definable in
arithmetic, and which has a priori nothing to do with physics.
Bruno Marchal
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.
