On 30 Jul 2014, at 14:36, David Nyman wrote:
On 30 July 2014 09:03, Bruno Marchal <[email protected]> wrote:
I think (maybe pace David) that materialism explains well
consciousness, by using comp. The problem is that such explanation
makes *matter* incomprehensible.
Well I must confess I'm not entirely pacified yet. Surely the whole
point is that if the second sentence is true it contradicts the
first! If the assumption that materialism can use comp to explain
consciousness in fact leads to the absurd conclusion you describe,
then materialism *cannot* use comp to explain consciousness, well or
otherwise.
What I meant is that a materialist can apply the Teaetetus definition
on the material machine, and get the correct (G, G*, S4Grz) theory.
Of course, we know better, and we say that if *that* works for the
mind, then it has to work for matter, on the sigma_1 (UD) restriction,
with the "p" and "<>p" nuances.
The materialiste can be in that sense fair with comp for accepting the
brain transplant, and accept strong AI, but still just ignore that if
this capture consciousness, then matter should follow from the logic
of sigma_1 based self-reference.
We agree, the position would be incoherent with respect to the mind-
body problem, but would still be correct as far as those material
machines are concerned.
The trouble is, there are a lot of nuances that tend to obscure the
logical steps of the argument, particularly in assumptions about the
scope of "reasonable" explanation. Brent and others point to
parallel accounts of neural activity and conscious self-reporting
and ask "what more could be required in the way of explanation?".
AUDA may indeed give a clue to the direction in which explanation
could be extended, beyond ostensive parallelism of this kind,
particularly with respect to the central logical puzzles of mutual
reference between 1p and 3p regimes. But this would require us to
relinquish any prior commitment to primitive materialist assumptions.
My recent forays into thought experiment have been an attempt to
articulate my own (still persisting) intuitions about the intrinsic
limitations of reductionist explanation under strictly materialist
assumptions (i.e. without either tacit or explicit reliance on
supernumerary posits). ISTM that one of the problems in reaching any
kind of stable agreement (or even disagreement) on these issues is
equivocation over the terms of reference. Consequently I've tried to
make my own view of the reductionist assumption clear: i.e. that
explanation of phenomena at any level whatsoever can in principle be
reduced without loss to accounts of the action of a finite class of
"primitive" physical entities and relations. Of course, this tends
to lead to disputation over the sense of "without loss", but I'll
come to that in due course.
Stated thus baldy and strictly eschewing equivocation, reductionism
entails that it is misleading to consider any derivative phenomenon,
above the level of the chosen explanatory basement, as having
"independent existence".
But this is a big ambiguous to me. As you know I make clear, at some
point, that at the ontological level, we need only a tiny part of the
arithmetical reality, which corresponds to the effective part of it.
Here 0 is clearly primitive. Would you say that s(0), s(s(0)), ... are
no more primitive, as they are derivative construct made possible by
the use of the successor axioms.
I use "primitive" in the sense of what is assumed, then what exist is
the truth or falsity between all possible arithmetical relations.
A priori, there too much. That's why there is that measure problem.
Under our substitution level, that we share (thanks to Everett!) we
are confronted with a battle between all universal numbers, somehow.
Strictly speaking (and strictness is essential for the succes of the
argument) such phenomena are both explanatorily and ontologically
dispensable. It's just that in extenso the proofs are a little
longer! I've offered analogies in terms of such things as societies
and football teams (you can easily supplement these with your own)
in terms of which this consequence of reductionism is rather obvious.
But for some reason it stops being "obvious" in the matter of matter
itself. On reflection, the reason is not so elusive: i.e. we
"directly experience" such higher-level phenomena in an unreduced
form. Hence none of us (and that includes Professor Dennett) can
avoid the fact of encountering, and discoursing in terms of, a
"reality" in unreduced high-level terms, even though our "best"
explanations actually rule out the other-than-metaphorically-
independent significance of any such levels. If you doubt the degree
of cognitive dissonance this engenders, consider the general tenor
of disputes over "free will".
OK.
This is the point at which the parallel with any other reductionist
analogy breaks down. Nobody would seriously claim, beyond a manner
of speaking, that football teams amount to anything other than the
aggregate action of the persons that constitute them. But on the
other hand almost everyone (pace Daniel Dennett) would claim direct
access to a reality that is something (even if we can't agree
exactly what) that is, at least, categorically distinct from any
description of the aggregate action of the material processes of the
brain. The same distinction, however, can't be claimed for
"computation", on the assumption of material reduction. Just as in
the case of the football team no instance of computation can escape
reduction to material tokens that have been contrived, under
suitable interpretation, to embody the necessary physical action.
OK. Leibniz saw this, I think.
There isn't even the saving grace that we directly perceive
computation in unreduced form.
Computation is a mathematical, actually arithmetical notions.
What we actually perceive are macroscopic physical devices that, by
assumption, produce all their effects entirely in terms of basic
material processes that are fully subject to reductive explanation.
Every explanation we give in terms of computation can in principle
be replaced without loss by a description of a physical process.
Yes, because the physical reality is itself supporting many turing
universal systems, like us, with comp, and like computers (without the
need of comp!).
This is the underlying reason that Alice's net behaviour can persist
unaltered even after disruption of any putatively "computational"
organisation of her brain.
Yes. And we can still ascribe consciousness, like we can do when
looking a movie, and entering it, somehow. But the consciousness is
not in this or that particular states, it is in all the stories, as
described by the sigma_1 complete part of truth.
Under physicalist assumptions, Alice is first, last and always a
physical device. Indeed, were that not the case, it would be
difficult to see how any "physical computer" could ever be
manufactured! On this analysis then, it can hardly be coherent to
claim that any association between consciousness and matter obtains
"qua computatio". If any such association were to obtain under these
assumptions, it would perforce be "qua materia".
I think that's the reductio ad absurdum, showing that consciousness is
more a semantical fixed point for machine (mathematical being) self-
observation.
The worrying thing (and it worried me even more in the light of your
recent objections) is not that this analysis lacks power, but rather
that the "flesh-eating microbe" of reductionism might devour even
the computational structures putatively derivable from simpler
number relations and hence nullify comp.
It would, if it was not protected by "Church's thesis", as Kleene
called. After this, to nullify comp, you need to sent all computers in
jails.
They are born universal dissidents, like some parents can witness.
My own intuition (still under review I admit) is that in an
important sense realism about these structures depends ultimately on
their generalised epistemic consequences. Is 17 prime independent of
anyone's knowing it? My response to this oft-posed question is that
if such putatively significant truths do not actually entail
somebody's knowing them, their possible consequence may be moot.
But with comp, "17 is prime" and similars does entail the existence of
many knowers on this. The solution of the universal diophantine
polynomial emulates all our dreams already, and dreams are structured,
they obeys laws.
Certainly any mathematics devoid of such epistemic consequence could
play no role in comp.
The epistemic arise in computer science (and thus in arithmetic) by
the self-referential ability of the numbers. They already know,
somehow, that as far as they are "correct" their soul is not reducible
to anything 3p, from its own view.
Arithmetic is postulated in the first instance as an ontology whose
truth is entirely "independent of us" but then the (truly
surprising) discovery is elucidated that such truths turn out
directly to entail our knowing them!
That's what might happen, apparently, if the observable obeys the
arithmetical quantization, which seems to be the case with QM.
If this is indeed the case it would surely settle the matter of the
"platonic" existence of mathematics in the most astonishing (but
satisfactory) manner.
I think it is made obligatory, but it force physics to arise from the
intensional variant of self-reference, making the, perhaps naive,
classical knowledge base description of matter testable.
It's not lost on me, by the way, that my "strict" account of
physicalism could still lay claim, prima facie, to an analogous a
posteriori epistemic justification in terms of a conscious knower.
It's just that - on the assumption of *strict* material reductionism
- the further supervention of computation on physical tokens could
play no role in the explanation.
But then how could it play a role in arithmetic? It is just that with
comp, we have to extend's Everett on *all* computations, on the full
sigma_1 reality, and justify the wave from a sort of competition
between all universal machines.
The knower is redeemed due to the arithmetical reality to distinguish
what the machine can justify about herself and what is true about
herself, and all the nuance brought by incompleteness. Like truth,
the knower has no name. When Löbian, not only it has no name, but it
can refute any names or 3p description, doing so it transforms itself.
Bruno
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.
