Re: What's the answer? What's the question?

[Bruno Marchal]

On 06 Jul 2014, at 05:18, meekerdb wrote:
On 7/5/2014 5:08 AM, David Nyman wrote:
On 5 July 2014 06:27, meekerdb <meeke...@verizon.net> wrote:

Ok, maybe it's mostly a matter of semantics. I don't exclude  
things as not
existing just because they are not part of the primitive ontology.
But of course I haven't been saying these things don't exist. On the
contrary, I've been labouring to differentiate, for the sake of this
discussion, two distinct senses of "existence". The first sense picks
out the basic ontology of a theory and the second refers to whatever
can (putatively) be derived from that ontology on the basis of  
further
epistemological considerations. And I've been pointing out that
nothing whose "existence" is picked out only in the the latter sense
can be claimed as having any independent relevance (ex hypothesi
reductionism) in the evolution of states defined in terms of the
former.

In physics the stuff that is most "primitive" in a model is also  
stuff who's
existence is least certain, e.g. strings, super-symmetric particles,
space-time quanta,...  While the stuff you would say doesn't  
really exist is
the most certain - including the instruments used to infer the  
primitive
stuff and records of the data taken.
That's simply not relevant to the point under discussion. Of course
what's most certain is whatever is directly epistemologically
available, but we're discussing what our theories may actually imply
if we take them *seriously*. The hypothesised ontology may be far  
less
certain, but we are persuaded to entertain it in the first place
precisely because we conjecture that the things of which we *are*
certain are ultimately its complex epistemological derivatives.

But this ceases to
be the case when we propose a second-order relation like  
computation
as the "physical correlate" of consciousness, precisely because it
vitiates the idea that such relations can be anything other than a
manner of speaking, in terms of the *ontology* of a reductive  
physical
theory. Hence, to attribute the ability to evoke conscious states  
to
such imaginary or virtual relations would seem to invoke a sort of
ontological magic.
I don't see it as any more magic than making mountains out of  
rocks.  You
seem to be invoking an argument from incredulity: Consciousness  
just can't
be made out of physical stuff or processes.
It doesn't take magic to make mountains out of rocks; it requires  
only
that we respect the distinction in a theory between an ontology and
its epistemological derivatives. What I was saying is that, under
physical reductionism, we can't coherently claim the "ontology of
consciousness" to be computation, because the relata picked out by
computation simply can't be justified as being independently
effective. The relata of computation are, ex hypothesi,
epistemologically derivative, abstractions of a *uniquely effective
micro-physical ontology*: "strings, super-symmetric particles,
space-time quanta". This ineluctable degeneracy to the basement
entails that, if you want consciousness to be "made out of physical
stuff or processes", that's where you'll find your parts kit.

But in any case it is surely rather obvious that we have no business
looking for consciousness to be "made out of stuff", unless we are
willing to take the stuff alone (i.e. the 3p part) to be  
consciousness
tout court. That would be to eliminate the 1p part and if we stick
closely to our theoretical analysis such elimination is obviously
incoherent on its face, as the very claim invokes the epistemological
derivatives it purports to exclude. Of course I know this isn't what
you mean: you accept the 1p part. But the point of my analysis has
been to argue that unless we are prepared to consign the latter to
mystery (i.e. inaccessibility to further explanation) we should
perhaps question whether we aren't being misled by very idea of
consciousness being "made out of something".

Arithmetic, in the first instance, is simply
posited as the minimal ontological assumption for the  
construction of
an explicit epistemology (i.e. a theory of knowledge and  
knowers); IOW
what physics explicitly eschews at the outset. From that point the
explanatory thrust hinges on epistemological considerations and  
hence
can no longer be straightforwardly reduced to the first-order
ontology.
Yes, that's an interesting aspect of Bruno's theory.  He identifies
"provable" with "believes".  But the the same kind of thing can be  
done in a
physical theory: "believes" = "acts as if it were true". There's  
even a
whole theory of Bayesian inference based on bets.  It may not be  
right, but
it's a theory of epistemology.
Yes, but it's a theory of epistemology "after the physical fact". It
assumes without further justification what it wishes to prove,
No, it defines a certain kind of belief, just as Bruno identifies  
belief with "provable in some axiomatic system" (which you must  
admit is not a standard meaning of "belief")


To get the consequences, you need only to acknowledge that *you* do  
believe in elementary arithmetic. We can identify "belief" with 3p- 
provability for any machines which is ideally correct in the  
representation of its belief (and "rich enough" to be Turing universal/ 
Löbian). It works thanks to the fact that []p -> p is not provable for  
all arithmetical p, despite it is true.

But this is not epistemology, which is concerned with knowledge/ 
consciousness. In this case it cannot be identify with any 3p notion,  
and this for logical reason. Somehow physicalism use an implicit  
identity between body ([]p) and consciousness ([]p & p). Belief is not  
epistemic, it is doxastic.




one can identify belief with certain actions in context.   I don't  
know what you mean by "after the physical fact".  If it's a physical  
theory of belief then of course it's explained in terms of physical  
facts.  You seem to reject this as though it's obviously wrong.  It  
may be inadequate or Bruno's theory may be better, but you seem to  
think it's somehow heretical to have a physical idea of what  
constitutes belief.

What might be "heretical" is not the explanation of belief with 3p  
things (physical or arithmetical), but the identity between  
"knowledge" (or truth, consciousness, etc.) with 3p things.




as when
we speak of "acting as if it were true" we have *already abstracted*
both ourselves and our actions from the framework of an exhaustive
physical reduction that has been hypothesised at the outset as being
uniquely effective.

Sure, it's a physical theory so it's in terms of physics.

That's the whole point and burden of my analysis.
So on that basis I disagree that "the same kind of thing can be done
in a physical theory". On the comp assumption, by contrast, belief in
a truth and "acting as if true", in conjunction with their truth
content, are alike derived from the outset as consequences of a
fundamentally epistemological theory.

A consequence of a hypothetical definition.

A consequence of the most standard definition in the field. I am not  
sure "hypothetical" can be applied to "definition".




And I don't see that comp has been able to derive anything about  
human actions.

On the contrary, it explains the human theological trap, and it  
classifies the error in fundamental studies by the confusion between  
the hypostases.
But then I remind that comp is not an explanation: it is a problem. It  
provides a purely mathematical version of the mind-body problem or  
some of its subproblems.




Not to say that can never do so, but I don't see it as a de facto  
virtue.

The de facto virtue of comp is that it is simple, leads to precise  
technical problems, is testable, and has no real opponents today  
(except literal fairy tales, but those are not "scientific opposition").




One might therefore say that
action, belief and truth are hypothesised as being complementary or
co-effective, rather than hierarchical-reductive, in relation.

Truth in comp only refers to mathematical truth of the form Exf(x).  
It's a long way to connect that to "I see a dog."

"i see a dog" is more like []("there is a dog") & <>("there is a dog")  
& there is a dog (in my mental space).





I think in the case of consciousness, explanation as opposed to
engineering has to take foundational questions of knowledge and
reference as seriously as those of physical phenomenology. And I  
also
think comp at least provides a possible model of how progress  
could be
made in this direction.
But not a physical, neuroscience based model.  You reject that, in  
spite of
the fact that it has had considerable success.
I don't reject it at all. I'm only going so far as saying that
ultimately it may lack the explanatory moxie to take us as far as we
could go under more comprehensive assumptions.
Ok, so long as you say "may".

Your stance, I think,
is that wherever a future neuroscience is capable of taking us is as
far as any reasonable person could expect explanation to reach. To  
put
it more baldly, whatever can't be known in this way just isn't
knowable. All I've really been trying to argue for in this thread is
that this limit may in the end turn out to be an artefact of a
particular explanatory strategy.

Maybe.  Maybe not.  Even in Bruno's theory not everything true is  
knowable.

It is indeed not knowable in the 3p way, but the point is that the  
sould is not Löbian. Machine's knowledge cannot have the same logic as  
machine's belief. the first person, although related to an infinity of  
machines (in the mind of god), cannot be a machine from the machine  
first person point of view.

Bruno



As to success, I don't recall Einstein being taken to task for  
dissing
the "considerable success" of Newton's theory. In the end any
paradigm, however successful, may suffer the fate of being subsumed
into a more comprehensive one if certain puzzles stubbornly resist it
for long enough. And I'm suggesting that one early but perhaps
worrying symptom that this fate may lie somewhere up the road for
physical reductionism is a persistent tendency to trivialise,  
mystify,
or eliminate epistemological puzzles that seem to resist capture
within the framework of that theory.

So what does an answer look like?  Is Turing completeness enough -  
that's
what Bruno says.  But apparently it makes you and a jumping spider  
equally
conscious.  I don't think that's a very good answer.
Where is it written that I and a jumping spider are equally  
conscious?
That is, on the assumption that I'm not actually a jumping spider,
which I guess is not a matter for complete certainty, in the
cyberverse. I think that comp implies only that I and a jumping  
spider
may share a particular species of self-referentiality that is
representable in any Turing complete system (indeed, that this is the
basis of the "I" shared by all such subjects). Beyond that
commonality, the spectrum of subjectivity (i.e. its possible
"objects") would extend asymptotically towards infinity, I guess, but
always according to the specifics of the logic and statistics
extractable from comp. At least, that is the hypothesis and the
project.

OK, I can buy that.
Brent

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 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.