On 20 Oct 2014, at 00:56, David Nyman wrote:
On 19 October 2014 17:48, Bruno Marchal <[email protected]> wrote:
On 19 Oct 2014, at 15:26, David Nyman wrote:
On 19 October 2014 02:10, Stathis Papaioannou <[email protected]>
wrote:
Whether I find it satisfactory or not is a different question. The
point I was making is that people who find it satisfactory express
this belief idea by claiming that consciousness does not exist.
Assuming that you don't, in fact, find it satisfactory, I'd be
interested in your reasons. Given the assumption of the exhaustive
adequacy of physical reduction, Graziano would appear to be quite
correct in his assessment that the idea of any "left over"
phenomenon, after correlation of conscious states with the relevant
physical processes, is "physically incoherent". On the same
assumptions, we clearly cannot cite any *judgement* to the contrary
as evidence of any such supernumerary phenomenon, as any such
judgement must likewise be nomologically entailed by physical law.
If so, what reason can you cite for believing that there is any
such thing?
Very good question, of course: the hard question.
Note that up to some point, we can "eliminate consciousness" to, in
appearance.
Yes, because what is empirically available is restricted (by
definition) to the physics of appearance.
(I need to be technical on this, but I might try to translate latter,
or to simplify, or to criticize)
Actually I was thinking about something else, like reducing the mind
body problem to the body problem, like in the UDA, and then extracting
physics from that problem (by finding the right statistic on the
relative personal diaries).
This might have made sense, if the quantum logics were appearing in
the S4Grz1, Z 1or X1 logics, and forgetting all about the S4Grz1*,
Z1*, and X1* star. (I can argue this would have lead to solipsism (no
first person plural discourse) and to a Quantum mechanics with
collapse, in fact superposition would not be contagious to the
*conscious* observer. This would have led to a QM with a consciousness
reducing the wave packet, but only in the diaries, and that could be
taken as an illusion in some coherent way.
It is avery funny theory: it is a non-collapse QM, where consciousness
describe itself as the collapser of the wave. There is no collapse,
only because that consciousness does not exist!
(Not sure it makes sense for a non zombie tough!)
If QM appears only at S4Grz1the idea above would make some more sense,
but still hard to swallow for any non-zombie entity). The reason here
is that S4Grz1 = S4Grz1*
But now the qualia and quanta appear at the star pov (Z1*, X1*,
S4Grz1*), and all the differences between S4Grz, Z, and X (and the
comp one S4Grz1, X1, Z1) come from the G/G* splitting, so the true non
justifiable invites itself in the picture, keeping the many nuances
brought by those different logics).
In fact I revive some of your old critics of comp, like it can be
considered as eliminating consciousness too, because it might make
logical sense, in case physics did not appear exclusively in the star
logic. A good thing, because it makes Everett-QM confirming the
sharing of the histories by many people. This is better than Albert-
Loewer theory, which get multi-solipsist.
The default assumption, as Graziano succinctly notes, is that the
details of this apparent physical mechanism (at least in some ideal
form) exhaust both the ontological and the epistemological catalogues.
Only with actual infinite magic in the details. I guess you saw this,
then formally it is even clearer, but a bit of this could have been
the case, if physics appear in the non star logics. Not sure the
computation can interfere in that setting. We get all isolated in such
setting, from each others.
Consequently, in this sense, appeals to the putative existence of
anything over and above such an exhaustive account must be
"physically incoherent". If one takes a sufficiently hard line (and
I do!) it becomes apparent that this mode of explanation gobbles up
competitors like some inexorable flesh-eating microbe. Anything
"meta-physical" (such as "computation", under these assumptions)
merely degenerates, under observation, into one or another physical
approximation.
Yes, and even physics is no more clear, because the theories use
explicitly arithmetical relations. Indeed in string theory you need to
believe that some infinite sum of all natural numbers is equal to
-1/12, independently of you, to get the right mass of the photon!
Then computation is a notion which exists even more as it does not
need any axiom of infinity, and the existent computations are provably
existent, even already in RA. No need of induction axiom.
In fact, for a logician, to believe in the physical laws, together
with the belief that they apply here and now, is equivalent with a
*very* strong induction axiom. Induction are self-limitation axiom.
When you define an object per induction, like defining the number by 0
is a number, and if x is a number s(x) is a number, and then you want
to say that only those things are number, and nothing else: you can
add induction axioms, up to make the notion haing a unique
interpretation, but in that case you will need to induce on
transfinities of object, and the theory is no more proof-checkable. In
some context you can been led to contradiction. may be the primitive
universe is a contradictory notion, even without comp.
Sorry for being perhaps too much technical, but it helps me to think a
bit aloud, inspired by your remark.
With comp, asserting the existence of a reality leads to a
contradiction, in case there is a reality!
Only religion studies or assume realities, and as such are
contradictory with comp at the start. But the original theologian were
not assuming a reality, only interrogating it, and that is present in
many part of religion, but usually not at the per-authority mainstream
levels (for historical reasons).
Take the UDA: the first person, there (unlike AUDA) admit a pure
third person description: the content of the diary that the person
takes with her in the multiplication (in arithmetic or in the QM
universal matrix, whatever).
The magic is described in the diaries, the person get information
(in shannon sense) from apparently nowhere (they don't feel the
split).
To share that information with other, we need the first person
plural, and the hope that the computations which makes us interact
are among the winning one. That might already suggest universal
group like the unitary group.
But up to now, this picture does not yet address the hard question,
despite it explains the content of the personal diaries.
What will specifically address that question is the Theatetus idea.
By definition, the owner of the diary has the content of the diary
as experience, making them true at that self-observation level:
So there is, as it were, a primary "level of truth" - that of self-
observation itself - that is given *by definition*?
The truth is always the truth of one or more arithmetical proposition,
interpreted in the "well known" (N, 0, +, *) structure.
Now, the machine (a 3p notion, definable in arithmetic) can refer to
itself, and to computable transformation of itself. This is explained
by the Dx = "xx" method, or relies on Kleene's recursion theorem. All
this correspond to 3p truth, and indeed provable by theories like PA
or ZF, and satisfied in their model (but for ZF the notion of standard
model is more doubtful, ZF might be believing already too much.
But what happen now, is that the machine cannot link in a general way,
the justifiable 3p truth ([]p), with the truth of p, that with the
fact that p. So []p and []p & p does not obey the same logic. It makes
the "[]p" into a belief, but it makes also the []p & p" into a
knowledge, or subjective truth, having no definition available for the
machine (only a fuzzy pointer, like "I" or "God", or "Universe", etc.)
I mean a level that is, in itself, distinct from whatever relative
truths may be implied by the content, or what-is-observed.
It is more elementary I think. It is the truth of arithmetical
proposition, and machine uttering or accessing such proposition. When
true, the corresponding "[]p & p" (the knower associate to the
machine) obeys a different logic than the logic of the utterer as seen
from a 3p point of view. Accepting classical ancient definition: self-
referential machines have soul.
it guatantie the link between provability and truth,
Guarantee in what sense?
In a very naïve sense, that by defining knowledge, not just by its
company with truth (after all, as we limit ourself to correct machine,
this would again make []p and ([]p when p) the same, but by its
formal company with truth, making the individual []p into the non
definable (by the machine) []p & p. This obeys a very different logic,
which describes well the phenomenological account of the first person
(no name, the owner of a body, its S4 knowledge logic, ...).
The machine cannot say "I know that snow is white", but the machine
can say "snow is white, and I believe that snow is white". For each
individual she can assert such knowledge, and as we know she is
correct (assuming arithmetic, of course) we can see that this obeys a
different logic than the 3p-correct-justifiability, and indeed a
knower logic, even an non definable knower logic.
but by incompleteness, you have to make explicit in the definition
of the "knowability" even for the correct prover/believer (which is
an amazing consequence of Gödel Löb theorems). This entails the
splitting between the truth accessible to the machine from its
perspective, and its ability to see that too,
Do you mean: to see that the accessible truth is restricted to its
(the machine's) particular perspective?
I mean that the machine is made unable to see that she is a machine.
She can bet she is a machine proving a theorem, but she cannot related
its "truth feeling" for that theorem, with any description of the 3p
way used to get that theorem.
Graziano seems to talk like he knew the substitution level (in UDA
term). We can't know that. By betting on a substitution level, we
equate the []p and the []p & p, acting like our own G*, but we can't
justify our correctness, which would be needed for doing this
scientifically.
Graziano is dangerous, because he seems to believe that we could
offer artificial brain to a patient guarantying the survive, where in
fact it has to be always a bet. The choice of a level, is a religion
instantiating computationalism.
Computationalism will split into two main forms; the low-level-ist and
the high-level-list. That will mirror the non-comp/comp division, but
also the materialist and the mystic (which here will means those
saying yes to the doctor, because they are attached to the material
reality, and those saying no to the doctor, because they want take
some rest. That is for tomorrow.
which is the case for machines believing in some induction axioms.
What axioms in particular?
If 0 is even, and if (n is even implies that s(n) is even), then all
numbers n are even.
If y = 1 is the derivative of y = x, and if (the derivative of x^n is
n*(x^(n-1) implies that the derivative of x^(n+1) =(n+1)*(x^n)), then
the derivative of y = x^n is equal to (x^(n-1) for all n).
That is all the formula of the type
[ P(0) & (P(n) -> P(n+1)) ] -> for all n P(n). With P a property
definable in the language of the machine.
It is not inductive inference, but the mathematical induction axioms.
We can get non computable "theories" or "super-machine" by allowing
stringer form of induction axiom. We get the "arithmetical god"
formally by replacing the P(n) above by S(n) with the meaning that is
an arbitrary set of numbers S. In that case, the theorems is no more a
mechanically generable set, and we are out the working hypothesis. In
fact when we do that we do analysis, or second order arithmetic. That
remain Löbian, but is no more machine. of course Löbian machine can do
theories about that, like us, and arithmetic contains a lot of those
no-machine entities. They are like Oracle machines, super-machine,
etc. They exist in arithmetical truth, but above the sigma_1 level
(the machine accessible level).
I agree with John Clark that such things are not really plausible on
the physical plane, but that is an obvious complex open problem.
Actually, machine cannot distinguish gods from machine much more
complex than themselves. Like spudboy suggested, we can't distinguish
a god with an alien much more advanced than us, in case he want play
with us..
That splitting entails also the very existence of all the nuances
between the (8) points of view. The ideally correct universal
machine is born in arithmetic with already psychological and
theological internal conflicts.
Can you make a distinction here between what you consider
psychological, and what theological?
Accepting that the psyche is mainly the same as the soul, the
psychological is included properly in the theological. The theological
is concerned with the first person or third person truth that we
cannot communicate, or only with special caution (like I explained
once with the notion of consistency).
Basically the psychological is in both G and G*, S4Grz, Z and Z*.
The proper theological part is in G* \ G (the set subtraction of G
from G*, truth minus provability), and its intensional variant, Z* \
Z, X* \ X.
Crazily enough S4Grz* \ S4Grz is empty. It is not obvious to prove,
but as Boolos says, it corroborates some intutionistic confusion
between proof and truth. It is the solipsist condition of the soul of
the correct machine, apparently.
But the more it introspects, the more it get the picture of the
abyssalness of its ignorance, making it naturally humble in front of
the possible truth, and humble in front of the bridges between
truth, belief, observation, sensations, and knowledge.
Its ignorance can hardly be more abysmal than my own!
yeah ...
With computationalism, you need to come back to math. It is needed for
the computer science.
(I guess that is the bad news).
UDA pers se address only the hard matter appearance problem, I think
from some of your post you understood that AUDA does address the
hard question, and gives perhaps the most we can hope for when
assuming computation: the theory of consciousness and person is
similar as the theory of god: it is a negative theory: you are not
this, nor that, etc.
Yes, I have some sense of how the "something over and above" can be
intelligibly situated in terms of AUDA, although I am painfully
aware of the limitations in my grasp of the detail.
I can try to sum up, but sane04 already sum it up in the second part
of the paper.
The real bomb here are the discovery of the (Turing) universal
machine, and Church's thesis, and the closure of the semi-computable
for diagonalization, then Gödel's theorem, and then to get at least
the soundness theorem of G and G* for formal theories or correct
machines (the easy part of Solovay theorem).
I recommend Boolos 1979, but it needs some course in logic before. But
you don't need to understand all the details. Only a part of it. Ask
any question. (But I am in teaching period, so expect delays)
Paradoxically (or perhaps not so much) what has helped me is
sticking to a hard line on physicalist, or (which is the same)
physical-computationalist, modes of explanation, because eventually
it becomes so apparent that "something over and above" is made
incoherent by the limiting assumptions of these schemas. For this
reason, I find honest physicalists like Graziano, or Churchland more
helpful than those who fudge the issues and hence make the
contradictions harder to expose.
I like them for that too. They are clear and honest. Then they confirm
that physicalism is not coherent with the type of comp I describe,
because what I say needs the concept of consciousness: comp is mainly
the idea that my consciousness is invariant for some digital
substitution. If consciousness does not exist, then I can say yes to
the doctor who suggests to put *anything* at the place of my brain.
But Graziano and Churchland illustrates also the danger of such view.
It makes the individual into a zombie. I don't like that at a time
when politician stop to invest in education, and makes laws which take
away responsibility or people.
Many apology for getting technical, but that is what AUDA is, and I
feel its solved some possible critics you made. I just realize that
all this is indeed rather subtle, and I might need to say more on each
hypostases.
But if I am too clear, people will believe I do religion, when I am
just translating what machines explains when looking inward.
Truth is in the head of all universal machine, and the Löbian are
those who somehow found it already. Of course they will never tell
you, and they want not talk about that, so you need to do some
arithmetical surgery, or to promise them that you will never claiming
anything as true, which I do with that "assuming comp", "limiting
ourselves to correct machines", etc.
There is a sort of theological trap when forgetting that we do
theories. We fall into it more than once with the human religions, and
I think that a lack of caution can make us fall again, even in the
machine theology.
In arithmetic, it is also a fact that you have the dreams of the
machines which fall in the theological trap all the time. That lack of
cautiousness might be the reason why souls fall, in the neoplatonist
theologies. I don't know, of course. Is it temporary? Is it necessary?
Bruno
David
Bruno
Bruno
David
--
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 everything-
[email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
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.
--
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.
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.
