On 09 Mar 2015, at 15:41, Telmo Menezes wrote:
On Sun, Mar 8, 2015 at 7:42 PM, Bruno Marchal <[email protected]>
wrote:
On 08 Mar 2015, at 17:23, Telmo Menezes wrote:
On Sun, Mar 8, 2015 at 3:42 PM, John Clark <[email protected]>
wrote:
On Sun, Mar 8, 2015 at 7:36 AM, Telmo Menezes
<[email protected]> wrote:
>> I have generally been inclined to agree with JKC that natural
selection can't act on consciousness, only on intelligence; so
consciousness is either a necessary byproduct of intelligence or
it's a spandrel.
> If you assume materialism.
In science if you don't assume materialism then one theory works as
well (and as poorly) as any other theory.
It seems you are confusing materialism with falsifiability. A
theory can be falsifiable without assuming materialism. In fact,
most scientific theories do not have to assume materialism at all.
They just make successful predictions about future observations.
That's why all non-materialistic theories of consciousness are such
a colossal waste of time, they're so bad they're not even wrong.
Just like the materialistic ones.
Well, the materialist theories just fail.
I wouldn't say they fail, I would say they don't exist in the
Popperian sense. I have never seen a materialist theory of mind that
is falsifiable in 3p.
I have seen a materialist theory of mind that is trivially
falsifiable in the 1p: the theory that consciousness doesn't exist.
The fact that people seriously propose this latter theory makes me
take the possibility of philosophical zombies more seriously....
That *is* the mind-body problem. That is the hard problem of
consciousness. The first hard thing to do to solve it, when assuming
mechanism, is to abandon the concept of matter, which is really only
a "god-of-the-gap" in both the explanation of mind and of matter (or
to abandon the notion of consciousness, of course)
I agree that -- unless someone finds a flaw in your argument --
assuming mechanism makes any materialist theory of mind fail.
The problem of the current institutionalized religions is that they
take for granted Aristotle's primary matter. But even for Aristotle
himself, this was a ... focus of attention on what happens nearby,
the greeks already got the "reversal". Some like Xeusippes and the
(neo)pythagoreans were open to a (simple?) mathematical reality.
Materialism fails, because there is no evidence for primitive
matter, not does the notion of primary explains anything. Matter,
like God (when used in argument) are concept equivalent with "now
shut up and obey the rules".
Matter can be seen as a simplifying assumption, formally equivalent
to a strong physical induction, which in particular is violated a
priori with comp. They introduce a simplifying identity link between
1p and 3p, which is not verify in arithmetic, nor in the SWE actually.
And I use "materialism" in its weaker sense of metaphysical or
theologic doctrine assuming a primitive material reality. By
"primitive" I always mean something which is estimated as having to
be assumed.
Quite the contrary with elementary arithmetic or Turing equivalent.
It is a Turing universal structure, and in a simple sense, it can be
shown that it has indeed to be postulated. You cannot reduce it at
anything which is not already Turing equivalent with it, and with
Church thesis, this is of a considerable generality, yet a highly
non trivial obeying precise theoretical laws.
The second recursion theorem of Kleene provides an abstract biology,
an abstract psychology/theology, and an abstract physics, on a
plateau.
I have a strong feeling there is a relation between recursion and
the arrow of time, I think we discussed this before. Could you
elaborate on the abstract biology?
Descartes who want animal being machine search all his life how to
build a machine capable to produce a machine identical to itself: in
other word: to self-reproduce. Driesch, an early embryologist will try
too, and conclude that this is impossible (and argue that we are not
machine from that).
I will find the solution in Watson's book on molecular biology, but it
was still unclear if the possibility was not taking into account
unknown properties of biochemical components, so molecular biology
will be an evidence that machine can reproduce, but it did not give a
clear conceptual understanding of that issue.
Well, I will found that conceptual and clean solution in the book by
Nagel & Newman, and that is what decided me to do math instead of
biology. From that point I knwe that life, mind, and all that, does
not need special property of matter, other than being able to
implement Turing machine, or Lisp programs, combinatores, etc.
The basic idea is very simple: if D applied to X duplicate X into XX,
then DD will produce itself. It is the usual Cantoirian (double)
diagonalization.
Unfortunately, like the mocking bird, that is Smullyan's name for the
combinators M which on x gives xx, MM does not stop, and functionally
has typically no output.
The solution of reproduction needs a specific way to not evaluate the
output, and this can be done with the procedure of quoting: if D"x"
gives "x"x"", then D"D" gives D"D". That is the correct, conceptually,
solution of the problem.
What is still not clean here, is that the quote procedure is not well
defined, and there is a bit of a treachery, as we substitute a
variable (x) by "D", in the scope of a quote. This is not a big
problem, and it can be solved by a special substitution function, like
the subs-except-quote (subst-sauf-quote, in french) that I implemented
for my paper "Amoeba, Planaria and Dreaming machine" (and detailed in
"Conscience et Mécanisme".
The totally clean and conceptually clear solution can be done using
the phi_i and the W_i, but I keep this for later, as most people are
not yet familiar with them. The planaria case is more difficult, and I
use some generalization of the second recursion theorem of Kleene,
which provided all what we need. This has been seen and exploited by
John Case, also. You can look at the reference in the paper cited above.
The machine already explains all this, the problem comes more from
the humans who don't listen.
This refusal to listen must eventually be part of the machine's
explanation, correct?
I am not sure. It comes from the mechanizability of the
diagonalization procedure. Gödel already saw that a machine (or an
axiomatizable theory) once Löbian, can prove its own Gödel's theorem.
So when Lucas and Penrose pretend to know something that the machine
does not know, they are simply wrong, the machine knows that if she is
consistent then she cannot prove its own consistency, nor even taking
it as a new axiom. It is just that no Löbian machine at all can know
that they are correct or consistent, and unless Lucas can prove he is
correct, his proof amounts to saying that he can apply the
diagonalization, but that is what the machine can do also.
Judson Web already saw that machines can refute the Gödelian
argument by Lucas, "against mechanism". Penrose argument is a
variant of Lucas, and again the machine can refute it.
Going on a slight tangent: when I read Penrose many years ago, I
wondered about randomness.
What happens when you augment a Turing machine with input from
something like http://random.org ?
Theoretically, such machines can solve more problems, but not compute
more functions. Here is the reference:
KURTZ S. A., 1983, On the Random Oracle Hypothesis, Information and
Control, 57, pp.
40-47.
Even a bigger tangent, motivated by a bit of Internet history that I
came across recently:
http://www.patrickcraig.co.uk/other/compression.htm
There is something weird about the fact that true randomness cannot
be generically compressed by any conceivable algorithm, nor can it
be reliable produced by a turing machine.
Well, that is almost the definition of true randomness. But I tend to
believe that the digits of PI are also "truly random" (it passes all
the statistical test of randomness, at least empirically) despite
being very much compressible.
If you iterate the WM-duplication, you can understand that the number
of programs will not grow as much as the number of sequences, and so,
like almost all functions from N to N are non computable, almost all
sequences of length n, with n big, are random: they just don't have a
much shorter algorithm than "print 0010111 ...".
Then there is the fact that what neural networks do has a lot in
common with data compression
What gives?
Expression like F=ma, or the SWE are incredible data compression. Data
compression is what theories do, and it is what rational machines do
when extrapolating, or inferring inductively. Of course, in the
details, this can be nuanced a lot, but it is the idea. You can see
all programs computing a function from N to N as compressing an
infinite amount of data: the input output of the function computed. of
course the inductive inference machine go from the data phi_ to the i
or the j such that phi_i = phi_j, when the universal machine go from
the i to the phi_i. Competence, programming, and learning are the
quasi inverse operation of computing. Again, up to some nuance, as
computation does not exceed the sigma_1, but theorizing is not bounded
by complexity a priori.
Bruno
Telmo.
Logicians knows that when they care about the question.
Bruno
Telmo.
John K Clark
--
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.
--
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.
