The Simulacra Theory of Mind
Eray Ozkural exa
"The Simulacra Theory of Mind":
The mind is a universal cognitive procedure which can faithfully
approximate a subset of the history of entire universe. This
reflective-process is at the root of subjective experience and
intelligence.
Many thanks to Schmidhuber for expressing similar concerns in his
amazing work. He has been a constant source of inspiration for me.
This view is quite possibly consistent with (the digital variant of)
Multism.
Regards,
--
Eray Ozkural
David Longley
Ozkural exa <er...@bilkent.edu.tr> writes
This looks like the (all too commonly opaque) seeping
Ozkurian/Rickertian grasp that environmental contingencies shape
behaviour phylogenetically and ontogenetically through selection &
variation (mutation/recombination).
I wonder whether trivia historians of the 20th century will ever bother
to draw the parallel between the products known as "Cognitive Science"
and those which one could pick up in Taiwan etc (e.g. "Guggi" handbags
and "Rolix" watches etc).
Give yourself an education and carefully work through "On Having A
Poem" as you have been advised:
http://www.bfskinner.org/media/Having_a_Poem.ram
--
David Longley
Eray Ozkural exa
> In article <fa69ae35.0403...@posting.google.com>, Eray
> Ozkural exa <er...@bilkent.edu.tr> writes
> >Please consider my new and somewhat radical theory of mind. I give you
> >"The Simulacra Theory of Mind":
> >
> > The mind is a universal cognitive procedure which can faithfully
> >approximate a subset of the history of entire universe. This
> >reflective-process is at the root of subjective experience and
> >intelligence.
> >
> >Many thanks to Schmidhuber for expressing similar concerns in his
> >amazing work. He has been a constant source of inspiration for me.
> >
> >This view is quite possibly consistent with (the digital variant of)
> >Multism.
> >
> >Regards,
> >
> >--
> >Eray Ozkural
>
> This looks like the (all too commonly opaque) seeping
> Ozkurian/Rickertian grasp that environmental contingencies shape
> behaviour phylogenetically and ontogenetically through selection &
> variation (mutation/recombination).
>
No. It's a realization of how important the concept of a universal computer is.
> I wonder whether trivia historians of the 20th century will ever bother
> to draw the parallel between the products known as "Cognitive Science"
> and those which one could pick up in Taiwan etc (e.g. "Guggi" handbags
> and "Rolix" watches etc).
>
No, because that is wrong.
> Give yourself an education and carefully work through "On Having A
> Poem" as you have been advised:
>
> http://www.bfskinner.org/media/Having_a_Poem.ram
No, I don't think it's relevant.
David Longley
>David Longley <Da...@longley.demon.co.uk> wrote in message
>news:<pwSBj3Do...@longley.demon.co.uk>...
>> In article <fa69ae35.0403...@posting.google.com>, Eray
>> Ozkural exa <er...@bilkent.edu.tr> writes
>> >Please consider my new and somewhat radical theory of mind. I give you
>> >"The Simulacra Theory of Mind":
>> >
>> > The mind is a universal cognitive procedure which can faithfully
>> >approximate a subset of the history of entire universe. This
>> >reflective-process is at the root of subjective experience and
>> >intelligence.
>> >
>> >Many thanks to Schmidhuber for expressing similar concerns in his
>> >amazing work. He has been a constant source of inspiration for me.
>> >
>> >This view is quite possibly consistent with (the digital variant of)
>> >Multism.
>> >
>> >Regards,
>> >
>> >--
>> >Eray Ozkural
>>
>> This looks like the (all too commonly opaque) seeping
>> Ozkurian/Rickertian grasp that environmental contingencies shape
>> behaviour phylogenetically and ontogenetically through selection &
>> variation (mutation/recombination).
>>
>
>No. It's a realization of how important the concept of a universal computer is.
Important for what Eray? It's only important to the extent that it is
useful surely - and it's that that you and others like you keep
neglecting. What is the concept useful for?
>
>> I wonder whether trivia historians of the 20th century will ever bother
>> to draw the parallel between the products known as "Cognitive Science"
>> and those which one could pick up in Taiwan etc (e.g. "Guggi" handbags
>> and "Rolix" watches etc).
>>
>
>No, because that is wrong.
Perhaps you haven't really thought it through?
>
>> Give yourself an education and carefully work through "On Having A
>> Poem" as you have been advised:
>>
>> http://www.bfskinner.org/media/Having_a_Poem.ram
>
>No, I don't think it's relevant.
But like other things that you neglect, how do you know that? Have you
ever really thought that one through?
--
David Longley
JXStern
exa) wrote:
>Please consider my new and somewhat radical theory of mind. I give you
>"The Simulacra Theory of Mind":
>
> The mind is a universal cognitive procedure which can faithfully
>approximate a subset of the history of entire universe.
Subset, I agree.
"Faithfully approximate"???
With any luck, the mind can faithfully approximate (?) things that are
not a subset of the history of the entire universe, contingent future
events, imaginary scenarios that cannot be realized, erroneous stories
about the past, etc. And then what?
> This
>reflective-process is at the root of subjective experience and
>intelligence.
Seems rather essentialist.
I generally count that as empty of explanatory power.
>Many thanks to Schmidhuber for expressing similar concerns in his
>amazing work. He has been a constant source of inspiration for me.
???
>This view is quite possibly consistent with (the digital variant of)
>Multism.
Doesn't do a thing for me.
I don't see any necessary link between simulation and cognition. You
might argue for one, but I haven't seen that yet.
I barely tolerate ideas of representation or description, but even
granting them fully, they do not for me comprise cognition, much less
conciousness, qualia, etc.
YMMV.
J.
Eray Ozkural exa
> On 27 Mar 2004 17:18:06 -0800, er...@bilkent.edu.tr (Eray Ozkural
> exa) wrote:
> >Please consider my new and somewhat radical theory of mind. I give you
> >"The Simulacra Theory of Mind":
> >
> > The mind is a universal cognitive procedure which can faithfully
> >approximate a subset of the history of entire universe.
>
> Subset, I agree.
>
> "Faithfully approximate"???
>
> With any luck, the mind can faithfully approximate (?) things that are
> not a subset of the history of the entire universe, contingent future
> events, imaginary scenarios that cannot be realized, erroneous stories
> about the past, etc. And then what?
I considered this of course. The Simulacra Theory is meant to be an
interesting argument like my other views, I'm trying to explain a
small part. Your objection is that the mind can be creative, it can
generate computational histories which have nothing to do with the
actual computational history of the universe, in the past or in the
future. But I wonder, would not it simply mean that it is in the
present? Since the mind itself is part of the entire universe, then it
can at least "exist". That I can grant fully, and that it can also
approximately reproduce past events or future events, either internal
or external, I can grant as well.
Now, my argument is in fact a more abstract mathematical claim, but I
wanted to cast it into something tangible. I think the mind can be
possibly reduced to function approximation and other functional
analysis procedures. What do you think?
> > This
> >reflective-process is at the root of subjective experience and
> >intelligence.
>
> Seems rather essentialist.
>
> I generally count that as empty of explanatory power.
Thank you. What is your alternative view that is not essentialist
then?
> >Many thanks to Schmidhuber for expressing similar concerns in his
> >amazing work. He has been a constant source of inspiration for me.
>
> http://www.idsia.ch/~juergen/
>
> ???
Yes. Did you study his new work?
> >This view is quite possibly consistent with (the digital variant of)
> >Multism.
>
> Doesn't do a thing for me.
>
> I don't see any necessary link between simulation and cognition. You
> might argue for one, but I haven't seen that yet.
>
If you don't then you may wish to explain to me whether a universal
computer has any distinction from a particular (not universal!)
computer. My argument is supposed to emphasize this distinction.
> I barely tolerate ideas of representation or description, but even
> granting them fully, they do not for me comprise cognition, much less
> conciousness, qualia, etc.
I think we can agree that much of cognition is computational, can we?
I have another theory for qualia, but I doubt you will like it: it
reduces qualia to computation. Consciousness is of course not simply
qualia (subjective experience)... It is more than that.
Regards,
--
Eray Ozkural
JXStern
exa) wrote:
>> >Please consider my new and somewhat radical theory of mind. I give you
>> >"The Simulacra Theory of Mind":
>> >
>> > The mind is a universal cognitive procedure which can faithfully
>> >approximate a subset of the history of entire universe.
>>
>> Subset, I agree.
>>
>> "Faithfully approximate"???
>>
>> With any luck, the mind can faithfully approximate (?) things that are
>> not a subset of the history of the entire universe, contingent future
>> events, imaginary scenarios that cannot be realized, erroneous stories
>> about the past, etc. And then what?
>
>I considered this of course. The Simulacra Theory is meant to be an
>interesting argument like my other views, I'm trying to explain a
>small part. Your objection is that the mind can be creative, it can
>generate computational histories which have nothing to do with the
>actual computational history of the universe, in the past or in the
>future. But I wonder, would not it simply mean that it is in the
>present? Since the mind itself is part of the entire universe, then it
>can at least "exist". That I can grant fully, and that it can also
>approximately reproduce past events or future events, either internal
>or external, I can grant as well.
OK, if you mean "can simulate" in the sense of "is capable of".
There's a lot of metaphysics of math and language one can get into in
and around there. I just want to find it interesting that math is so
powerful it can be either right or wrong, and leave the detailed
admiration to others.
>Now, my argument is in fact a more abstract mathematical claim, but I
>wanted to cast it into something tangible. I think the mind can be
>possibly reduced to function approximation and other functional
>analysis procedures. What do you think?
I put more emphasis on the identity and agency of a computational
instantiation, and less on the details of what it does.
>> > This
>> >reflective-process is at the root of subjective experience and
>> >intelligence.
>>
>> Seems rather essentialist.
>>
>> I generally count that as empty of explanatory power.
>
>Thank you. What is your alternative view that is not essentialist
>then?
Oh, various ugly things - reductionism, causal theories, deflationary
theories of truth. But, I may be misunderstanting you again.
Certainly at *some* point the mind is interesting in large part
because it is somewhat accurately reflecting the real world. It's
just that there are so many philosophers running around who put
correspondence with truth (sic) as their highest (only) criterion, I
get twitchy about any claims that look to mirror the world in the
mind.
>
>> >Many thanks to Schmidhuber for expressing similar concerns in his
>> >amazing work. He has been a constant source of inspiration for me.
>>
>> http://www.idsia.ch/~juergen/
>>
>> ???
>
>Yes. Did you study his new work?
Just glanced at the page. It seemed to point in many directions.
>> >This view is quite possibly consistent with (the digital variant of)
>> >Multism.
>>
>> Doesn't do a thing for me.
>>
>> I don't see any necessary link between simulation and cognition. You
>> might argue for one, but I haven't seen that yet.
>
>If you don't then you may wish to explain to me whether a universal
>computer has any distinction from a particular (not universal!)
>computer. My argument is supposed to emphasize this distinction.
I like the distinction, but my focus is on the particular.
>I think we can agree that much of cognition is computational, can we?
>I have another theory for qualia, but I doubt you will like it: it
>reduces qualia to computation. Consciousness is of course not simply
>qualia (subjective experience)... It is more than that.
I think it's all reducible to physical instantiations of computation.
Can't get more hard-core than that!
J.
Eray Ozkural exa
> On 29 Mar 2004 05:44:59 -0800, er...@bilkent.edu.tr (Eray Ozkural
> exa) wrote:
> >I considered this of course. The Simulacra Theory is meant to be an
> >interesting argument like my other views, I'm trying to explain a
> >small part. Your objection is that the mind can be creative, it can
> >generate computational histories which have nothing to do with the
> >actual computational history of the universe, in the past or in the
> >future. But I wonder, would not it simply mean that it is in the
> >present? Since the mind itself is part of the entire universe, then it
> >can at least "exist". That I can grant fully, and that it can also
> >approximately reproduce past events or future events, either internal
> >or external, I can grant as well.
>
> OK, if you mean "can simulate" in the sense of "is capable of".
> There's a lot of metaphysics of math and language one can get into in
> and around there. I just want to find it interesting that math is so
> powerful it can be either right or wrong, and leave the detailed
> admiration to others.
>
Yes, I certainly mean so. In fact, that was something I got the
inspiration from the fly papers. The flynculus has much dealing in
reality.
> >Now, my argument is in fact a more abstract mathematical claim, but I
> >wanted to cast it into something tangible. I think the mind can be
> >possibly reduced to function approximation and other functional
> >analysis procedures. What do you think?
>
> I put more emphasis on the identity and agency of a computational
> instantiation, and less on the details of what it does.
>
Still, to fully specify the work of agency we must identify the
function of individual components. Otherwise, I suspect we may fall
prey to subtle conceptual errors, which are themselves hard to
identify. We might be exercising them without even suspecting there is
an error...
What do you mean by identity?
> >> > This
> >> >reflective-process is at the root of subjective experience and
> >> >intelligence.
> >>
> >> Seems rather essentialist.
> >>
> >> I generally count that as empty of explanatory power.
> >
> >Thank you. What is your alternative view that is not essentialist
> >then?
>
> Oh, various ugly things - reductionism, causal theories, deflationary
> theories of truth. But, I may be misunderstanting you again.
> Certainly at *some* point the mind is interesting in large part
> because it is somewhat accurately reflecting the real world. It's
> just that there are so many philosophers running around who put
> correspondence with truth (sic) as their highest (only) criterion, I
> get twitchy about any claims that look to mirror the world in the
> mind.
The above view originated from a causality graph I was sketching on a
piece of paper. I was trying to reformulate logic as a learning
problem, an idea I've had for a long time but could not communicate to
anybody. So, I tend to think it has much to do with causal theories
and reductionism. I didn't have a theory of truth on my mind....
>
> >> >Many thanks to Schmidhuber for expressing similar concerns in his
> >> >amazing work. He has been a constant source of inspiration for me.
> >>
> >> http://www.idsia.ch/~juergen/
> >>
> >> ???
> >
> >Yes. Did you study his new work?
>
> Just glanced at the page. It seemed to point in many directions.
>
Yes, and directions worth considering.
> >> >This view is quite possibly consistent with (the digital variant of)
> >> >Multism.
> >>
> >> Doesn't do a thing for me.
> >>
> >> I don't see any necessary link between simulation and cognition. You
> >> might argue for one, but I haven't seen that yet.
> >
> >If you don't then you may wish to explain to me whether a universal
> >computer has any distinction from a particular (not universal!)
> >computer. My argument is supposed to emphasize this distinction.
>
> I like the distinction, but my focus is on the particular.
>
Why is it?
> >I think we can agree that much of cognition is computational, can we?
> >I have another theory for qualia, but I doubt you will like it: it
> >reduces qualia to computation. Consciousness is of course not simply
> >qualia (subjective experience)... It is more than that.
>
> I think it's all reducible to physical instantiations of computation.
>
> Can't get more hard-core than that!
No, you can't I guess :) But, there is more to it: what if there are
"things" that cannot be instantiated (here) at all? My theory of
qualia seems to explicitly allow this, given a universe with bounded
entropy. Note that in a digital universe, the total algorithmic
entropy will *not* increase.
Regards,
--
Eray Ozkural
JXStern
exa) wrote:
>> OK, if you mean "can simulate" in the sense of "is capable of".
>> There's a lot of metaphysics of math and language one can get into in
>> and around there. I just want to find it interesting that math is so
>> powerful it can be either right or wrong, and leave the detailed
>> admiration to others.
>
>Yes, I certainly mean so. In fact, that was something I got the
>inspiration from the fly papers. The flynculus has much dealing in
>reality.
Refers to thread a couple of months ago? Seems to refer to various
connectionist principles, and I assume Brooks' stuff.
>> >Now, my argument is in fact a more abstract mathematical claim, but I
>> >wanted to cast it into something tangible. I think the mind can be
>> >possibly reduced to function approximation and other functional
>> >analysis procedures. What do you think?
>>
>> I put more emphasis on the identity and agency of a computational
>> instantiation, and less on the details of what it does.
>
>Still, to fully specify the work of agency we must identify the
>function of individual components. Otherwise, I suspect we may fall
>prey to subtle conceptual errors, which are themselves hard to
>identify. We might be exercising them without even suspecting there is
>an error...
A couple of comments. First, functionalism is all very fine, but it
is a very slippery concept. It is necessarily a non-physical concept,
an abstraction like type or universal. All of these are
idealizations, and I always look for particular cases and nominalist
reasoning. Second, I want to allow any cognitive agent freedom to
err. In fact, I don't ever want to ask if a concept is in error, as
that too is an idealization.
Have you read the work of Ruth Garrett Millikan? Teleosemantics.
Very, very interesting stuff. Deals a lot with the issue of error. I
find half of RGM's work fantastically insightful - and the other half
fantastically wrong, but the wrongness can be fixed and the insightful
stuff enhanced, I think. I'd start with her more recent book, "On
Clear and Confused Ideas", much more readable than the more complete
"Language, Thought, and Other Biological Categories". Or you can tour
her web site:
http://vm.uconn.edu/~wwwphil/millikan.html
>What do you mean by identity?
To be an agent, a thing first has to have identity. This is implicit
in any kind of computationalism, I think, but mainstream philosophy
today seems to be neo-Fregean logicism and multi-world modal logic,
which is all ontological commitment on the one hand and
that-which-is-true idealization on the other. Identity is the trivial
(?!) property that any physical object has, in order to be a physical
object. It is a very problematic idea, philosophically, but then so
much the worse for such philosophy.
>> >> > This
>> >> >reflective-process is at the root of subjective experience and
>> >> >intelligence.
...
>The above view originated from a causality graph I was sketching on a
>piece of paper. I was trying to reformulate logic as a learning
>problem, an idea I've had for a long time but could not communicate to
>anybody. So, I tend to think it has much to do with causal theories
>and reductionism. I didn't have a theory of truth on my mind....
Let logic be logic, it sounds like what you're working on is something
else, probably what I would call cognition, or agency, or a
combination thereof. A cognitive agent would seem to need to do a lot
of generate(-hypothesis)-and-test, right?
>> >> >Many thanks to Schmidhuber for expressing similar concerns in his
>> >> >amazing work. He has been a constant source of inspiration for me.
>> >>
>> >> http://www.idsia.ch/~juergen/
>> >
>> >Yes. Did you study his new work?
>>
>> Just glanced at the page. It seemed to point in many directions.
>
>Yes, and directions worth considering.
I'll try to look around it some, but feel free to point out one or two
papers you think best.
>> >If you don't then you may wish to explain to me whether a universal
>> >computer has any distinction from a particular (not universal!)
>> >computer. My argument is supposed to emphasize this distinction.
>>
>> I like the distinction, but my focus is on the particular.
>
>Why is it?
I think computationalism can only be discussed in particularist,
nominalist cases. This is a non-traditional and difficult
methodological choice. But look. Searle has long ranted about how
syntax and formal systems cannot be identical to consciousness, and
y'know what, I think he has a point. Like most of his
anti-computational work it is a tendentious point founded on a
purposely perverse over-literal reading of computationalist
assertions, but even so. On the other hand, there seem to be orthodox
computationalists who say silly things like, "computers are not Turing
machines because a Turing machine is an abstract and infinite theory."
Well, they haven't read (or understood) Turing, but again, there is a
sort of a point there, known since Plato at least, that there is a
difference between the real and the ideal. And so, I concentrate on
the sublunar, realizable, particular.
Well, you asked.
:)
>> >I think we can agree that much of cognition is computational, can we?
>> >I have another theory for qualia, but I doubt you will like it: it
>> >reduces qualia to computation. Consciousness is of course not simply
>> >qualia (subjective experience)... It is more than that.
>>
>> I think it's all reducible to physical instantiations of computation.
>>
>> Can't get more hard-core than that!
>
>No, you can't I guess :) But, there is more to it: what if there are
>"things" that cannot be instantiated (here) at all? My theory of
>qualia seems to explicitly allow this, given a universe with bounded
>entropy. Note that in a digital universe, the total algorithmic
>entropy will *not* increase.
If it can't be instantiated, then it's not in my universe.
Qualia are a separate topic from cognition. Do you worry about these
things when dealing with "just" cognition? My intuition at this point
is that even qualia are computable, but I want to leave the issue for
another day. I have no concerns, when sober, about such things as
whatever "algorithmic entropy" might signify, probably the dust and
heat given off by the operation of Turing machines, right, and the
inevitable shaving off the corners of bits until their truth values
cannot reliably be distinguished, thereby deriving the uncertainty
principle! Excuse me while I twist the cap off another.
:)
Joshua Stern
Eray Ozkural exa
> On 30 Mar 2004 04:17:05 -0800, er...@bilkent.edu.tr (Eray Ozkural
> exa) wrote:
> >> OK, if you mean "can simulate" in the sense of "is capable of".
> >> There's a lot of metaphysics of math and language one can get into in
> >> and around there. I just want to find it interesting that math is so
> >> powerful it can be either right or wrong, and leave the detailed
> >> admiration to others.
> >
> >Yes, I certainly mean so. In fact, that was something I got the
> >inspiration from the fly papers. The flynculus has much dealing in
> >reality.
>
> Refers to thread a couple of months ago? Seems to refer to various
> connectionist principles, and I assume Brooks' stuff.
>
It doesn't really talk about connectionist principles, it's just
cutting edge cognitive science which decyphered the neural code in (a
small part of) the fly visual system. Their findings strongly suggest
that the computation and communication (as well as perception) in the
nervous system is near optimal, up to physical and computational
limits. It is so optimal that it is not wild speculation any more to
hypothesize a very accurate simulation of a subset of the universe
inside the brain any more. Their latest work characterizes the
computation in the integrative action of the fly's nervous system from
extraction features of visual input to computing the torque of the
wing. (A subspace of the input is correlated to the torque signal:
features) Their work is theoretically sound, and experimentally
grounded in huge volumes of data.
> >Still, to fully specify the work of agency we must identify the
> >function of individual components. Otherwise, I suspect we may fall
> >prey to subtle conceptual errors, which are themselves hard to
> >identify. We might be exercising them without even suspecting there is
> >an error...
>
> A couple of comments. First, functionalism is all very fine, but it
> is a very slippery concept. It is necessarily a non-physical concept,
> an abstraction like type or universal. All of these are
> idealizations, and I always look for particular cases and nominalist
> reasoning. Second, I want to allow any cognitive agent freedom to
> err. In fact, I don't ever want to ask if a concept is in error, as
> that too is an idealization.
I might offer Godel's realism in response. What if functions are
first-class entities in the world which await our affection just as
much as the more readily observable entities like your chair or
dishwasher? I can't really defend that view without contradicting
myself at some point I guess, but it's still an alternative. Realism
can avoid positing another realm of ideas.
> Have you read the work of Ruth Garrett Millikan? Teleosemantics.
> Very, very interesting stuff. Deals a lot with the issue of error. I
> find half of RGM's work fantastically insightful - and the other half
> fantastically wrong, but the wrongness can be fixed and the insightful
> stuff enhanced, I think. I'd start with her more recent book, "On
> Clear and Confused Ideas", much more readable than the more complete
> "Language, Thought, and Other Biological Categories". Or you can tour
> her web site:
>
> http://vm.uconn.edu/~wwwphil/millikan.html
>
I have heard of teleosemantics, but I have not read the work before.
Thanks for the link, I'm going to have a look (bookmarking).
> >What do you mean by identity?
>
> To be an agent, a thing first has to have identity. This is implicit
> in any kind of computationalism, I think, but mainstream philosophy
> today seems to be neo-Fregean logicism and multi-world modal logic,
> which is all ontological commitment on the one hand and
> that-which-is-true idealization on the other. Identity is the trivial
> (?!) property that any physical object has, in order to be a physical
> object. It is a very problematic idea, philosophically, but then so
> much the worse for such philosophy.
>
Identity is a curse on Multism, it was I think my primary concern to
explain the multitude of identities in the universe and to achieve
that I assigned every individual a universe of his own: hardly
ontological minimalism. Thus, it's a solution for digital philosophy
which successfully negates it, for nobody would like reality to be so
diversified only to explain subjectivity.
With regards to your (?!), I think identity is just as problematic for
any physical object as a mind. Why, what makes two particles
identical? If their quantum properties are identical, then are they
identical? But are not quantum properties "just numbers"? Or should we
look at identity at another level, such as behavioral?
> >The above view originated from a causality graph I was sketching on a
> >piece of paper. I was trying to reformulate logic as a learning
> >problem, an idea I've had for a long time but could not communicate to
> >anybody. So, I tend to think it has much to do with causal theories
> >and reductionism. I didn't have a theory of truth on my mind....
>
> Let logic be logic, it sounds like what you're working on is something
> else, probably what I would call cognition, or agency, or a
> combination thereof. A cognitive agent would seem to need to do a lot
> of generate(-hypothesis)-and-test, right?
My view is that logic is a particular formal axiomatic system which is
one among many. Therefore I tend to think it is an invention, and I
wanted to explore by which mechanism a cognitive agent may have
invented it.
> >> >> >Many thanks to Schmidhuber for expressing similar concerns in his
> >> >> >amazing work. He has been a constant source of inspiration for me.
> >> >>
> >> >> http://www.idsia.ch/~juergen/
> >> >
> >> >Yes. Did you study his new work?
> >>
> >> Just glanced at the page. It seemed to point in many directions.
> >
> >Yes, and directions worth considering.
>
> I'll try to look around it some, but feel free to point out one or two
> papers you think best.
>
The New AI paper:
http://www.idsia.ch/~juergen/newai/newai.html
And the OOPS talk slides:
http://www.idsia.ch/~juergen/oops2002/
He is surely one of the most hardworking people in the field.
> >> >If you don't then you may wish to explain to me whether a universal
> >> >computer has any distinction from a particular (not universal!)
> >> >computer. My argument is supposed to emphasize this distinction.
> >>
> >> I like the distinction, but my focus is on the particular.
> >
> >Why is it?
>
> I think computationalism can only be discussed in particularist,
> nominalist cases. This is a non-traditional and difficult
> methodological choice. But look. Searle has long ranted about how
> syntax and formal systems cannot be identical to consciousness, and
> y'know what, I think he has a point. Like most of his
> anti-computational work it is a tendentious point founded on a
> purposely perverse over-literal reading of computationalist
> assertions, but even so.
I agree that he has a point! We can evaluate some auxiliary criticism
of his as quite useful. AI cannot avoid explaining subjective
experience, it must ultimately explain what qualia is, what subjective
experience is, and what the function of consciousness is!
> On the other hand, there seem to be orthodox
> computationalists who say silly things like, "computers are not Turing
> machines because a Turing machine is an abstract and infinite theory."
> Well, they haven't read (or understood) Turing, but again, there is a
> sort of a point there, known since Plato at least, that there is a
> difference between the real and the ideal. And so, I concentrate on
> the sublunar, realizable, particular.
What do you think of Aaron Sloman's paper about the irrelevance of TM?
Did you read it?
> Well, you asked.
>
> :)
>
It was something!
> >No, you can't I guess :) But, there is more to it: what if there are
> >"things" that cannot be instantiated (here) at all? My theory of
> >qualia seems to explicitly allow this, given a universe with bounded
> >entropy. Note that in a digital universe, the total algorithmic
> >entropy will *not* increase.
>
> If it can't be instantiated, then it's not in my universe.
>
Realist enough!
> Qualia are a separate topic from cognition. Do you worry about these
> things when dealing with "just" cognition?
Yes, I do. Because I think there is something central about qualia,
for it underlines the difference in thought. (Mental differences)
> My intuition at this point
> is that even qualia are computable, but I want to leave the issue for
> another day.
My view is that qualia is a set of instances of halting problems
solvable by a particular computer.
> I have no concerns, when sober, about such things as
> whatever "algorithmic entropy" might signify, probably the dust and
> heat given off by the operation of Turing machines, right, and the
> inevitable shaving off the corners of bits until their truth values
> cannot reliably be distinguished, thereby deriving the uncertainty
> principle! Excuse me while I twist the cap off another.
>
> :)
Hmm.
1. Shannon's information theory can be derived from algorithmic
information theory by modelling random sources.
2. In turn, entropy in thermodynamics is identical to Shannon entropy.
3. Therefore algorithmic entropy signifies all the examples that you
give.
(4. That is why, the cognitive universals are relevant for physics!
Cognition is physical!)
Regards,
--
Eray Ozkural
Eray Ozkural exa
reply. Please excuse me, I've just woken up.
David Longley
Ozkural exa <er...@bilkent.edu.tr> writes
>There are some grammar mistakes in the first paragraph of my previous
>reply. Please excuse me, I've just woken up.
Not half as much as you will when you hit the real world and have to
abandon 99% of the "grammar mistakes" which you've been so eager to
share with us in this newsgroup over the past year.
Why you don't just try to listen and learn is one of the profound
tragedies of life.
--
David Longley
JXStern
exa) wrote:
>It doesn't really talk about connectionist principles, it's just
>cutting edge cognitive science which decyphered the neural code in (a
>small part of) the fly visual system. Their findings strongly suggest
>that the computation and communication (as well as perception) in the
>nervous system is near optimal, up to physical and computational
>limits. It is so optimal that it is not wild speculation any more to
>hypothesize a very accurate simulation of a subset of the universe
>inside the brain any more. Their latest work characterizes the
>computation in the integrative action of the fly's nervous system from
>extraction features of visual input to computing the torque of the
>wing. (A subspace of the input is correlated to the torque signal:
>features) Their work is theoretically sound, and experimentally
>grounded in huge volumes of data.
I'm sure it's all wonderful, but my concern is with cognition and
symbolic computation, which is probably intertranslatable with any
other sort, but if not, whatever.
>> A couple of comments. First, functionalism is all very fine, but it
>> is a very slippery concept. It is necessarily a non-physical concept,
>> an abstraction like type or universal. All of these are
>> idealizations, and I always look for particular cases and nominalist
>> reasoning. Second, I want to allow any cognitive agent freedom to
>> err. In fact, I don't ever want to ask if a concept is in error, as
>> that too is an idealization.
>
>I might offer Godel's realism in response. What if functions are
>first-class entities in the world which await our affection just as
>much as the more readily observable entities like your chair or
>dishwasher?
Then I am completely mistaken, that's what.
> I can't really defend that view without contradicting
>myself at some point I guess, but it's still an alternative. Realism
>can avoid positing another realm of ideas.
That's not my definition of "realism", that's reification gone wild.
Godel was a raving Platonist, and I'm a raving anti-Platonist.
>> http://vm.uconn.edu/~wwwphil/millikan.html
>
>I have heard of teleosemantics, but I have not read the work before.
>Thanks for the link, I'm going to have a look (bookmarking).
Enjoy.
>With regards to your (?!), I think identity is just as problematic for
>any physical object as a mind. Why, what makes two particles
>identical? If their quantum properties are identical, then are they
>identical? But are not quantum properties "just numbers"? Or should we
>look at identity at another level, such as behavioral?
By identical I mean the identity of a single object over time, ship of
theseus (sp?), not the value identity of two things. BTW, Millikan
makes a lot out of re-cognition identities, she goes on at some length
about it, losing me about halfway through.
>> >> >> >Many thanks to Schmidhuber for expressing similar concerns in his
>> >> >> >amazing work. He has been a constant source of inspiration for me.
>> >> >>
>> >> >> http://www.idsia.ch/~juergen/
>The New AI paper:
>
>http://www.idsia.ch/~juergen/newai/newai.html
>
>And the OOPS talk slides:
>http://www.idsia.ch/~juergen/oops2002/
>
>He is surely one of the most hardworking people in the field.
Well, glancing at these two, and considering your flynculus concerns,
you and Schmidhuber seem concerned with efficiency. I'm more
concerned with what there is to be efficient *at*. You assume a
framework first, in order to worry about efficiency. I worry about an
agent that never really understands the framework but acts on just a
few hints and misunderstandings about the world around it.
Methodological solipsism, if you will.
>I agree that he has a point! We can evaluate some auxiliary criticism
>of his as quite useful. AI cannot avoid explaining subjective
>experience, it must ultimately explain what qualia is, what subjective
>experience is, and what the function of consciousness is!
Searle's Chinese Room and Worstar stories do not concern qualia.
Well, not exactly, not in most people's interpretations - not in
Searle's own, I think.
>> On the other hand, there seem to be orthodox
>> computationalists who say silly things like, "computers are not Turing
>> machines because a Turing machine is an abstract and infinite theory."
>> Well, they haven't read (or understood) Turing, but again, there is a
>> sort of a point there, known since Plato at least, that there is a
>> difference between the real and the ideal. And so, I concentrate on
>> the sublunar, realizable, particular.
>
>What do you think of Aaron Sloman's paper about the irrelevance of TM?
>Did you read it?
Probably at some time, usually like his stuff. What's the exact
title, publication, is it online, is it on my bookshelf?
http://www.cs.bham.ac.uk/~axs/misc/talks/turing-irrel-slides.pdf
Partly based on this paper
The irrelevance of Turing machines to AI,
in Matthias Scheutz, Editor,
Computationalism: New Directions, MIT Press, 2002.
(Also at http://www.cs.bham.ac.uk/research/cogaff/),
http://www.cs.bham.ac.uk/research/cogaff/sloman.turing.irrelevant.pdf
Yes, I agree with the kinds of things he is saying here. Except that
I think he is letting the criticism force him to make the same mistake
he says the critics make. Strictly speaking you may not need to know
anything about TM's to discuss AI, as the limits of TM's are
irrelevant to AI. However, knowing about TMs as a methodology and
abstraction is very useful, even if we never approach the limits.
I see another title here, and a quick browse suggests I agree with the
content, too:
"Must Intelligent Systems Be Scruffy?"
http://www.cs.bham.ac.uk/research/cogaff/sloman.scruffy.ai.pdf
He and I both agree the answer is "Yes!"
>> Qualia are a separate topic from cognition. Do you worry about these
>> things when dealing with "just" cognition?
>
>Yes, I do. Because I think there is something central about qualia,
>for it underlines the difference in thought. (Mental differences)
OK, but I don't.
>> My intuition at this point
>> is that even qualia are computable, but I want to leave the issue for
>> another day.
>
>My view is that qualia is a set of instances of halting problems
>solvable by a particular computer.
Well, whatever is computable is a set of instances of halting
problems, if you want to look at it that way. Even simple stuff.
>> I have no concerns, when sober, about such things as
>> whatever "algorithmic entropy" might signify, probably the dust and
>> heat given off by the operation of Turing machines, right, and the
>> inevitable shaving off the corners of bits until their truth values
>> cannot reliably be distinguished, thereby deriving the uncertainty
>> principle! Excuse me while I twist the cap off another.
>>
>> :)
>
>Hmm.
>
>1. Shannon's information theory can be derived from algorithmic
>information theory by modelling random sources.
>2. In turn, entropy in thermodynamics is identical to Shannon entropy.
>3. Therefore algorithmic entropy signifies all the examples that you
>give.
>
>(4. That is why, the cognitive universals are relevant for physics!
>Cognition is physical!)
Cognition is physical, yes, and I suppose it's thermodynamically
describable, but my concerns are elsewhere. There's just a ton of
unpacking to do in even the simplest cases of cognition, I think, and
that's all the scope I can handle.
J.
bob
> abandon 99% of the "grammar mistakes" which you've been so eager to
> share with us in this newsgroup over the past year.
>
> Why you don't just try to listen and learn is one of the profound
> tragedies of life.
do you remember the saying 'consider the source'. that says it all
when it comes to listening to others. you rave at everyone not in yer
little club longomore. when you stop doing that you will have leanred
something yerself. yer BRILLI-ant parrot life.
Eray Ozkural exa
Non-sense. Off-topic. Ad-hominem attack.
You are a clown.
David Longley
<okams...@yahoo.com> writes
People challenge people who don't know - it's called education.
Something you're in dire need of you idiot!
--
David Longley
Glen M. Sizemore
according to Eray.
Eray Ozkural: Glen has no idea what he is talking about, he is a fucking
idiot just
like you!
Eray: I don't care about teaching Glen anything, he is an incompetent
mindless rat, why should I even try? He won't understand a thing! Look
at all the nonsense he speaks of neuroscience and computer science!
The guy is an imbecile! And so are you!
Get out of here you lamers!
Eray Ozkural: You are the USENET fuckwit
#1, up there in the crackpots list together with David Longley and
following you two Arthur T. Murray. And besides, you are the single
most abusive person on USENET, not me.
You are insanely confused. You don't even realize that you deserve
such response! As I said, you deserve insults and torments till the
end of your unnecessary life. You are nothing. Showing that I insulted
you does no harm to me, you are only an ignorant, foolish troll who
does nothing but disrupt the course of discussion on this group. You
have single-handedly cut down the signal-to-noise ratio to a quarter,
you fucking bastard.
Why do you complain when your own style of argumentation is used
against you? Do you think you will get away with insulting everybody
on this newsgroup? Do you think I am going to tolerate you forever?
Fucking moron!
Eray Ozkural: You and your retarded pet Glen are the only ignorant and
deluded
idiots on this group except for the trolls (namely Bloxy's, Moniahnr,
Psicock, whatever).
You are an idiot even if you don't realize that you are one.
You have no authority on the subject, and you cannot assert anything.
You can only try to assert things but you keep on living in a world of
fantasy in which Quine is God and Skinner is a prophet. Because of
your foolish and ignorant ravings, many people on this group,
including me, have started looking at the philosophy of Quine with
disgust. I assume that was *not* your goal. In fact, you don't seem to
have any rationality, so it's not even right to assume you can have
any goals. You are a hallucinating maniac and a sociopath with noscientific
credibility, morals, or personality in general. I despise
you and your lowly kind.
Down with your behaviorist religion. I have had enough!
Fuck off and die, Longley-bot!
Eray Ozkural: I do think you two are fucked up morons and that this planet
would be better off without you.
And
Eray Ozkural: I don't say this as only an insult that you deserve. You
deserve
insults and torment till the end of your unnecessary lives. I say this
with good intentions as a scholar who has a hint about AI and wants to
do you a favor. Don't make a fool of your worthless selves any more.
Get the hell out of here, you hopeless fucked up morons.
"Eray Ozkural exa" <er...@bilkent.edu.tr> wrote in message
news:fa69ae35.04033...@posting.google.com...