Re: consciousness


"Don Geddis" <don@xxxxxxxxxx> wrote in message
news:87r6qg6ryf.fsf@xxxxxxxxxxxxx
"Alpha" <OmegaZero2003@xxxxxxxxx> wrote on Thu, 19 Apr 2007:
"Don Geddis" <don@xxxxxxxxxx> wrote in message
news:877is99v8d.fsf@xxxxxxxxxxxxx
If your PC is currently running Microsoft Word, at the end of the day
everything that is happening is "just" electrical signals in
transistors.
Much like all of consciousness is really "just" neural signals in the
brain.

That "just" is belied by observations like Hofstader's that it is the
patterns at higher levels of abstraction that constitute the cognitive
functions.

I actually agree with that statement, and don't see how it is a
contradiction.

We're all (except perhaps Sizemore) definitely interested in these
processes
described at the high levels of abstraction.

What we're trying to do first is agree whether we have identified an
implementation substrate that is sufficient to account for the high-level
behavior. As a computationalist, I think we have already, and there is
no need to do further neurobiochemistry and explore quantum effects or
microtubules.

But you're absolutely right that merely concluding "Microsoft Word could
in
theory be implemented by voltages and transistors" is but the merest first
step to actually being able to design and implement the thing. We still
have
to invent memory, and microprocessors, and operating systems, and virtual
memory, and GUIs, and mice, and Windows, and proportionally-spaced fonts,
etc. There's lots of work to come.

But what we _don't_ have to do is look to physics for something beyond
logic
gates as a low-level substrate to host the final solution.

Your confusion is believing that the fact that these abstractions are
useful, is somehow evidence that the actual phenomena are somehow
distinct
from the obvious neural implementation.

Hofstader (and I) believes fervently that not only are the abstractions
useful, they are the onl;y thing worth talking about WRT human cognition.
Read his book and you will find that to be the case. He give many
analogies/examples.

But I don't disagree. I wonder why you think I would.

By far the most interesting part of the work to come is in designing
actual
computational algorithms.

All this discussion about substrate is merely because some (most?) people
(including YOU) have denied that consciousness is (or could be)
computational,
that logic programming using neural temporal spikes and axon connections
is
sufficient to support all we know about human cognition.

If you'd just accept that, then we could move on to the more interesting
questions of exactly what are the algorithms that enable this very
interesting cognitive behavior.

I think it is worth a try of course. But since I think intelligence
functions and consciousness are different beasts, therin lies the issue.

I also think there are far more difficult problems to be addressed in the
quest for a C machine. For example, in the case of the brain, just what
properties of the AP or patterns of APs, suffice to generate C!? I.e., when
I look at all the properties of a pulse (or set of pulses), they all have a
time-varying magnitude, are caused by electrochemical interactions, repeat
in certain ways, are composed of electrical activity and so forth. I see no
"redness" when a purported set of APs gives rise to the redness I see in my
mnd's eye, or directly when I see an apple.


But that does NOT mean that Word is somehow a different "kind" of thing
from voltages and transistors.

Sure it does. Per Hofstader these are abstrractions/Word is an abstraction
that has little to do with the types of things that correlate with the
abstraction (e.g., APs or atoms careening about, or in the case of Word,
voltages and transitorrs.

Word *is* a differnt type, or class of thing than a voltage or a transiter.
I wonder why you would think otherwise?

Is love (an emotion) differnt than atoms careening about? Of course. Is
the idea of a dog different than atoms creenning about. Of course. That is
one of the main points of DH's book!



Of course it is; it is a higher level abstraction and thence a differnt
class or type of thing.

And nobody supposes that a novel somehow has causal power on top of its
layout of ink and paper.

Of course it does; it is the meaning of the words that convey causal
power,
NOT the ink properties or the paper properties.

The question is whether you need to look down to subatomic (quantum)
physics
in order to find a sufficient substrate to explain the high-level
behavior.

Well, you specified voltages and transistors as being the same type of thing
as the functional unit that we call WOrd. And I claim there is a type
difference. And OO analyst would never model a voltage the same as he would
model a program like Word. The differences in propeties and methods are so
large that to label them the same type of thing is to msunderstand
classification/typing theory altogether..


With MS Word, and with books, the answer is no. It's immediately clear
that
there is a firm substrate at a low level of abstraction, and the high
level
properties of books and software can be fully implemented by ink/paper and
voltages/transistors respectively.

That isn't the point. The question was whether voltages and trnasistors ar
the same type of thing as Word. I claim they are not.


There is no need to search for other kinds (quantum?) of physics,
especially
for some new physics (ESP, souls) that isn't even part of the standard
model
of subatomic physicists today.

That is not clear, as we do not know how the existing collections of atoms,
molecules, neurons, neronal grooups etc., give rise to emergent patterns
that result in an "I" and an "I" being cognizant of other patterns
(exogenous or endogenous).


It is up to real AI researchers to find out wht those emergent mechanisms
are; to figure out how, perhaps, some assenbly of "just" transistors can,
with the right patterns of interconnections, may produce truely
intelligent
behavior.

Well, gosh, I agree with you.

You're starting to sound like a computationalist...

I am, only not as sure of the philosophical underpinnings as you seem to be.
And I raise the cautionary flag whenever I see people using the "just" or
"merely" trerms to descibe just (hehe) what C or intelligence functions are
based upon physiologically.
I think there is much more to the story and that we will find out that
molecular activities (thus implying QM effects) and other medium-scale
processes/objects such as microtubules play a part in the orchestration of
C-, if not I functions. Indeed, it is my opinion that emergence is
substrate dependent; new operators (that are emergent properties/processes)
spring forth due to the *particular* manner in which low-level
processes/objects interact, all the way down to sub-atomic "particles".
Without that particular arrangement and substrate, who knows if the
emergence would occur! We sure have not seen it in other substrates.


-- Don
_______________________________________________________________________________
Don Geddis http://don.geddis.org/
don@xxxxxxxxxx
All brontosauruses are thin at one end, much MUCH thicker in the middle,
and then thin again at the far end.
-- New dinosaur theory by "Miss Anne Elk" (Monty Python)



--
Posted via a free Usenet account from http://www.teranews.com

.

Relevant Pages

  • Re: consciousness
    ... described at the high levels of abstraction. ... theory be implemented by voltages and transistors" is but the merest first ... But what we _don't_ have to do is look to physics for something beyond ... gates as a low-level substrate to host the final solution. ...
    (comp.ai.philosophy)
  • Re: consciousness
    ... everything that is happening is "just" electrical signals in transistors. ... described at the high levels of abstraction. ... But what we _don't_ have to do is look to physics for something beyond logic ... gates as a low-level substrate to host the final solution. ...
    (comp.ai.philosophy)
  • Re: consciousness
    ... at multiple levels of abstraction. ... electrical signals in transistors. ... Curt's conscious rocks was one of saying a conscious ... It's the same with consciousness. ...
    (comp.ai.philosophy)
  • Re: Dos and donts
    ... Fast (with small transistors), nicely analytical ... voltages to create, revealing its simplicity, NOT. ... The "clint" circuit is nice because it is simple, ... pF across the opamp fixes it. ...
    (sci.electronics.design)
  • Re: Dos and donts
    ... Fast (with small transistors), nicely analytical ... voltages to create, revealing its simplicity, NOT. ... The follower clamp is perfectly stable as long as the opamp is ... The "clint" circuit is nice because it is simple, ...
    (sci.electronics.design)