Chalmers's book
Drew McDermott
I have now read David Chalmers's book, "The Conscious Mind." It's very good, by
which I mean it's very lucid and honest. However, in my opinion it will be
viewed as a reductio ad absurdum of dualism rather than the defense the author
intends. I'm thinking especially of his "paradox of phenomenal judgment," which
is that you can account for all judgments about experience without appeal to
experience. In other words, a zombie could take part in a wine tasting and
report with complete sincerity its appraisals of the tastes of various wines, in
a way that wine experts would find appropriate and perhaps even insightful,
without actually experiencing anything. That's because judgements (or, anyway,
utterances) can be explained entirely in terms of physical events in the brain.
In my opinion, there's no way to rule out the hypothesis that we are zombies of
this sort. In particular, as far as I can tell, I could be a zombie of that
sort. Hence explaining what zombies *think* they experience might very well be
explaining experience itself.
Comments? (And please, do not let the conversation degenerate into endless
bickering about each other's intellectual respectability and methodological
assumptions.)
-- Drew McDermott
Michael Schuerig
Drew McDermott <mcdermo...@cs.yale.edu> wrote:
> I have now read David Chalmers's book, "The Conscious Mind." It's very
> good, by which I mean it's very lucid and honest. However, in my opinion
> it will be viewed as a reductio ad absurdum of dualism rather than the
> defense the author intends. I'm thinking especially of his "paradox of
> phenomenal judgment," which is that you can account for all judgments
> about experience without appeal to experience. In other words, a zombie
> could take part in a wine tasting and report with complete sincerity its
> appraisals of the tastes of various wines, in a way that wine experts
> would find appropriate and perhaps even insightful, without actually
> experiencing anything. That's because judgements (or, anyway, utterances)
> can be explained entirely in terms of physical events in the brain. In my
> opinion, there's no way to rule out the hypothesis that we are zombies of
> this sort. In particular, as far as I can tell, I could be a zombie of
> that sort. Hence explaining what zombies *think* they experience might
> very well be explaining experience itself.
According to Chalmers you could be a zombie and would have no way of
knowing it. Chalmers himself might be a zombie notwithstanding his
claims to the contrary. He acknowledges that his zombie twin is
psychologically (in DC's sense, i.e. excluding the "phenomenal")
identical to him, but might nevertheless lack any phenomenal
experiences. Both of them are indistinguishable. And as far as I can
tell Chalmer's doesn't have an argument that he's the one with the
"feelings".
I think Chalmers's stabs at fundamental proto-phenomal properties are
particularly curious. The rationale for non-physical fundamental
properties is that apparently the physical properties don't suffice to
explain high-level mental phenomena such as consciousnes. Thus there
have to be (proto-)phenomenal properties down to the bottom. Out of
these properties consciousness is constructed in an almost functionalist
fashion -- one really wonders why he's not simply a functionalist.
[Sorry, if this is too terse, unintelligible or wrong, but it's too
late. I'll be back...]
Michael
Have a look at: "Commentary on Chalmers: Facing Backwards on the Problem
of Consiousness" and "The Unimagined Preposterousness of Zombies" by
Daniel Dennett
<http://www.tufts.edu/as/cogstud/pubslist.htm>
---
Michael Schuerig
mailto:uzs...@uni-bonn.de
http://www.rhrz.uni-bonn.de/~uzs90z/
Peter_...@uk.ibm.com
In <57feam...@ADEN.AI.CS.YALE.EDU>, mcdermo...@cs.yale.edu (Drew McDermott) writes:
>
>I have now read David Chalmers's book, "The Conscious Mind." It's very good, by
>which I mean it's very lucid and honest. However, in my opinion it will be
>viewed as a reductio ad absurdum of dualism rather than the defense the author
>intends. I'm thinking especially of his "paradox of phenomenal judgment," which
>is that you can account for all judgments about experience without appeal to
>experience. In other words, a zombie could take part in a wine tasting and
>report with complete sincerity its appraisals of the tastes of various wines, in
>a way that wine experts would find appropriate and perhaps even insightful,
>without actually experiencing anything. That's because judgements (or, anyway,
>utterances) can be explained entirely in terms of physical events in the brain.
>In my opinion, there's no way to rule out the hypothesis that we are zombies of
>this sort. In particular, as far as I can tell, I could be a zombie of that
>sort. Hence explaining what zombies *think* they experience might very well be
>explaining experience itself.
>
>bickering about each other's intellectual respectability and methodological
>assumptions.)
I'm curious, because I too think that consideration of zombies leads me
to think of qualia as essentially functional and hence falling under the
rubrick of functionalism.
What the zombie reminds us is that, if we take qualia to be essentially
non-functional, then our claims to have qualia (whether true or not)
would be true for the wrong reasons. It would be a sort of Gettier case.
That is, we'd claim that we had qualia, it might be true that we had
qualia, but the *reason* we claimed that we had qualia would be
the same as the *reason* the zombie has, who hasn't got qualia.
This seems to me to make the idea of qualia as essentially non-functional
out of step with what I take qualia to be. I take it that when I claim that
I have qualia it is part of the *reason* that I claim that I have qualia that I do,
indeed, have qualia. So, if we want our claims to have qualia to
be justified by our *actually* *having* *qualia*, it appears to me to
make it essential (and not just accidental) to what qualia are that
they be functional.
Cheers,
Pete Lupton
David Longley
In article <57feam...@ADEN.AI.CS.YALE.EDU>
mcdermo...@cs.yale.edu "Drew McDermott" writes:
>
> I have now read David Chalmers's book, "The Conscious Mind." It's very good, by> which I mean it's very lucid and honest. However, in my opinion it will be
> viewed as a reductio ad absurdum of dualism rather than the defense the author
> intends. I'm thinking especially of his "paradox of phenomenal judgment," which> is that you can account for all judgments about experience without appeal to
> experience. In other words, a zombie could take part in a wine tasting and
> report with complete sincerity its appraisals of the tastes of various wines, in> a way that wine experts would find appropriate and perhaps even insightful,
> without actually experiencing anything. That's because judgements (or, anyway,
> utterances) can be explained entirely in terms of physical events in the brain.
> In my opinion, there's no way to rule out the hypothesis that we are zombies of
> this sort. In particular, as far as I can tell, I could be a zombie of that
> sort. Hence explaining what zombies *think* they experience might very well be
> explaining experience itself.
>
> Comments? (And please, do not let the conversation degenerate into endless
> bickering about each other's intellectual respectability and methodological
> assumptions.)
>
> -- Drew McDermott
>
Hmmm.....join the club Drew. I'm apparently a zombie too, but I'm
also a "disturbed" and "Fragmented" zombie...To some folk, once
can't be one without the other <g>.
--
David Longley
Neil Rickert
>I have now read David Chalmers's book, "The Conscious Mind." It's very good, by
>which I mean it's very lucid and honest. However, in my opinion it will be
>viewed as a reductio ad absurdum of dualism rather than the defense the author
>intends. I'm thinking especially of his "paradox of phenomenal judgment," which
>is that you can account for all judgments about experience without appeal to
>experience. In other words, a zombie could take part in a wine tasting and
>report with complete sincerity its appraisals of the tastes of various wines, in
>a way that wine experts would find appropriate and perhaps even insightful,
>without actually experiencing anything. That's because judgements (or, anyway,
>utterances) can be explained entirely in terms of physical events in the brain.
>In my opinion, there's no way to rule out the hypothesis that we are zombies of
>this sort. In particular, as far as I can tell, I could be a zombie of that
>sort. Hence explaining what zombies *think* they experience might very well be
>explaining experience itself.
>Comments? (And please, do not let the conversation degenerate into endless
>bickering about each other's intellectual respectability and methodological
>assumptions.)
Let me see if I have Chalmers' position straight.
In our ordinary activity, we pick up a rich stream of information
with our eyes, our ears, and our other senses. We act upon this
information, as is evidenced in our behavior.
We are supposed to believe that this flow of information is
possible while at the same time there is no experience whatsoever.
In the meantime we also have experience, but apparently not as a
result of the input we are receiving. Nevertheless, the experience
seems to be highly correlated with the information.
Somehow, it seems to me that this reveals a highly confused notion of
`experience.' It seems to me that we ought to simply say that the
experience just is the picking up of information.
My conclusion is that we ought not trust arguments based on what is
said to be logically possible.
Ian P Wright
I entirely agree with both your comments on David
Chalmers' book. First, it is a very good book
precisely because it is so clear and intellectually
honest. It also has the merit of demonstrating
that a dualist position has no practical consequences
whatever.
The error, I believe, consists in thinking that
`consciousness' refers to an essence or property
rather than a collection of functional capabilities.
It is a `reification' error, logically similar to
the vitalist error of thinking that `life' refers
to an essence rather than a collection of
functional capabilities (such as reproducing,
homeostatis, and so forth).
Chalmers begins with a reified notion of `consciousness'
then discovers that this essence is always missing
from functional explanations. Therefore, he posits
`bridging laws' that link functional processes with
`phenomenological' processes. Such bridging laws
are only necessary if you begin with a reified
conception ofconsciousness as a non-functional,
`ineffable' essence. His bridging laws shore up
his conceptual scheme, rather than referring to
real mechanisms in the world.
When I first read Chalmers' book I was convinced by
it and thought it superior to other approaches.
Thinking about it further made me aware of these
philosophical errors. Therefore, although I know
think Chalmers is wrong, I am grateful to the book
for making me think. It is wrong in very interesting
ways.
-Ian.
-----------------------------------------------------------------
Ian Wright email: i...@cs.bham.ac.uk
School of Computer Science www: http://www.cs.bham.ac.uk/~ipw
University of Birmingham Tel: +44 (0) 121 4143734
Edgbaston Fax: +44 (0) 121 4144281
Birmingham B12 2TT, UK
Ian P Wright
I entirely agree with both your comments on David
Chalmers' book. First, it is a very good book
precisely because it is so clear and intellectually
honest. It also has the merit of demonstrating
that a dualist position has no practical consequences
whatever.
The error, I believe, consists in thinking that
`consciousness' refers to an essence or property
rather than a collection of functional capabilities.
It is a `reification' error, logically similar to
the vitalist error of thinking that `life' refers
to an essence rather than a collection of
functional capabilities (such as reproducing,
homeostatis, and so forth).
Chalmers begins with a reified notion of `consciousness'
then discovers that this essence is always missing
from functional explanations. Therefore, he posits
`bridging laws' that link functional processes with
`phenomenological' processes. Such bridging laws
are only necessary if you begin with a reified
conception of consciousness as a non-functional,
David Longley
> I entirely agree with both your comments on David
> Chalmers' book. First, it is a very good book
> precisely because it is so clear and intellectually
> honest. It also has the merit of demonstrating
> that a dualist position has no practical consequences
> whatever.
>
> The error, I believe, consists in thinking that
> `consciousness' refers to an essence or property
> rather than a collection of functional capabilities.
> It is a `reification' error, logically similar to
> the vitalist error of thinking that `life' refers
> to an essence rather than a collection of
> functional capabilities (such as reproducing,
> homeostatis, and so forth).
>
> Chalmers begins with a reified notion of `consciousness'
> then discovers that this essence is always missing
> from functional explanations. Therefore, he posits
> `bridging laws' that link functional processes with
> `phenomenological' processes. Such bridging laws
> are only necessary if you begin with a reified
> conception ofconsciousness as a non-functional,
> `ineffable' essence. His bridging laws shore up
> his conceptual scheme, rather than referring to
> real mechanisms in the world.
>
> When I first read Chalmers' book I was convinced by
> it and thought it superior to other approaches.
> Thinking about it further made me aware of these
> philosophical errors. Therefore, although I know
> think Chalmers is wrong, I am grateful to the book
> for making me think. It is wrong in very interesting
> ways.
>
> -Ian.
>
> -----------------------------------------------------------------
> Ian Wright email: i...@cs.bham.ac.uk
> School of Computer Science www: http://www.cs.bham.ac.uk/~ipw
> University of Birmingham Tel: +44 (0) 121 4143734
> Edgbaston Fax: +44 (0) 121 4144281
> Birmingham B12 2TT, UK
>
Yet the "reification error" you speak of has been forcefully made
at least since 1948 when Ryle wrote "The Concept of Mind" and it
is a clear signature of the positivist movement generally.
What is sad is that "Cognitive Science" has just led (wasted?) a
generation of muddled scientists from many areas to think there
*is* something to be gained by dabbling with "the mind". If
there's one thing characteristic of scientists (forced upon them
by the PH.D. if nothing else), it's that they are highly focal
and specialised. As a consequence they generally find the
multiple facets and skills/levels approach just *inconceivable*.
To even begin to think that way just *feels* thought disordered.
The real work is no longer in understanding this fact (which I
accept *is* a difficulty notion to hold onto) but to work as a
behaviour scientist within what must inevitably be accepted to be
the extensional stance.
http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
--
David Longley
Lewis Jorgenson
Da...@longley.demon.co.uk wrote:
>In article <57feam...@ADEN.AI.CS.YALE.EDU>
>mcdermo...@cs.yale.edu "Drew McDermott" writes:
>> In particular, as far as I can tell, I could be a zombie of
>> that sort. Hence explaining what zombies *think* they
>> experience might very well be explaining experience itself.
>> Comments? (And please, do not let the conversation degenerate
>> into endless bickering about each other's intellectual
>> respectability and methodological assumptions.)
I doubt that this one can approach this problem without making
certain methodological assumptions. Certainly, we can avoid
questions of motive and honesty on the part of participants.
Is it your assertion that a person is not happy and that they
just think they are happy?
Would your position require that the observation:
"I am not Napoleon Bonaparte-I just think I am Napoleon"
to be equivalent to the observation:
"I am not Lewis Jorgenson-I just think I am Lewis Jorgenson"?
The relationship of thinking to experience may be equivalence.
This does not require that sensation be a thought. The ability
to discern between knowing and being is not just an arbitrary
logical discernment. There is a world of difference between
having a toothache and knowing the meaning of the word toothache.
Those differences depict a two-way street where knowing impacts
being and being impacts knowing. It is on that street where
logic and justice and faith and beauty travel.
My understanding of zombies is that they do not have a mind of
their own and so I am confused by your approach. The evidence
suggests that the structure and function and contents of each
brain is an individual matter. That recognition has led us
to speak of individuals changing their minds, choosing of their
own free will, making up their own minds and other references
to the personal powers of cognition held by individuals. What
is the point of denying the person powers of cognition held
by individuals?
Air pressure can be viewed as an emergent property of the
brownian motion of air molecules. So too can mind be viewed
as an emergent property of the interaction of human body
components. You are certainly free to examine the activity
of one atom of air in a ballon and create formulas of how that
particular atom directly or indirectly acts to keep the ballon
expanded. Reductionism is an appropriate means of research
and yet it still only a reduction of the actual state of affairs.
Lewis Vance Jorgenson
andrew babian
Chalmers could just be wrong about the logical possiblity of zombies.
Consciousness could be essentially physical and functional. But
one tricky thing about consciousness or qualia is that it is not
communicable by language. Language recreates memories of experiences,
but these are different from that actual experience. Perhaps this
lack of ability to directly create and control qualia (surely a
consequence of brain structure and function) could lead to the idea that
consciousness is some separate thing, with a special status.
Anders N Weinstein
In article <57l2rn$2...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
> [Paraphrasing Chalmers]
> We are supposed to believe that this flow of information is
> possible while at the same time there is no experience whatsoever.
> In the meantime we also have experience, but apparently not as a
> result of the input we are receiving. Nevertheless, the experience
> seems to be highly correlated with the information.
>
>Somehow, it seems to me that this reveals a highly confused notion of
>`experience.' It seems to me that we ought to simply say that the
>experience just is the picking up of information.
Hear, hear.
On Chalmer's behalf, however, one might distinguish outer experience --
"thick" experience of (or as if of) an intersubjective world of
objects, such as tables, chairs and their visible surfaces arrayed in a
common space which anyone might view -- from "inner" experience,
cognizance of one's own mental states, including sensory states
(feelings), such as a splitting headache, and cognitive states, such as
thoughts, beliefs, intentions (including, btw, those "first-order"
perceptual states that constitute thick experience of the
intersubjective world).
Chalmers is mainly interested in inner experience, I think. Like many
philosophers Chalmers construes inner experience as "thin" experience
i.e. as a preconceptual given. Then he winds up with the problem that
the experiential state -- the having of a splitting headache -- comes
apart from the judgment or assertible content expressed by "Ouch, I
have a splitting headache", which employs terms that were acquired from
and have a role in the public practices of operating with that language.
I gather he winds up in the absurd position of suggesting that it could
perfectly well *seem* to you that you have a splitting headache -- if
you find yourself caused by your brain to incline to making that
judgment, say -- but that you not have a headache at all. [Perhaps you
need the home headache test advertised in a commercial parody on
Saturday Night Live ("it's turning blue -- good, so I don't have a
headache").] Ditto for moaning, holding your head, going to buy aspirin,
seeing a doctor about it, being unable to concentrate on your work,
and any other manifestation of pain one might imagine.
It's pretty clear to me that on this conception of a headache, as
something that dangles independently of any judgment or other expression
one might issue concerning it, it is a wheel that moves though nothing
else turns with it. We can and should Quine it (dispense with it in our
ontology) posthaste, and deny that there is any intelligible possibility
of zombies such as Chalmer's imagines.
I think the main mistake is the idea that such experiential states are
ontologically independent of the possibility of their finding
expression in language, behavior, and otherwise, in the natural forms
of expression available equally to non-linguistic creatures.
Wittgenstein taught us that pain is identified as what it is not by its
functional role inside the head or even the body, but by its role and
connections to other forms of behavior in human life (which is a
context "larger" than anything most functionalists consider). We cannot
so much as identify a verbal expression as a pain *judgment* apart from
these connections, so the idea that the judgments could conceivably
come apart from the states is confused.
BTW There is a very interesting paper by John McDowell called "One
Strand in the Private Language Argument" (published in Grazer
Philosophiche Studien of a few years ago), which makes the case that
Wittgenstein was extending the critique of the non-conceptual given
even to "inner" experience, such as experience of toothaches and
tickles. For linguistic creatures, he suggests, these too are
conceptually informed states.
I also think it's a mistake to think there could be a physicalist
account of normatively correct judgment as non-rationally induced
(caused) by brain states, but that's another story.
Hermital
Hello, Andrew:
You are correct. Most references to consciousness refer only to
phenomenological consciousness while totally ignoring ontological
consciousness, the underlying ground of being itself. Nevertheless,
egoless pure consciousness, unconditioned pure energy, uncreated
absolute pure being pre-exists: All else is supervenient.
--
Alan
Causes invariably precede effects absolutely in all frames of reference
in time as well as space.
Neil Rickert
In <57v9oa$4...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>In article <57l2rn$2...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>> [Paraphrasing Chalmers]
>> We are supposed to believe that this flow of information is
>> possible while at the same time there is no experience whatsoever.
>> In the meantime we also have experience, but apparently not as a
>> result of the input we are receiving. Nevertheless, the experience
>> seems to be highly correlated with the information.
>>Somehow, it seems to me that this reveals a highly confused notion of
>>`experience.' It seems to me that we ought to simply say that the
>>experience just is the picking up of information.
>Hear, hear.
>On Chalmer's behalf, however, one might distinguish outer experience --
>"thick" experience of (or as if of) an intersubjective world of
>objects, such as tables, chairs and their visible surfaces arrayed in a
>common space which anyone might view -- from "inner" experience,
>cognizance of one's own mental states, including sensory states
>(feelings), such as a splitting headache, and cognitive states, such as
>thoughts, beliefs, intentions (including, btw, those "first-order"
>perceptual states that constitute thick experience of the
>intersubjective world).
Outer experience, as I describe it, is the the picking up of external
information. Experience such as pains, itches, and headaches, is the
picking up of proprioceptive information (information of internal
physical states). Introspective experience, I argue, is the picking
up of information which was emitted by output neurons -- I like to
describe this as introspective information. Roughly speaking, it has
to do with what Gerald Edelman refers to as reentrant neurons. So I
claim that all experience is the picking up of information. What
varies between the different types of experience, is the source of
that information. Thoughts can involve all kinds of information, but
particularly involve proprioceptive and introspective information.
Schizophrenia, I suspect, has something to do with a loss of ability
to adequately distinguish between external and internal information.
>Chalmers is mainly interested in inner experience, I think.
As I have described experience, anything that picks up information
should have some kind of 'experience'. Thus we might say that the
computer on my desk is picking up information as I tap the keyboard,
so in some sense is having 'experience'. But presumably we would not
credit the computer with having any awareness of this 'experience'.
One might say that it is the existence of inner 'experience' which
allows one to have the experience of having 'experience', and thus of
being aware of the 'experience' (and there we have Edelman's
reentrant neurons, once again). So Chalmers is right in being
particularly interested in inner experience. His mistake, I think,
is in failing to consider inner experience in the context of other
types of experience (and 'experience').
Anders N Weinstein
In article <581to7$c...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>In <57v9oa$4...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>
>>On Chalmer's behalf, however, one might distinguish outer experience --
>>cognizance of one's own mental states, including sensory states
>>(feelings), such as a splitting headache, and cognitive states, such as
>>thoughts, beliefs, intentions
>
>information. Experience such as pains, itches, and headaches, is the
>picking up of proprioceptive information (information of internal
>physical states).
Some functionalists also take this line but it seems to me a tough
row to hoe. Prima facie a headache does not represent anything about
your neurons or anything else, does not have conditions of satisfaction
or informational content, even if it might be correlated with all sorts
of things in your body.
There are some people who do not feel pain, for example, and only learn
that their hand is on a hot burner by smelling the burning flesh.
Perhaps they could learn to detect these states via inner information
processing, but it would not thereby constitute pain. You have to
account for the difference.
Also, the issue concerned the relation between judgeable or conceptual
content expressed with "I have a headache" and the headache. On the
Wittgensteinian view it makes no sense to suppose such a judgment might
be mistaken -- the possibility of an error here is just not provided for
in the language game of 1st person expressions of subjective states,
his folowers might say. On the rival inner perception theory, however,
the judgment can come apart from the state and be mistaken.
>Introspective experience, I argue, is the picking
>up of information which was emitted by output neurons -- I like to
But the intentional content of a judgment like "I have a pain
in my arm" does not match the informational content you are attributing.
It might be well correlated with damage to the neurons, but then again, it
might not.
>As I have described experience, anything that picks up information
>should have some kind of 'experience'. Thus we might say that the
>computer on my desk is picking up information as I tap the keyboard,
>so in some sense is having 'experience'.
I would not want to say the computer has any experience.
A lot depends on how the norms constituting the information as contentful
are determined. For artifacts like a computer system, the norms
derive only from our purposes and the information is really nothing at all to
the system itself. This is why we don't consider computers as genuine
subjects of experience, even if (our) information is floating around in them.
With autonomous living organisms (whether natural or artificial) things
are different. The context of life provides the setting within which
we can talk about experience and natural information in a non-metaphorical
sense.
>credit the computer with having any awareness of this 'experience'.
>One might say that it is the existence of inner 'experience' which
>allows one to have the experience of having 'experience', and thus of
>being aware of the 'experience' (and there we have Edelman's
>reentrant neurons, once again). So Chalmers is right in being
>particularly interested in inner experience. His mistake, I think,
>is in failing to consider inner experience in the context of other
>types of experience (and 'experience').
The 'cognitive' theory of awareness is widespread enough in the
functionalist literature. The idea in broad terms seems to be this: if
we can provide an adequate account of first-order representational (or
informational or cognitive) content, then we can try to define
consciousness or awareness as the ability to have second order,
reflective representational states which represent among other things
those same old first-order states (or conditions like bodily damage).
I think there are problems with it, however. First, it seems possible
to have first-order consciousness without reflective consciousness,
e.g. someone running for a bus whose whole consciousness is directed
outward, at the bus needing to be caught (an example from Sartre). This
is clearly a state of consciousness, whether or not there is the
possibility of reflective awareness of it. It thus differs from, e.g.
the representation in my nervous system of the angles of my tendons or
whatever, which is no part of my conscious mental life. We might say
the latter is merely information that is in me but not in information for
me, and we need some way of making this distinction.
Second was the claim above, the sensory states lack representational
content altogether. And thirdly is the idea that this theory allows the
possibility of a mistake or misrepresentation of one's own mental
state where none really makes sense.
Jim Balter
In article <57feam...@aden.ai.cs.yale.edu>,
Drew McDermott <mcdermo...@cs.yale.edu> wrote:
>
>I have now read David Chalmers's book, "The Conscious Mind." It's very good, by
>which I mean it's very lucid and honest. However, in my opinion it will be
>viewed as a reductio ad absurdum of dualism rather than the defense the author
>intends. I'm thinking especially of his "paradox of phenomenal judgment," which
>is that you can account for all judgments about experience without appeal to
>experience. In other words, a zombie could take part in a wine tasting and
>report with complete sincerity its appraisals of the tastes of various wines, in
>a way that wine experts would find appropriate and perhaps even insightful,
>without actually experiencing anything. That's because judgements (or, anyway,
>utterances) can be explained entirely in terms of physical events in the brain.
>In my opinion, there's no way to rule out the hypothesis that we are zombies of
>sort. Hence explaining what zombies *think* they experience might very well be
>explaining experience itself.
>
>Comments? (And please, do not let the conversation degenerate into endless
>bickering about each other's intellectual respectability and methodological
>assumptions.)
Chalmers is incorrect that the logical possibility of zombies is inconsistent
with physicalism. What is inconsistent with physicalism is that zombies are
logically possible *and* we are not zombies. Daniel Dennett made the point
that we can be a certain sort of complex zombie, "a zimboe", in _Consciousness
Explained_ and again in a JCS article. Since the only difference between the
zombie world and our world are Chalmers' psychophysical laws that have no
consequences (ex hypothesi), but which serve only to satisfy Chalmers'
requirement that "something else needs to be explained", Ockham serves, as
always, to reject dualism for the more parsimonious physicalism. (Chalmers'
notion of laws as prescriptions that can be tacked onto worlds willy nilly,
rather than as descriptions of regularities, is a common dualistic
confusion, but that's a whole book in itself.)
Chalmers deems all the problems of zombies "easy problems" (by which he
doesn't mean that they are easy in the ordinary sense, only that they aren't
what he calls "the hard problem"). But if we were to solve all the problems
of zimboes, then we would have solved the problems of systems that parallel in
every detail ourselves. Once we have explained systems that have qualiaz and
are awarez and are consciousz, it will be perverse(z) to insist(z) that there
is something else to explain(z). Of course, perversity is always an option.
The philosopher Robert Kirk invented the notion of "philosophical zombie"
years ago, and viewed it as a serious challenge to physicalism. His
investigations for the last 25 years have led him to the opposite conclusion,
and his take on the subject is worth serious consideration. His short but
sharp review of _The Conscious Mind_ appears in the latest JCS (Vol 3,
no. 5/6, 1996, pp. 522-523).
Chalmers' book has considerable value in a couple of ways. One is that it
lays out some of the conceptual and logical challenges that must be met by a
physicalist explanation, such as Frank Jackson's color-deprived Mary. There
is definitely some epistemological clarity needed here. That's of no solace
to dualists though, since rejection of physicalism does not help solve the
problem (in fact, this is a general problem with Chalmers' thesis: he shows
many problems with various sorts of explanation, but his own sort of
explanation doesn't escape any of these problems.) OTOH, Chalmers presents
some strong arguments for "organizational invariance", which, when freed of
his dualistic baggage, simply is functionalism. His arguments in favor of
Strong AI and his demolition of Searle's Chinese Room, his responses to Ned
Block's Chinese Nation and Humongous Lookup Table, etc. distinguish him from
your run of the mill dualist.
Since you brought an Chalmers's alleged honesty, I think this deserves
comment. Chalmers employs a certain subtle sort of ad hominem when he says,
in the book and in some of his responses to critics, that he is only appealing
to those who think that there is "something further to explain". The problem
here is that, for many, the problem is not that there is nothing further to
explain, but rather what that something is and what sort of explanation is
required. Those who think that we may be zimboes may think that explaining
how notions of "self" come about in a society of zimboes is a very difficult
thing to explain, without discarding physicalism. But for Chalmers, if you
don't agree that "trying to define conscious experience in terms of more
primitive notions is fruitless", as he says as early as page 4, thereby
committing himself to his anti-physicalist conclusion, then you don't "take
consciousness seriously". On page xii we find that this is basically a
commitment to not viewing conscious experience as "an illusion". But those
who think that conscious experience may be a very deep, subtle, "illusion",
where "illusion" is just a metaphorical term because our normal notion of
illusion already entails notions of experience and perception that have yet to
be unpacked, take consciousness quite seriously, without making the sorts of
ad ignorantiam commitments Chalmers does, and for him to dismiss these people
as not "taking consciousness seriously" is *not* honest.
This view, that only certain committed positions are respectable, shows up
elsewhere, such as when Chalmers, in Tucson, pointed his "consciousness
detector" hairdryer at Patricia Churchland and detected only "low levels" of
consciousness, and pointed it at Dan Dennett and found none at all. Yes, this made
some people laugh, but this is ad hominem propaganda of a very disturbing sort.
The same tendency can be seen in Chalmers' responses in
sci.psychology.consciousness to Gregory Mulhauser's detailed review that was
posted to sci.psychology.journals.psyche . Chalmers makes unsupported and
unsupportable claims about Mulhauser's motives in writing the review, such as
that he was "out to grind axes" and "in his haste to show that everything in
the book is wrong, old, or trivial (preferably all three at once)". Of
course, many writers who are passionate about their theses go irrational when
faced with criticism, but it is most distressing when we are so eager to see
what sorts of objections might be raised and how they might be handled.
--
<J Q B>
Neil Rickert
In <582am4$j...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>In article <581to7$c...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>>Outer experience, as I describe it, is the the picking up of external
>>information. Experience such as pains, itches, and headaches, is the
>>picking up of proprioceptive information (information of internal
>>physical states).
>Some functionalists also take this line but it seems to me a tough
>row to hoe. Prima facie a headache does not represent anything about
>your neurons or anything else, does not have conditions of satisfaction
>or informational content, even if it might be correlated with all sorts
>of things in your body.
I don't see why this is applicable. Functionalists usually take the
position that experience has to do with how the information is
processed, and the picking up of information is taken as trivial. I
can see how your objection might apply to such a functionalism, but I
don't see its applicability to what I wrote.
>There are some people who do not feel pain, for example, and only learn
>that their hand is on a hot burner by smelling the burning flesh.
If they do not feel pain, it may be that they are not picking up the
appropriate information.
>Perhaps they could learn to detect these states via inner information
>processing, but it would not thereby constitute pain. You have to
>account for the difference.
The information that there is burning flesh in the vicinity is not
the same as the information that damage is being done to my flesh.
The difference seems obvious enough.
>Also, the issue concerned the relation between judgeable or conceptual
>content expressed with "I have a headache" and the headache. On the
>Wittgensteinian view it makes no sense to suppose such a judgment might
>be mistaken -- the possibility of an error here is just not provided for
>in the language game of 1st person expressions of subjective states,
>his folowers might say. On the rival inner perception theory, however,
>the judgment can come apart from the state and be mistaken.
I don't see that as a problem. I might be mistaken in saying "I am
seeing a blue sky", for I might instead be watching television, or
hallucinating. However it is not at all clear that I could be
mistaken in saying "I am having a blue sky experience", for the
hallucination could still count as such an experience. I suggested
that the experience is the picking up of the information. There is a
separate question, as to whether the information picked up was
actually misinformation. It seems to me that the question of whether
I could be mistaken has to do with whether what I picked up was
misinformation. An assertion such as "I am seeing a blue sky" is,
among other things, a claim about the source of the information
(i.e. about it not being misinformation). However "I have a
headache" is only a claim about picking up information, rather than
about the source of the information.
>>Introspective experience, I argue, is the picking
>>up of information which was emitted by output neurons -- I like to
>But the intentional content of a judgment like "I have a pain
>in my arm" does not match the informational content you are attributing.
I don't see the problem. I realize that you always want to argue
about informational content before you will consider it information,
whereas I want to look at the information before I try to decide what
is the informational content. It does often seem to me that your
perspective on information is solipsistic, and perhaps that is why
you see these problems.
>>As I have described experience, anything that picks up information
>>should have some kind of 'experience'. Thus we might say that the
>>computer on my desk is picking up information as I tap the keyboard,
>>so in some sense is having 'experience'.
>I would not want to say the computer has any experience.
I did use scare quote ("'experience'" rather than "experience") to
make a distinction.
>A lot depends on how the norms constituting the information as contentful
>are determined.
That dependency is built into your view of information, but is not a
concern for me. It still seems solipsistic.
We start life as solipsists. We create a solipsistic world. Based
on this solipsistic world, we establish norms for the content of
information. Only then can we decide what is information. And
only then can we open our eyes to look at the world. But if we do
this, we are sentenced to forever see only the solipsistic world,
for that is all our information could be about.
In my opinion, our idea of information cannot be based on norms.
Information is whatever we can get out of the informing signal, and
we must never allow some so-called norms to limit us in the
information we can extract. We are not solipsists, because we accept
information without preconceived ideas (norms, for example) as to
what is the content of that information. Therefore we are freed from
any solipsistic world and are able to use the information to discover
the kind of world that we are actually in.
> For artifacts like a computer system, the norms
>derive only from our purposes and the information is really nothing at all to
>the system itself.
Sure. But that is because we design how the computers will pick up
information. They have no ability to pick up information, except on
the terms that we defined their ability in their input devices.
> This is why we don't consider computers as genuine
>subjects of experience, even if (our) information is floating around in them.
What it boils down to, is that the computers are only picking up
information as our agents, and on our behalf, and are only picking up
information which fits our prescriptions. The problem with the
computer, is that it relies on a concept of informational content
which is based on norms (our norms), and these norms limit its
ability to pick up information.
>With autonomous living organisms (whether natural or artificial) things
>are different. The context of life provides the setting within which
>we can talk about experience and natural information in a non-metaphorical
>sense.
Is this a plea for vitalism? I wouldn't expect an amoeba to have
much experience either, although it might have 'experience.' Surely,
the difference is that cognizing organisms are not limited in their
information pickup by norms imposed upon them.
>>One might say that it is the existence of inner 'experience' which
>>allows one to have the experience of having 'experience', and thus of
>>being aware of the 'experience' (and there we have Edelman's
>>reentrant neurons, once again). So Chalmers is right in being
>>particularly interested in inner experience. His mistake, I think,
>>is in failing to consider inner experience in the context of other
>>types of experience (and 'experience').
>The 'cognitive' theory of awareness is widespread enough in the
>functionalist literature. The idea in broad terms seems to be this: if
>we can provide an adequate account of first-order representational (or
>informational or cognitive) content, then we can try to define
>consciousness or awareness as the ability to have second order,
>reflective representational states which represent among other things
>those same old first-order states (or conditions like bodily damage).
That is your 'cognitive' theory of awareness. It is a solipsist's
theory of awareness. According to that theory, all you could ever be
aware of, is a world which fits the norms you use to establish the
informational content. In my view, you must abandon the attempt to
provide "an adequate account of first-order representational (or
informational or cognitive) content." For any such account of
informational content could only serve as a way of setting limits on
what you can be aware of. A more complete conception of awareness
requires that there should be no such limits.
>I think there are problems with it, however. ...
> And thirdly is the idea that this theory allows the
>possibility of a mistake or misrepresentation of one's own mental
>state where none really makes sense.
The possibility of mistake only exists when information is evaluated
according to norms. We must avoid a norm-based theory. That does
not imply that we cannot set tentative norms for some of our
information, and have a possibility of mistakes for these types of
information. But we must not allow ourselves to be limited by
norms.
Anders N Weinstein
In article <584l95$f...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>In <582am4$j...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>>In article <581to7$c...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>
>>>Outer experience, as I describe it, is the the picking up of external
>>>information. Experience such as pains, itches, and headaches, is the
>>>picking up of proprioceptive information (information of internal
>>>physical states).
>
>>Some functionalists also take this line but it seems to me a tough
>>row to hoe. Prima facie a headache does not represent anything about
>>your neurons or anything else, does not have conditions of satisfaction
>>or informational content, even if it might be correlated with all sorts
>>of things in your body.
>
>I don't see why this is applicable. Functionalists usually take the
>position that experience has to do with how the information is
>processed, and the picking up of information is taken as trivial. I
>can see how your objection might apply to such a functionalism, but I
>don't see its applicability to what I wrote.
The question is (again) what the information picked up is when I have
a headache. The usual assumption behind the objection is that whatever
this information is, it is something that need not feel any particular
way, for example, it need not hurt.
As always, it might help to to distinguish the sensory state of having a
splitting headache itself from the intentional state of thinking or
judging or asserting or believing that I have a splitting headache.
The latter might plausibly be said to bear information about the
former. Still don't know what the former bears information about, or
how it can be explained entirely as information detection.
>>There are some people who do not feel pain, for example, and only learn
>>that their hand is on a hot burner by smelling the burning flesh.
>
>If they do not feel pain, it may be that they are not picking up the
>appropriate information.
>
>>Perhaps they could learn to detect these states via inner information
>>processing, but it would not thereby constitute pain. You have to
>>account for the difference.
>
>The information that there is burning flesh in the vicinity is not
>the same as the information that damage is being done to my flesh.
>The difference seems obvious enough.
Sure, but even after such a man recognizes that, in his words, "damage
is being done to *my* flesh", he is still not feeling pain. He might be
completely dispassionate about it if for some reason he has lost
interest in preserving his flesh, for example.
>I don't see that as a problem. I might be mistaken in saying "I am
>seeing a blue sky", for I might instead be watching television, or
>hallucinating. However it is not at all clear that I could be
>mistaken in saying "I am having a blue sky experience", for the
>hallucination could still count as such an experience. I suggested
Right, but if this requires that information be present inside your
brain, one wonders how you could be so infallible at it. Oughtn't you
to ask a neuroscientist whether such information is genuinely in there?
>that the experience is the picking up of the information. There is a
>separate question, as to whether the information picked up was
>actually misinformation. It seems to me that the question of whether
>I could be mistaken has to do with whether what I picked up was
>misinformation. An assertion such as "I am seeing a blue sky" is,
>among other things, a claim about the source of the information
>(i.e. about it not being misinformation). However "I have a
>headache" is only a claim about picking up information, rather than
>about the source of the information.
Let us say that first-order information pickup concerns the outer
world. Then what is picked up may be information or misinformation.
Now on the theory in question, reflective or second-order information
pickup would be the same sort of thing, detection of information, only
concerned with information about the state of the first-order
detectors.
Now I don't think this theory can account for first-person authority.
There is no reason at all why the second-order or reflective detectors
might not also receive misinformation. But what would that be like?
In such a case, perhaps, you might be expected to report,
non-commitally, that you are having an experience as if of something
blue and *still be wrong* in the claim. You could be wrong if your
reflective, or self-scanning information detectors received
*misinformation* about the state of your first-order information
detectors. Perhaps your blue-detector is not in fact activated, but
your second-order blue-detector-detector is getting the (mis)
information that it is.
In such a case I guess we would be forced to say that it seems to you
as if it seems to you that there is something blue before you, but this
is mistaken, for it does not in fact seem to you that there is
something blue before you.
But I think this is nonsense, it makes no sense to speak of a mistake
here. The "seems" operator doesn't significantly iterate, it's idempotent
as it were.
There are cases of such things as blindness denial, but I think these
are rather cases in which there is no way to coherently describe what
is going on using these person-level epistemic concepts, and we do better
to abandon them in favor of the sub-personal level of description.
>>>Introspective experience, I argue, is the picking
>>>up of information which was emitted by output neurons -- I like to
>
>>But the intentional content of a judgment like "I have a pain
>>in my arm" does not match the informational content you are attributing.
>
>I don't see the problem. I realize that you always want to argue
>about informational content before you will consider it information,
>whereas I want to look at the information before I try to decide what
>is the informational content. It does often seem to me that your
Well a brief sketch of what the information content is would be
helpful.
>perspective on information is solipsistic, and perhaps that is why
>you see these problems.
I bet I am the most anti-solipsistic thinker you will ever encounter.
I think conceptual thinkers are essentially members of communities, for
example. So the idea that intentional content could be determined
solipsistically, in independence of the environment and the culture
is another thing that makes no sense to me.
>That dependency is built into your view of information, but is not a
>concern for me. It still seems solipsistic.
Again I wonder, why 'solipsistic'?
> We start life as solipsists. We create a solipsistic world. Based
> on this solipsistic world, we establish norms for the content of
> information. Only then can we decide what is information. And
> only then can we open our eyes to look at the world. But if we do
> this, we are sentenced to forever see only the solipsistic world,
> for that is all our information could be about.
This makes no sense at all to me. Training in pre-existing practices is
what enables us to take our place in the community as subjects concious
of a common world. Sure a child can open its eyes at birth, but it
cannot see that Jones was called out on strikes from birth, it has to
acquire the relevant concepts. If Wilfrid Sellars ("Experience and the
Philsophy of Mind") is right, a developing concept user can't even see
that something is green until he or she has mastered the logical space
of color terms, and the distinction between looking green and being
green, and acquired a conception of him or her self as a reporter to
whom things might look green without being green.
Anyway, there is simply no step of "creating a solipsistic world" in
the story nor of an individual's establishing conceptual norms ex
nihilo all by himself.
I suspect it is your focus on what goes on inside the individual
brain that leads you to project your own solipsism onto me.
>In my opinion, our idea of information cannot be based on norms.
How about "conceptual information" which can expressed in a public
language? Surely there are norms here, e.g. about what can correctly
be called a chair.
>Information is whatever we can get out of the informing signal, and
Agreed, I think.
>we must never allow some so-called norms to limit us in the
>information we can extract. We are not solipsists, because we accept
But without norms there is nothing, at least nothing that could show
up as an object for us. It would just be James' bloomin' buzzin'
confusion. If all we had to do to cognize the world was soak our
surfaces in the photon spray, there would be no need for any sort of
development or cultivation of an understanding. Even you don't believe
that is sufficient.
Put another way, the norms in question are not really limits.
>information without preconceived ideas (norms, for example) as to
>what is the content of that information. Therefore we are freed from
Why not: we are not solipsists because we can criticize the current
conceptual norms and possibly hope to change them for the better
(a paradigm shift)?
>any solipsistic world and are able to use the information to discover
>the kind of world that we are actually in.
It is a mistake to think of the norms in question as a "prison-house of
language". Perhaps you are ascribing to me the belief that these norms
are fixed once and for all.
I do suspect that for animals they might be.
>> This is why we don't consider computers as genuine
>>subjects of experience, even if (our) information is floating around in them.
>
>What it boils down to, is that the computers are only picking up
>information as our agents, and on our behalf, and are only picking up
>information which fits our prescriptions. The problem with the
OK, but what would it be for the computer itself to generate the
prescriptions? Only if the computer led a life to which the information
could be relevant would such language get a grip on it. And then we
would be talking about an autonomous robot, not merely a computer.
>computer, is that it relies on a concept of informational content
>which is based on norms (our norms), and these norms limit its
>ability to pick up information.
I would have thought the point is not to do away with the norms, it is
rather to find the context in which they could sensibly be said to be
autonomously determined.
>>With autonomous living organisms (whether natural or artificial) things
>>are different. The context of life provides the setting within which
>>we can talk about experience and natural information in a non-metaphorical
>>sense.
>
>Is this a plea for vitalism? I wouldn't expect an amoeba to have
Not at all. But surely a new class of predicates are applicable to
living things that are not applicable to non-living things. For example
"behaves", but also more determinate ones: hunts, chases, grabs, jumps,
seeks, flees, and a million more.
>That is your 'cognitive' theory of awareness. It is a solipsist's
>theory of awareness. According to that theory, all you could ever be
>aware of, is a world which fits the norms you use to establish the
>informational content. In my view, you must abandon the attempt to
>provide "an adequate account of first-order representational (or
>informational or cognitive) content." For any such account of
>informational content could only serve as a way of setting limits on
>what you can be aware of. A more complete conception of awareness
>requires that there should be no such limits.
Don't get it -- I thought it was *you* who were pushing the idea that
information detection counts as awareness. That points towards an
account of informational content, doesn't it? And certainly information
detection requires selection or discrimination (proto-conceptual
abilities) in which some aspects of the world are detected and others
not (a kind of limits).
As to apriori limits, I don't quite know why you ascribe them to me.
I am strong for paradigm shifts, changes in the provisional apriori.
>>I think there are problems with it, however. ...
>> And thirdly is the idea that this theory allows the
>>possibility of a mistake or misrepresentation of one's own mental
>>state where none really makes sense.
>
>The possibility of mistake only exists when information is evaluated
>according to norms. We must avoid a norm-based theory. That does
Then you have the problem that you cannot say what information content
is in any language at all. Perhaps its something ineffable, I can only
gesture ineffectually at it and hope someone else has "the same" inside
him, but never describe it in words? Or what?
>not imply that we cannot set tentative norms for some of our
>information, and have a possibility of mistakes for these types of
>information. But we must not allow ourselves to be limited by
>norms.
I thought *you* were pushing the idea that certain formal relations
defining a representation system functioned as norms for determining
the results of possible measurements. This would certainly be a
provisional a priori, an imposition of a normative structure, it seems to
me.
I would say that what makes for openness in the growth of knowledge is
the appreciation of the fact that such structures themselves are up for
grabs, revisable in the quest of a best overall system. And so they are
not really a priori limitations at all. But all inquiry has to start
somewhere, with a stock of concepts (representational systems) that are
simply inherited.
Neil Rickert
In <5852an$2...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>In article <584l95$f...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>>I don't see why this is applicable. Functionalists usually take the
>>position that experience has to do with how the information is
>>processed, and the picking up of information is taken as trivial. I
>>can see how your objection might apply to such a functionalism, but I
>>don't see its applicability to what I wrote.
>The question is (again) what the information picked up is when I have
>a headache. The usual assumption behind the objection is that whatever
>this information is, it is something that need not feel any particular
>way, for example, it need not hurt.
Perhaps the usual assumption is wrong.
>>The information that there is burning flesh in the vicinity is not
>>the same as the information that damage is being done to my flesh.
>>The difference seems obvious enough.
>Sure, but even after such a man recognizes that, in his words, "damage
>is being done to *my* flesh", he is still not feeling pain. He might be
>completely dispassionate about it if for some reason he has lost
>interest in preserving his flesh, for example.
Right. But he is still not picking up the same information. There
is a certain absurd theory, which seems to be subscribed to by many
philosophers. According to that theory, language is the basis of
all. Thus if what is being picked up by feeling pain, and by
smelling the burning flesh can both be described as "damage is being
done to my flesh", then the identical information must be being
picked up in both cases. I think the theory is obviously absurd, and
the case you raise is an obvious example of that absurdity. In this
case "damage is being done to my flesh" is a description of the
information. But language is a crude tool, and correspondingly it is
an imprecise description. More precise information is being picked
up, than is described by that verbal assertion. And at the greater
precision of actual information pickup, what is being picked up in
the two cases is not identical.
>>I don't see that as a problem. I might be mistaken in saying "I am
>>seeing a blue sky", for I might instead be watching television, or
>>hallucinating. However it is not at all clear that I could be
>>mistaken in saying "I am having a blue sky experience", for the
>>hallucination could still count as such an experience. I suggested
>Right, but if this requires that information be present inside your
>brain, one wonders how you could be so infallible at it. Oughtn't you
>to ask a neuroscientist whether such information is genuinely in there?
Why would a neuroscientist know? The neuroscientist can examine the
signals being processed, but he cannot determine what those signals
inform me about.
>Let us say that first-order information pickup concerns the outer
>world. Then what is picked up may be information or misinformation.
>Now on the theory in question, reflective or second-order information
>pickup would be the same sort of thing, detection of information, only
>concerned with information about the state of the first-order
>detectors.
But if my information detectors are part of the physical world, then
information about them is still first-order information. I think
this attempt to distinguish first-order and second-order information
tends to lead to confusion.
>Now I don't think this theory can account for first-person authority.
Of course not. You insist on a second person meaning for
"information", and as long as you interpret information in that way,
it cannot account for first-person effects. All the more reason for
you to reconsider your restrictive notion of "information."
>There is no reason at all why the second-order or reflective detectors
>might not also receive misinformation. But what would that be like?
Let's consider an analogy. Suppose that I program my computer so
that whenever I type a number in the keyboard, the computer treats
that as a Fahrenheit temperature, and displays the corresponding
Celsius temperature. If I type in 40, it displays 0.0, and if I type
in 77, it displays 25.0 . Now I look at my wristwatch. It says that
it is 10 minutes after the hour. I duly type in 10 to the computer,
and the computer displays -12.2 . The computer has picked up
Fahrenheit temperature information, and acted accordingly. But it
was misinformation, for the number typed in was not from a
temperature reading at all. However, it cannot be wrong that the
computer picked up Fahrenheit information, because the information
content is set by 'norms', and the behavioral response of the
computer to its input (i.e. its programming) establishes those
'norms'.
Earlier you asked why a headache hurts. That it hurts affects my
behavior. My behavioral response to the input acts as a 'norm' to
establish the information content that was picked up. It might be
that misinformation was picked up, but the information content that
was picked up, whether or not it was misinformation, is determined by
the behavioral response, which you have agreed is tightly connected
to the nature of the experience.
>I bet I am the most anti-solipsistic thinker you will ever encounter.
So you say. No doubt you believe it. But it is surely possible to
take an anti-solipstic stance at the surface level, yet to hold to
deep principles which imply solipsism.
>I think conceptual thinkers are essentially members of communities, for
>example.
In my view, that is a confused idea, or at least an excessively
restrictive idea, of "conceptual."
>> We start life as solipsists. We create a solipsistic world. Based
>> on this solipsistic world, we establish norms for the content of
>> information. Only then can we decide what is information. And
>> only then can we open our eyes to look at the world. But if we do
>> this, we are sentenced to forever see only the solipsistic world,
>> for that is all our information could be about.
>This makes no sense at all to me. Training in pre-existing practices is
>what enables us to take our place in the community as subjects concious
>of a common world.
But on your view of knowledge, the training could never get off the
ground. The training can only occur if information is received. But
on your view of information, this seems not possible until after the
training has taken at least some partial effect.
> Sure a child can open its eyes at birth, but it
>cannot see that Jones was called out on strikes from birth, it has to
>acquire the relevant concepts.
Yes, a child can open its eyes at birth. But what, if anything, can
the child see? It is my position that, at birth, the child can see
very little. It must first learn how to see. Experiments have been
done where kittens are raised in a visually impoverished
environment. They never acquire proper visual capabilities. I think
this evidence supports my view that vision requires learning.
>Anyway, there is simply no step of "creating a solipsistic world" in
>the story nor of an individual's establishing conceptual norms ex
>nihilo all by himself.
>I suspect it is your focus on what goes on inside the individual
>brain that leads you to project your own solipsism onto me.
That cannot be, since my primary focus is on how information is
picked up from the external world. What goes on inside the
individual brain is a mere implementation detail. By contrast, your
insistence on conceptual norms as the standard for information
amounts to an emphasis on what goes on inside.
>>In my opinion, our idea of information cannot be based on norms.
>How about "conceptual information" which can expressed in a public
>language? Surely there are norms here, e.g. about what can correctly
>be called a chair.
How about abandoning this insistence on public language
expressibility? By your insistence, you in effect prescribe a
specific representation system, and you only allow information that
can be accomodated by that representation system. Therefore the only
type of world you could be informed about, is the type of world that
your representation system can accomodate. There could be no
guarantee that your representation system is suitable for the world
we live in, and that is why I accuse you of solipsism. It seems to
me that you cannot avoid this problem, unless you are willing to
adopt a notion of information which is not tied to any specific
representation system.
>>Information is whatever we can get out of the informing signal, and
>Agreed, I think.
Ok. That is a starting point. One way of getting information out of
the signal, is by representing it with whatever representation system
we are using. In essence, the nature of that representation system
sets the norms that you consider so important.
A second method is to modify the representation system. Thus we can
have situations such that whereas the prior unmodified representation
system could not accomodate the information, the newly modified
representation system can. It seems to me that the term 'norm' is
not appropriately applied to a representation system which we can
continuously modify so as to change what information can be
represented.
>>we must never allow some so-called norms to limit us in the
>>information we can extract. We are not solipsists, because we accept
>But without norms there is nothing, at least nothing that could show
>up as an object for us. It would just be James' bloomin' buzzin'
>confusion.
No, I don't think so. It seems to me that the bloomin' buzzin'
confusion arises when we try to apply norms to information which does
not fit those norms.
> If all we had to do to cognize the world was soak our
>surfaces in the photon spray, there would be no need for any sort of
>development or cultivation of an understanding. Even you don't believe
>that is sufficient.
But why must you jump to such an extreme alternative?
>Why not: we are not solipsists because we can criticize the current
>conceptual norms and possibly hope to change them for the better
>(a paradigm shift)?
But we could only validly criticize current norms on the basis of
information, and if the norms determine what is information, then
they could not be a basis for criticism. If data is theory laden (or
norm laden), then data acceptable by the current theory (or norms)
could not possibly be used to challenge that theory. You have to
collect data under the proposed new theory or norms. But, with your
way of defining information, I don't see how that is possible. To
put it differently, I don't have any problem with norms to establish
what is acceptable data, but I want "information" to be separate from
those norms. If you like, I can accept a more formal notion of data,
but I need a more informal notion (not tied to norms or theories) for
information.
>It is a mistake to think of the norms in question as a "prison-house of
>language". Perhaps you are ascribing to me the belief that these norms
>are fixed once and for all.
No, I am not ascribing such a belief. But I am ascribing beliefs
which have that as a consequence. If the norms are a filter for what
is to be considered information, then there never could be
information that would cause you to question those norms. If you
have a broader view of information, then the possibility of
information that challenges existing norms arises.
>I do suspect that for animals they might be.
I doubt that, although it depends on which animals. Animals are
physiologically tied to limited environments, and it is the
limitation on possible environments which establishes limits (or
prison houses) for such animals.
>>What it boils down to, is that the computers are only picking up
>>information as our agents, and on our behalf, and are only picking up
>>information which fits our prescriptions. The problem with the
>OK, but what would it be for the computer itself to generate the
>prescriptions? Only if the computer led a life to which the information
>could be relevant would such language get a grip on it. And then we
>would be talking about an autonomous robot, not merely a computer.
That, by itself, would not be sufficient. The robot would have to be
equipped with flexible input systems, so that it could control the
type of information it chooses to pick up. Current day computers
mainly get their input from fixed sensors which limit what
information they can gather. The Turing machine model is essentially
a model of a fixed input system connected to a flexible data
processing system for data that has been picked up. We need more of
the flexibility to be in the input system.
>I would have thought the point is not to do away with the norms, it is
>rather to find the context in which they could sensibly be said to be
>autonomously determined.
I am not trying to do away with norms. But I want to give them a
more limited role than that you assign them. I see norms as useful
tools to deal with known types of information, but not as limits on
what can be considered information.
>>>With autonomous living organisms (whether natural or artificial) things
>>>are different. The context of life provides the setting within which
>>>we can talk about experience and natural information in a non-metaphorical
>>>sense.
>>Is this a plea for vitalism? I wouldn't expect an amoeba to have
>Not at all. But surely a new class of predicates are applicable to
>living things that are not applicable to non-living things. For example
>"behaves", but also more determinate ones: hunts, chases, grabs, jumps,
>seeks, flees, and a million more.
I would have thought that many of these predicates could be
applicable to non-living robots, but not to living trees.
>>That is your 'cognitive' theory of awareness. It is a solipsist's
>>theory of awareness. According to that theory, all you could ever be
>>aware of, is a world which fits the norms you use to establish the
>>informational content. In my view, you must abandon the attempt to
>>provide "an adequate account of first-order representational (or
>>informational or cognitive) content." For any such account of
>>informational content could only serve as a way of setting limits on
>>what you can be aware of. A more complete conception of awareness
>>requires that there should be no such limits.
>Don't get it -- I thought it was *you* who were pushing the idea that
>information detection counts as awareness.
Actually, I said that it counts as experience (or at least as
'experience'), rather than as awareness, and that awareness also
requires some sort of reflexivity in the type of information that is
considered.
> That points towards an
>account of informational content, doesn't it?
Sure. But not the type of account that you want.
> And certainly information
>detection requires selection or discrimination (proto-conceptual
>abilities) in which some aspects of the world are detected and others
>not (a kind of limits).
But there, it seems to me, you have a problem. For one would require
information as a basis for selection or discrimination, and on your
account that information would not be available until after the
selection or discrimination has been carried out.
>As to apriori limits, I don't quite know why you ascribe them to me.
>I am strong for paradigm shifts, changes in the provisional apriori.
But you take a position which rules out the possibility of there
being information on which to base a decision ot make a paradigm
shift.
>Then you have the problem that you cannot say what information content
>is in any language at all. Perhaps its something ineffable, I can only
>gesture ineffectually at it and hope someone else has "the same" inside
>him, but never describe it in words? Or what?
Isn't this always the problem when conceptual change is involved?
Those proposing alternative paradigms often have great difficulty in
persuading others to even consider their new alternative.
It is not that the information cannot be described. It is that in
general it cannot be described in any fixed representation system.
New information requires a change to the representation system, and
the content may not be describable to those who have not adopted the
new system. Incidently that is where GOFAI fails. GOFAI is full of
talk about knowledge representation systems, and the apparent hope is
that some representation system will be found in which everything can
be represented.
>I thought *you* were pushing the idea that certain formal relations
>defining a representation system functioned as norms for determining
>the results of possible measurements.
Sure. Again, I am not opposed to norms. I am opposed to the way you
want to use them. The norms always define the conventional wisdom.
If science can advance, it must be able to challenge the accepted
norms. And if there is to be an informational basis for scientific
change, then we cannot tie our notion of information to those norms
we propose to challenge.
Anders N Weinstein
In article <57k9k9$j...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
>
>I'm curious, because I too think that consideration of zombies leads me
>to think of qualia as essentially functional and hence falling under the
>rubrick of functionalism.
I would emphasize that "functionalism" can be used in a very broad
sense for any sort of relationist theory, as well as a narrow one which
denotes a particular type of reductive account ("causal role
functionalism"). In the broad sense Wittgenstein's claim that the
concept of pain is constituted in part by its connections to other
elements in the stream of human life activity, including above all
expressive behavior, might be a "functionalist" notion of pain, insofar
as it emphasizes conceptual connections to manifestations and
background context. On the other hand it is not at all physicalist, nor
reductive, nor is it concerned with causal relations to "inputs" and
"outputs".
It really is very doubtful to me whether there can be a physicalist or
reductive functionalist theory of claiming to have a headache, or of
claiming anything else for that matter. (I would say the same for
rule-following behavior).
Certainly claiming is an activity human beings can come to engage in,
and human beings are made entirely of physical stuff. But I think we
understand that such a weak concession all by itself won't suffice to
bring the normative into registration with the mechanical -- after all,
every painting is a physical object, yet one would doubt that we could
reduce the terms of aesthetic criticism to basic physics. At best it
entails that every *particular* event we can classify and explain as an
instance of claiming (or rule-following) also falls under universal
laws of physics. But the normative predicates might "cross-classify"
(Fodor's term) the same events or serve a very different explanatory
function.
Once one appreciates this, I don't see the motivation for pursuing
reductive physicalism of any variety, including causal role
functionalism. Reductive physicalism is false, as even Jerry Fodor
argued, but non-reductive physicalism (everything else "weakly
supervenes" on the physical) is nearly toothless. At any rate I think
it is a very difficult to say how physicalism should operate as a
constraint on higher-level vocabularies.
>This seems to me to make the idea of qualia as essentially non-functional
>out of step with what I take qualia to be. I take it that when I claim that
>I have qualia it is part of the *reason* that I claim that I have qualia that I do,
>indeed, have qualia. So, if we want our claims to have qualia to
>be justified by our *actually* *having* *qualia*, it appears to me to
>make it essential (and not just accidental) to what qualia are that
>they be functional.
At any rate, essential to such states that they be manifestable by
overt linguistic and non-linguistic expressions. Which is functional
in the broad sense but not necessarily the narrow.
Anders N Weinstein
In article <587d0e$i...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>In <5852an$2...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>Right. But he is still not picking up the same information. There
>is a certain absurd theory, which seems to be subscribed to by many
>philosophers. According to that theory, language is the basis of
>all. Thus if what is being picked up by feeling pain, and by
>smelling the burning flesh can both be described as "damage is being
>done to my flesh", then the identical information must be being
>picked up in both cases. I think the theory is obviously absurd, and
>the case you raise is an obvious example of that absurdity. In this
>case "damage is being done to my flesh" is a description of the
>information. But language is a crude tool, and correspondingly it is
>an imprecise description. More precise information is being picked
>up, than is described by that verbal assertion. And at the greater
>precision of actual information pickup, what is being picked up in
>the two cases is not identical.
Actually, in the case of pain, language might be finer grained -- pain
is often pretty unsubtle. For example, I understand that being seriously
stabbed with a knife feels more like being punched than being cut, heart
attacks are often felt as aches in the arm, etc.
>>Now I don't think this theory can account for first-person authority.
>
>Of course not. You insist on a second person meaning for
>"information", and as long as you interpret information in that way,
I don't know what "second person meaning" means here. It is crucial to
my outlook that conceptual information be communicable.
This raises some very difficult issues about the individuation of
intentional contents. In some intuitive sense we want to say that what
I express by "My right arm is raised" is the same content that you
might express by saying "Anders' right arm is raised". After all, if
you assert the negation of the latter, then we have a disagreement on
a common topic. On the other hand, the cognitive role of the first-person
mode of presentation is special, since the role of the former sentence
is not equipollent in my reasoning with the latter -- I might have
forgotten my name, for example. Reconciling these two demands is a
tricky philosophical project, I think.
>>There is no reason at all why the second-order or reflective detectors
>>might not also receive misinformation. But what would that be like?
>
>Let's consider an analogy. Suppose that I program my computer so
>that whenever I type a number in the keyboard, the computer treats
>that as a Fahrenheit temperature, and displays the corresponding
>Celsius temperature. If I type in 40, it displays 0.0, and if I type
>in 77, it displays 25.0 . Now I look at my wristwatch. It says that
>it is 10 minutes after the hour. I duly type in 10 to the computer,
>and the computer displays -12.2 . The computer has picked up
>Fahrenheit temperature information, and acted accordingly. But it
>was misinformation, for the number typed in was not from a
>temperature reading at all. However, it cannot be wrong that the
>computer picked up Fahrenheit information, because the information
>content is set by 'norms', and the behavioral response of the
>computer to its input (i.e. its programming) establishes those
>'norms'.
I think the analogy limps because the computer is not "aware" in your
sense, it does not issue reflective or introspective reports.
>>I bet I am the most anti-solipsistic thinker you will ever encounter.
>
>So you say. No doubt you believe it. But it is surely possible to
>take an anti-solipstic stance at the surface level, yet to hold to
>deep principles which imply solipsism.
Of course it is possible, but the charge seems without foundation to
me. And bizarre, since the whole philosophical idea behind some of what
I say is that we must make communicability between speakers via use of
the public language (shareability of content) the foundation or
starting point for any philosophizing about mental representation, and work
backwards from there to whatever we can say about an individual mind
or thinker as secondary.
>>This makes no sense at all to me. Training in pre-existing practices is
>>what enables us to take our place in the community as subjects concious
>>of a common world.
>
>But on your view of knowledge, the training could never get off the
>ground. The training can only occur if information is received. But
>on your view of information, this seems not possible until after the
>training has taken at least some partial effect.
I would have thought you have the same problem.
I really think the difficulty is this: you are interested in a notion
of information as "thin" (pre-conceptual) experience. It would be the
raw or unprocessed data that serves as the basis for conceptual
change. Call this stuff "raw" information.
Whereas I am speaking of experience or data only in the "thick" sense,
as the conceptualized result or upshot of such processing --
"processed" data. I take that unless the data is conceptualized, it's
occurence does not count as an episode in my mental or epistemic life,
and must therefore be something as it were alien to my mental life. In
this sense, the preconceptual given is not information *for* me, and
processing it is not an activity I perform. I grant there might be a
sense in which it could be said to present at my sensory surfaces and
the "processing" could be done by subsystems in my brain.
But I am still a little puzzled by the role of this raw information in
your theory. In Gibson's sense the information that Jones has just
uttered the English words "Eschew obfuscatory terminology" is present
in the pattern of stimulus imposed on the eardrum of a new born
infant. But it is not detected as such, indeed, the infant might not
be able to even discriminate the acoustic blast into distinct words or
even phonemes without further development.
Now are you talking about raw information as something that is there at
the surface even if the detectors needed to pick it up have not
developed yet? If so, it can hardly serve as the basis for acquiring
new perceptual capacities -- it is after all not detected as such at
all.
On the other hand, if you require that this information be *detected*,
then you are very close to my view, since then you are talking about
information as something which must be the *result* of some sort of
processing which requires "education of the senses". Which is similar
to my notion of acquiring a concept. (Actually I would say the ability
to merely discriminate these things is only one *part* of the
collection of abilities that would constitute having a concept of such
things as English words -- a protoconceptual ability a pigeon might
acquire).
So I don't really understand the role of raw, undetected information in
your account.
>> Sure a child can open its eyes at birth, but it
>>cannot see that Jones was called out on strikes from birth, it has to
>>acquire the relevant concepts.
>
>Yes, a child can open its eyes at birth. But what, if anything, can
>the child see? It is my position that, at birth, the child can see
>very little. It must first learn how to see. Experiments have been
>done where kittens are raised in a visually impoverished
>environment. They never acquire proper visual capabilities. I think
>this evidence supports my view that vision requires learning.
I can't tell if you appreciate that we are in *complete* agreement here.
In the human case, I think the acquired ability to detect *conceptual*
information comes with the acquisition of language. Only thus, for
example, could I see *that* Jones was called out on strikes.
But certainly development is required to see. (Although Gibson himself
argued that on his view information might be detected from birth).
>That cannot be, since my primary focus is on how information is
>picked up from the external world. What goes on inside the
>individual brain is a mere implementation detail. By contrast, your
>insistence on conceptual norms as the standard for information
>amounts to an emphasis on what goes on inside.
But rules or conceptual norms are not inside the head. One might better
say they live outside in the world, in such things as customs or
practices or techniques. What's literally inside the head is circuitry
subserving the ability to engage in these things, a topic of little
relevance to philosophy of mind, I think.
>>>In my opinion, our idea of information cannot be based on norms.
>
>>How about "conceptual information" which can expressed in a public
>>language? Surely there are norms here, e.g. about what can correctly
>>be called a chair.
>
>How about abandoning this insistence on public language
>expressibility? By your insistence, you in effect prescribe a
>specific representation system, and you only allow information that
>can be accomodated by that representation system. Therefore the only
>type of world you could be informed about, is the type of world that
>your representation system can accomodate. There could be no
It is true enough on anybody's view -- including your own, I would have
thought -- that the only type of world one can be informed about is the
type of world that your representation system can accomodate. For example,
the child at birth cannot be and be informed about English phrases and Jones
being called out on strikes. Neither, perhaps, can a fully developed
Bushman.
I still don't understand the complaint. You are not suggesting that the
child is informed about Jones being called out *before* the relevant
detectional capacities have been developed. Yet you think raw
information (which the child is not informed about either) must be the
basis for this development.
On the other hand, if you are talking about what the child can be
informed about *after* this development is begun, then I can talk
about such things as well -- the child acquired new concepts, and so
had its eyes and ears opened to a new world of facts. But the new
concepts were not developed by the child because he or she *first*
detected the raw information -- since that is not something the child
is never informed about (unless he grows up and learns acoustics and
the physics of light)
>guarantee that your representation system is suitable for the world
>we live in, and that is why I accuse you of solipsism. It seems to
>me that you cannot avoid this problem, unless you are willing to
>adopt a notion of information which is not tied to any specific
>representation system.
But like you, I believe the representation system can change, so that
new sorts of facts can come to show up for one. E.g someone who
cannot perceptually discriminate English words can come to be able to,
and even come to understand what they mean. But this cannot be done
on the basis of first hearing or detecting information about subtle
acoustic features, since most of us never come to detect that
information at all.
>>>Information is whatever we can get out of the informing signal, and
>
>>Agreed, I think.
>
>Ok. That is a starting point. One way of getting information out of
>the signal, is by representing it with whatever representation system
>we are using. In essence, the nature of that representation system
>sets the norms that you consider so important.
>
>A second method is to modify the representation system. Thus we can
>have situations such that whereas the prior unmodified representation
>system could not accomodate the information, the newly modified
>representation system can. It seems to me that the term 'norm' is
I think I can agree with this entirely. But now it seems that this
information you are talking about can't merely be "raw" information
(e.g. about the low-level physical properties of the acoustical
blast which is there prior to the ability to detect it) but must rather
be processed information (e.g. information about words, phrases, and
the like). Again it seems it is only information for someone once they
have the capacity to detect it.
> It seems to me that the term 'norm' is
>not appropriately applied to a representation system which we can
>continuously modify so as to change what information can be
>represented.
Why not? The rules of a game are normative, yet we can change the
rules.
>> If all we had to do to cognize the world was soak our
>>surfaces in the photon spray, there would be no need for any sort of
>>development or cultivation of an understanding. Even you don't believe
>>that is sufficient.
>
>But why must you jump to such an extreme alternative?
I am trying to point out a generic sort of parallelism between what
you say and what I say. We both accept that mere immersion of the sensory
surfaces in patterned energy ("raw data") does not itself count as
information detection (a kind of experience). We both believe that something
must develop in the individual in order that the information that is
available count *as* information that the organism can detect.
I suggest that in the human case this something is conceptual abilities
whiche are tied to linguistic expressibility, and thus constitute a form
of information that no non-linguistic creatures detect. You have a
different view which includes the capacities of non-linguistic animals.
But: the general schema of something like conceptual relativism would
seem to threaten equally *both* our positions. It is still true on
your view that something can't show up for a creature as information
unless they have the detectors for it, so they are trapped in a
"solipsist" world until they do. You say in effect that they can evolve
new detectors, I say people can acquire new concepts, rules, explanatory
practices. Now what's the difference in point of solipsism?
>>Why not: we are not solipsists because we can criticize the current
>>conceptual norms and possibly hope to change them for the better
>>(a paradigm shift)?
>
>But we could only validly criticize current norms on the basis of
>information, and if the norms determine what is information, then
>they could not be a basis for criticism. If data is theory laden (or
>norm laden), then data acceptable by the current theory (or norms)
>could not possibly be used to challenge that theory. You have to
Sometimes this is in fact true, I think. But of course now you are
talking about relations between one conceptualized domain (data) and
another (theory). Of course data in this sense might be neutral
with respect to theory, and can show up as something that the theory
fails to fit. That does not yet get us to non-conceptualized data.
>collect data under the proposed new theory or norms. But, with your
>way of defining information, I don't see how that is possible. To
>put it differently, I don't have any problem with norms to establish
>what is acceptable data, but I want "information" to be separate from
>those norms. If you like, I can accept a more formal notion of data,
>but I need a more informal notion (not tied to norms or theories) for
>information.
But if you are talking science, don't you need a language (conceptual
scheme, representational system) in which to capture this datal information?
>>It is a mistake to think of the norms in question as a "prison-house of
>>language". Perhaps you are ascribing to me the belief that these norms
>>are fixed once and for all.
>
>No, I am not ascribing such a belief. But I am ascribing beliefs
>which have that as a consequence. If the norms are a filter for what
>is to be considered information, then there never could be
>information that would cause you to question those norms. If you
>have a broader view of information, then the possibility of
>information that challenges existing norms arises.
Well certainly the norms of a paradigm of scientific theory construction
are such that they do not always fit observational data perfectly.
Again this is a relation between conceptualized data and conceptualized theory.
Whereas the child who has not learned to hear distinct words is not in
any way analogous to a scientist confronted with data that doesn't fit
a theory -- the child cannot even become aware of any failure. In some
ways more akin a blind man in the world of the sighted, simply
oblivious to what he is missing.
>>Not at all. But surely a new class of predicates are applicable to
>>living things that are not applicable to non-living things. For example
>>"behaves", but also more determinate ones: hunts, chases, grabs, jumps,
>>seeks, flees, and a million more.
>
>I would have thought that many of these predicates could be
>applicable to non-living robots, but not to living trees.
Right you are. Obviously I am more concerned with sentient animal life
than with plants or amoeba.
>> And certainly information
>>detection requires selection or discrimination (proto-conceptual
>>abilities) in which some aspects of the world are detected and others
>>not (a kind of limits).
>
>But there, it seems to me, you have a problem. For one would require
>information as a basis for selection or discrimination, and on your
>account that information would not be available until after the
>selection or discrimination has been carried out.
But one does not require information as the *basis* for the
discrimination. Rather the discrimination itself *is* the basis, it
*is* detetection of information.
>>As to apriori limits, I don't quite know why you ascribe them to me.
>>I am strong for paradigm shifts, changes in the provisional apriori.
>
>But you take a position which rules out the possibility of there
>being information on which to base a decision ot make a paradigm
>shift.
Not in the case of scientific theories. Yes indeed in the case of the
child acquiring concepts, since that is not based on prior information
detection *by the child*. E.g. the child doesn't come to hear word
boundaries because the child has *first* consciously detected some raw
information that can usefully be categorized that way.
>>Then you have the problem that you cannot say what information content
>>is in any language at all. Perhaps its something ineffable, I can only
>>gesture ineffectually at it and hope someone else has "the same" inside
>>him, but never describe it in words? Or what?
>
>Isn't this always the problem when conceptual change is involved?
>Those proposing alternative paradigms often have great difficulty in
>persuading others to even consider their new alternative.
That make it *difficult* to communicate. But if private, non-conceptual
information were required to be transferred, it would be impossible to
ever have any confidence that one had communicated. I couldn't compare
my preconceptual given with someone else's to ensure it's the same, I
can only check the manifestations.
>It is not that the information cannot be described. It is that in
>general it cannot be described in any fixed representation system.
>New information requires a change to the representation system, and
>the content may not be describable to those who have not adopted the
>new system.
But I can agree with this.
>>I thought *you* were pushing the idea that certain formal relations
>>defining a representation system functioned as norms for determining
>>the results of possible measurements.
>
>Sure. Again, I am not opposed to norms. I am opposed to the way you
>want to use them. The norms always define the conventional wisdom.
>If science can advance, it must be able to challenge the accepted
>norms. And if there is to be an informational basis for scientific
>change, then we cannot tie our notion of information to those norms
>we propose to challenge.
Perhaps *scientific* theory is based on conceptualized observations of
the objects of the common-sense world which is scientific theory neutral,
but not independent of conceptual norms.
andrew babian
I am concerned when people use information as a metaphor for brain signals,
because it seems likely that brain signals are very specific in their
location while information usually means the same thing wherever it is
actually encoded. This can be important to consciousness if a signal
is conscious by virtue of being in some location or structure in the
brain or from interaction with elements that have a certain location.
Peter_...@uk.ibm.com
In <587ek0$d...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>
>Certainly claiming is an activity human beings can come to engage in,
>and human beings are made entirely of physical stuff. But I think we
>understand that such a weak concession all by itself won't suffice to
>bring the normative into registration with the mechanical -- after all,
>every painting is a physical object, yet one would doubt that we could
>reduce the terms of aesthetic criticism to basic physics. At best it
>entails that every *particular* event we can classify and explain as an
>instance of claiming (or rule-following) also falls under universal
>laws of physics. But the normative predicates might "cross-classify"
>(Fodor's term) the same events or serve a very different explanatory
>function.
>
>Once one appreciates this, I don't see the motivation for pursuing
>reductive physicalism of any variety, including causal role
>functionalism.
What? Once one appreciates that reductive explanations *might*
be false, motivation for seeking them is removed? I just don't see
how to get from the *possibility* of an investigation failing to the
idea that such investigations are unmotivated. I would have thought
the possibility that the investigation might succeed is ample motivation
for the investigation.
>Reductive physicalism is false, as even Jerry Fodor
>argued, but non-reductive physicalism (everything else "weakly
>supervenes" on the physical) is nearly toothless.
I'd be interested in discussing this dichotomy: on the one hand,
a certain position; on the other hand, something 'nearly toothless'.
It doesn't sound like a plausible dichotomy to me. It sounds like
a dichotomy specially constructed for a purpose: rubbishing
physicalism.
Couldn't we just say: the versions of reductive physicalism we
currently have are false and the less constrained versions we
have don't seem to have a stopping point. Physicalism needs
new insights in order to make progress?
>At any rate I think
>it is a very difficult to say how physicalism should operate as a
>constraint on higher-level vocabularies.
Perhaps, even in your terms, there is a motivation for investigating
reductive explanations, even if you're convinced they can't succeed.
Perhaps, by understanding better exactly how and where they fail,
one understands better the relationships between vocabularies at
different levels.
>>This seems to me to make the idea of qualia as essentially non-functional
>>out of step with what I take qualia to be. I take it that when I claim that
>>I have qualia it is part of the *reason* that I claim that I have qualia that I do,
>>indeed, have qualia. So, if we want our claims to have qualia to
>>be justified by our *actually* *having* *qualia*, it appears to me to
>>make it essential (and not just accidental) to what qualia are that
>>they be functional.
>
>At any rate, essential to such states that they be manifestable by
>overt linguistic and non-linguistic expressions. Which is functional
>in the broad sense but not necessarily the narrow.
I don't accept the above. I don't accept the doctrine of 'manifestability',
since I don't see that the peculiar connotations of the word 'manifest'
need apply. I see nothing at all incoherent about the idea that some-one
may be experiencing and yet be unable to manifest that they are
experiencing. And, apparently, nor does anyone else.
Cheers,
Pete Lupton
CDJ
On 5 Dec 1996, Neil Rickert wrote:
> But language is a crude tool, and correspondingly it is
> an imprecise description.
Yeah, that's what we need: a nonlinguistic description.
CDJ
Neil Rickert
>I am concerned when people use information as a metaphor for brain signals,
>because it seems likely that brain signals are very specific in their
>location while information usually means the same thing wherever it is
>actually encoded.
Are you telling me that when I next fill out my tax return, I can
enter my income in the space for deductions, and leave the income
field blank? Somehow I think the IRS would have a strong opinion
about whether the information means the same thing, independent of
its location.
Anders N Weinstein
In article <588rm7$n...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
>In <587ek0$d...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>>
>>Once one appreciates this, I don't see the motivation for pursuing
>>reductive physicalism of any variety, including causal role
>>functionalism.
>
>What? Once one appreciates that reductive explanations *might*
>be false, motivation for seeking them is removed? I just don't see
>how to get from the *possibility* of an investigation failing to the
>idea that such investigations are unmotivated. I would have thought
>the possibility that the investigation might succeed is ample motivation
>for the investigation.
Ditto for the project of transmuting base metals into gold, I guess.
Of course you are strictly correct that it *might* be possible (though
I doubt it). As I see it, the orientation that produces the hysterical
urgency behind the philosophical movement I detest ("naturalism") is
NOT: "hey, it's crazy, but it just might work, so let's devote our
lives to trying". It is rather something like "it MUST be possible. The
only alternative is something horrible and mysterious like Cartesian
substance dualism or vitalism in biology. Why if it's not possible, the
whole 'scientific world picture' collapses, and we would be forced to
believe in spooky immaterial souls and the like".
Whereas I take it that for Wittgenstein, Ryle, Anscombe, Strawson,
others, the irreducibility of the intentional per se is simply a
non-issue, a don't-care that is no more threatening to science than the
irreducibility to basic physics of "is a table", "can get you on the
subway" or "is hunting its prey". The real issue is the *mystification*
that comes from the failure, when philosophizing, to consider
intentional activity as situated in its proper context.
I think once the source of the urgency -- either we reduce it,
eliminate it, or else some great intellectual disaster follows -- is
exploded, I think it is fair to say that the motivation for pursuing
the reductive project is seriously undermined. I don't think many
people would try to claim that a computer might be a subject of mental
states unless they believed they *had* to for some other reason, that
it was intellectually *obligatory*, not optional, given some sort of
due respect for the achievements of science. Of course we *may* still
pursue reductive projects simply because we can, but that is not, as I
see it, the root impulse behind philosophical "naturalism". (I should
add that I take the human potentiality for coming to engage in
intentional activity to be natural as it stands, without reduction to
something else.)
>>Reductive physicalism is false, as even Jerry Fodor
>>argued, but non-reductive physicalism (everything else "weakly
>>supervenes" on the physical) is nearly toothless.
>
>I'd be interested in discussing this dichotomy: on the one hand,
>a certain position; on the other hand, something 'nearly toothless'.
>It doesn't sound like a plausible dichotomy to me. It sounds like
>a dichotomy specially constructed for a purpose: rubbishing
>physicalism.
Right you are. I should admit it is not toothless in that it does seem
to rule out immaterial souls and vitalism. But that is not the same as
saying that the predicates and explanations of biology, say, get
reduced.
>Couldn't we just say: the versions of reductive physicalism we
>currently have are false and the less constrained versions we
>have don't seem to have a stopping point. Physicalism needs
>new insights in order to make progress?
In sober moments I would agree. My problem is I don't really know how
to formulate physicalism in a way that has bite against the
irreducibility of the intentional (or the normative, or the social, or
the biological). I do admit that this is a philosophical project.
John Haugeland wrote a nice paper called "Weak supervenience"
(published in American Philosophical Quaterly, I think around 1978, but
I could be wrong) in which the requirement of supervenience upon the
physical was argued to be very weak indeed. In conversation, he has
suggested he thinks the version in that paper was too weak, but I am at
a loss how to strengthen it.
As I see it, it is perfectly possible that when the mathematician Mary
inscribes a proof of a theorem on a paper -- even, perhaps, if she is
"intuiting the truth of the Godel sentence", it is *also* possible that
every elementary particle in the light cone of this activity -- in her
body and in the air and in the pencil and paper and around the moons of
Jupiter -- might be found to have evolved in accordance with the basic
laws of physics. So physical science might explain this massive
evolution, but not use the concepts of "proving a theorem". We can see
what's going on *as* a proving of a theorem or *as* a complex physical
event, but there is no reason we can get the two sorts of facts in
registration (just as we can treat an artwork as an aesthetic or
cultural object, or as a physical object when we are interested in
moving it).
In particular, I take it basic physics seeks explanations which subsume
in the same fashion incorrect moves in the proof, crossing outs,
back-trackings, furrowing the brow, crumpling the paper and tossing it
in the wastebasket, etc. It as it were is blind, or cares nothing for
the concept of a person's proving a theorem (or struggling
unsucessfully to prove a theorem) and the sort of behavior that
manifests such an attempt. That requires use of a different vocabulary
even for the very description of the explanandum, it seems
to me.
So it might even be that Penrose is right about our mathematical
abilities (although I am skeptical) but wrong that new *physics* is
required. Perhaps the existing physics is complete within its own
sphere, nevertheless.
>>At any rate I think
>>it is a very difficult to say how physicalism should operate as a
>>constraint on higher-level vocabularies.
>
>Perhaps, even in your terms, there is a motivation for investigating
>reductive explanations, even if you're convinced they can't succeed.
>Perhaps, by understanding better exactly how and where they fail,
>one understands better the relationships between vocabularies at
>different levels.
Agreed.
>>>This seems to me to make the idea of qualia as essentially non-functional
>>>out of step with what I take qualia to be. I take it that when I claim that
>>>I have qualia it is part of the *reason* that I claim that I have qualia that I do,
>>>indeed, have qualia. So, if we want our claims to have qualia to
>>>be justified by our *actually* *having* *qualia*, it appears to me to
>>>make it essential (and not just accidental) to what qualia are that
>>>they be functional.
>>
>>At any rate, essential to such states that they be manifestable by
>>overt linguistic and non-linguistic expressions. Which is functional
>>in the broad sense but not necessarily the narrow.
>
>I don't accept the above. I don't accept the doctrine of 'manifestability',
>since I don't see that the peculiar connotations of the word 'manifest'
>need apply. I see nothing at all incoherent about the idea that some-one
>may be experiencing and yet be unable to manifest that they are
>experiencing. And, apparently, nor does anyone else.
And maybe rocks might be experiencing horrible pain and we can never
know it.
Of course a human being might be paralyzed and so temporarily
unable to manifest his or her pain. Still we are talking about a
creature with natural and acquired forms of expression, not a rock.
Neil Rickert
In <587qko$f...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>In article <587d0e$i...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>>Right. But he is still not picking up the same information. There
>>is a certain absurd theory, which seems to be subscribed to by many
>>philosophers. According to that theory, language is the basis of
>>all. Thus if what is being picked up by feeling pain, and by
>>smelling the burning flesh can both be described as "damage is being
>>done to my flesh", then the identical information must be being
>>picked up in both cases. I think the theory is obviously absurd, and
>>the case you raise is an obvious example of that absurdity. In this
>>case "damage is being done to my flesh" is a description of the
>>information. But language is a crude tool, and correspondingly it is
>>an imprecise description. More precise information is being picked
>>up, than is described by that verbal assertion. And at the greater
>>precision of actual information pickup, what is being picked up in
>>the two cases is not identical.
>Actually, in the case of pain, language might be finer grained -- pain
>is often pretty unsubtle. For example, I understand that being seriously
>stabbed with a knife feels more like being punched than being cut, heart
>attacks are often felt as aches in the arm, etc.
Language more fine grained than pain? You would have to explain how
that could work. It would seem to say that I could use language to
distinguish between a severe pain and a very severe pain, whereas I
couldn't actually feel any difference in my experience with the
pain. In that case, what is the linguistically represented
distinction telling us?
>>Of course not. You insist on a second person meaning for
>>"information", and as long as you interpret information in that way,
>I don't know what "second person meaning" means here. It is crucial to
>my outlook that conceptual information be communicable.
Sorry -- I meant third person, or objective.
>This raises some very difficult issues about the individuation of
>intentional contents. In some intuitive sense we want to say that what
>I express by "My right arm is raised" is the same content that you
>might express by saying "Anders' right arm is raised".
As long as you depend on "in some intuitive sense", that sounds ok.
Beyond that, I can't see that it much matters. I don't see how we
could ever be sure that what I mean by "Al Gore's right arm is
raised" is identical with what you mean by "Al Gore's right arm is
raised." It is not even clear that it is meaningful to talk about
whether the meanings are identical. I think it highly unlikely that
we have identical concepts corresponding to most of the words we we
can share.
>> But it is surely possible to
>>take an anti-solipstic stance at the surface level, yet to hold to
>>deep principles which imply solipsism.
>Of course it is possible, but the charge seems without foundation to
>me. And bizarre, since the whole philosophical idea behind some of what
>I say is that we must make communicability between speakers via use of
>the public language (shareability of content) the foundation or
>starting point for any philosophizing about mental representation, and work
>backwards from there to whatever we can say about an individual mind
>or thinker as secondary.
And yet the picture you have just painted does not assign a role to
connecting the mental representations with the world. So how does it
argue against solipsism?
>>But on your view of knowledge, the training could never get off the
>>ground. The training can only occur if information is received. But
>>on your view of information, this seems not possible until after the
>>training has taken at least some partial effect.
>I would have thought you have the same problem.
No, because I don't see training as the basic element. We are able
to learn even in the absence of any training.
>I really think the difficulty is this: you are interested in a notion
>of information as "thin" (pre-conceptual) experience. It would be the
>raw or unprocessed data that serves as the basis for conceptual
>change. Call this stuff "raw" information.
>Whereas I am speaking of experience or data only in the "thick" sense,
>as the conceptualized result or upshot of such processing --
>"processed" data.
However there may be much processed data which is still not
conceptualized. So I think your emphasis on conceptualized data
misses most of what is happening, so that you finish up with a theory
which is largely disconnected from reality.
> I take that unless the data is conceptualized, it's
>occurence does not count as an episode in my mental or epistemic life,
>and must therefore be something as it were alien to my mental life.
Presumably that is a matter of the definition of 'mental' and
'epistemic'. It is all the more reason to be skeptical as to the
importance of matters 'mental' and 'epistemic', for they are defined
so as to deliberately exclude so much from consideration.
> In
>this sense, the preconceptual given is not information *for* me, and
>processing it is not an activity I perform.
You couldn't even keep your balance while walking down the street,
were it not for that which you refuse to count as being information
for you. You don't normally control your balance by means of mental
and epistemic feats.
>But I am still a little puzzled by the role of this raw information in
>your theory. In Gibson's sense the information that Jones has just
>uttered the English words "Eschew obfuscatory terminology" is present
>in the pattern of stimulus imposed on the eardrum of a new born
>infant. But it is not detected as such, indeed, the infant might not
>be able to even discriminate the acoustic blast into distinct words or
>even phonemes without further development.
But if Jones is the infant's father, perhaps the infant does detect
the presence of a familiar and soothing tonality. Information is
being detected, but what is detected depends on the person. This is
why I see information as necessarily subjective.
>Now are you talking about raw information as something that is there at
>the surface even if the detectors needed to pick it up have not
>developed yet? If so, it can hardly serve as the basis for acquiring
>new perceptual capacities -- it is after all not detected as such at
>all.
What is not detectable I might refer to as potential information, but
not as information. But there is information being picked up, and
that can form the basis for detecting new information, initially seen
as patterns of variation in the currently detectable information.
>On the other hand, if you require that this information be *detected*,
>then you are very close to my view, since then you are talking about
>information as something which must be the *result* of some sort of
>processing which requires "education of the senses".
But it seems to me that your view of information excludes the
possibility of explaining how the senses could be educated.
>>Yes, a child can open its eyes at birth. But what, if anything, can
>>the child see? It is my position that, at birth, the child can see
>>very little. It must first learn how to see. Experiments have been
>>done where kittens are raised in a visually impoverished
>>environment. They never acquire proper visual capabilities. I think
>>this evidence supports my view that vision requires learning.
>I can't tell if you appreciate that we are in *complete* agreement here.
>In the human case, I think the acquired ability to detect *conceptual*
>information comes with the acquisition of language.
Then our agreement is far from complete. For it seems to me that you
have excluded the possibility that there could be evidence the child
might use to acquire language, and that leaves it as a complete
mystery that language could be acquired. I think the ability to
detect conceptual information is a prerequisite to acquisition of
language.
>>That cannot be, since my primary focus is on how information is
>>picked up from the external world. What goes on inside the
>>individual brain is a mere implementation detail. By contrast, your
>>insistence on conceptual norms as the standard for information
>>amounts to an emphasis on what goes on inside.
>But rules or conceptual norms are not inside the head. One might better
>say they live outside in the world, in such things as customs or
>practices or techniques. What's literally inside the head is circuitry
>subserving the ability to engage in these things, a topic of little
>relevance to philosophy of mind, I think.
Well that is why philosophers so often sound like dualists or
solipsists. Heads, brains, etc, are dispensible, as are trees,
birds, flowers. All we need is a purely abstract world with detached
concepts floating around.
>>How about abandoning this insistence on public language
>>expressibility? By your insistence, you in effect prescribe a
>>specific representation system, and you only allow information that
>>can be accomodated by that representation system. Therefore the only
>>type of world you could be informed about, is the type of world that
>>your representation system can accomodate. There could be no
>It is true enough on anybody's view -- including your own, I would have
>thought -- that the only type of world one can be informed about is the
>type of world that your representation system can accomodate.
Unfortunately, there is some ambiguity in the expression "can be
informed about". If it refers to the information I currently have
available to me, of course this is right. If it refers to the future
possibility of acquiring information, then it is wrong, for I may be
able to change my representation system so as to allow it to accept
that future information.
>I still don't understand the complaint. You are not suggesting that the
>child is informed about Jones being called out *before* the relevant
>detectional capacities have been developed. Yet you think raw
>information (which the child is not informed about either) must be the
>basis for this development.
Either raw information is the basis, or Berkeleyan ideas are directly
inserted into souls by some God.
>On the other hand, if you are talking about what the child can be
>informed about *after* this development is begun, then I can talk
>about such things as well -- the child acquired new concepts, and so
>had its eyes and ears opened to a new world of facts. But the new
>concepts were not developed by the child because he or she *first*
>detected the raw information -- since that is not something the child
>is never informed about (unless he grows up and learns acoustics and
>the physics of light)
This depends on what you mean by "the child". If what you mean, is
the ghost in the machine, then it may be reasonable to say that the
child is never informed about raw information. But if you mean the
flesh and blood child, whose sensory organs are regularly picking up
that raw information, then your assertion is surely wrong.
>>guarantee that your representation system is suitable for the world
>>we live in, and that is why I accuse you of solipsism. It seems to
>>me that you cannot avoid this problem, unless you are willing to
>>adopt a notion of information which is not tied to any specific
>>representation system.
>But like you, I believe the representation system can change, so that
>new sorts of facts can come to show up for one. E.g someone who
>cannot perceptually discriminate English words can come to be able to,
>and even come to understand what they mean. But this cannot be done
>on the basis of first hearing or detecting information about subtle
>acoustic features, since most of us never come to detect that
>information at all.
Again, all you can mean is that the ghost in the machine never
detects that information (about acoustic features) at all. But the
flesh and blood person behaves in such a way as to indicate that the
information was detected and acted upon.
>>A second method is to modify the representation system. Thus we can
>>have situations such that whereas the prior unmodified representation
>>system could not accomodate the information, the newly modified
>>representation system can. It seems to me that the term 'norm' is
>I think I can agree with this entirely. But now it seems that this
>information you are talking about can't merely be "raw" information
>(e.g. about the low-level physical properties of the acoustical
>blast which is there prior to the ability to detect it) but must rather
>be processed information (e.g. information about words, phrases, and
>the like). Again it seems it is only information for someone once they
>have the capacity to detect it.
There is nothing mere about the raw information. The most highly
processed information is derived from the raw information. If the
raw information is "mere", then so should be the words and phrases
derived from it.
>> It seems to me that the term 'norm' is
>>not appropriately applied to a representation system which we can
>>continuously modify so as to change what information can be
>>represented.
>Why not? The rules of a game are normative, yet we can change the
>rules.
It is usually considered bad practice to change the rules in the
middle of the game.
>>> If all we had to do to cognize the world was soak our
>>>surfaces in the photon spray, there would be no need for any sort of
>>>development or cultivation of an understanding. Even you don't believe
>>>that is sufficient.
>>But why must you jump to such an extreme alternative?
>I am trying to point out a generic sort of parallelism between what
>you say and what I say. We both accept that mere immersion of the sensory
>surfaces in patterned energy ("raw data") does not itself count as
>information detection (a kind of experience).
Likewise, a mere immersion in words and phrases does not count for
information detection. It seems to me you are making the wrong
distinctions.
A tree is immersed in patterned energy which impact its sensory
surfaces (its leaves). The distinguishes us from trees, is not that
we are also immersed in mere words and phrases. Rather, it is that
we are not merely immersed in the patterned energy. Unlike the tree,
we have the ability to seek out new sources of patterned energy, and
to modify the manner in which the patterned energy impacts our
surfaces. As Gibson insisted, we are not passive, but active
participants in the pickup of information.
> We both believe that something
>must develop in the individual in order that the information that is
>available count *as* information that the organism can detect.
>I suggest that in the human case this something is conceptual abilities
>whiche are tied to linguistic expressibility, and thus constitute a form
>of information that no non-linguistic creatures detect. You have a
>different view which includes the capacities of non-linguistic animals.
And you thereby appoint us gods, different from all other creatures.
And our cognition is then necessarily irreducible, for there are no
physiological differences which would make us into gods rather than
animals. But surely this is dualism, for the difference which makes
us gods must be the presence of a ghost in the machine.
>But: the general schema of something like conceptual relativism would
>seem to threaten equally *both* our positions.
It does not threaten my position at all.
> It is still true on
>your view that something can't show up for a creature as information
>unless they have the detectors for it, so they are trapped in a
>"solipsist" world until they do.
I have no such problem, for I have available to me information based
processes which will modify the detection ability to accomodate
previously undetected information. Perhaps you also have such
processes, but you have not explained what they could be, and your
views as to what constitutes information would seem to rule them
out.
> You say in effect that they can evolve
>new detectors, I say people can acquire new concepts, rules, explanatory
>practices. Now what's the difference in point of solipsism?
The difference is that I base the development of new detection
abilities on information, while you seem to exclude the possibility
of their being information on which to base such a development.
>But if you are talking science, don't you need a language (conceptual
>scheme, representational system) in which to capture this datal information?
When Einstein wrote of his general relativity, he presented some
thought experiments. In one, a person was travelling in an elevator
car, but the car could be in an elevator shaft or could be moving in
free space. Why did Einstein use this method for describing his
theory? My view is that he had no language in which to represent the
information he wanted to convey. Therefore he needed to paint a word
picture, which would provide the raw information his audience could
use to construct new representational abilities. Only after these
new abilities had been constructed, could he hope to convey the
information about general relativity.
More generally, I think we use literature and drama in the same way,
so as to convey what we cannot directly represent in language. And
often this may require that the conditions be created for generating
new detection and representation abilities in the audience.
>>But there, it seems to me, you have a problem. For one would require
>>information as a basis for selection or discrimination, and on your
>>account that information would not be available until after the
>>selection or discrimination has been carried out.
>But one does not require information as the *basis* for the
>discrimination. Rather the discrimination itself *is* the basis, it
>*is* detetection of information.
But if there is no raw information, then the discriminator cannot be
created. Once we have a stream of raw information, we can attempt to
divide the stream in various ways, so as to increase our ability to
discriminate. As we do so, we increase the amount of information we
extract from the stream.
>>But you take a position which rules out the possibility of there
>>being information on which to base a decision ot make a paradigm
>>shift.
>Not in the case of scientific theories. Yes indeed in the case of the
>child acquiring concepts, since that is not based on prior information
>detection *by the child*.
Presumably by this you mean that it is not based on prior information
detection by the ghost in the machine. Surely it is based on
information detection by the flesh and blood (including neurons).
> E.g. the child doesn't come to hear word
>boundaries because the child has *first* consciously detected some raw
>information that can usefully be categorized that way.
If you are only willing to consider what is conscious, then it seems
to me that you are committed to dualism.
>>Isn't this always the problem when conceptual change is involved?
>>Those proposing alternative paradigms often have great difficulty in
>>persuading others to even consider their new alternative.
>That make it *difficult* to communicate. But if private, non-conceptual
>information were required to be transferred, it would be impossible to
>ever have any confidence that one had communicated.
I don't see why? Those training guide dogs for the blind seem to
have confidence in their ability to assess whether the dogs have
acquired the abilities. Yet on your language-based view of the
conceptual, I don't see how you could say that this was based on
other than non-conceptual information.
> I couldn't compare
>my preconceptual given with someone else's to ensure it's the same, I
>can only check the manifestations.
It is only an illusion that you can compare your conceptual
information with somebody else's conceptual information. For you
cannot compare your concepts, and thus you cannot be certain that
what you take to be the conceptual information is the same as what
the other person takes to be the conceptual information. So you are
still left with checking the manifestations.
Anders N Weinstein
In article <589rt1$l...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>In <587qko$f...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>>Actually, in the case of pain, language might be finer grained -- pain
>>is often pretty unsubtle. For example, I understand that being seriously
>>stabbed with a knife feels more like being punched than being cut, heart
>>attacks are often felt as aches in the arm, etc.
>
>Language more fine grained than pain? You would have to explain how
>that could work. It would seem to say that I could use language to
>distinguish between a severe pain and a very severe pain, whereas I
>couldn't actually feel any difference in my experience with the
>pain. In that case, what is the linguistically represented
>distinction telling us?
I should have said: language, in particular scientific language, makes
possible representation of finer grained information than could possibly
be carried by sensory states like a feeling of pain. For example,
localizing the damage precisely, measuring the wavelength of light, etc,
things which can't be distinguished by our normal sensory means.
Not for nothing did Leibniz call perception a "confused" representation
of the structure of reality.
Of course if we are talking about representing the headache itself -- not
some sort of cause of the pain -- the issue is different.
>>This raises some very difficult issues about the individuation of
>>intentional contents. In some intuitive sense we want to say that what
>>I express by "My right arm is raised" is the same content that you
>>might express by saying "Anders' right arm is raised".
>
>As long as you depend on "in some intuitive sense", that sounds ok.
>Beyond that, I can't see that it much matters. I don't see how we
>could ever be sure that what I mean by "Al Gore's right arm is
>raised" is identical with what you mean by "Al Gore's right arm is
>raised." It is not even clear that it is meaningful to talk about
>whether the meanings are identical. I think it highly unlikely that
>we have identical concepts corresponding to most of the words we we
>can share.
I suppose objective discourse on a common topic is impossible, and
there is no point ever arguing with anyone about anything, since all
meanings are ultimately solipistic, private and unshareable and on your
view.
Look, if you put forward the sentence and I put forward the negation
and we treat ourselves as having a disagreement, then we are thereby
taking ourselves to be "samesayers", speakers who have a common topic
of debate. Perhaps certain further evidence could show that this is
an illusion -- you mean the completely different Al Gore who lives
next door to you -- and then the appearance of disagreement vanishes.
But the (not very formal)idea of a common content should be treated
as a kind of construction out of the idea of a topic of rational
agreement or disagreement. And to abandon the latter idea is to cease
to be a thinker.
>>I say is that we must make communicability between speakers via use of
>>the public language (shareability of content) the foundation or
>>starting point for any philosophizing about mental representation, and work
>>backwards from there to whatever we can say about an individual mind
>>or thinker as secondary.
>
>And yet the picture you have just painted does not assign a role to
>connecting the mental representations with the world. So how does it
>argue against solipsism?
But intentional contents *are* modes of directedness at the world.
For example, the content expressed by "Vice President Al Gore's right
arm is now raised" is directed towards the state of a flesh and blood
appendage in (I assume) Washington D.C. It answers for its correctness
to that reality. The intentional content I might express by pointing
at the wastebasket in my office and saying "that is empty" is directed
at the state of the wastebasket in public space.
The world is what we use our words to talk *about*. So there is
no problem about the connection with reality.
I don't recall mentioning mental representations however. I was
speaking rather of intentional contents of mental states expressible
by language. Perhaps you could think of the public symbols as
representations.
>>>But on your view of knowledge, the training could never get off the
>>>ground. The training can only occur if information is received. But
>>>on your view of information, this seems not possible until after the
>>>training has taken at least some partial effect.
>
>>I would have thought you have the same problem.
>
>No, because I don't see training as the basic element. We are able
>to learn even in the absence of any training.
Some things, of course. I don't think most of us could learn to use the
intellectual tools -- e.g. new symbolic notations, like that of the
calculus -- without training. At best we could say a few rare geniuses
might hit upon them on their own, but even they started somewhere, with
some intellectual or research tradition.
>>Whereas I am speaking of experience or data only in the "thick" sense,
>>as the conceptualized result or upshot of such processing --
>>"processed" data.
>
>However there may be much processed data which is still not
>conceptualized. So I think your emphasis on conceptualized data
>misses most of what is happening, so that you finish up with a theory
>which is largely disconnected from reality.
OK, so is it fair to say you are working with a trio: "raw"
(unprocessed) stimulation, *detected* information, and conceptualized
stuff. Where conceptualization is supposed to effect a kind of
selection or abstraction from the *detected* information, which is too
rich in content to be captured entirely in the net of language without
loss. Correct?
Then it is still true, as I said, that there is a parallel between
the situation with respect to your detected information and my
conceptual information, namely both require cultivation to obtain,
which cannot itself be accounted for in these terms.
I his Locke lectures _Mind and World_, John McDowell alludes to a
similar sort of position in the work of his former colleague Gareth
Evans' work (which McD, who edited Evans' posthumous book, generally
regards as brilliant). McDowell argues that it still fails under the
critique of the Myth of the Given, since the transition from information
to an epistemic state like a belief cannot account for the latter's
standing as justified.
Of course this polemic depends on an epistemological orientation towards
the philosophy of perception, which you have rejected. But I have
never understood the basis of the rejection; I think it is a neutral,
mundane, everyday fact that people have cognitive attitudes concerning
the world which can be assessed in epistemic terms.
>> I take that unless the data is conceptualized, it's
>>occurence does not count as an episode in my mental or epistemic life,
>>and must therefore be something as it were alien to my mental life.
>
>Presumably that is a matter of the definition of 'mental' and
>'epistemic'. It is all the more reason to be skeptical as to the
>importance of matters 'mental' and 'epistemic', for they are defined
>so as to deliberately exclude so much from consideration.
As long as there *are* matters mental and epistemic that is all I need.
>> In
>>this sense, the preconceptual given is not information *for* me, and
>>processing it is not an activity I perform.
>
>You couldn't even keep your balance while walking down the street,
>were it not for that which you refuse to count as being information
>for you. You don't normally control your balance by means of mental
>and epistemic feats.
Yes, pretty obviously that is not a mental feat and the relevant
information is not obtained by me. So what?
>>On the other hand, if you require that this information be *detected*,
>>then you are very close to my view, since then you are talking about
>>information as something which must be the *result* of some sort of
>>processing which requires "education of the senses".
>
>But it seems to me that your view of information excludes the
>possibility of explaining how the senses could be educated.
No, it only precludes explaining it at the level at which we speak
of information as detected.
I have been trying to suggest that you have a parallel problem.
>>But rules or conceptual norms are not inside the head. One might better
>>say they live outside in the world, in such things as customs or
>>practices or techniques. What's literally inside the head is circuitry
>>subserving the ability to engage in these things, a topic of little
>>relevance to philosophy of mind, I think.
>
>Well that is why philosophers so often sound like dualists or
>solipsists. Heads, brains, etc, are dispensible, as are trees,
>birds, flowers. All we need is a purely abstract world with detached
>concepts floating around.
Unfair. Compare the custom or practice of using a signpost (one of
Wittgenstein's favorites). Well in one sense a signpost is concrete --
it's nothing over and above a hunk of wood, after all. But the wood considered
all by itself seems dead, inert, devoid of significance.
On the other hand the signpost has a socially constituted significance
-- as Wittgenstein suggested, it has a kind of life: it *points*. But
it only takes on this feature in the context of a custom, a human
practice of using signposts in a certain way. Which practice, on the
"subjective" side, exists as skill or know-how on the part of the
users, an ability to use signposts correctly, to detect when others do
or do not, to train others in the proper use, etc. There need be no
include intellectual mental representations *of* the norms of the
practice.
In this way we understand socially constituted significance can emerge in
the world at its own level, without mediation of mental representations.
Now the signpost is thoroughly material. But where, Witt might lead us
to ask, is the *use* we make of the signpost, according to which it
*points*? We want to reply with terms like "convention", "custom",
"practice", "institution", "technique (of operating with)". That's
find, but what sort of entities are *these*? In some sense they are
abstract. Yet they are thoroughly embodied in activities with material
objects, a bit like Aristotle's forms in matter.
I think this model points the way out of the worry you hint at, about
"a purely abstract world". We can say that concepts too are embodied
in uses we make of linguistic tokens. Then they do not "float around" in
complete abstraction from what human beings do in the spatio-temporal
world with signs. But they are not material objects or attributes.
>>It is true enough on anybody's view -- including your own, I would have
>>thought -- that the only type of world one can be informed about is the
>>type of world that your representation system can accomodate.
>
>Unfortunately, there is some ambiguity in the expression "can be
>informed about". If it refers to the information I currently have
>available to me, of course this is right. If it refers to the future
>possibility of acquiring information, then it is wrong, for I may be
>able to change my representation system so as to allow it to accept
>that future information.
And here is the parallel, for I can say one can change one's concepts.
>>I still don't understand the complaint. You are not suggesting that the
>>child is informed about Jones being called out *before* the relevant
>>detectional capacities have been developed. Yet you think raw
>>information (which the child is not informed about either) must be the
>>basis for this development.
>
>Either raw information is the basis, or Berkeleyan ideas are directly
>inserted into souls by some God.
Well Berkeley's concepttion of ideas as images is confused.
But again, I thought on your view too the child is not informed about
raw information. So it is not a basis the child can go on.
>>On the other hand, if you are talking about what the child can be
>>informed about *after* this development is begun, then I can talk
>>about such things as well -- the child acquired new concepts, and so
>>had its eyes and ears opened to a new world of facts. But the new
>>concepts were not developed by the child because he or she *first*
>>detected the raw information -- since that is not something the child
>>is never informed about (unless he grows up and learns acoustics and
>>the physics of light)
>
>This depends on what you mean by "the child". If what you mean, is
>the ghost in the machine, then it may be reasonable to say that the
Of course I don't mean any ghost in the machine. The child stands to
his or her functional parts in a fashion similar to Ryle's University
and its colleges -- one is the organized whole comprised of the latter.
The child is certainly no ghost -- he or she is the creature you see
before you. But still the child is not a subsystem of the child, but
the whole.
>child is never informed about raw information. But if you mean the
>flesh and blood child, whose sensory organs are regularly picking up
>that raw information, then your assertion is surely wrong.
I would have thought you also agree that just because the sense organs
are stimulated does not mean the child is informed.
I wonder if you are dropping my attempted distinction between "raw"
information -- there at the surface whether or not it is detected --
and detected or obtained information. I think you are referring now to
information that is detected by a subsystem of the child, even if it is
not detected by the child itself (as information relevant to regulation of the
autonomic nervous system might be detected in the cerebellum, without
being detected by me).
>>But like you, I believe the representation system can change, so that
>>new sorts of facts can come to show up for one. E.g someone who
>>cannot perceptually discriminate English words can come to be able to,
>>and even come to understand what they mean. But this cannot be done
>>on the basis of first hearing or detecting information about subtle
>>acoustic features, since most of us never come to detect that
>>information at all.
>
>Again, all you can mean is that the ghost in the machine never
>detects that information (about acoustic features) at all. But the
The person is not a ghost, but the flesh and blood creature
you see before you.
>flesh and blood person behaves in such a way as to indicate that the
>information was detected and acted upon.
But not to indicate that the person got the information.
>>> It seems to me that the term 'norm' is
>>>not appropriately applied to a representation system which we can
>>>continuously modify so as to change what information can be
>>>represented.
>
>>Why not? The rules of a game are normative, yet we can change the
>>rules.
>
>It is usually considered bad practice to change the rules in the
>middle of the game.
Depends entirely on the kind of game.
>A tree is immersed in patterned energy which impact its sensory
>surfaces (its leaves). The distinguishes us from trees, is not that
>we are also immersed in mere words and phrases. Rather, it is that
>we are not merely immersed in the patterned energy. Unlike the tree,
>we have the ability to seek out new sources of patterned energy, and
>to modify the manner in which the patterned energy impacts our
>surfaces. As Gibson insisted, we are not passive, but active
>participants in the pickup of information.
I can agree. However I am not quite sure a child learning language
is exactly an active participant -- I expect the child will acquire the
language whether it seeks to or not, owing to the sub-personal
processes in its brain.
>> We both believe that something
>>must develop in the individual in order that the information that is
>>available count *as* information that the organism can detect.
>>I suggest that in the human case this something is conceptual abilities
>>whiche are tied to linguistic expressibility, and thus constitute a form
>>of information that no non-linguistic creatures detect. You have a
>>different view which includes the capacities of non-linguistic animals.
>
>And you thereby appoint us gods, different from all other creatures.
>And our cognition is then necessarily irreducible, for there are no
>physiological differences which would make us into gods rather than
>animals. But surely this is dualism, for the difference which makes
>us gods must be the presence of a ghost in the machine.
I don't see any problems with my form of conceptual dualism. I certainly
don't see any need for ghosts as immaterial components. It is just the
*description* of what we are doing when we obey a rule, or follow a
rational norm, that is irreducible to physiology.
To take one example, physiology can't explain what it is to be a
*correct* mathematical proof, and it similarly won't explain what it is
for a mathematician to be able to recognize or strive after this
property. Even if all such recognitions or strivings also have explanations
in the language of basic physics.
>> It is still true on
>>your view that something can't show up for a creature as information
>>unless they have the detectors for it, so they are trapped in a
>>"solipsist" world until they do.
>
>I have no such problem, for I have available to me information based
>processes which will modify the detection ability to accomodate
>previously undetected information. Perhaps you also have such
Not clear on this: how does the undetected information have an
influence if it is not detected?
>>But if you are talking science, don't you need a language (conceptual
>>scheme, representational system) in which to capture this datal information?
>
>When Einstein wrote of his general relativity, he presented some
>thought experiments. In one, a person was travelling in an elevator
>car, but the car could be in an elevator shaft or could be moving in
>free space. Why did Einstein use this method for describing his
>theory? My view is that he had no language in which to represent the
>information he wanted to convey. Therefore he needed to paint a word
>picture, which would provide the raw information his audience could
I'm not getting the difference between "painting a word picture" and
conveying information through language.
>More generally, I think we use literature and drama in the same way,
>so as to convey what we cannot directly represent in language. And
>often this may require that the conditions be created for generating
>new detection and representation abilities in the audience.
In general, I would agree, but would say that literature and drama and
the like can be the goad to the formation of new concepts and
understandings.
>>>But there, it seems to me, you have a problem. For one would require
>>>information as a basis for selection or discrimination, and on your
>>>account that information would not be available until after the
>>>selection or discrimination has been carried out.
>
>>But one does not require information as the *basis* for the
>>discrimination. Rather the discrimination itself *is* the basis, it
>>*is* detetection of information.
>
>But if there is no raw information, then the discriminator cannot be
>created. Once we have a stream of raw information, we can attempt to
Of course it can be created, but it can't be created by first detecting
what ex hypothesi went undetected. I.e. the move from not detecting
the information to detecting the information can't be accounted for
in informational terms.
>divide the stream in various ways, so as to increase our ability to
>discriminate. As we do so, we increase the amount of information we
>extract from the stream.
>
>>>But you take a position which rules out the possibility of there
>>>being information on which to base a decision ot make a paradigm
>>>shift.
>
>>Not in the case of scientific theories. Yes indeed in the case of the
>>child acquiring concepts, since that is not based on prior information
>>detection *by the child*.
>
>Presumably by this you mean that it is not based on prior information
>detection by the ghost in the machine. Surely it is based on
I wish you'd drop this nonsense about ghosts.
>information detection by the flesh and blood (including neurons).
No doubt our neurons detect lots of information that we don't. Without
this we would die.
>> E.g. the child doesn't come to hear word
>>boundaries because the child has *first* consciously detected some raw
>>information that can usefully be categorized that way.
>
>If you are only willing to consider what is conscious, then it seems
>to me that you are committed to dualism.
Why? I certainly think intentional states of consciousness are natural
attributes of living organisms. For example I can now look around and so
enjoy consciousness of my computer, desktop, and office. I can grant that
these states might be realized in neural material without supposing that
neuroscience explains their intentional attributes (the relationship to
my office).
It seems to me that if you never get to the point where you
consider states of consciousness, then something must be left out.
>>>Isn't this always the problem when conceptual change is involved?
>>>Those proposing alternative paradigms often have great difficulty in
>>>persuading others to even consider their new alternative.
>
>>That make it *difficult* to communicate. But if private, non-conceptual
>>information were required to be transferred, it would be impossible to
>>ever have any confidence that one had communicated.
>
>I don't see why? Those training guide dogs for the blind seem to
>have confidence in their ability to assess whether the dogs have
>acquired the abilities. Yet on your language-based view of the
>conceptual, I don't see how you could say that this was based on
>other than non-conceptual information.
Yes, I think skills as such are not conceptual. Training a guide dog is
not exactly transfering information to the dog. Perhaps something like
non-conceptual information that this is right, that is wrong, this is
better, and you've been a good doggie does get transmitted to the dog
in the course of the training.
>> I couldn't compare
>>my preconceptual given with someone else's to ensure it's the same, I
>>can only check the manifestations.
>
>It is only an illusion that you can compare your conceptual
>information with somebody else's conceptual information. For you
>cannot compare your concepts, and thus you cannot be certain that
Why not?
>what you take to be the conceptual information is the same as what
>the other person takes to be the conceptual information. So you are
>still left with checking the manifestations.
Of course we must check the manifestations. But that means we must not
think of the states of understanding or grasping as separable from the
manifestations. (This sounds like behaviorism, but I think the catch is
the language we use for describing the manifestations).
David Longley
In article <58d204$5...@usenet.srv.cis.pitt.edu>
'Once it is shown that a region of discourse is not
extensional, then according to Quine, we have reason to
doubt its claim to describe the structure of reality.'
C. Hookway
Logic: Canonical Notation and Extensionality
Quine (1988)
'..there's a tradition which argues that - epistemology
to one side - it is at best a strategic mistake to
attempt to develop a psychology which individuates
mental states without reference to their environmental
causes and effects...I have in mind the tradition which
includes the American Naturalists (notably Pierce and
Dewey), all the learning theorists, and such
contemporary representatives as Quine in philosophy and
Gibson in psychology. The recurrent theme here is that
psychology is a branch of biology, hence that one must
view the organism as embedded in a physical environment.
The psychologist's job is to trace those
organism/environment interactions which constitute its
behavior.'
J. Fodor (1980) ibid. p.64
Even if you two agreed on something, it wouldn't make any more
difference to anything than if you disagreed. You just aren't
discussing anything which could have any practical implications.
If one of you makes and assertion, and the other negates that
assertion, what is it which determines which one of you holds the
assertion or its negation true in the first place? There seems
little concern for any scientific evidence, just personal
preference.
That's why these exchanges are so futile.
Contrast this with what I have been advocating and requesting.
I started with practical problems in an area of applied behaviour
science. I then search PSYCLIT, ie the abstracts of all research
literature in psychology, reviewed what I found in conjunction
with the problem I encountered and then put the material up
hoping that others familiar with the literature in this area (and
perhaps others) might read what I have summarised and contribute
further.
What's wrong with classic philosophical discussion is precisely
what seems to happen in so many of these threads. They have no
consequence, and invariably end up in little more than folk
exchanging personal abuse.
--
David Longley
Neil Rickert
>That's why these exchanges are so futile.
>Contrast this with what I have been advocating and requesting.
Ah, yes. You have been advocating that humans should become mindless
automatons, whose only contact with reality is via the automatic
citation machine.
Neil Rickert
In <58d204$5...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>In article <589rt1$l...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>>Language more fine grained than pain? You would have to explain how
>>that could work. It would seem to say that I could use language to
>>distinguish between a severe pain and a very severe pain, whereas I
>>couldn't actually feel any difference in my experience with the
>>pain. In that case, what is the linguistically represented
>>distinction telling us?
>I should have said: language, in particular scientific language, makes
>possible representation of finer grained information than could possibly
>be carried by sensory states like a feeling of pain. For example,
>localizing the damage precisely, measuring the wavelength of light, etc,
>things which can't be distinguished by our normal sensory means.
There would still seem to be something wrong with this, since we can
only use scientific language by using sensory states. It seems to me
that you me experiential states, rather than sensory states.
>Not for nothing did Leibniz call perception a "confused" representation
>of the structure of reality.
Perhaps this is just a matter of Leibniz being confused about
perception. We have structured ways of talking about reality, as in
our scientific theories. But I think it a mistake to assume that
the structure in our scientific theories is the structure of
reality. Perception requires only a suitably structured way of
dealing with reality to suit the requirements of the human organism.
We should not think it a failing of perception if the way it
represents reality is not identical with the way our scientific
theories represent reality.
>Of course if we are talking about representing the headache itself -- not
>some sort of cause of the pain -- the issue is different.
I had thought that was what you were talking about when you brought
up the question of headaches.
>>>This raises some very difficult issues about the individuation of
>>>intentional contents. In some intuitive sense we want to say that what
>>>I express by "My right arm is raised" is the same content that you
>>>might express by saying "Anders' right arm is raised".
>>As long as you depend on "in some intuitive sense", that sounds ok.
>>Beyond that, I can't see that it much matters. I don't see how we
>>could ever be sure that what I mean by "Al Gore's right arm is
>>raised" is identical with what you mean by "Al Gore's right arm is
>>raised." It is not even clear that it is meaningful to talk about
>>whether the meanings are identical. I think it highly unlikely that
>>we have identical concepts corresponding to most of the words we we
>>can share.
>I suppose objective discourse on a common topic is impossible, and
>there is no point ever arguing with anyone about anything, since all
>meanings are ultimately solipistic, private and unshareable and on your
>view.
Why do philosophers jump to such absurd conclusions. You are making
the same mistake as those who criticized Kuhn's incommensurability
thesis. You are making the same mistake as Putnam, in his "Twin
Earth" argument.
I suppose you base this reaction on a fallacious theory of language.
It is at least arguable that given two formal languages using the
same set of formal symbols, if the symbols of the two languages are
not identical in what they represent, then there can be no common
communication between users of those two formal languages. But
natural language is not a formal language, and natural language words
are not symbols. Natural language communication is quite possible
even if concepts are not identical.
It might be that your concept of time is based on your clock which
happens to be 1 minute fast, while mine is based on my clock which is
2 minutes slow. In that case our concepts of time would not be
identical, but we could still communicate about time. At worst, our
communication might be limited to an accuracy of no better than 3
minutes error. But that is quite adequate for most communication
about time.
You raised the question about individuation of concepts. My point is
just that we can never expect content to be the same, but we can
communicate provided that my content and your content is sufficiently
similar.
>But the (not very formal)idea of a common content should be treated
>as a kind of construction out of the idea of a topic of rational
>agreement or disagreement. And to abandon the latter idea is to cease
>to be a thinker.
But my whole point is that we can have common content, without having
the same content. If there is sufficient similarity, then there can
still be a large amount of common content. Provided our
communication rests primarily on the common content, rather than on
the disagreements, there should be no serious problems.
>>>I say is that we must make communicability between speakers via use of
>>>the public language (shareability of content) the foundation or
>>>starting point for any philosophizing about mental representation, and work
>>>backwards from there to whatever we can say about an individual mind
>>>or thinker as secondary.
>>And yet the picture you have just painted does not assign a role to
>>connecting the mental representations with the world. So how does it
>>argue against solipsism?
>But intentional contents *are* modes of directedness at the world.
But then you should not make language the starting point. Rather,
you need to start with the processes that provide that
intentionality. To make language the starting point is to gloss over
the problem of intentionality, and pretend that it does not exist.
In effect you argue for a solipsistic world of language, detached
from reality. But whenever challenged you utter the magic word
"intentionality", and with a flourish of your magic wand you expect
all of the difficulties to go away. But they will not go away until
you are willing to analyze the processes which provide that
intentionality. For you cannot fully understand what is
intentionality without examining those processes.
>For example, the content expressed by "Vice President Al Gore's right
>arm is now raised" is directed towards the state of a flesh and blood
>appendage in (I assume) Washington D.C. It answers for its correctness
>to that reality. The intentional content I might express by pointing
>at the wastebasket in my office and saying "that is empty" is directed
>at the state of the wastebasket in public space.
No, that won't do. You cannot explain how words connect to reality,
by simply explaining how words relate to other words. That still
leaves you in a solipstic world where there is only language.
>The world is what we use our words to talk *about*.
Right. And as long as you make words the basis, then it doesn't
matter if that world actually exists. That is why I consider your
theory to be solipsist.
> So there is
>no problem about the connection with reality.
There is no problem, if you are happy with reliance on magic.
>>>>But on your view of knowledge, the training could never get off the
>>>>ground. The training can only occur if information is received. But
>>>>on your view of information, this seems not possible until after the
>>>>training has taken at least some partial effect.
>>>I would have thought you have the same problem.
>>No, because I don't see training as the basic element. We are able
>>to learn even in the absence of any training.
>Some things, of course. I don't think most of us could learn to use the
>intellectual tools -- e.g. new symbolic notations, like that of the
>calculus -- without training. At best we could say a few rare geniuses
>might hit upon them on their own, but even they started somewhere, with
>some intellectual or research tradition.
I do think that self education is different from training. One can
be self educated without the imposition of a training regimen, yet
not have to reinvent everything by oneself.
>>>Whereas I am speaking of experience or data only in the "thick" sense,
>>>as the conceptualized result or upshot of such processing --
>>>"processed" data.
>>However there may be much processed data which is still not
>>conceptualized. So I think your emphasis on conceptualized data
>>misses most of what is happening, so that you finish up with a theory
>>which is largely disconnected from reality.
>OK, so is it fair to say you are working with a trio: "raw"
>(unprocessed) stimulation, *detected* information, and conceptualized
>stuff. Where conceptualization is supposed to effect a kind of
>selection or abstraction from the *detected* information, which is too
>rich in content to be captured entirely in the net of language without
>loss. Correct?
I suppose that is roughly it. I think, however, it is a little too
simple. It would seem to me that stimulation is, by definition,
detected, for otherwise it would not stimulate. There can be
information which is detected, but the content of which has not yet
been discovered. It seems to me that you leave this out of the
equation. You might think of the work of the code breakers at NSA.
They work with information that has been detected. But until they
have succeeded in breaking the code, they still have not determined
what is the content of the information. The methods they use to try
to break the code are very much information based.
>Then it is still true, as I said, that there is a parallel between
>the situation with respect to your detected information and my
>conceptual information, namely both require cultivation to obtain,
>which cannot itself be accounted for in these terms.
But, unlike you, I am not willing to just take this 'cultivation' for
granted. The processes involved need to be investigated.
>I his Locke lectures _Mind and World_, John McDowell alludes to a
>similar sort of position in the work of his former colleague Gareth
>Evans' work (which McD, who edited Evans' posthumous book, generally
>regards as brilliant). McDowell argues that it still fails under the
>critique of the Myth of the Given, since the transition from information
>to an epistemic state like a belief cannot account for the latter's
>standing as justified.
As I have often indicated, I think the whole epistemic game of
justification is little more than a word game played by
philosophers.
>Of course this polemic depends on an epistemological orientation towards
>the philosophy of perception, which you have rejected. But I have
>never understood the basis of the rejection; I think it is a neutral,
>mundane, everyday fact that people have cognitive attitudes concerning
>the world which can be assessed in epistemic terms.
The question is whether that assessment in epistemic terms is
anything more than a "Just So" story. The type of arguments about
justification that I see in treatises on epistemology have little to
do with the ways scientists develop their theories, for example.
Perhaps epistemology puts justification in the wrong place. What
should matter is whether we are justified in our decision making.
But that has only a weak relationship to ascribed beliefs and whether
they are justified. The stories told by epistemologists in effect
make the justification of a belief an all-or-nothing affair. But a
belief might be well enough supported for justifying some types of
decisions, but not well enough supported for justifying its use in
other types of decisions.
>>> In
>>>this sense, the preconceptual given is not information *for* me, and
>>>processing it is not an activity I perform.
>>You couldn't even keep your balance while walking down the street,
>>were it not for that which you refuse to count as being information
>>for you. You don't normally control your balance by means of mental
>>and epistemic feats.
>Yes, pretty obviously that is not a mental feat and the relevant
>information is not obtained by me. So what?
Then surely that makes you a dualist. You body walks and maintains
balance, with the use of a great deal of information. Walking would
not be possible without that information. If that information is not
obtained by you, then when your body walks, it must be that you are
not walking. This seems to make you a detached spirit residing in
the near vicinity of the body, but somehow separate from it.
>>>On the other hand, if you require that this information be *detected*,
>>>then you are very close to my view, since then you are talking about
>>>information as something which must be the *result* of some sort of
>>>processing which requires "education of the senses".
>>But it seems to me that your view of information excludes the
>>possibility of explaining how the senses could be educated.
>No, it only precludes explaining it at the level at which we speak
>of information as detected.
>I have been trying to suggest that you have a parallel problem.
I think you have confused yourself by creating the detected/not
detected dichotomy. If I pick up a newspaper written in Sanskrit, I
can detect that it has information, but I would not have any idea as
to what is that information. The problem is not in whether the
information is detected or undetected, but in our ability to extract
the information from what is detected.
>>>But rules or conceptual norms are not inside the head. One might better
>>>say they live outside in the world, in such things as customs or
>>>practices or techniques. What's literally inside the head is circuitry
>>>subserving the ability to engage in these things, a topic of little
>>>relevance to philosophy of mind, I think.
>>Well that is why philosophers so often sound like dualists or
>>solipsists. Heads, brains, etc, are dispensible, as are trees,
>>birds, flowers. All we need is a purely abstract world with detached
>>concepts floating around.
>Unfair. Compare the custom or practice of using a signpost (one of
>Wittgenstein's favorites). Well in one sense a signpost is concrete --
>it's nothing over and above a hunk of wood, after all. But the wood considered
>all by itself seems dead, inert, devoid of significance.
Sure. But to continue the metaphor, the words on the signpost are
abstract, detached from reality and useless, unless the signpost is
there to hold them up.
>Now the signpost is thoroughly material. But where, Witt might lead us
>to ask, is the *use* we make of the signpost, according to which it
>*points*? We want to reply with terms like "convention", "custom",
>"practice", "institution", "technique (of operating with)". That's
>find, but what sort of entities are *these*? In some sense they are
>abstract. Yet they are thoroughly embodied in activities with material
>objects, a bit like Aristotle's forms in matter.
>I think this model points the way out of the worry you hint at, about
>"a purely abstract world". We can say that concepts too are embodied
>in uses we make of linguistic tokens. Then they do not "float around" in
>complete abstraction from what human beings do in the spatio-temporal
>world with signs. But they are not material objects or attributes.
Sure. But all of this still depends on you waving your magic wand
and saying the word "intentionality" at the right time, perhaps with
a little puff of blue smoke for effect. You still have given no
explanation of what is involved in this embodiment of concepts in
their uses, except in a question begging way (the uses themselves are
the embodiment in their uses).
>>Unfortunately, there is some ambiguity in the expression "can be
>>informed about". If it refers to the information I currently have
>>available to me, of course this is right. If it refers to the future
>>possibility of acquiring information, then it is wrong, for I may be
>>able to change my representation system so as to allow it to accept
>>that future information.
>And here is the parallel, for I can say one can change one's concepts.
Right. But then you also argue against conceptual relativism, and
you cannot give a procedure by which concepts are changed.
>>Either raw information is the basis, or Berkeleyan ideas are directly
>>inserted into souls by some God.
>Well Berkeley's concepttion of ideas as images is confused.
Sure. But that was several hundred years ago. It was a bold attempt
for its time.
>But again, I thought on your view too the child is not informed about
>raw information. So it is not a basis the child can go on.
The child is not consciously informed about raw information. But why
should we restrict our attention to that which is conscious?
>>This depends on what you mean by "the child". If what you mean, is
>>the ghost in the machine, then it may be reasonable to say that the
>Of course I don't mean any ghost in the machine. The child stands to
>his or her functional parts in a fashion similar to Ryle's University
>and its colleges -- one is the organized whole comprised of the latter.
>The child is certainly no ghost -- he or she is the creature you see
>before you. But still the child is not a subsystem of the child, but
>the whole.
Yet the child's body walks and maintains balance using information
which, by your claim, does not inform the child.
>>child is never informed about raw information. But if you mean the
>>flesh and blood child, whose sensory organs are regularly picking up
>>that raw information, then your assertion is surely wrong.
>I would have thought you also agree that just because the sense organs
>are stimulated does not mean the child is informed.
This only makes sense if you mean "consciously informed". And that
seems to amount to treating the conscious person as a spirit detached
from the body.
>I wonder if you are dropping my attempted distinction between "raw"
>information -- there at the surface whether or not it is detected --
>and detected or obtained information.
Perhaps your distinction was never very clear. Surely if the raw
information is picked up by sensory organs, that should count as
being detected.
> I think you are referring now to
>information that is detected by a subsystem of the child, even if it is
>not detected by the child itself (as information relevant to regulation of the
>autonomic nervous system might be detected in the cerebellum, without
>being detected by me).
But if the child is the whole, as you seemed to say above, then we
should say that information detected by a subsystem is information
detected by the child as a whole. Yet apparently you want to deny
this.
>>Again, all you can mean is that the ghost in the machine never
>>detects that information (about acoustic features) at all. But the
>The person is not a ghost, but the flesh and blood creature
>you see before you.
>>flesh and blood person behaves in such a way as to indicate that the
>>information was detected and acted upon.
>But not to indicate that the person got the information.
Then the person must not be the flesh and blood creature, contrary to
what you just said. It seems to me that you are trying to make
dualist distinctions, while claiming that you are not making those
distinctions.
>>A tree is immersed in patterned energy which impact its sensory
>>surfaces (its leaves). The distinguishes us from trees, is not that
>>we are also immersed in mere words and phrases. Rather, it is that
>>we are not merely immersed in the patterned energy. Unlike the tree,
>>we have the ability to seek out new sources of patterned energy, and
>>to modify the manner in which the patterned energy impacts our
>>surfaces. As Gibson insisted, we are not passive, but active
>>participants in the pickup of information.
>I can agree. However I am not quite sure a child learning language
>is exactly an active participant -- I expect the child will acquire the
>language whether it seeks to or not, owing to the sub-personal
>processes in its brain.
Well perhaps the detached disembodied spirit of the child is not an
active participant, but the attached neural and sensory system is
very much an active participant.
>>And you thereby appoint us gods, different from all other creatures.
>>And our cognition is then necessarily irreducible, for there are no
>>physiological differences which would make us into gods rather than
>>animals. But surely this is dualism, for the difference which makes
>>us gods must be the presence of a ghost in the machine.
>I don't see any problems with my form of conceptual dualism. I certainly
>don't see any need for ghosts as immaterial components. It is just the
>*description* of what we are doing when we obey a rule, or follow a
>rational norm, that is irreducible to physiology.
Your description is irreducible, because you insist on erasing from
it anything that could be involved in a reduction.
>To take one example, physiology can't explain what it is to be a
>*correct* mathematical proof, and it similarly won't explain what it is
>for a mathematician to be able to recognize or strive after this
>property.
There is no need for physiology to explain what it is to be a
mathematical proof, since that is not a physiological question. It
is not at all obvious that a sufficiently detailed physiology could
not explain what it is for a mathematician to strive after correct
proofs. It seems to me that there ought to be a physiological aspect
to that question.
>>I have no such problem, for I have available to me information based
>>processes which will modify the detection ability to accomodate
>>previously undetected information. Perhaps you also have such
>Not clear on this: how does the undetected information have an
>influence if it is not detected?
Again, 'undetected' is the wrong restriction. The signal is being
detected but the information is not being extracted from that
signal. Since the signal is being detected, it can have in
influence. Because we are active participants in perception, we can
search for other signals which are correlated to this one, as a way
of attempting to recover the information from the signal.
>>When Einstein wrote of his general relativity, he presented some
>>thought experiments. In one, a person was travelling in an elevator
>>car, but the car could be in an elevator shaft or could be moving in
>>free space. Why did Einstein use this method for describing his
>>theory? My view is that he had no language in which to represent the
>>information he wanted to convey. Therefore he needed to paint a word
>>picture, which would provide the raw information his audience could
>I'm not getting the difference between "painting a word picture" and
>conveying information through language.
The thought experiment purports to be a fictitious story. That is, it
uses language to directly convey misinformation.
>>>But one does not require information as the *basis* for the
>>>discrimination. Rather the discrimination itself *is* the basis, it
>>>*is* detetection of information.
>>But if there is no raw information, then the discriminator cannot be
>>created. Once we have a stream of raw information, we can attempt to
>Of course it can be created, but it can't be created by first detecting
>what ex hypothesi went undetected. I.e. the move from not detecting
>the information to detecting the information can't be accounted for
>in informational terms.
Again, this seems to be a confusion based on the unfortunate use of
the word 'detection'. Astronomers could not have known anything
about the information from pulsars until those signals were
detected. But when the signals were detected, the information in
them could still not be interpreted. They were originally called
LGMs (for "little green men"). It was clear from the nature of the
signal that they bore information, but it took additional research to
find the nature of that information. That additional research, it
seems to me, can be accounted for on informational terms.
>>>Not in the case of scientific theories. Yes indeed in the case of the
>>>child acquiring concepts, since that is not based on prior information
>>>detection *by the child*.
>>Presumably by this you mean that it is not based on prior information
>>detection by the ghost in the machine. Surely it is based on
>I wish you'd drop this nonsense about ghosts.
And I wish you would stop attempting to distinguish between whether
the information is available to conscious awareness.
>>information detection by the flesh and blood (including neurons).
>No doubt our neurons detect lots of information that we don't. Without
>this we would die.
This sort of view seems to be the basis of your dualism. You want to
detach what is conscious from everything else, and use that detached
portion to form a philosophy of mind. But it seems to me that such
an approach is necessarily solipsistic. For the intentionality that
connects our conscious thoughts to reality involves that information
that you allow for the neurons, but not for us. So we are left
floating around in an abstract world, divorced from our bodies and
from reality. It seems to me that we must understand the whole
person, including neurons, first. Only then can we hope to form a
coherent theory of consciousness (if it is important to do so).
> For example I can now look around and so
>enjoy consciousness of my computer, desktop, and office. I can grant that
>these states might be realized in neural material without supposing that
>neuroscience explains their intentional attributes (the relationship to
>my office).
It is not up to neuroscience to explain intentional attributes. But
it is up to science, including neuroscience, to consider the complete
picture of information flows. Before that science is complete, most
of what is said about intentionality is premature and should be a
place for skepticism.
>It seems to me that if you never get to the point where you
>consider states of consciousness, then something must be left out.
It seems to me that there is no sharp distinction between what is
conscious, and what is not. I am not attempting to leave anything
out, but I will not buy into bogus distinctions just to avoid the
accusation.
>>>That make it *difficult* to communicate. But if private, non-conceptual
>>>information were required to be transferred, it would be impossible to
>>>ever have any confidence that one had communicated.
>>I don't see why? Those training guide dogs for the blind seem to
>>have confidence in their ability to assess whether the dogs have
>>acquired the abilities. Yet on your language-based view of the
>>conceptual, I don't see how you could say that this was based on
>>other than non-conceptual information.
>Yes, I think skills as such are not conceptual. Training a guide dog is
>not exactly transfering information to the dog. Perhaps something like
>non-conceptual information that this is right, that is wrong, this is
>better, and you've been a good doggie does get transmitted to the dog
>in the course of the training.
If a guide dog's abilities were no better than that suggests, they
would be of little use.
>>> I couldn't compare
>>>my preconceptual given with someone else's to ensure it's the same, I
>>>can only check the manifestations.
>>It is only an illusion that you can compare your conceptual
>>information with somebody else's conceptual information. For you
>>cannot compare your concepts, and thus you cannot be certain that
>Why not?
How do your compare your concepts?
David Longley
> In <850057...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley)
> writes:
>
> >That's why these exchanges are so futile.
>
> >Contrast this with what I have been advocating and requesting.
>
> Ah, yes. You have been advocating that humans should become mindless
> automatons, whose only contact with reality is via the automatic
> citation machine.
>
>
Neil......just read some of the literature. You are making the
mistake of using your own common-sense (folk psychology) as a
reference system. Not only does that only offer a rather limited
conception of the world, but, taking remarks such as the above as
representative, it's doing you very limited service....stick to
the mathematics...
--
David Longley
Peter_...@uk.ibm.com
In <589no9$m...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>In article <588rm7$n...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
>>In <587ek0$d...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>>>
>>>Once one appreciates this, I don't see the motivation for pursuing
>>>reductive physicalism of any variety, including causal role
>>>functionalism.
>>
>>What? Once one appreciates that reductive explanations *might*
>>be false, motivation for seeking them is removed? I just don't see
>>how to get from the *possibility* of an investigation failing to the
>>idea that such investigations are unmotivated. I would have thought
>>the possibility that the investigation might succeed is ample motivation
>>for the investigation.
>
>Ditto for the project of transmuting base metals into gold, I guess.
Since this project has both been very fruitful (giving rise to modern
chemistry) and has also turned out to be possible (it is routine that
nuclear power stations produce small quantities of gold), this
would seem a particularly unfortunate analogy.
>Of course you are strictly correct that it *might* be possible (though
>I doubt it). As I see it, the orientation that produces the hysterical
>urgency behind the philosophical movement I detest ("naturalism")
If your position amounts to little more than an unhappiness with
'hysteical urgency', then what not simply say that in the first place?
>... is
>NOT: "hey, it's crazy, but it just might work, so let's devote our
>lives to trying". It is rather something like "it MUST be possible.
Well, perhaps it must. I simply don't think we know enough to
be able to say, yet.
>The
>only alternative is something horrible and mysterious like Cartesian
>substance dualism or vitalism in biology. Why if it's not possible, the
>whole 'scientific world picture' collapses, and we would be forced to
>believe in spooky immaterial souls and the like".
>
>Whereas I take it that for Wittgenstein, Ryle, Anscombe, Strawson,
>others, the irreducibility of the intentional per se is simply a
>non-issue, a don't-care that is no more threatening to science than the
>irreducibility to basic physics of "is a table", "can get you on the
>subway" or "is hunting its prey". The real issue is the *mystification*
>that comes from the failure, when philosophizing, to consider
>intentional activity as situated in its proper context.
There does seem to be an on-going problem with accounting
for the causal relevance of non-reductive accounts. Which is
why, I think, reductionism will remain attractive. If the causal
story can be told in terms of property dualism, where the
dualistic property associated with, say, qualia is somehow
causally effective qua dualistic property, then it'd be a help.
Since I haven't had any such account, it is premature to consider
this a closed book just yet.
>I think once the source of the urgency -- either we reduce it,
>eliminate it, or else some great intellectual disaster follows -- is
>exploded, I think it is fair to say that the motivation for pursuing
>the reductive project is seriously undermined.
Modulo the success of this project, something as yet unfulfilled, I'd
have to agree.
>I don't think many
>people would try to claim that a computer might be a subject of mental
>states unless they believed they *had* to for some other reason, that
>it was intellectually *obligatory*, not optional, given some sort of
>due respect for the achievements of science.
Don't understand what you're talking about. Are you talking about
a robot? If we had robots who could behave in ways which gripped
us as the sort of ways peoplebehave, I think *most* people would
consider them to be the subject of mental states - it would be rejected
by those who have some religious (or philosophical) axe to grind.
>Of course we *may* still
>pursue reductive projects simply because we can, but that is not, as I
>see it, the root impulse behind philosophical "naturalism". (I should
>add that I take the human potentiality for coming to engage in
>intentional activity to be natural as it stands, without reduction to
>something else.)
Well, of course not. Philosophical naturalism is an 'ism'.
However, reductive strategies are typical in science and in many
area of investigation. These enquiries can take place with or without
any expectation of success. Anyhow, this isn't what you said, although
it makes more sense. Probably the root of naturalism is the recognition
that it is very difficult to understand how the causal story is to be
told without some sort of reductive linkage.
>>>Reductive physicalism is false, as even Jerry Fodor
>>>argued, but non-reductive physicalism (everything else "weakly
>>>supervenes" on the physical) is nearly toothless.
>>
>>I'd be interested in discussing this dichotomy: on the one hand,
>>a certain position; on the other hand, something 'nearly toothless'.
>>It doesn't sound like a plausible dichotomy to me. It sounds like
>>a dichotomy specially constructed for a purpose: rubbishing
>>physicalism.
>
>Right you are. I should admit it is not toothless in that it does seem
>to rule out immaterial souls and vitalism. But that is not the same as
>saying that the predicates and explanations of biology, say, get
>reduced.
Well I'd say that Darwin's theory gives a naturalistic account of
the origin of species. Period. Now notice that I'm not saying that
Darwin gives necessary and sufficient conditions for what a Jaguar
is, nor am I saying that Darwin gives necessary and sufficient
conditions for the cicumstances under which Jaguars evolve.
But I think Darwin gave a naturalistic account of the origin of
species nevertheless.
So I'd say that what Darwin did was not only to naturalise a
certain area of science, but also to introduce a novel (?)
form of naturalisation. And the move Darwin made is to step
beyond the detail to the system as a whole. Instead of focussing
on the fitness of the bee for its environment and trying to explain
*that*, what Darwin did was to focus upon fitness and change
per se. Now I think this is a perfectly acceptable naturalising move,
one that we can exploit in other areas too.
That is, the argument that we can't give naturalistic necessary
and sufficient conditions for "is a table" is neither here nor there
from the point of view of success at naturalising the mental.
What is needed is not anything so ridiculous, but something
deeper but probably easier: an account of why and how such
classifications occur. That is, an account, not of the detail, but
of the system as a whole. I would think that the scope for such
approaches are pretty good, if for no other reason that they simply
don't seem to have been tried.
>>Couldn't we just say: the versions of reductive physicalism we
>>currently have are false and the less constrained versions we
>>have don't seem to have a stopping point. Physicalism needs
>>new insights in order to make progress?
>
>In sober moments I would agree. My problem is I don't really know how
>to formulate physicalism in a way that has bite against the
>irreducibility of the intentional (or the normative, or the social, or
>the biological). I do admit that this is a philosophical project.
Fair enough.
>John Haugeland wrote a nice paper called "Weak supervenience"
>(published in American Philosophical Quaterly, I think around 1978, but
>I could be wrong) in which the requirement of supervenience upon the
>physical was argued to be very weak indeed.
Supervenience (particularly global supervenience) is no good
for anything much. Some people seem to have latched onto it
as saying something important. Not me.
>In conversation, he has
>suggested he thinks the version in that paper was too weak, but I am at
>a loss how to strengthen it.
>
>As I see it, it is perfectly possible that when the mathematician Mary
>inscribes a proof of a theorem on a paper -- even, perhaps, if she is
>"intuiting the truth of the Godel sentence", it is *also* possible that
>every elementary particle in the light cone of this activity -- in her
>body and in the air and in the pencil and paper and around the moons of
>Jupiter -- might be found to have evolved in accordance with the basic
>laws of physics. So physical science might explain this massive
>evolution, but not use the concepts of "proving a theorem".
Because of your use of the word "explain", I agree with you.
>We can see
>what's going on *as* a proving of a theorem or *as* a complex physical
>event, but there is no reason we can get the two sorts of facts in
>registration (just as we can treat an artwork as an aesthetic or
>cultural object, or as a physical object when we are interested in
>moving it).
Or then, again, perhaps there is a reason we can get the two sorts
of facts to register.
Appealing to aesthetics and ethics in the context isn't likely to
be treated kindly by me, since they merely beg, in different
contexts, all the same issues.
>In particular, I take it basic physics seeks explanations which subsume
>in the same fashion incorrect moves in the proof, crossing outs,
>back-trackings, furrowing the brow, crumpling the paper and tossing it
>in the wastebasket, etc. It as it were is blind, or cares nothing for
>the concept of a person's proving a theorem (or struggling
>unsucessfully to prove a theorem) and the sort of behavior that
>manifests such an attempt.
Well, I'm not sure this a credible view of basic physics, if basic physics
includes thermodynamics. As soon as we include the question of
reversibility and irreversibility, we talk in a very different way.
It could be, for example, that the sort of behaviours we tend to associate
with rationality tends to be associated with reversibility, or at least, with
less irreversibility than otherwise.
I am reminded here of the cooling of beams of particles stored in a
D-ring. These particles travel at near the speed of light: but they
do travel on the circumference. They are cooled by measuring at
one point and then sending information across the diameter, ready
to bump the particles into shape. Crucial to this process is the fact
that it takes longer around the circumference that across the diameter.
This thermodynamic process also relies, presumably, on the fact
that the information is *correct*. Indeed, Maxwell's Demon does seem
to forge a link between correctness and reversibility or near-reversibility.
What I'm pointing out is that physicalism has more stings to its bow
than the rather sorry tale above would seem to imply. I find this
rather strange because Anders has told the same tale before and I have
pointed out that physicalism is a mansion with more rooms before now.
But, apparently, Anders is happier to trot out the same tired old
nags than to adjust anything he might say. We seem to be involved
in 'memoryless' exchanges as a result. Each interaction starts from the
same place it started from last time. Why this is, I simply don't know.
Better ask Anders.
At any rate, until we know better the relationship between levels
of language and how they're formed, I still don't see how you
intend to show what this 'blindness' amounts to. That is, is it
to do with physicalism per se, or is it merely to do with a particular
blinkered view of what 'basic physics' consists of? How do you
propose to answer such questions, if not by understanding better
the relationship between these languages and their levels?
>That requires use of a different vocabulary
>even for the very description of the explanandum, it seems
>to me.
As I say, basic physics comes in two favours: forces and fields
on the one hand and thermodynamics on the other. Don't forget
that when one is talking about physicalism, they're both relevant.
If we knew how our vocabularies were formed, perhaps you would
do better at showing some sort of sharp divide between the
language of basic physics and the rest. However, since this isn't
something we know much about, we're not sure whether the
boundaries you wish to draw have any greater significance than
any other mark in the sand of philosophical speculation.
>So it might even be that Penrose is right about our mathematical
>abilities (although I am skeptical) but wrong that new *physics* is
>required. Perhaps the existing physics is complete within its own
>sphere, nevertheless.
I think it is clear enough that thermodynamics is still an active area
of theoretical research. The idea that thermodynamics has come to
an end (or might as well have for our purposes) is plainly false. New
insights continue to occur. What is now an increasingly interesting
project is the attempt to apply thermodynamic concepts to *macro*scopic
structure. The traditional theory of thermodynamics cares little for
such entropy, since it tends to be numerically very small in comparison
with atomic entropy. However, the attempt to ignore macroscopic
entropy altogether appears to be a mistake. This gives rise to a quite
different notion of thermodynamics than previously thought.
So what I have is three major concerns:
1. Certain tests for physicalism ("is a table", etc.) that Anders
fixes on probably aren't important to physicalism at all.
2. The relationship between physicalism and reductionism
viewed as giving necessary and sufficient conditions as
finite formulae is bogus.
3. Anders' view of what counts towards physicalism seems
to be unduly limited.
All in all I see a rather hasty dismissal of something which has a
great deal of kick left in it.
Cheers,
Pete Lupton
Anders N Weinstein
In article <850057...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
|Even if you two agreed on something, it wouldn't make any more
|difference to anything than if you disagreed. You just aren't
|discussing anything which could have any practical implications.
|That's why these exchanges are so futile.
I seem to recall seeing the wish that this thread not degenerate
into meta-criticisms.
|what seems to happen in so many of these threads. They have no
|consequence, and invariably end up in little more than folk
|exchanging personal abuse.
Longley's scrupulous avoidance of personal abuse stands as a model
for us all.
Anders N Weinstein
In article <58frbh$o...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>In <58d204$5...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>>Not for nothing did Leibniz call perception a "confused" representation
>>of the structure of reality.
>Perhaps this is just a matter of Leibniz being confused about
>perception. We have structured ways of talking about reality, as in
No doubt he was.
>our scientific theories. But I think it a mistake to assume that
>the structure in our scientific theories is the structure of
>reality. Perception requires only a suitably structured way of
Fine with me -- talk of *the* structure of reality is a bit pompous, but
does not really denote anything, since there is no one structure.
>dealing with reality to suit the requirements of the human organism.
>We should not think it a failing of perception if the way it
>represents reality is not identical with the way our scientific
>theories represent reality.
It should be clear from what I've written that I agree whole-heartedly.
I tend however to think that judgements expressed using the concepts of
common-experience and those of scientific theories can often both be
objectively true. Pragmatists on the other hand tend to say that neither
is true in that sense, both are merely useful tools for this or
that purpose.
>>Of course if we are talking about representing the headache itself -- not
>>some sort of cause of the pain -- the issue is different.
>
>I had thought that was what you were talking about when you brought
>up the question of headaches.
Actually you were the one who implied that having a headache was a kind
of information detection. I was wondering what information it was, and
speculating it was something concering the neural cause of the pain.
>>>Beyond that, I can't see that it much matters. I don't see how we
>>>could ever be sure that what I mean by "Al Gore's right arm is
>>>raised" is identical with what you mean by "Al Gore's right arm is
>>>raised." It is not even clear that it is meaningful to talk about
>>>whether the meanings are identical. I think it highly unlikely that
>>>we have identical concepts corresponding to most of the words we we
>>>can share.
>
>>I suppose objective discourse on a common topic is impossible, and
>>there is no point ever arguing with anyone about anything, since all
>>meanings are ultimately solipistic, private and unshareable and on your
>>view.
>
>Why do philosophers jump to such absurd conclusions. You are making
>the same mistake as those who criticized Kuhn's incommensurability
>thesis. You are making the same mistake as Putnam, in his "Twin
>Earth" argument.
Of course I think these are good criticisms and good arguments.
Kuhn's formulation strikes me as a bit confused.
I am not sure how Twin Earth is relevant, that seems to be orthogonal.
Although I certainly am an externalist about intentional content.
>I suppose you base this reaction on a fallacious theory of language.
...
>It might be that your concept of time is based on your clock which
>happens to be 1 minute fast, while mine is based on my clock which is
>2 minutes slow. In that case our concepts of time would not be
>identical, but we could still communicate about time. At worst, our
>communication might be limited to an accuracy of no better than 3
>minutes error. But that is quite adequate for most communication
>about time.
I guess there are two issues here -- the setting of the origin of the scale,
and the precision with which we measure time. Certainly if the context
calls for determination to within the minute, than we will want to translate
each other's time measurements by a constant, after which indeed it is
no different than Farenheit and Celsius. It is a little hard for me
to imagine having any kine of a "concept of time" according to which different
watches could not by calibrated in that way.
>You raised the question about individuation of concepts. My point is
>just that we can never expect content to be the same, but we can
>communicate provided that my content and your content is sufficiently
>similar.
Ok, but how do we determine similarity of content?
The problem it seems to me is that reflecting on our differing standards
of precision seems to point to the idea of a more fine-grained background
language against which different procedures are assessed. As if that were
the more accurate representation of the structure of reality against
which our coarse-grained common-sense representations is compared so as
to determine similarity of content.
But I thought you wanted to avoid such an assumption.
>>But the (not very formal)idea of a common content should be treated
>>as a kind of construction out of the idea of a topic of rational
>>agreement or disagreement. And to abandon the latter idea is to cease
>>to be a thinker.
>
>But my whole point is that we can have common content, without having
>the same content. If there is sufficient similarity, then there can
>still be a large amount of common content. Provided our
>communication rests primarily on the common content, rather than on
>the disagreements, there should be no serious problems.
I am wondering about the scale against which we rank overlap
of content. Doesn't this require commensurating the incommensurable?
>>>And yet the picture you have just painted does not assign a role to
>>>connecting the mental representations with the world. So how does it
>>>argue against solipsism?
>
>>But intentional contents *are* modes of directedness at the world.
>
>But then you should not make language the starting point. Rather,
>you need to start with the processes that provide that
>intentionality.
I am not sure I believe in "processes that provide the intentionality".
But in any case the processes that are relevant are not self-contained --
they are linked to expressibility in a language one has mastered.
> To make language the starting point is to gloss over
>the problem of intentionality, and pretend that it does not exist.
Depends what the problem is. In Wittgenstein's slab language game, the
term "slab" has a certain significance, which connects it to slabs in
the world (or actions of bringing slabs). It has this significance
before an initiate comes to be able to play the game. In coming to play
it, say to give the orders for his own purposes, a learner comes
to appropriate a pre-exisiting semantic significance.
For most of us, it is similar with "mass" or "photon" or "Al Gore". In
coming to master the use of this language we come also to be able to
think the thoughts expressible with these pre-existing terms, i.e. to
have intentional states that are about mass or photons or Al Gore. The
individual thinker can of course come to reshape the pre-existing public
language, but starts as a creation of it.
Now where exactly is this "problem of intentionality" again?
>In effect you argue for a solipsistic world of language, detached
>from reality. But whenever challenged you utter the magic word
>"intentionality", and with a flourish of your magic wand you expect
>all of the difficulties to go away. But they will not go away until
>you are willing to analyze the processes which provide that
>intentionality. For you cannot fully understand what is
>intentionality without examining those processes.
Not sure I understand this. I believe the intentionality of thoughts
is not reducible, but is not autonomous either, in that it is internally
related to linguistic expressibility. One can also speak of the impersonal
intentionality of the use of the word in a community, like the function of
bow and arrows in a society. I am not sure what *processes* one needs
to analyze to account for intentionality: do you mean the process of
using the words "it's raining" to say that it's raining? Or to express
the thought that its raining -- which is thereby connected to reality
(its true only if precipitation is falling outside).
>>For example, the content expressed by "Vice President Al Gore's right
>>arm is now raised" is directed towards the state of a flesh and blood
>>appendage in (I assume) Washington D.C. It answers for its correctness
>>to that reality. The intentional content I might express by pointing
>>at the wastebasket in my office and saying "that is empty" is directed
>>at the state of the wastebasket in public space.
>
>No, that won't do. You cannot explain how words connect to reality,
>by simply explaining how words relate to other words. That still
>leaves you in a solipstic world where there is only language.
But I didn't connect the words to other words -- I said they were used
to express thought contents which connect them with things in the world.
In a certain sense I believe the language makes the thoughts possible.
>>The world is what we use our words to talk *about*.
>
>Right. And as long as you make words the basis, then it doesn't
>matter if that world actually exists. That is why I consider your
>theory to be solipsist.
Not if it doesn't make sense to try to account for significant words
solpisistically.
>> So there is
>>no problem about the connection with reality.
>
>There is no problem, if you are happy with reliance on magic.
Where is the reliance on magic? In my dialect of English, the name
"Al Gore" is conventionally used to denote the current Vice President.
In coming to master this language (and related skills), I came to be able
to enjoy "Al-Gore thoughts" -- intentional states directed at that man.
Such a thought is indeed intrinsically directed at that man. But the
possibility of my being in such a state is not something magic, that could
just as well exist in a brain in a vat.
>I do think that self education is different from training. One can
>be self educated without the imposition of a training regimen, yet
>not have to reinvent everything by oneself.
Fair enough. As long as one is still responsible to pre-existing standards
of right usage, I need not emphasize training.
>The question is whether that assessment in epistemic terms is
>anything more than a "Just So" story. The type of arguments about
>justification that I see in treatises on epistemology have little to
>do with the ways scientists develop their theories, for example.
I am not mainly interested in the treatises, but in the mundane
everyday uses of the concepts. I don't see that you have given support
for anything like the dismissive attitude you express, i.e. you haven't
shown that we should just completely dispense with the idea of a belief
which may be assessed as justified or not.
>Perhaps epistemology puts justification in the wrong place. What
>should matter is whether we are justified in our decision making.
>But that has only a weak relationship to ascribed beliefs and whether
>they are justified. The stories told by epistemologists in effect
>make the justification of a belief an all-or-nothing affair. But a
>belief might be well enough supported for justifying some types of
>decisions, but not well enough supported for justifying its use in
>other types of decisions.
Fine, but the McDowell argument is based on the problem of
how non-conceptual experience could stand in *any* sort of genuine
justificatory relation to a conceptual state like a belief.
>>>You couldn't even keep your balance while walking down the street,
>>>were it not for that which you refuse to count as being information
>>>for you. You don't normally control your balance by means of mental
>>>and epistemic feats.
>
>>Yes, pretty obviously that is not a mental feat and the relevant
>>information is not obtained by me. So what?
>
>Then surely that makes you a dualist. You body walks and maintains
I would admit I am some kind of dualist, but only in the same way that
I am a dualist about economic properties and physical properties of
dollar bills.
>balance, with the use of a great deal of information. Walking would
>not be possible without that information. If that information is not
>obtained by you, then when your body walks, it must be that you are
>not walking. This seems to make you a detached spirit residing in
>the near vicinity of the body, but somehow separate from it.
Wouldn't it be better to deny that my body is the agent of my walking?
*I* walk, and when I do, my body moves. Sub-personal information
processing in my brain makes this possible: if not for that, I couldn't
walk. But the brain or neural circuits themselves don't do any walking.
>I think you have confused yourself by creating the detected/not
>detected dichotomy. If I pick up a newspaper written in Sanskrit, I
>can detect that it has information, but I would not have any idea as
But detecting *that it has information* is one thing, detecting
the information contained therein another. The latter is not detected.
>>>Well that is why philosophers so often sound like dualists or
>>>solipsists. Heads, brains, etc, are dispensible, as are trees,
>>>birds, flowers. All we need is a purely abstract world with detached
>>>concepts floating around.
>
>>Unfair. Compare the custom or practice of using a signpost (one of
>>Wittgenstein's favorites). Well in one sense a signpost is concrete --
>>it's nothing over and above a hunk of wood, after all. But the wood considered
>>all by itself seems dead, inert, devoid of significance.
>
>Sure. But to continue the metaphor, the words on the signpost are
>abstract, detached from reality and useless, unless the signpost is
>there to hold them up.
I don't see how this continues the metaphor.
Witt's response for the case of signs was: in *use* it is alive. I.e.
that is what breathes life into it -- a custom, practice or institution.
>>I think this model points the way out of the worry you hint at, about
>>"a purely abstract world". We can say that concepts too are embodied
>>in uses we make of linguistic tokens. Then they do not "float around" in
>>complete abstraction from what human beings do in the spatio-temporal
>>world with signs. But they are not material objects or attributes.
>
>Sure. But all of this still depends on you waving your magic wand
>and saying the word "intentionality" at the right time, perhaps with
>a little puff of blue smoke for effect. You still have given no
>explanation of what is involved in this embodiment of concepts in
>their uses, except in a question begging way (the uses themselves are
>the embodiment in their uses).
I didn't say tautologously that the uses are the embodiment of the uses.
I said the meaning derives from the use, and "uses" (like rules), although
abstract, are themselves embodied sorts of things. They are not separate
and independent of human activities with material objects like signposts.
Now there is one attempt to work out this line of thought that we
find in Quine (and Sellars and others). It suggests that the relevant
uses must be described in impoverished behavioristic terms, as dispositions
to emit noises or such like. I think that if one starts with such an
impoverished basis and tries to work back to meaning, the right response is:
you can't get there from here.
So I guess I think the intentional is irreducible in that we have to
use intentional vocabulary to adequately characterize how we operate
with meaningful signs and symbols. But I don't think that by itself
makes it abstract and detached -- it is only if you sever its connection
to human practices that it comes to seem mysterious and ghostly.
>>Of course I don't mean any ghost in the machine. The child stands to
>>his or her functional parts in a fashion similar to Ryle's University
>>and its colleges -- one is the organized whole comprised of the latter.
>>The child is certainly no ghost -- he or she is the creature you see
>>before you. But still the child is not a subsystem of the child, but
>>the whole.
>
>Yet the child's body walks and maintains balance using information
>which, by your claim, does not inform the child.
Better to say the child walks. Your phrasing suggests a kind of
sleepwalking to me -- "my body is walking, I just can't control it".
>>>child is never informed about raw information. But if you mean the
>>>flesh and blood child, whose sensory organs are regularly picking up
>>>that raw information, then your assertion is surely wrong.
>
>>I would have thought you also agree that just because the sense organs
>>are stimulated does not mean the child is informed.
>
>This only makes sense if you mean "consciously informed". And that
>seems to amount to treating the conscious person as a spirit detached
>from the body.
No more so than treating the University as a spirit detached from its
colleges etc..
>>I wonder if you are dropping my attempted distinction between "raw"
>>information -- there at the surface whether or not it is detected --
>>and detected or obtained information.
>
>Perhaps your distinction was never very clear. Surely if the raw
>information is picked up by sensory organs, that should count as
>being detected.
I thought more was required -- the information that Jones was out at
first base is perhaps implicit in the pattern of light stimulating
the retina. I would not want to say it is picked up by anything just
because the retina is stimulated.
>> I think you are referring now to
>>information that is detected by a subsystem of the child, even if it is
>>not detected by the child itself (as information relevant to regulation of the
>>autonomic nervous system might be detected in the cerebellum, without
>>being detected by me).
>
>But if the child is the whole, as you seemed to say above, then we
>should say that information detected by a subsystem is information
>detected by the child as a whole. Yet apparently you want to deny
>this.
I am not sure whether I want to say it. In some cases it seems to me
the information picked up by the person must be rather different in
kind than the information processed by sub-personal systems. E.g. information
only expressible using an expressions with a demonstrative, like the information
that the book is *there*. It seems to me neural circuitry cannot use
a demonstrative mode of presentation, so cannot have the sort of contents
that can only be expressed in that way.
In any case, even if some of the sub-personal information is also
person-level information, not all of it will be, e.g. information involved
in the regulation of the autonomic nervous system. So something more
is required than detection inside an organism's body for it to count as
detection by the organism.
>Then the person must not be the flesh and blood creature, contrary to
>what you just said. It seems to me that you are trying to make
>dualist distinctions, while claiming that you are not making those
>distinctions.
I am saying that the person might be informed of a content even if that
content is not explicitly represented by a sub-system of the person. I
do not see that this makes the person an immaterial entity. I do think
the level at which we talk about the information available to the whole
organism is governed by different criteria than the level at which we
analyze the organism into functional subsystems.
I think a similar case is this: the person might be said to believe that
the floor will support his weight when he unreflectively steps off the
elevator, in that if he didn't believe it, he wouldn't step so confidently.
This does not mean that information is explicitly represented inside his
body in any form -- merely that he has the dispositional capacity to wonder
about the solidity of a floor on occasions.
The point is just the standards governing ascription of content at the
two levels are different, so the contents can potentially come apart.
>>To take one example, physiology can't explain what it is to be a
>>*correct* mathematical proof, and it similarly won't explain what it is
>>for a mathematician to be able to recognize or strive after this
>>property.
>
>There is no need for physiology to explain what it is to be a
>mathematical proof, since that is not a physiological question. It
>is not at all obvious that a sufficiently detailed physiology could
>not explain what it is for a mathematician to strive after correct
>proofs. It seems to me that there ought to be a physiological aspect
>to that question.
I think the last phrase is a little weaselly -- I guess I could grant
the weak claim that there is a physiological *aspect* to the question.
>>>I have no such problem, for I have available to me information based
>>>processes which will modify the detection ability to accomodate
>>>previously undetected information. Perhaps you also have such
>
>>Not clear on this: how does the undetected information have an
>>influence if it is not detected?
>
>Again, 'undetected' is the wrong restriction. The signal is being
>detected but the information is not being extracted from that
>signal. Since the signal is being detected, it can have in
>influence. Because we are active participants in perception, we can
>search for other signals which are correlated to this one, as a way
>of attempting to recover the information from the signal.
OK, I am getting clearer on what you mean. I guess the
processes that "search for other correlated signals" and "attempt to
recover the information" are not really the processes that detect the
information itself. E.g. the process that leads to the development of an
affords-mailing detector is not itself an affords-mailing detector, but
some kind of analyzer or decoder of meaningless patterns.
I am still a little bit puzzled as to how you avoid the Fodor problem.
E.g. how do those learning processes 'decide' what information content to
ascribe to those meaningless patterns if they don't have the relevant
concepts to begin with?
>>I'm not getting the difference between "painting a word picture" and
>>conveying information through language.
>
>The thought experiment purports to be a fictitious story. That is, it
>uses language to directly convey misinformation.
OK. But the ability to represent non-actual states is one capability
of linguistic representation. We don't have to go to non-conceptual
information for that.
>Again, this seems to be a confusion based on the unfortunate use of
>the word 'detection'. Astronomers could not have known anything
>about the information from pulsars until those signals were
>detected. But when the signals were detected, the information in
>them could still not be interpreted. They were originally called
>LGMs (for "little green men"). It was clear from the nature of the
>signal that they bore information, but it took additional research to
>find the nature of that information. That additional research, it
>seems to me, can be accounted for on informational terms.
OK, but as always I want to distinguish between a move from one sort
of conceptualized information to another, which is fine, and the move
from pre-conceptual stuff to conceptual information, which is problematic.
>>>>Not in the case of scientific theories. Yes indeed in the case of the
>>>>child acquiring concepts, since that is not based on prior information
>>>>detection *by the child*.
>
>>>Presumably by this you mean that it is not based on prior information
>>>detection by the ghost in the machine. Surely it is based on
>
>>I wish you'd drop this nonsense about ghosts.
>
>And I wish you would stop attempting to distinguish between whether
>the information is available to conscious awareness.
But why? It is not as though you don't understand the distinction --
you seem to grasp it just fine.
>>No doubt our neurons detect lots of information that we don't. Without
>>this we would die.
>
>This sort of view seems to be the basis of your dualism. You want to
>detach what is conscious from everything else, and use that detached
>portion to form a philosophy of mind. But it seems to me that such
Not exactly detached, just a distinguishable aspect. Again, you seem to
have no problem recognizing the distinction. I think conscious life must
be viewed as one aspect to the life of an embodied organism in a material
and social environment, so I don't see that it could exist on its own, as
Descartes and other soullists thought.
>an approach is necessarily solipsistic. For the intentionality that
>connects our conscious thoughts to reality involves that information
>that you allow for the neurons, but not for us.
The genuine intentionality for us is different than the variety of
intentionality found the sub-personal level. Strictly, I would not
say neurons are intentional, they do not really have a perspective
on the world as we do.
> So we are left
>floating around in an abstract world, divorced from our bodies and
>from reality.
I think this is like saying that because my net worth is not a property
of my neurons, it must float around in an abstract world divorced from
our bodies and from reality.
>> For example I can now look around and so
>>enjoy consciousness of my computer, desktop, and office. I can grant that
>>these states might be realized in neural material without supposing that
>>neuroscience explains their intentional attributes (the relationship to
>>my office).
>
>It is not up to neuroscience to explain intentional attributes. But
>it is up to science, including neuroscience, to consider the complete
>picture of information flows. Before that science is complete, most
>of what is said about intentionality is premature and should be a
>place for skepticism.
Must disagree. We already operate with these concepts correctly a
million times a day, and they are not placholders for future scientific
theories -- they are rather like "table", "office", "country". What we
need is only a clearer view of the concepts we already work with.
>>It seems to me that if you never get to the point where you
>>consider states of consciousness, then something must be left out.
>
>It seems to me that there is no sharp distinction between what is
>conscious, and what is not. I am not attempting to leave anything
>out, but I will not buy into bogus distinctions just to avoid the
>accusation.
You seemed to grant that I can be conscious of my office in a way
I cannot be conscious of information about stimulation of my retina.
>>>I don't see why? Those training guide dogs for the blind seem to
>>>have confidence in their ability to assess whether the dogs have
>>>acquired the abilities. Yet on your language-based view of the
>>>conceptual, I don't see how you could say that this was based on
>>>other than non-conceptual information.
>
>>Yes, I think skills as such are not conceptual. Training a guide dog is
>>not exactly transfering information to the dog. Perhaps something like
>>non-conceptual information that this is right, that is wrong, this is
>>better, and you've been a good doggie does get transmitted to the dog
>>in the course of the training.
>
>If a guide dog's abilities were no better than that suggests, they
>would be of little use.
But the abilities can be much better than the information explicitly
transmitted to the animal in the course of training. Since neither the
teacher nor the learner need employ a representation of the behavior to
be achieved in order for one to teach it and the other to learn it.
>>>It is only an illusion that you can compare your conceptual
>>>information with somebody else's conceptual information. For you
>>>cannot compare your concepts, and thus you cannot be certain that
>
>>Why not?
>
>How do your compare your concepts?
By seeing how they apply the words in public circumstances, what
conclusions the draw from them and perhaps other things that they say
about them.
Jim Balter
Actually, this is even more true, in fact ultimately true, of formal
languages, by the very meaning of "formal". Whatever P and Q refer to,
(P -> Q) communicates a relationship between them, and (P and P -> Q) |- Q
communicates a general claim. In the empirical world, we communicate things
like "P" with some empirical referent E, and Anders seems to fear that that
could be taken as referring to F instead. But that would only follow if we
did not share many axioms based upon being in the same world together. The
non-solipsist has this faith, or working assumption. But Anders *fears*
solipsism, and so comes up with all sorts of philosophical absurdities in
order to avoid it. This very fear is a form of solipsism; it accepts
solipsism as an option that must be avoided at all cost. It is a bit like a
former Catholic who has to prove that God does not exist, lest God strike her
down for her sins. The true atheist recognizes that the notion of God is
incoherent or unworkable, and thus doesn't have this fear. A physicalist does
not *fear* dualism, but rather recognizes dualism as being unparsimonious.
Only a closet solipsist fears that meanings *could* be private and unsharable.
Once one overcomes this fear as a driving force behind one's philosophy (as
one must in order to truly be a philosopher rather than a sophiphobe) and lets
the chips fall where they may, formal equivalences and isomorphisms become
acceptable and worth considering. This gives us a way to understand the
errors in Putnam's Twin Earth scenario, the errors in Harnad's notion of
symbol grounding, and numerous errors by Searle. The only significant
difference between our world, our thoughts, our brains, and those that are
functionally equivalent to them, whether alternate worlds, instantiated
programs, simulations, whatever, is that the former are *of us*, which is in
some sense not significant at all.
--
<J Q B>
Jim Balter
I'm not sure what you are agreeing with here. Hopefully there is no tacit
agreement with Anders' mischaracterization of the physicalist enterprise. It
seems to me that the urgency and perception of threat is on his side. He is
the one who says his life would be fulfilled if only he could get people to
cease their search for reductive explanations (or the transmutation of base
metals into gold). He is the one that seems to foresee some intellectual
disaster if this project is continued; in fact fear, whether fear of
reductionism, fear of physicalism, or fear of solipsism seems to be Anders'
driving force.
Another aspect of Anders' mischaracterization is the category error he
continues to make in regard to what it is that is being explained. The issue
is not a search for scientific justification for claims such as "this is a
table" or "this is a way to get onto the subway" (although there are
topological considerations) or "this animal is hunting its prey" (although
science can show that some such claims are mistaken, which makes me suspect
that Anders' scientific scope really is extremely narrow, and suggests that
Anders is confused even within this category), but rather the issue is an
exploration and explanation of these *claims* *as human behavior*. The
distinction here is between use (when people say these things) and mention
(when we refer to people saying these things).
When Anders attempts to offer explanations at that level, he talks about
language games and norms and roles within communities and so on. This is all
fine and well, but doesn't answer fundamental causal questions concerning the
*mechanisms* at work in the individuals that make up these communities, or in
the environment at large (including the comprising individuals) that bring
about these communities. And when pushed, Anders says that he wasn't offering
*causal* explanations. But that's the point; causal explanations are what are
being asked for, and non-causal explanations are no *substitute*. But Anders
seems to think that, by supplying non-causal explanations, he has discharged
his duty, and has shown that causal, reductive explanation is a "non-issue", a
"don't care".
Yet, this is an AI group, and presumably a fundamental question of concern is
"how to build an AI". And any good engineer recognizes that one path toward
approaching this question is to consider how an HI is built. For this, the
engineer wants to look both inside and outside. She wants to consider the
language games, roles, and norms that are imposed on this thing, and, when
possible, she wants to look inside and see how it is (statically) made and how
it (dynamically) functions when playing this game. And she would be amazed if
this examination were to reveal sawdust or empty air inside. This would be
anything but a "non-issue" or "don't care".
But for some reason Anders seems to see such engineeering as a threat, and
seems to feel some urgency in dissuading people from carrying out the project.
>What I'm pointing out is that physicalism has more stings to its bow
The sting of arrows, or the sting of snapping strings? :-)
>than the rather sorry tale above would seem to imply. I find this
>rather strange because Anders has told the same tale before and I have
>pointed out that physicalism is a mansion with more rooms before now.
>But, apparently, Anders is happier to trot out the same tired old
>nags than to adjust anything he might say. We seem to be involved
>in 'memoryless' exchanges as a result. Each interaction starts from the
>same place it started from last time. Why this is, I simply don't know.
>Better ask Anders.
This is why I stopped talking to him. However, compared to some other
examples of proof by repetition we've seen hereabouts .... Even at his worst,
Anders' confusions are quite deep, and worthy of response.
--
<J Q B>
David Longley
In article <58i2sg$7...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
>
> Longley's scrupulous avoidance of personal abuse stands as a model
> for us all.
>
Perhaps it should!
I have remarked frequently that a number of people in this group
succumb to soplipsism or don't base what they have to say on
any empirical research evidence. That's not abuse. I've also
pointed out that a number of folk don't know enough of the
relevent literature. That's not abuse either.
My point is that whter two folk agree or disagree is not relevent
to these issues.
Truth is not determined by convention.
--
David Longley
David Yeo
On Tue, 10 Dec 1996, Jim Balter wrote:
> This gives us a way to understand the
> errors in Putnam's Twin Earth scenario, the errors in Harnad's notion of
> symbol grounding, and numerous errors by Searle.
Apart from a lack of detail re: how iconic and categorical representations
obtain, what are "the errors in Harnad's notion of symbol grounding"? Do
you not accept Harnad's notion that symbol meaning cannot be grounded in
yet more (i.e. other) meaningless symbols?
- David Yeo (Applied Cognitive Science, University of Toronto)
Neil Rickert
In <850224...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley) writes:
>In article <58i2sg$7...@usenet.srv.cis.pitt.edu>
> ande...@pitt.edu "Anders N Weinstein" writes:
>> Longley's scrupulous avoidance of personal abuse stands as a model
>> for us all.
>Perhaps it should!
I guess you missed the sarcasm.
>I have remarked frequently that a number of people in this group
>succumb to soplipsism or don't base what they have to say on
>any empirical research evidence. That's not abuse.
It is abusive when it wrong, and excessively repeated.
> I've also
>pointed out that a number of folk don't know enough of the
>relevent literature. That's not abuse either.
That too is abusive when it wrong, and excessively repeated.
>My point is that whter two folk agree or disagree is not relevent
>to these issues.
>Truth is not determined by convention.
Except perhaps by David Longley's convention, whereby he sets himself
up as a standard of truth and accuses everyone else of falsity.
Anders N Weinstein
>
>Since this project has both been very fruitful (giving rise to modern
>chemistry) and has also turned out to be possible (it is routine that
>nuclear power stations produce small quantities of gold), this
>would seem a particularly unfortunate analogy.
No fair, it seems to me. Nuclear reactions are not an application
of alchemical principles. Modern nuclear physics has abandoned the
conceptual or explanatory framework which made the alchemical project
appear viable and natural. If it turns out to be possible according to
quite different principles, that is just being right by accident
for the alchemists.
>>Of course you are strictly correct that it *might* be possible (though
>>I doubt it). As I see it, the orientation that produces the hysterical
>>urgency behind the philosophical movement I detest ("naturalism")
>
>If your position amounts to little more than an unhappiness with
>'hysteical urgency', then what not simply say that in the first place?
It was unhappiness with what I take to be the *source* of what I called
the "hysterical urgency".
For example, if you look at a certain body of literature in soi-disant
"naturalist" philosophy of mind, you see a rather limited body of
options: eliminative materialism, Cartesian dualism, behaviorism,
cognitivism. The views of Wittgenstein and Ryle (and Aristotle) are
simply not on this menu at all. So I think something is clearly wrong
with the nearly apriori orientation -- the forehaving, as Heideggereans
might call it -- of these philosophers which makes it look as those
these great labelled ism's exhaust the possibilities, and we are condemned
to pick and choose from this list.
I think a lot of the confusion derives from thinking of the mind
Descartes-wise as the locus of the inner causal antecedents of mindful
conduct, which bestow upon it its character as intelligent or not. In
this way *all* the options partake of similar conceptual errors at
bottom. Having rejected the idea of the Cartesian soul or ego, one is
driven to locate this source inside the body, in the brain
(cognitivism), or else deny it altogether (behaviorism). And then it
also looks as though neuroscience might fail to find anything like it
inside the brain (eliminative materialism, a la Churchland).
Whereas I think these are positions are all fundamentally confused
about the nature of pyschological language. I think we learn from
Wittgenstein and Ryle that what we think of as psychological often
involves *no such thing* as that inner causal source, that a great many
psychological statements simply do not point inside the body for their
truth to some unseen reality which theoretical science might get to
know better. So the study of the actual physical mechanisms inside the
body, although interesting and important in its own right, is largely
irrelevant to these conceptual questions about intentionality and the
like.
>There does seem to be an on-going problem with accounting
>for the causal relevance of non-reductive accounts. Which is
True. On the other hand, views like the token event identity have their
own sorts of problems -- the problem being that the intentional
attributes of the states come to seem epiphenomenal, irrelevant to the
causal efficacy of the state. I.e. when one neural state -- which, as
it happens is a thinking that I need milk for my breakfast tomorrow --
causallly gives rise to another -- which happens to be an intention to
pick some up at the convenience store on my block on the way home from
work -- the content of the states seems to have nothing to do with the
causal process.
I think a big source of the problem is the casual reification of
states as causes. "Why are you going that way?" "I need to stop by the
store and pick up some milk for my breakfast" So far, all is clear.
But now comes the scientistic philosopher: Doesn't this mean there must
be some *inner* state -- my desire to have milk with my breakfast, say --
which *caused* me to walk this way. And where could this state be but inside
my brain? But how could such a state have the consequences of a desire or
intentiona? And now we are off down the garden path.
I don't think it is so innocent to move from everyday formulations like
"I went to the store because I needed milk" to such things as "my state
of desiring milk *caused* me to go to the store" (unless the latter is
taken to be equivalent to the former). Perhaps we can have the former
as a variety of causal explanation without reifying states as causes
that operate, as it were, behind the person's back.
>>I don't think many
>>people would try to claim that a computer might be a subject of mental
>>states unless they believed they *had* to for some other reason, that
>>it was intellectually *obligatory*, not optional, given some sort of
>>due respect for the achievements of science.
>
>Don't understand what you're talking about. Are you talking about
>a robot?
No. I said "computer" and I meant it. I was referring to the literature
explicitly occupied with the question whether a computer might think.
> If we had robots who could behave in ways which gripped
>us as the sort of ways peoplebehave, I think *most* people would
>consider them to be the subject of mental states - it would be rejected
>by those who have some religious (or philosophical) axe to grind.
But I agree a robot could be a an artificial person and so a fit subject
of mental states. It might also take out a mortgage loan, but you would not
try to explain that state solely in terms of inner information processing.
>However, reductive strategies are typical in science and in many
>area of investigation. These enquiries can take place with or without
>any expectation of success. Anyhow, this isn't what you said, although
>it makes more sense. Probably the root of naturalism is the recognition
>that it is very difficult to understand how the causal story is to be
>told without some sort of reductive linkage.
There is a body of philosophical work organized around the slogan
"reasons are not causes". I think this may be a somewhat misfiring way of
expressing a good idea; certainly I think it deserves to be taken
extremely seriously, and is far from slain and buried by Donald Davidson
as most naturalists assume.
>>>It doesn't sound like a plausible dichotomy to me. It sounds like
>>>a dichotomy specially constructed for a purpose: rubbishing
>>>physicalism.
>>
>>Right you are. I should admit it is not toothless in that it does seem
>>to rule out immaterial souls and vitalism. But that is not the same as
>>saying that the predicates and explanations of biology, say, get
>>reduced.
>
>Well I'd say that Darwin's theory gives a naturalistic account of
>the origin of species. Period. Now notice that I'm not saying that
>Darwin gives necessary and sufficient conditions for what a Jaguar
>is, nor am I saying that Darwin gives necessary and sufficient
>conditions for the cicumstances under which Jaguars evolve.
>But I think Darwin gave a naturalistic account of the origin of
>species nevertheless.
Fine by me. Note also that he doesn't tell you, metaphysically speaking,
what it is to be a species, or what role being an instance of a
reproducible form has in the concept of being a living thing at all.
[I happen to think the story about the latter might be the same even
if natural selection were not true, for example if gods just created
all the beasts in their present form. I also think our confidence that
the heart is -- has as its proper function to be -- a natural pump
should not be made to depend on facts about evolution.]
>That is, the argument that we can't give naturalistic necessary
>and sufficient conditions for "is a table" is neither here nor there
>from the point of view of success at naturalising the mental.
>What is needed is not anything so ridiculous, but something
>deeper but probably easier: an account of why and how such
>classifications occur. That is, an account, not of the detail, but
>of the system as a whole.
I guess one might look for this as well in Kant's architechtonic of
pure reason or Hegel's phenomenology of Geist or Heidegger's
existential analytic of Dasein, or Husserl's transcendental
phenomenology or whatever other philosophical work floats your boat.
Why does it have to be *natural science* that answers such questions?
>Well, I'm not sure this a credible view of basic physics, if basic physics
>includes thermodynamics. As soon as we include the question of
>reversibility and irreversibility, we talk in a very different way.
>It could be, for example, that the sort of behaviours we tend to associate
>with rationality tends to be associated with reversibility, or at least, with
>less irreversibility than otherwise.
I tend to expect that intelligence will not give us an escape from the
second law. I would expect the behavior we describe as acquiring
information about the world and intervening in the world so as to act
rationally on it will also turn out, when described in physical terms
applied to the constituents of the body and brain that make it possible,
to be in accordance with the second law.
>What I'm pointing out is that physicalism has more stings to its bow
>than the rather sorry tale above would seem to imply. I find this
>rather strange because Anders has told the same tale before and I have
>pointed out that physicalism is a mansion with more rooms before now.
I have not seen a well worked out new room. I guess I admit it is
a project to work it out. My beef is with the likes of Fodor or Churchland,
who seem to presuppose some largely inarticulate version of physicalism.
>But, apparently, Anders is happier to trot out the same tired old
>nags than to adjust anything he might say. We seem to be involved
>in 'memoryless' exchanges as a result. Each interaction starts from the
>same place it started from last time. Why this is, I simply don't know.
>Better ask Anders.
Although I know some physics, I am simply not qualified to evaluate
arguments based on thermodynamics. Again, I really don't see the likes
of David Chalmers worrying about *that* particular issue or anything
like it, nor does it come up in the philosophical discussions of
"naturalism" that exercise me.
Take me as wanting to make fun of the inchoate scientism I find in a
certain incestuous backwater of the English speaking philosophical
scene. If you do have a more sophisticated view, I hereby exempt you
from my target area.
>At any rate, until we know better the relationship between levels
>of language and how they're formed, I still don't see how you
I think we already know *how* to use the psychological language quite
well. What we lack is a clear reflective understanding of what we do a
thousand times a day. The examples we need to consider, however, are
already open to view.
>As I say, basic physics comes in two favours: forces and fields
>on the one hand and thermodynamics on the other. Don't forget
>that when one is talking about physicalism, they're both relevant.
>
>If we knew how our vocabularies were formed, perhaps you would
I don't understand this emphasis on how they were formed. I think we
can understand, e.g. the form of life of the a particular animal
synchronically, wihout looking to how it was formed. Ditto, perhaps for
such things as the custom of using signposts, and for our vocabularies.
Even if we do need to look to history to understand these things, I don't
see that that will answer the question about relations between different
vocabularies.
>do better at showing some sort of sharp divide between the
>language of basic physics and the rest. However, since this isn't
>something we know much about, we're not sure whether the
>boundaries you wish to draw have any greater significance than
>any other mark in the sand of philosophical speculation.
The Kantian distinction between acting in accordance with rules and
acting according to a conception of rules seems to me to be a good one.
Or more generally the distinction between normative and non-normative.
Anders N Weinstein
In article <jqbE26...@netcom.com>, Jim Balter <j...@netcom.com> wrote:
>In article <58habu$10...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
>>In <589no9$m...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>>>In article <588rm7$n...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
> The issue
>is not a search for scientific justification for claims such as "this is a
>table" or "this is a way to get onto the subway" (although there are
>topological considerations) or "this animal is hunting its prey" (although
>science can show that some such claims are mistaken, which makes me suspect
>that Anders' scientific scope really is extremely narrow, and suggests that
>Anders is confused even within this category), but rather the issue is an
I am not sure what you are referring to here. I would hardly deny that claims
that an animal is hunting its prey can be overturned by empirical inquiry.
As usual, my idea would be that you can not explain what it is for an
animal to be hunting its prey by looking only at the causally relevant
structures inside its body, or discover from such an inquiry that there
really is no such thing as hunting prey. I don't see that that inquiry
stands to "hunting its prey" as, I concede to Kripke/Putnam, atomic
theory might stand to "is gold".
"Hunting its prey" is also interesting to me in this regard since it
is all of: behavioral, intentional, natural.
>exploration and explanation of these *claims* *as human behavior*. The
>distinction here is between use (when people say these things) and mention
>(when we refer to people saying these things).
I am quite clear on the distinction, thank you.
I would point out that an awful lot of what you suggest depends on what
descriptive vocabulary we should use for characterizing the human use of
concepts, the making of the claims as *human behavior*. I think that it
is a form of psychologism to explain them solely as natural phenomena
in terms of their causes. One *can* do that, of course. But to do that is not
to take them as evaluable *claims*, beliefs we might ourselves want
to share, or take issue with, or argue about. The concepts of meaning
etc. are rooted in the latter sort of fundamentally evaluative stance,
I think.
>When Anders attempts to offer explanations at that level, he talks about
>language games and norms and roles within communities and so on. This is all
>fine and well, but doesn't answer fundamental causal questions concerning the
>*mechanisms* at work in the individuals that make up these communities, or in
>the environment at large (including the comprising individuals) that bring
>about these communities. And when pushed, Anders says that he wasn't offering
>*causal* explanations. But that's the point; causal explanations are what are
>being asked for, and non-causal explanations are no *substitute*.
Asked for by whom?
>seems to think that, by supplying non-causal explanations, he has discharged
>his duty, and has shown that causal, reductive explanation is a "non-issue", a
>"don't care".
For philosophical purposes, yes. Not absolutely of course.
>Yet, this is an AI group, and presumably a fundamental question of concern is
>"how to build an AI". And any good engineer recognizes that one path toward
OK, but another fundamental question is: what is the relation between
this question and philosophy of mind? Should philosophy of mind -- or,
for that matter, philosophy of social science (deals with such things
as norms and roles in communities) simply get handed off in toto to the
control system engineers?
>But for some reason Anders seems to see such engineeering as a threat, and
>seems to feel some urgency in dissuading people from carrying out the project.
I do not think I am threatened by engineering, nor do I want to
dissuade people from carrying it out. You might not have noticed, but
my target has always been philosophers. In particular, those who are
enthusiasts for the cognitive science perspective on the mind.
I suppose that does threaten me in a fairly concrete way: not just by
making for bad philosophy, but by making a very inhospitable
environment in the academic world for the sort of philosophy I would be
interested in talking about.
Neil Rickert
I'm keeping this response shorter than usual, since I am supposed to
be grading exams.
In <58ield$9...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>In article <58frbh$o...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>>You raised the question about individuation of concepts. My point is
>>just that we can never expect content to be the same, but we can
>>communicate provided that my content and your content is sufficiently
>>similar.
>Ok, but how do we determine similarity of content?
We have no need for such a comparison. To the extent that we are
successful in communicating, we must be communicating about what is
common. To the extent that we talk past each other, our linguistic
expressions must be hinging on areas where there are significant
conceptual differences.
>The problem it seems to me is that reflecting on our differing standards
>of precision seems to point to the idea of a more fine-grained background
>language against which different procedures are assessed.
No, I am not suggesting that.
>I am wondering about the scale against which we rank overlap
>of content. Doesn't this require commensurating the incommensurable?
We have no scale, and we don't normally try to measure overlap.
>Now where exactly is this "problem of intentionality" again?
In all of the stuff you glossed over. You simply waved your hands,
used a flourish of your magic wand, and claimed that the problems had
disappeared. But how does "slab" have any significance to a new
initiate before he comes to play the game? How does a learner come
to appropriate a pre-existing semantic significance? You skip past
all such questions, as if there were nothing to be explained.
>Not sure I understand this. I believe the intentionality of thoughts
>is not reducible, but is not autonomous either, in that it is internally
>related to linguistic expressibility.
>...
>>There is no problem, if you are happy with reliance on magic.
>Where is the reliance on magic?
In your use of "intentionality".
I take it that "not reducible" is pretty much a synonym for
"magical". Anything not magical should be reducible to physics. I
think most physicists would agree. Those physicists who think that
intentionality is not reducible to known physics are prone to talk
about the need for new physics which would allow it to be reducible.
This is part of the motivation for quantum consciousness arguments.
>>The question is whether that assessment in epistemic terms is
>>anything more than a "Just So" story. The type of arguments about
>>justification that I see in treatises on epistemology have little to
>>do with the ways scientists develop their theories, for example.
>I am not mainly interested in the treatises, but in the mundane
>everyday uses of the concepts.
As far as I can tell, in their ordinary every day use of language,
most people don't spend time worrying about whether they are
justified in holding beliefs. So again, I think that epistemology
has very little relevance to how we acquire and use knowledge.
>everyday uses of the concepts. I don't see that you have given support
>for anything like the dismissive attitude you express, i.e. you haven't
>shown that we should just completely dispense with the idea of a belief
>which may be assessed as justified or not.
What am I supposed to give support to? As far as I can tell, there
is little evidence that beliefs are the driving force behind
behavior. I have not seen any persuasive arguments to suggest
otherwise although I have seen many bogus arguments.
As far as I can tell, most philosophical talk of beliefs can only
make sense when it has to do with ascribed beliefs. I can agree that
an ascribed belief is a mental state. But if you ascribe a belief to
me, then it seems quite clear that the mental state involved is your
mental state as the ascriber, rather than any state that I might or
might not have.
>Wouldn't it be better to deny that my body is the agent of my walking?
>*I* walk, and when I do, my body moves. Sub-personal information
>processing in my brain makes this possible: if not for that, I couldn't
>walk. But the brain or neural circuits themselves don't do any walking.
So the body receives all of the information required for balance, but
you do not receive it. Somehow you manage to walk without the
information. But the body does not walk, so presumably it has no
need for the information. The body just comes along for a free ride
when you walk. Somehow this sounds as wierd as Chalmers' zombies.
>>Again, 'undetected' is the wrong restriction. The signal is being
>>detected but the information is not being extracted from that
>>signal. Since the signal is being detected, it can have in
>>influence. Because we are active participants in perception, we can
>>search for other signals which are correlated to this one, as a way
>>of attempting to recover the information from the signal.
>OK, I am getting clearer on what you mean. I guess the
>processes that "search for other correlated signals" and "attempt to
>recover the information" are not really the processes that detect the
>information itself. E.g. the process that leads to the development of an
>affords-mailing detector is not itself an affords-mailing detector, but
>some kind of analyzer or decoder of meaningless patterns.
More or less, although I would strike the word "meaningless". I
don't think we normally consider patterns that occur in nature to be
meaningless.
>I am still a little bit puzzled as to how you avoid the Fodor problem.
>E.g. how do those learning processes 'decide' what information content to
>ascribe to those meaningless patterns if they don't have the relevant
>concepts to begin with?
Again, I object to the "meaningless". Once patterns of correlation
have been discovered, there is some sort of meaning. At the very
least one of the correlated signals can be used to improve the
ability to predict others of those signals. That already gives some
kind of primitive meaning to the signal.
Suitable concepts can usually be constructed to accomodate the
patterns of correlation that were found. Certainly this has happened
often enough in science that it should not come as a surprise. For
example, the concept of gene came from the combinatorial properties
of the pattern of simple inheritance noticed by Mendel. It certainly
did not require knowing the structure of the DNA, or even that there
was such a thing as DNA.
Drew McDermott
To summarize: everyone in this newsgroup agrees that Chalmers is wrong. I had
hoped for more controversy.
This sort of thing doesn't count as controversy:
In article ... X writes
|> In article ... Y writes:
|>
|> > In ... X writes:
|> >
|> > >That's why these exchanges are so futile.
|> >
|> > >Contrast this with what I have been advocating and requesting.
|> >
|> > Ah, yes. You have been advocating that humans should become mindless
|> > automatons, whose only contact with reality is via the automatic
|> > citation machine.
|> >
|> >
|> Y, .....just read some of the literature.
Sigh.
-- Drew McDermott
Peter Lupton
In article: <Pine.SOL.3.91.961210083406.5819D-100000@tortoise> David Yeo <dyeo@tortoise> writes:
>
> On Tue, 10 Dec 1996, Jim Balter wrote:
>
> > This gives us a way to understand the
> > errors in Putnam's Twin Earth scenario, the errors in Harnad's notion of
> > symbol grounding, and numerous errors by Searle.
>
> Apart from a lack of detail re: how iconic and categorical representations
> obtain, what are "the errors in Harnad's notion of symbol grounding"? Do
> you not accept Harnad's notion that symbol meaning cannot be grounded in
> yet more (i.e. other) meaningless symbols?
I would say that there ar two ways (at least) to view a computation.
1. As a sort of mechanism, intracting with the world.
2. As a sort of process of symbol manipulation, where the symbols
are assigned values by the user.
In case 2 there is no question of symbol grounding - the symbols
are grounded by we humans before the first move by the computer is made.
In case 1, what we have to give an account of is not symbol grounding,
but the generation of symbols and representational states generally.
Either way, there doesn't seem to be a symbol grounding problem per se.
Perhaps the 'symbol grounding problem' occurs because one mixes pictures 1
and 2. That is, one thinks of a computer as though it were an autonomous
machine, whilst imposing *our* symbols on it. *Then* you have the problem of
how-come *our* symbolic assignment is grounded by the machine.
Cheers,
Pete Lupton
David Longley
In article <58k8ik$t...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
<DL>
> >I have remarked frequently that a number of people in this group
> >succumb to soplipsism or don't base what they have to say on
> >any empirical research evidence. That's not abuse.
>
> It is abusive when it wrong, and excessively repeated.
>
Do you *know* what it means to cite evidence and references?
> > I've also
> >pointed out that a number of folk don't know enough of the
> >relevent literature. That's not abuse either.
>
> That too is abusive when it wrong, and excessively repeated.
Where are these references? Where is the research that you make
reference to?
>
> >My point is that whter two folk agree or disagree is not relevent
> >to these issues.
>
> >Truth is not determined by convention.
>
> Except perhaps by David Longley's convention, whereby he sets himself
> up as a standard of truth and accuses everyone else of falsity.
>
This is just typically adolescent. I said "Truth is not
determined by convention". Now that clearly means I don't hold
that view. So what do you think it might mean?
Not enough folk have been kind enough to you in the past to let
you know how uninformed most of your remarks in this field are.
It is this kind of abuse, along with sweeping dismissals of good
scientific data which will deter other readers from making useful
contributions to this newsgroup. I really think you should give
some thought to that.
--
David Longley
Neil Rickert
In <850273...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley) writes:
>In article <58k8ik$t...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
><DL>
>> >I have remarked frequently that a number of people in this group
>> >succumb to soplipsism or don't base what they have to say on
>> >any empirical research evidence. That's not abuse.
>> It is abusive when it wrong, and excessively repeated.
>Do you *know* what it means to cite evidence and references?
Yes, you fucking ignoramous, I do know how to cite, and I am quite
capable of doing so where it is appropriate. Shitheads like you
shouldn't be accusing others of abuse. Why don't you switch off your
bloody computer, and finish that Ph.D. which you never completed --
if you can find any professor willing to put up with a closed minded
dogmatist like you.
[Yes, I can dish out abuse when appropriate.]
>> > I've also
>> >pointed out that a number of folk don't know enough of the
>> >relevent literature. That's not abuse either.
>> That too is abusive when it wrong, and excessively repeated.
>Where are these references? Where is the research that you make
>reference to?
Look, you fucking idiot, I am depending on the same evidence that you
use, together with much that you discard. When I am disagreeing with
the usual interpretation of well known evidence, I am do not need to
cite results which counter that evidence. I need only show that the
interpretation I am criticizing is riddled with holes, or lacks the
evidence to make it credible.
If you don't care for what I write, you can always go back to
psychology newsgroups. For AI, we require something more than
handwaving claims that there are heuristics. For AI folk, there
ain't a heuristic until there is an explicit program implementing
it. And that means that there ain't a heuristic. If you are such a
fucking brilliant know-it-all, just post in full detail the godammed
program that implements these heuristics. It will need to be an
extensional (GOFAI) program, since an ANN system cannot be seen to
satisfy these heuristics except by handwaving arguments that are not
persuasive. But of course we all know that you will not produce any
explicit heuristic programs. You will just make more handwaving
arguments, and perhaps mention ANNs.
All of your bloody talk about extensionalism is little more than
hypocracy. If you were half the extensionalist you claim to be, then
you would not allow the use of "heuristic" in such a non-extensional
handwaving manner. Talk about a heuristic is not extensional unless
you can produce the program that implements it. The plain fact is
that I practice extensionalism far more rigorously than you. You
only preach it. If you understood the extensionalism you preach, you
might realize why I have reservations about R-W learning theory. For
that theory is not sufficiently extensional for my liking.
>> >Truth is not determined by convention.
>> Except perhaps by David Longley's convention, whereby he sets himself
>> up as a standard of truth and accuses everyone else of falsity.
>This is just typically adolescent. I said "Truth is not
>determined by convention". Now that clearly means I don't hold
>that view. So what do you think it might mean?
Oh, bullshit. If you were sufficiently intelligent, I might argue
with you about where Quine went wrong in his "Truth by convention."
But as it is, you are a bloody closed minded dogmatist, who accepts
only truth by the conventional wisdom.
If you want to debate issues, start doing so. Judging by your
postings over the last 18 months, I doubt that you are capable of
doing so. But just get of this goddammed sermonizing. You are only
a bloody psychologist, after all. You have no right to be lecturing
to computer scientists and AI workers on how they should discuss
their subject.
>Not enough folk have been kind enough to you in the past to let
>you know how uninformed most of your remarks in this field are.
Translation: That idiot Longley doesn't agree with me.
>It is this kind of abuse, along with sweeping dismissals of good
>scientific data which will deter other readers from making useful
>contributions to this newsgroup. I really think you should give
>some thought to that.
On top of being a complete shithead, you are an out and out liar.
Your accusation that I have dismissed good scientific data is
a damnable falsehood.
Now how about you stop your fucking campaign of lies that your
are leveling against me.
David Longley
In article <545069...@luptonpj.demon.co.uk>
lup...@luptonpj.demon.co.uk "Peter Lupton" writes:
If one adopts a Rescorla-Wagner framework (which for AWs benefit,
I think generates something very close to a Heideggerian
worldview) where significant connections are made between classes
of stimuli on the grounds of spatial and temporal contiguity and
'surprisal', symbols miht better be conceived as 'marks' or
discriminated patterns of stimulation.
The transformations of these marks by a computer according to
logical rules cuts through all talk of symbols and meaning when
understood in this way.
When the symbol processing notion was first introduced it was
really *only* to point out that computers could be used to
process more than *just* numbers. All sorts of mischief has been
created since...
--
David Longley
David Yeo
On 11 Dec 1996, Neil Rickert wrote:
> For AI, we require something more than
> handwaving claims that there are heuristics. For AI folk, there
> ain't a heuristic until there is an explicit program implementing
> it. And that means that there ain't a heuristic. If you are such
I think there are two senses of "algorithm" being interwoven here. While
it is true that heuristics are procedures (hence algorithms), algorithms
usually also carry the implication that they will achieve the solution to
a problem in a finite number of steps. To quote Rosenberg (1986):
Algorithm: a finite set of well-defined rules for the solution of a
problem in a finite number of steps; for example, a full
statement of an arithmetic procedure for evaluating the sin
X to a stated precision.
(Dictionary of Artificial Intelligence and Robotics)
Heuristics (rules of thumb) do not offer this degree of certainty; in fact
they don't guarantee they'll achieve the solution to a problem at all. I
suggest that it's this latter (non-procedural) sense of algorithm that AI
researchers imply when distinguishing "algorithms" from "heuristics".
> !@#$%& brilliant know-it-all, just post in full detail the godammed
> program that implements these heuristics. It will need to be an
> extensional (GOFAI) program, since an ANN system cannot be seen to
> satisfy these heuristics except by handwaving arguments that are not
> persuasive.
Why do heuristics necessarily require GOFAI? I may be missing some subtle
point you are making (your posting carries considerable baggage from past
discussions with David Longley, so I may be inappropriately jumping into a
longstanding, and apparently venomous, debate), but it would seem that
ANN's are the embodiment of "rules of thumb". That is, as the product of
induction ANN's would clearly seem to be compatible with many standard
definitions of the term "heuristic". To again quote Rosenberg (1986):
Heuristic: pertaining to exploratory methods of problem solving in
which solutions are discovered by the evaluation of progress
made toward the final result. (ibid)
Cheers,
Anders N Weinstein
In article <850301...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>
>If one adopts a Rescorla-Wagner framework (which for AWs benefit,
>I think generates something very close to a Heideggerian
>worldview) where significant connections are made between classes
>of stimuli on the grounds of spatial and temporal contiguity and
>'surprisal', symbols miht better be conceived as 'marks' or
>discriminated patterns of stimulation.
The main problem is that we have to connect our concept use to worldly
objects, *not* to "proximal" stimuli. When one reaches unreflectively
to pick up pen and paper to write, say, one has not classified
"stimuli" -- one has classified the pen and paper *themselves* as tools
suitable for writing with. One can find oneself called out by the world
on this classification -- perhaps the cylindrical object turns out to really be
a walking-stick insect! -- *even though* there was never any occurent
intellectual act of recognition "in the mind".
Perhaps there was no representation anywhere inside the brain of the
object as a writing instrument, beyond the circuitry which realizes the
disposition to pick it up in these circumstances. But it is not *just*
an atomistic stimulus-response effect either, since this sort of
classificatory act requires all the other things which constitute
understanding of the role of pencils, paper, writing instruments, in
various human practices in which they play a role. A pigeon trained to
pick up a pencil and even write with it is not thereby classifying it
in the way we are, as equipment relevant to our purposes. Neither is a
neural net that can sort pencil-like stimuli.
In general, the Dasein-level, like Gibson's ecological level (which
applies also to animals), concerns relations between a certain sort of
intentional activity of a whole person and the things in the world that
are caught up in or play a role in that activity. The "proximal"
stimuli that causally mediate between the two are transparent at this
level of description.
Of course Gibson's ecological level gives a prominent role to proximal
patterns in the stimulus, but re-describes them in rich,
organism-involving terms, as natural sources of information about
features of the world that are relevant to that organism's life. In
this way the stimuli patterns are not considered epistemic
intermediaries or bases which stand between the cognizant subject ("the
mind") and the cognized objects and properties ("the world"). Indeed,
at this level there are in a sense no causal intermediaries between the
mind on one side of a causal boundary and the world on the other.
Rather there is the single organismic activity of extracting relevant
information about the world (and about itself as an object in it).
I don't see that a reductive behaviorism of stimuli and responses can
do justice for our intentional relations to the ("distal") world at all
-- at best we would be in a solipsistic world in which we only classify
stimuli, and our thought never really does answer to distal objects for
its correctness at all. And it is even harder to see how it can handle
the phenomenological immediacy with which the world presents itself to
our understanding.
I think that doing justice to these things requires modifying or
abandoning the causal picture of our perceptual relation to the world
that was bequeathed to us by modern science via Descartes and Galileo.
David Longley
In article <58lld4$5...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>
> Look, you fucking idiot, I am depending on the same evidence that you
> use, together with much that you discard. When I am disagreeing with
> the usual interpretation of well known evidence, I am do not need to
> cite results which counter that evidence. I need only show that the
> interpretation I am criticizing is riddled with holes, or lacks the
> evidence to make it credible.
Do you not think I (and others) think it somewhat incredible that
the ONLY references you cite (by name only) are the ones I have
documented specifically? You make it quite clear by what you
write that you have not read the material before, nor have you a
grasp of the overall literature which it is part of. Now that's
not abuse, just fact.
>
> If you don't care for what I write, you can always go back to
> psychology newsgroups. For AI, we require something more than
> handwaving claims that there are heuristics. For AI folk, there
> ain't a heuristic until there is an explicit program implementing
> it. And that means that there ain't a heuristic. If you are such a
> fucking brilliant know-it-all, just post in full detail the godammed
> program that implements these heuristics. It will need to be an
> extensional (GOFAI) program, since an ANN system cannot be seen to
> satisfy these heuristics except by handwaving arguments that are not
> persuasive. But of course we all know that you will not produce any
> explicit heuristic programs. You will just make more handwaving
> arguments, and perhaps mention ANNs.
More nonsense and abuse - and anyone who has read "Fragments"
will see it as exactly that.
>
> All of your bloody talk about extensionalism is little more than
> hypocracy. If you were half the extensionalist you claim to be, then
> you would not allow the use of "heuristic" in such a non-extensional
> handwaving manner. Talk about a heuristic is not extensional unless
> you can produce the program that implements it. The plain fact is
> that I practice extensionalism far more rigorously than you. You
> only preach it. If you understood the extensionalism you preach, you
> might realize why I have reservations about R-W learning theory. For
> that theory is not sufficiently extensional for my liking.
>
The above is just gibberish. I have made very explicit references
to the work in decision theory by Gluck and Bower on this matter,
and I have repeated the oft stated remark that the R-W model is
the most influential model over the past 25 years or so. Now IF
you understood the literature, and if you understood what I was
saying, you would see the relationship between what I have said
about the Extensional Stance (as opposed to the intentional
stance), the use of normative descriptive statistics (including
regression and its variants in pattern recognition), the problems
of biases of natural assessments (which can be modelled by ANNs)
and a host of other key areas of research which are covered in
"Fragments". But you don't. And the reason you don't is because
you don't have enough of a background in psychology, behaviour
science and AI.
No abuse - just fact.
> >> >Truth is not determined by convention.
>
> >> Except perhaps by David Longley's convention, whereby he sets himself
> >> up as a standard of truth and accuses everyone else of falsity.
>
> >This is just typically adolescent. I said "Truth is not
> >determined by convention". Now that clearly means I don't hold
> >that view. So what do you think it might mean?
>
> Oh, bullshit. If you were sufficiently intelligent, I might argue
> with you about where Quine went wrong in his "Truth by convention."
> But as it is, you are a bloody closed minded dogmatist, who accepts
> only truth by the conventional wisdom.
>
It's hardly conventional..
> If you want to debate issues, start doing so. Judging by your
> postings over the last 18 months, I doubt that you are capable of
> doing so. But just get of this goddammed sermonizing. You are only
> a bloody psychologist, after all. You have no right to be lecturing
> to computer scientists and AI workers on how they should discuss
> their subject.
Do yourself a favour Neil. Go and study some of the relevant
literature. I quite understand why you have an interest in the
area, and mathematics has an important part to play - but at the
moment it's quite clear to me that you just don't know enough of
the subject matter to really appreciate where the *important*
problems are. Pointing this out to you, in such detail is really
quite positive (though no doubt you'll read it as abuse).
You should look at how you write. It is generally just a report
of how you personally react to what you think others say. You
don't seem to know how to analyse it as a scientist.
>
> >Not enough folk have been kind enough to you in the past to let
> >you know how uninformed most of your remarks in this field are.
>
> Translation: That idiot Longley doesn't agree with me.
Well, take it that way if you wish - I mean to be helpful.
>
> >It is this kind of abuse, along with sweeping dismissals of good
> >scientific data which will deter other readers from making useful
> >contributions to this newsgroup. I really think you should give
> >some thought to that.
>
> On top of being a complete shithead, you are an out and out liar.
> Your accusation that I have dismissed good scientific data is
> a damnable falsehood.
Og hear - you're getting hysterical again.
>
> Now how about you stop your fucking campaign of lies that your
> are leveling against me.
>
There's no campaign - I'm just assessing your knowledge in this
field and letting you know how limited it is.
And if you feel the need to produce another diatribe like the
above again, do me and others the courtesy of sending it via e-
mail (after thinking carefully whether it is worth it).
--
David Longley
David Longley
In article <58lld4$5...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>
> Look, you fucking idiot, I am depending on the same evidence that you
> use, together with much that you discard. When I am disagreeing with
> the usual interpretation of well known evidence, I am do not need to
> cite results which counter that evidence. I need only show that the
> interpretation I am criticizing is riddled with holes, or lacks the
> evidence to make it credible.
>
> psychology newsgroups. For AI, we require something more than
> handwaving claims that there are heuristics. For AI folk, there
> ain't a heuristic until there is an explicit program implementing
> it. And that means that there ain't a heuristic. If you are such a
> fucking brilliant know-it-all, just post in full detail the godammed
> program that implements these heuristics. It will need to be an
> extensional (GOFAI) program, since an ANN system cannot be seen to
> satisfy these heuristics except by handwaving arguments that are not
> persuasive. But of course we all know that you will not produce any
> explicit heuristic programs. You will just make more handwaving
> arguments, and perhaps mention ANNs.
>
> hypocracy. If you were half the extensionalist you claim to be, then
> you would not allow the use of "heuristic" in such a non-extensional
> handwaving manner. Talk about a heuristic is not extensional unless
> you can produce the program that implements it. The plain fact is
> that I practice extensionalism far more rigorously than you. You
> only preach it. If you understood the extensionalism you preach, you
> might realize why I have reservations about R-W learning theory. For
> that theory is not sufficiently extensional for my liking.
Given the above, I recommend that you re-read this *carefully*
Full text is available at:
http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
Connectionist systems, it is claimed, do not represent knowledge as
production rules, ie as well-formed-formulae represented in the syntax
of the predicate calculus (using conditionals, modus ponens, modus
tollens and the quantifiers), but as connection weights between
activated predicates in a parallel distributed network:
'Lawful behavior and judgments may be produced by a
mechanism in which there is no explicit representation
of the rule. Instead, we suggest that the mechanisms
that process language and make judgments of
grammaticality are constructed in such a way that their
performance is characterizable by rules, but that the
rules themselves are not written in explicit form
anywhere in the mechanism.'
D E Rumelhart and D McClelland (1986)
Parallel Distributed Processing Ch. 18
Such systems are function-approximation systems, and are
mathematically a development of Kolmogorov's Mapping Neural Network
Existence Theorem (1957). Such networks consist of three layers of
processing elements. Those of the bottom layer simply distribute the
input vector (a pattern of 1s and 0s) to the processing elements of
the second layer. The processing elements of this middle or hidden
layer implement a *'transfer function'* (more on this below). The top
layer are output units.
An important feature of Kolmogorov's Theorem, is that it is not
constructive. That is, it is not algorithmic or 'effective'. Since the
proof of the theorem is not constructive, we do not know how to
determine the key quantities of the transfer functions. The theorem
simply tells us that such a three layer mapping network must exist. As
Hecht-Nielsen (1990) remarks:
'Unfortunately, there does not appear to be too much
hope that a method of finding the Kolmogorov network
will be developed soon. Thus, the value of this result
is its intellectual assurance that continuous vector
mappings of a vector variable on the unit cube
(actually, the theorem can be extended to apply to any
COMPACT, ie, closed and bounded, set) can be implemented
EXACTLY with a three-layer neural network.'
R. Hecht-Nielsen (1990)
Kolmogorov's Theorem
Neurocomputing
That is, we may well be able to find weight-matrices which capture or
embody certain functions, but we may not be able to say 'effectively'
what the precise equations are which algorithmically compute such
functions. This is often summarised by statements to the effect that
neural networks can model or fit solutions to sample problems, and
generalise to new cases, but they can not provide a rule as to how
they make such classifications or inferences. Their ability to do so
is distributed across the weightings of the whole weight matrix of
connections between the three layers of the network. The above is to
be contrasted with the fitting of linear discriminant functions to
partition or classify an N dimensional space (N being a direct
function of the number of classes or predicates). Fisher's
discriminant analysis (and the closely related linear multiple
regression technology) arrive at the discriminant function
coefficients through the Gaussian method of Least Mean Squares, each b
value and the constant being arrived at deductively via the solution
of simultaneous equations. Function approximation, or the
determination of hidden layer weights or connections is based on
recursive feedback, elsewhere within behaviour science, this is known
as 'reinforcement', the differential strengthening or weakening of
connections depending on feedback or knowledge of results. Kohonen
(1988) commenting on "Connectionist Models" in contrast to
conventional, extensionalist relational databases, writes:
'Let me make it completely clear that one of the most
central functions coveted by the "connectionist" models
is the ability to solve *simplicitly defined relational
structures*. The latter, as explained in Sect. 1.4.5,
are defined by *partial relations*, from which the
structures are determined in a very much similar way as
solutions to systems of algebraic equations are formed;
all the values in the universe of variables which
satisfy the conditions expressed as the equations
comprise, by definition, the possible solutions. In the
relational structures, the knowledge (partial
statements, partial relations) stored in memory
constitutes the universe of variables, from which the
solutions must be sought; and the conditions expressed
by (eventually incomplete) relations, ie, the "control
structure" [9.20] correspond to the equations.
Contrary to the conventional database machines which
also have been designed to handle such relational
structures, the "connectionist" models are said to take
the relations, or actually their strengths into account
statistically. In so doing, however they only apply the
Euclidean metric, or the least square loss function to
optimize the solution. This is not a very good
assumption for natural data.'
T. Kohonen (1988)
Ch. 9 Notes on Neural Computing
In Self-Organisation and Associative Memory
Throughout the 1970s Nisbett and colleagues studied the use of
probabilistic heuristics in real world human problem solving,
primarily in the context of Attribution Theory (H. Kelley 1967, 1972).
Such inductive as opposed to deductive heuristics of inference do
indeed seem to be influenced by training (Nisbett and Krantz 1983,
Nisbett et. al 1987). Statistical heuristics are naturally applied in
everyday reasoning if subjects are trained in the Law of Large
Numbers. This is not surprising, since application of such heuristics
is an example of response generalisation - which is how psychologists
have traditionally studied the vicissitudes of inductive inference
within Learning Theory. As Wagner (1981) has pointed out, we are
perfectly at liberty to use the language of Attribution Theory as an
alternative, this exchangeability of reference system being an
instance of Quinean Ontological Relativity, where what matters is not
so much the names in argument positions, or even the predicates
themselves, but the *relations* (themselves at least two-place
predicates) which emerge from such systems.
Under most natural circumstances, inductive inference is irrational
(cf. Popper 1936, Kahneman et al. 1982, Dawes, Faust and Meehl 1989,
Sutherland 1992). This is because it is generally based on
unrepresentative sampling (drawing on the 'availability' and
'representativeness' heuristics), and this is so simply because that
is how data in a structured culture often naturally presents itself.
Research has therefore demonstrated that human inference is seriously
at odds with formal deductive logical reasoning, and the algorithmic
implementation of those inferential processes by computers (Church
1936, Post 1936, Turing 1936). One of the main points of this paper is
that we generally turn to the formal deductive technology of
mathematico-logical method (science) to compensate for the heuristics
and biases which typically characterise natural inductive inference.
Where possible, we turn to *relational databases and 4GLs* (recursive
function theory and mathematical logic) to provide descriptive, and
deductively valid pictures of individuals and collectives.
This large, and unexpected body of empirical evidence from decision-
theory, cognitive experimental social psychology and Learning Theory,
began accumulating in the mid to late 1970s (cf. Kahneman, Tversky and
Slovic 1982, Putnam 1986, Stich 1990), and began to cast serious doubt
on the viability of the 'computational theory' of mind (Fodor
1975,1980) which was basic to functionalism (Putnam 1986). That is,
the substantial body of empirical evidence which accumulated within
Cognitive Psychology itself suggested that, contrary to the doctrine
of functionalism, there exists a system of independent, objective
knowledge, and reasoning against which we can judge human, and other
animal cognitive processing. However, it gradually became appreciated
that the digital computer is not a good model of human information
processing, at least not unless this is conceived in terms of 'neural
computing' (also known as 'connectionism' or 'Parallel Distributed
Processing). The application of formal rules of logic and mathematics
to the analysis of behaviour solely within the language of formal
logic is the professional business of Applied Behaviour Scientists.
Outside of the practice of those professional skills, the scientist
himself is as prone to the irrationality of intensional heuristics as
are laymen (Wason 1966). Within the domain of formal logic applied to
the analysis of behaviour, the work undertaken by applied scientists
is impersonal. The scientists' professional views are dictated by the
laws of logic and mathematics rather than personal opinion
(heuristics).
Applied psychologists, particularly those working in the area of
Criminological Psychology, are therefore faced with a dilemma. Whilst
many of their academic colleagues are *studying* the heuristics and
biases of human cognitive processing, the applied psychologist is
generally called upon to do something quite different, yet is largely
prevented from doing so for lack of relational systems to provide the
requisite distributional data upon which to use the technology of
algorithmic decision making. In the main, the applied criminological
psychologists as behaviour scientist is called upon to bring about
behaviour change, rather than to better understand or explicate the
natural heuristics of cognitive (clinical) judgement. To the applied
psychologist, the low correlation between self-report and actual
behaviour, the low consistency of behaviour across situations, the low
efficacy of prediction of behaviours such as 'dangerousness' on the
basis of clinical judgment, and the fallibility of assessments based
on interviews, are all testament to the now *well documented
unreliability of intensional heuristics (cognitive processes) as data
sources, and we have already pointed to why this is so.* Yet
generally, psychologists can rely on no other sources, as there are in
fact, inadequate Inmate Information Systems. Thus, whilst applied
psychologists know from research that they must rely on distributional
data to establish their professional knowledge base, and that they
must base their work with individuals (whether prisoners, governors or
managers) on extensional analysis of such knowledge bases, *they
neither have the systems available nor the influence to have such
systems established, despite powerful scientific evidence (Dawes,
Faust and Meehl 1989) that their professional services in many areas
depend on the existence and use of such systems.* What applied
psychologists have learned therefore is to eschew intensional
heuristics and look instead to the formal technology of extensional
analysis of observations of behaviour. The fact that training in
formal statistical and deductive logic is difficult, particularly the
latter, makes this a challenge, since most of the required skills are
only likely to be applicable when sitting in front of a computer
keyboard (Holland et al 1986). It is particularly challenging in that
the information systems are generally inadequate to allow
professionals to do what they are trained to do.
Over the past five years (1988-1993), a programme has been developed
which is explicitly naturalistic on that it seeks to record
inmate/environment (regime) interactions. This system is the
PROBE/Sentence Management system. It breaks out of solipsism by making
all assessments of behaviour, and all inmate targets *RELATIVE to
predetermined requirements of the routines and structured activities
defined under function 17 of the annual Governors Contract*. It is by
design a 'formative profiling system' which is 'criterion reference'
based.
The alternative, intensional heuristics, which are the mark of natural
human judgement (hence our rich folk psychological vocabulary of
metaphor) have to be contrasted with extensional analysis and
judgement using technology based on the deductive algorithms of the
First Order Predicate Calculus (Relational Database Technology). This
is not only coextensive with the 'scope and language of science'
(Quine 1954) but is also, to the best of our knowledge from research
in Cognitive Psychology, an effective compensatory system to the
biases of natural intensional, inductive heuristics (Agnoli and Krantz
1989). Whilst a considerable amount of evidence suggests that training
in formal logic and statistics is not in itself sufficient to suppress
usage of intensional heuristics in any enduring sense, ie that
generalisation to extra-training contexts is limited, there is
evidence that judgement can be rendered more rational by training in
the use of extensional technology. The demonstration by Kahneman and
Tversky 1983, that subjects generally fail to apply the extensional
conjunction rule in probability that conjunctions are always equal or
less probable than its elements, and that this too is generally
resistant to counter-training, is another example, this time within
probability theory (a deductive system) of the failure of extensional
rules in applied contexts. Careful use of I.T. and principles of
deductive inference (e.g. semantic tableaux, Herbrand models, and
Resolution methods) promise, within the limits imposed by Godel's
Theorem, to keep us on track if we restrict our technology to the
extensional.
Before leaving the concept of Methodological Solipsism, here's how one
commentator reviewed the situation in the context of the work of
perhaps psychology's best known radical behaviourist:
'Meanings Are Not 'In the Head'
Skinner has developed a case for this claim in the book,
VERBAL BEHAVIOR (1957), and elsewhere, where he
maintains that meaning, rather than being a property of
an utterance itself, is to be found in the nature of the
relationship between occurrence of the utterance and its
context. It is important enough to put in his own words.
..meaning is not properly regarded as a property either
of a response or a situation but rather of the
contingencies responsible for both the topography of
behavior and the control exerted by stimuli. To take a
primitive example, if one rat presses a lever to obtain
food when hungry while another does so to obtain water
when thirsty, the topographies of their behaviors may be
indistinguishable, but they may be said to differ in
meaning: to one rat pressing the lever 'means food'; to
the other it 'means' water. But these are aspects of the
contingencies which have brought behavior under the
control of the current occasion. Similarly, if a rat is
reinforced with food when it presses the lever in the
presence of a flashing light but with water when the
light is steady, then it could be said that the flashing
light means food and the steady light means water, but
again these are references not to some property of the
light but to the contingencies of which the lights have
been parts.
The same point may be made, but with many more
implications, in speaking of the meaning of verbal
behavior. The over-all function of the behavior is
crucial. In an archetypal pattern a speaker is in
contact with a situation to which a listener is disposed
to respond but with which he is not in contact. A verbal
response on the part of the speaker makes it possible
for the listener to respond appropriately. For example,
let us suppose that a person has an appointment, which
he will keep by consulting a clock or a watch. If none
is available, he may ask someone to tell him the time,
and the response permits him to respond effectively...
*The meaning of a response for the speaker* includes the
stimulus which controls it (in the example above, the
setting on the face of a clock or watch) and possibly
aversive aspects of the question, from which a response
brings release. *The meaning for the listener* is close
to the meaning the clock face would have if it were
visible to him, but it also includes the contingencies
involving the appointment, which make a response to the
clock face or the verbal response probable at such a
time..
One of the unfortunate implications of communication
theory is that the meanings for speaker and listener are
the same, that something is made common to both of them,
that the speaker conveys an idea or meaning, transmits
information, or imparts knowledge, as if his mental
possessions then become the mental possessions of the
listener. There are no meanings which are the same in
the speaker and listener. Meanings are not independent
entities...
Skinner, 1974, pp.90-2
One does not have to take Skinner's word alone, however,
for much current philosophical work also leads to the
conclusion that meanings are not in the head. The issue
extends beyond the problem of meaning construed as a
linguistic property to the problem of intensionality and
the interpretation of mentality itself. While the
reasoning behind this claim is varied and complex,
perhaps an analogy with machine functions can be helpful
here. A computer is a perfect example of a system that
performs meaningless syntactic operations. The
electrical configuration of the addressable memory
locations is just formal structures, without semantic
significance to the computer either as numbers or as
representations of numbers. All the computer does is
change states automatically as electrical current runs
through its circuits. Despite the pure formality of its
operations, however, the computer (if designed and
programmed correctly) will be truth-preserving across
computations: ask the thing to add 2 + 2 and it will
give you a 4 every time. But the numerical meanings we
attach to the inputs and outputs do not enter into and
emanate from the computer itself. Rather, they remain
outside the system, in the interpretations that we as
computer users assign to the inputs and outputs of the
machine's operations. Now, if one is inclined to a
computational view of mind, then by analogy much the
same thing holds for the organic computational systems
we call our brains. Meanings are not in them, but exist
in the mode through which they in their functioning
stand to the world.
Ironies begin to mount here. Brentano's claim that
'Intentionality' is the mark of the mental is now widely
accepted. Intentionality in its technical sense has to
do with the meaningfulness, the semantic context of
mental states. But the argument is now made that
cognitive operations and their objects are formal and
syntactic only, and do not themselves have semantic
context (e.g. see Putnam, 1975; Fodor, 1980; and Stich,
1983, for a range of contributions to this viewpoint).
Semantic issues do not concern internal mental
mechanisms but concern the mode of relation between
individuals and their worlds. Such issues are not really
psychological at all, it is claimed, and are relegated
to other fields of inquiry for whatever elucidation can
be brought to them. For example, while belief is a
canonical example of a mental, intentional state, Stich
says, 'believing that p is an amalgam of historical,
contextual, ideological, and perhaps other
considerations' (1983, p.170). The net result of these
recent moves in cognitive psychology and the philosophy
of mind seems to be that the essence of mentality - its
meaningfulness - is in the process of being disowned by
modern mentalism! But Stich's ashbin of intentionality -
historical and contextual considerations - is exactly
what behaviorism seeks to address. Can it be that
BEHAVIORISM will be the instrument called for final
explication of Brentano's thesis of the mental? One's
head spins to think it.
R. Schnaitter (1987)
Knowledge as Action: The Epistemology of Radical Behaviorism
In B. F. Skinner Consensus and Controversy
Eds. S. Modgil and C. Modgil
The reawakening of interest in connectionism in the early to mid 1980s
can indeed be seen as a vindication of the basic principles of
behaviourism. What is psychological may well be impenetrable, for any
serious scientific purposes, not because it is in any way a different
kind of 'stuff', but because structurally it amounts to no more than
an n-dimensional weight space, idiosyncratic and context specific, to
each and every one of us.
SECTION B: Methodological (Evidential) Behaviourism
The Perspective From 'The Extensional Stance
'Suppose that each line of the truth table for the
conjunction of all [of a person's] beliefs could be
checked in the time a light ray takes to traverse the
diameter of a proton, an approximate "supercycle" time,
and suppose that the computer was permitted to run for
twenty billion years, the estimated time from the "big-
bang dawn of the universe to the present. A belief
system containing only 138 logically independent
propositions would overwhelm the time resources of this
supermachine.'
C. Cherniak (1986)
Minimal Rationality p.93
'Cherniak goes on to note that, while it is not easy to
estimate the number of atomic propositions in a typical
human belief system, the number must be vastly in excess
of 138. It follows that, whatever its practical benefits
might be, the proposed consistency-checking algorithm is
not something a human brain could even approach. Thus,
it would seem perverse, to put it mildly, to insist that
a person's cognitive system is doing a bad job of
reasoning because it fails to periodically execute the
algorithm and check on the consistency of the person's
beliefs.'
S. Stich (1990)
The Fragmentation of Reason p.152
'I should like to see a new conceptual apparatus of a
logically and behaviourally straightforward kind by
which to formulate, for scientific purposes, the sort of
psychological information that is conveyed nowadays by
idioms of propositional attitude.'
W V O Quine (1978)
In the extract from Cherniak, the point being made is that as the
number of discrete propositions increase, the possible combinations
increases dramatically, or, as Shafir and Tversky 1992 say:
'Uncertain situations may be thought of as disjunctions
of possible states: either one state will obtain, or
another....
Shortcomings in reasoning have typically been attributed
to quantitative limitations of human beings as
processors of information. "Hard problems" are typically
characterized by reference to the "amount of knowledge
required," the "memory load," or the "size of the search
space"....Such limitations, however, are not sufficient
to account for all that is difficult about thinking. In
contrast to many complicated tasks that people perform
with relative ease, the problems investigated in this
paper are computationally very simple, involving a
single disjunction of two well defined states. The
present studies highlight the discrepancy between
logical complexity on the one hand and psychological
difficulty on the other. In contrast to the "frame
problem" for example, which is trivial for people but
exceedingly difficult for AI, the task of thinking
through disjunctions is trivial for AI (which routinely
implements "tree search" and "path finding" algorithms)
but very difficult for people. The failure to reason
consequentially may constitute a fundamental difference
between natural and artificial intelligence.'
E. Shafir and A. Tversky (1992)
Thinking through Uncertainty: Nonconsequantial Reasoning
and Choice
Cognitive Psychology 24,449-474
From a pattern recognition or classification stance, it is known that
as the number of predicates increase, the number of linearly separable
functions becomes proportionately smaller as is made clear by the
following extract from Wasserman (1989) when discussing the concept of
linear separability:
'We have seen that there is no way to draw a straight
line subdividing the x-y plane so that the exclusive-or
function is represented. Unfortunately, this is not an
isolated example; there exists a large class of
functions that cannot be represented by a single-layer
network. These functions are said to be linearly
inseparable, and they set definite bounds on the
capabilities of single-layer networks.
Linear separability limits single-layer networks to classification
problems in which the sets of points (corresponding to input values)
can be separated geometrically. For our two-input case, the separator
is a straight line. For three inputs, the separation is performed by a
flat plane cutting through the resulting three-dimensional space. For
four or more inputs, visualisation breaks down and we must mentally
generalise to a space of n dimensions divided by a "hyperplane", a
geometrical object that subdivides a space of four or more
dimensions.... A neuron with n binary inputs can have 2 exp n
different input patterns, consisting of ones and zeros. Because each
input pattern can produce two different binary outputs, one and zero,
there are 2 exp 2 exp n different functions of n variables.
As shown [below], the probability of any randomly selected function
being linearly separable becomes vanishingly small with even a modest
number of variables. For this reason single-layer perceptrons are, in
practice, limited to simple problems.
n 2 exp 2 exp n Number of Linearly Separable Functions
1 4
2 16 14
3 256 104
4 65,536 1,882
5 4.3 x 10 exp 9 94,572
6 1.8 x 10 exp 19 5,028,134
P. D. Wasserman (1989)
Linear Separability: Ch2. Neural Computing Theory and Practice
In later sections evidence is presented in the context of clinical vs.
actuarial judgment that human judgement is severely limited to
processing only a few variables. Beyond that, non- linear fits become
more frequent. This is discussed later in the context of connectionist
'intuitive',inductive inference and constraints on short-term or
working memory span (c.f. Kyllonen & Christal 1990 - "Reasoning
Ability Is (LIttle More Than) Working-Memory Capacity?!"), but it is
worth mentioning here that in the epilogue to their expanded re-print
of their 1969 review of neural nets 'Perceptrons - An Introduction to
Computational Geometry', after reiterating their original criticism
that neural networks had only been shown to be capable of solving 'toy
problems', ie problems with a small number of dimensions, using 'hill
climbing' algorithms, Minsky and Papert (1988) effectively did a
'volte face' and said:
'But now we propose a somewhat shocking alternative:
Perhaps the scale of the toy problem is that on which,
in physiological actuality, much of the functioning of
intelligence operates. Accepting this thesis leads into
a way of thinking very different from that of the
connectionist movement. We have used the phrase "society
of mind" to refer to the idea that mind is made up of a
large number of components, or "agents," each of which
would operate on the scale of what, if taken in
isolation, would be little more than a toy problem.'
M Minsky and S Papert (1988) p266-7
and a little latter, which is very germane to the fragmentation of
behaviour view being advanced in this volume:
'On the darker side, they [parallel distributed
networks] can limit large-scale growth because what any
distributed network learns is likely to be quite opaque
to other networks connected to it.'
ibid p.274
This *opacity* of aspects, or elements, of our own behaviour to
ourselves is central to the theme being developed in this volume,
namely that a science of behaviour must remain entirely extensional
and that there can not therefore be a science or technology of
psychology to the extent that this remains intensional (Quine
1960,1992). The discrepancy between experts' reports of the
information they use when making diagnoses (judgments) is reviewed in
more detail in a later section, however, research reviewed in Goldberg
1968, suggests that even where diagnosticians are convinced that they
use more than additive models (ie use interactions between variables -
which statistically may account for some of the non-linearities),
empirical evidence shows that in fact they only use a few linear
combinations of variables (cf. Nisbett and Wilson 1977, in this
context).
As an illustration of methodological solipsism (intensionalism) in
practice consider the following which neatly contrasts subtle
difference between the methodological solipsist approach and that of
the methodological or 'evidential' behaviourist.
Several years ago, a prison psychologist sought the views of prison
officers and governors as to who they considered to be 'subversives'.
Those considered 'subversive' were flagged 1, those not considered
subversive were flagged 0. The psychologist then used multiple
regression to predict this classification from a number of other
behavioural variables. From this he was able to produce an equation
which predicted subversiveness as a function of 4 variables: whether
or not the inmate had a firearms offence history, the number of
reports up to arrival at the current prison, the number of moves up to
arrival where the inmate had stayed more than 28 days, and the number
of inmate assaults up to arrival.
Note that the dependent variable was binary, the inmate being
classified as 'subversive' or 'not subversive'. The prediction
equation, which differentially weighted the 4 variables, therefore
predicted the dependent variable as a value between 0 and 1. Now the
important thing to notice here is that the behavioural variables were
being used to predict something which is essentially a propositional
attitude, ie the degree of certainty of the officers beliefs that
certain inmates were subversive.
The methodological solipsist may well hold that the officer's beliefs
are what are important, however, the methodological behaviourist would
hold that what the officers thought was just *an approximation of what
the actual measures of inmate behaviour represented*, ie his thoughts
were just vague, descriptive terms for inmates who had lots of
reports, assaulted inmates and had been moved through lots of prisons,
and were probably in prison for violent offences. What the officers
thought was not perhaps, all that important, since we could just go to
the records and identify behaviours which are characteristic of
troublesome behaviour and then identify inmates as a function of those
measures (cf. Williams and Longley 1986).
In the one case the concern is likely to be with developing better and
better predictors of what staff THINK, and in the other, it becomes a
matter of simply recording better measures of classes of behaviour and
empirically establishing functional relations between those classes.
In the case of the former, intensional stance, one becomes interested
in the *psychology* of those exposed to such factors (ie those exposed
to the behaviour of inmates, and what they *vaguely or intuitively
describe it as)*. From the extensional stance (methodological
behaviourist) defended in these volumes, such judgments can only be a
**function** of the data that staff have had access to. From the
extensional stance, one is simply interested in recording *behaviour*
itself and deducing implicit relations. Ryle (1949) and many
influential behaviourists since (Quine 1960), have, along with Hahn
(1933) suggested that this is our intellectual limit anyway:
'It is being maintained throughout this book that when
we characterize people by mental predicates, we are not
making untestable inferences to any ghostly processes
occurring in streams of consciousness which we are
debarred from visiting; we are describing the ways in
which those people conduct parts of their predominantly
public behaviour.'
G. Ryle
The Concept of Mind (1949)
Using regression technology as outlined above is essentially how
artificial neural network software is used to make classifications, in
fact, there is now substantial evidence to suggest that the two
technologies are basically one and the same (Stone 1986), except that
in neural network technology, the regression variable weights are
opaque to the judge, cf. Kosko (1992):
'These properties reduce to the single abstract property
of *adaptive model-free function estimation*:Intelligent
systems adaptively estimate continuous functions from
data without specifying mathematically how outputs
depend on inputs...A function f, denoted f: X Y, maps
an input domain X to an output range Y. For every
element x in the input domain X, the function f uniquely
assigns the element y to the output range Y.. Functions
define causal hypotheses. Science and engineering paint
our pictures of the universe with functions.
B. Kosko (1992)
Neural Networks and Fuzzy Systems: A Dynamical Systems
Approach to Machine Intelligence p 19.
The rationale behind Sentence Management as outlined in the paper
'What are Regimes?' (Longley 1992) and in section D below, is that the
most effective way to bring about sustained behaviour change is not
through specific, formal training programmes, but through a careful
strategy of apposite allocation to activities which *naturally require
the behavioural skills* which an inmate may be deficient in. This
depends on standardised recording of activity and programme behaviour
*throughout sentence* which will provide a *historical and actuarial,
record of attainment.* This will provide differential information to
guide management's decisions as how best to help inmates lead a
constructive life whilst in custody, and, hopefully, after release.
Initially, it will serve to support actuarial analysis of behaviour as
a practical working, inmate, and management, information system. In
time, it should provide data to enable managers to focus resources
where they are most required (ie provide comprehensive regime
profiles, which highlight strong and weak elements). Such a system is
only interested in what inmates 'think' or 'believe' to the extent
that what they 'think' and 'believe' are specific skills which the
particular activities and programmes require, and which can therefore
be systematically assessed as criteria of formative behaviour
profiling. What is required for effective decision making and
behaviour management is a history of behavioural performance in
activities and programmes, much like the USA system of Grade Point
Averages and attendance. All such behaviours are the natural skills
required by the activities and programmes, and all such assessment is
criterion reference based.
The alternative, intensional approach, of asking staff to identify
risk factors from the documented account of the offence, and
subsequently asking staff to look out for them in the inmate's prison
behaviour may well only serve to shape inmates to inhibit
(conditionally suppress) such behaviour, especially if their
progression through the prison system is contingent on this. However,
from animal studies of acquisition-extinction-reacquisition, there is
no evidence that such behaviour inhibition is likely to produce a
*permanent* change in the inmate's behaviour in the absence of the
inmate *learning new behaviours*. Such an approach is also blind to
base rates of behaviours. Only through a system which encouraged the
acquisition of *new* behaviours can we expect there to be a change in
risk, and even this would have to be *actuarially* determined. For a
proper estimate of risk, one requires a system where inmates can be
assessed with respect to standard demands of the regime. The standard
way to determine risk factors was to derive these from *statistical*
*analysis,* not from *clinical (intensional) judgement*.
Much of the rationale for this stance can be deduced from the
following. Throughout the 20th century, psychologists' evaluation of
the extent to which reasoning can be formally taught has been
pessimistic. From Thorndike (1913) through Piaget (see Brainerd 1978)
to Newell (1980) it has been maintained that:
'the modern.....position is that learned problem-solving
skills are, in general, idiosyncratic to the task.'
A. Newell 1980.
Furthermore, it has been argued that whilst people may in fact use
abstract inferential rules, these rules can not be formally taught to
any significant degree. They are learned instead under natural
conditions of development and cannot be improved by formal
instruction. This is essentially Piaget's position.
The above is, in fact, how Nisbett et al (1987) opened their SCIENCE
paper '*Teaching Reasoning*'. Reviewing the history of the concept of
formal discipline which looked to the use of latin and the classics to
train the 'muscles of the mind', Nisbett et. al provided some
empirical evidence on the degree to which one class of inferential
rules can be taught. They describe these rules as 'a family of
pragmatic inferential rule systems that people induce in the context
of solving recurrent everyday problems'. These include "causal
schemas", "contractual schemas" and "statistical heuristics". The
latter are clearly instances of inductive rather than deductive
inference.
Nisbett et. al. clearly pointed out that the same can not be said for
the teaching of deductive inference (i.e. formal instruction in
deductive logic or other syntactic rule systems). With respect to the
teaching of logical reasoning, Nisbett et. al. had the following to
say:
'Since highly abstract statistical rules can be taught
in such a way that they can be applied to a great range
of everyday life events, is the same true of the even
more abstract rules of deductive logic? We can report no
evidence indicating that this is true, and we can
provide some evidence indicating that it is not.....In
our view, when people reason in accordance with the
rules of formal logic, they normally do so by using
pragmatic reasoning schemas that happen to map onto the
solutions provided by logic.'
ibid. p.628
Such 'causal schemas' are known as 'intensional heuristics' (Agnoli
and Krantz 1989) and have been widely studied in psychology since the
early 1970s, primarily by research psychologists such as Tversky and
Kahneman (1974), Nisbett and Ross (1980), Kahneman, Slovic and Tversky
(1982), Holland et. al (1986) and Ross and Nisbett (1991).
A longitudinal study by Lehman and Nisbett (1990) looked at
differential improvements in the use of such heuristics in college
students classified by different subject groups. They found
improvements in the use of statistical heuristics in social science
students, but no improvement in conditional logic (such as the Wason
selection task). Conversely, the natural science and humanities
produced significant improvements in conditional logic. Interestingly,
there were no changes in students studying chemistry. Whilst the
authors took the findings to provide some support for their thesis
that reasoning can be taught, it must be appreciated that the findings
at the same time lend considerable support to the view that each
subject area inculcates its own particular type of reasoning, even in
highly educated individuals. That is, the data lend support to the
thesis that training in particular skills must look to training for
transfer and application within particular skill areas. This is
elaborated below in the context of the system of Sentence Management.
Today, formal modelling of such intensional processes is researched
using a technology known as 'Neural Computing' which uses inferential
statistical technologies closely related to regression analysis.
However, such technologies are inherently inductive. They take samples
and generalise to populations. They are at best pattern recognition
systems.
Such technologies must be contrasted with formal deductive logical
systems which are algorithmic rather than heuristic (extensional
rather than intensional). The algorithmic, or computational, approach
is central to classic Artificial Intelligence and is represented today
by the technology of relational databases along with rule and
Knowledge Information Based System (KIBS) which are based on the First
Order Predicate Calculus, the Robinson Resolution Principle (Robinson
1965,1979) and the long term objectives of automated reasoning (e.g.
Wos et. al 1992 and the Japanese Fifth Generation computing project) -
see Volume 2 and 3.
The degree to which intensional heuristics can be suppressed by
training is now controversial (Kahneman and Tversky 1983; Nisbett and
Ross 1980; Holland et al. 1986; Nisbett et al 1987; Agnoli and Krantz
1989; Gladstone 1989; Fong and Nisbett 1991; Ploger and Wilson 1991;
Smith et al 1992). In fact, the degree to which they are or are not
may be orthogonal to the main theme of this paper, since the main
thrust of the argument is that behaviour science should look to
deductive inferential technology, not inductive inference. Central to
the controversy, however, is the degree to which the suppression is
sustained, and the degree of generalisation and practical application
of even 'statistical heuristics'. For example, Ploger and Wilson
(1991) said in commentary on the 1991 Fong and Nisbett paper:
'G. T. Fong and R. E. Nisbett argued that, within the
domain of statistics, people possess abstract rules;
that the use of these rules can be improved by training;
and that these training effects are largely independent
of the training domain. Although their results indicate
that there is a statistically significant improvement in
performance due to training, they also indicate that,
even after training, most college students do not apply
that training to example problems.
D. Ploger & M. Wilson
Statistical reasoning: What is the role of inferential rule training?
Comment on Fong and Nisbett.
Journal of Experimental Psychology General; 1991 Jun Vol
120(2) 213-214
Furthermore, Gladstone (1989) criticises the stance adopted by the
same group in an article in American Psychologist (1988):
'[This paper]' criticizes the assertion by D. R. Lehman
et al. that their experiments support the doctrine of
formal discipline. The present author contends that the
work of Lehman et al. provides evidence that one must
teach for transfer, not that transfer occurs
automatically. The problems of creating a curriculum and
teaching it must be addressed if teachers are to help
students apply a rule across fields. Support is given to
E. L. Thorndike's (1906, 1913) assessment of the general
method of teaching for transfer.'
R. Gladstone (1989)
Teaching for transfer versus formal discipline.
American Psychologist; 1989 Aug Vol 44(8) 1159
What this research suggests is that whilst improvements can be made by
training in formal principles (such as teaching the 'Law of Large
Numbers'), this does not in fact contradict the stance of Piaget and
others that most of these inductive skills are in fact learned under
natural lived experience ('erlbnis' and 'lebenswelt' Husserl 1952, or
'Being-in-the-world' Heidegger 1928). Furthermore, there is evidence
from short term longitudinal studies of training in such skills that
not only is there a decline in such skills after even a short time,
but there is little evidence of application of the heuristics to novel
problem situations outside the training domain. This is the standard
and conventional criticism of 'formal education'. Throughout this
work, the basic message seems to be to focus training on specific
skills acquisition which will not so much generalise to novel
contexts, but find application in other, similar if not identical
contexts.
Most recently, Nisbett and colleagues have looked further at the
criteria for assessing the efficacy of cognitive skills training:
'A number of theoretical positions in psychology
(including variants of case-based reasoning, instance-
based analogy, and connectionist models) maintain that
abstract rules are not involved in human reasoning, or
at best play a minor role. Other views hold that the use
of abstract rules is a core aspect of human reasoning.
The authors propose 8 criteria for determining whether
or not people use abstract rules in reasoning. They
examine evidence relevant to each criterion for several
rule systems. There is substantial evidence that several
inferential rules, including modus ponens, contractual
rules, causal rules, and the law of large numbers, are
used in solving everyday problems. Hybrid mechanisms
that combine aspects of instance and rule models are
considered.'
E. E. Smith, C. Langston and R. E. Nisbett:
The case for rules in reasoning.
Cognitive Science; 1992 Jan-Mar Vol 16(1) 1-40
We use rules, it can be argued, when we apply extensionalist
strategies which are of course, by design, domain specific. Note that
in the history of logic it took until 1879 to discover Quantification
Theory. Furthermore, research on deductive reasoning itself suggests
strongly that the view developed in this volume is sound:
'Reviews 3 types of computer program designed to make
deductive inferences: resolution theorem-provers and
goal-directed inferential programs, implemented
primarily as exercises in artificial intelligence; and
natural deduction systems, which have also been used as
psychological models. It is argued that none of these
methods resembles the way in which human beings usually
reason. They [humans] appear instead to depend, not on
formal rules of inference, but on using the meaning of
the premises to construct a mental model of the relevant
situation and on searching for alternative models of the
premises that falsify putative conclusions.'
P. N. Johnson-Laird
Human and computer reasoning.
Trends in Neurosciences; 1985 Feb Vol 8(2) 54-57
'Contends that the orthodox view in psychology is that
people use formal rules of inference like those of a
natural deduction system. It is argued that logical
competence depends on mental models rather than formal
rules. Models are constructed using linguistic and
general knowledge; a conclusion is formulated based on
the model that maintains semantic information, expresses
it parsimoniously, and makes explicit something not
directly stated by the premise. The validity of the
conclusion is tested by searching for alternative models
that might refute the conclusion. The article summarizes
a theory developed in a 1991 book by P. N. Johnson-Laird
and R. M. Byrne.'
P. N. Johnson-Laird & R. M. Byrne
Precis of Deduction.
Behavioral and Brain Sciences; 1993 Jun Vol 16(2) 323-
380
That is, human reasoning tends to focus on content or intension. As
has been argued elsewhere, such heuristic strategies invariably suffer
as a consequence of their context specificity and constraints on
working memory capacity.
--
David Longley
Anders N Weinstein
In article <Pine.SOL.3.91.961211083356.13273A-100000@tortoise>,
David Yeo <dyeo@tortoise> wrote:
|I think there are two senses of "algorithm" being interwoven here. While
|it is true that heuristics are procedures (hence algorithms), algorithms
|usually also carry the implication that they will achieve the solution to
|a problem in a finite number of steps. To quote Rosenberg (1986):
|
| Algorithm: a finite set of well-defined rules for the solution of a
| problem in a finite number of steps; for example, a full
| statement of an arithmetic procedure for evaluating the sin
| X to a stated precision.
|
| (Dictionary of Artificial Intelligence and Robotics)
|
|Heuristics (rules of thumb) do not offer this degree of certainty; in fact
|they don't guarantee they'll achieve the solution to a problem at all. I
|suggest that it's this latter (non-procedural) sense of algorithm that AI
|researchers imply when distinguishing "algorithms" from "heuristics".
I have noticed this usage in the AI world -- does anyone know
whence it derives? My guess would be Newell and Simon.
It has always struck me as very unfortunate, and occasionally even a
barrier to communication. In this narrower sense, for example, Deep
Thought does not have an "algorithm" for choosing chess moves. Yet we
pretty clearly also want to say that Deep Thought and every other
symbolic computer program are by definition algorithms in the broader
sense. Wouldn't it be better to use a different term, like "decision
procedure" (more generally, maybe "solution procedure") for algorithmic
routines that meet this further criterion?
On another note: I heard Dan Dennett once claim that natural selection
is in effect employing an algorithm to search design space, perhaps one
using a randomizing element. I believe this claim is also made in his
Darwin book. I simply do not know what to make of this claim, but am
very strongly inclined to doubt its meaningfulness. Any opinions on
*that* usage?
Peter Lupton
Anders N Weinstein writes:
> Peter Lupton writes:
> >There does seem to be an on-going problem with accounting
> >for the causal relevance of non-reductive accounts. Which is
>
> True.
So. There are problems with non-reductive property-dualistic
theories of mind. Which haven't been solved. In which case,
I'd suggest it is a little premature to start trumpeting the
early demise of reductive approaches.
> >That is, the argument that we can't give naturalistic necessary
> >and sufficient conditions for "is a table" is neither here nor there
> >from the point of view of success at naturalising the mental.
> >What is needed is not anything so ridiculous, but something
> >deeper but probably easier: an account of why and how such
> >classifications occur. That is, an account, not of the detail, but
> >of the system as a whole.
>
> I guess one might look for this as well in Kant's architechtonic of
> pure reason or Hegel's phenomenology of Geist or Heidegger's
> existential analytic of Dasein, or Husserl's transcendental
> phenomenology or whatever other philosophical work floats your boat.
> Why does it have to be *natural science* that answers such questions?
People are trying to make sense of the idea that mental states
can be causal whilst accepting the causal closure of natural
science.
Seems to me thumptingly obvious that mental states, qua mental states
can be causes. You may accuse me of a certain lack of innocence,
but there you go. I reckon it is pretty damn obvious that some
mental states cause some behaviour.
> >What I'm pointing out is that physicalism has more stings to its bow
> >than the rather sorry tale above would seem to imply. I find this
> >rather strange because Anders has told the same tale before and I have
> >pointed out that physicalism is a mansion with more rooms before now.
>
> I have not seen a well worked out new room. I guess I admit it is
> a project to work it out. My beef is with the likes of Fodor or Churchland,
> who seem to presuppose some largely inarticulate version of physicalism.
Well, why didn't you simply say so?
> >At any rate, until we know better the relationship between levels
> >of language and how they're formed, I still don't see how you
>
> I think we already know *how* to use the psychological language quite
> well. What we lack is a clear reflective understanding of what we do a
> thousand times a day. The examples we need to consider, however, are
> already open to view.
Which is merely to change the subject from 'the relationship between'
to 'the examples we need to consider'.
> >As I say, basic physics comes in two favours: forces and fields
> >on the one hand and thermodynamics on the other. Don't forget
> >that when one is talking about physicalism, they're both relevant.
> >
> >If we knew how our vocabularies were formed, perhaps you would
>
> I don't understand this emphasis on how they were formed. I think we
> can understand, e.g. the form of life of the a particular animal
> synchronically, wihout looking to how it was formed. Ditto, perhaps for
> such things as the custom of using signposts, and for our vocabularies.
>
> Even if we do need to look to history to understand these things, I don't
> see that that will answer the question about relations between different
> vocabularies.
>
> >do better at showing some sort of sharp divide between the
> >language of basic physics and the rest. However, since this isn't
> >something we know much about, we're not sure whether the
> >boundaries you wish to draw have any greater significance than
> >any other mark in the sand of philosophical speculation.
>
> The Kantian distinction between acting in accordance with rules and
> acting according to a conception of rules seems to me to be a good one.
> Or more generally the distinction between normative and non-normative.
I think I'll simply repeat what I already said above.
--
Cheers,
Pete Lupton
Neil Rickert
>Full text is available at:
> http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
>Connectionist systems, it is claimed, do not represent knowledge as
>production rules, ie as well-formed-formulae represented in the syntax
>of the predicate calculus (using conditionals, modus ponens, modus
>tollens and the quantifiers), but as connection weights between
>activated predicates in a parallel distributed network:
That is just a way of saying that we do not understand how
connectionist systems represent knowledge. It follows that we do not
know exactly what they do represent, nor do we know whether they are
following any particular heuristic.
What it amounts to, is that (A) you don't know what the brain is
doing, and (B) you don't know what an ANN is doing. From that you
are jumping to the conclusion that they must both be doing the same
sort of thing. Even if the conclusion is correct, you still do not
have any evidence to support claims that people are following the
availability heuristic, for example. As an extensionalist, you ought
to be insisting on having the rules for that heuristic written down.
Then you ought to insist on programming it, and testing whether the
performance of the program is comparable to the performance of
humans. Without such testing, the evidence for these heuristics is
excessively weak.
Neil Rickert
In <Pine.SOL.3.91.961211083356.13273A-100000@tortoise> David Yeo <dyeo@tortoise> writes:
>On 11 Dec 1996, Neil Rickert wrote:
>> For AI, we require something more than
>> handwaving claims that there are heuristics. For AI folk, there
>> ain't a heuristic until there is an explicit program implementing
>> it. And that means that there ain't a heuristic. If you are such
>I think there are two senses of "algorithm" being interwoven here. While
>it is true that heuristics are procedures (hence algorithms), algorithms
>usually also carry the implication that they will achieve the solution to
>a problem in a finite number of steps. To quote Rosenberg (1986):
> Algorithm: a finite set of well-defined rules for the solution of a
> problem in a finite number of steps; for example, a full
> statement of an arithmetic procedure for evaluating the sin
> X to a stated precision.
> (Dictionary of Artificial Intelligence and Robotics)
>Heuristics (rules of thumb) do not offer this degree of certainty; in fact
>they don't guarantee they'll achieve the solution to a problem at all. I
>suggest that it's this latter (non-procedural) sense of algorithm that AI
>researchers imply when distinguishing "algorithms" from "heuristics".
This seems a little confused. I'll agree that heuristics are not
guaranteed to produce a solution. But it does not follow that they
are non-procedural. It only follows that the procedure is not
guaranteed to succeed.
>> !@#$%& brilliant know-it-all, just post in full detail the godammed
>> program that implements these heuristics. It will need to be an
>> extensional (GOFAI) program, since an ANN system cannot be seen to
>> satisfy these heuristics except by handwaving arguments that are not
>> persuasive.
>Why do heuristics necessarily require GOFAI? I may be missing some subtle
>point you are making (your posting carries considerable baggage from past
>discussions with David Longley, so I may be inappropriately jumping into a
>longstanding, and apparently venomous, debate), but it would seem that
>ANN's are the embodiment of "rules of thumb".
A heuristic that cannot be programmed is useless to AI. If there are
rules of thumb, they ought to be statable, and programmable. If you
cannot state what they are well enough to program them into a
repeatable procedure, then the claim that there is a heuristic is
empty talk.
Why would it seem that the ANNs are an embodiment of rules of thumb?
I don't get it. How can we be sure that there be rules of thumb, if
we don't know what those rules are?
Longley wants to insist on extensional talk. I want to know what is
the extension of these heuristics. It is possible to make an
attribution which claims that the ANN is following rules of thumb.
But the claim is empty, in the sense that there is no means to
establish whether the ANN is actually following those rules. In a
GOFAI system, you can be sure whether or not the system is following
the rules.
If we could really write down explicitly these rules of thumb that
people are supposedly following when they supposedly use a heuristic,
we should be able to program them. Then we should be able to carry
out controlled experiments, comparing the judgement of the computer
systems following these rules of thumb, and the judgements of the
people who are said to be following something like the availability
heuristic. It is my conjecture that, if we could ever do this, we
would find that the human judgement is better than that of the
presumed heuristics. But of course the controlled experiment cannot
be carried out, unless you can reduce the presumed heuristic to a set
of rules to follow, and then automate the following of those rules.
> That is, as the product of
>induction ANN's would clearly seem to be compatible with many standard
>definitions of the term "heuristic". To again quote Rosenberg (1986):
> Heuristic: pertaining to exploratory methods of problem solving in
> which solutions are discovered by the evaluation of progress
> made toward the final result. (ibid)
However, this definition does not work unless you know in advance
what is the final result you want, and unless you have some means of
evaluating the progress made. If you have a clear way of evaluating
progress made, and if you know what is the final result wanted, you
should be able to use GOFAI methods. It seems to me that Longley is
using the term "heuristic" where the above definition could not be
applicable.
Neil Rickert
>On another note: I heard Dan Dennett once claim that natural selection
>is in effect employing an algorithm to search design space, perhaps one
>using a randomizing element. I believe this claim is also made in his
>Darwin book.
Yes, he does make some such claim in his "Darwin's Dangerous Ideas."
> I simply do not know what to make of this claim, but am
>very strongly inclined to doubt its meaningfulness. Any opinions on
>*that* usage?
Unless I missed it, Dennett never explicitly states what the
algorithm is. In that sense, I think you are right that it lacks
meaning. It suffers from the same inexplicitness problem as
Longley's use of "heuristic".
Neil Rickert
In <850323...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley) writes:
>In article <58lld4$5...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>> When I am disagreeing with
>> the usual interpretation of well known evidence, I am do not need to
>> cite results which counter that evidence. I need only show that the
>> interpretation I am criticizing is riddled with holes, or lacks the
>> evidence to make it credible.
>Do you not think I (and others) think it somewhat incredible that
>the ONLY references you cite (by name only) are the ones I have
>documented specifically? You make it quite clear by what you
>write that you have not read the material before, nor have you a
>grasp of the overall literature which it is part of. Now that's
>not abuse, just fact.
You make it clear that you are a rigid dogmatist. You cannot
distinguish between disagreement and failure to read. I have not
gone around with large lists of citations, because the articles are
not particularly relevant. Many of them document that human behavior
cannot be explained as the execution of a formal logic based on
beliefs. It is certainly valuable that psychologists have documented
this. But it says no more than should have been the common
observation of anybody with his eyes open. That is, it should have
been common knowlege all along, although perhaps never formally
documented. One should not need to provide large citation lists when
dealing with what should be common knowledge.
>> If you don't care for what I write, you can always go back to
>> psychology newsgroups. ...
>More nonsense and abuse - and anyone who has read "Fragments"
>will see it as exactly that.
The abuse is in response to your abuse.
>The above is just gibberish. I have made very explicit references
>to the work in decision theory by Gluck and Bower on this matter,
>and I have repeated the oft stated remark that the R-W model is
>the most influential model over the past 25 years or so.
I really don't care how influential it is. It is still not
extensional by my standards. You cannot tell what a stimulus is, or
what a response is, without having one of those defective (i.e.
intensional or intentional) humans make the judgement. The theory
cannot account for a Copernicus, or a Newton, a Darwin or an
Einstein.
> Now IF
>you understood the literature, and if you understood what I was
>saying, you would see the relationship between what I have said
>about the Extensional Stance (as opposed to the intentional
>stance), the use of normative descriptive statistics (including
>regression and its variants in pattern recognition), the problems
>of biases of natural assessments (which can be modelled by ANNs)
>and a host of other key areas of research which are covered in
>"Fragments".
That was a long sentence. What it boils down to is: If I understood
the literature, I would agree with David Longley.
There is a natural corollary. If Copernicus had understood the
literature, he would have been a Ptolemaic. If Darwin had understood
the literature he would have been a creationist. If Wegener had
understood the literature, he would have known that continents do not
drift.
So why don't we just close all of the science labs. If we all know
the literature, we will realize that the conventional wisdom could
never be wrong, and that there is no point in doing any more
research.
>> >> >Truth is not determined by convention.
Except when Longley decides which are the conventions.
David Longley
In article <Pine.SOL.3.91.961211083356.13273A-100000@tortoise>
dyeo@tortoise "David Yeo" writes:
> On 11 Dec 1996, Neil Rickert wrote:
>
> > For AI, we require something more than
> > handwaving claims that there are heuristics. For AI folk, there
> > ain't a heuristic until there is an explicit program implementing
> > it. And that means that there ain't a heuristic. If you are such
>
> I think there are two senses of "algorithm" being interwoven here. While
> it is true that heuristics are procedures (hence algorithms), algorithms
> usually also carry the implication that they will achieve the solution to
> a problem in a finite number of steps. To quote Rosenberg (1986):
>
> Algorithm: a finite set of well-defined rules for the solution of a
> problem in a finite number of steps; for example, a full
> statement of an arithmetic procedure for evaluating the sin
> X to a stated precision.
>
> (Dictionary of Artificial Intelligence and Robotics)
>
> Heuristics (rules of thumb) do not offer this degree of certainty; in fact
> they don't guarantee they'll achieve the solution to a problem at all. I
> suggest that it's this latter (non-procedural) sense of algorithm that AI
> researchers imply when distinguishing "algorithms" from "heuristics".
>
> > !@#$%& brilliant know-it-all, just post in full detail the godammed
> > program that implements these heuristics. It will need to be an
> > extensional (GOFAI) program, since an ANN system cannot be seen to
> > satisfy these heuristics except by handwaving arguments that are not
> > persuasive.
>
> Why do heuristics necessarily require GOFAI? I may be missing some subtle
> point you are making (your posting carries considerable baggage from past
> discussions with David Longley, so I may be inappropriately jumping into a
> longstanding, and apparently venomous, debate), but it would seem that
> ANN's are the embodiment of "rules of thumb". That is, as the product of
> induction ANN's would clearly seem to be compatible with many standard
> definitions of the term "heuristic". To again quote Rosenberg (1986):
>
> Heuristic: pertaining to exploratory methods of problem solving in
> which solutions are discovered by the evaluation of progress
> made toward the final result. (ibid)
>
> Cheers,
>
> - David Yeo (Applied Cognitive Science, University of Toronto)
>
>
Thanks David.
That's how I believed I treated heuristics in "Fragments", and
how I thought I'd elaborated on that treatment when I posted an
extract from Gluck and Bower's work a few months back when
discussing the merits of the Rescrola-Wagner type of model as an
ANN antecedent. The Tversky & Kahneman (1973;1982) heuristics are
not normative, and I think the same can be said of the basic
mechanisms of classicial conditioning. They are too prone to the
biasing effects of stimulus intensity and local characteristics
of data distributions (context efffects).
I discussed some of these issues (which I consider to be
"Naturalised Epistemology") with a highly respected Quinean
philosopher earlier in the year when I was working on digitising
the 12 hours of "In Conversation" interviews with Quine. At that
time I felt what what I had to say was pretty consistent with the
direction which research on these issues was taking.
Clearly Rickert does not agree. But then, he seems to make a
point of not agreeing with much that I have to say. I mean no
offence to him when I say that I do not believe he understands
the relevant literature.
--
David Longley
Anders N Weinstein
In article <25632...@luptonpj.demon.co.uk>,
Peter Lupton <lup...@luptonpj.demon.co.uk> wrote:
>Anders N Weinstein writes:
>
>> Peter Lupton writes:
>
>> >for the causal relevance of non-reductive accounts. Which is
>>
>> True.
>
>theories of mind. Which haven't been solved. In which case,
>I'd suggest it is a little premature to start trumpeting the
>early demise of reductive approaches.
Perhaps I should have said: the *appearance* of problems.
Which might require therapy to dissolve, rather than
constructive philosophy to solve.
BTW I should say the very same problems appear to afflict, e.g.
computational accounts in psychology. Computationalist are certainly
non-reductive property-dualistic as far as I can tell, and ought to
have the same problems about causal relevance of computational properties.
(This is why I feel I can take it as read that physicalism is false --
because it's too strong to allow computational accounts as an
autonomous higher level of explanation. I just happen to believe
organismic intentionality involves a higher level still, but accept a lot
of the general anti-reductionist schema.)
>People are trying to make sense of the idea that mental states
>can be causal whilst accepting the causal closure of natural
>science.
If I understand the position of Jerry Fodor in the "Special Sciences"
paper reprinted in his _Language of Thought_, then the natural sciences
are *not* causally closed. It is only explanation at the most basic
level that is "causally closed". The laws of special sciences are
gappy, and the exceptions are a shapeless class from their vantage point.
Similarly, there can be intrusions into the higher level from below,
e.g. hardware events which interfere with the software, which are
non-systematic from the higher-level vantage point.
>Seems to me thumptingly obvious that mental states, qua mental states
>can be causes. You may accuse me of a certain lack of innocence,
>but there you go. I reckon it is pretty damn obvious that some
>mental states cause some behaviour.
Of course this is right -- worrying excessively about something
might simply cause me to lose my concentration and perform badly or
distractedly on some present task. That is a case of what we might
call non-rational causation by mental states.
I'm not sure rational causation -- persons acting or trying to act for
good reasons -- is the same sort of phenomenon. I would reiterate: what
is be "damn obvious" is that I sometimes go to the store *because* I
want to buy milk. What is not "damn obvious" to me that we should reify
my state of wanting milk and enquire about what it might be, or how it
could cause my behavior. Nor is it damn obvious that this requires
causal *laws*.
>> >At any rate, until we know better the relationship between levels
>> >of language and how they're formed, I still don't see how you
>>
>> I think we already know *how* to use the psychological language quite
>> well. What we lack is a clear reflective understanding of what we do a
>> thousand times a day. The examples we need to consider, however, are
>> already open to view.
>
>Which is merely to change the subject from 'the relationship between'
>to 'the examples we need to consider'.
OK.
I think the relation is similar to that between speaking of ink marks
and speaking of meaningful letters and words, between speaking of
paintings as furniture and as aesthetic objects, between speaking of
hammers as hunks of wood and metal and speaking of them as tools with a
conventionalized use, between speaking of bodily organs as aggregates
of atoms and speaking of them as things with a proper function to
perform, between speaking of a rod as a piece of metal and speaking of
it as a lever, between speaking of pieces of paper and speaking of
money, between considering a forest as a forest and as, say, a hiding
place or a rich natural resource.
[BTW I would recommend Stephen Mulhall's _On Being-in-the-World:
Wittgenstein and Heidegger on Seeing Aspects_ for an interesting
discussion of the treatment of "seeing-as" in these two philosophers.]
Whether or not these are all the same, instances of Aristotle's
form-matter structure, still I guess second sort of aspect involves
relationship to a larger organized whole and also a relationship to
life which the first sort does not. Perhaps if we need a label we can
dub the non-reductive philosophy "contextualism". One idea to consider
is that the paradigmatic methods of the natural sciences are based on
considering properties of objects which are independent from these
contexts ("decontextualizing"). Biological function would be a
borderline case, although its status as natural science as
been controversial.
Jim Balter
In article <58kjoh...@aden.ai.cs.yale.edu>,
Drew McDermott <mcdermo...@cs.yale.edu> wrote:
>
>To summarize: everyone in this newsgroup agrees that Chalmers is wrong. I had
>hoped for more controversy.
I saw a couple of postings agreeing with him. And Chalmers has defended
himself and has been defended in sci.psychology.consciousness, which is archived at
http://www.ai.sri.com/~connolly/psyche-list
Also, several supporting articles have been published in JCS (Journal of
Consciousness Studies). But if you are looking for a defense of Chalmers, I
don't know how you could find anything better than his book itself.
As for why many people agree that Chalmers is wrong, perhaps that's because
Chalmers *is* wrong. :-) If you accept Chalmers' premise that conscious
experience is a primitive essence, that if you don't have it then "all is dark
inside" and "there is nothing it is like to be" you, then you will reach the
conclusion that conscious experience is a primitive essence, and then it is a
matter of coming up with explanations of it. At least Penrose's microtubule
explanation is motivated by what he takes to be an informal proof (which you
have rebutted as well as anyone) that some non-algorithmic mechanism is
necessary (which is not Chalmers' conclusion; I am comparing methodology, not
equating arguments) . But all Chalmers has shown is that, if you sneak in
assumptions that "the mental doesn't logically supervene on the physical"
(e.g., he holds that zombies are logically possible because "I can't discern a
contradiction"), then of course you will conclude same.
There is really a more fundamental problem with Chalmers: his dualistic thesis
has no bite; nothing can possibly come it. If zombies are logically possible,
then we can do the exact same science and engineering that our zombie
counterparts would do, and come up with exactly the same results. Since
physical laws describe the physical world, Chalmers' psychophysical laws
describe nothing additional, aside from personal, subjective, "conscious
experience"; but by his thesis of organizational invariance, the very fact
that his psychophysical laws apply in this world (according to him), doing the
right things physically automatically brings about consciousness. Of course,
by Ockham's Razor, we might as well ignore Chalmers' thesis, since we would do
just as well without it (our zombie counterparts certainly would, since we
know Chalmers' thesis is false in their world). But there's no need to cut up
his book :-), since he covers the subject well and does a good job of
defending functionalism (under the guise of "organizational invariance") and
strong AI.
As Daniel Dennett writes in "Facing Backwards on the Problem of Consciousness",
Until Chalmers gives us an *independent* ground for contemplating the drastic
move of adding 'experience' to mass, charge, and space-time, his proposal is
one that can be put on the back burner.
--
<J Q B>
Peter_...@uk.ibm.com
In <58ntg7$5...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>In article <25632...@luptonpj.demon.co.uk>,
>Peter Lupton <lup...@luptonpj.demon.co.uk> wrote:
>>Anders N Weinstein writes:
>>
>>> Peter Lupton writes:
>>
>>> >for the causal relevance of non-reductive accounts. Which is
>>>
>>> True.
>>
>>theories of mind. Which haven't been solved. In which case,
>>I'd suggest it is a little premature to start trumpeting the
>>early demise of reductive approaches.
>
>Perhaps I should have said: the *appearance* of problems.
>Which might require therapy to dissolve, rather than
>constructive philosophy to solve.
>
>BTW I should say the very same problems appear to afflict, e.g.
>computational accounts in psychology. Computationalist are certainly
>non-reductive property-dualistic as far as I can tell,
I simply don't see that this should be taken as anywhere near
certain. Since it is part of physicalism to incorporate the structures
of thermodynamics, there is, on this account, an over-arching rather
*abstract* and, essentially, information-theoretic structure which falls
under the rubric of physicalism. Now. Given that, it is by no means *certain*
that the sorts of structures typical of computation *must* fall outside
of the physicalist scheme.
What you say Anders, is simply far too *hasty*.
> and ought to
>have the same problems about causal relevance of computational properties.
I would agree with you that an unduly restrictive account of cause
will cause problems. However, it seems obvious that any argument
that takes as a hypothesis the claim that causes are nothing but
equations of motions operating on initial conditions is, very likely,
in error. And most of the arguments which go against the causal
relevance of higher level structure are of that sort.
Cause for me is very much more closely related to information.
Which will, one can expect, productively engage with the
thermodynamical aspect of physicalism. So I expect that these
insights (which are quite modern) will be seen as increasingly
important.
>(This is why I feel I can take it as read that physicalism is false --
>because it's too strong to allow computational accounts as an
>autonomous higher level of explanation. I just happen to believe
>organismic intentionality involves a higher level still, but accept a lot
>of the general anti-reductionist schema.)
Probably what you're describing as 'physicalism' *is* false.
However, if by 'physicalism' one means 'that which is part of
the science of physics', then there a whole other area of
exploration to be performed. Since this area of exploration
(the thermodynamic area) has only just been developed (the
way in which thermodynamics should be extended to large-scale
and complex structure has only recently been started), you are
simply being far too *hasty*.
>>People are trying to make sense of the idea that mental states
>>can be causal whilst accepting the causal closure of natural
>>science.
>
>If I understand the position of Jerry Fodor in the "Special Sciences"
>paper reprinted in his _Language of Thought_, then the natural sciences
>are *not* causally closed. It is only explanation at the most basic
>level that is "causally closed". The laws of special sciences are
>gappy, and the exceptions are a shapeless class from their vantage point.
>Similarly, there can be intrusions into the higher level from below,
>e.g. hardware events which interfere with the software, which are
>non-systematic from the higher-level vantage point.
There is a certain amount of what Fodor says in that paper
that I agree with. However, Fodor simply hasn't got to grips
with information (that is, information theory) and its role
in cause generally and its role in natural science. So Fodor has,
like most people, a limited view of what natural science is
and what the causal closure of natural science amounts to.
However, I would agree with you that *if* we take 'natural
science' to mean 'inital conditions and equations of motion of
fundamental particles', *then* I'd agree with you and Fodor
that *that* isn't causally closed. But then I would say that's
because 'natural science' simply isn't restricted in that way.
>>Seems to me thumptingly obvious that mental states, qua mental states
>>can be causes. You may accuse me of a certain lack of innocence,
>>but there you go. I reckon it is pretty damn obvious that some
>>mental states cause some behaviour.
>
>Of course this is right -- worrying excessively about something
>might simply cause me to lose my concentration and perform badly or
>distractedly on some present task. That is a case of what we might
>call non-rational causation by mental states.
>
>I'm not sure rational causation -- persons acting or trying to act for
>good reasons -- is the same sort of phenomenon. I would reiterate: what
>is be "damn obvious" is that I sometimes go to the store *because* I
>want to buy milk. What is not "damn obvious" to me that we should reify
>my state of wanting milk and enquire about what it might be, or how it
>could cause my behavior. Nor is it damn obvious that this requires
>causal *laws*.
But it is our ordinary way of speaking, is it not, that a false belief,
is a *cause* of a person doing something apparently foolish, say.
And that the receipt of a piece of information can *cause* a person's
subsequent change of behaviour.
So I think these causal links *do* require further philosophical
investigation. And I don't think it is true to say that this relationship
has been adequately accounted for by Ryle, Witt'n and Anscombe.
And until that day - which may never happen - I reckon alternatives
are going to be investigated. That is, you really can't castigate others
for investigating areas while holding to a view that, in at least one
substantial area, is woefully incomplete. To observe that other areas
are incomplete at the same point is no defence, it seems to me. Work
needs to be done, we simply don't know where and how answers
will be found, different people will think that different avenues of
exploration will turn up trumps.
My own suspicion is that unless we can produce a reductive account of
beliefs and desires, we'll find that we can't be realist about the mind
as cause. Which, for me, makes your views a rather implausible
avenue of investigation. However, when I look around at the various
reductive accounts, they seem to be enormously crude in comparison
to parsimony as a reductive approach. And it does appear that parsimony
makes contact with physicalism through the other, relatively unexplored
(at least wrt this philosophical area) branch of physics: thermodynamics
and entropy. Now *this* branch of physicalism is overdue as a means to
augment our physicalist understanding of *cause*, which seems to have
become stuck in the 'equations of motion' arm. Thermodynamics and entropy
describe an intrinsically more abstract structure, one that is involved in
state per se and how that state per se tends to evolve in specific, yet
typically aggregative ways, rather than with the detail of modes of
interactions.
Cheers,
Pete Lupton
Jim Balter
In article <850224...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <58i2sg$7...@usenet.srv.cis.pitt.edu>
> ande...@pitt.edu "Anders N Weinstein" writes:
>
>>
>> Longley's scrupulous avoidance of personal abuse stands as a model
>> for us all.
>>
>Perhaps it should!
>
>succumb to soplipsism or don't base what they have to say on
>relevent literature. That's not abuse either.
I feel abused by this affirmation of the consequent.
No amount of citation of that which arguably is not abuse
will do as a demonstration that Longley is not abusive.
The first posting I ever saw from Longley was a lengthy scurrilous attack on
Aaron Sloman, filled with lies and hypocrisy. He has repeated similar abuse
against others over and over again. Perhaps this would be more tolerable if
he had interesting things to say, but he's got one note, which he repeats over
and over again at great length and with repeated long quotations, which is a
form of abuse in its own right. He isn't interested in any sort of exchange
or interaction, yet continues to snipe. When Drew McDermott explicitly asked
to avoid Longley's sort of sniping, how many expected Longley to do so anyway?
I certainly did, and sure enough Drew ended up commenting on it. Basically
Longley is a sociopath. What do you do with someone like that in a free forum
like this? You can ignore him, or you can return fire. I've tried both, and
neither is very satisfying.
--
<J Q B>
David Longley
> In <850325...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley)
> writes:
>
> >Full text is available at:
>
> > http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
>
> >Connectionist systems, it is claimed, do not represent knowledge as
> >production rules, ie as well-formed-formulae represented in the syntax
> >of the predicate calculus (using conditionals, modus ponens, modus
> >tollens and the quantifiers), but as connection weights between
> >activated predicates in a parallel distributed network:
>
> That is just a way of saying that we do not understand how
> connectionist systems represent knowledge. It follows that we do not
> know exactly what they do represent, nor do we know whether they are
> following any particular heuristic.
>
> What it amounts to, is that (A) you don't know what the brain is
> doing, and (B) you don't know what an ANN is doing. From that you
> are jumping to the conclusion that they must both be doing the same
> sort of thing. Even if the conclusion is correct, you still do not
> have any evidence to support claims that people are following the
> availability heuristic, for example. As an extensionalist, you ought
> to be insisting on having the rules for that heuristic written down.
> Then you ought to insist on programming it, and testing whether the
> performance of the program is comparable to the performance of
> humans. Without such testing, the evidence for these heuristics is
> excessively weak.
>
>
You just keep telling me you don't understand. You are starting
to see the questions that folk in this domain are asking but you
clearly don't see that *your* questions to me are very probably
only a consequence of having a vague grasp of what I have already
told you here. Apropos the above, the work of Gluck and Bower
(1988;1990) is referenced in section 9 of "Fragments".
I have been drawing this picture for a long time....
--
David Longley
David Longley
In article <58n7up$7...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>
> Longley wants to insist on extensional talk. I want to know what is
> the extension of these heuristics. It is possible to make an
> attribution which claims that the ANN is following rules of thumb.
> But the claim is empty, in the sense that there is no means to
> establish whether the ANN is actually following those rules. In a
> GOFAI system, you can be sure whether or not the system is following
> the rules.
>
I do not insist on extensional talk. What I have said is that for
science and technology the empirical evidence is in favour of
extensionally based systems over intensional heuristics. I have
then tried to outline the dynamic relation between the two.
> If we could really write down explicitly these rules of thumb that
> people are supposedly following when they supposedly use a heuristic,
> we should be able to program them. Then we should be able to carry
> out controlled experiments, comparing the judgement of the computer
> systems following these rules of thumb, and the judgements of the
> people who are said to be following something like the availability
> heuristic. It is my conjecture that, if we could ever do this, we
> would find that the human judgement is better than that of the
> presumed heuristics. But of course the controlled experiment cannot
> be carried out, unless you can reduce the presumed heuristic to a set
> of rules to follow, and then automate the following of those rules.
The above is a muddle and just shows that you have NOT read the
literature, nor paid attention to what I have said in the past.
There are programs which end up producing the same biases in
classification that one finds in natural human decision making. I
have not only reviewed that literature but cited long extracts
from some of the recent work by Gluck and Bower just to bring the
point home with respect to the place of conditioning as a modus
operandi of folk psychological decision making.
Now *that* should surprise a few folk!
--
David Longley
Jim Balter
In article <58n1t5$s...@usenet.srv.cis.pitt.edu>,
Anders N Weinstein <ande...@pitt.edu> wrote:
>In article <Pine.SOL.3.91.961211083356.13273A-100000@tortoise>,
>David Yeo <dyeo@tortoise> wrote:
>|it is true that heuristics are procedures (hence algorithms), algorithms
>|usually also carry the implication that they will achieve the solution to
>|a problem in a finite number of steps. To quote Rosenberg (1986):
>|
>| Algorithm: a finite set of well-defined rules for the solution of a
>| problem in a finite number of steps; for example, a full
>| statement of an arithmetic procedure for evaluating the sin
>| X to a stated precision.
>|
>| (Dictionary of Artificial Intelligence and Robotics)
>|
>|Heuristics (rules of thumb) do not offer this degree of certainty; in fact
>|they don't guarantee they'll achieve the solution to a problem at all. I
>|suggest that it's this latter (non-procedural) sense of algorithm that AI
>|researchers imply when distinguishing "algorithms" from "heuristics".
>
>whence it derives? My guess would be Newell and Simon.
>
>It has always struck me as very unfortunate,
It would be, if it were true.
>and occasionally even a
>barrier to communication. In this narrower sense, for example, Deep
>Thought does not have an "algorithm" for choosing chess moves. Yet we
>pretty clearly also want to say that Deep Thought and every other
>symbolic computer program are by definition algorithms in the broader
>sense. Wouldn't it be better to use a different term, like "decision
>procedure" (more generally, maybe "solution procedure") for algorithmic
>routines that meet this further criterion?
There is a confusion here. The uncertainty of a heuristic is not in its
reaching a conclusion, but its being a solution to a given problem. Deep
Thought has an algorithm for choosing chess moves, as does every chess
program, but it does not have an algorithm for winning chess, although its
algorithm is "closer" to being an algorithm for winning the game of chess than
most chess programs have. Likewise, there are linear (or some low degree)
heuristics for solving the traveling saleman problem, even though the problem
is NP-complete. The point here is that the linear algorithm doesn't actually
solve the traveling salesman problem, but it does get the right answer for
many (most?) "interesting" cases, and gets close in others. So it's an
algorithm for a problem which approximates the traveling saleman problem.
Another example is "zero-knowledge proofs", where someone can "prove" they
know something without revealing it, by answering questions where they would
have a .5 chance of getting the right answer to. Enough questions "prove" as
closely as desired. In formal terms, this algorithm is no proof procedure at
all, but so much the worse for formalisms (and Lucas/Penrose/Godel).
The confusion about "heuristic" seems, as so often, to stem from intensional
biases. An algorithm is an algorithm, even when it isn't an algorithm for the
problem *we* wanted solved. Dennett explores this idea in some detail in
_Darwin's Dangerous Idea_.
>On another note: I heard Dan Dennett once claim that natural selection
>is in effect employing an algorithm to search design space, perhaps one
>using a randomizing element. I believe this claim is also made in his
>Darwin book. I simply do not know what to make of this claim, but am
>very strongly inclined to doubt its meaningfulness. Any opinions on
>*that* usage?
Perhaps if you read the book .... Certainly the processes of mutation,
genetic combination, ontogeny, etc. appear algorithmic (well, there could be
QM via microtubules getting into the act, but ...) and natural selection is a
matter of designs that fit environmental conditions succeeding and designs
that don't not succeeding, which amounts to a search. People are writing
genetic algorithms these days that certainly seem to do something of this
sort. Perhaps you could say what part of the idea you are having trouble
understanding.
--
<J Q B>
Jim Balter
In article <58na6q$7...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>In <58n1t5$s...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>
>>On another note: I heard Dan Dennett once claim that natural selection
>>is in effect employing an algorithm to search design space, perhaps one
>>using a randomizing element. I believe this claim is also made in his
>>Darwin book.
>
>
>>very strongly inclined to doubt its meaningfulness. Any opinions on
>>*that* usage?
>
>algorithm is. In that sense, I think you are right that it lacks
>meaning. It suffers from the same inexplicitness problem as
>Longley's use of "heuristic".
I think I can safely say that it means something to say that Window '95
employs an algorithm, even though I can't state what it is and it definitely
seems to have random elements.
--
<J Q B>
Jim Balter
In article <58n7up$7...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>In <Pine.SOL.3.91.961211083356.13273A-100000@tortoise> David Yeo <dyeo@tortoise> writes:
>>On 11 Dec 1996, Neil Rickert wrote:
>
>>> For AI, we require something more than
>>> handwaving claims that there are heuristics. For AI folk, there
>>> ain't a heuristic until there is an explicit program implementing
>>> it. And that means that there ain't a heuristic. If you are such
>
>>it is true that heuristics are procedures (hence algorithms), algorithms
>>usually also carry the implication that they will achieve the solution to
>>a problem in a finite number of steps. To quote Rosenberg (1986):
>
>> Algorithm: a finite set of well-defined rules for the solution of a
>> problem in a finite number of steps; for example, a full
>> statement of an arithmetic procedure for evaluating the sin
>> X to a stated precision.
>
>> (Dictionary of Artificial Intelligence and Robotics)
>
>>Heuristics (rules of thumb) do not offer this degree of certainty; in fact
>>they don't guarantee they'll achieve the solution to a problem at all. I
>>suggest that it's this latter (non-procedural) sense of algorithm that AI
>>researchers imply when distinguishing "algorithms" from "heuristics".
>
>guaranteed to produce a solution. But it does not follow that they
>are non-procedural. It only follows that the procedure is not
>guaranteed to succeed.
They aren't guaranteed to succeed at *what we wanted of them*.
Intensionality rules.
>>> !@#$%& brilliant know-it-all, just post in full detail the godammed
>>> program that implements these heuristics. It will need to be an
>>> extensional (GOFAI) program, since an ANN system cannot be seen to
>>> satisfy these heuristics except by handwaving arguments that are not
>>> persuasive.
>
>>Why do heuristics necessarily require GOFAI? I may be missing some subtle
>>point you are making (your posting carries considerable baggage from past
>>discussions with David Longley, so I may be inappropriately jumping into a
>>longstanding, and apparently venomous, debate), but it would seem that
>>ANN's are the embodiment of "rules of thumb".
>
>A heuristic that cannot be programmed is useless to AI. If there are
>rules of thumb, they ought to be statable, and programmable. If you
>cannot state what they are well enough to program them into a
>repeatable procedure, then the claim that there is a heuristic is
>empty talk.
I don't see this at all. If we were to succeed in programming something
equivalent to a human being, it would certainly learn and employ heuristics
that hadn't been programmed.
>Why would it seem that the ANNs are an embodiment of rules of thumb?
>I don't get it. How can we be sure that there be rules of thumb, if
>we don't know what those rules are?
I know some rules of thumb that some people use, even though I didn't program
them. But perhaps that's the point: given a system complex enough to do the
sorts of things we want of it, it will start employing heuristics that we
didn't give it. And some of those may well be the sorts of heuristics that
Longley rails against, but which prove so effective in human affairs.
At some point, we don't know what's going on inside the machine.
Ho, intensionality!
>Longley wants to insist on extensional talk. I want to know what is
>the extension of these heuristics. It is possible to make an
>attribution which claims that the ANN is following rules of thumb.
>But the claim is empty, in the sense that there is no means to
>establish whether the ANN is actually following those rules. In a
>GOFAI system, you can be sure whether or not the system is following
>the rules.
Right; Longley cannot both escape the intensional and escape the narrow
confines of explicit algorithms. And for some of us, the whole idea of AI is
to go beyond a fancy calculator.
>If we could really write down explicitly these rules of thumb that
>people are supposedly following when they supposedly use a heuristic,
>we should be able to program them. Then we should be able to carry
>out controlled experiments, comparing the judgement of the computer
>systems following these rules of thumb, and the judgements of the
>people who are said to be following something like the availability
>heuristic. It is my conjecture that, if we could ever do this, we
>would find that the human judgement is better than that of the
>presumed heuristics. But of course the controlled experiment cannot
>be carried out, unless you can reduce the presumed heuristic to a set
>of rules to follow, and then automate the following of those rules.
Well, we perhaps could program a machine to learn to develop heuristics and
employ them, and we could check that it is really employing them. Of course,
these again aren't heuristics we gave it, and could well have all the "flaws"
of our own.
>> That is, as the product of
>>induction ANN's would clearly seem to be compatible with many standard
>>definitions of the term "heuristic". To again quote Rosenberg (1986):
Sounds like an affirmation of the consequent.
>> Heuristic: pertaining to exploratory methods of problem solving in
>> which solutions are discovered by the evaluation of progress
>> made toward the final result. (ibid)
This makes little sense; how does an evaluation of progress make for the
discovery of a solution? Presumably what is meant is that a measure of
progress is used to guide the exploration, hopefully reaching the solution.
However, this is only one sort of heuristic, and it is a notoriously
non-robust one. Think of monkeys trapped by holding a goody through a small
hole in a coconut and not being able free their closed fist (is this
apocryphal? it doesn't matter). Even adding backtracking doesn't generally
lead to solutions in a reasonable amount of time.
>However, this definition does not work unless you know in advance
>what is the final result you want, and unless you have some means of
>evaluating the progress made. If you have a clear way of evaluating
>progress made, and if you know what is the final result wanted, you
>should be able to use GOFAI methods. It seems to me that Longley is
>using the term "heuristic" where the above definition could not be
>applicable.
The goals can be general and context dependent, and the heuristics can be
dynamically discovered during the search. But then we wouldn't know what the
heck was going on inside that thing. Ooops; sounds downright intensional.
--
<J Q B>
David Longley
> In <58n1t5$s...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein)
> writes:
>
> >On another note: I heard Dan Dennett once claim that natural selection
> >is in effect employing an algorithm to search design space, perhaps one
> >using a randomizing element. I believe this claim is also made in his
> >Darwin book.
>
> Yes, he does make some such claim in his "Darwin's Dangerous Ideas."
>
> > I simply do not know what to make of this claim, but am
> >very strongly inclined to doubt its meaningfulness. Any opinions on
> >*that* usage?
>
> Unless I missed it, Dennett never explicitly states what the
> algorithm is. In that sense, I think you are right that it lacks
> meaning. It suffers from the same inexplicitness problem as
> Longley's use of "heuristic".
>
It's inexplicit basically *because* it is intensional and content
bound. In the end, I think it may be helpful for folk to think of
"heuristic" in the same way as adventious reinforcement or just
conditioning.
This is why I keep making the point that logical extensional
(effective) analysis which is normative, has little to do with
psychology.
--
David Longley
Jim Balter
In article <850396...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <58n7up$7...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>>
>> the extension of these heuristics. It is possible to make an
>> attribution which claims that the ANN is following rules of thumb.
>> But the claim is empty, in the sense that there is no means to
>> establish whether the ANN is actually following those rules. In a
>> GOFAI system, you can be sure whether or not the system is following
>> the rules.
>>
>
No, you just *abuse* people when they use "intensional idioms".
*This* response, for instance, is the sort that you frequently
take as "evidence" that I "haven't read the literature"
or have various sorts of misunderstandings. When anyone
points out that this is hypocritical or dishonest,
*that too* gets an abusive response.
>What I have said is that for
>science and technology the empirical evidence is in favour of
>extensionally based systems over intensional heuristics. I have
>then tried to outline the dynamic relation between the two.
Often you say "All I have said is ...". Pointing out that this is an
outrageous lie receives the same sort of abuse.
This has gone one for over a year; Longley responds to "you do X" with
"No, I do Y", as if one contradicted the other. This sort of dishonesty
is abusive.
>> If we could really write down explicitly these rules of thumb that
>> people are supposedly following when they supposedly use a heuristic,
>> we should be able to program them. Then we should be able to carry
>> out controlled experiments, comparing the judgement of the computer
>> systems following these rules of thumb, and the judgements of the
>> people who are said to be following something like the availability
>> heuristic. It is my conjecture that, if we could ever do this, we
>> would find that the human judgement is better than that of the
>> presumed heuristics. But of course the controlled experiment cannot
>> be carried out, unless you can reduce the presumed heuristic to a set
>> of rules to follow, and then automate the following of those rules.
>
>The above is a muddle and just shows that you have NOT read the
>literature, nor paid attention to what I have said in the past.
It shows no such thing, you abusive twit.
>There are programs which end up producing the same biases in
>classification that one finds in natural human decision making.
So what? Other than as an affirmation of the consequent.
>I
>have not only reviewed that literature but cited long extracts
>from some of the recent work by Gluck and Bower just to bring the
>point home with respect to the place of conditioning as a modus
>operandi of folk psychological decision making.
So what? Other than as an affirmation of the consequent.
>Now *that* should surprise a few folk!
*What* should surprise them? If your grammmar weren't so contorted perhaps
you wouldn't have so much cause to think that people haven't "paid attention"
to what you write.
--
<J Q B>
Neil Rickert
In <jqbE2B...@netcom.com> j...@netcom.com (Jim Balter) writes:
>In article <58n7up$7...@ux.cs.niu.edu>, Neil Rickert <ric...@cs.niu.edu> wrote:
>>A heuristic that cannot be programmed is useless to AI. If there are
>>rules of thumb, they ought to be statable, and programmable. If you
>>cannot state what they are well enough to program them into a
>>repeatable procedure, then the claim that there is a heuristic is
>>empty talk.
>I don't see this at all. If we were to succeed in programming something
>equivalent to a human being, it would certainly learn and employ heuristics
>that hadn't been programmed.
It is not at all certain that it would learn and employ heuristics
that could then be reduced to an explicit procedure. And if we could
not reduce the actions of the system to an explicit procedure, then
we would be merely guessing in saying that it was using a heuristic.
>>Why would it seem that the ANNs are an embodiment of rules of thumb?
>>I don't get it. How can we be sure that there be rules of thumb, if
>>we don't know what those rules are?
>I know some rules of thumb that some people use, even though I didn't program
>them.
There really is a problem with taking human uses of rules of thumb as
examples of heuristics. The problem is, that we sometimes use these
rules of thumb, and we sometimes don't. You would need another rule
that says when to apply the rule of thumb, before you could use it as
a heuristic.
Neil Rickert
In <850398...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley) writes:
>In article <58na6q$7...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>> Unless I missed it, Dennett never explicitly states what the
>> algorithm is. In that sense, I think you are right that it lacks
>> meaning. It suffers from the same inexplicitness problem as
>> Longley's use of "heuristic".
>It's inexplicit basically *because* it is intensional and content
>bound. In the end, I think it may be helpful for folk to think of
>"heuristic" in the same way as adventious reinforcement or just
>conditioning.
In other words, your meaning for the word "heuristic" is very
different from that used in AI and computer science. Since this is
an AI newsgroup, I think you should stick to the AI usage of the
term. Perhaps that would reduce the confusion.
Anders N Weinstein
In article <jqbE2A...@netcom.com>, Jim Balter <j...@netcom.com> wrote:
>In article <58n1t5$s...@usenet.srv.cis.pitt.edu>,
>Anders N Weinstein <ande...@pitt.edu> wrote:
>>is in effect employing an algorithm to search design space, perhaps one
>>I simply do not know what to make of this claim, but am
>>very strongly inclined to doubt its meaningfulness.
>
Well I looked at it in the bookstore. But it's not the kind of book I pay
money for.
>genetic combination, ontogeny, etc. appear algorithmic (well, there could be
>QM via microtubules getting into the act, but ...) and natural selection is a
>matter of designs that fit environmental conditions succeeding and designs
>that don't not succeeding, which amounts to a search. People are writing
>genetic algorithms these days that certainly seem to do something of this
>sort. Perhaps you could say what part of the idea you are having trouble
>understanding.
The problem is, I understand the abstract conception of a computation or
algorithm applied to formal or symbolic elements, such as strings on a
TM tape. I have trouble when one tries to see how to apply it to
arbitrary physical objects.
I would have thought that Niagara Falls is *not* employing an algorithm
to get water to the bottom, the stomach is not employing an algorithm
to digest food, the moon is not employing an algorithm to stay in its
orbit etc. Of course the states of these processes can be mapped onto
states of a formal computation trace, indeed they can be mapped onto an
infinite number of algorithms if we want. But that seems to me a
Pickwickian notion of instantiating an algorithm.
So I am wondering if one should say that evolution instantiates an
algorithm in the Pickwickian sense in which molecules in a cup of coffee
instantiate the Wordstar program, or in some more substantial sense.
I am also wondering about the glib treatment of randomness. I don't
think the issue is mainly about the possibility of indeterminism at the
micro-level. Processes at the biological level might be
non-law-governed when characterized in those terms, even if processes
described at the lower level are deterministic. Just as the disturbance
of a computation in my machine by a power dip might be a result of
deterministic laws of physics but not systematically explicable at the
algorithmic level.
Again Dennett's claim seems to blur the distinction I would want
between a non-systematic effect on a process, and a systematic
procedure that happens to employ a randomizing element.
David Yeo
On 11 Dec 1996, Neil Rickert wrote:
> In <Pine.SOL.3.91.961211083356.13273A-100000@tortoise> David Yeo writes:
> >I think there are two senses of "algorithm" being interwoven here. While
> >it is true that heuristics are procedures (hence algorithms), algorithms
> >usually also carry the implication that they will achieve the solution to
> >a problem in a finite number of steps. To quote Rosenberg (1986):
>
> > Algorithm: a finite set of well-defined rules for the solution of a
> > problem in a finite number of steps; for example, a full
> > statement of an arithmetic procedure for evaluating the sin
> > X to a stated precision.
>
> > (Dictionary of Artificial Intelligence and Robotics)
>
> >Heuristics (rules of thumb) do not offer this degree of certainty; in fact
> >they don't guarantee they'll achieve the solution to a problem at all. I
> >suggest that it's this latter (non-procedural) sense of algorithm that AI
> >researchers imply when distinguishing "algorithms" from "heuristics".
>
> This seems a little confused. I'll agree that heuristics are not
> guaranteed to produce a solution. But it does not follow that they
> are non-procedural. It only follows that the procedure is not
> guaranteed to succeed.
I did not mean to imply that heuristics are non-procedural (see my second
sentence - "While it is true that heuristics are procedures ... ." ), I
only meant that the feature which distinguishes algorithms and heuristics
is the extra burden algorithms carry re: achieving a solution in a finite
number of steps. The confusion you note arises from the tendency of some
to inappropriately confine/reduce algorithms to their procedural dimension.
> >Why do heuristics necessarily require GOFAI? I may be missing some subtle
> >point you are making (your posting carries considerable baggage from past
> >discussions with David Longley, so I may be inappropriately jumping into a
> >longstanding, and apparently venomous, debate), but it would seem that
> >ANN's are the embodiment of "rules of thumb".
>
> A heuristic that cannot be programmed is useless to AI. If there are
> rules of thumb, they ought to be statable, and programmable. If you
> cannot state what they are well enough to program them into a
> repeatable procedure, then the claim that there is a heuristic is
> empty talk.
While I agree with your first sentence, I don't think that it follows that
the rules of thumb must be statable. Indeed, that is one of the strengths
(and weaknesses) of ANNs; the heuristics (rules of thumb) that they derive
are often both unanticipated and difficult to summarise.
> Why would it seem that the ANNs are an embodiment of rules of thumb?
> I don't get it. How can we be sure that there be rules of thumb, if
> we don't know what those rules are?
Because ANNs discover (learn) solutions to the problems with which they
are presented as they explore the problem space. Moreover, the feedback
which underlies an ANNs weight adjustments can, I think, be validly seen
as evaluation of progress made toward the final result. This accords well
with the (cited) definition of a "heuristic" offered by Rosenberg.
<snip>
> If we could really write down explicitly these rules of thumb that
> people are supposedly following when they supposedly use a heuristic,
> we should be able to program them. Then we should be able to carry
> out controlled experiments, comparing the judgement of the computer
> systems following these rules of thumb, and the judgements of the
> people who are said to be following something like the availability
> heuristic.
Again, the inability to "write down explicitly these rules of thumb" does
not preclude the ability to "carry out controlled experiments". Everyday
many (in my view) interesting experiments are carried out using ANNs.
> heuristic. It is my conjecture that, if we could ever do this, we
> would find that the human judgement is better than that of the
> presumed heuristics. But of course the controlled experiment cannot
> be carried out, unless you can reduce the presumed heuristic to a set
> of rules to follow, and then automate the following of those rules.
Sorry, but I don't agree. In many ways your view would seem to be a
variant of what Ryle (1949) called the "intellectualist legend":
According to the legend, whenever an agent does anything intelligently,
his act is preceded and steered by another internal act of considering
a regulative proposition appropriate to his practical problem.
(The Concept of Mind, p. 31)
David Longley
In article <jqbE2A...@netcom.com> j...@netcom.com "Jim Balter" writes:
>
> The first posting I ever saw from Longley was a lengthy scurrilous attack on
> Aaron Sloman, filled with lies and hypocrisy. He has repeated similar abuse
> against others over and over again. Perhaps this would be more tolerable if
> he had interesting things to say, but he's got one note, which he repeats over
> and over again at great length and with repeated long quotations, which is a
> form of abuse in its own right. He isn't interested in any sort of exchange
> or interaction, yet continues to snipe. When Drew McDermott explicitly asked
> to avoid Longley's sort of sniping, how many expected Longley to do so anyway?
> I certainly did, and sure enough Drew ended up commenting on it. Basically
> Longley is a sociopath. What do you do with someone like that in a free forum
> like this? You can ignore him, or you can return fire. I've tried both, and
> neither is very satisfying.
> --
> <J Q B>
>
It's time you grew up and learned the difference between creative writing and
science.
--
David Longley
David Longley
> No amount of citation of that which arguably is not abuse
> will do as a demonstration that Longley is not abusive.
It isn't a matter of how you "feel" - it's a matter of whether
what I have said is true or not.
>
> The first posting I ever saw from Longley was a lengthy scurrilous attack on
> Aaron Sloman, filled with lies and hypocrisy. He has repeated similar abuse
> against others over and over again. Perhaps this would be more tolerable if
> he had interesting things to say, but he's got one note, which he repeats over
> and over again at great length and with repeated long quotations, which is a
> form of abuse in its own right. He isn't interested in any sort of exchange
> or interaction, yet continues to snipe.
That may well be how you construe what I have presented - it
certainly isn't everyone's. I'm not trying to be "interesting"
either. It is true however, that I am not interested in exchanges
with folk such as yourself. You just don't know enough of the
relevent literature yet.
> When Drew McDermott explicitly asked
> to avoid Longley's sort of sniping, how many expected Longley to do so anyway?
> I certainly did, and sure enough Drew ended up commenting on it. Basically
> Longley is a sociopath. What do you do with someone like that in a free forum
> like this? You can ignore him, or you can return fire. I've tried both, and
> neither is very satisfying.
I think the above just just illustrates the rather perverse way
you partcipate in this newsgroup. Do you have anything original
or useful to contribute? If so, perhaps it's about time we saw
some of it.
--
David Longley
Neil Rickert
In <Pine.SOL.3.91.961212072729.20314A-100000@tortoise> David Yeo <dyeo@tortoise> writes:
>On 11 Dec 1996, Neil Rickert wrote:
>> In <Pine.SOL.3.91.961211083356.13273A-100000@tortoise> David Yeo writes:
>> A heuristic that cannot be programmed is useless to AI. If there are
>> rules of thumb, they ought to be statable, and programmable. If you
>> cannot state what they are well enough to program them into a
>> repeatable procedure, then the claim that there is a heuristic is
>> empty talk.
>While I agree with your first sentence, I don't think that it follows that
>the rules of thumb must be statable. Indeed, that is one of the strengths
>(and weaknesses) of ANNs; the heuristics (rules of thumb) that they derive
>are often both unanticipated and difficult to summarise.
It is possible of course, to make an attribution which claims that
the ANN behaves as it does because it is following rules of thumb.
But it seems to me that if the presumed rules can not be given
explicitly, there are ample grounds for skepticism as to the validity
of that attribution.
>> Why would it seem that the ANNs are an embodiment of rules of thumb?
>> I don't get it. How can we be sure that there be rules of thumb, if
>> we don't know what those rules are?
>Because ANNs discover (learn) solutions to the problems with which they
>are presented as they explore the problem space.
Why does it follow from this that they are acting on rules of thumb,
or on rules of any kind? Presumably it can be said to be following
the rule of doing what it does. But if that is all you mean, I don't
see that the claim of following rules or heuristics has any useful
content.
Let me say it differently. One can think of an attribution of rule
following as prescriptive or as descriptive. If it is prescriptive,
then there ought to be explicit rules, at least in principle, which
constitute the prescription. If it is descriptive, then it seems to
me that it is no description at all unless the rules are part of the
description.
> Moreover, the feedback
>which underlies an ANNs weight adjustments can, I think, be validly seen
>as evaluation of progress made toward the final result. This accords well
>with the (cited) definition of a "heuristic" offered by Rosenberg.
I was not objecting to the idea of feedback. My concern is whether
you are using "rules" in an explanation which does not explain.
>> If we could really write down explicitly these rules of thumb that
>> people are supposedly following when they supposedly use a heuristic,
>> we should be able to program them. Then we should be able to carry
>> out controlled experiments, comparing the judgement of the computer
>> systems following these rules of thumb, and the judgements of the
>> people who are said to be following something like the availability
>> heuristic.
>Again, the inability to "write down explicitly these rules of thumb" does
>not preclude the ability to "carry out controlled experiments". Everyday
>many (in my view) interesting experiments are carried out using ANNs.
I agree that one can carry out many interesting experiments. But it
seems to me that there is still a question as to how best to explain
what the ANN is doing, and saying "it is following rules, but we
don't know what the rules are" just does not explain anything much at
all.
If talk about rule following is useful in discussions, I have no
problem with that. My real problem is when Longley uses the
attribution of heuristics as the final answer, and criticizes others
who think that it is no real answer at all.
>> It is my conjecture that, if we could ever do this, we
>> would find that the human judgement is better than that of the
>> presumed heuristics. But of course the controlled experiment cannot
>> be carried out, unless you can reduce the presumed heuristic to a set
>> of rules to follow, and then automate the following of those rules.
>Sorry, but I don't agree. In many ways your view would seem to be a
>variant of what Ryle (1949) called the "intellectualist legend":
Surely you are not suggesting that the carrying out of controlled
experiments is the intellectualist legend.
> According to the legend, whenever an agent does anything intelligently,
> his act is preceded and steered by another internal act of considering
> a regulative proposition appropriate to his practical problem.
I do not suggest that at all.
Neil Rickert
In <850396...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley) writes:
>In article <58n7up$7...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>> Longley wants to insist on extensional talk. I want to know what is
>> the extension of these heuristics. It is possible to make an
>> attribution which claims that the ANN is following rules of thumb.
>> But the claim is empty, in the sense that there is no means to
>> establish whether the ANN is actually following those rules. In a
>> GOFAI system, you can be sure whether or not the system is following
>> the rules.
>I do not insist on extensional talk.
No, you only insist on it for others, but never for yourself.
>> If we could really write down explicitly these rules of thumb that
>> people are supposedly following when they supposedly use a heuristic,
>> we should be able to program them. Then we should be able to carry
>> out controlled experiments, comparing the judgement of the computer
>> systems following these rules of thumb, and the judgements of the
>> people who are said to be following something like the availability
>> heuristic. It is my conjecture that, if we could ever do this, we
>> would find that the human judgement is better than that of the
>> presumed heuristics. But of course the controlled experiment cannot
>> be carried out, unless you can reduce the presumed heuristic to a set
>> of rules to follow, and then automate the following of those rules.
>The above is a muddle and just shows that you have NOT read the
>literature, nor paid attention to what I have said in the past.
I will thank you to not post your irrational beliefs about what I
have read. If I said something wrong, document that. Keep to the
evidence.
>There are programs which end up producing the same biases in
>classification that one finds in natural human decision making.
So what? Does it follow that humans are executing out these
particular programs? These programs are not doing what I called for
in my "muddle" above.
Neil Rickert
>Clearly Rickert does not agree. But then, he seems to make a
>point of not agreeing with much that I have to say. I mean no
>offence to him when I say that I do not believe he understands
>the relevant literature.
I would have thought it an ordinary courtesy to base communication on
what I have written, rather than on your own private intensional
beliefs about me.
Jim Balter
In article <850412...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <jqbE2A...@netcom.com> j...@netcom.com "Jim Balter" writes:
>>
>> The first posting I ever saw from Longley was a lengthy scurrilous attack on
>> Aaron Sloman, filled with lies and hypocrisy. He has repeated similar abuse
>> against others over and over again. Perhaps this would be more tolerable if
>> he had interesting things to say, but he's got one note, which he repeats over
>> and over again at great length and with repeated long quotations, which is a
>> form of abuse in its own right. He isn't interested in any sort of exchange
>> to avoid Longley's sort of sniping, how many expected Longley to do so anyway?
>> I certainly did, and sure enough Drew ended up commenting on it. Basically
>> Longley is a sociopath. What do you do with someone like that in a free forum
>> like this? You can ignore him, or you can return fire. I've tried both, and
>> neither is very satisfying.
>> <J Q B>
>>
>
>It's time you grew up and learned the difference between creative writing and
>science.
Which of the two is that sentence, you stupid fuck?
--
<J Q B>
Jim Balter
In article <850414...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <jqbE2A...@netcom.com> j...@netcom.com "Jim Balter" writes:
>>
>> No amount of citation of that which arguably is not abuse
>> will do as a demonstration that Longley is not abusive.
>
>It isn't a matter of how you "feel" - it's a matter of whether
>what I have said is true or not.
*What* is not a matter of how I feel? Strange how intensionality only works
in one direction for Longley. When I point out that he is a hypocrite, he
seems to be little concern as to whether what I have said is *true or not*.
But, it certainly isn't a matter of how I feel that citation of that which is
arguably not abuse will not do as a demonstration that Longley is not abusive,
it is a matter of elementary logic. And it certainly *not true* that it is a
matter of whether what he has said is true or not; that too is elementary
logic.
>> The first posting I ever saw from Longley was a lengthy scurrilous attack on
>> Aaron Sloman, filled with lies and hypocrisy. He has repeated similar abuse
>> against others over and over again. Perhaps this would be more tolerable if
>> he had interesting things to say, but he's got one note, which he repeats over
>> and over again at great length and with repeated long quotations, which is a
>> form of abuse in its own right. He isn't interested in any sort of exchange
>> or interaction, yet continues to snipe.
>
>That may well be how you construe what I have presented - it
>certainly isn't everyone's.
So it isn't a mater of how I feel, or how Aaron felt, but how "everyone"
feels.
>I'm not trying to be "interesting"
>either.
You are succeeding admirably.
>It is true however, that I am not interested in exchanges
>with folk such as yourself. You just don't know enough of the
>relevent literature yet.
I guess folk like myself means "non-sociopathic humans". Such folk would at
least be able to entertain the possibility thsat there can be a question as to
just what literature is relevant.
>> When Drew McDermott explicitly asked
>> to avoid Longley's sort of sniping, how many expected Longley to do so anyway?
>> I certainly did, and sure enough Drew ended up commenting on it. Basically
>> Longley is a sociopath. What do you do with someone like that in a free forum
>> like this? You can ignore him, or you can return fire. I've tried both, and
>> neither is very satisfying.
>
>I think the above just just illustrates the rather perverse way
>you partcipate in this newsgroup. Do you have anything original
>or useful to contribute? If so, perhaps it's about time we saw
>some of it.
Read what *I* wrote in response to Drew's question about Chalmers, you fucking
one-note hypocritical lying asshole.
--
<J Q B>
Jim Balter
In article <Pine.SOL.3.91.961212072729.20314A-100000@tortoise>,
David Yeo <dyeo@tortoise> wrote:
>On 11 Dec 1996, Neil Rickert wrote:
>
>> In <Pine.SOL.3.91.961211083356.13273A-100000@tortoise> David Yeo writes:
>
>> >it is true that heuristics are procedures (hence algorithms), algorithms
>> >usually also carry the implication that they will achieve the solution to
>> >a problem in a finite number of steps. To quote Rosenberg (1986):
>>
>> > Algorithm: a finite set of well-defined rules for the solution of a
>> > problem in a finite number of steps; for example, a full
>> > statement of an arithmetic procedure for evaluating the sin
>> > X to a stated precision.
>>
>> > (Dictionary of Artificial Intelligence and Robotics)
>>
>> >Heuristics (rules of thumb) do not offer this degree of certainty; in fact
>> >they don't guarantee they'll achieve the solution to a problem at all. I
>> >suggest that it's this latter (non-procedural) sense of algorithm that AI
>> >researchers imply when distinguishing "algorithms" from "heuristics".
>>
>> This seems a little confused. I'll agree that heuristics are not
>> guaranteed to produce a solution. But it does not follow that they
>> are non-procedural. It only follows that the procedure is not
>> guaranteed to succeed.
>
>I did not mean to imply that heuristics are non-procedural (see my second
>sentence - "While it is true that heuristics are procedures ... ." ), I
>only meant that the feature which distinguishes algorithms and heuristics
>is the extra burden algorithms carry re: achieving a solution in a finite
>number of steps. The confusion you note arises from the tendency of some
>to inappropriately confine/reduce algorithms to their procedural dimension.
This is even more confused. Algorithms are procedures, but procedures are not
necessarily algorithms, since algorithms terminate, whereas procedures need
not. At least, that's the technical usage as found in, say, Hopcroft and
Ullman. If you expect anyone to know what you mean by "procedural dimension",
you had best be explicit; if you mean "terminates", you are using the wrong
term. In any case, the issue of heuristics has nothing to do with
termination; it has to do with *intent*. A heuristic is a procedure that
doesn't necessarily achieve the *desired* result.
>Because ANNs discover (learn) solutions to the problems with which they
>are presented as they explore the problem space. Moreover, the feedback
>which underlies an ANNs weight adjustments can, I think, be validly seen
>as evaluation of progress made toward the final result. This accords well
>with the (cited) definition of a "heuristic" offered by Rosenberg.
Unless we know the desired result and are just testing the ANN, what is the
norm for the *right* final result?
Rosenberg's definition may be of the sort of ANN you describe, but it is quite
inaccurate as a definition of "heuristic", either in its traditional meaning
of "discovery technique" or its modern meaning of "approximate method".
--
<J Q B>
Jim Balter
In article <58pfgi$b...@usenet.srv.cis.pitt.edu>,
Anders N Weinstein <ande...@pitt.edu> wrote:
>Well I looked at it in the bookstore. But it's not the kind of book I pay
>money for.
Well bully for you. So beg, borrow, or steal a copy.
>The problem is, I understand the abstract conception of a computation or
>algorithm applied to formal or symbolic elements, such as strings on a
>TM tape. I have trouble when one tries to see how to apply it to
>arbitrary physical objects.
Like a physical computer made of silicon? This is just your problem, and I
suspect you just have it because it justifies your ideology.
>I would have thought that Niagara Falls is *not* employing an algorithm
>to get water to the bottom, the stomach is not employing an algorithm
>to digest food, the moon is not employing an algorithm to stay in its
>orbit etc.
I suggest that you reread what Dennett has written about the design stance,
and then perhaps you will get it. If you want to. And perhaps you will see
that digestion is a process that bears an "implemented by" relationship to the
stomach. No such relationship can be found between "orbit" and "moon" or
"fall" and "water".
>Of course the states of these processes can be mapped onto
>states of a formal computation trace, indeed they can be mapped onto an
>infinite number of algorithms if we want. But that seems to me a
>Pickwickian notion of instantiating an algorithm.
The state of my computer can be mapped into an infinite number of algorithms
in the same manner, but there is good reason to see my computer as
implementing a telnet session where we don't see the burned out motherboard in
the corner doing the same; but I'm sure you have seen these arguments before.
>So I am wondering if one should say that evolution instantiates an
>algorithm in the Pickwickian sense in which molecules in a cup of coffee
>instantiate the Wordstar program, or in some more substantial sense.
No Turing machine has ever instantiated a Wordstar program, so I can't imagine
where you are getting notions about such a "program". Or at least I couldn't
if I were playing the same sort of silly game as you are. But perhaps you
learned to play this silly "I can't make any sense of Dennett" game in one of
those books of the sort you are willing to pay for.
>I am also wondering about the glib treatment of randomness. I don't
>think the issue is mainly about the possibility of indeterminism at the
>micro-level. Processes at the biological level might be
>non-law-governed when characterized in those terms, even if processes
>described at the lower level are deterministic. Just as the disturbance
>of a computation in my machine by a power dip might be a result of
>deterministic laws of physics but not systematically explicable at the
>algorithmic level.
I find remarkable your ability to fail to absorb or even acknowledge anything
said to you. For instance, the bit about abstract vs. physical "algorithms"
above; you may *disagree* with what has been said, but remarkable is your
apparent (or feigned) lack of *familiarity* with the arguments. And here,
despite my repeatedly pointing out that laws are *descriptive*, so that the
idea of "non-law-governed" processes is incoherent, you don't even seem to
consider this. If the power were to dip in my machine, then if we *were* to
characterize the behavior of the system at the algorithmic level, the
description would certainly be of *some* algorithm, but it would not be the
one we had intended. There is *no* description at that level that isn't the
description of *some* algorithm (procedure, actually, since it need not
terminate). If the machine behaves in some way that, say, isn't describable
in the machine language of that machine, then there *is no* characterization
at that level, but we usually can find some other level, say the bit level, at
which we can find such a description. But the idea of an algorithmic
description that isn't algorithmic is incoherent. If what you mean is that no
description at the biological level is possible, that makes some sense, but we
can make it describable by adding randomness as a primitive, just as we can
describe a program that employs a quantum-mechanical timer.
>Again Dennett's claim seems to blur the distinction I would want
>between a non-systematic effect on a process, and a systematic
>procedure that happens to employ a randomizing element.
Perhaps he blurs it because such a distinction is incoherent. A
non-systematic effect on a process would seem to be a description at a certain
level that isn't a description at that level. Dennett has thoroughly covered
the notions of stances and attribution of descriptions, perhaps in books that
you wouldn't pay for. Perhaps you should to get a real job on the side.
--
<J Q B>
Neil Rickert
In <850439...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley) writes:
>> In other words, your meaning for the word "heuristic" is very
>> different from that used in AI and computer science. Since this is
>> an AI newsgroup, I think you should stick to the AI usage of the
>> term. Perhaps that would reduce the confusion.
>I'm using heuristic in the same way that Tversky and Kahneman use
>it.
Then that is not appropriate usage for an AI discussion.
> In that way it belongs in the repertoire of rules of thumb
>studdied within Attribution Theory. Have a look at the Journal of
>Personality and Social Psychology from abouut 1996 onwards just
>to see how much has been done on that.
As I have said in the past, I am not criticizing the creditable work
that Tversky and Kahneman have done. I am criticizing the overly
broad conclusions that you draw from their work.
>Next have a look at the dozen or so journals on Learning Theory,
>Memory and Cognition.....only THEN will you perhaps start to
>realise how VAST a literature I am alluding to.
I know there is a vast literature. It is a literature which
convincingly shows that the problem of learning has not been solved,
although small inroads have been made. Again, I am not criticizing
what has been done on learning. I am critical of the sweeping
conclusions that you draw -- most particularly your apparent
conclusion that it is a solved problem.
>What you have to remember is that up until the mid 70s "Computer
>Science" was a rather second rate academic discipline in
>universities.
No, that is false. It was a small discipline, present as a
department in only a few universities, but that did not make it
second rate.
> Most of the real work of interest in AI has ALWAYS
>been done by psychologists.
Balderdash.
> In fact the first computers were
>designed explicitly to be emulations of human thought.
Utter nonsense. The first computers were designed by engineers and
mathematicians, of whom von Neumann is one of the best known. They
were concerned with numerical calculations, rather than with human
thought.
>If you do NOT appreciate that, just just don't know your computer
>science history any better than you know your history of
>psychology.
In your near complete ignorance of the history of computer science,
you have the gall to proclaim yourself an expert.
>I think it is good that today, so much of the research in AI is
>being undertaken in Electrical Engineering Departments - I just
>wish they'd lose their reverence for the human brain and dreams
>about consciousness etc.
Perhaps it was once different on Mars, or whatever alien place it is
that you come from.
David Longley
> In <850398...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley)
> writes:
> >In article <58na6q$7...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:>
> >> Unless I missed it, Dennett never explicitly states what the
> >> algorithm is. In that sense, I think you are right that it lacks
> >> meaning. It suffers from the same inexplicitness problem as
> >> Longley's use of "heuristic".
>
> >It's inexplicit basically *because* it is intensional and content
> >bound. In the end, I think it may be helpful for folk to think of
> >"heuristic" in the same way as adventious reinforcement or just
> >conditioning.
>
> In other words, your meaning for the word "heuristic" is very
> different from that used in AI and computer science. Since this is
> an AI newsgroup, I think you should stick to the AI usage of the
> term. Perhaps that would reduce the confusion.
>
>
I'm using heuristic in the same way that Tversky and Kahneman use
it. In that way it belongs in the repertoire of rules of thumb
studdied within Attribution Theory. Have a look at the Journal of
Personality and Social Psychology from abouut 1996 onwards just
to see how much has been done on that.
Next have a look at the dozen or so journals on Learning Theory,
Memory and Cognition.....only THEN will you perhaps start to
realise how VAST a literature I am alluding to.
What you have to remember is that up until the mid 70s "Computer
Science" was a rather second rate academic discipline in
universities. Most of the real work of interest in AI has ALWAYS
been done by psychologists. In fact the first computers were
designed explicitly to be emulations of human thought.
If you do NOT appreciate that, just just don't know your computer
science history any better than you know your history of
psychology.
I think it is good that today, so much of the research in AI is
being undertaken in Electrical Engineering Departments - I just
wish they'd lose their reverence for the human brain and dreams
about consciousness etc.
--
David Longley
David Longley
This may be helpful. See http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
for the context.
'In contrast to formal theories of belief, intuitive
judgments of probability are generally not extensional.
People do not normally analyse daily events into
exhaustive lists of possibilities or evaluate compound
probabilities by aggregating elementary ones. Instead,
they use a limited number of heuristics, such as
representativeness and availability (Kahneman et al.
1982). Our conception of judgmental heuristics is based
on NATURAL ASSESSMENTS that are routinely carried out as
part of the perception of events and the comprehension
of messages. Such natural assessments include
computations of similarity and representativeness,
attributions of causality, and evaluations of the
availability of associations and exemplars. These
assessments, we propose, are performed even in the
absence of a specific task set, although their results
are used to meet task demands as they arise. For
example, the mere mention of "horror movies" activates
instances of horror movies and evokes an assessment of
their availability. Similarly, the statement that Woody
Allen's aunt had hoped that he would be a dentist
elicits a comparison of the character to the stereotype
and an assessment of representativeness. It is
presumably the mismatch between Woody Allen's
personality and our stereotype of a dentist that makes
the thought mildly amusing.. Although these assessments
are not tied to the estimation of frequency or
probability, they are likely to play a dominant role
when such judgments are required. The availability of
horror movies may be used to answer the question "What
proportion of the movies produced last year were horror
movies?", and representativeness may control the
judgement that a particular boy is more likely to be an
actor than a dentist.
The term JUDGMENTAL HEURISTIC refers to a strategy -
whether deliberate or not - that relies a natural
assessment to produce an estimation or a prediction.
.Previous discussions or errors of judgement have
focused on deliberate strategies and on
misinterpretations of tasks. The present treatment calls
special attention to the processes of anchoring and
assimilation, which are often neither deliberate nor
conscious. An example from perception may be
instructive: If two objects in a picture of a three-
dimensional scene have the same picture size, the one
that appears more distant is not only seen as "really"
larger but also larger in the picture. The natural
computation of real size evidently influences the (less
natural) judgement of picture size, although observers
are unlikely to confuse the two values or to use the
former to estimate the latter.
The natural assessments of representativeness and
availability do not conform to the extensional logic of
probability theory.'
A. Tversky and D. Kahneman
Extensional Versus Intuitive Reasoning:
The Conjunction Fallacy in Probability Judgment.
Psychological Review Vol 90(4) 1983 p.294
'Meehl's classic book, published in 1954, summarised
evidence for the conclusion that simple linear
combinations of cues outdo the intuitive judgements of
experts in predicting significant behavioural criteria.
The lasting intellectual legacy of this work, and of the
furious controversy that followed it, was probably not
the demonstration that clinicians performed poorly in
tasks that, as Meehl noted, they should not have
undertaken. Rather, it was the demonstration of a
substantial discrepancy between the objective record of
people's success in prediction tasks and the sincere
beliefs of these people about the quality of their
performance. This conclusion was not restricted to
clinicians or to clinical prediction:
People's impressions of how they reason, and how well
they reason, could not be taken at face value.'
D. Kahneman, P. Slovic & A. Tversky (1982)
Judgment Under Conditions of Uncertainty: Heuristics and
Biases
'A basic principle of probability is the conjunction rule,
p(B) >= p(A&B). People violate this rule often, particularly
when judgements of probability are based on intensional
heuristics such as representativeness and availability.
Through other probabilistic rules are obeyed with increasing
frequency as people's levels of mathematical talent and
training increase, the conjunction rule generally does not
show such a correlation. We argue that this recalcitrance is
not due to inescapable "natural assessments"; rather, it
stems from the absence of generally useful problem-solving
designs that bring extensional principles to bear on this
class of problem. We predict that when helpful extensional
strategies are made available, they should compete well with
intensional heuristics. Two experiments were conducted, using
as subjects adult women with little mathematical background.
In Experiment I, brief training on concepts of algebra of
sets, with examples of their use in solving problems, reduced
conjunction-rule violations substantially, compared to a
control group. Evidence from similarity judgements suggested
that use of the representativeness heuristic was reduced by
the training....
...We conclude that such intensional heuristics can be
suppressed when alternative strategies are taught.
The development of formal thought does not culminate in
adolescence as Piaget (1928) held; rather, it depends on
education (Fong, Krantz, & Nisbett, 1986, Nisbett, Fong,
Lehmann & Cheng 1987) and may continue throughout adulthood.
Probabilistic reasoning has been an especially useful domain
in which to study the impact of training in adulthood on
formal thought. Probabilistic principles are cultural
inventions at most a few centuries old (Hacking 1975).....
Tversky and Kahneman (1983) focused on processes in which
people substitute intensional for extensional thinking. In
the latter mode, concepts are represented mentally in the
same way as sets, hence, rules of logic and probability are
followed in the main. By contrast, intensional thinking
represents concepts by prototypes, exemplars, or relations to
other concepts (Rosch, 1978, Smith & Medlin 1981). Processing
is affected strongly by imaginability of prototypes,
availability of exemplars, etc., and its results are not
constrained as strongly by logical relations. A prime example
is the representativeness heuristic (Kahneman & Tversky
1972), in which probability of a outcome is judged in terms
of the similarity of that outcome to a prototype.
Tversky and Kahneman (1983) drew far reaching conclusions
from the fact that, in most of their tests, the prevalence of
conjunction errors was not affected by statistical education.
They developed the concept of "natural assessment", a
computation that is 'routinely carried out as part of the
perception of events and the comprehension of messages......
even in the absence of a specific task set.' They defined a
"judgmental heuristic" as a 'strategy that relies on a
natural assessment to produce an estimation or a prediction.'
They compared such mechanisms to perceptual computations, and
cognitive errors to perceptual illusions. In their view,
people well trained in mathematics nonetheless perform
natural assessments automatically. The results of these
mental computations strongly influence probability judgement.
Therefore, statistics courses presumably affect probability
judgements, in problems such as "Linda," no more than
geometry courses affect geometric visual illusions, i.e.,
scarcely at all.
Agnoli & Krantz (1989)
Suppressing Natural Heuristics by Formal Instruction:
The Case of the Conjunction Fallacy [my emphasis]
Cognitive Psychology 21, 515-550 (1989)
--
David Longley
David Longley
This really is a waste of time - but of course it's based on what
you have written. You are not going to make *any* headway on
*ANY* of these issues by just talking about what you think might
be possible. You need to look at the empirical research base and
see what is being done and how it all fits together.
Why don't you just try what I suggest? it's professional advice.
--
David Longley
Peter_...@uk.ibm.com
>The problem is, I understand the abstract conception of a computation or
>algorithm applied to formal or symbolic elements, such as strings on a
>TM tape. I have trouble when one tries to see how to apply it to
>arbitrary physical objects.
We're not talking about arbitrary physical objects. We're talking about
DNA as the genome giving rise to phenotypes which are fit or not. Although
the analogy is not exact, seeing sexual reproduction as a means of
searching fitness space doesn't seem a bad way of looking at evolution.
>I would have thought that Niagara Falls is *not* employing an algorithm
>to get water to the bottom, the stomach is not employing an algorithm
>to digest food, the moon is not employing an algorithm to stay in its
>orbit etc.
I agree with the above, but I'd worry about the 'etc.'.
>Of course the states of these processes can be mapped onto
>states of a formal computation trace, indeed they can be mapped onto an
>infinite number of algorithms if we want. But that seems to me a
>Pickwickian notion of instantiating an algorithm.
I certainly won't object to the observation that arbitrary 1-to-1
functions aren't what anyone means by 'instantiating an algorithm'.
>So I am wondering if one should say that evolution instantiates an
>algorithm in the Pickwickian sense in which molecules in a cup of coffee
>instantiate the Wordstar program, or in some more substantial sense.
Well, you can wonder away if you wish. But I think that once you stop
wondering, you'll see that seeing the genome as a means of exploring
fitness-space and seeing that this can, to some extent, be understood
using many of the same concepts we apply to heuristic programs,
makes the case rather different from the Pickwickian one.
For example, we might ask why sexual reproduction has evolved.
And the answer seems to be, because sexual reproduction is
good at exploring fitness space. Which, I think, brings in the notions
we often do associate with heuristic AI-type programs.
What I see here is a couple of red herrings. First the 'arbitrary
physical object' red herring. Second the 'arbitrary 1-to-1 mapping'
red herring. Let us set these to one side and discuss the question
at hand.
Cheers,
Pete Lupton
David Longley
In article <58qfeg$e...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>
> >What you have to remember is that up until the mid 70s "Computer
> >Science" was a rather second rate academic discipline in
> >universities.
>
> department in only a few universities, but that did not make it
> second rate.
I stand by what I said.
>
> > Most of the real work of interest in AI has ALWAYS
> >been done by psychologists.
>
> Balderdash.
???
>
> > In fact the first computers were
> >designed explicitly to be emulations of human thought.
>
> Utter nonsense. The first computers were designed by engineers and
> mathematicians, of whom von Neumann is one of the best known. They
> were concerned with numerical calculations, rather than with human
> thought.
Go and re-read the history.
>
> >If you do NOT appreciate that, just just don't know your computer
> >science history any better than you know your history of
> >psychology.
>
> In your near complete ignorance of the history of computer science,
> you have the gall to proclaim yourself an expert.
>
It's not that long a history, nor is it a very complex one.
--
David Longley
Neil Rickert
In <850469...@longley.demon.co.uk> Da...@longley.demon.co.uk (David Longley) writes:
>In article <58pk6q$c...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>> I would have thought it an ordinary courtesy to base communication on
>> what I have written, rather than on your own private intensional
>> beliefs about me.
>This really is a waste of time - but of course it's based on what
>you have written.
You quoted one paragraph above. That, of course, is based on what I
have written. In fact, it is what I have written.
Given your frequently repeated objections to indirect quotation, I am
shocked that you would claim that your own subjective memory, not
even supported by indirect quotation, has anything at all to do with
what I have written.
> You are not going to make *any* headway on
>*ANY* of these issues by just talking about what you think might
>be possible.
That's good, because I have not spent much time talking about that.
It is a measure of your confusion that you think otherwise.
> You need to look at the empirical research base and
>see what is being done and how it all fits together.
I have been looking at the empirical research base. That is why
I reject so much of Longley's claims.
>Why don't you just try what I suggest? it's professional advice.
What profession?
Hans Hartmann
Neil Rickert wrote:
>
> In <850323...@longley.demon.co.uk> Da...@longley.demon.co.uk
(David Longley) writes:
> >In article <58lld4$5...@ux.cs.niu.edu> ric...@cs.niu.edu "Neil
Rickert" writes:
>
> >> When I am disagreeing
with
> >> the usual interpretation of well known evidence, I am do not
need to
> >> cite results which counter that evidence. I need only show that
the
> >> interpretation I am criticizing is riddled with holes, or lacks
the
> >> evidence to make it credible.
>
> >Do you not think I (and others) think it somewhat incredible that
> >the ONLY references you cite (by name only) are the ones I have
> >documented specifically? You make it quite clear by what you
> >write that you have not read the material before, nor have you a
> >grasp of the overall literature which it is part of. Now that's
> >not abuse, just fact.
>
> You make it clear that you are a rigid dogmatist. You cannot
> distinguish between disagreement and failure to read. I have not
> gone around with large lists of citations, because the articles
are
> not particularly relevant. Many of them document that human
behavior
> cannot be explained as the execution of a formal logic based on
> beliefs. It is certainly valuable that psychologists have
documented
> this. But it says no more than should have been the common
> observation of anybody with his eyes open. That is, it should
have
> been common knowlege all along, although perhaps never formally
> documented. One should not need to provide large citation lists
when
> dealing with what should be common knowledge.
>
> >> If you don't care for what I write, you can always go back to
> >> psychology newsgroups. ...
>
> >More nonsense and abuse - and anyone who has read "Fragments"
> >will see it as exactly that.
>
> The abuse is in response to your abuse.
>
> >The above is just gibberish. I have made very explicit references
> >to the work in decision theory by Gluck and Bower on this matter,
> >and I have repeated the oft stated remark that the R-W model is
> >the most influential model over the past 25 years or so.
>
> I really don't care how influential it is. It is still not
> extensional by my standards. You cannot tell what a stimulus is,
or
> what a response is, without having one of those defective (i.e.
> intensional or intentional) humans make the judgement. The theory
> cannot account for a Copernicus, or a Newton, a Darwin or an
> Einstein.
>
> > Now IF
> >you understood the literature, and if you understood what I was
> >saying, you would see the relationship between what I have said
> >about the Extensional Stance (as opposed to the intentional
> >stance), the use of normative descriptive statistics (including
> >regression and its variants in pattern recognition), the problems
> >of biases of natural assessments (which can be modelled by ANNs)
> >and a host of other key areas of research which are covered in
> >"Fragments".
>
> That was a long sentence. What it boils down to is: If I
understood
> the literature, I would agree with David Longley.
>
> There is a natural corollary. If Copernicus had understood the
> literature, he would have been a Ptolemaic. If Darwin had
understood
> the literature he would have been a creationist. If Wegener had
> understood the literature, he would have known that continents do
not
> drift.
>
> So why don't we just close all of the science labs. If we all
know
> the literature, we will realize that the conventional wisdom could
> never be wrong, and that there is no point in doing any more
> research.
>
> >> >> >Truth is not determined by convention.
>
> Except when Longley decides which are the conventions.
From all the messages I read today, I liked that one the most. Sort
of AI-translation, reducing the word count.
(except for the last reply)
Interesting to observe how emotional logic can become. The latter is
a reply to most of the messages read today (g).
Regards Hans
Anders N Weinstein
In article <jqbE2B...@netcom.com>, Jim Balter <j...@netcom.com> wrote:
>In article <58pfgi$b...@usenet.srv.cis.pitt.edu>,
>Anders N Weinstein <ande...@pitt.edu> wrote:
>
>>algorithm applied to formal or symbolic elements, such as strings on a
>>TM tape. I have trouble when one tries to see how to apply it to
>>arbitrary physical objects.
>
No I do not have trouble *there*. In that case it is the social
context that determines what the devices correct description is, not
its internal physical structure alone. In another context a similarly
structured physical object might constitute a room heating device.
The trouble is to say somewhat explicitly why not everything is
a computer or an algorithm. You are right that this issue has been raised
in print by others, namely Searle and, I believe, Putnam (in
_Representation and Reality_).
>>I would have thought that Niagara Falls is *not* employing an algorithm
>>to get water to the bottom, the stomach is not employing an algorithm
>>to digest food, the moon is not employing an algorithm to stay in its
>>orbit etc.
>
>I suggest that you reread what Dennett has written about the design stance,
>and then perhaps you will get it. If you want to. And perhaps you will see
>that digestion is a process that bears an "implemented by" relationship to the
>stomach. No such relationship can be found between "orbit" and "moon" or
>"fall" and "water".
So you are saying that the stomach *is* implementing an algorithm?
>>Of course the states of these processes can be mapped onto
>>states of a formal computation trace, indeed they can be mapped onto an
>>infinite number of algorithms if we want. But that seems to me a
>>Pickwickian notion of instantiating an algorithm.
>
>The state of my computer can be mapped into an infinite number of algorithms
>in the same manner, but there is good reason to see my computer as
>implementing a telnet session where we don't see the burned out motherboard in
>the corner doing the same; but I'm sure you have seen these arguments before.
I agree, although I am not aware of a thoroughly convincing answer to
the challenge on the part of advocates of computationalism or naturalism.
I take it to be certain social facts that make for the "good reason", and so
not something intrinsic to the device as a natural scientific object.
>I find remarkable your ability to fail to absorb or even acknowledge anything
>said to you. For instance, the bit about abstract vs. physical "algorithms"
>above; you may *disagree* with what has been said, but remarkable is your
>apparent (or feigned) lack of *familiarity* with the arguments. And here,
Maybe some citations would help -- I guess have some familiarity with "the
arguments" but can not say that the issue has been resolved by any
conclusive response.
>despite my repeatedly pointing out that laws are *descriptive*, so that the
>idea of "non-law-governed" processes is incoherent, you don't even seem to
>consider this. If the power were to dip in my machine, then if we *were* to
You were right above: I do not consider it mainly because I disagree
with the claim.
>characterize the behavior of the system at the algorithmic level, the
>description would certainly be of *some* algorithm, but it would not be the
>one we had intended. There is *no* description at that level that isn't the
>description of *some* algorithm (procedure, actually, since it need not
>terminate). If the machine behaves in some way that, say, isn't describable
Isn't this a vacuous notion of algorithm (or "law-governed")? All you
are saying is that whatever in fact happens in the actual world, we
can, ex post facto, fit some description to it. That has no
counterfactual force at all, it's not even a *generalization*.
There must be more to the claim that something instantiates an algorithm
than that some description can be fitted to it, on pain of trivialization.
>terminate). If the machine behaves in some way that, say, isn't describable
>in the machine language of that machine, then there *is no* characterization
>at that level, but we usually can find some other level, say the bit level, at
>which we can find such a description. But the idea of an algorithmic
>description that isn't algorithmic is incoherent. If what you mean is that no
I agree that *that* is incoherent. All I am saying is that description as
algorithmic is normalizing, it explains by reference to an ideal, it
operates at its own level from which interferences from the lower level
are not systematic.
(I would say the same about the relation between intentional stance and
computational psychology, btw).
>description at the biological level is possible, that makes some sense, but we
>can make it describable by adding randomness as a primitive, just as we can
>describe a program that employs a quantum-mechanical timer.
But a quicksort program that is disturbed by quantum rays is *not* a
program that employs a quantum-mechanical timer.
>>Again Dennett's claim seems to blur the distinction I would want
>>between a non-systematic effect on a process, and a systematic
>>procedure that happens to employ a randomizing element.
>
>Perhaps he blurs it because such a distinction is incoherent. A
>non-systematic effect on a process would seem to be a description at a certain
>level that isn't a description at that level. Dennett has thoroughly covered
Actually it is one of those inter-level descriptions that I usually
ignore.
Anders N Weinstein
In article <58sa33$n...@usenet.srv.cis.pitt.edu>, I <ande...@pitt.edu> wrote:
>But a quicksort program that is disturbed by quantum rays is *not* a
>program that employs a quantum-mechanical timer.
I meant "cosmic" rays. I guess I like the suggestion that the source or
cause of the interference stems is quite remote in the universe from the
computer, although this is not necessary for the claim.
Jim Balter
In article <58squ5$q...@usenet.srv.cis.pitt.edu>,
>>program that employs a quantum-mechanical timer.
I never said it was; you 've mixed things up.
>I meant "cosmic" rays. I guess I like the suggestion that the source or
>cause of the interference stems is quite remote in the universe from the
>computer, although this is not necessary for the claim.
The operations carried out are still describable by some algorithm, even if
not an algorithm for quicksort. But then, I already said that. OTOH, a
quicksort program is a quicksort program regardless of disruptions of some
instantiation of it.
--
<J Q B>
David Longley
'Research on clinical versus statistical judgement has
had little impact on everyday decision making,
particularly within its field of origin, clinical
psychology......The interview remains the sine qua non
of entrance into mental health training programs and is
required in most states to obtain a license to
practice. Despite the studies that show that clinical
interpretation of interviews may have little or no
predictive utility, actuarial interpretation of
interviews is rarely if ever used, although it is of
demonstrated value.
Ultimately, then, clinicians must choose between
their own observations or impressions and the
scientific evidence on the relative efficacy of the
clinical and actuarial methods. The factors that
create difficulty in self-appraisal of judgmental
accuracy are exactly those that scientific procedures,
such as unbiased sampling, experimental manipulation
of variables, and blind assessment of outcome, are
designed to counter. Failure to accept a large and
consistent body of scientific evidence over
unvalidated personal observation may be described as
a normal human failing or, in the case of
professionals who identify themselves as scientific,
plainly irrational.
Finally, actuarial methods - at least within the
domains discussed in this article - reveal the upper
bounds in our current capacities to predict human
behavior. An awareness of the modest results that are
often achieved by even the best available methods can
help to counter unrealistic faith in our predictive
powers and our understanding of human behavior. It may
well be worth exchanging inflated beliefs for an
unsettling sobriety, if the result is an openness to
new approaches and variables that ultimately increase
our explanatory and predictive powers.
What is needed is the is the development of actuarial
methods and a measurement assurance program that
maintains control over both judgement strategies so
that their operating characteristics in the field are
known and an informed choice of procedure is
possible. Dismissing the scientific evidence or
lamenting the lack of available methods will prove
much less productive than taking on the needed work.'
Dawes, Faust & Meehl (1989)
Science v243, p1668-1674
Clinical Versus Actuarial Judgement
--
David Longley
David Yeo
Let me try to dispel your confusion. Since I argue that both algorithms
and heuristics are procedures then does it not follow that procedures are
not necessarily algorithms (i.e. they may be heuristics).
While the above is a nice turn of phrase, it rather misses the point. The
issue that Neil and I were discussing is whether heuristic = algorithm. I
note that, since both algorithms and heuristics are procedures, in a sense
he is correct. And I do agree AI requires the formalisation of heuristics.
But I then went on to argue that algorithms shoulder the further burden of
termination ... which introduces the second point you raise (see below).
> since algorithms terminate, whereas procedures need
> not. At least, that's the technical usage as found in, say, Hopcroft and
> Ullman. If you expect anyone to know what you mean by "procedural dimension",
> you had best be explicit; if you mean "terminates", you are using the wrong
> term. In any case, the issue of heuristics has nothing to do with
> termination; it has to do with *intent*. A heuristic is a procedure that
> doesn't necessarily achieve the *desired* result.
I thought that it was obvious to everyone that terminate, in the current
context, meant terminate with the desired (intended) result. Apparently
I was wrong. Thanks for the clarification.
> >Because ANNs discover (learn) solutions to the problems with which they
> >are presented as they explore the problem space. Moreover, the feedback
> >which underlies an ANNs weight adjustments can, I think, be validly seen
> >as evaluation of progress made toward the final result. This accords well
> >with the (cited) definition of a "heuristic" offered by Rosenberg.
>
> Unless we know the desired result and are just testing the ANN, what is the
> norm for the *right* final result?
You seem to be trying to work both sides of the fence on this one Jim.
The *right* final result is the desired (intended) result.
> Rosenberg's definition may be of the sort of ANN you describe, but it is quite
> inaccurate as a definition of "heuristic", either in its traditional meaning
> of "discovery technique" or its modern meaning of "approximate method".
> --
> <J Q B>
You are mistaken. Rosenbergs definition is in line many (most). Precisely
in what way do you see it as inaccurate? What is your definition?
Neil Rickert
>I remind you that I've taught university level courses in this
>area and when it comes to the history of computing, we are
>talking of a very short history.
Having taught courses is not a proof of competence.
> On the other hand, you do
>continue to show a very poor grasp of what *I* have been trying
>to make *you* understand.
If you were much of a psychologist, you should have realized that you
cannot make me understand anything. Understanding is something I
have to do, not something you can shove down my throat. As it
happens, I understand well enough how thoroughly confused are the
ideas you are pushing.
There ought to be opposition
> Finally, actuarial methods - at least within the
> domains discussed in this article - reveal the upper
> bounds in our current capacities to predict human
> behavior.
Unlike you, the authors recognize that they have not demonstrated
their point beyond certain limited domains.
Jim Balter
In article <Pine.SOL.3.91.961213154454.17272B-100000@tortoise>,
David Yeo <dyeo@tortoise> wrote:
>On Thu, 12 Dec 1996, Jim Balter wrote:
>
>> In article <Pine.SOL.3.91.961212072729.20314A-100000@tortoise>,
>> David Yeo <dyeo@tortoise> wrote:
>> >On 11 Dec 1996, Neil Rickert wrote:
>> >
>> >> In <Pine.SOL.3.91.961211083356.13273A-100000@tortoise> David Yeo writes:
>> >> >it is true that heuristics are procedures (hence algorithms), algorithms
>> This is even more confused. Algorithms are procedures, but procedures are not
>> necessarily algorithms,
>
>Let me try to dispel your confusion. Since I argue that both algorithms
>and heuristics are procedures then does it not follow that procedures are
>not necessarily algorithms (i.e. they may be heuristics).
I'm not confused, but you are making a silly logical error, or being a
politician. It doesn't follow from that fact that you claim that 1+1=2 or
that apple pie is good, but this just shows that you aren't consistent enough
to remember what it is you *have* claimed. The word "hence" in your "it is
true that heuristics are procedures (hence algorithms)" clearly implies that
procedures are necessarily algorithms. Make up your mind.
>While the above is a nice turn of phrase, it rather misses the point. The
"it" cannot miss a point. *I* have not missed the point, if you read all of
what I said.
>issue that Neil and I were discussing is whether heuristic = algorithm. I
>note that, since both algorithms and heuristics are procedures, in a sense
>he is correct. And I do agree AI requires the formalisation of heuristics.
>But I then went on to argue that algorithms shoulder the further burden of
>termination ... which introduces the second point you raise (see below).
>
>> since algorithms terminate, whereas procedures need
>> not. At least, that's the technical usage as found in, say, Hopcroft and
>> Ullman. If you expect anyone to know what you mean by "procedural dimension",
>> you had best be explicit; if you mean "terminates", you are using the wrong
>> term. In any case, the issue of heuristics has nothing to do with
>> termination; it has to do with *intent*. A heuristic is a procedure that
>> doesn't necessarily achieve the *desired* result.
>
>I thought that it was obvious to everyone that terminate, in the current
>context, meant terminate with the desired (intended) result. Apparently
>I was wrong. Thanks for the clarification.
You had to be wrong, because "the further burden of termination" that
algorithms shoulder is *not* "with the desired result", but merely termination
at all; the halting problem is not the problem of determining whether TMs
halt "with the desired result". My point (missed, it seems) is that it isn't
termination that is relevant, but rather "with the desired intent".
Procedures guaranteed to halt (algorithms) and procedures not guaranteed to
halt can both be heuristics. "termination" is a red herring, pure and simple.
>> >Because ANNs discover (learn) solutions to the problems with which they
>> >are presented as they explore the problem space. Moreover, the feedback
>> >which underlies an ANNs weight adjustments can, I think, be validly seen
>> >as evaluation of progress made toward the final result. This accords well
>> >with the (cited) definition of a "heuristic" offered by Rosenberg.
>>
>> Unless we know the desired result and are just testing the ANN, what is the
>> norm for the *right* final result?
>
>You seem to be trying to work both sides of the fence on this one Jim.
Oh, you mean that one is required to come in with an ideology, and fight for
it? "Working both sides of the fence" is what rationality is all about.
I'm simply asking a question.
>The *right* final result is the desired (intended) result.
Look, do me the favor of opening your mind and your eyes. I just said "unless
we know the desired result" .... That translates into "If we don't know the
desired result, what is the norm for the *right* desired result?" You
answering "the desired result" is liable to lead me to be frustrated with you
and start justifiably questioning your intelligence and skills. If you don't
want that, then pay more attention.
The point, which you have missed by a mile, is that if you *know* the desired
result, there is no need to build an ANN. Again I ask you, if you aren't just
testing to see whether an ANN can reach a result that you already know, what's
the norm? If you have a problem the solution of which you do not know, so
that the *correct* final result is unknown, what does it mean to refer to
"feedback ... validly seen as evaluation of progress made toward the final
result"? If we go hill climbing, how do we know we are on the right hill?
>> Rosenberg's definition may be of the sort of ANN you describe, but it is quite
>> inaccurate as a definition of "heuristic", either in its traditional meaning
>> of "discovery technique" or its modern meaning of "approximate method".
>> --
>> <J Q B>
>
>You are mistaken. Rosenbergs definition is in line many (most). Precisely
>in what way do you see it as inaccurate? What is your definition?
"Naysaying is not an argument". It really won't do to just claim without
support that I am mistaken, that Rosenberg's definition is "in line", and to
ignore what I wrote. I just gave you two definitions; don't you know one when
you see one? "approximate method" is equivalent to the standard CS
definition, often expressed as "rule of thumb" or "educated guess".
Rosenberg's definition is inaccurate because it isn't a definition of
"heuristic" at all, but rather a description of *one* sort of heuristic, which
is hill climbing. As I have pointed out, it is not robust, since it doesn't
deal with dead ends. For that you must add backtracking, but full
backtracking is computationally expensive, so one generally adds other types
of heuristics, such as branch sorting heuristics. These speed the search by
examining the "most likely" alternatives first. "Most likely" is approximate,
an educated guess, i.e., heuristic. Algorithms need not depend upon examining
likely possibilities first; they will terminate in any case. And procedures
not guaranteed to terminate are still not guaranteed to terminate if the "most
likely" branches are examined first. This is basic CS.
"This liberty [that can be taken when designing algorithms] involves the
introduction of rules of thumb, educated guesses of sorts, or, to use the
accepted term, heuristics". -- David Harel, _Algorithmics_
"The use of common sense in problem solving, as contrasted to formal logic, is
a form of 'impure' mathematics known as heuristics". -- Donald Fink,
_Computers and the Human Mind_.
"There are many instances of mathematical results which can be obtained by
means of what might be called 'heuristic principles', where such a principle
would not supply a /proof/ of the required result, but it would lead one to
anticipate that the result indeed ought to be true." -- Roger Penrose,
_Shadows of the Mind_
"... heuristics, or rules of thumb ..." -- Douglas Hofstadter, GEB
"heuristic: Providing assistance in discovering (or in presenting) a truth or
solving a problem, for example a model or a useful hypothesis. Contrasted
with proof.
heuristic principle: A principle that is neither asserted nor evaluated as
true but that is assumed for specific purposes at hand (such as to inquire
into, explain something) because of its previous success or usefulness as an
investigative tool.
heuristics: The name for the discipline that studies the methods by which
truth (fact, ideas, etc.) are discovered and (sometimes) communicated."
-- Peter Angeles, _Dictionary of Philosophy_
"... a simple heuristic method ... the record is moved to the front of the array ..." --
Gonnet and Baeza-Yates, _Handbook of Algorithms and Data Structures_.
"Heuristic search, or the discovery of a solution to a problem ... through the
disciplined generation and evaluation of a number of promising possibilities
..." -- Encyclopedia Brittanica.
The latter is the closest to Rosenberg, but that is *heuristic search*, not
*heuristic*, and is no more a definition of "heuristic" than is "move to
front". Is a definition of "binary search" a definition of "binary"? I don't
think so.
Before you start spouting off that I am mistaken, and making claims of what is
in line with what, do me, yourself, and the rest of the world a favor: round
up some supporting evidence, and leave your ideology at home.
--
<J Q B>
Jim Balter
In article <Pine.SOL.3.91.961213154454.17272B-100000@tortoise>,
David Yeo <dyeo@tortoise> wrote:
>> Ullman. If you expect anyone to know what you mean by "procedural dimension",
>> you had best be explicit; if you mean "terminates", you are using the wrong
>> term. In any case, the issue of heuristics has nothing to do with
>> termination; it has to do with *intent*. A heuristic is a procedure that
>> doesn't necessarily achieve the *desired* result.
>
>I thought that it was obvious to everyone that terminate, in the current
>context, meant terminate with the desired (intended) result. Apparently
>I was wrong. Thanks for the clarification.
One more point about your confusion, which I missed in my last response: I was
commenting what "procedural dimension" means, not what "terminates" means.
The former is what you needed to clarify. Please stop reading so sloppily.
Again, termination is a red herring. The central issue of "heuristic" is its
lack of a *guarantee*, its failure to be a *proof* procedure, its connection
with "common sense", "rule of thumb", "approximation", "educated guess".
--
<J Q B>
Jim Balter
In article <58sa33$n...@usenet.srv.cis.pitt.edu>,
Anders N Weinstein <ande...@pitt.edu> wrote:
>In article <jqbE2B...@netcom.com>, Jim Balter <j...@netcom.com> wrote:
>>In article <58pfgi$b...@usenet.srv.cis.pitt.edu>,
>>Anders N Weinstein <ande...@pitt.edu> wrote:
>>
>>>The problem is, I understand the abstract conception of a computation or
>>>algorithm applied to formal or symbolic elements, such as strings on a
>>>TM tape. I have trouble when one tries to see how to apply it to
>>>arbitrary physical objects.
>>
>>Like a physical computer made of silicon?
>
>No I do not have trouble *there*. In that case it is the social
>context that determines what the devices correct description is, not
>its internal physical structure alone. In another context a similarly
>structured physical object might constitute a room heating device.
Not an efficient one. In some social contexts, Anders' head would make a good
football. Perhaps that's even an efficient use of it. Adios.
--
<J Q B>
David Longley
> In your near complete ignorance of the history of computer science,
> you have the gall to proclaim yourself an expert.
>
I remind you that I've taught university level courses in this
area and when it comes to the history of computing, we are
talking of a very short history. On the other hand, you do
continue to show a very poor grasp of what *I* have been trying
to make *you* understand. All I ask is that you try to read it
*and* understand that there *is* enormous *prejudiced* opposition
to it:
Finally, actuarial methods - at least within the
domains discussed in this article - reveal the upper
bounds in our current capacities to predict human