Why "Cognitive Science" is Pseudo-Science
David Longley
Some have responded with indignation to my charge that "Cognitive Science"
is a pseudo-science. This short note makes my case from a slighly different
point of view.
Whilst the behaviourist simply regarded his scientific domain the behaviour
of animals, and as a scientist, regarded his observations to be fallible
simly as a matter of course, the "Cognitive Scientist" came to study what
they believed to be a representational process itself, namely "cognition".
When this all began, it looked, erroneously in my view to the computer as
a model. Throughout the late 50s and all the way up to the mid 80s, the
models which proliferated in journals like "Memory and Cognition", etc
were essentially information flow models based on formal logic.
What was wrong here is that an implictly methodologically solipsistic
stance had been adopted. The concern was with the representations, ie how
the world was constructed from the point of view of the animal, not how
the world actually may be. Thus, "Cognitive Science" became a bizzare
formal model of how animals (including rats), think about the world.
The alternative is to recognize that observers need to collect lots of
observations and test them, as observations are often only samples of
the real world. If one wants a science of behaviour, one has to begin
by appreciating that the immediate observations have to be treated as
samples of the world.
Kant made this point in contrasting the noumenal and phenomenal world.
Husserl did much the same with his notion of noema and noemata, but
however one wants to label it, the fact remains that science does not
concrn itself with singular observations perspectives or ocassion
sentences. What it pursues are relations in the form of observation
categoricals or conjunctions.
There is a PROFOUND difference between a science of behaviour and the
popular study of "cognition" as a conseqence. The former provides data
which has little to do with how we *know* ourselves simply because it is
based on data which is not generally available to us from a first person
perspective. It only comes out in the analyses of observations (plural).
What the descriptive study of "cognition" reveals is the diversity of
our points of view. It also reveals, in various languages, the biases
which come from only sampling the world from idiosyncractic perspectives.
This is what's familiar - and it's because it's familiar that it's
currently popular, not because it actually succeeds where behaviourism
apparently "failed".
In reality, behaviourism has not failed. In some respects it has just
metamorphosed.
There has, in my view, been a serious misappropriation of the Church-Turing
thesis by "Cognitive Science". It is a thesis which is in fact entirely
consistent with the basic tenets of Logical Empiricism, and therefore
behaviourism.
That this is received by some as a heresy is just a reflection of the sad
state of "Cognitive Science" and poor understanding of behaviourism as
advocated by the Logical Empiricists. I have outlined this more fully in
"Fragments", and have pointed interested readers beyond that to recent
books by Quine.
I am fully aware of the fact that what I am saying is radical... I just ask
that this critique is considered in the context of the full case developed in
Fragments....(which requires more than a passing familiarity with the work
of Quine, and a familiarity with the history of psychology this century).
--
David Longley
Neil Rickert
>Some have responded with indignation to my charge that "Cognitive Science"
>is a pseudo-science.
And rightly so.
> This short note makes my case from a slighly different
>point of view. ^^^^^^^^^^^^^
I doubt it. There is no case to be made.
>Whilst the behaviourist simply regarded his scientific domain the behaviour
>of animals, and as a scientist, regarded his observations to be fallible
>simly as a matter of course, the "Cognitive Scientist" came to study what
>they believed to be a representational process itself, namely "cognition".
>When this all began, it looked, erroneously in my view to the computer as
>a model.
You are guilty of misuse of terminology. As normally used "cognitive
science" is a term for interdisciplinary studies. It includes
psychology, linguistics, artificial intelligence, and other areas.
Behaviorism is part of cognitive science. If cognitive science is
pseudo-science, then it follows that behaviorism is pseudo-science.
As David Yeo pointed out, you are confusing "cognitive science" with
"cognitism" which is a particular set of theories about cognition.
Cognitivism may be the predominant paradigm in cognitive science, but
it is by no means the universally accepted paradigm.
> Throughout the late 50s and all the way up to the mid 80s, the
>models which proliferated in journals like "Memory and Cognition", etc
>were essentially information flow models based on formal logic.
Whereas I suppose you use models based on formal logic, but without
the information? Now that would be a real pseudo-science.
Anders N Weinstein
In article <835211...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>Some have responded with indignation to my charge that "Cognitive Science"
>is a pseudo-science.
It would seem that this is too strong a claim in view of your other claims.
For example, disciplines such as cultural anthropology and ethnography are
not "pseudo-sciences" by your lights are they?
Isn't your beef rather that the study of of varying subjective
perspectives on the world is not thereby a normative theory of
*correct* reasoning about the world? This is a far cry from saying
these disciplines do not themselves discover objective truths about
subjective perspectives.
Doesn't your own position *require* that there be some objective
scientific truths about subjective perspectives? For example, the
(alleged) objective truth that the folk often rely on the
representativeness heuristic? You wrote somewhere that the study of
folk thought processes can itself be scientific. But that means that
this study is not a "pseudo-science" by your lights. Indeed you seem to
believe that Kahneman et. al. are *right* about unconscious processes
involved in folk judgments.
In one way, then, I guess am more radical than you are -- I would urge
deep skepticism about cognitivist attributions of unconscious mental
processes (as does Rickert, apparently). Yet you seem to accept these
cognitive scientific hypotheses at face value, all the while bashing
cognitive science as pseudo-science. I don't get it.
>What was wrong here is that an implictly methodologically solipsistic
>stance had been adopted. The concern was with the representations, ie how
>the world was constructed from the point of view of the animal, not how
>the world actually may be. Thus, "Cognitive Science" became a bizzare
>formal model of how animals (including rats), think about the world.
Let me try to understand your position. It seems to me to be something
like this: there *is* a genuine realm of subjective phenomena which
constitute subjects' representation of the world. In effect, this may
often constitute a kind of folk theory, or confabulation, or, at worst,
hallucination. For example, it is a fact that Don Quixote's
representation of the world contains dragons, the Azande representation
of the world contains an all-pervasive magic. In the same spirit, I
think, one could say that Isaac Newton's story of the world contained
the fictional characters absolute space and time, that Lorentz's at one
time contained the mythical figure of the lumineferous either, and so
on.
In addition to these folk conceptions of the external world, there is
also a kind of second-order fiction, the subjects folk conception of
themselves and their own mental states and processes. This equally is
largely unscientific projection or myth-making.
Now your position *should* be: one *can* study these things, all right, in
an anthropological spirit. There are facts to be discovered here, so the
disciplines that study them are *not* pseudo-sciences. Rather, they are
simply not normative accounts of objectively valid reasoning.
When one is concerned with reality as it is independent of subjective
perspective, on this view, one must shift to the disciplined scientific
theory of the world. Scientific method as it were enables one to penetrate
through the subjectively generated haze and attain a representation of
things as they are in themselves, free of subjectively imposed
projection, etc.
In particular, a disciplined scientific theory of human behavior can
penetrate through the haze of the folk-theory of mental processes to
attain a representation of human behavior as it is in itself, stripped
of the subjective constructions we impose on it when we introspect.
The idea that scientific discipline manages to put us in touch with the
world as it is in iself, apart from subjective impositions, seems to
come naturally to scientifically trained intellectuals. For
example, Noam Chomsky expresses a similar outlook in his recent
lectures published in Mind. Chomsky bashes ordinary language philosophy
and the philosophy concerned with folk notions like intentionality, etc. as
concerned only with culturally variably folk conceptions which have
little role in scientific psychology. In particular, Chomsky
delights in informing us -- with ostentatious dismissiveness -- that the
concept of language has no status in modern scientific linguistics.
Interestingly, of course, for Chomsky, the scientifically real entities
-- "things in the world", he calls them, by contrast to things like the
Chinese language or the city of Beijing, which only exist in our
models of the world -- are *not* behaviors. Rather they are such
entities as internally represented formalisms, which are, of course,
not accessible to introspection either.
It is interesting to me that the same sort of scientistic mythology can be
used both to dismiss and to valorize cognitive science. I would follow
Putnam in urging rejecting of the dualism between the world as it is in
itself and the world as represented in our subjective perspectives.
It is folly to suppose that scientific understanding can magically get at
reality "neat", without the mediation of a conceptualization embodied
in language and culture.
Both common-sense and science are on equal footing insofar as both
require the application of concepts. One can take this to mean that
both alike involve projection or construction or imposition, but I
would rather say that neither does. Neither can be said to get at
reality as it is in itself, apart from description using some
conceptual resources, of course. But that is really an unintelligible
requirement.
David Longley
In article <4q9meq$h...@mp.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>
> You are guilty of misuse of terminology. As normally used "cognitive
> science" is a term for interdisciplinary studies. It includes
> psychology, linguistics, artificial intelligence, and other areas.
> Behaviorism is part of cognitive science. If cognitive science is
> pseudo-science, then it follows that behaviorism is pseudo-science.
>
There is a fundamental difference between behaviour science and "cognitive
science", there are a lrge number of people who regard the latter as no
more than a desperate bid by a relatively small group of estranged academics
to stay in business.
> As David Yeo pointed out, you are confusing "cognitive science" with
> "cognitism" which is a particular set of theories about cognition.
> Cognitivism may be the predominant paradigm in cognitive science, but
> it is by no means the universally accepted paradigm.
I have said before, I am not *confusing* anything. I've ben a professional
psychologist for over 20 years. I know the issues and the journals. If you
read the literature (even the book that Yeo cites) you'll find many of the
people who were responsible for the development of "Cognitive Science"
expressing similar reservations - cf Bruner, Neisser...
>
> > Throughout the late 50s and all the way up to the mid 80s, the
> >models which proliferated in journals like "Memory and Cognition", etc
> >were essentially information flow models based on formal logic.
>
> Whereas I suppose you use models based on formal logic, but without
> the information? Now that would be a real pseudo-science.
>
Instead of "supposing" and "guessing"...just read the literature......
You clearly don't know the literature to be making comments like the above.
--
David Longley
David Longley
In article <4qa2b1$t...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> In article <835211...@longley.demon.co.uk>,
> David Longley <Da...@longley.demon.co.uk> wrote:
> >Some have responded with indignation to my charge that "Cognitive Science"
> >is a pseudo-science.
>
> It would seem that this is too strong a claim in view of your other claims.
> For example, disciplines such as cultural anthropology and ethnography are
> not "pseudo-sciences" by your lights are they?
The latter are descriptive disciplines like history.
>
> Isn't your beef rather that the study of of varying subjective
> perspectives on the world is not thereby a normative theory of
> *correct* reasoning about the world? This is a far cry from saying
> these disciplines do not themselves discover objective truths about
> subjective perspectives.
*Is* it such a far cry?
>
> Doesn't your own position *require* that there be some objective
> scientific truths about subjective perspectives? For example, the
> (alleged) objective truth that the folk often rely on the
> representativeness heuristic? You wrote somewhere that the study of
> folk thought processes can itself be scientific. But that means that
> this study is not a "pseudo-science" by your lights. Indeed you seem to
> believe that Kahneman et. al. are *right* about unconscious processes
> involved in folk judgments.
>
I believe that Tversky & Kahneman are doing behaviour science. The field
is known as Behavioual Decision Theory.
> In one way, then, I guess am more radical than you are -- I would urge
> deep skepticism about cognitivist attributions of unconscious mental
> processes (as does Rickert, apparently). Yet you seem to accept these
> cognitive scientific hypotheses at face value, all the while bashing
> cognitive science as pseudo-science. I don't get it.
>
It's a paradox.....
> >What was wrong here is that an implictly methodologically solipsistic
> >stance had been adopted. The concern was with the representations, ie how
> >the world was constructed from the point of view of the animal, not how
> >the world actually may be. Thus, "Cognitive Science" became a bizzare
> >formal model of how animals (including rats), think about the world.
>
> Let me try to understand your position. It seems to me to be something
> like this: there *is* a genuine realm of subjective phenomena which
> constitute subjects' representation of the world. In effect, this may
> often constitute a kind of folk theory, or confabulation, or, at worst,
> hallucination. For example, it is a fact that Don Quixote's
> representation of the world contains dragons, the Azande representation
> of the world contains an all-pervasive magic. In the same spirit, I
> think, one could say that Isaac Newton's story of the world contained
> the fictional characters absolute space and time, that Lorentz's at one
> time contained the mythical figure of the lumineferous either, and so
> on.
>
> In addition to these folk conceptions of the external world, there is
> also a kind of second-order fiction, the subjects folk conception of
> themselves and their own mental states and processes. This equally is
> largely unscientific projection or myth-making.
>
> Now your position *should* be: one *can* study these things, all right, in
> an anthropological spirit. There are facts to be discovered here, so the
> disciplines that study them are *not* pseudo-sciences. Rather, they are
> simply not normative accounts of objectively valid reasoning.
>
OK so far....
> When one is concerned with reality as it is independent of subjective
> perspective, on this view, one must shift to the disciplined scientific
> theory of the world. Scientific method as it were enables one to penetrate
> through the subjectively generated haze and attain a representation of
> things as they are in themselves, free of subjectively imposed
> projection, etc.
>
Still on track...
> In particular, a disciplined scientific theory of human behavior can
> penetrate through the haze of the folk-theory of mental processes to
> attain a representation of human behavior as it is in itself, stripped
> of the subjective constructions we impose on it when we introspect.
>
Yes...under the limitations of empirical exploration yes..
> The idea that scientific discipline manages to put us in touch with the
> world as it is in iself, apart from subjective impositions, seems to
> come naturally to scientifically trained intellectuals. For
> example, Noam Chomsky expresses a similar outlook in his recent
> lectures published in Mind. Chomsky bashes ordinary language philosophy
> and the philosophy concerned with folk notions like intentionality, etc. as
> concerned only with culturally variably folk conceptions which have
> little role in scientific psychology. In particular, Chomsky
> delights in informing us -- with ostentatious dismissiveness -- that the
> concept of language has no status in modern scientific linguistics.
>
I think Chomsky may be coming round to a better appreciation of Skinner
and Quine.......one sees as much in the film "Manufacturing Consent"...
--
David Longley
Bennett
> In article <4q9meq$h...@mp.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert"
writes:
> >
> > You are guilty of misuse of terminology. As normally used "cognitive
> > science" is a term for interdisciplinary studies. It includes
> > psychology, linguistics, artificial intelligence, and other areas.
> > Behaviorism is part of cognitive science. If cognitive science is
^^^^^^^^^^^
> > pseudo-science, then it follows that behaviorism is pseudo-science.
>
> There is a fundamental difference between behaviour science and "cognitive
^^^^^^^^^^^^^^^^^
> science", there are a lrge number of people who regard the latter as no
> more than a desperate bid by a relatively small group of estranged academics
> to stay in business.
Do you equate behaviourism and behaviour science? I see an analogy
between these terms and cognitism/congnitive science.
I believe that, as Neil pointed out, cognitive science refers just to a
particular group of areas of study. I infer from your postings that you
regard it as more than this. You seem to think of cognitive science as
either standing for some particular (dubious) theory of cognition, or
being fundamentally based on certain (dubious) assumptions.
So you seem to be saying, not that cognitive science (the study of
cognition) is pseudo-science, but rather that cognitive science (a large
part of the work that is done today under the banner of cognitive science)
is pseudo-science.
Am I right? Or do you object to the idea of studying cognition at all?
> > As David Yeo pointed out, you are confusing "cognitive science" with
> > "cognitism" which is a particular set of theories about cognition.
> > Cognitivism may be the predominant paradigm in cognitive science, but
> > it is by no means the universally accepted paradigm.
>
> I have said before, I am not *confusing* anything. I've ben a professional
> psychologist for over 20 years. I know the issues and the journals. If you
> read the literature (even the book that Yeo cites) you'll find many of the
> people who were responsible for the development of "Cognitive Science"
> expressing similar reservations - cf Bruner, Neisser...
Forgive me, but I still get the impression that your postings are indeed
confusing cognitive science and cognitism, your credentials
notwithstanding. Why not start with a clear definition of cognitive
science, so we all know what you're talking about. This is a pretty
important issue -- if cognitive science really is pseudo-science, then I
think we should be told!
(I wonder why I always get so verbose when posting on serious topics.)
Bennett.
--
/ mce...@uow.edu.au / I reserve the right to be wrong.
David Longley
The following is a partial answer to ..... It's the best I can offer outside
what I have provided at length in "Fragments.."..
The "Cognitive Sciences" are supposed to be:
Psychology
Artificial Intelligence
Philosophy
Linguistics
Anthropology
and even Neuroscience!
Dealing with neuroscience first, at a guess, 95% of neuroscience is done on
rats and cats and lower animals, most of it at the anatomical and chemical
level. Most neuroscientists have little interest in anything outside their
own specialism, which I reckon in a overwhelming proportion of labs has
absolutley nothing to do with psychology of any sort (including behavioural).
AI is what I have been focusing on in these threads. I have been arguing
that any attempts to model "intelligence" are misdirected if they are
concerned with the classic intensional idioms. On the other hand, if it
is directed to the development of better technology in decision theory,
which is normative, that is classic behavioural work and sits neatly with
tghe work which Frege, Carnap, Church, Turing, Post, Kleene etc began as
the logicist/formalist tradition.
I have emphasised Quine's work since I (and many others) regard him as the
most influential living philosopher, he may turn out to be this century's
most influential philosopher. In a nutshell, he reduces epistemology to
a chapter of empirical (behavioural) psychology.
Again, there is a quite radical and potentially devastating Quinean critique
of contemporary linguistics, which is of a piece with his challenge to one
of the cenral concerns of Cognitive Science, namely "meaning" and the
notion of internal representation.
When it comes to psychology, I have outlined what is, I think, quite an
accurate picture of the nature of contemporary psychology. It is in fact
a descriptive account of the vicissitudes of folk psychological judgment.
This is studied under different names, "conditioning", "attribution theory",
"Personal Construct Theory", "natural assessments/heuristics" and so on.
All of this is interesting stuff, but basically useless as a source of
material for the applied psychologist. It is like the last element of this
odd fusion of subjects "anthropology".
What I am explictly criticisin is the muddle this all results in. It leads
to daft notions like behavioural approaches being subsumed to "cognitive".
The result is that people are totally thrown when they learn of the work
on "Clinical vs Actuarial Judgement", itself an outgrowth of the frankly
descriptive study of how people use rules of thumb when they do not have
normative rules to turn to.
It leads to people claiming that the effective components of programmes
for prisoners are "cognitive" and that "abstract reasoning" is something
to be striven for in education....
What I have endeavoured to point out is:
1) The above is all empirically WRONG
2) How and why these enormous mistakes have come about...
Although not entirely the answer - people do not want a science of behaviour,
it's too much hard work...they seem to want a "good read instead", and that's
what they are getting in "Cognitive Science" - at the expense of a professional
science and technology of behaviour (See Dawes 1994 "House of Cards: Psychology
and Psychotherapy Built on Myth").
--
David Longley
Anders N Weinstein
In article <835299...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>When it comes to psychology, I have outlined what is, I think, quite an
>accurate picture of the nature of contemporary psychology. It is in fact
>a descriptive account of the vicissitudes of folk psychological judgment.
>This is studied under different names, "conditioning", "attribution theory",
>"Personal Construct Theory", "natural assessments/heuristics" and so on.
So none of these studies is "pseudo-science", on your view. By that I
mean that they all arrive at objective truths about their subject matter,
as far as you are concerned.
>All of this is interesting stuff, but basically useless as a source of
>material for the applied psychologist.
I wonder why you say this. If you know people are prone to certain
errors, say to ignore base rates, isn't that a useful fact for applied
psychologists ? You know then to teach people to pay attention to base
rates where relevant, for example.
Another question: why not say that the articulation of normative
standards of rationality is the job of logic and philosophy, not
"behavior science". Clearly "behavior science" is every bit as
non-normative as any of the other disciplines you criticize. It studies
how people actually behave, not how they rationally ought to behave. As
such, behavior science is irrelevant to any epistemological concern.
It's not even as relevant as anthropology, which you criticize.
On this anti-psychologistic view the hopeful idea of "psychology" should
really be split in two: empirical or naturalistic psychology can study
actual courses of reasoning; logic and philosophy can attempt the
articulation of correct principles of reasoning.
But it is only confusion to think there is an epistemologically
interesting natural science of rational processes. That would be to
confuse causal regularities (brute tendencies) with genuine norms. You
can't get there from here.
The idea I like is that the irreducibility of intentional notions to
the ontology of the natural sciences is just a by-product of the
irreducibility of normative to the non-normative. This irreducibility
should not be very alarming; still, Quine's naturalism depends on missing
the point.
You on the other hand want it both ways. How do you think any normative
force ever gets into the naturalist picture? Isn't behavior science
equally a norm-free, purely descriptive discipline?
David Longley
In article <mcelwee-2006...@ppp17.cc.uow.edu.au>
mce...@uow.edu.au "Bennett" writes:
> Forgive me, but I still get the impression that your postings are indeed
> confusing cognitive science and cognitism, your credentials
> notwithstanding. Why not start with a clear definition of cognitive
> science, so we all know what you're talking about. This is a pretty
> important issue -- if cognitive science really is pseudo-science, then I
> think we should be told!
>
I am not going to give a clear definition of "Cognitive Science", since
even Gardner fails to do this adequately. What MUST be acknowledged is
the fact that it was premised on perceived FAILURE of behaviourism, and
it looked to the computer as a model for thinking. Now there are very
simple and very clear things which can be said about this. First of all,
behaviourism has NEVER been in decline in a scientific sense, in fact,
behavioural principles and techniques are about the ONLY things which
are of any use professionally. The decline of behaviourism is really just
a fanstasy of those who long for a return to the familiarity of mentalism.
SECONDLY, and somewhat ironically, the computer is a perfect vindication
of logical positivism. Turings 1937 paper was about making the BEHAVIOUR
of a computer (a person) explictly stateable in such a way that it could
be done by a machine. Very shortly after the design of the computer and
the publication of Shannon's paper on Information Theory, psychologists
started using the language which Wiener, Shannon etc were using to
talk about information, and then memory...and because they had physical
models which they could point to as operating, mechanical or electrical
systems. It's ironic, because Frege, Wittgenstein and all who followed,
including Carnap were explicit in their efforts to DESPSYCHOLOGIZE, ie
DE-COGNITIVIZE what they had to say about logic, which is a formal
normative discipline.
Today we have folk like Johnson-laird building mental models of how we
might think deductively, and others doing the same for induction, but
they may as well be building models of how we do flower-arranging,
bicycle riding etc etc.. They are building models of the SKILLS people
learn......but WHY????
Skinner made this point long ago - asking WHY do this? Wittgenstein did
the same, pointing out that nothing was really hidden here....
All in all, I think the last 30 to 40 years have just been a period of
fanciful creative science fiction.....little or nothing of scientific
value has come of it....and if anyone seriously thinks otherwise, I'd
really like to hear what they think it is. As an applied psychologist,
I have to look very hard for anything of practical import whatsoever.
Imagine, if you will, a team of applied psychologists who are being
asked to help managers manage difficult inmates.....what does Cognitive
Science have to offer?
--
David Longley
David Longley
In article <4qcfgj$8...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> I wonder why you say this. If you know people are prone to certain
> errors, say to ignore base rates, isn't that a useful fact for applied
> psychologists ? You know then to teach people to pay attention to base
> rates where relevant, for example.
But we are already doing that in statistics etc, we KNOW that - else
why bother developing statistucal methods... My point is that studying
the intricacies of what happens when we DON'T use normative techniques
is not, when the chips are down much use except to motivate people to use
normative methods. And, to do that you might as well just point out that
knowing how to use normative techniques is better than not doing so -
a bit like saying maths is useful when it comes to bridge design!
--
David Longley
Peter_...@uk.ibm.com
In <4qcfgj$8...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>
>On this anti-psychologistic view the hopeful idea of "psychology" should
>really be split in two: empirical or naturalistic psychology can study
>actual courses of reasoning; logic and philosophy can attempt the
>articulation of correct principles of reasoning.
>
>But it is only confusion to think there is an epistemologically
>interesting natural science of rational processes. That would be to
>confuse causal regularities (brute tendencies) with genuine norms. You
>can't get there from here.
Could you explain the reasoning a little more closely? I'll just offer a
possibility that you might like to respond to. Suppose we thought that
rationality was grounded in evolutionary pressures to do with how
an organism should make use of brute regularities to survive. And that
in order to achieve this, the organism would evolve to represent such
regularities since (as Popper puts it) it is survival-inducing to let our
ideas die (or fail) in our stead. This could be given, as I understand it,
a naturalistic account. If so, then certain other things follow - such as,
in such and such circumstances and, wrt the role of reason, such and
such behaviour would tend to be survival-inducing and so rational.
Now you might respond that the rationality of the desire to survive
is part of the normative story, and that I've squirreled this in illicitly.
But is it so? Do we know that the survival of the agent is incidental
to normative behaviour? It seems to me a difficult argument to claim
that the survival of norm-holders is just incidental to the norms that
they hold.
As I see it, this makes it possible for a very great part of what rational
norms are to fall within the sphere of science. Indeed, if you ask me
which part of the structure falls wholly within the realm of logic and
philosophy, I'd be hard pressed to say what. Seems to me that the
boundaries are blurring.
>The idea I like is that the irreducibility of intentional notions to
>the ontology of the natural sciences is just a by-product of the
>irreducibility of normative to the non-normative. This irreducibility
>should not be very alarming; still, Quine's naturalism depends on missing
>the point.
I see the relationship you give, that the irreducibility of the normative
implies the irreducibility of the intentional. What is not clear to me is
what particular aspect of normativity is irreducible - surely not the
whole shebang? There seems to me to be a very greal deal in and around
the area of normativity that could well be reducible. So what is this
irreducible core and what scope does science have with the rest?
The question of how evolving organisms exploit the structure of brute
regularity to further their own survival prospects strikes me as a wholly
naturalistic investigation, which science will do, and which is probing the
structure of rationality. That some philosophers feel this is their territory is
neither here nor there.
>You on the other hand want it both ways. How do you think any normative
>force ever gets into the naturalist picture? Isn't behavior science
>equally a norm-free, purely descriptive discipline?
The above questions were directed at Longley, not to me. What I've
done is to try to explain how normative force could get into the
naturalistic picture, through the investigation of what organisms *ought*
to do to compete and so survive. I'm not sure why this isn't a perfectly
good answer to the first of Anders' questions.
Cheers,
Pete Lupton
David Allan Lewis
David Longley (Da...@longley.demon.co.uk) wrote:
: I am not going to give a clear definition of "Cognitive Science", since
: even Gardner fails to do this adequately. What MUST be acknowledged is
: the fact that it was premised on perceived FAILURE of behaviourism, and
: it looked to the computer as a model for thinking. Now there are very
: simple and very clear things which can be said about this. First of all,
: behaviourism has NEVER been in decline in a scientific sense, in fact,
: behavioural principles and techniques are about the ONLY things which
: are of any use professionally. The decline of behaviourism is really just
: a fanstasy of those who long for a return to the familiarity of mentalism.
What also MUST be acknowledged is that the "decline of behaviourism"
MUST be regarded from a RELATIVE standpoint. ie, It is no longer
the absolute paradigm for psychological understanding, but is used as
a tool for a more COMPLETE comprehension and exploration of the human
creature. So you are in once sense accurate, that behaviourism's
scientific value hasn't declined (at least, not to irrelevancy). But
its decline is not a "fantasy" by some researchers wandering in a
haze of mysticism. It has become a tool, not an imperative.
[...]
: Today we have folk like Johnson-laird building mental models of how we
: might think deductively, and others doing the same for induction, but
: they may as well be building models of how we do flower-arranging,
: bicycle riding etc etc.. They are building models of the SKILLS people
: learn......but WHY????
: Skinner made this point long ago - asking WHY do this? Wittgenstein did
: the same, pointing out that nothing was really hidden here....
Do you mean NOTHING was hidden, or nothing was HIDDEN? There's a
large difference.
If you're not interested in anything more than the statistical behaviours
of inmates for the Ministry, then it's not surprising you find little
value in Cognitive Science. You can keep asking WHY?? and it's not
going to get any clearer, because you're not asking very good
questions. Maybe if you asked, "Is there any difference in the flower
gathering methods of difficult inmates and managable inmates? Are
these same differences observable in their mealtime behaviours? What
does this tell me about the difficult qualia?" But you're not.
: All in all, I think the last 30 to 40 years have just been a period of
: fanciful creative science fiction.....little or nothing of scientific
: value has come of it....and if anyone seriously thinks otherwise, I'd
: really like to hear what they think it is. As an applied psychologist,
: I have to look very hard for anything of practical import whatsoever.
: Imagine, if you will, a team of applied psychologists who are being
: asked to help managers manage difficult inmates.....what does Cognitive
: Science have to offer?
It is this applied view that biases you away from comprehension of the
value of Cognitive Science. If all you want to do is "manage
difficult inmates", collect your paycheck, and go home, then the
current research-orientation of Cogntive Science will OF COURSE have
little value to you. Pavlov's empirical studies of behaviourism bore
this same element. The fact is, you haven't TRIED to apply any of the
discoveries of Cognitive Science, you want the researchers to hand
your ideas to you. But they're busy pursuing understanding (or grants
:) instead of writing philosophical treatises for you to base your own
thoughts on. You ask what Cognitive Science has to offer. I ask if
you are looking for anything. It seems not, and thus it's not
surprising you've found nothing "of practical import whatsoever."
When there are artificial management agents observing, predicting, and
advising about behaviour instead of you, will you still wonder what
Cognitive Science has to offer you in looking for a job?
dave, more full of hot air than he thought.
+++++
David Lewis | dal...@sccs.swarthmore.edu
Swarthmore College | http://www.sccs.swarthmore.edu/~dalewis/
500 College Ave. | (610)-604-4533 Until August 31.
Swarthmore, PA 19081-1397 | (610)-690-2736 After August 31.
'It never is overcome in some people, whose work, to the end of life, gets
done in the interstices of their mind-wandering."--William James,
_Principles of Psychology_
+++++
Anders N Weinstein
In article <4qdoil$19...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
>In <4qcfgj$8...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>>
>>But it is only confusion to think there is an epistemologically
>>interesting natural science of rational processes. That would be to
>>confuse causal regularities (brute tendencies) with genuine norms. You
>>can't get there from here.
>
>Could you explain the reasoning a little more closely? I'll just offer a
Maybe not. It was part of a broad-brush ad hominem against Longley.
Longley pushes anti-psychologism about certain norms of rationality
as a criticism of certain sorts of research. An example I guess would be
Johnson-Laird's sort of study of "how people actually reason". This seems
like a fair view to me.
Yet Longley goes on to defend naturalism and "behavior science". But behavior
science equally obliterates all normative force. This was all I meant by
the original point.
Now: what to say about "teleofunctionalism"? This is, I admit, a much more
serious form of naturalism. This label might apply to e.g. Fred
Dretske, Ruth Garrett Millikan, some essays of Dan Dennett, and, I gather,
yourself.
On this view, the dichotomy I was relying on -- between causal
regularities (tendencies) and norms -- is too crude. Evolutionary
biology is, it would seem, a natural science in good standing (though
doubts have often been raised on this score). Yet it is up to its
eyeballs in norms, namely, norms of proper function. So, this sort of
naturalism supposes that one *can* get to norms of rationality from
purely natural norms of proper function.
I should add, for fans of computationalism, that the very idea of a
computer is itself a normative one. The formal rules of a computer
program are in some sense idealizations which stand as norms -- not
mere tendencies or regularities -- for an actual physical device. A
computer can *mal*function, a planet going around its orbit cannot. So
computational explanation also has to seek the ground of its normative
force. So if, as I suppose, it is the mark of the domain of the
Physical that it's explanations are based on norm-free laws, then the
Biological as well as the Computational are equally irreducible domains
(though of course none the worse for that).
>possibility that you might like to respond to. Suppose we thought that
>rationality was grounded in evolutionary pressures to do with how
>an organism should make use of brute regularities to survive. And that
>in order to achieve this, the organism would evolve to represent such
>regularities since (as Popper puts it) it is survival-inducing to let our
>ideas die (or fail) in our stead. This could be given, as I understand it,
>a naturalistic account. If so, then certain other things follow - such as,
>in such and such circumstances and, wrt the role of reason, such and
>such behaviour would tend to be survival-inducing and so rational.
I can think of a couple of problems with this sort of approach.
it would seem that false beliefs might be more conducive to survival then
true ones in many circumstances. Yet we still think the true ones are the
norm. It might also be that cockroaches will yet outperform us as survival
machines yet have no reason, no representation, no true beliefs about the
world of any kind. So I think the norm of truth can come apart from that
of survival.
To return to anti-psychologism about logic. Is it your view
that we accept modus ponens because evolution has discovered it is more
useful than other rules? Isn't it rather that the explanation of its
utility depends on its being logically valid (truth-preserving)?
I don't think one can think of logical norms as simply a quirk of human
biology, so that there might be rational Martians who by nature
employed a different rule. Rather failure to display cognitive
sensitivity to certain minimal norms of logic simply means they are not
rational or proposition-using creatures at all (though they might be
otherwise well adapted to their niche in the merely animal way).
I guess I am what you might call a Fregean rationalist: these norms can
not be understood to be grounded in evolutionary biology, but are in
some sense timeless and impersonal, like, say, the number system. Not
to mention the fact that any biological inquiry or explanation of
"survival value" must itself make use of these norms.
I have been repeatedly told that the problem of reflexivity is old hat for
naturalists. But I don't myself see what's wrong with it -- I think norms
of logic and rationality are more like constitutive rules which make
possible the game of scientific inquiry, and so are not themselves capable
of investigation by that inquiry in any non-circular way.
>As I see it, this makes it possible for a very great part of what rational
>norms are to fall within the sphere of science. Indeed, if you ask me
>which part of the structure falls wholly within the realm of logic and
>philosophy, I'd be hard pressed to say what. Seems to me that the
>boundaries are blurring.
So evolutionary biology tells us why, necessarily 2 + 2 = 4? I don't
think so.
David Longley
> On this anti-psychologistic view the hopeful idea of "psychology" should
> really be split in two: empirical or naturalistic psychology can study
> actual courses of reasoning; logic and philosophy can attempt the
> articulation of correct principles of reasoning.
>
This *is* essentially what I have been proposing, except that I'd
say that Artificial Intelligence has taken over from
logic/philosophy. It certainly shouldn't be put under the
essentially psychologistic umbrella of "Cognitive Science". It's
a technology......cybernetics in fact.
> But it is only confusion to think there is an epistemologically
> interesting natural science of rational processes. That would be to
> confuse causal regularities (brute tendencies) with genuine norms. You
> can't get there from here.
>
Yes, that's why I think Johnson-Laird etc are basically doing
"sky-writing" - although his work may explain the errors which
humans manifest when trying to be logical. What it should make
clear is why it's wrong for AI folk to look to what *people*
do..(rather than what folk like Frege and Church/Turing have
*done*).
> The idea I like is that the irreducibility of intentional notions to
> the ontology of the natural sciences is just a by-product of the
> irreducibility of normative to the non-normative. This irreducibility
> should not be very alarming; still, Quine's naturalism depends on missing
> the point.
>
> You on the other hand want it both ways. How do you think any normative
> force ever gets into the naturalist picture? Isn't behavior science
> equally a norm-free, purely descriptive discipline?
>
I think, prima facie, this last point is true. However. it may
well turn out that the normative elements have their "roots" (as
Quine would say), in fragments of our behaviour. It may well be
that Husserl has a little of it right too, that we have basically
sifted out a core set of principles from behaviour, which,
because of natural contingencies (ie their complexity) are just
stretched beyond their tolerance level. One has to be quite
careful in one's reading of the research on human behaviour as a
consequence....It is a true example of Research AND Development.
In this way, artificial intelligence picks up on a core behaviour
and puts it into development....
From this perspective, it is just an inevitable consequence of
the confusion in this area that results in psychologists
(behaviour scientists) working on the normative principles of
decision making. Really, it has little to do with them, and
should be a technology developed by mathematicians or as
suggested above, and in "Fragments"....
ARTIFICIAL INTELLIGENCE
--
David Longley
David Longley
In article <4qec3l$4...@larch.cc.swarthmore.edu>
dal...@condor.sccs.swarthmore.edu "David Allan Lewis" writes:
>
> If you're not interested in anything more than the statistical behaviours
> of inmates for the Ministry, then it's not surprising you find little
> value in Cognitive Science. You can keep asking WHY?? and it's not
> going to get any clearer, because you're not asking very good
> questions.
But there *are* important technological developments as I have
said in "Fragments"..specifically, inmate programmes (sex
offender, cognitive skills, anger management - In the case of the
former, "cognitive skills" is really quite central. Yet, again as
I have reviewed in "Fragments", what the research on "Reasoning"
suggests is altogether something quite different from what the
folk running these programmes are actually doing (here. at least
the problem is primarily wit the latter folk, not the Cognitive
Scientists).
leaving prisons aside, the same points go for education, and
there's no reason why we should not include clinical work as
well.
> Maybe if you asked, "Is there any difference in the flower
> gathering methods of difficult inmates and managable inmates? Are
> these same differences observable in their mealtime behaviours? What
> does this tell me about the difficult qualia?" But you're not.
>
Well, actually....... leaving flower gathering,aside, you clearly
have NOT read "Fragments" have you.....The above is precisely
what PROBE *is* designed to do...please have a look...
--
David Longley
Neil Rickert
In <835375...@longley.demon.co.uk> David Longley <Da...@longley.demon.co.uk> writes:
>In article <4qcfgj$8...@usenet.srv.cis.pitt.edu>
> ande...@pitt.edu "Anders N Weinstein" writes:
>> On this anti-psychologistic view the hopeful idea of "psychology" should
>> really be split in two: empirical or naturalistic psychology can study
>> actual courses of reasoning; logic and philosophy can attempt the
>> articulation of correct principles of reasoning.
>We are making progress now I think...
>This *is* essentially what I have been proposing, except that I'd
>say that Artificial Intelligence has taken over from
>logic/philosophy. It certainly shouldn't be put under the
>essentially psychologistic umbrella of "Cognitive Science". It's
>a technology......cybernetics in fact.
In that case, I can almost agree also.
Put in those terms, my disagreement with David Longley is that he is
attempting to dictate what he takes to be a priori principles of
correct reasoning.
In the sense of this discussion, what I am doing is investigating
what should be the principles of reasoning. I remove the word
"correct" because I don't want to make any a priori assumptions that
there is a specific correct form of reasoning. Thus I see the
subject of appropriate forms of reasoning to be one that requires
research, rather than one of making dogmatic assertions. David, by
contrast, takes it as a given that Frege solved that problem and that
there is nothing more to be done.
Anders N Weinstein
In article <835375...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>> You on the other hand want it both ways. How do you think any normative
>> force ever gets into the naturalist picture? Isn't behavior science
>> equally a norm-free, purely descriptive discipline?
>
>I think, prima facie, this last point is true. However. it may
>well turn out that the normative elements have their "roots" (as
>Quine would say), in fragments of our behaviour. It may well be
>that Husserl has a little of it right too, that we have basically
>sifted out a core set of principles from behaviour, which,
>because of natural contingencies (ie their complexity) are just
>stretched beyond their tolerance level. One has to be quite
>careful in one's reading of the research on human behaviour as a
>consequence....It is a true example of Research AND Development.
I am glad to find a point of agreement. But I don't see that this
addresses the worry about inconsistency. I take it anti-psychologism
to entail, in broad terms, that rational activity (or semantics) is not
and cannot be made the object of a nomological science in the manner of
physics or chemistry.
Recall Frege making fun of the idea that I should
measure the amount of magnesium in my brain in order to judge if 2 + 2 = 4.
Yet many of your posts express great interest in such issues as brain
chemistry, issues which are completely irrelevant to *logical* matters
like the cogency of someone's arguments. Flaws in brain chemistry can
*interfere* with correct reasoning, sure, but chemistry cannot explain
rational force.
I think most of the errors and confusions in philosophy of mind come
from contemporary (as opposed to, say, neo-Aristotelian) naturalism.
Throughout this newsgroup and, alas, even in a lot of what passes for
professional philosophy of mind, one sees again and again the blinkered
attempt to try to fit rational processes into the procrustean bed of
scientific method -- e.g. to suppose there *must* be causal laws of
succession to be found under psychological descriptions.
The anti-psychologistic point is that norms of rationality are *not*
laws of succession in the psychic sphere, and that intentional
explanation is sui generis because it functions by locating its explanada
by reference to a normatively structured space ("the logical space
of reasons").
>In this way, artificial intelligence picks up on a core behaviour
>and puts it into development....
I would distinguish the anti-psychologistic point about normative force
from the Formalist Fantasy -- the idea that the norms in question are
rigid typographical rules, like Carnap's logical syntax, and could be
embodied in an algorithm. I suspect rather after Dreyfus that the norms
in question are soft, and live as it were mainly in the background of
unreflective practice. They need not be explicable as a system formal
rules, and there is no reason to suppose they can be imparted to a
mechanical calculator that doesn't share our (or any!) form of life. If
the recently departed Kuhn was right, this is the case with scientific
practice as well. Even the formal calculi only get their sense from
embedding in this background.
David Longley
> In <835375...@longley.demon.co.uk> David Longley <Da...@longley.demon.co.uk>
> writes:
> >In article <4qcfgj$8...@usenet.srv.cis.pitt.edu>
> > ande...@pitt.edu "Anders N Weinstein" writes:
>
> >> On this anti-psychologistic view the hopeful idea of "psychology" should
> >> really be split in two: empirical or naturalistic psychology can study
> >> actual courses of reasoning; logic and philosophy can attempt the
> >> articulation of correct principles of reasoning.
>
> >We are making progress now I think...
>
> >This *is* essentially what I have been proposing, except that I'd
> >say that Artificial Intelligence has taken over from
> >logic/philosophy. It certainly shouldn't be put under the
> >essentially psychologistic umbrella of "Cognitive Science". It's
> >a technology......cybernetics in fact.
>
> In that case, I can almost agree also.
>
Good, perhaps we should build on that.
> Put in those terms, my disagreement with David Longley is that he is
> attempting to dictate what he takes to be a priori principles of
> correct reasoning.
To acknowledge that the predicate calculus is a sound set of
principles which the laws of physics can be expressed in terms
of, and that to date we know of no better is not a dogmatic
assertion of a priori principles, but statement of empirical
fact. I have aligned what I have to say with Quine's critique of
analyticty in "Two Dogmas of Empiricism" often enough to make
that point clear. Attributing this "a priorism" to me is just
wrong, and indicative of a fundamental misunderstanding as I have
said many times. of Empiricism
>
> In the sense of this discussion, what I am doing is investigating
> what should be the principles of reasoning. I remove the word
> "correct" because I don't want to make any a priori assumptions that
> there is a specific correct form of reasoning. Thus I see the
> subject of appropriate forms of reasoning to be one that requires
> research, rather than one of making dogmatic assertions. David, by
> contrast, takes it as a given that Frege solved that problem and that
> there is nothing more to be done.
>
>
This is not true either, I'd very much like to hear of
developments in formal logic (other than that of modal logic).
I'd like to see more in the way of automated deduction
(reasoning) - does anyone know of any decent PC packages (I only
have OTTER).
--
David Longley
Neil Rickert
In <835387...@longley.demon.co.uk> David Longley <Da...@longley.demon.co.uk> writes:
>In article <4qepmb$4...@mp.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
[AW]
>> >> On this anti-psychologistic view the hopeful idea of "psychology" should
>> >> really be split in two: empirical or naturalistic psychology can study
>> >> actual courses of reasoning; logic and philosophy can attempt the
>> >> articulation of correct principles of reasoning.
[DL]
>> >This *is* essentially what I have been proposing, except that I'd
>> >say that Artificial Intelligence has taken over from
>> >logic/philosophy. It certainly shouldn't be put under the
>> >essentially psychologistic umbrella of "Cognitive Science". It's
>> >a technology......cybernetics in fact.
[NR]
>> In that case, I can almost agree also.
>Good, perhaps we should build on that.
>> Put in those terms, my disagreement with David Longley is that he is
>> attempting to dictate what he takes to be a priori principles of
>> correct reasoning.
>To acknowledge that the predicate calculus is a sound set of
>principles which the laws of physics can be expressed in terms
>of, and that to date we know of no better is not a dogmatic
>assertion of a priori principles, but statement of empirical
>fact.
No, sorry, it is dogma.
The predicate calculus may be a sound set of principles if:
the structure of the universe is defined by a finite set of
exact physical laws.
the world is predictable, given those laws.
we know what the laws are.
However, that is not the situation we find ourselves in. Rather we
live in an uncertain, and possibly unpredictable world. We do not
have complete and certain laws giving the structure of that
universe. There is no guarantee that the universe is definable in
terms of a finitely generated system of laws.
You are making a priori assumptions about the adequacy of today's
physics and about the complexity of the structure of the universe.
What we need are reasoning principles which apply the the uncertainty
we must deal with. There is no certainty that reasoning principles
designed for an ideal world are suitable.
> I have aligned what I have to say with Quine's critique of
>analyticty in "Two Dogmas of Empiricism" often enough to make
>that point clear.
I don't see how Quine's critique helps you here. If the laws of
logic are not analytic, then Hume's skepticism applies and they might
be the wrong laws. Even if you take logic as analytic, if the laws
of physics are not analytic, Hume's skepticism applies and there is
no certainty that the universe will obey those laws of physics.
> Attributing this "a priorism" to me is just
>wrong, and indicative of a fundamental misunderstanding as I have
>said many times. of Empiricism
Where am I wrong? Are you not making the a priori assumptions I
suggested? And if you are not making those assumptions, then what
basis is there to assume that the predicate calculus is an
appropriate reasoning method for the universe that we inhabit?
David Longley
In article <4qf0up$h...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> I am glad to find a point of agreement. But I don't see that this
> addresses the worry about inconsistency. I take it anti-psychologism
> to entail, in broad terms, that rational activity (or semantics) is not
> and cannot be made the object of a nomological science in the manner of
> physics or chemistry.
>
> Recall Frege making fun of the idea that I should
> measure the amount of magnesium in my brain in order to judge if 2 + 2 = 4.
> Yet many of your posts express great interest in such issues as brain
> chemistry, issues which are completely irrelevant to *logical* matters
> like the cogency of someone's arguments. Flaws in brain chemistry can
> *interfere* with correct reasoning, sure, but chemistry cannot explain
> rational force.
Well, that's an a priori legislative statement if ever there was
one!
As to the rest of your (essentially Husserian) post - I think
your "lifeworld" is somewhat different from mine <g>.
--
David Longley
David Longley
In article <4qf8v3$9...@mp.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>
> Where am I wrong? Are you not making the a priori assumptions I
> suggested? And if you are not making those assumptions, then what
> basis is there to assume that the predicate calculus is an
> appropriate reasoning method for the universe that we inhabit?
>
The predicate calculus, at least in the form of database
technology, and principles of set theory/mathematical logic as we
know them, are used all the time by engineers and other applied
technologists.
If called upon to use one's professional skills to solve a
problem, I think most would say that they trusted this technology
(often implemented in computer hardware/software) to help them
manage and hopefully solve their problems.
I'm basically looking for new techniques not the "furniture of
the universe". ANNs are probably NOT what's needed, as we already
have the different types of regression and discriminant analysis
where we can see in detail how the weights are worked out....
Perhaps we should leave it here now that there has been at least
a modicum of agreement.....
--
David Longley
Neil Rickert
In <835401...@longley.demon.co.uk> David Longley <Da...@longley.demon.co.uk> writes:
>In article <4qf8v3$9...@mp.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>> Where am I wrong? Are you not making the a priori assumptions I
>> suggested? And if you are not making those assumptions, then what
>> basis is there to assume that the predicate calculus is an
>> appropriate reasoning method for the universe that we inhabit?
>The predicate calculus, at least in the form of database
>technology, and principles of set theory/mathematical logic as we
>know them, are used all the time by engineers and other applied
>technologists.
I have never suggested otherwise. But your database technology is
only as good as the data recorded in that database. If you filled
your PROBE database with data on the number, color composition, and
length of the hairs of each prisoner's head, you would not find the
database very useful. Predicate calculus can only tell you how to
use what is in the database. It does not tell you what to put in
that database. Yet what data to use is surely part of any effective
reasoning method.
>If called upon to use one's professional skills to solve a
>problem, I think most would say that they trusted this technology
>(often implemented in computer hardware/software) to help them
>manage and hopefully solve their problems.
Agreed. The important words are "to help them ..." . The technology
cannot solve the problems alone.
>I'm basically looking for new techniques not the "furniture of
>the universe".
You won't find them, for your strict extensionalism is too committed
to a priori assumptions as to what is an extension.
>Perhaps we should leave it here now that there has been at least
>a modicum of agreement.....
It is a very small modicum indeed.
David Longley
In article <4qgohn$h...@mp.cs.niu.edu> ric...@cs.niu.edu "Neil Rickert" writes:
>
> >Perhaps we should leave it here now that there has been at least
> >a modicum of agreement.....
>
> It is a very small modicum indeed.
>
hope for. Since the technology I am looking for MUST be
implemented in computer hardware or software, I reckon my
extensionalism is quite safe. I had hoped to elicit a little more
in the way of practical, constructive pointers from those working
in Knowledge Engineering and expert systems more generally. I
think the AI folk and some of those working in Statistics ought
to join forces, and not just in pursuit of ANN technology.
--
David Longley
David Longley
"In the 1970's, Cognitive Science was formulated in
terms of Putnam's doctrine of functionalism, in which
mental states are the functional states of an abstract
digital computer, thinking is abstract symbol
manipulation as in the operation of a computer program,
and the mind's symbols get their meaning by denoting
things in the world. Functionalism has since become the
mainstream doctrine within the philosophy of mind. But
in the mid-1970's, overwhelming empirical evidence
against functionalism began to pile up, and in the late
1970's, Putnam himself found functionalism to be
logically incoherent. In this volume, Putnam brilliantly
reveals the philosophical fallacies in the doctrine he
founded, showing why functionalism must fail as a
philosophy of mind. The fall of functionalism has major
consequences for generative linguistics, artificial
intelligence, and cognitive and developmental
psychology".
Lakoff (1988) in review of H Putnam (1988)
Representation and Reality
My original criticism remains that "Cognitive Science" *was*
conceived on the assumption that the computer and deductive
reasoning more generally was a sound model for "the mind". (The
early criticisms of the anti-psychologists had just gone unheeded
or unrecognised by "psychologists" etc. working on cognitive
processes such as "remembering","seeing", and "thinking" (to put
it in propositional attitude terms).
However, it *was* upon this basis that functionalism was proposed
by Putnam in the 1960s and 70s, and it *was* with this in mind
that the so called rationality assumption was taken as a
justification for research into the presumed mechanics of
attribution, ie the propositional attitudes.
I think it is now indisputable that these assumptions have now to
be abandoned, and with it the whole notion of Cognitive Science
and much else besides. It is this widely unrecognised problem at
the heart of much of the research in Cognitive Science (which as
has been pointed out covers Psychology, Anthropology, AI,
Linguistics and Philosophy, as well as bit of speculative
neuroscience) which accounts, I claim for the meagre progress
made in this whole area.
From the responses I have seen in this newsgroup, I think it fair
to say that these points have not been widely understood within
"Cognitive Science itself. But, then, from the responses I have
seen to date, few of those comments have actually come from
people actively involved in the field itself.
I offer what is outlined in "Fragments" as a corrective...
http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
--
David Longley
Neil Rickert
>My original criticism remains that "Cognitive Science" *was*
>conceived on the assumption that the computer and deductive
>reasoning more generally was a sound model for "the mind".
How it was originally conceived seems of little relevance. What
counts is what it has become. After all chemistry (then call
alchemy) was conceived on the assumption that base metals could be
transmuted into gold.
>I think it is now indisputable that these assumptions have now to
>be abandoned, and with it the whole notion of Cognitive Science
>and much else besides.
That goes too far. There is still a role for interdisciplinary
study.
David Longley
In article <835454...@longley.demon.co.uk>
Da...@longley.demon.co.uk "David Longley" writes:
> "In the 1970's, Cognitive Science was formulated in
> terms of Putnam's doctrine of functionalism, in which
> mental states are the functional states of an abstract
> digital computer, thinking is abstract symbol
> manipulation as in the operation of a computer program,
> and the mind's symbols get their meaning by denoting
> things in the world. Functionalism has since become the
> mainstream doctrine within the philosophy of mind. But
> in the mid-1970's, overwhelming empirical evidence
> against functionalism began to pile up, and in the late
> 1970's, Putnam himself found functionalism to be
> logically incoherent. In this volume, Putnam brilliantly
> reveals the philosophical fallacies in the doctrine he
> founded, showing why functionalism must fail as a
> philosophy of mind. The fall of functionalism has major
> consequences for generative linguistics, artificial
> intelligence, and cognitive and developmental
> psychology".
>
> Lakoff (1988) in review of H Putnam (1988)
> Representation and Reality
>
>
> My original criticism remains that "Cognitive Science" *was*
> conceived on the assumption that the computer and deductive
> reasoning more generally was a sound model for "the mind". (The
> early criticisms of the anti-psychologists had just gone unheeded
> or unrecognised by "psychologists" etc. working on cognitive
> processes such as "remembering","seeing", and "thinking" (to put
> it in propositional attitude terms).
>
> However, it *was* upon this basis that functionalism was proposed
> by Putnam in the 1960s and 70s, and it *was* with this in mind
> that the so called rationality assumption was taken as a
> justification for research into the presumed mechanics of
> attribution, ie the propositional attitudes.
>
> I think it is now indisputable that these assumptions have now to
> be abandoned, and with it the whole notion of Cognitive Science
> and much else besides. It is this widely unrecognised problem at
> the heart of much of the research in Cognitive Science (which as
> has been pointed out covers Psychology, Anthropology, AI,
> Linguistics and Philosophy, as well as bit of speculative
> neuroscience) which accounts, I claim for the meagre progress
> made in this whole area.
>
> From the responses I have seen in this newsgroup, I think it fair
> to say that these points have not been widely understood within
> "Cognitive Science itself. But, then, from the responses I have
> seen to date, few of those comments have actually come from
> people actively involved in the field itself.
>
> I offer what is outlined in "Fragments" as a corrective...
>
> http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
>
> --
> David Longley
>
Elsewhere I have challenged the very notion of "Cognitive
Science", saying that not only logical analysis (Quine 1956;1960)
made the whole enterprise highly dubious, but that empirical work
WITHIN this *new* discipline throughout the 70s and 80s actually
undermined its very raison d'etre.
The fact that so many folk still preach the groundless slogan
that "behaviourism" has been superceded by various forms of
cognitivism is just rhetoric. I urge anyone considering serious
study or research in the area give some thought to the material
outlined in "Fragments", as it may well spare them some of the
confusion which is all too pervasive in the field. Gardner's
remarks on the objectives of the positivists are a clear example
of the confusion I am alluding to.
The extract below, drawn from the end of the book which sets out
the history of the "Cognitive Revolution", presents the facts as
a "paradox" - it's not - what it amounts to is a devastating
self-refutation as I have set out in "Fragments of Behaviour: The
Extensional Stance" available at:
http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
Extract from "The Minds New Science"
H Gardner
"..... Now the cognitive age, with its high powered
computational techniques, has called into question the
view of human beings as operating in precise fashion. I
do not mean, of course, that hurnan behavior is no
longer subject to study by computational or other
cognitive scientific techniques - indeed, Johnson-Laird
has shown us just how some such behavior can be
accurately simulated; but the digital and deductive
fashion in which humans have been alleged to think is
not viable The broader question remains whether various
forms of human irrationality -those documented by
clinicians like Sigmund Freud and Carl Jung or by
anthropologists like Clifford Geertz and Dan Sperberpcan
be elucidated by the methods of cognitive science.
Even negative lessons are important, and a science
cannot be responsible for the message it yields, whether
cheerful or gloomy. Yet challenging the model of man-as-
computer raises far-reaching questions of the extent to
which cognitive science has embraced the proper view of
human mentation and provided the proper methods for its
study. I turn to these questions in the final chapter of
the book, as there I revisit the major themes of
cognitive science in light of the histories I have
related and the particular lines of current research I
have presented. In conclusion, I present my views about
the extent to which cognitive science has lived up to
its initial promise and delineate the principal
paradoxes and challenges it confronts at the present
time.
Conclusion: The Computational Paradox and the Cognitive
Challenge
Surveying the scientific landscape at the beginning of
the century, a farsighted observer might have felt
justified in announcing the arrival of the mind's new
science. After all, building on the philosophical
tradition of the Greeks and the Enlightenment, and in
the wake of dramatic breakthroughs in physics,
chemistry, and biology, the solution to the mystery of
human mental processes seemed at hand. Moreover, toward
the end of the nineteenth century, a raft of new
disciplines concerned particularly with human thought
and behavior had been launched. Surely the opportunity
to look at individuals in many cultures, in the light of
the latest findings about the human nervous system and
with the powerful tools of logic and mathematics, should
sooner or later yield a bona fide science of the mind.
From a contemporary perspective, it seems evident that
at least three Conditions had to fall into place before
this dream could reach fruition. First of all, it was
necessary to demonstrate the inadequacies of the
behaviorist approach. Second, the particular limitations
of each social science had to be acknowledged. Finally,
the advent of the computer was needed to provide the
final impetus for a new cognitive science.
ln the preceding chapters, I have shown how each of these
conditions came to be met.
By 1948, when Karl Lashley gave his famous Hixon
Symposium address on the problem of serial order in
behavior, it had become apparent to many scientists that
the behaviorist approach to human intellective activity
was fatally flawed. By the same token, the limits of
other schools in the behaviorist orbitplogical
positivism, str tural linguistics, anthropological
functionalism, Pavlovian reflexolOgy were already
becoming apparent. A fresh approach to these issues waS
sorely needed.
Paralleling the discovery of the limitations of the
behaviorist stance was a growing realization that each
of the several human and behavioral sciences, practiced
alone, harbored distinct and possibly crippling
limitations. Whether it was philosophy's ambivalence
about the relevance of empirical data to long-standing
epistemological issues, or psychology's diffficulty in
adjusting its experimental approaches to large-scale
issues, or anthropology's problems in transcending the
single case study, or neuro science's ambitions for
dealing with capacities that defy reduction to the
neural level, these various sciences increasingly felt
the need for fertilize tion with neighboring
disciplines.
Finally, and perhaps most decisively, there was the
coalescence of various mathematical and logical
demonstrations (such as those of Shannon, Turing, and
von Neumann) with important technological breakthroughs,
which culminated around mid-century in the first
computers. Once the power of these machines for dealing
with symbolic materials had been demonstrated, many
researchers became convinced that a science of cognition
might be fashioned in the image of the computer. By
1956, psychologists such as George Miller and Jerome
Bruner, computer scientists such as Allen Newell and
Herbert Simon, and linguists such as Noam Chomsky had
carried out work that (in retrospect) was cognitive-
scientific in spirit. And thirty years later, building
on these pioneering efforts, researchers such as David
Marr and Stephen Kosslyn (working at the intersection of
perceptual psychology and artificial intelligence),
Eleanor Rosch (combining psychological and
anthropological concerns), and Philip Johnson-Laird
(synthesizing approaches drawn from philosophy,
psychology, linguistics, and artificial intelligence)
had demonstrated that clear progreSs could be made in
resolving long-standing philosophical and scientific
issues. Though work on the perceptual issues is further
along than research on classification or on rationality,
it seems reasonable to declare in 1985 that cognitive
science has come of age.
It is therefore opportune, in the life of the science as
well as in the course of this survey, to take stock: to
revisit the principal themes of cognitive science in
order to clarify what has been accomplished over thepast
few decades and to discern what remains to be
accomplished if cognitive science is to achieve its full
potential. This evaluation will entail a consideration
of the central concept in cognitive sciencepthat of the
representational levelpas well as a re-examination of
two themes introduced in the opening chapters of this
book, the computational paradox and the cognitive
challenge
The Centrality of Mental Representation
To my mind, the major accomplishment of cognitive
science has been the clear demonstration of the validity
of positing a level of mental representation: a set of
constructs that can be invoked for the explanation of
cognitive phenomena, ranging from visual perception to
story comprehension. Where forty years ago, at the
height of the behaviorist era, few scientists dared to
speak of schemes, images, rules, transformations, and
other mental structures and operations, these
representational assumptions and concepts are now taken
for granted and permeate the cognitive sciences.
While most researchers (and perhaps most readers) take
the representational level for granted, this form of
analysis must be situated with reference to competing
levels of description and analysis. It has long been
acceptable in empirical science to talk of the nervous
system and, more generally, of biological systems. These
can, after all, be seen and even dissected. While most
physical scientists have been unconcerned professionally
with cultural and historical matters, it has been
acceptable (and uncontroversial) among scholars in the
humanities and social sciences to offer explanations in
terms of social forces, cultural practices, historical
traditions, and the like. How else, after all, to deal
with macroscopic social phenomena? The triumph of
cognitivism has been to place talk of representation on
essentially equal footing with these entrenched modes of
discoursepwith the neuronal level, on the one hand, and
with the sociocultural level, on the other. Whoever
wishes to banish the representational level from
scientific discourse would be compelled to explain
language, problem solving, classification, and the like
strictly in terms of neurological and cultural analysis.
The discoveries of the last thirty years make such an
alternative most unpalatable.
Making the general case for representation is one thing,
making it with precision and power quite another. Any
number of vocabularies and Conceptual frameworks have
been constructed in an effort to characterize the
representational level- scripts, schemes, symbols,
frames, images, mental models, to name just a few. And
any number of terms describe the operations carried out
upon these mental entities - transformations,
conjunctions, deletions, reversals, and so on. Cognitive
science needs to pu its conceptual house in order and to
transcend slogans and "buzz" word the field must agree
upon a language for talking about a range of
representational phenomena - even if that language turns
out to harbor various dialects.
As a start, I would single out two varieties of
representation. One form is initially or eventually
built into the hardwarepbe it computer or brain Such a
form must be invoked in order to detail what happens to
information but this variety of representation does not
involve processes of which the organism is in any way
conscious or aware. For example, during the early stages
of visual processing described by Marr and his
colleagues, the visual system must create symbolic
representations of physical information and then operate
on these representations. But no organism has any
options about these steps, and they are accessible only
to a cognitive scientist.
A second variety of representation encompasses those
problem-solv ing and classificatory behaviors that
individuals carry out with some flexibility and some
degree of explicitness and awareness. In analyzing a
sentence or a story, in creating an image or
transforming it, one may well become aware of having
created some mental representationpor mental modelpand
then one carries out operations upon that model.
Explicit awareness is not necessary here, but it is at
least a possibility. Moreover, the individual has the
option of changing the mode of representation or the
kind of rule that is invoked. This mental activity is
appropriately described in terms of representational
language but clearly warrants a separate status (or
terminology) from the kinds of representations that are
automatic and possibly wired in.
It may well be that there are several varieties of
representation, or that there exists a continuum from
implicit to explicit, or from hard-wired to flexibly
programmed. But unless a taxonomy can be agreed upon,
discussion of representation will seem ad hoc and
unsatisfactory. If representation is indeed the linchpin
of cognitive science, it must ultimately be stated as
clearly and accepted as widely as quantum theory in
physics or the genetic code of the biochemical sciences.
Such clarity and consensus seem a long way off.
The Computational Paradox
Strictly speaking, one could have had cognitive science
without the computer. After all, computational theory
antedated the invention of the computer. And yet as a
matter of historical fact, cognitive science was
unlikely to have arisen when it did, or taken the form
that it has, without the emergence of the computer in
our time. Since the first generation of cognitive
scientists, the computer has served as the most
available and the most appropriate model for thinking
about thinking. And for most, it soon became
indispensable in their daily empirical and theoretical
work. Though the linking of computation and cognitivism
turns out to have been a contingent rather than a
necessary fact, the fate of cognitive science is closely
tied to the fate of the computer.
And this leads to that strange state of affairs I have
dubbed the computationat paradox. With the vigorous
tradition, since the time of the Greeks, of thinking
about human thought as an embodiment of mathematical
principles, it is hardly surprising that the first
generation of cognitivists _reared in the logical
positivist tradition - should have embraced a highly
rationalistic view of human thought. One of the major
results of the first years of cognitive science,
however, has been a challenge of that ready assumption.
To be sure, when it comes to elementary and relatively
"impenetrable" processes like visual perception or
syntactic analysis, an authoritative computational
account may some day be given. That is, the kinds of
descriptions that are legitimately offered in the terms
of a digital von Neumann computer may turn out to be
appropriate accounts of these human cognitive processes
as well. But as one moves to more complex and belief-
tainted processes such as classification of ontological
domains or judgments concerning rival courses of action,
the computational model becomes less adequate. Human
beings apparently do not approach these tasks in a
manner that can be characterized as logical or rational
or that entail step-by-step symbolic processing. Rather,
they employ heuristics, strategies, biases, images, and
other vague and approximate approaches. The kinds of
symbol-manipulation models invoked by Newell, Simon, and
others in the first generation of cognitivists do not
seem optimal for describing such central human
capacities.
The paradox lies in the fact that these insights came
about largely through attempts to use computational
models and modeling; only through scrupulous adherence
to computational thinking could scientists discover the
ways in which humans actually differ from the serial
digital computerpthe von Neumann computer, the model
that dominated the thinking of the first generation of
cognitive scientists.
I must again underscore one point. By insisting on the
computational paradox' I do not mean to assert that it
is impossible to arrive at a computational account of
human behavioral and thought patterns in all of their
perversity, irrationality, and subjectivity. Such
accounts may well be possible andpas has been known
since the time of Turing - are certainly POssible in
principle. Rather, the paradox suggests that the
portrait of human cognition emerging from cognitive
science is far removedpat leaSt at the molar levelpfrom
the orderly, precise, step-by-step image that dominated
the thinking of the founders of the field (and of those
who dreamed about it in the more distant past). Human
thought emerges as messy, intuitive, subject to
subjective representationspnot as pure and immaculate
calculation. These processes may ultimately be modeled
by a computer, but the end result will bear little
resemblance to that view of cognition canonically
lurking in computationally inspired accounts.
Entailed in the reliance on the computer as a pivotal
model of thought is another difficulty which has only
recently begun to be recognized. Invo cation of the
computer leads naturally to a concentration on logical
prob lem solving (a la Newell and Simon) or on orderly,
highly rule-governed analysis (a la Chomsky). But
evidence from neuropsychological and devel opmental
studies of mental processes has indicated that our
concepts Of cognition need to be considerably broadened.
Processes involved in musical and other artistic
activities, and, quite possibly, processes involved in
knowing other individuals and in knowing oneself merit
the modifier cogniEve (Gardner 1983). To the extent that
this position is valid, a thoroughgoing cognitive
science will need to account for these abilities as well
as for more familiar logical mathematical applications
of mind. Whatever their relevance for the study of human
rationality or problem solving, models derived from the
computer are even less likely to be adequate to account
for these other uses of mind.
It could be countered that cognitive science ought to be
satisfied with modeling logical thought and that these
other forms of thought ought to fall by the wayside.
Perhaps cognitive science should embrace a classical
computational account, even if humans do not much
resemble a classical kind of computer. To restrict
cognitive science to one form of cognition, however, is
to refashion the subject matter to fit the current tools
of study. By the same token, to accept an account just
because such an account can be given is a scientifically
weak move. After all, the purpose of science is not to
propose a possible analysis (of which there will always
be an infinite number) but rather to come up with the
analysis that is most appropriate' parsimonious, and
convincing. All cognitive phenomena could, after all, be
described in terms of atoms, or in terms of historical
factorspand in either event, a representational account
would not even be necessary. But now, at the very time
when a representational account has been accepted, it is
important to try to find the optimal representational
account. Representation without computation is one
possible outcome for certain regionS of cognitive
science.
The idea of representation has until this point been
closely tied to our current conceptions of computers.
But there is no way of determining a priori to what
extent the ways currently embraced for describing the
presentations of computers will prove germane to
organisms, be they paramecia or professors. The kinds of
representations favored by neoassociationists like
Geoffrey Hinton turn out to be radically different (and
much sparer) than those countenanced by Jerry Fodor or
Zenon pylyshyn; moreover, it may turn out that neither
is adequate or suited for describing an individual who
is dreaming, writing a poem, or listening to music.
Earlier models of thoughtpthe reflex arc, the hydraulic
engine, the telephone switchboardpare now seen to be
extremely limited. It is already clear that one kind of
computer does not suffice to model all thought. We must
face the alternative that humans may be an amalgam of
several kinds of computers, or computer models, or may
deviate from any kind of computer yet described.
Computers will be pivotal in helping us determine how
computerlike we are, but the ultimate verdict may be
"Not very much."
Even if computers emerge as viable models for certain
facets of human thought, the question arises about the
various aspects of human nature that have been bracketed
by cognitive scientists. As I noted in chapter 3, nearly
all cognitive scientists have conspired to exclude from
consideration such nontrivial factors as the role of the
surrounding context, the affective aspects of
experience, and the effects of cultural and historical
factors on human behavior and thought (see D. Norman
1980). Some take the position that this is only a
temporary move, until the relatively discrete aspects of
cognition have been unraveled; others take the stronger
positions that cognitive science should never deal with
these aspects or even that a cognitive-scientific
account will ultimately render unnecessary any account
of these "fuzzier" factors.
Even a brief consideration of each of these "bracketed"
topics would require many pages, and since cognitive
scientists have themselves steered clear of these
issues, there is little work within the disciplinary
tradition on which I can draw. My own belief is that,
ultimately, cognitive science will have to deal with
these factors in one of two ways. Either scientists will
propose a cognitive account of affect in which, for
example, affective states will be viewed as quantitative
values along a dimension, like happinesS or cruelty; or
researchers will opt for a complex explanatory framework
in which the interaction of traditional cognitive
factors with affective or cultural factors can somehow
be modeled. These will be important but enormously
difficult undertakings, for which traditional
computational considerations may provide scant help.
H Gardner (1987)
The Minds New Science: A History of The Cognitive Revolution
--
David Longley
David Longley
> In <835454...@longley.demon.co.uk> David Longley <Da...@longley.demon.co.uk>
> writes:
>
> >My original criticism remains that "Cognitive Science" *was*
> >conceived on the assumption that the computer and deductive
> >reasoning more generally was a sound model for "the mind".
>
> How it was originally conceived seems of little relevance. What
> counts is what it has become. After all chemistry (then call
> alchemy) was conceived on the assumption that base metals could be
> transmuted into gold.
Try to understand that the ONLY reason it came into being in the first
place was because of a misconception of the scope and achivement of
behaviourism. Once the rationality assumption is shown to be untenable
*there is no reason accept "cognitive" science in its own terms. We
should look more closely at what behaviour science etc has to offer.
That, I have said, is exactky what neuroscientists *actually* do.
Having said that, work in AI may learn a few engineering tips from
those working in neuroanatomy/physiology, but they would do better
to look to engineering/logic for pointers...
>
> >I think it is now indisputable that these assumptions have now to
> >be abandoned, and with it the whole notion of Cognitive Science
> >and much else besides.
>
> That goes too far. There is still a role for interdisciplinary
> study.
Yes - bt just within SCIENCE - there is no need for "cognitive" science.
--
David Longley
Peter Lupton
In article: <4qeiqv$f...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>
> In article <4qdoil$19...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
> >In <4qcfgj$8...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
> >>
> >>But it is only confusion to think there is an epistemologically
> >>interesting natural science of rational processes. That would be to
> >>confuse causal regularities (brute tendencies) with genuine norms. You
> >>can't get there from here.
> >
> >Could you explain the reasoning a little more closely? I'll just offer a
>
> Maybe not. It was part of a broad-brush ad hominem against Longley.
> Longley pushes anti-psychologism about certain norms of rationality
> as a criticism of certain sorts of research. An example I guess would be
> Johnson-Laird's sort of study of "how people actually reason". This seems
> like a fair view to me.
>
> Yet Longley goes on to defend naturalism and "behavior science". But behavior
> science equally obliterates all normative force. This was all I meant by
> the original point.
>
> Now: what to say about "teleofunctionalism"? This is, I admit, a much more
> serious form of naturalism. This label might apply to e.g. Fred
> Dretske, Ruth Garrett Millikan, some essays of Dan Dennett, and, I gather,
> yourself.
My relatoinship to teleofunctionalism is quite close - probably I could be
said to fit into some part of its structure. Narrowly, however, I was making
the point that normative issues can't readily be separated from science and
that there is at least a prima facie case that some part of our rational norms
ought to come within the remit of science. For me the question is how far
rather than whether.
> On this view, the dichotomy I was relying on -- between causal
> regularities (tendencies) and norms -- is too crude. Evolutionary
> biology is, it would seem, a natural science in good standing (though
> doubts have often been raised on this score). Yet it is up to its
> eyeballs in norms, namely, norms of proper function. So, this sort of
> naturalism supposes that one *can* get to norms of rationality from
> purely natural norms of proper function.
That is the question, together with how direct that linkage is. Papineau
is one who seems to want a very direct link. He observes that beliefs
and desires are linked (true beliefs tend to produce satisfied desires),
and he seeks to open this circle by cashing the conditions of satisfaction
of a desire in terms of the biological function of that desire.
I think that approache is mistaken - it doesn't acknowledge sufficiently
the autonomy of reason. Dretske, Dennett and Millikan all, I think,
maintain some distance between reason and teleofunction. So do I.
What i do reckon, however, is that the norm connecting the function of
the brain and competitiveness/survival is there to be exploited.
> I should add, for fans of computationalism, that the very idea of a
> computer is itself a normative one. The formal rules of a computer
> program are in some sense idealizations which stand as norms -- not
> mere tendencies or regularities -- for an actual physical device. A
> computer can *mal*function, a planet going around its orbit cannot. So
> computational explanation also has to seek the ground of its normative
> force.
Agreed. Rule-following is one of the norms to be included. I don't want
to be in the position of being forced to accept that what a broken calculator
does is, by fiat, arithmetic.
> So if, as I suppose, it is the mark of the domain of the
> Physical that it's explanations are based on norm-free laws, then the
> Biological as well as the Computational are equally irreducible domains
> (though of course none the worse for that).
Fair enough, but science includes both. And that's what *I'm* talking
about. Nor can we draw a ready line between physics and biology. The
physicist is not averse to asking how physical systems might develop.
If such development includes life+evolution, no comfortable line is
at hand.
> >..... Suppose we thought that
> >rationality was grounded in evolutionary pressures to do with how
> >an organism should make use of brute regularities to survive. And that
> >in order to achieve this, the organism would evolve to represent such
> >regularities since (as Popper puts it) it is survival-inducing to let our
> >ideas die (or fail) in our stead. This could be given, as I understand it,
> >a naturalistic account. If so, then certain other things follow - such as,
> >in such and such circumstances and, wrt the role of reason, such and
> >such behaviour would tend to be survival-inducing and so rational.
>
> I can think of a couple of problems with this sort of approach.
> it would seem that false beliefs might be more conducive to survival then
> true ones in many circumstances. Yet we still think the true ones are the
> norm.
Certainly the relationship between the norms of reason and survival are not
direct - I have my own tale to tell which involves a lengthy excursion through
parsimony. On this scheme, truth becomes something approximated by limiting to
the most parsimonious account. Whether this is good or bad, it shows, at least,
that quite a sophisticated normative structure can be erected on the brutal
foundations of mere competition and survival.
> ....It might also be that cockroaches will yet outperform us as survival
> machines yet have no reason, no representation, no true beliefs about the
> world of any kind. So I think the norm of truth can come apart from that
> of survival.
Not sure that truth can 'come apart'. Rather, what the cockroach does
shallowly, we do far more deeply. And in doing it more deeply, we produce
a novel emergent structure - the space of reason. In a sense, we don't
need survival to produce the space of reason - parsimony and its role in
will do that. What the scientist needs competition and survival for is to
give due to the *norm*, the *ought* of reason.
> To return to anti-psychologism about logic. Is it your view
> that we accept modus ponens because evolution has discovered it is more
> useful than other rules? Isn't it rather that the explanation of its
> utility depends on its being logically valid (truth-preserving)?
I agree with *both* of the above rhetorical questions - I'll answer
'Yes' to both. I accept that the laws of logical systems are picked
out by their role in truth-preservation in various contexts. I would
add that the notion of truth itself depends, so far as I can tell, on
the fact that, as competing, surviving machines, we seek parsiony and
so correct ourselves thus giving rise, ultimately to the notion of that
which stands as correct and so true. I also believe that the notion of
a formal system itself depends crucially upon parsimony - without parsimony,
anything goes and generally will. So I don't see logical systems as
wholly separable from the considerations I'm bringing into play.
> I don't think one can think of logical norms as simply a quirk of human
> biology,
Nor do I. But then I don't think the function of the brain is a quirk.
What the brain is doing is something that cuts deep into the structure
of the world - I would say, it has a scheme that seems to be quite
general in uncovering regularities.
> ...so that there might be rational Martians who by nature
> employed a different rule.
Agreed. But still there remains the question: 'Why ought one do this
rather than that?' I think the answer is that This-Doers survive and
That-Doers don't.
> Rather failure to display cognitive
> sensitivity to certain minimal norms of logic simply means they are not
> rational or proposition-using creatures at all (though they might be
> otherwise well adapted to their niche in the merely animal way).
You say that because you don't see the rather beautiful way in which
parsimony can give rise to the space of reason while not begging it.
> I guess I am what you might call a Fregean rationalist: these norms can
> not be understood to be grounded in evolutionary biology, but are in
> some sense timeless and impersonal, like, say, the number system.
For me, the contrast you draw is a false one. These norms are, I
think timeless and impersonal and grounded in the same structures
which make biology so interesting.
> Not
> to mention the fact that any biological inquiry or explanation of
> "survival value" must itself make use of these norms.
Indeed. Any attempt to account for reason must use reason. Any such
attempt must be circular in some sense. But then science makes use of
matter to investigate matter, etc. If science got cold feet every time
it hit something circular, science would never have started.
> I have been repeatedly told that the problem of reflexivity is old hat for
> naturalists. But I don't myself see what's wrong with it -- I think norms
> of logic and rationality are more like constitutive rules which make
> possible the game of scientific inquiry, and so are not themselves capable
> of investigation by that inquiry in any non-circular way.
No statement (even the claim of being constitutive) of the role of reason can
avoid being circular - we agree in that. I'm just pretty sure that this is
neither here nor there when it comes to the question of to what extent
norms can be accounted for by science. What one ends up with is a 'fixed point'.
If our norms are thus-and-so, we account for the physical world thus-and-so,
we account for the biological world thus-and-so, we then infer there must be
norms thus-and-so. We then check we came back where we started from. There is
nothing more nor less to be expected from science.
> >As I see it, this makes it possible for a very great part of what rational
> >norms are to fall within the sphere of science. Indeed, if you ask me
> >which part of the structure falls wholly within the realm of logic and
> >philosophy, I'd be hard pressed to say what. Seems to me that the
> >boundaries are blurring.
>
> So evolutionary biology tells us why, necessarily 2 + 2 = 4? I don't
> think so.
Well, it does seem to me that in order to make sense of 2+2=4, you'll
need to know various things, such as a notion of truth, and what rule-
following is, etc.. I reckon that we'll find these norms intimately
connected to competition and survival through the role of parsimony
in constructing the normative structure we call the space of reason.
In what sense would, on that account, 2+2=4 fall *wholly* outside of
the conceptual apparatus needed by science to understand the processes
of evolution? I find it difficult to say. Although I admit and welcome
'the autonomy of the rational domain', you can have too much of a good
thing, autonomy being no exception. It seems to me that we have, here,
a typical introspection problem. Parsimony works utterly mysteriously.
We have no introspective access to the machinery of parsimony (except,
perhaps, that sometimes we get a feeling of things slotting into place).
Generally, however, the guiding hand of parsimony is so subtle that it
is with a shock when a genius such as Wittgenstein (or Goodman) reminds
us that there are other possibilities.
Can I understand 2+2=4 as something detached from this deep undercurrent?
I don't think so - I don't even know what one would do to attempt to
separate a *proposition* from its enveloping fit (which I take to be
a parsimony-induced fit). "2+2=4" is not just a squiggle, and I don't know
how to separate things out without also making what in context is a
proposition, merely syntax.
--
Cheers,
Pete Lupton
Marvin Minsky
>Well, it does seem to me that in order to make sense of 2+2=4, you'll
need to know various things, such as a notion of truth, and what rule-
following is, etc.. I reckon that we'll find these norms intimately
connected to competition and survival through the role of parsimony
in constructing the normative structure we call the space of reason.
In what sense would, on that account, 2+2=4 fall *wholly* outside of
the conceptual apparatus needed by science to understand the processes
of evolution? I find it difficult to say. Although I admit and welcome
'the autonomy of the rational domain', you can have too much of a good
thing, autonomy being no exception. It seems to me that we have, here,
a typical introspection problem. Parsimony works utterly mysteriously.
We have no introspective access to the machinery of parsimony (except,
perhaps, that sometimes we get a feeling of things slotting into place).
Generally, however, the guiding hand of parsimony is so subtle that it
is with a shock when a genius such as Wittgenstein (or Goodman) reminds
us that there are other possibilities.
>Can I understand 2+2=4 as something detached from this deep undercurrent?
I don't think so - I don't even know what one would do to attempt to
separate a *proposition* from its enveloping fit (which I take to be
a parsimony-induced fit). "2+2=4" is not just a squiggle, and I don't know
how to separate things out without also making what in context is a
proposition, merely syntax.
Well, in
ftp://ftp.ai.mit.edu/pub/minsky/AlienIntelligence
you'll find a rather different theory about why we might have come to
see arithmetic as so natural. It's a parsimony theory of a srt:
assuming that brains evolve from simple to complex, I argue that
conventional arithmetic is almost surely the first and only thing
we'll find that's even remotely like arithmetic. It would appear that
the "alternatives" to counting are too bizarre for evolutionary
survival -- that is, until the creatures are smart enough to invent
those complicated non-standard arithmetics.
David Longley
In article <662884...@luptonpj.demon.co.uk>
lup...@luptonpj.demon.co.uk "Peter Lupton" writes:
>
> My relatoinship to teleofunctionalism is quite close - probably I could be
> said to fit into some part of its structure. Narrowly, however, I was making
> the point that normative issues can't readily be separated from science and
> that there is at least a prima facie case that some part of our rational norms
> ought to come within the remit of science. For me the question is how far
> rather than whether.
It does come within the remit of science once one accepts the
implications of "Two Dogmas of Empiricism". It enters science as
logic itself, and there is no reason why the logician should not
use the computer to help him expand his science (it's more
reliable than his natural skills).
--
David Longley
Theo Vosse
David Longley (Da...@longley.demon.co.uk) wrote:
: Yes - bt just within SCIENCE - there is no need for "cognitive" science.
You are hopelessly confusing matters: what is "within science"?
Cognitive science is just as broad a term as 'natural sciences' used to
be (including physics, chemistry, biology). So how come there's no need
for cognitive science?
You earlier argument involving the computer as a basis for science is
invalid for more than one reason.
: conceived on the assumption that the computer and deductive
: reasoning more generally was a sound model for "the mind". (The
It may very well be that these assumptions got cognitive science going,
but that hardly discredits a discipline as a whole.
: I think it is now indisputable that these assumptions have now to
: be abandoned
WHY? Who indisputes it? There is *ABSOLUTELY NOTHING WRONG* with using
the computer cq. deductive reasoning or any other computational mechanism
as a basis for cognitive modelling. Can you provide one phenomenon which
fundamentally excludes any computational mechanism from being able of
simulating that phenomenon???
I suppose you strongly dispute the Church-Turing hypothesis as well?
: and with it the whole notion of Cognitive Science and much else besides.
Ah, it's that time of the year again, is it? Someone at the asylum has
forgotten to log off, and now the inmates are playing on Internet.
Ok, this argument is known as 'ad hominem', and it is just as invalid as
the argument that 'computers stink and therefore cognitive science
sucks', which seems to be a pretty good (and short) summary of your
point sofar.
: ... for the meagre progress made in this whole area.
Of course there is little progress. What else did you expect? A miracle?
Have you ever looked at the progress in physics? There isn't very much,
at least not in a broad sense. For over 40 years they are shooting atoms
through increasingly bigger super-colliders in the hope to see another
particle show up, almost by accident, but although their machines cost
several billions only to build the results are pretty meagre. And where's
the enormous progress in mathematics? Philosophy? Linguistics?
Just try to read some of the books or journals on cognitive science, and
you'll see that there still is progress. Just stay away from people who
only want to solve conciousness, or the meaning of life, the universe and
everything (which, obviously, is 42, as we all know).
--
Theo Vosse
----------
Unit for Experimental Psychology
University of Leiden
The Netherlands
Peter_...@uk.ibm.com
In <4qjvcm$h...@life.ai.mit.edu>, min...@transit.ai.mit.edu (Marvin Minsky) writes:
>In article <662884...@luptonpj.demon.co.uk> lup...@luptonpj.demon.co.uk writes:
>
>need to know various things, such as a notion of truth, and what rule-
>following is, etc.. I reckon that we'll find these norms intimately
>connected to competition and survival through the role of parsimony
>in constructing the normative structure we call the space of reason.
>In what sense would, on that account, 2+2=4 fall *wholly* outside of
>the conceptual apparatus needed by science to understand the processes
>of evolution? I find it difficult to say. Although I admit and welcome
>'the autonomy of the rational domain', you can have too much of a good
>thing, autonomy being no exception. It seems to me that we have, here,
>a typical introspection problem. Parsimony works utterly mysteriously.
>We have no introspective access to the machinery of parsimony (except,
>perhaps, that sometimes we get a feeling of things slotting into place).
>Generally, however, the guiding hand of parsimony is so subtle that it
>is with a shock when a genius such as Wittgenstein (or Goodman) reminds
>us that there are other possibilities.
>
>>Can I understand 2+2=4 as something detached from this deep undercurrent?
>I don't think so - I don't even know what one would do to attempt to
>separate a *proposition* from its enveloping fit (which I take to be
>a parsimony-induced fit). "2+2=4" is not just a squiggle, and I don't know
>how to separate things out without also making what in context is a
>proposition, merely syntax.
>
>
>
> ftp://ftp.ai.mit.edu/pub/minsky/AlienIntelligence
>
> you'll find a rather different theory about why we might have come to
>see arithmetic as so natural.
I don't think we really disagree. At least, if we do, it'll take a bit more
work to winkle out where.
I read what you mention above some time ago now - I agree with the
general gist of it, that there are simple structures which are just 'first in
line' for picking out. It can be taken further, however. I think that the whole
notion of parsimony (which in the limit of unbounded computational resources
in both space and time will reduce to Algorithmic Complexity as your article
has it) constitutes a notion of fit.
Once we add the observation that, if we are attempting to predict
a sequence of bits, we make predictive gain when the parameters
required to describe the sequence are fewer than the length of the
sequence, we find the notion of fit is bound up with the ability
of organisms to predict and so compete and survive. We get to
Algorithmic Complexity by adding the requirement of universality
and by observing that composition, alternation and iteration (feed-back/fixed
points) are needed anyway. So our parameters become the bits
of a program. We get to Algorithmic Complexity without resource
bounds by just observing that resource constraints are important
for survival in the wold, but we humans have changed the rules -
the only limit to the computation we can throw at a problem is that
provided by our technical abilities and certain properties of the
universe.
This, I think, makes the notion of fit above selected for as a *norm*
for living things. Living thing *ought* to use this notion of fit since,
unless they do, they aren't living things (they die out).
>It's a parsimony theory of a srt:
>assuming that brains evolve from simple to complex, I argue that
>conventional arithmetic is almost surely the first and only thing
>we'll find that's even remotely like arithmetic.
We agree. I'm just observing that arithmetic has to find its place -
it isn't sufficient that it be concise. It must also do work in the world.
And that's why the relationship between parsimony and prediction
is stressed by me.
>It would appear that
>the "alternatives" to counting are too bizarre for evolutionary
>survival -- that is, until the creatures are smart enough to invent
>those complicated non-standard arithmetics.
Agreed. Counting and addition play a role in our world as quite
abstract, though terse, structures. Part of what I'm saying is that
what we take counting and arithmetic and the theory of numbers to
be is the linkage of those structures considered syntactically to their
use. And their use is such that they must fit in (that notion of fit again)
with the way we identify objects, etc.. The point that I'm making here
is that, if we understand how biology works, we are lead to certain
norms which then (involving the mechanisms you describe) give rise to
arithmetic, say, as a structure taking its place along with others (such
as logic). The scientist cannot avoid taking these steps (other than as
a curious act of myopia), so far as I can see.
Cheers,
Pete Lupton
Peter_...@uk.ibm.com
In <835555...@longley.demon.co.uk>, David Longley <Da...@longley.demon.co.uk> writes:
>In article <662884...@luptonpj.demon.co.uk>
>
>>
>> My relatoinship to teleofunctionalism is quite close - probably I could be
>> said to fit into some part of its structure. Narrowly, however, I was making
>> the point that normative issues can't readily be separated from science and
>> that there is at least a prima facie case that some part of our rational norms
>> ought to come within the remit of science. For me the question is how far
>> rather than whether.
>
>implications of "Two Dogmas of Empiricism". It enters science as
>logic itself, and there is no reason why the logician should not
>use the computer to help him expand his science (it's more
>reliable than his natural skills).
I think you'll find we are discussing, not whether what falls under
our rational norms ought be used by science (which no-one is
objecting to) but whether those norms in themselves can't/can/must
fall under the remit of science and, if so, to what extent.
Cheers,
Pete Lupton
David Longley
In article <4qljoh$r...@highway.leidenuniv.nl>
vo...@ruls41.fsw.LeidenUniv.nl "Theo Vosse" writes:
> David Longley (Da...@longley.demon.co.uk) wrote:
>
> : Yes - but just within SCIENCE - there is no need for "cognitive" science.
>
> You are hopelessly confusing matters: what is "within science"?
>
I am refering to statistical, mathematical, logical, and the technology
of the neurosciences and behavioural science.
> Cognitive science is just as broad a term as 'natural sciences' used to
> be (including physics, chemistry, biology). So how come there's no need
> for cognitive science?
I've covered this at length elsewhere:
Full text of 'Fragments of Behaviour: The Extensional Stance' is available
(along with a summary text and some critical reviews by colleagues) at:
http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
>
> You earlier argument involving the computer as a basis for science is
> invalid for more than one reason.
>
> : conceived on the assumption that the computer and deductive
> : reasoning more generally was a sound model for "the mind". (The
>
> It may very well be that these assumptions got cognitive science going,
> but that hardly discredits a discipline as a whole.
>
To the extent that it is coupled (and it invariably is) with the neglect
of observation and analysis of behaviour, there *is* a serious problem.
The fact that so many working in "Cognitive Science" are oblivious to
these fundamental problems does not make them less important. What I
am saying is that when one turns to look to the applied arms of the
profession, one quickly finds that the so-called research findings from
this "new" science, just are not there....
> : I think it is now indisputable that these assumptions have now to
> : be abandoned
>
> WHY? Who indisputes it? There is *ABSOLUTELY NOTHING WRONG* with using
> the computer cq. deductive reasoning or any other computational mechanism
> as a basis for cognitive modelling. Can you provide one phenomenon which
> fundamentally excludes any computational mechanism from being able of
> simulating that phenomenon???
Then you just have not read the literature.
>
> I suppose you strongly dispute the Church-Turing hypothesis as well?
No..but the above has little if nothing to do with human "cognition".
>
> : and with it the whole notion of Cognitive Science and much else besides.
>
> Ah, it's that time of the year again, is it? Someone at the asylum has
> forgotten to log off, and now the inmates are playing on Internet.
>
I think you might be surprised to learn just how many share my criticism
of "Cognitive Science" and its empty promises.
> Ok, this argument is known as 'ad hominem', and it is just as invalid as
> the argument that 'computers stink and therefore cognitive science
> sucks', which seems to be a pretty good (and short) summary of your
> point sofar.
>
It's a total misrepresentation of what I have said "so far". Read "Fragments"
and think again...
>
> Just try to read some of the books or journals on cognitive science, and
> you'll see that there still is progress. Just stay away from people who
> only want to solve conciousness, or the meaning of life, the universe and
> everything (which, obviously, is 42, as we all know).
>
From what you have written, it's highly likely that I am far more familiar
with the literature on Cognitive Science and psychology than are you. As I
say, please read "Fragments" and think again.
--
David Longley
David Longley
> In <835555...@longley.demon.co.uk>, David Longley
> <Da...@longley.demon.co.uk> writes:
> >In article <662884...@luptonpj.demon.co.uk>
> > lup...@luptonpj.demon.co.uk "Peter Lupton" writes:
> >
> >>
> My relatoinship to teleofunctionalism is quite close - probably I could be
> said to fit into some part of its structure. Narrowly, however, I was making
> the point that normative issues can't readily be separated from science and
> that there is at least a prima facie case that some part of our rational
> norms ought to come within the remit of science. For me the question is
> how far rather than whether.
What can it mean to say that what you hold true is "close" to or
can be said to "fit into part of its ("teleofunctionalism's")
structure"... And what's more, what do you mean by *narrowly*
making a point?, or that "normative issues can't be *readily*
separated from science" ?
This is just rhetoric - in fact, it doesn't mean anything. It's
this sort of thing that's running amok as "Cognitive Science".
I'm asking whether it amounts to anything more than turgid
creative writing.
<DL> on normative "reasoning"
> >
> >It does come within the remit of science once one accepts the
> >implications of "Two Dogmas of Empiricism". It enters science as
> >logic itself, and there is no reason why the logician should not
> >use the computer to help him expand his science (it's more
> >reliable than his natural skills).
>
The above refers to the normative principles of logic and
probability.
> I think you'll find we are discussing, not whether what falls under
> our rational norms ought be used by science (which no-one is
> objecting to) but whether those norms in themselves can't/can/must
> fall under the remit of science and, if so, to what extent.
Well, what do you mean by "falling under the remit of science"
then? Do you mean that the principles of logic, set theory and
probability should be subjected to empirical test? They are
certainly the subject of research and development.
This all basically comes down to what does and does not come into
the scope of science...and what does or does not make sense when
discussing that....
"This pattern for a scientific language is evidently
rather confining. There are no names of objects.
Further, no sentences occur within sentences save in
contexts of conjunction, negation, and quantification.
Yet it suffices very generally as a medium for
scientific theory. Most or all of what is likely to be
wanted in a science can be fitted into this form, by
dint of constructions of varying ingenuity which are
familiar to logic students. To take only the most
trivial and familiar example, consider the 'if-then'
idiom; it can be managed by rendering 'if p then q' as
'not (p and not q)'.
It may be instructive to dwell on this example for a
moment. Notoriously, 'not (p and not q)' is no
translation of 'if p then q'; and it need not pretend to
be. The point is merely that in the places where, at
least in mathematics and other typical scientific work,
we would ordinarily use the 'if-then' construction, we
find we can get on perfectly well with the substitute
form 'not (p and not q) ', sometimes eked out with a
universal quantifier. We do not ask whether our reformed
idiom constitutes a genuine semantical analysis,
somehow, of the old idiom; we simply find ourselves
ceasing to depend on the old idiom in our technical
work. lIere we see, in paradigm, the contrast between
linguistic analysis and theor',r construction.
The variables 'x', 'y', etc., adjuncts to the
notation of quantification, bring about a widening of
the notion of sentence. A sentence which contains a
variable without its quantifier (e.g 'Fz' or '(y)Fxy',
lacking '(x)') is not a sentence in the ordinary true-
or-false sense; it is true for some values of its free
variables perhaps' and false for others. Called an open
sentence, it is akin rather to a predicate: instead of
having a truth value (truth or alsity) it may be said to
have an extension, this being conceived as the class of
those evaluations of its free variables for which it ~
true. For convenience one speaks also of the extension
of a Closed sentence, but what is then meant is simply
the truth value. A compound sentence which contains a
sentence as a component clause is called an extens?o71al
context of that component sentence if, whenever you
supplant the component by any sentence with the same
extension, the compound remains unchanged in point of
its own extension. In the special case where the
sentences concerned are closed sentences, then, contexts
are extensional if all substitutions of truths for true
components and falsehoods for false components leave
true contexts true and false ones false. In the case of
closed sentences, in short, extensional contexts are
what are commonly known as truth functions.
It is well known, and easily seen, that the
conspicuously limited means which we have lately allowed
ourselves for compounding sentences - viz., 'and',
'not', and quantifierspare capable of generating only
extensional contexts. It turns out, on the other hand,
that they confine us no more than that; the only ways of
embedding sentences within sentences which ever ob-
trude themselves, and resist analysis by 'and', 'not',
and quantifiers, prove to be contexts of other than
extensional kind. It will be instructive to survey them.
Clearly quotation is, by our standards, non-
extensional; we cannot freely put truths for truths and
falsehoods for falsehoods within quotation, without
affecting the truth value of a broader sentence whereof
the quotation forms a part. Quotation, however, is
always dispensable in favor of spelling.......
Heraclitus said pi-alpha-nu-tau-alpha-space-rho-epsilon-
iota, and correspondingly for the other example, thus
availing ourselves of names of letters together with a
hyphen by way of concatenation sign. Now, whereas the
quotational version showed a sentence (the Greek one)
embedded within a sentence, the version based on
spelling does not; here, therefore, the question of
extensionality no longer arises.
Under either version, we are talking about a certain
object - a linguistic formpwith help, as usual, of a
singular term which refers to that object. Quotation
produces one singular term for the purpose; spelling
another. Quotation is a kind of picture writing,
convenient in practice; but it is rather spelling that
provides the proper analysis for purposes of the logical
theory of signs.
We saw lately that singular terms are never finally
needed. The singular terms involved in spelling, in
particular, can of course finally be eliminated in favor
of a notation of the sort envisaged in recent pages, in
which there are just predicates, quantifiers, variables,
'and', and 'not'. The hyphen of concatena- tion then
gives way to a triadic predicate analogous to the '~' of
uIV, and the singular terms 'pi', 'alpha', etc., give
way to general terms which "correspond" to them in the
sense of $IV.
A more seriously non-extensional context is indirect
discourse: "Heraclitus said that all is flux." This is
not, like the case of quotation, a sentence about a
specific and namable linguistic form. Perhaps, contrary
to the line pursued in the case of quotation, we must
accept indirect discourse as involving an irreducibly
non-extensional occurrence of one sentence in an- other.
If so, then indirect discourse resists the schematism
lately put forward for scientific language.
It is the more interesting, then, to reflect that
indirect discourse is in any event at variance with the
characteristic objectivity of science. It is a
subjective idiom. Whereas quotation reports an external
event of speech or writing by an objective description
of the observable written shape or spoken sound, on the
other hand indirect discourse reports the event in terms
rather of a subjective projection of oneself into the
imagined state of mind of the speaker or writer in
question. Indirect discourse is quotation minus
objectivity and precision. To marshal the evidence for
indirect discourse is to revert to quotation.
It is significant that the latitude of paraphrase
allowable in indirect discourse has never been fixed;
and it is more significant that the need of fixing it is
so rarely felt. To fix it would be a scientific move,
and a scientifically unmotivated one in that indirect
discourse tends away from the very objectivity which
science seeks.
Indirect discourse, in the standard form 'says
that', is the head of a family which includes also
'believes that', 'doubts that', 'is Surprised that',
'wishes that', 'strives that', and the like. The
subiectivity noted in the case of 'says that' is shared
by these other idioms twice over; for what these
describe in terms of a subjective projection of oneself
is not even the protagonist's speech behavior, but his
subjective state in turn.
Further cases of non-extensional idiom, outside the
immediate family enumerated above, are 'because' and the
closely related phenomenon of the contrary-to-fact
conditional. Now it is an ironical but familiar fact
that though the business of science is describable in
unscientific language as the discovery of causes, the
notion of cause itself has no firm place in science. The
disappearance of causal terminology from the jargon of
one branch of science and another has seemed to mark the
progress in understanding of the branches concerned.
Apart from actual quotation, therefore, which we
have seen how to deal with, the various familiar non-
extensional idioms tend away from what best typifies the
scientific spirit. Not that they should or could be
generally avoided in everyday discourse, or even in
science broadly so-called; but their use dwindles in
proportion as the statements of science are made more
explicit and objective. We begin to see how it is that
the language form schematized in uIV might well, despite
its narrow limitations, suffice for science at its
purest.
Insofar as we adhere to that idealized schematism,
we think of a science as comprising those truths which
are expressible in terms of 'and', 'not', quantifiers,
variables, and certain predicates appropriate to the
science in question. In this enumeration of materials we
may seem to have an approximation to a possible standard
of what counts as "purely cognitive." But the standard,
for all its seeming strictness, is still far too
flexible. To specify a science, within the described
mold, we still have to say what the predicates are to
be, and what the domain of objects is to be over which
the variables of quantification range. Not all ways of
settling these details will be congenial to scientific
ideals."
W. V. O . Quine
The Scope and Language of Science
pp239-242 "The Ways of Paradox and Other Essays" 1966;76
--
David Longley
David Longley
In article <4qeiqv$f...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> Yet Longley goes on to defend naturalism and "behavior science". But behavior
> science equally obliterates all normative force. This was all I meant by
> the original point.
>
It's worth mentioning (again) that the position I am advocating has
its roots not so much in the philosophy of Quine (I think his work
is just the closest I have come across to my own views and think credit
should always be given where there is precedent) - it is really based
on empirical work in the psychology of perception and attribution, coupled
with some very good work in neuroanatomy and the evolution of the CNS.
In psychology, I have cited Bruner's work (realy just as an example), but
there is all of the Gestalt psychlogist's work on similarity, closure, the
phi phenomenon etc, and Michotte's work on attribution which Heider based
Attribution Theory on - not to mention all of the work on conditining (just
think of Conditioned Taste Aversion learning, superstition, and so on). From
my knowledge of psychology I find it very difficult to understand how anyone
who has read "Fragements" and wo also knows the history of psychology does
not recognise the force of what I have said and can see immediately that the
examples I cite are just a few of the more famous reseach studies..
Psychlogy *is* methodologically solipsistic in the main.
The extensional stance has its roots in my work in neuroscience in the early
1980s. That work has its roots in the work of Herrick and Sherrington, ie
the phylogeny of brain development - which should make ione think of notions
like metemeric segmentation at the most primative level, and re-representation
of function as Jackson outlined this at the end of the last century. Some time
ago I gave a brief outline of the homology between the ventral striatum and
neo-striatum, and how the different monoamine transmitter systems patched
into those systems - the pinjt being the differing roles that dopamine might
be playing in the facilitation of telo-receptor based sensori-motor
coordination (in the neo-striatal system) and cruder, more diffuse systems
we called emotion, in the paleo-striatal system.
The key notion here is the degree to which the CNS works to *integrate*
what otherwise quite fragemented and more local behaviour sequences. It
is from this prspective that the force of the "Extensional Stance" must
be assessed...One has to appreciate that the integrated stance is not one
which is characteristic of animal behaviour itself.
Once this is understood clearly, I think much of what I say in "Fragments"
becomes much more comprehensible - to the point of *seeming* obvious..
For this to be appreciated, once has to appreciated how fundamental a
"fragmented" stance is, and how indispensible this is for biological
systems in the interests of fault tolerance and adaptability (plasticity).
--
David Longley
Anders N Weinstein
In article <835652...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <4qeiqv$f...@usenet.srv.cis.pitt.edu>
> ande...@pitt.edu "Anders N Weinstein" writes:
>
>> science equally obliterates all normative force. This was all I meant by
>> the original point.
>
>its roots not so much in the philosophy of Quine (I think his work
>is just the closest I have come across to my own views and think credit
>should always be given where there is precedent) - it is really based
>on empirical work in the psychology of perception and attribution, coupled
>with some very good work in neuroanatomy and the evolution of the CNS.
This seems rather besides the point to me. I was trying to suggest
it is inconsistent to go on and on about the normativity of logic, and
then push naturalism or behavior science or neuroscience. All these
disciplines aim at laws how events succeed one another in their respective
domains. All are "merely descriptive", not prescriptive. They say
nothing about how events *ought rationally* to succeed one another.
Consider the sum total of neuroscientific or behaviorist science, logically,
as one big deductively closed theory. Now: no normative conclusions follow
from this database. Rather, norms of logic are *presupposed* in order to
derive any conclusions at all.
So how can you possibly think you derive a rational "ought" from all of these
non-rational "is's"? I am happy with Frege's anti-psychologism about logic.
I don't find any support for "naturalism" in it, much less for the study of
brain chemistry. That's only explanatorily relevant in *failures* of
rationality, as an interfering factor in the otherwise free activity of
reason.
Admittedly this way of putting it is a bit provocative, but I thought
you endorsed Frege's anti-psychologism.
>Psychlogy *is* methodologically solipsistic in the main.
I would say that in the main it is neutral because in the main
psychological research ignores foundational questions. For good or ill,
only philosophers are specially concerned with them.
For example, consider the result that chess experts don't think about a
huge number of positions. This is, I take it, an objective and
non-obvious psychological fact. But to recognize it does *not* depend
on being a "methodological solipsist."
For example, I myself take Wittgenstein and others to have shown that
human conceptual thought is only possible in the context of a framework
cultural norms, in particular linguistic norms, which is prior to
individual thinkers and thinkings and makes them possible. This is part
of the metaphysical nature of conceptual thinking, in my view.
This view entails that a brain considered all by itself -- a brain in a
vat, say -- simply could not have the thoughts that a chess-player
does, because it could not have internalized the rules of chess from a
larger cultural milieu. Without this internal relation to a
super-individual practice, there is no basis for saying it is thinking
about chess as opposed to something else. Indeed I am doubtful that was
can find any sense in the ascription of any psychological predicates at
all to a mere brain -- brains don't think, people do, and people are
not brains but beings in the world.
Still, that said, of course it's true and interesting that expert
chess players don't think about large numbers of positions. This result
is independent of the culturalist metaphysics of conceptual thought.
Most experimental psychology is similarly neutral about metaphysics.
>Once this is understood clearly, I think much of what I say in "Fragments"
>becomes much more comprehensible - to the point of *seeming* obvious..
Perhaps you should beware when controversial theses seem obvious.
On the other hand, I have to admit it seems obvious to *me*, after soaking
up Kant, Frege, Husserl, Heidegger, Sartre, Ryle, Wittgenstein, Anscombe,
Searle, Gibson, Sellars, Davidson, McDowell ... that philosophy of mind has
little to learn about the nature of human understanding from the brain or
behavioral or so-called cognitive sciences. These inquiries are pitched
at least one level down from the level that should concern us in philosophy
of mind, the level of the person.
David Longley
In article <4qnbt6$a...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> In article <835652...@longley.demon.co.uk>,
> David Longley <Da...@longley.demon.co.uk> wrote:
> >In article <4qeiqv$f...@usenet.srv.cis.pitt.edu>
> > ande...@pitt.edu "Anders N Weinstein" writes:
> >
> >> Yet Longley goes on to defend naturalism and "behavior science". But behavior> >> science equally obliterates all normative force. This was all I meant by
> >> the original point.
> >
> >It's worth mentioning (again) that the position I am advocating has
> >its roots not so much in the philosophy of Quine (I think his work
> >is just the closest I have come across to my own views and think credit
> >should always be given where there is precedent) - it is really based
> >on empirical work in the psychology of perception and attribution, coupled
> >with some very good work in neuroanatomy and the evolution of the CNS.
>
> This seems rather besides the point to me. I was trying to suggest
> it is inconsistent to go on and on about the normativity of logic, and
> then push naturalism or behavior science or neuroscience. All these
> disciplines aim at laws how events succeed one another in their respective
> domains. All are "merely descriptive", not prescriptive. They say
> nothing about how events *ought rationally* to succeed one another.
What you consider "besides the point" is just a reflection of you basically
missing the point.
We apply physical laws of science normatively all the time, ie when building
bridges, piloting space shuttles, using a computer. These laws were arrived
at naturalisticaly through empirical observation and testing.
Some of the empirical laws we discover, are, as Quine says, at the core of
our web of holistic belief. Such principles are almost treated as regulative
principles (Kant made the mistake of thinking of them as transcendental). To
appreciate the focce of what I am saying one has to be thorough in rejecting
the analytic-synthetic distinction (Quine 1951).
Cue.. if you are going to understand something new in its own terms, you have
to abandon trying to understand it in terms of familiar (old) categories <g>..
--
David Longley
Theo Vosse
In article <835611...@longley.demon.co.uk>, Da...@longley.demon.co.uk wrote:
> http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
Would be cute, except that it's impossible to retrieve the document... The
ftp-server is unreachable.
> To the extent that it is coupled (and it invariably is) with the neglect
> of observation and analysis of behaviour, there *is* a serious problem.
Ok, I messed up my previous posting by pressing the wrong button at the
wrong time, but I'll try again.
Any serious cognitive scientist will look at human behaviour. However, in
order to do any serious research, behaviour must be measurable. So, in my
field (human syntactic processing), there is quite a lot of data regarding
several behavioural aspects of the processes involved in reading and
understanding. The next step is to make any sense of this data and come up
with a model that not only follows the data, but also tries to explain or
model the underlying processes. Such models, theories or explanations can
serve as a tool for predictions and further investigations. That's (part
of) cognitive science. And there's nothing wrong with it.
> > Can you provide one phenomenon which fundamentally excludes any
> > computational mechanism from being able of simulating that phenomenon???
>
> Then you just have not read the literature.
Then you must be able to provide a reference (preferably in a major
journal) to an article giving such an example. E.g., I cannot take the
Dretske (1980) quote serious. His example that is supposed to support that
a description of a cognitive state is non-extensional is dubious (S does
know 'my' uncle is getting married, he only doesn't know Tom is 'my'
uncle; it's bloody word play).
But that's not the main point. The main point is that the description of
a cognitive state is in itself not of any interest, at least not to every
single researcher who is concerned with cognitive science.
Furthermore, cognitive science in general does *NOT* accept that there is
something like 'language of thought' at face value. There is
circumstantial evidence in favor, there is circumstantial evidence
against; there is nothing hard. Whatever one's view on this point, nobody
will accept that there is a "universal mentalese", at least not if I
understand the word correctly. Saying that it must be possible to
represent something using propositions is quite another matter.
Another possible example: you wonder how difficult it is to teach
deductive reasoning skills effectively. You seem to think that if "the
classical, functionalist, stance of cognitive psychology is in fact true",
it automatically should follow that learning deductive reasoning skills
should be easy. Do you?
> > I suppose you strongly dispute the Church-Turing hypothesis as well?
> No..but the above has little if nothing to do with human "cognition".
It most certainly does. If you accept the Church-Turing hypothesis, then
any physical process, including human cognition, must be computable.
Hence, algorithms, deductive reasoning, first order logic, etc. are all
proper means of modelling human behaviour (in the same way as they are
proper means of modelling physical phenomena).
And now for the point at which I stopped reading: you say that
methodological solipsism accepts "that thinking is a purely formal,
syntactic, computational affair." It seems to me that you apply this
notion you attribute to Fodor to the whole of cognitive science. That is
not necessarily true, of course.
On the other hand, I do accept this notion. If you accept that our brain
is a computational device, you must accept that reasoning is ultimately
'purely formal', 'computational' and 'syntactic'.
But from this it does not follow that 'desires, beliefs, etc.' are atomic
particles of our cognition. This is only a representation. The
representation may be wrong, but that does not invalidate the use of
representations per se.
So, what is your problem with cognitive science?
- that it relies on computational mechanisms as a metaphor?
- that it takes a solipsist stance?
- that it doesn't provide the answers to your problems?
- that propositional logic is not a very good representation for most
cognitive problems?
- that it is reductionist?
> From what you have written, it's highly likely that I am far more familiar
> with the literature on Cognitive Science and psychology than are you.
So you are a librarian? <non-extensional pun intended>
David Longley
In article <vosse-25069...@news.LeidenUniv.nl>
vo...@ruls41.fsw.LeidenUniv.nl "Theo Vosse" writes:
> In article <835611...@longley.demon.co.uk>, Da...@longley.demon.co.uk wrote:
>
> > http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
>
> Would be cute, except that it's impossible to retrieve the document... The
> ftp-server is unreachable.
>
I've tried it from here - however, I've sent you a self-extracting archive
by e-mail.
> > To the extent that it is coupled (and it invariably is) with the neglect
> > of observation and analysis of behaviour, there *is* a serious problem.
>
> Ok, I messed up my previous posting by pressing the wrong button at the
> wrong time, but I'll try again.
>
> Any serious cognitive scientist will look at human behaviour. However, in
> order to do any serious research, behaviour must be measurable. So, in my
> field (human syntactic processing), there is quite a lot of data regarding
> several behavioural aspects of the processes involved in reading and
> understanding. The next step is to make any sense of this data and come up
> with a model that not only follows the data, but also tries to explain or
> model the underlying processes. Such models, theories or explanations can
> serve as a tool for predictions and further investigations. That's (part
> of) cognitive science. And there's nothing wrong with it.
>
Apart from the focus on language/syntax (which is behavioural and effective
is anything is), what you describe is scientific method per se. The term
"Cognitive" has nothing to do with it.
> > > Can you provide one phenomenon which fundamentally excludes any
> > > computational mechanism from being able of simulating that phenomenon???
> >
> > Then you just have not read the literature.
>
> Then you must be able to provide a reference (preferably in a major
> journal) to an article giving such an example. E.g., I cannot take the
> Dretske (1980) quote serious. His example that is supposed to support that
> a description of a cognitive state is non-extensional is dubious (S does
> know 'my' uncle is getting married, he only doesn't know Tom is 'my'
> uncle; it's bloody word play
If you are going to seriously study language, you have to give up on notions
like "word play". See Quine 1956 "Quantifiers and Propositional Attitudes"
(The Ways of Paradox and Other Essays, 1966;1972).
> But that's not the main point. The main point is that the description of
> a cognitive state is in itself not of any interest, at least not to every
> single researcher who is concerned with cognitive science.
>
When it comes to studying language, my point is that it has nothing to do
with "cognition" anyway. If you think you can study abstract propositions
then I think you are in trouble,
> Furthermore, cognitive science in general does *NOT* accept that there is
> something like 'language of thought' at face value. There is
> circumstantial evidence in favor, there is circumstantial evidence
> against; there is nothing hard. Whatever one's view on this point, nobody
> will accept that there is a "universal mentalese", at least not if I
> understand the word correctly. Saying that it must be possible to
> represent something using propositions is quite another matter.
>
That's part of the problem with "Cognitive Science" is it not - it is all
things to all folk.
> Another possible example: you wonder how difficult it is to teach
> deductive reasoning skills effectively. You seem to think that if "the
> classical, functionalist, stance of cognitive psychology is in fact true",
> it automatically should follow that learning deductive reasoning skills
> should be easy. Do you?
It should be easier yes. The fact is that research evidence casts doubt
on this very basic assumption. See the review by Nisbett et al. in Science
as referenced.
>
> > > I suppose you strongly dispute the Church-Turing hypothesis as well?
> > No..but the above has little if nothing to do with human "cognition".
>
> It most certainly does. If you accept the Church-Turing hypothesis, then
> any physical process, including human cognition, must be computable.
> Hence, algorithms, deductive reasoning, first order logic, etc. are all
> proper means of modelling human behaviour (in the same way as they are
> proper means of modelling physical phenomena).
>
You mis my point. The empirical evidence speaks strongly against the
rationality assumption. I have not quarrel with the notion that people
*can* behave reasonably, just as they can learn to ride bicycles, speak
languages, use computers etc. My point is that the natural assessments
which people use are not extensional (Kahneman & Tversky 1982)
> And now for the point at which I stopped reading: you say that
> methodological solipsism accepts "that thinking is a purely formal,
> syntactic, computational affair." It seems to me that you apply this
> notion you attribute to Fodor to the whole of cognitive science. That is
> not necessarily true, of course.
At that point I am discussing a particular paper by Fodor (1980). At that
time Cognitive Science was still being formalised as an entity. I think
Fodor did quite a good job at characterising what the stance of most
psychologists who would otherwise class themselves as "Cognitivists"
actually subscribe to. It also desribes the views which the majority of
graduates will tacitly hold....(albeit without critical unerstanding).
>
> On the other hand, I do accept this notion. If you accept that our brain
> is a computational device, you must accept that reasoning is ultimately
> 'purely formal', 'computational' and 'syntactic'.
>
I do, but that has nothing to do with psychology qua "cognition".
> But from this it does not follow that 'desires, beliefs, etc.' are atomic
> particles of our cognition. This is only a representation. The
> representation may be wrong, but that does not invalidate the use of
> representations per se.
But cognitive scientists do study remembering (memory), seeing (perception),
and attitudes/beliefs.....there is a problem there, in that these are
attributions when not studied as "effective" processes. Where they are
studied as effective processs, they are *not* psychological/cognitive
processes. What I have endeavoured to do is highlight a very pervasive
confusion..
>
> So, what is your problem with cognitive science?
> - that it relies on computational mechanisms as a metaphor?
> - that it takes a solipsist stance?
> - that it doesn't provide the answers to your problems?
> - that propositional logic is not a very good representation for most
> cognitive problems?
> - that it is reductionist?
>
Some of the above, but mainly that it's a muddle and certainly no
substitute or alternative or superordinate discipline to behaviour
science...(See Skinner 1987).
--
David Longley
Anders N Weinstein
In article <835680...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <4qnbt6$a...@usenet.srv.cis.pitt.edu>
> ande...@pitt.edu "Anders N Weinstein" writes:
>
>> then push naturalism or behavior science or neuroscience. All these
>> disciplines aim at laws how events succeed one another in their respective
>> domains. All are "merely descriptive", not prescriptive. They say
>> nothing about how events *ought rationally* to succeed one another.
>
>bridges, piloting space shuttles, using a computer. These laws were arrived
>at naturalisticaly through empirical observation and testing.
These are not normative in the sense at issue. It is *our* functional norm for
a bridge that it doesn't break. It is *our* functional norm for a circuit that
it doesn't short out. And so on. But a bridge that breaks or
a circuit that shorts out are still acting in complete accordance with
physical laws. It is
In the world as conceived by physicalistic science, events simply happen in
accordance with laws. All a physicalist can say is that they happen as
they do happen. There is no right or wrong about it. There can be
no such thing as a *mal*function or a mistake at a physicalistically
conceived level of description.
For example, if there are laws of behavior science, they simply dictate
how people in fact behave, not how they ought rationally to behave. If
people do not often reason correctly, then that must be the result of
behavior science, just as much as in cognitive science. To paraphrase
Frege: one action potential considered as such no more contradicts
another than one ripple in a pond contradicts another. One disposition
to emit a noise no more contradicts another then Jupiter in its orbit
contradicts Saturn.
So if cognitive science is by your lights "pseudo" science, for the same
reasons, I guess behavior science and neuro-science are also
pseudo-sciences which simply characterize how we in fact reason.
I would say instead that these are all perfectly *genuine* sciences of
their respective domains. Maybe not behavior science, since there may
not be any useful notion of "behavior" for it to study, but I'll
concede the possibility for argument's sake. But beacuse these
disciplines do not conceptualize their domains in norm-involving terms,
they are impotent to address the questions that concern us in
epistemology or philosophy of mind.
With respect to Minsky's post, I would say the interesting philsophical
issue is concerned with competence not performance. We
all are talking as if we ourselves have a handle on what constitutes
normatively correct reasoning -- say, logically valid, or consistent,
or number-theoretically sound reasoning. Now how shall we understand the
"psychological state" that constitutes this grasp or competence? Can
*that* state be characterized by a computer program or set of purely formal
rules. Must it instead be understood as a grasp of *semantic*
principlesx -- a spooky apprehension of Plato's forms, if you want
to mystify it? I gather Godel himself was hospitable to this conclusion.
I don't think it should be viewed as mystical, but I also doubt it
can be modelled as a set of purely typographical rules. But even more
interesting to me then the question of formalism is the philosophical idea of
"competence" itself. I don't think states of competence can be understood
as internal representations in one's brain either, mainly because of the
normative force involved in the notion of a state of competence.
In order to apply the concept of mathematical or rational competence to
another human being, we have to be able to see their overt conduct not
merely as *conforming* to principles of correct reasoning, but as
positively *striving* after them. Nothing in the physicalistically
conceived domains can do this. I am serious when I say that such
things as the expression on a mathematician's face as he or she
crumples up a piece of paper and tosses it in the wastebasket are
crucial to our sense that the mathematician is so striving.
Of course, when this happens, I don't think there's any violation of
physical law going on in the room (for what it's worth, which is not
very much). Still, it is expressions like these, not themselves proofs
or arguments, but the manner in which attempts at proof or argument are
conducted, which make manifest the fact that the mathematician has *him
or herself* undertaken a responsibility to conform to the norms. For
this reason, I think the competence concept is part of a set of
concepts which one can apply to living human beings, and perhaps one
day to robots -- if they are capable of suitably expressive behavior -- but
*not* to automatic calculators (or brains) considered as such.
Cathy Highet and Erick Gallun
> >
> > > http://www.uni-hamburg.de/~kriminol/TS/tskr.htm
> >
I finally managed to download this document and now feel confident in
commenting. Your dismissal of cognitive science in this newsgroup
doesn't really seem to match with your paper. I think that you are
attacking a strawman. Perhaps I am not in the majority, but your paper
seems to be well within the domain of current cognitive science. UC
Berkeley, where I am a grad student, was one of the first sites to get
funding for a "Cognitive Science" program, and we are far from unified
on what it is. The monolithic structure you attack has not shown up in
any of the classes or discussions I have participated in with the
cognitive psychologists at Berkeley. We certainly discuss these ideas,
but no one is saying that "mind is to brain as software is to hardware"
is a solid position. It is just a working hypothesis, and our
conception of "software" is certainly undergoing revision.
Kahneman and Tversky showed quite convincingly that the software is not
what would be designed by a careful engineer, but rather is the result
of "kludgie" evolution. Heuristics are quick and dirty solutions to
very long and complex problems. Nonetheless, this doesn't mean that the
mind is not best thought of as programs run on the brain. It also
doesn't mean that we should abandon cognitive science for behaviourism.
What would behaviourist methodology give us in understanding this?
Skinner didn't even care about mental processes. They were unneccesary
for describing the environment-organism system. This isn't really what
you think we should go back to, is it?
> >
>
> That's part of the problem with "Cognitive Science" is it not - it is all
> things to all folk.
I'm not sure what the problem is with this. I'm content to call myself a
psychologist even though my friends in the clinical program share
virtually none of my assumptions. Cognitive scientist is certainly a
more precise description than that.
> But cognitive scientists do study remembering (memory), seeing (perception),
> and attitudes/beliefs.....there is a problem there, in that these are
> attributions when not studied as "effective" processes. Where they are
> studied as effective processs, they are *not* psychological/cognitive
> processes. What I have endeavoured to do is highlight a very pervasive
> confusion..
Okay, I'm not sure what you're getting at here. In my study of auditory
perception, I'm constantly trying to pull apart what is peripheral and
what is central (i.e., cognitive). Is this the distinction you mean?
Somehow I don't think so, because "effective" doesn't really mean
peripheral, but rather seems to mean something like cognitively
impenetrable. I guess I am subject to this "very pervasive confusion".
Enlighten me.
> >
> > So, what is your problem with cognitive science?
> > - that it relies on computational mechanisms as a metaphor?
> > - that it takes a solipsist stance?
> > - that it doesn't provide the answers to your problems?
> > - that propositional logic is not a very good representation for most
> > cognitive problems?
> > - that it is reductionist?
> >
> Some of the above, but mainly that it's a muddle and certainly no
> substitute or alternative or superordinate discipline to behaviour
> science...(See Skinner 1987).
> --
> David Longley
Again, I'm not sure what you are attacking here. If it is the lack of
"observation and analysis of behaviour", then I wonder what you call the
experiments being done by cognitive psychologists. They are no less
rigorous than concentrating solely on the behaviour elicited by the
environment. The difference is in the sort of theories that can be
presented and tested.
Certainly many of the physiologically-minded auditory researchers still
rely on solely behaviouristic conceptions, but we have argued
convincingly that the human mind is best described by taking into
account the cognitive processes that result in the heterogeneity of
responses emitted to the same stimuli. If I can be so crude, imagine
your responses to your favorite piece of music. They are not always the
same, and they transcend the physical responses to the sound waves
fairly quickly.
Erick Gallun
Auditory Perception Laboratory
UC Berkeley
For more rigorous examples from research on sound localization, visual
capture, attention, and binaural hearing, see
David Longley
In article <4qp7bh$g...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
<DL>
> >We apply physical laws of science normatively all the time, ie when building
> >bridges, piloting space shuttles, using a computer. These laws were arrived
> >at naturalisticaly through empirical observation and testing.
>
> These are not normative in the sense at issue. It is *our* functional norm for
> a bridge that it doesn't break. It is *our* functional norm for a circuit that
> it doesn't short out. And so on. But a bridge that breaks or
> a circuit that shorts out are still acting in complete accordance with
> physical laws.
This is not what I said - we apply the laws of science in engineering, and
we do so 'ceteris paribus'. We apply those laws with degrees of tolerance
given the forces etc we expect to operate. That our technology can fail
is more a function of our failure to apply those scientific rules in practice
than anything else. That was my point of quoting Feynmann some time back over
Challenger and the absurd way in which "risk" of component failure was arrived
at - ie "intuitively" believe it or not!.
>
> In the world as conceived by physicalistic science, events simply happen in
> accordance with laws. All a physicalist can say is that they happen as
> they do happen. There is no right or wrong about it. There can be
> no such thing as a *mal*function or a mistake at a physicalistically
> conceived level of description.
>
But that in paractice is invariably NOT how the engineer always conceives
his world, economic and other factors intrude. For the sake of argument,
we can class those along with intensional heuristics.
<snip>
> So if cognitive science is by your lights "pseudo" science, for the same
> reasons, I guess behavior science and neuro-science are also
> pseudo-sciences which simply characterize how we in fact reason.
>
The reason I say that "cognitive science" is pseudo science is because of
what it sets out to quantify into/over. Now it may well be that there are
folk working under the umbrella of cognitive science who are in fact working
on behaviour, and just unselfconsciously accepting the rubric of "cognitive"
- with them I have no argument. It's with those who explictly argue that they
are working on cognition and NOT behaviour (even indirectly) that I am taking
exception to.
> I would say instead that these are all perfectly *genuine* sciences of
> their respective domains. Maybe not behavior science, since there may
> not be any useful notion of "behavior" for it to study, but I'll
> concede the possibility for argument's sake. But beacuse these
> disciplines do not conceptualize their domains in norm-involving terms,
> they are impotent to address the questions that concern us in
> epistemology or philosophy of mind.
>
Well, I obviously disagree - and I have said why - I think the positive way
forward is to recognise that we are collections of behaviours, and that whilst
some micro elements may be made sense of along normative lines, we naturally
flout overall consistency, because "we" just are more of a creative fiction
as integrated entities than we are generally able to recognize (unless studied
from an extensional stance *as* behaviours).
--
David Longley
Anders N Weinstein
In article <835732...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <4qp7bh$g...@usenet.srv.cis.pitt.edu>
> ande...@pitt.edu "Anders N Weinstein" writes:
>
>> These are not normative in the sense at issue. It is *our* functional norm for
>> a bridge that it doesn't break. It is *our* functional norm for a circuit that
>> it doesn't short out. And so on. But a bridge that breaks or
>> a circuit that shorts out are still acting in complete accordance with
>> physical laws.
>
>we do so 'ceteris paribus'. We apply those laws with degrees of tolerance
>given the forces etc we expect to operate. That our technology can fail
>is more a function of our failure to apply those scientific rules in practice
>than anything else.
That may well be, but it does not address the question: whence
the normative force in attributions of errors, malfunctions, etc?
>Well, I obviously disagree - and I have said why - I think the positive way
>forward is to recognise that we are collections of behaviours, and that whilst
>some micro elements may be made sense of along normative lines, we naturally
>flout overall consistency, because "we" just are more of a creative fiction
>as integrated entities than we are generally able to recognize (unless studied
>from an extensional stance *as* behaviours).
But you keep saying that deductive logic is concerned with how we *ought*
to reason. This means that you think you and I, fragmented collections of
behaviors that we are, are nonetheless capable of grasping this standard as
a norm and discussing it as such in this forum. My question is, how is
*that* possible in your world-view?
What "fragmented collection of behaviors" constitutes our apprehending
the norms of deductive logic, so that we can compare actual behavior
against this standard and find it wanting? Shouldn't philosophy of mind
be concerned with *that* capacity -- the person qua noumenal entity,
one might say, who can unify or hold together (synthesize) a
multiplicity of conceptual contents in a single complex propositional
judgment, and a multiplicity of propositional judgments in a single
inference? Evidently *that* capacity of ours is *not* a "creative
fiction", at least if deductive logic really does stand as a graspable
norm to which our reasoning might aspire, as you so often suppose.
David Longley
In article <31D03E...@sirius.com>
er...@sirius.com "Cathy Highet and Erick Gallun" writes:
"Fragments" was not written as an attack on "Cognitive Science", but as a
critique of cognitivism and clinical judgment in Applied Criminological
Psychology (and psychology more widely). However, I do think some of the
themes outlined in "Fragments", particularly those drawing on Quine's
critcism of propositions and other intensional idioms should give a number
of "Cognitive Scientists" some therapeutic headaches!
>
> Certainly many of the physiologically-minded auditory researchers still
> rely on solely behaviouristic conceptions, but we have argued
> convincingly that the human mind is best described by taking into
> account the cognitive processes that result in the heterogeneity of
> responses emitted to the same stimuli. If I can be so crude, imagine
> your responses to your favorite piece of music. They are not always the
> same, and they transcend the physical responses to the sound waves
> fairly quickly.
I'm no expert in auditory perception, but did spend some time having the
psychological stuffing knocked out of me when I was at NIMR in the early
1980s. One of these sparing partners was an auditory neurophysiologist who
also played a part in the work on the Matching Law in the 60s...So he knew
more operant work than I did, and could use it for behavioural assaying.
As Skinner made clear in a number of places, just because the analysis
of behaviour gets difficult does not mean we have to cognitivise it!
>
> For more rigorous examples from research on sound localization, visual
> capture, attention, and binaural hearing, see
>
> http://ear.berkeley.edu/auditory_lab/
>
I'll have a look, but even before looking, I doubt whether much of it is
"cognitive".
--
David Longley
Aaron Sloman
I agree with much of what is written by Anders (ande...@pitt.edu
(Anders N Weinstein)) but not all
> Date: 25 Jun 1996 00:29:26 GMT
> Organization: University of Pittsburgh
> ...[much deleted]....
> For example, I myself take Wittgenstein and others to have shown that
> human conceptual thought is only possible in the context of a framework
> cultural norms, in particular linguistic norms, which is prior to
> individual thinkers and thinkings and makes them possible. This is part
> of the metaphysical nature of conceptual thinking, in my view.
>
> This view entails that a brain considered all by itself -- a brain in a
> vat,
[Note: I assume you mean a brain grown in a vat, rather than one
rescued from a human body and then stored in the vat, somehow kept
alive and working]
> ....say -- simply could not have the thoughts that a chess-player
> does, because it could not have internalized the rules of chess from a
> larger cultural milieu.
[If it was originally the brain of a normal person it might have.]
> Without this internal relation to a
> super-individual practice, there is no basis for saying it is thinking
> about chess as opposed to something else. Indeed I am doubtful that was
> can find any sense in the ascription of any psychological predicates at
> all to a mere brain -- brains don't think, people do, and people are
> not brains but beings in the world.
I find it intriguing that there's a subset of philosophers, like
Anders, who believe that all mentalistic descriptions of an agent
(i.e. descriptions of what the agent thinks, believes, hopes, knows,
etc.) somehow depend on implicit assumptions about the cultural or
evolutionary history of the agent (or both), while others (like me,
and Roger Young (The Mentality of Robots, Proceedings Aristotelian
Soc. 1994)) think that the descriptions refer to what the agent can
do now and in the future.
Thus from the first standpoint a brain grown in a vat, or a complex
software process that is somehow assembled in a machine, could never
use some sort of language, could not formulate mathematical or
philosophical problems, could not invent a game structurally
equivalent to chess or Go and play against itself, could not learn
to count events or structures occurring within it, could not be
disappointed on discovering that a proof it had constructed
concerning properties of numbers had a flaw, and so on.
These descriptions would be ruled out, a priori, independently of
what sorts of structures, processes, interactions between processes,
etc. occurred inside the system, no matter how closely in detail
they mirrored (at an appropriate level of abstraction) the processes
in a human chess player or mathematician.
From the second standpoint (i.e. Young's and mine), whether
mentalistic descriptions were appropriate would depend not on the
history of the system but on what the processes were, what sorts of
causal interactions occurred, what new dispositional states were
created by earlier events, etc. I.e. it would depend on the
present capabilities of the system and how it used them.
Of course, if the system's internal structure manipulations were
totally causally isolated from the environment then there's no way
it could be thinking about any individual object in the environment,
such as the Eiffel Tower, and perhaps not about physical and other
properties of things in the environment, such as water, electrical
resistance, etc. (Only a methodological solipsist would say the
causal links are not necessary for such reference to succeeed.)
However I deliberately chose examples of contents of mental states
that do NOT depend on such specific external causal links.
When confronted with disagreements like this, after much fruitless
debate with people who hold the first view, I have come to the
conclusion that this is one of those many cases in philosophy where
there is, in the end, no disagreement of substance, only a semantic
disagreement: one lot implicitly or explicitly *defines* mental
state descriptions to include reference to a causal or cultural
history whereas others do not, but instead refer to the present and
future causal powers of the system.
So Anders has, for reasons that he can perhaps explain, adopted one
set of implicit definitions whereas I adopt another.
My reasons are that in principle if I come across a sophisticated
system with the capabilities of the brain in the vat or a software
system such as I described above, then I want a language for
describing the internal states and processes, and since the language
for describing human mental states and processes already goes a long
way to meeting the need, as long as you drop any historical
presuppositions, then it is far simpler and more elegant to use the
existing language than to invent a new language that duplicates all
that richness.
(However, I don't think ordinary language is the *last* word in
either case. As we learn more about the variety of information
processing architectures and what the capabilities and limitations
of different sorts are, we'll need to refine and extend our
mentalistic vocabulary, just as discoveries about the architecture
of matter led to refinement and extension of ordinary language
descriptions of physical kinds of stuff, events, processes etc.
Perhaps the most spectacular example is the derivation of the
periodic table of the elements from a theory of the architecture of
matter, followed by the derivation of structural descriptions of
actual and possible chemical compounds.)
I have no objection to Anders and others adhering to their
restricted usage of mental concepts, as long as they realise that
there is no reason at all why others like me should follow their
semantic preferences, and some good reasons why we shouldn't.
Cheers.
Aaron
===
--
Aaron Sloman, ( http://www.cs.bham.ac.uk/~axs )
School of Computer Science, The University of Birmingham, B15 2TT, England
EMAIL A.Sl...@cs.bham.ac.uk
Phone: +44-121-414-4775 (Sec 3711) Fax: +44-121-414-4281
David Longley
In article <4qpmhb$j...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> In article <835732...@longley.demon.co.uk>,
> David Longley <Da...@longley.demon.co.uk> wrote:
> >In article <4qp7bh$g...@usenet.srv.cis.pitt.edu>
> > ande...@pitt.edu "Anders N Weinstein" writes:
> >
> But you keep saying that deductive logic is concerned with how we *ought*
> to reason. This means that you think you and I, fragmented collections of
> behaviors that we are, are nonetheless capable of grasping this standard as
> a norm and discussing it as such in this forum. My question is, how is
> *that* possible in your world-view?
>
Well, the same way that we learn everything else. In "Fragments"
I briefly review the literature on how context and content
specific such learning is (which is why we are so bad at it - cf
Wason 1966; Nisbett et al 1990).
> What "fragmented collection of behaviors" constitutes our apprehending
> the norms of deductive logic, so that we can compare actual behavior
> against this standard and find it wanting? Shouldn't philosophy of mind
> be concerned with *that* capacity -- the person qua noumenal entity,
> one might say, who can unify or hold together (synthesize) a
> multiplicity of conceptual contents in a single complex propositional
> judgment, and a multiplicity of propositional judgments in a single
> inference? Evidently *that* capacity of ours is *not* a "creative
> fiction", at least if deductive logic really does stand as a graspable
> norm to which our reasoning might aspire, as you so often suppose.
>
Well, this itself is an empirical matter - but there are pointers
from the sort of research done by Johnson-Laird and colleagues.
I guess the merits of that work is that it brings out the point
that we do have to learn logic as a set of skills and that they
can be idiosyncratic models. There are probably lots of models of
logic learning logic and the learning of normative rules in
general (cf. INDUCTION by Holland et al) just as there are lots
of ways of teaching people to drive, speak a language etc.
I think, again, just on the basis of the evidence that has
accumulated, it is this content specificity, ie poor transfer of
training of these skills which is the interesting point of view
from the stance I am advocating. Taken out of the
training/application context, with all its familiar examples
(e.g. the Frege examples of Morning Star/Evening Start etc.) it
does not travel well...or far.......We are only dealing with a
fragment of our behaviour here, our reading and writing skills -
and then, only aspects/fragments of those.
--
David Longley
H. M. Hubey
>Not if the psychological ("mind") states are a quite *different* function
>of the past than the brain states; for example, a function of quite
>different aspects of the past. For example, if the psychological states
Well, if you are being some kind of a dualist in separating mind
states from brain states, I have nothing to say really.
You either believe or you don't. If there is no scientific
proof/evidence ,then we are left with extrapolation from what
we know, and you can extrapolate your way, and I'll do mine
my way.
>by looking at your bodily movements alone. E.g. "what I am doing"
>when my hand goes up is raising the bid (at the auction). Anyone can
>see this, but it's doubtful it could be determined by studying neural
>structures alone, without reference to the institutional world.
Sorry but you could, if things like "intentions" are also brain
states as well as mind states. There is no other place where
the reasons for you lifting your hand might reside except in
your brain. The neurons that gave the signal for your hand to
go up got their activation signals from other neurons.
>Myself, I would advise against the hopeful but confused locution
>"mind/brain". The brain is a thing. But on the Aristotelian/Rylean
>view I would recommend there really is no such *thing* or part inside
I guess, everything is a "thing" including the "mind".
--
Regards, Mark
http://www.csam.montclair.edu/~hubey hu...@pegasus.montclair.edu
David Longley
> or riding a bicycle, then the objection is moot, since there may simply
> be no inner intellectual process, hence no structure to abstract and
> formalize. At the personal level you find a person who understands the
> social world in the practical sense that he or she can operate within in;
> at the sub-personal level you might only find circuitry that couples with
> the world without representing it symbolically. At least I would like to
> see more work on how far one could go with such designs at the
> sub-personal level.
>
This is certainly the conception which I think has most empirical credibility.
The point is though that whilst this is most likely to be the way biological
systems adapt, the price paid are the biases documented by folk such as
Tversky & Kahneman. We need the compensatory infrastructure of a learned
(if easily forgotten at times/places) formal logic and all tat it supports
(acuarial analysis)..
--
David Longley
Theo Vosse
> is anything is), what you describe is scientific method per se. The term
> "Cognitive" has nothing to do with it.
That is not a proper argument. I am trying to show that cognitive science
simply behaves the way it should: it analyzes behaviour and tries to model
it. You cannot dismiss this argument on the basis that is has nothing to
do with the term 'cognitive'.
> If you are going to seriously study language, you have to give up on notions
> like "word play". See Quine 1956 "Quantifiers and Propositional Attitudes"
I have not read that, but I have read other formal arguments on language,
and they all have the same problem: logic, or rather: logicians. Logicians
try to map natural language sentences to formulae as short as possible. If
you ever read anything about Montague logic/Lambek calculus, you must have
seen to what lengths they go to describe sentences like "John thought he
saw a unicorn". The problem with the unicorn being that it doesn't exist.
Or does it?
So, cognitive science takes a different approach (or at least, part of
cogsci does). And currently, we are even having trouble making sense out
of word recognition processes and low-level semantics, and that's what I
would call 'seriously studying language' for the moment.
> When it comes to studying language, my point is that it has nothing to do
> with "cognition" anyway.
I beg your pardon? Language has nothing to do with cognition? What on
earth makes you think that? Besides, my arguments also apply to memory
research, and visual and auditive perception.
> If you think you can study abstract propositions then I think you are in
> trouble,
Which abstract propositions do I think I can study? Language?
> That's part of the problem with "Cognitive Science" is it not - it is all
> things to all folk.
You mean: everybody uses it in a private meaning? Well, that's nothing to
be surprised about. But if you mean that, why do you say *IT* is a
pseudo-science?
> > You seem to think that if "the classical, functionalist, stance of cognitive
> > psychology is in fact true", it automatically should follow that learning
> > deductive reasoning skills should be easy. Do you?
> It should be easier yes.
No, it should not be. If you had read Goedel, Escher, Bach, or studied
formal logic or theoretical computer science, you would have been familiar
with either Goedel's or Turing's theorems on the incompleteness of logic
(this is a simplification). What follows from these theorems, is that
learning logic is indecidable. So, it actually is a miracle that some
people do manage to learn it, rather than the other way around.
Let's reverse the example: we know for sure that computers *are* built on
elementary logic circuits. But have you every tried to write, e.g. a
Prolog interpreter? Then you would appreciate the difficulty of
transferring your knowledge of logic to a machine that was built to
execute it.
> You mis my point. The empirical evidence speaks strongly against the
> rationality assumption.
But it says nowhere that systems based on logic should behave rational.
There is no limit on their behaviour (with the well-known exceptions, cf.
Turing etc.). Anybody can make a computer program that gets all the
answers to elementary addition excersises wrong.
> My point is that the natural assessments
> which people use are not extensional (Kahneman & Tversky 1982)
I am afraid I cannot grasp your English completely. With assessments you
mean: judgement, and being extensional means: existing in some physical
sense, being perceptible? So you say that natural human judgements don't
exist physically?
> At that point I am discussing a particular paper by Fodor (1980). At that
> time Cognitive Science was still being formalised as an entity.
I don't think CogSci was ever formalised as an entity.
> It also desribes the views which the majority of
> graduates will tacitly hold....(albeit without critical unerstanding).
That's
1) without evidence;
2) an insult to the majority of the graduates, whoever they are.
Did you know that there are actually quite a lot psychologists who believe
in either astrology, re-incarnation, ghosts, UFOs, or the healing
influence of positive thinking? So do you think Louise Hay defines
psychology as a formal entity? Should we abondon psychology as a whole
because of it?
> I do, but that has nothing to do with psychology qua "cognition".
No, but it counters your argument that using computers and deductive
reasoning do not form a proper basis for psychological/cognitive
modelling.
> What I have endeavoured to do is highlight a very pervasive confusion..
What is that confusion? That some people take e.g. memory for granted,
while others study it as a process?
> > So, what is your problem with cognitive science?
> ... mainly that it's a muddle and certainly no
> substitute or alternative or superordinate discipline to behaviour
> science...(See Skinner 1987).
And nobody claims it is. But physics is also not a subsitute to
behavioural sciences; does that make it a pseudo-science?
Now for the more serious point: cognitive science has a lot more to offer
than most other disciplines tradionally seen as a part of psychology. In
your texts, you mention the use of statistics. But we all know that
statistics can only describe: it takes a model to make it do something, or
to understand what it means. And that model has to be made by someone.
Now, if you randomly generate a model, what's the likelihood that you get
workable knowledge out of it? Pretty low, I'd guess.
The main part of cogsci is about designing, building and verifying models
of tasks related to human cognition. That in order to do so, cognitive
scientists sometimes take resort to introspection and ad-hoc modelling, is
to be expected.
It is the same with statistics: statisticians just measure things. If I
measure the length of my table, I know if I can get it through the door or
not. If I measure 1000 different tables, I end up with an average and a
variance that are useless (except for estimating which proportion of all
tables would fit through my door). There are even people who would claim
that statistical results don't mean anything, because they are obtained
under laboratory conditions, or they change over time, or whatever. Does
this mean statistics is a pseudo-science?
If you want answers from cogsci, you can get them, as long as you realize
what they mean. There is data on processes, and there are models of
behaviour, but they are just proposals for mechanisms describing small
fragments. They have to be tested, and reworked, and thought over again,
but these things take time. So it's pretty unfair to claim that a 40 year
old discipline should have come already so far as to put you out of a
job...
David Longley
ande...@pitt.edu "Anders N Weinstein" writes:
> I also take Frege to show us how to see through reductive
> naturalism -- if you can find support for behaviorism in Frege, I will
> eat my computer.
>
With Frege's Begrifsschrifft we have the first illustration of an
artificial "programming" language. The Church-Turing Thesis was
another formative development in the history of 'effectivity.
"Behaviourism" explicates procedures which elsewhere, are all
too often presented in cognitive terms, mentalistic/intensional
terms.
Note. I am not speculating about what Frege may or may not have
"thought" - but what he achieved. I take Quine (1960) to have
made a nonsense of Frege's original notion of Sinn.
--
David Longley
Anders N Weinstein
H. M. Hubey <hu...@pegasus.montclair.edu> wrote:
>Can a person know-that if he doesn't know-how?
The point of the Rylean (pragmatist, call it what you will) outlook
is that knowing-how is more fundamental than knowing-that. So, no,
I wouldn't think that someone can know-that something if he or she
didn't have any know-how.
Mainly, I think knowing-that requires a rather special sort of
knowing-how, namely, knowing how to operate with (public) symbols,
put forward theses, support them with reasons, etc. This is more than
just parroting them.
>football players, etc etc. Under these circumstances can anyone
>then claim that they have made real choices as to deciding
>whose paintings, or music is "superior" (in some way) or are we
>all copycats who ape the fashion of the times?
We must start such activities -- including knowledge seeking -- from a
pre-existing tradition (cp. Neurath's boat). So at first, we do in a
sense simply ape it. Later we can try to change it on our own, but it
is perhaps impossible to be completely original, rebuild one's
world-view from scratch, cut free of the influence of tradition in the
manner of Descartes' foundational project.
Anders N Weinstein
>
>>Not if the psychological ("mind") states are a quite *different* function
>>of the past than the brain states; for example, a function of quite
>>different aspects of the past. For example, if the psychological states
>
>Well, if you are being some kind of a dualist in separating mind
>states from brain states, I have nothing to say really.
Pity. Sure Cartesian substance dualism is spooky and unscientific. But
there's nothing wrong with property or attribute dualism. If I said that the
economic properties of a dollar bill could not be read off from its
intrinsic structure, would you accuse me of some pernicious sort of
dualism? Hardly. But that's all I mean: mental states have a nature
more like that.
Materialism or mind-brain identity theory seems to me basically not worth
fighting for. It's just a misfiring attempt to respond to the problem.
Even functionalists or computationalist are property dualists, since
computational properties are not reducible to physical properties. I
agree with this, only I think person-level intentional properties are
not reducible to inner computational ones, in pretty much the same
fashion.
>You either believe or you don't. If there is no scientific
>proof/evidence ,then we are left with extrapolation from what
>we know, and you can extrapolate your way, and I'll do mine
>my way.
Don't get it. Would you say the same about economic properties?
I wouldn't say I'm extrapolating, rather reflecting on our concepts.
>Sorry but you could, if things like "intentions" are also brain
>states as well as mind states.
There is no reason to believe this.
> There is no other place where
>the reasons for you lifting your hand might reside except in
>your brain. The neurons that gave the signal for your hand to
>go up got their activation signals from other neurons.
The last sentence is true but irrelevant. Reasons do not
have to reside anywhere -- think about "knowledge compilation"
if you doubt this.
H. M. Hubey
>there's nothing wrong with property or attribute dualism. If I said that the
>economic properties of a dollar bill could not be read off from its
>intrinsic structure, would you accuse me of some pernicious sort of
>dualism? Hardly. But that's all I mean: mental states have a nature
>more like that.
I can't extrapolate from here. I can see how the neuron right in
the middle of someone's retina might not "mean" the same thing
as someone else's retina or that the Nth neuron connected to X's
inputs from the right ear don't mean the same as Y's ear but I can't
see how the analogy works in this case. I see neurons firing and the
results are electro-chemical, but because every brain is different
the result of firing of corresponding neurons in brains of different
brains "mean" different things. In that sense, we want a piece of
paper with a certain composition and coloring to "mean" that
it has value of "one dollar" is analogous.
But your idea is something like saying that memory (engram) works
like holography. There is too much for me to fill in. It is in the
same vein as the historic ones of making brains/minds like
clockworks, telephone exchanges, digital computers etc.
I think by now we should be up to ANN's.
>Even functionalists or computationalist are property dualists, since
>computational properties are not reducible to physical properties. I
>agree with this, only I think person-level intentional properties are
>not reducible to inner computational ones, in pretty much the same
>fashion.
I can't make this out either; "reducible" in what sense. In the
sense of derivable from the other? In that case, it might be
possible. Look at the derivation of "structure" of attractor from
the time series of a single variable in the study of chaos.
I believe that the "structure" is derivable from the "process" in
the same sense, and hence "reducible" (eventually). Certainly
it is structure that is responsible for creating the process.
Anders N Weinstein
>
>I can't extrapolate from here. I can see how the neuron right in
>the middle of someone's retina might not "mean" the same thing
>as someone else's retina or that the Nth neuron connected to X's
>inputs from the right ear don't mean the same as Y's ear but I can't
>see how the analogy works in this case. I see neurons firing and the
>results are electro-chemical, but because every brain is different
>the result of firing of corresponding neurons in brains of different
>brains "mean" different things.
Well that would be one version of the point, which probably even
holds at the level of neuron firings. I'm more interested in what people
can say or do then in what happens among their neurons.
If Mary asserts to me, say, that Bill Clinton is now in the White
House, I don't have to look inside her brain to assess the claim, agree
or disagree with it, or ascribe the corresponding belief to her.
Looking inside her brain would add nothing in the way of understanding
what was said and would really be kind of silly.
Even if I have to check whether we understand one another, I would
still have to look at how she uses words in public situtations to
establish this, and, if things go on in her brain while she speaks that
is still pretty irrelevant to this sort of inquiry.
I would say the ability to do such things as make assertions and
justify them is what we mean by "the Mind", or a large portion of it.
It's the meaning of public utterances that we should take to be
primary. The meaning of "inner" intentional states might be explained
derivatively, in terms of their potentiality for being expressed in a
public language.
>But your idea is something like saying that memory (engram) works
>like holography. There is too much for me to fill in. It is in the
>same vein as the historic ones of making brains/minds like
>clockworks, telephone exchanges, digital computers etc.
I am not interested in modelling the brain. I have expressed some --
highly speculative -- sympathy for Rodney Brooks style architectures,
mainly because they seem to make very obvious that knowing-how need not
be based on inner theories or representations. But the points don't
hinge on how that project turns out. The question about the nature of
intentional states is a higher level inquiry.
>I can't make this out either; "reducible" in what sense. In the
>sense of derivable from the other? In that case, it might be
>possible. Look at the derivation of "structure" of attractor from
>the time series of a single variable in the study of chaos.
>I believe that the "structure" is derivable from the "process" in
>the same sense, and hence "reducible" (eventually). Certainly
>it is structure that is responsible for creating the process.
This is besides the point. The question is how Mary's assertions and
even her "inner" thoughts get their connection to the distal *Clinton*,
many miles away in the White House (or not, as the case may be). This
sort of connection simply *cannot* be purely a matter of "structure" or
"process" or "dynamics" inside Mary's brain, it seems to me. You need
to see her activity in its relation to a larger context than that if you
want to make it intelligible how Mary's claim answers for its correctness
to how things are in the White House.
H. M. Hubey
>In article <hubey.8...@pegasus.montclair.edu>,
>H. M. Hubey <hu...@pegasus.montclair.edu> wrote:
>>ande...@pitt.edu (Anders N Weinstein) writes:
>>
>>I can't extrapolate from here. I can see how the neuron right in
>>the middle of someone's retina might not "mean" the same thing
>>as someone else's retina or that the Nth neuron connected to X's
>>inputs from the right ear don't mean the same as Y's ear but I can't
>>see how the analogy works in this case. I see neurons firing and the
>>results are electro-chemical, but because every brain is different
>>the result of firing of corresponding neurons in brains of different
>>brains "mean" different things.
>Well that would be one version of the point, which probably even
>holds at the level of neuron firings. I'm more interested in what people
>can say or do then in what happens among their neurons.
Regardless of where your interest lie, if the mind is the working
of the neurons, then all of your ideas have to somehow
take the neurons into consideration. Continuum mechanics is able
to give wonderful descriptions of elasticity, and fluid dynamics
although we all know about the discrete/quantized nature of
reality.
>If Mary asserts to me, say, that Bill Clinton is now in the White
>House, I don't have to look inside her brain to assess the claim, agree
>or disagree with it, or ascribe the corresponding belief to her.
>Looking inside her brain would add nothing in the way of understanding
>what was said and would really be kind of silly.
Putting together 100 billion artificial neurons might not be
out of the capability of humanity in a century or so. Of course,
we can dream on, and imagine that we can somehow find a way to
noninvasively and nondestructively examine and keep track of all
the neuron firings in Mary's brain when she makes her assertions.
Would you still say that looking into that would be useless in
adding to our understanding?
>Even if I have to check whether we understand one another, I would
>still have to look at how she uses words in public situtations to
>establish this, and, if things go on in her brain while she speaks that
>is still pretty irrelevant to this sort of inquiry.
I don't have any problems taking into account all facts, including
behavior, or even positing unobservables if the resulting explanation
is more successful than others. I don't have any problems with the
existence of emotions, attitudes, or whatnot. But in the final
analysis, I don't see how or why we can forget about the basis
of all this activity which is neuronal firings.
>I would say the ability to do such things as make assertions and
>justify them is what we mean by "the Mind", or a large portion of it.
>It's the meaning of public utterances that we should take to be
>primary. The meaning of "inner" intentional states might be explained
>derivatively, in terms of their potentiality for being expressed in a
>public language.
Fine. Let's use them to good effect.
>I am not interested in modelling the brain. I have expressed some --
>highly speculative -- sympathy for Rodney Brooks style architectures,
>mainly because they seem to make very obvious that knowing-how need not
>be based on inner theories or representations. But the points don't
>hinge on how that project turns out. The question about the nature of
>intentional states is a higher level inquiry.
But know-that seems to be simple memorization and know-how is
more along the lines of what we think should be taught in schools,
what constitutes a large part of science, and what probably
counts for "understanding".
>This is besides the point. The question is how Mary's assertions and
>even her "inner" thoughts get their connection to the distal *Clinton*,
>many miles away in the White House (or not, as the case may be). This
>sort of connection simply *cannot* be purely a matter of "structure" or
>"process" or "dynamics" inside Mary's brain, it seems to me. You need
How else can it be? I can conjure up the inside of my apartment
in great detail even when I am not there. Somehow my 3D version of
my aparment is stored in my neurons and I can activate it at will.
I can, if I want, imagine that a thief might have broken into it
and even imagine what he might be doing. I don't understand exactly
what you are referring to if not to the way in which all of this
is coded into the neurons.
>to see her activity in its relation to a larger context than that if you
>want to make it intelligible how Mary's claim answers for its correctness
>to how things are in the White House.
Aaron Sloman
lot of points that I thought were irrelevant to my attack on his claim
that a brain in a vat could not be described using intentional language.
ande...@pitt.edu (Anders N Weinstein) writes:
> Date: 26 Jun 1996 18:36:18 GMT
> Organization: University of Pittsburgh
> .....
I wrote
[AS]
> >I find it intriguing that there's a subset of philosophers, like
> >Anders, who believe that all mentalistic descriptions of an agent
> >(i.e. descriptions of what the agent thinks, believes, hopes, knows,
> >etc.) somehow depend on implicit assumptions about the cultural or
> >evolutionary history of the agent (or both), while others (like me,
> >and Roger Young (The Mentality of Robots, Proceedings Aristotelian
> >Soc. 1994)) think that the descriptions refer to what the agent can
> >do now and in the future.
Anders replied
[AW]
> "What an agent can do now" is of course ambiguous
There's no ambiguity. I am talking about what capabilities the mechanism
has, but I was not specifying them. That's no more an ambiguity than
making an existentially quantified statement without specifying which
object makes it true.
However, I gave some examples.
[AW]
> ....What I can do now might include:
> make a promise, write a check, invite someone out to dinner, pay a debt,
> vote, register for classes, signal for the check, raise the bid at an
> auction, force a mate in three, etc. These are all, in some sense, socially
> constituted actions which are only possible in a certain
> context.
So what?
It's a fallacy to argue
Some processes require cultural or biological causes
therefore All do.
I don't dispute the premise. To do so would be silly. It's the
conclusion I was arguing against.
You originally seemed to be saying that mentalistic language (about
beliefs, desires, aims, emotions, attention etc.) is applicable ONLY
to systems whose internal machinations are causally linked with a
culture.
My counter claim was that the use of such language without the cultural
presuppositions is perfectly OK provided that the system NOW has an
appropriate collection of capabilities.
I then went on to suggest examples of capabilities that involved
internal manipulations of various kinds of structures, and although I
gave a few illustrations, these were not meant to be exhaustive.
I certainly was not talking about making promises, registering for
classes, etc. all of which obviously require more than internal
machinations.
[AW]
> They thus bear, I would say, an internal relation to cultural
> history.
Your examples do. Nothing follows from that about other mental states.
[AW]
> I freely grant that in saying this I am taking for granted without any
> argument the untenability of a methodologically individualist program.
> That program would construct social norms out of individual mental
> representations, assumed to be what they are independently of the context.
I wasn't talking about social norms let alone trying to construct them
out of anything individual. Obviously social norms require a social
context. Nobody would dispute such a trivially true statement.
> ....
[AW]
> I favor the holistic alternative, that human concept-involving
> intentional states and social institutions emerge together, as elements in
> a package deal.
You may favour it. Moreover, empirically that may be what happened.
However, you seemed to be making a stronger claim, namely that
intentional states COULD NOT exist without the social institutions.
I claim that that particular necessity can only arise out of a semantic
preference: your refusal to decouple the language of intentional states
from the social context.
Others, like me, are happy to decouple them, requiring only present
capabilities not past influences to justify the use of mentalist
descriptions.
So we have two sets of concepts
believesC, thinksC, wantsC, intendsC (prsupposing Cultural links)
believesW, thinksW, wantsW, intendsW (without cultural links)
Apart from the extra cultural prerequisites in the first set, there's no
difference. I explained how the second set could be useful in a context
where the first set would be inapplicable.
[AW]
> This mainly because mastery of the norms of the social
> world need only be implicit in skills, not explicit in theories or
> representations. So mastery of these norms may be more like riding a bicycle
> or catching a fly ball, a matter of knowing-how rather than knowing-that.
None of this is relevant to the point.
(a) I wasn't talking about social norms. By definition, they are not
implicated in believesW, thinksW, wantsW, etc. So they are
irrelevant (except for someone who falls into the
"some...therefore...all..."
trap, which is common on this news group, but I don't think would
tempt you.
(b) I wasn't talking about a contrast between what is implicit and
explicit, only about a contrast between what is and what is not
culturally (or biologically) produced. What is implicit e.g.
implicit in a neural net, or implicit in a complex collection of
(internal)condition-(internal)action rules) need not have any
cultural connection.
The "*W" concepts may include all sorts of skills, compiled
procedures, neural implementations, etc. without having explicit
theories or representations. Or they may include explicit theories
or representations. The two distinctions *C/*W and implicit/explicit are
orthogonal.
Either way, the *W states don't have to have cultural antecedents in
order to exist NOW implemented in a brain or brain-like machine, in
which they produce processes of the sort I was talking about (e.g. a
system passionately seeking a more elegant proof of some theorem, or
attempting to devise a game that permits a wide range of competences
to be used depending on which concepts and patterns the player has
learnt, or playing such a game with itself).
[AW]
> I think it is well understood that while such skills require some sub-personal
> circuitry to implement, they needn't require a symbolic representation
> of the task or quasi-inferential processes.
I have no idea why you brought in this distinction, since I did not
say anything about it, unless it is because you believe this:
[AW]
> ....My bet is that all merely
> tacit theories, including linguistic metatheory, are artifacts of this
> public explicative activity, and are not internally represented in our
> nervous systems.
you can bet that if you like. I have no reason to believe it.
You can make it true by definition, if you like. I'll just use a
different language to talk about things that are indescribable in
your terms.
If a copy of your brain gets created by chance processes in some
brain factory, then the copy, like the original, will have many
capabilities that do not use explicit representations, and the copy
will have no cultural causes. I.e. what's implicit need not be cultural.
(Of course, some implicit things are.)
[AS]
> >Thus from the first standpoint a brain grown in a vat, or a complex
> >software process that is somehow assembled in a machine, could never
> >use some sort of language, could not formulate mathematical or
> >philosophical problems, could not invent a game structurally
> >equivalent to chess or Go and play against itself, could not learn
[AW]
> Bit of a fudge. I didn't say it couldn't 'invent' a game structurally
> equivalent to present-day chess. Of course it could. It could do lots
> of things that are "structurally equivalent" to thinking, reasoning,
> etc.
But you apparently don't want to use the words "thinking", "reasoning",
"wanting", and the like, where there are no links to a culture.
My point is that that is a purely semantic preference, with nothing
of any substance supporting it or following from it, just like deciding
that you will not call a circle an ellipse because you (unlike most
mathematicians) require an ellipse to have two non-coincident foci or
major and minor axes of different lengths.
[AW]
> That's precisely what automatic calculators are for -- to go through
> steps that are structurally equivalent to the steps a real person goes
> through when the person reasons.
Not really. Most calculators go far beyond what any real person can do
with numbers. E.g. mine can calculate the square root of 3.5276 almost
instantaneously. I can't, and when I try to do it I am unlikely to go
through the same steps as my calculator. E.g. I make mistakes, correct
them, etc.
So let's restrict ourselves to machines that happen to replicate types
of structure and process that do occur in human brains. (Since I don't
know what these are, I can't be specific. For example they may, for all
I know, require non-computable quantum gravity effects in microtubules,
as Penrose and Hammeroff claim, though I have no reason to believe the
claim. But whether the claim is true or not is not relevant. All that's
relevant is that replicas of brains might exist that do not owe their
existence or the details of their states, to a pre-existing culture.)
[AS]
> But formal structure is not all. There is also semantics --
> original, non-derivative intentionality -- to worry about, and I don't see
> that you can't get that from structure alone.
Worry about it if you will. I don't.
I (and several others) have argued previously that existence of semantic
capabilities is not an all-or-nothing affair but rather a very complex
collection of capabilities, different subsets of which may be present in
different animals and machines (See 'What enables a machine to
understand?' in IJCAI95).
Maybe SOME of those capabilities involve cultural links.
But requiring cultural links as preconditions for ALL intentional
descriptions is just a semantic decision: i.e. a refusal to use a
certain set of concepts (the *W concepts).
[AW]
> Of course we can study purely formal structures all we want,
> and even embody some of them in highly useful mechanical calculators.
Er, .... why are you talking about formal structures and mechanical
calculators? I said nothing about either.
Mechanical devices are not likely to have the flexibility required for
the capabilities I was referring to: that's why we use electronic
devices for building computers nowadays. If you mean "mechanical" to
include "electronic" then perhaps you should choose a different word.
If by "formal" you mean using some particular form of syntax such as a
logician might use, then I cannot see why you bring that in to a
discussion of whether cultural antecendents are necessary. There might
be both formal and non-formal classes of systems, within each of which
there are some with and some without cultural antecendents.
It's another orthogonal, irrelevant distinction.
Remember you started this by referring to a brain in a vat, not a
formal system in a vat.
[AW]
> I would agree it is very hard to understand what it would be for an
> automatic calculator as such to be disappointed.
And why are you talking about automatic calculators? I wasn't.
I was assuming systems with far richer architectures, i.e. with the sort
of richness of mechanisms that you might find in a human brain.
(You don't seem to think the brain is an automatic calculator. We
need to compare like with like.)
[AW]
> This depends on
> whether you take affective states to be "program-resistant" -- simply
> outside the domain of computational models -- or not.
I was talking about mechanisms whose states and structures are not
the products of cultures. Now how you get from there to
COMPUTATIONAL models I don't know.
The relevant class of mechanisms may or may not all be computational,
whatever that means. I am not sure it is a well defined concept. (See
below).
If the physical world includes some non-computational mechanisms then
maybe the brain has a need for them, in which case my disconnected
mathematician will need them, and will not be purely computational
(whatever that means).
Even so it need not owe anything to a culture.
I have an open mind on whether non-computational mechanisms are needed
if that means mechanisms that cannot be precisely modelled on a
computer.
All I am claiming is that cultural precursors are NOT needed.
Or rather, all I am claiming is that I would be happy to apply
intentional descriptions to a system with no cultural precursors,
provided that the system's internal states and processes have sufficient
richness (of the right kind).
I happen also to believe, or at least conjecture, that some aspects of
human emotionality (partial loss of control of thought processes, such
as I described during my visit to Pittsburgh) may be replicated in
architectures that are implementable on computers. But that's not
relevant to whether cultural causes are necessary.
[AW]
> ...My
> inclination is to favor the former, though I gather some take
> the heroic line that even affect can be understood in purely computational
> terms (e.g. as re-prioritizing of goals).
I don't know what you mean by "purely computational". The only
notion of "computation" that I find precisely definable has nothing
to do with kinds of mechanisms or causes, but is restricted to
classes of abstract structures about which one can prove various
kinds of theorems (e.g. about complexity, decidability, etc.) In
that sense a Godel number can be a computation.
The mechanisms I am talking about are complex mechanisms with causal
powers and causal interactions" they change their own and each
others' states.
[AW]
> Ok, but note there's a real intellectualist bias involved in focusing
> on digital and highly symbolic processes like playing chess or proving
> theorems.
If you are no longer saying that ALL mental processes require a cultural
origin, since the "intellectualist" ones don't, then that's fine. All I
was claiming was that the cultural links are not necessary in all cases.
Notice however, that although I talked about playing chess and
proving theorems I said nothing about the *mechanisms* used for the
tasks.
When a human brain is involved, all sorts of very subtle and
complex mechanisms play a role, and for all I know most of the
important ones don't match the "digital and highly symbolic"
processes you are talking about.
The nature of a game or task does not constrain the nature of the
mechanisms capable of playing the game or performing the task.
Do you know for a fact that human brains use only digital processes for
playing chess or solving problems in mathematics or logic? I doubt it.
(I am not saying one thing or the other: it's irrelevant to the
original topic.)
[AW]
> In these sorts of activities, of course a large amount of
> conscious and explicit ratiocination precedes action.
That's NOT true of a real expert. Both in mathematics and in chess
the more sophisticated the expert the less conscious and explicit
the ratiocination that leads to decisions.
I suspect that this is true for all skills (one day I'll give some
reasons, but it's not relevant to our discussion of whether cultural
precursors are necessary, since both conscious explicit processes and
unconscious implicit processes may or may not have cultural origins.)
> ... irrelevant stuff about non-intellectual processes deleted ....
[AS]
> >From the second standpoint (i.e. Young's and mine), whether
> >mentalistic descriptions were appropriate would depend not on the
> >history of the system but on what the processes were, ....
[AW]
> To refer to "what the processes were" in this way is perhaps question
> begging against the externalist.
well, I can't cash that till we know a lot more about what ACTUALLY
happens in human brains.
All I am saying is that whatever actually happens is enough to justify
the use of intentional descriptions in the "W" sense, even if there are
similar descriptiosn of the "C" type that would not be applicable (just
as in one sense of "ellipse" a circle is not an ellipse whereas in the
mathematical sense it is).
[AW]
> In a real chess player, "what the
> process is" is one of: planning an attack, worrying about the
> weakness on the queen-side, etc. On the view in question these processes
> do not occur in a calculator, although structural analogues do.
I don't know why you keep bringing in calculators. If calculators
are not rich enough, use some other mechanism that's closer in
richness and power to brains, but not connected to a culture.
[AS]
> >.....it would depend on the
> >present capabilities of the system and how it used them.
[AW]
> Once again, "present capabilities" threatens to beg the question.
Not the question I was discussing which was whether past influences are
required.
> On the historical view, a person has capabilities a calculator by
> itself does not, eg, the capability to discharge a debt by writing a check.
which is obviously true, but totally irrelevant to whether
intentional descriptions are applicable to a culturally disconnected
mechanism. Don't forget that "some" /= "all".
[AW]
> .....From the standpoint
> of the critique of intellectualism in Ryle, and in the phenomenologists
> who have influenced Dreyfus, it is something like a constitutive mistake
> to try to characterize understanding of the everyday world on the model
> of the formal and mathematical sciences.
Whenever I hear a philosopher use the word "constitutive" I suspect
that it's a symptom of a rigid adherence to a particular set of
semantic decisions. (Likewise "criterial".)
As for Ryle. If you read his work carefully (I had the benefit of
also attending his lectures and classes in the late 50s) you will
find that not all his dispositions are dispositions to produce
*external* behaviour. E.g. he talks about the process of singing a
tune to yourself (i.e. in your mind) as one in which there's a
succession of dispositions to produce new dispositions, all of whose
manifestations are purely mental.
What's more, I think he was quite right: a human-like mind is a huge
collection of coexisting interacting dispositional states constantly
generating large numbers of internal processes very few of which
ever produce external manifestations, and when they do they need not
always do so reliably.
Of course, he did not know nearly as much as we do about kinds of
machines with sophisticated architectures, that might be capable of such
elaborate internal processing without external manifestations.
[AW]
> It is much more plausible
> to say that formal structures are all there is to these domains. But
> not for practical mastery of knives, forks, plates, tables, chairs,
> kitchens, dinner-times, restaurants, eating out, etc.
It's now looking to me as if you want to qualify your original
statement, the one against which I reacted, i.e. this
[AW]
| For example, I myself take Wittgenstein and others to have shown that
| human conceptual thought is only possible in the context of a framework
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| cultural norms, in particular linguistic norms, which is prior to
^^^^^^^^^^^^^^^
| individual thinkers and thinkings and makes them possible. This is part
| of the metaphysical nature of conceptual thinking, in my view.
You now seem to be restricting this to human conceptual thought
about socially relevant or cultural things (dinner things, dinner
times, restaurants, etc.)
If that's what you meant in the first place I apologise for getting
it wrong. I thought you were more like the typical Wittgensteinian
who makes much stronger statements, of the sort I was criticising.
[AW]
> Again, "causal powers" seems question begging. I have the present causal
> power of bringing about a change in the US. economy by spending money,
> for example.
Another fallacious use of "some...therefore...all..." Or irrelevant?
[AS]
> >So Anders has, for reasons that he can perhaps explain, adopted one
> >set of implicit definitions whereas I adopt another.
[AW]
> I think this line is a little cheap. A disagreement is purely semantic when
> both sides agree on the integrity and utility of the concepts, and
> disagree only on the labels.
Ok: in your sense it is not "purely semantic." I was using "purely
semantic" in a different sense, namely we are in disagreement about
whether a collection of concepts should be used, and whether a
particular form of words should be used to express them.
You are saying that a certain set of concepts has no use unless
combined with extra conditions referring to cultural precursors,
whereas I am saying they are potentially useful when they able me to
talk about processes in objects with no cultural (or biological)
causes, without my having to invent some elaborate new family of
concepts similar to the old ones except for the cultural baggage.
[AW]
> In practice few philosophical disagreements
> are like this. They are more questions about which concepts are
> the right ones to use.
I cannot make any sense of this. There's no such thing as a "right"
set of concepts. A set of concepts may be more or less useful for
particular sets of tasks, recognizing food, avoiding predators,
finding shortcuts, storing information for future use, making
predictions, and possibly also communicating with other concept
users.
[AW]
> This makes fruitful discussion exceedingly difficult,
> no doubt about that, but it is not impossible to advance considerations that
> weigh for and against a certain way of talking.
As I did: if I were not allowed to use the existing concepts to describe
processes in the brain in the vat or in the disconnected mathematician,
I'd have to invent a whole isomorphic family of concepts, which seems
like a waste of effort if the existing ones work well when used among
consenting adults.
[AW]
> (If Quine is right, then
> every debate can equally be characterized, by semantic ascent, as a
> debate about how we ought to talk.)
Quine was a good logician. His philosophy is a different matter.
[AW]
> At a minimum, I would take myself to be trying to make explicit the
> ordinary psychological concepts we actually use.
^^^^
Who is "we"? Speak for yourself. Mine are different, as should be
very clear from this discussion. But feel free to go on using yours,
if that's what you wish.
You may even be speaking for the majority of English speakers. That
does not bother me, as the majority of people have never thought in
any depth about the possibility of a disconnected thinker, and they
have never had the chance to design one, or observe one.
[AW]
> If I got grand, as
> I occasionally do, I would say I would like to help us to a better, more
> explicit understanding of the sort of things we are. And this is *not*
> computer programs, though we might have some of those inside us.
[AS]
Well I am not a computer program either, you'll be glad to hear.
On the other hand none of this discussion seems to me to shed any light
on whether a sufficiently sophisticated computer-based architecture (not
just "a program" !) might or might not have so much in common with
what goes on in human brains that refusing to use common mentalistic
language to describe it is about as silly as trying to run a race with
your feet tied together.
It's an unnecessary, crippling, semantic restriction.
[AW]
> But I don't think it is right to make it sound as though the philosophers
> who think this way have ssimply adopted some kind of arbitrary convention.
Of course, they *think* they are saying something deep and factual,
like the philosophers who argue over whether mathematical results
are discovered or invented, or those who argue over whether
scientific theories say something true or false or are merely
instruments for making predictions.
I think that close examination of many of these disputes shows that
there's no disagreement on any point of substance, just a fight over
which set of semantic rules to allow. But the participants don't see
it that way.
>...
[AW]
> To take one example, consider that great mystery about how the
> brain could, all by itself, radiate consciousness or
> intentionality.
This is just a silly question. I have no reason to believe my brain
radiates any such thing, either all by itself or with help. I am not
aware of any "great mystery".
If my brain radiates anything it's electromagnetic energy and some
heat energy. If you start from such bizarre questions and have to
distort conceptual systems in order to find answers that satisfy you
then maybe we have nothing to discuss.
Of course, there are lots and lots of mysteries about our brains: how
they process various kinds of information, how they store information,
how they transform and interpret information (e.g. in using samples from
an optic array to arrive at detailed information about the structures,
properties and relationships of many objects both near and far), how
they generate new motives, how they select between conflicting motives,
how they sometimes control their thought processes and sometimes lose
control, how they learn about infinite structures, how they control the
hands of a concert pianist and many more.
But these are mysteries to be tackled by science, not philosophy,
least of all philosphical stipulations apriori that cultures have to
be involved in all cases.
[AW]
> Once
> one is clear that the nature even of conscious mental states
> is constituted by relation to external factors,
??????????
(a) I am not sure what "constituted" means here
(b) I suspect that whatever it means this merely expresses a
definitional stipulation.
> ...one can see clearly that
> intentionality is not brought about by the brain alone (Cartesian materialism
> is not any advance over Cartesian dualism).
Your sort of intentionality is not. (The *C kind). My kind is (The *W
kind).
[AS]
> But then it is
> folly to look to brain science for the answer to the supposedly
> great puzzle, as absurd as taking a dollar bill to bits to find the secret
> of its value. Brain science can answer lots of questions, but not that one.
^^^^^^^^^^^^^^
Well *that* one strikes me as an incomprehensible puzzle whose answer
will never be of any interest to me anyway. So I'll stick with the
other questions for now, and continue to talk about a brain in a vat
as having the ability to search for mathematical results and feel
pleased at successes and frustrated by failures.
> ....
[AW]
> ...If Ryle and cohort are right, then a lot of the
> psychological language does not in fact talk about inner states and
> processes, but about molar dispositional states, about manners in which
> overt activity is conducted, or about context-dependent features of
> overt activity, such as what someone intends by one of his words.
I don't know about the cohorts, but this is not the Ryle I know.
E.g. read (carefully) the chapter on Imagination in The Concept of Mind
and don't just believe what people say about Ryle.
What he is arguing against is the view that when you imagine a scene
there really is inside you some sort of ghostly analogue of the
real physical scene and an internal state of seeing that ghostly
analogue.
He is not saying that mental states and processes do not exist.
He is saying that their existence does not involve internal replicas of
the external referents. In discussing a person picturing his nursery he
says (page 248):
"He is not being a spectator of a resemblance of his nursery
but he is resembling a spectator of his nursery".
I.e. the person imagining a scene is partly like a person seeing the
real scene, but not because he is really "seeing" an imaginary
replica of the scene.
He was not opposed to mental states and processes. He was opposed to
bad theories about what those mental states and processes were,
especially theories that generated infinite regresses of inner objects.
[AW]
>...A key
> point in this philosophy is that reason-giving explanation is not
> primarily about physically inner events and states.
It depends entirely on what is being explained. If you want to
explain why I suddenly stopped and started walking in the opposite
direction the explanation may be that I suddenly remembered an
important unfinished task. That's an INTERNAL event that explains my
abandoning one action and starting another. It also gives my
reasons.
It's not a physical event, but it is IMPLEMENTED in physical mechanisms.
If you want an explanation of why the task is important, that may
take a totally different form.
[AW]
> ..(Anyone who
> understands the possibility of knowledge compilation, or the
> distinction between the knowledge level and the symbol level, ought to
> understand this)
I have argued elsewhere (against Newell and Dennett) that between the
knowledge level and the symbol level is an information level.
> So it's not clear this language can be bent to your
> purpose.
I can use it straight, i.e. unbent.
[AS]
> >I have no objection to Anders and others adhering to their
> >restricted usage of mental concepts, as long as they realise that
> >there is no reason at all why others like me should follow their
> >semantic preferences, and some good reasons why we shouldn't.
[AW]
> I must say the first clause comes out like a pretty backhanded concession,
> given what follows.
Yup!
The concession will not be enough for you, if you want to impose your
semantic preferences on me.
Enough. This is much too long already.
Cheers.
Aaron
Aaron Sloman
> Date: Wed, 26 Jun 1996 23:36:01 -0400
> Organization: Massachvsetts Institvte of Technology
> .....
> As to *why* there are those who want to think of function in
> terms of history. First of all, the history always seems to
> coincide with function, for those things whose history we know.
I don't understand this use of "coincide".
Anyhow, if I come across a complex system I can study it and work
out that certain parts control other parts, and certain things are
energy sources, and other things are sensors, and other things are
motors, and by watching it I can tell that some components have the
function of facilitating escape from certain potentially destructive
situations.
For all of this I do NOT need to know anything at all about the system's
history.
I admit that analysing this concept of "function" is not easy, and
there may not be any sharp and clear division between things whose
components have functions and the rest.
Of course, if you don't want to use the word "function" except where you
know there has been some particular sort of historical background that's
fine.
I won't force you to talk about my kind of function.
> Second of all, it offers the hope of grounding something that on
> the face of it seems hopelessly open to all sorts of interpretation,
> i.e., function and meaning, on something that is unalterable and
> objective, i.e., history. The thermostat is of course currently disposed to
> control the temperature, but it is also disposed to take up wall space,
> to break when hit with a hammer, etc. How do you propose to eliminate
> those as functions? An easy way to eliminate them is to note that
> the thermostat was not constructed the way it was because it was easy
> to break, and it was not placed on the wall to take up space.
In some cases I'll describe the function of a thermostat in relation
to the goals of a designer or user.
In other cases I'll study the system of which it is a part (e.g. the
physiology of an animal) and see how by regulating the temperature a
mechanism helps to maintain that system and enable other things to
function so that they can better maintain that system, e.g. by helping
it avoid things that try to eat it, keep it going through cold weather,
etc. etc.
I do NOT need to know anything about its evolutionary history in order
to carry out this investigation of how it works.
If you ask whether the function of the lower half of a rock is to
keep the upper half off the ground, then my instinctive response is
that at best this is just a limiting case of function where there is
very little cooperation between subsystems and no clear advantage to
anything in keeping the upper half off the ground.
Many concepts have difficult limiting cases, about which fruitless
arguments can continue forever. (Like arguments over whether
thermostats have beliefs and desires as McCarthy once claimed.)
[AS]
> > My reasons are that in principle if I come across a sophisticated
> > system with the capabilities of the brain in the vat or a software
> > system such as I described above, then I want a language for
> > describing the internal states and processes, and since the language
> > for describing human mental states and processes already goes a long
> > way to meeting the need, as long as you drop any historical
> > presuppositions, then it is far simpler and more elegant to use the
> > existing language than to invent a new language that duplicates all
> > that richness.
[EF]
> I just don't think this is going to happen. You won't "come across"
> a sophisticated system that just popped into existence. It's going
> to have a history.
I was objecting to what appeared to be the very strong claim by Anders
that IF such a thing popped into existence it could not be described in
mentalistic terms.
One can discuss hypothetical questions even if the hypothesis has a
probability close to zero. Otherwise the claim I was objecting to would
have been a pointless claim.
Anyhow, as you can see, I think that even where a complex system
does have a history that's not what I am studying when I try to find
out what the functions of its components are.
(I think this emphasis on history comes from a period when certain
hard-headed biologists thought that talking about function was
unscientific and "teleological" in some bad sense. Then Darwin came
along and gave them a new justification for talking about function
in terms of natural selection.
But I don't think they needed the Darwinian excuse, and the original
talk of function was not based on any unscientific or unjustified
teleology.
It was adequately justified by what they found in highly integrated
organisms. However, that's just another bit of history...)
Aaron
Ed Faith
> Ed Faith <epf...@aol.com> wrote
> > .....
> > As to *why* there are those who want to think of function in
> > terms of history. First of all, the history always seems to
> > coincide with function, for those things whose history we know.
>
> I don't understand this use of "coincide".
I only mean that things generally turn out to have been selected
(by nature, by a company for mass production, etc.) on accout of
X, where X is something we'll usually accept as its function,
even when we don't know the history.
> Anyhow, if I come across a complex system I can study it and work
> out that certain parts control other parts, and certain things are
> energy sources, and other things are sensors, and other things are
> motors, and by watching it I can tell that some components have the
> function of facilitating escape from certain potentially destructive
> situations.
>
> For all of this I do NOT need to know anything at all about the system's
> history.
This is true, as I pointed out. We don't actually know enough about
our evolutionary history directly, to infer the function of this or
that bodily organ. Rather, we infer both history and function from the
current form.
> I admit that analysing this concept of "function" is not easy, and
> there may not be any sharp and clear division between things whose
> components have functions and the rest.
My point is that both concepts of "function" in most real-world cases
should probably be compatible with the same discovered functions. The
only difference is that in the one case we would take ourselves as making
guesses about the function--we can only make guesses since the history
is not available to us directly. In the other case we either still
take ourselves as making guesses, or else we try to justify our choices
by defining function as a function current physical form. I don't myself
think the latter project has much hope, at least not if we want a strict
means whereby we can discover function, the way we discover weight or
temperature.
> Of course, if you don't want to use the word "function" except where you
> know there has been some particular sort of historical background that's
> fine.
Like I said, in reality one just supposes the background to have been
just-so.
> In some cases I'll describe the function of a thermostat in relation
> to the goals of a designer or user.
But this employs a historical notion. You're willing to adopt a
historical notion for human-made things. Why not for naturally
evolved things?
Peter_...@uk.ibm.com
>If Mary asserts to me, say, that Bill Clinton is now in the White
>House, I don't have to look inside her brain to assess the claim, agree
>or disagree with it, or ascribe the corresponding belief to her.
>Looking inside her brain would add nothing in the way of understanding
>what was said and would really be kind of silly.
Change the example. Suppose that some-one wanted to join the
Secret Service or MI5. You now want to check whether this
person is or is not attempting to be a double agent of some
foreign power. Are you now so sure that looking at their brain
(were we able to do this) would be silly? And are you now so
sure that 'Looking inside [the] brain would add nothing in the
way of understanding what was said'. If one could understand
that what is said is lies, wouldn't that add something?
I can understand how, by using my brain, I can hide something.
But I find it more difficult to understand how, by using my brain,
I can hide the fact that I'm using my brain to hide something.
Cheers,
Pete Lupton
Anders N Weinstein
H. M. Hubey <hu...@pegasus.montclair.edu> wrote:
>ande...@pitt.edu (Anders N Weinstein) writes:
>
>Regardless of where your interest lie, if the mind is the working
>of the neurons,
This is a slogan I would emphatically reject.
> then all of your ideas have to somehow
>take the neurons into consideration. Continuum mechanics is able
I can't see how it could possibly be necessary to take neurons into
consideration when trying to find out what someone means by a claim,
whether it is well-argued, worthy of credence, etc. The neurons only
seem to come into consideration in abnormal cases -- schizophrenia,
brain damage and the like, when we are forced to suspend to some degree
a fully personal stance towards our interlocutor.
> Continuum mechanics is able
>to give wonderful descriptions of elasticity, and fluid dynamics
>although we all know about the discrete/quantized nature of
>reality.
I think a better analogy would be this: suppose I ask, in an everyday
context, whether a certain table is solid. Clearly what I mean is
something like -- can I rest things on it, perhaps lean or stand on it
without its breaking, and so on. We can usually answer this question
by ordinary means, and find out quite conclusively that the table is
(or is not) what I meant by "solid".
Now physics can in some sense explain why it is has these powers in terms
of molecular structure. But, contra a famous blunder of Eddington's, physics
can never discover that the table is "not really solid" -- not if it manifestly
*is* solid in the only sense of the word that was at issue. We might take
a term from the computational school and say that material science gives
the implementation level for common-sense, macro-world solidity, but that
the latter concept is really implementation-independent -- part of
a different and autonomous level of description.
I think things are similar with the relation of brain and so-called
cognitive science to ordinary intentional explanation. Brain science is
concerned with how our psychological powers and abilities are
implemented. But it does not explain what it is to possess them --
that's more a matter for reflection on our common-sense concepts, which
are similarly implementation- independent.
I think the implementation-independence of common-sense intentional
concepts may be even greater than the implementation-independence of
ordinary macro-physical concepts, if the common-sense concepts provide
the terms we need for *normative* assessment of people's epistemic
conduct.
>>If Mary asserts to me, say, that Bill Clinton is now in the White
>>House, I don't have to look inside her brain to assess the claim, agree
>>or disagree with it, or ascribe the corresponding belief to her.
>>Looking inside her brain would add nothing in the way of understanding
>>what was said and would really be kind of silly.
>
>Putting together 100 billion artificial neurons might not be
>out of the capability of humanity in a century or so. Of course,
>we can dream on, and imagine that we can somehow find a way to
>noninvasively and nondestructively examine and keep track of all
>the neuron firings in Mary's brain when she makes her assertions.
>Would you still say that looking into that would be useless in
>adding to our understanding?
I meant "add to our understanding of what was said". Of course all
knowledge "adds to human understanding" in some sense. The claim was
it does not explain how it is that she can take a stand in the public
sphere which answers for its correctness to how things are in the White
House.
>I don't have any problems taking into account all facts, including
>behavior, or even positing unobservables if the resulting explanation
>is more successful than others. I don't have any problems with the
>existence of emotions, attitudes, or whatnot. But in the final
>analysis, I don't see how or why we can forget about the basis
>of all this activity which is neuronal firings.
Depends what you mean by "basis". In a causal sense the level of
molecules is the basis for macro or common-sense solidity, but it does not
explain what it is, and we can perfectly well find truths at the higher
level while ignoring the molecules.
>> The meaning of "inner" intentional states might be explained
>>derivatively, in terms of their potentiality for being expressed in a
>>public language.
>
>Fine. Let's use them to good effect.
I nowhere denied the existence of "inner" states. Just denying that
the "inner" states of intentional psychology are individuated according to
the interests of brain science. Rather, the carving up of the "inner"
into individual states is dependent on their potentiality for flowering
into overt expressions in particular situations.
>How else can it be? I can conjure up the inside of my apartment
>in great detail even when I am not there. Somehow my 3D version of
>my aparment is stored in my neurons and I can activate it at will.
>I can, if I want, imagine that a thief might have broken into it
>and even imagine what he might be doing. I don't understand exactly
>what you are referring to if not to the way in which all of this
>is coded into the neurons.
Perhaps your thought about your apartment can be said to have an
implementation, and that would then be in neurons. But when you imagine
your apartment you stand in an intentional relation to the apartment,
which is only possible becuase you have had interactions with the
apartment -- the neurons alone can't fix it that you are thinking about
your apartment rather than another identical one. That's the real
issue.
There are also many intentional states which lack neural implementation,
possibly, e.g., your belief that the floor is solid when you walk. So
looking to the neurons is no help with these.
Anders N Weinstein
>
>Change the example. Suppose that some-one wanted to join the
>Secret Service or MI5. You now want to check whether this
>person is or is not attempting to be a double agent of some
>foreign power. Are you now so sure that looking at their brain
>(were we able to do this) would be silly? And are you now so
>sure that 'Looking inside [the] brain would add nothing in the
>way of understanding what was said'. If one could understand
>that what is said is lies, wouldn't that add something?
Fair enough. But. What would you have to find to tell you that
they're lying? To adapt an example of Dennett's, suppose your
job is made easier by discovering actual sentences of English
inscribed in there. You look and find the sentence "I am a double
agent", somewhere. Well that looks promising, but a question remains:
is it something the agent *believes*, or is it perhaps sarcasm, something
he remembered from a spy movie, something he is fantasizing about?
Finding other sentences which cohere with the spy hypothesis
would be a plus, sure, but still wouldn't settle these questions.
What you would need to find out is whether these sentences have a certain
causal potentiality -- certain typical effects in appropriate situations.
e.g. whether they are such as to bring about the person's passing
secrets to other foreign agents in a certain situation where he thinks
no one is watching.
But once we recognize this, one sees that one can bypass the middleman,
and find out if someone has that or similar dispositions without looking
in their brain, but by arranging the situation or a simulacrum of it and
seeing what they do. Which, I take it, is how anti-espionage sometimes
proceeds.
So I concede we *might* one day be able to identify such a typical causal
potentiality with some brain state and so identify spies by brain scans.
But we also might not, and shouldn't presuppose that it is possible. The
concept of deception doesn't depend on it.
Ed Faith
This continues a previous reply.
> I do NOT need to know anything about its evolutionary history in order
> to carry out this investigation of how it works.
Nor do you need to know anything about its evolutionary history to
make conjectures about its evolutionary history.
> Many concepts have difficult limiting cases, about which fruitless
> arguments can continue forever. (Like arguments over whether
> thermostats have beliefs and desires as McCarthy once claimed.)
Of course they do. I'm talking about the "center", not the "limit".
The important question for me is, which is the simplest way of defining
function, which captures most of what we call function, and from which
the rest can be derived by the usual means (e.g., similarity and close
relation--thus a table has a leg by reason of similarity to animal
legs, and both a single written book and the various individual copies
of that book are called "book", by reason of close relation).
> I was objecting to what appeared to be the very strong claim by Anders
> that IF such a thing popped into existence it could not be described in
> mentalistic terms.
No problem--just use analogy, or even metaphor. So we are allowed to say,
not just, "it's as if the object thinks", but, "the object thinks"--which is
allowed by reason of poetic license. I have no problem with that. Nor do
I have a problem with the view that the object literally thinks--which
simply amounts to a natural expansion of the concept through similarity.
However, there is also IMO a "core" of the concept which can be reasonably
simply explained in terms of history. When we get around to explicitly
defining or explaining a concept, we should not worry about trying to account
for every conceivable expansion, but instead we should try to find the core,
if there is one.
> One can discuss hypothetical questions even if the hypothesis has a
> probability close to zero. Otherwise the claim I was objecting to would
> have been a pointless claim.
Because of our tendency to expand concepts beyond their original confines
(in the ways I alluded to just above), no matter what the original concept,
we will probably find things which do not fit it, and yet which are, say,
similar enough, or related enough, to things which do fit it, that we'll
want to apply the term to those things. Your "hypothetical question" brings
up such a case. It does not fall under the "core concept", but we still
want to apply the word to it. My general feeling about it is: so what?
All you've done is illustrate, with one concept, something that's true
of concepts in general.
> Anyhow, as you can see, I think that even where a complex system
> does have a history that's not what I am studying when I try to find
> out what the functions of its components are.
That is like saying: you don't study the history of the universe when
you try to find out what its history is. That is true--in the sense
that what you study is the current state (e.g., the current pattern of
light hitting your telescope), and you use that to make guesses about
the history. However, the purpose of your inquiry is to make discoveries
about the history of the universe, and in that sense you're studing
the history.
> (I think this emphasis on history comes from a period when certain
> hard-headed biologists thought that talking about function was
> unscientific and "teleological" in some bad sense. Then Darwin came
> along and gave them a new justification for talking about function
> in terms of natural selection.
The question is what "function" means. We aren't necessarily able
to describe what a word means--the capacity to use a word correctly,
doesn't depend on the capacity to describe its meaning. The claim
that I perceive as being made is not that we should start talking
about function in certain terms, but that this is what function was
about all along, though we weren't explicitly aware of it until now.
Aaron Sloman
Ed Faith <epf...@aol.com> writes:
> Date: Fri, 28 Jun 1996 08:47:39 -0400
> Organization: Massachvsetts Institvte of Technology
> ....
> I only mean that things generally turn out to have been selected
> (by nature, by a company for mass production, etc.) on accout of
> X, where X is something we'll usually accept as its function,
> even when we don't know the history.
I presume "generally turn out" is some sort of empirical claim about
what most cases are like. It leaves open the possibility of
exceptions that don't fit the general pattern. That's all I was
getting at.
> > Anyhow, if I come across a complex system I can study it and work
> > out that certain parts control other parts, and certain things are
> > energy sources, and other things are sensors, and other things are
> > motors, and by watching it I can tell that some components have the
> > function of facilitating escape from certain potentially destructive
> > situations.
> >
> > For all of this I do NOT need to know anything at all about the system's
> > history.
>
> This is true, as I pointed out. We don't actually know enough about
> our evolutionary history directly, to infer the function of this or
> that bodily organ. Rather, we infer both history and function from the
> current form.
We SOMEtimes infer history. My point is that we don't have to infer
it or even think about it in EVERY case.
If all you are saying is that the existentially quantified statement
is true, then we are in no disagreement, for nothing I said was
intended to claim that history is NEVER relevant.
>.....
> My point is that both concepts of "function" in most real-world cases
> should probably be compatible with the same discovered functions.
Your "most" and "probably" are compatible with my "not always". I
was originally objecting to a statement that history was *always*
involved.
>....
> > Of course, if you don't want to use the word "function" except where you
> > know there has been some particular sort of historical background that's
> > fine.
>
> Like I said, in reality one just supposes the background to have been
> just-so.
Sometimes yes and sometimes no.
For instance, if a RANDOM mutation occurs that enables an INDIVIDUAL
to produce some new enzyme that helps it cope better in a given
environment, then I can identify the function of the new enzyme by
examining the individual in which it is produced, without assuming
that it has already been selected by evolutionary pressures. They
may not have had time to operate, given that this individual was the
subject of the mutation.
> > In some cases I'll describe the function of a thermostat in relation
> > to the goals of a designer or user.
>
> But this employs a historical notion. You're willing to adopt a
> historical notion for human-made things. Why not for naturally
> evolved things?
When a statement is obvious true I acknowledge it. When some other
statement is not true I do not treat it as true simply because it is
analogous to another that is true.
NB: "Some" does not imply "all."
(It's amazing how many people who contribute to this news group seem
to be unable to grasp this.)
I do not say that history is NEVER relevant to naturally evolved
things.
I merely say that (a) I need not be interested in the
history in order to study the function of an organ or process
(b) there need not be any history if this is a case of random
mutation.
OK?
Aaron Sloman
for which thanks.
> Date: Fri, 28 Jun 1996 15:13:16 -0700
> Organization: MIT
> ...
> .....I'm talking about the "center", not the "limit".
> The important question for me is, which is the simplest way of defining
> function, which captures most of what we call function, and from which
> the rest can be derived by the usual means ...
>.....
> However, there is also IMO a "core" of the concept which can be reasonably
> simply explained in terms of history. When we get around to explicitly
> defining or explaining a concept, we should not worry about trying to account
> for every conceivable expansion, but instead we should try to find the core,
> if there is one.
OK I think I understand better what you are saying.
However, as far as I am concerned it is not a correct analysis of
what I understand by function. It just happens to be the case that
MY concept of "X having function Y" is primarily concerned with what
the causal powers of X are within a larger system Z in which X
(normally) achieves Y which in some way contributes to the needs of
Z or protects Z from damage or destruction.
Sometimes X was deliberately selected for this role by agent, e.g. a
designer of Z. But that is an extraneous fact about the function,
namely that it had this sort of cause.
I assure you that's how MY concept works, and until I started
reading your postings I thought that was how it was for everyone.
However if YOUR concept of function has this historical connection
that's fine with me. I have no desire to force you to use my type of
language.
We can say there are two senses of having a function, which we might
label differently as "functionh" (historical) and "functiona"
(ahistorical). You use "function" to mean "functionh". I use
"function" to mean "functiona".
It's a purely semantic difference, like the disagreement over
whether a circle is an ellipse.
End of debate.
(Incidentally, I don't regard dictionaries as necessarily telling us
what words actually do mean, but neither of the two dictionaries I
have to hand says anything about historical origin in specifying
what "function" means. That doesn't prove my usage is correct,
hwoever.)
> The question is what "function" means.
So we have the answer: what it means for you is different from what
it means for me.
Cheers.
Ed Faith
> However, as far as I am concerned it is not a correct analysis of
> what I understand by function. It just happens to be the case that
> MY concept of "X having function Y" is primarily concerned with what
> the causal powers of X are within a larger system Z in which X
> (normally) achieves Y which in some way contributes to the needs of
> Z or protects Z from damage or destruction.
This fails to pick out Z, assuming it instead. Whereas if you
think of evolution in, for example, Dawkins's terms (in terms of
selfish genes and selfish memes), then the function of genes is
whatever they do (or their ancestors did) that encourages their
reproduction, as against the reproduction of competing genes.
Which does not assume a Z. Rather, in the process of investigating
what it is they do, we discover the whole system which they
serve and protect, and in so doing serve themselves.
Another problem I have with your conception is that it doesn't
identify the function of malfunctioning parts, i.e., parts which have
a function but fail to discharge it. Those parts fail to have causal
powers which serve and protect Z, and so on your account have no
function, and so can't be said to malfunction. It also doesn't identify
the function of parts which do not discharge their function as well as
some other possible replacement part.
Let me say something about this last point. When we say "X causes Y",
we don't mean that Y follows X a certain proportion of the time. Y
can follow X as infrequently as you like and still be caused by X, or
as frequently as you like and be uncaused by X. What's important is,
roughly, that Y follows X more often than Y occurs without X. Let's
call the absence of X, X'. In the case of X as a part of Z, X'
can be seen as the collection of possible replacements for X, including
(but not privileging) empty space (if X is a physical thing). Now,
X may not be the part which is most frequently followed by Y. Suppose
some part V is more frequently followed by Y. Then with respect to
V, X actually prevents, rather than causes, Y. If preventing Y does
not serve Z, then X has no function, considered relative to V.
So whether X causes Y is relative to what X is compared with. Evolutionary
accounts include the comparison. X was in actual competition with certain
other parts. Those are the parts relative to which X causes Y.
Peter Lupton
>
> In article <4r086s$12...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
> >
> >Change the example. Suppose that some-one wanted to join the
> >Secret Service or MI5. You now want to check whether this
> >person is or is not attempting to be a double agent of some
> >foreign power. Are you now so sure that looking at their brain
> >(were we able to do this) would be silly? And are you now so
> >sure that 'Looking inside [the] brain would add nothing in the
> >way of understanding what was said'. If one could understand
> >that what is said is lies, wouldn't that add something?
>
> Fair enough. But. What would you have to find to tell you that
> they're lying? To adapt an example of Dennett's, suppose your
> job is made easier by discovering actual sentences of English
> inscribed in there. You look and find the sentence "I am a double
> agent", somewhere. Well that looks promising, but a question remains:
> is it something the agent *believes*, or is it perhaps sarcasm, something
> he remembered from a spy movie, something he is fantasizing about?
> Finding other sentences which cohere with the spy hypothesis
> would be a plus, sure, but still wouldn't settle these questions.
>
> What you would need to find out is whether these sentences have a certain
> causal potentiality -- certain typical effects in appropriate situations.
> e.g. whether they are such as to bring about the person's passing
> secrets to other foreign agents in a certain situation where he thinks
> no one is watching.
Absolutely. It may well be that the brain scan would enable us to resolve
counter-factuals far more efficiently, perhaps, than any other way.
Suppose that our spy was actually a sleeper, activated by a certain
coded message. Surely, it would be far easier to determine this by
a brain scan than by attempting the alternative - to attempt to
re-create for our spy an astronmically large number of credible
scenarios each containing one of the possible coded messges. Presumably,
our spy would smell a rat after the first few billion such tests.
> But once we recognize this, one sees that one can bypass the middleman,
> and find out if someone has that or similar dispositions without looking
> in their brain, but by arranging the situation or a simulacrum of it and
> seeing what they do. Which, I take it, is how anti-espionage sometimes
> proceeds.
Well, we might in theory - but surely not typically in practice. I would think
that, in such cases, 'brain-reading' could be *extremely* effective at
weeding out double agents, in comparison with other techniques.
> So I concede we *might* one day be able to identify such a typical causal
> potentiality with some brain state and so identify spies by brain scans.
> But we also might not, and shouldn't presuppose that it is possible.
I think it's obviously possible, unless, that is, the physical world is
very different from what we think it to be. Perhaps that's all you
mean by not presupposing the possibility?
> The concept of deception doesn't depend on it.
Ah! That *is* all you mean by it.
I'll just point out that you opened this with the assertion that it
was silly to consider looking inside some-one's head for mental
state. I think we can now say that, actually, it would be sort of silly
to think that looking inside some-one's head would *not* be effective.
The sort of silliness that considers empty head to be a real (as distinct
from a mere logical) possibility. When it comes to silliness and practical
matters, empty head is about as silly as things get, logical possibilities
notwithstanding.
--
Cheers,
Pete Lupton
Aaron Sloman
Ed Faith <epf...@aol.com> writes:
> Date: Fri, 28 Jun 1996 21:21:15 -0400
> Organization: Massachvsetts Institvte of Technology
>
> Aaron Sloman wrote:
>
> > However, as far as I am concerned it is not a correct analysis of
> > what I understand by function. It just happens to be the case that
> > MY concept of "X having function Y" is primarily concerned with what
> > the causal powers of X are within a larger system Z in which X
> > (normally) achieves Y which in some way contributes to the needs of
> > Z or protects Z from damage or destruction.
[EF responds]
> This fails to pick out Z, assuming it instead.
I don't know what I'm supposed to have failed to do.
If Z is a dog, or a tree, or a bacterium and X is part of the
physiology or metabolism of Z, what's the problem? Of course I start
by assuming that I have found an object to study (dog, tree,
bacterium or whatever). What's the problem?
I don't need to know anything about evolution to find dogs or trees
interesting and to wonder how they work, and to find out that the
functions of roots of a tree include keeping the tree upright and
providing it with water and nourishment.
Did Harvey know or care about evolution or origins when he
investigated the function of the heart and blood vessels?
> Whereas if you
> think of evolution in, for example, Dawkins's terms (in terms of
> selfish genes and selfish memes), then the function of genes is
> whatever they do (or their ancestors did) that encourages their
> reproduction, as against the reproduction of competing genes.
> Which does not assume a Z. Rather, in the process of investigating
> what it is they do, we discover the whole system which they
> serve and protect, and in so doing serve themselves.
I don't know what point you are making.
If you are saying that SOMETIMES we don't start with an
understanding of the whole system Z but discover it in the process
of investigating a part X, then I have no disagreement with you.
This not contradict my claim that the study of a function of a part
of a system does not NEED to take any account of what originally
brought it about that the part has that function.
I suspect there are plenty of examples in social science of cases
where we don't start off by knowing what the relevant larger whole
is. (That was one of the claims of Marx, I guess.) Maybe also in
ecology. This are all examples of what you are talking about, if
I've understood you. But I said nothing to deny such possibilities.
It's interesting to ask how we decide that we have found the whole
system Z of which X is a part, and within which X has a function Y.
That could be a good topic for another discussion.
[EF]
> Another problem I have with your conception is that it doesn't
> identify the function of malfunctioning parts, i.e., parts which have
> a function but fail to discharge it.
Again, I don't see the problem, though I wasn't trying to say
anything about this. By studying an animal using sophisticated
non-invasive techniques I may discover that the function of a heart
is to pump blood around the system as a means of transporting
various nutrients and waste products. I may then notice that
something goes wrong and the heart stops working properly the animal
no longer copes as well with its environment, or possibly even dies.
I.e. I find that the heart malfunctions and I don't need to know
anything about evolution of hearts to do this.
Similarly if I see a bird with a broken wing and I see that it can't
fly, or cannot fly very well (e.g. not well enough to get away from
another animal chasing it) I can infer that there's a malfunctioning
part.
What's the problem?
Sometimes I may have to do lots of experiments to find out that all
instances of a particular organ in a certain group of animals are
malfunctioning perhaps as a result of a recent increase in the
prevalence of a chemical in the environment. E.g. the experiment may
include removing that chemical from part of the environment of some
of the animals and then discovering that the particular organ
develops more fully and they are therefore better able to find food,
avoid predators, or whatever.
[EF]
> ...Those parts fail to have causal
> powers which serve and protect Z, and so on your account have no
> function, and so can't be said to malfunction.
I think you have not noticed the simple possibility of detecting
that something is broken or damaged either by looking at it, or by
comparing it with others in the same animal (comparing one wing with
another), or by comparing it with other animals of the same sort
or doing experiments of the sort described above.
I can then formulate true counter-factual conditionals about what
the component would achieve if it were not broken, deformed, or
whatever, i.e. what function it is failing to fulfil.
[EF]
> Let me say something about this last point. When we say "X causes Y",
> we don't mean that Y follows X a certain proportion of the time.
Agreed. Hume got that one wrong.
> ..Y
> can follow X as infrequently as you like and still be caused by X, or
> as frequently as you like and be uncaused by X. What's important is,
> roughly, that Y follows X more often than Y occurs without X.
Er... that's also wrong. The sun rises more often after it has
recently been dark than it rises when it has not recently been dark.
But darkness does not cause the sun to rise.
(There's a huge philosophical literature on what "cause" means. I
think it is one of the most complex concepts we have, and still not
satisfactorily analysed. I don't think any formula that you can fit
into a screenful of text is going to do justice to its complexity.
I once had a PhD student at the University of Sussex, Chris Taylor,
whose thesis explored some of the amazing intricacy of the concept
as ordinarily understood, though I don't think he said the last
word.
Taylor, C.N, 1992 {A Formal Logical Analysis of Causal
Relations} DPhil Thesis, Sussex University. Available as
Cognitive Science Research Paper No.257
>...detail omitted...
[EF]
> So whether X causes Y is relative to what X is compared with. Evolutionary
> accounts include the comparison. X was in actual competition with certain
> other parts. Those are the parts relative to which X causes Y.
Notice that I was able to talk about comparisons above without
having to bring in evolution.
I don't claim that evolution is always IRRELEVANT. I merely claim
that it is not always RELEVANT to the study of functions of parts of
living things.
Some investigations of function or malfunction may certainly appeal
to evolutionary history.
Just remember that "some" does not imply "all".
Cheers.
Ed Faith
>
> We seem to be having a communication problem, though I am not sure.
>
> Ed Faith <epf...@aol.com> writes:
>
> > Date: Fri, 28 Jun 1996 21:21:15 -0400
> > Organization: Massachvsetts Institvte of Technology
> >
> > Aaron Sloman wrote:
> >
> > > However, as far as I am concerned it is not a correct analysis of
> > > what I understand by function. It just happens to be the case that
> > > MY concept of "X having function Y" is primarily concerned with what
> > > the causal powers of X are within a larger system Z in which X
> > > (normally) achieves Y which in some way contributes to the needs of
> > > Z or protects Z from damage or destruction.
>
> [EF responds]
> > This fails to pick out Z, assuming it instead.
>
> I don't know what I'm supposed to have failed to do.
It's not you, it's your conception that fails to pick out Z. Your
conception of function is a relative one. The evolutionary conception
is an absolute one. That is, I can say, "the function of X is Y", and
finish it there. You have to say, "within the larger system Z, the
function of X is Y", because function for you is only defined in the
context of some Z (as I understand you).
> If Z is a dog, or a tree, or a bacterium and X is part of the
> physiology or metabolism of Z, what's the problem? Of course I start
> by assuming that I have found an object to study (dog, tree,
> bacterium or whatever). What's the problem?
There's no real problem because you reliably pick out things that
in fact have an evolutionary history behind them, even though you've
sworn up and down that this is of no concern to you.
> I don't need to know anything about evolution to find dogs or trees
> interesting and to wonder how they work, and to find out that the
> functions of roots of a tree include keeping the tree upright and
> providing it with water and nourishment.
That's a reasonable conjecture about the evolutionary history of the
tree.
> Did Harvey know or care about evolution or origins when he
> investigated the function of the heart and blood vessels?
Does a jeweler know or care about origins when he sells you an
amber necklace? Maybe, maybe not. Nevertheless, whether something
is amber depends on its history, not just its current state. Note
the distinction between what something depends on, and
the methods we use to check it. In fact, currently I take your conception
of function as being a recognition of methods to check function, and
a hasty reduction of function to those methods.
> I suspect there are plenty of examples in social science of cases
> where we don't start off by knowing what the relevant larger whole
> is.
"The relevant larger whole" depends on the history. You have failed
to present anything else that it depends on.
> It's interesting to ask how we decide that we have found the whole
> system Z of which X is a part, and within which X has a function Y.
"How we decide" could be interesting, but we could also ask, what
does Z's being the whole of which X is a part depend on? Which is
separate from the question of what methods we use to decide. It's
this latter question that seems to be answered by the evolutionary
conception of function, and not by yours.
> [EF]
> > Another problem I have with your conception is that it doesn't
> > identify the function of malfunctioning parts, i.e., parts which have
> > a function but fail to discharge it.
>
> Again, I don't see the problem, though I wasn't trying to say
> anything about this. By studying an animal using sophisticated
> non-invasive techniques
The animals asked me to thank you.
> Similarly if I see a bird with a broken wing and I see that it can't
> fly, or cannot fly very well (e.g. not well enough to get away from
> another animal chasing it) I can infer that there's a malfunctioning
> part.
>
> What's the problem?
The problem is there are nonflying bird species. If one of those failed to
escape a cat, then you'd be mistaken to conclude that the wing was
broken or malformed. But what determines whether an individual bird
is a member of a nonflying species? Its parentage.
> [EF]
> > ...Those parts fail to have causal
> > powers which serve and protect Z, and so on your account have no
> > function, and so can't be said to malfunction.
>
> I think you have not noticed the simple possibility of detecting
> that something is broken or damaged either by looking at it, or by
> comparing it with others in the same animal (comparing one wing with
> another), or by comparing it with other animals of the same sort
> or doing experiments of the sort described above.
Comparing it with other parts in the same animal is illuminating
because of a common history among parts. Comparing an animal with
other animals is illuminating because of a common ancestry. When
you say "of the same sort" you are presumably talking about close
relatives in the family tree of life.
> [EF]
> > Let me say something about this last point. When we say "X causes Y",
> > we don't mean that Y follows X a certain proportion of the time.
>
> Agreed. Hume got that one wrong.
>
> > ..Y
> > can follow X as infrequently as you like and still be caused by X, or
> > as frequently as you like and be uncaused by X. What's important is,
> > roughly, that Y follows X more often than Y occurs without X.
>
> Er... that's also wrong. The sun rises more often after it has
> recently been dark than it rises when it has not recently been dark.
> But darkness does not cause the sun to rise.
That doesn't matter. To blunt the point of the criticism, you
have to show the converse, i.e., that in some cases X causes Y even
though introducing X is a sure way to reduce the frequency of Y.
> Notice that I was able to talk about comparisons above without
> having to bring in evolution.
You did not bring it in explicitly, that's true.
H. M. Hubey
>>Regardless of where your interest lie, if the mind is the working
>>of the neurons,
>This is a slogan I would emphatically reject.
Yes, I see that, and I see that that's what lies at the roots
of all disagreements essentially between the dualists, and the
materialists/reductionists/realists/etc/etc...
>> then all of your ideas have to somehow
>>take the neurons into consideration. Continuum mechanics is able
>I can't see how it could possibly be necessary to take neurons into
>consideration when trying to find out what someone means by a claim,
>whether it is well-argued, worthy of credence, etc. The neurons only
>seem to come into consideration in abnormal cases -- schizophrenia,
>brain damage and the like, when we are forced to suspend to some degree
>a fully personal stance towards our interlocutor.
I can't see how you can even make such a claim. No doubt neurons
are involved in abnormal cases, but no doubt they are also
well involved in everyday normal working of the brain. The real
problem now is that we are not yet at the level of knowledge
of the neurons which will allow us to keep track of the firings
of huge masses of neurons (singly, I mean) so that we can
somehow try to correlate neuron firings with things like "well-
argued" ideas, or those "worthy of credence".
>I think a better analogy would be this: suppose I ask, in an everyday
>context, whether a certain table is solid. Clearly what I mean is
>something like -- can I rest things on it, perhaps lean or stand on it
>without its breaking, and so on. We can usually answer this question
>by ordinary means, and find out quite conclusively that the table is
>(or is not) what I meant by "solid".
Continuuum mechanics does exactly that. A fluid is defined as that
matter which will not sustain shear stresses and solids which can
do so. In that sense, continuum mechanics is intimately tied to
our senses or "ordinary means" as you say.
>Now physics can in some sense explain why it is has these powers in terms
>of molecular structure. But, contra a famous blunder of Eddington's, physics
>can never discover that the table is "not really solid" -- not if it manifestly
>*is* solid in the only sense of the word that was at issue. We might take
I don't see what you mean. At the level of atoms one definition of
solidity is that it have a crystalline structure. That is not
close to our "intuitive" grasp of reality, since there are "amorphous"
solids, and also even things like "supercooled liquids" (like glass).
But the idea of solidness is handled well by continuum mechanics
and since that's a part of physics, the matter is already solved.
>a term from the computational school and say that material science gives
>the implementation level for common-sense, macro-world solidity, but that
>the latter concept is really implementation-independent -- part of
>a different and autonomous level of description.
They are different scale/level descriptions of the same reality.
The "reality" of the philosophers of old (and maybe even new)
is that which is observable/feelable via our naked senses. That
keeps the very small and the very large out of our range/scale of
observation, and hence gives rise to mystical philosophies that
"reality" is not the same as "appearance". It is through instruments,
and science that we can peer into the regions which our naked
senses do not allow us.
In that sense, the study of the mind, has one advantage. Not only
can we look at it the same way as into all things we cannot
observe (i.e. the Black Box) but all of us has some kind of
access into at least one representative of such a black box, so that
it's not so black after all. We can almost have our cake and eat it
too. By this I don't mean to ignore the very real difficulties of
having to peer into brains noninvasively and nondestructively.
>I think things are similar with the relation of brain and so-called
>cognitive science to ordinary intentional explanation. Brain science is
>concerned with how our psychological powers and abilities are
>implemented. But it does not explain what it is to possess them --
>that's more a matter for reflection on our common-sense concepts, which
>are similarly implementation- independent.
Looking into parts of a car, or one of the 5 million transistors
of a CPu wouldn't tell you what it's supposed to do either if you
did not already have some idea of what you were looking for. This
is no different than what was just being discussed earlier, namely,
that, knowledge/information is dependent on the observer. I don't
see any mysticism here either because this is similar to the
problem of value in economics.
>I think the implementation-independence of common-sense intentional
>concepts may be even greater than the implementation-independence of
>ordinary macro-physical concepts, if the common-sense concepts provide
>the terms we need for *normative* assessment of people's epistemic
>conduct.
There is no reason why intentional mind/brain states cannot be
a part of a theory. One may include emotions here too. One must
be a real purist to ignore the observed effects of drugs on
emotions, cognition, and action. Why don't you then answer what
kinds of magical powers do certain drugs have that they can
somehow transcend the everyday common properties of the thing
called brain? How about it? If the mind is something so separate
and apart from the brain, then the drugs that effect the mind must
have some of this quality too. Where is it stored in the drug?
Is it a part of the molecules? Can it be measured? How does
it distort the everyday things like intentions, goals, emotions?
>I meant "add to our understanding of what was said". Of course all
>knowledge "adds to human understanding" in some sense. The claim was
Of course it can add to our knowledge. It is doing it already albeit
in a rather limited fashion. Finding brain abnormalities or some
damage to certain parts of the brain certainly adds to our knowledge
of why Mr.X behaved a certain way.
>it does not explain how it is that she can take a stand in the public
>sphere which answers for its correctness to how things are in the White
>House.
It's only a matter of scale/degree/intensity not a difference of
kind.
>Depends what you mean by "basis". In a causal sense the level of
>molecules is the basis for macro or common-sense solidity, but it does not
>explain what it is, and we can perfectly well find truths at the higher
>level while ignoring the molecules.
You are getting mystical. Can we say that we can look at an apple
see its color, its shape, describe its chemical composition but none
of this explains what it is. You are confounding different levels
and scales together.
>which is only possible becuase you have had interactions with the
>apartment -- the neurons alone can't fix it that you are thinking about
>your apartment rather than another identical one. That's the real
>issue.
I can't even understand what you are saying here. This is like
writing a whole book in which every sentences veers off from the
truth in a small imperceptible way. At the end, you have total
mishmash, and not one sentence which is implausible. Philosophers
seem to like to do this. I don't know why. I like using floating
point numbers and intensities, and I am beginning to abhor binary
logic which allows nothing in between and ideas which try to
reduce everything to binary logic.
>There are also many intentional states which lack neural implementation,
>possibly, e.g., your belief that the floor is solid when you walk. So
>looking to the neurons is no help with these.
Ridiculous. My assumption that the floor is solid is based on years
of experience. My brain goes on autopilot when I walk since I don't
expect the ground to turn to melting ice. This is normal. Whenever someone
gets to be good at something, the conscious effort at doing it seems to
disappear. WE notice this because, we can think about something else
while doing it. You might be doing it without knowing it all the time
but as soon as something changes, your brain takes notice. If you
step out your front door one day and step into horse dung, I bet
the next day, your brain will sound the alarm when you step outside.
It only goes to show you that the brain/mind is doing its job
correctly (or efficiently) since if you had to decide what to
do every moment of your life, nothing would get done. It is the
things that get done by the brain/mind automatically that we call
subsconscious etc. The more we can do this, the more free resources
are available to the brain to think about other more novel things.
Ed Faith
Aaron Sloman wrote:
> For instance, if a RANDOM mutation occurs that enables an INDIVIDUAL
> to produce some new enzyme that helps it cope better in a given
> environment, then I can identify the function of the new enzyme by
> examining the individual in which it is produced, without assuming
> that it has already been selected by evolutionary pressures. They
> may not have had time to operate, given that this individual was the
> subject of the mutation.
(For some reason this just arrived at my news server, though it
should have arrived a couple of days ago.)
I think that the above is a good "test case" of different
notions of function. The response is so easy it surprises
me: I would guess that this new gene will end up serving
that function. So I'm making a guess about the future.
In the current situation, I'd just say that the creature
is lucky--same as I would be if I won the Megabucks lottery.
Anders N Weinstein
In article <435820...@luptonpj.demon.co.uk>,
Peter Lupton <lup...@luptonpj.demon.co.uk> wrote:
>
>Well, we might in theory - but surely not typically in practice. I would think
>that, in such cases, 'brain-reading' could be *extremely* effective at
>weeding out double agents, in comparison with other techniques.
I have not heard of anyone using this technique today. I think something
is amiss when a fantasy for which there is presently no support is taken
to be obviously possible. Wittgenstein for example wanted to alert us
to just this sort of "must" -- "it *must* be possible to read off thoughts from
a brain", "there *must* be laws of psychological phenomena just as there are
laws of physical phenomena" and so on.
I take it Wittgenstein didn't really oppose the idea that there *might*
turn out to be such correlations. He was more concerned to oppose what
he thought of as a misleading picture that says that there *must* be.
He wrote:
No supposition seems to me more natural than that there is no
process in the brain correlated with associating or with thinking;
so that it would be impossible to read off thought-processes from
brain-processes. ... It is thus perfectly possible that certain
psychological phenomena *cannot* be investigated physiologically,
because physiologically nothing corresponds to them. (Zettel 608, 609)
Now once one appreciates the outlook in his philosophy of mind,
this *is* a perfectly intelligible possibility. A somewhat
domesticated version is in Dennett's philosophy -- for Dennett,
intentional states are individuated by the intentional stance,
physiological states are individuated by the design or phyiscal
stances, and there's no particular reason the phenomena you reveal
through the former should have analogues in those revealed through the
latter. Could happen, but maybe not. (see Dennett "Brain-writing and
mind-reading" in Brainstorms; also, perhaps, Ryle on shamming in Concept
of Mind). If someone insists it *must* be the case or else some disaster
follows, that is a pretty good sign they are thinking of the mental in
the wrong way on this view.
The example Wittgenstein floated there raises an interesting question -
was he thinking about a case where there would be a violation of
physical law, a failure of physical science on its own terms? He *was*
perfectly happy to expect this possibility too. But I like to think he
was not comitting himself to it. It might be that the order revealed
through physical explanation might be gap free, yet still omit "the
order that is there when actions are done with intention" (Anscombe).
So I think physical law need not be violated for Wittgenstein's
possibility to come to pass.
BTW Dennett's paper "Real Patterns", in which Anscombe's remark is
cited, may help make clearer what it means to say that there is an
order revealed through the intentional stance that may be invisible
from the physical stance. Unfortunately, in my view, he tries to dress
this up with a mathematical notion of pattern as non-randomness -- you
are much more suited than me to critique that! I think one should take
the idea without the pseudo-technical dressing.
Anders N Weinstein
>
>They are different scale/level descriptions of the same reality.
>The "reality" of the philosophers of old (and maybe even new)
>is that which is observable/feelable via our naked senses. That
>keeps the very small and the very large out of our range/scale of
>observation, and hence gives rise to mystical philosophies that
>"reality" is not the same as "appearance". It is through instruments,
>and science that we can peer into the regions which our naked
>senses do not allow us.
True enough, I just think that psychological phenomena find their
actuality in the common-sense macro-world, not in an unseen micro-world
we must theorize about. Since I don't think of intentional states of
others as theoretical entities posited to give causal explanations, I
reject analogies drawn from theoretical sciences.
I don't reject theoretical sciences and am neither a mystic nor a
mysterian.
>In that sense, the study of the mind, has one advantage. Not only
>can we look at it the same way as into all things we cannot
>observe (i.e. the Black Box) but all of us has some kind of
>access into at least one representative of such a black box, so that
>it's not so black after all. We can almost have our cake and eat it
This is the error. I recommend against thinking of the person as a
black box when interested in mental states. And we don't investigate
their nature by giving our own case a special status.
>Looking into parts of a car, or one of the 5 million transistors
>of a CPu wouldn't tell you what it's supposed to do either if you
>did not already have some idea of what you were looking for. This
>is no different than what was just being discussed earlier, namely,
>that, knowledge/information is dependent on the observer. I don't
>see any mysticism here either because this is similar to the
>problem of value in economics.
Right. There is no mysticism here.
>There is no reason why intentional mind/brain states cannot be
>a part of a theory. One may include emotions here too. One must
>be a real purist to ignore the observed effects of drugs on
>emotions, cognition, and action. Why don't you then answer what
>kinds of magical powers do certain drugs have that they can
>somehow transcend the everyday common properties of the thing
>called brain? How about it? If the mind is something so separate
>and apart from the brain, then the drugs that effect the mind must
I nowhere said the mind is something so separate and apart from the
brain. I guess I can be classed a property dualist, but that's it.
One can't give an account of the nature of psychological
attributes in terms of the brain alone. This is perfectly compatible
with recognizing that brain states can have causal effects on mental
states. One might say that mental states are realized in neural
matter, without their individuation conditions being those of neuro-science.
>of experience. My brain goes on autopilot when I walk since I don't
>expect the ground to turn to melting ice. This is normal. Whenever someone
>gets to be good at something, the conscious effort at doing it seems to
>disappear. WE notice this because, we can think about something else
>while doing it. You might be doing it without knowing it all the time
>but as soon as something changes, your brain takes notice. If you
>step out your front door one day and step into horse dung, I bet
>the next day, your brain will sound the alarm when you step outside.
No doubt -- in *that* case there might be something in your brain
corresponding to the belief. But in the former case there need not be.
>It only goes to show you that the brain/mind is doing its job
>correctly (or efficiently) since if you had to decide what to
>do every moment of your life, nothing would get done. It is the
>things that get done by the brain/mind automatically that we call
>subsconscious etc. The more we can do this, the more free resources
>are available to the brain to think about other more novel things.
Maybe, but the point is that the mature circuitry supporting the
"walk" routine doesn't have to do *anything* to represent the belief
that the floor is solid, it just has to walk. So there need be
nothing occurrent in you corresponding to that belief. There does need
to be a complex of dispositions.
H. M. Hubey
ande...@pitt.edu (Anders N Weinstein) writes:
>>In that sense, the study of the mind, has one advantage. Not only
>>can we look at it the same way as into all things we cannot
>>observe (i.e. the Black Box) but all of us has some kind of
>>access into at least one representative of such a black box, so that
>>it's not so black after all. We can almost have our cake and eat it
>This is the error. I recommend against thinking of the person as a
>black box when interested in mental states. And we don't investigate
>their nature by giving our own case a special status.
I think we do both. The BB is everywhere in science. It seems to
be intimately connected with behaviorism in that we have a BB,
an input (Stimulus) and an output (Response), but it is used even
by that guy who allegedly skinned Skinner :-). What is the
competence-performance distinction that Chomsky makes if not the
Black Box? Performance is output/behavior, competence has to do
with something internal. This is the simple BB model. It is
everywhere. Even the introspecters like Freud, Reich, Jung, Adler
and anyone else did so by combining introspection with observation.
It was more like the stepwise refinement technique of programming
or the classical scientific paradigm. In that sense, the conclusions
reached by them actually reveal to us their own internal
problems because their output was a function of their own internal
mind/brain/personality/character and what they observed in other
people.
They, of course, did not apparently realize that they were
"projecting" their own characters into their theories but that
they were creating them purely out of their observations. We do
have some kind of access to our own minds, but that too is
colored by what we are so that the access is not noiseless, and
even if it were noiseless, it could only see "ourselves" and
not the parts that are shared by everyone.
>I nowhere said the mind is something so separate and apart from the
>brain. I guess I can be classed a property dualist, but that's it.
>One can't give an account of the nature of psychological
>attributes in terms of the brain alone. This is perfectly compatible
Neither can anyone else. There are too many neurons :-)..
But some (like me) extrapolate and think that the functioning
of the neurons (and the electro-chemical system) and our
life histories can explain it, and that the behavior we can
observe in small ANNs is sufficient to extrapolate to the human
brain so that we don't need anything else.
>Maybe, but the point is that the mature circuitry supporting the
>"walk" routine doesn't have to do *anything* to represent the belief
>that the floor is solid, it just has to walk. So there need be
OF course it does. IT's there entrenched so rock solid that it
does not even intrude into my consciousness unless and until
something goes wrong. It is only when something goes wrong that
we notice that we had false beliefs. It happens to people all
the time. I don't think we have to walk around mentally
going over all of our "beliefs" in order to have them. They are
there. If some people want to name them something else like
"subconscious beliefs", quasi-belief states, subcortical neuronal
belief sets, that's fine with me, but the idea of some kind of
an implicit belief is fine.
>nothing occurrent in you corresponding to that belief. There does need
>to be a complex of dispositions.
There certainly is always something there driving the motor neurons.
And if the motor neurons are sending signals to my legs they
are doing it because barring some other signal from some other
place, they expect a flat ground. If we are absentminded we might
not notice that there are holes on the ground but even those who
notice those things might of a particular temperement (developed
via experience) so as to expect something wrong.
Anders N Weinstein
In article <4qvi14$r...@sun4.bham.ac.uk>,
Aaron Sloman <a...@cs.bham.ac.uk> wrote:
>Sorry about the length of this, but to my surprise Anders brought in a
>lot of points that I thought were irrelevant to my attack on his claim
>that a brain in a vat could not be described using intentional language.
Sorry to be so irrelevant. It can be hard to respond usefully to
someone if one does not appreciate their motivations. Some
philosophers, John Searle, for example, think that every intentional
state has to be possible for a brain in a vat. Some philosophers of
cognitive science, such as Jerry Fodor at the time of "Methodological
Solipsism... " and later "Psycho-semantics" argued that scientific
psychology *must* restrict itself to individualitically determined
notions. (because, roughly, only such notions can capture local causal
powers of an object).
But let no one ever forget that "some" is not "all." Happily, Sloman is
not one of these all-or-nothing thinkers, and it was something of a
mistake for me to assimilate his view to an individualism like theirs.
I take it then that Sloman is *not* a blanket individualist
("methodological solipsist") about intentional states; nor does he base
his view on the supposed localism of causal powers. In some ways,
then, Sloman must be *much* more hospitable to my externalism about
intentional states than Fodor or Searle or whoever else -- I presume he
is happy to recognize that some, perhaps a very major portion of
intentional states simply could not be enjoyed by a brain in a vat. If
so, that is then a point of agreement which outweighs the disagrement.
But some is not all. Sloman is also keenly interested in the "some"
that in his view do *not* depend on context and could be enjoyed by a
brain or a control system or a computer disconnected from context. I
presume this could not include "playing chess" -- no ground for saying
the computer is playing chess as opposed to, e.g. plotting military
strategy for a Martian domain that's fortuitously isomorphic to the
chess world -- but I guess it could include " 'playing' a game
isomorphic to chess". I presume it could not include "proving a
geometry theorem" -- if it uses a standard first order logic, there
would be no intrinsic reason to say it's not proving theorems about
numbers or about its own syntax-- but I guess it could include "doing
something isomorphic to proving a geometry/number theory/syntactic
theorem".
These are interesting cases, I suppose, in the following sense. Even if
its designers thought it was merely generating symbol structures for no
purpose at all besides testing a programming language say, still its
output would be as good for us a proof of a theorem when we interpreted
it. If it was novel, we could publish it as a discovery say. Such is
one benefit of formalization.
But other than that, I am still puzzled about why one should be so
interested in isolating these domains. I suppose we already have a
language for talking about the system of structural relations that
might be common e.g. to Chess and Martian war -- it is the language of
mathematics, for example such things as abstract algebra and the theory
of relation structures. So why should we use *intentional* language for
this? When Deep Blue does its stuff, the only connection to real
*chess* comes from the purposes of the designers. When Gary Kasparov
does his, the connection does not so depend. Something is different,
something I would say depends essentially on the context.
On to some points of detail:
>I wasn't talking about social norms let alone trying to construct them
>out of anything individual. Obviously social norms require a social
>context. Nobody would dispute such a trivially true statement.
True, nobody disputes that. They do dispute what is required for a
social context to exist, and how it relates to individual mental
states. Searle is an individualist, as for example is Chomsky (about
the priority of internal languages over public) or David Lewis's
account of the nature of social convention. These thinkers believe
mental states are prior to and independent of social institutions. Some
social holists hold to the contrary that the wider institutional world
is involved in what it is to be a person or, perhaps, in what it is to
have an intentional state. They therefore think the institutional world
is prior to the individual mind -- it is the source of the intellectual
norms to which individual trains of reasoning must conform, say. On
this view the social context makes individual thought possible -- not
just causally, but ontologically, just as the rules of baseball are,
in Searle's terms "constitutive" not "regulative" -- they make baseball
possible, but don't cause the games.
Now I only bring this up to make externalism about mental states seem
more plausible to people who find it outrageous. If one accepts this
form of social holism, the failure of individualism about a large class
of mental states is pretty obvious. It's very hard to know what it means for
a brain in a vat to aspire to be, e.g. a knight errant in the absence of
the social world which makes this possible.
But you evidently are not one of those philosophers and do not think
externalism about a large class of intentional states is outrageous.
Nothing I could say along these lines could refute your narrow claims
about the move from "some" vs. "all" -- fine. But do you really think it
isn't philosophically interesting to explore the possibility of a deep
ontological relation -- not merely a causal one -- between individual
thinkings and the world of social institutions?
>So we have two sets of concepts
>
> believesC, thinksC, wantsC, intendsC (prsupposing Cultural links)
>
> believesW, thinksW, wantsW, intendsW (without cultural links)
>
>Apart from the extra cultural prerequisites in the first set, there's no
>difference. I explained how the second set could be useful in a context
>where the first set would be inapplicable.
I guess you can generate the second set by subtraction. I can agree
there might be some use for these contrived concepts, mainly in talking
about abilities to manipulate formal structures, I suppose. But so
what? Is that really all you want? Aren't you interested in the nature
of mental states too?
>(b) I wasn't talking about a contrast between what is implicit and
>explicit, only about a contrast between what is and what is not
>culturally (or biologically) produced. What is implicit e.g.
>implicit in a neural net, or implicit in a complex collection of
>(internal)condition-(internal)action rules) need not have any
>cultural connection.
>
>The "*W" concepts may include all sorts of skills, compiled
>procedures, neural implementations, etc. without having explicit
>theories or representations. Or they may include explicit theories
>or representations. The two distinctions *C/*W and implicit/explicit are
>orthogonal.
You are right that the distinctions are orthogonal. My thought,
however, is, first, that what is merely implicit in a device simply
cannot be identified apart from the function of the device in a
super-individual context. Admittedly, this context may be social or it
may be simply the material environment. Either way, I take it you
cannot perform your imagined abstraction for what is merely implicit --
you can't detach the attribution of the merely implicit from the
context and find it in a detached device if the device only takes on
this "implicit" knowledge insofar as it is deployed its context, not
considered in itself.
Example: someone might say that the presupposition that most motions of
objects are rigid is implicit in the operation of my visual system.
Since this assumption need have no "control system reality", it is not
something you could carve off and ascribe to a computational device
that is not functioning as a visual system for a robot or organism in a
particular material environment.
So, wherever there is merely implicit representation like this, then, I
would think there is no prospect of forming one of your *W predicates.
If much of our understanding is implicit like this, then much of it
can't be translated into decontextualized *W predicates. It is, as it
were, a complete non-entity at the control system level, though visible by
taking into account the larger context.
Now what does this have to do with the social? Well, if our capacity
for conforming to social norms is largely implicit like this then we
have another avenue for the critique of methodological individualism
about the social -- if you can't find the mastery of the social world
represented inside the individual, except in context-dependent
attributions, then there is no basis for supposing individual states
are prior to and independent of the social context.
Again, to be sure, some does not imply all, but this seems to me still
to be an interesting position to consider with regard to the
metaphysics of the social world. Off your narrow point, I guess, but
always in the back of my mind, in view of, say, Dreyfus' Heidegger.
That view has it, for example, that what it is to be a person is
primarily a matter of tacit conformity to social norms, from which you
inherit your initial understanding of the world and yourself. Then the
phenomena of exercising such understanding and perhaps changing it is
only visible by reference to a super-individual context. Isn't this a
possible view of what it is to have intentional states?
>Either way, the *W states don't have to have cultural antecedents in
>order to exist NOW implemented in a brain or brain-like machine, in
>which they produce processes of the sort I was talking about (e.g. a
>system passionately seeking a more elegant proof of some theorem, or
I don't understand how to find passion apart from the characteristic
human expressions of passion or some analogue. This is why sci-fi must
resort to blinking lights and even having smoke come out of the computers.
I guess it's OK with me to say a machine is doing something isomorphic
to proving a theorem.
>[AW]
>> Bit of a fudge. I didn't say it couldn't 'invent' a game structurally
>> equivalent to present-day chess. Of course it could. It could do lots
>> of things that are "structurally equivalent" to thinking, reasoning,
>> etc.
>
>But you apparently don't want to use the words "thinking", "reasoning",
>"wanting", and the like, where there are no links to a culture.
There are actually several reasons for that. Mainly it's because I
think these terms are norm-involving -- thinking or reasoning can be
correct or incorrect, for example, and one wants to know where the
norms come from.
One answer might be: the individual derives them by
internalizing them from norms that are, in some impersonal way, abroad
in the culture. And that intentional explanation as it were plots
individual reasonings against this system of norms (compare Burge's
analogy on the role of a score in assessing musical performances).
You might be interested in the opening chapters of Bob Brandom's
_Making it Explicit_ for a nice statement of this view. Brandom takes
it that an intentional state just *is* a social status like having
promised -- a bit too strong, it seems to me, but salutory to
contemplate. Certainly if you read this work you will understand why
some Wittgensteinians say what they do, or why one would deny that
mental states are control system states.
The second point is that machines conform to norms in a different way
the people do. Machines malfunction, but people make mistakes; people can
manifest an ability to take responsibility for conforming their thought
to a norm in a way that machines can't. People are different sorts of
"loci of responsibility" than automatic machinery. The features
that support this attribution -- the facial expressions, for example, of
someone working on a proof -- themselves lie outside the level at which
we talk about person-level rule-following, but they still are necessary to
support that language.
What, if anything, happens inside people's brains is as may be, it
seems to me, since reason-giving explanations make no committments
regarding that. Our heads could be empty -- not that they are, but
they could be -- as far as the normalizing function of reason-giving
explanation is concerned.
I can concede that a human artifact might be able to acquire and
manifest conformity to the intellectual norms abroad in a culture, and
so be accepted as a person. But this seems to me very different then
what computers or brains or control systems do, in virtue of the
different sort of responsibility at issue.
>But requiring cultural links as preconditions for ALL intentional
>descriptions is just a semantic decision: i.e. a refusal to use a
>certain set of concepts (the *W concepts).
Again this presupposes that the *W concepts are clear and well-defined,
which is really the point at issue. And even if you can generate some
derivative *W predicates by subtraction from predicates ascribing real
mentation, what follows about the nature of real mentation?
>[AW]
>> Of course we can study purely formal structures all we want,
>> and even embody some of them in highly useful mechanical calculators.
>
>Er, .... why are you talking about formal structures and mechanical
>calculators? I said nothing about either.
Sorry about the broad use of "mechanical", although I hardly think it's
without precedent. I take a computer to be an automatic calculator and
am concerned to oppose the computational theory of the nature of
intentional states. You are however evidently interested in a wider
and somewhat indeterminate class of mechanisms, including anything that
might go on inside a human brain when we think and act rationally.
Fair enough, I will not ascribe computationalism to you.
Now what do I think is the importance of the distinction between the
explicit and conscious and the implicit and practical understanding in
skills?
I inferred that the sorts of domains for which you might be able to
abstract your *W predicates from the wider context are mainly the
highly formal and digital ones like chess or go or theorem-proving.
These are the ones that seem to be applicable even to a disembodied
intellect engaging in conscious and explicit chains of reasoning. Not
such things as knowing how to catch a fly ball, for example. In the
latter case, the neural circuitry itself cannot be said to support
fly-ball catching apart from its function in the larger context.
Admittedly this depends only on the body and material environment and
not the culture, but that is still a form of externalism.
So: if one focusses on the former, one tends to think of the mental as
the domain of a disembodied reasoning engine, it seems to me, and so to
think that a computer as automatic calculator might thereby reproduce
the essence of intelligent activity. If one focuses instead on the
predicates ascribing potentialities for engaging in skillful
interactive activity in a material and social environment, then one can
say instead: while there is control circuitry supporting these
potentialities, the portion of the causal loops that go through these
neural circuits need not form a detachable, autonomous domain of study
-- not in the way that purely intellectual activities might be.
Briefly, then, while you might be able to construct
"proving-a-theorem-W", I can't see any use at all for
"racing-to-catch-a-fly-ball-W", or even
"knows-how-to-catch-a-fly-ball-W" and the like. There is a
control-system story for these activities, sure, but not an intentional
one and not one that applies to brains in vats. And if I think the
latter sort are more important to what it is to be a person, I will not
be so excited by the idea of contriving a language for talking about
the powers of control systems apart from the context in which they
function.
Again, I guess this is not deductively relevant to your narrow "some"
vs. "all" point, but it does indicate where I think one should be
looking in order to better understand human understanding. Not to the
causal machinery inside an individual brain -- an implementaion level
-- but to the material and social environment with which the circuitry
has as its function to engage.
>The nature of a game or task does not constrain the nature of the
>mechanisms capable of playing the game or performing the task.
A bit too strong, I would have thought. Surely the nature of the
task puts very strong constraints on the mechanisms that can subserve
engaging in it. For example, if you can't process natural language
without recursion, say.
>You are saying that a certain set of concepts has no use unless
>combined with extra conditions referring to cultural precursors,
>whereas I am saying they are potentially useful when they able me to
>talk about processes in objects with no cultural (or biological)
>causes, without my having to invent some elaborate new family of
>concepts similar to the old ones except for the cultural baggage.
Not exactly. I would suggest that the function of ordinary intentional
explanation is usually not to map inner causal antecedents of action
but rather to assess the activity in the light of (what we take to be)
norms of rational conduct. If you want to take a set of concepts which
has its main function in this application to living human beings and
adapt it to invent a bunch of variant concepts for use applied to
disembodied circuitry or to brains, be my guest! I would never suggest
you couldn't do it, although it hardly seems necessary to me.
>As I did: if I were not allowed to use the existing concepts to describe
>processes in the brain in the vat or in the disconnected mathematician,
I have never seen a (living) brain in a vat nor a disconnected
mathematician, so I suspect we have little need for talking about these
far-fetched eventualities. Are you really engaged in some pursuit that
requires a language for talking about these things? I may well be
wrong, but I suspect that is a deception designed to cover up a
fundamental individualist prejudice.
>> At a minimum, I would take myself to be trying to make explicit the
>> ordinary psychological concepts we actually use.
> ^^^^
>Who is "we"? Speak for yourself. Mine are different, as should be
>very clear from this discussion. But feel free to go on using yours,
>if that's what you wish.
I expect (with Ryle and Wittgenstein) that you use psychological language
correctly when you are not theorizing about it. The goal of their sort
of philosophy is to get one to acknowledge the concepts one actually
operates with when not corrupted by theory. Admittedly I can not
put in the hard work needed to achieve this.
>On the other hand none of this discussion seems to me to shed any light
>on whether a sufficiently sophisticated computer-based architecture (not
>just "a program" !) might or might not have so much in common with
>what goes on in human brains that refusing to use common mentalistic
>language to describe it is about as silly as trying to run a race with
>your feet tied together.
I don't think the paradigmatic uses of psychological vocabulary makes
reference to states and events inside of brains; rather to conduct seen
in the light of an interest in normative assessment.
>[AW]
>> To take one example, consider that great mystery about how the
>> brain could, all by itself, radiate consciousness or
>> intentionality.
>
>This is just a silly question. I have no reason to believe my brain
>radiates any such thing, either all by itself or with help. I am not
>aware of any "great mystery".
Good for you. But do I have to list the works published by otherwise
very intelligent people that are concerned with the question as to
whether science can solve the great mystery of consciousness? To take
one timely example, Chalmers' book seems to be getting some play in the
press. (Don't ask me to explain why.)
>Of course, there are lots and lots of mysteries about our brains: how
>they process various kinds of information, how they store information,
>how they transform and interpret information (e.g. in using samples from
>an optic array to arrive at detailed information about the structures,
>properties and relationships of many objects both near and far), how
>they generate new motives, how they select between conflicting motives,
>how they sometimes control their thought processes and sometimes lose
>control, how they learn about infinite structures, how they control the
>hands of a concert pianist and many more.
>But these are mysteries to be tackled by science, not philosophy,
>least of all philosphical stipulations apriori that cultures have to
>be involved in all cases.
You may be right about a priori stipulations -- I don't make them. But
I don't say "maybe the number two is thinking" either. Some apparent
claims are really nonsense or confusion which call for therapy.
I disagree about the role of science here. I don't regard the illusory
mystery about consciousness or intentionality as a mystery about brain
function. I think there is nothing we don't already know about
intentional states that could solve those difficulties; rather, we just
need to understand the concepts we already have and operate with
correctly a zillion times a day.
Just to take one example, the fact that there are natural expressions
of mental states and that mental states can often be read directly in
people's faces is, for Wittgenstein, absolutely crucial to the
possibility of psychological language. If you think instead of
mentality as a property of inner control systems, you are already
looking in the wrong place, it seems to me; the decisive move in the
conjuring trick has been made. One is never going to find any role for
the concept of expression that is so crucial on Wittgenstein's view.
Once you see that what we consider the mental is only intelligible if
situated in its proper context -- that of an embodied human life in a
material and social environment -- the apparent mystery dissolves, or
so it seems to me.
David Longley
In article <4r9sug$8...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
Why the change from "solipsist" to "individualist"? The history of this
position in the present intellectual context goes back at least as far
as Carnap's "Aufbau" and perhaps even Goodman's "Ways of Worldmaking".
However, as I'm sure you know, Carnap gave up on this essentially
"phenomenalist" approach for what he called "physicalism", but which
you might call "externalism". But once embraces that, *and* give up on
the analytic-synthetic distinction, one might as well accept that one's
in Quine's extensionlist/behaviourist/logical empiricist/naturalist camp
(as I have tried to point out elsewhere).
One of the historical ironies is that the "positivists" generally gave up
Machian "positivism" for logical empiricism, very early on. Probably as
far back as 1928. Even Ayer was critical of "Positivism". One has to see
Quine's work as revisionary - and this is Quin's own view.
As to Sloman's "quantification" joke.......... One just has to ask whether
one is being methodologically "individualistic" or "externalist" here <g>.
There may well be "some" who claim to be "cognitive scientists" but not
subscribe to the position articulated by Fodor, but then, maybe they just
don't appreciate the force of what he presented.
I've yet to meet a "cognitive scientist" who does not (in terms of what
they actually say they subscribe to) fall under the blanket framework
outlined by Fodor. Some are so confused they don't know what they subscribe
to.
If Sloman's position translates into one having to accept that there may
be as many cognitive sciences as there are cognitive scientists, I think
I can rest my original case. Fortunately, he has computer science to turn
to, which is altogether uncontroversial in this sense, and of indisputable
scientific and technological value.
--
David Longley
Anders N Weinstein
In article <836310...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>
>Why the change from "solipsist" to "individualist"? The history of this
"Individualism" is a label used in the writings of Tyler Burge. His
classic article is "Individualism and the Mental"; one pertaining to
cognitive science is called "Individualism and Psychology". I can dig
up exact citations if you wish.
Burge took some trouble to try to define "individualism" so it would
apply both to materialists and non-materialists like Descartes. If we
stick to the materialist sort of version, individualism might have it
roughly that the nature and individuation of mental (intentional)
states depends solely on local "intrinsic" features of the brain.
According to an individualist, if two brains are structurally congruent
then there could not be a psycholological (intentional) difference
between the mental states of the subjects who possess them.
The opposite is usually called "externalism". It has been
supported by Hilary Putnam's Twin Earth cases and by Burge's thought
experiments. Other philosophers like Dretske who cite evolutionary history
as determinative of intentional content have a different set of arguments.
Others follow Wittgenstein, who is often taken to have thought of meanings
as rules for the use of expressions and to have shown that rule-following
itself can only exist in a social context.
I believe "individualism" is basically the same as "methodological
solipsism", although Jerry Fodor once attempted to distinguish the two
(in Psychosemantics).
It is not clear to me Quine is not a methodological solipsist at
heart. I know of nowhere where he addresses the Twin Earth cases --
they would seem to refute his theory of stimulus meaning. If one is
stalking about a system of dispositions to assent to uninterpreted
strings whose boundary conditions are nerve endings, then these would
be exactly the same in the Earth and Twin-Earth cases. But if Putnam
and Burge are right, the meanings of their words would be different;
and so would the truth conditions of their mental states -- because
intentional states are individuated by worldly conditions of
satisfaction, the two subjects would be thinking different thoughts.
I would think the issue only arises for philosophers who don't think
there's any problem about intentional vocabulary. If you avoid the notion
of meaning or the mental altogether of course it won't arise.
Anders N Weinstein
>does not even intrude into my consciousness unless and until
>something goes wrong. It is only when something goes wrong that
>we notice that we had false beliefs. It happens to people all
It's only when things go wrong that a representation of that content
needs to be constructed.
>the time. I don't think we have to walk around mentally
>going over all of our "beliefs" in order to have them. They are
>there. If some people want to name them something else like
>"subconscious beliefs", quasi-belief states, subcortical neuronal
>belief sets, that's fine with me, but the idea of some kind of
>an implicit belief is fine.
We can be said to have the belief -- true. But there need not be
any inner vehicle which bears that content. Perhaps you can say
that there is an absence or lack of any of the myriad events
that might inhibit the normal walking. But this loose and open-ended
condition does not define an inner actuality.
The only point is that the language of belief comes apart from the
language of control system representations.
>There certainly is always something there driving the motor neurons.
>And if the motor neurons are sending signals to my legs they
>are doing it because barring some other signal from some other
>place, they expect a flat ground. If we are absentminded we might
This sort of "expecting" need only consist in being adapted to couple
with a ground that is solid.
David Longley
In article <4rbm0g$e...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> In article <836310...@longley.demon.co.uk>,
> David Longley <Da...@longley.demon.co.uk> wrote:
> >
> >Why the change from "solipsist" to "individualist"? The history of this
>
> "Individualism" is a label used in the writings of Tyler Burge. His
> classic article is "Individualism and the Mental"; one pertaining to
> cognitive science is called "Individualism and Psychology". I can dig
> up exact citations if you wish.
Thnaks for that. I have the latter in Rosenthal's anthology "The Nature
of Mind" but haven't read it. From a quick glance it seems to cover some
of the same ground as Fodor's paper (which it antecedes by a year I see).
Fodor's paper was really a summary of a position taken by *many* working
in areas like cognitive psychology and personality and social psychology.
It is therefore quite a sound summary of what many researchers have been
studying within "Attribution Theory" dating form the early 50s and before.
As such it is not too distasteful for psychologists to read. To find a
philosopher (Burge) just arguing the point without reference even indirectly
to the literature (which *dominates journals like "Personality and Social
Psychology") is another matter.
<snip>
>
> It is not clear to me Quine is not a methodological solipsist at
> heart. I know of nowhere where he addresses the Twin Earth cases --
> they would seem to refute his theory of stimulus meaning. If one is
> stalking about a system of dispositions to assent to uninterpreted
> strings whose boundary conditions are nerve endings, then these would
> be exactly the same in the Earth and Twin-Earth cases. But if Putnam
> and Burge are right, the meanings of their words would be different;
> and so would the truth conditions of their mental states -- because
> intentional states are individuated by worldly conditions of
> satisfaction, the two subjects would be thinking different thoughts.
>
I take Putnam's position to be far more sympathetic to Quine's than you
make out. In fact, it is hard to read "Representation and Reality" and
not come away with that. The basic idea is that meaning is not inside
the head.
To suggest that Quine is a methodological solipsist is somewhat bizarre.
> I would think the issue only arises for philosophers who don't think
> there's any problem about intentional vocabulary. If you avoid the notion
> of meaning or the mental altogether of course it won't arise.
>
Well, such work is often cited in support of the thesis that all is not
well in intentional (cognitive) psychology - that's true...(see "Fragments"
parts 2 and 3).
--
David Longley
Anders N Weinstein
In article <836331...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
[Re Burge's "Individualism and the Mental" and Fodor on "Meth. solipsism"]
>Thnaks for that. I have the latter in Rosenthal's anthology "The Nature
>of Mind" but haven't read it. From a quick glance it seems to cover some
>of the same ground as Fodor's paper (which it antecedes by a year I see).
>Fodor's paper was really a summary of a position taken by *many* working
>in areas like cognitive psychology and personality and social psychology.
Not exactly. Fodor's view is quirky: he was and is concerned with a theory
of inner *computational* states and was exercised in that early paper
by the "Formality Condition". This presupposes a *syntactic* view
of cognitive processes, and that is completely foreign to, say, the
phenomenological tradition. Husserl's noemata, like Frege's senses or
John Searle's intentional contents are *semantic*, not formal. Again,
phenomenology is concerned with how things seem from the point of view
of a subject, but that also is a semantic, not a syntactic matter.
It might be right to say: if one assumes the bracketing (methodological
solipsism), then one destroys the very possibility of making sense of
intentionality or semantics at all, and the concept of thought as something
which answers for its correctness to a world outside of it will
degenerate into a blind, purely syntactic affair.
For example, many devotees of the computational theory get themselves
into a bind over how semantics or reference could matter at all, since
it would be wholly extrinsic to the blind formal manipulations they
posit. For the phenomenologists, this is a pseudo-problem, since
thought and reasoning are semantic from the get-go, and need not have a
syntax.
>It is therefore quite a sound summary of what many researchers have been
>studying within "Attribution Theory" dating form the early 50s and before.
>As such it is not too distasteful for psychologists to read. To find a
>philosopher (Burge) just arguing the point without reference even indirectly
>to the literature (which *dominates journals like "Personality and Social
>Psychology") is another matter.
I don't think you appreciate: Burge *denied* that cognitive science is
"methodologically solipsist". In particular, he *denied* that a theory
of vision in the style of David Marr's book was methodologically
solipisist. So he disagrees with Fodor on methodology of cognitive science.
I don't see that work about "attribution theory" is relevant to Burge's
point. I can't see that Burge thinks attributions of externally determined
propositional attitudes are typically errors or projections.
>> It is not clear to me Quine is not a methodological solipsist at
>> heart.
>
>I take Putnam's position to be far more sympathetic to Quine's than you
>make out. In fact, it is hard to read "Representation and Reality" and
>not come away with that. The basic idea is that meaning is not inside
>the head.
Putnam rejects the indeterminacy of translation and indeed takes
it to be a reductio ad absurdum of Quine's view of meaning.
>To suggest that Quine is a methodological solipsist is somewhat bizarre.
Not at all. Quine has made some scattered remarks on skepticism about
the external world. His idea is that it makes no difference because the
function of theory with its reifying devices -- its role as predictor
of future stimuli -- would be unaffected by a grand and ingenious
permutation of all the referents, as effected by proxy functions.
So it doesn't matter.
Quine's view clearly has it that the language of a brain in a vat will
have the same stimulus meanings as that of an embodied speaker on the
Earth or of one on Twin-Earth. This is really a view of meaning according
to which whatever -- if anything -- lies on the far side of
the nerve endings is irrelevant to the cognitive functioning of the epistemic
engine. And that's basically a neurologized version of the bracketing,
a methodologically solpisist view.
H. M. Hubey
ande...@pitt.edu (Anders N Weinstein) writes:
>We can be said to have the belief -- true. But there need not be
>any inner vehicle which bears that content. Perhaps you can say
You are losing me. There is something in behavior which seems
to need a name. That something in behavior we have called
belief. There is something inside the brain that causes such
behavior (unless behavior is random and not deterministic/causal).
We have agreed that at the microscopic level they are connected
with the firings of neurons. At middle levels we can talk about
firings of masses of neurons, and as we go up the hierarchy we
have more and more correlated firings of masses of neurons. It
must be so since something as simple as walking requires
correlated firings of masses of neurons controlling the legs,
the arms, the body, etc. If the legs don't move as if they
are expecting a drop (as either taking a large step to avoid
a hole or taking steps downwards or upwards to go down/up
stairs) then the masses of neurons that control these legs
are behaving as if they are expecting a flat/level ground. By
using the word "expect" I mean that there are no neuron firings
coming from someplace in the brain that tells the legs to lift
higher, or lower or take a large step to avoid a hole.
We agreed to give the behavioral manifestation of this mass firing
process the name "belief". How can there be no "inner vehicle" that "bears
that content"? We just gave the description of that "inner vehicle
that bears that content." The only difference is that we skipped
the neural firings business (at least I did) because it seemed
obvious that names like propositional attitudes, intentional
stances, beliefs are names for such mass firings; in other
words, they are names for upper level macroscopic phenomena
(in the neuron world of the brain).
That whole point was that large numbers of neurons acting in
concert in some way can be given descriptions in terms of
behavioral (or even introspective) terminology. The whole thing
is an exercise in phenomena that occur at different scales. I gave
examples before from gases. Here's one from solids. The fact is
that most solids we deal with do not have completely crystalline
structures, for example steel, bronze, etc etc. Things like glass
are not crystalline at all; glass is a supercooled liquid. In any
case, if you look at a specimen of steel or brass with a miscroscope
(flattening some section, and then polishing it smooth) you can
see large areas/sections which differ from each other. This is still
at a level much larger than in which we'd see molecules but
smaller than can be seen by the naked eye. The large areas (forgot
the name for them) are clumps of mixtures of the things that we add
(such as carbon or nickel) to create alloys (such as steel) with
the iron. And there are different ways in which these mixtures form
depending on the temperature to which they were heated and the
rate at which they were cooled. So there are areas (invisible to
the naked eye) in which different quasi-crystalline structures
might have formed (and they have names too, like austenite).So we have
layers like subatomic particles, atoms, molecules, larger groupings
(whose names I can't remember), and then we can finally have
lumped parameter systems like I-beam, cylinder, angle-iron etc.
The material is describable at various levels and to various
degrees of accuracy.
So too with the brain. It's a mistake to jump from firings
of a few neurons to the whole brain or mind, and ignore
everything in between. Much of the discussion taking place
now on consciousness etc. is about the functions of various
parts of the brain.
It is at this level that the connections between the
physical brain, and the upper-level macroscopic, introspective,
behavioral descriptions such as goals, motivations, needs,
drives, beliefs, intents, etc takes place. How can it be
ignored? This is the most exciting part. It is here that
the discussion of the brain starts to impinge upon and start
to answer/resolve the philosophical/psychological/linguistic/logical
questions/dilemmas/conundrums. It's a bad mistake to ignore it
or try to sweep it under the rug. This is where "it's all happening."
>that there is an absence or lack of any of the myriad events
>that might inhibit the normal walking. But this loose and open-ended
>condition does not define an inner actuality.
In light of what I wrote above, this makes no sense.
>The only point is that the language of belief comes apart from the
>language of control system representations.
I don't know if you are really referring to "control theory"
or making a general remark of some sort.
Anders N Weinstein
>are expecting a drop (as either taking a large step to avoid
>a hole or taking steps downwards or upwards to go down/up
>stairs) then the masses of neurons that control these legs
>are behaving as if they are expecting a flat/level ground. By
True. But if I asked you: point out the inner state that represents
the proposition "the ground is solid and level". Tell me for example what
sort of representational system this proposition is encoded --
English? Predicate calculus? a vector of activation weights?
you might not be able to give any answer. You might have to say:
its not represented in any system, rather its merely implicit in the fact
that, as you rightly say, the circuitry is moving the legs in a way
appropriate to flat solid ground, which it is wired up to do by default,
in the absence of inhibiting inputs.
>using the word "expect" I mean that there are no neuron firings
>coming from someplace in the brain that tells the legs to lift
>higher, or lower or take a large step to avoid a hole.
Again you are right that no such signals are coming. But can you say:
the *absence* of such an inhibiting signal on a nerve fiber is itself the
representation of the belief that the floor is solid? There might be
many reasons such an inhibiting signal might arise.
A possible view is that I have a higher level belief store which
contains a world model which has as its consequences that the ground
is flat and level. I think this misrepresents the nature of default
assumptions -- I think we often act first and articulate later.
So an explicit representation of the assumption need not exist anywhere
before some sort of observably fishy circumstance raises a doubt about the
solidity of the floor and we hold up.
>That whole point was that large numbers of neurons acting in
>concert in some way can be given descriptions in terms of
>behavioral (or even introspective) terminology. The whole thing
Fine. But why not then suppose that belief is not a phenomenon at the
level of mass neuron firings, but at an even bigger level, the level of
molar activity of a whole organism in its normal habitat (in which, say,
the ground is usually solid)? Strictly, it is not the sensori-motor
routines that belief the ground is solid after all -- they can't
do anything when doubt arises.
>>The only point is that the language of belief comes apart from the
>>language of control system representations.
>
>I don't know if you are really referring to "control theory"
>or making a general remark of some sort.
You might try looking into some of Dan Dennett's essays, say "Brain
Writing and Mind Reading" in his _Brainstorms_ to get a better feel
for what I am trying to illustrate as the looseness of the relation
between the language of belief and the language of neural
representations.
H. M. Hubey
ande...@pitt.edu (Anders N Weinstein) writes:
>True. But if I asked you: point out the inner state that represents
>the proposition "the ground is solid and level". Tell me for example what
>sort of representational system this proposition is encoded --
>English? Predicate calculus? a vector of activation weights?
>you might not be able to give any answer. You might have to say:
>its not represented in any system, rather its merely implicit in the fact
>that, as you rightly say, the circuitry is moving the legs in a way
>appropriate to flat solid ground, which it is wired up to do by default,
>in the absence of inhibiting inputs.
Sure it's encoded/represented and it is done so in some of the
100 billion neurons we have. OF this part I am sure, however
if you ask me to point out which neurons are doing it, I can
say flatly that it's a pointless question. It's like
asking which atoms/molecules of a gas are striking the
walls of a pressure meter that I inserted into a container
of gas. HOw is it so important? That the atoms' motions
cause the pressure readings is known (according to physics)
but pressure measurement is a macroscopic measurement,
and it is dependent on the velocities of the atoms in
the gas.
>Again you are right that no such signals are coming. But can you say:
>the *absence* of such an inhibiting signal on a nerve fiber is itself the
>representation of the belief that the floor is solid? There might be
>many reasons such an inhibiting signal might arise.
So what? I was, in fact, making the reverse arguement, namely that
we'd walk differently if we did suspect that the ground had
some unsuspected traps, say, as in some construction site at which
you can expect all sorts of objects all over the place. In order
to change behavior appropriate to that situation, we'd have to
either believe (be told) that such action is necessary or we'd
learn it the hard way by stepping into a hole or stepping on
something that caused problems.
The important thing is that there are expectations all the
time. IF there is a sameness (invariance) in our environment
we won't spend any time thinking (consciousness again) about it
and it would seem as if we are on auto-pilot. I've done this
many times. ONe day I showed up at work, and didn't realize it
was Saturday until I saw that the parking lot was empty. Most
of the time, when I arrive at work, I don't remember much
about the road because I am usually thinking about something
else, and driving is more like walking (i.e. automatic).
>A possible view is that I have a higher level belief store which
>contains a world model which has as its consequences that the ground
>is flat and level. I think this misrepresents the nature of default
>assumptions -- I think we often act first and articulate later.
IT doesn't matter. It's not necessary to consciously think about
what beliefs we have in order to have them. We learned about the
world, and learned to speak a language without even being
aware (consciously, again) of what we were doing. The neurons
and the inputs (stimulus) took care of all that.
>Fine. But why not then suppose that belief is not a phenomenon at the
>level of mass neuron firings, but at an even bigger level, the level of
>molar activity of a whole organism in its normal habitat (in which, say,
>the ground is usually solid)? Strictly, it is not the sensori-motor
>routines that belief the ground is solid after all -- they can't
>do anything when doubt arises.
We don't have any evidence that any other organs of our body
does anything but obey the signals that come from the brain.
If there is belief it's in the brain.
>You might try looking into some of Dan Dennett's essays, say "Brain
>Writing and Mind Reading" in his _Brainstorms_ to get a better feel
>for what I am trying to illustrate as the looseness of the relation
>between the language of belief and the language of neural
>representations.
I never claimed that I could point out a particular set of
neurons or a particular sequence of firings which would correspond
to various beliefs. Neither do I think it necessary to be able
to discuss it.
If someone punches you in the face, we know that none of the
molecules of his hand touched the molecules of your face. It's
impossible, since that would mean that the electrons of those
cells/molecules would have to have been stripped away, etc.
And yet, the nose will probably bleed just the same, and it
won't make any difference if you can't specify which of your
cells have broken and which of your veins have broken.
It is not even an appropriate question to ask at this level.
At this level we are talking about essentiall the lumped
parameter systems of physics so the objects and actions must
be appropriate to this level.
David Longley
In article <4rc9b7$h...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> Not exactly. Fodor's view is quirky: he was and is concerned with a theory
> of inner *computational* states and was exercised in that early paper
> by the "Formality Condition". This presupposes a *syntactic* view
> of cognitive processes, and that is completely foreign to, say, the
> phenomenological tradition. Husserl's noemata, like Frege's senses or
> John Searle's intentional contents are *semantic*, not formal. Again,
> phenomenology is concerned with how things seem from the point of view
> of a subject, but that also is a semantic, not a syntactic matter.
Fodor passed the buck with respect to a clear explication of what he meant
by the formality condition, and I must say that there are other points in
that paper where I just don't understand what he was saying. I provided a
summary of parts of it (in "Fragments" as I made sense of it) but am very
open to having it better explained. It was Geach's criticism which made
direct contact with my immediate area of work, not only his choice of
example, but his concise summary. Even though he ridicules the stance, I
think it is a fair summary of the naive psychological stance, ie the one
which empirical psychologists endeavour to provide a description of in
Personality & Social psychology. It is also the position assumed in most
of cognitive psychology.
At the time I first read it I was very much engrossed in Husserl and
Heidegger, without any professional experience as an applied psychologist.
It's only on the basis of the latter experience that it begins to look
inadequate (as a descriptive enterprise and as an account of how folk
psychology operates I think it fits the other data very well).
Do you think I have made the latter concern sufficiently clear (or is it
perhaps something which is only likely to be readily appreciated if one
does work in the field I have been pre-occupied with for the past 12 years?
I wonder...
<DL>
> >It is therefore quite a sound summary of what many researchers have been
> >studying within "Attribution Theory" dating form the early 50s and before.
> >As such it is not too distasteful for psychologists to read. To find a
> >philosopher (Burge) just arguing the point without reference even indirectly
> >to the literature (which *dominates journals like "Personality and Social
> >Psychology") is another matter.
>
> I don't think you appreciate: Burge *denied* that cognitive science is
> "methodologically solipsist". In particular, he *denied* that a theory
> of vision in the style of David Marr's book was methodologically
> solipisist. So he disagrees with Fodor on methodology of cognitive science.
>
I'll read it and come back to you on this.
> I don't see that work about "attribution theory" is relevant to Burge's
> point. I can't see that Burge thinks attributions of externally determined
> propositional attitudes are typically errors or projections.
>
Can you clarify this point? What is Burge's position re the propositional
attitudes?
> >> It is not clear to me Quine is not a methodological solipsist at
> >> heart.
> >
> >I take Putnam's position to be far more sympathetic to Quine's than you
> >make out. In fact, it is hard to read "Representation and Reality" and
> >not come away with that. The basic idea is that meaning is not inside
> >the head.
>
> Putnam rejects the indeterminacy of translation and indeed takes
> it to be a reductio ad absurdum of Quine's view of meaning.
>
What text are you implictly refering to when you make this claim?
> >To suggest that Quine is a methodological solipsist is somewhat bizarre.
>
> Not at all. Quine has made some scattered remarks on skepticism about
> the external world. His idea is that it makes no difference because the
> function of theory with its reifying devices -- its role as predictor
> of future stimuli -- would be unaffected by a grand and ingenious
> permutation of all the referents, as effected by proxy functions.
> So it doesn't matter.
>
I think you may be trading ease of reading for technical coherence here.
Ontological Relativity really comes down to "A rose by any other name.."
does it not?
> Quine's view clearly has it that the language of a brain in a vat will
> have the same stimulus meanings as that of an embodied speaker on the
> Earth or of one on Twin-Earth. This is really a view of meaning according
> to which whatever -- if anything -- lies on the far side of
> the nerve endings is irrelevant to the cognitive functioning of the epistemic
> engine. And that's basically a neurologized version of the bracketing,
> a methodologically solpisist view.
>
Bracketing is a concrete decision not to make use of knowledge, mainly
scientific knowledge.
Again, which text are you refering to? I haven't seen him refer to either
of the above thought experiements.
Which side is the "far side" here? What Quine thinks immaterial is the
internal details. What *is* important is what is public. The externalist
perspective is fundamental to Quine's break with phenomenalistic (Machian)
empiricism.
Are you sure you haven't made a major mistake here?
--
David Longley
Peter_...@uk.ibm.com
In <4r9g1r$7...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>In article <435820...@luptonpj.demon.co.uk>,
>Peter Lupton <lup...@luptonpj.demon.co.uk> wrote:
>>
>>Well, we might in theory - but surely not typically in practice. I would think
>>that, in such cases, 'brain-reading' could be *extremely* effective at
>>weeding out double agents, in comparison with other techniques.
>
>I have not heard of anyone using this technique today. I think something
>is amiss when a fantasy for which there is presently no support is taken
>to be obviously possible. Wittgenstein for example wanted to alert us
>to just this sort of "must" -- "it *must* be possible to read off thoughts from
>a brain", "there *must* be laws of psychological phenomena just as there are
>laws of physical phenomena" and so on.
I didn't say it *must* be possible. We're talking about what's sensible and
what's silly to believe. You keep sliding between what logically possible
(that it's loggically possible the brain doesn't produce behavour) to what's
reasonable to believe. We were talking about what's silly - not what's
logically possible.
>
>I take it Wittgenstein didn't really oppose the idea that there *might*
>turn out to be such correlations. He was more concerned to oppose what
>he thought of as a misleading picture that says that there *must* be.
Which is not what we were talking about. We were talking about looking
inside a spy's head to find out whether he was a double agent or not and
whether it was reasonable to expect that to be effective or whether to
think that was just silly - to use your term.
>He wrote:
>
> No supposition seems to me more natural than that there is no
> process in the brain correlated with associating or with thinking;
> so that it would be impossible to read off thought-processes from
> brain-processes. ... It is thus perfectly possible that certain
> psychological phenomena *cannot* be investigated physiologically,
> because physiologically nothing corresponds to them. (Zettel 608, 609)
Yeah. Sure. He also said the same about seeds. That you examine seeds
all you liked but couldn't find out what the next generation was going to
be - it just is the case it always turns out the same as the previous
generation.
He said this a few years before we discovered DNA. (I think Crick & Watson
uncovered the structure of DNA the year after Witt'n died.)
Just shows that what's logically possible isn't the same as what's natural,
despite Witt's sense of what's natural.
>Now once one appreciates the outlook in his philosophy of mind,
>this *is* a perfectly intelligible possibility.
Sure. But we weren't talking about what's intelligible - we were
talking about what's silly. It's perfectly intelligible to suppose
that volcanoes aren't brought about by deep events within the
earth. But pretty damn silly.
To respond to the logical point a little. Witt talks about things being,
in some sense, very different from how we usually take them to be.
He imagines regularities which have no visible means of support.
Now I can easily imagine all sorts of relationships which could obtain
and I think that my ability to think up examples is no worse than Witt's.
However, I think Witt misses something. The examples he gives are examples
of regularities. Why are they regular and not irregular? That is,
what is it about Witt's examples which makes it the case that we
consider them regular? Interestingly enough, in the context of the
sort of relationships Witt is considering, I don't think there is *anything*
he has mentioned which does or could account for their regularity/
irregularity. In fact, it seems to me, that the notion of regularity Witt
is using is, actually, derivative upon the ordinary notion of regularity which
*does* make use of ordinary physical law, what-have-you.
The issue is this: granted a background of physical law, stateful
transition, etc. etc., we can readily use this background structure
to account for regular/irregular relationships of the sort Witt
considers. But why should we think that if *all* we had was the
sort of relationship Witt considers there would *still* be a distinction
which was makable between what's regular and what's irregular?
No answer is to be found in Witt - the question isn't raised. Odd,
because elsewhere Witt is sensitive to the above sort of some/all
distinction. (That is, just because, from time to time, any old X
could be a P that doesn't mean that it could be the case that
all X's could be P).
>A somewhat
>domesticated version is in Dennett's philosophy -- for Dennett,
>intentional states are individuated by the intentional stance,
>physiological states are individuated by the design or phyiscal
>stances, and there's no particular reason the phenomena you reveal
>through the former should have analogues in those revealed through the
>latter. Could happen, but maybe not. (see Dennett "Brain-writing and
>mind-reading" in Brainstorms; also, perhaps, Ryle on shamming in Concept
>of Mind). If someone insists it *must* be the case or else some disaster
>follows, that is a pretty good sign they are thinking of the mental in
>the wrong way on this view.
>
>The example Wittgenstein floated there raises an interesting question -
>was he thinking about a case where there would be a violation of
>physical law, a failure of physical science on its own terms? He *was*
>perfectly happy to expect this possibility too. But I like to think he
>was not comitting himself to it. It might be that the order revealed
>through physical explanation might be gap free, yet still omit "the
>order that is there when actions are done with intention" (Anscombe).
>So I think physical law need not be violated for Wittgenstein's
>possibility to come to pass.
Do you mean the physical laws we know and love? Or some putative
physical laws? Our current physical laws are local - that, to my mind,
would pretty well fix the question.
>BTW Dennett's paper "Real Patterns", in which Anscombe's remark is
>cited, may help make clearer what it means to say that there is an
>order revealed through the intentional stance that may be invisible
>from the physical stance. Unfortunately, in my view, he tries to dress
>this up with a mathematical notion of pattern as non-randomness -- you
>are much more suited than me to critique that! I think one should take
>the idea without the pseudo-technical dressing.
I think the argument which says that the sort of processes scientists
go through to pick out physical regularities won't pick out intentional
regularities is pretty suspect. But then you know I think that - that's
been the burden of much of what I've been saying for some time, now.
Dennett *mentions* Algorithmic Complexity, which I take to be very
important in understanding how the space of reasons is constructed.
What Dennett *doesn't* do is to draw out the links between entropy
and Algorithmic Complexity. In failing to do this he misses a trick, which
is to link pattern (better: regularity) to competitiveness and survival.
Cheers,
Pete Lupton
Anders N Weinstein
In article <4re063$r...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
>In <4r9g1r$7...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>>>
>I didn't say it *must* be possible. We're talking about what's sensible and
>what's silly to believe. You keep sliding between what logically possible
>(that it's loggically possible the brain doesn't produce behavour) to what's
>reasonable to believe. We were talking about what's silly - not what's
>logically possible.
Sorry. The original thing that I thought was silly was looking inside
a normal person's brain to better find out what they mean by a remark.
I still think that's pretty damn silly -- even where they are
themselves unable to clarify what they mean. This because what is
hidden in that way is of no consequence to the communicative function of
language.
This recalls Quine's notorious claim that "in linguistics one has no
choice but to be a behaviorist". I have suggested the problem with that
claim is *not* that Quine's behaviorism ignores that insides of our
skulls. It is rather that he insists on artificially redescribing what
is (potentially) observable in overt linguistic activity in
impoverished mechanistic terms. He doesn't seem to appreciate as John
McDowell does that intentionality can be *both* irreducible *and*
observable.
>The issue is this: granted a background of physical law, stateful
>transition, etc. etc., we can readily use this background structure
>to account for regular/irregular relationships of the sort Witt
>considers. But why should we think that if *all* we had was the
>sort of relationship Witt considers there would *still* be a distinction
>which was makable between what's regular and what's irregular?
I'm afraid I don't understand your point here. Could you elaborate?
How do I invoke to the background of physical law if I employ a loose
everyday distinction between what's regular and what's irregular?
And note one lesson of Wittgenstein's is the difference between the
behavior that is following a rule -- with its normative distinction
between correct and incorrect -- and events that are merely regular,
in conformity with some rule, for example a statistical statement.
[Re my claim that an envisaged failure of psycho-physical parallelism
need not involve failure of physical law:]
>Do you mean the physical laws we know and love? Or some putative
>physical laws? Our current physical laws are local - that, to my mind,
>would pretty well fix the question.
Why? What if intentional descriptions are not locally supervenient?
That's precisely the burden of Twin Earth type cases, for example. In those
cases the local physical explanation of why Oscar's hand goes out is "the
same" in the relevant sense as the explanation of why Twin Oscar's hand
goes out. Yet the intentional psychological facts are different -- Oscar is
deliberately reaching for water (H20) because he wants to drink a glass
of water (H20), but Twin Oscar is deliberately reaching for Twater (the
XYZ he calls /water/) because he wants to drink a glass of Twater. (It
simplifies things to do this with something other than H20, which is
so important in our bodies, but I trust you will make nothing of that).
So what does the locality of mechanical explanation of motions have to do
with intentional explanations of action, explanation by reference to
semantically evaluable states? Seems to me there's a fairly unobjectionable
sense in which one might say the latter as such are just not in the world of
physics, without supposing them to be particularly mysterious.
Anders N Weinstein
In article <836389...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <4rc9b7$h...@usenet.srv.cis.pitt.edu>
>
>open to having it better explained. It was Geach's criticism which made
>direct contact with my immediate area of work, not only his choice of
>example, but his concise summary.
I don't get the Geach point. Fodor may well have room to mount a defense
against Geach's doggerel.
Fodor can just say the moral assessment of an action of trigger pulling
will depend on the way it was caused by an inner goal representation --
a sentence in the inner intention box, perhaps. If that box contained
[bring it about that] "Dead(John)" then pulling the trigger is murder.
If it said something else then pulling the trigger might be excusable,
even if it resulted in a death.
>Do you think I have made the latter concern sufficiently clear (or is it
>perhaps something which is only likely to be readily appreciated if one
>does work in the field I have been pre-occupied with for the past 12 years?
I think you have extrapolated from a modest but thoroughly sensible
view about intuitive biases and applied psychology to a grand
metaphysical theory about the mental which few should accept. Quine
nowhere cites Meehl's type of work in support of his philosophy, and I
expect Meehl's brief against intuitive judgment does not attack the
very idea of an intentional state.
>> I don't see that work about "attribution theory" is relevant to Burge's
>> point. I can't see that Burge thinks attributions of externally determined
>> propositional attitudes are typically errors or projections.
>>
>
>Can you clarify this point? What is Burge's position re the propositional
>attitudes?
I would say Burge is basically a realist about them; but you wouldn't
want to go by my attributions. I recall remarks in "Individualism
and Psychology" about how ontological preconceptions should be made
secondary to explanatory practices that are "paying their way".
I believe "Individualism and the Mental" also contains remarks suggesting
he does not think they are merely second class.
>> Putnam rejects the indeterminacy of translation and indeed takes
>> it to be a reductio ad absurdum of Quine's view of meaning.
>
>What text are you implictly refering to when you make this claim?
I thought it was in Representation and Reality, maybe more recent
work, I'll have to look it up. Putnam's views are in constant flux.
>Ontological Relativity really comes down to "A rose by any other name.."
>does it not?
I don't see it. For Quine "this rose" might be interpreted as denoting a
set of numbers. Not very Shakespearean.
>Again, which text are you refering to? I haven't seen him refer to either
>of the above thought experiements.
No, but that's the problem. (There is a brief reference replying to
Putnam in the Library of Living Philosophers vol. by Schillp)
Stimulus meaning is "solipsistic" not in the sense that it ignores
all scientific knowledge, but in the broader sense of Burge's "
individualism" -- it supposes links with external objects do not have
any deep role to play in the determination of meaning.
>Which side is the "far side" here? What Quine thinks immaterial is the
The side of the worldly objects.
>internal details. What *is* important is what is public. The externalist
>perspective is fundamental to Quine's break with phenomenalistic (Machian)
>empiricism.
Not a far enough break from my point of view. In spite of working from
a naturalistic point of view, he still thinks the "boundary" conditions
our beliefs have to meet are thin (non-conceptualized) experience.
This makes reference to an extra-mental reality a mystery.
Peter_...@uk.ibm.com
In <4rc9b7$h...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>
>It might be right to say: if one assumes the bracketing (methodological
>solipsism), then one destroys the very possibility of making sense of
>intentionality or semantics at all, and the concept of thought as something
>which answers for its correctness to a world outside of it will
>degenerate into a blind, purely syntactic affair.
>
>For example, many devotees of the computational theory get themselves
>into a bind over how semantics or reference could matter at all, since
>it would be wholly extrinsic to the blind formal manipulations they
I think reason computationalists don't have an answer to the question
of semantics/reference isn't that they adopt methodological solipsism,
it's that they can't see what's in front of their noses. To gain the required
notion of 'fit' needed to lift computationalism from a syntactic theory
to a semantic one, you have to see the role of parsimony. Some computations
are parsimonious; some aren't i.e. Some fit; some don't i.e. Some represent rightly;
some don't.
The reason why parsimony is so transparent is just that, being parsimony-
guided ourselves, we expect and therefore don't notice parsimony at work.
We find it difficult to imagine what a computational system would be like
were it *not* parsimonious, so extraordinary would such a thing be. Some
of the hardest things to see are those that are assumed.
So I don't think the computationalist has a special problem in incorporating
semantics. Indeed, I think the computationalist's problem in incorporating
semantics is exactly the same problem one has in *characterising* or
*specifying* (rather than just helping oneself to) rational behaviour.
It is one thing to treat rational behaviour as a black box, it is another to
give a specification of *which* black box which goes beyond the circular
definition - rational behaviour is what my black box says is rational behaviour.
Cheers,
Pete Lupton
David Longley
In article <4rf3qa$r...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> In article <836389...@longley.demon.co.uk>,
> David Longley <Da...@longley.demon.co.uk> wrote:
> >In article <4rc9b7$h...@usenet.srv.cis.pitt.edu>
> >
> >open to having it better explained. It was Geach's criticism which made
> >direct contact with my immediate area of work, not only his choice of
> >example, but his concise summary.
>
> I don't get the Geach point. Fodor may well have room to mount a defense
> against Geach's doggerel.
>
> Fodor can just say the moral assessment of an action of trigger pulling
> will depend on the way it was caused by an inner goal representation --
> a sentence in the inner intention box, perhaps. If that box contained
> [bring it about that] "Dead(John)" then pulling the trigger is murder.
> If it said something else then pulling the trigger might be excusable,
> even if it resulted in a death.
Then I think you might profit from spending some time talking to people
who have done just that to come to appreciate how odd people's reasoning
can really get and yet still be considered sane. It is probably impossible
for me to persude you. However (and partially for similar reasons) my bet
is on better behavour management (from the externalist perspective) rather
than directly reasoning with, or delivering cognitive skills programmes (as
explained elsewhere).
>
> >Do you think I have made the latter concern sufficiently clear (or is it
> >perhaps something which is only likely to be readily appreciated if one
> >does work in the field I have been pre-occupied with for the past 12 years?
>
> I think you have extrapolated from a modest but thoroughly sensible
> view about intuitive biases and applied psychology to a grand
> metaphysical theory about the mental which few should accept. Quine
> nowhere cites Meehl's type of work in support of his philosophy, and I
> expect Meehl's brief against intuitive judgment does not attack the
> very idea of an intentional state.
I have no interest in extrapolation to a grand theory. My concerns are
practical. The problem began with how difficult it is to report accurately
and simply draws on work in the philsophy of mind and decision theory/
attribution theory to give some account of why such difficulties are in
fact encountered.
After considering all that, the extensional stance is offered as a solution.
And it is a *technology* I am proposing, not a grand metaphysical theory
about the mental.
>
> >> I don't see that work about "attribution theory" is relevant to Burge's
> >> point. I can't see that Burge thinks attributions of externally determined
> >> propositional attitudes are typically errors or projections.
> >>
> >
> >Can you clarify this point? What is Burge's position re the propositional
> >attitudes?
>
> I would say Burge is basically a realist about them; but you wouldn't
> want to go by my attributions. I recall remarks in "Individualism
> and Psychology" about how ontological preconceptions should be made
> secondary to explanatory practices that are "paying their way".
> I believe "Individualism and the Mental" also contains remarks suggesting
> he does not think they are merely second class.
>
I've only read the first 4 pages (slowly!). I think it will take a few days
to do justice to. Thanks for pointing me to it. It's exactly the sort of
hep I was looking for.
> >> Putnam rejects the indeterminacy of translation and indeed takes
> >> it to be a reductio ad absurdum of Quine's view of meaning.
> >
> >What text are you implictly refering to when you make this claim?
>
> I thought it was in Representation and Reality, maybe more recent
> work, I'll have to look it up. Putnam's views are in constant flux.
>
I get the same impression. I get the impression he is gradually becoming
a Quinean (the preface to the above book would seem to suggest so).
> >Ontological Relativity really comes down to "A rose by any other name.."
> >does it not?
>
> I don't see it. For Quine "this rose" might be interpreted as denoting a
> set of numbers. Not very Shakespearean.
>
Isn't that exactly the point. The rose would still be a rose though. But
let's not puruse that.
> >Again, which text are you refering to? I haven't seen him refer to either
> >of the above thought experiements.
>
> No, but that's the problem. (There is a brief reference replying to
> Putnam in the Library of Living Philosophers vol. by Schillp)
>
I can't find it in Quine's reply to Putnam.
> Stimulus meaning is "solipsistic" not in the sense that it ignores
> all scientific knowledge, but in the broader sense of Burge's "
> individualism" -- it supposes links with external objects do not have
> any deep role to play in the determination of meaning.
>
> >Which side is the "far side" here? What Quine thinks immaterial is the
>
> The side of the worldly objects.
>
> >internal details. What *is* important is what is public. The externalist
> >perspective is fundamental to Quine's break with phenomenalistic (Machian)
> >empiricism.
>
> Not a far enough break from my point of view. In spite of working from
> a naturalistic point of view, he still thinks the "boundary" conditions
> our beliefs have to meet are thin (non-conceptualized) experience.
> This makes reference to an extra-mental reality a mystery.
>
I thought you were arguing from a phenomenological perspective originally?
Perhaps a short summary of your position would be helpful.
--
David Longley
H. M. Hubey
ande...@pitt.edu (Anders N Weinstein) writes:
>Sorry. The original thing that I thought was silly was looking inside
>a normal person's brain to better find out what they mean by a remark.
Why? Isn't objectivity worth some effort?
>I still think that's pretty damn silly -- even where they are
>themselves unable to clarify what they mean. This because what is
>hidden in that way is of no consequence to the communicative function of
>language.
Why would we assume that the person will tell the truth?
Even if he wants to tell the truth, why assume that he
knows the truth so that he can tell it? If we even assume that
he "knows" it, why assume that he can tell it?
These are all separate components of a process, and can
be thwarted by malfunctions at various levels. These have
been given names like lying, self-defense, projection,
white-lies, courtesy, psychoticism, neuroticism, etc etc.
Wouldn't it be nice to have a Lie Detector? :-)..
Peter Lupton
In article: <4rf1uj$r...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>
> >The issue is this: granted a background of physical law, stateful
> >transition, etc. etc., we can readily use this background structure
> >to account for regular/irregular relationships of the sort Witt
> >considers. But why should we think that if *all* we had was the
> >sort of relationship Witt considers there would *still* be a distinction
> >which was makable between what's regular and what's irregular?
>
> I'm afraid I don't understand your point here. Could you elaborate?
> How do I invoke to the background of physical law if I employ a loose
> everyday distinction between what's regular and what's irregular?
The point is that we have absolutely *no* *idea* *what's* called into
play when we talk about regularity. Nothing Witt'n says in this regard
shows otherwise. The observation that we use words like regularity
in ordinary everyday loose talk says *absolutely* *nothing* about
what those words mean, how much baggage is needed to cash the meanings
of those words. It is just a terrible confusion to think that, just
because we use them lightly, easily and shallowly they somehow must
mean something light, easy and shallow.
There is a great difference between using a word and giving a
specification for that word. If one were to attempt to construct
or to define a regularity-detector, there is no limit to what insights
may be needed to do so.
> And note one lesson of Wittgenstein's is the difference between the
> behavior that is following a rule -- with its normative distinction
> between correct and incorrect -- and events that are merely regular,
> in conformity with some rule, for example a statistical statement.
>
> [Re my claim that an envisaged failure of psycho-physical parallelism
> need not involve failure of physical law:]
>
> >Do you mean the physical laws we know and love? Or some putative
> >physical laws? Our current physical laws are local - that, to my mind,
> >would pretty well fix the question.
>
> Why? What if intentional descriptions are not locally supervenient?
>
> That's precisely the burden of Twin Earth type cases, for example.
Sure. We can all come up with something *not* supervenient. So what?
Are you claiming this lack of supervenience as an all-embracing dictum?
If so, then I think you're claiming something very unlikely to be true.
--
Cheers,
Pete Lupton
Anders N Weinstein
In article <4rg4ki$11...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
>In <4rc9b7$h...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
>>
>I think reason computationalists don't have an answer to the question
>of semantics/reference isn't that they adopt methodological solipsism,
>it's that they can't see what's in front of their noses. To gain the required
>notion of 'fit' needed to lift computationalism from a syntactic theory
>to a semantic one, you have to see the role of parsimony. Some computations
>are parsimonious; some aren't i.e. Some fit; some don't i.e. Some represent rightly;
>some don't.
OK, but is it a fit to proximal items -- retinal input, say -- or fit
to a distal item -- the desk in public space whose presence I
visually detect, say? It makes no difference to an inner computational
subsystem as such whether there's a desk there or whether it's operating
like a brain in a vat. It's so-called "representations" would be just as
parsimonious, in the latter case, wouldn't they? And yet they wouldn't be
representing anything distal.
So I'm not sure how the notion of parsimony helps with the problem. It
may be that you are already looking to a wider context of biological
function in an environment, but that's the move that's doing the work, then.
>It is one thing to treat rational behaviour as a black box, it is another to
>give a specification of *which* black box which goes beyond the circular
>definition - rational behaviour is what my black box says is rational behaviour.
I don't understand what you mean by "treating rational behavior as a
black box." Apart from purely intellectual exercises like proving
theorems, rational behavior typically takes place in the world. For
example, buying milk at the store on the way home so I can have some
with my breakfast. A black box can't do that.
Anders N Weinstein
>be thwarted by malfunctions at various levels. These have
>been given names like lying, self-defense, projection,
>white-lies, courtesy, psychoticism, neuroticism, etc etc.
If we want to know if someone is self-deceived, etc., we
do not want to know what's inside someone's brain. We are talking
about a peculiar schism between what they say and what they do.
To take one example: if someone you love is ruining their life through
alcohol abuse yet refuses to acknowledge it, one does not discover or
address their denial by looking for a sentence meaning "I have a drinking
problem" inside their brain. Rather one tries to present objective
evidence in such a way as to force a suitable acknowledgment. Perhaps
one could say they already have much of this same evidence but it is
not inferentially efficacious -- they are refusing to draw the appropriate
conclusion.
Many of the phenomena you cite are more like this. I expect the
unconscious motives of a psycho-analyst are simply not the unconscious
information-bearers postulated by a cognitive science. They pertain
rather to modes of conducting oneself in the world.
Anders N Weinstein
In article <150042...@luptonpj.demon.co.uk>,
Peter Lupton <lup...@luptonpj.demon.co.uk> wrote:
>In article: <4rf1uj$r...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>>
>> In article <4re063$r...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
>There is a great difference between using a word and giving a
>specification for that word. If one were to attempt to construct
>or to define a regularity-detector, there is no limit to what insights
>may be needed to do so.
I'm still not sure what use you're making of the point.
I would have thought Wittgenstein showed there is no absolute or
mathematical notion of "regularity". Uultimately an inquiry into the
meaning of "regularity", like any other inquiry into the meanings of
words, bottoms out at what is judged to be regular by human beings and
creatures like us. So it is at bottom only our "loose" "easy" natural
reactions that are the measure of what is regular; without it one can
always Goodmanize.
>> Why? What if intentional descriptions are not locally supervenient?
>>
>> That's precisely the burden of Twin Earth type cases, for example.
>
>Sure. We can all come up with something *not* supervenient. So what?
>Are you claiming this lack of supervenience as an all-embracing dictum?
>If so, then I think you're claiming something very unlikely to be true.
No, but it takes only one example to refute a generalization. Twin-Earth
entails that intentional states are not locally supervenient. So
Wittgenstein's possibility can be proven actual by a priori means: Accept
the Twin-Earth cases, and you accept that one can't in general read
intentional states off of brain states. You can't study these differences
physiologically, because physiologically nothing need correspond to them.
The original question was how the "locality" of physical law makes a
difference. For example, to someone who thinks of intentional
attributes of persons as more like their economic attributes, say.
Andrzej Pindor
In article <4rckun$i...@usenet.srv.cis.pitt.edu>,
Anders N Weinstein <ande...@pitt.edu> wrote:
>In article <hubey.8...@pegasus.montclair.edu>,
>H. M. Hubey <hu...@pegasus.montclair.edu> wrote:
>>
>>are expecting a drop (as either taking a large step to avoid
>>a hole or taking steps downwards or upwards to go down/up
>>stairs) then the masses of neurons that control these legs
>>are behaving as if they are expecting a flat/level ground. By
>
>the proposition "the ground is solid and level". Tell me for example what
>sort of representational system this proposition is encoded --
>English? Predicate calculus? a vector of activation weights?
>you might not be able to give any answer. You might have to say:
>its not represented in any system, rather its merely implicit in the fact
>that, as you rightly say, the circuitry is moving the legs in a way
>appropriate to flat solid ground, which it is wired up to do by default,
>in the absence of inhibiting inputs.
>
belief in the ground being flat is nowhere represented - it is exactly this
default wiring which represents the mentioned default belief. If we had
a tribe which lived in the trees, then in the process of childhood formation
a different default wiring of the motor circuitry would arise, the one which
would _represent_ different sort of environment, don't you see this? In case
of discovering environment different from the default one, 'inhibiting
signals' would be sent to the motor circuitry to modify the default behaviour,
right? If the behavior were to be proper for the new environment, the signals
in combination with the default wiring would have to _represent_ somehow the
new environment, wouldn't you think so?
Your claim that the belief as to the nature of the ground over which a person
is walking is nowhere represented in the brain amount to a claim that
the persons behavior, as a physical object, moving legs one way or another,
cannot be correlated with the physical (chemical etc) structure of this
object, i.e. that there is some nonmaterial thing (mind, soul, or like) which
makes this physical object move independantly of the object's physical
structure - dualism in its worst version!
Andrzej
>>using the word "expect" I mean that there are no neuron firings
>>coming from someplace in the brain that tells the legs to lift
>>higher, or lower or take a large step to avoid a hole.
>
>Again you are right that no such signals are coming. But can you say:
>the *absence* of such an inhibiting signal on a nerve fiber is itself the
>representation of the belief that the floor is solid? There might be
>many reasons such an inhibiting signal might arise.
>
>A possible view is that I have a higher level belief store which
>contains a world model which has as its consequences that the ground
>is flat and level. I think this misrepresents the nature of default
>assumptions -- I think we often act first and articulate later.
>So an explicit representation of the assumption need not exist anywhere
>before some sort of observably fishy circumstance raises a doubt about the
>solidity of the floor and we hold up.
>
>>That whole point was that large numbers of neurons acting in
>>concert in some way can be given descriptions in terms of
>>behavioral (or even introspective) terminology. The whole thing
>
>Fine. But why not then suppose that belief is not a phenomenon at the
>level of mass neuron firings, but at an even bigger level, the level of
>molar activity of a whole organism in its normal habitat (in which, say,
>the ground is usually solid)? Strictly, it is not the sensori-motor
>routines that belief the ground is solid after all -- they can't
>do anything when doubt arises.
>
>>>The only point is that the language of belief comes apart from the
>>>language of control system representations.
>>
>>I don't know if you are really referring to "control theory"
>>or making a general remark of some sort.
>
>You might try looking into some of Dan Dennett's essays, say "Brain
>Writing and Mind Reading" in his _Brainstorms_ to get a better feel
>for what I am trying to illustrate as the looseness of the relation
>between the language of belief and the language of neural
>representations.
>
--
Andrzej Pindor The foolish reject what they see and
University of Toronto not what they think; the wise reject
Information Commons what they think and not what they see.
pin...@breeze.hprc.utoronto.ca Huang Po
David Longley
In article <4rjkav$9...@usenet.srv.cis.pitt.edu>
ande...@pitt.edu "Anders N Weinstein" writes:
> In article <hubey.8...@pegasus.montclair.edu>,
> H. M. Hubey <hu...@pegasus.montclair.edu> wrote:
> >
What makes you think acknowledging the problem is anything more than a
correlate of doing anything about it? One of my reasons for being so
critical of cognitivism is that when it comes to getting people to
change, changing teir "thinking" is but one small fragment of what's
involved.
In fact, it's such a small fragment that I have serious doubts as to its
importance at all. Its current vogue may be no more than a reflection
of how verbally proficient educated folk are.
^^^^^^^^
I encourage anyone who has not done so already to read "House of Cards:
psychology and Psychotherapy Built on Myth" by R Dawes (1994).
The EVIDENCE is very much against the efficacy of such interventions. What
does work is behavioural, even if it is somewhat mispresented as having a
"cognitive" component.
o oo o oooo
--
David Longley
Anders N Weinstein
In article <Du32x...@gpu.utcc.utoronto.ca>,
Andrzej Pindor <pin...@gpu.utcc.utoronto.ca> wrote:
>In article <4rckun$i...@usenet.srv.cis.pitt.edu>,
>Anders N Weinstein <ande...@pitt.edu> wrote:
>>
>>True. But if I asked you: point out the inner state that represents
>>the proposition "the ground is solid and level". Tell me for example what
>>sort of representational system this proposition is encoded --
>>English? Predicate calculus? a vector of activation weights?
>>you might not be able to give any answer. You might have to say:
>>its not represented in any system, rather its merely implicit in the fact
>>that, as you rightly say, the circuitry is moving the legs in a way
>>appropriate to flat solid ground, which it is wired up to do by default,
> ^^^^^^^^^^^^^^^^^^^^
>>in the absence of inhibiting inputs.
>>
>It is amazing that you state the above and at the same time claim that the
>belief in the ground being flat is nowhere represented - it is exactly this
>default wiring which represents the mentioned default belief. If we had
All I said was that one could not point to a structured vehicle which
codes this content in some representational system. This has the
consequence, for example, that you could not "read off" this
presupposition from the neural circuitry alone, without regard to the
context in which it operates; perhaps indeed the history that produced
it -- in a different context the same neural structures might
"represent" a very different default assumption or none at all.
You can say that it is represented implicitly in the circuitry if you like,
as long as you don't suppose that for each content one can ascribe from
the intentional stance, there must be a discriminable code that bears it.
>a tribe which lived in the trees, then in the process of childhood formation
>a different default wiring of the motor circuitry would arise, the one which
>would _represent_ different sort of environment, don't you see this? In case
>of discovering environment different from the default one, 'inhibiting
>signals' would be sent to the motor circuitry to modify the default behaviour,
>right? If the behavior were to be proper for the new environment, the signals
>in combination with the default wiring would have to _represent_ somehow the
>new environment, wouldn't you think so?
I didn't deny what you say. I asked: what is the code or system in which
the presupposition is represented -- is it like English, or FOPC or like
representation by neural activation vectors. The answer was: none of those,
the representation is not coded in any language, notation, system of any
kind, although it is implicit in the structure.
>Your claim that the belief as to the nature of the ground over which a person
>is walking is nowhere represented in the brain amount to a claim that
>the persons behavior, as a physical object, moving legs one way or another,
>cannot be correlated with the physical (chemical etc) structure of this
>object, i.e. that there is some nonmaterial thing (mind, soul, or like) which
>makes this physical object move independantly of the object's physical
>structure - dualism in its worst version!
This strikes me as a pretty stupid charge given what I said. I didn't
say that there was no physical structure inside the organism which made
walking possible -- just the opposite, I alluded (sketchily) to one.
All I said was one couldn't find a vehicle which codes the content "the
ground is solid" in some system or language or notation, and that one
couldn't identify this belief without looking at the larger context.
Anyway I'm not afraid of property dualism. If I said there was no way
to identify your economic states from your neurons alone, you would not
call me a horrible dualist. Why couldn't mental attributes be like
economic ones: objectively real but non-physyical properties of the
person? Reason-giving explanation is not mainly concerned with the
person "as physical object" but has a rather different function, I
think. Still no laws of physics need be violated when a person walks.
Anders N Weinstein
In article <836597...@longley.demon.co.uk>,
David Longley <Da...@longley.demon.co.uk> wrote:
>In article <4rjkav$9...@usenet.srv.cis.pitt.edu>
> ande...@pitt.edu "Anders N Weinstein" writes:
>
>correlate of doing anything about it?
I didn't really say that it was. Even here the acknowledgment
can precede action and is not identical to it.
>critical of cognitivism is that when it comes to getting people to
>change, changing teir "thinking" is but one small fragment of what's
>involved.
I never said the whole solution should be sought in "changing their
thinking". But if you don't get the acknowledgment, you are not going
to get anyone seeking to modify their behavior either -- or are you
recommending forced behavior modification? (Too much time working
on prisoners?)
The point was to *distinguish* the behavioral sense in which one might
"unconsciously know" one has a problem without acknowledging it from
the cognitivists sense in which information might be unconsciously
represented in one's brain. It was not to recommend a talking cure as
opposed to a behavioral one.
Peter Lupton
In article: <4rjh0p$8...@usenet.srv.cis.pitt.edu> ande...@pitt.edu (Anders N Weinstein) writes:
>
> In article <4rg4ki$11...@zen.hursley.ibm.com>, <Peter_...@uk.ibm.com> wrote:
> >In <4rc9b7$h...@usenet.srv.cis.pitt.edu>, ande...@pitt.edu (Anders N Weinstein) writes:
> >>
> >I think reason computationalists don't have an answer to the question
> >of semantics/reference isn't that they adopt methodological solipsism,
> >it's that they can't see what's in front of their noses. To gain the required
> >notion of 'fit' needed to lift computationalism from a syntactic theory
> >to a semantic one, you have to see the role of parsimony. Some computations
> >are parsimonious; some aren't i.e. Some fit; some don't i.e. Some represent rightly;
> >some don't.
>
> OK, but is it a fit to proximal items -- retinal input, say -- or fit
> to a distal item -- the desk in public space whose presence I
> visually detect, say?
I would say we achieve distal fit *through* the process of finding
proximal fit. And that without the proximal process, the very idea of
distal fit would be forfeit.
> .........It makes no difference to an inner computational
> subsystem as such whether there's a desk there or whether it's operating
> like a brain in a vat. It's so-called "representations" would be just as
> parsimonious, in the latter case, wouldn't they? And yet they wouldn't be
> representing anything distal.
>
> So I'm not sure how the notion of parsimony helps with the problem.
It depends which problem you have in mind, doesn't it? In the case of a
brain in a vat versus the person as a whole, there are two quite specific
problems we have. One problem is to make sense of what the brain in the
vat is about, the other is what a person would be about and why we take one
to have true thoughts, the other not. The assumption is, I take it, that
the brain in a vat is hooked up to a computer, say, producing inputs which
fool the brain into thinking it is seeing trees, etc., when plainly it isn't.
Well, something's going on, and that something has its own logic, its own
reason and - I'd like to say - it's own reference/content/etc., its own notion
of error, truth/falsehood. I really don't see why the story the computationalist
can tell in that context isn't quite adequate to show what's relevantly common
in the two cases.
Certainly the brain in a vat case is not the case of the person. But there
is something very similar about the two cases. In one case reference is made
to trees, in the other case reference is made to tree-ish patterns of some sort.
Why should it be that we take one to be veridical and the other not? I think it
is just this: we take ourselves to be in a certain relationship to other
intelligences and this doesn't include being part of their experiments. So
that if the sky opened up and we were confronted by a mischievous demon, we'd
feel that everything was, to some extent, a lie.
Now I don't see why such a content can't be understood computationally.
A computational device is, after all, embedded in a causal structure.
A computational device also has access to indexicality wrt that very
structure. The expression 'I'm not part of some other intelligence's
experiment' is, I should think, perfectly expressible by such a
computational device.
Notice, as an aside, that I'm no internalist. Although internalism
and externalism are, by definition, binary alternatives, I don't
think that they play that role in a more interesting sense. Rather I think
that, starting from internalism, one will, through the same reasons
Putnam gives in the 'Meaning of "Meaning"', come to embrace externalism.
One is, after all, attempting to achieve parsimony. This comes with its
own need for retraction and correction - error, in short. Once the
possibility of error is factored into the system, the possibility is
that the way one latches onto things isn't the most parsimonious - it is
just one doesn't know it. So one wants to place meanings to some extent
at arms length - deferring to those who know better, etc.. This is plainly
a path an internalist can take and it leads to externalism.
To put it another way, once the conceptual framework of internalism and
learning is in place, the result will evolve to externalism. To put it another
way, take care of internalism and externalism will take care of itself. Thus
I reckon that the internalist notion of content is the thing to go for.
So, although internalism and externalism are logically at odds, the
relationship is actually that the one develops from the other.
One whose meanings are all internalist is an unstable position, and will
slide into externalist meanings.
> It
> may be that you are already looking to a wider context of biological
> function in an environment, but that's the move that's doing the work, then.
You'll have to be more specific about *what* work.
> >It is one thing to treat rational behaviour as a black box, it is another to
> >give a specification of *which* black box which goes beyond the circular
> >definition - rational behaviour is what my black box says is rational behaviour.
>
> I don't understand what you mean by "treating rational behavior as a
> black box." Apart from purely intellectual exercises like proving
> theorems, rational behavior typically takes place in the world. For
> example, buying milk at the store on the way home so I can have some
> with my breakfast. A black box can't do that.
You always treat rational behaviour as a black box - which just means:
never look inside. The problem with treating rationality as a black box
is that there is a nasty circularity, here. When one says: He's behaving
rationally (and so treats rationality as a black box), the judgement
that there is rationality is being made by - another rational agent.
The black box definition of rationality is just what black boxes say
is rationality. This is no definition of any sort: what contribution
is made by the contents of the black boxes is entirely up for grabs.
As I say, the problem of stating (without circularity) *what* behaviour
counts as rational (without begging the question) just is the problem
the computationalist is concerned with.
--
Cheers,
Pete Lupton
H. M. Hubey
ande...@pitt.edu (Anders N Weinstein) writes:
>address their denial by looking for a sentence meaning "I have a drinking
>problem" inside their brain. Rather one tries to present objective
First, if we don't do so it is probably due to the fact that
we can't do it and not because it would not be useful.
It's impossible to get people to "recognize" their problems. If it
were so easy, why would they have become alcoholics in the first
place.
So it's easier to see why it is so difficult if we think about
the neurons being changed. Since new memories are built on top
of old ones, everything we learn at any time depends on and makes
use of what we already learned. It's difficult to teach an old
dog new tricks, as they say, and they are right.
>Many of the phenomena you cite are more like this. I expect the
>unconscious motives of a psycho-analyst are simply not the unconscious
>information-bearers postulated by a cognitive science. They pertain
>rather to modes of conducting oneself in the world.
It's a complicated problem. One of the ways of discussing it at the
upper layers is via positing things like ego, id, etc. It's far
from neuron firings. But if we bear in mind that it's difficult
to make new memories and that all of our memories are tucked away
someplace, it's easier to see how problems can arise at later
dates, and why it could be difficult for people to get the "Aha"
feeling and suddenly recognize their problems or change their
ways. They can't and they usually don't. If someone has come
to learn over many years that "people can't be trusted" or
that "everybody's out to take advantage" , or "you have to
take before they take from you", it's going to be difficult
to get them to change. That's probably why people who eventually
make it to jail, tend to return.