Zombies and PCT

[From Rick Marken (01.02.12.2000 PST)]

Bruce Nevin (01.02.12 20:14 EST)

Given that there is already a straightforward account of
language as a self-organizing structure (a structure of constraints,
above all dependence) growing out of social experience, it is also
simpler to put it that way around (Ockham).

I prefer to view language the way I view all other behavior: as the
control of perception. I'm not prepared to abandon this view simply
because there is "already a straightforward account of language as a
self-organizing structure...growing out of social experience" (what-
ever that may be). By that test I would have to give up my view of,
say, operant behavior as the control of reinforcement simply because
there is already a straightforward account of operant behavior as
action controlled by its consequences (the reinforcements themselves).

(talking, listening, writing poems and research papers, etc) is people
controlling various perceptual variables. It would be great if you,
an expert on language, would do some reseach to test this notion.
An example of the kind of research you might do is the kind the WORD
grammar checking program does on me everytime I run it. The WORD
grammar checker wants me to avoid the use of passive voice. So when
it finds sentences written in the passive voice it suggests that
I rewrite the sentence in active voice. I would do this if I were
controlling for perceiving sentences written in the active voice.
The passive voice sentences discovered by WORD would be as much of
a disturbance as are the misspellings discovered by the spell checker
(and sometimes even by myself). But these sentences are not a
disturbance as evidenced by the fact that I do nothing about them
when they are discovered. So I am not controlling for writing in
active voice; but I am controlling (believe it or not) for spelling
correctly. Voice and spelling are two potential "linguistic"
controlled variables. I bet you could think of a lot more, as well
as some simple and clever ways to test to determine whether or
not they are actually being controlled.

Best regards

Rick

···

from my perspective, what we are seeing when we see language behavior
--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: marken@mindreadings.com
mindreadings.com

[From Bruce Nevin (01.02.12 20:14 EST)]

Rick Marken (01.02.12.1545)--

You need a category level to control categories. An important
aspect of language is the ability to use one type of perception (a
word) to represent a category of perceptions. You have to be able
to perceive categories in order to be able to control for this
very central aspect of language.

This is assuming to be true what has been supposed hypothetically.

The only evidence for categories correlated with language is language itself. Absent independent evidence, it is just as plausible to say that the socially established structures of language delimit a categorization of experience. Given that there is already a straightforward account of language as a self-organizing structure (a structure of constraints, above all dependence) growing out of social experience, it is also simpler to put it that way around (Ockham). There are various other kinds of evidence for this, for example where institutionalized word usage contradicts perceptual categories (a tomato is not a fruit, but ask a speaker of a non Indo-European language who knows no English to group into categories various items including a head of lettuce, a tomato, and a guava).

In any case, when something in language is correlated with a perception, neither the language structure nor the perception is necessarily a category. "That brightness to the left of where you were looking" is not a category, nor is the intensity perception to which it refers, nor does it make of the latter a "category perception" just because on another occasion it could refer to another such perception, any more than your finger is a category, nor the rabbit that it points to, however many other rabbits (or "point-at-able objects") it might point to on other occasions. And here we are talking only of denotation, the most primitive form of how language "means" things.

         Bruce Nevin

         Bruce Nevin

···

At 03:47 PM 02/12/2001 -0800, Richard S. Marken wrote:

[From Richard Kennaway (2001.02.13:0914 GMT)]

Rick Marken (01.02.12.1545):

I seem to recall talking [with Gazzaniga]
about whether the non-verbal brain was aware of the object it
had perceived. One indication that it was is the fact that it
will point to the correct object when asked (apparently the
right brain can _understand_ language).

How does that constitute evidence of awareness? What do you mean when you
ascribe awareness to something?

-- Richard Kennaway, jrk@sys.uea.ac.uk, http://www.sys.uea.ac.uk/~jrk/
   School of Information Systems, Univ. of East Anglia, Norwich, U.K.

[From Bruce Nevin (01.02.12 01:37 EST)]

Rick Marken (01.02.12.2000 PST)--

Bruce Nevin (01.02.12 20:14 EST)

> Given that there is already a straightforward account of
> language as a self-organizing structure (a structure of constraints,
> above all dependence) growing out of social experience, it is also
> simpler to put it that way around (Ockham).

I prefer to view language the way I view all other behavior: as the

control of perception. I'm not prepared to abandon this view ...

You are not being asked to abandon this view. The self-organizing structure is of course a byproduct of the control of perceptions by individuals in social settings, just as the rings and arcs of the crowd program are. And just as in real life an individual may perceive the arc formed by others and not go closer than others to the point of interest, though without conflict with controlling some perception of drawing attention to oneself, or the like, their reference for proximity to that point would be a smaller distance (I've experienced that, and I expect most of us have), and just as a "normal distance for conversation" becomes customary and institutionalized among a group of people (differently e.g. for Americans and for Arabs), in the same way the much more complex structures of language become institutionalized and are social realities that children learn to perceive and control even as they use them. No change to PCT is required. All that is required is a willingness to look at the data and an interest in it. You are interested in other things, so by all means, let be.

         Bruce Nevin

···

At 08:04 PM 02/12/2001 -0800, Rick Marken wrote:

[From Bruce Gregory (2001.0213.0622)]

Rick Marken (01.02.12.1545)

It is clear to me from our exchange that I simply don't understand HPCT
well enough to contribute anything useful to this topic.

BG

[From Rick Marken (01.02.13.0720)]

Me:

I seem to recall talking [with Gazzaniga] about whether the
non-verbal brain was aware of the object it had perceived.
One indication that it was is the fact that it will point
to the correct object when asked (apparently the right brain
can _understand_ language).

Richard Kennaway (2001.02.13:0914 GMT) --

How does that constitute evidence of awareness? What do you
mean when you ascribe awareness to something?

Great questions. As far as the first goes, I only meant that
_pointing_ at a banana when asked "which of these fruits was
shown earlier" seems to constitute the same evidence of
awareness as _saying_ "it was the banana" when asked the same
question. In the blindsight discussion it was assumed that
the ability to answer questions about what you saw indicated
an awareness of what you saw. I consider pointing as one way
of answering questions.

I have a pretty informal notion of awareness. I guess I ascribe
awareness to something that is able to perceive what it is
perceiving. Being able to answer questions about what one
is perceiving would, therefore, constitute behavioral evidence
of awareness, from my perspective.

Best regards

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
MindReadings.com mailto: marken@mindreadings.com
www.mindreadings.com

[From Rick Marken (01.02.13.0750)]

Bruce Nevin (01.02.12 01:37 EST) --

The self-organizing structure [of language] is of course a byproduct
of the control of perceptions by individuals in social settings, just
as the rings and arcs of the crowd program are.

But you said that this self organizing structure gave a "straight-
forward _account_ of language". that suggests that the self organizing
structure is a _model_ of language behavior. Now you are saying
that this structure is a byproduct (side effect) of control, like
the rings in the CROWD program. If you do see the "self-organizing
structure of language" as a noticeable but basically irrelevant side
effect of control (like the rings in the the CROWD program) then
we're right on track.

No change to PCT is required. All that is required is a willingness
to look at the data and an interest in it.

As long as the data were collected with an understanding that the
behaving systems might be engaged in the process of control of
perception then I agree. What data are you talking about?

You are interested in other things, so by all means, let be.

I am also interested in language. I would love to see someone doing
PCT based research on language. So I think our interests overlap
considerably. It would be great if you (a linguist) would do some
PCT based research on language. I think that would really help
get PCT some visibility in the cognitive science community.

Best regards

Rick

···

---
Richard S. Marken Phone or Fax: 310 474-0313
MindReadings.com mailto: marken@mindreadings.com
www.mindreadings.com

[From Bill Powers (2001.02.13.0835 MST)]

Bruce Gregory (2001.0212.1642)--

Again, I find the term "verbal levels"
unclear. Aren't all levels potentially verbal? For example, "I see a red
patch." Is this not a verbal report of a rather low-level perception?

I don't claim to be clear about this subject, either. But as I think of
"the verbal levels," they are the levels where symbols are manipulated
rather than the analog perceptions for which they stand. Symbols, sequences
of symbols (statements), and logical or at least rule-driven manipulations
of statements make up what I think of as the verbal levels. Obviously, I'm
using these levels right now, the symbols being configurations on a screen
rather than sound-configurations. That's a long way from understanding
language and thought, of course.

It's quite possible to control perceptions (at lower levels) without
involving the verbal systems. You can match a red patch to paint chips by
controlling for a lack of change in color perception as you switch your
gaze from one location to another. You don't have to be saying to yourself,
"That's a red patch, and now this is another red patch."

In fact, you could never match one color sample to another using the verbal
systems. The term "red" refers to a _category_ of sensations, so all sorts
of samples would be called "red" even if we can see they don't match. Also,
you can match color samples even if you don't know the name of the color.

This applies to all kinds of perceptions. Any time continuous control is
involved, you have to use the lower-level control systems and leave the
verbal levels out of it. That's because we don't have nearly enough words
to distinguish all the states of the analog perceptions that we must
discriminate to achieve smooth control.

Why is the category level any more verbal than the relationship level?

To my mind, it's because relationships are perceived and controlled in an
analog fashion. We can name certain relationships, like "above" and "below"
and "on", but that isn't sufficient to give smooth control during, say, a
tracking task. We aren't controlling those three words when we track; we're
controlling a perceived continuum of relative distances. We pick out some
crudely defined situations and attach words to them, but a word like
"above" doesn't tell us enough quantitatively to allow good control.

More generally, verbal systems use symbols, which are discrete rather than
continuously variable. I have proposed that symbols are created by
including them in the same class with the things symbolized: that is, the
symbol "dog" (as both a sound and a visual configuration) is arbitrarily
made a member of the perceptual class that also includes such
configurations as Great Danes, Chihuahuas, huskies, terriers, and so forth.
Any example of those breeds, plus the auditory or visual word, is reported
at the category level as the same perceptual signal. Note that this creates
the same effect as memory association, but without invoking that mechanism
(whatever it is).

A judge at a dog-show can't be using just the verbal levels to distinguish
among dogs, even within a breed. Instead, dogs are judged on a continuous
scale of perceptions, on which one dog can be closer than another to the
ideal representative of the breed or type (as the judge sees it). The judge
looks back and forth between the best two terriers, and finally picks the
one that seems the most terrier-like to the judge. All that is nonverbal.

Exactly how would you [Rick] propose to distinguish
between an awareness problem and a verbalization problem. For example, if I
say I can't see anything, but I am manifesting blind sight, how do we know
if the problem is awareness or verbalization?

I think the distinction would be more properly that between visualization
and verbalization. Is it that the person can't see, or that the person
can't describe in words what is seen? Awareness, I think, is much harder to
establish, since either visual control or verbalization can, in principle,
take place without awareness.

If you want to find a PCT explanation, the test would be in whether you can
control something the state of which can't be perceived without sight. If
you control, say, the position of a vase on a table so it's exactly between
two other vases, then the only known way to do this is to be able to see
the three vases. If you succeed in this control task, then inside you there
must be a control system using visual information. If, at the same time a
person does this, that person claims not to be able to see the vases, then
we have to conclude that the person's conscious (or at least verbal) self
is cut off, in some way, from the control systems in the brain that are
using the visual information at a lower level. Of course we can investigate
physiologically to see if the visual systems (retina, optic nerve, midbrain
structures) are intact; in the cases that are most interesting, I believe
they usually are.

So we can establish the existence of visual perceptions by showing control
of variables that can be observed only by using vision. If the person still
seems unable to describe anything visual, we must then suspect that there
is something wrong with the
symbolizing/categorizing/sequencing/rule-following systems. Or, perhaps,
something is preventing awareness from receiving information from the lower
system -- although anything having to do with awareness is pretty hard to
investigate objectively.

In another post, you propose that awareness may be able to get information
without relying on the neural perceptual systems. I think that proposal is
far harder to believe than the explanation I offer (of course, otherwise I
would offer the same explanation as yours). In all cases where people claim
to get information about the world without using the senses, I feel that
almost any alternative explanation would be more believable; an "ordinary"
explanation would have to be pretty incredible to be more incredible than
positing an information channel that nobody has ever been able to
demonstrate. I think that's a pretty good rule of thumb to follow (I didn't
invent it). For example, which is more incredible: that 100 people saw a
space ship from another planet hovering in the sky exactly where Venus
should have been, but wasn't, or that 100 people saw Venus and were
ignorant of what they were looking at? I'll take the latter any day. It
doesn't take much to convince me that 100 people, or for that matter 1000,
can be totally and simultaneously mistaken about what they're looking at.

"Blindsight," I assume, does not involve any actual knowledge about the
world that does not use the senses. I would like to see the experiments
done with the person actually blindfolded (something an experimenter might
not think of doing to someone who claims to be blind). The contrary
postulate goes beyond the bounds of my credulity.

Best,

Bill P.

[From Bill Powers (2001.02.13.1134 MST)]

Bruce Nevin (01.02.12 20:14 EST)--

The only evidence for categories correlated with language is language
itself. Absent independent evidence, it is just as plausible to say that
the socially established structures of language delimit a categorization of
experience.

I wonder about that. The idea of category, it seems to me, implies
interchangeability among the members of a category; that is, when we look
at a Chihuahua, we say it is "a dog," but we would say the same thing if it
were a Great Dane. So in some sense we perceive Chihuahuas and Great Danes
as being the same in some way even _before_ we say "dog" -- otherwise, how
would we know what to say? I know that there are alternative models, in
which each perception is separately "associated" with the category-word,
but that gets awkward when a new member shows up and is perceived as
similar enough to be called by the same word.

Suppose I'm trying to lift the end of something heavy. What I need is
something to pry with. How many things are in that category, "something to
pry with?" How do I detect objects in the immediate endironment that I
could use for prying? This suggests categoryness to me: a twig, a
screwdriver, a cane, a metal bar, a 2 x 4, and so on -- all equivalent in
that they can be seen as things to pry with.

Words do designate categories, but do they really _make_ categories? I
dimly remember having had this discussion with you before. I think we may
have an example of synecdoche here: using a part of something to stand for
the whole, as in "10 sail" meaning 10 ships. You say

Given that there is already a straightforward account of
language as a self-organizing structure (a structure of constraints, above
all dependence) growing out of social experience, it is also simpler to put
it that way around (Ockham).

But how can anything, especially a _product_ of the brain, organize itself?
I ran into this concept in the 1950s when the idea of "self-organizing
systems" got popular. I could never see how a system could organize
_itself_, which would entail not only bringing into being the properties it
ends up with, but creating the ability to organize itself, too. It always
seemed to me that the ability to organize something had to exist first, and
that only _other_ aspects of the system would become organized as a result.
Thoughts like these were what made me end up postulating a "reorganizing
system" rather than saying that the brain reorganized "itself."

If all we needed for language was its self-organizing structure and social
experience, then all animals would have language, wouldn't they? What's
left out of this account is what the brain needs to be able to do in order
to create language. Brains that can do whatever-it-is end up with language;
brains that can't don't get language even though they have social
experiences, and even if language is self-organizing as you propose.

There are various other kinds of evidence for
this, for example where institutionalized word usage contradicts perceptual
categories (a tomato is not a fruit, but ask a speaker of a non
Indo-European language who knows no English to group into categories
various items including a head of lettuce, a tomato, and a guava).

This by no means settles the question of whether the word or the category
comes first. By my definition of a category, we can group things any way we
please to form categories: the things in my pocket, for example. We just
"or" together the perceptual signals we want to be in the same category. I
can group pears and tomatoes into a category which I call "fruit," and at
the same time admit that technically, the word "fruit" does not include the
word "tomato." But I can _perceive_ that apples and tomatos are
interchangeable in some respects: they're not meat or potatoes.

In any case, when something in language is correlated with a perception,
neither the language structure nor the perception is necessarily a
category. "That brightness to the left of where you were looking" is not a
category, nor is the intensity perception to which it refers, nor does it
make of the latter a "category perception" just because on another occasion
it could refer to another such perception, any more than your finger is a
category, nor the rabbit that it points to, however many other rabbits (or
"point-at-able objects") it might point to on other occasions.

But these are precisely what I call categories. Any time you precede a noun
by "a" or "an" you're referring to a category. Any time the referent of a
word is not a specific, unique, present-time perception (essentially all
the time), it's referring to a category. This is why, when you want to
refer verbally to the patch of sunlight reflecting off bottom of a spoon
over on the left side of your current orientation, you have to mention many
categories: the specific perception you're trying to indicate is in their
intersection. In your case, the intersection of "brightness," "left:,
"where", and "looking."

It is very hard to find denotative words in the sense of words that point
to one and only one perception.

Best,

Bill P.

[From Bill Powers (2001.02.13.1532 MST)]

Bruce Gregory (2001.0212.1153)

HPCT describes a world of zombies to the extent that awareness is not an
integral feature of the HPCT model. Awareness somehow floats above the
control system and observes it. (This is a purely verbal "explanation"
quite independent of the model.)

Yes, it's independent of the model. It's verbal in the sense that it's an
attempt to put a certain kind of experience into words. The words mainly
point to experiences that people will recognize if they've noticed some of
the subjective properties of awareness, such as its changeable content.

The blindsight phenomenon suggests to me,
not there is something wrong with or missing from PCT, but that we cannot
assume that the perceptual world of our awareness is identical with the
perceptual signals in the PCT model. I find that interesting. Apparently no
one else does. So be it.

It would be interesting if there were any reason to think that we can be
aware of the outside world by any means other than neural signals. Should I
take suggestions of that possibility seriously? I don't think the role of
neural signals in perception is PCT-specific.

Best,

Bill P.

[From Bill Powers (2001.01.13.1540 MST)]

Bruce Gregory (2001.0212.1337)]

This I do not understand. The persons report seeing _nothing_. Your
description would seem to apply if the person saw patches of color or
patterns, but could not identify those patterns, or possibly even
characterize them as patterns. Is this correct? This is the reason that I
conjecture that the control machinery and the "awareness" machinery might
be parallel rather than identical.

The problem here is in the meaning of "the person". In PCT, the brain is
modular, both within and across levels. So a whole person is the sum of a
lot of systems which control autonomously except for sending and receiving
reference signals and copies of perceptual signals. If we consider only the
envelope of a person, everything the person does or experiences is done and
experienced by "the person." But when we model the person, there is not
just a single system: there are higher and lower levels, and even within a
level there are different control systems, all acting and perceiving at the
same time. We are conscious of _some_ of those activities, but by no means
all.

Sometimes when Mary says something to me, I will be so engrossed in what
I'm doing that I don't hear her. I would report hearing _nothing_. Yet it
often happens that by pausing and reflecting in the right way, I can
realize that some words were in fact said, and sometimes I can actually
play them back and _this time_ "hear" them. "Oh, the garbage? I was _just_
about to take it out." So what is hearing, and who is it that hears? I can
also ask, what is seeing, and who is it that sees?

In a hierarchical model, we can have control of perceptions happening at
one level while higher levels are neither receiving signals from nor
sending signals to the active systems. This description applies to some
quite common experiences that people have, such as hearing but not hearing,
as above. It is quite common for one person to stare right at another, and
even to respond to facial expressions like smiles, and a moment later to
say, "Sorry, what were you saying? Did you want something? I was thinking
about why my check didn't arrive." Blindsight, I think, is surely just an
exaggerated example of this not-unfamiliar sort of occurrance.

Right back at you, fella!

Attaboy.

Best,

Bill P.

[From Bill Powers (2001.02.13.1604 MST)]

Bruce Gregory (2001.0212.1350)--

Rick:

Did the authors really determine
that these people are blind?

Bruce:

Depends of course on what you mean by blind. They determined that the
people were unlikely to be lying or confabulating.

That's not quite what I was suggesting earlier -- did they blindfold the
people to make sure they were not using visual information? If a person
could not perform a control task while blindfolded, it is reasonable to
assume that visual information was used when the task was done while not
blindfolded, even if not used by those subsystems of which the person was
aware.

If so, how do they think these
people manage to fit the lock in the keyhole with no fumbling:
by extrasensory perception?

No, by "seeing". (Presumably if they wore blindfolds they would be unable
to perform the acts successfully.

That's what I was asking about. Did they check this?

Rick:

It seems to me like Bill's explana-
tion is the obvious one -- PCT or not. These people can perceive
and control the perceptions involved in putting a key into
a lock. They are just not able to report on those perceptions.

Bruce G:

That is, they are unaware that that they are perceiving anything.

Yes. Remember that we aren't talking about whole people here, but the
subsystems that make up a whole person. We shouldn't expect the parts to
behave like the whole system. It's perfectly possible for a conscious
person to be concentrating attention at one level or in one group of
subsystems, while other subsystems go on perceiving and controlling as
usual -- but outside awareness. This happens all the time. In a control
model, of course, all actions are involved in controlling some perceptions
or other. But not all perceptual signals are in consciousness all the time.

I suggest that it can only be "an awareness problem". I am not sure what
you mean by "higher level systems involved in controlling for reporting on
these perceptions." At what level do we control for reporting?

At the levels which organize and control verbal communication. Those are
not the levels where, for example, you manage to stay balanced while
walking on the deck of a rolling ship.

I agree with you that it's hard to tell from objective evidence whether a
perceptual signal is missing, or present but not connected to awareness.
The latter is suggested when the person goes on controlling the perceptual
signal in the normal way but reports being unaware of doing so (or of
having done so).

I agree with just about everything Rick has been saying on this subject --
and no, we haven't been consulting offline.

Best,

Bill P.

[From Bruce Nevin (01.02.13 15:44 EST)]

Rick Marken (01.02.13.0750)--

Bruce Nevin (01.02.12 01:37 EST) --
> The self-organizing structure [of language] is of course a byproduct
> of the control of perceptions by individuals in social settings, just
> as the rings and arcs of the crowd program are.

But you said that this self organizing structure gave a "straight-
forward _account_ of language". that suggests that the self organizing
structure is a _model_ of language behavior.

The structures of language and the use of those structures ("language behavior") are not the same thing.

How those structures themselves are perceived and how those perceptions are controlled has not been determined, for the most part, but it is clear that language is structured (indeed, in a sense is structure, i.e. dependencies among entities identified by the Test, and dependency upon dependency), and that structures in language correlate with information that speakers/writers intend and that hearers/readers understand, though this is only part of the meanings associated with speaking or writing as with all other human activities.

If you do see the "self-organizing
structure of language" as a noticeable but basically irrelevant side
effect of control (like the rings in the the CROWD program) then
we're right on track.

In actual life situations side effects of controlling some perceptions may be perceived and those perceptions may themselves be controlled, and then, though they are still irrelevant to the systems of which they are side effects, they are not irrelevant to the other systems that are controlling them. I gave an example. The Crowd program is in this respect an idealization that shows some important things about social behavior but (quite justifiably) leaves out other things.

As long as the data were collected with an understanding that the
behaving systems might be engaged in the process of control of
perception then I agree. What data are you talking about?

Data from various forms of testing for controlled variables, plus methods for identifying dependencies among those controlled variables, tests and methods that have been in use in descriptive linguistics (though not always described or explained or even understood as such by all practitioners) for the past 50 years. Some references:

Z. S. Harris. 1951a[1946]. Methods in Structural Linguistics. Chicago: Univ. of Chicago Press. Reprinted as Structural Linguistics. [Original title: Methods in Descriptive Linguistics. The publisher wanted the buzz-word.]

Z. S. Harris and Charles F. Voegelin. 1951b. "Methods for Determining Intelligibility among Dialects of Natural Languages". Proceedings of the American Philosophical Society 95.322-329.

Z. S. Harris and Charles F. Voegelin. 1953. "Eliciting in Linguistics". Southwestern Journal of Anthropology 9:1.59-75. (Reprinted in Harris 1970.769-774.)

Z. S. Harris. 1954. "Distributional Structure". Word 10:2/3.146-162. (Also in Linguistics Today: Published on the occasion of the Columbia University Bicentennial, ed. by Andre Martinet & Uriel Weinreich, 26-42. New York: Linguistic Circle of New York, 1954.-Repr. in The Structure of Language:
Readings in the philosophy of language ed. by Jerry A[lan] Fodor & Jerrold J[acob] Katz, 33-49. Englewood Cliffs, N.J.: Prentice-Hall, 1964, and also in Harris 1970a.775-794, and in Harris 1981.3-22.)

Here are two of the cited places where the 1953 and 1954 papers are reprinted:

Z. S. Harris. 1970. Papers in Structural and Transformational Linguistics. Dordrecht/ Holland: D. Reidel.

Z. S. Harris. 1981. Papers on Syntax. Ed. by Henry Hiz. (=Synthese Language Library, 14.) Dordrecht/Holland: D. Reidel.

I recommend starting with _Methods_ (possibly the most systematically misconstrued book in the history of science), then "Eliciting" for more detail on the Test and "Distributional Structure" for more detail on the analysis of dependencies. There's lots more, of course, but this is a good start. For questions about information, the origin of language, language learning and loss, and so on there are other references.

I am also interested in language. I would love to see someone doing
PCT based research on language. So I think our interests overlap
considerably. It would be great if you (a linguist) would do some
PCT based research on language. I think that would really help
get PCT some visibility in the cognitive science community.

I have many pages of records of the results of disturbing variables controlled by a speaker of the Pit River (Achumawi) language and identifying reference values for those variables. Eventually I will organize these into a specification of what the controlled variables and their references are from which a careful reader, to be sure with much practice to develop skills, could reproduce examples given, and probably could generate new utterances in the language, though it would not be possible to see if these were disturbances to variables controlled by native speakers because there are none left. This will take the form of a grammar and dictionary of the language, with analyzed texts for exemplification and practice. All of this will not happen soon, but I hope to achieve it before I die. My dissertation is a start.

         Bruce Nevin

···

At 07:52 AM 02/13/2001 -0800, Richard S. Marken wrote:

[From Bruce Gregory (2001.0214.1053)]

Bill Powers (2001.02.13.1604 MST)

That's not quite what I was suggesting earlier -- did they blindfold the
people to make sure they were not using visual information? If a person
could not perform a control task while blindfolded, it is reasonable to
assume that visual information was used when the task was done while not
blindfolded, even if not used by those subsystems of which the person was
aware.

I know of no one who believes or suggests that the people were not using
visual information, only that they were not aware of using visual
information. As far as the subjects were concerned, they could see nothing.
This, it seems to me, is quite different from not being able to describe
what you are seeing. We normally think that what we are observing, in a
tracking task for example, is part of the control loop. The blind sight
results suggest to me an alternative interpretation in which the perceptual
system branches. One branch goes to the control mechanism postulated by PCT
(the parietal lobe) and the second branch goes to the part of the brain
associated somehow with awareness 9the temporal lobe). This interpretation,
together with the split-brain results, suggests to me support for the
self-perception model proposed by Bem (Bem, D.E., Self-perception theory.
In Berkowitz (Ed.) _Advances in experimental social psychology (Vol. 6)_.
New York: Academic Press (1972)) in which people infer their beliefs from
watching themselves behave. I see no conflict between PCT and
self-perception theory. I am aware that this interpretation is at variance
with your interpretation in which the observer can causally alter the
control system. I prefer my interpretation because it avoids the need to
invoke what appears to me to be a very hazy mechanism.

At the levels which organize and control verbal communication. Those are
not the levels where, for example, you manage to stay balanced while
walking on the deck of a rolling ship.

I think of verbal behavior as no different than any other behavior. The
same hierarchy controls verbal behavior as controls one's ability to stay
balanced while walking on the deck of a rolling ship. I'm sure you agree.

I agree with you that it's hard to tell from objective evidence whether a
perceptual signal is missing, or present but not connected to awareness.
The latter is suggested when the person goes on controlling the perceptual
signal in the normal way but reports being unaware of doing so (or of
having done so).

So we have no differences.

I agree with just about everything Rick has been saying on this subject --

Why am not surprised?

BG