[From: Bruce Nevin (Tue 920419 09:00:07)]
(Bill Powers (920514b) ) --
The only reference signals (and perceptions) that can't be changed freely
as required by higher levels are system concepts. And the only reason we
can't vary our reference signals and perceptions at that level with
complete freedom is that there seems to be no place to stand except another
system concept -- if there is a higher viewpoint, it's impossible to put
into words or systematize. If there's free will, the only place it can work
is at the top, because everything else is dependent and interconnected. And
even at the top, we're free only to be human.
I suggested some time ago a level we might call conversion or (Kuhnian)
revolution, though inspiration might be a better term. From this to us
giddy point of view we shift from one set of systems concepts to
another. It is a point of view that at least some teaching traditions
have sought to cultivate, e.g. Sufism.
Rick Marken (920515 14:00)
Control is all around -- maybe the problem is
that it's too much around -- we take it for granted. Bill P once said that
feedback is like the air we breath; I think this is true of control too.
Because it is everywhere, it is invisible, unless you know what to look
for (like the answer that's "a blowin' in the wind").
This is what I mean when I say "talk to a fish about water."
(Avery Andrews 920515.1432) --
Actually, in the case of grammatical
generations, I suspect that there really are things corresponding to them
(ie., for the generalizations that lead people to postulate noun-phrases,
there is a noun-phrase detector).
From my perspective, a noun phrase, verb phrase, etc. are byproducts of
something simpler and more basic, even when we limit ourselves to
structures in language-as-artifact (rather than considering what control
systems do and how they do it, resulting in those structures).
(Cynthia Cochran (Fri, 15 May 1992 18:17:52) ) --
Cynthia took this as a springing-off point:
Bruce says that
<more than one from of behavioral outputs can accomplish the same
<controlled perception
in speech and some other social/psychological behaviors. This is indeed
Bill read this as misperception of the basics:
Bill Powers (920515.2000)]
A basic concept of CT is that you VARY your actions in order to keep
producing the SAME perceived result. You have to do this because the
environment keeps changing (and your own actions keep changing) and
disturbing the controlled result.
Just for the record, what I was saying (Wed 920413 09:17:47) assumes
the above and looks for something more. I think Cynthia was responding
appropriately to what I actually said, though I emphatically agree with
(what I read as) Bill's intent, that you can't get at this "something
more" about control for social conformity without first understanding
control in less complex cases.
Here's a replay of what I actually said:
Speech is different from the control of a pointing finger in a way that
I think is important for all the social sciences. In the usual case,
behavioral outputs are incidental byproducts of control. They are not
themselves controlled. Some other perception is controlled, and the
behavioral outputs are variable means, whatever it takes in a
disturbance-prone environment to make the controlled perception match a
reference perception in memory or imagination. With speech, however
(and with any conventionalized behavior) the form of the behavioral
outputs is itself subject to control, concurrently with the perceptions
the control of which the behavioral outputs are the variable means.
This is possible whenever there is "free" variability that is not
constrained by the contingencies of control--more than one form of
behavioral outputs can accomplish the same controlled perception. Then
choice among alternatives (or in the range of free variability) itself
is exploited as an aspect of self image, or social standing, or
relationship to others involved in the transaction, etc.
Even pointing with the finger can have a personal style, or a manner
associated with a particular community.
To accomplish this, the behavioral outputs involved in effecting control
of one perception must themselves be monitored and controlled with
respect to particular choices among their range of free variability.
I'd appreciate thoughts on this. So far, Cynthia's is the only response
I've seen.
In simple cases of control, behavioral outputs are byproducts of control.
In control of "manner" the form of the behavioral outputs is itself
controlled. One controls the perception of the relationship of
fingertip to target (basic) and concurrently controls the manner of
carrying out the pointing gesture. My boss's boss has positively
serpentine mannerisms. If he had enough clout and prestige, I would
expect to see others taking on those mannerisms, with little or no
awareness of doing so.
Bill, you may be responding to me obliquely, by way of your response to
Cynthia, when you refer to practical difficulties of modelling:
In general, higher systems can control their perceptions not only by
varying the amount of reference signal sent to a given set of lower
systems, but by changing WHICH LOWER SYSTEMS are provided with reference
signals by that higher system. This is changing the kind of behavioral
action rather than just the amount. This is tricky to implement in a model
when you don't already have a lot of control systems available in the
model, which is about where we are.
I recognize the difficulty. This is why we linguists aren't instantly
working up models. An enormous number of different perceptions
on different levels are under concurrent control in even simple social
transactions. These include negotiation or affirmation of agreements
about "subjective" perceptions such as self image and relationship
between the parties, and "objective" perceptions such as objects and
events, with an intermediate area of things like goals, tasks, roles,
etc. (It appears that the perceptions that seem most "objective" are
those that are subject to the most firmly institutionalized agreements,
which fits well with the notion of intersubjective agreement.)
I have no objection to starting simple. It would be silly and futile
for me to object! What I do object to is premature foreclosure.
It is simply not the case that linguists' concern with the form of
language behavioral outputs necessarily betrays prejudices inherited
from BP behavioral science. (BP: before PCT, natch.) Linguists and
linguistics have been remarkably free, even ostentatiously free, from
S-R and behaviorist theories of psychology. This is because these
theories have been obviously incapable of accounting for language
behavioral outputs. It is often claimed that American structuralists
were behaviorists. I have argued previously in this forum that this is
a canard based on a shallow and partisan reading of the work of
Bloomfield and others.
What I am trying to put forward is a way of thinking about the
complexity of social transactions, including communication and the use
of language, in PCT terms. Only given such a framework is it sensible,
I think, to go for simplified situations in which to apply the test for
control and in which to start developing models. I am perfectly happy
to put lots of the complexity on a shelf. There must first be a shelf
on which to put it.
(Bill Powers (920515.2000) ) --
Reorganization is fundamentally a
cybernetic process, of course, but it isn't necessarily totally "random" or
"statistical" in nature. All that's meant by saying that a process is
random is that we know of no algorithm that will predict what it will do
next. A random reorganizing system is powerful because it doesn't take
anything for granted. But reorganization could be quite systematic in some
way that is too subtle, or too advanced, for us to find order in it.
With the e. coli demo you have shown that a random reorg is sufficient.
I think you have commented elsewhere that this does not preclude
reorganization being guided (controlled). I have some ideas about that
(below).
(Dag Forssell (920516-1) ) --
Bill Powers (920512.0930)
Influences should be thought of as disturbances. That is, you can
perform an act that by itself would alter the other's perceptual world
if it were the only influence.....
It makes sense to me to see influences as disturbances. Can you see
information as disturbances also?
I like this a lot.
For Gregory Bateson, the elementary unit of information (or of mind) is
a difference that makes a difference, i.e. that is transformed into
another difference, such that the sequence of transformations proceeds
in a loop.
* No disturbance => no error => no change
* Disturbance => error => change in reference signal
=> change in output of effectors
Deliberate influence is disturbance controlled in a way that is
contingent upon behavioral outputs of the one disturbed, as they in turn
disturb perceptions that the influencer is controlling.
Define:
<agent> := I, M
I = the influencer
M = the mark (the one influenced)
o.<agent> = behavioral outputs of <agent>
p.<agent> = perceptual input to <agent>
r.<agent> = reference perception in <agent>
e.<agent> = r.<agent> - p.<agent>
E = some perception in the environment, can be o.<agent>
I perceives o.M | p.I != r.I ==> e.I ==> o.I
Behavioral output o.I affects E in some way.
M perceives E | p.M != r.M
M has some choices in the face of disturbance:
(1) p.M != r.M ==> e.M ==> o.M
This is called control.
(2) p'.M = r.M
where p'.M by the imagination connection overrides p.M
This is called ignoring.
(3) p.M = r'.M
where r'.m by the imagination connection overrides r.M
This is called adapting, or being influenced.
The acquisition of a new reference perception need not by by a
reorganization process reaching into the random. It may be by choice of
a perceptual universe, in memory and imagination, in which the error
does not occur.
Choice (3) is attractive e.g. if the emotional value of p.M (or of some
imagined opposite, anti-p.M) is greater than that associated with r.M.
Choice (2) is attractive e.g. if e.M is likely to be transient.
(Bill Powers (920515.2000) ) --
I've now heard ouches from two owners of linguistic toes.
Or was that a heel?
You stopped one quote short in your "you said, I said" summary. You
left out the one in which I loudly agreed with your sequel in this
message. To recap again:
You:
If it seems that there is structure in language, then a model that
explains this phenomenon should not contain that structure, but only
components that lead to phenomena which can be seen as having that
structure.
Me:
If it seems that there are words [word dependencies, etc.] in
language, then a model that explains this phenomenon should not contain
words [etc], but only components that lead to phenomena which can be
seen as being words [etc].
Me (missing quote):
The structure is there. The interpretation of it, or an account of how
perceptual control systems bring it about, is up to us.
You (apparently ignoring this):
By remaining at the level of phenomena, we can only observe and record
apparent cooccurrances and dependencies of particular things. We can't
explain why they are related as they are. The method of modeling attempts
to go beneath that level to the level of underlying operations, looking for
operations that could produce both the phenomena and the observed
dependencies among them.
I think I'm agreeing with you. You apparently think I'm not. One of us
is missing something.
When
I examine words closely, they seem to be just familiar chunks of sound or
objects on paper. The meanings they seem to have (for those words that can
individually designate meanings) turn out to be ordinary perceptions. So
both words and their meanings are ordinary perceptions.
Let's be careful here. What about those words that cannot "individually
designate meanings?" And of those that can, I have provided evidence in
prior posts indicating that the association of words to nonverbal
perceptions is not a simple matter, but is "dirtied" by all sorts of
historically contingent social conventionalization. For example, what
is the nonverbal perception associated with "on" or "by"? Is there just
one?
Avery (920515) opined that
grammar does not seem to be real-world
interactive in the manner that most things studied in psychology are,
so the structures in grammar can't be coming from simple interactions
with a complex environment.
I agree with you, Bill, that
grammar is real-world interactive during the time it develops.
In Generativist theory, the structure in language springs full-blown
from DNA, like Athena from Zeus's headache. The argument from paucity
of data says (1) that language is too complex, and (2) what children
hear around them exemplifies too little of that complexity, and anyway
is too fraught with performance error, for (3) the limited cognitive
capacities of infants to make anything of it, so it must be biologically
innate. The response is (1) that language is not all that complex
(though Generativist descriptions of it are), (2) that what children
experience is carefully structured in the conventional frameworks of a
Language Acquisition Support System (LASS), and (3) that the cognitive
capacities of infants are demonstratedly much much greater than
estimated by Piaget and others on whom Chomsky et al. depended.
Operator grammar addresses point 1, Bruner addresses point 2, and PCT
should have a great deal to say about point 3 in addition to the
considerable body of conventional work that has been published on
cognitive capacities of infants and children. I want PCT to embrace
points 1 and 2 as well, and that is the end to which much of my writing
here has been bent.
No
skill, after it becomes habitual, is real-world interactive to the same
extent it is while it's being learned. We reduce principles and reasoning
to slogans at the drop of a hat. We reduce slogans to slurred and run-
together events even more readily: ISWEARTOTELLLTHETHRUT
HANDNOTHINGBUTTHETRUTHSOHELPMEGOD. By the time we're adults, "grammar" is
just how we say things; most people say the same things the same way every
time, without considering whether it's grammatical or not. When you lose
the real-world interactiveness of grammar, you've also lost grammar.
There is an element of truth in this, and it underlies a range of pervasive
phenomena of language change encompassing lenitions (e.g. "slurred
pronunciation"), cliches, morphemization or grammaticalization (e.g.
"building" as a concrete noun alongside the verb form--"he was building
that building" vs. "he was swinging that swinging"), and so forth.
However, you underestimate the creative aspect of language use.
If you
hear "Every man for itself," the "itself" isn't grammatically wrong, it's
wrong because you said the memorized phrase wrong; you made the wrong sound
at the end. We no longer think of "Every man for himself" as a sentence in
which the words have individual meaning, sexist or otherwise. It's just
something you emit under certain circumstances. It's like a shaped grunt.
This is just plain naive.
Everyone often says or writes novel sentences that she or he has never
said or written before and is unlikely ever to say or write again. This
post is filled with examples. So are yours. Sure, there are frozen
expressions of all kinds, ranging from familiar quotations like "now is
the time for all good men to come to the aid of their party," to cliches
like "every man for himself," to idioms like "take the bull by the
horns," and so on. But they are by no means all that people produce and
hear, even abstracting away errors.
I can't remember when I've said, read, or heard "every man for himself"
in the past five years.
This is probably a good answer to your objection about treating rules as
controlled perceptions. As long as there is a possibility of making
mistakes, the rules are perceived and the utterances are adjusted until the
right rules are perceived. But once a person has settled on utterances that
will reliably fit the rules, the utterances are reduced to phrases and no
longer are treated as having internal structure. Then those utterances are
no longer rule-driven. They're just the way you talk.
This is a reasonable description of the fate of frozen expressions.
They become treated almost as complex words. But except for quotations,
most of them are partially productive, that is, subject to grammatical
operations in some parts but not in others, as in "He really took the
bull by the horns," "Take the bull by the horns," "You must take the
bull by the horns," "You must learn to be brave and take the proverbial
bull by the horns," and so on.
It is interesting that socially "safe" conversation departs little from
a relatively narrow range of frozen expressions that are conventional
tokens for local definitions of membership. More intimate conversation
often involves more creative use of language. A complex issue--there
are many forms of communication that do not use language, avoidance of
"comfortable" cliches can be socially distancing, etc.
Got to run.
Bruce
bn@bbn.com