The Test requires gradual disturbances

CSGnet Archive CSG2003
1

[From Bruce Nevin (2003.03.14 17:10 EST)]

Bill Powers (2003.03.13.0719 MST) --

It's _essential_, for doing the Test, to give the other person a means for
making a correction when you disturb something. In fact, no actual change
of the variable being disturbed should occur: your disturbance should be
small enough for the other person to cancel its effect completely, and slow
enough so no significant error ever develops.

In other words, the Test can only apply to continua. Above the Category Level (according to that hypothesis) no gradual disturbance is possible, except for inducing uncertainty or inconsistency at the margins of category membership. Either it is or it isn't. Therefore disturbances are quantal and instantaneous, not gradual, and resistance cannot maintain the imperceptibility of the disturbance.

         /Bruce Nevin

···

At 10:18 AM 3/13/2003, Bill Powers wrote:

2

[Fromn Bill Powers (2003.03.14.1630 MST)]

Bruce Nevin (2003.03.14 17:10 EST)--

In other words, the Test can only apply to continua. Above the Category
Level (according to that hypothesis) no gradual disturbance is possible,
except for inducing uncertainty or inconsistency at the margins of category
membership. Either it is or it isn't. Therefore disturbances are quantal
and instantaneous, not gradual, and resistance cannot maintain the
imperceptibility of the disturbance.

Yes, thanks, I see what's wrong with what I said. To rephrase the
condition, the Test should not apply any disturbance that causes a
_significant_ amount of error, an amount that can't be corrected with an
ordinary variation in the person's actions. An example would be the way
Dick Robertson and David Goldstein applied self-image disturbances. They
merely said something that contradicted a subject's stated self-image, and
then recorded the action (verbal) taken by the subject. The action, in very
nearly every case, was a correction of the "mistake," so that while an
error was apparently introduced by the experimenter's evaluation, there was
no impediment to immediately correcting it, and nearly all the subjects
immediately did so.

By the way, I didn't mean to say "imperceptibility," if I did.

Best,

Bill P.

3

[From ]

Bill Powers (2003.03.14.1630 MST)--

Bruce Nevin (2003.03.14 17:10 EST)--

In other words, the Test can only apply to continua. Above the Category
Level (according to that hypothesis) no gradual disturbance is possible,
except for inducing uncertainty or inconsistency at the margins of category
membership. Either it is or it isn't. Therefore disturbances are quantal
and instantaneous, not gradual, and resistance cannot maintain the
imperceptibility of the disturbance.

Yes, thanks, I see what's wrong with what I said. To rephrase the
condition, the Test should not apply any disturbance that causes a
_significant_ amount of error, an amount that can't be corrected with an
ordinary variation in the person's actions. An example would be the way
Dick Robertson and David Goldstein applied self-image disturbances. They
merely said something that contradicted a subject's stated self-image, and
then recorded the action (verbal) taken by the subject. The action, in very
nearly every case, was a correction of the "mistake," so that while an
error was apparently introduced by the experimenter's evaluation, there was
no impediment to immediately correcting it, and nearly all the subjects
immediately did so.

By the way, I didn't mean to say "imperceptibility," if I did.

You didn't use that word. It's my paraphrase of "no actual change of the variable being disturbed should occur: your disturbance should be small enough for the other person to cancel its effect completely, and slow enough so no significant error ever develops." [Bill Powers (2003.03.13.0719 MST)]

I think that all that remains of your first statement is "It's _essential_, for doing the Test, to give the other person a means for making a correction when you disturb something", to which now is added that they should be able to correct it "with an ordinary variation in [their] actions".

Perhaps a better way to get at the "slightness" of the disturbance is that it should disturb only the one variable. A contradiction to some feature of their self-image should not change their perception of your attitude toward them, or of the nature or even viability of their relationship to you.

Consider in a certain country "the relationship between adults (especially parents) and children. Typically, the mother will start a small flirtation with the child, pulling its penis or otherwise stimulating it to interpersonal activity. This will excite the child, and for a few moments cumulative interaction will occur. Then just as the child, approaching some small climax, flings its arms around the mother's neck, her attention wanders. At this point the child will typically start an alternative cumulative interaction, building up toward temper tantrum. The mother will either play a spectator's role, enjoying the child's tantrum, or, if the child actually attacks her, will brush off his attack with no show of anger on her part."

The claim is that this sort of interaction is typical between the child and any adult, and between the adult and any child.

It appears that the adult is disturbing a number of variables that the child is controlling: some (the "flirtation") as a result of which the child pays attention to her and begins some interplay with her building toward a peak; and the other a disengagement as a result of which the cumulation of energy and intensity in the interaction is redirected or thwarted - her attention wandering, as a result of which the child works up a tantrum, or her adopting a non-interactive spectator stance, as a result of which the tantrum subsides ineffectually.

So also for "competitive and rivalrous behavior": "The mother will, for example, tease the child by suckling the baby of some other woman and will enjoy her own child's efforts to push the intruder from the breast."

Enter a hypothetical European or American visitor - a missionary, someone in the family of a colonial administrator, a tourist. Not an anthropologist. Any of these scenes is likely to disturb some variables that they are controlling, regarding the welfare of children, or fairness, or the like. We may imagine their subsequent attempts to resist these disturbances.

Remove this ethnocentric foil. Move to some other, like scene, unsullied by outside intrusion. Understand of course that this is not constant, but recurrent and typical, and interactions like this may involve any adult and any child at any time. Enter a PCT anthropologist. A certain amount of helpful generalization has already been provided in this account, so assume that you've seen enough of these scenes to see them as instances of the same kind of thing going on, whatever it is. Some such interaction is in process, or has just occurred (pick one, or both in succession). What does he or she do to understand what is going on? By slight and easily corrected disturbances.

         /Bruce Nevin

···

At 06:39 PM 3/14/2003, Bill Powers wrote:

4

[From Bill Powers (2003.03.15.1059 MST)]

I think that all that remains of your first statement is "It's _essential_,

for doing the Test, to give the other person a means for making a
correction when you disturb something", to which now is added that they
should be able to correct it "with an ordinary variation in [their] actions".

I think we can simplify it even more: it's essential, in testing to see
what another person is controlling, that the other person continue to
control. If your disturbance makes control impossible or even
inconveniently difficult, you will not be able to do the test correctly.

Perhaps a better way to get at the "slightness" of the disturbance is that
it should disturb only the one variable. A contradiction to some feature of
their self-image should not change their perception of your attitude toward
them, or of the nature or even viability of their relationship to you.

The test is not sensitive to what other variables are being controlled at
the same level. If the variable in question is protected by actions against
disturbances, it can be accepted, pro tem, as controlled. However, I
suspect that you mean to refer to higher-level variables. If controlling
one variable entails disturbing some higher-level variable, then it's
likely that the reference level will change in the middle of the Test, or
even that the control system will be turned off (gain set to zero). One of
the conditions for doing the test is that the person keep controlling the
same variable relative to essentially the same reference level long enough
for the test to be done. Excessive disturbance is likely to cause the kinds
of change you imply and alter what the participant is controlling.

Consider in a certain country "the relationship between adults (especially
parents) and children. Typically, the mother will start a small flirtation
with the child, pulling its penis or otherwise stimulating it to
interpersonal activity. This will excite the child, and for a few moments
cumulative interaction will occur. Then just as the child, approaching some
small climax, flings its arms around the mother's neck, her attention
wanders. At this point the child will typically start an alternative
cumulative interaction, building up toward temper tantrum. The mother will
either play a spectator's role, enjoying the child's tantrum, or, if the
child actually attacks her, will brush off his attack with no show of anger
on her part."

Sounds pretty weird to me -- like the sort of thing they do in Boston. Also
it sounds made-up, or at least heavily laden with interpretations. This
"typically" expression with the detailed descriptions that follow speaks to
me of overgeneralization; another term for "typically" is "sometimes" or "I
saw it happen once". As you can see, I am highly distrustful of "facts"
like these, even though I do not doubt that such interactions can occur --
what imaginable human interactions are there that _never_ occur?

The claim is that this sort of interaction is typical between the child and
any adult, and between the adult and any child.

I hope you'll forgive me for saying that I want the claim to be justified
before I spend a lot of time on trying to understand it. For all I know,
the exact scenario described above never happened even once -- it could be
a made-up story meant to give the flavor of the kind of interactions the
observer is convinced she saw and is trying to communicate. Any actual
scientic description would include much more of "I saw the following" and
much less interpretation of invisible feelings and motivations, as well as
fewer claims that general categories of behavior are represented by the
specific instances being observed.

As to doing the test, it is quite possible to do it by watching for
naturally-occurring disturbances, as Frans Plooij did with free-living
chimpanzees. If the child does something and the mother does nothing to
prevent or change it, we can conclude that at the moment the something was
not a controlled variable for the mother. On the other hand, if the mother
consistently intervenes to restore or create some condition (like keeping
the child within a five-foot circle), that would be a candidate controlled
variable. When independent control systems interact, there are plenty of
disturbances just from the fact that they behave independently. The
observer doesn't necessarily have to introduce artificial disturbances.

So also for "competitive and rivalrous behavior": "The mother will, for
example, tease the child by suckling the baby of some other woman and will
enjoy her own child's efforts to push the intruder from the breast."

My initial suspicion would be that this arises from the general enjoyment
we get from controlling things, especially living things. "Getting a
reaction" is the evidence that we're causing something to happen. If I were
generalizing, I'd test other hypotheses involving pets and food animals,
trying to see what there is about the situation that is being recreated by
the person. I know there are people who view babies and animals -- and
indeed, strangers -- as not particularly human, and who can prod and poke
at them in various ways without ever seeming to have any empathy.

Remove this ethnocentric foil. Move to some other, like scene, unsullied by
outside intrusion.

Well, when you introduce the missionaries, there is plenty of opportunity
to observe disturbances and attempts to correct their effects, on both
sides of the equation, without the PCT observer's having to intrude at all.

Understand of course that this is not constant, but
recurrent and typical, and interactions like this may involve any adult and
any child at any time. Enter a PCT anthropologist. A certain amount of
helpful generalization has already been provided in this account, so assume
that you've seen enough of these scenes to see them as instances of the
same kind of thing going on, whatever it is. Some such interaction is in
process, or has just occurred (pick one, or both in succession). What does
he or she do to understand what is going on? By slight and easily corrected
disturbances.

What the PCT observer has to watch for is not the recurrent and typical
interactions so much as what happens when there are modest deviations from
the recurrent and typical patterns (assuming that we can confidently state
what we should recurrently and typically observe). If nobody makes any move
to limit or correct such deviations, then we can say they are uncontrolled
consequences of other behaviors. If someone reacts to the deviations by
correcting them, we can reasonably guess that there is a controlled
variable at least closely coupled to what we see being altered and then
restored, if we are not looking at the variable itself.

We also must not forget to observe individual cases before making any
generalizations. Some people will control a variable that others do not
control. Some mothers, in fact, may not enjoy teasing their children. We
then have to explain why some do and some don't -- the reason clearly can't
be cultural if everyone is supposed to be in the same culture. You can't
use the same explanation for both A and not-A.

Best,

Bill P.

5

from David Wolsk (2003.03.15.21.47PST)

[From Bill Powers (2003.03.15.1059 MST)]>

I think we can simplify it even more: it's essential, in testing to see
what another person is controlling, that the other person continue to
control. If your disturbance makes control impossible or even
inconveniently difficult, you will not be able to do the test correctly.

As to doing the test, it is quite possible to do it by watching for
naturally-occurring disturbances, as Frans Plooij did with free-living
chimpanzees. If the child does something and the mother does nothing to
prevent or change it, we can conclude that at the moment the something was
not a controlled variable for the mother. On the other hand, if the mother
consistently intervenes to restore or create some condition (like keeping
the child within a five-foot circle), that would be a candidate controlled
variable. When independent control systems interact, there are plenty of
disturbances just from the fact that they behave independently. The
observer doesn't necessarily have to introduce artificial disturbances.

In comparing these two approaches, I feel the second has much more to offer.
I've always felt that active testing for the CV opens a can of worms.

David
Victoria Canada (where we're using our worms to catch herring this week)

···

Subject: Re: The Test requires gradual disturbances

6

[From Bruce Nevin (2003.03.15 13:13 EST)]

Bill Powers (2003.03.15.1059 MST)

Sounds pretty weird to me […] it sounds
made-up […]

Any actual scient[if]ic description would include much more of “I
saw the following” and much less interpretation of invisible
feelings and motivations, as well as fewer claims that general categories
of behavior are represented by the specific instances being
observed.

Your disbelief is understandable, even predictable - but mistaken. This
was in a small country far from here, which still exists, albeit much
changed today.

It is true that I am quoting from a publication whose point is to state
some generalizations, but the basis for these generalizations is far
broader than the specific cases selected as illustrations. I mentioned
that child-adult interactions of that sort were described as pervasive
and typical. Statements of the specificity you wish were made in other
publications, including a detailed photographic analysis. The
investigators (a husband and wife team) learned the language and spent an
extended period of time living with these people, so they had ample
opportunity to determine what was recurrent and characteristic vs. what
was accidental or fortuitous, and to do so was what they were there for.
They previously had done the same sort of research in other places, and
so had in mind not only their own ethnocentric expectations but also the
differently ethnocentric expectations of tribes in other places.

As to doing the test, it is quite possible to
do it by watching for

naturally-occurring disturbances, as Frans Plooij did with
free-living

chimpanzees. […] The observer doesn’t necessarily have to introduce
artificial disturbances.

Given the resistance, I thought it important that you bring this up
rather than I. Naturalistic observation is essential for this kind of
research. Participant observation relies almost entirely on observation
of naturally-occurring disturbances and resistance to disturbance as they
occur among natives of the culture.

Something is judged to be a disturbance to the state of a controlled
variable not because it changes the state of that variable but because we
expected it to change the state of the variable and it did not, and it is
from this that we infer that the variable is being controlled. But in
this work what we expect may not be relevant, the disturbance that a
person refrains from causing can be most telling, and it may be that we
expect neither the disturbance nor the refraining.

when you introduce the missionaries, there is
plenty of opportunity

to observe disturbances and attempts to correct their effects, on
both

sides of the equation, without the PCT observer’s having to intrude at
all.

Yes, but such disturbances are extraneous to the culture. We want to
understand these people’s perceptions of (among other things) what is
proper and what is improper, in their own terms. What disturbances arise
and are resisted in the course of their ordinary interactions among
themselves?
The motivation for this may be obscure for you, or maybe you wish me to
justify this methodological claim. In any science, it is important to
keep the investigator’s preconceptions out of the analysis of
data. This field, as also the study of language, imposes a prior
requirement to keep the investigator’s preconceptions out of the data
themselves, that is, out of the interactions that are being observed
and recorded. This is exactly analogous to Heisenberg’s indeterminacy
propositions, only worse because much higher up in the perceptual
hierarchy, hence more pervasive and less predictable. Yes, surely, one
learns something from the fact that such and such is contrary to one’s
culture-laden expectations, but what is learned thus is ipso facto
relative to those expectations - a reflection of the culture you brought
with you in your head - and what we are after is the culture on its own
terms.

The mother-child interactions that I described earlier are not due to
some pathology or predilection of the individual mother. The
justification given for this and for the generalizations that the
interactions illustrated was that this sort of interaction is commonplace
and not even limited to parent-child pairs. The generalization has a
broader basis than that. Here are two other kinds of interactions:

In historical memory, the people of different villages had engaged
sometimes in warfare. “… [W]ar was thought of as containing large
elements of mutual avoidance. The village of B______ G___ was surrounded
by an old vallum and foss, and the people explained the functions of
these fortifications in the following terms: ‘If you and I had a quarrel,
then you would go and dig a ditch around your house. Later I would come
to fight with you, but I would find the ditch and then there would be no
fight’ - a sort of mutual Maginot Line psychology. Similarly the
boundaries between neighboring kingdoms were, in general, a deserted
no-man’s land inhabited only by vagrants and exiles.”

“[These people employ] definite techniques for dealing with
quarrels. Two men who have quarrelled will go formally to the office of
the local representative of the [king] and will there register their
quarrel, agreeing that whichever speaks to the other shall pay a fine or
make an offering to the gods. Later, if the quarrel terminates, this
contract may be formally nullified. Smaller - but similar - avoidances
(pwik) are practiced, even by small children in their quarrels. It is
significant, perhaps, that this procedure is not an attempt to influence
the protagonists away from hostility and toward friendship. Rather, it is
a formal recognition of the state of their mutual relationship, and
possibly, in some sort, a pegging of the relationship at that state. If
this interpretation is correct, this method of dealing with quarrels
would correspond to the substitution of a plateau for a
climax.”

Resistance to disturbance may (often does) constitute a reciprocal
disturbance. Among the people that one of the investigators had studied
in the immediately preceding years, this characteristically led to a
runaway process. One sort was symmetrical - competition, rivalry, and so
on. The other sort was asymmetrical - exhibition/spectatorship,
aid-giving/dependence, dominance/submission. And indeed this sort of
runaway process is familiar in the temper tantrums of children, quarrels,
arms races, and in our conventions of entertainment, instruction, rank as
institutionalized authority, and so on. What was striking was that among
the people presently under discussion runaway processes of either kind
were always and consistently damped to equilibrium.

“The principal hierarchical structures in the society - the caste
system and the hierarchy of full citizens who are the village council -
are rigid. There are no contexts in which one individual could
conceivably compete with another for position in either of these systems.
An individual may lose his membership in the hierarchy for various acts,
but his place in it cannot be altered. Should he later return to
orthodoxy and be accepted back, he will return to his original position
in relation to the other members.”

This has the appearance of a kind of answer to what must be a universal
question, what to do about interpersonal conflict. Obviously no one
deliberated over this question and proposed this answer, nor did anyone
say, as it were, “OK, let’s try this and see how well it
works,” but that is the appearance. So an obvious question is, how
does this development of a “way things are done” come
about?

It appears to be a general theme even where the above functional
explanation does not seem to apply. Other forms of cultural expression
have an analogous damping any crescendo to a plateau.

“In general, the lack of climax is characteristic for [the] music,
drama, and other art forms. The music typically has a progression,
derived from the logic of its formal structure, and modifications of
intensity determined by the duration and progress of the working out of
these formal relations. It does not have the sort of rising intensity and
climax structure characteristic of modern Occidental music, but rather a
formal progression.” [I think Gamelan music of Indonesia is like this. I’m sure you must have heard Gamelan at one time or another.]

“The formal techniques of social influence - oratory and the like -
are almost totally lacking … To demand the continued attention of an
individual or to exert emotional influence upon a group are alike
distasteful and virtually impossible; because in such circumstances the
attention of the victim rapidly wanders. Even such continued speech as
would, in most cultures, be used for the telling of stories does not
occur [here]. The narrator will, typically, pause after a sentence or
two, and wait for some member of the audience to ask him a concrete
question about some detail of the plot. He will then answer the question
and so resume his narration. This procedure apparently breaks the
cumulative tension by irrelevant interaction.”

Perhaps a general fabric of “how things go” is itself the
controlled perception? Difficult to test, since any disturbance at that
high a level is by means of specific disturbances e.g. of telling a story
and expecting to be listened to.

In this broader context, the anthropologist witnessing recurrent
interactions of adults and small children, in which some
“cumulative” engagement is invited and then thwarted or
dissipated, might suppose that these might be the occasions of some
learnings by these children that are essential to their successfully
becoming adults who do all these many various things in this
characteristic, climaxless, endless plateau kind of way, without
deliberation or coercion, because it feels like the right way that things
should be. If not a controlled perception, this must be a parallelism or
congruence or analogical likeness of a great many controlled
perceptions.

“As is usual in anthropology, the data are not sufficiently precise
to give us any clue as to the nature of the learning processes involved.
Anthropology, at best, is only able to raise problems of this order. The
next step must be left for laboratory experimentation.”

But the method of participant observation does make us aware of what
questions to ask.

I have omitted names because they are irrelevant here, but I can provide
references for all of this if you want.

    /Bruce

Nevin

···

At 02:00 PM 3/15/2003, Bill Powers wrote:

7

[From Bill Powers (2003.03.16.1305 MST)]

Bruce Nevin (2003.03.15 13:13 EST)--

Your disbelief is understandable, even predictable - but mistaken. This
was in a small country far from here, which still exists, albeit much
changed today.

I didn't actually disbelieve your description, but thought I would give the
expected response first.

However, I still disbelieve to this extent: I doubt that the participants
in this culture always showed the "typical" behavior all of the time, so
there is an inherent falsehood in saying "participants" did this or that. I
know I am flying in the face of all social-science studies as they are done
now, but I'm sure you recognize that what I say is true, even if you don't
agree that it's an important truth.

>Something is judged to be a disturbance to the state of a controlled
variable not >because it changes the state of that variable but because we
expected it to change the >state of the variable and it did not, and it is
from this that we infer that the >variable is being controlled. But in this
work what we expect may not be relevant, the >disturbance that a person
refrains from causing can be most telling, and it may be that >we expect
neither the disturbance nor the refraining.

If our expectations are drawn from observations within the culture, then
the problem you raise here (which relates to _prior_ expectations) is not
relevant. The ethnologist reports, as you wrote the other day, that the
mother typically behaves this way and that way relative to her children and
other children. Fine: let us make that our expectation.

If every mother behaves exactly as described with respect to every child,
then we will obtain no further information about this interaction, and
cannot even find out what variables are being controlled (through passive
observation). But I would not expect this to be the case. I would expect to
see deviations from this typical pattern, and if this pattern is actually
under control, I would expect to see variations in the actions of the
people involved which have the effect of restoring the pattern to its
"typical" form. If the child, for example, failed to become excited or
aroused in the typical way, I would expect the mother to increase the level
of disturbance until the child was once again behaving in what the observer
thinks of as the typical way. I would expect this, that is, under the
hypothesis that the typical pattern that I have in mind is being perceived
and controlled by the mother.

>The motivation for this may be obscure for you, or maybe you wish me to
justify this >methodological claim. In any science, it is important to keep
the investigator's >preconceptions out of the analysis of data. This field,
as also the study of language, >imposes a prior requirement to keep the
investigator's preconceptions out of the data >themselves, that is, out of
the interactions that are being observed and recorded. This >is exactly
analogous to Heisenberg's indeterminacy propositions, only worse
because >much higher up in the perceptual hierarchy, hence more pervasive
and less predictable.

You don't have to justify it. The hypotheses for the Test do not have to
come from cultural prejudices; it's better if they don't. The ethnologist's
report is a sufficient starting point.

>Yes, surely, one learns something from the fact that such and such is
contrary to one's >culture-laden expectations, but what is learned thus is
ipso facto relative to those >expectations - a reflection of the culture
you brought with you in your head - and what >we are after is the culture
on its own terms.

Yes, of course. I shouldn't indulge in too-subtle humor -- it's almost
always misunderstood. I _take it for granted_ that any prejudices that are
brought into an ethnological study, even in Boston, are detrimental. What
else? I am interested in the question you raise as to how we can perform
the Test to discover controlled variables in a strange culture, and my
answer is, the same way you discover them in the Coin Game or any other
application of the Test. You look for what seem to be stable patterns, and
then you ask what makes them stable. You detect stability by seeing the
effects of disturbances, little deviations from the seemingly stable
pattern, and you check to see if the stability is maintained because of
_systematic resistance_ to those disturbances.

The mother-child interactions that I described earlier are not due to some
pathology or predilection of the individual mother. The justification
given for this and for the generalizations that the interactions
illustrated was that this sort of interaction is commonplace and not even
limited to parent-child pairs. The generalization has a broader basis than
that. Here are two other kinds of interactions:

...

It appears to be a general theme even where the above functional
explanation does not seem to apply. Other forms of cultural expression
have an analogous damping any crescendo to a plateau.

This should pose no problems for the Test. What we want to find out is
whether this culture has mechanisms for damping out any conflicts so that
escalation doesn't happen. Indeed, we want to find out whether the limiting
of conflicts (or their main effects) is a conscious purpose of people in
this culture, or whether it's just an accidental side-effect of otherwise
meaningless habits of behavior. We can do this by observing small
deviations from the typical patterns, and then seeing how people act when
such deviations occur. If their actions have only random effects on the
size of the deviations, as often making them greater as smaller, we can
scratch the hypothesis that people in this culture are controlling for the
pattern we thought we observed. We would have to conclude that the pattern
is stable only in a statistical sense and by accident. But I strongly doubt
that this is what we would find, considering the observations you present.

Best,

Bill P.

8

[From Bruce Nevin (2003.03.18 10:54 EST)]

Bill Powers (2003.03.16.1305 MST)--

Yes, of course. I shouldn't indulge in too-subtle humor -- it's almost
always misunderstood.

Irony & other subtleties get lost so commonly in email that it's worth the effort to avoid them or to flag them explicitly. (BTW, I'm not in Boston.)

I am interested in the question you raise as to how we can perform
the Test to discover controlled variables in a strange culture, and my
answer is, the same way you discover them in the Coin Game or any other
application of the Test. You look for what seem to be stable patterns, and
then you ask what makes them stable. You detect stability by seeing the
effects of disturbances, little deviations from the seemingly stable
pattern, and you check to see if the stability is maintained because of
_systematic resistance_ to those disturbances.

What we want to find out is
whether this culture has mechanisms for damping out any conflicts so that
escalation doesn't happen. Indeed, we want to find out whether the limiting
of conflicts (or their main effects) is a conscious purpose of people in
this culture, or whether it's just an accidental side-effect of otherwise
meaningless habits of behavior. We can do this by observing small
deviations from the typical patterns, and then seeing how people act when
such deviations occur. If their actions have only random effects on the
size of the deviations, as often making them greater as smaller, we can
scratch the hypothesis that people in this culture are controlling for the
pattern we thought we observed. We would have to conclude that the pattern
is stable only in a statistical sense and by accident. But I strongly doubt
that this is what we would find, considering the observations you present.

Perhaps there are three options:

1. The pattern in the coin game is controlled by the individual being tested.

2. The Pattern of rings and arcs in the Gather/Crowd programs is a byproduct of a number of individuals each controlling some variable(s) at about the same values. The pattern itself is not controlled by any individual. Yet it is a controlled outcome. Disturb a ring somehow and you simultaneously disturb one or more individual's control of proximity. So in an interesting sense the rings and arcs are collectively controlled.

Here, you say "it's just an accidental side-effect of otherwise meaningless habits of behavior." But "habits of behavior" are never meaningless. It's just that their meaning (i.e. the purposes being controlled) is of a different kind (logical type) from the meaning of the pattern, and has nothing to do with it.

3. The pattern, as with the rings and arcs, is indeed of a different logical type from the phenomena that exemplify and constitute the pattern - the perceptions controlled by individuals - the pattern is necessary in order to enable the individuals to control those perceptions: the latter cannot be controlled without fitting into and constituting the former. They are the coin of cooperation, and like any coin must persist in their status as "legal tender" as they change hands. The elegantly structured pattern of phonemic contrasts of a language, perhaps a byproduct of maximizing the perceptual distance between those points at which utterances are comparable and therefore capable of contrast; but as well as that the sound system of a language must be consistent from one speaker of the language to another if they are to communicate. The pattern of maintaining a steady state is served by thwarting or diverting runaway processes as means. This may be a byproduct of avoiding conflict, but it is generalized across all relations, and farther than is needful for that purpose. Whatever the reason, the same characteristic climaxlessness is seen in Gamelan music, in the continuities of batik design, and so on, where interpersonal conflict can scarcely be at issue, and where indeed there is evidently no need for the maneuvers of thwarting or diverting. Perhaps individuals control for a consistent pattern that connects, or perhaps we control for consistency between ourselves and our fellows (there is evidence for this) and so converge on a consistent pattern without perceiving it as such.

Such a perception would be disturbed by juxtaposition with people whose ways of living together do not exemplify and help to constitute the same pattern that ours do. One or both of us changes, or one or both of us avoids the juxtaposition. One strategy to avoid the juxtaposition is to encapsulate: that kind of people are different from our kind of people. Each poke fun at the other, and this affirms a recognition as well as a separation or rejection as unsuited for myself. If there isn't cause for conflict (scarce resources, etc.) this can be good natured over many generations, as in aboriginal California or around the Gulf of Mexico, both areas enormously diverse ethnically and linguistically.

Good. OK, we are talking about pervasive patterns, such that a given pattern is expressed in diverse forms that one might not at first see as connected. To figure out how this works it's best to work in a settled culture and leave our multicultural tumult here for later. (Even so, in practically any "simple" society today you have the effects of colonialism and conflict to reckon with.)

A simple hypothesis is that people living together in communities converge on Principle and System Concept variables and reference values because doing so facilitates cooperation in control of lower-level variables. These Principle and System Concept variables are not usually represented directly by overt verbal formulations. As has been observed here before, if everybody's already controlling a given variable at a certain reference level there's no need to tell them to do so, and exhortations & admonitions are for the things that people are not doing, hence the commonplace disparity between precept and example in the communities familiar to us. So we can't expect reliably to get at Principles and System Concepts by asking people or ourselves for the precepts by which we believe we live. Especially not in our complex cultural situation.

The possibility of variables at the highest levels being socially created and interactively reconstituted in the course of usage resolves a problem of timing. Variables and reference values at the highest levels are extremely slow to change. If the highest reference values were set by error from intrinsic control systems such as hunger, core body temperature, and the like, one would starve or die of hypothermia before reorganizing them suitably for survival. Intrinsic error is more timely remediated at lower levels. Consequently, intrinsic error never has a chance to affect the highest levels, because by the time it could be effective there the organism is either dead or incapacitated.

         /Bruce Nevin

···

At 03:48 PM 3/16/2003, Bill Powers wrote:

9

[From Bill Powers (2003.03.18.1243 MST)]

Bruce Nevin (2003.03.18 10:54 EST)--

Perhaps there are three options:

1. The pattern in the coin game is controlled by the individual being tested.

2. The Pattern of rings and arcs in the Gather/Crowd programs is a
byproduct of a number of individuals each controlling some variable(s) at
about the same values.

3. The pattern, as with the rings and arcs, is indeed of a different
logical type from the phenomena that exemplify and constitute the pattern -
the perceptions controlled by individuals - the pattern is necessary in
order to enable the individuals to control those perceptions: the latter
cannot be controlled without fitting into and constituting the former. They
are the coin of cooperation, and like any coin must persist in their status
as "legal tender" as they change hands.

Well, I can tell which of the three possibilities you favor, but the third
one may not turn out to be significantly different from what is already
proposed in HPCT, once we get past the use of metaphors. What you're
working up to proposing here seems to be that the higher-order pattern
(once established) has an existence independent of the lower-order control
processes, and facilitates the lower order control processes, in fact
making them possible. A way I can make sense of your comment is to say that
once the higher-level perception exists _in other people_, it somehow
facilitates the growth of the same higher-order organization in a newcomer.
At least that's not counter-intuitive.

But there is a logical problem here, because before any higher-order
perception can exist in any individual, the perceptions of lower order of
which it is a function must already exist. So before a newcomer can be
taught to control the higher-order pattern, there must be a higher-order
pattern to perceive, and in order for that pattern to be perceived, there
must be an input function receiving signals from lower levels and
extracting that pattern. That input function can't be constructed without
some minimum of perceptual signals from lower levels to serve as input.
Therefore the lower-level perceptions must exist first, before the higher
can be formed.

I'm ready to admit that the formation of a new input function and thus a
new higher-order perceptual signal could well be facilitated by the
existence of input functions at that level in other individuals. The
influence wouldn't be direct, but would be a consequence of living in a
world where others are controlling particular kinds of variables at the
higher level. A simple example that comes to a simple mind is
word-meanings. If people use a word like "honesty," the question a newcomer
would have is what experience this word indicates. Through continued
interaction with those who already have meanings for this word, the
newcomer would find out what procedures are consistently considered honest
and not honest -- low error or high error. Of course others would do their
best to make it easier to form the right higher-order perception, but
basically the newcomer has to do it alone on the basis of considering many
experiences which he is informed have known values of honesty. Eventually
the newcomer would be able to predict when an action would be termed honest
or dishonest, after which we would say he has acquired the "same"
higher-level controlled variable that the others control.

My strong inclination is to avoid making it any more complicated than that
until forced to do so.

Perhaps individuals control for a consistent
pattern that connects, or perhaps we control for consistency between
ourselves and our fellows (there is evidence for this) and so converge on a
consistent pattern without perceiving it as such.

Why "without perceiving it as such?" I think we perceive many things even
at high levels, and control those perceptions, without having symbolized
them and worked them into any formal system of understanding. We can even
control perceptions without being conscious of doing so. Is that consistent
with what you mean?

I've been meaning to comment supportively on Tom Hancock's differentiation
between "ideal" and "actual" goals, which is another way of approaching
this subject. The "ideal" goal is a goal that has been put into words and
made part of a logical structure of thought, but it might be very different
from the actual reference signal that (supposedly) is being described. In
doing the Test, we can discover with considerable confidence the actual
pattern being controlled, but we may represent it in a way very different
from what is actually under control. In the coin game, one person may say
that the controlled pattern is a "zig-zag" while an other says it is the
letter N on its side, or a Z. But whoever controls this pattern is
producing a specific configuration, whatever it is called. One can _draw_
the configuration, and everyone will agree that that is the right one --
while continuing to disaqree about what to call it.

My point is that the fact that a person is unable to describe in words the
cultural pattern that is under control by no means indicates that the
person is not both perceiving and controlling that pattern. The Test is
particularly useful in this regard, because it can pick up the signs that a
control system is present even when a person (or an animal) is unable to
communicate a description of what is being controlled to someone else or
even to himself.

At to the remainder of your post, I could comment but this is getting too
long and we need to discuss this much. I'm sure the more important
neglected points will come up again.

Best,

Bill P.

10

[From Bruce Nevin (2003.03.18 21:43 EST)]

Bill Powers (2003.03.18.1250 MST)--

Well, I can tell which of the three possibilities you favor,

Actually I don't "favor" any one, I think that all three exist. They are complementary to each other.

1. Controlled variables.
2. Collective outcomes of individual control, probably irrelevant to it.
3. Collective outcomes of cooperative control, which facilitate it or may even make it possible in some instances. A pre-existing heritage for any newborn. (Origins in prehistory another discussion.)

[paraphrasing me]

the higher-order pattern
(once established) has an existence independent of the lower-order control
processes, and facilitates the lower order control processes, in fact
making them possible.
once the higher-level perception exists _in other people_, it somehow
facilitates the growth of the same higher-order organization in a newcomer.

This is the situation when the newcomer is an infant born into the community.

But there is a logical problem here [...] the lower-level perceptions must exist first, before the higher
can be formed.

The pattern need not be among the perceptions controlled by individuals, any more than the rings and arcs are. The difference is that the rings and arcs have no (or very little) value in facilitating individuals' control of the variables that they are controlling, and that is because that control does not involve communication, cooperation, etc. The sound pattern of a language is not controlled as such by the speakers of the language, but the tightly organized systematicity of it is what makes a complex vocabulary possible. It makes possible the sudden and astonishingly rapid explosion of vocabulary that begins at about 18 months to age 2, it enables one "accent" to be understood by the speaker of a different one, indeed enables a speaker to shift from one dialect or register unconsciously as circumstances change.... It is a shared organizing framework, immanent in the conventionalizations of "the way things are done", with respect to which the controlled perceptual variables of human interactions are controlled.

Such patterns are sustained between the people, as it were - immanent in the conventionalized manner of their interactions - without being variables that any individual controls or even necessarily perceives. (Each convention is a controlled perception, but their consistency with one another in a coherent pattern is not.) Just as the rings and arcs are sustained between the individual agents, immanent in their actions, without being variables that any agent controls or (in the simulation) even perceives. People can and do perceive them sometimes, but cannot control a ring or arc simply because to do so requires the cooperation of the other people in the formation. People can and do perceive cultural patterns sometimes, but while these patterns are used as means for individuals to cooperate in controlling other variables, they cannot themselves be subject to modification. They cannot themselves be subject to modification by an individual because to make such a change is to lose mutual intelligibility. They cannot themselves be subject to modification collectively or by individuals cooperating in controlling them because that would require another framework with respect to which the collective change to the pattern would be coordinated.

There are also Principle- and System-Concept-level perceptions that individuals develop, such as you mention. These indeed do depend, as you say, upon a prior capacity to control the lower-level perceptions that enter the input functions in which these higher perceptions are constructed.

         /Bruce Nevin

···

At 03:52 PM 3/18/2003, Bill Powers wrote: