contrasts, control of perception

[From Rick Marken (931006.0830)]

Martin Taylor (931015 19:10) --

All I'm trying to get you to see is that the levels in
the hierarchy of perceptual control behave differently.

I think it was my mistakes to bring up levels; I talked as
though I knew "spin" was an event level perception that was
the same event as "sbin"; I was also talking as though the
/p/ and /b/ sounds are sensations that can be discriminated.
I shouldn't have used "level like" words becuase I don't know
what levels are involved -- or whether they are levels. All
I know is that when I say "spin" and "sbin" they sound like
the same word but they are also discriminably different. I'm
trying to describe my experience, not my theory.

The experiences I am describing seem pretty robust; "sbin"
sounds to me like "spin" spoken with an accent; so the two
words sound different AND the same. I conceive of this in
terms of levels of perception -- but I don't think that that
conception is necessary in order to have the experience.

I said:

When a person tries to
say "spin" they apparently don't care whether the sound after the "s" is
aspirated or not; this is not part of the perception being controlled.

You reply:

Yes it is. If it weren't, the word would sometimes be spoken with an
aspirated p sometimes without and sometimes with something in between.

So now you are rejecting (or simply ignoring) the results of the "pair
test" which is supposedly a test for controlled variables (which I
think it is). The results of that test show (I believe) that virtually
every native speaker of english experiences no error when "sbin" is
heard as an alternative to "spin". Your justification for asserting
that aspiration is controlled when people say "spin" is based on a mistaken
notion of what would indicate that a variable is controlled. The fact
that a variable (aspiration) happens to be repeated at the same level
(as it typically is when people say "spin") is (as you know) not, in
itself, evidence of control; aspiration could be repeated as an irrelevant
side effect of other controlled actions (just as the people in the
CROWD program consistently form an arc around the "guru"; even though the
arcs are not controlled). A variable is controlled if there is resistance
to disturbances that would move the variable from it's intended state.
The "pair test" shows that when you disturb aspiration in the word "spin"
(by saying "sbin" instead) there is NO resistence. This is pretty good
evidence that aspiration is NOT controlled in this case. The same test
reveals that aspiration IS controlled when the intended word is "pin"
because the absence of aspiration (as in "bin") is a disturbance that
meets with resistance (the subject says "different").

But you are rarely conscious of well-controlled perceptions, so when
you speak or listen, you rarely hear the difference without some training.

And, thus, we have "the test" to determine objectively (for all to see)
which variables are controlled and which are not.

Avery Andrews (930610.1711)--

I don't think he's [Burgess] has quite gotten there, but is headed somewhat
in the right direction. He doesn't cite Bill, but does cite someone
called MacKay (1982) `The Motor System Contols What it Senses'
(BBS 5:557), who does.

Holy cow! Could you take a look at the MacKay paper and see what it says
and give a brief report. How could we have missed this?!?! This would be
a GREAT birthday present for Greg -- someone who really got PCT in parallel
with Bill. What would really be neat is if he reports some research based on
something like a test for controlled variables.

But before anyone gets too excited, I have to warn you that there have been
isolated examples of pretty PCT-like statements made by people who really
didn't "get it" (in terms of controlled variables and the research strategies
required to detect them). For example, in an article I am reading (written
by Gavan Lintern) there is a reference to the fact that "Gibson, Olum and
Rosenblatt (1955) have argued that a pilot guides an aircraft toward the
landing aimpoint by controlling the point of optic outflow so that it
corresponds with the aiming point". Sounds like a hypothesis about a
controlled perceptual variable to me. Unfortunately, there were no real
tests to determine whether this variable is controlled. What is typically
done to test this hypothesis is to see whether removal of "cues" to
this variable (like outflow patterns) affect performance. This only
gives vague hints about whether a variable is controlled (what if the
pilot switches to controlling another variable which results in observed
measures of performance remaining the same or better)? The right way to
test this hypothesis is to apply disturbances that influence the state
of the variable and see if they are resisted by the action of the
operator.

Best

Rick

[Martin Taylor 931006 12:10]
(Rick Marken 931006.0830)

Rick, earlier

When a person tries to
say "spin" they apparently don't care whether the sound after the "s" is
aspirated or not; this is not part of the perception being controlled.

I replied:

Yes it is. If it weren't, the word would sometimes be spoken with an
aspirated p sometimes without and sometimes with something in between.

Rick, now:

So now you are rejecting (or simply ignoring) the results of the "pair
test" which is supposedly a test for controlled variables (which I
think it is). The results of that test show (I believe) that virtually
every native speaker of english experiences no error when "sbin" is
heard as an alternative to "spin". Your justification for asserting
that aspiration is controlled when people say "spin" is based on a mistaken
notion of what would indicate that a variable is controlled. The fact
that a variable (aspiration) happens to be repeated at the same level
(as it typically is when people say "spin") is (as you know) not, in
itself, evidence of control; aspiration could be repeated as an irrelevant
side effect of other controlled actions (just as the people in the
CROWD program consistently form an arc around the "guru"; even though the
arcs are not controlled).

Quite true. Point well taken. Further test is required, but it is not as
you follow:

The "pair test" shows that when you disturb aspiration in the word "spin"
(by saying "sbin" instead) there is NO resistence. This is pretty good
evidence that aspiration is NOT controlled in this case. The same test
reveals that aspiration IS controlled when the intended word is "pin"
because the absence of aspiration (as in "bin") is a disturbance that
meets with resistance (the subject says "different").

What this shows is that the perception of aspiration is irrelevant to
the perception of the WORD, not that aspiration is uncontrolled when the
word is spoken.

How would one test whether aspiration were a controlled perception in speaking
"spin?" I think it would be very hard to disturb that CEV directly.
Perhaps it would be possible to do it using synthesized voice, allowing
the subject use knobs to control the degree of aspiration (a short period of
relatively high-frequency noise after a brief silence) and the timing of
the voicing onset. I'm willing to bet, however, that the aspiration
will prove to be a controlled perception in "spin" as well as in "pin",
even though Rick correctly says that the evidence is not there.

Bill Powers (931006.0745 MDT) provided a description of what is going on
that seems to be pretty much a rewording of what I have been trying to say,
except that he does not concern himself with the discrete nature of
category-level perceptions. Other than that, I think I can say that if
you understand what he says, you are pretty close to understanding what
I mean.

Martin

[Avery Andrews 100793.0918]
  (Rick Marken (931006.0830))

> holy cow ...

It will be a while before I can check out the Mackay piece more
carefully, but it is my impression that he as the control of
perception idea, but not the consequent role of the Test.

Avery.Andrews@anu.edu.au