Tar baby

CSGnet Archive CSG1998
1

[From Oded Maler (980214)]

After a short moment of an illusion of soberity, comes again the
patriotic rhetoric:

[From Rick Marken (980213.1000)]

Bruce Abbott (980213.1145 EST) --

> All in all, I continue to be amazed at view being defended here
> that 150 years of research on the relationship between stimulation
> and perception has no relevance for a view which holds that behavior
> is the control of perception.

All in all, I continue to be amazed by your amazement. PCT has
bitten you on reinforcement, natural selection, research methods,
operant conditioning, causality, Ebbinghaus, anticipation and
prediction, model-based control, behavior modification, etc. Now
it's biting you on psychophysics. You probably figure that you've
won all these debates but that's irrelevant; the fact of the matter
is that you keep running up against PCT resistance to nearly
every one of your most cherished beliefs about psychological science.
How long will it take you to figure out that PCT is not the baby you
thought it was when you found it-- it's Rosemary's Baby;-)

[..]

Look at how Bill carries on about the Test.
Look at how he is critical of standard psychophysical research.
Those who think I'm driving all those great scientists away from
PCT should note that Bill hits a pretty good wood shot himself;-)

I think PCT suffers from a lack of sympathetic criticism (not the one you
get from rejection letters from the establishment). In a community, a
lot of things can be said about the outside world and accepted
noddingly by the adherents, without much thought. Even very wise
bright and original persons can say nonsense about certain topics.
No one is omniscient and no one has mental resources to think about
everything. This is one of the roles of scientific community - to
give critical feed-back, and this is what some people are trying to
do here, albeit the violent reaction of certain fundamentalists.

To the topic itself, if you accept the view that the human brain
is a large black box about whose internal structure you can have
only a crude idea (HPCT), that you have only a crude access to
its I/O ports, and that you yourself are a similar kind of
mechanism, but with a significantly different wiring and "reality
tunnel" (I borrow from Wilson), you should know the limits of any
method to infer something about it function and mechanism.

Even when a component is usually embedded in a closed loop situation,
it might be useful to know its I/O characteristics, which puts limits
on its performance in *any* context. You cannot, of course, isolate
such "gate" neither in the lowest-level (intensity detector) not on
the higher level (concept detector) but it seems to me psychophysics
is trying to do the best possible to have a clue about these
characteristics. The insight that a closed-loop situation might
lead sometimes to "wrong" conclusion for certain experiments, important
as it maybe to the developmental history of certain individuals, is
perhaps not so significant to the whole picture and can be considered
as an additional noise in the experimental setting, which might be
marginal compared to the problems inherent in experimenting with
"other minds", which are shared by all approaches, including PCT.

--Oded

2

[From Rick Marken (980214.0820)]

Oded Maler (980214)

I think PCT suffers from a lack of sympathetic criticism

I think it suffers from a lack of understanding, as witness your
own comments on the matter.

Even when a component is usually embedded in a closed loop
situation, it might be useful to know its I/O characteristics,
which puts limits on its performance in *any* context.

This was not the issue. The issue was _how_ you measure the I/O
characteristics of an intact closed loop system. PCT shows that
the I/O relationships measured in standard psychophysical studies
are actually O/I relationships (that's the behavioral illusion).
Also, the "I" variable manipulated in these experiments is not
necessarily the same as the perceptual input to the system (this
was Bill's point), so, again, the I/O relationship measured is
not necessarily the same as the actual I/O characteritics of the
system. The solution to this problem, provided by PCT, is to Test
to determine the actual I (controlled) variable in the experiment.

Best

Rick

ยทยทยท

--

Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken/

3

[From Bill Powers (980214.102q1 MST)]

Oded Maler (980214)]--

To the topic itself, if you accept the view that the human brain
is a large black box about whose internal structure you can have
only a crude idea (HPCT), that you have only a crude access to
its I/O ports, and that you yourself are a similar kind of
mechanism, but with a significantly different wiring and "reality
tunnel" (I borrow from Wilson), you should know the limits of any
method to infer something about it function and mechanism.

I look at it the other way around: the environment is the big black box
(including the physics and chemistry of brain and body), while perception
is always open to immediate inspection. We live in the world of experience,
in direct contact with the outputs of our perceptual functions, inside our
brains. All the inference we have to do is about the physical world, not
about ourselves.

Best,

Bill P.

4

[From Bruce Gregory (980214.1557 EST)]

Bill Powers (980214.102q1 MST)

I look at it the other way around: the environment is the big black box
(including the physics and chemistry of brain and body), while perception
is always open to immediate inspection. We live in the world of experience,
in direct contact with the outputs of our perceptual functions, inside our
brains. All the inference we have to do is about the physical world, not
about ourselves.

A commendable economy of expression!

Bruce

5

[Martin Taylor 980215 18:50]
Rick Marken (980214.0820) to Oded Maler (980214)

I think PCT suffers from a lack of sympathetic criticism

I think it suffers from a lack of understanding, as witness your
own comments on the matter.

Even when a component is usually embedded in a closed loop
situation, it might be useful to know its I/O characteristics,
which puts limits on its performance in *any* context.

This was not the issue. The issue was _how_ you measure the I/O
characteristics of an intact closed loop system.

Actually, this was _precisely_ the issue. The _components_ of a loop
have I/O characteristics that can be measured. The intact closed loop
system does not, at least not in the everyday sense of I/O characteristics.

When you put together a set of components for which you know the I/O
characteristics, you can make a loop for which you can compute the
behaviour at its two outputs when its two inputs are varied. I don't
think you can do the reverse (determine the I/O characteristics of
the components when you know the four-way relationships among the
loop inputs and outputs.

PCT shows that
the I/O relationships measured in standard psychophysical studies
are actually O/I relationships (that's the behavioral illusion).

Wrong, as usual. But it would be easier to dispute such claims if they
were accompanied by reasons rather than simple re-asserted.

Also, the "I" variable manipulated in these experiments is not
necessarily the same as the perceptual input to the system

The input to any function is not the same as its output. So what's your
point?

(this
was Bill's point)

And what was his point?

, so, again, the I/O relationship measured is
not necessarily the same as the actual I/O characteritics of the
system.

"The system"? What "system?"

What is measured in the typical psychophysical study is a set of partial
derivatives of a function relating the physical variable to some
unspecifiable variable internal to the organism. In PCT we call such an
internal variable a "perceptual variable."

The solution to this problem, provided by PCT, is to Test
to determine the actual I (controlled) variable in the experiment.

I still have great difficulty in understanding what it would benefit
PCT theory or practice to discover that the subject was controlling for
a self-image as a competent subject (or for satisfying the experimenter,
or ...).

I have no difficulty in understanding how it benefits PCT to learn that
(approximately) the ability to discriminate a change in acoustic intensity
is proportional to the base intensity level, rather than being a fixed
number of microwatts per square meter (or whatever the apporpriate unit
is). That tells you a lot about what control is possible for any
perception that has acoustic intensity as an intrinsic input variable.

I'm unclear why you think it benefits PCT to know that subject A is
controlling a perception we might call "pride" whereas subject B is
controlling for an imagined perception of more money at the end of the
experiment, when both control systems use the same mechanism--getting
the most possible correct in a set of forced-choice trials.

And I'm unclear why you think it does not benefit PCT to discover that
both Subject A and Subject B can detect, say, acoustic intensity increments
at least as well as 0.5 dB when the level is 30dB _and_ when the level
is 70 dB, but seem (if they are really trying) not to be able to do
better than that at either base intensity.

Martin

6

From Oded Maler (980216)-2]

[From Bill Powers (980214.102q1 MST)]

I look at it the other way around: the environment is the big black box
(including the physics and chemistry of brain and body), while perception
is always open to immediate inspection. We live in the world of experience,
in direct contact with the outputs of our perceptual functions, inside our
brains. All the inference we have to do is about the physical world, not
about ourselves.

Yes, but we are talking about the special activity called "studying
human behavior" in which the parts of the physical world we investigate
are populated by entities (presumably) like us.

--Oded