[From Rick Marken (930110.1030)]
Martin Taylor (930108 14:00) --
Quite apart from that, I have a real problem with how a PCT approach to
psychophysics could differ from a conventional approach. In a typical
psychophysical experiment, a subject is asked to discriminate between
two possibilities for which event was presented on a particular trial, or
to react as fast as possible with one of two discriminative responses to
a pattern that might be of type A or of type B.
I agree with this description of the typical psychophysical experiment.
You go on to say:
There's no opportunity
for the subject to reduce the error of a perception.
What error of perception? Error is inside the subject (if it exists).
This statement implies that you think of "error" as an objective
phenomenon; but it is not (for you or the subject). Error (for the subject)
will exist if the subject is controlling a perception that cannot be
made to equal the subject's reference specification for it. What
perception(s) is the subject controlling in a psychophysical experiment?
Well,that's the question -- it can only be answered by research -- testing
for the controlled variable. As it sits, we can only guess what the subject
MIGHT be controlling in such experiments. One likely possibility (in
the "discrimination" experiment, for example) is a RELATIONSHIP, between
stimulus perception and the imagined meaning of the "discriminative"
response. I would imagine that the subject's experince of how well s/he
is controlling that variable is pretty poor -- and, indeed, the
relationship between stimuli and responses is INTENDED by the experimenter
to BE noisy -- the subject SHOULD have poor control (an exception is
magnitude estimation, where it is MUCH earier for the subject to control
the relationship between stimulus and response -- and the data are
correspondingly better, though not perfect -- probably because this is
a difficult relationship to control, if that IS what the subject is
controlling -- in magnitude estimation it seems like that is what the
subject is trying to control).
The action is
performed, the experimenter records which action it was, and the whole
cycle starts again with no informational connection between the last
action and the next presentation.
This is your description of the situation as an observer; if you could
see the variable controlled by the subject things would look QUITE
different (but then you would not be a conventional psychologist
anymore). A person does not become an S-R device just because you
put them in an experiment where (relative to YOUR goals) it appears
that they MUST act like one. If people are organized as control systems,
then they are always controlling -- and S-R relationships are just
evidence of disturbance compensation. The more reliable the S-R relationship,
the clearer the evidence of a controlled variable because the subject is
reliably compensating for disturbances to that variable. That's why I think
that the variable controlled in magnitude estimation is the relaitonship
between S (as perceived) and R (in terms of its imagined representation of
magnitude -- if it's a number) -- it's because the quality of the relation-
ship between S and R is so high. The S-R relationship in magnitude
estimation is like the relationship between the cursor and the target
in pursuit tracking -- as the subject controls the relationship between
S (target) and R (cursor) (keeping S-R = 0) , R is highly correlated
with S (as in magnitude estimation). The same is true of rating data
(stimulus discrimination); I got pretty high correlations between S (sound
burst enery) and R (ratings of signal presence) because the subject
was controlling the relationship between these two variables.
The only feedback that can do anything
is an indication as to whether the action was right or wrong,
Feedback doesn't really DO anything -- it is always there. This sentence
suggests that you buy into the notion of feedback adopted in most
psychological studies -- as something that is GIVEN to people (here,
have some feedback, dear, it's good for you). That is not the meaning
of feedback in control engineering OR PCT. The feedback you are talking
about is just an independent "stimulus" event -- it is a DISTURBANCE
(or a potential disturbance; it is a disturbance is if it influences
the perceptual variable that the subject is controlling).
and that
feedback can only contribute to reorganization, not to the actions or
perceptions on the next trial.
Not at all. Since the feedback is a potential disturbance it might be
related to the responses that the subject is using to control whatever
perception is being controlled; and indeed it does influence the subject's
responses, and rather systematically. The subject is apparently trying to
hear the word "correct" said ALL THE TIME -- perhaps as an indication that
the relationship variable is being controlled to the experimenter's
satisfaction (another controlled variable -- "being a good subject"). The
influence of this kind of feedback is clear in discrimination experiments
where the subject is given "false feedback"; the subject's responding changes
in the way that you would expect if one variable the subject was controlling
for was "hearing the word 'correct'". Of course, all this data is fairly
noisy so it's not very useful for determining controlled variables -- but
it does make the psychophysicist happy because s/he thinks that variance
reveals something interesting about the workings of the sensory systems.
It reveals very little about what is being controlled.
Why should taking a PCT approach make any difference to the way psychophysical
experiments are done or interpreted?
I hope I have given you some clues. Psychophysical experiments are based
on the assumption that responses are caused by stimuli -- rather than the
assumption that people control sensory input. The best you can get out
of such experiments (if people really DO control sensory input) is a
relatively noisy indication of disturbance-output relationships; but
not nearly as noisy as those seen in other kinds of experiments (and
psychophysical relationships --like the one's I found in my thesis --are
often based on the anaysis of one subject at a time -- which is very good
from a PCT modelling perspective); so the results of psychophysical
experiments are not chopped liver -- but they are not haute cuisine either.
In response to Greg Williams I said:
The PCTer's
problem with the typical S-R model is that it says that behavior (R) is
the last step in a causal chain that starts in the environment, at the
sensory surface or in the brain (the latter being the reason we know that
cognitive "models" of behavior are really S-R).
Greg (920109) replies
As I said above, I'd like some documentation on this claim. Maybe you can
direct me to some typical behaviorists to ask whether they think there can
never be ... S-R-S-R-S... situations?
You seem to treat the writings of Skinner and other behaviorists as though
they were religious texts. I personally don't much CARE what these people
SAY; you can find a quote in Skinner to show that he knew all about PCT
as easily as you can find a quote in genesis to show that the Biblical
author (God?) knew all about evolution or quantum physics.
What I care about is what people DO. There are several psychologists
running around saying that they are PCT theorists; they SAY all
the right things in the introductions to their articles and books. But
I know they are NOT PCT theorists becasue they don't DO PCT research.
(They also don't do modelling but I'm being lenient). The same holds
for all conventional psychologists. Many of these psychologists talk
about studying purposeful behavior, intentionality, etc etc. They
RAIL against the impoverishment of behavioral psychology and claim to
be setting off to look at behavior from a bold new perspective. BUT
they do the SAME OLD KIND of research, which assumes that behavior is a
DEPENDENT VARIABLE and sensory input is an INDEPENDENT VARIABLE; the
bold new perspective turns out to be Descartes's reflexology in modern
dress.
So I'm not going to waste my time trying to find quotes in the
scripture of conventional psychology (or life science in general)
to support my claim that they believe in an output generation
model of behavior. You can prove ME wrong about my claim by
pointing to ONE -- just ONE -- study in conventional psychology
that involves testing an organism to determine what perceptual
variable it is controlling for; just one.
Best
Rick