NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/bit/listserv/csgl/2056

Text File | 1993-01-06 | 8.9 KB | 177 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!zaphod.mps.ohio-state.edu!malgudi.oar.net!news.ysu.edu!psuvm!auvm!VAXF.COLORADO.EDU!POWERS_W
Return-Path: <@VMD.CSO.UIUC.EDU:[e-mail]>
X-Envelope-to: [e-mail]
X-VMS-To: @CSG
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Message-ID: <[e-mail]>
Newsgroups: bit.listserv.csg-l
Date:         Tue, 5 Jan 1993 16:34:43 -0700
Sender:       "Control Systems Group Network (CSGnet)" <[e-mail]>
From:         "William T. Powers" <[e-mail]>
Subject:      Devil's Advocate
Lines: 161

[From Bill Powers (930105.1530)]

Greg Williams (920105) --

>For the tracker to "respond" to the "discriminative stimuli,"
>all that is necessary is for him/her to be able to see the
>cursor movement, NOT to "tell... WHAT THE DISTURBANCE IS." If
>the cursor is seen to be moving away from the target position
>-- due to the net COMBINATION of handle position and net
>disturbance, of course -- then the tracker responds by moving
>the handle in the direction (determined previously in practice,
>via "reinforced" learning) which moves the cursor in the
>direction toward the target position.

How about "the tracker sees an error between the cursor position
and its intended position, and responds by moving the handle at a
velocity proportional to the amount of the difference and a
direction corresponding to the direction of the difference?" This
is a verbal description of the organization of the control
system.

Your way of putting this assumes that the intended position of
the cursor relative to the target is AT the target. It is
perfectly possible to move the cursor so it remains a fixed
distance to either side of the target. This makes the definition
of a disciminative stimulus somewhat difficult, because at that
specified distance from the target, most of the time, one can see
-- nothing. The stimulus now has to be defined as the distance
between the cursor and an arbitrarily-located empty place in
space, or alternatively as the distance of the target from that
empty place minus the distance of the cursor from that empty
place. No matter how you put it, the discriminative stimulus has
to be imaginary.

This mistake has been made many times in the past -- the view
assumes that some "salient" (meaning obvious-to-me) aspect of the
situation is the reference condition, forgetting that this
condition is just one point on a scale, and therefore not
realizing that control could take place relative to any position
on that scale. This is how people have concluded that reference
signals come from the environment. That's another myth that got
launched in the '50s.

The cursor position and velocity always reflect the ongoing
behavior of the disturbance PLUS the ongoing behavior of the
handle. If the cursor begins moving slowly to the right, this
could indicate that the disturbance has started pushing it to the
right a little faster than the handle is pushing it to the left,
or that the handle has started pushing it to the left a little
slower than the disturbance is pushing it to the right. The
information required to make even this qualitative judgment is
not contained in the cursor position or velocity. You must
perceive your own handle movements directly and estimate how the
cursor would be moving and where it would be positioned if your
handle were the only influence.
In Demo1 there is a phase in which the difference between
compensatory and control behavior is illustrated. In compensatory
behavior the "cursor" on the screen shows the disturbance
magnitude, not the actual cursor position. The task is to
estimate where to put the handle at each instant so that the
effect on the now-invisible cursor would keep it from being
disturbed. This is impossible on the face of it, so the
demonstration shows a trace of what happened to the real cursor
during the run (afterward), and you can also alternate with
controlling the real cursor so you can pay attention to how your
hand moves and learn how much it needs to move and where the
center of movement is. By using all this (higher-level feedback)
information over may trials, you can actually improve your
performance in the compensatory phase quite a lot. You can, with
a lot of practice, get the RMS error in the invisible cursor
position down to only about 10 times what it is when you can't
see the disturbance but can see the cursor.

Even with all this practice, you can't estimate your handle's
effect on the cursor well enough to achieve the kind of control
you get without having to pay attention to the handle at all and
without any direct information about the disturbance magnitude.
In the compensatory case you don't have to estimate the amount of
disturbance by comparing felt handle position with seen position
of the cursor. You are given an exact quantitative picture of the
disturbance magnitude. And you still can't achieve the
performance of a control system within less than a factor of 10
worse.

Furthermore, when the disturbance is shown on the screen as a
pointer AT THE SAME TIME YOU ARE CONTROLLING A VISIBLE CURSOR,
your tracking performance is not measurably different in most
cases when the disturbance information is eliminated (this is not
in Demo1 but I have done the experiment). The only case in which
some measurable difference can be seen is when the participant
pays attention to the disturbance information and tries to use it
to improve control. In that case, the quality of control
deteriorates sharply until the person ceases to pay attention to
the disturbance information.

All these demonstrations, which I have actually done and which
are easily reproducible, show that the person is not making any
use of information about the disturbance, either directly when it
is available on the screen, or indirectly by estimation of
expected handle effects on the cursor.

These facts about control are easily demonstrated, but to
understand them you have to think quantitatively. As long as you
are allowed to talk about "cursor movements" and "handle
movements" and "target movements" you can gloss over these
quantitative facts because the language doesn't specify HOW MUCH
movement there is in relation to other movements. In a subtle
way, you're using the known outcome of the experiment to provide
just the meaning for the general terms that is needed to make  them fit the
 observations. It would not be possible to go the
other way: if you didn't know how the experiment came out,
describing the relationships in terms of "movements" and
"positions" and other such qualitative notions wouldn't tell you
anything about which way the cursor would move. The cursor
movement is a small difference between two variables that are
changing over a range of perhaps 20 times as much (RMS
comparisons). Just saying that the disturbance moves the cursor
to the right while the handle tends to move it to the left leaves
the direction of the actual cursor movement undefined.

I tell you: the disturbance is changing to push the cursor to the
left while the handle is moving to push it to the right. Which
way, pray tell, do you predict that the cursor will be moving?

In your Devils's Avocacy, you have brought out exactly what is
wrong with conventional objections to PCT. It is the qualitative
nature of the arguments that makes it seem that alternative
explanations fit the facts. Qualitative descriptions are crude
enough to allow for any outcome at all -- or its opposite.
Qualitative explanations, if cleverly enough constructed, are
unfalsifiable. This is the attraction of generalization and
qualitative description: you can't be wrong.
-----------------------------------
>One last time: don't take high correlations as THE sign of
>stimulus-response relationships.

Why not? That's what THEY do, isn't it? This brings out the main
reason we can't talk to conventional behavioral scientists. If
you show them a tracking experiment, the first thing they will do
is look for high correlations: this behavior is a response to
that stimulus. When we carefully set up the experiment to show
what actually happens to the correlations, what do they do? Do
they say "Oh, migosh, it looks as though I have the wrong
explanation!"?

In a pig's eye. They immediately back off, and say that this
situation is more complex than it appeared, and requires a
different explanation. They abandon the simple analysis of simple
experimental data and start talking about vague effects of
discriminative stimuli and reinforcements, all of which somehow
have exactly the effects required to account for the experiment
-- for the cursor going up a little here, down a little there,
and wiggling just so in between while the handle traces out an
almost perfect mirror image of the invisible disturbance.

The real problem is that such people don't have any idea of what
a real explanation amounts to. They have given up on science.
---------------------------------------------------------------
Best,

Bill P.