[From Bruce Abbott (951018.1530 EST)]
Bill Powers (951018.0615) --
RE: Rick Marken (951017.2200) --
Well, Rick has finally said the magical words: _relevant_ side effects.
I suddenly see that the core of this whole argument is one of those
language misconstructions that seems to accompany the decline of
precision. The term "irrelevant side effects" is being construed in the
same way as "a terrible disaster" or "an unpleasant stench." The first
adjective is taken as merely an emphasis of the following noun:
disasters are terrible, stenches are unpleasant. By this interpretation,
side-effects are irrelevant.
Well I, for one, would like to know whose interpretation this is. It
certainly isn't mine, although I did argue that others not familiar with the
technical meaning of this term within control theory might misconstrue its
meaning. My recollection is that this was Rick's view, at least at an
earlier time.
The term "irrelevant side-effect" is being taken in a way that doubles
this misinterpretation. To say "side-effect of behavior" is taken to
mean that all effects of behavior are side-effects. Morover, all side-
effects are being taken as irrelevant. So to say "irrelevant side-
effects of behavior" is, according to this construction, to say that all
effects of behavior are not only side-effects, but are irrelevant.
Yes, my point exactly. Now if you could just pursuade this person to stop
using the term in this erroneous way, we'll all be happier.
What Rick has finally brought out is that behavior has many effects: a
main effect relative to the control system under immediate consideration
(through what we call the environmental feedback function), and two
classes of side-effect: those relevant to some other control system in
the same organism, and those irrelevant to any control system in the
same organism. The latter are what we term irrelevant side-effects; that
is, effects which are side-effects because they do not relate to the
control system under immediate consideration, and which are irrelevant
because they do not relate to any other control system in the same
organism, at least in any immediate way. We might even add a third
category: those effects of behavior that are relevant to some other
organism's control systems, which does not tell us whether they are the
main effect, a relevant side-effect, or an irrelevant side effect for
the organism producing the behavior. However, when we speak of the
relevance of effects of behavior, we are normally speaking from the
point of view of the behaving system and what it can know.
Rick must have read my previous couple of posts, wherein I argued this point
(that much observable behavior is not a side-effect of control, as here
defined). I can only say I'm delighted to see that he--and you--agree with
me on this.
Rick has pointed out that the activities of termites may have two kinds
of effect, one exemplified by the laying down of pheremones and the
other by the creating of a mound that maintains their body temperature
at 70 degrees (a relevant side-effect). There is, in addition, at least
one irrelevant side-effect that could be misconstrued as relevant: an
termite on top of the mound could see farther than a termite on the
ground, and thus see enemies approaching from a greater distance. I
assume that termites identify enemies mainly by smell and that their
vision is ineffective over any but very short distances, so that taking
a position on top of the mound wouldn't do the termite any good
concerning the approach of enemies, even if termites could understand
the concept of an "approaching enemy." So even though it's true that
being on top of the mound provides lines of sight between the termite
and the surroundings, that side-effect of producing the mound is
irrelevant from the standpoint of the termite. It might prove to be
happily relevant from the standpoint of an anteater.
I couldn't have said it better.
So--
Effects of behavior:
a. Main effect on controlled variable of system directly producing
the behavior (relevant).b. Side-effects of same behavior (all relevant):
1. Effect on another system's controlled variable in the same
organism at the same level, a disturbance (relevant
to other system, or both if there is a reciprocal effect)2. Effect on controlled variable of higher-level system in
same organism (relevant to higher-level system).3. Effect on intrinsic variable in same organism (relevant to
reorganizing system).c. Irrelevant side-effects of same behavior
1. Effects on environment or other organisms that do not
alter any controlled variable in the system producing
the behavior.
Wow, my position exactly. So, now that we all agree, let's get back to my
reason for arguing for these distinctions. SOMEONE (certainly not me!) has
asserted what can be reduced to the following syllogism:
1. The phenomena studied by "conventional" behavioral scientists are
merely the irrelevant side-effects of control.
2. Irrelevant side-effects of control are of no interest in PCT.
3. Therefore, phenomena discovered by "conventional" psychologists (like
those labeled "reinforcement," "attachment," "modal action pattern,"
etc.) are of no importance or relevance to the problem of understanding
the complex biological control systems of animals and people.
But we've just seen that much observable behavior cannot be construed as
"irrelevant side-effects of control," as now defined. Therefore, not all
behavioral phenomena studied by "conventional" science take the form of
"irrelevant side-effects," and the above syllogism is proven false. My
thanks to Rick for providing that excellent insight.
Yer honor, the defense rests.
Regards,
Bruce