[Avery Andrews 961003]
I'm still not to happy with Kent's treatment of `conflictive control',
for various reasons.
1) It's extremely wasteful of output energy, too wasteful, I suspect, to
be ecologically viable.
2) If Cindy and Avery are conflictively controlling, say, the position
of a table, they are doomed to stand there forever tugging at it. They
can't put it down and walk away until their references for its position
are harmonized. Such conflictive control locks are pretty rare, amongst
grownups, at least for low-level perceptions. Even amongst young kids
they don't really last that long.
3) Many aspects of organized life are much easier to destroy or disrupt
than to create (dams, skyscrapers, courses given at a regular time and
place). Suppose we have a conflict-ridden university department, where
prof. A wants to put on a course about the Deconstructive Semantics
of Neonatal Infant Cries, while prof. B is violently opposed to this.
If B looses the committee fight, he could go on to sabotage the course
in various ways (stealing the handouts, cancelling the Bookshop's order
for the textbook, hiring thugs to beat up prof. A., etc.) but mostly
this doesn't happen. So B *does not have* a reference level for the
course not being given, at least for that semester (he might be fighting
mightily to get it off the course list for next year, but that's a
different matter).
So I'd say that when you see something getting collectively controlled,
that's because the references for the relevant perceptions are
harmonized (conflict tending to subvert control, in various ways that
Kent points out).
The problem of harmonizing perceptions, i.e. getting `the same
reference' for a perception, is conceptually tricky. Cindy and
Avery have different perceptual systems, how can their reference levels
for the position of the table `be the same'?
Trying to think about this, it seems to me that perceptions (scalar
signals, and maybe in addition logical TRUE/FALSE values) are either
linked (locally, in a given situation) or not linked. Two perceptual
signals (in the same head, or different heads) are linked if a change
in one implies a change in the other; otherwise they're not linked
(independent). In the case of many of the continuous perceptions that
are important at the lower levels of the perceptual hierarchy,
perceptions can be linked in a somewhat stronger way in that each can
be locally approximated as a linear transformation of the other,
e.g. small changes in one map approximately linearly into small changes
in the other.
Linkage is highly local and relative to particular situations. For
example, if Cindy and I are looking at a ball from different angles,
her perception of its size might be locally linked to my perception of
its transverse location in my visual field; if she moves around to
a different position, her perception of its horizontal postion
will be linked to mine, etc. Linkage can be enforced by basic facts
of geometry and perception, or various kinds of mechanical linkages.
Consider for example a situation where an object is moving along a
J-shaped track, (mirror image J, let's say), with the stem of the
(mirrored) J sitting along the y-axis, the bottom of the J touching
the x-axis, at, let's suppose, x=1.
Let critter A be able to perceive the position of the object along the
y axis, and be able to exert forces parallel to that axis, while
critter B can perceive the position of the object on the x-axis and
exert force parallel to that axis. Now if the object is on the stem,
A will have full control, and B will be able to do nothing (B's
perception of the object will be `stuck'), but if A pushes the object
into the curved region, suddenly B will be able to exert some control,
and conflict will be possible. If B is smart, B will notice that
his/her control is pretty good when the perception is 1, but
increasingly lousy near 0 and 2, and that around 0 it can be lost
completely, for long periods of time, and then come back, for no
apparent reason. A might similarly notice that near 0, control gets
worse, and that after an excursion past 0 it can get into or out of
a state (the short upcurve of the J) where the maximimum value is quite
limited. From the point of view of A and B, this is just wierd and
unaccountable properties of the perceptual variables they are trying
to control; I think people find themselves in this kind of situation
when playing with complicated mechanical linkages that they don't
really understand (like sofa beds), or trying to get members of the
opposite sex into a receptive mood, etc.
At any rate, when two perceptions are linked, it's pretty easy to tell
when they are being controlled at conflicting reference levels, and my
belief is that social control happens not because of conflictive
control, but because people are good at harmonizing their references
for linked perceptions. What the mechanisms are I don't really know,
but it's probably basically the same bag of tricks that are used to
deal with immovable objects and other forms for `getting stuck' in
the physical world. E.g. if you can't acheive a higher level goal
with one combination of lower-level references, try a different one.
Avery.Andrews@anu.edu.au