[From Bruce Abbott (991123.1130 EST)]
Rick Marken (991122.0900)
1) The number of unintended consequences of any control
action is large, probably infinite. Any one of these myriad
consequences could be something another person wants the
controlled to take responsibility for.
I agree. However, the number of unintended consequences that a given
individual may actually be held responsible for by others is, relative to
the number possible, small. Most of them fall under the heading of social
rules, which everyone grows up learning -- and generally obeying. Because
they are part of the environment in which the individual develops, it is
likely that the control hierarchy that the person develops will already have
been optimized with their effects and side-effects already taken account of.
2) Taking responsibility for _any_ consequence means setting
up a system to control a perceptual representation of that
consequence. It also implies that the consequence will be
controlled at a particular level. For example, one unintended
consequence of opening the refrigerator door may be a
squeaking sound. One takes responsibility for this consequence
by controlling for it. But "taking responsibility" also implies
controlling the consequence at a particular level. I would
not be seen as "taking responsibility" if I controlled for a
loud squeak; the assumption is that I am "taking responsibility"
when I am controlling the consequence (squeak) at the level
desired by the person wanting me to "take responsibility".
I agree, except for the following: Taking responsibility does not only mean
accepting responsibility for seeing that your actions do not disturb some
variable another person is controlling, it also can mean accepting
responsibility for having disturbed it in the first place. "Rick, you
stepped on my foot!" Rick: "I'm sorry, I didn't mean to."
3) When a person is asked to take control (and set a particular
reference for) any particular unintended consequence of a
control action, they are being asked to set up a new control
system (with a fixed reference for the controlled variable)
in an existing hierarchy of control systems. Based on my models
of hierarchies of control systems, inserting a control system
with a fixed reference into this hierarchy is almost certain
to lead to conflict between control systems. For example, if one
tried to insert the squeak control system (with a reference of
zero, say) into their existing hierarchy of control systems,
it is likely that the actions that had to be taken to keep
the squeak at zero (very slow movement of the door) would
conflict with other systems that are controlling for things
like getting a glass of milk.
Unlike your spreadsheet demo, real people normally are able to quickly and
easily resolve or accommodate to such conflicts, as Chuck Tucker alluded.
As you note, to reduce the intensity of the squeek, you simply open the door
more slowly. This is done by resetting the reference for the speed of
door-opening -- no problem unless a fast speed is required by some
higher-level system (e.g., you're in a hurry). Or, as Chuck Tucker notes,
you can get the can of WD-40 out and oil the squeeky hinge. This eliminates
the conflict, because now opening the door no longer produces the offending
squeek.
The reorganization model seems to imply that _any_ change in the control
hierarchy will have consequences that will require quite a bit of random
reorganization to iron out, but real human beings evidently have much more
efficient methods than random reorganization for accommodating such changes
, as Bill has noted in the past.
4) So my conclusion, based on my understanding of what is
involved in "taking responsibility for unintended consequences
of one's actions" and of Bill's theory of behavior, is that
it is not wise to expect that asking a person to "take
responsibility for the _unintended_ consequences of their
actions" will necessarily increase their power (in terms
of the number of variables they are controlling). In fact,
it is far more likely (given the large number of unintended
consequences that one might want a person to become responsible
for and their arbitrary relationship to that person's existing
control hierarchy) that "taking responsibility for the
_unintended_ consequences of one's actions" is more likely
to reduce one's power (breadth of control) by creating internal
conflict that will reduce one's ability to control variables one
was previously able to control successfully.
Which of these outcomes actually happens in a given situation will depend on
quite a number of factors, probably the most important being whether the
individual being asked to "take responsibility" perceives the request as
reasonable. It would appear that much effort is expended in RTP to foster
that perception in the students.
Regards,
Bruce A.
