From Tom Bourbon (930616.1020)
Replying to Hans Bloom [Hans Blom 930616] and Oded Maler [From
Oded Maler (930616)].
Both of you are pressing the idea that cells in bodies are like people in
social systems. That may be true in all cases, in some cases, or in no
cases. It would be easier for some of us (me, in particular) to imagine
what you are saying were you to present your claims in the form of PCT
models, even if they are only preliminary drawings, as I suggested
yesterday. Drawings like that are a first step to nearly all of my attempts
at modeling; they are my first try at "checking out" my ideas,
intuitions, and inspirations. The majority of my brilliant ideas about
modeling have died at that stage -- ideas that somehow "felt" right simply
could not be turned into workable models. On the other hand, things I
originally thought I could not model, like some of my interactive tasks,
turned out to be easy, once I had thought my way through drawing the
relationships.
I know that both of you are familiar with this first step in modeling.
Could the two of you go at least that far toward clarifying your thoughts?
At least one of us out here would have a better idea what you mean.
As an example, it may be true that there are Social Reference Signals
(Great Big P*s), developed and handed down over thousands of years, or even
over the brief life span of a fashion fad. If so, where do I draw P* in a
diagram that shows, in a simple case, two people who are affected by
P*? And where do I draw the arrows from P* to the simple loop representing
each of the people? If I can't imagine that preliminary step in
modeling, there is no chance at all that I can go further and produce a
formal model and simulations.
As another example, I can easily imagine at least two ways aggregates
of cells in a body might act in concert. They also happen to be two ways
aggregates of people might act in concert. It is possible to produce
diagrams of very simple examples of the two cases and those diagrams can
be turned into extremely simple models that can be run in simulations.
CASE A. Each entity (cell, person, model, device, etc) has its own ref.
sig. (p*) and controls its perception (p) specified in p*. In the process,
each entity also affects at least one variable controlled by at least one of
the other entities in its neighborhood. (This is like the relationships in
CROWD-GATHER and in my articles on interference among control systems.) In
this case, through its actions, each entity controls its perceptions and
interferes with variables affected by its neighbors, which also control
their own (disturbed) perceptions. It is easy enough to diagram two or
three entities in such an interaction.
How might sources other than the entities or the variables they control
affect this relationship? There are at least two possibilities. First, a
"locally universal" disturbance might affect the variables controlled by all
of the entities (smoke suddenly fills the room, the experimenter jostles
the petri dish, toxins flood the region, the earth shakes beneath our feet,
the programmer set the program so that at t120 the positions of all
stationary people in CROWD-GATHER are randomly reassigned ...). Easy enough
to diagram and (in simple cases) to model. Second, if the proper
connections exist, a "remote" source might "reset" the reference signals
inside each of the entities. Easy enough to diagram and model, assuming the
proper connections. (Diagramming begins to force the issue of the
plausability of connections we assume.)
CASE B. Each entity (cell, person, model, device, etc.) at level 0 has its
reference signal set by error signals from entities at level 1. The p* at
level 1 specifies a perception that does not exist at level 0. Similarly,
the error signal from level 1 is the p* for a perception that does not
exist at level 1. Diagrams? Models?
Unlike CASE A, when I try to draw the diagrams for this case, I have some
problems. If the entities at level 0 are, say, muscle cells, everything is
fine; if they are people, I encounter big trouble.
Hans and Oded, could you give us at least a few simple diagrams of the
relationships you are asking us to consider? If they look manageable,
someone might try to produce models and simulations. If the ideas passed
that step, then it would certainly be true that people in social groups and
cells in bodies can be modeled by the same PCT systems. If we don't at
least try to reach that step, we will merely exchange opinions on the
subject. That might be enjoyable for all of us, but it will do nothing to
resolve the important question.
Until later,
Tom Bourbon