KFT, Rick on Springs

[Avery Andrews 930416.1048]

Noting the reservations that various people have expressed (there are of
course always reasons why what you propose to do might not work), here's
roughly what I have in mind to try.

a) present hierarchical systems of ECSs as an implementation model
    for coordinative structures.

b) argue that there are various different ways in which a given
    c.s. can be implemented, and that in principle, at least some
    of these can be empirically distinguished by quantitative
    experimentation.

c) counter the KFT (Kugler, Fowler & Turvey) philosophical line by
    arguing that it's better to have a definite tho possibly wrong
    proposal about implementations than none at all.

It goes without saying that it's today's young innocent faces who are
the target audience, not established KFT figures.

The reason I'm thinking `Synthese' is that it is read by people with a
diverse range of interests, rather than being the preserve of one of
the established psychological subdisciplines (it's my impression that
the best and the brightest look these things over and decide to do
something different).

Enough plans & philosophy. Here is a little nit to pick with Rick
Marken (having been a parent of young children, I now see nitpicking
as a fairly serious activity):

In `Degrees of Freedom in Behavior' (Mind Readings, p. 187), Rick
says that coordinative structures and mass-spring systems can't
explain resistance to continuously varying disturbances:

`.. a continous disturbance, such as an oscillating force applied
  to the mass, eliminates stability--the position of the mass changes
  continuously until the disturbance stops.'

It seems to me that this depends entirely on the spring constant. If
the spring in question is a piece of battleship armour, the disturbance
exerted by my arm muscles, the changes in position of the mass will be
difficult to observe (unless you're using diffraction-grating-making
equipment, of course). If the time scales are large relative to
the transport lag around the loop, a control system looks pretty much
like a mass spring system, the main difference being that it is easier
to tune the gain, the extent to which it is linear, etc. What Rick
is doing is comparing a soft spring with a high-gain control system,
which is not a valid comparson.

What I see as wrong with the point-attractor stuff is that they want to
regard questions of implementation as beyond the domain of inquiry,
which see as a built in requirement that the investigation lead nowhere
(another great Chomsky phrase), since it's exactly the fiddly details
of implementation that produce quantitative predications that can be
tested.