The other similar conjectures are in historical order:
Pask's P-individuals, probably the clearest exposition is in his
book Conversation Cognition and Learning . (1975) Elsevier.
The idea is that agents engage in comunicative exchanges with
respect to some commonly accessible object, and when they achieve
a large degree of coherence in their representation of their object
both participants together with their representations constitute
a meta-participant for meta conversations. Th participant individuals
may be localised intra-body entities, or distributed trans-body
entities.
The agents in Marvin Minsky's society of mind are rather low-level
gofer sorts of things compared to Pask's P-individuals.
A physiological basis for your mimickers or either of the above, might
well be provided by Geral Edelman's extensive neuronal groups.
(see Bright Air Brilliant Fire (1993)).
Rom Hare's Ethogenic Psychology involves the social/conscious/subconscious
sandwich of actiors and agents rather nicely, see chapter 2 of
Motives & Mechanisms. Rom Harre', David Clarke, Nicola DeCarlo.
London, Methien (1985).
My own view is given in my paper in the Pask Festschrift issue of
Systems Research vol 10 ,4 I think (1993).
Gary Boyd.
.