1.2.2. The M model is a canonical form of representation with two salient features. The first of these is an „hierarchy of formal metalanguages“ or „a stratified linguistic structure“ in which we embed descriptions of concepts and cultural constraints and intelligent communication. Now organisms do not actually use formal languages. They communicate in terms of open and referentially mixed languages. But a description of their activity in these terms is ambiguous. So the hierarchy of meta languages is introduced to avoid this ambiguity and (as in 2.2.4) to achieve a unified and simple image of the mechanism that is responsible for the communication process.
Next, the M model has two components, namely a mechanistic and a descriptive component. When the M model is experimentally identified (when it acts as an M system) these components are associated with incompletely comparable sequences of observations that (as in 2.1.4.) refer to distinct ontological classes. However, as the model is abstracted from reality the calibre of the distinction changes. When the M system is identified with a computer programme, for example, the mechanical component becomes the intensive definition of part of a formal language or, as Gorn (1962) points out, of the data processing devices that act upon it. The descriptive component becomes the extensive definition of this language. (This interpretation is pursued in 2.1.6.). At the most abstract possible level the distinction between these components is simply a distinction between the intension and the extension of relevant terms in the linguistic structure (this is the calibre of the distinction in an M model, devoid of an identification).
No attempt is made to particularise the M model but a few special cases (chiefly of learning models) have been worked out and described in the literature. Nor does this paper contain any mathematics though the bones of suitable kinds of calculus have been described by Watanabe (1962), Martin (1963) and others. The chief objective of the present discussion is to exhibit the canonical representation ‚M Model‘, to examine its identification with matters of fact in ‚M Systems‘ and to demonstrate that although more elaborate structures may be needed to describe individual ‚learning‘ and ‚mentation‘ no lesser construct would be sufficient.
2.1. Basic models
2.1.1. One of the broadest concepts shared between Cybernetics and Control Engineering is the idea of an ultrastable or an hierarchically organised and goal directed adaptive control mechanism (Ross Ashby, 1960; Pask, 1963e). To avoid unnecessary symbolism it will be convenient to adopt the graphical convention in Fig. 1 where adaptation is brought about by changes in the parametric coupling between the levels in the hierarchy. These changes are designed to satisfy a dispositional relation, F, named as a goal F. Thus Fig. 1 represents a single level adaptive control mechanism, that interacts with its environment to bring about some condition (normally a dynamic equilibrium or stationary state) that is characterised by an invariant called F.