What ontological categories does cybernetics need?
Ontology is the branch of metaphysics concerned with the study of what categories of things exist and how they may be arranged in menus to which we can choose to be committed.
Throughout history, most people have believed that reality ultimately consists of two categorially different kinds of entities, namely bodies and minds (or bodies and spirits). This idea is reflected in the forked, or furcated, ontology offered by ancient metaphysics, with its single discontinuity between two concrete forms of reality, giving rise to two branches of study: on the one hand cosmology, or the philosophy of nature, and on the other psychology, or the philosophy of mind.
The furcated menu of ancient metaphysics
It can be argued that the natural outcome of modern discoveries and ways of thinking is a rotation of the ancient furcated structure, giving instead a layered, or stratified, ontology, and that the menu is now about choosing how many layers are needed, and which ones.
It follows that we ought to be using a stratified ontology in cybernetics. I have proposed that we construct such a stratified structure by making use of two discontinuities, each of which was already separately established in the literature, but which had not previously been used together in this way (Stewart, 1989).
The stratified menu
Obviously, we will require the stratified menu to deal with the issues formerly raised by the furcated menu, including: the place of mind, the place of imparity, the kinds of construction of new categories of entity, the kinds of independence that such entities may achieve, and the kinds of reduction possible.
We regard cybernetics as being a branch of science, and so expect it to consist of a structure of propositions, in which those of lesser generality are built upon a few propositions of great generality. It may be a useful exercise to ask ourselves which are the propositions of great generality in cybernetics? What are cybernetics' most general principles?
The primary domain
Perhaps your answer would include Newton's Three Laws of Motion, the First Law of Thermodynamics, and all of the rest of classical physics. From those laws, and the entities with which they are concerned, one can construct what we shall call the primary domain. A primary system is regarded as being closed to energy, and a significant attribute of it is the quantity of matter or of energy involved in it. Thus a primary machine would be one designed to process energy, such as an electric fire or a car engine.
The secondary domain
Perhaps, among the most general principles of cybernetics, you would also want to include the Second Law of Thermodynamics, or Shannon's Tenth Theorem, or Ashby's Law of Requisite Variety. If you follow Ashby's view that cybernetics is concerned only with systems that are open to energy but closed to information (Ashby,1956), then you might mention only this second group—in other words, the laws of the type that were added to physics by the advent of thermodynamics—as being the only ones relevant to cybernetics. Let us call those laws, and the entities with which they are concerned, the secondary domain. A secondary machine would be one of which the important capability is that of processing information, and it does not matter—within reason—what quantity of energy is involved. A television set and a computer would be examples of secondary machines.
The primary and secondary domains, as we have just defined them, include everything that is usually regarded as necessary to complete most branches of science. However, at this point we have to ask whether the entities of the primary and the secondary domains are all that we need for the ontology of a complete cybernetics. Or are more required—more than are needed for any other branch of science? Is cybernetics special in this respect?
The tertiary domain
Any relationship between systems which involves a manager or a designer of any kind—or any other sort of agent system—raises the question as to whether his purposiveness is still adequately dealt with by the traditional two-domain cybernetic account. This is the question which was originally raised by Taylor (Taylor, 1950a) (Taylor, 1950b) as an objection to the mechanistic conception of purposiveness put forward by Rosenblueth, Wiener and Bigelow (Rosenblueth et al,1943), and it has not so far been satisfactorily answered.
I have argued that purpose requires imparity, and that imparity is separated from information by an unbridgable discontinuity (Stewart, 1989). Therefore, mechanisms have to be conceived of as ternary in structure if they are to be used in the cybernetics of management or of design. It follows that, in order to accommodate ternary mechanisms, we need to have a ternary conceptual structure. The implication is that we need a third, or tertiary, domain.
Another reason for taking the view that classical physics, plus the addition of thermodynamics, does not give a sufficient ontological infrastructure for cybernetics is our evident continued worry about the ontological problem of consciousness—which nowadays usually takes the form of controversy about the mind-brain relationship.
Surface and deep dualism
Most people continue to adopt the ancient ontology—of matter and mind—and think of the menu as offering a choice of whether to eliminate one of them and, if so, which one. Even if they manage to avoid a surface dualism by denying one form of reality, they may still fall into deep dualism by accepting the validity of the menu. Those who have rejected a surface dualism, while retaining a deep one, are usually much occupied with questions of reductionism. I have shown that, due to the operation of Hume's Law, although it is in an important sense mechanistic, the proposed theoretical structure based on ternary mechanisms—including imparity as well as energy and information—is not reductionist.