This paper takes a critical look at the deepest foundations of cybernetics and particularly at our thinking about what sorts of things really exist, and which of them are needed for a complete cybernetics.
On a diagram of the main quadrants of cybernetics, it is obvious that much progress has been made with systems that are governoid or designoid, by action internal to themselves-examples would be in artificial intelligence and in evolutionary biology. But far less progress has been made with systems that are governed or designed by an agent external to themselves-examples would be in management cybernetics and in the cybernetics of politics.
The heuristic process is held up by the lack of a proper way of comparing theorising with reality - and a lack of proper foundations to provide heuristic productivity.
Mechanisms and principles of dynamics are the main entities in cybernetics (constructions from nexus).
Ontology is the branch of metaphysics concerned with the study of what sorts of things exist and how they may be arranged in categories.
Technically speaking, my main theme will be about menus of ontological commitment and why we should be using a stratified rather than the old furcated ontology. It starts with an account of the relation of scientific concepts to reality (the topic that got people so excited at the recent Conference) quoting the usual four views: positivism, instrumentalism, idealism, and realism. It goes on to show why the subject of cybernetics, because it uses the concept of 'information', has from its earliest days required a new ontological menu but has often failed to recognise the fact. If we do not understand the implications of needing a modern ontology, we condemn ourselves to repeat the mistakes of an ancient one. In passing, I use Ternality Theory as an example-by showing that it is essentially instrumentalist, and uses the new ontology.
Most people keep with the ancient ontology (matter and mind) and think of the menu as being whether to eliminate one of them. This brings problems of reduction. However, the natural outcome of modern discoveries and ways of thinking is to rotate the structure to give a layered ontology, and the menu is about choosing which layers to add.
A modern ontology requires categories for memes, policies, conventions, and other entities involved in mechanisms of purposive or directed activity. A stratified , menu will require that the issues raised by the furcated menue be dealt with: the place of mind, the place of imparity, the kinds of reduction, the kinds of constructions and independence.
My contribution has been to take the ontology implied by the arrival of informtion concepts, show that it involved the addition of a second 'domain', and show how most of our problems can be solved by adding a third domain.