Notebooks

Cybernetics

16 Oct 2002 15:27

Old draft from the 1990s, last touched some time before 16 October 2002

A science which seems to have dissolved into the others. A lot of good science was done under this banner; it just doesn't seem to hold together. Cybernetics helped give rise to some new fields, like cognitive science; it disseminated about a dozen ideas and bits of applied math which have proved useful (in, e.g., neurobiology); but what else? As a study of abstract machines in general, it becomes identical with dynamics, or computation theory, or some amalgam of both; algebra, even. As a more limited science of "communication and control" it suffers from the fact that communication and control in animals is, when you get down to blood and guts, rather different from communication and control in machines, and neither resembles the mechanisms of C&C in society. This is not to say that there are no similarities; of course there are; but they're at the very general level of things like "feedback" and "you must have an information channel," and pretty much exhausted by ideas which are now common currency in many particular fields. And even then, animals have control without feedback. It may be that we haven't exhausted the potential of a science of communication and control, but I think at this point the burden of proof would be on the optimists.

Dissolved? Not entirely. There's an old joke that if you're not part of the solution, you're part of the precipitate, and not everything associated with cybernetics has gone into solution. Caked on the bottom of the reaction vessel we find: A prefix which seems indispensible to marketroids; the occassion for a great deal of vaporizing in the social sciences and humanities; and a peculiarly navel-gazing sub-sect of systems theory, which isn't exactly God's gift to the advancement of learning in the first place.

History --- Wiener, Rosenblueth, Ashby, Herbert Simon. Origins in mathematical logic, physiology, engineering, statistical mechanics. Predecessors --- Rashevsky, Lotka, Cannon, Sherrington (?). Popular and semi-popular views. Metaphorical uses. Appropriation. Descendants --- artificial intelligence, systems theory, dynamics, complexity, Prigoginism. And the Whole Earth Catalog.

Reflections, 6 August 2021

Well, that was a little harsh, wasn't it? But the place where it might be unfair was my conviction that ideas like "you need an efficacious feedback channel" were, in fact, internalized common currency wherever they were useful. This was, in retrospect, overly optimistic.

permanent link for this note   RSS feed for this note

Notebooks: