It is argued that choosing which state of affairs to prefer or to aim for is both logically and psychologically distinct from being able to discriminate one state of affairs from another, and it is a mistake to think of them as both being subsumed under the concept of information. Stewart (1) defined the former, when it is a transmissible component of an event, as imparity.
An early insight of cybemetics was that, in the description of feedback systems, it is the information retumed to the input, rather that the energy, that is important, and it is now accepted that cybemetics is an information science, in the sense that there are two domains of scientific concept - the classical one of matter and energy, and the secondary one of information - and that cybernetics is mainly concerned with the latter.
Although, from the beginning, cybernetics was recognised as proposing a new mechanistic conception of purposiveness, this was never mechanistic in the classical sense. Certainly a one-domain mechanism - concerned only with matter and energy - is inconsistent with teleology. However, it is a mistake to regard this problem as having been solved when the second domain was added. Evidence of this is the fact that people in the outside world, even after fifty years of cybernetics, continue to insist that no mechanistic conception can be adequate for dealing with human affairs.
A third domain needs to be included in our conception of mechanism, because cybernetics is fundamentally about controlling, managing, governing, and designing - it is always about systems in which evaluation has to take place. Somewhere, the system needs to have a third kind of input, and the imparity in a situation or a mechanism needs to be separated out and made a recognised part of the process of cybernetic analysis or design. Unfortunately, our theoretical or practical procedures have often not made the ternary structure explicit.
It is well known to philosophers that there is a logical gap between information and imparity. That ought cannot be derived from is was pointed out by Hume, and this logical law is not confined to imparity in the moral sphere. Interestingly, Hume's Law implies that considerations occurring in the domain of imparity cannot even in principle, be reduced to those in the domain of information. This ensures that a ternary cybernetic conception of mechanism cannot be reductionist, and so makes it distinct from both the classical, and the usual cybernetic, mechanistic conception.
(1) Stewart, D J, Kybernetes 18, 4, 19-28, 1989.