Practical implications of the discontinuity
Previous pages have been about theoretical issues, and have argued that the discontinuity between information and imparity exists, and needs to be recognised. However, there are some practical implications for cybernetics, which can no longer be ignored.
For example, in expert systems, the discontinuity which ultimately derives from Hume's Law implies that the imparity in the system has to be supplied by the human designer, and to hand this function over to the computer is quite a different enterprise from handing over the processing of energy or of information—and may in principle be impossible. Conventional discussions of the allocation of function between human and machine have ignored this problem. This has often been because the component of imparity, when it has only a small amount of information content, has seemed trivial. For example, a chess-playing computer program cannot itself originate what it means to win a game of chess. This is determined by the designers of the whole situation in which the computer is being used, but the amount of information required for the determination is very small and so the point has been overlooked.
Many of the fields in which cybernetics can be applied to human affairs involve a large element of what you might call 'managerial work'. Most discussions of managerial work in the literature concentrate on the perceptual and cognitive aspects of what they call 'management decision'. The implication drawn is that the difficult part is to see the need for intervention and to decide what to do. But an analysis of this work in terms of the three domains soon shows that management is not a purely second-domain process. If it were, it could eventually be dispensed with and replaced by computers and expert systems.
As mentioned in the case of choosing wallpaper, if it is possible to settle upon a formula for reliably expressing some of the imparity in terms of discriminative information, then some portion of the exercise of taste and judgement can be replaced by purely discriminative work.
It follows from this that calculating and reasoning are abilities that can eventually be handed over to computers and expert systems, but that to incorporate taste and judgement into an expert system is a different sort of project altogether. This is where the interesting cybernetic work is going to be in the future, and this project requires us to be very clear about the distinction between information and imparity.