A tentative classification of goal-seeking behaviours

[M. P. Schiitzenberger](https://en.wikipedia.org/wiki/Marcel-Paul_Sch%C3%BCtzenberger), 'A tentative classification of goal-seeking behaviours', _Journal of Mental Science_, vol. 100 (1954), pp. 97-102. https://doi.org/10.1192/bjp.100.418.97 Snippets of text (from the beginning, maybe the middle, and the end) appear below, to give a sense of the content for the chapter. For more depth, the original source is cited, above. ---

Goal-seeking behaviour, once a great mystery, is now beginning to be understood. In its simplest forms it is, in fact, understood today almost completely. Thus the theory of the simple regulator, such as the thermostat, not only includes an extensive repertoire of techniques, but the elementary principles, of the necessity for negative feedback for instance, are becoming scientific commonplaces. In getting to know, however, about these simple systems and their principles, we should not make the mistake of thinking that there is nothing more to be learned. On the contrary, in real life many an important goal is to be achieved only through some quite complex pattern of behaviour, a pattern for which the simple concept of 'negative feedback' is quite inadequate. It is of these more complex patterns that I wish to speak today.

One way of studying the subject is by way of actual experiment but I shall make little reference to actual experiments today. The fact is that before such experimentation can be undertaken with any usefulness there must be a preliminary period of study and thought. Before we can experiment we must be clear about what questions we want to ask, and why these questions are significant, and what are to be the interpretations of the experiment's various possible outcomes. Before we can usefully start experimenting, in other words, we must have a well-developed theory. Such a theory must inevitably, if it is to be precise, be mathematical; but I hope to show in this paper that what is necessary, at least at first, is the logic and precision of mathematical thought rather than its more advanced techniques. If then we are to explore the properties of the more complex forms of goal-seeking behaviour we must first construct some suitable mathematical models. [p. 205] [....]

## Basic Concepts In order to make the ideas as clear as possible I shall start with a very simple example. Let us suppose a man is on the top of a hill and that he wishes to get to a house in the valley; let us assume that the 'goal' is his arrival there in the _shortest possible time+. Between him and the house are many causes of delay: boulders, marshes, escarpments, and so on. Travelling in a bee-line is out of the question. Let us consider his possible modes of behaviour. [p. 206] [....]

## Strategy and Tactic Our next step is to see more clearly what is the relation between these two -- to show them as derivatives of a single concept. Let us go back to the man on the hill. [p. 208] [....]

## Classification of Imperfect Tactics Having considered the optimal path, let us turn next to consider the case of the path that is grossly non-optimal, to that, say, of a stone rolling down towards the valley under the action of gravity. [....] [p. 209] [....]

## The Stochastic Environment What we have done so far is to show that the 'strategy' is simply one of the tactics: it is that extreme tactic based on the best function as given by the isochronal curves. It is now instructive to show conversely (so close is the relation between them) that any tactic may be viewed as some sort of strategy. [....] [p. 211] [....]

## Conclusion In this paper I have attempted to show something of what is implied by 'goal-seeking' when the whole situation is more complex than that occurring in, say, a simple thermostat. [p. 212] [....]

## Summary When 'goal-seeking' behaviour is considered in situations of more than the most elementary type, the problems that arise are related to those of strategies and tactics. I have attempted to show that clear-cut principles are involved, capable of mathematical treatment. It appears likely that among the factors of special importance are those of 'span of foresight' and 'degree of flexibility'. The case has also been considered in which the organism faces an environment that can be characterized only in terms of probability. [p. 213]