Inevitably, an account of how enterprises are to be controlled is an extraordinarily complex business. This is the reason why Figure 48, which reduces the matter to its absolute essentials, is none the less complicated and difficult to study. However, if all that has gone before has been well comprehended, the task is not insupportable. As usual, the study begins with a (or in this case, rather, the) world situation. For the enterprise, this world situation may be understood in the form of two fairly well contra-distinguished aspects. The first is the internal world situation, which is the one that the enterprise is; the second is the external world of the enterprise environment: that part of nature which has a direct bearing on the internal state. As has been recognized before, these two systems are not absolutely separable; they do not have absolutely clear-cut boundaries. One of the reasons for the problem of separability is that each is influenced by a set of coenetic variables to be found in the world at large. All this is illustrated at the top of the diagram, where it will be seen that a set of preferred states is marked off (as a circle) from the totality of states represented by the phase space. In these two small pictures, as in those others which follow, the present state is hown within a preferred set of states. To go back to the drawing: the input from the set of coenetic variables is shown helping to determine this happy state of affairs. Each of these world situations may be observed to be amplifying variety. This plethora of information is seen cascading from the side of each picture. As we know, the first problem of control is to become aware of this variety, and in each case data is shown cascading on to the lid of a black box. But before attending to its subsequent use, it will be noted that the internal and the external world situations are depicted as interacting through a self-vetoing homeostat (H1). This is no more than a formalization of the actual state of affairs, for whatever the management does or fails to do about controlling the situation, it will in some sense control itself through the interactions which its two halves necessarily have. Management, however, will wish to influence this interim interaction for motives of its own, and the model now continues by saying what is to happen.