DESIGN FOR A BRAIN 7/11 



We assume that we have before us some system that has the 

 diagram of immediate effects shown in Figure 7/5/1. Some 

 variable, or several, called ' essential ', is given to act on a system 

 S so that if the variable (or all of them) is within given limits, 

 S is unchanging; but if it is outside the limits, S changes always. 

 (An. adequate variety of values is assumed possible to S so that 

 it does not develop, for instance, simple cyclic repetitions.) A 

 system called * environment ' interacts with another system R. 

 Environment has some effect on the variable called essential, and 

 S has some effect on R. Given this, and nothing more, does it 

 follow that the system R will change from acting non-adaptively 

 towards the environment, to acting adaptively towards it? (At 

 the moment I wish to add no further assumption; in particular 

 I do not wish to restrict the generality by making any assumption 

 that R is composed of parts resembling neurons.) 



7/11. Because the whole consists of two parts coupled — on the 

 one side the environment and reacting part R, and on the other 

 the essential variables and S — we can use the veto-theorem of 

 S. 6/9. This says that the whole can have as states of equilibrium 

 only such states as allow a state of equilibrium in both the essen- 

 tial variables and S. Now S is at equilibrium only when the 

 essential variables are within the given limits. It follows that all 

 the possible equilibria of the whole have the essential variables 

 within the given limits. So if the whole is started at some state 

 and goes along the corresponding line of behaviour, then if it 

 goes to an equilibrium, the equilibrium will always be found to 

 be an adapted one. 



Thus we arrive at the solution of the problem posed at the 

 end of Chapter 1 ; the mechanism has been shown to be necessary 

 by S. 7/8, and sufficient by the present section. 



7/12. This solution, however, is severely abstract and leaves 

 unanswered a great number of supplementary questions that are 

 apt to be asked on the topic. Further, it leaves one with no 

 vivid imaginative or intuitive conception of what is going on when 

 a system (one as complex as a human being, say) goes about its 

 business. The remainder of the book will therefore be concerned 

 with expanding the solution's many implications and specialisa- 

 tions. 



86 



