9/8 



DESIGN FOR A BRAIN 



The control of aim 



9/8. The ultrastable systems discussed so far, though develop- 

 ing a variety of fields, have sought a constant goal. The homeo- 

 stat sought central positions and the rat sought zero grill-potential. 

 In this section will be described some methods by which the 

 goal may be varied. 



If the critical states' distribution in the main-variables' phase- 

 space is altered by any means whatever, the ultrastable system 



10 



5- 



IO 



2Q 



U 



Figure 9/8/1. 



will be altered in the goal it seeks. For the ultrastable system 

 will always develop a field which keeps the representative point 

 within the region of the critical states (S. 8/7). Thus if (Figure 

 9/8/1) for some reason the critical states moved to surround B 

 instead of A, then the terminal field would change from one which 

 kept x between and 5 to one which kept x between 15 and 20. 

 A related method is illustrated by Figure 9/8/2. An ultra- 

 stable system U interacts with a variable A. 

 E and R represent the immediate effects which 

 U and A have on each other ; they may be 

 thought of as A's effectors and receptors. If 

 A should have a marked effect on the ultra- 

 stable system, the latter will, of course, develop 

 a field stabilising A ; at what value will depend 

 markedly on the action of R. Suppose, for 

 instance, that U has its critical states all at 

 values and 10, so that it always selects a field 

 stabilising all its main variables between these values. If R 

 is such that, if A has some value a, R transmits to U the 

 value 5a — 20, then it is easy to see that U will develop a field 

 holding A within one unit of the value 5 ; for if the field makes 



120 



A 



Figure 9/8/2. 



