9/3 DESIGN FOR A BRAIN 



of all six groups of main variables. These changes of field will 

 continue to occur as long as the high excitation in the brain 

 persists. They will cease when, and only when, the linkages at 

 stage 6 transform an excitation of skin receptors into such a 

 bodily movement as will cause, through the pedal, a reduction 

 in the excitation of the skin receptors ; for only such linkages 

 can stop further encounters with critical states. The system 

 that is, will change until there occurs a stable field. The stability 

 will be shown by an increase in the voltage on the grill leading 

 to changes through skin, brain, muscles, and pedal that have 

 the effect of opposing the increase in voltage. The stability, in 

 addition, has the property that it keeps the essential variables 

 within physiological limits ; for by it the rat is protected from 

 electrical injury, and the nervous system from exhaustion. 



It will be noted that although action 3 has no direct connec- 

 tion, either visually in the real apparatus or functionally in the 

 diagram of immediate effects, with the site of the changes at 6, 

 yet the latter become adapted to the nature of the action at 3. 

 The subject was discussed in S. 5/15. 



This example shows, therefore, that if the rat and its environ- 

 ment formed an ultrastable system and acted purely automati- 

 cally, they would have gone through the same changes as were 

 observed by Mowrer. 



9/3. The two examples have taken a known fact of animal 

 behaviour and shown its resemblance to the behaviour of the 

 ultrastable system. Equally, the behaviour of the homeostat, 

 a system known to be ultrastable, shows some resemblance to 

 that of a rudimentary nervous system. The tracings of Figures 

 8/8/4 and 8/8/5 show its elementary power of adaptation. In 

 Figure 8/8/5 the reversal at R x might be regarded as the action 

 of an experimenter who changed the conditions so that the i aim ' 

 (stability and homeostasis) could be achieved only if the ' organ- 

 ism ' (Unit 1) reversed its action. Such a reversal might be 

 forced on a rat who, having learned a maze whose right fork 

 led to food, was transferred to a maze where food was to be 

 found only down the left fork. The homeostat, as Figure 8/8/5 

 shows, develops a reversed action in Unit 1, and this reversal 

 may be compared with the reversal which is usually found to 

 occur in the rat's behaviour. 



108 



