DESIGN FOR A BRAIN 9/19 



might therefore prove advantageous; and this advantage, though 

 slight, might be sufficient to establish the mutation as a species 

 characteristic. Then a second mutation might continue the pro- 

 cess. The change from the original system to the ultras table 

 can therefore be made by a long series of small changes, each 

 of which improves the chance of survival. The change is thus 

 possible under the action of natural selection. 



Summary 



9/19. The solution of the problem of Chapter 1 is now completed 

 in its essentials. It may be summarised as follows: 



In the type-problem of S. 1/17 the disturbances that come to 

 the organism are of two widely different types (the distribution 

 is bi-modal). One type is small, frequent, impulsive, and acts on 

 the main variables. The other is large, infrequent, and induces a 

 change of step-function form on the parameters to the reacting 

 part. Included in the latter type is the major disturbance of 

 embryogenesis, which first sends the organism into the world with 

 a brain sufficiently disorganised to require correction (in this 

 respect, learning and adaptation are related, for the same solution 

 is valid for both). 



To such a distribution of disturbances the appropriate regulator 

 (to keep the essential variables within physiological limits) is one 

 whose total feedbacks fall into a correspondingly bi-modal form. 

 There will be feedbacks to give stability against the frequent 

 impulsive disturbances to the main variables, and there will be a 

 slower-acting feedback giving changes of step-function form to 

 give stability against the infrequent disturbances of step-function 

 form. 



Such a whole can be regarded simply as one complex regulator 

 that is stable against a complex (bi-modal) set of disturbances. 

 Or it can equivalently be regarded as a first-order regulator 

 (against the small impulsive disturbances) that can reorganise 

 itself to achieve this stability after the disturbance of embryo- 

 genesis or after a major change in its conditions has destroyed this 

 stability. When the biologist regards the system in this second 

 way, he says that the organism has ' learned ', and he notices that 

 the learning always tends towards the better way of behaving 



136 



