16/7 DESIGN FOR A BRAIN 



be found to occur by the fast process ; and we can make this 

 prediction without any reference to the particular physical or 

 chemical details of the particular change. 



The principle is well known in chemical dynamics. Thus there 

 is a reaction whose initial and final states are described by the 

 equation 



6FeCl 2 + KCIO3 + 6HC1 = 6FeCl 3 + KC1 + 3H 2 0. 

 There are at least two processes leading from the initial to the 

 final states : one corresponding to the reaction (of the thirteenth 

 order) as written above, and one composed of a series of reactions 

 of low order of which the slowest is the reaction (of the third 

 order) 



2FeCl 2 + Cl a = 2FeCl 3 . 

 The first is slow, for it has to wait for an appropriate collision 

 of thirteen molecules, while the second is fast. We can predict 

 that the fast will be preferred ; and direct testing has shown that 

 the reaction occurs by the second, and not the first, process. 



From this we may draw several deductions. First, the multi- 

 stable system will similarly tend to adapt by its fast rather than 

 by its slow process. Secondly, since the fast process, by S. 12/4, 

 is that of adaptation by a series of small independent parts, any 

 multistable system will behave as if it 4 preferred ' to adapt by 

 many small independent adaptations rather than by a few com- 

 plex adaptations : it ' prefers ' to adapt piecemeal if this is 

 possible. Finally, by using the fast process, the time it takes 

 in getting adapted will tend to the moderate T 2 (of S. 12/3) 

 rather than to the immoderate T v It is therefore at least partly 

 free from the fault of excessive slowness described in S. 11/7. 



178 



