14/6 AN INTRODUCTION TO CYBERNETICS 



to the conditioners is sufficient to keep the conditioners effectively in 

 control of the two thousand rooms. Thus the regulation that could 

 not be done in one stage may, if the conditions are suitable, be 

 possible in two. 



The quantities of communication (the channel capacities) involved 

 in these regulations could be measured to any desired accuracy, 

 and the exact degree of any amplification ascertained. Thus if 

 amplification had actually occurred, the reality of the fact could be 

 demonstrated beyond dispute. 



Whence (in the last example) comes the supplementation? In 

 general, from whatever supplies the other inputs. In the example 

 just given, these include the other factors that contributed to the 

 machines' design and manufacture, and also the environment itself, 

 which communicates to the conditioner, and not to the mechanic, 

 what is the temperature and humidity at each moment. As a 

 result, these sources of information play a part in the total regulation, 

 without using the mechanic as a channel. 



The example just given shows two levels of regulation, but there is 

 no reason why the number should stop at two. A doctor who 

 looks after the set of mechanics and keeps them healthy and able to 

 work might claim, so far as the rooms were concerned, to be a regu- 

 lator at the third level. The matter need not be pursued further 

 once the principle is clear, especially since many cases will probably 

 not show the various regulators arranged in a simple hierarchy. 



14/6. Amplification in the brain. We can now understand 

 quantitatively why this indirect method has proved superior — why 

 it is the method used by those organisms that have the most powerful 

 resources for regulation — it allows amphfication. 



The gene-pattern, as a store or channel for variety, has limited 

 capacity. Survival goes especially to those species that use the 

 capacity efficiently. It can be used directly or indirectly. 



The direct use occurs when the gene-pattern is used directly to 

 specify the regulator. The regulator is made (in the embryo) and 

 the organism passes its life responding to each disturbance as the 

 gene-pattern has determined. Amplification does not occur (from 

 our present point of view, though some advantage is gained (S.13/9) 

 if the disturbances recur frequently in the organism's lifetime). 



The indirect use occurs when the gene-pattern builds a regulator 

 (i?i) whose action is to build the main regulator (R2), especially if 

 this process is raised through several orders or levels. By achieving 

 the ultimate regulation through stages, the possibility of large-scale 

 supplementation occurs, and thus the possibility of an ultimate 



270 



