11/10 THR FULLY -JOINED SYSTEM 



adapted at essential variables 1 and 2. (More precisely: the 

 disturbance that comes to the environment finds such a set of 

 values on the step-mechanisms S that in the reaction to it, essential 

 variables 1 and 2 never go outside their proper limits.) The 

 reaction does, however, send essential variable 3 outside its limit. 



We now take as given that the system will go through a process 

 like that of S. 7/23, and that after it is over the step-mechanisms 

 will be changed to new values such that now essential variable 3 

 is kept within limits, and also that the other two still react in the 

 same way as before (this being necessary if we are to assume that 

 adaptations once formed are held, as we wish to). 



To see what this implies, let S 3 be the set of those step-mechan- 

 isms that ended with changed values after the trials due to 3's 

 adaptation (some such there must be, or the change of 3's response 

 to the disturbance is an effect without a cause). Now apply the 

 operational test of S. 4/12: as the step-mechanisms in S 3 are 

 changed in value, but the behaviours showing at 1 and 2 are not, 

 it follows that, whatever is shown in Figure 11/10/1, there can 

 be no effective channel of communication from the step- 

 mechanisms S 3 (in S) through R and the environment to 1 and 2. 



Next, let M 12 represent all those parts, in R and the environment, 

 that play a part in determining how the disturbance eventually 

 affects 1 and 2. By M 3 represent similarly the parts on the channel 

 from S 3 through R and the environment to 3. (Nothing is assumed 

 about how M 12 and M 3 are related.) Now M 3 cannot be void 

 (or S 3 would have no channel to affect 3, and the final adaptation 

 of 3 could not have occurred). Similarly, neither can M 12 be 

 void. Finally, the earlier deduction that there is no channel from 

 S 3 (through R and environment) to 1 or 2 implies that there are 

 no common variables to M 12 and M 3 , nor any channel from M 3 to 

 M 12 (for otherwise changes at S 3 would show at 1 and 2, contrary 

 to hypothesis). 



It has thus been shown that, for adaptations to accumulate, 

 there must not be channels from some step-mechanisms (e.g. S 3 ) 

 to some variables (e.g. M 12 ), nor from some variables (e.g. M 3 ) 

 to others (e.g. M 12 ). Thus, for the accumulation of adaptations 

 to be possible the system must not be fully joined. The idea so 

 often implicit in physiological writings, that all will be well if only 

 sufficient cross-connexions are available, is, in this .context, quite 

 wrong. 



155 



