6/21 AN INTRODUCTION TO CYBERNETICS 



who faces a complete set of dials. Often, however, this is not so — 

 some of the dials are hidden, or missing — and an important part of 

 Black Box theory is concerned with making clear what peculiarities 

 appear when the observer can observe only certain components of 

 the whole state. 



The theoretical developments are large, and little explored. They 

 will almost certainly be of importance in psychology; for, to the 

 psychologist, the individual subject, whether a neurotic person or a 

 rat in a maze, is largely a system that is not wholly observable; for 

 the events in the subject's brain are not directly observable at the 

 clinical or experimental session. 



It should be noticed that as soon as some of a system's variables 

 become unobservable, the "system" represented by the remainder 

 may develop remarkable, even miraculous, properties. A common- 

 place illustration is given by conjuring, which achieves (apparently) 

 the miraculous, simply because not all the significant variables are 

 observable. It is possible that some of the brain's "miraculous" 

 properties — of showing "foresight", "intelligence", etc. — are miracu- 

 lous only because we have not so far been able to observe the events 

 in all the significant variables. 



6/21. As an example of the profound change that may occur in 

 the observer's opinion about a mechanism if part of it becomes 

 inaccessible to direct observation, consider the following example. 

 The observer is assumed to be studying a Black Box which consists 

 of two interacting parts, A and Z. Both are affected by the common 

 input I. (Notice that ^'s inputs are / and Z.) 



Suppose the important question is whether the part A does or does 

 not show some characteristic behaviour B (i.e. follow trajectory B). 

 Suppose this is shown (followed) only on the simultaneous occurrence 



of 



(1)/ at state a 



and (2) Z at state y. 



Suppose that Z is at state y only after I has had the special value ii. 

 We (author and reader) are omniscient, for we know everything 

 about the system. Let us, using full knowledge, see how two 

 observers (One and Two) could come to different opinions if they 

 had different powers of observation. 



114 



