THE BLACK BOX 6/5 



and /S, and possible output states /, g, h and j, a typical protocol 

 might read (and be yet another transformation!): 



Time: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 



State: ag aj af af af j3/ ^h jS/z ah aj ^f ah )8/ j3/ ah pj «/ 



(Parentheses have been omitted for brevity.) 



This form, though it may seem artificial and unnatural, is in fact 

 typical and general. It will represent anything from the investiga- 

 tion of an electrical network by putting in a sinusoidal voltage and 

 observing the output, to a psychiatric interview at which questions 

 a, /3 were put and answers g,f, h,j ehcited. 



Thus, the primary data of any investigation of a Black Box consists 

 of a sequence of values of the vector with two components: 



(input state, output state). 



(The possibility is not excluded that each component may itself 

 be a vector (S.3/5).) 



From this there follows the fundamental deduction that all 

 knowledge obtainable from a Black Box {of given input and output) 

 is such as can be obtained by re-coding the protocol; all that, and 

 nothing more. 



Ex.: Tabulate the transitions observed in the system that started at ag. Find 

 some regularities in them. 



6/4. It will be noticed that nothing has been said about the skill of 

 the experimenter in manipulating the input. The omission was 

 deliberate, for no skill is called for! We are assuming, remember, 

 that nothing is known about the Box, and when this is so the method 

 of making merely random variations (e.g. guided by throws of a die) 

 on the input-switches is as defensible as any other method; for no 

 facts yet exist that could be appealed to as justification for preferring 

 any particular method. With terrestrial machinery — industrial, 

 biological, neuronic — -the experimenter has often had previous 

 experiences with Boxes of the same class. When this is so he may 

 be able to use a method that explores what he does not know about 

 the present Box more efficiently than some other method. (These 

 matters, of exploring a partly known system, lead into questions of 

 altogether more advanced type, and their consideration must be 

 postponed; a little is said on the subject in S.13/5 and onwards.) 



6/5. Absoluteness. When a generous length of record has been 

 obtained, the experimenter will look for regularities, for repetitive- 

 ness in the behaviour (S.7/19). He may notice, for instance, in 



89 



