THE SYSTEMS CONCEPT REDEFINED 299 



be they random or constant errors. No measurement is likely to be perfect. 

 We are always faced with this probability of error, and of (biological) varia- 

 tion in the thing being measured. 



The human machine is subject to error in measurement, just as is any 

 other machine. It is no accident that athletic competitions, especially by 

 professional athletes, are described as "games of inches," the differentiating 

 factor being the ability to estimate distance under great psychological stress. 



In summary, it is a measurement which is fed back into a computer to 

 guide it in making corrections to its actions. This measurement is ot the dif- 

 ference, A, between where one is and where one wants to be — that is, of the 

 error. The error is increased by noise. 



Noise 



The subject of the detection of a signal of information (energy) over back- 

 ground was discussed in Chapter 3 in the discussion of sensitivity of a detec- 

 tor and the VVeber-Fechner Law. The principles introduced there apply also 

 to the detection of information to be fed to a computer. If the source provides 

 a strong signal over background, the detector will feed a correspondingly 

 strong signal to the computer. If the background noise is high (i.e., the 

 signal-to-noise ratio is low) the signal sent to the computer may not be in- 

 telligible (discernible from the background). Strength of the signal, back- 

 ground noise, and degradation of the information by noise introduced in the 

 detector determine what the computer receives as input. 



Unfortunately there usually are many strong signals entering a detector, 

 only some of which are useful. Those which are not useful are also noise, 

 like the background. The machine must be able to classify signals: to accept 

 the information and by-pass the rest. One of the most useful systems yet 

 built to separate information from noise is the EEG analyzer, a machine 

 which scans the information and sorts the rather complex total waves into 

 their three or four main components. 



Continuously confusing the control circuits of a human being is an un- 

 remitting input of noise — disordered, and perhaps not even useful informa- 

 tion. Noise can take several forms. First of all it may be of either external 

 or internal origin. External noise comes in from the environment through 

 the senses. It is probably better to call it incomplete rather than disordered, 

 for there is order and regularity reaching our senses from everywhere about 

 us in nature. The trouble arises because we have only a limited eapa< it\ 

 or interest — subject to, or determined by our freely-chosen goals in life. In 

 other words, what is useful, interesting information to one man is noise to 

 another; and for one man, what is noise at ( ) P. M. may not be so at 9 A. M. 

 (traffic information, for example). This is unfortunate, but nevertheless true. 

 It is unfortunate because it means that two men with a common interest in 



