144 EMPIRICAL TOOLS AND EMPIRICISM 



unaided senses. To go still furdier— to an elaborate, automatic, 

 pointer-reading device— may win us gains in sensitivity and conven- 

 ience, but can afford us no further fundamental gain in clarity of 

 denotation. 



Consider the analogous case of "temperature." Our thermal sensi- 

 bility, quite limited at best, is also notoriously unreliable; it is, for 

 example, heavily dependent on the recent history of the sensing 

 surface. The denotation of temperature is immensely clarified and 

 refined with the aid of a thermometer. At first sight this instrument 

 (unlike the equal-ann balance) seems to involve no comparison with 

 an external standard. The balance is a null-reading instrument in- 

 volving an explicit comparison, but scale-reading instruments like 

 the thermometer normally involve implicit comparisons. The scale is 

 constructed by comparison ( direct, indirect, or by way of colligati^'e 

 relations) with certain external standards. Thus whenever I read a 

 temperature from a thermometer scale, I make an implicit compari- 

 son with, say, tlie temperatures of melting ice and boiling water. This 

 state of affairs remains unaltered when the instrument is more deeply 

 involved, in amplifying and/or converting an otherwise undetectable 

 signal. An ohmeter gives a direct pointer reading of "resistance," but 

 ultimately it yields a comparison of this resistance with that of a 

 particular spool of manganin wire. 



Denotations depending on "absolute" judgments— that is, judg- 

 ments made relative to more or less uncertain internal standards— suf- 

 fice for many purposes, e.g., counting. For some purposes, e.g., psy- 

 chiatric diagnosis, they may well be indispensable. But, in general, 

 denotations are most firmly established when we need make only 

 relative judgments, i.e., comparisons with external standards. Thus, 

 for example, the human eye is thoroughly ineffectual in judging the 

 absolute intensity of illumination, but highly adept at distinguishing 

 small differences in intensities that occur, or can be brought together, 

 in the same visual field. Given one "standard candle," comparisons 

 founded on use of even so primitive a device as the Bunsen grease- 

 spot photometer in\'est "light intensity" with a well-defined denota- 

 tion. Hebb's comment amply suggests that we here encounter no 

 trivial shortcoming of optic physiology but, rather, a general psy- 

 chological phenomenon: 



Man or animal tends to perceive relative rather than absolute in- 

 tensity, extent, or frequency. One can readily train an animal to 



