EMPIRICAL TOOLS AND EMPIRICISM 143 



Even today in certain applications the human eye offers us discrim- 

 inatory capacities unmatched by any instrument or combination of 

 instruments. Why should we hesitate to make the fullest possible 

 use of the sensory faculties with which we are endowed? A vain lust- 

 ing after reduction to instrumental measurements in general, and to 

 pointer readings in particular, seems to rest on complete misconcep- 

 tion of what we gain with their aid. Not categorical denotation, not 

 the complete elimination of human judgment, not "objectivity." The 

 fundamental achievement is this: denotative clarity is enormously 

 enhanced by adoption of an "external" rather than an "internal" 

 standard of reference. Though instruments and pointer readings are 

 generally sufficient to ensure this great advance, they are not always 

 necessary to ensure it. 



THE EXTERNAL STANDARD OF REFERENCE 



How shall I establish the denotation of "weight"? That denotation 

 remains highly uncertain for just so long as it is "fixed" only by an 

 internal standard, i.e., the sense of muscular strain felt in lifting. In 

 that case I have always to remember what it felt like to lift other 

 "weights," and my memory is not wholly reliable. Nor are my sensa- 

 tions unaffected by my recent personal history: things may well seem 

 heavier when I am tired. My own estimates of the weight of a given 

 object vary from time to time, and disagree with the estimates made 

 by other persons who rely on their internal standards. 



The self-consistency of my own reports of weight, and the concur- 

 rence of different reporters, are both enormously improved when we 

 all agree to adopt one particular set of standard weights. "Weighing" 

 is then reduced to making a comparison between one or more of 

 these and the unknown weight. I may, for example, seek to make 

 some combination of standard weights that "feels the same" when it 

 is picked up immediately before or after the unknown weight— the 

 magnitude of which can then be expressed in terms of the stand- 

 ard ( s ) . As already indicated ( p. 51 ) , I can even go somewhat further. 

 Making comparison not through trials by lifting but with an equal- 

 arm balance, I exclude secondary complications that might arise 

 when unknown and standard differ sharply in shape, texture, etc. In 

 this way the role of human judgment, though never completely elim- 

 inated, is still further diminished and simplified— while at the same 

 time we acquire a delicacy of discrimination wholly unknown to the 



