38 SCIENCE (and common sense) 



of the concepts used by all obsen^ers— an eflFect particularly marked 

 for the tactile and visual data with which both science and common 

 sense prefer to work. Moreover, for many of the purposes of both, 

 perfectly concordant reports are— however attractive abstractly— 

 simply unnecessary. Finally, if we obtain some 99% of reports con- 

 cordant within the range of acceptability we are in practice quite 

 content to reject the 1% of discrepant reports as due to observational 

 error or some hitherto undetected abnormalitv of the observers. But, 

 even after all these dispensations, still the criterion of concurrence 

 is not reduced to a usable form. For we never could, or do, require 

 a convention of all normal men to attest their concurrence in a "fact." 

 Ordinarily we are satisfied to accept the reports of only a few, and 

 commonly of but one, observer—// the observation involved is one we 

 judge simple and unambiguous. Indeed from a single observer we 

 judge particularly competent we will accept even reports of observa- 

 tions not at all simple and unambiguous. Polanyi tells us that: 



. . . members of the Fifth International Botanical Congress [1931] 

 declared that "the concept of most species must rest on the judgment 

 and experience of the individual taxonomist," . . . Reflecting on this 

 discussion on the definition of a species, S. C. Hai'land recalled how 

 in Fanny s First Flay, by Bernard Shaw, the dramatic critic replies to 

 the question whether the play was a good play, that if the play was 

 by a good author, then it was a good play. "The situation would ap- 

 pear to be somewhat similar," writes Harland, "in regard to what 

 constitutes a species." 



In principle the criterion of concurrence is enforced absolutely; 

 in practice we leave large scope to individual judgment— which or- 

 dinarily assumes concurrence— and there is yet no indication that, in 

 the vast majority of cases, we have erred in doing so. Common sense 

 seems to make generally sound judgments of the "simple and unam- 

 biguous" observations in which the report of one "normal" observer 

 is simply repeated when additional witnesses are called in. Science 

 does as well— and even better in that often it can so simplify the 

 observations to be made that the possibility of observational error is 

 minimized. At the extreme of simplification, observation is reduced 

 to reading the position of a needle that moves across a graduated 

 scale until it points to "the result." Judgments of brightness made 

 witii the unaided eye are notably variable, and unreliable, as every 



