EMPIRICAL TOOLS AND EMPIRICISM 147 



Other concepts. Asked to decide whether a given specimen is a "hving 

 organism," the microbiologist apphes various culturing techniques to 

 obtain results that ordinarily support an unequivocal answer. But 

 when the crystalline viruses were first encountered, a latent indefi- 

 niteness of the denotation of "living organism" became quite pain- 

 fully manifest. 



Colligative relations as standards. Although we have discussed 

 separately different species of external standards, most typically 

 what we loork with is a mixed standard. Beyond "things" and tech- 

 niques a mixed standard involves colligative relations, which may in- 

 deed constitute the ultimate foundation of the entire standard. If an 

 instrument of comparison is used, such relations determine its design 

 and operating protocol. "Things" accepted as standards are selected 

 with a view to the relations in which they figure, and techniques ac- 

 cepted as standards are shaped by such relations. Ultimately the 

 denotation of "living organism" is established on the group of colliga- 

 tive relations expressing certain uniformities in the behavior of "living 

 organisms" in various circumstances. (Perhaps this is the useful sense 

 of the otherwise completely circular "law of definite proportions": A 

 compound contains its component elements in an invariant pro- 

 portion. ) 



Ordinarily I can safely rely on my judgment, i.e., on an internal 

 standard, in deciding that a particular object is a "lever" or a par- 

 ticular spot of light is a "planet." But in doubtful cases a suitable re- 

 lation can by itself constitute an exemplary external standard. One 

 or a very few trials suflBce to show whether the object does act in 

 accordance with the law of the lever; a relatively small number of 

 observations suffice to show whether the spot of light moves in accord- 

 ance with Kepler's first and second laws. Assuming affirmative indica- 

 tions, we identify the object as a lever and the light as a planet. 

 Serving thus as reference standards, the relations are not in any sense 

 "lost" to us: with them we can now go on to hazard predictions be- 

 yond what we have observed. 



PRIMARY AND SECONDARY DENOTATIONS 



There is only one standard kilogram, but many kilogram (and sub- 

 kilogram) "weights" are actively in use. All these are, ultimately, 

 referred to the mass of a particular platinum cylinder kept in the 

 vaults of the International Bureau of Weights and Measures, near 



