118 COLLIGATR'E RELATIONS AND SCIENTIFIC LAWS 



soluble in carbon disulfide, burning in oxygen to give a gas . . . and 

 so on. Is this any more than a definition of what we mean by "sulfur"? 

 Meyerson teaches us to see that the definition yields the following per- 

 fectly genuine colligative relation: If a substance displays many of 

 the following properties . . . , then probably it will display all the 

 others. Such a relation oflFers multiple predictions of which many 

 have been confirmed repeatedly. Can any such prediction ever fail? 

 Yes! The definition notwithstanding, every chemist well knows that 

 the sulfur of experience may be quite diflFerent from the ideal sulfur 

 of the definition. "Real" sulfur may be not yellow but white; not hard 

 and crystalline but rubbery or amorphous; melting not at 119.0°C 

 but at some other temperature dependent on its previous history, 

 purity, and isotopic composition— and so on. 



What to do? Restrict application of the relation to some standard 

 sulfur prepared by a specified procedure? Certainly not! Just as we 

 guard the generality of such concepts as "water" and "lever," we in- 

 sist on keeping "sulfur," and its defining relations, conveniently ap- 

 plicable to the myriad specimens of real sulfur— each with its own 

 impurities and its own history. Seeking efficiency, we make the char- 

 acteristic compromise, as is perhaps most evident when we ask: 

 How "many" properties must we check before identifying a substance 

 as sulfur and predicting its other properties? Obviously, the more 

 properties we check the less likely we are to make false identifica- 

 tions and false predictions. Ordinarily, howev^er, we are content to 

 check only one or a few properties— perhaps even none beyond the 

 label on the container. 



A COLLIGATIVE RELATIOxN: BOTH INVENTION AND DISCOVERY 



We invent the concepts weight and distance; we invent the ideal 

 lev^er to which the law of the lever applies by definition. But the law 

 acquires value only through a simultaneous discovery: there are in 

 nature distinguishable systems predictably conformable to the law. 

 Similarly, the concept sulfur is an invention, scientifically meaningful 

 only because we often encounter recognizable specimens sufficiently 

 approximating "ideal" sulfur that useful predictions can be drawn 

 from the sulfur relation. These complementary elements of inven- 

 tion and disco\^ery are evident even in the more abstract scientific 

 concepts. Some writers dispute whether Thomson invented or dis- 

 covered the electron. This futile issue is easily resolved: the electron 



