56 WHAT IS SCIENCE? 



by this invariability that they are capable of being 

 experienced by everybody, and consequently are capable 

 of that universal assent which makes them the proper 

 subject-matter of science. 



Nevertheless, there is one particular form of relation 

 involved jn laws which can be distinguished from others, 

 and on which emphasis may be laid once more. This 

 relation is that which characterizes^ what we have called 

 the law of the properties of a substance, or a kind of 

 system, the law, namely, which asserts that there is 

 such-and-such a substance or such-and-such a kind of 

 system, steel or magnets, for example. These laws, in 

 an elementary and imperfect form, are the earliest laws 

 of science, and they retain their peculiar significance 

 through much of its consequent development. The 

 recognition that there are such laws, and that they are 

 laws just as much as others more generally recognized 

 and called by that name, enables difficulties to be over- 

 come which have troubled some of those who have tried 

 to explain the nature of science. One of these difficulties 

 is connected with the " classificatory " sciences, such as 

 the older zoology and botany, or mineralogy. Such 

 sciences seem at first sight to state no laws at all ; they 

 simply describe the various animals, plants, or minerals, 

 and arrange them in groups according to their resem- 

 blances or differences, but do not state about them any 

 laws that are usually recognized as such. But, if such 

 sciences do not state laws, why are they regarded as 

 sciences ? We can answer now : They 



laws of the kinfi which assorts tfre pi 



system. In establishing that there is such an animal as a 

 cow and pointing out accurately its differences from a 

 sheep, or in investigating the differences between quartz 

 and rock-salt, zoology or mineralogy is discovering 

 laws, and laws of the kind that are of the most funda- 

 mental importance. The_ " ckssificatory " sciences differ 



