i The Study of Nature 7 



be definite in order to be clear. If he wishes to be definite 

 in thought the student must never rest content with the 

 dubious "I think" or the vague "about," but endeavour 

 after the clear " I know " and the precise " with a prob- 

 able error of." Vagueness and indecision are utterly 

 foreign to the scientific method. It often happens that 

 there is no definite knowledge concerning some fact ; then 

 all that the scientific method of description permits is 

 to say, " There is no information," and to wait until the 

 scientific method of observation has found out something. 

 The difficulty is not overcome by guessing, or by calling 

 the unknown unknowable. There is a place for specula- 

 tion and imagination in the scientific method ( 18), but it 

 is a place apart, which must be shut off, for if speculations 

 are not kept in strict quarantine they are certain to infect 

 our conceptions of facts with their own fatal vagueness. 



1 2. Scientific Terms. Definite words are necessary for 

 the expression of definite ideas, hence scientific terms 

 have to be employed. A term has one definite meaning 

 which does not change with time. The rush of affairs 

 drifts words from their original meanings, as ships drag 

 their anchors in a gale, but terms sheltered from common 

 use hold to their moorings for ever. The word let, for 

 example, has drifted in 200 years from meaning hinder 

 until now it means permit ; but the term bisect has remained 

 unaltered in significance for twenty centuries. Many 

 scientific terms are derived from the Greek and have an 

 unfamiliar appearance ; a list of all those employed in this 

 book, together with their derivation, is given in Appendix III. 



1 3. Classification of the facts and processes of Nature 

 is necessary before we can form definite 'ideas concerning 

 them ; but the definiteness of classification is an artificial 

 restriction. In Nature one thing merges into another by 

 imperceptible degrees, and although, for example, we can 

 readily class typical metals and non-metals, typical igneous 

 and sedimentary rocks, typical plants and animals, there are 

 in each of these pairs of classes many cases which cannot 

 be referred with certainty to either side of the dividing 

 line. Nature is discrete only within certain limits, and its 



