38 WHAT IS SCIENCE? 



observe are different. Thus, in our example, it is im- 

 possible for every one to observe that a particular book 

 fell to the floor and made a noise on striking it ; JDUI^ jjs 

 /possible for every one to observe that, if a book is pushed ' \ 

 /over the edge of the table, it will fall to the floor and/ 

 will make a noise on striking it. Concerning that judg- 

 ment there can be universal agreement ; and that agree- 

 ment will not be upset, even if somebody has never 

 actually observed a book fall ; so long as he agrees when 

 at last he is placed in the necessary circumstances, that 

 a book will fall and that, when it falls, it will make a 

 noise, then universal agreement is secured. 



If we could imagine ourselves without any experience 

 of the external world derived from our senses, we might 

 doubt whether there actually are such relations concerning 

 which universal agreement can be obtained ; we might 

 expect that it would be as impossible to find universal 

 relations between events as to find universal events. 

 But we all know from our experience that there are such 

 ^relations and we know of what kind these relations are. 

 ^They are of the kind that have just been indicated ; the 

 universal relations that we can state are between events 

 which are such that, if one event happens, then another 

 event happens."] Again, there might conceivably be other 

 relations betwran events of a different kind, yet of the 

 same universality ; actually there are not at any rate 

 if we interpret the relation just stated correctly. There 

 is a certain class of relations between events for which 

 universal agreement can be obtained, which is thereby 

 distinguished from other classes for which it cannot be 

 obtained. Indeed, we might almost say that it is only 

 this class which can be the subject of universal agreement ; 

 for the necessity that all men, even if they live at different 

 times, should agree imposes limitations on the form of the 

 relation. But we need not inquire into this abstruse 

 matter ; /all that is necessary for our purpose i> to 



