60 WHAT IS SCIENCE? 



that we can ask relevantly is why we are as certain that 

 the sun will rise to-morrow as we can be of any future 

 event and why we are so much less certain that it will 

 not rain to-morrow. 



It is obvious that our certainty in one case and our 

 uncertainty in the other are derived from our previous 

 experience of the happening of similar events, and that 

 the difference in knowledge is due to a difference in that 

 previous experience. Of course this statement does not 

 help us to solve our problem, for since laws are undoubtedly 

 derived from previous experience, it is clear that it 

 is there that the foundation and evidence for them 

 must be found. But the form in which the problem has 

 been put enables us to avoid altogether a question to 

 which those who have discussed the matter have usually 

 devoted most of their attention. They have asked how 

 it is that previous experience gives any knowledge of 

 future experience and what justification there can be 

 for asserting in any case whatever that we have such 

 knowledge. The point of view that I tried to suggest 

 in the last few paragraphs is that this question is essen- 

 tially unanswerable because it is based on the neglect 

 of the fundamental distinction between different kinds 

 of knowledge. Our " knowledge " of future events 

 simply is something based on our knowledge of past 

 events ; when we say that we know something about 

 the future we only mean that we have a mental attitude 

 based on past experience ; and it is absurd to ask why 

 it is based on past experience, for, if it were not so based, 

 it would be something quite different. In my opinion 

 (though the reader should be warned that others would 

 dissent strongly) it can only lead to confusion of thought 

 to attempt to compare this knowledge with other kinds 

 of knowledge and to ask how they stand in relative 

 certainty. And yet some comparison of knowledge of 

 future events with other kinds of knowledge is always 



