i2o PHYSICAL BASIS OF CIVILIZATION 



corroborated by observations or experimental tests. 

 Is it therefore untrustworthy? It may be, but is 

 by no means necessarily so! For the mental process 

 which connects known antecedents with known 

 subsequents, by means of links obtained from the 

 imagination, has supplied a large number of the 

 most reliable inferences and conclusions in science 

 and in thought, and, as explained in the third para- 

 graph of the preface, the value and validity of such 

 conclusions is not impaired by the impossibility 

 of obtaining observations or experimental tests in 

 support of them. For illustration, let it be sup- 

 posed that on a desolate island which is hundreds 

 of miles away from other lands, and whose surface 

 consists of granite and the products of the decom- 

 position of this mineral, some articles of pottery 

 and clothing, and nothing else indicating human 

 origin, are found. The granite and the products of 

 its decomposition are the indisputable evidence 

 that the island has existed before human beings 

 had begun to make pottery and clothes. Yet 

 even this indisputable statement is derived from 

 the imagination, and cannot be corroborated by 

 observation or experimental tests. The presence 

 there of the articles named is the only fact evidenced 

 by sense experience, and corroborative by observa- 

 tion and experimental tests. The conclusion is 

 then reached, that a person or a number of persons 

 were out in a large or small craft, or on a raft, and 

 had with them there these articles; that from 



