[ 42 ] 



evidence, but hitherto without any explicit recog- 

 nition of the nature of the procedure. Let us begin 

 by inquiring what it is to which the test applies. 

 Evidence at first hand is sometimes supposed to be 

 simply a record of the sensations of the witness ; 

 but this never is so in practice, for we have not 

 words to describe many simple sensations, and if 

 we had the words, the use of them would still in- 

 volve the syntaxis of reminiscences of previous 

 sensations with which to compare those described. 

 Thus, nothing could appear to be more directly a 

 record of sensation than the statements of a patient 

 to the doctor at his bedside ; yet the simplest of 

 these involves a syntax which may be misleading 

 even when the statement is made in all good faith. 

 Many a patient has complained of pain in the toes 

 of a foot which the doctor knows to have been 

 amputated one more proof, the ancient Greeks 

 might have said, how little truth there is in the 

 " lying witness of the senses." A clearer arrange- 

 ment of ideas is obtained by imagining a separation 

 between the sensation and the syntax with which it 

 is associated ; the sensation itself is not misleading, 

 but it is connected in the nervous organism of the 

 patient with reminiscences of former sensations 

 which he had learnt to refer to his toes ; later, 

 when he has acquired a fresh set of reminiscences, 

 if the same pain be felt it will be referred with 

 equal confidence to the stump of the amputated 



