through the stones; yet from these observed flashes and 
     tappings he becomes, or seems to become, apprised of 
     unobserved football matches, flower gardens and eclipses 
     of the sun. How then does he learn the ciphers in which 
     his signals are arranged, or even find out that there are 
     such things as ciphers? How can he interpret the 
     messages which he somehow deciphers, given that the 
     vocabularies of those messages are the vocabularies of 
     football and astronomy and not those of flickers and 
     tappings?

          This model is of course the familiar picture of the 
     mind as a ghost in a machine, about the general defects 
     of which nothing more need be said. But certain particular 
     defects do need to be noticed. The use of this sort of 
     model involves the explicit or implicit assumption that,