REASON AND LANGUAGE. 123 



know that human idiots, devoid of consciousness, learn 

 movements in the same way. But we also know that 

 fully conscious men and women often adopt through 

 distinct agreement (it may be tacitly) certain special 

 movements as "signs." These latter are, of course, 

 truly conventional signs, but not the former, which — as 

 having been nevertheless acquired — may be distin- 

 guished as " acquisitional" signs. 



Mr. Romanes continues : * *' The subdivision of con- 

 ventional signs may further be split into those which 

 are due to past associations, and those which are due to 

 inferences from present experience. A dog which 

 ' begs ' for food, or a parrot which puts down its head 

 to be scratched, may do so merely because past experi- 

 ence has taught the animal that by so doing it receives 

 the gratification it desires ; here is no need for reason — 

 i.e. inference — to come into play. But if the animal has 

 had no such previous experience, and therefore could 

 not know by special association that such a particular 

 gesture, or sign, would lead to such a particular con- 

 sequence, and if under such circumstances a dog should 

 see another dog beg, and should imitate the gesture on 

 observing the result to which it led ; or if under such 

 analogous circumstances a parrot should spontaneously 

 depress its head for the purpose of making an expres- 

 sive gesture, — then the sign might strictly be termed a 

 rational one." 



Now, there is, proverbially, great virtue in an " if," 

 and much unequivocal evidence would be needed to 

 show that such acts ever occur in animals. Granting, 



* p. 86. 



