Sensory Discrimination: Methods of Investigation 55 



from the ear, but no one can tell us how to judge from the 

 structure of the ear what range and fineness of pitch dis- 

 criminations exist in its possessor's mind. No investi- 

 gator has yet succeeded in relating the different quahties 

 of smell and taste to differences in the end organs. 



§ 12. Behavior ^s Evidence of Discrimination 



The atSMm ent fr om behavior is as follows : If an animal 

 reacts in a different way to two qualitatively unlike stimuh, 

 then, providing that it is conscious at aU, it may be sup- 

 posed to receive qualitatively unlike sensations from them. 

 If it al ways rea cts in the same way to both, then both may be 

 su^osed to"Eeaiccompanied by the same sensation quality. 

 Obviously these statements need further discussion. For 

 one thing, it may be urged that in our own case the same 

 external reaction is often made to stimuH that are never- 

 theless consciously discriminated. A man may eat with 

 reUsh and without observable difference in behavior, for 

 example, foods that yet give him perfectly distinguish- 

 able smell and taste sensations. Precisely this objection 

 holds against a method of experimentation, formerly a 

 good deal used, which may be called the Preference Method 

 of testing discrimination. Vitus Graber, for instance, 

 attempted to find whether animals belonging to a variety 

 of species could discriminate colors, by offering them the 

 choice of two compartments illuminated each with a dif- 

 ferent color. Clearly, if the animals chose one compart- 

 ment as often as the other, it would be rash to conclude 

 that the two lights produced for them indistinguishable 

 sensation qualities. There might simply be the absence 

 of any preference, along with perfect discrimination. 

 The fact is that in all experiments upon animals, whether 



