STUDIES ON THE MENTAL LIFE OF ANIMALS. 547 



re(;ent writei> on animal i)syfhokigy to jiraii^e the abilities of aninialy. This can not 

 but lead to partiality in deductions from facts and more especially in the choice of 

 facts for investigation. How can scientists who write like lawyers, defending 

 animals against the charge of having no power of rationality, l)e at the same time 

 impartial judges on the bench?" 



Finally, even the writers who might have won valuable results have 

 contented themselves with arguing- against theories of the eulogists. 

 They have not yet made investigations of their own. 



Further, animal psychology has been hitherto too nuich derived 

 from anecdotes; authors cite only those facts that arc exceptional, 

 extraordinary, or considered as such, instead of which the normal and 

 simplest cases should always be reported. Anecdotes also have the 

 disadvantage of being rarely given at first hand. Finally, by their 

 very definition they are unverifiable. 



II. 



The method adopted should be an exclusively experimental one, 

 and that alone is what Mr. Thorndike has used. It is essentially a 

 method of observation and experiment, submitted, as in all such cases, 

 to a certain lunnber of constant precautions. It was necessar}' here 

 more than elsewhere to av-oid making generalizations from the indi- 

 vidual to the species on observations of only a single case, to regu- 

 late in a precise manner the conditions of each experiment as far as 

 possible, and to use in these experiments only animals whose previous 

 historj^ was known, all of which precautions seem to have been 

 ignored by preceding psychologists. This objective attitude once 

 adopted, there were still some details to be arranged. It was necessary 

 to formulate a plan- to establish a certain number of constant points 

 which, in the whole series of experiments, would serve as guides for 

 the observer. Mr. Thorndike reduces these points to three, fornni- 

 lated as the three following questions, which constitute the logical 

 structure of the method: 

 , 1. What is it the animals inider observation do? 



2. How do they do it? 



3. What do they feel while they thus act?'' 



The theoretical aspect being thus settled, it became necessary to 

 determine the material and psychological conditions of the experi- 

 ments. The principle kept in view was to select an experiment or 

 series of experiments which should be at once simple and at the same 

 time significant. From a psychological point of view it was necessar}^ 

 to manage so as to have only a restricted num])er of known })sycho- 

 logical elements. Our author adopted the following ingenious strat- 

 agem: Using for his experiments cats, dogs, and chickens, he deprived 



a Thorndike, pp. 3 and 4. b Thorndike, p. 5. 



