July 14, 1898] 



NATURE 



249 



ANIMAL INTELLIGENCE : AN EXPERI- 

 MENTAL STUDY} 

 TVyTANY are the writers on animal intelligence, but 

 ■^'■*- few have made comparative psychology a subject 

 of scientific investigation by the methods of careful ob- 

 servation and of experiment under conditions allowing 

 of some control. Right welcome, therefore, is Mr. 

 Thorndike's experimental study, of which a brief pre- 

 liminary notice appeared in Nature a few weeks ago 

 (vol. Ivii. p. 372). 



This careful research goes far to confirm the con- 

 clusion, to which the present writer has been led, 

 that the method of animal intelligence is one of un- 

 discriminating trial and error, of profiting by chance 

 experiences, and one which depends on the establish- 

 ment of direct associations — a conclusion which is in 

 close accord with that reached by Prof. Wundt. Mr. 

 Thorndike is, however, somewhat severe in his criti- 

 cisms of previous writers in the same field, complains 

 that they have made no observations of their own, 

 and says that most of the books do not give us a 

 psychology, but rather a eulogy of animals. " They 

 have all been about animal intelligence, never about 

 animal stupidity." One of the previous writers has, 

 however, said : "And then, as Mr. P. G. Hamerton well 

 remarks, we have to take into account the immensity of 

 the ignorance of animals." Ignorance and stupidity are, 

 of course, by no means synonymous. But it is the 

 former rather than the latter that is so abundantly ex- 

 emplified in animal life. 



In many of his experiments Mr. Thorndike's method 

 was as follows. Very hungry kittens were shut up in 

 box-cages, 20 inches long by 15 broad and 12 high, and 

 food was placed outside within the animals' sight. To 

 get out the kitten had either to pull down wire loops 

 placed in different positions in different cages, or turn a 

 broad button, or press an ordinary thumb-latch, or push 

 down a small platform, or simply pull a string stretched 

 across the roof. These devices (each in its separate 

 cage) were so arranged that on the fitting push or pull 

 the door opened ; and fish was the reward of success. 

 In other cages two or three distinct actions on the part 

 of the kitten were required before the door opened. In 

 yet other experiments the kitten was released and fed 

 directly she either licked herself or scratched herself. 

 The object of the investigation was to watch and record 

 the establishment of associations ; and the results are 

 plotted in curves, giving the time-intervals between im- 

 prisonment and escape in successive experiments. 



The curves are far from smooth, as is indeed to be 

 expected where the internal factors are necessarily some- 

 what inconstant, and where the difficulties to be overcome 

 by the subjects are different in different cases ; but they 

 bear out the contention that the method of animal in- 

 telligence is to profit by chance experience, and is de- 

 pendent on the gradual establishment of direct associa- 

 tions. I have endeavoured to extract from some of Mr. 

 Thorndike's carefully plotted data a mean curve for the 

 method of trial and error, and though it does not come 

 out very well, it does serve to indicate that gradual 

 sweep towards rapid and assured success, which would 

 theoretically result on this method. In contradistinction 

 to this the curve of rational procedure is quite different. 

 I plotted some curves of this type a few months ago, 

 after reading Dr. Lindley's dissertation on "A Study of 

 Puzzles" {Amer.Journ. of Psych., vol. viii. No. 4). They 

 were for ordinary wire-puzzles, and show a sudden leap 

 from failure to success when the trick of the puzzle was 

 discovered and understood, and after that some slight 

 improvement in rapidity of success as the manipulative 

 details were mastered. 



i "Animal Intelligence: an Experimental Study of the Associative 

 Processes in Animals." By Edward L. Thorndike, AM. (Monograph 

 Supplement to the /'o'fA<'/<'^'Vn/.^rt//Vw, June 1898.) 



NO. 1498, VOL. 58] 



Passing reference may here be made to Dr. Lindley's 

 interesting study above mentioned. He finds by observ- 

 ation that the method of the young child is largely that 

 of the aninial. Trial and error, chance success, and 

 direct association are predominant. In older children, 

 who are beginning to generalise the results of their 

 experience, rational procedure based on a considered 

 scheme or plan, makes itself more and more felt. Further 

 observation on similar lines will serve to link such 

 results as Mr. Thorndike's with the human psychology of 

 the text-books. 



To return to Mr. Thorndike's research. The con- 

 ditions of his experiments were perhaps not the most 

 conducive to the discovery of rationality in animals if it 

 exist. The sturdy and unconvinceable advocate of 

 reasoning (properly so-called) in animals may say that to 

 place a starving kitten in the cramped confinement of one 

 of Mr. Thorndike's box-cages, would be more likely to 

 make a cat swear than to lead it to act rationally. And 

 he may further urge that where the string passes out of 

 sight and the bolt is hidden from view, the opportunities 

 of understanding the situation are excluded. All the 

 kitten could think would be : here's something loose and 

 unnecessary to the normal constitution of a box ; I'll try 

 that on chance. But although I do not deem Mr. 

 Thorndike's method so conclusive for the anti-rationalist 

 view as observation under more natural, and, I may add, 

 more sympathetic conditions, yet the form of his curves 

 affords no particle of evidence for reasoned behaviour. 



We may pass over his experiments on dogs and chicks 

 with the barest mention. They serve to support the same 

 conclusions with some differences of detail. 



When we come to his psychological explanation of the 

 nature of the associations involved, I find much to agree 

 with but somewhat to dissent from. Where he argues 

 that animals form no free ideas, I am heartily with him. 

 I have myself contended that they are incapable of 

 analysing a situation. And if in interpreting the facts of 

 observation one's language may seem to imply that the 

 sight of an object and its taste are analysed out and then 

 associated, this is due to the inevitable analytic form 

 which the use of words entails. Animals, in my opinion, 

 do not analyse in this way, and do not form " free " ideas. 

 The utmost that we can allow is that certain elements in 

 a complex situation may, under given circumstances, 

 predominate in consciousness over others ; and this, not 

 through any process of abstraction, but from the inter- 

 play of the nature of the animal and the circumstances 

 of the case. 



But when Mr. Thorndike says that " the groundwork 

 of animal association is not the association of ideas, but 

 the association of idea with impulse," I for one, as at 

 present advised, am not prepared to follow him. 

 " Impulse," he defines as "the consciousness accompany- 

 ing a muscular innervation apart from that feeling of the 

 act which comes from seeing oneself move, from feeling 

 one's body in a different position, &c." Now in the first 

 place this involves the assumption that physiological 

 innervation is accompanied by a specific form of con- 

 sciousness here termed "impulse." The question is still 

 sub judice. But there is, at any rate, much to be said in 

 favour of the view that consciousness is directly stirred 

 only by afferent nerve-currents, and that the innervation 

 process is itself unconscious, though its effects are com- 

 municated to consciousness by an afferent back-stroke 

 from the motor organs as they move. This alternative 

 view should, I think, have been mentioned, at all events 

 in criticising one who provisionally holds it. On this 

 view the efferent impulse (apart from its effects) cannot 

 be psychologically associated with anything, since it 

 is physiological and unconscious. In the second place, 

 to suppose that one who holds the impulse as such to be 

 purely organic, holds also that " an animal whenever it 

 thinks of an act can supply an ' impulse ' to do the act," 



