Various theories of selection and elimination in learning are reviewed and criticized-(1) the pleasure-pain theory, (2) the confirmation and inhibition theory (Hobhouse); (3) the congruity theory (Holmes), (4) the completeness of response theory (Peterson), (5) the intensity theory (Carr), (6) the frequency-recency theory (Watson), and (7) the drive or motor-set theory (Woodworth, Perry, Tolman). The names given in parentheses are those used by the author, who makes little attempt to see similarities and to trace historical developments of the theories or to analyze them. All but the last named of the theories are dismissed as being only descriptive of the facts and not explanatory, while the last is accepted in toto apparently as explanatory, the author overlooking the fact that this theory fails to show the neural mechanism of learning or to suggest why an act is eliminated and not repeated in successive trials. The question of learning is not essentially one of how the animal persists in the activity till it reaches the consummatory act, this being a different problem though somewhat closely related to learning; the real question concerns the mechanism of elimination. This fact is clearly evident in the theories rejected by Kuo as "merely descriptive." It is important to emphasize this point in order that experiments may be directed at the real crux of the problem. It is probably admitted by all investigators now that the inward "drives " to behavior, of which the consummatory reactions are an expression, are not faculties with arbitrary powers but metabolic changes, internal secretions, interstimulation of various part processes in the organism, etc. These factors serve to make the organism react selectively to certain of the numerous exteroceptive stimuli, but how do the erroneous responses-errors from the standpoint of the consummatory act become suppressed and eliminated? It is the how of which we are yet so ignorant, and ignoring it is no solution of the problem. The pleasure-pain theory is hardly stated in scientific terms according to the criteria mentioned in a foregoing paragraph. The same thing is true of the "law of effect," so frequently utilized by some writers; it does not get at the how. The inhibition theory assumes what is to be explained. The intensity theory needs more clarification, and it probably reduces, so far as intensity of response is concerned, to some sort of consistency or congruity conception. The frequency-recency theory is susceptible of experimental investigation of a definite quantitative nature, and in the light of the accumulating evidence it is becoming apparent that frequency-recency factors are in themselves but fixing, not selective, agents in learn

ing; that they tend to fix wrong as well as right responses. Several writers have seen this clearly and various drainage theories have been attempted, but the chief difficulty has been in most cases that the theories have not been amenable to experimental test.

As the sponsor to one of the theories rejected by Kuo as descriptive, I wish to point out that this author has apparently overlooked the two experimental studies supporting the completeness of response theory.2 The second of these articles analyzes quantitatively the merits of the frequency-recency theories of learning. My own results agree entirely with those of Kuo, but this author seems to miss the nucleus of the difficult problem of learning. The congruity theory seems to be going in the right direction, and I acknowledge much help from its early statement both by J. M. Baldwin and by Holmes; it places the emphasis on the relative consistency of the several impulses to action. This is what I have aimed at in the completeness of response theory, suggesting that the advantage lies with the impulses which give the individual-or the dominant determining tendencies, if this is preferred-the completest expression. It seems to me, however, that to account for the inhibition of the "ill-adaptive acts" we must assume some sort of overlapping of the several stimulus effects so that they can operate simultaneously; and this view certainly has much support from observation-both objective and introspective-of the higher deliberative acts. There is nothing in the motor-set theory accepted by Kuo that is inconsistent with this conception or new to it, but it does not go far enough to meet the real issue. I find much that is good in the motor-set conception of behavior, and accept all of it so far as I know, but cannot regard it as a theory of the selection phenomena in learning, though it may explain selection in response.

2 PETERSON, J. The Effect of Length of Blind Alleys on Maze Learning: An Experiment on Twenty-Four White Rats. Behav. Monog., 1917, 3, No. 4; and Frequency and Recency Factors in Maze Learning by White Rats. J. of Animal Behav., 1917, 7, 338-364. A study on "Learning When Frequency and Recency Factors are Negative" will appear in the J. of Exper. Psychol. In this forthcoming article frequency and recency of response are distinguished from frequency and recency of stimulation.

SPECIAL REVIEWS

S. SMITH & E. R. GUTHRIE. General Psychology in Terms of Behavior. New York: Appleton, 1922. Pp. xii+270.

The authors say in the preface that "an attempt is made to state in terms of behavior the facts and principles of general psychology. The facts set forth are those which lead the student to a systematic explanation of his own conduct and that of his fellows." Not all writers who have in the past promised to state "psychology" in terms of behavior have fulfilled their promise. Smith and Guthrie really do what they promise. The seven chapters of this book, comparatively small and yet full of content, are entitled "The Elements of Behavior, Instinct, Learning, Coenotropes, Perception, Human Motives, Social Psychology."

It is the business of a reviewer to state his difference of opinion where the matter concerned seems important; but that must not be interpreted as belittling the value of the whole book, which will prove to be an unusually valuable text.

The authors say (p. 4) that "along the nerve structures pass nervous impulses." The reviewer believes that nervous functioning is of the nature of streaming, that the term "impulse," suggesting something momentary, is for that very reason very misleading and ought to be especially avoided in the introductory teaching of psychology. The authors say (p. 5) that “The great number of neurones and the complexity of their connections account for the fact that an impulse leaving a particular sense organ may find its way to one group of muscles at one time and to another group of muscles at another time." The reviewer believes that this fact can not be explained ("accounted for ") by the mere use of the vague term complexity," that it would be compatible even with relative "simplicity." Further, "Accommodation for near vision gives us a sensation of muscle strain. In distant vision and dim light such sensations are practically absent" (p. 11). The reviewer feels. compelled to conclude that both authors are "far-sighted" and therefore do not know better.

66

All actions are classified by the authors so far as "external' actions are concerned under three groups as (1) orientation, (2)

locomotion, and (3) intervention. "Man, by means of his elaborate intervention responses, so constructs the world about him as to lessen his dependence upon movements of orientation and locomotion, with a resulting increase of convenience and safety." The reviewer would go a little further than this mere three-fold division. The authors add, then, two "internal" responses, (4) "movements of visceral muscles," and (5) "glandular secretions." The authors use this occasion for talking of "emotion" in the traditional way (p. 37): "Emotional expression is made up of overt bodily movements and of characteristic internal responses accompanying them." The reviewer asks if anyone knows of any (more than momentary) life activity whatsoever, emotional or unemotional (blacksmithing, for example), that is not so made up. When will psychologists come to recognize that a distinction between emotion and other action cannot be based on the ever present internal secretions? But why distinguish at all? Is it really necessary for psychology to talk of emotions just because poetry talks of them?

The authors deserve much praise for making it thoroughly clear in their text that "the instincts of babies are more numerous and less complex than are the instincts of the young of other species. . The elaborate instincts of lower animals in man

all are learned." Whether the emphasis placed by the authors on the distinction of "precurrent and consummatory" responses is equally praiseworthy, the reviewer is not so sure. "The final response that removes the maintaining stimuli is called a consummatory response." One example given by the authors is this: "When the trail is lost, the dog is left with an abortive tendency to respond to an odor stimulus." To the reviewer the authors have not made clear what that "abortive tendency" in consummatory responsiveness is. Is it something (yet unknown to psychology) which now steps in and plays the rôle of "the maintaining stimulus "? It looks, further, as if the dog's last (unfortunate) jump (the final response) which did break the trail were in accordance with the authors' definition "the consummatory response." Unless we introduce teleology it must be so. Further, is it really true that food seeking is always terminated by eating? In the squirrel, in the honey bee? Combat always ends when the foe is routed? It is never ended by both parties before? And it is only in the mythological Battle of the Huns that the fighting motions continue even afterwards? The notion of the consummatory response ought not to be overemphasized. However, the treatment as given by the

authors has pedagogical value even though it may lack the explanatory feature. It is certainly true, as the authors say (p. 64), "that the consummatory response in any chain reflex is a convenient basis for classifying instincts." They classify animal instincts ("in the case of man, these utilities are served for the most part by learned acts") under five heads: Swallowing Food, Copulation, Securing Shelter, Defeating Antagonists, Cleanliness.

"Psychology," the authors rightly say, "must explain how habits are developed." They explain it in the following two-fold way: (1) "Through use, the tendency of a response to follow its stimulus becomes better established." The reviewer does not believe that that is a general truth. Do all our original reflexes, daily exercised, become better and better established all through life? Do we sneeze the better the older we get? He has stated his own opinion elsewhere, in his own textbook, also years ago in his book on " Behavior," properly distinguishing between lower and higher centers. (2) "The conditioned response.-How may a response be provoked by a new stimulus?" The reviewer holds that this second "law" of habit formation is the only one that really exists, but that it includes, in a certain limited sense, the fact mentioned under (1), although that first fact, plain and simple, is not a law of habit formation. The "explanation" of habit formation, as given by the authors, seems thus to the reviewer a little too superficial, lacking a penetrating analysis of all the factors. Still, this superficiality is a usual feature of virtually all psychology texts and should not be held out against this one in particular.

“Habits which men universally share, we call coenotropes. They are the products of original nature and commonly shared environment." Example (p. 146): "Everybody knows how to use a stick as a weapon, but no one does so instinctively." Under coenotropes the authors discuss also play, which they call both an "impractical " conduct and also "an incomplete act given in response to an incomplete situation." The latter, the incompleteness, does not seem to the reviewer an essential characteristic of what anybody would call play. The play with a high-powered automobile, for example, not infrequently completes a whole life circle. And kissing games are not always, contrary to what the authors (p. 152) tell us, supervised by adult chaperones. On the other hand, the fox in the fable, when he finally ceases to jump at the grapes hanging too high, cannot, because he never consummates the feast, be said to be playing. Does he mean to call himself a playful animal when he pronounces the