ies can ever undermine our work upon habit formation, THE AIM OF BEHAVIOR 53 efficiency of training methods, the reactions of human be- ings to one another (ethical problems, etc.), the retention of habits, etc., is hardly likely. The tendency towards too great simplification is manifested in another way. Because it is found in a few cases (or in many) that animals respond positively or negatively to a physical stimulus without training it is assumed that the mechanics of response have been put upon a simple physico-chemical basis, that there is no longer any pressing need for a study of the reactions of animals, and that further work along such lines must be looked upon as routine.
No one believes more thoroughly in the complete physico- chemical nature of all response, from the simplest to the most complex, than the author. We are prepared to state our belief in the view that we shall ultimately be able to trace the complete set of physico-chemical changes (quan- titative energy transformations) from the moment of in- cidence of the stimulus to the end of the movement in the muscle. In fact it is one of the goals of behavior to assist in making this possible. But again we insist that all of the facts about response be considered. In very few cases have we reduced reactions to their simplest terms. When we have done so we shall have no c[uarrel with the label which is at- tached to them, be it that of "' simple reflex " or '' tropism." Analysis of reaction must end somewhere. The simple reflex is a convenient hypothesis. It may not be a final goal any more than was the atom in chemistry or than the electron of the physicist may be twenty-five years from now. Both the electron and the simple reflex w^ork well enough at present as convenient means of expressing the results of scientific analysis. The present end of analysis, then, in behavior wall be the reduction of complex con- genital (instinct) and acquired (habit) forms of response to simple reflexes. Having reached the goal it follows that complete predictability, both of stimulus and response, will be afforded.
Furthermore, behavior is interested in synthesis as well as in analysis.^ The building up of habits from simple re- ^ It must be remembered that synthesis is essential to complete analysis of even tropism-like responses. Cole, Holmes, and others have 54 SOME PROBLEMS ENUMERATED flexes for practical and theoretical (and ethical) purposes is as important as tearing them down. We should be able ultimately to formulate laws of habit formation (effect of punishment, of different training methods, retention, etc.)- In the invertebrate realm w^onderful progress has been made towards reducing complex responses to more elementary forms. Progress has been much slower in the case of the responses of the higher vertebrates. That some advance has been made appears clearly in our studies of instinct, and of sensory and motor habits.
Summary. — In this chapter we have attempted to throw the problems of behavior into three great classes, — sense- organ functions, instinctive functions, and learning. We have enumerated some of the separate problems in each division. It is interesting to consider now for a time what we would have before us if a large number of such prob- lems were solved: a group of isolated facts or a system of behavior? As was pointed out in Chapter I, psychology, as it is at present adhered to, would say that the results would lack coherence and that the needed coherence could be obtained only by interpreting the facts in terms of con- sciousness. But introspection likewise gives us isolated facts (?). Titchener * points out that introspection is only a method. A method, even a legitimate one, can give us only facts. Inferences, generalizations, theories, hypothe- ses, all are necessary to a science. But these are open to us. We shall obtain first a systematic grouping of results which will constitute the phylogeny of behavior; second, a similar grouping to constitute the ontogeny of behavior; and thirdly, correlations between structure and function Avhich will be worth while. These three strata of facts will furnish the neurologist, trained in physical and physiological chemistry, with a set of phenomena which is tangible enough to guide research. Finally the raw material will shown that in the case of certain insects and frogs, the tropism-like responses to light (positive or negative reactions to size and intensity differences) fail to appear when definite objects are in the range of the visual receptors.
* American Journal of Psychology, XXIII, p. 447.
THE AIM OF BEHAVIOR 55 be at hand whereby behavior can be regulated and pre- dicted. The behaviorist will have his theories, his inferences, and his generalizations. A limited set of observations upon the visual responses of animals possessing a fovea compared with a similar set upon animals possessing only a modified retinal area may lead to the generalization that only ani- mals with a fovea are capable of responding to differences in the wave-length of light. We may then hazard an ex- planation of the phenomenon in terms of the lack of devel- opment of certain sensory structures, or of certain photo- chemical substances. In general, we have ahead of us prob- lems which, while not sharply separated from those of the experimental zoologist, the physiologist, and the physicist either in subject-matter or in methods, are yet nevertheless distinct and independent. We have divided their labor and have marked off our problems. They are not likely to object to this or ever to claim any portion of the province which we are likely to want. Rather, they will gladly re- linquish all claim upon many of the problems which are ordinarily classed as belonging to their fields. There should follow a constant interchange, among these sciences, of behavior material, data on evolution, neural structure, and physiological chemistry. The behaviorist has no need to fear that this close contact with other biological and physical sciences will injure him. His field is too indepen- dent and too necessary to the other sciences for him to have misgivings about its future. If he becomes lost in a general biological mirage some other cause than his rejection of a psychology based upon introspection must be advanced to account for it.^ BIBLIOGRAPHY Thorndike, E. L.. Animal Intelligence. New York, Macmillan, 1911. Washburn, M, F., The Animal Mind. New York, Macmillan, 1908. Yerkes, E. M., The Dancing Mouse. New York, Macmillan, 1907.
' It needs to be emphasized here that training in behavior should be accompanied by training in histology, physiology, and experimental zoology, and preceded by training in chemistry and physics. It is hoped that the day is past for considering philosophy to be the proper subject to pursue along with behavior. In saying this we do not mean to underestimate the value of philosophy in any way.
CHAPTER III APPARATUS AND METHODS Introduction. I. Stimulus to general activity: stimuli to locomotion. — (a) General stimuli positively responded to. — (&) General stimuli negatively responded to. II. Methods of studying the receptors of animals: forcing the formation of sensory habits. — Control box. — Pawlow's method. — Methods dependent upon in- stinctive response. — Control and auxiliary methods. — Criticisms of the methods for determining the sensitivity of receptors. III. Apparatus for obtaining specific stimuli: apparatus for ob- taining monochromatic light. — Use of apparatus. — The selenium cell. — Device for securing a purified spectrum, — Apparatus for testing response to white light, form, and size. — Apparatus for producing auditory stimulation: apparatus for obtaining con- stant air supply. — The Helmholtz system of tandem-driven forks. — Animal control box for work on audition. — Yerkes' apparatus for testing hearing in frogs. — Apparatus for obtaining olfactory and cutaneous stimuli. — Yoakum's temperature apparatus. IV. Methods of studying motor habits: introduction. — The problem box method. — Description of boxes. — Description of maze experi- ments.— Apparatus for the study of the " delayed reaction."
Introduction. — In this chapter only the more general methods and apparatus nsed in the study of behavior can be considered. The books of Jennings, of Mast, and of Loeb give adequate descriptions of the technical methods and the apparatus employed in the study of invertebrates. The presentation of the technique used in investigations of the behavior of vertebrates must necessarily be meager and incomplete. This field has been developed rapidly and the motives for development have been very varied. These differences in standpoint have had their effect upon the types of apparatus which have been devised and upon the methods employed in conducting experiments. In past years only the roughest kind of preliminary work has been attempted. In more recent times there has grown up a strong tendency to do more careful work with better instru- ments and by methods which are capable of being standard- 56 STIMULI TO ACTION 57 ized. This is shown by the general tendency on the part of investigators to standardize their stimuli and to adopt iron-clad rules concerning the physical control of the ani- mal and the objective recording of results. Up to the present time we have had to be content with a small begin- ning in the various laboratories. There has not been sufficient general interest in the subject to justify large expenditures for apparatus. It is hoped that the next few years will bring decided improvement in the equipment and housing of behavior laboratories. What is said here about apparatus and methods must be looked upon as suggestive. It is frankly confessed that probably too much considera- tion is given to those used in the personal work of the writer.
I. The Stimulus to General Activity ^ Stimuli to locomotion. — The first and most obvious need in behavior is to provide some form of stimulation which will make the animal move in some way. Unless the animal will work steadily we are powerless to force habits upon him. We have as yet explored only a few of the various forms of stimuli which may be used for the purpose of calling out desirable general activity. As will be shown in Chapter VII, we have attempted to establish habits on the basis of providing " satisfiers " or " dissatisfiers " for the animal. This point of view is anthropomorphic and one which cannot be applied, or rather it is one which no one would apply in the realm of animals below the vertebrates. In many of the lower forms general activity of the kind we obtain by using food as a stimulus with vertebrates, can be called out successfully by many different kinds of stimuli. Light, e.g., can be so adjusted in intensity that we ^ We necessarily use the term stimulus rather loosely in be- havior at present. We employ certain objects to make the animal move. Let us call such objects general stimuli. While moving in response to the general or primary stimuli {e.g., food or electric shock) we find that we can make him respond positively or nega- tively to light stimuli, sound stimuli, etc.. that we can make him learn to open a problem box, etc. We shall call these latter objects specific or secondary stimuli.
58 APPARATUS AND METHODS can call out at will (1) general but undirected activity (shock movements), (2) an immediate positive response, or (3) an immediate negative response. We never think in such cases that the animal goes towards the light because it is a satisfier or away from it because it produces dis- satisfaction. For this reason we prefer to speak of stimuli positively and negatively reacted to in discussing the means by which we evoke general response in the animal of a kind necessary to the formation of habits. We shall consider under two heads the more common forms of stimuli which we have found to work in a satisfactory way. (A) General stimuli to which the animal responds positively; e.g., food and objects connected with the sex life, etc. (B) Those to which the animal gives a negative response; e.g., the elec- tric shock, extremes of temperature, intense light, etc.
(A) General stimuli positively responded to. — Food has been the most commonly used form of stimulus for induc- ing general locomotion. The use of food requires a rather careful technique. If a large amount of food is given on one day the animal will be sluggish on the following day. If too little food is offered the animal becomes too eager. Long-continued experimentation with light feeding pro- duces a starved and unsatisfactory animal. The most thorough way to go about the problem of food control is to surround the animal with the food which is necessary for its well being and then determine the length of feed- ing period required properly to maintain metabolism. The maintenance of metabolism can be determined by weighing the animal. It is not fair to talk of the cruelty and in- humanity of keeping the animal hungry, as has been done by several writers, until there is some factual support for the charge. There is not the slightest difficulty in keeping the animal in perfect condition and at the same time hun- gry enough to work properly. We have found no animal (as used under ordinary laboratory conditions) which does not work well when food is used as the general stimulus. Yerkes states that he was unable to use it satisfactorily with the dancing m_ouse. We repeated the maze experiment on the dancer with food as the stimulus. So far as we STIMULI TO ACTIVITY 59 could judge the method was as satisfactory, from the standpoint of the rapidity of learning and from that of the well being of the animal, as the punishment method of Yerkes. Observers have stated that certain reptiles and amphibia are not commonly in a physiological state where food is positively reacted to; e.g., some of the reptiles eat as infrequently as once a month. Furthermore, in work- ing in the field it often happens that the animal under observation is not under our control to the extent where we can regulate its food supply. Under these conditions some other form of stimulation must be employed. Sex objects will certainly work well as a general stimulus in the case of some animals. This was shown by Dr. Kimball at the University of Chicago in his work upon rabbits. (This work was never published.) We have gone far enough in repeating his experiment to show that the female can be put in the problem box in the place of food. The male will attempt to get access to her by the same methods which he will employ in the case of the food. The male monkey will attempt not only to get at the female, but also, if at all aggressive, to reach other males in the same way. The difficulty with using sex objects as a general stimulus is due to the fact that sensitivity to such objects is as a rule seasonal. It is extremely difficult to use such a form of stimulation i;i the case of most animals where experi- mentation is to extend over a long period of time. The rabbit forms a notable exception to this rule. Even with that animal only the male can be experimented upon in this way. With some birds, not directly under the con- trol of the experimenter, and possibly with other animals as well, objects connected with the reproductive function are the only ones which can conveniently be used for mo- tivation during the breeding cycle. This is especially true with the terns. It was found that during the nesting period the nest and eggs could be used in place of food. The activity induced was fully as great, apparently, as that ob- tained in other animals by using food.
(B) General stimuli negatively responded to. — A few years ago Yerkes advanced the view that " punishment " 60 APPARATUS AND METHODS formed by far the best single method of motivation. Pun- ishment was obtained by means of an induction shock ad- ministered by a metal grill connected with an inductorium. This punishment method has not worked any too well. It has been criticised by Hamilton who found that it made his dogs restless and hesitant; by Lashley, who found that it made rats, where association was difficult, after a time refuse to work. Under ideal conditions, where the problem offered is easy, it apparently does hasten association. Hoge and Stocking found that discrimination between a 16 c.p. and a 2 c.p. light was hastened by administering food as a reward, and punishment for failure. I.e., a com- bination of the food and punishment method, in their ex- periments, seemed to be most satisfactory.^ In discrimina- tion work generally troublesome position habits are very soon set up. At first glance it would appear as though punishment were extremely useful in eliminating such habits. In actual practice it does not always have this meritorious effect — the animal becoming sulky under con- tinued punishment. Animals differ greatly in their sen- sitivity to the induced current. Rats are oftentimes badly disturbed by a current which is too weak for the human being to feel. Other animals are extremely resistant even to fairly high currents. The rabbit is not disturbed in the smallest degree by a current which is unbearable to the human being. Even when all possible objections are urged against the use of punishment it must be ad- mitted that in some problems (especially in the formation of sensory habits) where the animal is neither too resistant nor too sensitive, the use of the induction shock is to be recommended.^ In addition to the electric shock, other forms of stimuli which are reacted against have been tried. In sensory habits (with monkeys) where two specific stimuli are emploj^ed we have tried the experiment of putting sharp needlepoints (not visible) upon the handle of the food box placed immediately underneath the color to be reacted ^ For a discussion of habits arising under controlled incentive, ^ On the quantitative use of faradic stimulation, see Martin.
FORMATION OF SENSORY HABITS 61 against. This method does not work as well as the electric shock. The reason is not hard to see. The avoiding reac- tion should be aronsed at such a point that the animal ivhen turning away from the grill can immediately turn towards the right stimidus. This makes the right response on each successive trial the one which was last executed (recency) and also the one most often executed (frequency). The role played by recency and frequency in the fixation of arcs in habits is discussed in Chapter VII.
II. Methods of Studying the Receptors of Animals Forcing the formation of sensory habits. — The sensory range of animals and the sensitivity of their receptors can best be studied by forcing the formation of sensory habits. The method for establishing these has taken several dif- ferent forms. In forcing such habits it has been usual to confront the animal simultaneously with not more than two specific stimuli, although this method has by no means been universally employed. Multiple stimuli methods are sure to be found useful in certain types of work.* The general procedure is as follows: the animal is placed' in an experiment box. This box has a home compartment * Multiple stimuli methods have not been worked out as yet in any great detail. Hamilton has made a series of tests upon mammals (p. 111). His method places the subject in an experimental situa- tion which may be reacted to in many different ways and with varying degrees of satisfactoriness or adequacy. The subject of the experiment is placed in a small room on one side of which there are four doors. From experience he learns that he may escape by one of the doors and one only, but which of the four to choose is his problem, for it is the plan of the experimenter to lock, in a given trial, the door through which the animal escaped in the previous trial and tAvo others. Any one, then, of the three doors may be unlocked in a given trial. The animal has absolutely no way of predicting which is unlocked. Yerkes's method is similar in purpose to that of Hamilton but it offers a somewhat wider range of useful- ness. It consists essentially in the presentation to the subject — bird or mammal; young or old; normal or abnormal — of a bank of twelve keys numbered from left to right, one to twelve. The subject is given to understand verbally, or through actual experience with the apparatus, that pressing some one of the twelve keys will yield certain desired results such, e.g., as the displaying of a picture, the presentation of food, the ringing of a bell. Success in the experiment means simply pressing the key which brings the desired result. The 62 APPARATUS AND METHODS where the animal is restrained until the experimenter is ready to make the test. At the will of the experimenter the door of the home box is opened and the animal is then con- fronted with both stimuli simultaneously; e.g., a circular and a square translucent plate illuminated from behind, each by lights of known candle-power. The circle and the square are separated from each other usually by a distance of 25 cm. A partition divides one end of the stimulus chamber into two compartments. The animal, leaving the home box, may respond by entering the compartment illumi- nated by the square or the one illuminated by the circle. Usually several preliminary responses are taken to see if there is any instinctive tendency to respond positively either to the one or the other of the two stimuli. Where experiments upon retention are to be carried out later, there is general agreement that the animal must be trained against his positive tendency, — i.e., he must be trained to respond positively to the stimulus primitively reacted against.