ZING- YANG Kuo

If we accept mechanistic and non-Darwinian con-cepts of behavior as basic to behaviorism, we mustconsider both Jacques Loeb (Loeb, 1889, 1918) andJohn B. Watson as the pioneer behaviorists. Theword "behaviorism" was coined by Watson (1913,1914), and the movement was, at first, a method-ological revolt rather than an attempt to build a newsystem of psychology. At that time, psychology wasstill the handmaiden of philosophy, and Watsonhimself had been trained by philosophically-oriented psychologists, although his own researcheswere mainly on animal behavior. (In this respect,Loeb, as an experimental biologist, was freer fromthe bondage of philosophy.) Soon such philoso-phers as Bertrand Russell and Ralph Barton Perrybecame interested in behaviorism, and psycholo-gists like Tolman, Weiss, Lashley, and even Watsonhimself attempted to give behavioristic interpreta-tions of the mental entities - thinking, memory, andthe like - losing sight of the original aim of the

behavioristic movement. To call thought processessubvocal speech (Watson, 1914, 1919); to give be-havioristic interpretations of consciousness, ideas,memory, and the like (e.g., Tolman, 1951); to sub-

stitute for mental faculties various new names of"hypothetical constructs"; to devise ingenious andelaborate schema of cerebral structures, applyingthe principles of electronic computers or informa-tion processing; to explain reasoning, intelligence,memory, etc., or to rename "volition" or "will" asdrive or motivation (with physiological implica-tions): all these are philosophical speculations atbest. They hardly help us solve any basic problemsin behavior. Instead, the behaviorist should concen-trate his efforts on behavioral problems, leaving allthe problems of mental concepts alone; these, as amatter of fact, shQuld be imperceptibly and ulti-mately absorbed, never to reappear, as our knowl-edge of behavior advances and broadens and as ourexperimental techniques, especially the biophysicaland biochemical ones, are improved.

The fundamental error of the behaviorists,therefore, was that they paid too much attentionto the claims of philosophers and philosophically-oriented psychologists and tried to use objectiveor physiological constructs as replacements for thenames of the old mental entities. Watson's greatestcontribution to the behavioristic movement was nothis hypothetical constructs, but his repudiation ofintrospection as a scientific method, his pioneerwork in the study of the behavior of newborn in-fants, and this statement in the forceful and coura-geous preface to his 1919 book Psychology from theStandpoint of a Behaviorist:

From Zing-Yang Kuo, The Dynamics of Behavior Develop-ment (New York: Plenum Press, 1967), pp. 3-26.

FROM W.:\TSONIAN BEHAVIORISM TO BEHAVIOR EPIGENETICS

The key which will unlock the door of any otherscientific structure will unlock the door of psychol-ogy. The differences among the various sciencesnow are only those necessitated by the division oflabor. Until psychology recognizes this and dis-cards everything which cannot be stated in the uni-versal terms of science, she does not deserve herplace in the sun. . . nor does the author claim be-havior psychology as a creation of his own. It hashad rapid development and is a direct outgrowth ofthe work on animal behavior. It is purely an Ameri-can production. . . .

On the other hand, the concepts of tropism(taxis) reflex and of instinct and learning and thestimulus and response (S-R) formula, which mayappear to us now as extreme oversimplifications oras obsolete, refer to real behavioral problems and, assuch, must be carefully examined and investigatedby the behaviorists before they can be discardedand replaced by newer and broader concepts ofbehavior.

procedures, and (2) mathematical-mindedness. Thestatistical requirements for any animal experimenthave become so rigid and so sophisticated that noeditor of an American technical periodical will ac-cept any paper on animal behavior without ad-equate statistical data. These research techniquesconstitute a very commendable contribution, andno coming student of behavior, whatever his scien-tific outlook, can afford not to master them beforehe begins his research on animals. And manyyounger European workers on animal behavior havejust begun to appreciate this scientific achievementof the American animal psychologists.

However, it must be admitted that much Ameri-can experimental work, especially the learningresearch on rats, has been carried out in a verynarrowly confined environment so that conclusionsfrom the results of such experiments may not beaccepted without a great many reservations. More-over, the stress on mathematics has gone beyond thebounds of scientific proportion. Although recentprogress in mathematical physics and computer en-gineering has been rapid, the phenomena ofbehav-ior are far more complex than physical events, andthe science of behavior is still in its infancy in com-parison with the very advanced stage of the physicalsciences. For these reasons, I feel strongly that it ispremature to devise "mathematical models" for theprediction of behavior. If many learning theoristshave apparently succeeded in their mathematicalpredictions, that can be explained by the fact thatthey have ignored the great complexity and variabil-ity of behavior and have reduced behavioral phe-nomena to an extremely few simple parameters. Forexample, the behavior of a fly would be a very com-plicated affair if we took into account the complex-ity, variability, and interplay of such determiningfactors as morphology, biophysics, biochemistry,developmental history, environmental context andthe various characteristics of the stimulating ob-jects. However, if we simply put a fly in a very smallblackened box with two small openings both pen-etrated by outside light and send a stream of cold airthrough one opening and a stream of warm airthrough the other, we could predict with almostmathematical accuracy by which of the two open-ings the fly would escape from the box and in howmany trials. But, in so doing, we must ignore otherdetails of the fly's activities inside the black box. Inother word~- the ~lIrre~" nf ~lIrh "n pvnpr;mpnt,,)

The Descendants ofWatsonianBehaviorism

Watson's S-R formula and his almost unquestion-ing acceptance of the concept of conditioned reflexas the key to the study of behavior were ratherunfortunate because of their oversimplification ofthe phenomena of behavior. However, with someminor variations such as Tolman's "purposive be-haviorism," the S-R formula and conditioned reflexconcept have become the standard-bearer ofbehav-iorism and the keynote of American animal psy-chology for the last fifty years. Its main thesis isanimal learning. Its basic principle is conditioning-Pavlovian or Skinnerian. And the animal used inlearning experiments has been mainly, if not exclu-sively, the albino rat. Theories of learning may vary,but none has been diverted from Watson's originalS-R formula.

Whether or not one agrees with the fundamentalviews of the psychologists of learning, one mustadmit that the descendants ofWatsonian behavior-ism have evolved what has now become an Ameri-can tradition in psychology. There are two aspectsof this tradition: (1) The strictly experimental atti-tude which has resulted in the development oflabo-ratory techniQues with rilZid standards and elaborate

58

prediction is due not to the accuracy of mathemati-cal theories or models, but rather to the oversimpli-fication of the experimental situation in animal

learning.

The Unsolved Problemof Behaviorism

Because the behaviorists have confined themselvesto methodological approaches without broadeningtheir outlook on behavior and raising their scientifichorizon beyond the simple S-R formula and theconcept of conditioned reflex, they have been forcedto fall back on the problem of the "psyche." If evenlearning cannot be explained without postulatingmotivation, learning set, or modification of cerebralstructures, no behaviorist can avoid the problem ofintervening variables. To call them physiologicalcounterparts of behavior neither explains behaviornor satisfies the demands of the mentalists. Or toadmit the existence of neural processes, while deny-ing the relevance of the neural connection betweenstimulus and response, as does the operant behav-iorist, is a scientific evasion at best. The protestationagainst behaviorism from philosophers, mentalisticpsychologists, or even social scientists (cf. the sym-posium on "The Limits of Behaviorism in PoliticalScience," American Academy of Political and So-cial Science, Philadelphia, 1962) may be ignored.But when a physiologist like Adrian complains thathe cannot understand "thought" without referenceto the mind or when the psychologist of animallearning insists that he, too, needs some such con-cepts as motivation or learning set in order to ex-plain the learning process, it seems that the time hascome for us to re-examine the current scientificoutlook of physiology and reassess the basic scien-tific value of the behaviorists' S-R formula. Perhapsthe merger of physiology with the science of behav-ior, with a much more broadened prospectus, maylead to the eventual dissolution of the problem ofintervening variables.

as a countermovement against it before World WarII. The Gestalt psychologists' criticism of the over-simplification of the S-R formula and the conceptsof reflex and chain reflexes was valid. Nevertheless,I feel compelled to part company with them inpractically every phase of their treatment of animalbehavior. This, even though my concepts ofbehav-ioral gradients and environmental context may ap-pear to some as another holistic approach. As amatter of fact, both concepts are not only experi-mentally analyzable, but can be analyzed only by theatomistic methods that have been anathematized bythe Gestalt psychologists.

It should also be noted that, besides beingpredeterministic concerning development, the Ge-stalt concept of behavior is not free from vitalistic

implications.

The Protest of Ethology

Although C. O. Whitman and O. Heinroth andvarious others have been credited as pioneers in thefield, it was Konrad Z. Lorenz and Niko Tinbergenwho systematically conceptualized ethology. Ethol-ogy was unknown in America before World War II,yet its impact on the thinking and research ofAmerican behavioral scientists was no less dramaticthan Gestalt psychology had been in the late 1920sand the 1930s. Like the Gestalt psychologists, theethologists' chief target was American behaviorismin the current sense. The main principles of ethol-ogy consist of revival of the concept of instinct andan emphasis on (1) naturalistic observation, (2) so-cial aspects of behavior, (3) species-specificity, (4)the adaptive value of behavior for the survival of thespecies, and (5) the need for a diversification ofanimal species in behavior studies. These are allaimed at American learning psychologists and per-haps rightly so, especially at those who had confinedthemselves to running rats in mazes.

Behaviorism Reconsidered

Gestalt Psychology as a ProtestAgainst Behaviorism

In reviewing the history of the behavioristic move-ment one most important aspect has been long over-looked: the ontogenetic implications. These wereclear to me in the basic tenets of behaviorism, inWatson's own later interest in the behavior of hu-

In reviewing the history of behaviorism it may befitting to say a few words about Gestalt psychology

59FROM WATSONIAN BEHAVIORISM TO BEHAVIOR EPIGENETICS

man neonates, in the vigorous anti-instinct cam-paign in the 1920s, and in Watson's effort (1924) tocatch up with the anti-instinct trend. An objectiveand unbiased historian of science should have beenable in the late 1920s to forecast the future trend ofdevelopment of the original Watsonian behavior-ism, namely, the ontogenetic approach to the studyof animal behavior. Despite theoretical differencesthat were an inevitable historical outcome, the greatproliferation of studies of prenatal behavior of vari-ous vertebrates and of newborn infants in the 1930swas not merely a coincidence. The studies of vari-ous aspects of the ontogeny of postnatal behaviorafter World War II must be regarded partly as acontinuation of the 1930s trend and partly as silent(in most cases) protest against the rat-learning psy-chologists from Tolman to Hull, Skinner, and theirfollowers, who seem to have deviated from theoriginal tenets of Watsonian behaviorism.

"The basic aim of this book is to present theepigenetic point of view as a logical crystallization ofthe continuous and truly behavioristic trends initi-ated by Loeb and"Watson; it is an attempt to replacethe versions of Tolman, Hull, Skinner, and thelearning theorists on the one hand and the mainconcepts of behavior advocated by the ethologistson the other." In what follows we shall discuss thevarious aspects of this viewpoint, illustrated, when-ever possible, with experimental or observationalevidence; at the same time, we shall suggest a neworientation for future behavioral studies.

The Meaning of Epigenesisin Behavior

We shall define behavioral epigenesis as a con-tinuous developmental process from fertilizationthrough birth to death, involving proliferation, di-versification, and modification of behavior patternsboth in space and in time, as a result of the continu-ous dynamic exchange of energy between the devel-oping organism and its environment, endogenousand exogenous. The ontogenesis of behavior is acontinuous stream of activities whose patterns varyor are modified in response to changes in the effec-tive stimulation by the environment. In these epige-netic processes, at every point of energy exchange, anew relationship between the organism and the en-vironment is established; the organism is no longer

the same organism and the environment no longerthe same environment as they were at the previousmoment. Thus, in ontogenesis, both patterns ofbehavior and patterns of the environment affecteach other and are therefore in a constant state offlux; that is, changes in the environmental patternsproduce changes in behavior patterns which in turnmodify the patterns of environment. The epigeneticview of behavior is bidirectionalistic (Gottlieb,1970) rather than environmentalistic, as it considersevery behavior pattern as a functional product of thedynamic relationship between the organism and itsenvironment, rather than as a passive result of envi-ronmental stimulation. In other words, the epigen-esis of behavior is a continuum of the dynamicprocess of interlocking reactions between the organ-ism and the environment, resulting in the reorgani-zation or modification of the existing patterns ofboth the behavioral gradients and the environmen-tal context. At the same time, it does not violate thecurrent views of heredity held by physiological ge-neticists and experimental morphologists, accord-ing to whom ". . . the phenotype is more and moreconsidered not as a mosaic of individual gene-con-trolled characters but as the joint product of a com-plex interacting system, the total epigenotype"(Mayr, 1963, p. 6). The epigenetic behaviorist ac-cepts the morphological structures and their func-tional capabilities at a given stage as the end resultof development in the preceding stages, but consid-ers them merely as one of the five groups of deter-miners of behavior. The higher the phyletic level(especially, when there is a complex social life) andthe more advanced the morphogenesis, the moreimportant are the other determining factors inshaping or modifying behavior patterns. In verte-brates, and in birds and mammals in particular, asontogenesis progresses and as the developing ani-mal widens its environmental contacts, especiallythrough sensory inputs from the distance receptors,more new patterns of behavior are required fromtime to time to meet the new demands of the con-tinuing changes of the environment. Such new pat-terns are determined mainly by the developmentalhistory or historical antecedents of the animal, theexisting pattern of behavioral gradients, and theenvironmental context; morphological structuresand their functional capabilities act as determiningfactors of behavior only in a negative way, that is,they merely set a limit to certain body movements

ZING-YANG KUO6U

(for example, a dog can only snarl at or bite itsenemy but cannot throw a stone at him).

Some Salient Points on BehavioralEpigenesis

The following comments should help to clarify theepigenetic view of behavior and serve as a guide forfuture investigations of behavior:

1. The behavior of animal and man is a continu-ous stream of activity from fertilization to death. Insuch a continuous process of energy exchange be-tween the organism and the environment, behaviornever comes to an end. If there is an end in behav-ior, that end must be death for death is the end of allends. However, in this continuum, behavior pat-terns are in a process of perpetual change - from

moment to moment, from sleep to wakening, fromeating and drinking to defecation and micturition,from combat to copulation, from laying eggs tobrooding and feeding the young, and so on. But forthe sake of scientific convenience, we often takeonly a very small fraction of such ceaselessly chang-ing events for investigation. We are justified in do-ing so only if we are fully aware that we are notreading any teleological meaning into such a smallfraction of ever-changing behavioral events. Nei-ther courtship, threat, fighting, hoarding, nor learn-ing aT-maze to discriminate visual patterns is an"intended " "driven " or "motivated" behavior for

" ,no behavior pattern or series of patterns is an entity,nor does it have any "goal" to reach. When we saythat the animal has reached its "goal" (e.g., the foodbox or copulation), it is an anthropomorphic state-ment of the fact that the sequence of behavioralevents under observation has undergone a shift. Inthis situation, the behaving organism itself has nopurpose, but its stream of activities is arbitrarilysectionalized or fractionalized to serve the purposeor the observer or experimenter. Furthermore, inontogenesis, no movement is a preparatory act toserve a future purpose of the animal. Thus the beakmovements of the chick embryo are no more a pre-paratory action for postnatal food getting than forcrying in fright or pain, crowing, grooming, orfighting, etc. The epigeneticist does not entertainany concept of finalism in behavior. In short, behav-ior begins with heartbeats and ends when all cardiac

movement stops; this is the true and only finalism ofbehavior.

2. Both the American ~nimal psychologists andthe European ethologists have made two basicassumptions: the uniformity of nature (environ-ment) and the uniformity of behavior. Bothassumptions are based on inadequate observation.

As far as environment is concerned, light,sound, temperature, humidity, wind force, and thechemical composition of the air are in a state ofconstant change, except in an elaborately and me-chanically controlled laboratory room. When socialenvironment is added, the variation is far greater.Even the effectiveness of the stimulating object var-ies from moment to moment. If we include intra-organic or endogenous stimulation as part of theanimal's environment, the environmental variabil-ity becomes far more complex. And when we cometo the animal's environment in nature or in the field,environmental changes are far beyond the manipu-lation of the observer. I have been a bird-watchersince childhood, yet I have not seen two birds in thesame nest grow up under the same natural condi-tions; as a result, no two birds react to the samestimulating object in an identical manner. But bi-ased by an a priori assumption of uniformity, stu-dents of behavior are often apt to overlook or ignorethe variability of environment and its effects onbehavior.

To emphasize the significance of the variabilityof behavior let us illustrate it somewhat in detailwith a single case from our unpublished records onfighting behavior in the dog.

The dog in question was a smooth-haired ShanChow named Bobby who was selected for trainingas a fighter. He had all the anatomical qualities of agood fighter. After training, Bobby would alwaysattack any dog in sight, regardless of sex. However,his fighting behavior manifested the followingvariations:

. If the dog attacked offered no resistance, Bobby

would grab him by the back of the neck, shakehim for a minute or less, and then let him loose.

. If the victim happened to fall to the ground on

his back, Bobby would grab his throat and shakehim until the dog was motionless or dead.

. If the other dog offered resistance, fightingwould be very fierce. Bobby would never stopuntil the other dog was either killed or so injured

61FROM WATSONIAN BEHAVIORISM TO BEHAVIOR EPIGENETIC$

. When Bobby was between ten and twenty daysold, the trainer used the sound of a Chineserattle to frighten him. After he grew up to be afighter, the sound of the rattle would alwaysstop him when he was rushing after anotherdog. However, once his attack was already inprogress, the rattle became ineffective.

. Soon after a heavy engagement in fighting,Bobby was reluctant to get involved in anotherbattle.

that he became almost motionless, or until thefighting was interrupted by the trainer.If the other dog's resistance was not easily over-come, fighting became furious and continueduntil the victim was killed or both dogs wereseriously wounded and exhausted, or until theywere stopped by the trainer.If the first fight did not end decisively, Bobbyand his opponent would fight so ferociously onthe second encounter that the trainer and hisassistant would be unable to separate them ex-cept by ducking both dogs into a pond. Climb-ing up from the pond, both would rush to fightagain until Bobby finally succeeded in killing orsubduing his enemy.When Bobby was attacking another dog andsome other dogs who were his friendly compan-ions came to his assistance, he would give up hisvictim and walk away.When Bobby was a small pup, he had been at-tacked by another fierce dog. Later, after he hadgrown up to be the strongest dog in the labora-tory, Bobby would always walk away from hisold attacker whenever he saw him in the dis-tance. However, on one occasion, he was tooclose to avoid him; the other dog then attackedBobby but proved to be no match for him.Fighting was ended by the trainer's interven-tion. From this time on, Bobby would rush afterhis old enemy whenever he was in sight.Bobby was trained to attack any female on sight.But, on one occasion, he rushed to attack a fe-male in heat, and he was changed into a totallydifferent dog. After several faulty attempts,Bobby succeeded in his first copulation. Fromthat time on, Bobby never attacked any bitchwhether in heat or not.When feeding was two to three hours overdue,Bobby was very reluctant to engage in any

fighting.After two or three copulations in a day, Bobbywould not initiate any fighting unless he wasattacked.His tendency to attack other dogs was reducedby about a third during extremely cold or hot

days.If Bobby was five or more pounds heavier thanhis usual weight, he was much less aggressive,his fighting was much less fierce, and it did notlast as long as usual.

This account of the variation in the fighting pat-terns of this dog has been confined to the level ofgross activities. We have not touched upon varia-tions in the biophysical and biochemical com-ponents of the patterns. These would makeour description of the fighting behavior even more

complex.In fact, this description of Bobby's fighting

patterns is an extreme oversimplification of thevariability of behavior. Bobby never attackedthe same dog twice in the same manner. He mightmake a frontal attack, leap on the other's back, grabits neck, or ear, or push it to the ground andgrab and seize any part of the victim's body, etc.Furthermore, his attack varied with the reactionof his victim. A really stereotyped responsepattern, especially in the higher vertebrates, hardlyexists.

The seemingly uniform responses we may some-times observe are more apparent than real. Not onlydoes the same organism not make two identical re-sponses to an identical stimulating object, but evenin such a seemingly simple movement as the liftingof the head of a four or five day old chick embryo, itsamplitude and extent are never twice the same. Thenumber of muscle fibers and their neural connec-tions involved in the head lifting, the extent of themuscular contractions, the energy consumption,and other physical and chemical changes vary frommoment to moment. We must take all these varia-tions into consideration if the science of behavior isnot to remain on the crude descriptive level of grossbodily movement.

3. The developmental history of the animal is ofutmost importance.

4. In view of the variability of behavior and ofenvironment and the diversification of historicalantecedents among individuals of the same species,

62 ZING-YANG KUO

we believe that no two individuals of the samespecies acquire in their life history the same behav-ioral repertoire, except those motor activities whichare the direct results of morphological structuresand their functional capabilities. In animals of lowerphyletic levels and in embryos or fetuses, in whichenvironmental variations within the shell or uterusare more or less standardized and anatomical andchemical factors playa comparatively more impor-tant role in determining behavior, there are greatersimilarities in behavior among the growing animalsof the same species. In such instances there may besome justification in calling such prenatal behavioror behaviors of lower organisms, species-typical orspecies-characteristic (both are better terms than"species-specific," as used by many ethologists andAmerican animal psychologists with the implicationof predeterminism).

On the other hand, when we come to postnataldevelopment, except for certain movements, suchas modes of locomotion or eating or drinking whichare directly determined by anatomical structures ortheir functional differences, the organization of be-havior patterns is so varied and so diversified thatwe can hardly group them into categories and callthem species-specific unless we project our subjec-tive, purpose-directed aims into the end results ofthe patterns, ignoring the variables of such patterns.In other words, unless we project human ends intothe classification of animal behavior and disregardall the variables in environment and developmentalhistory, it does not seem possible for students ofbehavior to arrive at a behavioral inventory or rep-ertoire common to all vertebrate animals of thesame species. Our objective should be to discoverthe behavioral repertoire of the individual animaland its causal factors rather than that of the species.On the same grounds we shall raise questions withregard to behavioral genetics.

5. In methodology, the behavioral epigeneticistinsists on strict laboratory procedure and, at thesame time, makes use of field or naturalisticobservations as a means of discovering problems ofbehavior to be solved in the laboratory. However,we must point out emphatically that it is a danger-ous procedure to arrive at any scientific generaliza-tions from naturalistic observations, especiallywhen such observations are of short duration, be-cause the observer may not have seen enough of the

variability of behavior and environment in the field.Lorenz's observations on the fighting behavior

of wolves and chow dogs is a case in point. Hereports that fighting within either of these speciesalmost always ends harmlessly for both combatants.Whenever two animals are engaged in combat, theweaker one soon lies on the ground and exposes itsthroat while the stronger one climbs over it, andboth bark. The victor then walks away with an ap-pearance of dignity and makes no further attack onthe vanquished. Lorenz concludes that these behav-ior patterns result from an evolutionary process andare of great value for the preservation of the species.I have never observed wolves fighting. However,since my childhood I have been observing thefighting behavior of chows, and since 1928 I haveconducted a large number of controlled experi-ments on the fighting behavior of these animals.Apart from such experimental fights, I have kept norecord, but I have undoubtedly seen over one thou-sand of these fights. Therefore, on the basis of ob-servations and experiments on these animals as wellas on other species, I have concluded that animalfighting is so complex and variable, and involves theinterwoven reactions of so many factors, that itwould be most unscientific to follow Lorenz andsingle out one pattern of fighting and chivalry andexplain it in Darwinian terms while ignoring manyother fighting patterns. The case of Bobby previ-ously described is evidence enough. Suffice it tostate here that there are indeed certain situations inwhich the weaker dog saves itself from injury bylying on its back, a truly successful application ofGandhi's philosophy of nonresistance or Russell'spacifism through submission: "Better Red thandead." However, there are also many circumstancesin which such a response is an invitation to certaindeath, as it gives the top dog an opportunity to grabthe underdog's throat and shake it until the under-dog is dead or at least shows no sign of bodilymovement. I have seen many cases in the field andin my laboratory of truly pacifist and nonresistantchow dogs dying after they had behaved in the fash-ion Lorenz described. A remarkable product ofadaptive behavior through evolution and naturalselection, indeed!

As far as my own observations are concerned,when dogs fight in the manner Lorenz describes,both combatants belong to the class of what I usedto call the Gandhi variety. Both the top dog and the

'ROM WATSONIAN BEHAVIORISM TO BEHAVIOR EPIGENETICS

specific as far as my observation is concerned; theethologists would, therefore, say that this behaviormust be taken as genetically determined throughnatural selection. However, the type of behaviorpattern on the roof or tree top and the type in the airseen incompatible. One favors the notion ofinterspecies coexistence, the other interspeciesstruggle. But why have both types been preserved?Have they survival value at different moments? Tome this is not a question of natural selection, nordoes it have anything to do with the question ofsurvival value for either species. It is a question ofthe ontogeny of behavior during which differentenvironmental complexes have brought about bothtypes of rather incompatible patterns. This is anexperimental problem to be tested in the laboratoryso as to determine the actual environmental-ontogenetic factors responsible for the developmentof such incompatible patterns. If contradictory pat-terns of behavior exist in two species even in natureas is the case between the magpie and the eagle, it atleast indicates that animal life in nature is far morecomplex and variable than the naturalists have

thought.In sum, naturalistic observations are valuable

only insofar as they help the student of behaviorbroaden his outlook so that in devising experimen-tal programs he will look beyond the narrow con-fines of his laboratory animal.

underdog lack the anatomical elements to be good

fighters. As a matter of fact, the top dog of this typehas never engaged in a real fight in its life and isdisqualified for training in fighting. Whenever itmeets a stronger and more aggressive dog, it imme-diately becomes submissive and readily accepts therole of the underdog, although it will reverse its rolewhen confronted with a still weaker and more sub-missive one. Chow dogs of this type in the Chinesevillages are mostly underfed, look unhealthy, andhave a great variety of skin diseases and verminousinfestation. All such dogs are true to the type ob-served and reported by Lorenz. The chow dogs heobserved are probably family pets; unlike the Chi-nese peasants, no Western dog lover will allow hispup to fight with his neighbors' dogs to the finish.One who was familiar with the developmental his-tory and the anatomical make-up of the dogsLorenz observed would probably have arrived at adifferent conclusion. Personally, I have witnessedscores of dog fights which, unless forcibly inter-ruptedby men, would always end in either death orsevere injury to one or both combatants. Thesecases do not include the dogs with special trainingfor fighting in my laboratory.

Let me illustrate my point with another natural-istic observation. For some decades I have beenobserving the behavioral relationship between thecommon magpie and common eagle in SouthChina. The magpie is much less than one-half thesize of the eagle, and its wing span is hardly morethan one-third of the eagle's. Moreover, its beakand claws lack the sharpness of the eagles' hook-likebeak and talons. Nonetheless, every time I hear aneagle cry in pain, I am sure, even without looking upin the sky, that the eagle must be flying low and thata magpie must be making diving attacks on the backof its head or neck. The eagle is absolutely defense-less until it climbs up in the air beyond the reach ofthe magpie. On the other hand, when an eagle isresting on a roof top or the top of a tree, a magpiemight come along, hopping back and forth on itslong legs in front of the eagle, making typical mag-pie noises, and often coming within one foot of theeagle. However, neither bird appears to be con-cerned about the presence of the other. There is nosign of territorial dispute, neither threat nor hostil-ity, nor friendliness. Each simply "ignores" thepresence of the other. In both cases, the behaviorof the two birds appears stereotyped and species-

Five Groups of Determining Factors

Despite the great variability and complexity ofbehavior, for the convenience of investigation anddescription we may classify the determining factorsof behavior into the following five categories:

. Morphological factors

. Biophysical and biochemical factors

. Stimulating objects. Developmental history. Environmental context

Every response of the animal is the functionalproduct of the combined effects of the interwovenreactions of these five factoral groups through in-puts and feedbacks and their organic trace effects(Schneirla, 1965). Our present knowledge concern-ing any of these factors and their interrelations is

64 ZING-YANG KUO

cellular animals or animals without social life, itmay require many years, even decades - especially

when we deal with the species known as Homo sapi-ens with its complicated language capacities and so-cial cultures - before any attempt at a mathematicalformula for the prediction of behavior or even as-pects of behavior would be justified.

While it is fully justifiable for the behavioralscientist to limit his investigations to one level, oreven to certain aspects of one level, of behavior (forexample, various aspects of learning or condition-ing; effects of early experiences on later life; "im-printing"; effects of cortical lesions on learning;so-called critical periods, etc.), we must not forgetthe fact that our knowledge of behavior based on theresults of such investigations is negligible. Any at-tempt to draw conclusions from such studies and touse them as principles of behavior in general or asthe basis for theories of behavior is bound to behampered immediately by a large number of prob-lems that may be insoluble, unless we broaden andreorient the outlook of behavioral studies.

Conclusion

Thus, from the standpoint of the epigenetic behav-iorist, the relationship between the behaving organ-ism and its environment is an extremely complexand variable dynamic process. It goes deeper andbeyond the molar level. As we shall see more clearlyin the following chapters, behavior is far more thanthe visible muscular movements. Besides suchmovements, the morphological aspect, the physi-ological (biophysical and biochemical) changes, thedevelopmental history of the animal, and the ever-changing environmental context are interwovenevents which are essential and integral parts of be-havior. In our study of behavior all such eventsmust be investigated in a coordinated way. Thebehavioral process is not just a stimulus-responserelationship, a conditioning process, nor just a rev-elation of innate actions in the form of "courtship,""threat," food-begging, egg-rolling in the nest, andthe like. In other words, the study of behavior is asynthetic science. It includes comparative anatomy,

comparative embryology, comparative physiology(in the biophysical and biochemical sense), experi-mental morphology, and the qualitative and quanti-t"tiv.. "n"lv!:i!: nf the dvnamic relatinn!:hin between

not adequate to devise any useful mathematical for-mula for the prediction of behavior. The behavioralepigeneticist hopes for the day when he will be able,like the theoretical physicist, to arrive at such aformula even though he may have to wait for somedecades. His hope is based on the fact that, notwith-standing such great variability and complexity,there are to be found some common factors in be-havior such as those due to some common morpho-logical characteristics of the species. For example,morphological structures of the limbs determinethe modes of locomotion; the oral structure deter-mines the modes of eating and drinking, thevocal apparatus determines the characteristics ofvoice and singing. Moreover, there are certaincommon characteristics both in the developmentalhistory (e.g., in prenatal life most animals developunder a more or less standardized environment) andin the environmental context. We can use suchcommon characteristics as the base upon which weshould be able to build our correlations with thebiochemical and biophysical factors as well aswith the varying characteristics of the stimulatingobjects (all of which also have certain constant char-acteristics) in order to arrive at a tentative formula.This would probably be an appropriate and rela-tively fast approach toward the goal of predictionand control of behavior and ultimately towardthe creation of behavioral neo-phenotypes, that is,novel behavior patterns, unknown or unobservedbefore.

I have italicized the words functional product, tostress the fact that behavior is not merely an accu-mulated or arithmetical sum of these five categoriesof determinants. Thus, if we were to designate Behto represent behavior; A to represent biophysicaland biochemical factors; B, morphological factors;C, developmental history; D, stimulating objects;and E, environmental context, this would be theformula for behavior: Beh = (A + B + C + D +

E). But this formula would be much too simple tomake any scientific sense. We must bear in mindthat each category of the determining factors com-prises an enormous number of variables. The inter-woven reactions among these variables themselvesare complex enough. But when those reactions ofthe variables of all five categories are brought to-gether, it may be far beyond the capability of ordi-nary statistical methods to arrive at an adequatefnrmll!" T rn!"~~ w" ti",,! with th" h"h"v;nr nf 11n;-

FROM WATSONIAN BEHAVIORISM TO BEHAVIOR EPIGENETICS 65

Referencesthe organism and the external physical and socialenvironment.

The chief objective of the epigenetic behavioristis to seek order out of such complex behavioralphenomena in order to formulate laws of behaviorwithout resorting to vitalism, either explicitly orimplicitly. He or she has two main tasks: to obtain acomprehensive picture of the behavioral repertoireof the individual and its causal factors from stage tostage during development; and to explore the po-tentials and limitations of new behavior patterns

("behavioral neo-phenotypes") that are not com-monly observed or do not exist in "nature" so as topredict or control the evolution of behavior in thefuture.

This is the main theme of radical behaviorismfrom the epigenetic standpoint. In the light of thistheme, facts derived from experiments and observa-tions are presented and interpreted, and certainmore or less unconventional concepts of behaviorare developed and discussed in the followingchapters.

Gottlieb, G. (1970) Conceptions of prenatal behavior. InL. R. Aronson et al. (eds), Development and Evolution of

Behavior. San Francisco: Freeman.

Loeb, ]. (1889) Der Heliotropismus der Tiere und seineUebereinstimmung mit dern Heliotropismus der Pflanzen.Wiirzburg: Hertz.

Loeb,]. (1918) Tropisms, ForcedMovements, and Animal Con-duct. Philadelphia: Lippincott.

Mayr, E. (1963) Animal Species and Evolution. Cambridge,Mass.: Harvard Univ. Press.

Schneirla, T. C. (1965) Aspects of stimulation and organiza-tion in approach-withdrawal processes underlying verte-

brate behavioral development. In D. S. Lehrman, R. A.

Hinde, and E. Shaw (eds), Advances in the Study of Behav-ior, vol. I, New York: Academic Press.

Watson,]. B. (1913) Psychology as the behaviorist views it.

Psychol. Rev. 20: 158-77.Watson,]. B. (1914) Behavior: An Introduction to Comparative

Psychology. New York: Holt.Watson,]. B. (1919) Psychology from the Standpoint of a Be-

haviorist. Philadelphia: Lippincott.Watson,]. B. (1924) Behaviorism. New Ynrk: Nnrtnn