critical periods in the development of social behavior in

Critical Periods in the Developmentof Social Behavior in Puppies

J. P. SCOTT, Ph.D.

PART OF A PROGRAiM for thestudy of heredity and social behavior in dogs,the author and his colleagues have made athorough study of the development of socialbehavior in puppies, with the idea of findingout the times at which heredity was mostlikely to exert its effects. As we did so, we alsoobserved that there were certain periods inwhich environmental factors were particu-larly likely to affect behavior. As a result wehave, from time to time, reported evidenceon what we have called the critical periodhypothesis.16'18 This idea is one which isbasically related to certain clinical ideas con-cerning the effect of early experience on latermental health and behavioral adjustment. Itis therefore important that it be clearlyunderstood.

Critical Period Hypothesis

What is the critical period hypothesis? Inthe first place it is in certain respects nolonger a hypothesis but a well-established gen-eralization which can be stated as follows: Allhighly social animals which have been so farstudied show early in life a limited period inwhich the group of animals with which theindividual will form positive social relation-ships is determined. To take a few of manyexamples, the slave-making ants raid the nestsof other species for eggs and larvae. As thecaptive ants grow up, they become attached

From the Division of Behavior Studies, R. B.Jackson Memorial Laboratory, Bar Harbor, Maine.

Presented at the Annual Meeting of the AmericanPsychosomatic Society, May 5, 1957, Atlantic City,N. J.

Received for publication May i, 1957.

to their captors and take care of their young,and no longer recognize their own species.The experiments of Lorenz9 with the newlyhatched greylag geese which quickly form asocial bond with the first moving object theysee, whether goose or human, have drama-tized the findings of Heinroth and others thatcontact with the young birds in the properstage of development establishes a strongsocial relationship regardless of the speciesconcerned. Lambs that are taken at birth andraised on the bottle form social relations withpeople rather than other sheep and become asa result quite unsheeplike in many respects.13

The dog is particularly interesting becausethe process of socialization with humanbeings is a normal part of its life as a domesticanimal. Dogs are more closely attached topeople than are many animals, and developa relationship which is in many ways similarto the human parent-child relationship. Fur-thermore, the critical period for socializationin the dog does not begin at birth but approxi-mately three weeks later.13

The existence of critical periods for theprocess of primary socialization can thereforebe taken as established. Other parts of thecritical period hypothesis, namely, that thereexist certain periods of sensitivity to psycho-logical damage, still remain as hypotheses andneed a great deal more experimental evidencebefore they are accepted.

The existence of critical periods of anysort implies certain subsidiary hypotheses.The first of these is that the critical periodhas a physical basis and results from the stateof anatomical, psychological, and physiologi-cal development of the animal. The secondis that hereditary variability between species

PSYCHOSOMATIC MEDICINE

S C O T T

will affect the course of development. Wehave found in a limited survey of differentfoims that such differences are great, andthey do not consist simply of the condensa-tion or elongation of a standard type of de-velopment. The order of developmentalevents may even be reversed in differentspecies. It does look, and this is the thirdhypothesis, as if the social development ofany particular species is strongly correlatedwith the social organization of the adult.14

For example, in dogs there is a close associa-tion between mother and puppies during thefirst 3 weeks in life, but since the permanentsocial relations of the puppy are formed afterthis period and at a time when the motherleaves the litter for long periods, the result isthat the strongest relationships are formedwith the litter mates. This relationship is inturn the basis of pack organization of adultdogs and wolves. A final hypothesis is thatthere should be .genetic variability in thecourse of development within a species, whichmeans that the time of onset of critical periodsand their relative sensitivity should vary fromindividual to individual. This paper will bechiefly concerned with the evidence which wehave been able to gather regarding the physi-cal basis of the critical period for primarysocialization in the puppy.

Normal Development

I shall first describe the normal course ofsocial development in the puppy. The new-born pup is a very immature animal, beingboth blind and deaf and unable to move ex-cept in a slow crawl. Its movements are slowand shaky, and its reflexes sometimes occurseconds after stimulation. In spite of beingdeaf it whines loudly if cold or without food.Its chief needs appear to be warmth, milk,and elimination, and these are taken care ofby reflex behavior. If moved away from themother the puppy will crawl, usually in acircle, throwing the head from side to side.If it comes into contact with the mother itattempts to nurse. When it succeeds it pusheswith its head and forepaws and also with thehind feet. In doing this it touches and so

VOL. xx, NO. i, 1958

43

stimulates the other puppies in the litter. Anytouch will initiate exploratory movements.The mother also frequently stimulates thepuppies by nosing them and licking the geni-tal and anal areas. This produces reflexelimination as well as stimulating the pups toexploratory movement and nursing.

This is the typical behavior of puppies inthe neonatal -period. No immediate change isobserved in their behavior. They graduallygrow somewhat stronger and quicker untilthe opening of the eyes, which typicallyoccurs just before 2 weeks of age. This marksthe beginning of the transition "period, one ofvery rapid change and development. At itsbeginning the eyes open, and at its end thepuppy gives a startled reaction to sound. Itcan walk instead of crawl, and can movebackward as well as forward. Its first teethappear, it takes solid food, and it begins tourinate and defecate without assistance bythe end of the period, which typically occursjust before 3 weeks of age.

At r.pproximately 3 weeks of age the puppyfirst begins to notice other individuals at adistance and shows evidence of conditioningand habit formation. This is the beginning ofthe period of socialization, in which its pri-mary social relationships are formed. At al-most the same time the mother begins to leavethe puppies unattended, so that there is atendency for the strongest relationships to bedeveloped with the litter mates rather thanthe mother. During the next few weeks it iseasy for the human observer to form a positivesocial relationship with the puppy. Thisperiod of primary socialization comes to anend somewhere around 7 to 10 weeks of age,which is the normal period of weaning. Thisdoes not mean that the puppies are self-sufficient with regard to food. Studies ofwolves show that the pups are not able tohunt at all until 4 months of age, and are notreally independent until 6 months of age.10

In the domestic dog, food is normally sup-plied by man, but in the wild state it is pro-vided by the hunting parents. This period,which ends with the sexual maturity of thepuppy, is called the juvenile period. In someof our domestic breeds estrus of the females

44 S O C I A L B E H A V I O R I N P U P P I E S

occurs as early as 5 months, but in the ances-tral wolves this does not happen until the endof the second year.

We can see that there might he severalcritical periods affecting the development ofnew social relations, that of primary socializa-tion at 3 weeks, the time when sexual rela-tions are established at the end of the juvenileperiod, and that when the parental relation-ships are established as a new generation ofpups is born. We are here concerned pri-marily with the first of these periods, theperiod of primary socialization, because wecan make the assumption that the end of theprimary social relationships will strongly af-fect the degree of adaptation and adjustmentin later relationships.

Anatomical Basis for Social Behavior

Function is of course related to form, andform in turn to the process of growth. Let usexamine the anatomical changes which ac-company changes in social behavior. In theyoung pups there are three sorts of externalchanges which are easy to follow, the erup-tion of the teeth, the opening of the eyes, andthe opening of the external ear canal. Thelast is somewhat harder to follow and is prob-ably less reliable than the other two.

All the following data are based on observa-tions of purebred puppies reared under thestandard conditions of our long-term experi-ment on genetics and social behavior.17 Thetotal numbers differ slightly from table totable because complete information was notgathered on all animals in the early part ofthe experiment.

The first teeth to appear are the canines,followed shortly by the incisors, which allcome in together except that the correspond-ing upper teeth come in before the lower. Noanimals show teeth at 2 weeks, nearly halfhave some teeth by 3 weeks, and nearly allhave at least the upper canines through thegums by 4 weeks of age. An eruption of theteeth therefore coincides with the beginningof the period of socialization.

The teeth erupt the earliest in basenjis andbeagles (Table 1), and wirehaired terriers

TABLE 1. ERUPTION OP UPPER CANINE TEETH

Breed

BasenjiBeagleCocker spanielSheltieWirehaired fox

terrierTOTAL

No.

51

546730

31233

% Erupted at2 weeks

0000

00

3 weeks

79742230

1447

4 weeks

100100100100

8999

are definitely slower than all the rest. Theshelties are unusual in that the lower teethdevelop relatively fast and, in certain animals,the lower canines actually appear before theupper.

The eyes typically open about 2 weeks ofage or slightly before. Very few are partlyopen at 1 week, the majority are open at 2weeks, and all animals have the eyes com-pletely open by 3 weeks of age (Table 2).The opening of the eye is the first externalsign of the transition period. Eighty-five percent of the purebred puppies have the eyespartly open at 2 weeks, and we can estimatethe average time at 13 days. There is, how-ever, individual variability and differencesbetween breeds. The eyes open earliest incocker spaniels and beagles, a little later inbasenjis, and much slower in shelties andwirehaired terriers.

The opening of the ears is a more pro-longed process and comes a little later thanthat of the eyes. Over half the animals showthe ears at least partly open at 2 weeks, andall are completely open at 4 weeks. The dif-ferences between strains are not great. Theshelties appear to be the most rapid, and the

TABLE 2. OPENING OP THE EYB

% Completely open atBreed No. 1 week 2 weeks 3 weeks

Basenji 43 0 65 100Beagle 49 0 94 100Cocker spaniel 51 2 94 100Sheltie 25 0 31 100Wirehaired fox

terrier 27 0 11 100TOTAL 195 0.5 67« 100

a Average with equal weight for breeds = 59.


SCOTT

TABLE 3.

Breed


terrierTOTAL

PUPILLARY REFLEX (EYE OPEN ORPARTLY OPEN)

No.

42455126

24188"

%1

week

0000

00

Animals2

weeks

57788838

1362»

respondi3

weeks

9898

10096

8396

ing at4

weeks

100100100100

100100

TABLE 4.

45

FUNCTION OF THE EARAnimals giving startle responses to some sort

Breed


terrierTOTAL

No.

43495724

27200

UI suunu

2 weeks

0020

41

Age3 weeks

72846162

9274

4 weeks

100100100100

100100

a Not observed m 8 additional animals: 1 basenji,4 beagles, 3 terriers.

* Average weighted equally for breeds = 59%.

wirehaiied terriers the slowest, with the otherbreeds being intermediate and showing onlyslight differences.

The variability of any of these eventswhich can be precisely timed, such as theeruption of the teeth or opening of the eyes,has a range of approximately i week in nor-mal animals. The wirehaired fox terriersappear to be definitely slower in all respects,but in the other breeds the speed of development is not correlated in the different traits.It looks as if there can be separate variabilityin any one of these characters. One might ex-pect that the development of all the teethwould be correlated, but even here theshelties show a relatively more rapid develop-ment of teeth in the lower jaw than do theother strains.

Function of the Sense Organs

Histological studies of the development ofthe puppy eye1 indicate that the retina is notfully developed at the time of the opening ofthe eyes, nor even by 3 weeks of age (Table3). However, it can be shown that the eyeresponds to light at a much earlier age. Somepuppies will give a winking reflex to light atbirth, long before the eyes open. This appearsto happen most frequently in those breedswhich have red hair color and light skinpigment.

As soon as the eyes open and the pupil canbe clearly seen we can demonstrate a pupil-lary reflex when a strong light is shown into

VOL. xx, NO. 1, 1958

the eye. At the same time the nystagmus re-flex can also be observed. If the puppy's headis moved slowly sideways the eyes roll or flickback and forth. This, however, is probably areflex controlled by the nonauditory portionof the acoustic nerve rather than by sight.Incidentally, puppies do not shown nystag-mus in reaction to a moving object or rotatingcylinder held in front of them.

The above reflexes concerned with eyefunction appear early in development, but itis probable that these are responses to lightand darkness, and that the capacity to per-ceive images is not developed until 4 or 5weeks of age.

The onset of function in the ears is muchmore definite (Table 4). Only 1 per cent ofthe puppies give a startle reaction to sound by2 weeks of age, and 74 per cent give a reactionat 3 weeks of age. We may therefore estimatethat the average time is about 19.5 days.There is considerable genetic variability be-tween the breeds, and the wirehaired terriershave the highest percentage of animals whichgive a definite startle reaction.

As with the eyes, there is no evidence thatthe puppies use their ears in finer ways atfirst. The tendency to startle to all sorts ofloud sounds persists for a week or two.

All the puppies gave some reaction to odorsat birth. They give reliable avoidance reflexesto two substances, oil of anise and a proprie-tary drug used as a dog repellant, which is acompound related to citronella. Eighty-threeper cent of all puppies gave the avoidanceresponse to the repellant, and a smaller num-ber to oil of anise. Harman found that the

46 SOCIAL BEHAVIOR IN PUPPIESTABLE 5. APPEARANCE OF THE "WINK REFLEX"

TO TOUCH

Breed


terrierTOTAL

No.

43495126

27196

% Responding byBirth

86967469

6380

1 week

1009898

100

10099

2 weeks

100100

100

parts of the brain connected with olfactionwere unmyelinated at birth, and it is probablethat these reactions to odor are largely con-nected with the sense of taste rather thantrue olfaction. In the human subject bothsubstances can be detected in both the noseand throat. There is some observational evi-dence that hungry puppies react to the smellof milk in the neonatal stage, but we have nodefinite evidence on this point.

With regard to other senses, puppies defi-nitely react to pain and touch at birth. Themajority of all newborn animals give a "wink-ing reflex" to touch and nearly all of themshow this by i week of age (Table 5).

In genera], the function of the sense organs

at birth is quite limited, but all senses are atleast partially functional at approximately 3weeks. Prior to the beginning of the period ofsocialization it is impossible for the animal tobe stimulated by many environmentalchanges. Those stimuli which aie effectiveare ones which set off the reflexes connectedwith eating and other vital piocesses. Ineffect, the very young puppy is insulated frommany sorts of environmental stimulation.

Development of the CentralNervous System

Observation of puppies in the neonatal andtransition stages gives no indication that theanimals learn in the way that adults do. Forexample, a puppy placed on the scales maycrawl to the edge and fall off. When he is putin the situation repeatedly he does the samething time and again, with no improvementin adjustment.

The only change in behavior for some timeafter birth is that the puppies become some-what faster and stronger in their simple be-havioral reactions. Fuller et al.B found thatthe avoidance behavior of puppies could be

Beo»- - J.

k - i -4-n-J«- m.SOCIALIZATIONEYES OPEN^ CONDITIONING

/ BEGINS

* K — H , JUVENILE-

ADULT EEGs- MOTHER OUT

I I13 14 15 16

Fig. 1. Average heart rates at different periods in development. Note that all breeds show alower rate during the early part of the period of socialization than they do either earlier or laterThe heart rate change probably measures the puppy's emotional reaction to being picked up, andindicates unusual emotional sensitivity at this age.


S C O T T

easily conditioned to sounds shortly before3 weeks of age. This could not be done pre-viously, and the change in an individual ani-mal's behavior occurred from one day to thenext. This change is, of course, related to thedevelopment of the function of hearing, butno conditioning to other sensory stimuli couldbe obtained at earlier ages. Responses weregiven to taste and touch, but did not produceconditioning. James and Cannon7 confirmedFuller's results and found that the avoidancereaction to a mild electric shock is restrictedto the part stimulated by 28 days of age, indi-cating that psychological development is stillgoing on.

Harmanu has made histological examina-tion of the brains of young puppies and findsthat the newborn brain is myelinated in veryfew places, those corresponding to the partsassociated with the observed reflex patterns ofbehavior. This lack of myehnation may ac-count for the slow movements of the youngpuppies, as contrasted with their more rapidicsponses at later ages.

A measure of actual function of the devel-oping nervous system was obtained byCharles and Fuller8 by taking the EEG ofpuppy brains at different ages. At birth thepuppy brain has almost no waves at all, andthere is no differentiation between sleepingand waking states. At 3 weeks, correspondingto the other differences in behavior notedabove, the sleeping and waking states aredifferentiated, and the amplitude of waves isincreased. The adult form of the EEG isachieved at approximately 8 weeks of age,which is shortly after the earliest time atwhich mothers normally wean their pups.

Another measure of nervous developmentis the heart rate (Fig. 1). At first glance thismight be considered a purely physiologicalresponse. However, the heart is actually avery sensitive indicator of both body activityand various kinds of emotions. As will be seenon the graph, the heart rate of the newbornpuppies is very high and stays this waythrough the second week. Then it takes avery decided dip at from 3 to 6 weeks, com-ing back up to the early level around 7 weeksof age Thereafter the heart rate slowly de-

VOL. xx, NO. 1, 1958

47

clines toward the adult level. These generalchanges seem to be independent of breed.The first change occurs at the beginning ofsocialization, and its end occurs at 7 weeks,coinciding approximately with the time of theadult EEG. We can suppose that this is theperiod when complete cortical connections areestablished with the hypothalamus. We canconclude that the period from 3 to 7 weeks isan especially sensitive one for emotional re-actions, which corresponds to observation ofovert behavior. We might also speculate that,since the cortex is not completely developed,emotional reactions during this time might beless permanently learned. On the other hand,they might be more disturbing because com-plete cortical control has not been established.We have here a fascinating field for furtherprecise experiments on the effects of earlyexperience.

The development of the nervous systemand its associated psychological abilities goesthrough several stages. The puppy definitelydoes not come into the world with all its psy-chological abilities developed. They are, infact, quite immature at birth and for the next3 weeks. We would expect and have found10

that it would be extremely difficult to pro-duce psychological trauma upon very youngpuppies, and that any future effects on theirbehavior produced at this time would have tobe made by physiological or anatomicalinjury.

These results raise the question of what thesituation is in the human infant. Human de-velopment is obviously different from that ofthe puppy, but we have every indication thatthe neonatal human infant has a nervous sys-tem which is decidedly undeveloped. Weneed to know more about the origin of learn-ing abilities in human infants before we cantalk authoritatively about the effects of earlyexperience. It is not too much to suggest thatit would be contrary to the general law ofbiological adaptation to find that the nervoussystem was highly sensitive to psychologicaldamage at such a period as birth. This, how-ever, does not invalidate the possibility ofanatomical birth injuries.

SOCIAL BEHAVIOR IN PUPPIES

Social Behavior

The most obvious social behavior of youngpuppies is vocalization. This response is obvi-ously related to various sorts of social contacts.The type of behavior involved is et-epimeletic,or calling for care and attention. In any situa-tion in which the puppy is unable to adapt,the puppy substitutes this reaction for anyattempt at adjustment on its own part. New-bom puppies whine repeatedly until they be-gin to nurse. Fredericson4 has shown that,besides hunger, the sensation of cold is thestimulus most likely to produce whining. Inaddition, puppies will whine loudly if acci-dentally hurt. All these things can be allevi-ated by social contact.

During the early stages of development thenumber of whines made by the puppies whilebeing weighed for a period of i minute wascounted. In general the response decreases,so that by 4 weeks of age most of the puppiesmake no noise at all. The response is chieflydue to contact with the cold scales, and there

is evidence of breed differences, the beaglesreacting with relatively small numbers ofnoises.

During the period of socialization the pup-pies begin to whine in response to beingplaced in a strange environment no matterwhether warm or cold. Fredericson4 hasshown that this response is considerably les-sened if another puppy from the same litteris placed with it. We are now gathering dataon the development of this response, and itappears that isolation in a strange environmentalways produces a stronger reaction than isola-tion in the home pen.

This brings up another experimental prob-lem which has been scarcely touched. Thereis considerable evidence21 that there is a proc-ess of primary "localization," in which ayoung animal becomes psychologically at-tached to a particular physical environment.There may be a critical period in developmentfor this as well as for the analogous processof socialization.

8 0 -

60-

4 0 -

2 0 -

-1-»Un-4«— m , SOCIALIZATION I Z , JUVENILE

• 1ST SOLID FOOD

FINAL WEANING

MOTHER OUT

10 II 12 13 14 15 16

Fig. 2. Weekly percentage gain of weight in male cocker spaniel and basenji pups. Note thechange in the slope of the curve at 3 weeks. This reflects both the taking of solid food andincreased food motivation resulting from learning and habit formation. The difference betweenthe two breeds reflects a genetic difference in palatability of the food supplied.


SCOTT

Developmental Changes inEating Behavior

General changes from sucking to eatingsolid food have already been described. Ourmost objective measurement of eating be-havior is an indirect one. The gain in weightof the animal reflects the amount eaten. Ifwe chart the weekly gains of puppies we find,as we might expect, that the rate of gain de-clines week by week. However, there is adefinite change in the nature of this curve at3 weeks of age. The curve before this time isentirely dependent upon the milk supply ofthe mother. Afterwards the puppy has thepossibility of eating solid food, and we wouldexpect that the decline in the growth wouldbe less rapid as soon as solid food was avail-able. However, there is probably a psychologi-cal factor involved also. Puppies pay no atten-tion to solid food if they get plenty of milkfrom the mother, even beyond the age of 3weeks. This was definitely the case in our F2hybrids which were fed by Fx mothers havingan abundance of milk, even in excess of whatthe pups could use. In their case also the rateof decline of growth was halted after 3 weeks,indicating that the animals were taking inmore food (Fig. 2). We may conclude thatbecause of their ability to be conditioned theanimals now learn to eat, increasing theirfood motivation.

This idea is supported by observations onhand-fed puppies. During the neonatal stagethey are quite difficult to feed by ordinarymeans. It is hard to get them to take a nursingbottle, and they have a tendency to stop be-fore the stomach is filled. They can be labori-ously fed by a dropper, but the easiest andmost practical way to insure adequate feedingis to inject milk into the stomach through atube. The explanation is that natural feedingis stimulated by specific primary stimuli orreleascrs which activate certain reflexes andsimple patterns of behavior. Puppies becomemore active and more responsive to externalstimuli if hungry, but hunger has little rela-tion to the amount of food taken.

During the transition period, between 2and 3 weeks, the pups are much more easy

VOL. XX, NO. I, I958

49

to feed. They will readily take a bottle and,if placed near a dish, they can clumsily lapthe food. However, they will not take ade-quate amounts if the dish of warm milk issimply left with them. They still have to bestimulated by fresh food and handling severaltimes a day in order to be properly nourished.

By contrast, puppies older than 3 weeks ofage are relatively easy to feed, particularly be-cause a supply of food can be left with themand they will continue to eat adequateamounts.

We can draw several conclusions fromthese observations. One is that puppies can-not be satisfactorily fed on a demand schedulebefore 3 weeks of age. Maternal care, togetherwith the reflex behavior of the puppies, willprovide adequate nutrition, but the puppyhas not yet become a self-regulating organismwith regard to eating. He needs externalstimulation. Later on, when the ability toform associations and habits is developed, thehunger mechanisms, together with motivationto eat produced by learning, are sufficient toregulate the eating of the puppy in a satisfac-tory way. This, of course, poses certain ques-tions regarding the reasons back of thenecessity for handling human infants, a prob-lem which has been raised by Spitz,20

Ribble,11 and others.

Feeding and Socialization

The simplest theory of primary socializa-tion is that this relationship is built upthrough the association of food with a particu-lar individual, and that the relationship candevelop further based on this original bond.Brodbeck2 tested this by rearing cocker spanielpuppies which were mechanically fed, andcomparing their behavior with that of puppiesfed by hand. In both cases the experimenterspent an equal amount of time playing withthe pups. The result was that the pups whichhad never been fed by people still showedstrong attraction to the experimenter. Thisindicates that feeding is not the only elementin the formation of a social bond.

On various occasions we have observedpuppies which through some accident to the

50 SOCIAL BEHAVIOR IN PUPPIESmother or deficiency in her milk supply wereunderfed. Young pups that are well-fed arefat and lethargic. In contrast, the underfedanimals are active, show more interest inpeople, and seem to show an earlier expres-sion of many behavior traits than do thenormal ones. An opportunity to check thisfinding occurred recently in connection withan experiment of John A. King, in which hefed two groups of basenji puppies by handtwice a day during the period of socialization.One group was given as much as they couldeat, and food was left in the dish so thatthey could eat later. The other group wasgiven considerably less and their weight wasmuch lower. Both groups were given a han-dling test which measures responses of thepuppy to a human handler.

In puppies raised in a different situation inwhich most of the food comes from themother but in which there is considerablehandling from 5 weeks on, there is a big de-cline of the timidity score between 5 and 7weeks, which may be taken as a measure ofthe process of socialization. In the experi-mental pups, whose contacts with people werestarted at 4 weeks and confined to feeding,the timidity score at 5 weeks averaged verynearly as low as the other animals at 7 weeks.The average score stayed slightly higher insubsequent weeks than the controls, but notsignificantly so. At an early age, and using agenetic type of animal susceptible to the de-velopment of timidity, animals which aretotally dependent for food show a greater de-gree of socialization than those only partiallydependent.

Another obvious result was that there was adifferentiation between the reaction to thepeison who did the feeding and another per-son who did not. This difference appearedmore strongly as the pups grew older. All thisindicates that the process of feeding does con-tribute importantly to social relationships butdoes not constitute the whole process.

The difference between the hungry andnonhungry animals was not as great as hadbeen anticipated. There was a great deal ofindividual variability in behavior. However,hungry animals did show a consistently

greater proportion of investigation and foodbegging in their responses than did thecontrols.

Sucking Frustration

Levy's experiment on puppiess indicatedthat the canine equivalent of thumb suckingcould be produced by sucking frustration.Levy's experiment extended over a relativelylong period in the development of the puppy,including the tiansition period and the earlypart of the period of socialization. Because ofthe importance of sucking in early develop-ment and its disappearance later, there shouldbe a limited period in which this effect couldbe produced.

Ross, Fisher, and King12 expeiimentedwith this possibility using the technique ofsudden weaning, after which the pups couldobtain food only by lapping or eating solidfood. The experiment was done on a group of34 experimental and 24 control animals.None of the experimental puppies developedspontaneous body sucking, but they did showan increased tendency to suck a finger of theexperimenter. When this was analyzed in re-lation to age of the puppy, it was found thatall the puppies weaned between the ages of10 and 18 days showed more fingei suckingthan the controls. After this date many experi-mental animals were as low as the controls,although there were some indications ofsmaller effects. There were no effects at all bythe time the animals were 5 weeks of age.

We can draw two conclusions. One is thatthere is an obvious change in the puppies'behavior with regard to sucking at the begin-ning of the period of socialization. Up to thistime a puppy deprived of its mother willreadily suck objects which are put into itsmouth. After this point it shows very littletendency to suck. The simplest explanationis that the power to discriminate between ob-jects which give nourishment and thosewhich do not has been developed.

The other conclusion is that the simple andsudden deprivation of the opportunity tonurse does not by itself produce the neuroticform of body sucking. Whether there is a


SCOTT

TABLE 6. ESTIMATED DEVELOPMENTAL AGES INDAYS FOR SIGNIFICANT EVENTS

Mean and Standard DeviationTeeth

Eyes Ears Eruption,completely startle upper

Breed open to sound canines

Basenji 13.1 19.6 18 6Beagle 11.0 18.7 19.1Cocker spaniel 11.Oil.9 20.2±2.9 23.3Sheltie 15 2 20.3 22.6Wirehaired fox

terrier 16.9 17.8+2.2 24.3+3.0TOTAL 13.0+2.3 19.5+2.3 20.8+2.9

critical peiiod for this latter phenomenon isyet to be determined.

Hereditary Variability

As can be seen in Table 6, there is con-siderable evidence that hereditary variabilityin development exists, both between indi-viduals and between breeds. For example,cocker spaniels are the first to open their eyes,are slow to develop the startle response tosound, and are intermediate between theother breeds with regard to the developmentof teeth. The wirehaired terriers were theslowest to develop in every trait observed ex-cept the startle reaction to sound, in whichthey are the fastest. There is every indicationthat the different sense organs and anatomicalcharacteristics vary in the speed of develop-ment independently of each other. There isno one highly correlated pattern of develop-ment which is slowed down or speeded upas a unit.

Assuming that the age of onset of each de-velopmental event falls into a normal curve,we can estimate the average time and vari-ability of each event. As seen by Table 6,there is approximately 3 days' difference be-tween the fastest and slowest breeds. Eighty-six per cent of all animals should fall withina range of 1 week. If anything, these estimatesof variability are probably too great, sincethey are based on the 1 per cent or so of ani-mals which are as much as a week away fromthe normal.

The species as a whole shows a develop

VOL. XX, NO. I , I958

51

mental pattern which is not departed fromexcept in cases of gross abnormality. Eventsoccur in the general order described in theearly part of this paper, so that we can speakof definite periods of development. We canset the beginning of the transition period at

OO XT

13 days with a normal range of 3 or 4 dayson either side. Similarly, we can place thebeginning of the period of socialization atapproximately 19.5 days with a similar rangeof variability. This means that when experi-mental procedures are employed in whichtime is the experimental variable, close atten-tion must be paid to the breed and the stateof individual development. Two litters ofpuppies of exactly the same chronological agecould give completely different results.

Conclusions

At the beginning of this paper 1 statedthat there is a definite critical period for theprocess of socialization, and that it must reston a physical basis. The data which I havejust presented show that there is a definitephysical basis in the dog, and that there aretwo important points where sudden changesoccur. One of these is the point just before 3weeks, where there are changes in the abilityto be conditioned, in the EEG, in the emo-tional responses indicated by the heart rate,in the ability to hear, and in the growth rateand method of nutrition. The other point at7 to 8 weeks is so far defined only by the adultEEG and the change in emotional reactionmeasured by the heart rate. Final weaningsometimes occurs as early as this date.

To summarize, the puppy before 3 weeksof age is highly insulated from its environ-ment by the immature state of developmentof the sense organs, by the lack of ability forconditioning, and by maternal care. From 3 to7 weeks the puppy is in an extremely inter-esting stage in which its sense organs andcerebral cortex are not yet completely devel-oped but in which it has extremely sensitiveemotional reactions and is capable of makingassociations. This is the time when primarysocialization normally takes place and duringwhich it is easiest for a dog owner to establish

52 SOCIAL BEHAVIOR IN PUPPIESa strong social bond. These facts provide uswith an experimental opportunity to analyzesome of the theories of the effects of earlyexperience on later social adjustment andmental health.

Comparing these with results on humanbabies, we see that we need more funda-mental facts about human development. Weknow a great deal about babies from birthuntil 10 days of age. After this time mothersand babies leave the hospital and disappearinto the home where there is little oppor-tunity for scientific study. Facts and informa-tion are very scarce until about the age of 2years when the babies begin to emerge fromthe home and appear in nursery school. It isprecisely this period in which we are mostlikely to find the period corresponding to theprimary period of socialization in the dog, ifit exists. As I mentioned above, one of themost important basic facts yet to be ade-quately described is the development of sim-ple learning ability in young infants.

One of the biggest problems in predictinghuman development is the element of indi-vidual variability in behavior. The study ofdevelopment in the different dog breeds,where we should get the maximum possiblevariability in development, gives us some hintas to what we might expect in human beings.There is considerable individual variability,and variability between dog strains. However,the timing does not vary a great deal in termsof the length of the period of development.The changes at 3 weeks of age in puppiesappear to take place within a week for allanimals. If we assume that the life span of adog is one sixth that of a human, we mightexpect a human range of variability of 6weeks. Actually, the possible range in earlydevelopment is probably considerably lessthan this.

The existence of a critical period for pri-mary socialization is so widespread in the ani-mal kingdom that there is every reason forsuspecting that a similar period exists inhuman development. If so, we are faced witha number of questions:

1. When? In other animals, the criticalperiod may occur immediately after birth or

hatching, or fairly late in development, as itdoes in the puppy. The human infant is some-what more mature than a puppy at birth, butfar less so than a lamb. We would expect thatthe critical period would not begin immedi-ately after birth, and the evidence that wehave indicates that it may begin as early as amonth or 6 weeks, or as late as 5 or six months.However, any social relationship depends onthe behavior of two individuals, and the pe-riod immediately after birth may be a criticalone for the mother, although the evidencefrom adoptions indicates that mothers canform a strong social relationship at other times.

2. How long? The period of primary social-ization certainly lasts as long as the peiiod ofcomplete dependency on the mother, up to\Yi or 2 years, and possibly even longer.

3. By what means? The positive behavioralmechanisms which establish primary socialrelationships, such as feeding, contact, han-dling, and the like, are difficult to understandwithout experimental data. Likewise, wewould expect to find negative mechanismswhich prevent the socialization of infants tostrangers, such as the anxiety or fear reactionto strangers described by Spitz.19 Their under-standing would be immensely important as ameans for bringing about a better identifica-tion with other human beings, and a broaderrange of tolerance.

4. Why? This brings us down to the pri-mary data in this paper, and raises the ques-tion of the physical basis of critical periods inthe human infant. As I have indicated above,animal experiments indicate the type of thingwe should look for: the development of sen-sory and motor organs, the development ofthe central nervous system, and the powerof simple learning, the development of socialbehavior patterns, and the development ofindividual hereditary differences. We havemuch information on the first two, and farless on the rest. It is this kind of evidencewhich is the easiest to get on human subjects,and without this firm foundation we shallnever be on safe ground.

Seen in this light, our study on the dogprovides the opportunity for the experimentalanalysis of environmental and hereditary in-


SCOTT

fluence on an undeveloped nervous system.The dog is an animal which is capable offorming a type of social relationship withpeople very similar to the human parent-childrelationship. We should be able to find outwhether early emotional experiences producelasting effects. Without evidence, it is just aslogical to suppose that an immature nervoussystem would be less severely affected. It willtake a long time and will be a laborious andexpensive job to obtain the needed facts, butwe should eventually be able to bring thephenomena of early social experience out ofthe realm of conjecture into that of estab-lished scientific fact.

Summary

1. The existence of a critical period for theestablishment of primary social relationshipsis a well-established phenomenon in socialanimals. This paper has dealt with the physi-cal and hereditary basis of the critical periodin the dog.

2. Normal social development in thepuppy can be divided into several periodsbased on changes in social relationships. Sev-eral of these may be critical, but the most im-portant is that of primary socialization, begin-ning about 3 weeks of age.

3. The beginning of this period is accom-panied by certain anatomical changes: theeruption of the teeth and the opening of theears. Eyes open at an earlier date.

4. Its beginning is also closely associatedwith the appearance of the function of theears. The senses of touch and taste (includingtasting gases) are present at birth. The eye issensitive to light before it opens, but not com-pletely functional until some time after.

5. Changes in the heart rate appear whichcorrespond to changes in the EEG and abilityto be conditioned.

6. At 3 weeks there is a change in the rateof growth, which may be attributed to bothadditional food and to a psychological changein food motivation.

7. Experimental evidence indicates thatfeeding is an important factor in socialization,but not the only factor.

VOL. xx, NO. 1, 1958

53

8. There is a marked change in the finger-sucking response of weaned puppies at anage corresponding to the beginning of thecritical period.

9. Hereditary variability of the exact timeof onset of the critical period exists both be-tween individuals and between breeds. How-ever, the functional variability appears to besmaller than that in the accompanying ana-tomical changes.

10. In the period immediately followingbirth the puppy is strongly protected frompsychological influences. During the criticalperiod it becomes highly sensitive, at a timewhen the sense organs and nervous systemare still not completely developed. The exacteffects of experience during this time are stillto be determined.

11. These data suggest facts which must beascertained in order to establish the existenceand duration of a similar critical period inhuman infants. They also suggest ways inwhich important clinical theories can be ex-perimentally tested.

Hamilton StationP.O. Box S47

Bar Harbor, Maine.

References

1. BLUME, D. A study of the eye of the dog.In preparation.

2. BRODBECK, A. J. An exploratory study on theacquisition of dependency behavior in puppies.Bull, Ecol. Soc. Am. 35:73, 1954.

3. CHARLES, M. S., and FULLER, J. L. Develop-mental study of the electroencephalogram of thedog. Electroencephalog. & Clin. Neur&physiol.8:645, 1956.

4. FREDERICSON, E. Perceptual homeostasis anddistress vocalization in puppies. /. Person. 20 •472, 1952.

5. FULLER, J. L., EASLER, C. A., and BANKS, E. M.Formation of conditioned avoidance responses inyoung puppies. Am. ]. Physiol. 3:462, 1950.

6. HARMAN, P. J. Private communication.7. JAMES, W. T., and CANNON, D. J. Condi-

tioned avoiding responses in puppies. Am. ].Physiol. 168:251, 1952.

8. LEVY, D. M. Experiments on the sucking re-flex and social behavior of dogs. Am. J. Ortho-psychiat. 4:203, 1934.

54 SOCIAL BEHAVIOR IN PUPPIES

9. LOHENZ, K. Der Kumpan in der Umvelt desVogels. /. Orntthol. 83:137, 1935.

10. MUME, A. The Wolves of Mt. McKinley.Washington, D.C., G.P.O., 1944.

11. RIBBLE, M. A. Infantile experience in rela-tion to personality development. In HUNT,J. M. (Ed.): Personality and the behavior dis-orders. New York, Ronald Press, 1944.

12. Ross, S., FISHER, A., and KING, D. The ef-fect of early enforced weaning on sucking be-1 . r • T ^ °

havior ot puppies. In preparation.13 SCOTT, J. P. Social behavior, organization and

leadership in a small flock of domestic sheep.Comp. Psychol. Monogr. 18:1, 1945.

14. SCOTT, J. P. The relationships between de-velopmental change and social organizationamong mammals. Anat. Rec. 111.73, 1951.

15. SCOTT, J. P. "The Process of Socialization inHigher Animals." In Interrelations Between the

Social Environment and Psychiatric Disorders.New York, Milbank Mem. Fund, 1953.

16. SCOTT, J. P., FREDEBICSON, E., and FULLER,J. L. Experimental exploration of the criticalperiod hypothesis. /. Person. 1:162, 1951.

17. SCOTT, J. P., and FULLER, J. L. Manual ofDog Testing Techniques. Bar Harbor, Maine,Jackson Laboratory, 1950.

18. SCOTT, J. P., and MARSTON, M. V. Criticalperiods affecting the development of normal andmaladjustive social behavior of puppies. J.Genet. Psychol. 77:25, 1950.

19. SPITZ, R. A. Anxiety in infancy: A study ofits manifestations in the first year of life. In-ternat. ]. Psycho. Analysis 21:1, 1950.

20. SPITZ, R. The psychogenic diseases in infancyPsychoanalyt. Study Child 6:255, 1951.

21. THORPE, W. H. Learning and Instinct in Ani-mals. London, Methuen, 1956.

Public Health FellowshipsTwo Public Health fellowships are available at the Peter Bent Brigham Hospital

to individuals who have had two years of training in a psychiatric hospital. The posi-tion offers an opportunity in a general hospital setting to a person interested in seniorsupervision by psychoanalysts especially interested in problems having to do withpsychophysiological disturbances. The setting and the supervision provide experiencein developing good working relationships with internists and surgeons in the care oftheir patients. The American Board of Psychiatry and Neurology gives credit to thisprogram for one year of training.

Applicants for the year July 1, 1958, to July 1, 1959, should send curriculum vitaeto Dr. Henry M. Fox, Associate Clinical Professor of Psychiatry, who is in charge ofthe program. The address is. Peter Bent Brigham Hospital, 721 Huntington Ave.,Boston 15, Mass.


critical periods in the development of social behavior in

Documents