Psychological ReviewVol. 65, No. 4, 1958

HEREDITY, ENVIRONMENT, AND THE QUESTION "HOW?"1

ANNE ANASTASI

Fordham University

Two or three decades ago, the so-called heredity-environment questionwas the center of lively controversy.Today, on the other hand, many psy-chologists look upon it as a dead issue.It is now generally conceded that bothhereditary and environmental factors en-ter into all behavior. The reacting or-ganism is a product of its genes and itspast environment, while present envi-ronment provides the immediate stimu-lus for current behavior. To be sure, itcan be argued that, although a giventrait may result from the combined in-fluence of hereditary and environmentalfactors, a specific difference in this traitbetween individuals or between groupsmay be traceable to either hereditary orenvironmental factors alone. The de-sign of most traditional investigationsundertaken to identify such factors,however, has been such as to yield in-conclusive answers. The same set ofdata has frequently led to opposite con-clusions in the hands of psychologistswith different orientations.

Nor have efforts to determine theproportional contribution of hereditaryand environmental factors to observedindividual differences in given traits metwith any greater success. Apart fromdifficulties in controlling conditions, suchinvestigations have usually been basedupon the implicit assumption that he-reditary and environmental factors com-bine in an additive fashion. Both ge-neticists and psychologists have repeat-edly demonstrated, however, that a moretenable hypothesis is that of interaction(15, 22, 28, 40). In other words, the

1 Address of the President, Division of Gen-eral Psychology, American Psychological As-sociation, September 4, 1957.

nature and extent of the influence ofeach type of factor depend upon thecontribution of the other. Thus theproportional contribution of heredity tothe variance of a given trait, ratherthan being a constant, will vary un-der different environmental conditions.Similarly, under different hereditaryconditions, the relative contribution ofenvironment will differ. Studies de-signed to estimate the proportional con-tribution of heredity and environment,however, have rarely included measuresof such interaction. The only possibleconclusion from such research wouldthus seem to be that both heredity andenvironment contribute to all behaviortraits and that the extent of their re-spective contributions cannot be speci-fied for any trait. Small wonder thatsome psychologists regard the heredity-environment question as unworthy offurther consideration!

But is this really all we can find outabout the operation of heredity and en-vironment in the etiology of behavior?Perhaps we have simply been asking thewrong questions. The traditional ques-tions about heredity and environmentmay be intrinsically unanswerable. Psy-chologists began by asking which typeof factor, hereditary or environmental,is responsible for individual differencesin a given trait. Later, they tried todiscover how much of the variance wasattributable to heredity and how muchto environment. It is the primary con-tention of this paper that a more fruit-ful approach is to be found in the ques-tion "How?" There is still much to belearned about the specific modus oper-andi of hereditary and environmentalfactors in the development of behavioral

197

198 ANNE ANASTASI

differences. And there are several cur-rent lines of research which offer promis-ing techniques for answering the ques-tion "How?"

VARIETY OF INTERACTION MECHANISMS

Hereditary factors. If we examinesome of the specific ways in which he-reditary factors may influence behavior,we cannot fail but be impressed by theirwide diversity. At one extreme, we findsuch conditions as phenylpyruvic amen-tia and amaurotic idiocy. In thesecases, certain essential physical pre-requisites for normal intellectual de-velopment are lacking as a result ofhereditary metabolic disorders. In ourpresent state of knowledge, there is noenvironmental factor which can com-pletely counteract this hereditary deficit.The individual will be mentally defec-tive, regardless of the type of environ-mental conditions under which he isreared.

A somewhat different situation isillustrated by hereditary deafness, whichmay lead to intellectual retardationthrough interference with normal socialinteraction, language development, andschooling. In such a case, however, thehereditary handicap can be offset byappropriate adaptations of training pro-cedures. It has been said, in fact, thatthe degree of intellectual backwardnessof the deaf is an index of the state ofdevelopment of special instructional fa-cilities. As the latter improve, the in-tellectual retardation associated withdeafness is correspondingly reduced.

A third example is provided by in-herited susceptibility to certain physi-cal diseases, with consequent protractedill health. If environmental conditionsare such that illness does in fact de-velop, a number of different behavioraleffects may follow. Intellectually, theindividual may be handicapped by hisinability to attend school regularly. Onthe other hand, depending upon age of

onset, home conditions, parental status,and similar factors, poor health mayhave the effect of concentrating the in-dividual's energies upon intellectual pur-suits. The curtailment of participationin athletics and social functions mayserve to strengthen interest in readingand other sedentary activities. Con-comitant circumstances would also de-termine the influence of such illnessupon personality development. And itis well known that the latter effectscould run the gamut from a deepen-ing of human sympathy to psychiatricbreakdown.

Finally, heredity may influence be-havior through the mechanism of socialstereotypes. A wide variety of in-herited physical characteristics haveserved as the visible cues for identify-ing such stereotypes. These cues thuslead to behavioral restrictions or op-portunities and—at a more subtle level—to social attitudes and expectancies.The individual's own self concept tendsgradually to reflect such expectancies.All of these influences eventually leavetheir mark upon his abilities and in-abilities, his emotional reactions, goals,ambitions, and outlook on life.

The geneticist Dobzhansky illustratesthis type of mechanism by means ofa dramatic hypothetical situation. Hepoints out that, if there were a culturein which the carriers of blood group ABwere considered aristocrats and those ofblood group O laborers, then the blood-group genes would become importanthereditary determiners of behavior (12,p. 147). Obviously the association be-tween blood group and behavior wouldbe specific to that culture. But suchspecificity is an essential property ofthe causal mechanism under considera-tion.

More realistic examples are not hardto find. The most familiar instancesoccur in connection with constitutionaltypes, sex, and race. Sex and skin pig-

HEREDITY, ENVIRONMENT, AND THE QUESTION "How?" 199

mentation obviously depend upon he-redity. General body build is stronglyinfluenced by hereditary components,although also susceptible to environ-mental modification. That all thesephysical characteristics may exert apronounced effect upon behavior withina given culture is well known. It isequally apparent, of course, that in dif-ferent cultures the behavioral correlatesof such hereditary physical traits maybe quite unlike. A specific physical cuemay be completely unrelated to indi-vidual differences in psychological traitsin one culture, while closely correlatedwith them in another. Or it may beassociated with totally dissimilar behav-ior characteristics in two different cul-tures.

It might be objected that some of theillustrations which have been cited donot properly exemplify the operation ofhereditary mechanisms in behavior de-velopment, since hereditary factors en-ter only indirectly into the behavior inquestion. Closer examination, however,shows this distinction to be untenable.First it may be noted that the influenceof heredity upon behavior is always in-direct. No psychological trait is everinherited as such. All we can ever saydirectly from behavioral observations isthat a given trait shows evidence ofbeing influenced by certain "inheritableunknowns." This merely defines a prob-lem for genetic research; it does notprovide a causal explanation. Unlikethe blood groups, which are close to thelevel of primary gene products, psy-chological traits are related to genesby highly indirect and devious routes.Even the mental deficiency associatedwith phenylketonuria is several stepsremoved from the chemically defectivegenes that represent its hereditary ba-sis. Moreover, hereditary influencescannot be dichotomized into the moredirect and the less direct. Rather dothey represent a whole "continuum of

indirectness," along which are found alldegrees of remoteness of causal links.The examples already cited illustrate afew of the points on this continuum.

It should be noted that as we proceedalong the continuum of indirectness, therange of variation of possible outcomesof hereditary factors expands rapidly.At each step in the causal chain, thereis fresh opportunity for interaction withother hereditary factors as well as withenvironmental factors. And since eachinteraction in turn determines the di-rection of subsequent interactions, thereis an ever-widening network of possibleoutcomes. If we visualize a simple se-quential grid with only two alternativesat each point, it is obvious that thereare two possible outcomes in the one-stage situation, four outcomes at thesecond stage, eight at the third, andso on in geometric progression. Theactual situation is undoubtedly muchmore complex, since there will usuallybe more than two alternatives at anyone point.

In the case of the blood groups,the relation to specific genes is so closethat no other concomitant hereditary orenvironmental conditions can alter theoutcome. If the organism survives atall, it will have the blood group deter-mined by its genes. Among psycho-logical traits, on the other hand, somevariation in outcome is always possibleas a result of concurrent circumstances.Even in cases of phenylketonuria, intel-lectual development will exhibit somerelationship with the type of care andtraining available to the individual.That behavioral outcomes show pro-gressive diversification as we proceedalong the continuum of indirectness isbrought out by the other exampleswhich were cited. Chronic illness canlead to scholarly renown or to intel-lectual immaturity; a mesomorphicphysique can be a contributing factorin juvenile delinquency or in the at-

200 ANNE ANASTASI

tainment of a college presidency! Pub-lished data on Sheldon somatotypesprovide some support for both of thelatter outcomes.

Parenthetically, it may be noted thatgeneticists have sometimes used theterm "norm of reaction" to designatethe range of variation of possible out-comes of gene properties (cf. 13, p.161). Thus heredity sets the "norm"or limits within which environmentaldifferences determine the eventual out-come. In the case of some traits, suchas blood groups or eye color, this normis much narrower than in the case ofother traits. Owing to the rather differ-ent psychological connotations of boththe words "norm" and "reaction," how-ever, it seems less confusing to speak ofthe "range of variation" in this context.

A large portion of the continuum ofhereditary influences which we have de-scribed coincides with the domain ofsomatopsychological relations, as dennedby Barker et al. (6). Under thisheading, Barker includes "variations inphysique that affect the psychologicalsituation of a person by influencing theeffectiveness of his body as a tool foractions or by serving as a stimulus tohimself or others" (6, p. 1). Rela-tively direct neurological influences onbehavior, which have been the tradi-tional concern of physiological psychol-ogy, are excluded from this definition,Barker being primarily concerned withwhat he calls the "social psychology ofphysique." Of the examples cited inthe present paper, deafness, severe ill-ness, and the physical characteristicsassociated with social stereotypes wouldmeet the specifications of somatopsycho-logical factors.

The somatic factors to which Barkerrefers, however, are not limited to thoseof hereditary origin. Bodily conditionsattributable to environmental causes op-erate in the same sorts of somatopsy-chological relations as those traceable

to heredity. In fact, heredity-environ-ment distinctions play a minor part inBarker's approach.

Environmental factors: organic. Turn-ing now to an analysis of the role ofenvironmental factors in behavior, wefind the same etiological mechanismswhich were observed in the case ofhereditary factors. First, however, wemust differentiate between two classesof environmental influences: (a) thoseproducing organic effects which may inturn influence behavior and (b) thoseserving as direct stimuli for psychologi-cal reactions. The former may be illus-trated by food intake or by exposure tobacterial infection; the latter, by tribalinitiation ceremonies or by a course inalgebra. There are no completely satis-factory names by which to designatethese two classes of influences. In anearlier paper by Anastasi and Foley (4),the terms "structural" and "functional"were employed. However, "organic"and "behavioral" have the advantage ofgreater familiarity in this context andmay be less open to misinterpretation.Accordingly, these terms will be used inthe present paper.

Like hereditary factors, environmentalinfluences of an organic nature can alsobe ordered along a continuum of indi-rectness with regard to their relation tobehavior. This continuum closely paral-lels that of hereditary factors. One endis typified by such conditions as mentaldeficiency resulting from cerebral birthinjury or from prenatal nutritional in-adequacies. A more indirect etiologicalmechanism is illustrated by severe mo-tor disorder—as in certain cases of cere-bral palsy—without accompanying in-jury to higher neurological centers. Insuch instances, intellectual retardationmay occur as an indirect result of themotor handicap, through the curtail-ment of educational and social activi-ties. Obviously this causal mechanism.

HEREDITY, ENVIRONMENT, AND THE QUESTION "How?" 201

corresponds closely to that of hereditarydeafness cited earlier in the paper.

Finally, we may consider an environ-mental parallel to the previously dis-cussed social stereotypes which weremediated by hereditary physical cues.Let us suppose that a young womanwith mousy brown hair becomes trans-formed into a dazzling golden blondethrough environmental techniques cur-rently available in our culture. It ishighly probable that this metamorpho-sis will alter, not only the reactions ofher associates toward her, but also herown self concept and subsequent be-havior. The effects could range all theway from a rise in social poise to a dropin clerical accuracy!

Among the examples of environmen-tally determined organic influenceswhich have been described, all but thefirst two fit Barker's definition of soma-topsychological factors. With the ex-ception of birth injuries and nutritionaldeficiencies, all fall within the socialpsychology of physique. Nevertheless,the individual factors exhibit wide di-versity in their specific modus opemndi—a diversity which has important prac-tical as well as theoretical implications.

Environmental factors: behavioral,The second major class of environmen-tal factors—the behavioral as contrastedto the organic—are by definition directinfluences. The immediate effect ofsuch environmental factors is always abehavioral change. To be sure, someof the initial behavioral effects maythemselves indirectly affect the indi-vidual's later behavior. But this rela-tionship can perhaps be best concep-tualized in terms of breadth and per-manence of effects. Thus it could besaid that we are now dealing, not with acontinuum of indirectness, as in the caseof hereditary and organic-environmentalfactors, but rather with a continuum ofbreadth.

Social class membership may serve

as an illustration of a relatively broad,pervasive, and enduring environmentalfactor. Its influence upon behavior de-velopment may operate through manychannels. Thus social level may deter-mine the range and nature of intellec-tual stimulation provided by home andcommunity through books, music, art,play activities, and the like. Even morefar-reaching may be the effects upon in-terests and motivation, as illustrated bythe desire to perform abstract intellec-tual tasks, to surpass others in competi-tive situations, to succeed in school, orto gain social approval. Emotional andsocial traits may likewise be influencedby the nature of interpersonal relationscharacterizing homes at different socio-economic levels. Somewhat more re-stricted in scope than social class, al-though still exerting a relatively broadinfluence, is amount of formal schoolingwhich the individual is able to obtain.

A factor which may be wide or narrowin its effects, depending upon concomit-ant circumstances, is language handi-cap. Thus the bilingualism of an adultwho moves to a foreign country withinadequate mastery of the new languagerepresents a relatively limited handicapwhich can be readily overcome in mostcases. At most, the difficulty is one ofcommunication. On the other hand,some kinds of bilingualism in childhoodmay exert a retarding influence uponintellectual development and may un-der certain conditions affect personalitydevelopment adversely (2, 5, 10). Acommon pattern in the homes of immi-grants is that the child speaks one lan-guage at home and another in school,so that his knowledge of each languageis limited to certain types of situations.Inadequate facility with the language ofthe school interferes with the acquisitionof basic concepts, intellectual skills, andinformation. The frustration engenderedby scholastic difficulties may in turnlead to discouragement and general dis-

202 ANNE ANASTASI

like of school. Such reactions can befound, for example, among a number ofPuerto Rican children in New York Cityschools (3). In the case of certaingroups, moreover, the child's foreignlanguage background may be perceivedby himself and his associates as a sym-bol of minority group status and maythereby augment any emotional malad-justment arising from such status (34).

A highly restricted environmental in-fluence is to be found in the oppor-tunity to acquire specific items of in-formation occurring in a particular in-telligence test. The fact that suchopportunities may vary with culture,social class, or individual experientialbackground is at the basis of the testuser's concern with the problem ofcoaching and with "culture-free" or"culture-fair" tests (cf. 1, 2). If theadvantage or disadvantage which suchexperiential differences confer upon cer-tain individuals is strictly confined toperformance on the given test, it willobviously reduce the validity of the testand should be eliminated.

In this connection, however, it is es-sential to know the breadth of the en-vironmental influence in question. Afallacy inherent in many attempts todevelop culture-fair tests is that thebreadth of cultural differentials is nottaken into account. Failure to considerbreadth of effect likewise characterizescertain discussions of coaching. If, incoaching a student for a college admis-sion test, we can improve his knowledgeof verbal concepts and his reading com-prehension, he will be better equippedto succeed in college courses. His per-formance level will thus be raised, notonly on the test, but also on the cri-terion which the test is intended to pre-dict. To try to devise a test which isnot susceptible to such coaching wouldmerely reduce the effectiveness of thetest. Similarly, efforts to rule out cul-tural differentials from test items so as

to make them equally "fair" to subjectsin different social classes or in differentcultures may merely limit the useful-ness of the test, since the same culturaldifferentials may operate within thebroader area of behavior which the testis designed to sample.

METHODOLOGICAL APPROACHES

The examples considered so far shouldsuffice to highlight the wide variety ofways in which hereditary and environ-mental factors may interact in thecourse of behavior development. Thereis clearly a need for identifying ex-plicitly the etiological mechanism where-by any given hereditary or environmen-tal condition ultimately leads to a be-havioral characteristic—in other words,the "how" of heredity and environment.Accordingly, we may now take a quicklook at some promising methodologicalapproaches to the question "how."

Within the past decade, an increasingnumber of studies have been designedto trace the connection between specificfactors in the hereditary backgrounds orin the reactional biographies of indi-viduals and their observed behavioralcharacteristics. There has been a defi-nite shift away from the predominantlydescriptive and correlational approachof the earlier decades toward more de-liberate attempts to verify explanatoryhypotheses. Similarly, the cataloguingof group differences in psychologicaltraits has been giving way gradually toresearch on changes in group charac-teristics following altered conditions.

Among recent methodological devel-opments, we have chosen seven as be-ing particularly relevant to the analysisof etiological mechanisms. The first rep-resents an extension of selective breed-ing investigations to permit the identifi-cation of specific hereditary conditionsunderlying the observed behavioral dif-ferences. When early selective breedinginvestigations such as those of Tryon

HEREDITY, ENVIRONMENT, AND THE QUESTION "How?" 203

(36) on rats indicated that "maze learn-ing ability" was inherited, we were stilla long way from knowing what wasactually being transmitted by the genes.It was obviously not "maze learningability" as such. Twenty—or even ten—years ago, some psychologists wouldhave suggested that it was probablygeneral intelligence. And a few mighteven have drawn a parallel with the in-heritance of human intelligence.

But today investigators have beenasking: Just what makes one group ofrats learn mazes more quickly than theother? Is it differences in motivation,emotionality, speed of running, generalactivity level? If so, are these behav-ioral characteristics in turn dependentupon group differences in glandular de-velopment, body weight, brain size, bio-chemical factors, or some other organicconditions? A number of recent andongoing investigations indicate that at-tempts are being made to trace, at leastpart of the way, the steps whereby cer-tain chemical properties of the genesmay ultimately lead to specified behav-ior characteristics.

An example of such a study is pro-vided by Searle's (31) follow-up ofTryon's research. Working with thestrains of maze-bright and maze-dullrats developed by Tryon, Searle demon-strated that the two strains differed ina number of emotional and motivationalfactors, rather than in ability. Thusthe strain differences were traced onestep further, although many links stillremain to be found between maze learn-ing and genes. A promising methodo-logical development within the samegeneral area is to be found in the recentresearch of Hirsch and Tryon (18).Utilizing a specially devised techniquefor measuring individual differences inbehavior among lower organisms, theseinvestigators launched a series of stud-ies on selective breeding for behavioralcharacteristics in the fruit fly, Dro-

sophila. Such research can capitalizeon the mass of available genetic knowl-edge regarding the morphology of Dro-sophila, as well as on other advantagesof using such an organism in geneticstudies.

Further evidence of current interestin the specific hereditary factors whichinfluence behavior is to be found in anextensive research program in progressat the Jackson Memorial Laboratory,under the direction of Scott and Fuller(30). In general, the project is con-cerned with the behavioral character-istics of various breeds and cross-breedsof dogs. Analyses of some of the datagathered to date again suggest that"differences in performance are pro-duced by differences in emotional, mo-tivational, and peripheral processes, andthat genetically caused differences incentral processes may be either slightor non-existent" (29, p. 225). In otherparts of the same project, breed dif-ferences in physiological characteristics,which may in turn be related to be-havioral differences, have been estab-lished.

A second line of attack is the explo-ration of possible relationships betweenbehavioral characteristics and physio-logical variables which may in turn betraceable to hereditary factors. Re-search on EEC, autonomic balance,metabolic processes, and biochemicalfactors illustrates this approach. Alucid demonstration of the process oftracing a psychological condition to ge-netic factors is provided by the identifi-cation and subsequent investigation ofphenylpyruvic amentia. In this case,the causal chain from defective gene,through metabolic disorder and conse-quent cerebral malfunctioning, to feeble-mindedness and other overt symptomscan be described step by step (cf. 32;33, pp. 389-391). Also relevant arethe recent researches on neurologicaland biochemical correlates of schizo-

204 ANNE ANASTASI

phrenia (9). Owing to inadequatemethodological controls, however, mostof the findings of the latter studies mustbe regarded as tentative (19).

Prenatal environmental factors pro-vide a third avenue of fruitful investi-gation. Especially noteworthy is therecent work of Pasamanick and his as-sociates (27), which demonstrated atie-up between socioeconomic level, com-plications of pregnancy and parturition,and psychological disorders of the off-spring. In a series of studies on largesamples of whites and Negroes in Balti-more, these investigators showed thatvarious prenatal and paranatal disor-ders are significantly related to the oc-currence of mental defect and psychi-atric disorders in the child. An impor-tant source of such irregularities in theprocess of childbearing and birth is tobe found in deficiencies of maternal dietand in other conditions associated withlow socioeconomic status. An analysisof the data did in fact reveal a muchhigher frequency of all such medicalcomplications in lower than in highersocioeconomic levels, and a higher fre-quency among Negroes than amongwhites.

Direct evidence of the influence ofprenatal nutritional factors upon subse-quent intellectual development is to befound in a recent, well controlled ex-periment by Harrell et al. (16). Thesubjects were pregnant women in low-income groups, whose normal diets weregenerally quite deficient. A dietary sup-plement was administered to some ofthese women during pregnancy and lac-tation, while an equated control groupreceived placebos. When tested at theages of three and four years, the off-spring of the experimental group ob-tained a significantly higher mean IQthan did the offspring of the controls.

Mention should also be made of ani-mal experiments on the effects of suchfactors as prenatal radiation and neo-

natal asphyxia upon cerebral anomaliesas well as upon subsequent behavior de-velopment. These experimental studiesmerge imperceptibly into the fourth ap-proach to be considered, namely, the in-vestigation of the influence of early ex-perience upon the eventual behavioralcharacteristics of animals. Research inthis area has been accumulating at arapid rate. In 19 54, Beach and Jaynes(8) surveyed this literature for the Psy-chological Bulletin, listing over 130 ref-erences. Several new studies have ap-peared since that date (e.g., 14, 21, 24,25, 35). The variety of factors coveredranges from the type and quantity ofavailable food to the extent of contactwith human culture. A large numberof experiments have been concernedwith various forms of sensory depriva-tion and with diminished opportunitiesfor motor exercise. Effects have beenobserved in many kinds of animals andin almost all aspects of behavior, includ-ing perceptual responses, motor activity,learning, emotionality, and social reac-tions.

In their review, Beach and Jaynespointed out that research in this area hasbeen stimulated by at least four distincttheoretical interests. Some studies weremotivated by the traditional concernwith the relative contribution of matu-ration and learning to behavior develop-ment. Others were designed in an ef-fort to test certain psychoanalytic theo-ries regarding infantile experiences, asillustrated by studies which limited thefeeding responses of young animals. Athird relevant influence is to be foundin the work of the European biologistLorenz (23) on early social stimulationof birds, and in particular on the spe-cial type of learning for which the term"imprinting" has been coined. A rela-tively large number of recent studieshave centered around Hebb's (17) the-ory regarding the importance of earlyperceptual experiences upon subsequent

HEREDITY, ENVIRONMENT, AND THE QUESTION "How?" 205

performance in learning situations. Allthis research represents a rapidly grow-ing and promising attack on the modusoperandi of specific environmental fac-tors.

The human counterpart of these ani-mal studies may be found in the com-parative investigation of child-rearingpractices in different cultures and sub-cultures. This represents the fifth ap-proach in our list. An outstanding ex-ample of such a study is that by Whit-ing and Child (38), published in 1953.Utilizing data on 75 primitive societiesfrom the Cross-Cultural Files of theYale Institute of Human Relations,these investigators set out to test anumber of hypotheses regarding the re-lationships between child-rearing prac-tices and personality development. Thisanalysis was followed up by field ob-servations in five cultures, the results ofwhich have not yet been reported (cf.37).

Within our own culture, similar sur-veys have been concerned with the di-verse psychological environments pro-vided by different social classes (11).Of particular interest are the study byWilliams and Scott (39) on the as-sociation between socioeconomic level,permissiveness, and motor developmentamong Negro children, and the explora-tory research by Milner (26) on therelationship between reading readinessin first-grade children and patterns ofparent-child interaction. Milner foundthat upon school entrance the lower-class child seems to lack chiefly two ad-vantages enjoyed by the middle-classchild. The first is described as "a warmpositive family atmosphere or adult-re-lationship pattern which is more andmore being recognized as a motivationalprerequisite of any kind of adult-con-trolled learning." The lower-class chil-dren in Milner's study perceived adultsas predominantly hostile. The secondadvantage is an extensive opportunity

to interact verbally with adults in thefamily. The latter point is illustratedby parental attitudes toward mealtimeconversation, lower-class parents tend-ing to inhibit and discourage such con-versation, while middle-class parents en-courage it.

Most traditional studies on child-rear-ing practices have been designed interms of a psychoanalytic orientation.There is need for more data pertainingto other types of hypotheses. Findingssuch as those of Milner on opportunitiesfor verbalization and the resulting ef-fects upon reading readiness representa step in this direction. Another pos-sible source of future data is the ap-plication of the intensive observationaltechniques of psychological ecology de-veloped by Barker and Wright (7) towidely diverse socioeconomic groups.

A sixth major approach involves re-search on the previously cited soma-topsychological relationships (6). Todate, little direct information is avail-able on the precise operation of thisclass of factors in psychological devel-opment. The multiplicity of ways inwhich physical traits—whether heredi-tary or environmental in origin—mayinfluence behavior thus offers a rela-tively unexplored field for future study.

The seventh and final approach tobe considered represents an adaptationof traditional twin studies. From thestandpoint of the question "How?"there is need for closer coordination be-tween the usual data on twin resem-blance and observations of the familyinteractions of twins. Available dataalready suggest, for example, that close-ness of contact and extent of environ-mental similarity are greater in the caseof monozygotic than in the case of di-zygotic twins (cf. 2). Information onthe social reactions of twins towardeach other and the specialization ofroles is likewise of interest (2). Espe-cially useful would be longitudinal stud-

206 ANNE ANASTASI

ies of twins, beginning in early infancyand following the subjects throughschool age. The operation of differen-tial environmental pressures, the devel-opment of specialized roles, and otherenvironmental influences could thus bemore clearly identified and correlatedwith intellectual and personality changesin the growing twins.

Parenthetically, I should like to adda remark about the traditional applica-tions of the twin method, in which per-sons in different degrees of hereditaryand environmental relationships to eachother are simply compared for behav-ioral similarity. In these studies, at-tention has been focused principallyupon the amount of resemblance ofmonozygotic as contrasted to dizygotictwins. Yet such a comparison is par-ticularly difficult to interpret because ofthe many subtle differences in the en-vironmental situations of the two typesof twins. A more fruitful comparisonwould seem to be that between dizygotictwins and siblings, for whom the he-reditary similarity is known to be thesame. In Kallmann's monumental re-search on psychiatric disorders amongtwins (20), for example, one of the mostconvincing bits of evidence for the op-eration of hereditary factors in schizo-phrenia is the fact that the degrees ofconcordance for dizygotic twins and forsiblings were practically identical. Incontrast, it will be recalled that in in-telligence test scores dizygotic twins re-semble each other much more closelythan do siblings—a finding which re-veals the influence of environmental fac-tors in intellectual development.

SUMMARY

The heredity-environment problem isstill very much alive. Its viability isassured by the gradual replacement ofthe questions, "Which one?" and "Howmuch?" by the more basic and appro-priate question, "How?" Hereditary in-

fluences—as well as environmental fac-tors of an organic nature—vary along a"continuum of indirectness." The moreindirect their connection with behavior,the wider will be the range of variationof possible outcomes. One extreme ofthe continuum of indirectness may beillustrated by brain damage leading tomental deficiency; the other extreme, byphysical characteristics associated withsocial stereotypes. Examples of factorsfalling at intermediate points includedeafness, physical diseases, and motordisorders. Those environmental factorswhich act directly upon behavior can beordered along a continuum of breadthor permanence of effect, as exemplifiedby social class membership, amount offormal schooling, language handicap,and familiarity with specific test items.

Several current lines of research offerpromising techniques for exploring themodus operandi of hereditary and envi-ronmental factors. Outstanding amongthem are investigations of: (a) heredi-tary conditions which underlie behav-ioral differences between selectively bredgroups of animals; (b) relations be-tween physiological variables and indi-vidual differences in behavior, especiallyin the case of pathological deviations;(c) role of prenatal physiological factorsin behavior development; (d) influenceof early experience upon eventual be-havioral characteristics; (e) culturaldifferences in child-rearing practices inrelation to intellectual and emotionaldevelopment; (/) mechanisms of soma-topsychological relationships; and (g)psychological development of twins frominfancy to maturity, together with ob-servations of their social environment.Such approaches are extremely variedwith regard to subjects employed, na-ture of psychological functions studied,and specific experimental procedures fol-lowed. But it is just such heterogeneityof methodology that is demanded bythe wide diversity of ways in which he-

HEREDITY, ENVIRONMENT, AN£> THE QUESTION "How?" 207

reditary and environmental factors in-teract in behavior development.

REFERENCES

1. ANASTASI, ANNE. Psychological testing.New York: Macmillan, 1954.

2. ANASTASI, ANNE. Differential psychology.(3rd ed.) New York: Macmillan, 1958.

3. ANASTASI, ANNE, & CORDOVA, F. A. Someeffects of bilingualism upon the intelli-gence test performance of Puerto Ricanchildren in New York City. /. educ.Psychol, 1953, 44, 1-19.

4. ANASTASI, ANNE, & FOLEY, J. P., JR. Aproposed reorientation in the heredity-environment controversy. Psychol. Rev.,1948, 55, 239-249.

5. ARSENIAN, S. Bilingualism in the post-war world. Psychol. Bull, 1945, 42,65-86.

6. BARKER, R. G., WRIGHT, BEATRICE A.,MYERSON, L., & GONICK, MOIXIE R.Adjustment to physical handicap andillness: A survey of the social psychol-ogy of physique and disability. Soc.Sci. Res. Coun. Bull., 1953, No. 55(Rev.).

7. BARKER, R. G., & WRIGHT, H. F. Mid-west and its children: The psychologi-cal ecology of an American town.Evanston, 111.: Row, Peterson, 1955.

8. BEACH, F. A., & JAYNES, J. Effects ofearly experience upon the behavior ofanimals. Psychol. Bull., 1954, 51, 239-263.

9. BRACKBIIX, G. A. Studies of brain dys-function in schizophrenia. Psychol.Bull., 1956, 53, 210-226.

10. DARCY, NATALIE T. A review of the lit-erature on the effects of bilingualismupon the measurement of intelligence./. genet. Psychol., 1953, 82, 21-57.

11. DAVIS, A., & HAVIGHURST, R. J. Socialclass and color differences in child rear-ing. Amer. social. Rev., 1946, 11, 698-710.

12. DOBZHANSKY, T. The genetic nature ofdifferences among men. In S. Persons(Ed.), Evolutionary thought in Amer-ica. New Haven: Yale Univer. Press,1950. Pp. 86-155.

13. DOBZHANSKY, T. Heredity, environment,and evolution. Science, 1950, 111,161-166.

14. FORGUS, R. H. The effect of early percep-tual learning on the behavioral organi-zation of adult rats. J. comp. physiol.Psychol., 1954, 47, 331-336.

15. HALDANE, J. B. S. Heredity and politics.New York: Norton, 1938.

16. HARRELL, RUTH F., WOODYARD, ELLA, &GATES, A. I. The effect of mothers'diets on the intelligence of the offspring.New York: Bur. Publ., Teach. Coll.,Columbia Univer., 1955.

17. HEBB, D. O. The organization of behav-ior. New York: Wiley, 1949.

18. HIRSCH, J., & TRYON, R. C. Mass screen-ing and reliable individual measurementin the experimental behavior genetics oflower organisms. Psychol. Bull., 1956,53, 402-410.

19. HORWITT, M. K. Fact and artifact in thebiology of schizophrenia. Science, 1956,124, 429-430.

20. KALLMANN, F. J. Heredity in health andmental disorder; Principles of psychi-atric genetics in the light of compara-tive twin studies. New York: Norton,1953.

21. KING, J. A., & GURNEY, NANCY L. Effectof early social experience on adult ag-gressive behavior in C57BL10 mice. /.comp. physiol. Psychol., 1954, 47, 326-330.

22. LOEVINGER, JANE. On the proportionalcontributions of differences in natureand in nurture to differences in intelli-gence. Psychol. Bull., 1943, 40, 725-756.

23. LORENZ, K. Der Kumpan in der Unwelldes Vogels. Der Artgenosse als aus-16'sendes Moment sozialer Verhaltungs-weisen. /. Orn., Lpz., 1935, 83, 137-213; 289-413.

24. LUCHINS, A. S., & FORGUS, R. H. The ef-fect of differential postweaning environ-ment on the rigidity of an animal's be-havior. /. genet. Psychol., 1955, 86,51-58.

25. MELZACK, R. The genesis of emotionalbehavior: An experimental study of thedog. J. comp. physiol. Psychol., 1954,47, 166-168.

26. MILNER, ESTHER A. A study of the rela-tionships between reading readiness ingrade one school children and patternsof parent-child interaction. Child De-velpm., 1951, 22, 95-112.

27. PASAMANICK, B., KNOBLOCH, HILDA, &LILIENFELD, A. M. Socioeconomic statusand some precursors of neuropsychiatricdisorder. Amer. J. Orthopsychiat., 1956,26, 594-601.

28. SCHWESINGER, GLADYS C. Heredity andenvironment. New York: Macmillan,1933.

208 ANNE ANASTASI

29. SCOTT, J. P., & CHARLES, MARGARET S.Some problems of heredity and socialbehavior. /. gen. Psychol., 1953, 48,209-230.

30. SCOTT, J. P., & FULLER, J. L. Researchon genetics and social behavior at theRoscoe B. Jackson Memorial Labora-tory, 1946-1951—A progress report. /.Hered., 1951, 42, 191-197.

31. SEARLE, L. V. The organization of he-reditary maze-brightness and maze-dull-ness. Genet. Psychol. Monogr., 1949,39, 279-325.

32. SNYDER, L. H. The genetic approach tohuman individuality. Sci. Man., N. Y.,1949, 68, 165-171.

33. SNYDER, L. H., & DAVID, P. R. The prin-ciples of heredity. (5th ed.) Boston:Heath, 1957.

34. SPOERL, DOROTHY T. Bilinguality andemotional adjustment. J. abnorm. soc.Psychol, 1943, 38, 37-57.

35. THOMPSON, W. R., & MELZACK, R. Earlyenvironment. Sci. Amer., 1956, 194(1), 38-42.

36. THYON, R. C. Genetic differences in maze-learning ability in rats. Yearb. mat.Soc. Stud. Educ., 1940, 39, Part I,111-119.

37. WHITING, J. W. M., et al. Field guidefor a study of socialization in five so-cieties. Cambridge, Mass.: HarvardUniver., 1954 (mimeo.).

38. WHITING, J. W. M., & CHILD, I. L. Childtraining and personality: A cross-cul-tural study. New Haven: Yale Uni-ver. Press, 1953.

39. WILLIAMS, JUDITH R., & SCOTT, R. B.Growth and development of Negro in-fants: IV. Motor development and itsrelationship to child rearing practices intwo groups of Negro infants. ChildDevelpm., 1953, 24, 103-121.

40. WOODWORTH, R. S. Heredity and envi-ronment: A critical survey of recentlypublished material on twins and fosterchildren. Soc. Sci. Res. Coun. Bull.,1941, No. 47.

(Received September 18, 1957)