Reviews

Nature-Nurture and the Two Realms ofDevelopment: A Proposed Integration withRespect to Mental Development

Robert B. McCallBoys Town Center for the Study of Youth Development

MCCAIX, ROBERT B Nature-Nurture and the Two Realms of Development A Proposed Inte-gration with Respect to Mental Development CHILD DEVELOPMENT, 1981, 52, 1-12 It isargued that developmental psychologists need attitudes, methods, and conceptual schemes thatintegrate the distinctive contnbutions of nature and nurture to study developmental change aswell as consistency in both developmental functions and mdividual differences in behaviors ofinterest A conceptual scheme for early mental development is described that synthesizes thesedisparate orientations This theoretical viewpoint deals with 2 sets of seenungly contradictoryfacts (1) the early porhon of the developmental function is largely maturational, while indi-vidual differences are unstable and not highly correlated with either genetic or environmentalfactors, and (2) as nature's hold on the developmental function declines with age, individualdifferences correlate more stronglj with both genetic and environmental factors

Young sciences often emerge out of dog-ma, and the vesbges of dogma sometimes lingerm the form of allegiances to narrow concep-tions about the fundamental nature of behaviorand to insular methodological strategies Sucha division characterizes psychology in generaland developmental psychology in particular

Specifically, Cronbach (1957) warned usmore than 20 years ago about the "two disci-plines of scientific psycbology"—experimentalmanipulation of variables, on the one hand, andthe consistency of individual differences acrosscontexts, behaviors, and time, on the otherThe distinction has many partial confoundsexperimental versus observational research,group means versus individual difiFerences,analysis of vanance versus correlations, etc

In the history of our own subdisciplme,the old "child developers," correlationists bynature, were joined by a few animal research-ers who brought with them a fervent belief ingroup differences and the experimental methodThese newcomers failed to convert the corre-lationists, although they did manage to makedevelopmental psychology more "scientifically

respectable " But coexistence is not mtegrationThose who emphasize group differences andexpenmental research and students of individ-ual differences are still essentially separate sects(McCall 1977a)—the two realms of develop-mental psychology

This dichotomy partly influences what westudy Consider the immortal nature-nurturedebate Behavior geneticists pursue the geneticcontribution to a behavior through the studyof individual differences, while environmental-ists attempt to uncover experiential influencesm any way they can For the most part, neithergroup talks to the other, and both schools ig-nore Anastasi's (1958) plea, also of more than20 years ago, to forget these allegiances and towork together to understand how heredity andenvironment jointly contribute to behavioralcharacteristics

Superimposed on these differences is thefact that few developmental psychologists ofany stnpe actually conduct a great deal of re-search on development per se— t̂hat is, changewithm organisms across age (McCall 1977a,Wohlwill 1973) Rather, contemporary devel-

This article is based parti) on tnatenal contained in "Transitions in the Growth of Mind"presented by the author as part of the symposium, "Consistency and Change m Behayior OverTune and Across Situations," at the meetings of the Amencan Psychological Association inToronto, August 1978 I thank Jerome Kagan, Robert Plomin, and Joachim Wohlwill for theirhelpful comments on an earlier draft of this manuscript Requests for repnnts should be sent toR B McCall, Center for the Study of Youth Development, Boys Town, Nebraska 68010

ICktld Detelopment, 1981,52, 1-12 © 1981 by the Society for Research in Child Development, Inc0009 3920/81/5201-OOOItOl 00)

2 Child Development

opmental psychology is mamly the parametricstudy of immature organisms, it is not pnmarilythe study of longitudinal changes

The ecumenism of Cronbach, Anastasi,and Wohlwill has largely fallen on deaf earsIt should be resurrected as often as necessary,because I believe our understandmg of manybehaviors is limited by these allegiances Eachorientation potentially has something unique tocontribute, but these approaches also must bewoven together into a single conceptual fabricbefore real progress will be made in under-standing behayior and its development

This paper offers a modest step towardintegratmg the two disciplines of psychologyaround the nature-nurture issue in the early de-velopment of mental behavior It begins witha descnption of the major issues sketchedabove, pomtmg out the limitations of each andthe fragmented and sometimes contradictorypicture of behavior that results from relying ononly one perspective Then a conceptualizationof mental development is proposed as an illus-tration of how an integrated viewpoint can beapplied to a behavioral domain

This theoretical scheme helps to reconcileapparently conflicting facts about early mentaldevelopment For example, how is it possiblethat much of early mental development is un-der maturational control, yet individual differ-ences are unstable and do not correlate highlywith either genetic or environmental factors'"And as the maturational influence declines withage, why do correlates with both genetic andenvironmental factors increase?

But the importance of this theoretical pro-posal IS not that it represents a detailed modelof mental development It does not, and thescheme quickly breaks down when pushed tomatch the details of the literature Rather, itstands as one example of how a behavioral do-main can be conceptualized across traditionalboundaries to yield a fuller, more comprehen-sive understanding of its development

The Cost of Separatism

Specialization is often necessary But thecost IS sometimes lack of integration, minimumcross-fertilization of ideas and approaches, andnarrowness Developmental psychology suffersthese consequences on many fronts, and anycomprehensive theory or approach to a prob-lem must integrate several of these otherwiseseparate stances

I have argued before in these pages (Mc-Call 1977a) that developmental psychology israrely actually developmental, and that ourdiscipline does not pursue the study of changenearly as vigorously as it searches after consis-tency These and other allegiances limit thebreadth and social contribution of our disci-phne Therefore, our approach to major issuesmust be developmental and must explainchange as well as consistency

The Ttvo Realms of DevelopmentChange and consistency over age can oc-

cur m both of Cronbach's (1957) two realmsSuppose figure 1 is a plot of the growth of anattribute over age—verbal fluency, for exam-ple The heavy line depicts the average of agroup of subjects, and the thinner lines repre-sent each of the Rye individuals m the hypo-thetical sample Statistically, the distinction issimply the difference between the absoluteyalue of the average curve at different ages,on the one hand, and the stabihty of relativerank orderings of lndiyiduals from one age tothe next, on the other The average absoluteyalue of a trait m a group is independent ofthe relatiye stability of mdiyidual differencesbecause the correlation between two sets ofscores is independent of the means of those <two distributions

Developmental functions—The measuredvalue of a given attribute plotted across agedefines the developmental function of that char- ^actenstic In figure 1, the thinner lines are de-velopmental functions for individuals, while theheavier line is the developmental function forthe sample If the sample is representative ofa species, then the group curve is an estimate ',of the species-general deyelopmental functionDeyelopmental functions are what expenmen- "talists and those concemed about the nature ofthe species in general (eg , Piaget) tend tostudy

Deyelopmental functions for mdiyiduals,groups, or species are either continuous or dts- 'continuous (Emmerich 1964) By one defini- 'tion, a developmental function is continuouswhen changes are quantitatiye rather than ]qualitatiye, that is, when the fundamental na-ture of the attribute remains the same oyer age !̂A plot of height over age would be continuous tbecause the fundamental character of height is .^,essentially the same at every age, even thoughthe average measured yalue changes The de-yelopmental function for yocabulary would alsobe continuous In contrast, a plot of Piagetiansensorimotor deyelopment would be discontm-

uous because its specific behavioral characterIS different from one stage to the next Obvi-ously, whether a trait is continuous or discon-tinuous depends partly on how it is measured(eg, the IQ score portrays a continuous func-tion for a characteristic that may actually bediscontinuous)

Individual differences —Development canalso be assessed in another realm—in terms ofthe relative consistency of individual differ-ences over age That is, do individuals main-tain the same relative rank ordering withmtheir group at two different ages, or does therelative rank ordermg change from one age tothe next' The term stabtlUy refers to the rela-tive consistency of such individual differencesIn the hypothetical plot in figure 1, individualdifferences are not stable during the earlyyears but become more so later

The costs of separate disciplines—Re-search on development as well as other areasof psychology tends to be earned out in oneor the other of these realms but not simulta-neously m both For example, Piaget was con-cerned only with the species-general functionfor mental development—he hterally did notcare about individual differences On the otherhand, much research on early mental behaviorhas revolved around correlates of early mentaltest performance and predictions to later IQ

AGE

FIG 1 — A hypothetical plot of the develop-mental function of a given behavior for five indi-viduals (thin lines) and the developmental functionof the group (heavy line) (Reprmted from McCallet al [1977], with permission of the Society forResearch m Child Development )

Rohert B. MeCall 3

( l e , individual differences) Such studies fre-quently rely on DQ or IQ, indices which essen-tially obviate any analysis of the qualitabveand quantitative nature of the species-generaldevelopmental function

Researchers tend to ignore the potentialindependence of these realms and draw conclu-sions about a behayior in general on the basisof eyidence from only one sphere One classicillustration is Bloom's (1964) suggestion thatby the time one is 4 years old, 50X of one'sadult intelligence is deyeloj)ed This conclusionwas based on the statistical fact that the corre-lation between IQ at 4 years and at 17 yearsIS approximately 71 ( 71^ = 50) But thisclaim for the development of intelligence ingeneral is based solely on the stabihty of indi-vidual differences—it completely ignores thefact that the average child's mental age ( l e ,the "amount" of mental skill displayed) m somesense will multiply several times during thisinteryal Relying solely on indiyidual differ-ences IS like studying the consistency of a dif-ference of a few inches m the heights of giantsequoia trees from seedlings to maturity whileIgnoring the issue of hovv' all the trees grow tobe over 300 feet tall

Another illustration of the failure to keepthese orientations separate is the recent preoc-cupation with controlling for secular change,time of measurement, and testing experienceMuch research on individual differences hasbeen maligned or dismissed because it lackedsuch controls But the data showang secularchange and repeated testing effects pertainalmost totally to the developmental function,relatively little evidence suggests that thesefactors infiuence the stabibty or pattern of indi-vidual differences (McCall 1977a)

Potenttal or real independence^—Whilethe two realms are potentially independent,they may not be independent m nature Butthis IS an empirical question My point is thatit IS rarely asked One such attempt (McCall,Eichorn, & Hogarty 1977) found that dips mthe level of cross-age correlations for mentaltest performance coincided with qualitativechanges in the fundamental nature of mentalbehavior assessed by that instrument But thisapparently happy convergence may not neces-sarily be the case, nor have we giyen muchthought to the conceptual relationship betweenstage changes ( l e , discontinuities) and sta-bihty of mdiyidual differences (see Uzgiris1977)

4 Child Development

The oyemdmg point is that major strate-gies of inquiry must consider developmentalchange and consistency m both developmentalfunction and individual differences

Nature-NurtureThe nature-nurture issue is one of the

most persistent, celebrated, and heated contro-versies in psychology It also suffers from al-legiances and the failure to distinguish betweendevelopmental function and individual differ-ences

A prime illustration is the interpretationof the Skodak and Skeels study (1949, Honzik1957) on the IQs of adopted children and theirbiological and foster parents From the stand-point of individual differences, the IQs of theadopted children correlated 38 with those oftheir biological parents but essentially zerowith an estimated index for their rearing par-ents However, the ayerage IQ of the childrenwas 21 points higher than the ayerage of theirbiological parents and nearly identical with theestimated ayerage of their rearing parentsHereditanans tend to emphasize the individual-difference result, enyironmentalists concentrateon the mean difference (l e, deyelopmentalfunction) Both obseryations are useful but dis-tinct pieces of information, and they are notcontradictory (Jensen 1973) But we rarelyyiew them as two pieces of the same puzzle

From another standpoint, essentially allmethods of population genetics used to deter-mine the heritability of a characteristic arebased on individual differences This fact posestwo major limitations First, genetic or environ-mental factors that influence individual differ-ences may or may not infiuence the develop-mental function of that sample For example,circumstances that permit almost all infants towalk and run may not play any role m howwell, how fast, or how far one individual canwalk or run relative to another And factorsthat produce precocity in the attainment of ob-ject permanency, thus producing a spurt m thedevelopmental function for these individuals,may have nothing to do with how all childrenultimately acquire essentially comparable pro-ficiency m the basic elements of this skill

Second, since all methods of assessinghentabihty rely on individual differences, thereIS almost no way to assess the hentabihty ofa species-general deyelopmental function be-cause there are no mdiyidual differences bydefinition Race or other subgroup differencescould be used, but the races may not differ m

the basics of sensorimotor development or earlylanguage acquisition, for example Further, fac-tors that determine differences between racesmay not determine their similanties—which iscrucial to the definition of the species-generaldeyelopmental function

Lest the enyironmentalists feel smug, theyare no better off The enyironmental as well asthe genetic factors necessary to produce funda-mental characteristics m the species are avail-able to almost everyone we study As a result,the only way we can study the importance ofcertain major environmental factors for devel-opment IS to take advantage of tragedies—children reared m closets, bom blind and latergiyen sight, or fed from birth through a fistula

In fact, there may be a good eyolutionaryexplanation for the possibility that one set offactors influences mdiyidual differences and an-other set operates on the species-general de-yelopmental function If a giyen trait fayorssuryiyal and reproduction, then such an attri-bute would tend to become characteristic ofeach member of the species That is, this ge-netic trait would characterize the species-gen-eral deyelopmental function But genetic yan-ability on this attribute would be reduced, con-sequently, mdiyidual differences m this attn-bute would be caused by environmental cir-cumstances (McClearn & DeFnes 1973, Plomin& Rowe 1979) Perhaps, then, it is not sur-prising that some anthropologists and geneti-cists (King & Wilson 1975, Washbum 1978)haye suggested that 99% of our genetic materialproduces species-general characteristics, not m-diyidual differences (but see Plomin & Kuse1979)

This IS not to say that genetic factors mmdiyidual differences are not important, quitethe contrary, society is much more concemedwith differences between people than with howhumans differ from the apes The point, how-eyer, is that nearly the entire nature-nurtureargument resides m the realm of individualdifferences and ignores the species-general de-yelopmental function Moreoyer, it is quite pos-sible for the species deyelopmental funcbonto be almost totally under genetic control(giyen certam necessary environmental circum-stances) but for individual differences m thatbehavior to possess no hentabdity and perhapsno obvious correlations with stable environ-mental attributes m most research populationsIndeed, this is very nearly the case for earlymental development (McCall 1979a, 1979b)Moreover, deyelopmental functions must be

RoheH B. McCall 5

studied as functions, and there are distinctlimits on our abihty to lnyestigate the factorsthat contribute to species-general developmen-tal functions But this is no reason to dismissthem from our theorizing or to confuse themwith other aspects of deyelopment

Toward a Broader View—the Case ofMental Development

The thesis to this point has been thatallegiances and fayorite methodological ap-proaches contribute to a narrow, incomplete,and occasionally distorted understanding ofsome of the behayiors we study We need atti-tudes, methods, and conceptual schemes thatintegrate the distinctiye contributions of natureand nurture to the study of deyelopmentalchange as well as consistent m deyelopmen-tal functions and individual mfferences

The following scheme for early mental de-velopment IS offered as a step m this directionI caution that it is only a single step Newanswers to major questions are not proposedInstead, I have tried to understand the appar-ent contradictions m the data already ayail-able Further, what follows is not a detailedtheoretical model that must yield a host oflnnoyatiye testable deductions or must be rigor-ously eyaluated in terms of the evidenceRather, this is an illustration of a way ofthinking, it is an example of what one suchintegrated approach to early mental develop-ment might look like m broad outline Whilethe general strategy could be applied to thestudy of almost any behavior, most of the spe-cifics focus on mental deyelopment and maynot generalize to social, personal, and other be-hayioral domains

CanaltzattonThe concept of canahzation, described by

Waddington (1957) and recently apphed tomental development by Scarr-Salapatek (1976),imphes a species-typical path, called a creod,along which nearly all members of the speciestend to develop However, a characteristic fol-lows the creod only as long as species-typicalappropriate enyironments predominate Whensuch environments exist, development proceeds'normally", when such circumstances deviatemarkedly, deyelopment can go awry There-fore, the utility of the concept of canalizationdepends on the breadth of enyironments desig-nated as species typical When deyelopment is"highly canalized," mdividuals follow the spe-cies creod under a wide range of diyerse en-vironments and exhibit strong self-nghtmg ten-

dencies following exposure to seyerely atypicalenyironments When deyelopment is "less can-alized," individuals do not follow a commondevelopmental path as uniformly, variability menvironments and genetic circumstances pro-duce more frequent and more seyere deflec-tions from the sj>ecies ayerage, and there isless tendency to self-right toward that normfollowing such deyiations

Canalization is a piyotal concept becauseIt has lmphcations for the several aspects ofdeyelopment discussed aboye These implica-tions may be seen more clearly m the schemeof early mental deyelopment described below

A Canahzation Scheme of EarlyMental Development

A fundamental proposition is that earlymental deyelopment is highly canahzed duringthe first 18-24 months of life but thereafterbecomes less canalized That is, infants pro-ceed along the species-typical path under awide range of enyironments, and there is astrong self-nghting tendency should extremecircumstances deflect an mfant from this creodHowever, beginning at approximately 18-24months, mental development becomes progres-sively less canalized with age, and it is afterthis point that the self-nghtmg tendency weak-ens and individual differences become morestable

The scoop approach—This notion is pic-tured in figure 2 as a scoop or trough Thescoop represents the creod or species-generaldevelopmental function for mental deyelop-ment The designs on its mside signify differ-ent qualitatiye stages of mentality The pre-dominant character of mental behavior changesduring infancy, and these stages are disconti-nuities m the nature of the deyelopmental func-tion of early lntelhgence The grooyes or chan-

FiG 2-The "scoop" approach to thinkingabout mental deyelopment

6 Child Development

nels that appear in late infancy and branch intochildhood are mdiyidual differences m geneticdisposition They allow for deyelopmg varia-tions and divergence in genetic disposition mthe population

Now imagine that a hghtweight ball rep-resents an mdiyidual who begins life at theleftmost point of the scoop The ball rolls downthe scoop oyer deyelopmental time The mclineof the scoop represents necessary enyironmen-tal contnbutors to the species-general functionHoweyer, in the course of deyelopment, enyi-ronmental winds blow oyer the scoop and pro-duce mdiyidual differences in the rate of de-yelopment (if the winds blow up or down thescoop) or m the nature of the individual's de-velopmental function (if the winds blow acrossthe scoop)

To summarize With respect to the de-yelopmental function, the basic path definedby the scoop represents heredity's creod, theforce of grayity that pulls the ball (l e , lndi-yidual) down the scoop which slants to thenght IS the enyironmental component Both arenecessary for species-general deyelopmentWith respect to mdiyidual differences, the di-yergmg grooyes in the scoop are mdiyidualyariations m qualitatiye genetic disposition,while winds along the scoop that can blow theball from stage to stage faster or slower orwmds across the scoop that can blow the ballinto new qualitatiye emphases are the enyiron-mental component

As a conceptual scheme, this approach isan attempt to combine Piaget's (1954/1966)stage description of the species-general de-yelopmental function with what is known aboutmdiyidual differences m mental performanceat a giyen age or across age A ball startmgout at the left is tightly contained by the nar-row creod which it follows under a wide rangeof enyironmental winds Eyen when a stronggust does blow the ball somewhat off course,the steep sides of the scoop force it to rightitself Of course, there are limits, and temblyseyere enyironmental or genetic factors canproduce permanent deflectaons m the deyelop-mental course But, under a wide range of cir-cumstances, infants will follow basically thesame deyelopmental path Nature is governingthe developmental function, and individual dif-ferences have minimum longitudinal stabiht)'and only minor contemporaneous correlationswith genetic or environmental variations

The mental behaviors that emerge duringthis period tend to be fundamental character-istics which all members of the species attaincompletely with relatively insignificant individ-ual differences (Horn [1968] has called themanlagen) Every infant acquires object perma-nency, the abihty to walk, and basic symbohsmunder a wide range of genetic and environ-mental circumstances Indeed, in their mostelementary form, these attributes are developedessentially all-or-none We can barely imagineadult individual differences m proficiency ofbasic figure-ground We are concerned thatinfants acquire object permanency but not howwell they will ultimately achieve it

Mental development begins to be lesscanalized at approximately 18-24 months, al-though it IS a gradual process Characteristicsand abilities now emerge that all infants willacquire, but some will eventually attain muchdifferent levels of performance from othersAlmost all humans acquire language, but somebecome more facile and fluent than others

Three factors of the scheme change simul-taneously First, the creod becomes more dif-ferentiated (1 e , the grooyes representing indi-yidual differences in genetic dispositions) Sec-ond, the sides of the scoop fall off, making theindiyidual more yulnerable to the differentialand yanable winds of experience Notice thatboth unique genetic and enyironmental circum-stances now haye a greater potential impact onan individual, and correlations with genetic andenvironmental factors are likely to increase andremain more stable over time now that canali-zation IS weakening Third, the ball is pickingup speed This represents the cumulatiye effectof experience and the fact that humans arepartly able to select their own environments—a circumstance that produces longitudinal sta-bility and correlations between enyironmentaland genetic circumstances

These changes complicate the theoreticaldynamics considerably, but they unquestion-ably occur m some form On the one hand, thespeed of the ball lends stability to the system(1 e , enyironmental breezes may haye less ef-fect than they might otherwise), and preyiouslyestablished charactenstics tend to persist Onthe other hand, notice that the lowered wallspotentially make the organism more subject toenyironmental influences, although the windsmust be strong enough to overcome the ball'sinertia But, if the ball is blown from its previ-ous course, its gathering speed means tliat it

IS more likely to persist m the new directionThe indiyidual organism is now potentiallymore yulnerable to both enyironmental andgenetic influences than when younger, but theimpact of these factors may be resisted by theinertia of deyelopment ( eg , cumulative ex-perience, environment selection) or, once effec-tive, amplified by these same factors Earlygenetic and environmental factors can nowhaye a more noticeable and prolonged effecton indiyidual differences because of their cu-mulatiye character and the fact that peopleselect enyironments that tend to remain rela-tiyely stable oyer age But at the same time,the potential for change is also greater becauseof genetic differentiation and the lowering ofnature's protectiye walls Which force wins outin any mdiyidual case requires more speciflcinformation than this scheme contains, but it islikely that ages 2-4 (approximately) may bethe most "sensitiye" period for mental deyelop-ment

Implications of the SchemeThe scoop approach suggests seyeral prin-

ciples which heretofore appeared contradictorybut may now be seen to be diverse facets ofthis broader conception of mental development

The developmental function m the first 18months—The scheme suggests that infantsshould follow a relatively common sequence ofstages during the first 18 months of life andthat most infants who deviate from that courseshould return to the norm once the deflectingcircumstance is removed

Piaget (1954/1966) proposed a stage-sequence theory of early mental developmentthat has received wide acceptance Althoughrecent formulations have made slight changesin the Piagetian model, there is more agree-ment than disagreement For example, severalresearchers agree that major stage boundariesoccur at approximately 2, 7-8, 13, and 21months of age ( eg , Fischer 1980, McCallet al 1977, Piaget 1954/1966, Uzgins 1976)Although individual investigators differ on whatbasic mental attribute underlies each stage,there is much more agreement about the timingand sequence of specific behavioral events(e g , the exploration of objects and perceptualcontingencies, object permanency, entity-entityassociative relations, imitation, vocabulary, two-word sentences, symbolism) Readers interestedm the specific attributes of these stages are re-ferred to the above references

Rohert B. McCall 7

More important from the standpoint ofcanalization is whether self-nghtmg is a pos-sible phenomenon Infants who haye sufferedmajor adyersity—including nutritional depriya-tion, prematurity, anoxia, and other neonatalinsults that produce contemporaneous depres-sion in infant test scores—tend to recoyer andreturn to essentially normal deyelopment withm3 to 6 years if they are reared m adequate en-yironments (Honzik 1976, Hunt 1976, Sam-eroff & Chandler 1975, Scarr-Salapatek 1976)Howeyer, if such infants are reared m marked-ly inferior enyironments, the effects of earlyinjury can persist Therefore, infants who arereturned to species-typical environments aftera variety of major insults can return to thecreod

Individual differences tn the first 18months —While infants closely follow the bio-logical creod, most individual differences willbe neither as stable across time nor as corre-lated with genetic or environmental factors asthey will later (Honzik 1976, McCall 1979a,1979b, McCall et al 1977, McCall, Hogarty,& Hurlburt 1972, Scarr-Salapatek 1976) Therelative instability of mdividual differences de-rives from the strong self-righting tendencycharacteristic of highly canalized developmentPerhaps the mtraindividual variabihty charac-teristic of this period has the adaptive functionof increasing the likelihood of eventuallymatching behavior with environmental circum-stances

There has been some debate about thegenetic correlates of individual differences ininfant test performance Wilson (1978) pointsto high within-pair concordances among mono-zygotic (MZ) and dizygotic (DZ) twins forsingle mental assessments and for developmen-tal profile Others (McCall 1972, McCall,Appelbaum, & Hogarty 1973, Scarr-Salapatek1976) have argued that hentabihty is basedon the difference between withm-pair correla-tions for MZs versus DZs, and that, while thesedifferences are significant m Wilson's data, theyare not large In fact, hentabihties for singleassessments in the Louisville Twin Study ayer-age 25 between 3 months and 5 years (McCall1979b), a result consistent with other datademonstrating modest kinship correlations pnorto 4-6 years of age ( e g , Beckwith 1971, Cas-ler 1976, Honzik 1957)

Age-to-age changes show some hentabihtyin the first 2 years (50), but apparently thisdeclines thereafter (McCall 1970, 1972, 1979b,

8 Child Development

Wilson 1978) This early hentabihty for de-yelopmental profile among twins has not beendemonstrated for singleton sibhngs (McCall1970, 1972, McCall et al 1973) This differ-ence may deriye from pre- and pennatal cir-cumstances more likely shared by twins mgeneral—and by MZ twins in particular—which decline m influence oyer the first yearsof life A similar interpretation based upontemporary permatal factors may explain Cas-ler's (1976) observation of modest correlationsbetween biological mothers' IQ and Cesell sub-test scores of their orphanage-reared infants at2 months but not at 9 or 15 months of age

Occasionally, a presumed environmentalcircumstance obseryed in the first year of lifepredicts later mental performance to a sur-prisingly high degree Howeyer, it is usuallynot clear whether this factor actually influencedthe child during the first year with enduringconsequences, or whether it is an early pre-dictor of more functional enyironmental yari-ables that haye an influence on mental perfor-mance only later m childhood For example,one might expect that highly yerbal parentswould talk to their children a great deal, be-ginning in early mfancy Developmental re-searchers might obserye a correlation betweenearly parental language and later child intelh-gence and be prone to infer that talking to a6-month-old infant lmproyes the child's intelli-gence at 3 years of age However, languagestimulation at 6 months may actually haye noeffect on the child, but it is the same parentwho talks to the 6-month-old who also talks tothe 3-year-oId—at which tune such languagestimulation actually does haye a causal effectThis possibility is anathema to researchers, be-cause m the study of the natural deyelopmentof mental competence it is often impossibleto discriminate between the early precursorand the early dysfunctional correlate of a latercausal factor (except under certain conditions,e g, adoption and unusual enyironmentalchanges in the liyes of young children)

By and large, howeyer, age-to-age stabih-ties and correlations of mental test performancewith genetic and environmental factors, whilenot zero, are quite modest and definitely lowerthan analogous relationships found after 18-24months Presumably, this fact derives from thehigh canalization of early mental deyelopmentthat keeps infants on the species-general de-yelopmental creod and retums wayward indi-yiduals to that path

This characterization of early mental de-yelopment may leave traditional developmentalresearchers frustrated What does one do witha mentality that is both discontinuous and un-stable^ The deyelopmental function does notplot continuously on the same axis, and thereare few correlations to find Even mdiyidualdifferences in age-at-stage-entrance might notbe consistent from stage to stage

But if this IS the way nature has createdIts children, let not scientists cast them intoanother image Deyelopment of this kmd isproblematic for research only if scientists nar-rowly focus on continuity, stabihty, and sim-plistic dichotomies of heredity and enyiron-ment This state of affairs forces us to studychange—change in the species' deyelopmentalfunction, change m indiyidual differences indeyelopmental function, change in the corre-lates of mentality at different ages, factors nec-essary for change to occur, the behavioral con-sequences of change, the fundamental natureunderlying each distinct stage, and the breadthof honzontal decalage

The developmental function after 2 years—After 18-24 months of age, canalizationgradually weakens Stages characterizing thespecies-general developmental function becomemore widely spaced, and traits emerge ( e g ,symbohsm, yerbal fluency) for which profi-ciency leyels will ultimately vary widely be-tween individuals Pnor to this point, all de-velopments were essentially species-general andpossessed few endurmg individual differencesAfter this point, while some fundamental spe-cies-general anlagen may still emerge ( eg ,at 5-7 years of age, see Cruen & Doherty[1977], White [1965]), greater diyersity occurswithin species-general stages as a function ofboth genetic and enyironmental factors

Individual differences after 2 years —As aresult of the decline of canalization and theincreased yanabihty, indiyidual differences mmentahty become more stable across age andshow higher correlations with both enyiron-mental and genetic circumstances (McCall1979b) Apparently, with the emergence ofcertain symbolic capabilities, predictions tolater IQ increase rapidly (McCall et al 1977),predictions from mental test scores to adulteducational and occupational success approachasymptote (McCall 1977b), kinship similantiesincrease and level off at 5-7 years (Honzik1957), and correlations with presumed environ-mental factors also increase (McCall 1979b)

Rohert B. McCall 9

PlasttcityDuring the past 2 decades, some psychol-

ogists have emphasized the malleability of men-tal development dunng the early years of life(eg . Hunt 1961, White 1976) The fact thatindividual differences in IQ stabilize and ge-netic correlates reach asymptote at this timehas led some to suspect that environmental in-fluences would have their greatest impact be-fore 6 or eyen before 3 years of age It wasonly a small step to assume that experiencesprior to this point were permanent and thatsubsequent experiences were less influential oreyen inconsequential In yiew of the presentmodel, this attitude is oyersimplified and pos-sibly misleading There is a difference betweensaying that many or eyen most people do notchange greatly oyer age and saying they arelocked in and that change under any circum-stance IS impossible (see Rutter [1979], for areview)

For example, consider the potential forchange given contemporary, species-typical en-vironments Contrary to the prevailing notion,the scoop approach suggests that the potentialfor change, especially improvement, m mentalperformance is actually greater in early child-hood than in infancy The organism wouldappear most susceptible to variations in envi-ronment m early childhood, after the sides ofthe scoop begin to decline, but before thespeed of the ball increases to the point whereindividuals select their own enyironments andresist or are denied opportunities for improve-ment

Longitudinal correlations —Perhaps wehave been seduced into beheving m the per-manence of mental abihty after age 6 by longi-tudinal stability coeflBcients of 85-95 andby heritabilities of 70-80, figures we maymentally round up to 1 00 Eyen so, neitherdata set should compel us to believe in the im-mutability of mental performance First, theyare both based on individual differences andare therefore potentially mdependent of ayer-age changes m the groups assessed (as hasbeen illustrated by the Skodak and Skeelsdata) But, in addition, eyen year-to-year cor-relations of 90 permit substantial, gradualchange in relatiye performance for some mdi-yiduals and not others Iiymg in species-typicalenyironments In the Fels Longitudinal Studyof middle-class children m Ohio, the ayeragechild showed a tested IQ range of 28 5 IQpoints between 2« and 17 years of age, and

one m seyen children displayed changes of 40points or more (McCall et al 1973) Increasesof more than 70 points are not unknown(Hmdley & Owen 1978, McCall et al 1973)These shifts in relative performance were notrandom error but rather gradual changes overseyeral years with inflection points at 5—7 yearsand approximately 10 years Such deyelopmen-tal profiles were not more similar among sib-lings than among unrelated children matchedfor year of birth, sex, and SES (McCall 1970,McCall et al 1973) Therefore, while approxi-mately half the Fels sample displayed rela-tiyely unchanging IQ pattems oyer age, per-haps because of the stability of their environ-ments, the other children recorded rather sub-stantial, meaningful, and presumably environ-mentally based changes m relative mental per-formance in species-typical environments, de-spite the high year-to-year correlations for thetotal sample

Similarly, hentabihties of 70-80 do notobviate substantial enyironmental changes eyenwithm such samples Jensen (1973) has point-ed out that the enyironmental standard devi-ation for IQ will be 6-9 IQ points, and thatthe 21-point difference between the means ofthe adopted children and their biological par-ents in the Skodak and Skeels (1949) data isquite consistent with a hentabihty of 70- 80and with the presumed improvement in homeenvironment provided by their adoptive par-ents

Longitudinal correlations and hentabihtiesproyide some information about the effective-ness of differences m environments representedin the samples studied But the environmentsnot represented in those samples also have im-plications for the potential to change Clarkeand Clarke (1976) have recently reviewedsome unusual cases of children subjected tomiserably restrictiye circumstances for the firstyears of life Of course, these naturalistictragedies do not afford the expenmental ngorwe would like (Sroufe 1977), and not all chil-dren recoyer completely from such debilitatingcircumstances But that some do oyercomeseyere deprivations during the first 6 years oflife indicates that the potential for massivechange under some circumstances for somechildren and across some portions of the men-tal scale IS possible dunng childhood (Rutter1979)

Factors that produce change —But whatenvironmental circumstances contribute to

10 Child Development

change-* What shall we look for? Because ofthe wide range of environments in the popu-lation and their correlation with mental testscores, we have traditionally concentrated ourattention on between-family environmental fac-tors ( eg , general intellectual climate of thehome, opportunities for education and enrich-ment, general language environment, encour-agement for mental accomphshment, etc ) Noquestion, these factors have an effect

But we have ignored another sphere ofpotential environmental influence, that is,within-family or developmental environmentalfactors—an odd omission for our subdisciplineIn a group of subjects assessed at any one age,the environmental vanance can be conceptuallypartitioned into between-family and within-lndividual components If the hentabihty isassumed to be 70- 80, then the standard de-viation associated with environmental factorswill be approximately 6-9 IQ points Butlongitudinal studies of children remaining mtheir middle-class families suggest an mtra-mdiyidual standard deyiation between 2/s and17 years of age of 7 8 IQ points, and the formof this yanation is not more similar among sib-hngs than among unrelated children (McCallet al 1973) While it is impossible to conclu-siyely attnbute all this lntramdiyidual yanationto enyironmental factors plus error, there isreason to belieye that a good part of it is en-yironmentally based For example, a reyiew ofrecent data on the hentabihty of intelhgencesuggests a more substantial role for withm-family enyironmental factors than previouslysuspected (Plomin & DeFnes 1980), and muchof that yariability may be deyelopmentalchange within individuals

Therefore, a major domain of environ-mental circumstances not often mentioned orinvestigated for its potential impact on mentalperformance includes specific environmentalevents that are "matched" with the child's in-tellectual and motivational disposition at a spe-cific tune For example, a trip to Cape Canav-eral may have a profound influence on themental development of an 8-year-old who hap-pens to be studying airplanes and space andhas a teacher who is a pnvate pilot But thesame experience may haye no influence on thatchild's older brother who is heayJy into basket-ball, girls, and rock music

LtmttattonsThe scoop approach has profound limita-

tions It IS incapable of predicting how muchenvironmental wmd is sufficient to blow an

mdiyidual off the creod at what age It doesnot predict when an enyironmental wind willoyercome the increasing momentum of a roll-ing ball It does not specify the characteristicsof the stages of the species-general functionacross age, and it does not deal with the factsof mental performance among the aged

Another possible problem is that the ap-proach has emerged from and is intended toexplain the literature on the nature of mentaldeyelopment primarily as reflected in IQ scoresIt IS possible that measures of more specificmental abilities would giye quite a differentpicture of enyironmental and genetic corre-lates, deyelopmental stabihty and instability,and the changing degree of canalization oyerage

Further, many of my colleagues have sug-gested a wide variety of changes or additionsto the scoop to help it handle one or anothersjDecific aspect of mental deyelopment It isunderstandably tempting to see how many dif-ferent phenomena can be accommodated by amodel

But I haye not tried to present a modelof mental deyelopment which will make de-cisiye predictions about parameters of mentalperformance which can be forthnghtly testedMy claims for this approach are much moremodest I haye tned to sketch a way of think-ing that has helped me see the relationshipbetween heredity and environment and theirdifferential influences on the developmentalfunction and individual differences Stability ofindividual differences and continuity of devel-opmental function are potentially distinctrealms of deyelopment, and they may be underthe influence of different factors at the sameor different ages While holding these realmsof development conceptually distinct, this ap-proach permits an understanding of how bothare woven into the same fabric of mental com-petence and how they change over age Thisconceptualization has also shown me how itIS possible for nature to hold its young closeto a single developmental creod wnile individ-ual differences are unstable and relabvely un-correlated with both enyironmental and genebccircumstances It helps me appreciate thatstrong early mfluences may have endunngeffects because of the branching nature of thecreod and the dynamic character of mdividuals

rrating on their own environments But onother hand, the potential for change is

nevertheless present, perhaps more stron^y mchildhood than in infancy, at least under cer-

tain circumstances It stimulates me to searchfor more specific lntramdividual environmentalfactors that are matched with an mdividual'sabilities and motivations

As IS often true m the growth of knowl-edge, what seems like a clarification may actu-ally comphcate our pursuit of understandingBut the job of describing and explaining thedeyelopment of mentality is a fundamental re-sponsibility of our discipline If nature hasmade it more complicated than we thought,then let us be about our business with renewedvigor

ReferencesAnastasi, A Heredity, environment, and the ques-

tion "how''" Psychological Review, 1958, 65,197-208

Beckwith, L Relationships between attnbutes ofmothers and their infants IQ scores ChildDe\.elopment, 1971, 42, 1083-1097

Bloom, B S Stability and change m human char-acteristics New York Wiley, 1964

Casler, L Maternal intelligence and institutional-ized children s developmental quotients a cor-relational study Development^ Psychologif,1976, 12, 64-67

Clarke, A M , & Clarke, A D B Early experiencemyth and evidence London Open Books,1976

Cronbach, L J The two disciplines of scientificpsychology Amencan Psychologist, 1957, 12,671-684

Emmench, W Continuity and stability m early so-cial development Child Development, 1964,35, 311-332

Fischer, K W A theory of cogmtwe developmentthe control and construction of hierarchies ofskills Psychological Review, 1980, 87, 477-531

Gruen, G E , & Doherty, J A constnictivist viewof a major developmental shift m early child-hood In I C Uzgins & F Weizmann (Eds ),The structuring of experience New YorkPlenum, 1977

Hindley, C B , & Owen, C F The extent of indi-vidual changes in IQ for ages between 6months and 17 years, in a Bntish longitudinalsample Journal of Child Psychology and Psy-chiatry, 1978, 19, 329-350

Honzik, M P Developmental studies of parent-child resemblance m mtelhgence Child Devel-opment, 1957, 28, 215-228

Honzik, M P Value and limitations of mfant testsan overview In M Lewis (Ed ), Ongtns ofintelligence New York Plenum, 1976.

Rohert B. McCall 11

Horn, J L Organization of abilities and the devel-opment of lntelhgence Psychological Review,1968, 75, 242-259

Hunt, J McV Intelligence and experience NewYork Ronald, 1961

Hunt, J V Environmental nsk in fetal and neo-natal life and measured infant intelligence InM Lewis (Ed ), Origins of intelligence NewYork Plenum, 1976

Jensen, A R Let's understand Skodak and Skeels,finally Educational Psychologtst, 1973, 10,30-35

King, M C, & Wilson, A C Evolution at twolevels m humans and chimpanzees Science,1975, 188, 107-116,

McCall, R B IQ pattern oyer age compansonsamong siblings and parent-child pairs Science,1970, 170, 644-648

McCall, R B Similarity m developmental profileamong related pairs of human infants Science,1972, 178, 1004-1005

McCall, R B Challenges to a science of develop-mental psychology Child Development, 1977,48, 333-344 (a)

McCall, R B Childhood IQ's as predictors of adulteducational and occupational status Science,1977, 197, 482-483 (b)

McCall, R B Quahtative transitions in behavioraldevelopment in the first three years In M HBomstein & W Kessen (Eds ), Psychologicaldevelopment from infaney HiUsdale, N JErlbaum, 1979 (a)

McCall, R B The development of intellectual func-tioning in mfancy and the prediction of laterIQ In J D Osofsky (Ed ), Handbook of in-fant development New York Wiley, 1979(b)

McCall, R B , Appelbaum, M I , & Hogarty, P SDevelopmental changes in mental performanceMonographs of the Society for Research tnChild Development, 1973, 38(3, Serial No150)

McCall, R B , Eichom, D H , & Hogarty, P STransitions in early mental developmentMonographs of the Society for Research tnChild Development, 1977, 42(3, Serial No171)

McCall, R B , Hogarty, P S , & Hurlburt, N Tran-sitions m mfant sensorimotor development andthe prediction of childhood IQ American Psy-chologtst, 1972, 27, 728-748

McCleam, C E , & DeFnes, ] C Introduction tobehavtor genetics San Francisco W H Free-man, 1973

Piaget, J The ongtns of intelligence tn chtldrenNew York Intemational Universities Press,1966 (Onginally published 1954.)

12 Child Development

Plomin, R, & DeFnes, J C Genetics and lnteUi-gence recent data Intelbgence, 1980, 4, 15-24

Plomm, R, & Kuse, A R Genetic differences be-tween humans and chimps and among humansAmerican Psychologist, 1979, 34, 188-190

Plomm, R, & Rowe, D C Genetic and environ-mental etiology of social behavior m infancyDevelopmental Psychology, 1979, 15, 62-72

Rutter, M Maternal depnvation, 1972-1978 newfindings, new concepts, new approaches ChtldDevelopment, 1979, 50, 283-305

Sameroff, A J , & Chandler, M J Reproductive nskand the contmuum of caretaking casualty InF D Horowitz (Ed ), Review of chdd devel-opment research Vol 4. Ghicago Universityof Chicago Press, 1975

Scarr-Salapatek, S An evolutionary perspective oninfant lntelhgence species patterns and indi-vidual vanations In M Lewis (Ed ), Ongifwof tnteUtgence New York Plenum, 1976

Skodak, M , & Skeek, H M A final follow-up studvof 100 adopted children Joumal of GeneticPsychology, 1949, 75, 85-125

Sroufe, L A Early experience evidence and mythContemporary Psychology, 1977, 22, 878-880

Uzgins, I C Orgamzabon of sensonmotor inteUi-gence In M Lewis (Ed ), Ongtfw of intelli-gence New York Plenum, 1976

Uzgiris, I C Commentary In R B McCall, D HEichom, & P S Hogarty Transitions m earlymental development Monographs of the Soci-ety for Research in Child Development, 1977,42(3, Senal No 171)

Waddmgton, C H The strategy of the genes Lon-don Allen & Son, 1957

Washbum, S L Human behavior and the behaviorof other animals American Psychologist, 1978,33, 405-418

White, B L The first three years of life Engle-wood Chffs, N J Prentice-Hall, 1976

White, S H Evidence for a hierarchical arrange-ment of leammg processes In L P Lipsitt &C C Spiker (Eds ), Advances tn chdd devel-opment and behavior Vol 2. New York Aca-demic Press, 1965

Wilson, R S Synchronies m mental developmentan epigenetic perspective Sctence, 1978, 202,939-948

Wohlwill, J F The study of behavioral develop-ment New York Academic Press, 1973