schlinger ap

7/28/2019 Schlinger AP

1/15

Theory in Behavior AnalysisAn Application to Child Development

Henry D. Schlinger, Jr. Western New England College

Theory in behavior analysis is described, and its appli-cation to developmental psychology is illustrated usingexamples from the literature on the development of infantmemory and perceptual size constancy. Traditional ap-proaches to development are briefly described and jux-taposed with a behavioral approach. Behavior-analytictheory is evaluated according to several criteria of goodscientific theories. It is concluded that modern behavior-analytic theory, which grew largely out of the work ofB. F. Skinner, succeeds well at explaining many of thefacts of infant memory and perceptual size constancy andmay also succeed at explaining facts in other areas ofdevelopmental psychology. Behavior analysis may thus beable to provide som e theor etical unifor mity in the field ofdevelopmental psychology.

The General ProblemIn the past several decades, the field of developmentalpsychology has grown and generated a vast amount ofresearch. Despite this growth, the field remains theoret-ically fractionated, with very little of its research unitedby a common theoretical orientation. Not surprisingly,textbo oks in the field reflect this state of affairs. A lthoughmany developmental textbooks are written from a gen-erally cognitive perspective, none adopts a unitary theo-retical approach. Based simply on the amount of infor-mation in the field, these textbooks give the impressionthat developmental psychologists know a lot about thebehavioral changes that constitute development. On theother hand, the am ou nt of contrasting evidence for man yphenom ena and the num erous explanations and theoriesseem to suggest that developmental psychologists are un-able to make much sense of this information. They maybe able to tell us what children are likely to do and atapproxim ately what age but not how or why they are ableto do it. At the very least, developmental psychology is afield with the appearance of confusion. Unlike the naturalsciences, it seems to be accepted practice to assume thatmany different theories and explanations are needed toaccount for essentially the same subject matter, in thiscase, the behavioral changes that are observed over timein the life of an individual. This general practice resultsin the existence of a large body of factual inform ationbut one lacking theoretical consistency or unity. Studen tsof developm ental psychology may learn a lot of individualfacts, and they may be better able to predict the average

age at which certain behavioral changes emerge, but theyare not given the theoretical tools with which to under-stand or explain these changes within a unified frame-work. In addition, those who work with children in ap-plied settings gain little practical knowledge that can beused to reliably change behavior. Behavior analysts whoare interested in development find this state of affairstroubling.Although the theoretical systems within develop-mental psychology are largely unsatisfactory, at least forbehavior analysts, some of the research may be valuableand therefore worthy of consideration. However, we neednot throw the baby out with the bath water. Behavioranalysts ought to ask whether it is possible to make senseof this accumulation of apparently unrelated data withoutnecessarily accepting the accompanying explanations andtheories. Over the years, there have been some attemptsby behavior analysts to apply behavior-analytic theory tochild development (e.g., Bijou, 1976; Bijou & Baer, 1978).However, behavior analysts by and large have not ad-dressed the extensive research in the area conductedlargely by psychologists who are not behavior analystsalthough there are some exceptions (e.g., Gewirtz, 1972a,1972b; Gewirtz & Pelaez-Nogueras, 1990, 1991). Becausemuch of the existing research in developmental psychol-ogy suggests a strong environmental component to be-havioral development, behavior-analytic theory is in anideal position to interpret the behavioral changes that de-fine this development.

Although it is possible to show that behavior-analytictheory succeeds as a theory of behavioral development,its treatment in most developmental textbooks indicatesthat it is regarded as being relatively unimportant for theunderstanding of development. In some cases behavior-analytic theory is relegated to an almost historical placein developmental psychology (e.g., Scarr, Weinberg, &Levine, 1986). In other books, it is never credited withbeing able to explain anyth ing m ore than trivial behaviors(e.g., Collins & K uczaj, 1991). Still other textboo ks offerrather lengthy sections on learning (not explicitly behav-ior-analytic theory) in which th e principles of respondentand operant conditioning are detailed; however, subse-quent discussions of behavioral changes attributed toThe au thor is grateful to Edward K. Mo rris and David C. Palmer fortheir helpful comments on earlier versions of this article.Correspondence concerning this article should be addressed toHenry D. Schlinger, Jr., Department of Psychology, Western New EnglandCollege, 1215 Wilbraham Rd., Springfield, MA 01119.

1396 November 1992 American PsychologistCopyright 1992 by the American Psychological Association, Inc. 0003-066X/92/S2 00Vol.47, No. II, 1396-1410


2/15

learning rarely refer back to these principles (e.g., Dwor-etzky, 1991). In man y textboo ks the discussion of learn-ing, not to mention behavior-analytic theory, is almostnonexistent (e.g., Collins & Kuczaj, 1991; Santrock,1990). Finally, many standard textbooks in developmentalpsychology omit the significant empirical and theoreticalcontributions of behavior analysts to the scientific un-derstanding of child behavior. This sizable literature in-cludes, but is not limited to, notable work by Donald M.Baer, Sidney W. Bijou, Jac ob L. Gew irtz, Lewis P. Lipsitt,and Carolyn Rovee-Collier and their respective colleagues.At the very least, this omission does a disservice to stu-dents of developmental psychology by depriving them ofvaluable empirical a nd theoretical analyses of behavioraldevelopment in children.

In the present article, I clarify a behavior-analyticposition on scientific theory and then demonstrate howbehavior-analytic theory can be used to interpret certainfacts in developmental psychology. Throughout thisarticle, a behavior-analytic theory of development iscounterposed with traditional structural theoretical ap-proaches. The purpose of the article is to show that in-terpreting developmental research according to behavior-analytic theory can provide theoretical cohesiveness todevelopmental psychology.Conception of Behavior in DevelopmentalPsychologyIt has been standard practice in psychology in general,and developmental psychology in particular, to differen-tiate behavior largely according to its form. In develop-mental psychology, this practice has resulted in such pre-sumably distinct categories as motor behavior, social be-havior, emotion al b ehavior (emotion), cognitive behavior(cognition), perceptual behavior (perception), languagebehavior, and so on. Even within these categories struc-tural distinctions are made. For example, cognitive-de-velopmental psychologists distinguish among imitation,conservation, and object constancy as if each of thesebehavioral classes was a different creature requiring itsown distinct explanatory mechanism. The assumption istha t each of these types of behav ior are fundam entallyand functionally different.

Once behavior has been distinguished largely on thebasis of structural characteristics, it becomes easier toassum e that different theories or explanations are requiredfor the developm ent of each type of behavior. For exam ple,Piaget's theory in tends to explain cognitive d evelopment,Freud's theory intends to explain social development,both Piaget and Kohlberg have theories of moral devel-opment, and there are different theories of emotional de-velopment, language development, and so on. Even withinsingle theoretical systems, structurally differentiated be-haviors are afforded different explanations. In Piaget'stheory, for example, there are different mental structuresto explain object permanence, conservation, egocentrism,and so on. In fact, the entire field of child developmentimplies a domain that is functionally distinct from adultpsychology, as if each w ere governed by different psycho -

logical principles. Students often wonder what to makeof this theoretical eclecticism, which is often presentedin developmental psychology textbooks as if it were astrength.The im plications of a structural ap proach in devel-opmental psychology cannot be overstated. It assumesthat classes of behavior, distinguished primarily on thebasis of structural properties, are separate functionalclasses requiring different explanatory mechan isms. Un-fortunately, all too often, the theories and explanationsused to account for the development of these behaviorsare arrived at by way of logical error tha t is often circularin n ature . First, the behavioral class is given a n am e. T hename is then reified and itself becomes the object of study.Finally, the name is used as the explanation of the ob-served behavior. Much of the Piagetian approach to de-velopment illustrates this. For example, by about 24months, infants are able to behave effectively toward ob-jects under a wide range of conditions of their absence.Piaget and other developm ental psychologists have term edthese observed behav ior-enviro nm ent interactions objectpermanence. Wh en studying the development of the phe-nomenon, it is said that they are studying object per-man ence, or the em ergence of the concept of object per-manence, which implies that what is being studied issomething other than the observed behavior-environmentrelations. Then, when infants are finally able to behaveappropriately with respect to objects that are out of sight,it is said they can do so because they now possess th econcept of object permanence or the mental capacity torepresent objects. Object permanence, which began asthe name for certain observed behavioral relations, hasbecome a thing (a structure or process) located inside thechild that is said to be responsible for the observed be-havior. Although the logical errors in this type of theo-rizing have been noted by psychologists who are not be-havior analysts (e.g., Dworetzky, 1991; Flannagan, 1991),it has done little to diminish such practices.

In addition to the number of different theories, de-velopmental psychology is also noteworthy for the largenum ber of facts researchers have collected. Unfortunately,man y of the facts have been generated by nonex perimen-tal, mostly correlational, research, and therefore they donot represent the types of basic facts that characterize anatural science approach, namely, functional relationsbetween controlled variables. Also, because much of thecorrelational research involves the child's age as one ofthe variables, it is often implied that functional relationshave been uncovered when they have not. In such ac-counts, the independ ent v ariable is said to be the child'sage or some invented cognitive structure that developsand is conveniently said to be age specific. For example,Piagetian psychologists describe the developmen t of objectperm anence durin g several substages of the sensorimotorperiod. Each change in children's behavior with respectto hidden or displaced objects is correlated with changesin their ages. Th us , during Stages 1 and 2 (1 to 4 months),objects that are out of sight are literally out of mind; inother words, infants will not continue to behave toward

November 1992 American Psychologist 1397


3/15

such objects. However, by Stage 6 (18 to 24 months) in-fants will continue to respond to hidden objects.Even though age is the variable correlated with thebehavioral changes, the real independent variable is saidto be a developing mental structure that enables the childto respond to hidden objects. Because the mental struc-ture presum ably develops maturationally, age is naturallya logical reflection of changes in it. Such structural nor-mative approaches form the basis for man y of the major"theories" in the area of human development and haveencouraged the invention of stage theories of behaviorchange in developmental psychology. Correlational re-search does allow for a modicum of prediction, but theactual control of the relevant variables that is the h allmarkof a natural science approach is absent and so too is sci-entific prediction and understanding derived from thiscontrol. Even when true experimental research is con-ducted in developmental psychology, the resulting dataare rarely interpreted according to any particular theory.Instead, the data are left without interpretation or theyare simply described using inferential circular explana-tions. For example, moral behavior is said to result frommoral understanding, attachment behaviors are said toresult from emotional bonds between child and parent,and other behaviors, called cognitive, are said to resultfrom mental processes such as concepts, memories, orperceptions. Finally, the purpose of much correlationalnormative research seems to be nothing more than dem-onstrating the presence or absence of some behavior inchildren and the approximate corresponding ages but notho w the behavior came to be.

The concept of stages and structures in develop-men tal psychology is not totally without merit; it has aidedin the organization of data and in communication be-tween researchers (Lipsitt, 1981). However, the problemsgenerated by structural normative approaches far out-weigh these benefits. Lipsitt listed several of these prob-lems. First, structural approache s to behavior posit struc-tures that are not directly observable. Classic examplesare Freud's ego, super-ego, and id; Piaget's cognitivestructures, schemata, concepts, and rules; and the input,storage, and retrieval systems of the information-pro-cessing model. A second problem noted by Lipsitt is thatthe emphasis on hypothetical structures and processeshas tended to obfuscate other, perhaps mo re central, un-derlying behavioral processes. Third, as I have alreadynoted, structural normative approaches tend to confusedescription with explanation. For example, according toLipsitt, the transitional natu re of behavior change, whichis an important behavioral process in its own right, isoverlooked when norm ative descriptions lead us to believethat we have explained a transition by saying, for examp le,that the child is now in "the preoperational stage." Afourth problem with structural normative orientations isthat inferred "constitutional-maturational determinants"of behavior change are emphasized at the expense of mo redirect environmental or physiological variables. A fifthproblem is that structu ral concepts that em phasize stagesreduce optimism ab out the potential of hum ans to benefit

from environmental manipulations, especially those ineducation. If behavioral development is indeed con-strained by the stagewise development of underlyingstructures and processes, then, in principle, it should beimpossible to accelerate that d evelopmen t. However, thereis sufficient evidence to force serious reconsideration ofthis interpretation of development so much so that theconcept of maturational stages as important determinantsof behavior is seriously threatened (see Dworetzky, 1991,pp. 262-266; Lipsitt, 1981). Finally, structural normativeinterpretations easily fall into the logical verbal trap ofreification. According to Lipsitt,The em pirical hazard is that we come to regard those stages asreal conditions of the organism rather than as artifacts of ourobservational procedures and methodologies. Conceptualiza-tions of development in terms of stages are usually followedclosely by the adoption of a structural view of the mind. Thepostulation of structures is usually based upon behavioral ob-servations, to be sure, but the language quickly becomes met-aphorical. The special words, initially devised merely to abbre-viate complex behavior patterns, now become taskmasters andslaves, (pp. 31 -32)The problem of reification leads to another trouble-some verbal practice in developmental psychology. Lo-cating the causes of behavior inside the child makes iteasier to describe the child as the originator of his or heractions. Thus, after contact with environmental stimu-lation, the child is said to "sense," "perceive," "remem-ber," "think," "judge," "decide," and so on as if theseverbs referred to real actions. However, once describedin this way, behavior can be more easily attributed tothese presumably internal activities. The most conspic-uous examples of this are found in cases of clear-cut op-erant or respondent conditioning. For example, accordingto Kalnins and Bruner (1973), conditioning studies withinfants provide evidence that infants are not only able"to anticipate an outcom e," but they are able to "cho ose"and "deploy" "means for achieving it." Moreover, infantsare said to have the "ability to correct and coordinatebehaviour in a fashion that would lead it more directlyto the goal" (p. 307). However, what are the actions towhich the verbs anticipate, choose, deploy, correct, an dcoordinate refer? Although these are the w ords of comm onusage and are easily understood by everyone, they divertattention away from the controlling variables of the be-haviors in question. In their study, Kalnins and Brunershowed with infants that when sucking a nonnutritivenipple at a certain rate produced a focused picture, suck-ing rates increased, and, conversely, when a blurred pic-ture was produced, sucking rates decreased. The facts areclear enough and the interpretation is simple enough:Seeing a focused picture strengthened the sucking ratesthat produced the focused picture; the process is calledreinforcement. Conversely, the process of weakeningsucking rates by presenting a blurred picture is calledpunishment. Saying that the child is "able to carry outskilled, voluntary-controlled, adaptive behavior that leadshim to goals" adds nothing to the facts and diverts controlof sucking away from the environment (i.e., the clarity

1398 November 1992 American Psychologist


4/15

of the picture) to some hy pothetical agent inside the child.Again, although such language seems benign enough, itis not completely objective, it states more than the factsallow, and it adds little to the prediction and nothing tothe control of the behavior. In a word, it is unnecessary.Moreover, it reduces the likelihood that the true range ofenvironm ental d etermin ants of the behavior will be iden-tified. The conclusion from this brief presentation is thatthe general theoretical approach found in developmentalpsychology does not succeed well at understanding be-havioral develop men t. In this article, I am suggesting tha tbehavior-analytic theory can do better.Theory in Behavior AnalysisSkinner on TheoryThe role of theory in behavior analysis has been a subjectof debate in psychology especially since B. F. Skinner'swell-known 1950 article, "Are Theories of LearningNecessary?" Misunderstandings of Skinner's positionhave been widespread; Skinner has been portrayed asbeing antitheory and behaviorism as being atheoretical.In his often cited article, "Are Theories of Learning Nec-essary?," Skinner argued against only certain types oftheories: namely, those that appealed to inferred internalconstructs as explanations of behavior. For Skinner, suchtheories and the hypothetico-deductive methods used toformulate them were actually detrimental for a scienceof behav ior because they represen ted a diversion from th eanalysis of real behavior in th e real world. Scientists usingthis approach would be more concerned with testingtheorems than actually searching for orderly relations asis the practice in the natural sciences. As for the title ofthe article, Skinner simply meant that he and others hadalready begun to discover orderly behavioral relations us-ing a research program "with a minimum of theoreticalconstructs and in the absence of hypothetico-deductivemethods" (Zuriff, 1985, p. 88), and thus exp lanations atother levels of analysis did not seem necessary. Therefore,Skinner was not opposed to all theories in psychology,only those with explanations that appealed to inferredevents when more objective events were available.

Although critics apparently reacted against Skinn er'santitheory stance, it is more likely that what they reallyobjected to were his interpretations. In other words, criticsdid not like Skinner's behavioristic theory (e.g., Baars,1986). However, there may have also been misunder-standings about Skinner's views on theory in science andpsychology. Let us, therefore, look at how Skinner usedthe term theory to define his own approach. Before doingso , it might be instructive to examine some of the defi-nitions of theory found in the dictionary. According toWebster's Seventh New Collegiate Dictionary (1965), the-ory is (a) "the analysis of a set of facts in their relationto one another;" (b) "the general or abstract principlesof a body of fact" or a "science;" or (c) "a plausible orscientifically acceptable general principle or body ofprinciples offered to explain phenomena." To paraphrase,a theory includes some already established body of fact,

abstract principles inductively derived therefrom, and theuse of such principles to explain new facts.Skinner frequently discussed his position on scien-tific method and theory (e.g., 1938, 1947, 1950, 1956).However, Zuriff(1985, p. 89) sum marized Skinner's (andthe behavior-analytic) position. According to Zuriff,Skinnerian theory consists of concepts that express "em -pirical functional relationships between behavioral andenvironmental variables." More specifically,Theory, in the acceptable sense, evolves in the attempt to presentthe collection of empirical fact in a formal and economical way.A formulation using a minimal number of terms to representa large number of experimental facts is a theory. As the theorydevelops, it integrates more facts in increasingly more econom-ical formulations. Theoretical concepts thus merely collate ob-servations and do no t refer to nonbehavioral processes. A Skin-nerian theory is, therefore, a simple, comprehensive, and abstractdescription of a corpus of data. (p. 89)

As can be seen from Zuriff's description, scientifictheory for Skinner essentially mirrors the statementsfound in the dictionary: namely, a collection of empiricalfacts (specifically functional relations) and the formulationof principles derived inductively from those facts with aminimal number of terms and concepts. Some mightsuggest that Skinner's approach is nothing more thanrandom fact gathering; however, this would be incorrect.In elucidating his position on the relation between factand theory, Skinner (1947) wrote,A theory . . . has nothing to do with the presence or absenceof experimental confirmation. Facts and theories do not standin opposition to each other. The relation, rather, is this: theoriesare based upon facts; they are statements about organizationsof facts. The atomic theory, the kinetic theory of gasses, thetheory of evolution and the theory of the gene are examples ofreputable and useful scientific theories. They are all statementsabout facts, and w ith proper operational care they need be noth-ing more than that. But they have a generality which transcendsparticular facts and gives them a wider usefulness. Every scienceeventually reaches the stage of theory in this sense, (p. 28)

What is implied here, and in fact is stated directlyby Skinner elsewhere, is that these theoretical principlesshould then be used to interpret (i.e., understand) novelempirical observations (e.g., Skinner, 1957). The inter-pretation of novel behavioral relations using establishedprinciples has been a consistent practice in behavioranalysis for man y years (e.g., Bijou, 1976; Bijou & Baer,1978; Cata nia, 1984; Palmer, 1991 ; Skinner, 1957;Whaley & Malott, 1973). It is part of a long tradition inthe na tural sciences of using theoretical principles indu cedfrom scientific facts to understand novel phenomena thatare usually more complex than what can be studied inthe laboratory. According to Palmer,Interpretation has served, and con tinues to serve, an honorablerole in science, so honorable that we often fail to distinguishbetween an interpretation and an experimental analysis. New-ton's explanation of ocean tides is an interpretation based onhis experimental analysis of phenomena such as the motion ofpendulum s and colliding balls of wool, glass and cork. No one,least of all, Newton, has attempted to establish experimental



5/15

control over the tides. Yet Newton's principles (to a reasonableapproximation) are sofirmlyestablished and the extrapolationto this phenomenon so plausible, that we accept his interpre-tation as if it were the direct outcome of an experimental analysis,(p. 261)The key to a plausible interpretation is that it proceedsfrom a foundation of well-established principles derivedfrom a rigorous program of basic research. This is theadvantage that behavior analysis brings to the task of un-derstanding that part of hu ma n behavior that has not yetbeen directly subjected to an experimental analysis.In general then, behavior-analytic theory is con-cerned with lawful relations between observable events,namely, behavior and its environmental determinants.Consequently, the theory is maximally descriptive andminimally inferential. This is one reason why behavioranalysts are so strict (some would say picky) about theiruse of language in describing behavioral events. Specifi-cally, behavior-analytic theory is composed of the lawsand principles, derived from the experimental analysis ofbasic un its of behavior, tha t describe known functionalrelations between behavior and environment. Behavior-analytic theory includes, but is not limited to, the lawsof respondent (Pavlovian) conditioning as well as the op-erant laws of reinforcement (and punishment) and stim-ulus control, and all of their related principles and re-spective parameters.

Basic Units in Behavior AnalysisThe key to the emergence of laws of behavior has beenthe discovery of functional analytic units of behavior. Th esuccess of any science depends on whether that sciencehas discovered "prop er u nit s" of analysis. As Zeiler (1986)noted,A well-defined unit clarifies the way phenomena are concep-tualized and thereby guides research and theory. Isolation of aunit brings order to otherwise discrepant phenomena; invalidunits easily lead to confusion as to the meaning and significanceof the data. (p. 1)

The generic units of modern behavior analysis werefirst described by Sk inner (1935 ). Skin ner defined stim uliand responses not as independent structural units but asfunctional classes. Thu s, stimuli were not defined by theirphysical energy and responses were not defined by theirtopography. Rather, both were defined functionally, thatis, by their respective effects on one another, and, as aresult, funda men tal un its of the analysis of behavior werediscovered. Several behavior analysts have described theimportance of this discovery for a science of behavior(e.g., Branch, 1977; Cat ania , 1973; Zeiler, 19 86). Ac-cording to Zeiler,The fundamental units (operants, respondents, discriminativeoperants) are the smallest entities that display the full charac-teristics of adaptive behavior. The previous structural entities(stimuli and responses in isolation) now become componentsof the basic units, analogous to nuclei and cytoplasm as com-ponents of cells. Research can involve the variables determininghow generic classes are constructed and the factors responsible

for particular forms of coordinated behavior, but never is itnecessary to move to a nonbehavioral level of analysis, (pp. 4-5)This last statement is, of course, Skinner's (and the be-havior-analytic) position on the role of explanation (ortheory) in the study of behavior. In the present article, Iargue that it is because of these fundamental units thatbehavior analysis can bring order and clarity to devel-opmental psychology, which, borrowing from the earlierquo te by Zeiler (1986), can be characterized as possessing"discrepant phenomena" with "invalid units," all ofwhich produce "confusion as to the meaning and signif-icance of the data." Just as the discovery of the cell, as abasic unit of biology, integrated the d isciplines of anatomy ,embryology, botany, and zoology into the unified fieldthat we now call the biological sciences (see Zeiler, 1986),so too can the basic units of behavior analysis help tounify the various topical subareas of developmental psy-chology.

The concept of basic units of analysis has implica-tions for the distinction between stru ctural and functionalapproaches to behavior. Behavior analysts do not denythat behavioral units can have structure. Rather, they as-sert that th e units m ust be functionally defined. As Branch(1977) put it,The crux of the issue is not whether units can have structure;they do. The important question is, to what is the structure due?A structural account points to structural aspects of the behaver. . . whereas a functional account will emphasize the role ofmanipulable variables in the formation of the units, (p. 172)For behavior analysis, the "manipulable variables" re-sponsible for the formation of units of behavior are to befound largely in the individual's environment.Behavior-Analytic View of the EnvironmentThe concept of environment is a common one in psy-chology; however, the conception of environment in be-havior analysis differs from that found in other psycho-logical approaches; therefore, it warrants clarification.Most standard views of the environment are not onlystructural in nature but molar as well. In child develop-mental psychology, the environment is typically definedby people (e.g., parents, teachers, friends) and institutions(e.g., schools, comm uni ties; Collins & Kuczaj, 1991) oras the experiences a child has (Scarr et al., 1986). Forexample, in the debate about the effects of early adverseexperiences, the environment is described as either beingenrich ed or deprived o r restricted (e.g., Collins & Kuczaj,1991). With regard to social development, the context issaid to be im porta nt. T he context of social developmentusually is said to include the family, its socioeconomicstatus, the number of siblings, parents' employment sta-tus, and so on. Although allusions to specific interactionsare sometimes m ade, the environm ent usually is viewedas a structural en tity with properties that can be quantifiedand thus included in correlational calculations with thechild's age or developmental progress. These conceptionsof environment may aid in conducting correlationalstudies, but they fall prey to some of the same problems



6/15

described earlier concerning structural conceptions ofbehavior: Namely, they tend to obscure more fundamen-tal and molecular environmental variables, and they be-come easily involved in circular explanations. Finally,standard psychological conceptions follow from the layvernacular that the environment is something that sur-rounds us.The behavior-analytic view of the environment is,by contrast, a functional one. Behavior analysts view theenviron men t as consisting of energy changes (i.e., stimuli)of various sorts that affect not only the sensory receptorsof organisms but also their behavior. Thus, stimuli, aswell as other events (e.g., motivational operations), andbehavior can enter into orderly functional relations witheach other. These functional relations form th e basic unitsof behavior from which the laws and principles of behavioranalysis have been induced. A functional conception ofthe environment implies, among other things, that thedepiction of the environment as something that surroundsus is incomplete at best. If, in fact, the environment isdenned functionally as all of the events that enter into

functional relations with an organism's behavior at anyone time, then we must ex pand our view of the environ-ment to include events both inside as well as outside thebody. Hence, the skin is no longer an arbitrary stru cturalboun dary, and the env ironmen t may now be viewed froma thoroughly functional perspective.Behavior-Analytic View of Locus of Co ntrolAny discussion of environment raises the problem ofwhere the determinan ts of behavior lie. Th e issue of locusof control is central in behavior analysis and is one of thefeatures that most distinguishes it from other psycholog-ical approaches to behavior. Previously, I described howsome developmental researchers interpreted a relativelysimple demonstration of operant conditioning in infantsby placing the determ iners of the behavior inside the child.In this typical account of behavior, the child is exposedto an environm ent bu t it is the child who "atte nds ," "de-cides," "chooses," and, thus, "behaves." It is not sur-prising that such interactions are described as voluntarybehavio r in wh ich th e child is seen as the agent (i.e., orig-inator or initiator) of his or her own actions.

Behavior analysts register several objections to suchagent-action accounts of behavior (e.g., Hineline, 1990).The most obvious objection is that, in many cases, thecontrol of behavior is clearly not inside the child but inthe environment. The Kalnins and Bruner (1973) studyis a perfect example. The rates of sucking changed when-ever the experimenters changed the outcome of sucking,that is, the clarity of the picture. Thus, changes in theenvironment and not the child were responsible for thechanges in sucking behavior. It would be as though anexperimenter repeatedly operantly conditioned and thenextinguished a rat's lever press but then attributed thechanges in behavior to the rat 's own decision mak ing an dvoluntary control. At the very least, such descriptions areredundant and add nothing new to the understanding ofbehavior. At worst, they lead researchers away from the

real controlling variables and in search of elusive q ualitiesthat can only be inferred after first manipulating the sub-ject's environment.In short, behavior analysis attributes hu ma n behav-ior to a constellation of variables including the individual'sgenes and physiological make up bu t primarily to the en-vironment, both past and present. Consequently, differ-ences in behavior between individuals are more likely tobe attributed to differences in the functional aspects oftheir environments. When behavior can be unequivocallytraced to environmental interactions, as it can often bein a controlled laboratory, such an account is strength-ened.What follows is an illustration of how behavior-an-alytic theory might be applied to two research areas inchild development that are not typically presented asbeing related. A behavior-analytic interpretation relatesthem by examining them as examples of adaptive be-havior, that is, behavior as a function of environmentalvariables and subject to the same natural laws.Illustrative Examples of DevelopmentalPhenomenaInfant MemoryThe importance of the concept of memory for cognitiveapproaches to psychology cannot be overstated. To theextent that the study of mem ory can not be separated fromthe general study of wha t is called informatio n processing(Cohen & Gelber, 1975), it is possible to view it as theessence of modern cognitive psychology. As a result, avast and complicated theoretical network has emergedarou nd th e concept of mem ory, and th is raises difficultiesfor a behavior-analytic interpretation of infant memory.The main difficulty is that behavior analysis and at leastsome versions of cognitive psychology hold d iametricallyopposed views of reality. According to M cDowell (1991 ),"behavior analysis presupposes a materialistic ontology,according to which the world consists of material objectsand e vents," whereas (cognitive) "psychology presup posesa non materialistic ontology, according to which the worldconsists of non -ma terial objects and ev ents as well as ma -terial objects and ev ents" (p. 29). Hence, trying to m akesense out of concepts such as memory traces, storage lo-cations, and retrieval systems using behavior-analytictheory is like comp aring apples an d oranges. Theoreticalquestions in the memory literature also reveal the diffi-culties behavior analysis has working within a differentframework. So, for example, questions such as "Is thereonly one storage system in which traces are set down?"or is there "a limited capacity short-term memory inwhich items are held for only brief periods of time, anda high-capacity long-term memory in which items arecomplexly coded thereby greatly reducing informationloss du e to forgetting o r retrieval failure?" (Cohen & Gel-ber, 1975, pp. 34 8-349 ) are unanswerable because theproposed theoretical entities are not accessible to empir-ical methods. As I stated earlier, rather than disregardingbehavior because a theory is untestable, the behavior-



7/15

analytic approach attempts to understand the behavioraccording to the established laws and principles of be-havior analysis and to regard the particular inferred theo-retical constructs as unnecessary. I have already describedbehavior analysts' opposition to inferred internal eventsas explanations of observed behavior. Behavior analysts'opposition to inferences, however, is not an absolute oneto all inferences but rather only to those that are unver-ifiable. The task of behavior analysis when confronting apsychology based in large part on inferred internal eventsis to examine the behavior from which the internal eventsare inferred.This may seem inadequate for cognitively orientedpsychologists, and it may be partially responsible for thecriticism that, by concentrating primarily on observablebehavior, behavior analysis ignores what is most inter-esting about human psychology (McDowell, 1991). Theinadequacy, however, may arise from the selection of theparticular behavior not by the behavior analyst but bythe cognitive psychologist. It is this point that raises asecond difficulty for a behavior-analytic interpretation of(infant) memory. Attempting to interpret units dennedby cognitive psychologists means buying into the generalcognitive strategy of identifying units of analysis in ad-vance of their measurem ent "merely on the basis of theirputative structure or topography" (Branch, 1977, p. 171).The concept of memory and all of its components (e.g.,short-term and long-term memory and so on) qualifiesas such a unit. Recall that for cognitive psychology be-havior seems to be important largely insofar as it reflectsinferred internal events. So cognitively oriented psychol-ogists are less likely to define behav ior on the basis of itsfunctional relation to environmental variables than bywhat it suggests to them about underlying structures andprocesses. The behavior analyst who attempts to interpret

data derived from this fundamental assumption, then,does so at some risk. Notwithstanding these difficulties,let us look briefly at how behavior-analytic theory mightapproach some of the facts that comprise the topic ofinfant memory.Standard description of (infant) memory. One de-velopmental textbook defines memory as "a person'smental record of an event" (Scarr et al., 1986, p. 232).This definition brings with it all of the attend ant problem sof structural appro aches to behavior described previously.Briefly, behaviors are observed in certain situations andthey are given a nam e, in this case, memory. T he behav-iors are reified as a "mental record of an event." Themental record is never directly observed. Finally the con-cept of memory or of the "mental record of an event" isused to explain the very behaviors they are said to rep-resenta circular explanation . For exam ple, the relativelypoor memory of preschoolers has been attributed to alack of meaningful organization of their memories, theirlimited information-processing capacity, and an inade-qu ate kn owledge base (Sca rr et al., 1986). Of course, lackof meaningful organization, limited information-pro-cessing capacity, and inadequate knowledge bases are allinferred from the behaviors to be explained in the first

place. Descriptions such as these create surrogates for thechild's behavior, which are then moved inside the childwithout any independent empirical confirmation. Scarret al. (1986) provided two examples in their respectivediscussions of the poor m emo ry organization and limitedinformation-processing capacity of preschoolers:One might compare the storage area of the human mind to avast library, filled with endless shelves of information. To recalla particular experience or retrieve certain information, the childmust search the shelves to find what she needs. If her memorycollection is organized in a meaningful way, if it is cataloged,the search is much easier, (pp. 233-234)a n d ,When faced with a memory task, an older child automaticallybegins grouping item s in his mind , rehearsing the lists of words,and applying other strategies to the task. He has a program orscheme for memorization. The younger child does not havesuch a scheme. She has to th ink a bou t how she is going to learnthe pictures or words. This uses up m ental "spac e," leaving lessroom for actually storing information, (p. 234)The first quote is clearly analogical, but the second quoteimplies the existence of mental structures (schemes) andprocesses. Such descriptions do make communicationeasier; everyone kn ows wh at a library is and wh at a (com-puter) program is, and we are all familiar with searchingthe shelves of a library for a book and with storage "s pace "in computers. In these descriptions, however, things donot happen in the real world of behavior and environmentbut in another world of inferred events called mind orcognition. Although all of this appears relatively benign,it contributes little to the understanding of the behaviorfrom which all of it is inferred. Moreover, concentratingon mem ories and other inferred mental structures directsresearchers away from investigating more fundamentalunderlying processes such as the reinforcement contin-gencies responsible for the observed behaviors ("mem-ory"). A behavioral analysis takes a different app roach .Experimental facts of infant memory. It has beensaid that imitation, object perm anenc e, attachment, con-ditioning, and preference for novel stimuli in infants allimply and require m emo ry (Cohen & Gelber, 1975).Studies of memory in infants are often concerned withwhat is called recognition memory. Visual recognitionmemory is inferred "when infants respond differentiallyto familiar and novel stimuli" (Werner & Siqueland, 1978,p. 79). Investigators have used both habituation and nov-elty preference procedures to test for such recognitionmemory. Consider a typical habituation procedure. Asingle pattern (e.g., consisting of geometrical shapes orpictures of faces) is presented repeatedly until visual fix-ation time decreases, and then in test trials both noveland familiar stimuli are presented. The dependent mea-sure is the difference in fixation time between the twotypes of stimuli. The novelty preference proced ure, whichis a variation on the habituation procedure, involves si-multaneously presenting two identical patterns to infants,one on the left side and the other on the right side. Afterthis has been done repeatedly, a test trial is presented in



8/15

which a novel stimulus is presented with the familiarstimulus. Again, the dependent measure is the proportionof visual fixation time to the novel versus the familiarstimulus (Cohen & Gelber, 1975). Results of both typesof procedures show generally that visual fixation behavioris proportionally longer to the novel stimuli (Miranda &Fantz, 1974). There are at least two questions concerningboth procedures: What is being measured? How shouldthe results be interpreted? The answers provided by tra-ditional developmental psychology and behavior analysisreveal the differences in their respective approaches. Fordevelopmental psychologists, what is purportedly beingmeasured is recognition memory, which is itself inferredfrom the differential fixation times. That is, if there aredifferences in fixation times, recognition memory exists,and if there are n o differences, then there is no recognitionmem ory. The standard explanation of the behavior of theinfants under these tests of "short-term memory" seemsto be the same as the memory itself and, thus, circular.For example, Miranda and Fantz (1974) concluded that"th e infant is able to take in and retain inform ation froma visual stimulus and thereby to differentiate b etween tha tstimulus and an unfamiliar one" (p. 651). However, isthis saying anything mo re than there are different fixationtimes to different stimuli? In fact, the only evidence forthe taking in and retention of information and the dif-ferentiation between stimuli is the very behavior to beexplained: differential fixation times.

A behavior analysis of infant memory. Recall thatbehavior analysts might not conduct this specific type ofresearch, and therefore they are at some disadvantagewhen attempting to interpret the results. Nevertheless,for behavior-analytic theory, the dependent measure inthe habituation and novelty preference studies wouldprobably be the amount of time the infant's eyes fixateon a particular pattern (or, more molecularly, the actualeye movements to the stimuli). The behavior-analytic in-terpretation of the results would very likely be in termsof stimulus control as predicted by the extant experi-men tal literature on habituation rather than appealing toinferred internal events (i.e., recognition memory). Suchtheoretical parameters as rate of habituation, the inter-stimulus interval, generalization, spontaneous recovery,and dishab ituation may be brought to bear. For behavior-analytic theory, it is sufficient to account for infants' be-havior on the basis of what is already known about ha-bituation, the facts of which occupy part of behavior-analytic theory. Of course, habituation may be understoodat another levelthat of underlying cellular processes(e.g., Kandel, 1975)although this does not negate theprinciples of habituation at the level of behavior and en-vironment. None of this, however, requires appealing toinferred internal entities or processes. Incidentally, ha-bituation has been studied in many other species, in-cluding those with relatively simple nerv ous systems, suchas the marine invertebrate Aplysia (see Kandel, 1975),without inferring hy pothetical cognitive events to ac coun tfor the facts. It is also possible th at these so-called stu diesof habituation are not really studies of habituation at all

but rather studies of operant stimulus control. That is,the actual m ovem ents of the eyes described by the phrase"visual fixation" may be functionally related to the con-sequence of seeing the stimulus display. This might beespecially true in infants because it is known that manytypes of visual stimuli seem to function either as (uncon-ditioned) positive reinforcers for the behavior that pro-duces them (e.g., Kalnins & Bruner, 1973; Siqueland &DeLucia, 1969) or as unconditional eliciting stimuli(Fantz, 1961). In any case, a behavior analysis would beable to tease apart the real controlling variables by sys-tematically examining the functional relations betweenthem.

Interestingly, especially for the pre sent article, severalexperiments have investigated visual memory or retentionin infants using operant conditioning procedures (e.g.,Rovee & Fagen, 1976; Sullivan, Rovee-Collier, & Tynes,1979; Watson, 1967). In one experiment, Rovee and Fa-gen (1976) studied memory or "retention" of operantfoot kicks in 3-month-old infants. Each infant's right footwas attached with a ribbon to a mobile hanging abovethe infant's crib so that when the infant's foot moved,the mobile would immediately move. The more vigor-ously the infant's foot moved, the m ore the mobile wouldmove. The subjects were presented with daily sessionslasting 15 minutes, the last 12 minutes of which right-foot kicks were followed by the m oving m obile for 9 min -utes followed by a 3-minute extinction period. There wassignificant and specific response carryover from one dayto the next, prom pting the investigators to conclude that"clearly, 3-month-old infants are capable both of inte-grating simple perceptual-motor responses and of direct-ing them selectively at later intervals" (p. 10). However,it wasn't the infants who integrated "simple perceptual-mo tor responses." It was the environm ent created by theexperimenters.Unlike most studies of infant memory, the experi-ments by Rovee-Collier and her colleagues (e.g., Rovee& Fagen, 1976; Rovee-Collier & Sullivan, 1980; Rovee-Collier, Sullivan, Enright, Lucas, & Fagen, 1980; Sullivanet al., 1979) demonstrated how a research program intoa stronghold of cognitive psychologym emory can bederived from behavior-analytic theory. Moreover, the re-sults of these experiments not only challenge the conclu-sions about infant memory generated by more "tradi-tional paradigms" (Rovee-Collier, 1983), but the resultsare easily interp reted in term s of the principles of ope rantstimulus control. However, cognitively oriented psychol-ogists can take even straightforward demonstrations ofoperant conditioning in infants and interpret them as ev-idence for inferred events. So, for example, Cohen andGelber (1975) interpreted operant conditioning in infantsby saying that "w hatever is stored . . . must include someinformation about the appropriate response and thereforemay be an instance of enactive me mo ry" (p. 351). How-ever, is it really necessary to infer storage, information,and enactive mem ory to understand, predict, and controlthe behavior? Watson (1967) likened the mem ory requiredfor operant conditioning in infants to a "computer-like



9/15

process" in which "the mem ory records of stimulus inputand response output are scanned under the guidance ofa ' learning instruction' reading: 'Find and repeat the re-sponse which preceded the reception of the reward stim-ulus' " (p. 55). Needless to say, the view of the presentarticle is that these invented processes are unnecessaryobstacles to understanding the behavior from which theyare inferred. From a behavior-analytic perspective, theresults from the operant conditioning studies in infantscan be predicted and understood according to the exten-sive literature on operan t conditioning. For exam ple, thefinding by Rovee and Fagen (1976) that response strengthwas so high on Day 2 "after just 9 minutes of reinforcedrespond ing" on Day 1 could have been predicted. Be-havior analysts already know that considerable responsestrength can be generated after just one or a few reinforcedresponses (Neuringer, 1970; Skinner, 1938). The fact thatfoot kicks occurred on days after train ing (even after b riefextinction periods at the conclusion of the preceding ses-sion) may be predicted by the p rinciples of reinforcementand operant stimulus control. That is, the functions ofthe stimuli associated with the experiment (e.g., the sightof the mob ile and the ribbon attached to the infants' legsand so on), especially at the beginning of the session,evoked leg kicks because they had in the past been cor-related with reinforcement for the behavior. In fact, op-erant stimulus control has been clearly demonstrated invery young infants when, for example, changes in thevisual characteristics of the prevailing stimuli were re-flected in changes in th e respo nse rates of infant leg kicks(Rovee-Collier & Capatides, 1979).

In short, in the particular experimental environ-men ts found in the experim ents of Rovee-Collier and hercolleagues, foot kicks were adaptive. Of course, neitherthe experiments nor the present interpretation explainwhy the sight of the moving mobile functions as a rein-forcerif we are interested in the parameters of rein-forcers for human infants, that is another experimentalquestionbut that it does is sufficient to understand theoccurrence of the behavior. Any analysis of operant footkicks in terms of infant memory is unnecessary becausewe can account for the observed changes in behaviorwithout it. For behavior analysis, the operant behavior isimportant in its own right and not as an indicator ofinferred underlying structures or processes. A behavior-analytic interpretation takes noth ing away from the infant;on the contrary, as the studies by Rovee-Collier and herassociates showed, a behavior-analytic approach dem-onstrates a way of more thoroughly investigating the be-havioral capabilities of young children.Perceptual Size ConstancyStandard treatment of size constancy. Another area ofstudy that demonstrates the amenability of phenomenato behavioral interpretation is perceptual behavior. Per-ception is often contrasted with sensation. The term sen-sation is usually used to refer to the basic effects of stimulion the sense organs. Perception, however, has been cog-nitivized as

The process by which animals gain knowledge [italics added]about their environment and about themselves in relation tothe environment. It is the beginning of knowing, and so is anessential part of cognition. More specifically, to perceive is toobtain information [italics added] about the world throughstimulation. (Gibson & Spelke, 1983, p. 2)Dworetzky (1991) used the term perception to refer tohow the organism interprets the sensory experience. Sizeconstancy is an example of one type of experience thatorganisms are said to interpret.The rule about the constancies of objects goes some-thing like this: Even though the size, shape, and positionof objects change with respect to the image projected onthe retina, the actual size, shape, and position of the ob-jects in the environm ent rem ain co nstant. Size constancyis said to refer to "the stability of apparent size, despitechanges in projected retinal image size" (Banks & Sala-patek, 1983, p. 523 ). However, what is it that is stable?Dworetzky (1991) described an exam ple of size constancyas follows:When you look at your parked car as you are walking awayfrom it, your visual sensory system sends a message to yourbrain. First the image of the car is projected through the lensonto an area at the back of each of your eyes called the retina.As you walk away (still looking at your car), the image projectedonto the retina gets smaller and smaller as you get farther fromyour car. . . . Although the sensory image of your car is shrink-ing rapidly, you don't perceive that your car is changing size.Instead, you perceive that your car is simply becoming moredistant, (p. 134)However, what is this perception?Dworetzky (1991) defined size constancy as "thelearned perception that an object remains the same size,despite the fact that the size of the image it casts on theretina varies with its distance from the viewer" (p. 134).What is the "perception" and how is it "learned"? It issometimes said that we know that an object is the samesize even though it appears smaller. It is even said thatalthough the object casts a smaller image on the retina,we still see it as being the same size. Developmental re-searchers characteristically have been interested in theage at which children "acquire size constancy" (e.g., Day& M cKenzie, 1981; Yonas, Gra nru d, & Petterson, 1985).However, what exactly is acquired?These confusing ways of talking about the ph enom -enon pose all of the problem s of any structural in terpre-tation of behavior. Just as with memory, behaviors areobserved in certain situations and they are given a name,in this case, perception, or more specifically, size con-stancy. The name is reified as a type of knowledge orcognition about the world, although the knowledge orcognition is never directly observed. In th is case, the childis said to "possess size constancy." This implies that, asa result of certain experiences, a structure is created insidethe child that enables the child to correctly interpret thesizes of objects. Con centrating on this and other inferredstructures prevents researchers from investigating morefundamental behavioral processes, like the nature of theexperiences themselves (e.g., the reinforcement contin-



10/15

gencies that contribute to the observed behaviors ["per-ceptions"]). The emphasis is not on the child's behaviorbut on inferred surrogates of the behavior, called inter-pretations or recognitions, the only evidence of which isthe very behaviors they are said to explain. In othe r words,the behavior of the child in the particular environm entalcontext that defines the perception in the first place isoverlooked in lieu of inventing internal surrogates, whichare said to interpret the sensory information.

Experimental facts on depth perception/size con-stancy. Researchers have determined the approximateages when sensitivity to the cues for depth, and thus size,perception first appears. T here is general agreement thatcontrol by the binocular cue of retinal disparity is firstevident in infants between 2 and 3 months of age and isfully present by 5 months of age (Banks & Salapatek,1983). Several studies indicated that depth-appropriateresponding occurs at abou t 5 month s (Bechtoldt & Hutz,1979; Gord on & Yonas, 1976; Yonas, Oberg, & No rcia,1978). For example, using a stereoscopic display thatsimulated the approach of an object on a collision coursewith the infant's face, Yonas et al. (1978) showed that 5-month-olds but not 31/2-month-olds exhibited morereaching, head withdrawal, and blinking.

Interestingly, control by monocular cues appearslater. Yonas, Cleaves, and Petterson (1978) used a trap-ezoidal window presented parallel to the frontoparallelplane to investigate control by the mo noc ular cue of linearperspective. Wh en one eye was covered, both adults an d7-month-old infants (incorrectly) reached more to thelarger ("nearer") side of the trapezoid, wh ereas 5-m onth-olds reached ab out equally often to b oth sides. In an otherstudy, Yonas, Petterson, and Granfud (1982) comparedthe control by familiar and unfamiliar size cues in 5- and7-month-olds. The familiar objects were two different-size pictures of the infan ts' m others; the unfam iliar objectswere two checkerboards of different sizes. All stimuli werepresented at the same distance and were viewed mon-ocularly. Results showed that the 7-month-olds reachedmore often to the larger face but equally often to the twocheckerboards, whereas the 5-month-olds reached equallyto all stimuli, suggesting that the 7-month-olds but notthe 5-month-olds exhibited size-appropriate respondingto familiar objects.

Although researchers have a good idea about whatinfants do and when they do it with respect to size con-stancy, there has been no attempt to relate these facts tothe wealth of other facts of child development. Th e reasonis developmental psychologists generally believe that nosingle theory can accomplish such a task (Hetherington& Parke, 1986). Moreover, the normative facts of theemergence of size constancy encourage the assumptionof maturatio nal processes, even though there is little directevidence for them. Another, more parsimonious, expla-nation of the relevant behavior is in term s of its adaptivevalue. In short, for behavior analysts the question mightbe "What typically happens to infants between about 5and 7 months of age that produces a transition betweennon-size-appropriate and size-appropriate behavior?

A behavior analysis of size-appropriate behavior. Abehavioral theory of perceptual size constancy first de-scribes behavior in its context and then interprets it ac-cording to established principles. Specifically, perceptualbehavior seems to be behavior under complex stimuluscontrol (e.g., Knapp, 1987; Malott & Whaley, 1981;Nevin, 1973 ;Schoen feld&C um ming, 1963). Perceptualstimulus control is usually of a visual or an auditory na-ture, but there is no reason why it cann ot be d emo nstratedwith other sensory systems. A behavioral approa ch doesnot dismiss questions about the chronology of certainbehavioral changes; however, it looks for answers in theinteractions between the behavior and the environment.Hence, we might ask why control by monocular cuesappears to develop later than control by binocular cues.To answer this, behavior analysts would want to under-stand the nature of the infant's early visual experiences.They might discover, for example, that from the earliestdays after birth and for the first several months, the in-fant's visual environment consists largely of objects thatmove toward the infant (e.g., people's faces, bottles, pac-ifiers, and so on), which produce the binocular cues ofconvergence and retinal disparity. In the presence of theseevents, certain consequences (e.g., being touched, havingsomething placed into the mouth, or simply seeing theobject) immediately follow the infant's behaviors (e.g.,looking or reaching). Some of the infant's reactions (e.g.,head withdrawal and blinking) are probably initially partsof unconditional reflexes and are elicited by objects thatmov e quickly toward th e infant's face; however, other be -havior (e.g., looking or reaching) becomes operant andthus comes under the stimulus control of the binocularcues. Reaching in the absence of approaching objects isnot followed by the sam e consequences, whereas reachingwhen these things approach is reinforced either by tou ch-ing and grasping the object or some form of attentionfrom the parent. Many of the mo nocula r cues (e.g., linearperspective, interposition, an d texture gradients) proba blybecome more important once the infant is in a sittingposition,.which typically occurs between 5 and 8 mo nthsof age. The qu antity and quality of potentially functionalstimuli increases dramatically when the infant's body po-sition changes from the prone or supine to the sittingposition and then again from the sitting to the standingposition. Just as with behavior that we speak of as mem -ory, behavior-analytic theory asks abou t the variables thatdeterm ine changes in the behavior of infants under thesefunctional perceptual cues. Thus, in part by asking dif-ferent questions, behavior-analytic theory offers an ob-jective and parsimo nious way of describing the ph enom -ena as well as a parsimonious explanation of the changesin behavior. However, what about size constancy?

First, let us be clear about what is seen when thedistance of objects changes. The image cast on the retinabecomes smaller as objects become more distant andlarger as objects get closer. If we move away from ourautomobile, the image on the retina gets smaller and,however we describe it, we never see it as anything butsmaller. When it is said that we perceive it as being theNovember 1992 American Psychologist 1405


11/15

same size, what is meant is that we react to it as beingthe same size. W hat a re these reactions? They are beh aviorlike any other behavior. For example, as the retinal imageof our automo bile gets smaller, we do n ot shriek that ourcar is shrinking. We still behave toward our automobilein most of the ways we always have. We still call it ourautomobile. We may reach in our pocket for our keys,and we will walk toward it when we want to go som ewhereand so on. So what is size constancy? An objective de-scription of size constancy is to say that even though weactually see objects as smaller or larger (in terms of retinalimage size), we contin ue to behave toward them in m ostof the ways we always have. Size constancy, like all objectconstancies, is a property of objects in the real world.That is, the sizes of objects are constant. It is not necessaryto infer cognitive structures, behavioral surrogates, orperceptions to explain the size-constant behaviors. In ad-dition, it is not necessary to say that the constant sizes ofobjects are taken in and acquired. For psychology, itshould be sufficient to say that our behavior comes underthe stimulus control of objects and either varies or doesnot vary as the sensory charac teristics of their size, shape,and position change depending on the operative contin-gencies. How does this happen ? Many accounts of depthperception and the related perception of size constancyallude to experiential factors but offer no interpretations(e.g., Banks & S alapatek, 1983). Behavior analysis is inan ideal position to offer an interpretation.

A stimulus con trol accou nt of size constancy requiresthe identification of the stimulus characteristics involvedwhen the distance of objects varies. The ability to reactappropriately to distant objects (or depth) depends onseveral types of cues. There are two general classes ofdepth cues: binocular cues and monocular cues (Banks& Salapatek, 1983). The binocular cues include conver-gence and retinal disparity (or binocular parallax). Themonocular cues can be classified according to those thatare static (often called pictorial cues because they areused by artists to create the impression of depth) andthose that are kinetic and require motion by either theobject or the observer (Banks & Salapatek, 1983). Thestatic monocular cues that are important for depth per-ception include linear perspective, interposition, texturegradients, relative and familiar size, aerial perspective,and, to a lesser degree, visual accom mo dation . Th e kineticmonocular depth cues include changes over time in theretinal image produced either by the movements of theobserver's head and body (e.g., motion parallax) or bythe movements of objects (e.g, optical expansion or per-spective transformation; Banks & Salapatek, 1983). Thecritical question is "How do these cues come to controlbehavior?"One problem in assessing perceptual stimulus con-trol is the actual behavior to be measured. Perceptualbehavior is whichever behavior is being investigated und erthe control of perceptual cues. However, it is possible toargue that, with some behaviors such as visually guidedreaching, infants may not be physically capable of exhib-iting the behavior, but some behavior may still be brought

under the control of perceptual cues. In some cases, thisprob lem has been overcome by using other measures suchas changes in heart rate or sucking rate (e.g., Campos,Langer, & Krowitz, 1970; Kalnins & Bruner, 1973). Theimp lication is that if stimuli d o no t yet differentially con-trol some behavior, then we cannot speak of perception.In other words, psychologists do not speak of perceptionin the absence of some measurable (perceptual) responsejust as they don't speak of recognition memory in theabsence of differential fixation times.I previously described research by Yonas et al. (1982)regarding the emergence of size constancy in infants.These results are interesting not only because they indi-cate that size-appropriate responding is typically presentby 7 months of age but also that they are amenable to abehavior-analytic interpretation. Accordingly, objects be-come "familiar" because of a history of reinforcementthat has produced perceptually appropriate respondingto them . However, a behavior analysis must do more thanoffer post hoc explanations; it must offer plausible mech-anisms. First, a behavior-analytic interpretation ap-proaches the transition from non-size-appropriate to size-appropriate responding as adaptive behavior. Ideally, thenext step would involve an experimental demonstrationin which size-appropriate responding is generated in anorganism without such behavior. Lacking such a dem-onstration, behavior analysts would offer a plausible in-terpretation according to the laws and principles of be-havior analysis. Let me suggest how one might interpretsuch a behavioral transition in infants in their "natural"environment.One response in infants that comes under the ap-prop riate c ontrol of object size and de pth cu es fairly earlyis (visually guided) reaching . Obviously, operant reachin gcanno t occur until the underlying musculoskeletal struc-tures have matured, although such development appar-ently can be accelerated by environm ental manipu lation(Wh ite, Castle, & Held, 1964; W hite & Held, 1966). Con -sider an infant lying in its crib. By the time reaching isphysically possible, one of the first contingencies to whichthe infant is exposed must involve the sight of the parentas a stimulus in whose presence reaching out is successful;that is, reaching produces some type of contact or inter-action with the parent. Such a stimulus is termed a dis-criminative stimulus (S D), and its control over behavioris evidenced when, on subsequent occasions, it evokes thebehavior that was previously successful. In the presentexample, however, successful reaching is possible onlywhen the retinal image of the parent is relatively large(e.g., when the parent is leaning over the crib). When theretinal image is small (i.e., the parent is farther away),reaching out, which probably occurs during the initialstages of discrimination learning, goes unreinforced. Inthis situation, retinal image size is probably the most im-portant S D (or S delta, that is, a stimulus in whose presencebehavior is unsuccessful) for the size-appropriate behavior.Consider another scenario. An infant is sitting onthe floor of an average-size room. There are two stuffedanimals of different sizes. The smaller of the two is next



12/15

to the infant, and the larger one is on the other side ofthe room ; however, both stuffed an imals pro duc e the sameretinal image size. There are also other toys in the room,some of which produ ce the sam e retinal image size as thefarthest stuffed animal even though they are closer to theinfant and vice versa. What visual cues are provided bythe two stuffed animals and the other toys that determinereaching behavior to the stuffed animals? Traditional ap-proach es might ask how the infant " kn ow s" which stuffedanimal is really closer, how the infant "calculates" dis-tances, or whether the infant has "acquired" size con-stancy? The behavior analyst asks what experiences (i.e.,interactions with environment) produce size-constantbehavior? Recall that both stuffed animals produce thesame retinal image size, so retinal image size cannot bea functional stimulus for size-appropriate reaching. Ini-tially, the infant reaches for both of the stuffed animals.Reaching for the closer one is successful, and reachingfor the farthest one is unsuccessful. The binocular cuesnotwithstanding , most of the other static mon ocular cuescome into play. For example, the closer stuffed animal orother close objects tend to ob scure the m ore distan t stuffedanimal (interposition). Moreover, any receding lines inthe roo m will app ear to converge near the farthest stuffedanim al (linear perspective). Also, the m ore distan t stuffedanimal and other equally distant objects produce a finergrained texture than closer objects. In the presence of thisstimulus complex, then, reaching for the closer stuffedanimal is successful (i.e., it is reinforced); and this par-ticular complex becomes an S D and will under similarcircumstances in the future evoke reaching for closer ob-jects. Reaching for the farthest stuffed animal is not suc-cessful; the same stimulus complex becomes an S delta(i.e., a stimulus that is correlated with failure or non-reinforcement of responses) and will not evoke reachingfor farther objects. Of course, determining the particularproperties of the environment that become functionallyrelevant requires an experimental analysis; the presentinterpretation can only suggest possible candidates.

Visually guided reaching is not the only example ofsize-constant behavior. Once verbal behavior is present,that too becomes part of the class of responses controlledby depth cues. For most people, verbal responses con-trolled by the sight of objects are usually acquired undera variety of condition s in which th e sensory characteristicsof the objects vary. One of the verbal responses that con-tinues to be appropriate despite these changes is called atact (Skinner, 1957), which is roughly equivalent to thenam e of an object. Most verbal comm unities maintain theconstancy of tacts regardless of sensory changes in othercharacteristics of objects. The acquisition of this verbalbehavior might be expedited through stimulus generaliza-tion from the stimulus characteristics present when thetact was first conditioned, or it might require more directtraining. For example, the tact "m am a" m ight be acquiredwhen th e m other is very close, thus creating a large retinalimage. If, however, the mother moves away, her physicalfeatures, which remain constant, may still evoke "mama,"which would normally be quickly reinforced by the m other

reacting to the infant's speaking. If it does not, pro mp tingmay be necessary. More likely, however, tacts, like mostother behavior, are acquired un der varied conditions; thus,stimulus generalization is built in.I have presented a brief and general account of abehavior-analytic theory of infant memory and of per-ceptual size constancy. Some readers may claim that thethrust of my objections to standard developmental ac-counts is primarily semantic. It is true that some of myobjections are largely semantic, but that does not renderthem trivial or unimportant. Behavior analysts, withSkinner in the forefront, have often pointed out the "pit-falls associated with the use of terms that lead easily tothe postulation of inner, hypothetical structures" (Branch,1977, p. 178). Other readers may object that the presentaccount of infant m emo ry and perceptual size constancyis not based on direct behavior-analytic research of thepheno mena , and they may, therefore, question the valueof such an enterprise. Its value, however, depends on thepower of the the ory a nd th e laws it mak es up. What followsis a brief evaluation of behavior-analytic theory.

Behavior-Analytic Theory EvaluatedTheory in behavior analysis is grounded in the law ofeffect or the principle of reinforcement. What leads us tocall reinforcement a scientific law? McCain and Segal(1988) listed four criteria that must be obtained beforeany statement can qualify as a scientific law:(1) The statement must be about kinds of events and not directlyabout any singular event. (2) The statement must show a func-tional relation between two or more kinds of events ("kind ofevent" refers either to things or to properties of things). (3) Theremust be a large amount of data confirming the law and little ornone disconfirming it. (4) The relation should be applicable tovery different events (although there may be limiting conditions).(P. 52)Th us, a scientific law "is a collection of facts grouped intoa consistent body of knowledge, from which it is possibleto make predictions" (Bachrach, 1972, p. 49). This isindeed consistent with Skinner's (1947) conception of therelation between fact and theory in science. According tothese criteria, then , reinforcem ent qualifies as a scientificlaw. As Bijou and Baer (1978) wrote,It is a summary of many, many well-proved facts, and it is alsoan induction that goes beyond them to suggest that the unifor-mity with which they are found to be true suggests strongly tha tthey are generally (but not universally) true. In tha t the inductiongoes further than proven facts, it is a statement of theory; inthat it goes beyond fact only to suggest that an observed gen-erality is probably more general than the cases observed so far,it is empirically based and characteristic of a natural scienceapproach, (p. 8)However, what about the larger issue of theory?Psychological theories can be evaluated accordingto several criteria. There have been various attempts toevaluate developmental theories (e.g., Green, 1989;Tho mas , 1985). Although these treatmen ts suggest some-what different standards, they agree on some. For ex-



13/15

ample, a good scientific theory should (a) be testable, (b)be internally consisten t, (c) possess predictive validity, (d)possess external validity, and (e) be theoretically econ om -ical. With respect to these criteria, behavior analysis israted comparatively high as a theory of development(Green, 1989; Thomas, 1985), even though it is not anexplicit theory of development. It is possible to abstracta subset of important criteria for the evaluation of psy-chological theories in general (see Poling, Schlinger,Starin, & Blakely, 1990; also Bachrach, 1972). They in-clude empirical support, logical support, generality, par-simony, and utility.Criteria for Evaluating Theories in PsychologyEmpirical support. Any theory must have observationsthat relate to it and support it, and the theory must notbe so general as to be able to account for any possible setof observations. Behavior-analytic theory has a largenumber of supporting observations in the form of em-pirical functional relations between behavioral and en-vironm ental events with num erou s species, including h u-mans, and under myriad experimental conditions, whichnot only supp ort th e theory but, as we saw earlier, actuallytogether comprise the theory. Thus, they do well in in-terpreting or explaining the possible environmental de-terminants of behavior.

Logical support. The mechanisms proposed by atheory m ust be plausible and mu st not involve reification,circularity, teleology, or nominal fallacy. The plausibilityof explanatory mechanisms is enhanced by beinggrounded in empirical observations. Clearly, then, themechanisms proposed by behavior-analytic theory areplausible. They a re plausible because they are generalitiesof objective scientific facts. However, they do not bythemselves escape from the logical errors of reification,circularity, teleology, or nominal fallacy; those are syn-tactic errors. However, the primary explanatory mecha-nisms in behavior-analytic theoryrespondent and op-erant unitsare directly testable without inferring eventsat other levels; hence, they are more resistant to logicalerror.Generality. Generality refers to the range of behaviorsa theory claims to explain and the range of conditionsund er w hich it does so. Behavior analysis is, by definition,the study of ontogenetically adaptive behavior. Any be-havior that undergoes change as a result of interactionswith environm ental variables is considered adaptive andis thus po tentially u nderstanda ble with behavior-analytic

theory. Behavior analysts have shown that other devel-opmental phenomena can be understood according tobehavior-analytic theory, for example, exploratory be-havior (Bijou, 1976), cognitive abilities (Bijou, 1976;Malott & Whaley, 1981), social referencing (Gewirtz &Pelaez-Nogueras, 1991), attachment (Gewirtz, 1991), andmoral behavior (Bijou, 1976; Gewirtz & Pelaez-Nogueras1991). If behavior-analytic theory can provide plausibleaccounts of what is termed infant mem ory and perceptualbehavior, then it will have enhanced its generality.Parsimony. Good theories invoke simple explana-

tory mechanisms first. A simple explanation is one thatis potentially observable and has both empirical and log-ical support. Not all adequate ex planations possess thesecharacteristics, but parsimo ny d ictates that the ones thatdo should be invoked first. Behavior-analytic theory ap-peals to established laws and principles concerning re-lationships between observable behavior and objectiveenvironmental events. If these explanations are sufficientto account for behavior, then more complex onesthatis, those that appeal to unverifiable events and processesmay not be necessary.Utility. One feature of the natural sciences that has"sold" them is not only their power to understand naturalphen ome na bu t their power to change them. As Schwartz(1989) put it, "science has delivered the goods" (p. 3).Behavior analysts have consistently maintained that thegoals of a science of behavior are prediction and control.Perhaps it would be more accurate to say control andprediction, because the prediction in behavior analysis ispredicated on being able to control the subject matter.This relation between prediction and control is the sam ewithin biology, physics, and chemistry. The appliedbranches of those sciencesfor example, medicine andengineeringare to a certain degree accountable for theirclaims by being able to effect technological changes. Sotoo can behavior analysis. In fact, some cognitive psy-chologists have noted the successes of the applied fieldsof behavior analysis while questioning the methods andtheory (e.g., Baars, 1 986). However, this misses the po intthat the scientific methodology and theory of behavioranalysis derived from years of basic research have madethe applied successes possible.

Thus, it appears that behavior-analytic theory notonly meets these criteria of good scientific theories, but,I argue, that in so doing it may be able to offer moreplausible interpretations of behavioral phenomena thanother psychological theories.ConclusionI have presented only two brief examples of the appli-cation of behavior-analytic theory to problems of childdevelopment. However, they reflect the general strategyof analyzing behavior in terms of the basic units of be-havior analysis and the laws and principles that sum-marize them. Similar applications of behavior-analytictheory to other problems of child development shouldencourage o ptimism that the science of behavior analysiscan, if given a chance, po tentially unite the d isparate sub-disciplines of developmental psychology, thus not onlyenabling a better understanding of the behavioral changesthat define development but also enhancing the effective-ness of applications to parenting, educational, and ther-apeutic endeavors. The ultim ate evaluation of the behav-ior-analytic app roach , according to Bijou and Baer (1978),"will depend on the adequacy w ith which it accounts forthe psychological development of hum an s" (p. 8). In thisarticle, I have attempted to offer a glimpse of how thataccount might proceed.



14/15

REFERENCESBaars, B. J. (1986). The cognitive revolution in psychology. New York:Guilford Press.Bachrach, A. J. (1972). Psychological research: An introduction (3rded.). New York: Random House.Banks, M. S., & Salapatek, P. (1983). Infant visual perception. In P. H.Mussen, M. Haith, & J. Campos (Eds.), Handbook of Child Psy-

chology: Vol. 2. Infancy and developm ental psychobiology (4th ed.,pp . 436-571). New York: Wiley.Bechtoldt, H. P., & Hutz, C. S. (1979). Stereopsis in young infants andstereopsis in infants with esotropia. Journal of Pediatric Ophthal-mology, 16, 49-54 .Bijou, S. W. (1976). Child development: The basic stage ofearly childhood.Englewood Cliffs, NJ: Prentice-Hall.Bijou, S. W., & Baer, D. M. (1978). Behavior ana lysis of child develop-ment. Englewood Cliffs, NJ: Prentice-Hall.Branch, M. N. (1977). On the role of "memory" in the analysis ofbehavior. Journal of the Experimental Analysis of Behavior, 28, 1 7 1 -179.Campos, J. J., Langer, A., & Krowitz, A. (1970). Cardiac responses onthe visual cliff in prelocomotor human infants. Science, 170, 196-197.Catania, A. C. (1973). The concept of the operant in the analysis ofbehavior. Behaviorism, 1, 103-115.Catania, A. C. (1984). Learning (2nd ed.). Englewood Cliffs, NJ: Prentice-Hall.Cohen, L. B., & Gelber, E. R. (1975). Infant visual memory. In L. B.Cohen & P. Salapatek (Eds.), Infant perception: From sensation tocognition: Vol. 1. Basic visual processes, (pp. 347-403). San Diego,CA: Academic Press.Collins, W. A., & Kuczaj, S. A. (1991). Developmental psychology:Childhood and adolescence. New York: Macmillan.Day, R. H., & McK enzie, B. E. (1981). Infant p erception of the inva riantsize of approaching and receding objects. Developmental Psychology,77 ,670-677 .Dworetzky, J. P. (1991). Introduction to child development (4th ed.). St.Paul, MN : West.Fantz, R. (1961). The origin of form perception. Scientific American204, 66-72 .Flannagan, O. (1991). The science of the mind (2nd ed.). Cambridge,MA: MIT Press.Gewirtz, J. L. (1972a). Attachment, dependence, and a distinction interms of stimulus control. In J. L. Gewirtz (Ed.), Attachment anddependency (pp. 139-177). Washington, DC: Winston.Gewirtz, J. L. (1972b). On the selection and use of attachment anddependence indices. In J. L. Gewirtz (Ed.), Attachment and dependency(pp. 179-215). Washington, DC: Winston.Gewirtz, J. L. (1991). Identification atta chmen t, and their developmentalsequencing in a conditioning frame. In J. L. Gewirtz & W. M. Kur tines(Eds.), Intersections with attachment (pp. 247-255). Hillsdale, NJ:Erlbaum.Gewirtz, J. L., & Pelaez-Nogueras, M. (1990). Infant social referencingas a learned process. In S. Feinman (Ed.), Social referencing and thesocial construction of reality in infancy (pp. 1-11). New York: PlenumPress.Gewirtz, J. L., & Pelaez-Nogueras, M. (1991). Proximal mechanismsunderlying the acquisition of moral behavior patterns. In W. M. Kur-tines & J. L. Gewirtz (Eds.), Handbook of moral behavior and devel-opment: Vol. 1. Theory (pp. 153-182). Hillsdale, NJ: Erlbaum.Gibson, E. J., & Spelke, E. S. (1983). The development of perception.In P. H. Mussen, J. H. Flavell, & E. M. Markman (Eds.), Handbookof child psychology: Vol. 3. Cognitive development (4th ed., pp. 1-76).New York: Wiley.Gordon, F. R., & Yonas, A. (1976). Sensitivity to binocular depth in-formation in infants. Journal of Experimental Child Psychology, 22 ,413-422.Green, M. (1989). Theories of human development: A comparative ap-proach. Englewood Cliffs, NJ: Prentice-Hall.Hetherington, E. M., & Parke, R. D. (1986). Child psychology: A con-temporary viewpoint (3rd ed.). New York: McGraw-Hill.Hineline, P. N. (1990). The origins of environment-based psychological

theory. Journal of the Experimental Analysis of Behavior, 53 , 3 0 5 -320.Kalnins, I. V., & Bruner, J. S. (1973). The coordination of visual ob-servation and instrumental behavior in early infancy. Perception, 2,307-314.Kandel, E. R. (1975). Nerve cells and behavior. In R. F. Thompson(Ed.), Progress in psychobiology (pp. 281-2 92). San Francisco: Free-man.

Knapp, T. J. (1987). Perception and action. In S. Modgil & C. Modgil(Eds.), B. F. Skinner: Consensus and controversy (pp. 283-294). NewYork: Falmer.Lipsitt, L. P. (1981). Sensorimotor development: What infants do andhow we think about what they do. In I. E. Sigel, D. M. Brodzinsky,& R. M. Golinkoff (Eds.), New directions in Piagetian theory andpractice (pp. 29-37). Hillsdale, NJ: Erlbaum.Malott, R. W, & Whaley, D. L. (1981). Psychology. Kalamazoo, MI:Behaviordelia.McCain, G., & Segal, E. M. (1988). The game of science. Pacific Grov e,CA: Brooks/Cole.McDowell, J. J. (1991). Irreconcilable differences and political realityin these dark ages. The Behavior Analyst, 14, 29-33.Miranda, S. B., & Fantz, R. L. (1974). Recognition memory in Down'sSyndrome and normal infants. Child Development, 45, 651-660.Neuringer, A. J. (1970). Superstitious key pecking after three peck-pro-duced reinforcements. Journal of the Experimental Analysis of Be-havior, 13, 127-134.Nevin, J. A. (1973). Stimulus control. In J. A. Nevin (Ed.), The studyof behavior: Learning, motivation, emotion, and instinct (pp. 114-152). Glenview, IL: Scott, Foresman.Palmer, D. C. (1991). A behavioral interpretation of memory. In L. J.Hayes & P. N. Chase (Eds.), Dialogues on verbal behavior, (pp. 261 -279). Reno, NV: Context Press.Poling, A., Schlinger, H., Starin , S., & Blakely, E. (1990). Psychology: Abehavioral overview. New York: Plenum Press.Rovee, C. K., & Fagen, J. W. (1976). Extended conditioning and 24-hour retention in infants. Journal of Experimental Child Psychology,21, 1-11.Rovee-Collier, C. K. (1983). Infants as problem solvers: A psychobiol-ogical perspective. In M. D. Zeiler & P. Harzem (Eds.), Advances inanalysis of behaviour (Vol. 3; pp. 6 3-1 01) . New York: Wiley.Rovee-Collier, C. K ., & Capatid es, J. B. (1979). Positive behavioral con-strast in 3-month-old infants on multiple conjugate reinforcementschedules. Journal oj the Experimental Analysis o f Behavior, 32, 15-27 .Rovee-Collier, C.

schlinger ap

Documents