B - ROSALES-RUIZ,(1997O) - Behavioral Cusps, A Developmental and Pragmatic Concept (Olhar)

7/30/2019 B - ROSALES-RUIZ,(1997O) - Behavioral Cusps, A Developmental and Pragmatic Concept (Olhar)

1/12

533

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 1997, 30, 533544 NUMBER 3 (FALL 1997)

BEHAVIORAL CUSPS: A DEVELOPMENTAL ANDPRAGMATIC CONCEPT FOR BEHAVIOR ANALYSIS

JESUS ROSALES-RUIZUNIVERSITY OF NORTH TEXAS

AND

DONALD M. BAERUNIVERSITY OF KANSAS

Most concepts of development explain certain behavior changes as products or markersof the invariable succession of emerging periods, stages, refinements, or achievements thatdefine and order much of an individuals life. A different but comparable concept can bederived from the most basic mechanisms of behavior analysis, which are its environmentalcontingencies, and from its most basic strategy, which is to study behavior as its subjectmatter. From a behavior-analytic perspective, the most fundamental developmental ques-tions are (a) whether these contingencies vary in any systematic way across the life span,and thus make behavior change in a correspondingly systematic way; and (b) whethersome of these contingencies and their changes have more far-reaching consequences thanothers, in terms of the importance to the organism and others, of the behavior classesthey change. Certain behavior changes open the door to especially broad or especiallyimportant further behavior change, leading to the concept of the behavioral cusp. Abehavioral cusp, then, is any behavior change that brings the organisms behavior intocontact with new contingencies that have even more far-reaching consequences. Of allthe environmental contingencies that change or maintain behavior, those that accomplishcusps are developmental. Behavior change remains the fundamental phenomenon of de-velopment for a behavior-analytic view; a cusp is a special instance of behavior change,a change crucial to what can come next.

DESCRIPTORS: development, developmental stages, pivotal behaviors, behaviortraps, behavior analysis, behavior change

Conceptualizing the development of be-havior over the life span has been an endur-ing problem in psychology. Organismic the-ories postulate an invariable succession ofemerging stages, periods, achievements, dif-ferentiations, refinements, or products; they

The authors are grateful to Sigrid Glenn, Joel Green-spoon, Hayne Reese, Wendy Roth, and John Wright forsympathetic, critical, careful, competent, detailed, andconstructive argument; but they should not be held re-sponsible for the arguments advanced here. The authorsare also grateful to the National Institute of Child Health

and Human Development for research support (HD18955).

Address correspondence to Jesus Rosales-Ruiz at the De-partment of Behavior Analysis, University of North Texas,P.O. Box 13438, Denton, Texas 76203 (E-mail: [email protected]) or to Donald M. Baer at the Depart-ment of Human Development and Family Life, Universityof Kansas, Lawrence, Kansas 66045-2133.

suppose that much behavior develops inobedience to that sequence. And, becausethe sequence is invariant, it requires an ex-planatory logic, which most often takes theform of its apparent goal, as if the sequence

were self-organizing: The individual is seenas traveling epigenetic roads to uniquelyadult stages of development, much like atrain stopping at various stations before itreaches its final, always scheduled destina-tion, or a butterfly passing through the em-

bryo-larva-pupa-imago stages to the inevita-ble fluttering forth (see Overton & Reese,1973; Reese, 1991; Reese & Overton, 1970;Spiker, 1966). Whereas the teleologicalsequence implied in such approaches is thatan embryo is just a butterflys way of makinganother butterfly, it is equally plausible to


2/12

534 JESUS ROSALES-RUIZ and DONALD M. BAER

argue that a butterfly is just an embryos wayof making another embryo. Perhaps the con-cept of development is sometimes a wayto ignore an arbitrary half of the evolution-

ary process.Behavior analysis is different; it has nocomparable guiding metaphor to explainpatterns of behavior change throughout thelife span. At least, none is intrinsic to itspresent logic. Of course, one or several suchmetaphors might be added. But that addi-tion would seem apt only if it were done inthe natural-science style that has guided thedevelopment of behavior analysis so far.That means it must be more than a meta-phor; the premises justifying it should be

verifiable.Behavior analysis currently offers its well-

known behavior-shaping contingencies as itsbasic analytical processes; and it offers them,so far, without specifying any distinctive, re-liable patterning of them over the life span.If a concept of development is to be added,that concept must posit a reliable pattern ofhow these contingencies are applied over thelife span. Stated this way, the possibility ofa reliable pattern of behavior-change pro-cesses over the life span becomes a matter of

facts to be determined rather than as a the-ory to be imposed. We can ask whether theapplication of these contingencies, by natureand by people, varies in any systematic way.

We can ask whether the behaviors to whichthey are applied vary in any systematic way,and if so, whether that is by nature or byidiosyncratic societal convention. Discerningthose kinds of systematic patterns of contin-gencies across the life span appears to be animplicit theme of two recent texts that de-scribe development from a behavior-analyticperspective (Novak, 1996; Schlinger, 1995).These texts are oriented toward undergrad-uate readers; their mission is to show howtraditional developmental topics are amena-ble to a behavior-analytic interpretation. But

we can also ask whether some of the resul-

tant behavior changes have more far-reach-ing consequences than others. Here, we ad-dress that question by describing the conceptof developmental cusps (Rosales-Ruiz &

Baer, 1996) and suggesting some criteria forfar-reaching.

A PRAGMATIC CONCEPT OFDEVELOPMENT FORBEHAVIOR ANALYSIS

Consider a cusp as a behavior change thathas consequences for the organism beyondthe change itself, some of which may be con-sidered important. That requires us to de-velop the criteria of importance. To ap-

proach those criteria, we must expand thedefinition of cusp: We take as axiomatic thatany behavior change results from changes inthe interaction between the organism and itsenvironment. What makes a behaviorchange a cusp is that it exposes the individ-uals repertoire to new environments, es-pecially new reinforcers and punishers, newcontingencies, new responses, new stimuluscontrols, and new communities of maintain-ing or destructive contingencies. When someor all of those events happen, the individuals

repertoire expands; it encounters a differen-tially selective maintenance of the new as

well as some old repertoires, and perhapsthat leads to some further cusps.

Consider, for example, what can happenas a result of learning to crawl. The babysuddenly has increased access to the environ-ment and its contingencies. Now the babycan get to toys, family, and other thingsmore easily, or can stumble into obstacles,all of which produce interactions that willfurther shape the babys behavior. Some ofthese interactions initiate the shaping of oth-er behaviors that will soon contribute to

walking, others will shape responsiveness tovisual cliffs (e.g., Campos, Bertenthal, &Kermoian, 1992), and still others will pro-duce a variety of parental contingencies,


3/12

535BEHAVIORAL CUSPS

some delighted, some dismayed, that willfurther shape how much more and howmuch less of the physical and social environ-ment will be open to the childs further in-

teraction. Thus, if walking, safety, and theimmediate next direction of socialization areimportant for that baby at that time, crawl-ing is a cusp.

This argument does not deny the devel-opment of the many small, sequential skillsthat culminate in crawling. Perhaps each ofthem is a prerequisite for the next, and thusfor crawling. But the important point hereis that none of these skills alone suddenlyopen the childs world to new contingenciesthat will develop many new, important be-

haviors. Instead, each of them opens thechilds world only to the next skill. Their endpoint, crawling, is a cusp.

By contrast, consider a child who has allthe prerequisites for walking, yet continuesto crawl; an early study by Harris, Johnston,Kelley, and Wolf (1964) dealt with such acase. They systematically shaped walking ina preschool girl who almost always crawled.

Walking made it possible for her to partici-pate in the upright, fast-moving games herpeers played, which was most of their games.

A host of new interactions typically will fol-low from walking. If leg strength and par-ticipation in peer socialization were impor-tant to that girl at that time, walking was acusp for her. Or perhaps it was a cusp forher parents and teachers: Perhaps the girlsbehavior showed that walking and peergames had little importance for her at thattime; it was her parents and teachers whobelieved that leg strength, coordination, andpeer participation would have consequencesthat would be important to her later.

Teach a child to read accurately and flu-ently, and suddenly and systematically a vastamount of further development, and a new,drastically more efficient method of teach-ing, are operative. If any of that is importantto the child or to the childs future, then

accurate, fluent reading is a cusp. Teach achild with developmental disabilities gener-alized imitation, and future expansion of thechilds repertoire can suddenly and system-

atically be as explosive as the social environ-ment cares to make it, simply by modelingnew skills, not necessarily intentionally. Ifany of that is important, to the child or tothose responsible for the child, generalizedimitation is a cusp. Teach an infant to dis-criminate between positive parent attentionand disapproving parent attention, and youend the paradoxical reinforcement of inap-propriate child behavior, which suddenlyand systematically will alter the childs andthe parents futures, especially their joint fu-

tures. If gentle social guidance is importantto the child at that time, then coming underthe conventional stimulus controls used nat-urally by almost every parent (and almostevery subsequent teacher) is a cusp. Giveyoung adults the first sizable, dependable,disposable income of their lives, and sud-denly, systematically, and enduringly, newsources of teaching will emerge that may al-ter and expand some of their criteria for andsome of their practice of what constitutesfood, housing, transportation, entertain-

ment, travel, family, and responsibility. Ifany of that is important to the young peo-ple, to their society, or to its economy, dis-posable income is a cusp. (The parallel ar-gument for elderly people who can retire

with a disposable income is obvious.)These examples show that the concept of

cusp always depends on the phrase, If thatis important . . ., as if the audience mustdecide if that is important. We suggest thatin these arguments, importance most oftenis indeed a social phenomenon. In biology,perhaps importance is unquestionably sur-vival. In development, survival is rarely clear,so importance is very often a matter ofsomething else, usually social validity. A cuspmay unquestionably open new environmentsfor a child, and we may view what those new


4/12


environments will produce as being impor-tant; but if we inquire, we often will findthat others do not. More than one preschoolteacher has told parents that their child is a

social isolate and that the teacher can re-mediate that, only to be told by the parentsthat they prefer their child to be a socialisolate, because the parents think that isola-tion is important to the childs artistic, in-tellectual, or political development.

Not all new cusps need be seen as positiveor desirable. Introducing a child to an ad-diction is an obvious example of a terriblecusp (for the great majority of us); teachinga child that the correct first response to anynew problem is to seek help rather than to

persist in independent tries is a more subtleexample (for many of us).

Sometimes changing only one behaviorwill create a cusp; sometimes it will be nec-essary to change a class of behaviors. A cuspmay be easy to accomplish, or it may bedifficult, tedious, subtle, or otherwise prob-lematic; yet if the cusp is not achieved, littleor no further change is possible in its realm(and perhaps in several other realms). But,

when the cusp is achieved, a set of subse-quent changes, important to someone, sud-

denly becomes easy or highly probable. Andwhen that cusp brings the developing organ-ism into contact with other, subsequent con-tingencies crucial to further, more complex,or more refined development in a therebysteadily expanding, steadily more interactiverealm, that will connote the conventional la-bel of developmental. In traditional theory,the connotations of increased complexity orrefinement often are put forward as causaland explanatory, in a teleological sense. Thecusp explains in a different way. It points outthat certain behavior changes cause subse-quent broad or important behavior changes,in the sense of making those subsequentchanges available. If we want to explainthose subsequent changes, we need to knowthe contingencies that shape them and the

cusp that makes them available for thatshaping.

The logic of cusps is implicit in earlierdiscussions by Baer and Wolf (1970) and

Baer, Rowbury, and Goetz (1976), who con-sidered behavioral traps and the responsesthat enter such traps (cf. Martin & Pear,1978; Stokes & Baer, 1977). A behavioraltrap is a community of reinforcement in thenatural environment that could maintainand potentially shape much new behavior ofits members. Preschools, universities, andother social organizations are traps waitingfor new members to enter and so, probably,to be shaped. To the extent that these trapsshape behavior beyond the entry responses,

and to the extent that those behaviors areimportant to someone at some time, the en-try responses are cusps. For example, achilds rudimentary social skills could betrapped in the natural community of peerssocial reinforcement by reinforcing responsesthat result in proximity to other children.The contingencies practiced by those peerson the behavior of anyone in steady contact

with them will differentiate, discriminate,schedule, and maintain a much larger, morerefined, and more complex set of social skills

(e.g., Allen, Hart, Buell, Harris, & Wolf,1964). In this example, the cusp is the be-havior change of being proximate to thegroup. That is a very small behavior changeand relatively easy to program; but it is alsoa cusp because of the extent and importanceof what happens next.

Some arguments by the Koegels and theircolleagues (Koegel & Frea, 1993; Koegel &Koegel, 1988; Koegel, Koegel, & Schreib-man, 1991) embody the cusp concept. Theycall pivotal any behavior changes that re-sult in collateral changes of other behaviorsas well (Koegel & Frea, 1993, p. 369).They suggest that many children with au-tism do not persevere in problems as do typ-ically developing children. But programmingmore reinforcement across a variety of prob-


5/12

537BEHAVIORAL CUSPS

lem-solving opportunities can remediatethat, and thereby increase the childrens rep-ertoires; when that happens, it widens therange of situations that evoke teaching from

the teachers. The result is new and improvedskills not specifically targeted by the initialprogram; Koegel and Koegel (1988) cite ty-ing shoes, buttoning clothes, and restaurantskills as examples. Similarly, Koegel and Frea(1993) report that effectively teaching stu-dents eye contact and appropriate facial ex-pressions may decrease some abnormal be-havior and increase effective conversation.To the extent that these collateral behaviorchanges prove to be important or introducethe organism to new shaping environments

that prove to be important, they are cuspsas well as pivotal behaviors. If, for example,the collateral behavior changes seem to beonly brief, stereotypic conversations aboutvery few topics, they remain collateral be-havior changes, but their importance to thechild or to others seems problematic, andthus they may not be cusps. Cusps are be-havior changes that systematically lead to ei-ther widespread further changes or to im-portant further changes.

Again, the criteria for importance are usu-

ally situational. Most often, they hinge onwhat the behavior changes are and on whattheir consequences are for that organism,not in their own right, but relative to whatthat organism wants, what its caretakers, ad-vocates, and teachers want for it, and whata disinterested audience sees as significantfor that organism, or for any organism intheir society or species. These wantingsmay be pragmatic, or they may reflect anallegiance, even an implicit one, to sometheory about what is important to any de-veloping organism. Behavior analysis is notsuch a theory, apart from its usual endorse-ment of evolution as an inevitable processand of survival as a near-universal reinforcerof exceptional importance. That is, behavioranalysis is a theory about how behavior is

changed, not about how it should bechanged. That it can be changed by proce-dures that are so prevalent in the natural

world, and that are so easily open to social

intervention, probably reflects great survivalvalue.Thus, cusps are behavior changes, some-

times simple, sometimes complex, that sys-tematically cause other, further, not formallyprogrammed behavior changes that are sig-nificant either because of their breadth orbecause of their importance to the organismor its species. That importance is seen some-times by the organism, or by parties con-cerned for that organism, or by its relevanceto the selection pressures of the environ-

ment, or all of those. Cusps often accom-plish that kind of extensive or important col-lateral behavior change because they increasethe organisms exposure to the relevantteaching contingencies.

Restated, the importance of cusps isjudged by (a) the extent of the behaviorchanges they systematically enable, (b)

whether they systematically expose behaviorto new cusps, and (c) the audiences view of

whether these changes are important for theorganism, which in turn is often controlled

by societal norms and expectations of whatbehaviors should develop in children and

whe n that developme nt should happen.Most of that is ultimately judged by survival,but ultimately is a long time and is ex-tremely difficult to predict in advance. It isthe third criterion, including our guessesabout survival, that often prompts us to seeonly certain behavior changes as develop-mental.

The cusp concept is focused on under-standing the importance of what happens af-ter any behavior change, in order to definedevelopment. Other approaches, by contrast,define development by asking what new lev-el of ability or complexity the behaviorchange represents. Yet, cusps can be simple:

Access to other environments sometimes re-


6/12


quires only a simple response, like dialingthe critical number, or keyboarding the crit-ical address, or extending the stimulus con-trol of an existing response. They can also

be as complex as the task analysis of conser-vation, seriation, transitivity, or self-instruc-tion. In other approaches, the ability to readmight be valued as developmental because ofthe time required to teach it, the extensiveskill it represents, or the mental functions itis inferred to represent. However, if teachingreading were to have little effect beyond theachievement of reading, it would, for thisbehavior-analytic view of development, beirrelevant to development; it would not be acusp. It would be typical of modern applied

behavior analysts to ask how to repair anenvironment in which reading did not leadto broad further changes. (It might be typ-ical of near-future applied behavior analyststo ask what behavior changewhat mediaskill?is, in that future world, better thanreading for producing those broad furtherchanges.)

As mentioned, cusps can range from quitelarge to quite small behavior changes. Anobvious example of a large cusp is general-ized imitation. An example of a small cusp

is seen in an anecdote from a parent rearinga child with profound retardation: Teachingthis child to manipulate the door latchesthat separated her from the outside fencedyard transformed her from a child who askedoften all day (and often unsuccessfully) fordoors to be opened for her into a child whocould manage them herself. The childs newskill greatly expanded her opportunities forlearning and activity from mainly indoorones. It obviously enhanced her control oversome of her daily life. It transformed herfamilys perception of her as an eternal prob-lem to a learner whose skill acquisitionscould improve everyones lifefrom some-one to be managed into someone who nowcould be taught more independence. A cusp

whose size is less easy to assess is chaining

the elements of verbal behavior (e.g., teach-ing the chunking of verbal messages; cf.Case, 1987). At the least, it transforms a lis-tener from one who must be spoken to with

slow-paced, one-word messages into onewho can respond correctly to ordinary sen-tences, which may not be seen as a very im-portant change. But, given enough of otherrelated skillsof other cusps passedit canalso transform that listener into an efficientstudent.

Normal children get through many cuspsto what follows in their various worlds, usu-ally by extensive if casual teaching (e.g., im-itation and spoken language), and aided byvarious skills acquired through prior cusps

that made them increasingly better at self-teaching (e.g., self-regulation). Less fortu-nate, less endowed, less skilled, and less well-taught children do not get through as manyof those cusps and become problems thatattract diagnostic labels and remedial teach-ing.

The point of these examples is that cuspscan vary in size, particularly in the length orintensity of their teaching programs, yethave similarly important consequences for

what can happen next. It is not their man-

agement, their complexity, or the complexityof the behavior they target but their behav-ior-change outcomes that define their im-portance. Thus, cusp transcendence is prag-matic, but pragmatics do not change thelaws of behavior or the principles of behaviormanagement. However, they may wellchange management tactics, because the na-ture of cusps is that the developing organ-isms situation changes in systematically im-portant ways.

SMALL CONVERGENCESOF TRADITIONAL AND

BEHAVIOR-ANALYTIC VIEWS

Organisms are always doing somethingand are always doing new things; there are


7/12

539BEHAVIORAL CUSPS

no holes in the stream of behavior (Bijou &Baer, 1961; Schoenfeld & Farmer, 1970;Skinner, 1953; Watson, 1926). The questionhas always been whether that stream has a

structure. Some developmentalists organizeit as a progression of stages, often accordingto what they call the complexity of behavior.They describe how behavior increases, notin amount but in complexity, during certainparts of the life span, from early and simpleto late and complex. In most arguments,that sequence is predictable and uniform.Thus, a stage of development is a portion ofthe organisms life, qualitatively differentfrom the preceding or subsequent stages,

whose content is often (but not necessarily)

described as a mental structure that guidesaction and is said to be universal, and is rel-evant to many outcomes, especially emo-tional, cognitive, and moral ones. Its timingis seen as modifiable, but only a little; andits sequence is seen as even more resistant tochange (e.g., Bickhard, Cooper, & Mace,1985; Flavell, 1982; Glasersfeld & Kelly,1982; Lerner, 1986; Overton & Reese,1973; Reese & Overton, 1970; Wohlwill,1973).

Stage concepts of development are often

challenged, even within the scientific com-munity that generated them. Piagets stagesof cognitive development (1971), Freudsstages of psychosexual development (1905),Kohlbergs stages of moral development(Kohlberg, Levine, & Hewer, 1983), and Er-iksons stages of psychosocial development(1950)four prominent exampleshavebeen criticized on many grounds, most of

which reflect the vagueness of three sets ofcriteria: those that define a stage; those thattell the theorist how many stages are neededto explain development; and those that de-fine the transition from one stage to the next(see Brainerd, 1978, for a heuristic exampleof these unresolved questions). These arecriticisms not of the stage strategy but of itstopical tactics. In effect, these criticisms as-

sume and applaud the stage strategy by ask-ing that it find better tactical criteria.

Some modern theories of development donot postulate a stage-specific mental struc-

ture that explains all developmental phe-nomena. Some theorists now see cognitivedevelopment, for example, as highly diverseand seamlessly continuous: Individuals usemultiple rules, strategies, hypotheses, and soforth, changing them from one kind ofproblem to the next; these structures rangesimultaneously from simple to complex; thecompetence of each one may change at anytime; and each one is more likely to be spe-cific to a small domain (e.g., speech percep-tion, reading, arithmetic, language, catego-

rization, or reasoning) rather than to be gen-eralized across them all (see Case, 1987;Fisher, 1980; Flavell, 1982, 1992; Howe &Pasnak, 1993; and Siegel, 1991, for reviewsof this shift in conceptualization).

The general stage concept is still used,even so. For example, Flavell postulates de-veloping capacities to process informationand to resist interference, which, if they ex-ist, should allow more complex cognitionacross all relevant domains (see Flavell,1982). Within a domain, though, it is levels

of skill competence rather than stages ofqualitative changes that are assumed to pro-ceed in an orderly sequence (Fisher, 1980;Fisher & Silvern, 1985; Siegler, 1981).

The thesis that developing an ability orcompetence will open a much larger realmto improvement is not new; like most the-oretical overreaches, it has seen its waves ofendorsement and rejection. As the 20th cen-tury began, educational psychologists oftensupposed that training any specific skill (e.g.,matching colored sticks) would educate thesenses and make them hospitable to manyuntrained discriminations, just as studyingany small discipline (e.g., Latin, mathemat-ics) would improve reasoning in general.Later, that thesis was refined: Not any train-ing or study would lead to generalized re-


8/12


sults; onlycertainkinds of training or studieswould do that. A specific ability would ben-efit a larger domain (mathematics would im-prove reasoning) only if both contained suf-

ficient common elements (see Thorndike,1903). That was the logic of transfer (e.g.,Grose & Birney, 1963). It automatically rec-ommended training in larger categories, soas to sample more of the elements that areoperative in the larger domains to be bene-fited. It also warned the teacher that the ben-efited domain would be no larger than thecommon elements justified. Clearly, the con-cept had moved behavior-ward. But it wasstill as vague as any stage theory in offeringcriteria for identifying common elements.

Thus, it was not long before a psycholo-gist like Ferguson (1954, 1956), apparentlyfollowing Spearman (1927) and Thurstone(1938), would see little use in constructs asgeneral as intelligence. These constructscould only denote subsets of more real abil-ities, which in turn were properties of theultimate reality, behavior. So Fergusons con-cept of development was to list the skill mas-teries that together would justify the termabilityand to catalog their transfer functions(generalizability) at different stages of learn-

ing and at different ages. The developmentalquestion had become: What prior learnedabilities transfer to what untaught abilities,and how, and under what conditions? Theconverse question became: What new abili-ties alter prior abilities, and in what ways?

Forty years later, cognitive scientistswould be asking: What prior abilities ,learned or otherwise, lead to what changesin development? They would answer thequestion of how by inferring cognitive me-diators such as memory access, informationorganization, inference itself, and strategiz-ing (Glaser, 1992, p. 249); they would an-swer the question of under what conditionsby inferring developing levels of function forthose inferred mediators.

From a behavior-analytic point of view,

the small domain to which these relativelynew cognitive-analysis tactics were applied

was admirable. Smaller arenas of analysis al-low a much more intimate interaction be-

tween research and data and allow more ofthe data to be experimental. More impor-tant, smaller domains of analysis allow, andalmost insure, at least a partial intersectionof the logic of behavior analysis and cogni-tive analysis: (a) We all analyze behavior,even when it is not the fundamental unit ofour theory; (b) behaviors are readily changedby environmental contingencies; and (c) weknow any behavior can contact different en-vironmental contingencies than other behav-iors do. These three points tell us that dif-

ferent behaviors can come under differentcontrol (even though some theories needsome very similar behaviors to be under sim-ilar control). To the extent that even similarbehaviors do come under different control,then an overarching stage-like organizationof great quantities of behavior is improbable,although not impossible to program. Our re-search ought to look first for regularity inmuch smaller domains, then seek experi-mental control of as much of that regularityas proves to be possible (and ethical), and

then ask if that control can be extended (ex-perimentally) to a domain large enough to

justify a stage concept.For behavior analysis, behavior classes as

large as intelligence have never seemeduseful, or even real. Response classes havebeen defined by the experimenters ability toprove that all members of the putative classare in fact under the same control (antece-dent, consequent, or both) and have beenunderstood by the experimenters ability tomake them. Similar response classes that re-sult from similar histories of programmingthen have been seen as possible events innatural development.

In behavior analysis, the stage conceptseems neither essential nor explanatory, butit is still heuristic. Bijou (1993, p. 46) argues


9/12

541BEHAVIORAL CUSPS

that it can guide analysis; he sketches a se-quence of foundational, basic, and societalstages, much as Kantor proposed in 1959(see Bijou, 1989, 1993). When the changes

described by a stage concept show great gen-erality across behaviors and contexts, formany children and for a specific period ofthe life span, then, and only then, does theconcept of stage become correspondinglyheuristic.

But, when a discipline knows, or thinksit knows, how to diminish or disassemble orhow to create or intensify some of the gen-erality described by the stage concept, and

when a discipline can do so by fairlystraightforward environmental interventions

(as has been done for at least some cases likeconservation skills, Kuhn, 1974, and gener-alized imitation, Baer, Peterson, & Sherman,1966), then the concept of stage becomescorrespondingly more fragile and arbitrary.Behavior analysis has always at least asked ifits processes could create or intensify or di-minish or disassemble that kind of general-ity, and has succeeded often enough to makethis argument viable.

In fact, learning to manage the detailedcomposition of stages may soon prove to be

more interesting than the generalities thestages describe. For stage theory, those gen-eralities are described rather than experimen-tally analyzed. By contrast, the managementof their components is almost always exper-imentally analyzed; that is what manage-ment means in behavior analysis. Then whynot shift interest to the often dramaticchanges in behavior that become possible

with experimental mastery of those compo-nents? For example, one way in which chil-dren expand their vocabularies is the dis-ambiguation effect described as part of themutual-exclusivity bias (Merriman & Bow-man, 1989). Around 2 years of age, mostchildren begin to learn new words when pre-sented with a novel name in the presence ofa novel object (one whose name has not

been learned) and a familiar object (onewhose name has been learned). These chil-dren typically select the novel object andthus learn the name of the novel object. Be-

fore the age of 2, children usually selectthe object whose name already controls theirbehavior. However, children as young as 2years also have demonstrated the disambig-uation effect when correction and reinforce-ment procedures are used. In behavior anal-ysis, this phenomenon has been experimen-tally investigated in persons with mental re-tardation, with both spoken and visualstimuli and with visual stimuli in matchingtasks (cf. Dixon, 1977; McIlvane & Stod-dard, 1981, 1985). It has been demonstrated

that learning by exclusion permits an eco-nomical way of expanding the repertoire ofindividualsa way that the teaching com-munity could use to produce almost error-free behavior changes, even when otherteaching methods have failed (e.g., de Rose,de Souza, & Hanna, 1996). Learning by ex-clusion is a cusp that along with other cuspsmay lead us to an understanding of the be-havior changes and the environments thatare required to produce the vocabulary ex-plosion typically seen at around 18 months

of age (Smith, 1926).Studies looking for cusps will eventually

produce a long list of organismenviron-ment interactions, some of small importancefor what can happen next, others of greatimportance for what can happen next, andstill others of importance conditional on

what other cusps have been attained. Thusa cusp may be universal, but it need not andrarely will be. Similarly, a cusp may have

wide generality, but need not. One childscusp may be another childs waste of time.

In metaphor, cusps often are steps in anorderly path. Perhaps more often they arelike the branches of a tree: They stem froman earlier branch or trunk, and new branch-es may stem from them, where their struc-ture in conjunction with the environment


10/12


allows for that. But their mutual order, size,and probability of twigs are not very thor-oughly predetermined. Sequences, whethernecessary or merely societal, can be essential

to this concept; but it is the cusps that needto be analyzed first. As we come to under-stand them, we will then be in a better po-sition to learn when their sequences are cru-cial or conditional.

As behavior changes that proved to becusps for one child or another, or many, arelisted, any reader is free to chunk that listaccording to the readers criteria, which maybe a predetermined notion of complexity, se-quence, or growth. Some readers no doubt

will chunk them exactly that way; others will

find a variety of alternative logics. However,a list of cusps, defined as they are here, is alist of teachable behaviors, a set of teachingprocedures that accomplish them, a shapingcommunity of reinforcement, and a descrip-tion of the systematic consequences of doingso, including the consequences of the con-sequences. Teachable cusps are susceptible toexperimental analysis, and experimentalanalysis allows us to identify their conse-quences. Nonteachable cusps are susceptibleonly to correlational analysis; correlational

analysis allows us to say only what their ac-companiments are. A truly developmentalanalysis needs more certainty about whatcauses what. An illustrative example of suchanalysis is the case of the disappearingstepping reflex. Newborns held upright withtheir feet on a surface display well-coordi-nated step-like movements; these responsesdisappear within the first few months andare seen again towards the end of the firstyear. These changes have been explained ascorrelates of the maturation of the voluntarycortical centers. We could suppose that thosecenters first inhibited subcortical or reflexivemovements and later facilitated them athigher levels of control (McGraw, 1943). Af-ter all, inhibition is what a center should doto make one of its skills disappear, and fa-

cilitation is what it should do to make thatskill reappear. This explanation remainedunchallenged for 40 years until Thelen andFisher (1982) demonstrated experimentally

that the disappearance of this reflex was dueto an increase in the babys weight and tothe changing mechanical demands of its pos-ture. They restored the stepping reflex bysubmerging infants in torso-deep warm wa-ter and inhibited it again by adding weights.Once again, the value of an inferred centralcontrol had varied inversely with the appli-cation of experimental analysis.

The cusp concept defined here is mostpowerful when it is limited to those changesthat can be experimentally taught and the

consequences of which can be experimental-ly verified. Correlational analyses that lookfor the sequelae of a cusp will not easily sep-arate cause and effect; experimental control

will be required to meet the definition. Astage theory may be as unverifiable as thetheorist wishes; then it can be made to em-brace everything the theorist needs to ex-plain. By contrast, to the extent that cuspassessment must be verifiable, cusp-based de-velopment will automatically be a set onlyof already-tested facts and procedures. This

cusp concept will not embrace everythingthat a developmental theory needs to ex-plain, because ethics and practicality bar ex-perimental analysis from many parts of thatdomain. Teaching reading to see its conse-quences fits the cusp concept; awaiting com-plete myelinization of the nervous system tosee its consequences does not. But if myeli-nization should ever become experimentallymanageable and ethically acceptable to man-age, it might then be tested for its cusp qual-ities.

REFERENCES

Allen, K. E., Hart, B. M., Buell, J. C., Harris, F. R.& Wolf, M. M. (1964). Effects of social rein-forcement on isolate behavior of a nursery schoolchild. Child Development, 35, 511518.


11/12

543BEHAVIORAL CUSPS

Baer, D. M., Peterson, R. F., & Sherman, J. A.(1966). The development of imitation by rein-forcing behavioral similarity to a model. Journalof the Experimental Analysis of Behavior, 10, 405416.

Baer, D. M., Rowbury, T., & Goetz, E. (1976). Be-havioral traps in the preschool: A proposal for re-search. In A. D. Pick (Ed.), Minnesota symposiaon child psychology (Vol. 10, pp. 327). Minne-apolis, MN: University of Minnesota Press.

Baer, D. M., & Wolf, M. M. (1970). The entry intonatural communities of reinforcement. In R. Ul-rich, T. Stachnik, & J. Mabry (Eds.), Control of human behavior: From cure to prevention (Vol. 2,pp. 319324). Glenview, IL: Scott, Foresman.

Bickhard, M., Cooper, R., & Mace, P. (1985). Ves-tiges of logical positivism: Critiques of stage ex-planations. Human Development, 28, 240258.

Bijou, S. W. (1989). Behavior analysis. In R. Vasta(Ed.), Annals of child development: Six theories of child development: Revised formulations and current

issues (pp. 6183). Greenwich, CT: JAI Press.Bijou, S. W. (1993). Behavior analysis of child devel-

opment. Reno, NV: Context Press.Bijou, S. W., & Baer, D. M. (1961). Child develop-

ment I: A systematic and empirical theory. NewYork: Appleton-Century-Crofts.

Brainerd, C. J. (1978). The stage question in cogni-tive-developmental theory. Behavioral and BrainSciences, 2, 173213.

Campos, J. J., Bertenthal, B. I., & Kermoian, R.(1992). Early experience and emotional develop-ment: The emergence of wariness of heights. Psy-chological Science, 3(1), 6164.

Case, R. (1987). Neo-Piagetian theory: Retrospectand prospect. International Journal of Psychology,

22, 773791.de Rose, J. C., de Souza, D. G., & Hanna, E. S.

(1996). Teaching reading and spelling: Exclusionand stimulus equivalence. Journal of Applied Be-havior Analysis, 29, 451469.

Dixon, L. S. (1977). The nature of control by spokenwords over visual stimulus selection. Journal of theExperimental Analysis of Behavior, 27, 433442.

Erikson, E. (1950). Childhood and society. New York:Norton.

Ferguson, G. A. (1954). On learning and humanability. Canadian Journal of Psychology, 8, 95112.

Ferguson, G. A. (1956). On transfer and the abilitiesof man. Canadian Journal of Psychology, 10, 121158.

Fisher, K. W. (1980). A theory of cognitive devel-opment: The control and construction of hierar-chies of skills. Psychological Review, 87, 477531.

Fisher, K. W., & Silvern, L. (1985). Stages and in-dividual differences in cognitive development.An-nual Review of Psychology, 36, 613648.

Flavell, J. H. (1982). Structures, stages, and sequencesin cognitive development. In W. A. Collins (Ed.),

The concept of development: The Minnesota sympo-sia on child psychology (pp. 128). Hillsdale, NJ:Erlbaum.

Flavell, J. H. (1992). Cognitive development: Past,present, future. Developmental Psychology, 28(6),9981005.

Freud, S. (1905). Three essays on the theory of sexuality.London: Hogarth Press.

Glaser, R. (1992). Learning, cognition, and educa-tion: Then and now. In H. L. Pick, P. V. D.Broek, & A. D. C. Knill (Eds.), Cognition: Con-ceptual and methodological issues (pp. 239265).

Washington, DC: American Psychological Asso-ciation.

Glasersfeld, E., & Kelly, M. (1982). On the conceptsof period, phase, stage, and level. Human Devel-opment, 25, 152160.

Grose, R., & Birney, R. (Eds.). (1963). Transfer of learning. Princeton, NJ: Van Nostrand.

Harris, F. R., Johnston, M. K., Kelley, C. S., & Wolf,

M. M. (1964). Effects of positive social reinforce-ment on regressed crawling of a nursery-schoolchild. Journal of Educational Psychology, 55, 3541.

Howe, M., & Pasnak, R. (1993). Shifting concep-tions of cognitive development. In M. L. Howe& R. Pasnak (Eds.), Emerging themes in cognitivedevelopment: Vol. 1. Foundations (pp. 267276).New York: Springer-Verlag.

Kantor, J. R. (1959). Interbehavioral psychology. Bloo-mington, IN: Principia Press.

Koegel, R., & Frea, W. (1993). Treatment of socialbehavior in autism through the modification ofpivotal social skills. Journal of Applied Behavior

Analysis, 26, 369377.

Koegel, R., & Koegel, L. (1988). Generalized respon-sivity and pivotal behaviors. In R. H. Horner, G.Dunlap, & R. L. Koegel (Eds.), Generalizationand maintenance: Life-style changes in applied set-tings (pp. 4166). Baltimore: Paul H. Brookes.

Koegel, R., Koegel, L., & Schreibman, L. (1991). As-sessing and training parents in teaching pivotalbehaviors. In R. J. Prinz (Ed.), Advances in behav-ioral assessment of children and families (Vol. 5, pp.6582). London: Jessica Kingsley.

Kohlberg, L., Levine, C., & Hewer, A. (1983). Moralstages: A current formulation and a response to crit-ics. Basel, Switzerland: Karger.

Kuhn, D. (1974). Inducing development experimen-tally: Comments on a research paradigm. Devel-

opmental Psychology, 10(5), 590600.Lerner, R. M. (1986). Concepts and theories of human

development (2nd ed.). New York: RandomHouse.

Martin, G., & Pear, J. (1978). Behavior modification:What it is and how to do it. Englewood Cliffs, NJ:Prentice Hall.

McGraw, M. B. (1943). The neuromuscular matura-


12/12


tion of the human infant. New York: ColumbiaUniversity Press.

McIlvane, W. J., & Stoddard, L. T. (1981). Acquisi-tion of matching-to-sample performances in se-vere mental retardation: Learning by exclusion.

Journal of Mental Deficiency, 25, 3348.McIlvane, W. J., & Stoddard, L. T. (1985). Complex

stimulus relations and exclusion in severe mentalretardation. Analysis and Intervention in Develop-mental Disabilities, 5, 307321.

Merriman, W. E., & Bowman, L. L. (1989). Themutual exclusivity bias in childrens word learning.Monographs of the Society for Research in Child De-velopment, 54 (34, Serial No. 220).

Novak, G. (1996). Developmental psychology: Dynam-ical systems and behavior analysis. Reno, NV: Con-text Press.

Overton, W., & Reese, H. (1973). Models of devel-opment: Methodological implications. In J. R.Nesselroade & H. W. Reese (Eds.), Life-span de-velopmental psychology: Methodological issues (pp.6586). New York: Academic Press.

Piaget, J. (1971). Biology and knowledge. Chicago:University of Chicago Press.

Reese, H. (1991). Contextualism and developmentalpsychology. In H. W. Reese (Ed.), Advances inchild development and behavior(Vol. 23, pp. 187230). New York: Academic Press.

Reese, H., & Overton, W. (1970). Models of devel-opment and theories of development. In L. R.Goulet & P. B. Baltes (Eds.), Life-span develop-mental psychology: Research and theory (pp. 115145). New York: Academic Press.

Rosales-Ruiz, J., & Baer, D. M. (1996). A behavior-analytic view of development. In E. Ribes & S.

W. Bijou (Eds.), Recent approaches to behavioraldevelopment (pp. 155180). Reno, NV: ContextPress.

Schlinger, H. (1995). A behavior analytic view of childdevelopment. New York: Plenum Press.

Schoenfeld, W. N., & Farmer, J. (1970). Reinforce-ment schedules and the behavior stream. In W.N. Schoenfeld (Ed.), The theory of reinforcementschedules (pp. 215245). New York: Appleton-Century-Crofts.

Siegel, L. (1991). On the maturation of developmen-tal psychology. In F. S. Kessel, M. H. Borstein, &

A. J. Sameroff (Eds.), Contemporary constructionsof the child: Essays in honor of William Kessen(pp.251264). Hillsdale, NJ: Erlbaum.

Siegler, R. (1981). Developmental sequences within

and between concepts. Monographs of the Societyfor Research in Child Development, 46(2), 113.

Skinner, B. F. (1953). Science and human behavior.New York: Free Press.

Smith, M. E. (1926). An investigation of the devel-opment of the sentence and the extent of vocab-ulary in young children. University of Iowa Studiesin Child Welfare, 3, 5.

Spearman, C. (1927). The abilities of man: Their na-ture and measurement. London: Macmillan.

Spiker, C. C. (1966). The concept of development:Relevant and irrelevant issues. In H. W. Stevenson(Ed.), Concept of development: A report of a con-ference commemorating the 40th anniversary ofthe Institute of Child Development, University of

Minnesota (pp. 4054). Monographs of the Societyfor Research in Child Development, 31 (5, SerialNo. 107).

Stokes, T. F., & Baer, D. M. (1977). An implicittechnology of generalization. Journal of AppliedBehavior Analysis, 10, 349367.

Thelen, E., & Fisher, D. M. (1982). Newborn step-ping: An explanation for a disappearing reflex.Developmental Psychology, 18, 760775.

Thorndike, E. L. (1903). The psychology of learning.In Educational psychology (Vol. 2, pp. 357464).New York: Teachers College, Columbia Universi-ty.

Thurstone, L. (1938). Primary mental abilities. Chi-cago: University of Chicago Press.

Watson, J. B. (1926). What the nursery has to sayabout instincts. In C. Murchison (Ed.), Psycholo-

gies of 1925 (pp. 135). Worcester, MA: ClarkUniversity Press.

Wohlwill, J. H. (1973). The study of behavioral de-velopment. New York: Academic Press.

Received August 9, 1996Initial editorial decision September 27, 1996Final acceptance April 3, 1997

Action Editor, David P. Wacker

Documents

B - ROSALES-RUIZ,(1997O) - Behavioral Cusps, A Developmental and Pragmatic Concept (Olhar)