THEORETICAL ARTICLE

The Role of Contingency Adduction in the Creative Act

Nolan Williams1

Accepted: 8 October 2020# Association for Behavior Analysis International 2020

AbstractThis article discusses the potential role of contingency adduction in creative behavior. Some have characterized creativity as thestudy of generativity. Generativity is the investigation of procedures that result in the occurrence of untrained, often composite,patterns from earlier trained components. An increasing number of applied programs are attempting to apply generative proce-dures in their design. Headsprout Early Reading®, for example, explicitly employed generative procedures to teach reading.There remains a lack of understanding about the role contingency adduction plays in the generative process. Contingencyadduction is defined when patterns shaped under one context are recruited by contingencies in another context for which thepattern was not originally shaped. Adduced patterns may be new sequences of repertoires, the combination of repertoires, or therepertoire may acquire a new function. The moment of reinforcement of these new patterns from previously established patternsmarks the moment of adduction. Thus, procedures that make such selection more likely may be fundamental to encourage whatmight be called creative behavior. Examples and nonexamples of contingency adduction involving both verbal and nonverbalprocedures in both animals and humans will be described, and their implications noted.

Keywords Adduction . Contingency Adduction . Creativity . Variability

Most societies value creative, novel, and innovative solutionsin art, sport, business, entertainment, and science and technol-ogy. An adequate theory of human behavior must address allaspects of human functioning—including behavior called cre-ative. Of course, creativity is an important topic to behavioranalysts. By extension, if behavior analysts could reliably en-gineer such behavior, the social value of our science wouldincrease.

To guide behavior analysts’ efforts in developing proce-dures to encourage creative performance, behavior analystsmust first determine what sorts of behaviors qualify as crea-tive. This determination may be aided by an analysis of theconditions which control the use of the tact “creative” in oursociety. To establish the boundary of the tact “creative,” itmay be useful to first identify the conditions which do notoccasion the tact’s use. When do verbal communities rein-force the use of tacts such as “uncreative,” “application,” or“derivative?” For example, saying dog in the presence of a

new dog, even though it has not previously occurred, is nottypically considered creative.

Popular media provides numerous examples of uncreativebehavior. Critics pan screenwriters for recycling plot lines.Fans of well-established bands, authors, and comedians groanwhen their favorite entertainers repeatedly produce materialwith similar sounds, themes, and punchlines. The electoratehas little patience for politicians who persist in courses ofaction that have failed to produce solutions to major societalproblems. Romantic partners often brandish the accusation ofthoughtlessness and lack of creativity when a once well-received anniversary gift becomes an annual or predictableoccurrence.

The common denominator in each example above isbehavior or a repertoire that was reinforced under one setof conditions is simply repeated under new conditions. Anadditional commonality is that the repeat performance isemitted under contingencies where repetition is unlikely tobe reinforced. These two conditions seem to define what itmeans to be uncreative. If behaviors that are extended toand repeated across sometimes novel situations are notcreative behaviors, under what conditions do we use thetact “creative?” What follows is a discussion of perfor-mances the verbal community often tacts as creative.These examples serve as the basis of an attempt to abstract

* Nolan [email protected]

1 University of North Texas, 103 East Park Place, JeffersonvilleIndiana 47130, USA

https://doi.org/10.1007/s40732-020-00440-z

/ Published online: 1 November 2020

The Psychological Record (2021) 71:543–551

what common features of these performances might con-trol that tact “creative.”

Artistic behavior is often called creative. Painters who em-ploy new brush strokes or who depict a subject in a novel wayare often said to be creative. Musicians who create novel ar-rangements of notes are said to be creative. Writers who tellnew stories, or who tell old stories in new ways are said to becreative. The behavior of scientists, engineers, and inventorsstand out as creative, when those behaviors result in novelsolutions to various problems.

Each of these instances of creativity seem to share an ele-ment of novelty. Each example involves a new topography ortopographies of behavior. These novel topographies must alsomeet some criterion for reinforcement, and in so doing, meetthe requirement in ways never directly reinforced. In addition,the label of creativity seems to be reserved for novel acts thatmeet criterion for reinforcement in a socially acceptable man-ner. That is, these novel behaviors do not also meet require-ments for punishment—such as hallucinations and other pat-terns considered pathological. The tact “creative,” therefore, isa pattern or patterns of behavior that meet criterion for rein-forcement and for which the individual has no direct history ofreinforcement. And these patterns also fail to meet criterionfor punishment.

Behavior analysts assume all behavior is the product oflearning history, contingencies of reinforcement, and stimuluscontrol. Given these assumptions, the emergence of noveltopographies, or topographies that have never produced rein-forcement for the organism before, presents a real mystery forbehavior analysts. The question that behavior analysts mustanswer is, what variables are responsible for the novel behav-iors that seem to be prerequisites for creative acts?

Variability is a critical component of any novel or creativebehavior, and therefore any behavior analytic account of cre-ativity must address the sources of such variability. One strat-egy for understanding variability in operant behavior is exem-plified in the literature on the reinforcement of variability(Page & Neuringer, 1985; Neuringer, 2002). Others have sug-gested that extinction, rather than reinforcement, is the likelysource of variability (Holth, 2012a; Kieta, 2017). In eithercase, the emerging behavior analytic account of creativitymay benefit from considering the of role contingency adduc-tion in the generation of novel behavior and in acts of creation.This article reviews the concept of contingency adduction andprovides examples of the role adduction plays in the produc-tion of novel instances of behavior.

What Is Contingency Adduction?

Contingency adduction describes a class of outcomes obtain-ed when patterns shaped in one context are recruited by con-tingencies in contexts different from the one for which the

pattern(s) was originally established (Andronis, Layng, &Goldiamond, 1997; Layng, Twyman, & Stikeleather, 2004).That is, when an organism is exposed to new contingencies ofreinforcement, some aspect or aspects of the new context willoccasion a response, or a combination of previously learnedresponses, that meets this new criterion for reinforcement,selecting this new stimulus control topography and/or patternof behavior (Ray, 1969). This selection of new behavioralvariants from previously shaped patterns may be a primarysource of what might be considered creative behavior.

It is important to distinguish between contingency adducedbehavior and shaped behavior. Shaping involves the selectionfrom variation along a single dimension of behavior that hasnot been previously reinforced, and that occurs in the samecontext. Consider the task of shaping a dog’s approach behav-ior. Each step in the shaping plan involves the samedimension—the movement of the dog toward the trainer—and the criterion for reinforcement is changed gradually, re-quiring successive decreases in the total distance between thetrainer and the dog.

Adduction describes the moment when “previouslyshaped” or “preestablished” behavior—often along multipledimensions—occurs in a new context and meets a new con-tingency requirement. For instance, after a trainer shapes a dogto approach on the command “come” and then shapes the dogto raise its right paw in the presence of a lifted right fist, theresponses may be combined or blended by presenting thestimulus for “come” along with a raised right fist that controlsraising the right paw. The combined behavior may result inthe dog limping towards the trainer. If the new behavior isreinforced, this first reinforcement of the new pattern repre-sents the moment of adduction.

Another important note, the prevailing contingencies areresponsible for the adduction of behavior. Organisms do notadduce behavior. Organisms do not select their own physicaltraits; the organism’s ecology selects its physical traits vianatural selection. Likewise, an organism is no more capableof adducing a new response into its behavioral repertoire thanit is capable of evolving an extra arm.

Several different outcomes can be categorized as adduc-tion. Different component responses can blend to form newresponse topographies. New sequences of responses can occurand be established as new functional units. New antecedentvariables can acquire control of the response. And new con-sequential and motivational variables can acquire control ofthe response. In summary, contingency adduction is charac-terized by:

& The recruitment of repertoires established under one set ofconditions by contingencies operating under another set ofconditions

& The patterns come “preshaped” or from species typicalbehavior

544 Psychol Rec (2021) 71:543–551

& May be a single pattern or combinations of patterns& Meets a new contingency requirement& May be derived from a variety of sources including

– Resurgence– Combining stimuli– Separating stimuli– Impossible discrimination– Schedule induced or adjunctive behavior (Andronis et al.,

1997; Layng et al., 2004)

Contingency Adduction as the Engine of Creativity

Given the generative nature of the outcomes of adduction,procedures that make adduction more likely could be pivotalto understanding and occasioning creative behavior. The roleof adduction in creativity can best be illustrated using exam-ples of adduced, creative performances of both nonhumananimals and humans. In some cases, the organisms understudy emitted previously learned responses in novel contexts(Andronis et al., 1997; Layng et al., 2004). In other examples,the organism emitted novel topographies (Schiller, 1957;Pryor, Haag, & O’Reilly, 1969; Epstein, 1985). In all cases,the responses met new contingency requirements. In each ex-ample, it is the first occasion where the previously establishedresponse or composite met the new contingency requirementthat represents the moment of adduction. After the initial oc-casion, subsequent instances of the adduced response repre-sent a simple operant. Contingency adduction simply refers tothe procedures used to induce novel performance and the mo-ment that such performances are selected via reinforcementand then become part of the organism’s repertoire. Adductionis a principle that may underlie much of the behavior consid-ered creative, and one that if overlooked makes the sources ofcreative performance seem mysterious.

Adduction in Nonhuman Animals

Creative Symbolic Aggression in Pigeons

The first laboratory investigation of adduction occurred at theBehavior Analysis Research Laboratory at the University ofChicago (Andronis, 1983; Andronis et al., 1997). Androniset al. (1997) report procedures that effectively adduced novelsymbolic aggression in pigeons from nonsocial componentbehaviors. Even though it had no effect on their own workschedule, pigeons came to peck keys that increased the workrequirements for a bird in an adjacent chamber, visiblethrough a transparent acrylic wall. By pecking a key, pigeons

produced a houselight change and increased schedule require-ments, from which they themselves had consistently escaped,for a bird in the adjacent chamber. Further, the birds wouldswitch to whichever side key produced that change. This wasachieved without direct training of the pattern and occurred asa novel recombination of earlier trained nonsocial componentbehaviors that involved:

1) Responding under three schedule values FR-10, FR-50,and FR-100 correlated with red, white, and green house-light colors, respectively (a three-ply multiple schedule)on a food key located above the food hopper;

2) Pecking transparent side keys mounted in a transparentacrylic wall to change the houselight color from white tored and lower schedule requirements

3) Switching between side keys when a peck to a side keyproduced a change to a green houselight that indicted anincreased schedule requirement to FR-100

The investigators then placed the birds in a chamberadjacent to the one in which they were initially trained;the original training chamber was clearly visible and emp-ty. Food key pecking was initially maintained by a fixed-ratio schedule. Pecks to the side keys quickly extinguishedonce lights changed in the empty adjacent chamber but hadno effect on their own work requirement. Side key peckingdid not return when a bird was subsequently placed in theadjacent chamber. The food key schedule was changed toFI-40s, a reinforcement schedule that reliably produces ag-gression in pigeons when a conspecific is present. Thebirds attempted to physically attack the adjacent bird butwere prevented from doing so by the acrylic wall. Thecomplex symbolic social pattern arose not as a result ofdirect training or shaping. Instead, it was a function ofexisting nonsocial behavioral components being adducedinto a symbolic, social pattern after the acrylic wallprevented physical attacks.

After 10 sessions, two clear patterns emerged. First, all fourreferent birds showed scalloping and earned food at the maxi-mum rates allowed by the FI schedules. Second, all four refer-ent birds consistently pecked whichever switching key raisedthe schedule requirement and changed the houselight color togreen (the FR-100) for the conspecific pigeon. Further, thepigeons tracked changes in the switching keys so that when akey failed to produce the schedule increase, they switchedresponding to the other key. At no time did pecking the sidekeys change their own work requirements. Side key peckingwas maintained by the change in houselight color and the sub-sequent increase in schedule requirement for another bird. Notaware of the experimental conditions, one might say the birds“creatively” used the experimental arrangements to deliver anovel attack they could not otherwise accomplish.

545Psychol Rec (2021) 71:543–551

Creative Problem Solving in Pigeons

Epstein (1985) provided another excellent example of adduc-tion in his famous box and banana study. This study demon-strates an example of adduction where the outcome is a newauto-chained sequence of responses that meets a new contin-gency requirement. In this study, pigeons learned to peck abanana to produce food. Once the pecking response was well-established, the experimenter attached the banana to the ceil-ing of the chamber, and pecks made while flying or jumpingfailed to produce food, resulting in the extinction of thoseresponses. At the same time, the experimenter added a boxto the chamber that was the right height for the pigeons tostand on and reach the banana. Epstein found that the birdsdid not make use of the box in their attempts to solve theproblem. After these observations, he removed the banana.Then he trained the pigeons to move the box to specifiedpoints in the cage. In particular, the birds learned to movethe box towards different colored dots in the chamber; placingthe box on the dot was reinforced with food. Epstein alsotrained the pigeons to climb on top of the box to produce food.Once the birds reliably pushed the box and climbed on top ofthe box, the banana was reintroduced. This time, the pigeonspushed the box under the banana and pecked the banana.

In this example, pecking the banana was made highly po-tent by food deprivation. This allowed for the auto-chaining ofthe pigeon’s response. That is, the deprivation combined withthe bird’s history with the banana resulted in the potentiationof reduced proximity to the banana as a conditioned reinforc-er. Thus, moving the box and then jumping atop was beingadduced moment by moment; each time the pigeon moved thebox and jumped on top, it resulted in a decreased proximity tothe banana. Even though the initial movement did not bringthe bird into contact with the banana, pushing and jumpingwere likely adduced by the reduction in proximity, and thuscontinued. With each push the bird got closer. Thus, adduc-tion selected the auto-chaining that eventually resulted in a“creative” solution to the problem, which was thenreinforced, and so adduced, by the delivery of food.

Creative Porpoises

Contingency adduction may also have been an importantcomponent of the creative behavior reported by Pryor et al.(1969) in their report on the creative porpoise. In this study,the experimenters trained two porpoises to perform a novelresponse at the beginning of each new training session. Theexperimenters reinforced only one response per session. Theybegan by reinforcing species typical responses, one at a time,until they had exhausted the animals’ repertoires. Each ses-sion, the porpoises tended to fall into stereotypic responsepatterns. When the animals became “stuck” on a response,the experimenters shaped a novel topography, usually a blend

of the previously reinforced topographies. Once a new topog-raphy was created, the trainers reinforced it several times toensure that the response was “strong.” After 16 sessions undersuch contingencies, the experimenters observed the emissionof novel responses from the animals (Pryor et al., 1969). Fromsession 16 on, the criterion for reinforcement was the produc-tion of a novel response at the beginning of each session.Subsequent reanalysis of the experiment (Holth, 2012b) sug-gests that the variation or novelty in the pattern may not havebeen shaped, but instead occurred because of prolonged pe-riods of extinction. Once the variation occurred it wasreinforced.

These criteria for reinforcement resemble the lag schedulesin Neuringer’s research (Page & Neuringer, 1985) in so far astopographies reinforced in previous sessions were not candi-dates for reinforcement in future sessions, but variants are.This schedule differed in that each new session required anew variation, at no time (after session 16) did the trainersreinforce responses from previous sessions.

In general, novel responses appeared after a few of thepreviously reinforced responses underwent extinction. Oncethe novel response (i.e., standing on the tail and spitting water)was reinforced, that same response was emitted exclusivelyfor the duration of the session. These novel responses oftenconsisted of components that had been previously learned andrecombined. The moment where the new response was rein-forced represents the moment of adduction. The new responsecombination was not shaped, it was “adduced” or selectedfrom previously established patterns by reinforcement andbecame established as part of the organism’s repertoire. In thiscase, the key to establishing creative performance in the por-poise was to expand the repertoire, and to utilize extinction toinduce variability in responding.

Creative Tool Building in Primates

Another example of contingency adduction comes from PaulSchiller’s (1957) work with apes and tool use. In these studies,the experimenter put chimpanzees into enclosures and placedfood just out of their reach, on the outside of the enclosure.Inside the cages, the chimps had two hollow sticks that couldbe connected. The solution to the problem was to connect thesticks and use them to pull the food into the cage. Schiller(1957) conducted several experiments to isolate the require-ments for the chimps to be able to successfully complete thetool use problem.

Adult apes in such environments, with no specific training,invariably solve this problem. However, young primates ex-posed to the same situations invariably fail to solve the prob-lem. Schiller (1957) questioned this discrepancy and begansearching for an explanation for why young primatesstruggled to solve problems that seemed easy for older ones.He formulated two general hypotheses. One hypothesis

546 Psychol Rec (2021) 71:543–551

suggested that the chimps needed some minimal set ofexperiences to solve the problem. The other hypothesispostulated that the problem solving was a product of somematurational factors interacting with instincts. Schiller(1952) was quickly able to rule out an explanation based uponthe maturational factor. Thus, he began his search for what hecalled the “necessary general experiences” that were crucialfor the final performance (Schiller, 1957).

Schiller described several behaviors that needed to be“readily available” to the apes before they could solve theirproblem. He discovered that many of these behaviors werespecies typical and developed during periods of free play,where the chimps manipulated the sticks in a wide variety ofways. Schiller found three critical components for producing asolution. First, the animals must learn to fluently connectsticks, one inside the other. Second, the animals needed todevelop skills manipulating the stick side to side, up anddown, and pointing the stick in various directions. Finally,the chimps needed to experience moving objects closer tothem using shorter sticks within their enclosures. If the chimpsfailed to demonstrate any of these responses prior to beingexposed to the problem, the final performance would not oc-cur (Schiller, 1957).

Schiller also noted the need to change contingency require-ments to adduce tool use. The chimps would consistently tryto reach for the food with their hands, with shorter sticks, andwould sometimes try throwing things at the food before thefinal solution emerged. If any of these operants succeeded,tool use did not emerge (Schiller, 1957).

In these studies, the first time the primates connected thesticks and pulled the bananas into the cage successfullymarked the moment of adduction. This example illustrates atype of creative behavior where previously trained responsesare resequenced. What Schiller described is analogous to re-quirements for some forms of generative behavior. That is,teachers must train component behaviors to fluency, makethose behaviors potent at the same time, and extinguish ormake unavailable previously established patterns.

Examples of Adduction in Human Animals

Creative Discovery in Headsprout Early Reading

The designers of the Headsprout Early Reading program usedan oddity to sample and the combined stimulus procedure in adiscovery exercise (Layng et al., 2004) to produce adductionof letter/sound combinations by children who could not pre-viously make such responses. In the oddity procedure, theexperimenter trained the subject to respond away from a pre-viously learned stimulus. Once this stimulus consistently oc-casions responding to the “other” stimulus, novel “other stim-uli” may be introduced and the previously established

response can enter new contingencies with different anteced-ent and consequential stimuli. A learner who has previouslylearned to click on sn when hearing the phonetically pro-nounced “sn” using a computer mouse, is presented with thenever before encountered letter n displayed alongside sn. Thelearner hears, click on the sound that is not “sn.” The learnerclicks on n and hears yes “n” (phonetically pronounced). Nextthe learner sees n, sn, and another letter displayed and hearsclick on “n.” The learner clicks on n in the presence of “n” andhears yes “n,” which marks the moment of adduction. Thespoken “n” now occasions selecting “n” from a three-letterarray.

In the combined stimulus procedure, the experimenters firsttaught nine different single letter/sound combinations. In thenext phase experimenters presented the single letters as twoletter blends. During each trial, a narrator pronounced one ofthe four blends and asked the learners to click on the corre-sponding letter blend. Without any training or previous expo-sure to the blends, the students selected letter blends thatcorresponded to the sound. In this example, experimenterstaught learners the needed components for the composite“blend” to occur. The resulting blends were adduced the mo-ment they produced reinforcement (Layng et al., 2004).

These examples demonstrate the important role that stimu-lus control can play in occasioning creative behavior. Theresearchers engineered either a response from a separatedstimulus or produce a composite performance by combiningstimuli that controlled the component responses. These proce-dures add to the behavior analytic account of the origins ofvariability in behavior.

Creativity at Morningside Academy

Morningside Academy is a laboratory school for typicallydeveloping elementary and middle school students with aver-age to above average intelligence test scores and who may ormay not have a diagnosed learning disability. Morningside isnot a school for students with significant emotional or behav-ioral problems or autism spectrum disorder diagnoses.Morningside Academy primarily serves students who struggleto acquire foundational skills in math, reading, and writing aswell as critical thinking and problem-solving skills (Johnson& Street, 2020). Morningside Academy’s model of generativeinstruction is an example of a program utilizing adduction(Johnson & Street, 2004). Morningside Academy focuses onthe identification of minimal repertoires or generative sets thatare most likely to recombine when contingencies in the learn-ing environment change. In this model, educators identify theminimal components that make up more difficult compositetasks, and they introduce them using either Mathetics or DirectInstruction (Gilbert, 1962; Adams & Engelmann, 1996).Then, students practice the component material until fluent.Once students demonstrate fluency, teachers introduce

547Psychol Rec (2021) 71:543–551

application situations that are carefully designed to make eachcomponent likely, but that require more complex compositesto meet the contingency requirement. These arrangements typ-ically result in the adduction of new patterns (Johnson &Layng, 1992; Johnson & Street, 2004).

Johnson and Layng (1992) provide an example of one ofthe programs from the model. Four students completed assess-ments of their ability to solve word problems involving frac-tions. None of the four students answered more than half ofthe 14 questions correctly on the assessment. Others answeredas few as three questions correctly. Different assessmentsshowed that the students also struggled with solving whole-number word problems as well as fraction computations. Inthe intervention program, the teachers focused exclusively onteaching component skills; the actual performance of solvingfraction-word problems was not taught.

Once component training was complete, the studentsshowed impressive gains. The worst performance of the fourstudents on a follow up 14-item assessment of the fractionalword problems was one incorrect response. These gains wereachieved exclusively by strengthening component skills andwere not the product of shaping by successive approximations(Johnson & Layng, 1992).

In this example, the experimenters identified two compo-nents (fractional computations and solving whole-numberword problems) needed to produce a more complex composite(solving fraction-word problems) and taught them to fluency(for component and composite analysis, see Binder, 1993;Weiss, 2001). Then, in a situation where both componentsneeded to be applied to solve a problem, the two componentsblended, and the resulting performance was adduced. Thisexample is another reminder of the importance of having theproper component skills to encourage adduction. Much of thework regarding generative sets focuses on finding such com-ponents to accelerate the learning process via adduction.

Occasioning Creative Behavior

There are several practical applications for adduction proce-dures. Pursuits in education, art, athletics, and science repre-sent just a few of the areas where adduction procedures can beused to occasion creative performance. The final section ofthis report will review some of these procedures.

Robert Epstein has contributed enormously to proceduresin this area. He lists four important competencies for improv-ing creative performances and making the adduction of novelrepertories more likely (Epstein, 1999). The first of these com-petencies is capturing, which includes a variety of methods forrecording new ideas as they come to you (a way to keepadduced behaviors in the repertoire). A second competencyis to seek out tasks that require performance beyond your skilllevel. These situations make multiple behaviors probable atthe same time, which increases the likelihood of new blends of

behavior being emitted. Some of these new blends may beuseful and become candidates for adduction by the prevailingcontingencies. Another of Epstein’s suggestions includesbroadening skills and knowledge. This encourages adductionby adding components to your repertoire that could later re-combine into a novel, creative performance. Diversity in therepertoire increase the likelihood of interesting and creativeblends that could then be adduced. Epstein’s fourth suggestionis to regularly change the physical and social environment,because new stimulus arrangements encourage new combina-tions of behavior to compete with one another.

Epstein’s four competencies increase the likelihood of theemergence and adduction of creative performance in any ofthe fields described previously. These represent excellentstrategies for individuals looking to boost their own creativeperformance. What about strategies for those tasked with fos-tering creativity in others? Teachers, trainers, and coachescould modify Epstein’s competencies so that they expose theirlearners to such contingencies. But other strategies also exist.

Presentation of multiple cues was frequently cited in theprevious examples of adduced creativity. This is a relativelyeasy way to produce new blends of behavior. Such strategiescould usefully be employed to teach complex athletic skills.For example, a coach could teach two component skills tofluency, such as running a post route in football and catchinga thrown football. Then, by combining the cues (i.e., “post”and a thrown ball) the young receiver would be more likely tosuccessfully catch the pass while running the post route. Thissame strategy could be used for a variety of athletic endeavors.In addition, multiple instructional stimuli could be presentedto an artist simultaneously, to induce novel brush strokes,color combinations, etc. Similar strategies could be imple-mented in music, dance, or even in scientific endeavors.

When teachers and trainers find themselves in situationswhere a creative response is essential, but they are unsure ofwhat the outcome needs to be, strategies can be used to encour-age maximum creative performance. Along the lines of Epstein’ssecond and third competencies, creating environments where oldpatterns stop producing reinforcement may lead to variants thatcan be reinforced. Much like the attacking pigeons in theAndronis et al. (1997) experiment, the primates in Schiller’sstudies, and the porpoises in Pryor et al.’s (1969) report, ifhumans continue to be reinforced for engaging in patterns thatpreviously produced reinforcement they are unlikely to emit anynovel behavior. In addition, new skill development must be en-couraged. Like Epstein and Andronis et al.’s pigeons, thelimping dog, and the children at Morningside Academy, onemust have the component skills needed for a creative perfor-mance in the repertoire before composite blends emerge(Epstein, 1985; Andronis, Layng, & Goldiamond, 1997;Johnson & Street, 2020). When those seeking to engender crea-tive performance in others do not know what the compositeperformance will need to be, they may be best served to seek

548 Psychol Rec (2021) 71:543–551

to expand the performer’s skills, providing as many behavioralcomponents as possible. The more component skills in the rep-ertoire, the more likely a novel recombination will occur thatmeets the contingency requirement. This may come in the formof practicing new skills, not typically thought necessary, it mayinclude trainings in the work places that focus on an unusual setof skills, it could come about by providing more opportunitiesfor continued education, collaborations with others who ap-proach the same problem from a different perspective or a slight-ly different problem with key similarities, to name just a fewpossibilities.

Finally, those delivering reinforcement should establishcontingencies that reinforce and capture any novel behaviorthat occurs. Each newly adduced pattern becomes part of theorganism’s repertoire and in turn can blend with other re-sponses, which may lead to the creative solution needed.

Areas for Future Research

Contingency adduction remains a rich area of research forbehavioral scientists. A few areas of investigation will bediscussed in this section, with an emphasis on questions withimplications for creativity. There are interesting questions forboth applied and experimental investigators, the answers towhich will extend our understanding of creative, novel, andgenerative behavior.

One interesting area of research is the investigation of therelationship between adduction, creativity, and fluency.Fluency is defined as the rate of performance that makes skillsuseful in everyday life and remembered, even after a signifi-cant period of no practice (Johnson & Street, 2020; Binder,1987, 1988; Haughton, 1972). It has long been known thattraining component skills to fluency expedites the learning ofcomposite behaviors made up of those components(Haughton, 1980; Gagne & Foster, 1949). But many questionsremain in terms of fluency and its relationship to novel, gen-erative, and creative behavior.

One question that might be asked is whether there can betoo much fluency. There are studies that suggest that in somecircumstances, less firmly established performance might bepreferable (Williams, Granzin, Engelmann, & Becker, 1979;Johnson & Layng, 1996). Is there a level of fluency at which aparticular repertoire is less likely to be adduced into a newcontingency, thus inhibiting creative performance? Couldhighly fluent repertoires block the adduction of less fluentrepertoires? If so, under what conditions might these phenom-ena occur?

On the other hand, many questions remain about increasedfluency and how it supports contingency adduction and crea-tive behavior. Precision teachers have reported fluency aimsfor several academic skills (e.g., Haughton, 1972; Binder,1987). At these levels of fluency, such skills are more readilyavailable for contingency adduction. This is exemplified in

Morningside Academy’s Model of Generative Instruction,which utilizes adduction extensively in its programming(e.g., Johnson & Street, 2020). However, such aims remainunidentified for athletic skills, artistic skills, and across a widerange of occupational skills to name but a few areas. Even inthe field of education, work remains in defining fluency aimsfor certain skills. For any given composite skill, the relation-ship between the level of fluency of the components and thelikelihood that these skills recombine into a more complexcomposite skills could be investigated. It could be that thereis a sweet spot between too much and too little fluency atwhich recombination is most likely.

Another interesting area of research related to contin-gency adduction and creative behavior comes from NevesFilho, Assaz, Dicezare, Knaus, and Garcia-Mijares(2020). These investigators have replicated Epstein andcolleagues’ 1985 box displacement problem, where pi-geons were taught two different behaviors, and a novelrecombination of those behaviors was required toproduce reinforcement. The replication was different inone key respect, the two component repertoires wereestablished using two different reinforcers. The firstrepertoire, directional pushing, was established usingfood. The second repertoire, climbing and pecking, wasestablished using water. The results revealed that pigeonsthat were taught the two component repertoires using thesame reinforcer all quickly produced the novel responsesequence necessary to produce reinforcement. On theother hand, the birds that received food for pushing andwater for climbing and pecking failed to produce thenovel response sequence required for reinforcement.Even after retraining using food to establish bothrepertoires, only half of the pigeons who initially failedto produce the solution went on to produce the novelresponse. Neves Filho et al. (2020) have demonstratedthe importance of a common consequential element inestablishing novel and creative behavior.

An interesting extension of Neves Filho et al. (2020) mightexplore the use of generalized conditioned reinforcers forteaching the repertoires to be combined. This would makeeach repertoire free from specific deprivations. Further, if thegeneralized conditioned reinforcers were of the same sensorymodality, such as two different frequencies of tones or twodifferent colors of lights, then the reinforcers wouldn’t involvedifferent physiological systems nor produce different intero-ceptive feedback, which the authors suggest might account forthe effects of the different reinforcers in disrupting the prob-lem solving behavior. The outcome of such a study wouldprovide valuable insight into best practices for teachingcreativity.

The investigators in the aforementioned study also con-ducted functional generalization tests with the birds whoeventually passed the box displacement test. This test

549Psychol Rec (2021) 71:543–551

included two new boxes that differed in appearance fromthe initial box, and one which was not functional, that is,wouldn’t support the pigeons climbing on it. All fourbirds in this phase failed the test (Neves Filho et al.,2020). This raises another interesting question: is there aprogram history that would allow the pigeons to passfunctional generalization tests? Perhaps a program similarto the concept teaching programs described by Layng andothers might produce a history where novel and creativesolutions generalize across functionally similar but phys-ically different stimuli (Layng, 2019; Tiemann & Markle,1985). Developing these programs would shed light onthe origins of creative performances as well as providedirections for those wishing to encourage creativebehavior.

Conclusion

Adduction may play a critical role in the production of crea-tive behavior. Procedures that increase the likelihood of ad-duction may be harnessed to engineer creative performance.In situations where teachers do not know what the topographyof the creative performance needs to look like, procedures canbe used to encourage a variety of component responses thatmay combine to form a composite response that meets thenew contingency requirement.

Response variability is vital to the process of adduction.Several procedures can be used to produce such variability(including the combined stimulus procedure, the oddityfrom sample procedure, the use of lag schedules, and ex-tinction). In situations where teachers are not sure of whatthe composite performance needs to look like, extinctioninduced variability may be the best way to produce theneed variety.

The role of establishing component skills in encouragingadduction has also been noted. For known outcomes, weshould conduct component analyses. For unknown outcomes,we should try to increase both related and seemingly unrelatedskills.

The moment of adduction is itself an instance of rein-forcement. As important as variability and componentskills are, it is equally important that novel response com-binations produce reinforcement when emitted so thatthey may be adduced and added to the repertoire. Thisis of critical importance, if one wants to add novel andcreative responses to a learner’s repertoire. As noted at theoutset, the process of creativity may seem mysterious.Lifting that veil of mystery may best be accomplishedby further investigating how contingency adduction mayprovide a basis for understanding the origins of creativeperformance.

Acknowledgements Special thanks to Dr. T. V. “Joe” Layng, Dr. PaulAndronis, Dr. Joanne Robbins, and Awab Abdel Jalil for invaluable con-tributions to the development and revision of this manuscript.

Data Availability Not applicable.

Compliance with Ethical Standards

Conflicts of Interest The authors have no conflicts of interest.

Code Availability Not applicable.

Human and Animal Rights and Informed Consent The present articledid not use human or animal subjects and did not involve the acquisitionof informed consent.

References

Adams, G. L., & Engelmann, S. (1996). Research on direct instruction:25 years beyond DISTAR. Educational Achievement Systems.Seattle, WA.

Andronis, P. T. (1983). “Symbolic aggression” by pigeons: Contingencyadduction (Unpublished doctoral dissertation). University ofChicago, Chicago, IL.

Andronis, P. T., Layng, T. J., & Goldiamond, I. (1997). Contingencyadduction of “symbolic aggression” by pigeons. Analysis ofVerbal Behavior, 14(1), 5–17.

Binder, C. (1993). Behavioral fluency: A new paradigm. EducationalTechnology, 33(10), 8–14.

Binder, C. V. (1987). Fluency-building: Research background.Nonantum, MA: Precision Teaching & Management Systems.

Binder, C. V. (1988). Precision teaching: Measuring and attaining aca-demic excellence. Youth and Policy, 10, 12–15.

Epstein, R. (1985). The spontaneous interconnection of three repertoires.The Psychological Record, 35(2), 131–141.

Epstein, R. (1999). Generativity theory. Encyclopedia of Creativity, 1,759–766.

Gagne, R. M., & Foster, H. (1949). Transfer of training from practice oncomponents in a motor skill. Journal of Experimental Psychology,39(1), 47.

Gilbert, T. F. (1962). Mathetics: The technology of education. Journal ofMathetics, 1(1), 7–74.

Haughton, E. C. (1972). Aims—Growing and sharing. Faculty ofEducation: York University. Arlington, VA.

Haughton, E. C. (1980). Practicing practices: Learning by activity.Journal of Precision Teaching, 1(3), 3–20.

Holth, P. (2012a). Variability as an operant? The Behavior Analyst, 35(2),243.

Holth, P. (2012b). The creative porpoise revisited. European Journal ofBehavior Analysis, 13(1), 87–89.

Johnson, K., & Street, E. M. (2004). The Morningside Model ofGenerative Instruction: An integration of research-based practices.In Evidence-based educational methods (pp. 247–265). AcademicPress. San Diego, CA.

Johnson, K., & Street, E. M. (2020). Generative responding throughcontingency adduction. In R. A. Rehfeldt, J. Tarbox, M. Fryling,& L. Hayes (Eds.), Applied behavior analysis of language andcognition. Oakland, CA: New Harbinger.

Johnson, K. R., & Layng, T. J. (1992). Breaking the structuralist barrier:Literacy and numeracy with fluency. American Psychologist,47(11), 1475.

550 Psychol Rec (2021) 71:543–551

Johnson, K. R., & Layng, T. J. (1996). On terms and procedures: Fluency.The Behavior Analyst, 19(2), 281–288.

Kieta, A. (2017). Reinforcing variability produces stereotypic behavior.(Unpublished master’s thesis), University of North Texas, Denton,TX.

Layng, T. J. (2019). Tutorial: Understanding concepts: Implications forbehavior analysts and educators. Perspectives on Behavior Science,42(2), 345–363.

Layng, T. J., Twyman, J. S., & Stikeleather, G. (2004). Engineeringdiscovery learning: The contingency adduction of some precursorsof textual responding in a beginning reading program. Analysis ofVerbal Behavior, 20(1), 99–109.

Neuringer, A. (2002). Operant variability: Evidence, functions, and the-ory. Psychonomic Bulletin & Review, 9(4), 672–705.

Neves Filho, H. B., Assaz, D. A., Dicezare, R. H. F., Knaus, Y. C., &Garcia-Mijares, M. (2020). Learning behavioral repertoires with dif-ferent consequences hinders the interconnection of these repertoiresin pigeons in the box displacement test. The Psychological Record,70(2), 1–9.

Page, S., & Neuringer, A. (1985). Variability is an operant. Journal ofExperimental Psychology: Animal Behavior Processes, 11(3), 429.

Pryor, K. W., Haag, R., & O’Reilly, J. (1969). The creative porpoise:Training for novel behavior 1. Journal of the Experimental Analysisof Behavior, 12(4), 653–661.

Ray, B. A. (1969). Selective attentions: The effects of combining stimuliwhich control incompatible behavior. Journal of the ExperimentalAnalysis of Behavior, 12(4), 539–550.

Schiller, P. H. (1952). Innate constituents of complex responses in pri-mates. Psychological Review, 59, 177–191.

Schiller, P. H. (1957). Innate motor action as a basis of learning. In C.Schiller (Ed.), Instinctive behavior: The development of a modernconcept. International Universities Press. New York, NY

Tiemann, P. W., & Markle, S. M. (1985). Analysing instructional con-tent: A guide to instruction and evaluation. Stipes Publishing.Champaign, IL

Weiss, M. J. (2001). Expanding ABA intervention in intensive programsfor children with autism: The inclusion of natural environment train-ing and fluency based instruction. Behavior Analyst Today, 2(3),182.

Williams, P., Granzin, A., Engelmann, S., & Becker, W. C. (1979).Teaching Language to the truly naive learner: An analog study usinga tactual vocoder. Journal of Special Education Technology, 2(3),5–15.

Publisher’s Note Springer Nature remains neutral with regard to jurisdic-tional claims in published maps and institutional affiliations.

551Psychol Rec (2021) 71:543–551

Psychological Record is a copyright of Springer, 2021. All Rights Reserved.