an analysis of autism as a contingency shaped disorder of verbal behavior-drash and tudor 2004

The Analysis ofVerbal Behavior 2004, 20, 5-23

An Analysis ofAutism as a Contingency-ShapedDisorder of Verbal Behavior

Philip W. Drash, Autism Early Intervention CenterRoger M. Tudor, Westfield State College

This paper analyzes autism as a contingency-shaped disorder of verbal behavior. Contingencies of rein-forcement in effect during the first to third year ofa child's life may operate to establish and maintain thosebehaviors that later result in a diagnosis of autism. While neurobiological variables may, in some cases,predispose some children to be more or less responsive to environmental variables than others, our analy-sis suggests that reliance on neurobiological variables as causal factors in autism is unnecessary. We presentsix paradigms that may play critical etiologic roles in the development of behaviors labeled as autistic.Recognizing these contingencies and their resulting behaviors during the first two years of a child's lifemay contribute substantially to earlier identification, more effective treatment and, quite possibly, to thedevelopment of Applied Behavior Analysis programs for the prevention of autism that could be imple-mented immediately. Conceptualizing autism as a contingency-shaped disorder of verbal behavior mayprovide a new and potentially more effective paradigm for behavioral research and treatment in autism.

Autism is widely regarded as one ofthe mostsevere of childhood behavioral disorders(Barton-Cohen, Allen, & Gillberg, 1992). Theeffects of autism are pervasive and interferewith the acquisition ofnormal behavioral rep-ertoires across almost every area of humanfunctioning. During the past 30 years, exten-sive research has been devoted to the develop-ment and implementation of effective behav-ioral treatments and to an analysis ofthe etiol-ogy of autism. Since Lovaas (1987) demon-strated that it is possible to achieve relativelytotal recovery in some young children diag-nosed with autism by using an intensive 40-hour-per-week behavioral intervention, signifi-cant progress has been made in the behavioraltreatment of children diagnosed with autism.These treatments have for the most part reliedon a functional analysis of behaviors that arelabeled as autistic (Charlop-Christy, & Kelso,

Earlier versions of this paper were presented atthe Association for Behavior Analysis, 25th annualmeeting, Chicago, IL, May, 1999, and at the 26thannual meeting, Washington, DC, May, 2000. Theauthors express their appreciation to Jack Michael,Dick Malott, Mark Sundberg, and Hank Schlingerfor their valuable recommendations on earlier ver-sions of this manuscript. Send reprint requests toPhilip W. Drash, Ph.D., BCBA, Autism Early Inter-vention Center, 2901 W. Busch Blvd., Suite 807,Tampa, FL, 33618; e-mail: [email protected].

1997; Maurice, Green, & Foxx, 2001; Maurice,Green, & Luce, 1996; Leaf, & McEachin,1999; Lovaas, 1977, 1981; Sundberg, &Partington, 1998).

Despite substantial progress in the treatmentof autism, determining the etiology of thosebehaviors that may later result in a diagnosisof autism continues to be an unresolved issue.The Advocate reported that statistics recentlyreleased by the Autism Society of America(2002, p. 6) stated that autism is increasing ata rate of 10 to 17% per year. As a result therehas been a strong advocacy for a substantialincrease in research into the causes of autism.Discovering the cause of autism is consideredby many professionals and autism advocatesas essential in the development ofmore effec-tive programs for the prevention and treatmentof autism. Lee Grossman, President ofthe Au-tism Society ofAmerica, and Robert Beck, Ex-ecutive Director, recently stated (2002), "Af-ter more than 50 years ofstudy, no causes havebeen identified. There needs to be a geometricincrease in research funding ... to determinethe causes of autism" (p. 7).

Currently there are at least two major hy-potheses regarding the causes of autism, thosethat are primarily neurobiological and thosethat are primarily behavioral. The neurobiologi-cal hypothesis attributes the cause ofautism toa presumed but as yet unidentified neurobio-logical disorder. Conversely, behavioral theo-

5

6 PHILLIP W. DRASH & ROGER M. TUDOR

ries ofautism describe how environmental con-tingencies operating during the first one to threeyears of a child's life may establish and main-tain those behaviors that later result in the di-agnosis of autism.

This paper presents a behavioral analysis ofthe etiology of those behaviors upon which adiagnosis ofautism is based. Our analysis sug-gests that reinforcement contingencies operat-ing during the first to third year ofa child's lifemay play a substantial role in the developmentof the behaviors that are subsequently diag-nosed as autism. We have previously publishedan analysis of contingencies of reinforcementthat may lead to delay in acquisition of verbalbehavior in typical young preschool children(Drash & Tudor, 1990, 1993). A similar analy-sis may also be applied to the shaping of be-haviors that result in the diagnosis of autism(Drash, High, & Tudor, 1999; Drash & Tudor,1999, 2000).We recognize that the prevailing opinion

among many professionals and parents, includ-ing many behavior analysts, is that a disorderas severe and debilitating as autism could onlyoccur through defective neurological or bio-logical mechanisms. We understand and respecttheir opinions. Moreover, we do not proposethat this analysis represents the only possibleexplanation for the occurrence ofbehaviors thatresult in the diagnosis of autism. We do notrule out the possibility that, in the future, re-searchers may discover specific neurological,biological or genetic factors that may contrib-ute directly to the development of behaviorsthat later result in a diagnosis of autism. How-ever, a behavioral analysis of the cause of au-tistic behaviors is in no way dependent uponthe presumption of such factors.Our goals are identical to those of parent

advocates and other professionals who call formore effective procedures for prevention, treat-ment, and possible cure of autism that can beimplemented immediately. Our operating as-sumption is that a behavioral analysis of thecontingencies that may shape those behaviorsupon which a diagnosis ofautism is based willultimately serve the best interests of childrendiagnosed with autism and their families. Webelieve that, in the long term, this analysis ofautism as a contingency-shaped disorder ofverbal behavior may contribute materially tothe development of more effective behavioralprograms for prevention, early intervention,

and treatment ofautism that will, in the future,benefit hundreds, and perhaps thousands, ofyoung children and their families.

The Neurobiological View ofthe Causes ofAutism

Many contemporary theories ofautism havelinked its cause to as yet unidentified neuro-logical or biological factors. In support of thisposition, Minshew, Sweeney, and Bauman(1997) in the Handbook ofAutism and Perva-sive Developmental Disorders (Cohen &Volkmar, 1997) stated, "Autism is now gener-ally accepted to be a disorder of brain devel-opment and hence of neurological origin"(p.344). Similarly in the same volume Dykensand Volkmar (1997) stated, "Researchers gen-erally agree that autism is the result of someneurobiological factor or factors" (p. 388).These popular assumptions about the causesof autism have spawned at least three nationalorganizations devoted to discovering medical,neurological, or biological causes of autism:CureAutism Now (CAN); DefeatAutism Now!(DAN!); Autism Research Institute, and theNational Alliance for Autism Research.Although neurobiological views of the

causes of autism are intuitively appealing,medical research has failed to provide conclu-sive evidence for a neurological, biological, orgenetic cause for autism. In the introduction toa special issue of the Journal ofAutism andDevelopmental Disorders devoted to contem-porary research in autism, Alexander, Cowdry,Hall, and Snow (1996) stated, "No consensusregarding causes or potential cures for autismis assumed. This is a problem that is not yetsolved" (p. 118). Likewise, Bailey, Phillips, andRutter (1996) stated, "a replicable, neurophysi-ological basis for autism has not yet been iden-tified" (p. 89). More recently other neurobio-logical researchers have reached similar con-clusions. Lauritsen, Mors, Mortensen, andEwald (1999) stated, "Infantile autism is a het-erogeneous disorder ofunknown etiology" (p.335). Trottier, Srivastava, and Walker (1998)reported, "The etiology of autism is complex,and in most cases the underlying pathologicmechanisms are unknown" (p. 103). Thus de-spite the prevalence ofthe neurobiological ex-planation, it is evident that no conclusive sci-entific evidence for a neurobiological cause forautism currently exists.

THE ETIOLOGY OF AUTISM 7

Behavioral Theories ofAutism

Behavior analytic explanations ofthe causesof autism are numerous and diverse. These in-clude the behavioral hypothesis of Ferster(1961), the contingency-shaped or behavioralincompatibility theory of Drash and Tudor(1993, 1999, 2000), the behavioral mismatchtheory of Lovaas and Smith (1989), the socialcommunication theory of Koegel, Valdez-Menchaca, and Koegel (1994), the stimuluscontrol theory of Spradlin and Brady (1999),and the behavioral interference theory ofBijouand Ghezzi (1999). These behavioral theoriesall incorporate the view that the behaviors ofchildren labeled as autistic can be analyzed interms ofthe concepts and principles ofAppliedBehavior Analysis (ABA), and that these con-cepts and principles can produce effective treat-ment programs for children diagnosed withautism (Ghezzi, Williams, & Carr, 1999). How-ever, these theories differ greatly with regardto the initiating cause ofbehaviors upon whicha diagnosis of autism is based.With the exception of the first two theories

(Drash & Tudor, 1993, 1999, 2000; Ferster,1961) these behavioral theories attribute theinitiating cause ofautism to a defective neuro-logical or biological process that interferes withthe normal developmental process. Lovaas andSmith (1989) postulate a mismatch between thenormal environment and the nervous systemof the child. The Koegel, Valdez-Menchaca,and Koegel (1994) theory postulates a defec-tive neurological process that may result in in-appropriate socialization and defective lan-guage development. Spradlin and Brady (1999)hypothesized that possible neurological limi-tations in children with autism make it moredifficult to establish stimulus control. Bijou andGhezzi (1999) postulated that young childrenwith autism have "abnormalities in their sen-sory equipment" that produce a "tendency toescape and avoid tactile and mild auditorystimuli" (p. 34). This avoidance behavior theninterferes with normal social and language de-velopment. These four theories can be termedreductionistic in that they attribute the initiat-ing cause of autism to an hypothesized butunidentified neurobiological process. In con-trast to these, only Drash and Tudor (1993,1999, 2000) and Ferster (1961) rely on a com-pletely behavioral analysis that attributes theearly development ofautistic behaviors to spe-

cific and identifiable contingencies of rein-forcement in the early environment ofthe child.One ofthe first significant behavioral analy-

ses ofautism was published by Ferster (1961).He presented a detailed analysis ofhow a vari-ety of contingencies of reinforcement operat-ing between parent and child during the earlyyears might establish and strengthen a reper-toire ofbehaviors typical ofchildren diagnosedas autistic. He observed that a child's disrup-tive behaviors may be maintained by their ef-fect on his parents or caregivers because theyfunction as an aversive stimulus that can beterminated ifthe caregiver supplies a reinforcer.Moreover, he also observed that over time suchaversive behaviors may be strengthened bycontinued reinforcement and become prepotentover other age-appropriate behaviors. Unfor-tunately, Ferster's analysis was regarded bysome as a behavioral version of the discardedpsychogenic theory which ascribed autism toparental personality traits. The implications ofthe article for research and treatment, includ-ing extensions to the communicative functionsofaberrant behavior, therefore, were never fullyanalyzed.

In a review of the Bijou and Ghezzi (1999)analysis, Hayes (1999) cautioned that attribut-ing psychological events to biological causes isunnecessary and is an impediment to the devel-opment of effective (behavioral) treatmentssince it leaves lingering doubt as to the possi-bility oftruly successful psychological interven-tion. Likewise, Schlinger (1995) stated that be-havioral or environmental contingencies thatmight account for a behavioral complex shouldbe ruled out first before attributing the behav-ior to neurological, biological, or genetic causes.

CATEGORIES OF BEHAVIOR THAT REQUIREANALYSIS IF AUTISM IS TO BE CONCEPTUALIZED

AS A CONTINGENCY-SHAPED DISORDER

The Diagnostic and Statistical Manual ofPsychiatric Disorders (DSM IV; AmericanPsychiatric Association, 1994) outlines threespecific categories of behavior that are con-sidered essential for an accurate clinical diag-nosis ofautism. Since there is no neurobiologi-cal or genetic test for autism, the diagnosis ofautism is based entirely on observed behavior.Based on our research and that of others wehave included a fourth category of behaviorthat we consider a central feature of autism.


In order to demonstrate that behavioral con-tingencies may be largely, if not completely,responsible for the behavioral complex diag-nosed as autism, it is necessary to analyze howcontingencies ofreinforcement may establish,shape, and maintain the behaviors that com-prise each ofthese four categories ofbehavior.

Qualitative Impairment in Communication asManifested by Little or No Spoken Language

Severe language deficiency is a classic fea-ture ofautism (Churchill, 1978; Richter, 1978;Rutter, 1974, 1978). Autism in young childrenis rarely diagnosed in the absence of a signifi-cant deficiency in spoken language. Indeed, itis the lack ofage-appropriate spoken languageat age 2 to 3 years that typically initiates theentire referral, diagnostic, and treatment pro-cess. Many children diagnosed with autism at2 to 3 years of age have little or no expressiveor receptive verbal behavior, while others haveminimal receptive repertoires but no expres-sive verbal behavior. Many 2 and 3-year-oldchildren when first diagnosed with autism arefunctioning at a 9 to 12 month level of lan-guage acquisition. Thus, serious deficiency inor lack ofspoken language repertoires may betheprimary and essential distinguishing char-acteristic of autism, since without deficiencyin spoken language such children quite prob-ably would not be diagnosed as autistic.

Qualitative Impairment in Social InteractionIncluding Marked Social Isolation andImpairment in Peer Relationships

Social isolation and delay in age-appropri-ate social behavior is a second major categoryof behavior typical of children with autism.Almost all children diagnosed with autism havesome deficits or impairment in their social rep-ertoires. Behaviors in this category includesocial isolation and aloneness, avoidance ofeye-contact, lack ofage-appropriate social play,lack of responsiveness to other persons, andlack of age-appropriate social-interactionalskills.

Markedly Restricted, Repetitive, and Stereo-typed Patterns ofBehavior, and LimitedResponsiveness to Environmental Stimuli

This category includes behaviors such as ste-reotyped body movements, hand flapping, per-sistent preoccupation with specific parts ofobjects, emotional responses to inconsequen-tial alterations in trivial aspects ofthe environ-ment, unreasonable insistence on followingspecified routines in precise detail, preoccu-pation with specific responses, such as smell-ing or spinning objects. The range and varietyofenvironmental stimuli that function as rein-forcers is also markedly restricted.

Moderate to Severe Disruptive Behaviors,Task-Avoidance, and Noncompliance

Disruptive behaviors, task-avoidance, andnoncompliance are not specified as distinctdiagnostic criteria for autism in the DSM IV.However, we view these behaviors as both typi-cal and critical components in most cases ofautism. Moreover, in his original papers onautism, Kanner listed a variety of disruptivebehaviors including temper tantrums, aggres-siveness, and destructiveness as characteristicof children with autism (Frith, 1991). There isalso overwhelming evidence in the researchliterature on autism that documents that chal-lenging behaviors are highly typical of chil-dren labeled as autistic (Carr & Durand, 1985,1986; Charlop-Christy & Kelso, 1997; Durand,1999; Iwata, Dorsey, Slifer, Bauman, &Richman, 1994; Leaf & McEachin, 1999;Lovaas, 1993; Repp, & Singh, 1990). Typicalbehaviors in this category include crying,screaming, temper tantrums, head-banging,kicking, biting, task-avoidance, non-respon-siveness, noncompliance, aggressive behavior,self-stimulatory behavior, and self-injuriousbehavior.

How CONTINGENCIES OF REINFORCEMENTMAY OPERATE TO ESTABLISH AND SHAPE THEBEHAVIORS THAT SUBSEQUENTLY RESULT IN A

DIAGNOSIS OF AUTISM

In summarizing their interference theory ofautism, Bijou and Ghezzi (1999) concluded,"... most ofthe abnormal behaviors ofchildrenwith autism serve to compensate for their defi-ciencies in social-emotional and verbal behav-ior" (pp. 39-40). Based on their analysis, defi-ciencies in social-emotional behavior and ver-bal behavior are alone sufficient to account formost ofthe behaviors that are observed in chil-


dren labeled as autistic. While agreeing withmuch of their analysis, we differ substantiallyin at least three respects.

First, we view autism primarily as a contin-gency-shaped disorder of verbal behavior thatoften coexists with a repertoire of avoidanceand other disruptive behaviors. In order to ana-lyze the causes of autism, it is first necessaryto analyze the causes of deficiency in verbalbehavior. As will be discussed below, the so-cial-emotional deficits ofchildren with autismcan be causally linked by a behavioral analy-sis to the deficiencies in verbal behavior andthe presence ofdisruptive avoidance behaviors.

Second, we view the presence of inappro-priate verbal behavior, that is, aversive vocalmanding (e.g., screaming, crying or whining)in combination with avoidance and other dis-ruptive behaviors, rather than age-appropriateverbal behavior (such as, pre-speech vocalsounds, words, phrases, etc.) asprimary causalfactors contributing to the shaping and main-tenance of other behaviors on which a diagno-sis of autism is based. Both experimental andclinical evidence details how repertoires ofaversive vocal manding and other disruptiveavoidance behaviors can be shaped by rein-forcement contingencies, and once established,are incompatible with the acquisition of func-tional verbal behavior (Drash, 1993; Drash,High, & Tudor, 1999; Drash & Tudor, 1993,Richter, 1978). Moreover, since many ofthesemore aversive behaviors terminate parent-childinteractions, they may also prevent or inhibitthe establishment of social-emotional bondingand other social behaviors. These two reper-toires, aversive vocal manding and other dis-ruptive behaviors, may thus be responsible formost of the other behavioral symptoms of au-tism.

Third, absence of age-appropriate verbalbehavior and the presence ofavoidance behav-ior can be explained as a result of contingen-cies ofreinforcement operating during the firstto third year of a child's life, especially duringthe first 12 to 18 months. A behavioral analy-sis of those contingencies explains the devel-opment or lack of development of verbal be-havior and the presence of disruptive andavoidance behavior without relying on hypo-thetical neurological variables to explain theiroccurrence.The fact that there may be subsets of chil-

dren diagnosed as autistic with accompanying

neurological, biological, or genetic abnormali-ties is not disputed. This would not be incom-patible with our contingency-shaped theory.Since the diagnosis of autism is based entirelyon a continuum of observed behaviors, i.e.,"autism spectrum disorders" (Wetherby &Prizant, 2000), the diagnosis will, on occasion,almost inevitably include subsets of atypicalchildren, who in addition to their principal dis-order, such as Down syndrome, Fragile X syn-drome, or Rett's disorder, may also displaybehaviors that are typical ofchildren diagnosedas autistic. In other cases, children diagnosedwith autism may later be found to have sei-zures, brain tumors, or other physiologicalproblems which may have contributed to thedevelopment of autistic behaviors.

BASIC PREMISES OF A BEHAVIORAL ANALYSISOF AUTISM AS A CONTINGENCY-SHAPED

DISORDER OF VERBAL BEHAVIOR

To analyze autism as a contingency-shapeddisorder of verbal behavior it is necessary toprovide a conceptual analysis showing howreinforcement contingencies may establish andmaintain a repertoire ofdeficient verbal behav-ior and disruptive avoidance behaviors duringthe first one to two years of life.The purpose ofthis paper is to present a con-

ceptual analysis that extends well establishedbehavioral principles to an analysis of the eti-ology of behaviors that are later diagnosed asautistic. All of the terms of this analysis referto potentially modifiable conditions in thechild's environment and directly observableand measurable aspects of his performance.Moreover, experimental evidence supportseach of the basic premises of this analysis.There are at least four major premises uponwhich this analysis is based. In summary, theseare: 1) The acquisition of verbal behavior, orthe lack thereof, by children labeled as autisticis primarily a function of reinforcement con-tingencies provided by caregivers and othersduring the first years of a child's life. 2)Caregivers and others may inadvertently shaperepertoires ofdisruptive and avoidance behav-iors in their infants and young children duringthe first one to three years of a child's life. 3)Disruptive and task-avoidance responses arefrequently present in young children diagnosedwith autism, or PDD. 4) When present, dis-ruptive behaviors may become incompatible


with and may prevent the acquisition of age-appropriate verbal behavior, as well as othersocial behaviors.The first question is whether and to what

degree contingencies of reinforcement pro-vided by caregivers and others may facilitateor impede the language acquisition of theirchildren. Empirical support for the concept thatparents strongly influence the language acqui-sition oftheir children from infancy forward isprovided by the longitudinal research of Hartand Risley (1995, 1999) on language develop-ment in children from 7 months to 3 years.Their research shows that the frequency andcomplexity, or lack thereof, of the child's ver-bal behavior at age 3 years is directly relatedto the frequency and complexity of verbal be-havior that occurs between parents and theirchildren from the first year of life forward.When the hourly frequency of talk betweenparent and child was high, the children devel-oped large vocabularies and spoke in complexsentences by age three. When the hourly fre-quency of talk between parent and child waslow, children developed much smaller vocabu-laries and spoke in much less complex sen-tences. This finding suggests that ifthere is verylittle or no interactive talk between parent andchild during the first year to three years, thechild's verbal behavior may be deficient at threeyears.

In keeping with our premise that acquisitionof verbal behavior is a function of contingen-cies of reinforcement, Hart and Risley (1999)concluded, "We propose that language devel-opment is governed by the same natural lawsas motor, social, and cognitive development"(p. 199).The second question is whether caregivers

and others, during the first year of a child'slife, may unintentionally shape disruptive in-fant response repertoires. In a series of experi-mental studies ofemotional behavior ofinfants,ages 6 to 12 months, and their parents, Gewirtzand Pelaez-Nogueras (1991, 1999) demon-strated that disruptive infant behaviors, suchas crying, whining, and screaming, can be in-advertently shaped by parents during the firstyear of life. Moreover, the parent-infant con-tingencies that shape the disruptive behaviors canbe identified and modified as early as 6 to 9months of age. In discussing how disruptive in-fant behaviors may be shaped, they concluded,"The infant adaptive problem behaviors actually

appear to be operants under the control ofoccasioning stimuli and consequences inadvert-ently provided ... by the responding of well-in-tentioned, loving parents" (1999, p. 272). Thus,without intending to do so, parents may shapedisruptive behaviors during the first year of lifewhich may interfere with the acquisition ofmoreadaptive responses.The third question is whether disruptive and

task-avoidance responses are frequentlypresent in young children diagnosed with au-tism or PDD. We have previously shown thataversive manding and other disruptive avoid-ance behaviors are frequently present in youngchildren diagnosed with autism or PDD. Ananalysis of all cases (N = 48) admitted during1992 and 1993 with a diagnosis of autism orPDD, revealed that in 41 of the 48 cases, or in85% ofthe cases, disruptive or task-avoidancebehavior was present and was a major factorinterfering with acquisition ofverbal behavior(Drash, 1993).The final question is whether it is possible

for aversive manding and other disruptive be-haviors, when present, to prevent the acquisi-tion of verbal behavior. We have shown thataversive manding and other avoidance behav-iors when present may be incompatible withthe acquisition of age-appropriate verbal be-havior. It is only after inappropriate verbal be-havior and other disruptive behaviors aregreatly reduced or eliminated by providing re-inforcement only for acceptable vocal mands(that is, extinction combined with differentialreinforcement ofacceptable vocal mands), thatshaping of appropriate vocal mands can pro-ceed. (Drash, High, & Tudor, 1999).The contingency-shaped theory of autism is

based entirely on behaviors that are readilyobservable, measurable, and modifiable, andthe analysis is thus subject to further experi-mental analysis and verification.

WHEN Do CONTINGENCIES OF REINFORCEMENTBEGIN TO SHAPE THE BEHAVIORS THAT ARE

LATER DIAGNOSED AS AUTISTIC?

Until recently the diagnosis of autism wasrarely made before a child was 2 to 3 years ofage. Despite this, accumulating evidence sug-gests that many parents have expressed con-cern to their pediatrician about their child's lan-guage and social delay by 18 months of age(Siegal, Pliner, Eschler, & Elliot, 1988). Other


research reported that 50% ofparents ofa childdiagnosed as autistic suspected a problem be-fore their child was 1 year old (Ornitz, Guthrie,& Farley, 1977). Barton-Cohen, Allen, andGillberg (1992) demonstrated that behaviorsthat are correlated with a later diagnosis ofautism can be accurately identified at 18months. More recent research has shown thatbehavioral correlates of a later diagnosis ofautism can be identified as early as 8 months(Werner, Dawson, Osterling, & Dinno, 2000)to 12 months (Osterling & Dawson, 1994).The fact that behaviors correlated with a later

diagnosis ofautism can be detected as early as8 to 12 months, in combination with parentalawareness and concern over developmentalproblems before a child is 12 months old, sug-gests that the contingencies producing lan-guage and social delays are in all likelihoodoperating during the first year oflife, and quitepossibly as early as 6 to 8 months, if not ear-lier. Therefore, a behavioral analysis shouldfocus on pinpointing those caregiver-infantinteractions that may establish and reinforceaversive vocal mands and other disruptive andavoidance behaviors that occur during the firstyear of life and that may later result in a diag-nosis of autism. The contingencies and result-ing behaviors, to be described below, in all like-lihood, have their origins during the first yearand are then further shaped, refined, andstrengthened as the child develops.

THE RELATIONSHIP BETWEEN MANDINGAND VERBAL DELAY IN YOUNG CHILDREN

LABELED AS AUTISTIC

Skinner's (1957) analysis of verbal behav-ior is particularly relevant to analyzing verbaldeficiency in young children labeled as autis-tic. Skinner's analysis indicates that verbalbehavior is acquired primarily because it pro-duces reinforcement through the mediation ofother persons. Skinner's (1957) identificationof the mand as the first verbal operant to beacquired is particularly critical to our analysis.If during the first year to three years of life aninfant is given all the essential, life sustainingreinforcement and nurture without a require-ment for age-appropriate vocal manding, thenit is quite possible that verbal behavior maynot develop. A detailed description of howSkinner's analysis of verbal behavior can beapplied to the analysis and treatment ofverbal

behavior in children diagnosed as autistic isprovided by Sundberg and Michael (2001) andSundberg and Partington (1998).

REINFORCEMENT PARADIGMS THAT MAYCONTRIBUTE TO THE DEVELOPMENT OFDEFICIENT VERBAL REPERTOIRES IN

CHILDREN WITH AUTISM

In a previous publication we presented ananalysis of reinforcement contingencies thatmay contribute to language delay in young pre-school children (Drash & Tudor, 1990, 1993).Our research has demonstrated that a similaranalysis applies equally well to language de-lay in young children with autism (Drash, High,& Tudor, 1999; Drash & Tudor, 1999, 2000).There are at least six reinforcement para-

digms that may contribute to significant defi-ciency in verbal behavior that we have identi-fied and analyzed in our verbal behavior re-search with children labeled as autistic over thepast thirty years. The observations upon whichthese analyses are based represent multiple rep-lications ofwithin-subject studies in which thecontingencies preventing acquisition ofverbalbehavior were repeatedly identified and ana-lyzed, and then systematically replaced withcontingencies that produced age-appropriateverbal behavior. Our standard data collectionsystem for continuous recording and analysisof verbal behavior during the shaping of ver-bal behavior has been previously described(Drash & Tudor, 1991). This data recordingsystem allowed us to meet the requirements fordrawing valid inferences from within-subjectcase studies, replicated across multiple sub-jects, as discussed by Kazdin (1982).

Each of these six reinforcement paradigmsmay contribute to the establishment ofa reper-toire of behavior that is incompatible with theacquisition ofage-appropriate verbal behavior.Several of these paradigms may concurrentlycreate a repertoire of avoidance responses.

1. ReinforcementforAversive VocalManding, such as Crying or Screaming, orOther Avoidance Behaviors that May BeIncompatible with Acquiring Age-Appropri-ate Verbal Behavior

Manding is the first type of verbal behavioremitted by infants (Drash, High, & Tudor,1999; Drash & Tudor, 1993; Schlinger, 1995;


Skinner, 1957). The first cries of an infant arerespondent in nature and gradually come un-der operant control as vocal mands when acaregiver responds to these cries. In most in-stances, the infant's cries and screams are trans-formed into more acceptable verbal behaviorthrough the parent's subtle shaping and a vari-ety of prohibitions against screaming, cryingor fussing (Bruner, 1983; Hart & Risley, 1995;Schlinger, 1999). Conversely, if caregivers in-advertently provide reinforcement for crying,fussing, or screaming to the exclusion of re-quiring age-appropriate vocal mands, a strongrepertoire of aversive vocal manding may beestablished (Ferster, 1961). For example, iftheparent has heard the infant produce the sound,"bababa" on multiple occasions, the parent mayprompt the child to say "baba" for bottle bypresenting the bottle with the verbal prompt,"Say, bababa." However, if the child has beenwithout food for some time, the response mightbe a cry or scream, the response that in the pasthas been reinforced. To escape the aversivecrying of the infant, the parent may quicklypresent the bottle without first requiring anechoic response to the prompt. The infant'slikelihood of crying in response to futureprompts for vocal behavior will have been fur-ther strengthened. Over a longer period oftimethe caregiver may completely avoid the aver-sive cries ofthe infant by providing food, milk,and other reinforcers without first requiring ap-propriate vocal responses. The aversive vocalmanding of the child can interfere with andultimately terminate the parents' languageteaching efforts and thereby prevent the acqui-sition of age-appropriate verbal behavior.

After the contingencies have shaped a strongrepertoire of aversive vocal manding, stimu-lus control over these responses will be present.The child will show an increased tendency toemit aversive vocal mands and other avoidancebehaviors in situations similar to those thatwere previously reinforced. For example, thechild might pull the parent to the refrigeratorand wait to be given juice or milk. Ifthe parentattempts to require that the child first say"juice" or "milk," the child may cry or screamuntil given the juice. The teaching efforts arethus terminated, and aversive manding is fur-ther strengthened.The majority of young children in our re-

search and clinical programs have had behav-ioral repertoires that reflect the influence ofthis

paradigm. An analysis of all cases (N = 48)admitted for treatment with a diagnosis of au-tism or PDD during 1992-1993 revealed that85% had disruptive or task-avoidance behav-ior that interfered with acquisition of age-ap-propriate verbal behavior (Drash, 1993). Typi-cal of children in this category is a 3 1/2-year-old child referred for treatment because of se-vere language delay. The initial evaluationshowed that the child produced no words orother age-appropriate verbal behavior, and heemitted severe oppositional behavior. Duringthe initial evaluation he screamed loudly whenprompted to produce a word or sound. Twenty-seven percent of his initial responses were ei-ther screams or task-refusals. The mother re-ported that at home whenever she promptedthe child to produce a word or sound, the childscreamed until she ceased prompting him forvocal behavior.A second child, who was first diagnosed as

at-risk for autism at 15 months ofage, engagedin similar behaviors. When evaluated by us at3 years of age, he had no expressive verbalbehavior, and his mother reported that he onlycommunicated by screaming. Fifty percent ofhis baseline responses were either screams ortask-refusals. Temper tantrums were his pre-dominant response when he did not immedi-ately obtain a reinforcer. During the initial in-terview he began violently kicking his mother.She reported that his tantrums often consistedof pinching, kicking, scratching, biting, andhead-butting. A speech therapist who begantherapy with the child at 18 months discontin-ued structured therapy as a result of these dis-ruptive behaviors.As demonstrated by both cases, a single vo-

calization, the aversive vocal mand, a cry orscream, functioned as a generic, all-purposemandthat the child emitted to obtain reinforce-ment or to escape or avoid aversive stimuli.The aversiveness of the child's behavior dis-couraged parental attempts to teach more ap-propriate verbal behavior.

There are other vocal behaviors that, whilenot as aversive, are functionally equivalent toaversive manding in that they are incompat-ible with and prevent the acquisition of age-appropriate vocal behaviors. In one instance, a2 1/2-year-old child with a diagnosis of apha-sia was referred for the treatment of languagedelay. Although the child had no age-appro-priate expressive verbal behavior, he could pro-


duce a variety of vocal sounds. But whenprompted to produce an echoic response, heprimarily emitted one sound, "eee." The par-ents were originally amused by this responseand inadvertently reinforced it, but they soonrealized that the child vocalized few othersounds. The predominance ofthe single sound"eee" prevented the parents from teaching thechild age-appropriate vocal behavior. It wasonly after this single sound was replaced withother more appropriate vocal mands throughverbal behavior therapy that the child began toacquire age-appropriate verbal behavior.

After a repertoire of aversive or competingvocal mands is established, it can be extremelyresistant to modification. In each of the abovecases, a repertoire ofaversive manding or otherincompatible vocal behaviors effectively pre-vented acquisition of a repertoire of age-ap-propriate vocal mands. In addition, these aver-sive behaviors may also prevent or inhibit theestablishment ofsocial-emotional bonding andother social behaviors. For a further discussionof this reinforcement paradigm see Malott,Malott and Trojan (2000, pp. 295-298).

2. Reinforcementfor Gestural Manding andOther Nonvocal Forms ofManding

This category includes behaviors such aslooking at, reaching toward, pointing to, stand-ing next to, or pulling the caregiver toward thedesired item. During our standard clinical in-terview we routinely ask parents of childrenlabeled as autistic the following question,"Since your child cannot talk, how do youknow what he or she needs or wants?" The mostfrequent response is that their child looks at,reaches toward, or pulls them to the desiredobject. Caregivers routinely reinforce andstrengthen a repertoire of nonvocal gesturalmanding by supplying reinforcing stimuli with-out first setting a contingency for acceptablevocal mands (Drash, High, & Tudor, 1999). Ifthe reinforcing stimulus item is not immedi-ately forthcoming, the child may respond withcrying or screaming until the parent presentsthe reinforcer. If the parents believe that theirchild is unable to speak, they may immediatelyprovide a reinforcer, thus combining the effectsofparadigms one and two. If nonvocal ges-tural manding continues to be reinforced untila child is three to four years, it will becomedeeply ingrained and pervasive in the child's

repertoire. This nonvocal repertoire becomesthe child's primary mode of obtaining rein-forcement, thereby preventing the acquisitionof age-appropriate verbal behavior.

3. Anticipating the Child's Needs and ThusReinforcing a Repertoire ofNonrespondingthat Prevents both Vocal and NonvocalMands

Although similar in some respects to para-digm number two, a significant difference ex-ists. In this paradigm, caregivers anticipate thechild's "needs and wants" and deliver rein-forcement noncontingently before the childmands either vocally or gesturally. This para-digm may over time establish a repertoire ofvery low rate behavior in which the child sim-ply waits passively for reinforcement withoutany form of manding, either vocal or gestural.Therapeutic attempts to prompt vocal behav-ior often produce temper tantrums or otherforms of task-avoidance.

Several parents reported that they anticipatedtheir child's needs to prevent their child frombecoming "frustrated." For example, the par-ents of a 2 1/2-year-old nonverbal child re-ported that they anticipated all their child'sneeds and never required him to speak. Theywere unaware that by providing noncontingentreinforcement they were strengthening a rep-ertoire of nonresponding.Over time this paradigm may produce a si-

lent or passive child who "appears to have littleor no interest in the environment," a character-istic typical of many children labeled autistic.

4. Extinction of Verbal Behavior

This paradigm is in effect whenever an in-fant is in an environment in which parents orother caregivers do not actively prompt, re-spond to, and reinforce the child's vocal utter-ances. Such environments may occur more fre-quently than is generally recognized. In today'sculture it is likely that both parents (or the singleparent) will be working and will leave the childin a day-care center or in the care of a relative,a baby sitter, or a nanny.Although the staff ofmany day-care centers

may provide excellent physical care for younginfants and toddlers, the staff may not havesufficient time or expertise to provide consis-tent and ongoing reinforcement on a moment


to moment basis for the verbal behavior ofeachindividual infant. Hart and Risley (1995) stated,"Quality out-of-home care can be provided forinfants and young children, even though itrarely is" (p. 207), moreover, "the most impor-tant aspect to evaluate in child care settings forvery young children is the amount of talk ac-tually going on moment to moment, betweenchildren and their caregivers" (p. xxi).

Extinction of verbal behavior may also occurat home when an infant is cared for by relatives,a baby sitter, or a nanny. The research of Hartand Risley (1995) indicates that the essential el-ement in language delay appears to be, "Howfrequently does the caregiver talk with the childeach hour?" If there is little or no talk betweencaregiver and infant each hour during the hoursof care, the verbal behavior of the child may beon extinction. In one case, the parents, both ofwhom worked long hours, left the child at homein the care ofa non-English speaking nanny. Thenanny was specifically instructed not to speak tothe child in her native language. The child de-veloped no language and was diagnosed withPDD at age 2 years. It is, of course, impossibleto draw a causal relationship without knowledgeof the major verbal and other contingencies thatmay have been in effect during the two-year pe-riod (e.g., How much did the parents talk to thechild when they were at home in the eveningsand weekends? Did the nanny acquire some En-glish and begin to speak to the child during thetwo year period?) This case, however, illustrateshow it might be possible for extinction to oper-ate in a seemingly normal home environment ifa high hourly rate of conversation betweencaregiver and child does not occur.

Relatively few children who spend time inalternative placements will be seriously lan-guage delayed, and even fewer will be diag-nosed with autism or PDD. However, as Hartand Risley (1995, 1999) have shown, degreeof language delay is relative and is directly re-lated to the frequency of talk between parentand child that infants and children receive dur-ing the first three years of life.Other environmental events that may disrupt

or reduce the frequency of reinforcement forverbal behavior or place it on extinction dur-ing the critical first two years of life have beendiscussed by Fowler (1990). Some of theseinclude prolonged physical or emotional illnessof one or both parents, death of a parent, mov-ing to a new residence, increased work de-

mands on the time of one or both parents, andplacing the child with a new caregiver or nanny.Extinction of verbal behavior might also oc-cur when children are allowed to sleep or re-main isolated in their crib without adult inter-action for inordinate amounts oftime each dayover an extended period of weeks or months.

5. Interaction between Organic or PresumedOrganic Factors and Behavioral Factors

Certain physical disabilities such as hearingloss, chronic ear infections, or prolonged ill-ness, especially when occurring during the firsttwo years of life, can directly interfere with theestablishment ofverbal behavior (Bijou, 1966,1983). Other physical disabilities may have nodirect effect on a child's ability to produce ver-bal behavior. However, it is the reaction oftheparents or caregivers to the disability or pre-sumed disability that may function to reducesubsequent requirements for verbal behaviordue to a fear of precipitating additional prob-lems. In one case, the parents of a 2 1/2-year-old child who was later diagnosed with autism,believed their child had chronic ear infectionsbecause he screamed and covered his earswhenever they spoke to him. They discontin-ued their efforts to teach language because theybelieved it caused him pain. In another case,the parents of a 3-year-old verbally delayed,asthmatic child decreased their attempts toteach language for fear ofprecipitating an asth-matic attack (Drash & Tudor, 1989).

6. Non-Suppression ofDisruptive Behaviorand Failure to Establish Early VerbalInstructional Control and Compliance

In contrast to the first five paradigms thatdetail how reinforcement contingencies mayoperate to create specific behavioral repertoiresconsistent with a diagnosis of autism, this sec-tion describes the absence ofspecific caregiverbehaviors that may contribute to the establish-ment ofdisruptive behaviors typically observedin children labeled autistic. From birth to threeyears ofage many typical children engage in avariety of behaviors designated as disruptive,oppositional, defiant, or noncompliant (i.e.,"the terrible twos"). During these years mostparents attempt to reduce or eliminate thosebehaviors and strengthen socially acceptableresponses.


In the case of children diagnosed with au-tism, the elimination of disruptive andnoncompliant behavior is even more critical.As stated by Charlop-Cristy and Kelso (1997),"The rationale behind compliance training issimple-if the child does not comply, then hewill not learn! Compliance plays a vital role inevery aspect of the learning situation" (p. 53).

Parents and caregivers of children labeledas autistic often demonstrate lack of controlover the disruptive and noncompliant behav-iors of their children. Such behaviors may in-clude screaming, severe temper tantrums, kick-ing, hitting, biting, throwing objects, jumpingon furniture, damaging property, and runningabout uncontrollably. These behaviors are quitefunctional for the child in at least two ways.First, they allow the child to obtain reinforce-ment, for example, by screaming until he isgiven a specific toy, food, or other reinforcer,and second, they allow the child to avoid orescape compliance with parental or caregiverdemands or requests. Parents often report thatthey believe these behaviors are an integralcomponent of their child's disability, and thattheir child is unable to control the behavior.Parents often do not provide the consequencesnecessary to reduce or eliminate these behav-iors, believing that doing so might cause addi-tional problems. (See also Paradigm 5 above.)

In some cases the non-confrontive behaviorof parents may have been shaped over a pe-riod ofmonths or years by the aversive contin-gencies of the child's behavior. After manyunsuccessful attempts to reduce or eliminatedisruptive and noncompliant behaviors, parentsmay simply discontinue attempting to disci-pline the child as they might a typical child.These disruptive and oppositional behaviorsbecome increasingly more severe and difficultto manage as the child becomes older and stron-ger.

In one extreme but illustrative case, a 2-year-old child was referred for failure to developlanguage. The parent's main concern, surpris-ingly, was not the child's language delay, butrather the child's severe temper tantrums andaggressive behavior. The mother reported thatshe was forced to carry him wherever she went.When she attempted to put the child down, hescreamed and attacked her viciously by biting,pinching, hitting, and pulling her hair until shepicked him up and cuddled him, thus furtherreinforcing the aggressive behavior. The

mother was adamant that these behaviors werepart of his disability and could not bring her-self to discipline him.

How THE PRESENCE OF DISRUPTIVE ANDAVOIDANT RESPONSES AND THE LACK OF VERBALBEHAVIOR MAY CONTRIBUTE TO DEFICIENCIES IN

SOCIAL-EMOTIONAL DEVELOPMENT

During the first two years of a child's life,parents provide the vast majority of stimula-tion, teaching, and reinforcement necessary forchildren to acquire verbal behavior, socialbonding, and a variety of social skills (Hart &Risley, 1995, 1999). During the first year, typi-cal infants receive ongoing informal trainingin social and pre-language skills from theirparents or caregivers on a daily basis (Bruner,1983). Parents routinely reinforce a variety ofsocial and pre-language behaviors includingeye-contact, responding to name, kissing, hug-ging, babbling, cooing, clapping, singing, smil-ing, laughing, looking at books, pointing topictures and objects, playing interactive games,and following simple instructions. In so doinga repertoire of receptive language (i.e., verbalinstructional control and compliance) is beingestablished, and the child's behavior is beingbrought under the stimulus control of verbalbehavior (i.e., "Show me your nose, eyes, ears,""Touch the apple," "Give me the ball," "Whereis your bottle, teddy bear?" etc.) Consequently,by 18 to 24 months the typical toddler has de-veloped an extensive and relatively complexrepertoire ofsocial behaviors, a relatively largereceptive vocabulary, is coming under thestimulus control of verbal behavior, and is be-ginning to develop expressive verbal behav-ior.However, if during the critical first 12 to 18

months, negative manding and disruptiveavoidance responses are reinforced andstrengthened, this predominant response rep-ertoire may become incompatible with the de-velopment of interactive parent-child socialbehaviors. Consequently when parents attemptto teach social and pre-language behaviors, pre-established patterns of aversive manding anddisruptive avoidance behavior may function todecrease the time parents spend teaching theirchild. This may ultimately produce large defi-cits in social-emotional and pre-language be-haviors normally established during the firsttwo years of life, and the child may begin in-


creasingly to avoid interactions with adults andother children.

Ifpervasive extinction, as described in Para-digm 4, is in effect rather than aversivemanding, then, by definition, limited teachingopportunities will have occurred, and very fewage-appropriate social behaviors will have beenestablished. Extinction may ultimately producea child whose behavior is relatively unrespon-sive to human interaction.Once deficits in the social-emotional reper-

toire occur, either through extinction or avoid-ance, it becomes increasingly difficult to en-gage the child in effective teaching interactions.Parental time that might ordinarily be spentteaching language and social behaviors, as wellas in establishing novel conditioned reinforc-ers, may be directed toward avoiding interac-tions that occasion disruptive or avoidancebehavior. This is illustrated by the example ofthe 2-year-old described in Paradigm 6. Thechild's disruptive behaviors were so severe thatwhenever the mother attempted to engage himin a teaching interaction, he immediately be-gan screaming and pulling her hair. The motherconsequently terminated her attempts to teachthe child.

How CONTINGENCIES MAY CONCURRENTLYSHAPE RESTRICTED, REPETITIVE, ANDSTEREOTYPED RESPONSE REPERTOIRES

Children who have not acquired age-appro-priate verbal behavior and social repertoires by2 years of age are necessarily restricted to avery limited set ofresponses typical ofpre-lin-guistic infants and children. These are prima-rily cause-and-effect activities that provideautomatically reinforcing sensory stimulation.Typical behaviors include thumb-sucking,mouthing and banging objects, finger flicking,hand flapping, spinning objects, inspectingspecific aspects oftoys, or rubbing the surfaceof objects. As the child grows older many ofthese behaviors may come under the controlof other environmental contingencies. For ex-ample, an automatically reinforcing behaviorsuch as pressing one's eye, may be uninten-tionally reinforced and shaped into a more se-rious self-injurious behavior by the attemptsof caregivers to prevent the behavior. In addi-tion to automatic reinforcement, research hasshown that self-injurious, repetitive, andperseverative behaviors may be reinforced and

maintained by positive reinforcement or byavoiding or escaping aversive contingencies(Durand, 1999; Iwata, Dorsey, Slifer, Bauman,& Richman, 1994).

Since the development in young children ofconditioned social reinforcers, such as playingball, tag, or hide and seek, is dependent in largemeasure on verbal behavior and verbal instruc-tional control, a lack ofverbal behavior neces-sarily restricts the child's repertoire to thoserepetitive and stereotyped behaviors typicallyassociated with younger children.

ADVANTAGES OF CONCEPTUALIZING AUTISMAS A CONTINGENCY-SHAPED DISORDER

OF VERBAL BEHAVIOR

Conceptualizing autism as a contingency-shaped disorder of verbal behavior representsa significant departure from the contemporaryneurobiological theories of its etiology andprovides a new paradigm (Kuhn, 1966) forbehavioral research in autism. It is relevant toask what changes may result from thisconceptualization and whether these changesmay represent a substantial improvement overthe current approaches to prevention, early in-tervention, and treatment of autism. A numberof the potential changes and concomitant ad-vantages that might result from thisconceptualization are discussed below.

Conceptualizing Autism as a Contingency-Shaped Disorder of Verbal Behavior Createsa New Paradigmfor Behavioral Researchand Treatment in Autism

Since autism has previously been viewed pri-marily as a neurobiological disorder, there hasbeen little incentive for behavioral researchersin the field of autism to conduct experimentalanalyses of the contingencies of reinforcementthat exist between parent and child from birthto one year, especially as those contingenciesrelate to the shaping of verbal behavior. More-over, there have been and continue to be sig-nificant "politically correct" pressures, bothfrom within and outside the profession, that havestrongly mitigated against conducting researchthat might implicate contingencies ofreinforce-ment between parents and children as signifi-cant factors in the etiology ofautism. However,when autism is viewed, not as a neurobiologi-cal disability or a disease entity, but as a con-


tingency-shaped disorder of verbal behavior,identifying the specific contingencies that mayprevent or hinder acquisition of verbal behav-ior becomes a first priority for behavioral re-search with the objective ofprevention and ear-lier intervention.There are at least two lines of research that

may be immediately productive in evaluatingthe effects of reinforcement contingencies inthe shaping of behaviors that may later resultin the diagnosis ofautism. Behaviors correlatedwith a later diagnosis of autism can be identi-fied at least as early as 8 to 10 months, andperhaps earlier (Werner, Dawson, Osterling, &Dinno, 2000). Consequently, behavioral re-search focusing on specific parent-child rein-forcement contingencies that establish the ini-tial stages ofverbal behavior between birth and12 months would be particularly useful inshowing how the precursors of verbal behav-ior are shaped and precisely what behaviorsmay be incompatible with or prevent the ac-quisition of verbal behavior. Further analysisof the role of automatic reinforcement as it re-lates to the shaping of verbal behavior duringthe first year would also be quite valuable(Bijou & Baer, 1965; Skinner, 1957; Smith,Michael, & Sundberg, 1996; Sundberg &Michael, 2001; Sundberg, Michael, Partington,& Sundberg, 1996).A second related line of research would be

to investigate the role ofaversive manding andother avoidance behaviors in preventing theacquisition ofverbal behavior. Since avoidancebehaviors and aversive manding are stronglyimplicated as casual factors in this analysis oflanguage delay, it would be important to ex-amine the early stages of language acquisitionto evaluate precisely how aversive mandingand disruptive avoidance behaviors are origi-nally established and how these behaviors mayfunction to prevent acquisition of appropriateverbal and social behavior. It would also beuseful to analyze why and how avoidance be-haviors are shaped in some children and notothers.

Preventing Autism Now: A Practical Behav-ioral Strategy that Can Be ImplementedImmediately

One of the major goals of both behavioraland biomedical research in autism during re-cent years has been to identify causal variables

that might lead to comprehensive programs forthe prevention or cure of autism that could beimplemented immediately (i.e. Cure AutismNow [CAN]; Defeat Autism Now! [DAN!]).This goal also has been strongly supported byparent advocacy groups nationwide (Grossman& Beck, 2002; Jacobson, 2000; Maurice, 1996,2001; Perry, 2001).

This analysis of autism as a contingency-shaped disorder of verbal behavior providesBehavior Analysts with the unique ability toanswer one ofthe most long-standing and chal-lenging questions of parents and parent advo-cates nationwide: "How can autism be pre-vented?" Based on the current analysis, itshould be possible to begin to prevent manycases ofautism immediately by identifying re-inforcement contingencies that might preventor inhibit the development of verbal behaviorduring the period between birth and 18 to 24months and replacing them with reinforcementcontingencies that could establish age-appro-priate verbal and social behavior.One important component ofABA preven-

tion programs would be to establish, as soonas feasible, a nation-wide network ofAppliedBehaviorAnalysis screening programs for par-ents and their infants between birth and 24months that would focus on identifying andmodifying contingencies and behaviors thatmay interfere with acquisition of age-appro-priate verbal behavior. Once these contingen-cies are identified, parents could be taught toreplace them with contingencies that will shapeage-appropriate verbal and social behavior byage 2 to 3 years. In those cases in which com-plete prevention might not occur, preventionefforts could lead directly into ABA early in-tervention programs. For a related discussionof behavioral prevention programs, see Drashand Tudor (1990, 1993).Based upon the success ofprevious preven-

tion and early intervention programs, there isa high probability that the more severe symp-toms of autism might never occur (Anderson,Avery, DiPietro, Edwards, & Christian, 1987;Begab, Haywood, & Garber, 1981; Bijou,1983; Birnbrauer & Leach, 1993; Drash, 1992;Drash & Raver, 1987; Drash, Raver & Murrin,1987; Fenske, Zalenski, Krantz, &McClannahan, 1985; Garber, 1988; Guralnick,1997; Harris, Handleman, Gordon, Kristoff, &Fuentes, 1991; Lovaas, 1987; Menolascino &Stark, 1988; Smith, Groen, & Wynn, 2000).


Developing a Contingency-based Strategyfor Earlier Intervention

One of the principal goals of contingency-based early intervention programs would be tobegin ABA intervention programs before therepertoires ofaversive manding and disruptiveavoidance behaviors are well established. Asshown by several of the cases presented previ-ously, the disruptive and avoidance behaviorsof children diagnosed with autism/PDD areoften well established by 18 months to twoyears.

If behavioral interventions that are focusedon the development of appropriate verbal andsocial behavior and elimination of disruptiveand avoidance behaviors are begun during thefirst year to 18 months, the probability for to-tal recovery may be greatly enhanced. Green,Brennan, and Fein (2002) recently reported ona case in which early intensive behavioral treat-ment of a 1-year-old child at high-risk for au-tism produced total recovery within a periodofthree years. This study demonstrated the ef-fectiveness ofABA intervention for autism ata younger age than has been previously re-ported.

Reducing the Number of Treatment Hoursper Week and the Total Length of Treatment

Since the degree ofpre-language or languagedelay is necessarily small at age 6 to 18 monthsand the incompatible avoidance behaviors usu-ally are not well established, it should be pos-sible to restore young children to "relativelynormal functioning" much more rapidly thanolder children who have well established rep-ertoires ofdisruptive and avoidance behaviors.The Lovaas (1987) program required 2 to 3years of 40-hours-per-week intensive indi-vidual ABA treatment to achieve recoverywhen children began treatment at an averageage ofthree years. Therefore, ifABA treatmentis begun between 6 and 18 months, it shouldbe possible to restore a child to relatively nor-mal functioning within one to two years.Our research with children diagnosed as au-

tistic, PDD, or at-risk in the age range of 18months to 2 1/2 years suggests that, for somechildren, a program of far less intensity than a40-hour-per-week program may be sufficientto produce substantial recovery within one totwo years. In one previously published case, a19-month-old, seriously language-delayed

child, diagnosed as at-risk, achieved relativelynormal language and behavioral functioningafter 10 months of in-office, ABA verbal be-havior therapy consisting of a total of only 52one-hour sessions (1 to 2 sessions per week)(Drash & Tudor, 1989, pp. 30-31; 1990, pp.199-201). In a second case, a 2 1/2-year-oldnonverbal child, diagnosed with aphasia,achieved relatively normal language and be-havioral functioning after 11 months of in-of-fice, ABA verbal behavior therapy consistingof 36 one-hour sessions (1 to 2 sessions perweek) (Drash, 2001).The total number of therapy hours provided

to each of these two children was only a frac-tion of that typically provided in a 40-hour-per-week, in-home ABA program over a com-parable duration. For the first child, the com-parison is 52 hours vs. 1,600 hours, or 3%, andfor the second child 36 hours vs. 1,760 hours,or 2%. At present these two cases clearly rep-resent the exception rather than the rule. How-ever, they suggest that, in some cases, both thelength and intensity of treatment for youngerchildren, especially those considered at-risk,may be considerably less than that required forolder children.

Making "Functional Recovery" a Routineand Expected Outcome in ABA Treatment ofYoung Children with Autism

As a result of recent advances in the field ofApplied Behavior Analysis, relatively total re-covery or cure is now recognized as a legiti-mate and obtainable outcome in the behavioraltreatment of autism (Maurice, 1993; Maurice,Green, & Foxx, 2001; Maurice, Green, & Luce,1996; McEachin, Smith, & Lovaas, 1993;Lovaas, 1987). At present, however, total re-covery as an outcome remains the exceptionrather than the rule in the treatment of autism.Moreover, some career experts in autism con-tinue to maintain that total recovery is impos-sible (Mesibov, 1997) and appear unwilling toconsider evidence that indicates children havetotally recovered.The dispute over the term "total recovery"

greatly detracts from the fact that hundreds ofchildren are daily making excellent progressin ABA treatment programs, and many are, forall practical purposes, substantially recoveringfrom autism (Maurice, 2001). To avoid thecontroversy surrounding the term "total recov-


ery" we propose the term "functional recov-ery" that would operationally define the con-dition ofmany children who have made excel-lent improvement in language, intelligence, andsocial behavior, but who may or may not bedescribed as "totally recovered." Objectivemeasures might include successful function-ing in regular grade, ability to interact appro-priately and independently in social situations,elimination of temper tantrums and other dis-ruptive or avoidance behaviors, and normalfunctioning on standardized tests of language,intelligence, social adjustment, and academicsubjects. This would allow for the establish-ment of an objective continuum of treatmentoutcomes, ranging from slight or no improve-ment, to partial recovery, tofunctional recov-ery, to total recovery. By describing each cat-egory in operational terms based on direct, rep-licable observations of behavior, much of thecontroversy regarding the effects ofABA treat-ment could be avoided, and the realistic ben-efits of ABA treatment could be more easilyrecognized and accepted.

CONCLUSION

Conceptualizing autism as a contingency-shaped disorder ofverbal behavior that beginsduring the first year oflife provides a new para-digm for behavioral research and treatment inthe areas ofprevention, earlier intervention, andrecovery from autism. By beginning ABAtherapy during the first 6 to 18 months and fo-cusing on modification of the specific contin-gencies of reinforcement that may prevent orinterfere with acquisition of verbal behavior,it may be possible to prevent many cases ofautism and reduce the debilitating effects ofmany others. The duration and intensity oftreatment may be reduced, and "functional re-covery" from autism might become the normand expected outcome ofABA treatment.

REFERENCES

Alexander, D., Cowdry, R. W., Hall, Z. W., &Snow, J. B. (1996). The state of science inautism: A view from the National Institutesof Health. Journal ofAutism and Develop-mental Disorders, 26, 117-119.

American Psychiatric Association. (1994). Di-agnostic and statistical manual of mental

disorders (4th ed.). Washington, D.C.: Au-thor.

Anderson, S. R., Avery, D. L., DiPietro, E. K.,Edwards, G. L., & Christian, W. P. (1987).Intensive home-based early interventionwith autistic children. Education and Treat-ment ofChildren, 10, 352-366.

Autism Society ofAmerica. (2002). Advocate(4th ed.) 35, 4. Bethesda, MD: Author.

Bailey, A., Phillips, W., & Rutter, M. (1996).Autism: Towards an integration of clinical,genetic, neuropsychological and neurobio-logical perspectives. Journal ofChild Psy-chology & Psychiatry & AlliedDisciplines,37, 89-126.

Barton-Cohen, S., Allen, J., & Gillberg, C.(1992). Can autism be detected at 18months? The needle, the haystack, and theCHAT. British Journal ofPsychiatry, 161,839-843.

Begab, M. J., Haywood, H. C., & Garber, H. I.(Eds.). (1981). Psychosocial influences inretardedperformance (Vols. 1 & 2). Balti-more: University Park Press.

Bijou, S. W. (1966). A functional analysis ofretarded development. In N. R.. Ellis (Ed.),International Review ofResearch in Men-talRetardation. New York: Academic Press.

Bijou, S. W. (1983). The prevention of mildand moderate retarded development. In F.J. Menolascino, R. Neman, & J. A. Stark(Eds.), Curative aspects ofmental retarda-tion: Biomedical and behavioral advances(pp. 223-241). Baltimore: Brookes.

Bijou, S. W., & Baer, D. M. (1965). ChildDe-velopment II: Universal stage of infancy.Englewood Cliffs, NJ: Prentice-Hall.

Bijou, S. W., & Ghezzi, P. M. (1999). The be-havior interference theory ofautistic behav-ior in young children. In P. M. Ghezzi, W.L. Williams, & J. E. Carr (Eds.), Autism:Behavior analyticperspectives (pp. 33-43).Reno, NV: Context Press.

Bimbrauer, J. S., & Leach, D. J. (1993). TheMurdoch early intervention program after2 years. Behaviour Change, 10, 63-74.

Bruner, J. (1983). Child's talk: Learning to uselanguage. New York: W. W. Norton.

Carr, E. G., & Durand, V. M. (1985). Reducingbehavior problems through functionalcommunication training. Journal ofAppliedBehavior Analysis, 18, 111-126.

Carr, E. G., & Durand, V. M. (1986). The so-cial-communicative basis of severe behav-

an analysis of autism as a contingency shaped disorder of verbal behavior-drash and tudor 2004

Health & Medicine