Embed Size (px)
Citation preview
Effects of Social Reinforcement on the Emission of Tactsby Preschoolers
Carly Moher Eby and R. Douglas GreerTeachers College, Columbia University
We conducted 2 experiments on the effects of social attention versus token contingen-cies on the emission of verbal operants by preschoolers, with and without a disabilitydiagnosis. Four participants, 3 females and 1 male, 3 to 4 years old, were selected toparticipate in Experiment 1 and 6 participants, 5 females and 1 male, 2 to 4 years old,in Experiment 2. Experiment 1 compared effects of the 2 contingencies on numbers ofchild-initiated tacts in 3 different settings using an alternating treatment design.Experiment 2, using a multielement design, compared the automated delivery of tokensversus adult attention on the percentage of peer-to-peer and adult conversational units.Participants in both experiments initiated more tacts with contingent social attentionthan with contingent tokens. Implications are that tacts and conversational units aremaintained more by social reinforcers than nonsocial generalized conditioned reinforc-ers (i.e., tokens). Social control of tacts may be essential to social verbal behavior.
Keywords: generalized conditioned reinforcement, social reinforcement, tacts, tokens,verbal behavior analysis
The acquisition of language begins with anenvironment rich in language and social inter-actions (Hart & Risley, 1995). A critical com-ponent of that environment is the presence ofspeaking adults and the interactions that occurbetween adult caretakers and children. For in-stance, in Hart and Risley’s study, parents whowere (a) more responsive to their children, (b)spent more time talking to their children, (c)used a greater variety in language, and (d) en-gaged in longer interactions, had children withgreater vocabulary growth, better use of vocab-ulary, and higher IQ scores.
In the study of language and its developmentin children, one central issue concerns the con-ditions under which children come to learn themeaning of words (Bloom, 2002; Hoff & Shatz,
2009; Tomasello & Bates, 2001). In Skinner’sanalysis of verbal behavior, a major componentof vocabulary use is known as a tact repertoire(Skinner, 1957). Skinner defined the tact as a“verbal operant in which a response of a givenform is evoked by a particular object or event orproperty of an object or event” (pp. 81–82).Tacts are responses in a given community thatare occasioned by objects, events, people, oractions that the speaker contacts in his or hercurrent environment. Learning a fluent tact rep-ertoire is the cornerstone of being a true speakerand is a necessary component of becoming trulyverbal (Greer & Speckman, 2009).
When Skinner (1957) first defined the tact asone of the elementary verbal operants, he iden-tified its functional reinforcer as generalizedconditioned reinforcement. By definition, gen-eralized conditioned reinforcers are reinforcersthat have been paired with unconditioned orconditioned reinforcers, and as a result the mo-mentary conditions of the organism (e.g., moti-vating operations, such as satiation or depriva-tion) are not likely to affect their efficacy.However, Skinner’s discussion about tacts andgeneralized conditioned reinforcers precededthe development of the token as a generalizedreinforcer in applied behavior analysis (Ayllon& Michael, 1959; Ayllon & Azrin, 1968).
This article was published Online First January 23, 2017.Carly Moher Eby and R. Douglas Greer, Department of
Health and Behavior Studies, Teachers College, and theGraduate School of Arts and Sciences, Columbia Univer-sity.
This paper is based on the first author’s doctoral disser-tation.
Correspondence concerning this article should be ad-dressed to Carly Moher Eby, who is now at The NewEngland Center for Children, 33 Turnpike Road, Southbor-ough, MA 01772. E-mail: [email protected]
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Behavioral Development Bulletin © 2017 American Psychological Association2017, Vol. 22, No. 1, 23–43 1942-0722/17/$12.00 http://dx.doi.org/10.1037/bdb0000043
23
Moreover, Skinner emphasized that one of thedefining characteristics of verbal behavior isthat it’s behavior that is socially mediated.
Generalized conditioned reinforcers, accord-ing to their use in applied behavior analysis, arenot necessarily social in nature. Money, tokens,and social attention are all classified together asgeneralized conditioned reinforcers, despitetheir vastly different physical dimensions andtheir different social functions. Money is a gen-eralized conditioned reinforcer or token for arange of behaviors because it can lead to manyunconditioned reinforcers (e.g., food, warmth)or conditioned reinforcers (e.g., travel, books,music) and so it is relatively independent ofmomentary deprivations (Catania, 2007; Coo-per, Heron, & Heward, 2007; Greer & Ross,2008; Kazdin, 1984; Moore, 2008; Skinner,1953, 1957). In educational research and appli-cations, token economies are commonly used asgeneralized conditioned reinforcers. Anothertype of generalized conditioned reinforcer, so-cial attention or approval, differs from bothmoney and tokens in terms of its stimulus prop-erties, how it’s delivered as a reinforcer, and theabsence of an exchange for a “back-up” rein-forcer (Skinner, 1953). In the case of attentionand approval of a listener, the reinforcement fora speaker is the direct outcome of the response,whereas nonsocial generalized conditioned re-inforcers are representative or tokens of rein-forcement. For example, the child who notices apassing cat and says to her mother, “I see a cat,”experiences a reinforcing consequence whenher mother replies, “Yes, I see the cat, too.There goes the cat.” This illustrates a socialgeneralized conditioned reinforcer. In contrast,the child who has learned that desirable class-room behavior affords her tokens that can beexchanged for activities or privileges, such asaccess to hall passes or extra recess time, expe-riences a reinforcing consequence when the to-kens have been exchanged for the special activ-ity or privilege, and this is arguably independentof any attention or approval that occur as part ofthat exchange, thus illustrating the nonsocialgeneralized conditioned reinforcer.
Some evidence suggests that the tact is fun-damentally a social operant that is controlled bysocial attention of others (Greer & Ross, 2008;Schmelzkopf, 2010). That is, if the tact andrelated social verbal behavior is indeed social,
then requiring social reinforcement as part ofthe objective for teaching or inducing the tactrather than other forms of generalized condi-tioned reinforcers (e.g., tokens) may be critical(Greer & Du, 2015). Research with infants andthe development of preverbal skills may pro-vide a better understanding of the role that so-cial reinforcement plays in the tact operant. In aseries of studies with infants, Pelaez, Virues-Ortega, and Gewirtz (2011a, 2011b) demon-strated that infant vocalizations increased withcontingent maternal vocalizations, and that con-tingent maternal vocalizations (either vocal im-itation or motherese speech) produced greatereffects (infant vocalizations increased more)than did noncontingent maternal vocalizations.
The available definitions and uses of the vo-cabulary pertaining to generalized conditionedreinforcers have varied widely, and subse-quently the applications in the verbal behaviorresearch of the use of generalized conditionedreinforcers have also shown inconsistencies.For instance, investigators have included oper-ations such as “non-specific” reinforcement in-cluding food and toys (Braam & Sundberg,1991; Stafford, Sundberg, & Braam, 1988),praise (Carroll & Hesse, 1987; Sundberg, Juan,Dawdy, & Arguelles, 1990), food and praise(Partington & Bailey, 1993), tokens and praise(Williams, Carnerero, & Perez-Gonzalez,2006), or opportunities to mand (Greer & Ross,2008). When it comes to the reinforcement oftacts, a necessary condition may be that there isa social dimension to the reinforcement thatmaintains them. The lack or low incidences ofinitiated language interactions in individualswith autism diagnoses may be a related issue.As outlined in the Diagnostic and StatisticalManual of Mental Disorders (5th ed.; AmericanPsychiatric Association, 2013), the diagnosticcriteria for autism spectrum disorder includes,“Deficits in social-emotional reciprocity, rang-ing, for example, from abnormal social ap-proach and failure of normal back-and-forthconversation; to reduced sharing of interest,emotions, or affect; to failure to initiate or re-spond to social interactions” (American Psychi-atric Association, 2013, p. 50).
Prior research has not addressed the possibledistinction between the various types of gener-alized conditioned reinforcers as they apply tosocial verbal behavior. This distinction is animportant one to make because it is critical that
24 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
a relevant reinforcer that controls a particularoperant is present. Catania (2007) noted that itis not only the behavior that is learned but alsothe entire three-term contingency including thereinforcer. Skinner (1957) proposed that verbalbehavior was essentially social behavior. Themand requires that another mediate the environ-ment for the speaker, but its function is onlyindirectly social. The social function for themand is that of a social contract where thelistener provides something other than socialinteraction. In fact, some have argued that themand is indistinguishable from nonverbal be-havior unless certain contextual factors are inplace (Barnes-Holmes, Barnes-Holmes, & Cul-linan, 2000). The tact appears to be the basicspeaker verbal operant that is directly controlledby social contingencies. The onset of the tact asa verbal behavioral developmental cusp maydepend on the onset of social attention as aconditioned reinforcer (Greer & Du, 2015).
The research reported here was designed toexamine whether and how young childrenwould respond differentially to social attention(such as adult praise or approval) versus non-social generalized reinforcers (such as tokens).In Experiment 1, we sought to determinewhether contingent experimenter delivered so-cial attention or contingent experimenter-delivered tokens differentially affected theemission of child-initiated tacts in an appliedsetting. In Experiment 2, the research was ex-tended to a laboratory analysis where we elim-inated adult social contact for token delivery bycomparing social reinforcers delivered by anadult versus tokens delivered through a chute,on the emission of participant-initiated tacts andother social verbal operants in social settings.
General Method
Overview
In both experiments, we investigated the ef-fects of adult social attention versus token con-tingencies on the emission of verbal operants.The method components that were common be-tween both experiments will be reported here,whereas components that were specific to onlyone of the experiments will be reported sepa-rately in the Method sections of each experi-ment.
Participants
Refer to Table 1 for detailed informationabout the participants. The participants in Ex-periment 1 were three girls and one boy, ages 3to 4 (Participants A, B, C, and D). The partici-pants in Experiment 2 were five girls and oneboy, ages 3 to 4 (Participants A, B, E, F, G andH). Two of the participants (A and B) wereselected for both experiments. Participants wererecruited from local preschools.
Dependent Variable and OperationalDefinitions
The primary dependent variable of interestwas frequency of tacts. The tact was defined bySkinner’s (1957) as “a verbal operant in whicha response of a given form is evoked by aparticular object or event or property of anobject or event” (pp. 81–82). Examples of tactsemitted in the experiments reported here in-clude verbal statements such as, “look at thesnow,” “this is a coconut tree,” and “I’m doinga puzzle.”
In addition to tacts, four other verbal re-sponses were recorded. These included mands,intraverbals, conversational units, and wh-questions. A mand, as defined by Skinner(1957), is “a verbal operant in which the re-sponse is reinforced by a characteristic conse-quence and is under the functional control ofrelevant conditions of deprivation or aversivestimulation” (p. 36). An example of a mandemitted in the current study was a participantstating “can we go inside the room now?” whenthe door was closed and she had been asked towait outside of the room. Intraverbal behavior isverbal behavior evoked by another verbal re-sponse, without correspondence between thetwo responses (Skinner, 1957). An example ofan intraverbal response emitted in the currentstudy was when one participant stated, “whatletter is it?” and a second participant stated, “S.”“Wh” questions refer to any “Who,” “What,”“Where,” “Why,” or “How” questions. Conver-sational units are verbal episodes in which atleast two individuals exchange roles of speakerand listener. That is, Person A emits a verbalresponse in the presence of a listener (Person B)and Person B responds as both a listener andspeaker, then Person A responds as a listener,completing the conversational unit (Donley &
25SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Table 1Experiment 1 and 2 Participant Characteristics
Participants Age/GenderLevel of verbal cusps/verbal
capabilities Diagnosis/Standardized test scores
A (Experiment 1 & 2) 3.4/Female - Listener/Speaker -Expressive Language Disorder andAdjustment Disorder
- Instructional control, independentmands/tacts with autoclitics,self-talk, conversational units,book stimuli conditionedreinforcement for observing
- Weschler Preschool & Primary Scalesof Intell.-III Full Scale SS:101;Performance SS:97; Verbal SS:104
B (Experiment 1 & 2) 4.0/Female - Listener/Speaker/Early reader/Early writer
- No diagnosis/“Typically developing”
-Instructional control, independentmands/tacts with autoclitics,naming, observational learning,self-talk, conversational units,say-do in speaker as ownlistener function, book stimuliconditioned reinforcement forobserving
C (Experiment 1 only) 3.9/Male -Listener/Speaker/Early reader/Early writer
-PDD-NOS
-Instructional control, independentmands/tacts with autoclitics,listener half of Naming,appropriate self-talk duringfantasy play, conversationalunits, book stimuli conditionedreinforcement for observing,emits some palilalia
-Preschool Language Scale-4:Expressive: SS:112; AuditoryComprehension: SS:110
D (Experiment 1 only) 4.0/Female -Listener/Emergent speaker -“Preschooler with a disability”-Instructional control, independent
mands with autoclitics, bookstimuli conditionedreinforcement for observing
- Preschool Language Scale-4
Auditory Comprehension SS:75;Expressive SS:73
- Weschler Preschool & Primary Scalesof Intell.-III Full Scale: SS:79;Performance: SS:76; Verbal:84
E (Experiment 2 only) 4:0/Male -Listener/Speaker/Early reader/Early writer
- No diagnosis/“Typically developing”
-Instructional control, independentmands/tacts with autoclitics,listener half of Naming,observational learning, self-talk,conversational units, say-do inspeaker as own listener function,book stimuli conditionedreinforcement for observing
F(Experiment 2 only) 3:11/Female -Listener/Speaker/Early reader/Early writer
- No diagnosis/“Typically developing”
-Instructional control, independentmands/tacts with autoclitics,self-talk, conversational units,say-do in speaker as ownlistener function, book stimuliconditioned reinforcement forobserving
26 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Greer, 1993; Greer & Ross, 2004; Lodhi &Greer, 1989).
Independent Variable
The independent variable in both experi-ments was the manipulation of the reinforce-ment contingency. Thus, the variable that wasmanipulated was either the delivery of tokens orthe delivery of social attention, depending onthe experimental condition in effect. In the at-tention condition, any tact emitted by partici-pants was followed by vocal and nonvocal so-cial attention from the experimenter. Forexample, if the participant said, “It’s a guitar,”the experimenter responded with a vocal re-sponse such as, “you’re right, that is a guitar!”as well as a nonvocal response, such as a smile,a nod, or a light pat on the back. This wasconsistent across both experiments.
In the token condition, any tact emitted by theparticipants was followed by delivery of a tokeninto the participant’s clearly labeled cup. InExperiment 1, tokens were experimenter-delivered. In Experiment 2, tokens were deliv-ered through a chute. Across both experiments,no vocal responses were given by the experi-menter in the token condition, and at the end ofthe session in the token condition, participantswere given the opportunity to exchange theirtokens for a preferred food or activity item.
Experiment 1
Method
Participants. Four participants, 3 girls and1 boy, were selected for this experiment (Par-ticipants A, B, C, and D). They were selectedfrom a preschool that served children, 2- to
5-years old, with and without disabilities. Theschool employed the Comprehensive Applicationof Behavior Analysis to Schooling (CABAS)model of education. In this model, scientificprocedures are applied to pedagogy, curriculumdesign, classroom management, staff training,and parent education. A description of the par-ticipants is presented in Table 1.
Setting and materials. Sessions were con-ducted in three noninstructional settings: freeplay, structured play, and transition. The firstsetting, free play, consisted of an area of theclassroom that measured 2.4 m by 3 m. The playarea contained a variety of age-appropriate toys,including blocks, puzzles, books, toy vehicles,toy animals, dolls, cause-effect toys, and musi-cal toys. In the second setting, structured play,participants were instructed to go to a specifictable within the classroom. Imaginative playtoys were present, including a play farm withtoy animals, a dollhouse with toy people, and acar garage with toy cars. In the third setting,transition, participants were asked to wait in aspecified hallway area before starting a newactivity. The hallway was an open space outsideof the classroom that measured approximately1.5 m by 2.7 m. It was a small “nook” at the endof a longer hallway. There were two classroomsand one office and the doors to these roomsremained closed during sessions. There weretwo bulletin boards hanging on the walls. Thebulletin boards were covered with colored paperand colorful borders. One bulletin board con-tained graphs depicting class-wide data. An-other bulletin board displayed a winter theme,with artificial snowflakes and photographs ofstudents. On one wall, there were two treesmade of paper taped to the wall. One tree re-sembled an apple tree. These stimuli were often
Table 1 (continued)
Participants Age/GenderLevel of verbal cusps/verbal
capabilities Diagnosis/Standardized test scores
G (Experiment 2 only) 4:1/Female - Listener/Speaker/Early reader/Early writer
- No diagnosis/“Typically developing”
- Instructional control, independentmands/tacts with autoclitics,self-talk, conversational units
H (Experiment 2 only) 2:10/Female - Listener/Speaker/Early reader - No diagnosis/“Typically developing”- Instructional control, independent
mands/tacts with autoclitics,self-talk, conversational units
27SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
what evoked tacts emitted by the participants inthis setting. This setting was set up to resemblea “waiting” condition, in which the participantswere waiting to enter the classroom.
Token cups and tokens were present in thetoken reinforcement sessions. The token cupswere plastic water coloring paint cups with lidsthat had a hole in the center (where tokens wereinserted). The participants were familiar withthe token cups as they were the same as thoseused for reinforcement of academic and socialresponding in their classrooms. The cups werelabeled with the participant’s names and thechildren could identify their names. Tokenswere small, colored, plastic disks. These werethe same tokens that the participants earned intheir classroom token economies already estab-lished. There was a video camera present duringall sessions. The video camera was a Flip digitalcamera measuring 10.5 cm � 5.5 cm. It wasmounted on an adjustable tripod that extendedto a maximum height of 134 cm.
Design. The design was an alternatingtreatment design (Kennedy, 2005). In this de-sign, the researcher alternates between two ormore experimental conditions from session tosession. The two alternating conditions weretoken reinforcement and social attention rein-forcement. Experimental conditions rotated dai-ly. The condition was randomly assigned eachday by “picking out of a hat” the name forwhich reinforcement operation to implement.Thus, two pieces of paper each with a conditionwritten down (i.e., “token” or “attention”) wereplaced in an opaque container and somebodyother than the primary experimenter selected apaper. If the selected paper said “token” thenthe token condition was implemented. Like-wise, if it said “attention” then the attentioncondition was implemented. A total of threesessions took place daily, one in each setting(i.e., transition, free play, structured play). Theexperimental condition (token or attention) re-mained the same for each setting on a givenday. If time ran out on a given day and fewerthan three sessions, or one in each setting, tookplace, then those sessions were carried over tothe next day. Once one session for each settinghad taken place under the most recently selectedcondition (token or attention), then the nextcondition was selected using the “pick out of ahat” method. A minimum of 18 sessions wasconducted for each participant. Thus, at least
three sessions were conducted in each setting/condition combination. For instance, there werethree transition/attention sessions, three freeplay/attention sessions, and three structured-play/attention sessions. The data for each of thethree sessions under each contingency conditionwere also blocked and reported as total re-sponses across all three settings.
Procedure. Two participants were broughtto the experimental setting prior to beginningthe session. Participant pairs varied across ses-sions. Thus, each of the six participants waspaired with every other participant at somepoint. However, the pairings were not system-atically programmed nor were they counterbal-anced. This variable was not controlled for inthis initial experiment; however, they were con-trolled for in Experiment 2. Both of the partic-ipants received the same experimental condi-tions at the same time. Although they wereobserved in pairs, each participant was treatedas a single participant and their responses werefollowed by individualized contingencies. Forinstance, if only Participant A emitted a tact,specific reinforcement was delivered to her in-dividually, such that either the experimenterdelivered vocal praise and also said her name(e.g., “that’s right, A!”), or a token was deliv-ered into Participant A’s clearly labeled cup. Ifboth participants emitted tacts at the same time,specific reinforcement was again delivered onan individual basis. That is, in the case of at-tention reinforcement, the experimenter deliv-ered specific vocal praise to each participant bysaying their names clearly (e.g., “Yes A, it’s acat! You’re right B, that’s a school bus!”). Or, inthe case of token reinforcement, a token wasdispensed into both of the cups as the onlyconsequences, one for each of the participants.Once the participants and the experimenter werein the experimental setting, the experimenterstarted the video camera. All sessions were fiveminutes in duration.
In the free play setting, sessions began withthe experimenter saying, “We are going to playin the toy area. You can talk about anything youwant while we play.” In the structured playsetting, sessions began with the experimentersaying, “We are going to play with toys at thetable and you can talk about anything you wantwhile we play.” In the transition setting, ses-sions began with the experimenter saying, “We
28 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
need to wait here while we get ready for thenext activity, but you can talk while we wait.”
All sessions were videotaped and the exper-imenter remained present during all sessions.The experimenter did not initiate conversationswith the participants in either condition, unlessit was absolutely necessary to give a vocal di-rection for safety purposes. An example of adirection that was given for safety purposes waswhen a participant stood on a chair and theexperimenter said “please keep your feet on thefloor.” She responded to any verbal operantsemitted by the participants according to theexperimental condition that was in effect. Forexample, when the attention condition was ineffect, tacts emitted by the participants resultedin a vocal response. Wh- questions also resultedin vocal responses during the attention condi-tion. Mands sometimes resulted in vocal re-sponses, such as when a participant emitted amand for something from the experimenter andits delivery was not possible. For example,when a participant said, “Can we go in theclassroom now?” the experimenter responded,“Not now, we need to wait.” Otherwise, mandswere not followed by their specified reinforcer,as a way to control for reinforcement of thetarget behavior, which was tacts.
In the token condition, the experimenter didnot initiate any social interactions with the par-ticipants, unless it was necessary to give a di-rection, such as for safety purposes. Tacts emit-ted by participants resulted in a token deliveredinto his or her token cup, but no vocal response.Wh- questions and mands did not result in vocalresponses during the Token condition. Mandswere not followed by their specified reinforcer.
Data collection. Data were collected on allverbal operants emitted in a session. Sessionswere videotaped and viewed at a later time toensure accurate data recording. Data were re-corded by marking an “X” in the appropriatecolumn to identify the type of verbal operantemitted. In the next column, the observer wrotethe exact word or phrase spoken. In cases wherecertain words were unintelligible, the observermarked this with the word unclear in parenthe-ses. In the next column, the observer recordedthe type of consequence that was delivered. Shewrote a T to denote token delivery, A to denoteattention, MR to denote mand reinforcement, orNR to denote no reinforcement. Next, the ob-server recorded the audience to whom the ver-
bal response was directed. The columns werelabeled S (Self), P (Peer), and A (Adult), byplacing an “X” in the appropriate column. Fi-nally there was a “Notes” column where theobserver recorded any additional information.In cases where a participant emitted a tact and awh- question within the same sentence, bothverbal operants were counted and recorded sep-arately. Similarly, any conversational units orintraverbals that also contained tacts werecounted and recorded as both. Refer to Table 2for an example of the verbal operants that wereobserved in one of the sessions in Experiment 1,and how they are recorded and consequated.
Interobserver agreement. A second ob-server, who was a graduate of a Master’s levelprogram in applied behavior analysis and whowas trained in the analysis of verbal behavior,independently viewed and transcribed the videorecorded sessions, and recorded data on thecustom-made data form. Point-by-point interob-server agreement was calculated by dividing thesmaller number of tacts measured by the largernumber of tacts measured, then multiplying by100. For Participant A, IOA was conducted in44% of sessions and mean agreement was87.2% (range: 83%–100%). For Participant B,IOA was conducted in 39% of sessions andmean agreement was 89.1% (range: 80%–100%). For Participant C, IOA was conductedin 11% of sessions and mean agreement was93.8% (range: 88.9%–100%). For ParticipantD, IOA was conducted in 27% of sessions andmean agreement was 83% (range: 81.8%–100%).
Experiment 1 Results
Figure 1 displays a bar graph showing a com-parison of total emissions of verbal operantsunder the conditions of social attention rein-forcement (A) and token reinforcement (T) oftacts in Experiment 1 for Participants A, B, C,and D. It is important to keep in mind that tactemission was the primary dependent variablethat was directly contacted by the tested vari-able. The additional behaviors that were mea-sured were conversational units, intraverbals,wh- questions, and mands. There was no pro-grammed reinforcement for these behaviors, ex-cept when a tact was emitted as part of a con-versational unit, intraverbal, or wh- question.Participant A emitted a total of 103 tacts underattention conditions, versus 71 tacts under token
29SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
conditions. Participant B emitted 120 tacts un-der attention conditions and 70 tacts under to-ken conditions. Participant C emitted 73 tactsunder attention conditions and 43 tacts under
token conditions. Participant D emitted 47 tactsunder attention conditions and 32 tacts undertoken conditions. Refer to Table 3 for the num-bers of verbal operants (tacts, conversational
Table 2Example of the Verbal Responses Recorded in One of the Transition–Attention Sessions in Experiment 1,in Which Tacts Were Reinforced With Social Attention From an Adult
Speaker Target audience Words spoken Verbal operant code Adult response
Participant B Adult “Who was eating in here? I see aDorito.”
WH, T “Oh you see a Dorito, B?”
Participant A Adult “I see a Dorito.” T, IV “You see it too, A?”Participant B Adult “It was raining when I came here.” T “It was raining when I
came, too.”Participant B Adult “When I came out of my mommy’s
car I felt a drip on my head.”T, IV, CU “You did? Whoa!”
Participant A Adult “I have my Tinkerbell umbrella.” T, IV “That’s a good thing, A.”Participant B Participant A “You got a Tinkerbell umbrella?” IVParticipant A Participant B “Yeah, Tinkerbell umbrella.” IV, CUParticipant B Adult “What are we waiting for?” WH, M “We’re waiting to go
inside the classroom.”Participant B Participant A “Who drived you here? Mommy or
the bus?”WH
Participant A Participant B “Mommy.” IV, T “That’s nice that you camewith your mom, A.”
Participant B Participant A “Me too, my mommy.” CU, IV, T “You too, B? Neat!”
Note. WH � Wh- question; T � tact; IV � intraverbal; CU � conversational unit; M � mand.
0
20
40
60
80
100
120
140
Tacts Conversational Units
Intraverbals Wh Questions Mands 0
20
40
60
80
100
120
140
Tacts Conversational Units
Intraverbals Wh Questions Mands
0
20
40
60
80
100
120
140
Tacts Conversational Units
Intraverbals Wh Questions Mands
E stnarepO l abre
V fo rebmu
N lat oTdeni b
moC sgnitteS ll
A nI dettim
Participant A Participant B
Participant C Participant D
Figure 1. The sum of verbal operants emitted across attention and token conditions forParticipants A, B, C, and D in Experiment 1.
30 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
units, wh- questions, intraverbals, and mands)emitted in each setting type (transition, struc-tured play, free play), under both conditions(attention condition and token condition).
Figure 2 is a line graph showing a compari-son of the numbers of tacts emitted in the at-tention versus token conditions. Each data pointrepresents the total number of tacts emitted in acombined 15-min session that was the sum ofall three 5 min settings. All participants emittedmore tacts in the attention condition than in thetoken condition. Participant A emitted a meanof 34 tacts in the attention condition and a meanof 23 tacts in the token condition, across exper-imental days. Participant B emitted a mean of40 tacts in the attention condition and a mean of23 tacts in the token condition. Participant Cemitted a mean of 24 tacts in the attentioncondition and 16 tacts in the token condition.Participant D emitted a mean of 15 tacts in theattention condition and a mean of 10 tacts in thetoken condition. In most cases the repeated ro-tation across the two contingencies resulted inmore distinctive differences in responding un-der the two conditions, with greater responsive-ness to the social attention condition.
Discussion of Experiment 1 and Rationalefor Experiment 2
The results of Experiment 1 showed that tactsoccurred more frequently under social attentionreinforcement conditions as compared with to-ken reinforcement conditions for four partici-pants for whom both adult attention and tokensfunctioned as reinforcers, although ParticipantD showed only slight differences. These resultssuggest that for these children tacts were rein-forced more by social verbal reinforcement.These data suggest that distinctions betweensocial generalized reinforcement and nonsocialgeneralized reinforcement may be important inverbal behavior development.
Two of the settings used in this experimentpresented with some limitations. The data weremore variable in the Structured Play and FreePlay settings than in the Transition setting for 3of the participants (A, C, and D). A possiblereason for the variable data observed in theStructured Play setting is that this setting hadanthropomorphic toys and there may havebeen more self-talk speaker-listener ex-changes (speaker-as-own-listener) involving
Table 3Numbers of Verbal Operants (Tacts, Conversational Units, Wh- Questions, Intraverbals, and Mands)Emitted in Each Setting Type (Transition, Structured Play, Free Play), Under Two Different Conditions(Attention Condition and Token Condition), for Participants A, B, C, and D, in Experiment 1
Participant
Tacts CU WH IV Mands
A T A T A T A T A T
Participant ATransition 46 42 24 4 5 7 6 0 4 2Structured play 27 19 8 6 0 0 2 0 7 3Free play 30 10 9 0 2 4 0 4 4 10Total 103 71 41 10 7 11 8 4 15 15
Participant BTransition 40 32 29 1 10 12 11 1 6 3Structured play 49 24 30 4 6 8 9 4 3 4Free play 31 14 9 1 3 1 3 3 14 4Total 120 70 68 6 19 21 23 8 23 11
Participant CTransition 47 23 21 4 0 1 15 9 4 8Structured play 10 9 4 1 1 0 1 2 5 6Free play 16 17 12 0 0 0 7 2 1 1Total 73 49 37 5 1 1 23 13 10 15
Participant DTransition 16 5 6 2 13 10 4 3 6 2Structured play 21 15 2 1 0 0 5 6 4 5Free play 10 8 4 0 3 4 1 0 3 5Total 47 28 12 3 16 14 10 9 13 12
Note. CU � conversational units; WH � Wh- questions; IV � intraverbals; A � attention condition; T � token condition.
31SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
fantasy play with the anthromorphic toys andfewer social interactions with the adult orpeer present. Thus, the structured play settingappeared to act as a setting event in which theaudience was the self as a listener rather thanthe peer or adult as a listener. Lodhi andGreer (1989) reported similar findings in theirinvestigation of the effects of anthropomor-phic toy play conditions on the age-appropri-ated self-talk of young children during fan-tasy play.
Another unique setting event was also no-ticed in the Free Play setting and this may haveaccounted for some variability in the data forParticipants C and D. The Free Play settingincluded toys that made noise, such as musicaltoys and talking toys. These noisy toys mayhave functioned as competing items for talking.When participants played with these toys, theyappeared less likely to talk, thus the noisy toysappeared to have an abolishing effect on thereinforcement available for verbal operants.
Figure 2. Number of tacts emitted in all three settings combined for Participants A, B, C,and D.
32 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
The data here show that, overall, each of thefour participants emitted a greater number oftacts when social attention was delivered thanwhen token reinforcement was delivered. Be-tween 30% and 51% more tacts occurred whensocial attention was delivered contingently.
It is, however, important to explore why tactscontinued to occur when contingent tokenswere delivered. One possible explanation is thatthere may have been carry-over effects createdby the alternating treatment design. To elimi-nate these effects, in Experiment 2 we used atoken cup as a discriminative stimulus to signala changeover schedule from one condition tothe next. Even with a changeover schedule inplace, carryover effects are common in alternat-ing treatment designs and may be impossible toeliminate completely (Kennedy, 2005). Finally,it is possible that the delivery of tokens, whichrequired a physical movement by the experi-menter, could have had an element of socialattention, albeit nonvocal, but social nonethe-less. In Experiment 2, we adopted a method of“automated” token delivery designed to elimi-nate the adult as much as possible. Zrinzo andGreer (2013) achieved this by using a tokenchute system, which effectively removed theadult from the token delivery process becausethe participant could not see the adult who wasdelivering the tokens through the chute.
The purpose of Experiment 2 was to continueto investigate the emission of tacts under con-ditions of social verbal reinforcement versustoken reinforcement. However, the methodol-ogy was revised and refined based on some ofthe limitations that were identified in Experi-ment 1. First, we selected a single setting inwhich to observe the participants, rather thanthree settings. This was done to achieve a morecontrolled experimental setting so that the ef-fects of certain extraneous variables could befurther eliminated. For example, in Experiment2, we attempted to better control the number ofopportunities to tact by presenting a slideshowof pictures. The setting for Experiment 2 was alaboratory setting designed to eliminate any so-cial contact in the delivery of tokens. We alsosimulated the transition setting from Experi-ment 1 because that was the setting in whichparticipants emitted the greatest numbers oftacts. Also of note, the participants in Experi-ment 2 were all typically developing childrenwith the exception of Participant A who still
carried a diagnosis of language and adjustmentdisorder. Finally, in Experiment 2, we strength-ened the research design with a multi elementdesign that included an alternating treatmentphase followed by repeated sessions under eachof the contingencies.
Experiment 2
Method
The dependent variables remained the samefor Experiment 2 as those used in Experiment 1.
Participants. Six participants were se-lected for Experiment 2. Five female and onemale preschooler participated. Two of the par-ticipants from Experiment 1, Participants A andB, also participated in Experiment 2. Partici-pants C and D from Experiment 1 did not par-ticipate in the second experiment. To be consis-tent and to avoid confusion, we did not assignthe labels C or D in this experiment. Partici-pants G and H were selected from the samepreschool classroom as Participants A and B.Participants J and I were selected from a differ-ent preschool. The preschool that Participants Jand I attended was characterized as a private,nonprofit, community-based, cooperative pre-school. The school was run out of a large grad-uate college located in a major metropolitancity. Parent members of the cooperative pre-school served as the teachers on a rotating cy-cle. In addition, Participants G and H weresisters. Table 1 displays participant information.
Setting and materials. The setting for Par-ticipants A, B, G, and H was a small therapyroom within the preschool that was typicallyused for physical therapy. It had a door, a win-dow looking out on a playground, and a closetdoor containing materials in it that remainedclosed. There were some materials present thatwere used for physical therapy, such as matsand therapy balls; however, these were removedduring experimental sessions. Materials thatwere used during experimental sessions in-cluded a table that measured 58 cm � 118 cmwith a custom-made partition mounted on it.The partition (shown in Figure 3) was con-structed using white foam filled tri-fold displayboard that measured 60.96 cm � 91.44 cm inthe middle panel and 30.48 cm � 91.44 cm ineach of the side panels. It was mounted on thetable using white duct tape. Two holes 5 cm in
33SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
diameter were cut in the foam board in themiddle panel using an Exacto knife. A whiteplastic flexible corrugated pipe (i.e., drain pipe)was inserted through each of the holes to forma chute through which tokens could be deliv-ered. Thus, there were two token chutes. Justbelow the chutes, identical plastic cups weremounted on the table and secured with Velcro.The cups were transparent and had no lids. Thechutes were positioned so that tokens wouldtraverse the chutes and into the participants’respective cups. The token chutes were presentonly when the experimental condition called fortoken reinforcement. During sessions when to-kens were not being delivered, the chutes wereremoved from the tri-fold board and the holeswere sealed with white duct tape. The settingfor Participants I and J was a living room in theprinciple investigator’s apartment. Sessionswere not conducted at their preschool becauseof scheduling conflicts that required their ses-sions to be conducted outside of school hours.The same partition described above was pres-ent.
An Apple iPad was mounted at the top of thetri-fold board, positioned in the center of the
middle panel (see Figure 3). Photos were dis-played in a slideshow format on the iPad usingthe iPhoto software. The photos were prese-lected and organized into five different sets.Each set contained 30 photos. Five categorieswere equally represented in each set. The cate-gories were Animals, Sports, Basic Shapes/Letters/Numbers, Cartoon/Book Characters,and Miscellaneous. These categories were se-lected because they contained items that wereage-appropriate and commonly known amongpreschoolers. The slideshow ran on an auto-matic timer. The timer was set so that eachphoto was displayed for 10 s, then it automati-cally changed to a new picture within the set.The slideshow function “shuffle” was turned onso that the order of pictures was randomized. Inaddition, if the set was viewed again within theexperiment, a new order of photos was dis-played. There were five sets and these werecounterbalanced across the two experimentalconditions (attention and token) so that each setwas displayed an equal number of times in eachexperimental condition.
Two child-size chairs were positioned infront of the table/partition. A white line waspositioned on the floor using duct tape in be-tween the two chairs and the table as a visualboundary for the participants to indicate thatthey should remain behind the line. A videocamera was placed at one end of the room,positioned so that both participants could becaptured in the frame. Tokens used in sessionsunder token conditions were stored behind thepartition, outside of the participants’ view
Design. The design was a multielement de-sign that included an initial alternating treat-ment phase, followed by reversal phases. Thisentailed rapid alternation between token rein-forcement and social attention reinforcement,followed by repeated exposure to both condi-tions in an A-B-A (Participants A and B) or B-A(Participants E, F, G, and H) sequence. In themultielement assessment, order of sessions wasa simple alternation if only one session per daytook place. For instance, on Day 1 the attentioncondition was implemented and on Day 2 thetoken condition was implemented. If, however,two sessions took place in one day, the order ofsessions was counterbalanced to minimize anysequence effects. Thus, on day one the sequencewas Attention, and then the Token condition; on
Figure 3. Apparatus used in Experiment 2.
34 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
day two the sequence was Token, then Atten-tion.
Dependent variable. The dependent vari-ables were the numbers of tacts emitted and thenumbers of peer-to-peer conversational units.As in Experiment 1, mands, intraverbals, con-versational units, and Wh- questions were alsomeasured. Frequency measures were used forall variables of interest.
Independent variable and procedure.Similar to the first experiment, two participantswere present in each session. However, in thesecond experiment, pairs were held constant.Thus, the pairs throughout the study were: Awith B, G with H, and I with J. Participantswere brought to the experimental setting andasked to sit in the chairs. The photo slideshowdescribed in the materials section was in prog-ress when they arrived. To communicate to theparticipants which experimental conditionwould be in effect, the experimenter deliveredan antecedent such as, “Look, what’s that?”while pointing to the picture on the screen. Ifthe participants responded by emitting a correcttact for the picture, either a token was deliveredthrough the token chute (by a second experi-menter located behind the partition not ob-served by the participants) or attention wasgiven by the primary experimenter, dependingon the experimental condition in effect for thatsession. This was done twice at the beginning ofeach session to ensure that the contingency wasestablished for each participant. Next, the ex-perimenter said, “Please wait in your chairswhile I get something ready for our next activ-ity,” and then the experimenter sat approxi-mately 1 m away from the participants andpretended to be occupied with her work. Thismarked the beginning of the 5-min “waiting”period that served as the experimental session.The participants were given no further instruc-tions and they were not asked to talk about thepictures in the slide show.
In the attention condition, any tacts emittedby a participant while he or she “waited” werefollowed by attention from the experimenter inthe form of vocal approvals (e.g., “That’s right!It is a leopard” or “I see Mickey Mouse too!”),nonvocal approvals, such as smiles, laughter, orthumbs up, and physical attention, such as tick-les and light pats.
In the token condition, any tacts emitted by aparticipant while he or she “waited” were fol-
lowed by the delivery of a token through thechute into his or her cup. For this condition, asecond experimenter sat behind the partition outof the participants’ view, but such that the sec-ond experimenter could see the primary exper-imenter, who sat 1 m from the participants. Theprimary experimenter used hand gestures to sig-nal to the second experimenter when to delivera token and into which chute. Prior to eachsession, the two experimenters established thehand signals to be used. The signals were asingle extended index finger to signal a tokenfor chute #1 (left side) and two extended fingersto signal a token for chute #2 (right side). Thus,if the participant sitting in the chair closest tochute #1 emitted a tact, the primary experi-menter discretely held up one finger and thesecond experimenter then deposited a token intochute #1. If both participants emitted a tact atthe same time, the primary experimenter ges-tured with one finger then two fingers, then thesecond experimenter deposited a token intoboth chutes. No vocal responses were given inthis condition.
In token sessions, the participants were giventhe opportunity to exchange their tokens for asmall snack (e.g., a Skittle, M&M, Starburst, orHershey chocolate kiss). Across both token andattention conditions, participants were asked toparticipate in a brief activity after the 5 min of“waiting” (in actuality, the “waiting” periodserved as the experimental session). The activ-ity was not included in any of the experimentalmeasures. It simply served as a “reason” for theparticipants to have to wait in the experimentalsetting. Examples of activities included moldingshapes using play-doh, drawing on dry-eraseboards, building with blocks, and completing apuzzle. After 2 to 3 min of the activity, theexperimenter ended the activity and the partic-ipants left the setting.
Interobserver agreement. A second ob-server independently viewed the video re-corded sessions and recorded data on the cus-tom-made data sheet. Transcripts of what wassaid were compared across observers. Inter-observer agreement (IOA) was calculated bydividing the numbers of agreements by thenumber of agreements plus disagreements,then multiplying by 100%. For Participant A,IOA was conducted in 30% of sessions andmean agreement was 91% (range: 82%–100%). For Participant B, IOA was conducted
35SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
in 30% of sessions and mean agreement was91% (range: 83%�97%). For Participant E,IOA was conducted in 37% of sessionsand mean agreement was 91% (range:86%�94%). For Participant F, IOA was con-ducted in 37% of sessions and mean agree-ment was 95% (range: 84%�100%). For Par-ticipant G, IOA was conducted in 31% ofsessions and mean agreement was 96% (range:89%�100%). For Participant H, IOA was con-
ducted in 31% of sessions and mean agreementwas 89% (range: 83%�94%).
Experiment 2 Results
Figure 4 shows total sum of the frequency oftacts emitted across all sessions by conditions.The number of tacts emitted was significantlyhigher under attention reinforcement than tokenreinforcement for all six participants.
Figure 4. The total sum of tacts, conversational units, intraverbals, wh- questions, andmands emitted across all sessions, when attention was delivered contingently (black bar) andwhen tokens were delivered contingently (grey bar), for Participants A, B, E, F, G, and H.
36 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Figure 5 displays line graphs for ParticipantsA and B, depicting the numbers of tacts emittedsession by session. Participant A’s graph showsan immediate differentiation in her tacts acrossthe rapidly alternating conditions. Tacts wereconsistently higher in the attention condition. Inthe second phase, attention reinforcement wasdelivered in repeated sessions. Participant Ashowed very stable responding, and a reversalwas implemented after three sessions. In thethird phase, only token reinforcement was de-livered. There was an immediate drop in thelevel of responding and after five sessions an-other reversal back to Attention was imple-mented. There was an immediate increase in
level from the last token session, however, thelevel of responding did not immediately returnto the response level observed in the initialphase.
Participant B’s graph shows that it took 3sessions before differentiation was observed.However, there was a clear separation in thedata paths for the remainder of the first phase ofalternating treatments, and tacts were clearlyoccurring more frequently under attention rein-forcement. In the second phase, only attentionreinforcement was delivered for repeated ses-sions. After three sessions, there was stability inher tacts, and a reversal was implemented. Inthe third phase, only token reinforcement was
0
10
20
30
40
50
60
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Session
Alternating Contingencies Attention Tokens Attention
Participant A
noi
ss
eS
nim
5 a
ni d
ettim
E s t
ca
T fo
re
bm
uN
Participant B
Figure 5. Numbers of tacts emitted under attention reinforcement (closed circle) and tokenreinforcement (open circle) for Participants A and B.
37SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
delivered for repeated sessions. The datashowed a descending trend and after five ses-sions, another reversal was implemented backto attention. The data in this phase show a sharpascending trend and a return to initial levels ofresponding.
Figure 6 shows graphs for Participants E andF. Participant E’s graph shows an immediatedifferentiation of responding between the twoconditions, with a higher number of tacts con-sistently occurring in the attention condition. Inthe second phase, only token reinforcement was
delivered and this resulted in a low, descendingtrend in the data. After four sessions and stableresponding, a reversal to attention reinforce-ment was implemented and there was an imme-diate increase in the level of responding.
Participant F had some initial variability inher responding; however, the two data pathsdid show some separation and tacts occurredmore frequently in the attention condition. Inthe second phase, only tokens were deliveredand Participant F’s tacts dropped to zero bythe third repeated session. In a reversal to
0
10
20
30
40
50
60
70
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Attention
Token
Alternating Contingencies Tokens Attention
0
10
20
30
40
50
60
70
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Session
Alternating Contingencies Token Attention
noi
ss
eS
nim
5 a
ni d
ettim
E st
ca
T fo
re
bm
uN
Participant E
Participant F
Figure 6. Numbers of tacts emitted under attention reinforcement (closed circle) and tokenreinforcement (open circle) for Participants E and F.
38 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
attention, Participant F showed an immediateincrease in the number of tacts emitted andthis remained stable across three consecutivesessions.
Figure 7 shows graphs for Participants G andH. Participant G showed some initial variabilityin her responding, however, differentiation ofthe reinforcement conditions was clearlyachieved after the first four sessions. In thesecond phase, tacts were maintained under at-tention reinforcement for four repeated ses-sions. In a reversal to token reinforcement, thenumber of tacts emitted showed an immediatedrop in level and declined over two consecutivesessions.
Participant H showed an immediate and cleardifferentiation in responding across the twoconditions. In the second phase, tacts weremaintained under attention reinforcement forfour repeated sessions. In a reversal to tokenreinforcement, the number of tacts emittedshowed an immediate drop in level and thendeclined over two consecutive sessions. For allsix participants a functional relation was dem-onstrated between contingent adult attentionand the emission of tacts.
The results of Experiment 2 revealed a func-tional relation between contingent adult atten-tion and the emission of tacts in typically de-veloping preschoolers. These data support the
0
5
10
15
20
25
30
35
40
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Attention
Token
Alternating Contingencies Attention Tokens
0
5
10
15
20
25
30
35
40
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Session
Alternating Contingencies Attention Tokens
sn
ois
se
S ni
m 5
a ni
detti
mE
stc
aT f
o r
eb
mu
N
Participant G
Participant H
Figure 7. Numbers of tacts emitted under attention reinforcement (closed circle) and tokenreinforcement (open circle) for Participants G and H.
39SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
theory that tacts are maintained by a specifictype of generalized conditioned reinforcement,social attention, and not by other nonspecifictypes (i.e., tokens).
General Discussion
We suggest that these data support the theorythat the advancement of the tact repertoire thatis basic to the initiation of verbal episodes orconversational units requires that the reinforce-ment be social (Donley & Greer, 1993; Lodhi &Greer, 1989). If the reinforcement operationsbeing applied to the instruction of tacts arefaulty, this will negatively impact the acquisi-tion of tact repertoires and, in turn spontaneousverbal episodes or conversational units by chil-dren (Schmelzkopf, 2010).
The results provide empirical support for thetheory that tacts function to recruit social atten-tion. In the current investigation, social rein-forcers were more effective than tokens in thereinforcement of tacts, suggesting that a—orthe—major function of tacts is to recruit socialreinforcement, such as attention or approval.When Skinner (1957) first defined the tact, heemphasized the importance of discriminativestimuli for the tact: “The tact emerges as themost important of verbal operants because ofthe unique control exerted by the prior stimu-lus” (p. 83). Skinner (1957) called this “theessence of the tact” (p. 82). However, the cur-rent findings suggest that the consequence rela-tion is equally important and contributes greatlyto the “essence” of the tact. Since Skinner’s(1957) early treatment of the tact, several re-searchers have suggested tacts are sociallylearned behavior (Delgado & Oblak, 2007;Greer & Du, 2010; Pistoljevic, 2008; Pistoljevic& Greer, 2006; Schauffler & Greer, 2006;Schmelzkopf, 2010). Alternately, there may bedifferent types of tacts where the differentiationbetween the types of tacts might be a function ofthe particular type of conditioned social rein-forcers as suggested by Greer and Du (2015).
The current findings have provided evidencethat preschoolers’ tacts are strengthened whenthey result in adult attention as compared withother types of nonsocial reinforcers. In Skin-ner’s (1953) treatment of social reinforcers, hesuggested that signs of approval and disap-proval become generalized conditioned rein-forcers and punishers when they are paired with
a variety of primary reinforcers and punishersduring early child development. These earlysocial experiences in conjunction with biologi-cally important events (e.g., food, warmth,physical contact with caregivers) would because for the strength and power of social rein-forcers later in life. Moreover, research points toa social learning role for the establishment ofnew reinforcers (Greer & Singer-Dudek, 2008;Singer-Dudek, Oblak, & Greer, 2011). Evi-dence that children with language delays as oldas three and four years of age may not bereinforced by attention and praise, but that it canbe established as reinforcers through observa-tional social learning conditions, was reportedby Greer, Singer-Dudek, Longano, and Zrinzo(2008) and replicated by Schmelzkopf (2010).The current findings contribute to the literatureby adding evidence for the strength of socialreinforcers specific to select types of verbalbehavior in young children.
The effectiveness of the social reinforcementdelivered in these experiments may also dependon the nonvocal properties, such as eye contact,facial expressions (e.g., smiles) and physicalcontact (e.g., pat on the back, high five). Otherresearchers have empirically demonstrated thiseffect. For example, Kazdin and Klock (1973)found that the use of smiles and physical contactenhanced the effectiveness of verbal approval.Park, Pereira-Delgado, Choi, and Greer (2008)found that playful physical contact enhancedthe reinforcement delivered during academicinstruction. One of the challenges of using so-cial stimuli as reinforcers is that they are diffi-cult to describe and quantify in the way thatother reinforcers, such as food or tokens, can be.Social reinforcers come in a variety of formsand are delivered in a variety of ways and not allthese forms of social reinforcement are equiva-lent. Consistent with this viewpoint, the type ofsocial reinforcement delivered in the currentstudies would be difficult to quantify and itseems that there may be a need to investigatethis further to enhance our understanding ofsocial reinforcement and its effects on verbalbehavior.
The participants in Experiment 1 representeda relatively wide range of levels of verbal ca-pabilities. Participant B was the only participantwho did not have any diagnosis or classifica-tion. Participant A did have a diagnosis of “ex-pressive language disorder” and “adjustment
40 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
disorder”; however, her IQ was tested in theaverage range and she had similar verbal capa-bilities to Participant B. Participants C and Dwere both selected from a different classroomthan the rest of the participants. It was a nonin-tegrated classroom setting; therefore, all of thestudents in their classroom carried a classifica-tion as “preschooler with a disability” and thelevels of verbal capabilities demonstrated byParticipant C and D were overall lower thanthose of the rest of the participants.
These differences in levels of verbal capabil-ity became apparent in the results from Exper-iment 1. Participants C and D emitted relativelyfew tacts overall as compared with the otherparticipants. Participants C and D emitted atotal of 73 and 47 tacts, respectively, whereasParticipants A and B emitted between 103 and120 tacts, respectively. In addition, their tactswere noticeably less varied and less “sophisti-cated” than the tacts emitted by the rest of theparticipants. For example, Participant C emittedmany of the same tacts across sessions, such as,“tickle,” “apple tree,” “a car,” and “a dinosaur.”Similarly, Participant D repeated tacts withinand across sessions, such as “a baby,” “babylion,” “a dirty baby.” As a comparison, some ofthe tacts emitted by the other participants weremore varied in terms of the range of words usedand their sophistication in terms of use of auto-clitics. For example, Participant B said, “It wasraining when I came here . . . When I came outof my Mommy’s car I felt a drip on my head . . .Who drived you here? Mommy or the bus?”
Although the results obtained from Partici-pants C and D were consistent with the rest ofthe participants in Experiment 1 in that tacts didoccur more frequently under social reinforce-ment than token reinforcement, the overall ef-fect was smaller for these two participants. Thestrength of social reinforcers was less powerfulfor these two participants and one possiblecause is their lower levels of verbal capabilities,as outlined in Greer and Speckman (2009).
In Experiment 2, all six of the participantswere typically developing, with the exception ofParticipant A; however, her verbal capabilitiesmatched those of the other participants. In con-trast to Experiment 1, the results in Experiment2 were very consistent across participants interms of the frequency of tacts emitted and thestrength of the functional relation between tactsand social reinforcement. Also, Experiment 2
eliminated the role of the adults in the deliveryof tokens because tokens were deliveredthrough a chute so that the participants couldnot see where they came from. One exceptionwas Participant F, who emitted fewer tacts, rel-ative to the rest of the participants. However,she was paired with Participant E, who emittedthe highest number of tacts across all partici-pants, so the relatively low emission of tacts byParticipant F may have been a direct effect ofParticipant E’s frequency of tacts. Individualparticipant characteristics and their impact onthe peer with whom they were paired were notinvestigated in this study. However the datafrom Participant E and Participant F are com-pelling and it would be interesting to investigatethe effects of participant characteristics such asage, gender, and the characteristics of the peerwith whom they are paired, on the outcomes ofthis type of study.
The inclusion of typically developing chil-dren in Experiment 2 was important for gaininga better understanding of how social reinforcersimpact language in the absence of any knownlanguage disorders. With a clearer picture ofhow language functions for the typically devel-oping children included in this study, this infor-mation can be used to investigate the samephenomenon in children with language disor-ders. With Participants C and D in mind, willsuch investigations reveal that for children withlanguage disorders, social reinforcers do notfunction to reinforce tacts, either to the samedegree as their typically developing peers, ornot at all? This may be an important futuredirection in this line of research.
In Experiment 2, participants were shownpictures on an iPad as a measure to provideopportunities to tact. The pictures were orga-nized into five different sets of stimuli and ex-posure to each set was counterbalanced so thatthe participants each set in both experimentalconditions an equal number of times. However,the repeated exposure to the same stimuli mayhave resulted in an overall decrease in partici-pant levels of responding over time. Skinner(1957) discussed the “novelty of the occasion”in relation to tacts: “A given object does notremain the inevitable occasion for the reinforce-ment of an appropriate response, and the prob-ability of response therefore comes to vary withthe occasion” (p. 85). If more novel stimuli hadbeen present during experimental sessions, tacts
41SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
may have been maintained at a higher level.However, the repeated exposure was an impor-tant control measure. Stimulus sets were coun-terbalanced across experimental conditions andthis controlled for any variability across condi-tions due to the experimental stimuli.
The current findings have implications forfuture research. As noted, there was a weakerrelation for some of the participants with thelowest levels of verbal behavior, relative to theirparticipant counterparts. This suggests thatthere is a need to extend this research to otherchildren who demonstrate lower levels of verbalbehavior, such as children with autism, forwhom the relation between tacts and social re-inforcers may be weak or not present at all. Oneof the best predictors of outcomes for childrenwith autism is the development of “spontaneouslanguage” before age 6 (Szatmari et al., 2003).The dilemma with this is that social reinforcersare often ineffective for children with autism(Greer et al., 2008), but they are necessary forthe acquisition of critical speaker repertoires.Thus, acquiring the social function of the tactappears to be an important verbal developmen-tal cusp. Acquiring and emitting language underthis reinforcement control is probably necessaryfor the acquisition of more advanced speakercapabilities, such as a fluent tact repertoire andNaming. Future research should further investi-gate social reinforcement as a critical verbaldevelopmental cusp.
References
American Psychiatric Association. (2013). Diagnos-tic and statistical manual of mental disorders (5thed.). Arlington, VA: Author.
Ayllon, T., & Azrin, N. H. (1968). The token econ-omy: A motivational system for therapy and reha-bilitation. New York, NY: Appleton-Century-Crofts.
Ayllon, T., & Michael, J. (1959). The psychiatricnurse as a behavioral engineer. Journal of theExperimental Analysis of Behavior, 2, 323–334.http://dx.doi.org/10.1901/jeab.1959.2-323
Barnes-Holmes, D., Barnes-Holmes, Y., & Cullinan,V. (2000). Relational frame theory and Skinner’sVerbal Behavior: A possible synthesis. The Behav-ior Analyst, 23, 69–84.
Bloom, P. (2002). How children learn the meaning ofwords. Cambridge, MA: MIT Press.
Braam, S. J., & Sundberg, M. L. (1991). The effectsof specific versus nonspecific reinforcement on
verbal behavior. Analysis of Verbal Behavior, 9,19–28.
Carroll, R. J., & Hesse, B. E. (1987). The effects ofalternating mand and tact training on the acqui-sition of tacts. Analysis of Verbal Behavior, 5,55– 65.
Catania, A. C. (2007). Learning (4th ed.). Cornwall-on-Hudson, NY: Sloan Publishing.
Cooper, J. O., Heron, T. E., & Heward, W. L. (2007).Applied behavior analysis (2nd ed.). Upper SaddleRiver, NJ: Pearson Education Inc.
Delgado, J. A. P., & Oblak, M. (2007). The effects ofdaily intensive tact instruction on the emission ofpure mands and tacts in non-instructional settingsby three preschool children with developmentaldelays. Journal of Early and Intensive BehaviorIntervention, 4, 392– 411. http://dx.doi.org/10.1037/h0100381
Donley, C. R., & Greer, R. D. (1993). Setting eventscontrolling social verbal exchanges between stu-dents with developmental delays. Journal of Be-havioral Education, 3, 387–401. http://dx.doi.org/10.1007/BF00961542
Greer, R. D., & Du, L. (2010). Generic instructionversus intensive tact instruction and the emissionof spontaneous speech. The Journal of Speech andLanguage Pathology – Applied Behavior Analysis,5, 1–19. http://dx.doi.org/10.1037/h0100261
Greer, R. D., & Du, L. (2015). Identification andestablishment of reinforcers that make the devel-opment of complex social language possible. In-ternational Journal of Behavior Analysis and Au-tism Disorders. Advance online publication.Retrieved from http://www.ijobasd.org/index.php/IJOBAS/article/view/24
Greer, R. D., & Ross, D. E. (2004). Verbal behavioranalysis: A program of research in the inductionand expansion of complex verbal behavior. Jour-nal of Early and Intensive Behavior Intervention,1, 141–165. http://dx.doi.org/10.1037/h0100286
Greer, R. D., & Ross, D. E. (2008). Verbal behavioranalysis: Inducing and expanding complex com-munication in children with severe language de-lays. Boston, MA: Allyn & Bacon.
Greer, R. D., & Singer-Dudek, J. (2008). The emer-gence of conditioned reinforcement from observa-tion. Journal of the Experimental Analysis of Be-havior, 89, 15–29. http://dx.doi.org/10.1901/jeab.2008.89-15
Greer, R. D., Singer-Dudek, J., Longano, J., &Zrinzo, M. (2008). The emergence of praise asconditioned reinforcement as a function of obser-vation in preschool and school age children. Re-vista Mexicana de Psicología, 25, 5–26.
Greer, R. D., & Speckman, J. (2009). The integrationof speaker and listener responses: A theory ofverbal development. The Psychological Record,59, 449–488.
42 EBY AND GREER
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Hart, B. M., & Risley, T. R. (1995). Meaningfuldifferences in the everyday life of America’s chil-dren. Baltimore, MD: Paul Brookes.
Hoff, E., & Shatz, M. (Eds.). (2009). Blackwell hand-book of language development. Malden, MA: Wi-ley-Blackwell.
Kazdin, A. E. (1984). Behavior modification in ap-plied settings (3rd ed.). Homewood, IL: DorseyPress.
Kazdin, A. E., & Klock, J. (1973). The effect ofnonverbal teacher approval on student attentivebehavior. Journal of Applied Behavior Analysis, 6,643– 654. http://dx.doi.org/10.1901/jaba.1973.6-643
Kennedy, C. H. (2005). Single-case designs for edu-cational research. Boston, MA: Pearson Educa-tion, Inc.
Lodhi, S., & Greer, R. D. (1989). The speaker aslistener. Journal of the Experimental Analysis ofBehavior, 51, 353–359. http://dx.doi.org/10.1901/jeab.1989.51-353
Moore, J. (2008). Conceptual foundations of radicalbehaviorism. Cornwall-on-Hudson, NY: SloanPublishing, LLC.
Park, H., Pereira-Delgado, J., Choi, J., & Greer, R. D.(2008). The effects of playful physical contact asan establishing operation on correct academic re-sponding of three preschool students. Journal ofEarly and Intensive Behavior Intervention, 5, 90–105. http://dx.doi.org/10.1037/h0100412
Partington, J. W., & Bailey, J. S. (1993). Teachingintraverbal behavior to preschool children. Analy-sis of Verbal Behavior, 11, 9–18.
Pelaez, M., Virues-Ortega, J., & Gewirtz, J. L.(2011a). Contingent and noncontingent reinforce-ment with maternal vocal imitation and motheresespeech: Effects on infant vocalizations. EuropeanJournal of Behavior Analysis, 12, 277–287.
Pelaez, M., Virues-Ortega, J., & Gewirtz, J. L.(2011b). Reinforcement of vocalizations throughcontingent vocal imitation. Journal of Applied Be-havior Analysis, 44, 33–40. http://dx.doi.org/10.1901/jaba.2011.44-33
Pistoljevic, N. (2008). The effects of multiple exem-plar instruction and intensive tact instructionalhistories on the acquisition of naming in pre-schoolers (Doctoral dissertation). Available fromProQuest Dissertations and Theses database. (UMINo. 3317598)
Pistoljevic, N., & Greer, R. D. (2006). The effects ofdaily intensive tact instruction on preschool stu-dents’ emission of pure tacts and mands in non-instructional setting. Journal of Early and Inten-sive Behavior Intervention, 3, 103–120. http://dx.doi.org/10.1037/h0100325
Schauffler, G., & Greer, R. D. (2006). The effects ofintensive tact instruction on audience-accuratetacts and conversational units. Journal of Earlyand Intensive Behavior Intervention, 3, 121–134.http://dx.doi.org/10.1037/h0100326
Schmelzkopf, J. (2010). A study of the relation be-tween acquisition of adult approvals as condi-tioned reinforcers and the emission of vocal verbaloperants for preschool students diagnosed withdevelopmental disabilities (Doctoral dissertation).Available from ProQuest Dissertations and Thesesdatabase. (UMI No. 3400577)
Singer-Dudek, J., Oblak, M., & Greer, R. D. (2011).Establishing books as conditioned reinforcers forpreschool children as a function of an observa-tional intervention. Journal of Applied BehaviorAnalysis, 44, 421–434. http://dx.doi.org/10.1901/jaba.2011.44-421
Skinner, B. F. (1953). Science and human behavior.New York, NY: The Macmillan Company.
Skinner, B. F. (1957). Verbal behavior. Cambridge,MA: Prentice Hall, Inc. http://dx.doi.org/10.1037/11256-000
Stafford, M. W., Sundberg, M. L., & Braam, S. J.(1988). A preliminary investigation of the conse-quences that define the mand and the tact. Analysisof Verbal Behavior, 6, 61–71.
Sundberg, M. L., Juan, B. S., Dawdy, M., &Argüelles, M. (1990). The acquisition of tacts,mands, and intraverbals by individuals with trau-matic brain injury. Analysis of Verbal Behavior, 8,83–99.
Szatmari, P., Bryson, S. E., Boyle, M. H., Streiner,D. L., & Duku, E. (2003). Predictors of outcomeamong high functioning children with autism andAsperger syndrome. Journal of Child Psychologyand Psychiatry, 44, 520–528. http://dx.doi.org/10.1111/1469-7610.00141
Tomasello, T., & Bates, E. (2001). Language devel-opment: Essential readings. Malden, MA: Black-well Publishing.
Williams, G., Carnerero, J. J., & Pérez-González,L. A. (2006). Generalization of tacting actions inchildren with autism. Journal of Applied BehaviorAnalysis, 39, 233–237. http://dx.doi.org/10.1901/jaba.2006.175-04
Zrinzo, M., & Greer, R. D. (2013). Establishment andmaintenance of socially learned conditioned rein-forcement in young children: Elimination of therole of the adult. The Psychological Record, 63,43–62. http://dx.doi.org/10.11133/j.tpr.2013.63.1.004
Received September 15, 2015Revision received October 11, 2016
Accepted October 12, 2016 �
43SOCIAL REINFORCEMENT OF TACTS
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
