establishing generalizations

A comparison of two teaching procedures to establish generalized intraverbal-tacting in children with autism

Francesca degli Espinosa Queen’s University, United Kingdom and ABA Clinic, United Kingdom

Kate Wolff and Sophie Hewett Queen’s University, United Kingdom

Previous research has investigated generalized intraverbal-tacting by teaching children with autism to respond using autoclitic frames. The present study compared the effectiveness and efficiency of a Frame and a No Frame procedure across counterbalanced stimulus sets with 4 children with autism. In the Frame condition, children were taught to respond using autoclitic frames (e.g., “Shape square,” “Number two,” “Color green,” “It’s mummy,” “S/he is drinking”) corresponding to the verbal antecedent (“What shape?”, “What number?”, “What color?”, “Who is it?”, “What is s/he doing?”). In the No Frame condition, intraverbal-tacting was established without the autoclitic frame. Irrespective of stimuli employed, 2 children acquired intraverbal-tacting only in the Frame condition. The other 2 children acquired intraverbal- tacting in both conditions, with the Frame procedure requiring fewer teaching trials for 1 child and producing greater generalization for the other. Implications for clinical practice and the role of additive intraverbal stimulus control of autoclitic frames are discussed. Key words: autoclitic frames, intraverbal-tacting, multiple verbal control, question

discrimination

Responding to questions is a fundamental skill that emerges early in children. Before the age of two, neurotypically developing toddlers show differential accurate responding to yes and no, and what and where questions (Goodwin et al., 2012; Seidl et al., 2003) in the presence of visual stimuli. By the age of three and a half, they comprehend a variety of questions framed with who, what, where, when, why and how in the context of social play and shared book- reading (Ervin-Tripp, 1970; Rhyner, 2007; Rowe et al., 2017). A consistent and growing body of evidence suggests that parental use of questions, particularly Wh-questions, is posi- tively associated with children’s vocabulary

acquisition (Goodwin et al., 2012; Hoff- Ginsberg, 1985, 1986; Rowe et al., 2017). Take, for example, a parent and 2-year-old

child looking at a book on trains. When the parent says, “Look at the train” the child ori- ents to the picture, points to it, says, “Train” and then looks at the parent for social approval and to continue the interaction. The parent says, “That’s right, it’s a train, what is it?”, to which the child replies, “Train.” The parent then says, “Choo choo,” to which the child laughs while repeating “Choo choo.” The par- ent says, “Yes, the train goes choo. What sound does the train make?”, to which the child repeats, “Choo choo.” Similarly, the parent may say, “Look, the train is blue.” While looking at the picture, the child (who by this age can echo short phrases) repeats “Train is blue,” which the parent follows with the ques- tion, “Yes! What color is it?”, to which the child answers, “Blue.” Through these social interactions, occurring multiple times daily

We have no conflicts of interest relevant to this article to disclose.

Address correspondence to: Francesca degli Espinosa, ABA Clinic, 40a Burgess Road, Southampton, Hamp- shire, SO16 7AH, UK; Email: [email protected]

doi: 10.1002/jaba.869

Journal of Applied Behavior Analysis 2021, 54, 1468–1487 NUMBER 4 (FALL)

© 2021 Society for the Experimental Analysis of Behavior (SEAB).

1468

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

across many contexts, the parent not only teaches new words and concepts (e.g., identity names, colors, object and animal sounds), but also teaches the child to produce these new words and concepts in response to the relevant corresponding question. In behavior-analytic terms, through frequent

social interactions involving verbal and nonver- bal stimuli across multiple exemplars, parents not only establish novel topographies as tacts but also bring the emission of these responses under the multiple control of both verbal (i.e., questions and comments) and nonverbal stimuli (e.g., colored trains). Bondy et al. (2004) use the term intraverbal-tacting to describe responding differentially to questions regarding some dimension of objects and peo- ple (e.g., “Who’s this?”, “What is he doing?”). Intraverbal-tacting is an example of multiply controlled verbal behavior that involves the convergent control of verbal stimuli and some aspect of the nonverbal physical environment (Michael et al., 2011). In typical development, intraverbal-tacting

develops without specialized instruction during day-to-day social interactions. However, this skill can be delayed (Goodwin et al., 2012) or impaired (Howlin, 1982) in children with autism spectrum disorders (ASD). A problem commonly encountered in clinical practice is the child who, despite having been taught the component tacts (e.g., the names of items and colors), when shown a colored object, such as a blue train, and asked, “What color?” responds “Train.” Although children with ASD show clear deficits in responding differentially to questions regarding visual stimuli, procedures to establish intraverbal-tacting have not been extensively investigated. Applied research has mainly focused on establishing specific intraverbal responses with respect to the fea- ture, function, and class of items (DeSouza et al., 2019; Ingvarsson et al., 2016; Jahr, 2001) and to two-component questions (e.g., “What’s an animal that’s red?”; Aguirre

et al., 2019; Kisamore et al., 2016) using a range of strategies, such as echoic to intraverbal or tact to intraverbal transfer procedures, multi- ple exemplar training, blocked trials, and dis- crimination training (see DeSouza et al., 2017; Stauch et al., 2017 for reviews). A number of studies have shown that an echoic Differential Observing Response (DOR) procedure, whereby children echo the critical part of the verbal antecedent prior to the target response, can facilitate differential intraverbal responding (Jahr, 2001; Kisamore et al., 2013, 2016). For example, Kisamore et al. (2016) taught children with ASD to respond to two-component ques- tions such as “What’s an animal that’s red?” or “What’s a yellow vehicle?” by repeating “Ani- mal red” before saying “Parrot” or “Yellow vehicle” before saying “Dump truck”.

While these studies on intraverbal responding focused on teaching specific responses to specific questions, two recent stud- ies (degli Espinosa et al., 2020; Meleshkevich et al., 2020) have focused on establishing gen- eralized discrimination of specific questions (i.e., “What is it?” vs. “What color?”) as tested on novel visual stimuli. In degli Espinosa et al. (2020) two children with ASD were tau- ght to say the verbal frames “It’s a” and “It says” when responding to the verbal anteced- ents “What is it?” and “What does it say?” when shown pictures of animals. They were also taught to say the verbal frames “Color,” “Number,” and “It’s a” when responding to the verbal antecedents “What color?”, “What number?”, and “What is it?” when shown col- ored numbers and objects. A series of steps were conducted to teach the

participants to use these verbal frames. The first step involved teaching echoics of the verbal frames and the corresponding variable terms (e.g., “Say color red.” “Say color blue,” “Say number four,” “Say number two”) in the absence of any visual stimuli. The second step employed simultaneous simple-discrimination procedures to teach tacts of component visual

1469Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

stimuli using a verbal frame (e.g., “It’s a car,” “Color red”) when presented with the anteced- ent verbal stimuli (“What it is?”, “What color?”). The same question was asked across multiple exemplars of the same stimulus class (e.g., “What is it?” for objects, “What color?” for colored swatches). The third step involved the consecutive presentation of both verbal antecedents (e.g., “What is it?”, “What color?”) in the presence of the compound stimuli (e.g., a red car, a blue spoon, a green bag, a yel- low hat) in a conditional-only arrangement (Grow et al., 2011). Children were prompted to use the frame when errors occurred, but the actual intraverbal-tact was not directly prompted. Following mastery of this final step, near errorless intraverbal-tacting and general- ized use of the frame to novel exemplars within and across stimulus classes was demonstrated for both children. Meleshkevich et al. (2020) established gener-

alized intraverbal-tacting in response to the ver- bal antecedents “What shape?” and “What number?” and pictures of shapes with numbers written inside them, and in response to “What is it?” and “What color?” and pictures of col- ored objects with three young children with ASD. The first two training steps from the degli Espinosa et al. (2020) procedure (echoic, simultaneous simple discrimination) were rep- laced with fully prompting the use of the target frame and the intraverbal tact through a conditional-only method of discrimination training. For two children, the conditional-only method was not immediately effective, and similar to degli Espinosa et al., brief exposure to a simultaneous simple-discrimination proce- dure to teach tacts with the verbal frame (e.g., “color yellow” in response to colored swatches) was implemented before a return to the conditional-only procedure. This proce- dural change produced accurate intraverbal- tacting. An additional variation was presenting one visual stimulus at a time, in isolation, and rotating the verbal antecedents, rather than

presenting visual stimuli in an array and rotating the verbal antecedents. Despite these procedural differences, similar results were obtained: improved levels of intraverbal-tacting on the teaching stimuli and within and across stimulus class generalization. In addition, chil- dren responded to novel visual stimuli by using the learned verbal frames. The observed gener- alized intraverbal-tacting to novel untaught compound stimuli may have resulted from the additive, multiple, and generalizable verbal con- trol exerted by teaching children to echo the conditional critical term within an autoclitic frame (e.g., “It’s a [cat], “It says [meow], “Color [brown]”) in response to the verbal antecedent (degli Espinosa et al., 2020; Meleshkevich et al., 2020). Autoclitic frames are verbal frames in which

the terms are organized in orderly sequences of intraverbally related fixed and variable terms (Palmer, 2007, 2016; Skinner, 1957). In the presence of a compound visual stimulus (e.g., a brown cat), the verbal antecedent, for example “What color?”, sets the occasion for the emis- sion of an echoic (the first term of an autoclitic frame) that, in turn, exerts additional intraverbal control over a class of variable responses (i.e., color names), of which the spe- cific member (i.e., the color tact) is the sample at hand. The resultant intraverbal-tacting then reflects a summation of control from the vari- ous verbal sources: the verbal antecedent, the echoic, and the intraverbal relation between frame elements. However, because the frame procedure was

the sole procedure used in the only two publi- shed studies on intraverbal-tacting (degli Espinosa et al., 2020; Meleshkevich et al., 2020), it is not possible to conclusively determine whether the frame was necessary for producing generalized responding, or whether more tradi- tional conditional discrimination training proce- dures would be sufficient to establish conditional control of the verbal antecedent over the relevant discriminative nonverbal property, irrespective of

Francesca degli Espinosa et al.1470

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

the verbal frame. Therefore, the present study sought to delineate the effect of autoclitic frames in generalized intraverbal-tacting by comparing frame procedures such as those previously described (degli Espinosa et al., 2020; Meleshkevich et al., 2020) to otherwise similar conditional-discrimination procedures without verbal frame training.

Method

Participants and Preassessment Four children from the UK participated in

the study: Adam (15 years), Alice (8 years), Jake (6 years), and Fred (5 years). All children had received a diagnosis of ASD from a multi- disciplinary team, could produce at least two- word echoics, could mand with spoken words for favorite items and activities, and possessed a repertoire of at least 200 spoken tacts that included the names of common nouns, shapes, numbers, colors, familiar people, and actions. They could all sit at a table receiving instruc- tions for at least 10 min and were regularly tau- ght using a token economy system to access tangible reinforcers in their day-to-day ABA sessions. None of the children displayed chal- lenging behavior. All children received ABA- based instruction from ABA instructors at home and/or school. Adam’s score on the VB-MAPP was

116, Alice’s was 112, Jake’s was 96, and Fred’s was 108. All children demonstrated skills within the mand, tact, listener, echoic, and intraverbal milestones of Level 2, and some Level 3 skills within the academic domains. No child could answer different questions in mixed order regarding visual stimuli (T11 milestone). Adam and Jake demonstrated intraverbal- tacting to the verbal antecedents “What color?” and “What is it?” regarding colored objects but did not show accurate discrimination of any other verbal antecedent and the corresponding visual stimulus. Alice and Fred had no intraverbal-tacting skills. Alice had received no

instruction in these. Although Fred had received some intraverbal-tact teaching on col- ored objects in response to the antecedents “What is it? and “What color is it?”, this was terminated 6 weeks before his participation in the study because it had not been effective. A preassessment of the children’s tact reper-

toire was carried out to evaluate whether they could accurately tact the stimuli that would be presented in the experiment. Only stimuli from the component sets were used. Each stimulus card was presented once, in isolation, with the corresponding verbal antecedent. For example, the instructor held up the picture of a familiar person and asked “Who is it?” or a number card and asked “What number?” These verbal antecedents were typically used when children were asked to tact those items. All children cor- rectly tacted all presented stimuli; no stimulus replacement was necessary. There were no programmed consequences for either correct or incorrect responding. To ensure cooperation, maintenance tasks

were interspersed every one, two, or three con- secutive target trials. Tokens were delivered for maintenance tasks only on a variable ratio (VR) 3 token production (three responses, on average, to produce a token) and fixed ratio (FR) 5 exchange production (five tokens to produce exchange opportunities). When Alice, Fred, and Jake’s responding met the exchange production schedule, the exchange period occurred, during which experimental stimuli were removed and a tray or a box containing their personal reinforcers was presented. The children then manded or picked up their pre- ferred item and consumed the reinforcer. If a toy or iPad was chosen, consumption was roughly 30 s. Food was placed in a small Tupperware box, which the children could open. Once the food was consumed, the box was replenished and replaced on the tray. Before each set of trials, Adam was given a pretask choice (Piazza et al., 1996) between favorite items. He manded for the item either

1471Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

spontaneously or when the adult asked “What shall we work for?” The chosen toy was then placed on the table in sight and Adam received it upon meeting the exchange production schedule. This was the method of preference assessment used regularly by Adam’s teachers.

Setting and Materials The procedures for all children were conducted

in a quiet room at home or school. The instruc- tor and the child sat opposite one another at a table. Three sets of stimuli were created for each

stage of both procedures using a matrix to ensure that all components were combined to derive compound stimuli: Shape/Number, Number/ Color, and Agent/Action. Figures 1 and 2 illus- trate the Shape/Number and Agent/Action stimu- lus sets. All visual stimuli were laminated and measured 13.3 cm x 10 cm. Each stimulus class contained two component sets (Component), a teaching compound set (Teaching), and three generalization compound sets (g1, g2, and G). For Shape/Number, the component sets were eight black outlined shapes printed on a white card and eight black numbers (1 to 8) printed on

Figure 1 Teaching and Generalization Stimuli for Shape/Number

Francesca degli Espinosa et al.1472

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

a white card (Number component and Shape component set). Four stimuli from each type were selected to make up sixteen compound teaching stimuli (Teaching set). The remaining stimuli were divided into three generalization groups (g1, g2, and G). The g1 group included one untaught number and one taught shape, g2 one taught number and one untaught shape, and G (Generalization group) combined two untaught components (see Figure 1). The Num- ber/Color stimuli were similarly constructed by combining eight colors (yellow, red, blue, pink, purple, orange, and brown) and 1-8 numbers.

The Agent component stimulus set comprised eight headshot photographs, each depicting a female (F1 to F4) or a male (M1 to M4) familiar to the child which s/he could tact (e.g., mummy, daddy, uncle’s name, sister’s name, female/male instructor’s name). The Action component set included eight photographs of unfamiliar people engaging in the target actions. The remaining sets were created by taking photographs of each of the familiar agents performing each of the target actions. Each group contained two pictures of female and two pictures of male agents (See Figure 2).

Figure 2 Teaching and Generalization Stimuli for Agent/Action

1473Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

Experimental Design and Dependent Variables An adapted alternating-treatments design

(Sindelar et al., 1985) was used in which two intervention conditions (Frame and No Frame procedure) were provided concurrently with respect to two related response classes: Shape/ Number and Agent/Action discriminations for Adam and Jake, and Number/Color and Agent/ Action discriminations for Fred and Alice. Base- line and postintervention generalization measures were identical for all children across all stimulus classes, except that in postintervention generaliza- tion, teaching stimuli employed during the inter- vention phase were excluded. During the intervention phase, the two conditions (Frame and No Frame) were counterbalanced across chil- dren and stimulus classes. Adam received the Frame procedure for Shape/Number and the No Frame procedure for Agent/Action. Jake received the Frame procedure for Agent/Action and the No Frame procedure for Shape/Number. Alice received the Frame procedure for Number/Color and the No Frame procedure for Agent/Action. Fred received the Frame procedure for Agent/ Action and the No Frame procedure for Number/Color. The primary dependent variable was percent-

age of correct unprompted intraverbal tacts, defined as (a) verbal responses that cor- responded to both the verbal antecedent and the visual stimulus, (b) occurring within 2 s of instruction delivery, and (c) irrespective of the use of the autoclitic frame. Additional data were also taken on the independent use of mat- ched autoclitic frames at baseline and in post- intervention generalization. This was defined as the child independently saying, “Shape [shape name]” to the verbal antecedent “What shape?” and “Number [number name]” to “What number?” when presented with the Shape/ Number visual stimuli, “Color [color name]” to “What color?” and “Number [number name]” to “What number?” with the Number/ Color visual stimuli, and “It’s [agent name]”

and “S/he is [actioning] to “Who is it?” and “What is s/he doing?” with the Agent/Action visual stimuli. Use of the autoclitic frame was recorded on each trial in baseline and post- intervention generalization, whether responding on the primary dependent variable (the intraverbal tact) was correct or incorrect in both conditions. Table 1 shows examples of correct and incorrect responses across the three stimu- lus sets (Agent/Action, Shape/Number, Color/ Number).

Procedures Baseline Sessions Baseline procedures and stimulus presenta-

tion were identical to those reported by degli Espinosa et al. (2020). A sequence generator computerized program provided the randomiza- tion for the visual stimuli positions and the order of verbal antecedents to be presented in each trial in each session. Two sessions per day were conducted: one baseline session for each stimulus set consisting of 20 trials (40 trials in total). Each session consisted of targets belong- ing to only one set of compound stimuli (e.g., Agent/Action only or Color/Number only). Trials across stimulus sets were not inter- spersed in each session. The instructor laid out four visual stimuli in a horizontal array. Five trials in which a different verbal antecedent for the four visual stimuli were presented. After five trials, the array was removed and a different array of four visual stimuli was presented and the next five trials commenced. This continued until all 20 trials were conducted and the ses- sion for that stimulus set ended (for Adam, Jake, and Alice). Each of Fred’s baseline ses- sions were conducted in the same manner, with the exception that he received 30 baseline trials per set in the first two sessions, and 20 baseline trials per set in the third session. Thus, the total number of baseline trials (160) was the same for all four children. In each session, both verbal antecedents related to the

Francesca degli Espinosa et al.1474

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

compound stimuli were presented an equal number of times. Figure 3 provides an example of five randomized trials in a four-stimuli array for each stimulus set. After laying out the four visual stimuli, the instructor presented one of the target verbal antecedents while pointing to the visual stimulus, about 1 cm from the top edge of the card. The instructor did not point on the actual card surface to avoid inadvertently pointing to a specific feature of the compound visual stimulus (e.g., the number or the shape, the face of the

person, or the hands performing the action). The child was given 2 s to respond. The response was recorded and the instructor moved to the next visual stimulus and verbal antecedent until all five trials for that array had been completed, at which point a new array of stimuli within the same set of compound stimuli was presented. As in the pre- assessment, programmed consequences for either correct or incorrect responding were not delivered. To ensure children’s cooperation, maintenance tasks were interspersed every one, two, or three

Table 1

Definitions of Correct and Incorrect Responses to Verbal Antecedents (VA) and Visual Stimuli (VS) Across Classes

Agent/Action Response type Who is it? What is she doing?

Correct Matched VA & VS “Mummy” “Drinking” “It’s mummy” “She is drinking”

Incorrect 1 Discrimination error (Unmatched VA & matched VS)

“Drinking” “Mummy”

2 Tact error (Matched VA & unmatched VS)

“Auntie” “It’s auntie” “Eating”

“She is eating” 3 Combined response “Mummy is drinking” “Mummy is drinking”

“Mummy drinking” “Mummy drinking” “Drinking mummy” “Drinking mummy”

4 Anticipatory response Any responding prior to VA

Any responding prior to VA

Shape/Number Response type What shape? What number?

Correct Matched VA & VS “Triangle” “Four” “Shape triangle” “Number four”

Incorrect 1 Discrimination error (Unmatched VA & matched VS)

“Four” “Triangle”

2 Tact error (Matched VA & unmatched VS)

“Square” “Three”

“Shape square” “Number three” 3 Combined response “Four triangle “Four triangle

“Triangle four” “Triangle font” 4 Anticipatory response Any responding prior

to VA Any responding prior to VA

Number/Color Response type What number? What color?

Correct Matched VA & VS “Three” “Red” “Number three” “Color red”

Incorrect 1 Discrimination error (Unmatched VA & matched VS)

“Red” “Three”

2 Tact error (Matched VA & unmatched VS)

“Five” “Green”

“Number five” “Color green” 3 Combined response “Three red” “Red three”

“Red three” “Three red” 4 Anticipatory response Any responding

prior to VA Any responding prior to VA

1475Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

consecutive target trials, as previously described. The baseline trials were never the final reinforced trial and were therefore never immediately followed by reinforcement.

Echoic Session After the last baseline session had been com-

pleted, one session of echoic training was con- ducted for each response that would be targeted in both intervention conditions, to ensure that participants could produce and accurately articu- late all responses. This was particularly relevant for the Frame procedure, as these word combina- tions were novel to all participants. Sixteen echoic trials were conducted (eight for each dis- crimination), once only. The instructor and the child sat opposite each other. The instructor said, “We are going to repeat some words” and then delivered the instruction (e.g., “Say ‘It’s mummy’” or “mummy”) depending on the con- dition to which that stimulus class had been assigned. Echoic trials of Frame targets (e.g., “Its’ a [agent name] and “She is [actioning]”) and No Frame targets were presented in separate and counterbalanced trial sessions, rather than inter- spersed with one another. There were no programmed consequences for a correct echoic; maintenance targets were interspersed, as previ- ously described, to maintain cooperation with the task. All children echoed the relevant verbal stimuli and no additional teaching procedure to shape clear articulation was necessary.

Intervention Sessions Each teaching day was divided into two ses-

sions, with one condition presented in the first session and the other in the second session on the same day. The two conditions were discrimination training with an autoclitic frame requirement (Frame procedure) and discrimination training without an autoclitic frame requirement (No Frame procedure). In both conditions, proce- dures employed a conditional-only (Grow et al., 2011) teaching method, 0-s prompt delay on the initial session only, and a 2-s prompt delay in each

session thereafter. Each condition was randomly assigned to either the first or second session of the day. There were 20 trials in each session (40 trials in total per day). Children received no more than two teaching sessions per day (one for each condi- tion), approximately five days per week. When one procedure led to skill mastery before the alter- native condition, teaching continued for five addi- tional sessions (100 trials) while postintervention generalization commenced for the condition in which mastery criterion had been met. If, after the additional 100 trials, mastery criterion was still not met, rather than continuing to expose the child to an unsuccessful intervention when its counterpart had shown to be effective, the effec- tive procedure was applied to that stimulus set. For Alice, given significant time constraints, the Frame procedure was introduced after 60 trials (three sessions) without mastery in the No Frame condition, rather than 100. Mastery criterion for both conditions was 95% of teaching trials with correct responding for three consecutive sessions. Stimulus presentation and randomization were

the same as during the baseline sessions (see

Figure 3 Examples of a Four-Stimuli Array

Francesca degli Espinosa et al.1476

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

Figure 3). Unlike baseline, however, token rein- forcement for correct responding was provided. The token reinforcement schedules were FR 1 (token production) and FR 5 (exchange produc- tion) for all children for all independent responses. Prompted responses were not reinforced, but were followed each time by an unprompted trial, which, if correct, was reinforced. Frame Procedure. The target correct

response was an intraverbal tact corresponding to the verbal antecedent and visual stimulus, produced within an autoclitic frame in which the fixed part (i.e., “Shape,” “Number,” “Color,” “It’s,” “S/he is”) matched the verbal antecedent (i.e., “What Shape?”, “What num- ber?”, “What color?”, “Who is it?”, “What is s/he doing?”). In the very first session only, the instructor pointed to a card, gave the verbal antecedent (e.g., “Who is?” or “What is she doing?”) and immediately provided a full 0-s prompt of the target response (e.g. “It’s mummy,” “She is drinking”). If the child cor- rectly echoed, a 2-s prompt delay trial was

immediately conducted with that visual stimulus in which the instructor delivered the verbal ante- cedent again and waited 2 s. If a correct response was made within 2 s, a token was delivered and the instructor moved on to the next trial. If a correct response did not occur, the instructor ret- urned to a 0-s prompt delay on the following trial. With Adam, on the Shape/Number trials in the Frame condition, when he did not respond within 2 s, the instructor gave a partial verbal prompt (e.g., “Sh” for “shape” or “Nu” for “number”) to facilitate the emission of the target response. When he responded correctly, the instructor immediately presented the same verbal antecedent and visual stimulus and waited 2 s for an independent response to occur. A token was delivered upon correct responding. The 0-s prompt delay and 2-s prompt delay

trial sequence continued until the child responded to the full verbal prompt correctly and responded correctly on each subsequent trial within 2 s for 10 consecutive trials. This criterion was met for all children in the first

Figure 4 Flowchart of Error Correction Procedures for the “Frame” Condition

1477Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

intervention session. In all subsequent sessions, all trials were presented with a 2-s prompt delay to allow the child to respond correctly and independently within 2 s (see Figure 4). The instructor presented the four cards and delivered the verbal antecedent while pointing to the rele- vant compound visual stimulus, allowing 2 s for the child to respond. If the child emitted the cor- rect response within 2 s, a token was delivered. If a discrimination error (e.g., saying, “She is drinking,” “Drinking,” or “It’s drinking” when the verbal antecedent was “Who is it?”) or a com- bined response error (e.g., saying, “Mummy is drinking”) occurred, the trial was re-presented with a 0-s prompt delay. If the child responded accu- rately to the 0-s prompt delay, a 2-s prompt delay trial followed. If the child omitted the frame from the response (frame omission) but demonstrated accurate discrimination (e.g., saying, “Drinking” to the verbal antecedent “What is she doing?”), the instructor presented the verbal antecedent again and immediately provided a “frame prompt.” This entailed re-presenting the verbal antecedent and saying the fixed term of the frame (e.g., “She is”) but not the target intraverbal-tact. If the child emitted the target response (e.g., “She is drinking”) after the frame prompt, a 2-s prompt delay trial was presented. If, on this trial, the child did not emit the target response and again omitted the frame, the instructor re-presented the verbal antecedent, followed by a full verbal prompt (e.g., “She is drinking”) and another 2-s prompt delay trial. If the child responded before the verbal antecedent (anticipatory response) was delivered, the instructor presented the verbal antecedent while pointing to the visual stimulus and waited 2 s for the response to occur. If any error occurred, depending on which error (discrimination, frame omission, combined response), the previously described relevant error correction procedures followed; if the response was correct, a token was given. Regardless of the error, any response that had been immediately preceded by a frame prompt or a full verbal prompt did not receive reinforcement and was always immediately

followed by a 2-s prompt delay trial. If the child responded correctly within 2 s prior to the prompt being given, a token was given. No Frame Procedure. The target correct

response was an intraverbal tact corresponding to the verbal antecedent stimulus (i.e., “What Shape?”, “What number?”, “What color?”, “Who is it?”, “What is s/he doing?”) and the visual stim- ulus (e.g., “Triangle,” “Four,” “Red,” “Mummy,” “Drinking”). In this condition, responses to the target stimuli were taught without the frame requirement. As shown in Figure 5, besides the response requirement, criterion to transition to the 2-s prompt delay trial, correction procedures for discrimination errors, combined response errors, anticipatory responses, and reinforcement schedule were identical to the Frame Procedure, with the one exception that no frame omission error was possible. In the very first session at the beginning of intervention, the instructor pointed to a card (e.g., a 4 within a triangle, a photograph of mummy drinking, or a red 4), stated the verbal antecedent (e.g., “What shape?”, “What is she doing?”, or “What color?”) and immediately pro- vided a 0-s delay full verbal prompt of the target response (e.g., “Triangle”, “Cutting”, “Red”). If the child correctly produced the target prompted response, a 2-s prompt delay trial was conducted in which the instructor delivered the verbal ante- cedent again and waited 2 s for the child to respond independently. If a correct response was given within 2 s, the instructor delivered a token and moved on to the next trial.

Postintervention Generalization Sessions Stimuli tested during baseline but not used

during the teaching interventions were pres- ented. This phase was otherwise identical to the baseline sessions.

Interobserver Agreement (IOA) and Treatment Integrity A second independent observer was present

in the room or watched a video of the session

Francesca degli Espinosa et al.1478

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

for all baseline trials and no less than 40% of teaching sessions for all the children (range, 40%-73%) to calculate IOA and treatment integrity. On each trial, the experimenter and observer noted if the response was correct, incorrect, or prompted, and a matched verbal frame had been emitted. IOA was 99.61% for Adam (range, 93%-100%), 99.71% for Alice (range, 93.75%-100%), 96% for Jake (range, 92%-100%), and 99.43% for Fred (range, 93.75%-100%). All six instructors (two of whom were the

second and third authors) conducting the inter- vention were taught to implement procedures using a Behavioral Skills Training package that consisted of explanation, modeling, and role- playing implemented until each instructor met 100% procedural fidelity. To calculate

treatment integrity, a second observer watched videos of the sessions and marked whether the instructor had performed all the steps for the trial, depending on the phase. A trial was cor- rect if the instructor engaged in all of the fol- lowing steps: (a) presented verbal and nonverbal stimuli following the datasheet, (b) provided the correct time delay before giv- ing a prompt, (c) used the specified prompting strategy when an error occurred, (d) presented a transfer trial after any prompted trial, (e) delivered tokens only after a correct transfer or initial unprompted trial, (f) provided token board exchange on a FR 5 schedule, and (g) did not provide feedback during baseline and generalization phases. A trial was scored as incorrect if an error was made in any of the steps. The number of correct trials was divided

Figure 5 Flowchart of Error Correction Procedures for the “No Frame” Condition

1479Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

by the total number of trials and the result was converted to a percentage for each session. Treatment integrity across instructors was 97.5% (range, 87.5%-100%).

Results

Figures 6 to 9 display the percentage of correct responding with respect to teaching and generali- zation sets across the different phases of the study (baseline, intervention, and postintervention gen- eralization). Table 2 displays the mean percentage of errors by component during baseline. During baseline, Adam, Alice, and Jake responded almost exclusively with the number tact, irrespective of whether the verbal antecedent was “What color?” for the Number/Color set or “What shape?” for the Shape/Number set. Fred was the exception, switching between the two responses. The error with respect to Agent/Action was similar for all children: a combined response. For example, say- ing the name of the agent and the action such as “mummy drinking” or “mummy is drinking,” regardless of the verbal antecedent, hence the near zero scores for all children. During intervention, Adam (Figure 6) and

Fred (Figure 7) acquired intraverbal-tacting on

the teaching stimulus sets and showed generalized responding to novel sets in both conditions. Both children extended the use of the frames that had been taught in the intervention phase of the Frame condition to novel stimuli in the post- intervention generalization phase of the Frame condition. Adam required 140 trials to achieve mastery in the Shape/Number Frame condition (7 sessions) and 220 trials in the Agent/Action No Frame condition (11 sessions). Fred met the mastery criterion in 120 trials in the Agent/ Action Frame condition (6 sessions), and in 100 trials in the Number/Color in the No Frame condition (5 sessions). Jake (Figure 8) and Alice (Figure 9) met

mastery criterion in the Frame condition for the teaching stimuli but not in the No Frame condition. When the Frame procedure was employed on the unmastered stimulus set, both participants mastered intraverbal tacts. Jake met mastery criterion for Agent/Action in the Frame condition in eight sessions (i.e., 160 tri- als) and showed generalized frame use and intraverbal-tacting to novel Agent/Action stim- uli. The No Frame condition with the Shape/ Number stimulus set was conducted for an additional five sessions with no progress. Thus,

Figure 6 Mean Percentage of Correct Responding for Adam Across Conditions

Francesca degli Espinosa et al.1480

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

the Frame procedure was introduced on the unmastered Shape/Number teaching set and mastery criterion was subsequently met in seven sessions (i.e., 140 trials). Alice met mas- tery criterion for Number/Color in the Frame condition in seven sessions (i.e., 140 trials) and showed generalized frame use and intraverbal- tacting to novel Number/Color stimuli. The No Frame condition with the Agent/Action

stimulus set was conducted for an additional three sessions with no progress. Thus, the Frame condition was introduced. This modification resulted in an immediate increase in accurate intraverbal-tacting on the Agent/Action teaching stimuli, reaching 100% correct discrimination in the fifth session. Unfortunately, generalized responding to novel Agent/Action stimuli could not be tested due to time constraints.

Figure 7 Mean Percentage of Correct Responding for Fred Across Conditions

Figure 8 Mean Percentage of Correct Responding for Jake Across Conditions

1481Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

Discussion

This study aimed to explore the potential effectiveness of two procedures (Frame and No Frame) on the emergence of generalized intraverbal-tacting for four children with ASD. At baseline, no differential responding to differ- ent verbal antecedents was observed. In the Frame condition, children were taught to emit the target response within a verbal frame. In the No Frame condition, the same procedures were employed, but the target response was taught without the verbal frame requirement. Following intervention with one set of stimuli, responding to novel exemplars of the taught stimulus class was assessed. By the end of the

study, generalized intraverbal-tacting was observed in all four children, but the effects of verbal frame teaching varied across the children. For two children, intraverbal-tacting and subse- quent generalized responding was observed only in the Frame condition. For the other two chil- dren, both procedures were effective, but the Frame procedure was more efficient than the No Frame procedure for one child. Thus, the Frame procedure facilitated acquisition in three of the four children. The study builds on prior research using

similar procedures to establish question dis- crimination with verbal frames (degli Espinosa et al., 2020; Meleshkevich et al., 2020), but is

Figure 9 Mean Percentage of Correct Responding for Alice Across Conditions

Table 2

Mean Percentages of Correct Responding Across Baseline Sessions

Baseline mean percentages by verbal antecedent Stimulus sets

Agent/Action Shape/Number Number/Color

Child Who is it?

What is s/he doing?

Combined response

What shape?

What number?

Combined response

What number?

What color?

Combined response

Adam 0 0 100 100 0 0 Jake 0 2.5 97.5 100 0 0 Fred 0 0 100 43 56 0 Alice 0 0 100 100 0 0

Francesca degli Espinosa et al.1482

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

the first to include a No Frame condition, enabling direct within-subject comparisons of the two procedures. Beneficial effects of the Frame procedure were observed with Jake, Alice, and Adam. Both Jake and Alice required the use of the Frame procedure to acquire dis- criminated intraverbal tacts. Although Adam acquired intraverbal-tacting in both conditions, he required more teaching trials in the No Frame condition compared to the Frame condi- tion, suggesting that while both procedures were effective, the Frame procedure was more efficient. However, for one child (Fred), the Frame and No Frame procedures were similarly effective and efficient. It may be worth noting that Fred was also the only child for whom base- line was slightly above chance level with the Number/Color stimuli. These were the stimuli that had been randomly selected for the No Frame condition, but Fred had a history of prior teaching of this type of discrimination, albeit unsuccessful. Although a 6-week teaching break had been imposed on this skill immediately prior to the study, the effects of this prior teaching his- tory may not have been completely overcome. While the Frame procedure did not facilitate learning the target intraverbal tacts relative to the No Frame procedure, accurate responding to novel stimuli was slightly greater in the Frame condition. This suggests that even when both procedures are similarly effective in the training phase, the Frame procedure may produce greater generalization. For all children, usage of the relevant verbal

frame acquired in one context (Frame interven- tion) was observed in novel but similar contexts (same response class and same verbal antecedent, but different exemplars). When effective frame usage did occur, it was likely due to multiple, converging, and additive intraverbal sources of control generated by the production of the frame. In the case of the stimulus sets involving Number/Color and Shape/Number, the frame response also entailed an echoic. Echoing the last word, which was also the critical conditional

term, may have exerted intraverbal control over the subsequent terms of the frame and the rele- vant visual dimension. This may have narrowed the control of one element of the nonverbal stimulus over incompatible responses, those that did not match the verbal antecedent and the his- tory of the fixed frame elements (i.e., color and color names rather than color and number names). In the stimulus set involving Agent/ Action, similar intraverbal relations were evident. Hearing “Who is it?” intraverbally evoked “It’s” and hearing “What is s/he doing?” evoked “S/he is verbing” that, in turn, exerted control over the subsequent variable terms of the agent and action response class. That no explicit point-to-point correspondence was involved, however, shows that effective frame-based teaching procedures include, but are not limited to, echoic-based training methods. The current procedures bear some resem-

blance to DOR procedures, in which partici- pants are prompted to repeat the verbal antecedent, which has been shown to facilitate listener (see Grow & LeBlanc, 2013 for a review) and intraverbal conditional discrimina- tions (Kisamore et al., 2016). In DOR accounts, repeating the verbal antecedent facili- tates contact with a given stimulus dimension, and this in turn results in accurate responding. While this is undoubtedly true at a descriptive level, an account that explains the DOR in terms of its controlling relations may provide further insight into how the product of the response (e.g., saying an aspect of the verbal antecedent aloud) exerts stimulus control over subsequent behavior. In this study, it is suggested that teaching the children verbal frames that matched the verbal antecedents likely generated two sources of intraverbal con- trol: (1) an intraverbal relation between the ver- bal antecedent (“What color?”, “What shape?”, “What number?”, “Who is it?”, “What is s/he doing?”) and the fixed frame element (“Color,” “Shape,” “Number,” “It’s,” “S/he is”), and (2) an intraverbal relation between the frame

1483Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

elements and the response class matching the frame, determined by both word order (i.e., the word “Color” before the word “Green”) and which terms followed which frame. In other words, saying “It’s,” “Number,” and “Shape” had a history of contiguous usage with identity names only. Saying “She is” and “He is” had a history of contiguous usage with verbs only in the presence of visual stimuli of people per- forming actions. Similarly, “Color” had a his- tory of contiguous usage with color names only in the presence of such stimuli. Irrespective of the stimuli employed, the frame taught, and whether or not it entailed a strictly defined echoic, the Frame procedure was likely effective because of these additional sources of stimulus control generated by producing additional ver- bal stimuli. Moreover, additional sources of stimulus con-

trol were revealed by the error patterns. The dom- inant error in the Number/Color and Shape/ Number sets for three children (excepting Fred) during baseline was saying the number. In con- trast, all children produced a combined response error for the Agent/Action set (e.g., saying “mummy drinking” or “mummy is drinking”). This specific error suggests that although verbal frames containing multiple tacts of a compound stimulus had been established, each tact was not emitted under the sources of control (verbal and visual) that typically evoke these responses in lan- guage acquisition. In typical development, toddlers learn new tacts in social contexts in which verbal and nonverbal stimuli are simultaneously pres- ented (Rowe et al., 2017). At the same time, as their echoic ability increases, they also learn to mand (“want ball”) and tact their environment using multiple terms arranged in structural order (“daddy eating”) by echoing their caregivers in those contexts (Brown, 1973). Because intraverbal-tacting and multiple-word combina- tions appear to occur in a similar developmental time frame, it is unclear how best to assess, sequence, and incorporate these two potentially related skills in ABA-based language intervention.

The combined response error shown by all partici- pants seems to indicate that introducing the skill of responding differentially to relevant verbal ante- cedents before teaching multiple tacts of a com- pound stimulus (e.g., adjective–noun combinations as in “red ball” or agent–action as in “mummy is drinking”) may be advantageous. Research on both the development of an assess- ment and teaching sequences of intraverbal-tacting would help practitioners program in a develop- mentally sensitive way, in which each subsequent skill builds on prerequisite repertoires. This, in turn, may reduce rote responding and the restricted stimulus control errors observed at base- line in the children in this and previous studies. Since the present methods involved several

interrelated components, future research should seek to better isolate those most effective. For example, it would be important to compare the relative efficacy of different discrimination training procedures (e.g., simple-conditional vs. conditional-only). degli Espinosa et al. (2020) used a simple-conditional procedure, whereas Meleshkevich et al. (2020) and this study used conditional-only procedures. While all procedures resulted in generalized skills, a systematic within- study comparison of the different training methods would be useful in determining methods with maximum efficacy. It would also be useful to explore the role of

different methods of stimulus presentation. Like degli Espinosa et al. (2020), but unlike Meleshkevich et al. (2020), stimuli were pres- ented in an array rather than one at the time. Although both presentation methods were effective, our rationale for presenting the stim- uli in an array was informed by previous research on the topic and by practical consider- ations. During intervention, FR 1 token pro- duction and FR 5 exchange production schedules were applied to expedite acquisition. Presenting the stimuli in an array provided stimuli correlated with proximity to reinforce- ment. Furthermore, this arrangement reduced intertrial intervals, as the instructor did not

Francesca degli Espinosa et al.1484

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

have to handle or replace new stimuli in each trial of the array, but instead replaced the entire array during reinforcement consumption. Finally, while more complex, the array arrange- ment may more closely resemble the conditions in which these types of questions are asked in typical language environments. For example, while looking at picture books, the young child shifts attention toward the visual stimulus to which the adult points, before answering the question. This more closely resembles the pre- sent stimulus-array method than presenting one visual stimulus at a time. These variations and combinations may be worth investigating in future research to identify the most efficient, effective, and preferred procedural arrange- ments to establish intraverbal-tacting in chil- dren with ASD. As more procedures are explored and become available, client prefer- ences could be evaluated through a concurrent- chains preference assessment (e.g., Geiger et al., 2012) as a measure of social validity. In future research, it may also prove benefi-

cial to explore structural properties of the stim- ulus sets used. In both the Number/Color and the Agent/Action stimuli the identity, adjective, or action dimensions were embedded within a single stimulus. In the Shape/Number set, each card depicted two visually distinct stimuli, rather than one compound stimulus. The Shape/Number stimuli consisted of two com- posite visual stimuli (and identity tacts) on the same card with each question matching one distinct and separate stimulus, and may have been inherently easier to learn than the other stimulus sets. If so, then differences in stimulus complexity may have interacted with the main independent variable (Frame vs. No Frame). Pre- vious research suggests that differences in stimulus complexity may play a role in the acquisition of intraverbal conditional discriminations. For exam- ple, Aguirre et al. (2019) taught intraverbal responding to verbal antecedents containing over- lapping (e.g., “What vehicle is small?”, “What toy is small?”) and nonoverlapping (e.g., “What

do you brush your teeth with?”, “What food is cold?”) stimuli, and found that it was advanta- geous to first establish responding in relation to nonoverlapping verbal antecedents prior to teaching responding to verbal antecedents with overlapping components. Similarly, Meleshkevich et al. (2020) found it took longer to teach object/ color intraverbal-tacting than it did to establish shape/number, despite the fact that the Frame procedure was used with both sets of stimuli. This may have occurred because of structural dif- ferences between the (overlapping vs. nonoverlapping visual elements) shape/num- ber and the object/color stimulus sets, as well as a difference in the learning history. Children may have had to overcome a longer reinforcement his- tory for saying the identity name of a colored object rather than the color, compared to seeing numbers embedded in shapes—a less familiar combination. Structural properties of teaching stimuli may be important for practitioners to con- sider when choosing the initial stimulus sets to establish intraverbal-tacting. Future research should also consider inter-

spersing generalization probes on untaught stim- uli during the intervention phase rather than at the end of the intervention. Such procedures would more accurately pinpoint the establish- ment or expansion of a response class. Future research might also profitably explore whether frame usage acquired with respect to one class of verbal antecedents generalizes to novel verbal antecedents with progressively fewer subsequent trials or no intervention. Such work would have important applied implications, perhaps reducing the need for prolonged structured discrimination training trials in favor of a transition to more nat- ural teaching contingencies. As with previous research of this kind, the

specific trained response topographies may raise concerns about generalization to the natural environment, because some of the established frames (“Color green” or “Shape triangle”) are grammatically uncommon. Whether specific procedures should be introduced either to fade

1485Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

these frames or to replace them with more con- ventional ones (e.g., “The color is green,” “The shape is a triangle”) should be addressed by future investigations. In our clinical experience, these frames fade over time, potentially because they require more effort and do not produce more reinforcement. Therefore, we would not recommend that specific procedures be employed to teach children to omit the less conventional frames, given their observed facili- tative function in remediating a fundamentally important skill deficit in many children with ASD. Regarding the teaching of a more extended frame, we would tentatively suggest that this may depend on the child’s echoic abil- ity. The children in this study could accurately echo two to three consecutive words at most and thus additional echoic shaping trials may have been necessary. In sum, this study adds to the emerging body

of evidence on methods to establish generalized intraverbal-tacting in children with ASD and to remediate restricted stimulus control errors. Applied research of this kind requires a shift in focus from elementary verbal relations (specific topographies under simple stimulus control) to multiply controlled verbal behavior (generalized response classes under complex stimulus control). In clinical practice, as was the case in this study, different children will likely require different pro- cedural adaptations depending on their existing verbal repertoire and response pattern. The more that is learned about the controlling variables for complex verbal behavior (including but not lim- ited to intraverbal-tacting), the more effective will be the tools available to ABA practitioners to establish generative verbal behavior in children who would not otherwise acquire it.

REFERENCES

Aguirre, A. A., LeBlanc, L. A., Reavis, A., Shillingsburg, A. M., Delfs, C. H., Miltenberger, C. A., & Symer, K. B. (2019). Evaluating the effects of similar and distinct dis- criminative stimuli during auditory conditional

discrimination training with children with autism. The Analysis of Verbal Behavior, 35(1), 21–38. https://doi.org/ 10.1007/s40616-019-00111-3

Bondy, A., Tincani, M., & Frost, L. (2004). Multiply controlled verbal operants: An analysis and extension to the picture exchange communication system. Behavior Analyst, 27(2), 247–261. https://doi.org/10. 1007/BF03393184.

Brown, R. (1973). A first language: The early stages. Har- vard University Press.

degli Espinosa, F., Gerosa, F., & Brocchin-Swales, V. (2020). Teaching multiply-controlled tacting to children with autism. European Journal of Behavior Analysis, 1–21. https://doi.org/10.1080/15021149. 2020.1737407

DeSouza, A. A., Akers, J. S., & Fisher, W. W. (2017). Empirical application of Skinner’s verbal behavior to interventions for children with autism: A review. The Analysis of Verbal Behavior, 33(2), 229–259. https:// doi.org/10.1007/s40616-017-0093-7

DeSouza, A. A., Fisher, W. W., & Rodriguez, N. M. (2019). Facilitating the emergence of convergent intraverbals in children with autism. Journal of Applied Behavior Analysis, 52(1), 28–49. https://doi. org/10.1002/jaba.520

Ervin-Tripp, S. (1970). Discourse agreement: How chil- dren answer questions. In J. R. Hayes (Ed.), Cogni- tion and the development of language (pp.79-107). Wiley.

Geiger, K. B., Carr, J. E., LeBlanc, L. A., Hanney, N. M., Polick, A. S., & Heinicke, M. R. (2012). Teaching receptive discriminations to chil- dren with autism: A comparison of traditional and embedded discrete trial teaching. Behavior Analysis in Practice, 5(2), 49–59. https://doi.org/10.1007/ bf03391823

Goodwin, A., Fein, D., & Naigles, L. R. (2012). Com- prehension of wh-questions precedes their production in typical development and autism spectrum disor- ders. Autism Research, 5(2), 109–123. https://doi.org/ 10.1002/aur.1220

Grow, L. L., Carr, J. E., Kodak, T. M., Jostad, C. M., & Kisamore, A. N. (2011). A comparison of methods for teaching receptive labeling to children with autism spectrum disorders. Journal of Applied Behavior Analysis, 44(3), 475–498. https://doi.org/10.1901/ jaba.2011.44-475

Grow, L., & LeBlanc, L. (2013). Teaching receptive lan- guage skills. Behavior Analysis in Practice, 6(1), 56–75. https://doi.org/10.1007/BF03391791

Hoff-Ginsberg, E. (1985). Some contributions of mothers’ speech to their children’s syntactic growth. Journal of Child Language, 12(2), 367–385. https:// doi.org/10.1017/S0305000900006486

Hoff-Ginsberg, E. (1986). Function and structure in maternal speech: Their relation to the child’s develop- ment of syntax. Developmental Psychology, 22(2),

Francesca degli Espinosa et al.1486

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense

155–163. https://doi.org/10.1037/0012-1649.22. 2.155

Howlin, P. (1982). Echolalic and spontaneous phrase speech in autistic children. Journal of Child Psychology and Psychiatry, 23(3), 281–293. https://doi.org/10. 1111/j.1469-7610.1982.tb00073.x

Ingvarsson, E. T., Kramer, R. L., Carp, C. L., Pétursd�ottir, A. I., & Macias, H. (2016). Evaluation of a blocked-trials procedure to establish complex stimulus control over intraverbal responses in chil- dren with autism. The Analysis of Verbal Behavior, 32 (2), 205–224. https://doi.org/10.1007/s40616-016- 0071-5

Jahr, E. (2001). Teaching children with autism to answer novel wh-questions by utilizing a multiple exemplar strategy. Research in Developmental Disabilities, 22(5), 407–423. https://doi.org/10.1016/S0891-4222(01) 00081-6

Kisamore, A. N., Karsten, A. M., & Mann, C. C. (2016). Teaching multiply controlled intraverbals to children and adolescents with autism spectrum disorders. Jour- nal of Applied Behavior Analysis, 49(4), 826–847. https://doi.org/10.1002/jaba.344

Kisamore, A. N., Karsten, A. M., Mann, C. C., & Conde, K. A. (2013). Effects of a differential observ- ing response on intraverbal performance of preschool children: A preliminary investigation. The Analysis of Verbal Behavior, 29(1), 101–108. https://doi.org/10. 1007/BF03393127

Meleshkevich, O., Axe, J. B., & degli Espinosa, F. (2020). Effects of time delay and requiring echoics on answering questions about visual stimuli. Journal of Applied Behavior Analysis, 54(2), 725-743. https:// doi.org/10.1002/jaba.790

Michael, J., Palmer, D. C., & Sundberg, M. L. (2011). The multiple control of verbal behavior. The Analysis of Verbal Behavior, 27(1), 3–22. https://doi.org/10. 1007/BF03393089

Palmer, D. C. (2007). Verbal behavior: What is the func- tion of structure? European Journal of Behavior

Analysis, 8(2), 161–175. https://doi.org/10.1080/ 15021149.2007.11434280

Palmer, D. C. (2016). On intraverbal control and the def- inition of the intraverbal. The Analysis of Verbal Behavior, 32(2), 96–106. https://doi.org/10.1007/ s40616-016-0061-7

Piazza, C. C., Fisher, W. W., Hagopian, L. P., Bowman, L. G., & Toole, L. (1996). Using a choice assessment to predict reinforcer effectiveness. Journal of Applied Behavior Analysis, 29(1), 1–9. https://doi. org/10.1901/jaba.1996.29-1

Rhyner, P. M. (2007). An analysis of child caregivers’ lan- guage during book sharing with toddler-age children. Communication Disorders Quarterly, 28(3), 167–178. https://doi.org/10.1177/15257401070280030601

Rowe, M. L., Leech, K. A., & Cabrera, N. (2017). Going beyond input quantity: Wh-questions matter for tod- dlers’ language and cognitive development. Cognitive Science, 41(1), 162–179. https://doi.org/10.1111/ cogs.12349

Seidl, A., Hollich, G., & Jusczyk, P. W. (2003). Early understanding of subject and object Wh-questions. Infancy, 4(3), 423–436. https://doi.org/10.1207/ S15327078IN0403_06

Sindelar, P. T., Rosenberg, M. S., & Wilson, R. J. (1985). An adapted alternating treatments design for instructional research. Education and Treatment of Children, 8(1), 67–76.

Skinner, B. F. (1957). Verbal behavior. Appleton-Cen- tury-Crofts.

Stauch, T., LaLonde, K., Plavnick, J. B., Savana Bak, M. Y., & Gatewood, K. (2017). Intraverbal training for individuals with autism: The current status of multiple control. The Analysis of Verbal Behavior, 33(1), 98–116. https://doi.org/10.1007/ s40616-017-0079-5

Received August 23, 2020 Final acceptance June 30, 2021 Action Editor, Alice Shillingsburg

1487Comparison of Intraverbal-Tacting Procedures

19383703, 2021, 4, D ow

nloaded from https://onlinelibrary.w

iley.com /doi/10.1002/jaba.869 by B

ehavior A nalyst C

ertification B oard, W

iley O nline L

ibrary on [13/01/2023]. See the T erm

s and C onditions (https://onlinelibrary.w

iley.com /term

s-and-conditions) on W iley O

nline L ibrary for rules of use; O

A articles are governed by the applicable C

reative C om

m ons L

icense