In Study 1, we assessed children’s Inc-BiN repertoires and categorized their repertoires. Next, each participant completed speaker and listener trainings and we tested for the emergence of corresponding listener and speaker responses, respectively. Our analysis focused on the correspondence between Inc-BiN subtypes (Inc-BiN, Inc-UniN, NiN) and derived speaker and listener literacy responses.

MethodParticipants

Thirteen preschoolers (10 males), with ages ranging from 3.7 to 4.9 years and a mean age of 4.7 years (SD = 0.54), participated in this study. The participants attended school for a duration ranging from 1.5 to 2.5 years (see Table 1). We selected participants based on the presence and absence of the Inc-BiN capability in their verbal repertoire (Inc-BiN assessment described below). Each participant had been receiving educational services at a private behavior analytic preschool utilizing the Comprehensive Application of Behavior Analysis (CABAS) model (Greer, 2002). At the school, educators determined curricula and assessed the presence of developmental cusps via the Preschool Inventory of Repertoires for Kindergarten (C-PIRK; Greer & McCorkle, 2003). Prior research has demonstrated that the C-PIRK curriculum effectively prepares preschoolers for general education classrooms by developing socialization, communication, and adaptive behaviors (Waddington & Reed, 2009). The current study utilized this curriculum because it outlined the prerequisite repertoires of verbal behavior.

Table 1 Demographics and Group Characteristics per Degree of BiN

As per the C-PIRK assessment, all 13 participants demonstrated the prerequisite repertoires to Inc-BiN, including reliably attending to instructors’ voices, faces, and instructional stimuli. All participants in the study had an Individualized Education Plan (IEP) and were educationally classified as a “preschooler with a disability” and received related services. Half of the participants attended a behavior analytic Early Intervention (EI) program prior to enrolling in the preschool where the study took place. Table 1 lists basic information about participants.

The performance of three participants met the criterion for Inc-BiN (criterion on listener and speaker responses). These participants had large verbal repertoires and could vocally mand and tact a variety of items. We defined this as participants emitting six or more vocal verbal operants (i.e., echoics, mands, tacts, and intraverbals) during a 5-min probe in a noninstructional setting. Prior to this study, researchers calculated a mean of 134 learning opportunities (i.e., discrete trials) across academic programs before mastering one objective for Inc-BiN participants. Programs included reading instruction and beginning-level math skills (e.g., one-to-one correspondence between 1 and 20 and the concept of more or less).

Seven participants’ performances met the criterion for Inc-UniN (criterion on listener responses, but not speaker responses). Participants emitting three or more vocal verbal operants (i.e., echoics, mands, tacts, and intraverbals) during a 5-min probe in a noninstructional setting, were defined as having a moderate verbal repertoire. At times, these participants gestured to mand, instead of vocalizing mands in novel situations. Inc-UniN participants required a mean of 228.28 learning opportunities before achieving one objective. Academic programming included early reading skills (e.g., textually responding to letter name and sound) and beginning-level math skills (e.g., number identification).

Three participants did not meet the criterion for speaker or listener incidental language acquisition and were thus deemed as having no incidental naming (NiN). Refer to Table 1 for inclusion criteria for participants. All participants could vocally mand and tact 25 or more items in contrived contexts. At times, these participants required teacher vocal prompts to help communicate their needs in natural settings. NiN participants required 330 learning opportunities before meeting the one academic objective. Each completed academic programming that included prereading skills (e.g., textually responding to letter name and sound) and beginning-level math skills (e.g., number identification).

Setting and Materials

All sessions were conducted in the participants’ respective classrooms. During all sessions, the participant sat next to the researcher at a U-shaped table.

During the derived relations training and testing sessions, participants learned letter-name-sound classes (e.g., A, “A”, “aa”) or sight words that were based on the Fry Sight Word list (e.g., who, pretty, live). Researchers presented the 3.81 by 3.17 cm letters via Microsoft PowerPoint on a 33.2 cm Macintosh laptop. During each trial, one to three letters were presented in a horizontal array, depending on whether the session was measuring speaker (one stimulus at a time) or listener responses (three stimuli simultaneously presented). Sight word instruction was presented in a similar manner, where one to three words appeared on the screen (refer to Fig. 1 for experimental sequence).

Fig. 1

The Experimental Sequence of Study 1 and 2 for All Participants

MeasurementBidirectional Naming

During the bidirectional naming assessment, the untaught topography of point, tact and intraverbal responses were measured. For point-to responses, three stimuli were presented in a horizontal array (two nonexemplars and one target exemplar) with the vocal antecedent of “point to (name of stimulus).” For tact and intraverbal responses, stimuli were presented in random order ensuring that target stimuli were not presented consecutively. Correct responses were those that corresponded to the discriminative stimulus and were emitted within 5 s of the discriminative stimulus. Incorrect responses were those that did not correspond to the discriminative stimulus or no response within 5 s. The assessment of each topography included 20 unconsequated trials and we calculated the percentage correct as number of correct responses divided by 20 and multiplied by 100.

All participants observed object-name relations (i.e., seeing and hearing the name) and were assessed for corresponding speaker (tact, intraverbal) and listener (point to) responses 2 hr later. Participants who scored 80% or higher on speaker and listener responses were categorized as having Inc-BiN. Participants who scored 80% or higher on listener responses, but below 80% on speaker responses were categorized as Inc-UniN. Participants who scored 80% or higher on speaker responses, but below 80% on listener responses were also categorized as Inc-UniN, although no participants in the current study demonstrated that form on Inc-UniN. Participants who scored below 80% for both speaker and listener responses we categorized as having no incidental naming (NiN).

Listener/Speaker Trainings and Derived Responses

Teaching data were collected during learn unit instruction, (i.e., interlocking three-term contingencies; at least two for a teacher and one for the student/learner; Albers & Greer, 1991), where correct responses were reinforced and incorrect responses were consequated with a correction procedure. The researcher collected data on learn units to criterion across each training protocol. The mean number of learn units to criterion (i.e., LUC) was calculated by adding the total number of learn units presented to meet mastery criterion and divided by the number of objectives achieved. We calculated learn units to criterion separately for listener and speaker protocols.

During letter name and sound instruction, the researcher administered 30 learn units (LUs; i.e., learning trials) per session. Half of the trials were letter names and half were letter sounds and this included five LUs for each of the three target responses—all presented in a randomized order. Both topographies (letter names and sounds) were plotted as separate data paths (see Figs. 2, 3 and 4). At the end of each session, correct responses were totaled and graphed as a percentage. In order to calculate the correct percentage, the number of correct LU responses were divided by the total number of LU opportunities presented then converted into a percentage (i.e., 5/15 * 100 = 33%).

Fig. 2

Training Data for Each Inc-BiN Participant in Study 1

Fig. 3

Training Data for Inc-UniN Participants in Study 1

Fig. 4

Training Data for Inc- NiN Participants in Study 1

During sight word instruction, the researcher administered 20 LU (i.e., learning trials) per session, that included four operants with five LU opportunities to respond to each operant—also presented in a randomized order. At the end of the session correct responses were recorded in a similar manner as letter name and sound instruction.

We measured the accuracy of derived responses following listener and speaker trainings. We measured derived responses in accordance with which repertoire was reinforced (i.e., measurement of speaker responses following the listener training protocol and measurement of listener responses following the speaker training protocol). A correct response was one that corresponded to the discriminative stimulus and was emitted within 5 s. For example, during the speaker training protocol, if the discriminative stimulus was a picture of the letter A, tacting the stimulus “A” was considered a correct speaker response. An incorrect response was one that did not correspond to the discriminative stimulus or no response within 5 s. Whereas, during the listener training protocol, if the discriminative stimulus was a field of three stimuli (i.e., two nontarget stimuli and one target stimulus) and the vocal antecedent of “point to the letter A” was given, pointing to the target stimulus “A” was considered a correct listener response. An incorrect response was one that did not correspond to the target stimulus or no response within 5 s.

Procedures

The study began with researchers assessing each participant’s degree of incidental bidirectional naming. Then, participants contacted both listener and speaker protocols (see Fig. 1). Below we describe each component in detail.

Inc-BiN assessment

The Inc-BiN assessment was composed of two components: exposure to name-object relations and the assessment of corresponding speaker and listener responses (Greer & Longano, 2010; Morgan et al., 2020). During the exposure, the researcher obtained the participant’s attention, provided identity matching trials, named the stimulus during the trials, and provided consequences contingent on the participant’s match response (not contingent on responses in accordance with the name, such as an echoic).The purpose of the matching trials was to provide an opportunity for the researcher to associate names with stimuli during a task that was in the participant’s repertoire.

At the start of each trial, the researcher laid out three stimuli in front of the participant (two nontarget exemplars and the target stimuli), presented the participant with the target exemplar and the vocal antecedent, “Match ____ to _____” (e.g., “Match Krumm to Krumm”). Contingent upon a correct matching response, the researcher provided social consequences such as vocal praise (i.e., “Great job Susan!”) and tickles (with learner’s consent). If the participant matched the target stimulus to a nonexemplar or did not respond within 3 s, the researcher implemented a correction procedure whereby she modeled the correct matching response and represented the antecedent to allow the participant the opportunity to independently respond. Each session was composed of 20 match responses (five stimuli with four LUs per stimulus). Inc-BiN exposure sessions were continued until a participant responded with 90% matching accuracy during one session (Greer et al., 2005). It is important to note that there were no reinforcement contingencies arranged for responding to the researcher’s modeling of object–name relations (e.g., echoics).

The second component of the Inc-BiN assessment was a test for corresponding listener and speaker responses. After the Inc-BiN exposure, the researcher assessed corresponding listener and speaker responses 2 hr later. Three responses were measured. The researcher assessed listener responses by presenting the participant with three stimuli (two nonexemplars and one target stimulus) and the vocal direction “Point to _____”. For speaker responses, the researcher assessed pure and impure tacts. For pure tacts, the researcher held one stimulus in front of the participant and waited up to 5 s for the participant to provide the corresponding tact (name). For impure tacts, the researcher conducted trials similar to pure tact trials and added the vocal direction, “What is this?” The researcher conducted 20 LUs for each response type (point to, pure tact, impure tact), which was composed of four LUs opportunities to respond to each of five stimuli. During this probe for listener and speaker responses, the researcher provided no accuracy feedback, but intermittently praised academically related behaviors such as sitting and attending.

Listener and Speaker Protocols

During each teaching phase, participants were taught either listener or speaker responses and then tested in the untaught topography (i.e., derived responses). Teaching phases were conducted as learn unit instruction (Albers & Greer, 1991). Upon emitting a correct response, the researcher delivered vocal praise, playful physical contact, or a token—the specifics were tailored to each participant and held constant across training conditions. The researcher engaged in a correction procedure contingent on incorrect responses. During derived response phases, the researcher provided no feedback.

Preassessments

Preassessements were conducted for both listener and speaker responses at the onset of the study to ensure we taught responses not already in each participant’s repertoire. The letter-name-and-sound assessment included the letter names and letter sounds for all 26 letters in the alphabet. During the preassessment, one sample stimulus was presented at a time, where the participant was instructed to respond to “What letter is this?” or “What sound does this make?” During the preassessment and probe sessions no consequences were delivered for correct and incorrect responses. The preassessment consisted of 52 responses for letter names/sounds. The researcher randomized the order of presenting letters for each participant. After completing the preassessment, the researcher chose letters to teach based on those the participant responded to incorrectly during the preassessment and the researcher eliminated letters for which a participant responded accurately.

Preliminary data indicated that five (Inc-BiN- 1, Inc-BiN-2, Inc-BiN-3, Inc-UniN-1, and Inc-UniN-2) of the selected participants had previously mastered all letter names and sounds. For these participants, a preassessment on 100 preprimer Dolch words were conducted in the same manner. During the preassessment, one sample was presented at a time, where the participant was instructed to “read.” During the preassessment and probe sessions no consequences were delivered for correct and incorrect responses. The researcher randomized the order of words for each participant. After completing the preassessment, the researcher chose 16 words to teach based on those the participant responded to incorrectly during the preassessment and the researcher eliminated words to which a participant responded accurately. It is important to note that the researcher selected the same 16 sight words for all participants.

Listener teaching protocol

The listener teaching protocol consisted of teaching the participant to identify target stimuli with a point response, followed by tests of the emergence of corresponding speaker responses. For letter names and sounds instruction, the teaching phase consisted of two experimentally defined responses (i.e., letter names and letter sound). For each session, instruction was presented in a rotated sequence. The participant received 15 LU opportunities to point to the letter name and 15 LU opportunities to point to the letter sound, for a total of 30 LUs. The criterion for each response was 90% accuracy for two consecutive sessions or 100% accuracy in one session. During each session, the researcher presented an array of three stimuli, one target exemplar and two nontarget exemplars. After presenting the stimuli, the researcher presented the vocal antecedent (e.g., “point to M”/ “point to /m/”) and allowed 5 s for the participant to respond. If the participant pointed to the target exemplar, the researcher provided social consequences (i.e., vocal praise, playful physical contact, tokens). If the participants emitted an incorrect response, the researcher implemented a correction procedure. For the correction procedure, the researcher modeled the correct response (i.e., pointed to the target stimulus), and re-presented the antecedent to give the participant an opportunity to emit the response independently. If the participant continued to emit the incorrect response after the correction procedure, the researcher modeled the response up to three times before proceeding to the next learn unit. The researchers provided no consequences for corrected responses.

Researchers taught sight words in the same manner as letter names and sounds, with minor exceptions. The researcher administered 20 LU per sessions, composed of five LU opportunities to point to each of four sight word. During each session, the researcher presented an array of three stimuli, one target exemplar and two nontarget exemplars. After presenting the stimuli, the researcher presented the vocal antecedent (e.g., “point to [word]”) and allowed 5 s for the participant to respond.

Regardless of letter-names-sounds or sight words instruction, after a participant’s responding met mastery criterion during the listener teaching protocol portion, the researcher assessed derived speaker responses. The researcher presented each stimulus on the laptop via Microsoft PowerPoint, said, “What letter is this? / “what sound does this make?” and waited up to 5 s for the participant to respond. For letter names and sounds derived responses, there were 12 LUs in a session, which was composed of three stimuli, two responses (e.g., letter name, letter sound), and two opportunities of each learn unit type. For sight word derived responses there were six unconsequated LUs in a session, which was composed of three stimuli and two opportunities of each learn unit type. Criterion for the demonstration of the derived responding was 80% accuracy in one session.Footnote 1

Speaker teaching protocol

The speaker teaching protocol consisted of teaching the participant to textually respond to target stimuli (e.g., selected sight, followed by tests of the emergence of corresponding listener responses). For letter names and sounds instruction, the training phase consisted of two experimentally defined responses (i.e., letter name and letter sound). For each session, instruction was presented in a rotated sequence. The participant received 15 LU opportunities to answer the intraverbal tact response of letter name (i.e., “What letter is this?”) and 15 LU opportunities to intraverbal tact response of letter sound (i.e., “What sound does this make?”) for a total of 30 LUs. The criterion for each response was 90% accuracy for two consecutive sessions or 100% accuracy across one session. During each session, the researcher presented the target exemplar on a laptop via Microsoft Office PowerPoint with the vocal antecedent and followed 5 s for the participant to respond. If the participant emitted the target response the researcher provided social consequences (i.e., vocal praise, playful physical contact, tokens). If the participants emitted an incorrect response, the researcher implemented a correction procedure. For the correction procedure, the researcher modeled the correct response (e.g., “/m/” or “m”) and re-presented the antecedent to give the participant an opportunity to emit the response independently. If the participant continued to emit the incorrect response after the correction procedure, the researcher modeled the response up to three times before proceeding to the next learn unit. The researchers provided no consequences for corrected responses. Data were collected and recorded in the same manner as stated above. The same procedure was implemented for the participants that were taught sight words.

Researchers taught sight words in the same manner as letter names and sounds, with minor exceptions. The researcher administered 20 LUs per session, composed of five LU opportunities to textually respond/tact each of the four sight words. During each session, the researcher presented one target exemplar on the laptop screen via Microsoft Office PowerPoint. After presenting the stimuli, the researcher presented the vocal antecedent “read” and allowed 5 s for the participant to respond.

After a participant’s responding met mastery criterion during the speaker training portion of the protocol, the researcher assessed derived listener responses. The researcher presented each stimulus on the laptop via Microsoft Office PowerPoint and said, “point to the letter m / “point to the letter that makes the /mm/ sound?” or “point to here” and waited up to 5 s for the participant to respond. For letter names and sounds derived responses, there were 12 LUs in a session, which was composed of three stimuli, two responses (e.g., letter name, letter sound), and two opportunities of each learn unit type. For sight word derived relations, there were six unconsequated LUs in a session, which was composed of three stimuli and two opportunities of each learn unit type. Criterion for the demonstration of the derived responding was 80% accuracy in one session. See Table 2 for acquisition probes and intervention rotation per participant.

Table 2 Sequence in Acquisition Probes and Intervention Decisions to Stop a Protocol

The researcher ended teaching portions of protocols contingent on a participant mastering the respective protocol. In addition, we implemented a decision protocol (Keohane & Greer, 2005) to make decisions about whether to stop instruction if learning was not occurring. We implemented the decision protocol as an ethical guard against keeping participants in a training phase when additional instructional supports were needed beyond the purview of this experiment. We analyzed data paths, or the line connecting two consecutive data points. “Stop” decisions in the decision protocol included 0% accuracy in one session, three consecutive descending data paths, or five variable and overall descending data paths. We terminated a training phases contingent on the data meeting two stop decisions. This rule only affected NiN participants.

Experimental Design

In the present study, repeated acquisition design was utilized (Kirby et al., 2021). Each participant underwent both listener and speaker protocols, along with corresponding derived responses tests. The order of the training protocols (listener or speaker) was counterbalanced across participants, ensuring variability, and controlling for potential order effects. For instance, Participant Inc-BiN-1 commenced with the listener training phase, whereas Participant Inc-BiN-2 started the study with the speaker training phase. In addition, the assignment of specific stimuli to protocols was counterbalanced across participants to minimize potential bias. For example, the set of stimuli assigned to Participant Inc-UniN-1’s listener protocol was systematically assigned to Participant Inc-UniN-2’s speaker protocol. Participants progressed through the designated conditions (listener training protocol or speaker training protocol) at their individual pace, continuing until their performance met the predefined teaching criterion within a given condition. Each participant rotated through both conditions (listener and speaker) twice, resulting in a total of four conditions for each participant.

Throughout the study, we assessed the association between the degree of Inc-BiN (including participants with varying degrees such as Inc-BiN, Inc-UniN, and NiN) and derived responses. We categorized participants based on their respective degrees and examined differences in derived listener and speaker responses among these groups. This analysis allowed us to explore potential patterns or distinctions in the relationship between the participants' degree of Inc-BiN and their performance on derived responses tasks for both listener and speaker protocols. The inclusion of participants with different degrees adds depth to our exploration, providing an understanding of the interplay between the degree of Inc-BiN and the acquisition of derived responses.

Interobserver Agreement and Procedural Fidelity

A second independent observer collected data for the purposes of calculating interobserver agreement (IOA) and treatment fidelity using the Teacher Performance Accuracy Rate Observation form (TPRA; Hranchuk & Williams, 2021; Ingham & Greer, 1992; Ross et al., 2005). The researchers calculated trial-by-trial IOA by counting the number of learn units on which researcher and independent observer agreed, divided that number by the total number of opportunities presented, and multiplied the resulting number by 100. IOA was collected for 56% of the Inc-BiN probe (i.e., untaught response topographies after Inc-BiN exposure) sessions with 100% agreement. We recorded IOA for 32% of the training protocol sessions (i.e., listener and speaker training phases) with 100% agreement and 100% of the derived responses sessions with 98% agreement (range, 95% to 100%).

The independent observer also collected procedural fidelity data using the TPRA form. Per observation, the independent observer rated the accuracy of researcher-delivered antecedents and consequences. Fidelity was calculated by dividing the number of correct components divided by the total number of components multiplied by 100. Treatment fidelity was conducted for 21% of all session with a mean fidelity of 98% (range, 95% to 100%).

Results and DiscussionListener and Speaker Protocol Performances

Each participant completed speaker and listener trainings and were tested for untaught derived responses. Below we categorize outcomes by level of Inc-BiN at the start of the study. Figures 2, 3, and 4 display the training and derived responses data for participants with Inc-BiN, Inc-UniN, and NiN, respectively. In each graph, all listener responses are grey, such that grey data points represent listener trainings and grey bars represent derived listener responses. All speaker responses are black, such that black data points represent speaker training and black bars represent derived speaker responses.

Participants with Inc-BiN

Figure 2 displays the training and derived responses data for three Inc-BiN participants. All Inc-BiN participants steadily achieved the acquisition criterion during the respective trainings and met criterion levels during all derived responses probes. Inc-BiN-1, Inc-BiN-2, and Inc-BiN-3 exhibited a rapid learning pattern, regardless of the specific training approach. On average, it took these participants 63 LUs to master speaker responses across two to four sessions, achieving derived listener responses with high accuracy between 98% to 100%. Likewise, the acquisition of listener responses was consistent, with participants requiring an average of 50 LUs across one to seven sessions. They also achieved derived speaker responses with an accuracy ranging from 94% to 100%. Overall, Inc-BiN participants readily acquired speaker and listener responses under direct reinforcement contingencies and derived corresponding listener and speaker responses.

Participants with Inc-UniN

Figure 3 displays the training and derived responses data for seven Inc-UniN participants. All Inc-UniN participants steadily achieved acquisition criterion during the respective trainings and met criterion levels during derived listener relation probes. On average, it took participants approximately 120–255 LUs to master speaker responses, with the number of sessions ranging from 3.5 to 8.5. However, accuracy rates in deriving listener responses after learning speaker responses ranged from 83.33% to 100%. Learning listener responses also varied, taking an average of about 30–180 LUs, over two to six sessions. After mastering listener responses, accuracy rates in deriving speaker responses ranged from 25% to 100%.

Participants with NiN

Figure 4 displays the training and derived responding data for three NiN participants. NiN participants completed each condition only one time due to the lack of prerequisite skills needed in order to complete the academic task, such as attending to letter stimuli. All NiN participant’s performances met the decision protocol rules to stop a training phase and no NiN participant met criterion levels during derived responses probes. Overall, NiN participants did not acquire derived listener or speaker responses under direct reinforcement contingencies for both training conditions.

Summary

Overall, participants categorized as Inc-BiN derived both listener and speaker responses, independent of the training protocol. In addition, they also had a faster rate of acquisition during both trainings; approximately half of their Inc-UniN counterparts. Inc-UniN participants derived listener responses after completing speaker trainings, but not vice versa. These results are consistent with verbal behavior development theory (VBDT), which states that listener behavior precedes speaker behavior (Greer & Speckman, 2009). That is, participants can readily acquire listener responses (i.e., matching and pointing) before emitting speaker behavior (i.e., tacting/labeling). Furthermore, all NiN participants had variable responding across trainings, and demonstrated a lack of derived responding for both listener and speaker responses. The outcomes are in alignment with Morgan et al. (2020), who showed similar relations. These outcomes extended previous work by quantifying outcomes in terms of meeting criterion for demonstrating reliable derived responses or not. This study also extends Morgan et al.’s findings on the emergence of speaker and listener responses following match-to-sample responses by examining both listener–speaker and speaker–listener derived responses.

Outcomes According to Degrees of Bidirectional Naming

In the following analysis, we compare outcomes based on each participant’s degree of bidirectional naming at the start of the study, categorized as Inc-BiN, Inc-UniN, or NiN. Descriptive statistics were conducted for three measures: (1) the mean number of LUC across all academic programs; (2) LUC across the listener training protocol; and (3) LUC across the speaker training protocol. These measures are depicted in Tables 1 and 3.

Table 3 Familiar Stimuli Used during MEI InterventionLearn Units to Criterion (LUC)

We compared the number of learn units required to master listener and speaker responses. We omitted data from NiN participants because these participants’ responding often met the stop decision protocol rule. Thus, below we report data on Inc-BiN and Inc-UniN participants (refer to Table 3). Overall, Inc-UniN participants had a higher average of overall LUC (M = 228.29, SD = 57.43) than Inc-BiN participants (M = 134, SD = 10.15) – meaning Inc-UniN participants required more learning opportunities to respond at criterion.

Derived Responses

Figure 5 displays the percentage of derived speaker and listener responses grouped by degrees of bidirectional naming. Overall, participants who demonstrated Inc-BiN derived the untaught topography, regardless of protocol, whereas participants who demonstrated Inc-UniN derived 95.92% of the listener responses (following speaker training) and 52.92% of the speaker listener (following listener training). Participants who demonstrated NiN, were unable to derive the untaught relation at criteria levels, which was not surprising given that these participants did not meet mastery criterion during trainings.

Fig. 5

Percentage of Derived Responses for Participants with Various Degrees of the Bidirectional Naming Capability in Study 1. Note. BiN (n = 3), UniN (n = 7), and NiN (n = 3)

Summary

The findings from Study 1 demonstrate associations between a participant’s degree of Inc-BiN and derived responses, similar to Morgan et al. (2020). However, Morgan et al. taught baseline relations using match-to-sample procedures. Thus, our findings expand the generality of the relation between Inc-BiN and derived responses to traditional verbal behavior research on the emergence of listener and speaker responses. In particular, participants who demonstrated NiN struggled to acquire both listener and speaker responses and subsequently failed to derive corresponding speaker and listener responses. Participants who demonstrated Inc-UniN acquired both listener and speaker responses under direct reinforcement conditions and derived listener, but not speaker responses. Last, participants who demonstrated Inc-BiN acquired both listener and speaker responses under direct reinforcement conditions and derived both speaker and listener responses. The current findings suggest a relation between bidirectional naming and derived responses and pinpoint how different subcategories of bidirectional naming are associated with different types of derived responses. Our findings replicate those of Morgan et al. (2020) and add a level of detail by parsing out differential effects on listener and speaker derived responses, in particular.

Eikeseth and Smith (1992) was one of the first studies to highlight the impact of naming on derived relations. Once naming was established, discrimination training such as match-to-sample procedure (i.e., listener training) produced emergent performance as a speaker, and vice versa. This effect was replicated by Howarth et al. (2015). Research focused on listener and speaker trainings demonstrate mixed findings, suggesting that these trainings may not be universally applicable, but should be tailored to individual learners based on their preexperimental repertoires and learning histories. These discrepancies prompt questions about the learner—for whom do these outcomes apply?

The results of Study 1 demonstrated an association between a learner’s degree of bidirectional naming and the acquisition of both listener and speaker responses, as well as derived speaker and listener responses (Morgan et al., 2020). However, we observed this association based on the degree of bidirectional naming that a participant possessed at the outset of the experiment. Thus, it remains to be seen how altering a participant’s Inc-BiN repertoire leads to corresponding changes in derived responses.