I thank Karola Dillenburger and the associate editor for their valuable feedback on earlier versions of this manuscript. This study is part of the author's doctoral dissertation. Portions of the data were presented at the sixth international convention of the Association for Behavior Analysis International (2011).
Tact training versus bidirectional intraverbal training in teaching a foreign language
Article first published online: 11 OCT 2013
© Society for the Experimental Analysis of Behavior
Journal of Applied Behavior Analysis
Volume 47, Issue 1, pages 165–170, Spring 2014
How to Cite
Dounavi, K. (2014), Tact training versus bidirectional intraverbal training in teaching a foreign language. Jnl of Applied Behav Analysis, 47: 165–170. doi: 10.1002/jaba.86
- Issue published online: 7 MAR 2014
- Article first published online: 11 OCT 2013
- Manuscript Accepted: 8 AUG 2013
- Manuscript Received: 12 NOV 2012
- emergent relations;
- foreign language acquisition;
The current study involved an evaluation of the emergence of untrained verbal relations as a function of 3 different foreign-language teaching strategies. Two Spanish-speaking adults received foreign-language (English) tact training and native-to-foreign and foreign-to-native intraverbal training. Tact training and native-to-foreign intraverbal training resulted in the emergence of a greater number of untrained responses, and may thus be more efficient than foreign-to-native intraverbal training.
There is increasing demand for individuals to speak a second language (European Union, 2006), but few behavior-analytic studies have focused on foreign-language instruction with typically developing populations. According to Skinner's (1957) analysis, learning a foreign language includes the acquisition of several different verbal repertoires, including listener responses, mands, tacts, and intraverbals. Of specific interest in the current study are three verbal operants commonly targeted in foreign language instruction: (a) the foreign language tact, in which the learner emits a foreign word as a response to a nonverbal stimulus; (b) the native-to-foreign (N-F) intraverbal, in which the learner emits a foreign word as a response to its native equivalent; and (c) the foreign-to-native (F-N) intraverbal, in which the learner emits a native word as a response to its foreign equivalent (Petursdottir & Haflidadottir, 2009).
Skinner (1957) viewed verbal operants as functionally independent (e.g., Miguel, Petursdottir, & Carr, 2005), but functional interdependence occurs under some conditions (e.g., Grannan & Rehfeldt, 2012). Some studies have suggested that when two relations share a common novel stimulus or response topography, teaching one relation is more likely to result in the emergence of the other relation (e.g., Petursdottir, Olafsdottir, & Aradottir, 2008). In a study with two typically developing preschool children, Petursdottir and Haflidadottir (2009) found that N-F intraverbal training, despite requiring more sessions, was more likely to result in emergent tacts and intraverbal responses compared to F-N intraverbal training. The authors suggested that this was due to the shared novel (i.e., foreign) topography between foreign tacts and N-F intraverbals. However, not all of their results were consistent with that interpretation. The goal of the present study was to evaluate foreign tact and bidirectional intraverbal training in teaching a foreign-language vocabulary to adults and further evaluate possible emergence of untrained responses following each type of training.
Participants and Setting
Two native Spanish speakers participated: Pedro, a 37-year-old man, and Carlota, a 29-year-old woman. Pedro had a high-intermediate and Carlota had a low-intermediate level of English language proficiency, as assessed through a brief online test (http://www.cambridgeesol.org/testyourenglish/index.php). Both participants were able to echo unknown words in English. Most sessions were conducted in the participants' homes.
The experimenter continuously recorded responses on a data sheet. Correct foreign tacts were defined as speaking the word in the foreign language (English) after presentation of the corresponding picture. Correct intraverbal responses were defined as vocalizing the foreign (English) or native (Spanish) word when presented with its equivalent and vice versa. A correct response was recorded only if it occurred within 3 s of the antecedent stimulus presentation. An incorrect vocalization, no response within 3 s, or more than one response were all recorded as incorrect. In addition, for F-N intraverbals, synonyms of the native target word were recorded as correct (e.g., participants translated “socket” as “enchufe,” “toma de corriente,” or both).
A second observer independently collected interobserver agreement data across all conditions. Data were recorded in 31% of sessions for Pedro and 33% of sessions for Carlota. An agreement was scored for each trial on which both observers recorded a correct or an incorrect response; otherwise, a disagreement was scored. Interobserver agreement for each session was calculated by dividing the number of agreements by the total number of trials and converting this ratio to a percentage. Mean agreement was 99.7% (range, 93% to 100%) for Pedro and 99.8% (range, 97% to 100%) for Carlota.
Stimulus sets and general procedure
The antecedent stimuli were pictures for tact probes and tact training and spoken words for intraverbal probes and intraverbal training. Stimuli were randomly assigned into four 30-stimuli sets and integrated into separate PowerPoint presentations. The pictures were retrieved from the Internet and depicted on a white background for 3 s. Spoken words were recorded and integrated into the PowerPoint presentation. For both participants, foreign tact training was conducted with Sets 1 and 2 and intraverbal training with Sets 3 and 4, with five to seven sessions conducted per day on consecutive days, and an intersession interval of 30 s to 5 min. Stimuli within each set were presented in the same order across sessions, simulating common printed materials (e.g., books with lists of translated words). The order of stimulus presentation within each set was randomly changed after the participant demonstrated 30 of 30 correct responses in one session (30 stimulus presentations) during training. For both the tact and intraverbal training evaluations, native-language tact probes were conducted after foreign tact or F-N intraverbal pretests and before the baseline probe of the second operant (either tact or F-N intraverbals). This was done to ensure the use of the same native word for both participants and to eliminate ambiguity regarding the illustrations (e.g., an adjustment was made when participants tacted a button in Spanish instead of the cuff).
Pretests were conducted to identify unknown stimuli to include in subsequent probe and training sessions. During foreign tact and F-N intraverbal pretests, the experimenter presented a PowerPoint for each 30-stimulus set. Before tact pretests, the experimenter instructed the participants to name the pictures in English. Before F-N intraverbal pretests, the participants were instructed to say the Spanish word when presented with the English word. Known words were substituted until a sufficient number of unknown words had been identified. If one participant gave a correct response on the pretest, that word was substituted for both participants.
Baseline and posttraining probes
During baseline, F-N intraverbal and foreign tact probes were conducted. Presession instructions were the same as in the pretests. In the foreign tact probes, the pictures in the designated stimulus set were presented via a PowerPoint presentation. F-N intraverbal probes were identical, except that spoken words were presented rather than pictures. The experimenter informed the participant of the number of correct responses at the end of each probe session; no other consequences were delivered. Identical probes, with the addition of an N-F intraverbal probe, were conducted posttraining to test for the emergence of untrained relations. The mastery criterion during posttraining probes was 30 of 30 correct responses. If a participant scored less than 30 but at least 27, a second probe session was conducted. For intraverbal probes, if less than 30 of 30 correct responses were obtained in the second probe or less than 27 of 30 on the first probe, the reverse intraverbal relation was trained.
During foreign tact training, the participants were taught to tact the pictures by emitting a response in the foreign language (English) following the presentation of the corresponding picture. During N-F intraverbal training, the participants were taught to emit the correct foreign (English) word after the presentation of the corresponding spoken word in their native language (Spanish). During F-N intraverbal training, the participants were taught to emit the correct native (Spanish) word after the presentation of the corresponding spoken word in the foreign language (English).
Training sessions were identical to probes except that correct responses were reinforced with the presentation of the correct answer in the form of a spoken word, which had been embedded in the PowerPoint and scheduled to appear 3 s after the presentation of the antecedent stimulus. A correction was automatically provided in the same way when an incorrect response was emitted; the participants were not required to repeat the correct response. Social praise was intermittently provided during sessions and was always delivered at the end of sessions for improved accuracy (i.e., if the participant had obtained a higher number of correct responses than in the previous sessions). Otherwise, verbal encouragement to try harder was provided (e.g., “You have learned the difficult words now, let's see if you can recall more this time.”). The mastery criterion during training was 30 of 30 correct responses for two consecutive sessions.
The effects of tact and bidirectional intraverbal training on the emergence of untrained intraverbal responses were evaluated via a pretest–posttest design with replication across stimulus sets and participants. The sequence of probe conditions was counterbalanced across participants and stimulus sets. For intraverbal training, the sequence of the different training conditions was counterbalanced across participants.
RESULTS AND DISCUSSION
Foreign tact training resulted in the emergence of bidirectional intraverbal relations that reached mastery level for both participants (Figure 1, top). Furthermore, intraverbal training in the N-F direction resulted in the emergence of the reverse intraverbal and the foreign tact relations at mastery level (Figure 1, bottom). Results were different when intraverbal training was provided in the F-N direction: Emergent responses occurred, but these were below mastery level. When training was provided for the reverse intraverbal relation (N-F, Spanish to English), foreign tacts emerged at mastery level. The fact that emergent relations (foreign tact relations) shared a response with the trained relation (N-F direction intraverbals) might account for the present results (Petursdottir et al., 2008; Polson & Parsons, 2000).
In addition to the emergence of untrained relations, the current results suggest differential efficiency of the training methods. For both participants, N-F intraverbal training was the most efficient (i.e., 17 sessions for Pedro and 18 sessions for Carlota), followed by foreign tact training (i.e., 19 and 21 sessions for Pedro and 24 and 23 sessions for Carlota). Because additional training (i.e., reverse intraverbal training) was required after F-N intraverbal training, this approach was the least efficient.
The present results differ from those obtained in studies that have examined transfer of stimulus control in children, in which no or few emergent responses have been found (Miguel et al., 2005; Petursdottir & Haflidadottir, 2009; Petursdottir et al., 2008). This may be due to the current study including adult participants, who are verbally more competent.
It would be interesting to examine how prompts provided in different sensory modalities affect the acquisition of verbal operants and whether the possible differences can be attributed to an individual's learning history with specific prompting techniques (Coon & Miguel, 2012). Alternatively, the prompts' formal relation to the response might be important (e.g., auditory prompts might facilitate the acquisition of vocal responses and textual prompts might facilitate the acquisition of textual responses). Finally, there might be differences in emergent relations when probes are conducted in a sensory modality that is different from training.
In the present study, foreign tact training was conducted prior to intraverbal training for both participants; thus, sequential confounding is a possibility. Also, N-F intraverbal probes were not conducted during baseline, which does not allow a direct comparison of pretest and posttest data. It was assumed that this probe would not provide additional information, because previous research had suggested that the acquisition of F-N intraverbals is faster than the acquisition of the reverse (Petursdottir & Haflidadottir, 2009). Therefore, pretesting of F-N intraverbals might provide sufficient information for the present purposes. In addition, given that F-N intraverbals and foreign tacts had been previously tested, participants might guess which foreign word corresponded to each native word during N-F intraverbal probes, and this could confound the probe results. On the other hand, including an N-F probe might reveal whether guessing is likely to occur, which could allow researchers to separate the effects of training from the effects of mere exposure to the training stimuli. Therefore, researchers might consider including N-F intraverbal probes in future studies. In addition, the type of feedback provided at the end of probe sessions might be sufficient for participants to deduce the correct answers. Future research could build on the current study by including baseline probes for all tested relations, a greater number of baseline probe sessions, and eliminating feedback during probes. Future research should attempt to replicate these findings across more participants, and maintenance and generalization measures should be collected.
A careful consideration of Skinner's (1957) analysis of verbal behavior and of observed emergent responses may prove to be useful in designing effective methods to teach a second language. Future research should explore the use of new technologies, such as the development of programmed self-instruction applications, to teach foreign-language vocabulary, to enhance fluency, and to deliver instruction with high procedural fidelity.
- 2012). The role of increased exposure to transfer-of-stimulus-control procedures on the acquisition of intraverbal behavior. Journal of Applied Behavior Analysis, 45, 657–666. doi: 10.1901/jaba.2012.45-657 , & (
- European Union. European Commission. (2006, February). Europeans and their languages (Special Eurobarometer 243/Wave 64.3: TNS Opinion and Social). Retrieved from http://ec.europa.eu/public_opinion/archives/ebs/ebs_243_en.pdf
- 2012). Emergent intraverbal responses via tact and match-to-sample instruction. Journal of Applied Behavior Analysis, 45, 601–605. doi: 10.1901/jaba.2012.45-601 , & (
- 2005). The effects of multiple-tact and receptive-discrimination training on the acquisition of intraverbal behavior. The Analysis of Verbal Behavior, 21, 27–41. , , & (
- 2009). A comparison of four strategies for teaching a small foreign-language vocabulary. Journal of Applied Behavior Analysis, 42, 685–690. doi: 10.1901/jaba.2009.42-685 , & (
- 2008). The effects of tact training and listener training on the emergence of bidirectional intraverbal relations. Journal of Applied Behavior Analysis, 41, 411–415. doi: 10.1901/jaba.2008.41-411 , , & (
- 2000). Selection-based versus topography-based responding: An important distinction for stimulus equivalence? The Analysis of Verbal Behavior, 17, 105–128. , & (
- 1957). Verbal behavior. New York, NY: Appleton-Century-Crofts. doi: 10.1037/11256-000 (
Action Editor, Einar Ingvarsson