The present investigation evaluated the effectiveness of multiple exemplar training (MET) on the facilitation of bidirectional symmetry-like relations (object-naming) for typically developing children (ages 3-4 yrs) whose first language was Spanish. Two experiments were conducted in which a multiple probe design was implemented to introduce exemplar training across 3 four-item stimulus sets. Participants were first trained in listener (name-object or A-B) relations via either conditional discrimination training (in Experiment 1), or a respondent-type training procedure (in Experiment 2). This training was followed by tests for derived symmetry-like relations (B-A relations). If participants failed these tests, MET was implemented in which B-A relations were explicitly taught with novel stimulus sets. Following MET, probes were conducted again with the original training set and MET continued with novel stimuli until participants met criterion for the emergence of derived B-A relations (Experiment 1) or derived A-B and B-A relations (Experiment 2 only); or after they had been exposed to a total of three different MET sets, whichever occurred first. Results from Experiment 1 indicated failed tests for B-A relations following direct training in A-B relations, and marked improvements in derived B-A relations following MET across all participants. Results from Experiment 2 demonstrated the emergence of derived A-B relations following a respondent-type training procedure (denoted as receptive symmetry training) and the emergence of B-A relations following MET across all participants. These results lend support for the use of a respondent-type training procedure to establish derived relations in typically developing children. Results and implication are discussed in light of a behavior analytic account of language development in general, and for second language acquisition specifically.

Many children diagnosed with autism spectrum disorder have difficulties identifying and labeling feelings and emotions of others. Three adolescent males all diagnosed with autism spectrum disorder participated in this study. In a multi-element design, the participants were trained to tact private events of others in context using novel video-based scenarios. Two of the three participants were able to increase and maintain their responding for all trained and derived relations, as well as a transformation of stimulus function task. The third participant required multiple exemplar training to novel stimuli to increase his responding for all the video-based scenarios. The results of the study support the utility of relational training for teaching children with autism to identify private events of others in context.

The purpose of this study was to evaluate the effectiveness of multiple exemplar training in establishing comparative relations of faster and slower as a generalized operant in children with autism. The current study employed a multiple baseline across participants to evaluate the effectiveness of the multiple exemplar training procedures. The participants used in the study were 3 children diagnosed with Autism. The baseline phase consisted of 4-trial blocks; this was done by presenting each stimulus once for AB- relation. Participants were trained comparative relations of faster and slower with one set of stimuli, and test probes were conducted with a novel, untrained set of stimuli to evaluate the degree to which participants responding generalized. This sequence was repeated for the duration of baseline. Y/Z relation remained at low levels during baseline and was only used to see if the skill generalized. The results suggest the procedures were effective for teaching comparative relations for 3 of the 3 participants. Additionally, 3 of the participants responded correctly to the test relations, supporting the use of the procedures in prompting generalized relational responding. An implication is that we can utilize these procedures to establish relational responding in children with autism.