Contextual control over equivalence and nonequivalence explains apparent arbitrary applicable relational responding in accordance with sameness and opposition

Abstract

We evaluated whether contextual control over equivalence and nonequivalence (i.e., selecting comparisons equivalent to the samples in the presence of a contextual cue, and excluding the selection of comparisons equivalent to the samples in the presence of another contextual cue) can account for apparent arbitrarily applicable relational responding (AARR) in accordance with the frames of sameness and opposition, as defined in relational frame theory (RFT). Three college students were trained to maintain previously established conditional discriminations in the presence of a contextual cue X1, and to reverse them in the presence of another contextual cue X2 (e.g., X1-A1B1, X1-A2B2, X2-A1B2, X2-A2B1). Subsequent tests demonstrated that X1 and X2 functioned as cues for equivalence and nonequivalence. Later on, X1 and X2 were demonstrated to be functionally equivalent to supposed contextual cues for the frames of sameness and opposition employed in RFT studies (i.e., SAME and OPPOSITE cues), in tests for arbitrary and nonarbitrary derived stimulus relations. The functional equivalence of X2 and OPPOSITE suggests that OPPOSITE worked as a cue for nonequivalence. Thus, the results in RFT studies with SAME and OPPOSITE can be explained just by contextual control over equivalence and nonequivalence. Therefore, the explanation that they actually demonstrated AARR in accordance with the frames of sameness and opposition can be questioned and replaced by a more parsimonious explanation, based on a few simple learning principles. We discuss the implications of this conclusion for the debate among competing theories about the origin of stimulus equivalence and other derived stimulus-stimulus relations.

The evidence base for the opposition frame: A reply to Stewart, Dymond, & Roche (2019)

“…These outcomes suggest X1 was a cue for equivalence-or sameness in RFT terms-and X2 was a cue for nonequivalence (i.e., for selecting comparisons from a different equivalence class as the samples). Therefore, they are compatible with Alonso-Alvarez and Perez-Gonzalez's (2017 thesis. X1 and X2 were then presented in a series of training and test trials analogous 2 to SAME-A1 [B1, B2, N1], SAME-A1 [C1, C2, N2], SAME-A2 [B2, B1, N1], and OPPOSITE-A1 [C2, C1, N2].…”

“…The hypothesis that the participants selected C2 by the exclusion of C1 and N2 still holds, however. Those derived relations can be explained by contextual control over equivalence, nonequivalence, and exclusion (Alonso-Alvarez & Perez-Gonzalez, 2017. However, this hypothesis can explain results that the RFT hypothesis cannot explain.…”

“…For instance, they trained SAME-short line [ , N2]. Thus, the participants selected C2 by the exclusion of both C1 and N2 (e.g., Alonso-Alvarez & Perez-Gonzalez, 2017. In addition, N2 was never reinforced during training.…”

“…In this experiment, pretraining with nonarbitrary relations was replaced by pretraining with arbitrary relations. Stewart et al (2019) argue that "Alonso-Alvarez and Perez-Gonzalez's (2017 thesis is incompatible with data generated with research protocols that have not employed their particular methodology" (p. 4). These outcomes suggest X1 was a cue for equivalence-or sameness in RFT terms-and X2 was a cue for nonequivalence (i.e., for selecting comparisons from a different equivalence class as the samples).…”

“…Stewart et al (2019) argue that "Alonso-Alvarez and Perez-Gonzalez's (2017 thesis is incompatible with data generated with research protocols that have not employed their particular methodology" (p. 4). Would the results be the same with the X1 and X2 cues used by Alonso-Alvarez and Perez-Gonzalez (2017)? Specifically, they mention three tests designated by Steele and Hayes as U, V, and W tests (p. 546).…”