Association for Behavior Analysis International

The roots of relational frame theory (RFT) can be traced back to an early conference paper on rule-governed behaviour in 1984. The seminal book-length treatment of RFT is now itself 20 years old. In that time the account has introduced many new terms, concepts and methods that would be unfamiliar to traditional behavior analysis. The current symposium presents four papers that involve critically reappraising this (RFT) work in an effort to determine its value, while also identifying ways in which to move forward. We argue that progress will likely involve being genuinely open to identifying potential weaknesses in analytic strategies, limitations in key concepts, and in a willingness to engage genuinely with alternative approaches to the study of human language and cognition within behavior analysis. Specifically, the four papers will consider (1) recent developments in the analysis of data from an RFT methodology, known as the implicit relational assessment procedure (IRAP); (2) the limited utility of the concepts of pliance, tracking and augmenting within RFT; (3) the use of a new framework in applied behavior analyses of language and cognition; and (4) the potential benefits of drawing on both RFT and Verbal Behavior Development Theory (VBDT) in the experimental analysis of human language and cognition.

A basic background in behaviour analysis is assumed.

The implicit relational assessment procedure (IRAP) was originally conceptualised as a method for assessing the strength of natural verbal relations as conceptualised within RFT. The method itself involved combining the relational evaluation procedure (REP), an RFT-based methodology, and an instrument developed within mainstream social cognitive psychology known as the implicit association test (IAT). The latter was designed to measure the strength of associations in memory and was therefore clearly a tool based on the assumptions of cognitive psychology. In combining the REP and IAT into the IRAP, an increasingly vigorous program of research emerged in which the IRAP was used as an instrument for assessing implicit cognition rather than the strength of natural verbal relations. Although the research program was not without value, in retrospect it was always going to be limited and the IRAP as such would fail to deliver on its original purpose. The current paper will review this retrospective narrative on the history of the IRAP and consider some of the more recent research that has focused on using it as a measure of the dynamics of relational framing itself.

Pliance, tracking and augmenting were defined as functionally distinct categories of rule-governed behavior in 1982. Since this time, however, the terms have rarely been used as the basis for conducting systematic experimental-analytic research, despite their theoretical centrality to the study of rule-governed behavior. 40 years later, it seems useful to reflect upon their place within the literature on the experimental analysis of human behavior, and relational frame theory in particular. In the current talk we evaluate their place within the literature and argue that they should be considered middle-level terms, which lack the relative precision of technical terms within the literature on relational frame theory (RFT). We explore the potential utility of conceptualizing rules as involving increasingly complex derived relational networks and focusing on various dimensions that impact such networks. Finally, we briefly consider a new program of research that has begun to take this approach in the context of up-dating RFT.

Demonstrating awareness of oneself and the states of others is argued to involve a highly complex behavior referred to as perspective taking. Before abstracting or inferring another person's perspective, one depends on a sufficient previously trained relational repertoire. The objective of the current study was to draw on the fundamental units of AARR, specifically with respect to deictic repertoires, using the MDML framework and explore a set of tasks to evaluate and train perspective-taking (PT). A set of non-arbitrary and arbitrary tasks were used to investigate relational repertoires at four levels of relational development (1-mutual entailment, 2-relational framing, 3-relational networking, 4-relating relations) for various generalised patterns of responding (coordination, difference, opposition, comparison, and hierarchy). Data from two children of similar developmental age (one with typical development and one with autism) were collected. The typical development child presented the expected level in abstract relations and showed success in the PT test. The child with autism initially failed the PT test but after an MDML-based intervention showed development in his relational repertoire and finally succeeded in the PT test. This study adds potentially valuable information about the minimal units required for deictic relational responding.

Conceptual developments in RFT, which have provided a general framework (Hyper Dimensional Multi-Level framework) and a dynamical unit of analysis (Relating, Orienting, and Evoking, ROE) have served to highlight clear points of contact and overlap between the analysis of naming and different levels and dimensions of derived relating, in general. Previous studies on naming have presented the object and its name simultaneously during both training and testing, and thus the training component may establish a transformation of function (ToF) directly between the object and the name. The aim of the current study was to test the emergence of speaker naming and entailed ToF with a non-simultaneous presentation technique and evaluate the effectiveness of Multiple Exemplar Training (MET) if deficits are observed. Five typically-developing toddlers participated in the study, and initially, none of the participants exhibited correct naming responses. Three participants received MET, which led to improvements in speaker naming for all. Of these, one needed additional training with simultaneous stimulus presentation trials. The remaining two participants were tested repeatedly, without MET, and did not show any consistent improvements in naming. The applications of the HDML framework to assess the strength of the levels/dimensions of naming are discussed.