Quiñones, J. (2008). Relational coherence and transformation of function in ambiguous and unambiguous relational networks. Unpublished Doctoral Dissertation, University of Nevada, Reno.
Two experiments were conducted to examine the relational coherence and transformation of function in two relational networks, one ambiguous and the other unambiguous. In both experiments, the participants learned two three- node networks (A1 LESS THAN B1, A1 GREATER THAN C1 and A2 GREATER THAN B2, C2 LESS THAN A2) using the precursor to the Relational Evaluation Procedure (pREP) (Cullinan, Barnes, & Smeets, 1998; Cullinan, Barnes-Holmes, & Smeets, 2000, 2001). Participants were presented with test trials to examine if they classified the combinatorial relations (B1 ↔ C1 and B2 ↔ C2) as GREATER THAN or LESS THAN and as SAME or DIFFERENT. Although the combinatorial relation of Network 1 is derivable (B1 GREATER THAN C1 and thus also B1 DIFFERENT C1), this same relation of Network 2 is ambiguous and therefore not derivable. Network 2 participants disambiguated the BC combinatorial relations using SAME or DIFFERENT in a coherent fashion, but in a way that differed participant to participant. Those who classified the relation as SAME did not consistently choose B2 as GREATER THAN or LESS THAN C2. Those who classified it as DIFFERENT did consistently choose B2 as either GREATER THAN or LESS THAN C2. In examining transformation of function, Experiment 1 included a respondent conditioning phase following the pREP, and pre and post tests for transformation of function were conducted using the Extrinsic Affective Simon Task (EAST) (De Houwer, 2003; De Houwer et al., 2001; De Houwer & Eelen, 1998). Unfortunately, the EAST analysis failed to be an effective measure of emotional valence (see De Houwer & De Bruycker, 2007 for similar findings) and therefore, an assessment of the functions transfers by the derived relations was not possible. However, the relational coherence was further explored in Experiment 2 by introducing a SAME versus DIFFERENT bias for ambiguous combinatorial relations using multiple exemplar training. When participants had the relevant history, the ambiguous network became unambiguous in a fashion that fit with the results in Experiment 1, with all the nodes connected based on the new relations of SAME or DIFFERENT. The findings provide support for the flexible nature of relational frames and the importance of history in establishing these frames.