Negative patterning and biconditional discriminations in human causal reasoning: a second look

Harris and Livesey (2008) reported that patterning discriminations were easier to learn than biconditional discriminations in a human causal reasoning task, consistent with the predictions of the attentional buffer model of Harris (2006) and inconsistent with the predictions of configural cue models. However, their evaluation of patterning and biconditional tasks also failed to find a positive patterning advantage, which has been a hallmark prediction of many configural cue theories. This failure raises the question of whether Harris and Livesey's method was appropriate for a general evaluation of the role of configural cues in learning complex discriminations. Using a design that does produce a positive patterning advantage, the study reported here shows negative patterning discriminations to be at least as difficult to learn as biconditional discriminations, consistent with configural cue models.