Knowledge-based judgements of causality: Contiguity, congruity, and direction of the causal arrow

Abstract

<p>Under what circumstances does knowledge of causal asymmetry and temporal delay influence causal judgements? We begin a series of thirteen experiments. by providing evidence that both high-level (causal reasoning) processes, and low-level (associative) processes influence causal assessment depending on what is asked about the events. Specifically, participants were more sensitive to causal structure in their ratings than in their prediction responses, on earlier rather than later trials, and when asked to provide an integrative causal rating. Emphasising the direction and nature of the causal relationship and the wording of the test question had no influence on participants' sensitivity to causal asymmetry. Next, we provide evidence that participants' ratings track conditional rather than uncondtional contingencies as predicted by the conditional P account as well as the Rescorla-Wagner model at asymptote. Our results suggest that participants tend to rate the influence of each cause conditional on the absence of the other cause. This tendency is not reflected by the Rescorla-Wagner model. Finally, we examine the role of temporal contiguity on judgments of contingency using a human analogue of the Pavlovian task. Our results suggest that knowledge of temporal delay modulates causal judgements.</p>