Evidence for deductive reasoning in blocking of causal judgments

We have recently demonstrated that pre-training of additivity (the outcome of two causal cues is larger than one causal cue) greatly enhances blocking. This manipulation could work by removing a ceiling effect on the outcome, as proposed by Cheng (1997). Alternatively, it could remove the logical ambiguity associated with blocking under non-additive conditions, thus permitting blocking as a deductive inference. We used a counterintuitive combination rule-subtractivity rather than additivity-to discriminate between these two accounts. Prior to a backward blocking causal judgment procedure (AB+, A+), we trained participants that a compound of one causal and one non-causal cue leads to the outcome, but that a compound of two causal cues leads to no outcome. This design allows the logical deduction that cue B is non-causal (blocking), but retains the ceiling effect on outcome magnitude. Blocking was strong following both subtractivity and additivity training, supporting the deductive reasoning account.