Abstract
Everyday conditional reasoning commonly occurs in a sequence, such as:
… If she leaves home early today then she will get to work quickly. If she gets to work quickly then she will get extra things done …
Experimental studies of conditional reasoning assume that people make the same amount of inferences from sequences as they do from simple arguments that contain just a single conditional premise. In a series of four experiments this assumption is shown to be wrong. Our first experiment showed that people make more inferences from transitive sequences, and fewer inferences from certain atransitive sequences, than from corresponding simple arguments. We established in the second experiment that these differences arise because people can construct a single representation of a sequence. The third experiment replicated the result that people make fewer inferences from certain atransitive sequences. In the final experiment we found that people make more inferences from transitive sequences, regardless of the presence of negatives. We argue that the inferences people make from common everyday sequences depends on their representation of the state of affairs that sequences describe.
