The search for agents in a causal gap

This study used a reversal experiment within a participant's frame of reference as a tool for studying human agency. Twenty-seven participants were given a sentence reading task, asked to formulate a rule for choosing which sentence to read, and instructed to fbollow and break the rule in an FBFB or BFBF sequence. Despite the difficulty they had in formulating an unambiguous rule, 23 participants both broke and followed the rule, and 20 broke it with a mutually exclusive opposite. There was evidence that a participant's genetic heritage, learning history, and rule, the experimental task, and the instructions were in no case sufficient to control the sentences he or she read. There was evidence of two causal gaps, one before and one after a participant's choice, and positive evidence of participants' systematic control over the sentences read in reversal experiments.