It is, of course, no more possible to define inferential thinking than to define life. Such matters can be elucidated only by a theory. It is useful, however, to give a working definition that delimits the domain of study. An inference is a process of thought that leads from one set of propositions to another.
-- Johnson-Laird, 1983, pp. 23-24
Premises do not lead to conclusions in themselves, independent of reasoners--they lead to conclusions only insofar as their meaning is grasped, and their implications are seen, by a human being situated in the world.
-- Johnson, 1987, p. 57
A variety of thinking tasks studied by cognitive scientists are of special interest because they represent the direct translation of relatively straightforward normative models into laboratory paradigms. These tasks include reasoning with propositional arguments or quantified syllogisms, simple probability judgments, revision of beliefs construed as subjective probabilities, and decision making with clear alternatives. Perhaps because the normative status of logic, probability theory, and utility theory has seemed clear, these tasks have often been taken as prototypes of rational thinking, presumably the highest of the higher mental processes.
Despite (or perhaps because of) this special status, research on these thinking tasks occupies a peculiar position in cognitive psychology. The literatures on these tasks are relatively insular, and they seem to have