Propositional Reasoning by Model?

Abstract

Two theories of propositional deductive reasoning are considered: Johnson-Laird's mental models and Braine's mental logic. The model theory is said to account for practically all of the known phenomena of deductive propositional reasoning, offer a general theory of conditionals, account for the most important aspects of Braine's theory, and predict new phenomena that rule theories cannot explain. I argue that (a) the model theory is flawed in a way that is difficult to overcome, (b) conditionals are seriously misrepresented, (c) the algorithms proposed to implement it either allow invalid inferences or are psychologically useless, (d) Braine's theory accounts for all of the new phenomena worth considering, and (e) the model theory can predict Braine's results only at the cost of self-refutation. I conclude that the mental model theory of propositional reasoning offers no reason to reject the program of mental logic.