Aristotle indicates four types of inferences whose conclusion does not necessarily follows its premises. In this paper, I review them from the mental models theory and try to show that, if this last framework is correct, people should note without a lot of effort that their conclusions are not necessary. I mainly base my analysis on the distinction between canonical and noncanonical models provided by Khemlani, Lotstein, Trafton, and Johnson-Laird (2015).
