Misdirection can be defined as the deliberate action of indulging some kind of misrepresentation in the perception of an agent, or of a group of agents. Such misrepresentations can be of a different nature: amongst other forms, they can be verbal (as the result of linguistic communication) but they can also be visual (as the result of observation). Various types of verbal misdirection have been investigated formally, regarding lies and strategies of bluffing for instance, based on existing analyses of those notions in the epistemological literature. However, no much attention has been paid to visual misdirection in particular. In this paper, we introduce a new framework based on dynamic epistemic logic, the aim of which is to model verbal misdirection that affects agents' beliefs, as well as visual misdirection on agents' observations of their environment. The logic contains two modalities for capturing verbal and visual perception, and a dynamic modality for executing actions either on the agent's beliefs or on her observations. We also provide a sound and complete axiom system for the logic, and illustrate the usefulness of the framework by formalizing a classical scenario of visual misdirection: the French Drop magic trick.
Presented at the Fifth PLM Workshop on Delusion in Language and Mind, in Amsterdam, October 23---24, 2020