An information theoretic approach to veridical hallucination

Abstract

David Lewis, in "Veridical Hallucination and Prosthetic Vision", outlines his views on seeing. He discusses, by way of several examples, unusual visual conditions and gives explanations of why one does or does not see in those conditions. However, it is not always clear exactly how Lewis' views apply to unusual cases. He also admits that he has made mistakes in applying his criteria to examples, in the Postscript to the original article. However, I think Lewis' ideas are worthwhile and would like to expound upon them. In what follows, I hope to provide clearer criteria that are compatible with Lewis' views, and show how such criteria do or do not apply to unusual circumstances. The criteria I will use in place of Lewis derive from a branch of signal theory, called Information Theory. Information Theory is a formal calculus for quantifying and computing the information content of a source or a signal carrying information about a source. It is an attempt to formalize an intuitive notion of information that we all work with. The goal will be to look for discrepancies between the information theoretic criteria and Lewis' conclusions, so cases where there is substantial agreement between Lewis and the information theoretic criteria will be only briefly glossed. Clarification of both views can be obtained by seeing how and why they differ and which view is plausibly correct about the case.