Inferring Ignorance from the Locality of Visual Perception

This paper presents a logical theory that supports high-level reasoning about knowledge and perception. We construct a formal language in which perception can be described. Using this language, we state some fundamental axioms, and we show that these are sufficient to justify some elementary but interesting inferences about perception. In particular, our axioms make it possible in some cases to infer that an agent does not know about a particular event, because he has had no way to find out about it.