Formal Humor Logic Beyond Second-Most Plausible Reasoning

Humor employs an essential false logic which masks the incongruity of two central meanings that are brought into overlap. Formalizing this false logic—if it exists, exists intersubjectively, and is indeed essential for humor—to a degree that is sufficient for computational detection and generation of humor has been a vexing problem for computational humor research. This paper will outline several such logics, in addition to the default of reasoning in a way that is one degree more implausibly than the most common-sense logic that can connect two meanings. The results are not least influenced by a pilot study asking participants to explain different types of jokes.