From comedy shows to everyday conversation, humor plays an important part in human interaction, whether it’s used to expedite courtship or enhance social bonding. And yet, if asked why we find certain things funny many of us will often struggle to express the reason. For something so prevalent in our daily lives, it is certainly bizarre to think that we don’t know what particularly it is that makes us laugh. For many, humor exists more as intuition than logic. For example, take the classic, “Why did the chicken cross the road?” joke. Many find this joke is funny, albeit trite and overused. But why do we find this joke funny? Or inversely, why don’t we?

The field of computational humor seeks to address this deceptively complicated question by introducing ways to break down humor systematically. In 1991, linguists Victor Raskin and Salvatore Attardo proposed the General Theory of Verbal Humor (GTVH) [1], which breaks humor into six Knowledge Resources (KRs):

  1. Script Opposition (SO): proposes that each joke can be broken into two “scripts,” with opposing, yet compatible semantics
  2. Logical Mechanism (LM): refers to the mechanism that connects the two scripts of any given joke
  3. Situation (SI): includes the context of the elements in the joke i.e. activities, props needed to make the joke coherent
  4. Target (TA): refers to the object of the joke
  5. Narrative Strategy (NS): The “theme” of the joke e.g. riddle, narrative, question answer, etc.
  6. Language (LA): concerned with the verbalization of the joke

These KRs can be used to model jokes and analyze the degree of similarity between them. While GTVH aims to categorize jokes of varying types, it has a major drawback – many of the aforementioned KRs serve more as a heuristic than a framework for joke categorization.

While other models exist for humor interpretation e.g. [2], [3], a high performance computational framework for humor interpretation is largely lacking. It is our belief that a logical structure, or even a mathematical formalization exists behind the abstract concept of humor. Once this structure is found we will be able to exploit computational resources to analyze and even construct humor. We may finally find out why the chicken crossed the road, and more importantly, why it was funny.


1. Salvatore Attardo (1994). Linguistic Theories of Humor, pp. 223 – 226. Mouton de Gruyter: Berlin, New York.

2. I.M.Suslov, Computer Model of “a Sense of Humour”. I. General Algorithm. Biofizika SSSR 37, 318 (1992) [Biophysics 37, 242 (1992)];

3.  P. Marteinson (2006) On the Problem of the Comic, Legas Press, Ottawa, ISBN 978-1-894508-91-9