The nature of irony: Toward a computational model of irony

Abstract In reasoning about irony, we propose the following question as a criterion for judging whether any proposed theory of irony is viable: Could that theory be used as the basis for a computer program that reasons about irony? That is, the theory must be a computational theory in the sense that it gives a description of the following three tasks that is explicit enough that a computer could be programmed to perform them: Distinguish irony from nonirony; describe why a situation is ironic or not; and generate descriptions of ironic situations. For the purposes of this paper we distinguish between: (1) an ironic situation, and (2) an ironic statement. The main goal of the paper is to identify and characterize the various types of ironic situations, not to identify and characterize types of ironic statements. An ironic statement is an utterance of a speaker which refers to certain aspects of an ironic situation to make a point, i.e. to achieve certain communicative goals held by the speaker. The treatment of ironic statements would, therefore, identify the aspects of ironic situations that are pointed out according to the goals of the speaker. After presenting some of the results of our work on the three tasks outlined above, we touch on some more general issues we have encountered in our work on irony, namely the appreciation of irony, the relationship between irony and sarcasm, the pragmatics of irony, and the relationship of studies of irony to studies of humor.