The information-conveying ability of humorous texts : an analysis of "semantic objects"

Abstract

This thesis examines how humorous texts may convey reliable information, despite the obvious problem that hearers regularly accept that joke tellers are not held to the truthvalue of what they utter. Only Zhao (1988) has claimed that jokes may convey objective, factual information. This thesis constitutes the fIrst attempt to answer the question how jokes and longer humorous texts may do this. Essentially the argumeht is this: when presented with a text of any kind (i.e., serious or non-serious), hearers compare their own background assumptions with the assumptions of the text. These "assumptions" are named herein, as a superordinate term, "semantic objects," the sum of which make up the "textual world" (TW). In serious communication, hearers assume that the text's semantic objects match their own background assumptions about what they take to obtain in the "real world" (RW); through Gricean implicature, new, reliable information may be learned when such texts do not match, but are compatible with, their background assumptions. In joke texts, however, hearers must sift through unreliable semantic objects to determine which should be accepted as reliable. Using a taxonomy of these semantic objects to create a version of the 1W for four sample jokes, it turns out that in joke texts, the vast majority of semantic objects are true for the RW. Furthermore, those semantic objects that are unreliable for the RW are either blatantly false or are directly associated with the joke's punch line, such that a hearer may determine the reliability of each semantic object (and thus learn new, reliable information) using two heuristics: 1) Is the semantic object incompatible with the RW as you believe it to be?, and 2) Is the semantic object logically necessary for the given punch line? If the answer to both of these questions is no, the semantic object may be accepted as reliable.