A Formalism and an Algorithm for Computing Pragmatic Inferences and Detecting Infelicities

Since Austin introduced the term infelicity, the linguistic literature has been ooded with its use. Today, not only performatives that fail are considered infelicitous but also utterances that are syntactically, semantically, or pragmatically ill-formed. However, no formal or computational explanation has been given for infelicity. This thesis provides one for those infelicities that occur when a pragmatic inference is cancelled. We exploit a well-known di erence between pragmatic and semantic information: since implicatures and presuppositions, i.e., the carriers of pragmatic information, are not specifically uttered, pragmatic inferences are defeasible, while most of semantic inferences are indefeasible. Our contribution assumes the existence of a ner grained taxonomy with respect to pragmatic inferences. It is shown that if one wants to account for the natural language expressiveness, she should distinguish between pragmatic inferences that are felicitous to defeat and pragmatic inferences that are infelicitously defeasible. Thus, it is shown that one should consider at least three types of information: indefeasible, felicitously defeasible, and infelicitously defeasible. The cancellation of the last of these determines the pragmatic infelicities. A new formalism has been devised to accommodate the three levels of information, called strati ed logic. Within it, we are able to express formally notions such as utterance u presupposes p or utterance u is infelicitous. Special attention is paid to the implications that our work has in solving some well-known existential philosophical puzzles. The formalism yields an algorithm for computing interpretations for utterances, for determining their associated presuppositions, and for signalling infelicities. Its implementation is a Lisp program that takes as input a set of strati ed formulas that constitute the necessary semantic and pragmatic knowledge and the logical translation of an utterance or set of utterances and that computes a set of optimistic interpretations for the given utterances. The program computes for each set of utterances the associated presuppositions and signals when an infelicitous sentence has been uttered. ii Acknowledgments First, I thank my supervisor, Graeme Hirst, for his competent guidance, patience, and humor in explaining and re-explaining so many things; and con dence that I can delineate and solve problems on my own. I also thank him for the way he taught me to weight and glue words into sentences, sentences into paragraphs, to put down ideas in a way that other people can understand them. The aws in this thesis show that I still have a long way to go until I come to know well everything he taught me. I want to thank Hector Levesque, my second reader, for the helpful comments he gave me. I thank Hector Levesque and Ray Reiter for teaching me how to tackle a logical problem and how to nd formal explanations for the phenomena around us. I thank Je Siskind for reminding me that the ultimate proof is a program that does something. I thank my colleagues in the natural language, cognitive robotics, and knowledge representation groups who shared with me their questions, problems, opinions, and thoughts. I am grateful for the nancial support I have received from the University of Toronto and from the Natural Sciences and Engineering Research Council of Canada. Because a thesis is not only words, programs, or formulas I thank Ed Klajman for being the friend with whom I was able to get into a world free of syntactic, semantic, or pragmatic constraints. I thank my parents for everything they have taught me, for the understanding and trust they have shown me every day since I have started this work. But most of all, I want to thank my wife, Oana, for her endless love, support, and strength in believing that one day, we will be again together, in a better world, every single minute. iii