An Experimental Study Of Scalar Implicature

Through experimental studies on Chinese quantifier yixie, this research aims toidentify the nature and characteristics of its scalar implicature during utteranceinterpretation, as well as to address the theoretical issues including thesemantics-pragmatics interface debates pertaining to post-Gricean pragmatics. In recentyears, as a type of generalized conversational implicature, scalar implicature has receivedincreased attention due to its two-sided nature. Among all the different theoretical modelsproposed in literature, default theory and relevance theory are two typical competingapproaches. Essentially, these models are all introspective philosophical theories requiringempirical support. Due to the limitation of traditional qualitative analysis measures, themore convincing quantitative investigation methods, such as psychology experiment, aresuggested, which utilize the objective modern statistics tools. Moreover, the existingconclusions regarding the nature of scalar implicature are not consistent across all thestudies. Although accumulated experimental conclusions come out with new ideas orsuggestions, still no results of scalar implicature based on Chinese data have been putforward at current stage. Under such circumstance, there is no doubt that my original studyin this dissertation will have significant theoretical implications for deeper understandingthe nature of scalar implicature and the relationship between semantics and pragmatics.In this psychology experimental study, a with-in subject design was adopted. Theparticipants were native Chinese speakers. Each experiment is consisted of two parts, i.e. atruth value judgment task and an open questionnaire task. In the first part, the participantswere asked to judge whether a sentence containing yixie matched with a given quantityscenario in a certain way. While in the second one, the participants were required to provideanswers to assigned questions. The experiment data, including the rate of positive answersto stimulus questions, the reaction times of making judgments, and the answers to the openquestions were analyzed by performing a series of mixed linear models and repeatedmeasures analysis in general linear model.The primary conclusions of this study are as follows:1. The semantic content of Chinese quantifier yixie denotes indefinite quantity, whoselower bound is somehow greater than “zero”, and its upper bound can, to some extent,include “all”. Within this range, however, the kernel meaning of yixie mostly denotessmaller proportional quantity while its upper bound “all” is peripheral. 2. The scalar implicatures of yixie can be largely generated in upper bound contexts,but with longer reaction time. It suggests that, from the perspective of processing speed,scalar implicatures are not dufalt.3. When the referent of yixie is getting closer to the quantity range of its kernelmeaning, a lower proportion of scalar implicatures are generated, and vice versa. Inaddition, the nature of scalar implicatures which is influenced by certain modifiers of scalaritems could vary when the semantic content of scalar expressions has different degree ofuncertainty. With a wider range of possibilities in semantic content, more pragmatic factorscould intrude into the calculation of scalar implicatures, indicating that the nature of them isbiased towards nonce pragmatic inferences.4. The nature of scalar implicatures demonstrates that there is strong interactionbetween semantics and pragmatics, and the semantic content of an utterance and itsenrichment present a continuum without a clear borderline.5. As the result, this study is prone to providing support to relevance theory, whereaschallenging the explanations of scalar implicatures in default theory. It can also be taken asexperimental evidence against other default models for implicatures.