An Empirical Study About The Chinese Irony Comprehension From The Perspective Of The Graded Salience Hypothesis

As a ubiquitous form of language, frequently used in everyday discourse, irony has been extensively discussed from different perspectives. Ever since the 1970s, irony finally became a focus approached by psychologists, linguists, and philosophers, with different approaches to irony being proposed. Most of the Chinese scholars studied irony from rhetoric and pragmatic perspectives, and recently some researches studied on the mechanism of irony comprehension in a psychological way. Revolving around the problem of irony comprehension, the thesis conducts an empirical study about irony comprehension from the perspective of the graded salience hypothesis. The spring up manifold views on the nature and comprehension of irony can be sorted out as three modes: the direct access view, the modular model and the graded salience hypothesis.The thesis confirms the validity of the graded salience hypothesis by comparing the reading times of target sentences and analyzing the mistakes types from the subjects answering comprehension questions about ironic target sentences. The experiment consists of two parts, each including 20 ironies and 20 literal utterances. The major findings are: (a) ironies obviously took longer to read than their literal counterparts, which is consistent with the prediction of the graded salience hypothesis as well as the modular view while contradicts the assumption of the direct access model that ironic and nonironic utterances including literal ones should take equal time to process. (b) According to Giora, comparisons were made between the mean reading times of familiar and unfamiliar ironies. The finding that all the familiar ironies took less time to read than unfamiliar ones is also consistent with the graded salience hypothesis. (c) as predicted by the graded salience hypothesis, familiar ironies took no longer to read than their literal counterparts, which is contrary to the prediction of the modular view that ironies should be processed longer than literal sentences. (d) unfamiliar ironies were found to take much more time to read than their literal counterparts. In addition, it was discovered that the role of the literal meaning led to the misinterpretation of ironies through the analyses of mistake types. Statistics show that out of the 51 mistakes in answering multiple-choice questions, 25 mistakes were caused by the misinterpretation of ironies as literal ones, and among the 28 mistakes made in answering the yes-or-no questions, 11 mistakes were derived from the misinterpretation of ironies as literal ones. Based on the above results, the thesis draw such a conclusion that the literal meaning is involved in irony comprehension whether in the initial phase or the integration phase, which is in conflict with the direct access view that only the contextually compatible meaning exists in both phases. Moreover, since the standard pragmatic model holds that the contextually incompatible meaning, that is, the literal meaning in a ironically biased context will be suppressed and replaced by the ironic meaning, it can not explain the above results why the existence of literal meanings led to wrong interpretation of ironies. However, according to the retention hypothesis, the literal meaning may either be retained or suppressed. The literal meaning may be suppressed only with the ironic meaning be retained for further processes, or both the literal and the ironic meanings may be maintained in the integration phase depending on their function in constructing the intended meaning. These can well account for why some mistakes were caused by the misinterpretation of ironies as literal utterances while some were not.In addition, the thesis recorded the answering times about ironies and literal utterances for reference and found similar results from the comparisons between the mean answering times about ironies and those of literal utterances. Later in discussions, the thesis discusses on the familiarity of irony and the testing methods to distinguish familiar ironies from unfamiliar ones.