Probe Into The Reanalysis Of English Garden-path Sentences

This thesis is an empirical study which is aimed to investigate Chinese English learners'comprehension of English garden path sentences. As temporarily ambiguous sentences, garden path sentences have been arousing enormous research interests of psycholinguists and linguists ever since they were recognized in 1970s. Through the examination of how parsers are first lured down to the garden path and then spend efforts to repair the initially incorrect interpretation and finally construct a globally licensed representation to their satisfaction, researchers try to gain an insight into the operation of human language processing system.Studies on garden path sentences by far can be categorized into three phrases according to their research emphases. Studies from 1970s to 1980s were conducted on the basis of psychology and psycholinguistics. Researchers advanced various sentence parsing models in order to offer an explanation for this phenomenon. Later on, ever since the end of 1980s, researchers began to explore the production of garden path effect through the grammatical analysis of the garden path structures themselves. It was assumed at that time that garden path effect was caused because parsers violated certain syntactic principle. Recently, more research attention has been switched to the investigation of the ultimate representations of garden path sentences. And researchers are more concerned whether comprehenders are able to derive globally licensed understanding of garden path sentences.Since recent studies have somewhat indicated that under some circumstance English natives may not be successful in understanding garden path sentences, we are quite interested in the extent to which Chinese English learners extract correct meanings from such sentences. In fact, we assume that Chinese English learners have the tendency to derive incomplete interpretations from garden path sentences, and this basic assumption can be further divided into the following three specific ones. We hypothesize that first, Chinese English learners tend to have partial analysis of garden path sentences so that they derive incomplete interpretations. Second, thematic roles assigned in the initial analysis cause the incorrect interpretation to linger. Third, semantic plausibility play an important part in affecting the chances that parsers achieve a full and complete reanalysis of garden path sentences. Apart from the assumptions above, we further assume that Chinese English learners'comprehension of garden path sentences bears positive correlation with their English language proficiency. In order to examine the validity of our assumptions, we designed a comprehension test to investigate how well Chinese English learners understand English garden path sentences.According to the data collected from the comprehension test, we had some important findings about the nature of Chinese English learners'ultimate representations of garden path sentences. Owing to the fact that there was a mean error percentage of 63.14% of responses to questions probing their understanding of the subordinate clause in the plausible garden path condition and another mean error percentage of 46.89% in the implausible garden path condition, in spite of the fact that subjects were doing perfectly well in giving responses to questions tapping into the main clause, with a percentage of 12.48 % incorrect responses in the implausible garden path condition and another even smaller proportion of incorrect responses of 9.87% were given in the plausible garden path condition, we found that first Chinese English learners had much difficulty in processing English garden path sentences, and instead, they were very likely to have incomplete representations from such sentences. In addition, results from the Paired-samples T-tests indicated that there was a significant difference between subjects'understanding of the subordinate clause and the main clause in both conditions, with the t =21.56, p (Sig=.000) <0.05 in the plausible condition and the t =15.88, p (Sig=.000) <0.05. In addition to the data that the percentage of incorrect"Yes"responses to questions about the subordinate clause in the plausible garden path condition was much larger (63.14%) than in the non-garden path condition (13.53%), and subjects had significantly different performance in their understanding of the subordinate clause in the garden path condition and non-garden path condition, with the t value of 26.47, p (Sig =.000) <0.05, there existed a higher percentage of incorrect"Yes"responses in the implausible condition (46.89%) than in the non-garden path control condition (13.53%),and again subjects had significantly different performance in their understanding of the subordinate clause in the garden path condition and non-garden path condition, with the t value of 10.85, p (Sig =.000) <0.05. We found then, once led down to the garden path, Chinese English learners had much trouble in erasing the misinterpretation resulting from the initial thematic role assignment.Moreover, subjects gave a comparatively larger proportion of incorrect"Yes"responses to questions about the subordinate clause in the former condition (63.14%) than in the latter condition (46.89%), and subjects performed significantly differently in giving responses to questions about the subordinate clause of plausible and implausible garden path sentences, with the t value of 7.71, p (Sig =.000) <0.05. We found that semantic plausibility had obvious influence on Chinese English learners'decisions to discard the misinterpretation and block their chances to have full analysis of the whole sentence.Finally, data from the Correlation Analysis between subjects'error percentage of giving responses to comprehension questions of garden path sentences and their results of the final English examination demonstrated that there was no significant correlation between subjects'chances of giving incorrect responses to questions about the subordinate clause of garden path sentences and their English proficiency, r=-0.174, Sig. (2-tailed) =0.165, p>0.05. But subjects'chances of giving incorrect responses to questions about the main clause of garden path sentences had significantly negative correlation with their English proficiency, r=-0.249, Sig. (2-tailed) =0.046, p<0.05.