【作者】 余莉莉；

【导师】 闫国利；

【作者基本信息】 天津师范大学 ， 应用心理学， 2014， 博士

【摘要】 研究主要使用中文中的熟语为研究材料,探讨了在熟语这种由两个或者两个以上的词构成的、比词大的特殊的语言单位中的整体加工和成分加工过程。研究使用的熟语材料中包括了四字成语和三字惯用语两种。两者具有一般熟语的共同特征：同时具有能够表达整体的比喻意义和能够表达成分的字面意义。两者也存在一定差异：成语的结构更加凝固,成分之间的限制性更强；惯用语在材料的匹配上具有更大的优势。研究一采用四字成语为实验材料,研究二采用三字惯用语为实验材料,共同探讨两个主要问题：第一,在熟语中是否都存在整体加工和成分加工；第二,熟语中的整体加工和成分加工是否受到熟语结构紧密程度的调节。实验1探讨了在命名任务中的成语的整体加工和成分加工。实验分别对四字成语和与其尾字相同的一般短语的尾字进行重复启动和无关启动。实验结果中显示了显著的启动效应,对成语命名的时间比对短语命名的时间短。成语条件下的启动量和一般短语条件下的启动量之间具有显著的差异,成语条件下的启动量大于短语条件下的启动量。实验2探讨了句子阅读中的成语整体加工和成分加工。实验采用了三种不同的启动方式对置于句子中的成语进行了启动,包括对成语整体比喻意义的启动,对成语尾字字面意义的启动和无关启动。实验结果显示,在成语区域的分析上,两个与意义有关的启动条件都促进了成语的加工,表现在总阅读时间指标上对意义相关启动条件下的成语的注视时间短于无关条件启动下的对成语的注视时间。实验1和实验2的结果显示了不同任务下(命名任务和句子阅读任务),对成语尾字重复启动和意义相关启动条件下成语的优势加工。这说明,在对成语的认知加工中存在成分加工。在实验1的结果中,成语的命名时间短于一般短语,并且成语条件下的启动量大于一般短语,这说明在成语的认知中也存在整体加工。这一点在实验2的对成语整体比喻义的启动使得对成语的注视时间变短这一结果上也有所体现。实验1和实验2的结果更符合熟语加工理论中的混合加工理论。实验3探讨了成语的语义透明度对成语整体和成分加工的影响。实验材料采用了语义透明度高和语义透明度低的词对,放置于句子当中。对成语以及成语后一个词的区域进行了分析。两个区域上的实验结果没有提示任何指标下的成语的语义透明度效应。这说明,成语的语义透明度并不影响成语的加工。这和混合加工理论的假设是不一致的。研究二中考察了在三字惯用语和三字匹配短语中的词N+2预视效应,对惯用语中的整体加工和成分加工进行探讨。实验4试图重复以往研究中的词N+2预视效应：即词N+2的预视效应只出现在词N+1为高频词的情况下；当词N+1为低频词时,没有观察到词N+2的预视效应。实验采用了Yang等人(2009,2012)实验中的部分材料。实验结果得到重复,但是,在N+1为低频词时,存在对词N+1和词N+2联合区域凝视时间上的词N+2的预视效应。实验5采用了和实验4的材料进行严格匹配的短语和惯用语(如,词N+1的频率、笔画,词N+2的频率等)为实验材料。考察了三字惯用语和匹配的短语中的词N+2的预视效应。实验结果发现,词N+2预视效应在联合区域(整个惯用语或短语)的所有注视时间上都产生,在词N+2区域的单一注视时间和凝视时间上也出现。这显示了在预视这一早期效应上的惯用语和短语的加工类似,存在对惯用语成分的加工。但在联合区域的凝视时间和回视路径阅读指标上,惯用语的注视时间呈现出比短语的注视时间短的趋势,这仍在一定程度上说明了惯用语中的整体加工优势。综上所述,不论是在四字成语还是在三字惯用语中都是既存在整体加工,又存在成分加工。比起与熟语语义相关的因素,如语义透明度,熟语中的整体加工或成分加工方式或程度更加受到熟语结构紧密性的调节。实验的结果部分支持了熟语加工中的混合加工理论和建构理论。更多还原

【Abstract】 The main purpose of this study is to explore the whole-word processing and component processing in Chinese idioms by using idioms as main stimuli in the study. Idiom is a specific linguistic unit that normally contains more than two words. The idioms used in the study include two different types, which were4-character idioms and3-character idioms. The two different types of idioms share the general property of idioms, which is, having the figurative meaning, which can express the meaning of the whole-idiom, and the literal meaning, which can express the meaning of the component of the idiom at the same time. There are also differences between these two types of idiom, the most important characteristic of which is, the linguistic structure of the4-character idioms is more compacted. Whist, the3-character idioms are more suitable for the psycholinguistic experiments due to its property of better control (for example, comparing with normal phrases).4-character idioms were used in the first study, and then3-character idioms were used in the second study. Two main questions were answered in the study:First of all, are there both whole-word processing and component processing in the idioms? Second, whether the compactness of the linguistic structure of the idiom will influence the whole-word and component processing in the idioms?Experiment1investigated these two questions by using the naming task. The repeated and unrelated priming of the final character were applied into both four-character idiom and compared normal phrases. The result showed that, idioms were named faster than normal phrases; naming reaction time was also faster in the repeated priming condition than the unrelated condition. Prime effect was larger in the idioms than in the phrases.Experiment2investigated the same questions by embedding the idioms in the sentences. Three different types of prime were used in this experiment, which were, semantic prime that related to the figurative meaning of the idiom, semantic prime which related to the literal meaning of the final character in the idiom, unrelated prime. Results showed that both semantic primes primed the idioms, caused shorter go path time in the fixations of whole-idioms.The component processing in the idioms was showed in Experiment1and2by priming the last character in the idiom, as well as the whole-word processing, indicated by the more efficient processing in idioms comparing to phrases (Experiment1), and the figurative meaning priming (Experiment2). Results of the first two experiments indicated the fitness of the hybrid hypothesis in idiom processing.The influence of transparency on the processing of idioms was tested in Experiment3. The paired transparent and opaque idioms were embedded in the same fragments of sentences. Results showed null effect of transparency, which was against to the assumptions of hybrid hypothesis.To answer the question of whether there are whole-word processing and component processing in3-character idiom, the Study2observed the word n+2effect in both3-character idioms and matched phrases.The observation of word n+2preview benefit in Chinese has some constraints, which are word n+2preview benefit only showed when word n+1was a high frequency character, so that when word n+1was a low frequency character, there would be no effect for the preview of word n+2. By using the selected version of stimuli from Yang et al.’s study (2009,2012), the result was replicated except in the analysis of the combined region of word n+1and word n+2, which showed a word n+2preview effect in gaze duration, even when word n+1was a low frequency character.Experiment5investigated the whole-word and component processing in3-character idioms by observing the n+2preview benefit in both3-character idioms and well-matched phrases, when word n+1was a low-frequency character. The valid n+2preview benefit from both idioms and phrases were showed in all fixation measures in the combined region, also on single fixation duration and gaze duration in region of word n+2, which indicating a similar processing of3-character idioms to normal phrases. However, idioms were marginally processed faster than idioms on gaze duration and go past time in combine region.To summarize, there are whole-word processing and component processing in both types of idioms. Comparing to the semantic related factors (i.e. semantic transparency), the compactness of the structure of idioms is more influencial on processing in idioms. The results from the study were partly support the assumptions of hybrid hypothesis and configurative hypothesis in idiom processing.更多还原