| 摘要: | 已有人提出，相較於「邏輯上的」選擇，第二語言學習者更接受量詞級距推理「語用」詮釋。比如，若我吃了所有的餅乾而要講「我吃了些餅乾」，這在「一些」作為「一些而非全部」的詮釋下不成立—是為語用詮釋；若我指更強烈的詞「全部」，我會講出來。然而，「一些」也能解釋成「一些且可能全部」。邏輯上的詮解，因為一些是全部的一個子集合，就像我已經吃了所有餅乾，然後我已經吃了一些餅乾這事也必定成立。這詮釋可以用在希望說謊的某人身上，就邏輯上，他們在講實話。兩個詮解在第一語言裡都有可能（雖然語用詮釋較為頻繁），但第二語學習者與第一語言相比，是否有閱讀上的偏好？相對於文獻，我提出第二語學習者的表現，就如同他們第一語言一樣。¬文獻裡導入結論的關鍵因素，乃部份不適合標量蘊義研究所使用的實驗題材。例如，某人認為「一些大象有象鼻」為偽，這是錯的，因有大象由於非法獵殺、天生缺陷等而沒有象鼻。測驗絕對事實當中，任誰都用諸如「一些水在攝氏100度煮沸」的命題，終究為偽。因此測驗語句得仔細設計。除了材料的問題，許多其他因素必須納入考量。惟獨測試「一些」之於「全部」會引起落差效果。而我們講「大多大象有象鼻」較能夠接受，也縮小異常子集合的範圍—因此顧及「邏輯」詮釋較佳。第二方面是測量受試者的工作記憶，因為與母語人士相比，（大腦）處理能力會限制第二語學習者所考量到的選項。第三，使用自主閱讀方法的反應時間則用以追溯這些處理效果。最後，文獻並未考量受試者的誠實觀點—人會因著謊言觀點不接受「一些」的邏輯詮解。最後，有了清楚的資料庫，我們足以回答理論的重要問題，因為neo-Gricean theory比Relevance theory做了更好的預測，而進行語用蘊義時與邏輯蘊義相左，兩派理論同樣做出相對的預測。亦即，語用蘊義將先產生，而若發現，邏輯性詮釋才有發展—或反之亦然？這些理論伴隨應用層面的問題，將在本研究全面探討。 ;It has been claimed that learners of a second language accept more ‘pragmatic’ interpretations of scalar implicature than their ‘logical’ alternatives. For example, if I ate all the cookies and were to say ‘I ate some of the cookies’, this would be false under the interpretation of ‘some’ as ‘some but not all’ – a pragmatic interpretation; if I meant the stronger term ‘all’, I would have said it. However, ‘some’ can also be interpreted as ‘some and possibly all’, the logical interpretation, as if I have eaten all the cookies, then it also must be true I have eaten some of the cookies, as some is a subset of all. This interpretation may be used by someone who wishes to lie, as logically, they are telling the truth. In the first language both interpretations are possible (although the pragmatic interpretation is the more frequent one), but do second language learners have a preference for either reading as compared to their first language? Contrary to the literature, I argue that second language learners behave just like their first language. The key factor that has led to the conclusion in the literature are some of the experimental materials used which are unsuited for scalar implicature study. For example, if one believes that ‘some elephants have trunks’ is false, this is incorrect as there may be elephants who do not have trunks due to poaching, birth defects, etc. In testing absolute truths, one must use such items like ‘some water boils at 100C’, which is patiently false. Thus test sentences must be carefully crafted. Apart from the issues with materials, several other factors must be taken into consideration. Only testing ‘some’ versus ‘all’ may induce a distance effect; it is more acceptable to say ‘most elephants have trunks’ which reduces the abnormal subset size – thus better allowing for a ‘logical’ interpretation. A second aspect is measuring the working memory of the participants, as processing capacity may limit how many alternatives a second language learner considers, compared to a native speaker. Thirdly, reaction timing using a self-paced reading methodology is introduced to trace these processing effects. Finally, the literature has not considered participants’ views of honesty – a person may not accept a logical interpretation of ‘some’ because they prefer not to lie. These aspects are dealt with to provide a valid and reliable study of scalar implicature. Finally, with a clear database, we will be able to answer important questions of theory, in that does a neo-Gricean theory make better predictions than Relevance theory in relation to when a pragmatic implicature is processed versus a logical implicature, as both theories make opposite predictions. That is, will a pragmatic implicature be generated first, and if found to be inconsistent with the context, a logical interpretation will be developed – or vice versa? These questions of theory, along with the applied aspects will all be addressed by this study. |
