在此三年計畫裡，我們將進行整合語料庫語言學與心理語言學方法之研究。計畫整體目標在於解釋概念隱喻中之概念資訊如何產生，並其如何表徵於單語或多語語料庫資料之中。基於文獻中仍待解答之問題，我們倡議採用以語料庫驅動之方法來定義概念隱喻中之概念知識。我們提出以下三個研究問題： a）跨語言中概念隱喻之概念領域如何相似或不同？ b）以語料庫驅動之方法如何有助於定義相關概念領域之關係？ c）人們對由計算語言學工具定義之概念領域如何反應？透過語料庫資料之操弄以及心理語言學實驗，我們期盼由詞彙資源所表徵之資訊類型能由心理語言學實驗獲得證實。專特來說，我們希望藉由檢驗概念隱喻之源域及其搭配詞資訊，此研究能解釋概念領域之認知基礎。此研究之成果將改進當前首重隱喻辨識過於界定領域資訊之自動隱喻辨識研究。 In this three-year proposal, we will conduct studies with integration of both corpus linguistic and psycholinguistic paradigms. The overall proposal aims to provide an explanation as to how conceptual information in conceptual metaphors is generated and how it is then represented both in monolingual and multilingual corpora data. Based on the questions unanswered by previous research, we propose to use a corpus-driven approach to help defining the conceptual knowledge of conceptual metaphors. The following three research questions are asked: a) How are conceptual domains of conceptual metaphors similar or different across languages? b) How can a data-driven approach help defining the relationships between related conceptual domains? c) How do people respond to the conceptual domains defined through computational methods? Through the manipulation of corpora data and psycholinguistic experiment, we hope that the types of information represented by lexical resources can be validated through psycholinguistic experiments. In particular, we hope that our study can explain cognitive underpinnings of conceptual domains through examining the source domains of conceptual metaphors and their collocation information. The results from our study will help refine the current work of automatic metaphor identification which focuses first on metaphor identification than determining domain information.