論後設語言語意表述之簡化：以本體論為本之研究

Simplifying Meaning Representation in Metalanguage: An Ontology-Based Approch

黃意晴(Yi-qing Huang)

指導教授 ： 吳俊雄

國立中正大學/文學院/語言學研究所/碩士(2013年)

摘要

本論文是要探討如何選擇適當的後設詞(meta-word)來簡化近義詞在電腦知識庫中的語意表述(meaning representation)，進而使電腦能夠辨識中文近義詞的關係。本研究收集了18篇關於中文近義詞的辨析，並將這些分析分成兩方面來看，一方面是句法功能，另一方面則是語意用法，本文即從這兩方面來判斷如何選出適當的後設詞。所謂後設詞是從後設語言(metalanguage)所衍生出來的概念，後設詞即為自然語言裡一個單字在電腦知識庫中的表述(representation)。
本文根據這18篇的近義詞辨析，列出四項句法功能為近義詞分析最常見的項目: (1)語法功能、(2)論元類型、(3)時貌類型，及(4)句型。為了使後設詞在處理中文近義詞時，不會導致句子不合文法，因此選擇擁有較多句法功能的詞做為適當的後設詞。在語意方面，列出兩項近義詞分析最普遍的做法，即是辨析(1)語意上的差異，以及(2)構詞共現。為求後設詞在電腦知識庫中能夠合理的轉換近義詞語意，在此仍是選擇擁有較多語意或是較廣泛的構詞共現者為最佳的後設詞。本文再根據從句法或是語意方面所選出的最佳後設詞，來示範如何運用在自然語言處理，使電腦能夠成功理解別兩組詞為近義詞，並討論若沒有統一的後設詞，則會造成電腦無法判斷兩者為近義詞關係。
本文研究提出，擁有較多功能者不論是在句法上抑或是語意上，比隨意選擇其中一個字為後設詞更為重要，因為這能使後設詞在電腦知識庫中，在句法上或是語意上都能順利地處理近義詞問題。若選擇較少功能的近義詞為後設詞，則可能使電腦在處理含有非後設詞的近義詞句子時出現錯誤。當此近義詞在轉換為後設詞時，若遇到該後設詞沒有的功能，則會導致整個語意表述公式不合文法，也無法有效的簡化近義詞在電腦知識庫中的語意表述。

關鍵字

語意表述 ； 後設詞 ； 近義詞

並列摘要

This study aims to explore how to find a proper meta-word so that simplify the meaning representations of near-synonyms, and make computer understand Chinese near-synonyms. We collect 18 lexical semantic papers about Chinese near-synonyms, then divide these analyses into two parts: one is syntactic functions, the other is semantic usage, then we discuss how to choose a proper meta-word from these two parts. A meta-word is an idea extended from the concept of metalangauge; it is the representation of a natural language word in the knowledge base.
These 18 studies show that four syntactic functions are examined commonly, that is, (1) grammatical function, (2) argument types, (3) aspectual types and (4) sentence types. To make computer process sentences successfully, we hope the meta-word is syntactically and semantically reasonable; thus we choose a word with more functions as meta-word. In semantic criterion, we list two common ways used to discriminate near-synonyms: (1) meaning differences, and (2) lexical collocation. In terms of our principle i.e. a word with more functions is qualified as meta-word, we also choose a word with more semantic usages as meta-word. We illustrate how to use meta-words to make computer understand Chinese near-synonyms, and discuss if there is no canonical form, then computer cannot recognize two words are synonymous.
We find that it is important to choose a word with more functions as meta-word because if we choose a meta-word with less functions, it may cause the meaning representations ungrammatical when we encounter a function that the meta-word does not have.

並列關鍵字

meaning representation ； meta-word ； near-synonym

參考文獻

Huang, S. L., &Chen, K. J.(2008). Knowledge Representation and Sense Disambiguation for Interrogatives in E-HowNet.Computational Linguistics and Chinese Language Processing,13(3),255-278.

Hung, W.T. (2010).Analyze the near-synonyms ‘Manufacture’ and ‘Produce’.Journal of Applied Chinese, 6, 223-246.

Tsai, M. C., Huang, C. R.,Chen, K. J., &Ahrens, K. (1998).Towards a Representation of Verbal Semantics – An Approach Based on Near Synonyms.Computational Linguistics and Chinese Language Processing,3(1): 61-74.

Blackburn, P., & Bos, J. (2005).Representation and Inference for Natural Language: A First Course in Computational Semantics, CSLI Publications.

Bos, J. (2011).A Survey of Computational Semantics: Representation, Inference and Knowledge in Wide-Coverage Text Understanding. Language and Linguistics Compass 5(6): 336–366.