TY  - THES
T1  - Syntactic-Semantic Analysis of Modern Chinese with Left-Associative Grammar
A1  - Feng,Qiuxiang
Y1  - 2012/03/09
N2  - In this dissertation, Left-Associative Grammar is applied and tailored for the syntactic-semantic analysis of modern Chinese. The analysis follows the principle of time-linearity and description. Left-Associative Grammar is one of the two bases of Database Semantics. The DBS graphs of semantic relations, which can be automatically generated from the LAG derivation result, provide a more explicit view of the content of a sentence. The part of speech signature, as a more abstract presentation, helps to generalize the sentence patterns in Chinese with a clearer view of the deep semantic relations.

To meet the requirements of automatic LAG syntactic-semantic analysis and ensure its efficiency, a bilingual dictionary is built up. Lexical items in this dictionary are stored as non-recursive feature structures called proplets. It is also supposed to help on the lexical level to reduce ambiguity caused by polysemy and temporary shift of part of speech. NOUN, VERB and ADJECTIVE are the three word classes in the LAG-Chinese dictionary. This word classification is different from that in the traditional Chinese grammar. The class of NOUN includes noun in a narrow sense as in traditional grammar, and pronoun, numeral, quantifier as well. The class of VERB includes common verbs with one, two or three valences, causative verbs, modal verbs, directional verbs, etc.. The Class of ADJECTIVE includes adjective in a narrow sense, adverb, preposition, conjunctive, auxiliary, etc.. Each word has its attribute-value pairs with its grammatical and semantic information.

Based on the study of modern Chinese grammar, various basic usages of NOUNs, VERBs and ADJECTIVEs are analyzed. Rules, operations and graphical representation of the analysis result are provided. According to the demand of semantic analysis, adding values for the semantic attribute is proposed to note semantic roles and pragmatic functions, such as agent, patient, experiencer, and so on. The derivation result is meant to represent the content and deep semantic relations of the sentence. From the perspective of the fundamental usage of nouns, verbs and adjectives, the structures of subject-predicator, coordination, predicator-object, etc, are studied and analyzed, as well as the variants of these basic structures, including object-fronting, semantic passive, formal passive, element omission and so on.

    With respect to the analysis of a function word, the algorithm of conditional absorption is proposed as an improvement of Left-Associative Grammar. In the absorption of a function word by a content word, the core value and/or the semantic value of the function word is maintained under certain
conditions, which helps to avoid possible back-tracing to a large extent in later language production and machine translation. In all the 36 possible fundamental intra-propositional part of speech signatures, 29 are found in modern Chinese. It demonstrates the flexibility of Chinese sentence patterns. That all the 29 relations are correctly analyzed shows the adaptability of the algorithm adopted and proposed in this research. Because of the variety of Chinese texts, the approach of automatic syntactic and semantic analysis based on improved Left-Associative Grammar should also be applied to other large-scale Chinese corpus in order to test its efficiency. In addition, language production and machine translation based on this research is also a possibility for the following-up study.
KW  - Chinesisch
KW  - syntaktisch-semantische Analyse
KW  - Verstehen natürlicher Sprache
KW  - Datenbank Semantik
CY  - Erlangen
PB  - Universitätsbibliothek der Universität Erlangen-Nürnberg
AD  - Universitätsstraße. 4, 91054 Erlangen
L2  - http://www.opus.ub.uni-erlangen.de/opus/volltexte/2012/3155
ER  -