A Morpho-Syntactic Analyzer of Controlled Japanese
Yukiko Sasaki Alam
Abstract
The proposed morpho-syntactic analyzer parses controlled Japanese
texts such as articles in newspapers, technical magazines and
professional journals and public documents that are transcribed
wherever applicable by using Joyo Kanji (frequently used
Chinese characters). The analyzer parses sentences in controlled
Japanese texts into morpho-syntactic units, further dividing them into
the content and the functional parts, and assigning a functional role
or roles to each unit in the sentences. As the system is not equipped
with a dictionary, the parsing algorithm is based on the orthographic
characteristics of words and morphemes, and the role assignment to
each unit is based on the functional elements located at the end of
the unit, which is a feature of a Head-final language like
Japanese. The system is a light-weight rule-based morpho-syntactic
analyzer that could be a useful tool for natural language
processing. As the system identifies syntactic units rather than
individual morphemes, together with the functional and/or syntactic
roles of the units, it would help a computational system understand
the syntactic and functional structures of sentences, and eventually
interpret the semantics of the sentences.
Proceedings of GEAF07; CSLI Publications On-line
Proceedings TOC
Proceedings as a single large pdf file