Welcome to MCG
Mandarin Chinese Grammar (MCG) is an open-source HPSG grammar for Mandarin Chinese developed at the Language Technology Lab of the German Research Center for Artificial Intelligence (DFKI GmbH), Saarbrüken, Germany.
The grammar is released under the LGPL-LR, a version of the LGPL adapted for linguistic resources. You can check out the latest version of the grammar via:
svn co https://mcg.opendfki.de/repos/trunk mcg
The grammar aims for broad-coverage and linguistically precise analysis of spoken and written Chinese sentences. At the moment, the grammar targets at the standard mainland Mandarin (Putonghua 普通话), although future extensions to other variants of Mandarin or other Chinese dialects are possible. Analyses are drawn in the framework of HPSG, while using Minimal Recursion Semantics (MRS) as the semantic representation.
The development of the grammar started back in May 2004, and the project was suspended in mid-2005 due to time constraints. The development resumed in July 2010, with some input from Nanyang Technological University.
An updated version of the MCG has been released on 08.08.2012.
svn co https://mcg.opendfki.de/repos/tags/20120808 mcg
Several challenging phenomena have been addressed in this release, among them are the following highlights:
- Topic-comment constructions
- Basic BA(把)-constructions
- Basic BEI(被)-constructions
- Various DE(的)-phrases, including the associative DE and different types of nominalizing DE
- Resultative verb compounds (RVC)
- Basic Noun Phrase structures, including the treatment of numeral-classifier phrases, demonstratives, and their interactions with adjective and prepositional modifications
- Various marking structures, including aspect markers, sentence final markers, ordinal numeral prefix
- Locative phrases
- Basic coordinated structures
- Various post-verb adverbial modifiers
- Certain serial verb constructions
The lexicon currently consists of a small number of entries (over 500) as representatives for various lexical categories. Automatic extension of the lexicon will be carried out in the near future. Also included in this release is a small Chinese phenomenon-oriented test suite with both positive and negative items (thanks to Xiangli Wang).
Yi Zhang, Rui Wang, and Yu Chen. 2012. Joint Grammar and Treebank Development for Mandarin Chinese with HPSG. In Proceedings of the eighth international conference on Language Resources and Evaluation (LREC), Istanbul, Turkey. PDF Poster(11M)
Main developers and contributors of MCG are:
- Yi Zhang <yzhang AT coli.uni-sb.de>
- Rui Wang <ruiwang AT dfki.de>
- Yu Chen <yuchen AT dfki.de>
- Francis Bond <bond AT ieee.org>
- Zhenzhen Fan <zhenzhen AT nus.edu.sg>
We thank Xiangli Wang for sharing her phenomena oriented test-suite with us.