==============================================================
Chinese corpus for the CoNLL-2009 shared task
"Syntactic and Semantic Dependencies in Multiple Languages"
http://ufal.mff.cuni.cz/conll2009-st/

Version 1.1: February 11, 2009

Organizer of this corpus:
Nianwen Xue, Brandeis University, xuen@cs.brandeis.edu
==============================================================

This document provides the basic information of this corpus. 

WARNING

This data is a subset of the Chinese Treebank 6.0 released by the Linguistic Data Consortium. 
(Catalog number: LDC2007T36). The semantic annotation in this data is released by the LDC as the Chinese Proposition Bank 2.0 (Catalog number LDC2008T07). The use of this data during and after the CONLL shared task is conditional on your agreement to the LDC special license for the 2009 CoNLL shared task.

(1) GENERAL

This distribution includes the data sets for Chinese language, part
of the CoNLL-2009 shared task. The corpus includes syntactic dependencies 
and semantic dependencies. 

(2) CONTENTS OF THE DISTRIBUTION 1.0

The following files are included in this distribution:

* README.TXT - this file

* Chinese-conll09-train.txt - the training corpus

* Chinese-conll09-dev.txt - the development corpus

* frames: the frame files used in the annotation of the Chinese Proposition Bank. The frame files 
  are named with a unique ID + pinyin, and the corresponding Chinese character(s) can be found inside 
  the frame files. 

* tagsets.txt - the document describing the tagsets
  The tagsets of parts of speech (POS), syntactic dependencies and semantic dependencies are 
  described.

(3) SPECIFICS OF THE CHINESE DATA SET

The special features of this corpus are:

Encoding:
* The encoding of the corpus files is UTF-8.

Segmentation and POS-tagging:
* Each line corresponds to a word. Words are correctly segmented and 
  the CoNLL-2009 shared task does not handle word segmentation. The POS tags in the PPOS field are 
  generated MXPOST retrained on the Chinese Treebank. The POS-tagger is trained on 
  half of the data to tag the other half. 

* The (P)FEAT field is left empty because of the lack of inflection moprhology in Chinese.

Syntactic dependency:

* The Chinese Treebank is a phrase structure corpus and the dependency structure is automatically 
 converted from the phrase structure representation. The conversion is based on the identification 
 of phrases that have four different kind syntactic relations: modification, predication, 
 complementation and coordination. It's worth noting that the head of a coordination structure is 
 the first conjunct, and the heads of other conjuncts are conjunctions immediately preceding them. 
 The conjunctions all "modify" the first conjunct. Some punctuation marks are treated as coordinating 
 conjunctions.


Semantic dependency (predicates and arguments):

* Only verbal predicates are annotated.

* The same word can be an argument to multiple predicates.

* The predicate senses correspond to those in the Chinese Propbank frame files.


The following tools have been used to generate the Predicted (P-) columns:

* The PPOS field is generated by MXPOST (ftp://ftp.cis.upenn.edu/pub/adwait/jmx/jmx.tar.gz) retrained on the Chinese Treebank.

* PHEAD and PDEPREL are generated using the MALT Parser 
  (http://maltparser.org/). The parser was trained using one half of the data to tag the other half.


REFERENCES

Nianwen Xue, Fei Xia, Fu-Dong Chiou and Martha Palmer. 2005. The Penn Chinese Treebank: Phrase 
Structure Annotation of a Large Corpus. Natural Language Engineering, 11(2):207-238

Nianwen Xue and Martha Palmer. 2009. Adding semantic roles to the Chinese Treebank. Natural Language 
Engineering, 15(1):143-172.

CONTRIBUTORS

The provider thanks everyone who has contributed to the Chinese Treebank and the Chinese Proposition Bank. The following individuals directly contributed to the Chinese Treebank (in alphabetic order): Mitch Marcus, Meiyu Chang, Fu-Dong Chiou, Shizhe Huang, Zixin Jiang, Tony Kroch, Martha Palmer, Fei Xia, Nianwen Xue. The contributors to the Chinese Proposition Bank include (in alphabetic order): Meiyu Chang, Gang Chen, Helen Chen, Zixin Jiang, Martha Palmer, Zhiyi Song, Nianwen Xue, Ping Yu, Hua Zhong.