CoNLL-2009 Shared Task:
Syntactic and Semantic Dependencies in Multiple Languages



CoNLL-2009 Paper Submission Format and Procedure

Format and length

Only electronic submissions (.pdf or .ps format only) will be accepted. Submissions must follow the two-column format of NAACL HLT 2009 proceedings and conform to the official NAACL HLT 2009 style guidelines; however, the submission for the CoNLL 2009 Shared Task system description paper is NOT anonymous. Papers should be at most six (6) pages in length, including references. We strongly recommend the use of the NAACL HLT 2009 LaTeX style files or Microsoft Word Style files from the NAACL-HLT 2009 conference website.

The system description papers must describe the approach taken to perform parsing and semantic role labeling, and provide an informative analysis of the results. We encourage authors to include information on training/evaluation times and memory requirements (if available). This will facilitate comparisons on several dimensions besides system performance. The system description papers need not describe the shared task itself. This will be covered in the task description paper (see below for a proper reference). The system description papers may report results of post-evaluation systems in addition to the official scores, as long as these results are marked as post-evaluation. Please report your official scores from the Results page.

In addition to the Shared Task overview paper, we ask authors to cite the proper references to the datasets used to give their authors credit; if not for the original effort, then certainly for the effort they had to devote to the conversion (often non-trivial) from the varying treebank formats the seven datasets were originally in to the common Shared Task format. The license that you have signed actually require this for all future references.

To save you some time, we provide here the complete set of all the seven references (the overview paper plus the seven language resources, with a common one for Catalan and Spanish) for you to use in your system description papers, both in the .bib and .doc formats. In order not to cut too much space for your paper when including these references, we have also increased the allowed paper length to six pages (from last year's five).


author = {Jan Haji\v{c} and Massimiliano Ciaramita and 
Richard Johansson and Daisuke Kawahara and 
Maria Ant\`{o}nia Mart\'{\i} and Llu\'{\i}s M\`{a}rquez and Adam Meyers and 
Joakim Nivre and Sebastian Pad\'{o} and Jan \v{S}t\v{e}p\'{a}nek
and Pavel Stra\v{n}\'{a}k and Mihai Surdeanu
and Nianwen Xue and Yi Zhang},
title = {The {CoNLL}-2009 Shared Task: Syntactic and Semantic Dependencies in Multiple Languages},
booktitle = {Proceedings of the 13th Conference on 
Computational Natural Language Learning ({CoNLL-2009}), June 4-5},
address = {Boulder, Colorado, USA},
year = {2009}

author = {Mariona Taul\'{e} and Maria Ant\`{o}nia Mart\'{\i} and Marta Recasens},
title = {{AnCora: Multilevel Annotated Corpora for Catalan and Spanish}},
booktitle = {Proceedings of the {6th International Conference on Language Resources and Evaluation (LREC-2008)}},
year = {2008},
address = {Marrakesh, Morroco}

author = {Martha Palmer and Nianwen Xue},
title = {{Adding semantic roles to the Chinese Treebank}},
journal = {{Natural Language Engineering}},
volume = {15},
number = {1},
pages = {143--172},
year = {2009}

author = { Jan Haji\v{c} and Jarmila Panevov\'{a} and 
Eva Haji\v{c}ov\'{a} and Petr Sgall and Petr Pajas and
Jan \v{S}t\v{e}p\'{a}nek and Ji\v{r}\'{\i} Havelka and 
Marie Mikulov\'{a} and Zden\v{e}k \v{Z}abokrtsk\'{y}},
title = "{Prague Dependency Treebank 2.0}",
booktitle = {CD-ROM},
number = {{Cat. No. LDC2006T01, ISBN 1-58563-370-4}},
publisher = {Linguistic Data Consortium, Philadelphia, Pennsylvania, USA. URL:},
year = {2006}

author = {Surdeanu, Mihai and Johansson, Richard and Meyers, Adam and 
M\`{a}rquez, Llu\'{i}s and Nivre, Joakim},
title = {The {CoNLL}-2008 Shared Task on Joint Parsing of Syntactic and Semantic Dependencies},
booktitle = {Proceedings of the 12th Conference on Computational Natural Language Learning ({CoNLL-2008})},
year = {2008}

author = {Aljoscha Burchardt and Katrin Erk and Anette Frank and 
Andrea Kowalski and Sebastian Pad\'o and Manfred Pinkal},
title =  {The {SALSA} corpus: a {G}erman corpus resource for lexical semantics},
booktitle = {Proceedings of the 5th {International Conference on Language Resources and Evaluation (LREC-2006)}},
year = {2006},
address = {Genoa, Italy}

author = {Daisuke Kawahara and Sadao Kurohashi and {K\^oiti} Hasida},
title = {Construction of a {J}apanese Relevance-tagged Corpus}, 
booktitle = {Proceedings of the {3rd International Conference on Language Resources and Evaluation (LREC-2002)}},
year = {2002}, 
address = {Las Palmas, Canary Islands},
pages = {2008--2013}

Usage in LaTeX: ... datasets used \cite{PDT20}, \cite{...}, ... The
task is described in \cite{CoNLL-2009-ST} ... (or similarly, well you
know LaTeX, don't you...)

Word .doc file containing the reference in the appropriate style for an easy cut & paste into Word-based submissions:



At least two reviewers will read your paper. We will select reviewers from the participants and from the organizers. They will check the format of the paper and compliance with the specific requirements as described above.

Massimiliano Ciaramita from this year's Shared Task organizing team will oversee the paper submission and review process.

Submission site

All submissions will be handled by a separate CoNLL 2009 Shared Task track within the NAACL-HLT 2009 START conference system submission site.

Deadline: March 31, 2009, midnight HST (April 1, 10am GMT)

Final Camera-ready version of the paper

After incorporating changes based on the reviews, please use the START conference upload mechanism for submitting the camera-ready version of the paper. Please check back here for any last-minute hints regarding the final version of your paper.

Deadline: April 15, 2009, 2pm HST (April 15, midnight GMT)