Parse and Query with Universal Dependencies
Gertjan van Noord, Gosse Bouma and Peter Kleiweg

The PaQu Parse-and-Query website is an online resource for uploading Dutch corpora for automatic parsing and querying the parsed output using XPath. It is meant to facilitate online access to and interaction with parsed corpora and manually corrected treebanks.

The Dutch treebanks and the Dutch Alpino parser provide dependency annotation in the CGN/Lassy/Alpino format. This format is the de-facto standard annotation format for Dutch, but in recent years, a novel standard format for dependency annotation has become widespread: Universal Dependency. This format is now the standard dependency annotation format, and treebanks for many languages are available in this format. Consider the listing at http://universaldependencies.org/.

In the new PaQuUd project (financed by Clariah+), we extend the PaQu tool with Universal Dependency annotation. This extension is possible because it is possible to map the analyses in the CGN/Lassy/Alpino format to the UD format fully automatically. In PaQuUd, we aim for the newer UD2.1 format, with extensions.