An Open Source Tool for Partial Parsing and Morphosyntactic Disambiguation

This article presents a formalism and an open source implementation of a new tool for simultaneous partial parsing and morphosyntactic disambiguation and correction. We argue that, contrary to the common pipeline approach, where morphosyntactic tagging is fully accomplished before shallow or partial parsing, both tasks are best approached in parallel. This has been suggested before, and formalisms which allow for the interweaving of partial parsing and morphosyntactic disambiguation have been proposed. Our approach is novel in that a fully uniform formalism is presented, and a single grammar rule may contain structure-building operations, as well as morphosyntactic correction and disambiguation operations. The formalism has been implemented in Java and is now available under the GNU General Public License.