DRDs Lexicons: conception, development and multilingual linking

The meeting on DRDs Lexicons: conception, development and multilingual linking will take place in Lisbon, Portugal, on 24-25 October 2016.  This is a special focus meeting, transversal to all working groups of the TextLink action.

During this meeting, we will review main issues related to the building of DRD Lexicons and their multilingual linking.  Some topics to address are the definition of Discourse-related Devices, their categorization in primary and secondary markers (or alternative lexicalizations), language-specific issues affecting subcategories of DRDs (such as adpositions and anaphoric elements) and specific categories of spoken language.

We will revise the existing Lexicons of DRDs and the options that were undertaken in terms of source data and methodology to create DRDs lists: grammar and coursebook inspection, translations from existing lists for other languages, corpus-driven or extraction from discourse treebanks, automatic identification in parallel corpora.

Another topic of discussion is the list of features to include in the Lexicons, such as category, rhetorical sense, representative corpus examples, equivalents in other languages, position, linear order of the conjuncts, restriction on the context (e.g., tense, mood, negation, aspect).

These questions will provide input to a reflection on the challenges of creating multilingual linked lexicons of DRDs in the scope of the action and on the minimum harmonization that we might try to reach for.

Our goal is to provide breaking ground to establish best practices and recommendations for the construction of DRDs by relying on the diversity of languages and frameworks represented in our network.