Dear members,
We invite you to contribute data for UniDive’s shared task on morpho-syntactic parsing, to be hosted as part of SyntaxFest 2025 (August, Ljubljana). Data for this task largely resembles dependency trees as in Universal Dependencies, but bypasses wordhood debates by adopting the distinction between content and function elements to fully characterize sentences. The systems participating in this task will be required to output relations between content nodes and ascribe them morpho-syntactic features. Data and systems in this data format will allow easier cross-lingual studies and natural treatment of many additional languages, especially of the agglutinative type.
We are looking for data contributors and annotators for data in various languages for this task. Our current team includes annotation teams in about a dozen of languages and we would like to expand the number and variety of the languages included. The data is set to be released around *February 2025*, and it can be either data tagged from scratch or semi-automatically converted from existing treebanks. Initial or partial data is to be available already *by the end of 2024*.
Researchers interested in joining are welcome to reply to one or all the organizers (omer.goldman(a)gmail.com, weissweiler(a)utexas.edu, reut.tsarfaty(a)gmail.com) to hear more details of the data structure, including annotation guidelines and English example.
Best regards,
Omer Goldman, Leonie Weissweiler, and Reut Tsarfaty