Oxford University Faculty of Classics

The PARSEME Ancient Greek Corpus

We are the PARSEME Ancient Greek working group! Welcome to our online presence.

PARSEME started as a COST-initiative in 2017. It is ‘an interdisciplinary scientific network devoted to the role of multi-word expressions (MWEs) in parsing’, specifically in Natural Language Processing. The annotated corpus currently contains 30 languages and 6 dialects from 10 language families, including a Modern Greek corpus. The annotated corpora are accessible here and guidelines for how to query them can be found here. The current annotation effort is part of the UniDive COST Action Working Group 1 (Corpus annotation). The Ancient Greek corpus is the first corpus-language corpus to join the PARSEME family (although efforts for Latin and Ancient Egyptian are also envisaged)..

We are currently focussing on verbal multi-word expressions (however efforts are under way to branch out to nominal and functional multi-word expressions in the near future). Verbal multi-word expressions are lexically speaking multi-lexemic units, e.g. to look up, to make a suggestion, and to take heart, the syntactic head of which is the verb. The semantic head is not necessarily the verb! Syntactically speaking, verbal multi-word expressions are complex predicates, in that they are monoclausal and in that all the lexemes contribute to the event structure. A multitude of frameworks have been applied to verbal multi-word expressions.

Verbal multi-word expressions sit at three interfaces, the syntax-lexicon interface (e.g. to take heart vs to make a suggestion), the syntax-semantics interface (e.g. to lack patience vs to have no patience), and the syntax-pragmatics interface (e.g. to make a comment vs to make a contribution). This hovering at interfaces contributes to the internal heterogeneity of this group of constructions.

Our goal is to make multi-word expressions as omnipresent in dictionaries, grammar books, and teaching resources as they are in the ancient texts.

This working group arose from the project Giving gifts and doing favours: Unlocking Greek support-verb constructions funded by the Leverhulme Trust (grant n. ECF-2020-181) at the University of Oxford (2020–2024).

As of September 2024, we are excited to present the Udemy course "Classical Greek Verbal Multi-Word Expressions: PARSEME-GRC" ((https://www.udemy.com/course/classical-greek-verbal-multi-word-expressions-parseme-grc/?referralCode=5F0B52286FC15909B21F).). If you are interested in joining as an annotator, please (1) complete the course, (2) contact the working group leader with your annotated sample, (3) attend a team meeting to get to know us all, and (4) get assigned your annotation portion. You can join for as short as a term or as long as you like. We only ask that you commit to a portion of text on joining.