Lemmatising Verbs in Middle English Corpora: The Benefit of Enriching the Penn-Helsinki Parsed Corpus of Middle English 2 (PPCME2), the Parsed Corpus of Middle English Poetry (PCMEP), and A Parsed Linguistic Atlas of Early Middle English (PLAEME)

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 7170–7178 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 7170 Lemmatising Verbs in Middle English Corpora: The Benefit of Enriching the Penn-Helsinki Parsed Corpus of Middle English 2 (PPCME2), the Parsed Corpus of Middle English Poetry (PCMEP), and A Parsed Linguistic Atlas of Early Middle English (PLAEME) Michael Percillier, Carola Trips University of Mannheim B6, 30–32, 68159 Mannheim percillier@uni-mannheim.de, ctrips@mail.uni-mannheim.de Abstract This paper describes the lemmatisation of three annotated corpora of Middle English — the Penn-Helsinki Parsed Corpus of Middle English 2 (PPCME2), the Parsed Corpus of Middle English Poetry (PCMEP), and A Parsed Linguistic Atlas of Early Middle English (PLAEME) — which is a prerequisite for systematically investigating the argument structures of verbs of the given time. Creating this tool and enriching existing parsed corpora of Middle English is part of the project Borrowing of Argument Structure in Contact Situations (BASICS) which seeks to explain to which extent verbs copied from Old French had an impact on the grammar of Middle English. First, we lemmatised the PPCME2 by (1) creating an inventory of form-lemma correspondences linking forms in the PPCME2 to lemmas in the MED, and (2) inserting this lemma information into the corpus (precision: 94.85%, recall: 98.92%, accuracy: 94%). Second, we enriched the PCMEP and PLAEME, which adopted the annotation format of the PPCME2, with verb lemmas to undertake studies that fill the well-known data gap in the subperiod (1250–1350) of the PPCME2. The case study of reflexives shows that with our method we gain much more reliable results in terms of diachrony, diatopy, and contact-induced change. Keywords: Lemmatisation, Middle English, verb argument structure 1. Introduction The project Borrowing of Argument Structure in Contact Situations1 (BASICS) investigates the contact situation be- tween Old French (OF) and Middle English (ME) which set in after the Norman Conquest (1066) and lasted until 1500. More specifically, we hypothesise that the copying of OF verbs with their predicate-argument structures (AS) has favoured and produced grammatical changes in ME. Instead of using the more traditional and more problem- atic term ‘borrowing’ we use Johanson’s (2002, 287f) term ‘copying’ as it allows for the non-identicality of original and copied material. We assume that verbs copied from OF to ME are global copies in Johanson’s terms, bringing along a block of properties (material, semantic, combinational in syntax and word structure, frequential). So far we have conducted item-oriented, class-oriented, and construction- oriented studies to gain new insights into this contact sit- uation and contact-induced changes in more general terms (see e.g. Percillier (2019), Trips and Stein (2018), Trips and Stein (2019)). To do so, we developed a method that allows us to systematically query verbs of French and non-French origin in the linguistically annotated corpora available for Middle English. In this article we will describe our method in detail, and by means of a corpus study show how it is successfully applied to phenomena relevant to our project. …

Content truncated to 500 words