Clause-Level Tense, Mood, Voice and Modality Tagging for German

Abstract

We present a language-independent clausizer (clause splitter) based on Universal Dependencies (Nivre et al., 2016), and a clause-level tagger for grammatical tense, mood, voice and modality in German. The paper recapitulates verbal inflection in German-always juxtaposed with its close relative English-and transforms the linguistic theory into a rule-based algorithm. We achieve state-of-the-art accuracies of 92.6% for tense, 79.0% for mood, 93.8% for voice and 79.8% for modality in the literary domain. Our implementation is available at https://gitlab.gwdg. de/tillmann.doenicke/tense-tagger.

Delexicalised Multilingual Discourse Segmentation for DISRPT 2021 and Tense, Mood, Voice and Modality Tagging for 11 Languages

“…Furthermore, a lot of the datasets in the shared task incorporate automatically created dependency trees (created by models trained on UD treebanks), which may lead to follow-up errors in the clause-splitting step. Dönicke (2020) reports an F1 of 81% for predicting clauses in a German text after preprocessing it with a spaCy model trained on the German UD treebanks. Even though this number only gives a rough estimate on how well our system identifies clauses, there is clearly room for improvement.…”

“…5 Algorithm 1 shows an updated version of the original algorithm that has been modified to work with a broader range of languages, specifically the languages in the shared task. In the following, the algorithm is briefly described, with a focus on the adaptions made for multiple languages (numbers in parentheses refer to lines in the pseudocode); for further explanations see Dönicke (2020).…”

“…35-37). In Dönicke (2020), the lemmas of the verbs are used but in our multilingual implementation, we map the lemmas to three categories of modal verbs (cf. Biber et al, 2002, p. 176): permission/possibility/ability (POS), obligation/necessity (OBL), and volition/prediction (VOL).…”

Exploring owner-based brand personality on Facebook and LinkedIn with Natural-Language-Processing (NLP)

“…The Python package spaCy offers a solution, but reports question its reliability for German text [23]. In fact, determining the tense in German text is almost a research project in itself [22]. Given the limited scope and resources of this research project, tense as an explanatory aspect is not further considered.…”

Modelling Speaker Attribution in Narrative Texts With Biased and Bias-Adjustable Neural Networks

“…Clause segmentation is performed with the clausizer presented inDönicke (2020). The manually created clause-level annotations are then automatically mapped to the detected clauses 4.…”