As part of the International Workshop of Embodied Interaction & Embodied Intelligence [EC]² in Essen on August 21st and 22nd, Thomas Schmidt, Anne Ferger, Felix Bergmann and Karola Pitsch presented a work-in-progress report on the ZuMult installation at the University of Duisburg-Essen.
The talk was titled Complex Interactional Corpora: Workflows and Platform for video-based Studies of Embodied Interaction and Human-Machine Interaction and introduced an international audience of interaction and human-machine communication researchers to GI-based workflows and the ZuMult platform.
At the workshop „“Corpus linguistics 2040: Which data, which methods, which models?“ which took place at the Leibniz-Institute for the German Language on July 10th and 11th, Thomas Schmidt, Anne Ferger and Elena Frick presented a poster on: Multimodal data and qualitative exploration of audiovisual language corpora – the ZuMult corpus platform in 2025 The poster is available on Zenodo.
As on many other occasions of a similar kind, we felt it was our duty to remind the audience of the relevance of oral data and audiovisual language data. We did so by pitching our poster with a little poem:
Corpus linguists, be awoken To the importance of the spoken token Billions of written words are well and good But language can be fully understood Only by looking at its use in interaction with speakers’ and listeners’ reactions, Including the empirical reality Inherent in multimodality, the use of words in conversation, The dialectal variation, Stages of language acquisition Under different conditions, and not to forget: phonology. For this, we need technology. We have therefore invented A platform that will be presented In a much more detailed fashion As part of our poster session.
As part of the IPRA conference panel Corpus pragmatics – Methodological and empirical findings on the combination of qualitative and quantitative research in pragmatics, organised by Martin Luginbühl and Stefan Hauser, Thomas Schmidt presented a joint paper by Karola Pitsch, Anne Ferger and himself, entitled:
Beyond words – a case study in corpus approaches to interaction and multimodality
The paper focussed on the use of ZuMult with corpora containing massive amounts of video data and summarised ongoing work on that topic at the University of Duisburg-Essen.
Lucky Luke Award for Margo Blevins and Hans C. Boas
Before the workshop, Thomas Schmidt honoured Margo Blevins and Hans C. Boas with the Lucky Luke Award, recognising their outstanding achievements in all things Texas German.
The platform is based on ZuMult and provides access to the Texas German Corpus which TGDP has been building up for more than twenty years. This new version has almost 500 open-ended interviews (166 hours) with speakers of Texas German, totalling more than 1.2 million tokens of transcribed text for this unique variety of German.
Der Beitrag befasst sich mit dem Vorkommen und der Frequenz von gesprochensprachlichen Verschmelzungsformen in wissenschaftlichen Vorträgen von Studierenden – einerseits mit Deutsch als Erstsprache (L1) (n=25) und andererseits mit Deutsch als Zweit- bzw. Fremdsprache (L2) (n=96). Es wird gezeigt, dass Studierende eine Reihe von verschiedenen Klitisierungen verwenden und sich dabei deutliche Unterschiede zwischen Studierenden mit Deutsch als L1 und L2 aber auch zwischen unterschiedlichen akademischen Kontexten bzw. Erhebungsorten nachweisen lassen. Die für den Beitrag durchgeführten Analysen erfolgten mit Hilfe des im ZuMult-Projekt entwickelten Tools ZuRecht.
Am 27.09.2024 haben Christian Fandrych, Matthias Schwendemann und Franziska Wallner im Rahmen der Online-Tagung „Korpora, datengeleitetes Lernen (DDL) & Konstruktionen – Anwendungsszenarien für die DaFZ-Vermittlungspraxis“ einen Workshop durchgeführt. Im Zentrum standen dabei die beiden Tools ZuMal und ZuViel und mögliche Ansatzpunkte für Konstruktionsdidaktik und DDL-Aktivitäten. Die Folien zum Workshop können hier abgerufen werden.
