Refine
Year of publication
Document Type
- Conference Proceeding (122)
- Article (109)
- Part of a Book (55)
- Book (11)
- Part of Periodical (3)
- Review (2)
- Other (1)
- Working Paper (1)
Language
- English (218)
- German (83)
- French (2)
- Multiple languages (1)
Keywords
- Korpus <Linguistik> (304) (remove)
Publicationstate
- Veröffentlichungsversion (232)
- Zweitveröffentlichung (53)
- Postprint (17)
- Ahead of Print (1)
Reviewstate
- Peer-Review (304) (remove)
Publisher
- IDS-Verlag (20)
- Institut für Deutsche Sprache (20)
- European Language Resources Association (18)
- de Gruyter (14)
- Association for Computational Linguistics (13)
- Leibniz-Institut für Deutsche Sprache (12)
- Linköping University Electronic Press (11)
- CLARIN (8)
- Erich Schmidt (8)
- Universitäts- und Landesbibliothek Darmstadt (8)
This contribution explores the relationship between the English CEFR (Common European Framework of Reference for Languages) vocabulary levels and user interest in English Wiktionary entries. User interest was operationalized through the number of views of these entries in Wikimedia server logs covering a period of four years (2019–2022). Our findings reveal a significant relationship between CEFR levels and user interest: entries classified at lower CEFR levels tend to attract more views, which suggests a greater user interest in more basic vocabulary. A multiple regression model controlling for other known or potential factors affecting interest: corpus frequency, polysemy, word prevalence, and age of acquisition confirmed that lower CEFR levels attract significantly more views even after taking into account the other predictors. These findings highlight the importance of CEFR levels in predicting which words users are likely to look up, with implications for lexicography and the development of language learning materials.
We investigate the optional omission of the infinitival marker in a Swedish future tense construction. During the last two decades the frequency of omission has been rapidly increasing, and this process has received considerable attention in the literature. We test whether the knowledge which has been accumulated can yield accurate predictions of language variation and change. We extracted all occurrences of the construction from a very large collection of corpora. The dataset was automatically annotated with language-internal predictors which have previously been shown or hypothesized to affect the variation. We trained several models in order to make two kinds of predictions: whether the marker will be omitted in a specific utterance and how large the proportion of omissions will be for a given time period. For most of the approaches we tried, we were not able to achieve a better-than-baseline performance. The only exception was predicting the proportion of omissions using autoregressive integrated moving average models for one-step-ahead forecast, and in this case time was the only predictor that mattered. Our data suggest that most of the language-internal predictors do have some effect on the variation, but the effect is not strong enough to yield reliable predictions.
In diesem Beitrag werden Komposita mit den relationalen Zweitgliedern Gatte und Gattin aus genderlinguistischer Perspektive untersucht, basierend auf manuell annotiertem zeitungssprachlichen Korpusmaterial. Frauen werden im analysierten Korpus ca. 12-mal häufiger in ihrer ehelichen Rolle versprachlicht als Männer. Statistische Analysen zeigen, dass sie dabei systematisch in ein possessives Verhältnis zum Ehemann gesetzt werden (Arztgattin = Gattin eines Arztes), während Ehemänner in den untersuchten Komposita tendenziell doppelt individualisiert werden (Arztgatte = Gatte, der Arzt ist). Neben den Zweitgliedern geben auch die Genera der beiden Konstituenten Aufschluss über die kodierte Bedeutungsrelation: Genusgleichheit (Kanzlergatte) führt zu einer qualifizierenden, Genusdivergenz (Kanzleringatte) zu einer possessiven Lesart. Die Analyse belegt außerdem die Existenz movierter Kompositumserstglieder – diese sind sogar die häufigste Form zur Benennung weiblicher Personen im Erstglied. Trotzdem herrscht bei der Bezugnahme auf Frauen eine größere Formenvarianz als bei Männern, welche fast ausschließlich mit maskulinen Erstgliedern versprachlicht werden. Damit zeigt die Studie, wie genderlinguistische Perspektiven auch im Bereich der Wortbildung einen neuen Analysezugang bilden.
OWID und OWIDplus – lexikographisch-lexikologische Online-Informationssysteme des IDS Mannheim
(2023)
Lexikographische und lexikalische Ressourcen zum Deutschen werden an vielen unterschiedlichen Institutionen erarbeitet, z. B. an Akademien der Wissenschaften oder in privatwirtschaftlichen Verlagen. Auch am Leibniz-Institut für Deutsche Sprache (IDS) in Mannheim werden solche Materialien erstellt und der (Fach-)Öffentlichkeit unter dem Dach von OWID, dem „Online-Wortschatz-Informationssystem Deutsch“ (owid.de), präsentiert.
Linguistische Studien arbeiten häufig mit einer Differenzierung zwischen gesprochener und geschriebener Sprache bzw. zwischen Kommunikation der Nähe und Distanz. Die Annahme eines Kontinuums zwischen diesen Polen bietet sich für eine Verortung unterschiedlichster Äußerungsformen an, inklusive unkonventioneller Textsorten wie etwa Popsongs. Wir konzipieren, implementieren und evaluieren ein automatisiertes Verfahren, das mithilfe unkorrelierter Entscheidungsbäume entsprechende Vorhersagen auf Textebene durchführt. Für die Identifizierung der Pole definieren wir einen Merkmalskatalog aus Sprachphänomenen, die als Markierer für Nähe/Mündlichkeit bzw. Distanz/Schriftlichkeit diskutiert werden, und wenden diesen auf prototypische Nähe-/Mündlichkeitstexte sowie prototypische Distanz-/Schrifttexte an. Basierend auf der sehr guten Klassifikationsgüte verorten wir anschließend eine Reihe weiterer Textsorten mithilfe der trainierten Klassifikatoren. Dabei erscheinen Popsongs als „mittige Textsorte“, die linguistisch motivierte Merkmale unterschiedlicher Kontinuumsstufen vereint. Weiterhin weisen wir nach, dass unsere Modelle mündlich kommunizierte, aber vorab oder nachträglich verschriftlichte Äußerungen wie Reden oder Interviews vollkommen anders verorten als prototypische Gesprächsdaten und decken Klassifikationsunterschiede für Social-Media-Varianten auf. Ziel ist dabei nicht eine systematisch-verbindliche Einordung im Kontinuum, sondern eine empirische Annäherung an die Frage, welche maschinell vergleichsweise einfach bestimmbaren Merkmale („shallow features“) nachweisbar Einfluss auf die Verortung haben.
We present a collection of (currently) about 5.500 commands directed to voice-controlled virtual assistants (VAs) by sixteen initial users of a VA system in their homes. The collection comprises recordings captured by the VA itself and with a conditional voice recorder (CVR) selectively capturing recordings including the VA-directed commands plus some surrounding context. Next to a description of the collection, we present initial findings on the patterns of use of the VA systems during the first weeks after installation, including usage timing, the development of usage frequency, distributions of sentence structures across commands, and (the development of) command success rates. We discuss the advantages and disadvantages of the applied collection-specific recording approach and describe potential research questions that can be investigated in the future, based on the collection, as well as the merit of combining quantitative corpus linguistic approaches with qualitative in-depth analyses of single cases.
This paper analyses intensification in German digitally-mediated communication (DMC) using a corpus of YouTube comments written by young people (the NottDeuYTSch corpus). Research on intensification in written language has traditionally focused on two grammatical aspects: syntactic intensification, i.e. the use of particles and other lexical items and morphological intensification, i.e. the use of compounding. Using a wide variety og examples from the corpus, the paper identifies novel ways that have been used for intensification in DMC, and suggests a new taxonomy of classification for future analysis of intensification.
Der vorliegende Beitrag vergleicht die Verwendung der anglizistischen Nomination old school und der nativen Entsprechung Alte Schule im Hip-Hop-Subkorpus des Songkorpus (Schneider 2020). Dieser Vergleich erfolgt auf zwei Ebenen: Zum einen wird die diskurs-spezifische Verwendung anhand eines adaptierten Analyse-Frameworks für Hip-Hop-Texte von Androutsopoulos und Scholz (2002) untersucht, zum anderen wird der syntaktische und morphologische Gebrauch in den Deutschraptexten analysiert. Dabei zeigt sich, dass es jeweils spezifische Verwendungstendenzen auf diskursiver Ebene gibt, die wesentlichsten Unterschiede aber in der syntaktischen und morphologischen Verwendung auftreten, allen voran in der höheren Produktivität der anglizistischen Nomination. Es wird dafür argumentiert, dass sich dies unter anderem auf sprachstrukturelle bzw. wortformale Spezifika des Englischen zurückführen lässt, wie den nicht vorhandenen Flexionssuffixen der Adjektive. Damit werden die in der Anglizismenforschung etablierten Überlegungen zu Verwendungsgründen um eine simple, aber gegebenenfalls folgenreiche Beobachtung ergänzt, die sich vor allem bei den sprachökonomischen Ansätzen einordnen lässt. Schließlich wird darüber auf diskursiver Ebene wiederum auch ein Bezug zu terminologischen Vorteilen hergeleitet: Trotz flexibler Verwendung wird das schriftliche Abbild bei Wortbildungen geschont (Oldschoolstyle, Oldschool-Aufnahmen, Oldschooler), was für die Wiedererkennbarkeit des Diskurselements – neben der zusätzlichen Auszeichnung durch die Eigenschaft ‚fremdsprachig‘ – zuträglich sein könnte.
Computational language models (LMs), most notably exemplified by the widespread success of OpenAI's ChatGPT chatbot, show impressive performance on a wide range of linguistic tasks, thus providing cognitive science and linguistics with a computational working model to empirically study different aspects of human language. Here, we use LMs to test the hypothesis that languages with more speakers tend to be easier to learn. In two experiments, we train several LMs—ranging from very simple n-gram models to state-of-the-art deep neural networks—on written cross-linguistic corpus data covering 1293 different languages and statistically estimate learning difficulty. Using a variety of quantitative methods and machine learning techniques to account for phylogenetic relatedness and geographical proximity of languages, we show that there is robust evidence for a relationship between learning difficulty and speaker population size. However, contrary to expectations derived from previous research, our results suggest that languages with more speakers tend to be harder to learn.
"Das im Januar 2022 gestartete Projekt "Sprachanfragen" (https://www.ids-mannheim.de/gra/projekte2/sprachanfragen/) verfolgt erstmalig das Ziel, Sprachanfragedaten zu erfassen, aufzubereiten und ein wissenschaftsöffentliches Monitorkorpus aus ihnen zu erstellen. Dazukommend wird eine Rechercheschnittstelle entwickelt, mit der die Sprachanfragen systematisch wissenschaftlich analysierbar gemacht werden. Das Poster gibt einen Überblick über das Projekt, zeigt erste Ergebnisse und bietet einen Ausblick auf Überlegungen zur Konzeption eines Chatbots zur automatisierten Beantwortung von Sprachanfragen." Ein Beitrag zur 9. Tagung des Verbands "Digital Humanities im deutschsprachigen Raum" - DHd 2023 Open Humanities Open Culture.
For many reasons, Mennonite Low German is a language whose documentation and investigation is of great importance for linguistics. To date, most research projects that deal with this language and/ or its speakers have had a relatively narrow focus, with many of the data cited being of limited relevance beyond the projects for which they were collected. In order to create a resource for a broad range of researchers, especially those working on Mennonite Low German, the dataset presented here has been transformed into a structured and searchable corpus that is accessible online. The translations of 46 English, Spanish, or Portuguese stimulus sentences into Mennonite Low German by 321 consultants form the core of the MEND-corpus (Mennonite Low German in North and South America) in the Archive for Spoken German. In addition to describing the origin of this corpus and discussing possibilities and limitations for further research, we discuss the technical structure and search possibilities of the Database for Spoken German. Among other things, this database allows for a structured search of metadata, a context-sensitive token search, and the generation of virtual corpora that can be shared with others. Moreover, thanks to its text-sound alignment, one can easily switch from a particular text section of the corpus to the corresponding audio section. Aside from the desire to equip the reader with the technical knowledge necessary to use this corpus, a further goal of this paper is to demonstrate that the corpus still offers many possibilities for future research.
We introduce DeReKoGram, a novel frequency dataset containing lemma and part-of-speech (POS) information for 1-, 2-, and 3-grams from the German Reference Corpus. The dataset contains information based on a corpus of 43.2 billion tokens and is divided into 16 parts based on 16 corpus folds. We describe how the dataset was created and structured. By evaluating the distribution over the 16 folds, we show that it is possible to work with a subset of the folds in many use cases (e.g., to save computational resources). In a case study, we investigate the growth of vocabulary (as well as the number of hapax legomena) as an increasing number of folds are included in the analysis. We cross-combine this with the various cleaning stages of the dataset. We also give some guidance in the form of Python, R, and Stata markdown scripts on how to work with the resource.
One of the fundamental questions about human language is whether all languages are equally complex. Here, we approach this question from an information-theoretic perspective. We present a large scale quantitative cross-linguistic analysis of written language by training a language model on more than 6500 different documents as represented in 41 multilingual text collections consisting of ~ 3.5 billion words or ~ 9.0 billion characters and covering 2069 different languages that are spoken as a native language by more than 90% of the world population. We statistically infer the entropy of each language model as an index of what we call average prediction complexity. We compare complexity rankings across corpora and show that a language that tends to be more complex than another language in one corpus also tends to be more complex in another corpus. In addition, we show that speaker population size predicts entropy. We argue that both results constitute evidence against the equi-complexity hypothesis from an information-theoretic perspective.
Neologisms, i.e., new words or meanings, are finding their way into everyday language use all the time. In the process, already existing elements of a language are recombined or linguistic material from other languages is borrowed. But are borrowed neologisms accepted similarly well by the speech community as neologisms that were formed from “native” material? We investigate this question based on neologisms in German. Building on the corresponding results of a corpus study, we test the hypothesis of whether “native” neologisms are more readily accepted than those borrowed from English. To do so, we use a psycholinguistic experimental paradigm that allows us to estimate the degree of uncertainty of the participants based on the mouse trajectories of their responses. Unexpectedly, our results suggest that the neologisms borrowed from English are accepted more frequently, more quickly, and more easily than the “native” ones. These effects, however, are restricted to people born after 1980, the so-called millenials. We propose potential explanations for this mismatch between corpus results and experimental data and argue, among other things, for a reinterpretation of previous corpus studies.
Poster des Text+ Partners Leibniz-Institut für Deutsche Sprache Mannheim präsentiert beim Workshop "Wohin damit? Storing and reusing my language data" am 22. Juni 2023 in Mannheim. Das Poster wurde im Kontext der Arbeit des Vereins Nationale Forschungsdateninfrastruktur (NFDI) e.V. verfasst. NFDI wird von der Bundesrepublik Deutschland und den 16 Bundesländern finanziert, und das Konsortium Text+ wird gefördert durch die Deutsche Forschungsgemeinschaft (DFG) – Projektnummer 460033370. Die Autor:innen bedanken sich für die Förderung sowie Unterstützung. Ein Dank geht außerdem an alle Einrichtungen und Akteur:innen, die sich für den Verein und dessen Ziele engagieren.
Die Darstellung von und Arbeit mit Transkripten spielt in vielen forschungs- und anwendungsbezogenen Arbeiten mit Daten gesprochener Sprache eine wichtige Rolle. Der im ZuMult-Projekt entwickelte Prototyp ZuViel (Zugang zu Visualisierung von Transkripten) knüpft an etablierte Verfahren zur Transkriptdarstellung an und erweitert diese durch neue Möglichkeiten des interaktiven Arbeitens mit Transkripten im digitalen Medium. Der Beitrag führt in diese neuen Möglichkeiten ein und erklärt, wie sie in didaktischen DaF/DaZ-Kontexten aber auch hinsichtlich forschungsbezogener Perspektiven angewendet werden können
Telephone-based remote interpreting has come into widespread use in multilingual encounters, all the more so in times of refugee crises and the large influx of asylum-seekers into Europe. Nevertheless, the linguistic practices in this mode of communication have not yet been examined comprehensively. This article therefore investigates selected aspects of turn-taking and clarification sequences during semi-authentic telephone-interpreted counselling sessions for refugees (Arabic–German). A quantitative analysis reveals that limited audibility makes it more difficult for interpreters to claim their turn successfully; in most cases, however, turn-taking occurs smoothly. The trouble sources that trigger queries are mainly content-related and interpreters vary greatly in the ways they deal with such difficulties. Contrary to what one might expect, the study shows that coordination fails only rarely during telephone-based remote interpreting.
This paper presents an extended annotation and analysis of interpretative reply relations focusing on a comparison of reply relation types and targets between conflictual pages and neutral pages of German Wikipedia (WP) talk pages. We briefly present the different categories identified for interpretative reply relations to analyze the relationship between WP postings as well as linguistic cues for each category. We investigate referencing strategies of WP authors in discussion page postings, illustrated by means of reply relation types and targets taking into account the degree of disagreement displayed on a WP talk page. We provide richly annotated data that can be used for further analyses such as the identification of interactional relations on higher levels, or for training tasks in machine learning algorithms.
ZuRecht steht für Zugang zur Recherche in Transkripten. Es handelt sich um eine prototypische Implementierung einer webbasierten grafischen Benutzeroberfläche, welche Zugriff auf Transkripte gesprochener Sprache aus dem Archiv für Gesprochenes Deutsch (AGD) des Leibniz-Instituts für Deutsche Sprache (IDS) bietet. Der Zugriff erfolgt über die neue, im Projekt „ZuMult“ entwickelte Schnittstelle zur Suche in mündlichen Korpora. ZuRecht dient einerseits der Demonstration der Möglichkeiten der neuen Schnittstelle, indem es komplexe Suchanfragen mit der speziell für die Korpusrecherche entwickelten Anfragesprache CQP auf Transkriptionen gesprochener Sprache erlaubt. Andererseits kommt ZuRecht als Erweiterung der Datenbank für Gesprochenes Deutsch (DGD) zum Einsatz und eröffnet den DGD-Nutzer:innen viele neue Forschungsmöglichkeiten, insbesondere auf den Gebieten der Gesprächsanalyse und der DaF/DaZ-bezogenen Forschung. Im Beitrag werden die Funktionalitäten von ZuRecht ausführlich vorgestellt und ihre Einsatzmöglichkeiten in den genannten Disziplinen exemplarisch vorgeführt.
Im vorliegenden Artikel wird ein Überblick über das von der DFG geförderte Projekt Zugänge zu multimodalen Korpora gesprochener Sprache – Vernetzung und zielgruppenspezifische Ausdifferenzierung (ZuMult) gegeben. Dabei wird zunächst auf die Sprachdaten und auf die technische Basis der Applikationen eingegangen, die dem Projekt zugrunde liegen. Im Anschluss werden die weiteren Beiträge in diesem Themenheft von KorDaF kurz vorgestellt. Übergeordnetes Thema von ZuMult ist die Verbesserung der Zugänglichkeit von digitalen mündlichen Sprachdaten für verschiedene Anwendungen und Zielgruppen, wobei der Fokus dieses Themenhefts auf Applikationen und Anwender:innen aus der Fremdsprachendidaktik und der DaF-/DaZ-Forschung und -Lehre liegt. Die einzelnen Beiträge beleuchten zentrale methodische und/oder technische Aspekte dieses Themas und beschreiben die Architektur und verschiedene prototypische Anwendungen, die das Projekt entwickelt hat.
Als Teil der NFDI vernetzt Text+ ortsverteilt verschiedenste Daten und Dienste für die geisteswissenschaftliche Forschung und stellt sie der wissenschaftlichen Gemeinschaft FAIR zur Verfügung. In diesem Beitrag beschreiben wir die Umsetzung beispielhaft im Bereich der Text+ Datendomäne Sammlungen anhand von Korpora, die in verschiedenen Disziplinen Verwendung finden. Die Infrastruktur ist auf Erweiterbarkeit ausgelegt, so dass auch weitere Ressourcen über Text+ verfügbar gemacht werden können. Enthalten ist auch ein Ausblick auf weitere zu erwartende Entwicklungen. Ein Beitrag zur 9. Tagung des Verbands "Digital Humanities im deutschsprachigen Raum" - DHd 2023 Open Humanities Open Culture.
Das Forschungs- und Lehrkorpus für GesprochenesDeutsch (FOLK) ist ein Korpus des gesprochenen Deutsch in natürlichen sozialen Interaktionen, das seit 2008 in der Abteilung Pragmatik am Leibniz-Institut für Deutsche Sprache in Mannheim aufgebaut wird. FOLK besteht aus Audio- und Videoaufzeichnungen natürlicher Gespräche aus verschiedensten gesellschaftlichen Bereichen (private, institutionelle und öffentliche Interaktionsdomäne), die durch Transkription, weitere Annotationen und Metadaten-Dokumentation für korpusgestützte Analysen erschlossen und zur wissenschaftlichen Nutzung bereitgestellt werden. FOLK wird auf vielfältige Weise für Untersuchungen zum gesprochenen Deutsch genutzt, insbesondere in der Gesprächsforschung, der Korpuslinguistik und anwendungsorientierten Zweigen der Linguistik.
This paper introduces the Nottinghamer Korpus deutscher YouTube-Sprache (‘The Nottingham German YouTube Language Corpus’ - or NottDeuYTSch corpus). The corpus comprises over 33 million words, taken from roughly 3 million YouTube comments published between 2008 and 2018, written by a young, German-speaking demographic. The NottDeuYTSch corpus provides an authentic and representative linguistic snapshot of young German speakers and offers significant opportunities for in-depth research in several linguistic fields, such as lexis, morphology, syntax, orthography, multilingualism, and conversational and discursive analysis.
The NottDeuYTSch corpus is a freely available collection of YouTube comments written under German-speaking videos by young people between 2008 and 2018. The article uses the NottDeuYTSch corpus to investigate how YouTube comments can be used to produce learning materials and how corpora of Digitally-Mediated Communication can benefit intermediate learners of German. The article details the effects of authentic communication within YouTube comments on teenage learners, examining how they can influence the psycholinguistic factors of motivation, foreign language anxiety, and willingness to communicate. The article also discusses the benefits and limitations of using authentic corpus material for the development of teaching material.
Unter Neologismen finden sich bedeutungsgleiche Ausdrücke (im weitesten Sinne Synonyme), die unter bestimmten Bedingungen sprachliche Unsicherheiten hervorrufen. Das liegt u. a. an ihrer semantisch-konzeptuellen Ähnlichkeit, an nicht abgeschlossenen Lexikalisierungsprozessen, aber es treten auch Zweifel auf, weil es Unterschiede zwischen der Allgemein- und der Fachsprache gibt. Für einige Neologismen ist es auch charakteristisch, dass mehrere morphologische Varianten gleichzeitig in den Wortschatz eintreten, sodass nicht immer klar ist, wann welche präferiert werden. Dass all diese Ausdrücke lexikalischem Wettbewerb und situationsgebundenen Gebrauchsbedingungen ausgesetzt sind und dass sie zu Zweifel führen können, wird in Onlineforen sichtbar. Dieser Beitrag beschäftigt sich mit der Frage, wie solche Paare/Gruppen korpusgestützt semantisch analysiert und wie sie in deskriptiven Wörterbüchern angemessen beschrieben werden können, um sowohl Gemeinsamkeiten als auch Unterschiede für Nachschlagende sichtbar zu machen. Dazu werden konkrete Beispiele und ein gegenüberstellendes Wörterbuchdarstellungsformat für neologistische Synonyme vorgeschlagen.
Dieser Werkstattbericht zeigt anhand verschiedener korpusbasierter Ressourcen, wie Fragen zu sprachlichen Phänomenen, die für Sprachlernende nicht oder nur unzureichend dokumentiert sind, empirisch beantwortet werden können. Besonderes Augenmerk wird dabei auf OWIDplusLIVE gelegt. Hierbei handelt es sich um ein Werkzeug zur tagesaktuellen Analyse von Token (einzelne Wortformen/Lemmata) und Bi-/Trigrammen (zwei bzw. drei direkt aufeinander folgende Token). Über eine Anbindung an KorAP können zudem Belege aus dem DeReKo (Deutsches Referenzkorpus) abgerufen und analysiert werden.
In many countries of the world, perspectives on gender equality and racism have changed in recent decades. One result has been more attention being devoted to traces of androcentric and racist language in society. This also affects dictionaries. In lexicography there are discussions about whether or to what extent social asymmetries are inscribed in dictionaries and if this is still acceptable. The issue of the nature of description plays an important role in this discussion. If sexist usages are often found in language use, i.e. in the corpus data on which the dictionary is based, does the dictionary also have to show them? How is this, in turn, compatible with the normative power of dictionaries? Do dictionaries contribute to the perpetuation of gender stereotypes by showcasing them under the banner of descriptive principles? And what roles do lexicographers play in this process? The article deals with these questions on the basis of individual lexicographical examples and current discussions in the lexicographic and public community.
It is well known that the distribution of lexical and grammatical patterns is size- and register-sensitive (Biber 1986, and later publications). This fact alone presents a challenge to many corpus-oriented linguistic studies focusing on a single language. When it comes to cross-linguistic studies using corpora, the challenge becomes even greater due to the lack of high-quality multilingual corpora (Kupietz et al. 2020; Kupietz/Trawiński 2022), which are comparable with respect to the size and the register. That was the motivation for the creation of the European Reference Corpus EuReCo, an initiative started in 2013 at the Leibniz Institute for the German Language (IDS) together with several European partners (Kupietz et al. 2020). EuReCo is an emerging federated corpus, with large virtual comparable corpora across various languages and with an infrastructure supporting contrastive research. The core of the infrastructure is KorAP (Diewald et al. 2016), a scalable open-source platform supporting the analysis and visualisation of properties of texts annotated by multiple and potentially conflicting information layers, and supporting several corpus query languages. Until recently, EuReCo consisted of three monolingual subparts: the German Reference Corpus DeReKo (Kupietz et al. 2018), the Reference Corpus of Contemporary Romanian Language (Barbu Mititelu/Tufiş/Irimia 2018), and the Hungarian National Corpus (Váradi 2002). The goal of the present submission is twofold. On the one hand, it reports about the new component of EuReCo: a sample of the National Corpus of Polish (Przepiórkowski et al. 2010). On the other hand, it presents the results of a new pilot study using the newly extended EuReCo. This pilot study investigates selected Polish collocations involving light verbs and their prepositional / nominal complements (Fig. 1) and extends the collocation analyses of German, Romanian and Hungarian (Fig. 2) discussed in Kupietz/Trawiński (2022).
Following the successes of the ninth conference in 2022 held in the wonderful Santiago de Compostela, Spain, we are pleased to present the proceedings of the 10th edition of International Conference on CMC and Social Media Corpora for the Humanities (CMC-2023). The focal point of
the conference is to investigate the collection, annotation, processing, and analysis of corpora of computer-mediated communication (CMC) and social media.
Our goal is to serve as the meeting place for a wide variety of language-oriented investigations into CMC and social media from the fields of linguistics, philology, communication sciences, media
studies, and social sciences, as well as corpus and computational linguistics, language technology, textual technology, and machine learning.
This year’s event is the largest so far with 45 accepted submissions: 32 papers and 13 poster presentations, each of which were reviewed by members of our ever-growing scientific committee. The contributions were presented in five sessions of two or three streams, and a single poster session. The talks in these proceedings cover a wide range of topics, including the corpora construction, digital identities, digital knowledge-building, digitally-mediated interaction, features
of digitally-mediated communication, and multimodality in digital spaces.
As part of the conference, we were delighted to include two invited talks: an international keynote speech by Unn Røyneland from the University of Oslo, Norway, on the practices and perceptions of
researching dialect writing in social media, and a national keynote speech by Tatjana Scheffler from the Ruhr-University of Bochum on analysing individual linguistic variability in social media and
constructing corpora from this data. Additionally, participants could take part in a workshop on processing audio data for corpus linguistic analysis. This volume contains abstracts of the invited talks, short papers of oral presentations, and abstracts of posters presented at the conference.
Funktionsverbgefüge stehen seit jeher in der Sprachkritik, die sich nun auch auf digitale Räume ausbreitet. Vertreten wird dort die These, Funktionsverbgefüge und ihre entsprechenden Basisverben seien äquivalent und könnten in allen Kontexten durch die verbalen Entsprechungen ersetzt werden. Dies kann durch die vorliegende korpusbasierte und textlinguistische Studie am Beispiel des Gefüges Frage stellen widerlegt werden. Anhand eines extensiven Datenmaterials aus den Wikipedia-Artikel-Korpora des IDS zeige ich die semantischen, grammatischen und textlinguistischen Unterschiede zwischen dem Basisverb und dem Funktionsverbgefüge im Gebrauch auf, die sich in der Anreicherung, Verdichtung, Perspektivierung, Gewichtung und Wiederaufnahme von Informationen im Text manifestieren.
In this article, we provide an insight into the development and application of a corpus-lexicographic tool for finding neologisms that are not yet listed in German dictionaries. As a starting point, we used the words listed in a glossary of German neologisms surrounding the COVID-19 pandemic. These words are lemma candidates for a new dictionary on COVID-19 discourse in German. They also provided the database used to develop and test the NeoRate tool. We report on the lexicographic work in our dictionary project, the design and functionalities of NeoRate, and describe the first test results with the tool, in particular with regard to previously unregistered words. Finally, we discuss further development of the tool and its possible applications.
In diesem Beitrag beschäftigen wir uns mit moralisierenden Sprachhandlungen, worunter wir diskursstrategische Verfahren verstehen, in denen die Beschreibung von Streitfragen und erforderlichen Handlungen mit moralischen Begriffen enggeführt werden. Auf moralische Werte verweisendes Vokabular (wie beispielsweise „Freiheit“, „Sicherheit“ oder „Glaubwürdigkeit“) wird dabei verwendet, um eine Forderung durchzusetzen, die auf diese Weise unhintergehbar erscheint und keiner weiteren Begründung oder Rechtfertigung bedarf. Im Fokus unserer Betrachtungen steht dementsprechend das aus pragma-linguistischer Sicht auffällige Phänomen einer spezifischen Redepraxis der Letztbegründung oder Unhintergehbarkeit, die wir als Pragmem auffassen und beschreiben. Hierfür skizzieren wir zunächst den in der linguistischen Pragmatik verorteten Zugang zu Praktiken der Moralisierung, betrachten sprachliche Formen des Moralisierens und deren kotextuellen und insbesondere pragma-syntaktischen Struktureinbettungen, um anschließend Hypothesen zu kontextuellen Wirkungsfunktionen aufzustellen. Darauf basierend leiten wir schließlich anhand von exemplarischen Korpusbelegen Strukturmuster des Moralisierens ab, die wir in dem Terminus „Pragmem“ verdichten und mittels qualitativer und quantitativer Analysen operationalisieren.
When comparing different tools in the field of natural language processing (NLP), the quality of their results usually has first priority. This is also true for tokenization. In the context of large and diverse corpora for linguistic research purposes, however, other criteria also play a role – not least sufficient speed to process the data in an acceptable amount of time. In this paper we evaluate several state of the art tokenization tools for German – including our own – with regard to theses criteria. We conclude that while not all tools are applicable in this setting, no compromises regarding quality need to be made.
This paper describes a method for extracting collocation data from text corpora based on a formal definition of syntactic structures, which takes into account not only the POS-tagging level of annotation but also syntactic parsing (syntactic treebank model) and introduces the possibility of controlling the canonical form of extracted collocations based on statistical data on forms with different properties in the corpus. Specifically, we describe the results of extraction from the syntactically tagged Gigafida 2.1 corpus. Using the new method, 4,002,918 collocation candidates in 81 syntactic structures were extracted. We evaluate the extracted data sample in more detail, mainly in relation to properties that affect the extraction of canonical forms: definiteness in adjectival collocations, grammatical number in noun collocations, comparison in adjectival and adverbial collocations, and letter case (uppercase and lowercase) in canonical forms. The conclusion highlights the potential of the methodology used for the grammatical description of collocation and phrasal syntax and the possibilities for improving the model in the process of compilation of a digital dictionary database for Slovene.
This paper presents a multilingual dictionary project of discourse markers. During its first stage, consisting of collecting the list of headwords, we used a parallel corpus to automatically extract units from texts written in Spanish, Catalan, English, French and German. We also applied a method to create a taxonomy structure for automatically organising the markers in clusters. As a result, we obtain an extensive, corpus-driven list of headwords. We present a prototype of the microstructure of the dictionary in the form of a standard XML database and describe the procedure to automatically fill in most of its fields (e.g., the type of DM, the equivalents in other languages, etc.), before human intervention.
In this paper we present Trendi, a monitor corpus of written Slovene, which has been compiled recently as part of the SLED (Monitor corpus and related resources) project. The methodology and the contents of the corpus are presented, as well as the findings of the survey that aimed to identify the needs of potential users related to topical language use. The Trendi corpus currently contains news articles and other web content from 110 different sources, with the texts being collected and linguistically annotated on a daily basis. The corpus complements Gigafida 2.0, a 1.13-billion-word reference corpus of standard written Slovene. Also discussed are the ways in which the corpus will be integrated into various lexicographic projects, helping not only in the identification of neologisms but also in monitoring changes in already identified language phenomena.
In this paper, we propose a controlled language for authoring technical documents and report the status of its development, while maintaining a specific focus on the Japanese automotive domain. To reduce writing variations, our controlled language not only defines approved and unapproved lexical elements but also prescribes their preferred location in a sentence. It consists of components of a) case frames, b) case elements, c) adverbial modifiers, d) sentence-ending functions, and e) connectives, which have been developed based on the thorough analyses of a large-scale text corpus of automobile repair manuals. We also present our prototype of a writing assistant tool that implements word substitution and reordering functions, incorporating the constructed controlled language.
This paper presents the project “The first Romanian bilingual dictionaries (17th century). Digitally annotated and aligned corpus” (eRomLex) which deals with the editing of the first bilingual Romanian dictionaries. The aim of the project is to compile an electronic corpus comprising six Slavonic-Romanian lexicons dating from the 17th century, based on their relatedness and the fact that they follow a common model in order to highlight the characteristics of this lexicographical network (the affiliations between the lexicons, the way they relate to the source, the innovations towards it, their potential uses) and to facilitate the access to their content. A digital edition allows exhaustive data extraction and comparison and link with other digitized resources for old Romanian or Church Slavonic, including dictionaries. After presenting the corpus, we point to the necessary stages in achieving this project, the techniques used to access the material and the challenges and obstacles we encountered along the way. We describe how the corpus was created, stored, indexed and can be searched over; we will also present and discuss some statistical analyses highlighting relations between the Romanian lexicons and their Slavonic-Ruthenian source.
This paper investigates the long-term diachronic development of the perfect and preterite tenses in German and provides a novel analysis by supplementing Reichenbach’s (1947) classical theory of tense by the notion of underspecification. Based on a newly compiled parallel corpus spanning the entire documented history of German, we show that the development in question is cyclic: It starts out with only one tense form (preterite) compatible with both current relevance and narrative past readings in (early) Old High German and, via three intermediate stages, arrives at only one tense form again (perfect) compatible with the same readings in modern Upper German dialects. We propose that in order to capture all attested stages we must allow tenses to be unspecified for R (reference time), with R merely being inferred pragmatically. We then propose that the transitions between the different stages can be explained by the interplay between semantics and pragmatics.
Inspired by GWLN 3, we take a look at the new words, meanings, and expressions that have been created during or promoted by the COVID-19 pandemic. The pandemic provides a rare opportunity to follow the rise, spread, and integration of words and expressions in a language that may serve as an illustration of how linguistic innovation in general works. Relevant words were selected from various lists, notably monthly and annual lists of prominent words attested in the corpus of The Danish Dictionary. Analysis of these lists gives an insight into the number of words that stand out month by month and what kinds of words are involved, both in terms of morphological type and of semantic category, with special attention given to neologisms. Finally, we discuss the criteria for selecting which words to include in the dictionary. With this study, Danish is added to the list of languages covered in the GWLN series on
COVID-19 neologisms.
This paper discusses an investigation of how senses are ordered across eight dictionaries. A dataset of 75 words was used for this purpose, and two senses were examined for each word. The words are divided into three groups of 25 words each according to the relationship between the senses: Homonymy, Metaphor, and Systematic Polysemy. The primary finding is that WordNet differs from the other dictionaries in terms of Metaphor. The order of the senses was more often figurative/literal, and it had the highest percentage of figurative senses that were not found. We discuss leveraging another dictionary, COBUILD, to re-order the senses according to frequency.
Vergleichbare Korpora für multilinguale kontrastive Studien. Herausforderungen und Desiderata
(2022)
This contribution aims to show the necessity of working in the development of multilingual corpora and appropriate tools for multilingual contrastive studies. We take the corpus of the lexicographical project COMBIDIGILEX as example to show, how difficultit is to build a suitable data basis to study and compare linguistic phenomena in German, Spanish and Portuguese. Despite the availability of big reference corpora for the three languages (at least for written language), it is not able to obtain a comparable data basis from, because the mentioned corpora are created according to different requirements and they are also powered by disparate information systems and analyse tools. To break the status quo, we plead for increasing research infrastructures by means of compatible language technology and sharing data.
The article investigates the hypothesis that prominence phenomena on different levels of linguistic structure are systematically related to each other. More specifically, it is hypothesized that prominence relations in morphosyntax reflect, and contribute to, prominence management in discourse. This hypothesis is empirically based on the phenomenon of agentivity clines, i.e. the observation that the relevance of agentivity features such as volition or sentience is variable across different constructions. While some constructions, including German DO-clefts, show a strong preference for highly agentive verbs, other constructions, including German basic active constructions, have no particular requirements regarding the agentivity of the verb, except that at least one agentivity feature should be present. Our hypothesis predicts that this variable relevance of agentivity features is related to the discourse constraints on the felicitous use of a given construction, which in turn, of course, requires an explicit statement of such constraints. We propose an original account of the discourse constraints on DO-clefts in German using the ‘Question Under Discussion’ framework. Here, we hypothesize that DO-clefts render prominent one implicit question from a set of alternative questions available at a particular point in the developing discourse. This then yields a prominent question-answer pair that changes the thematic structure of the discourse. We conclude with some observations on the possibility of relating morphosyntactic prominence (high agentivity) to discourse prominence (making a Question Under Discussion prominent by way of clefting).
So far, Sepedi negations have been considered more from the point of view of lexicographical treatment. Theoretical works on Sepedi have been used for this purpose, setting as an objective a neat description of these negations in a (paper) dictionary. This paper is from a different perspective: instead of theoretical works, corpus linguistic methods are used: (1) a Sepedi corpus is examined on the basis of existing descriptions of the occurrences of a relevant verb, looking at its negated forms from a purely prescriptive point of view; (2) a "corpus-driven" strategy is employed, looking only for sequences of negation particles (or morphemes) in order to list occurring constructions, without taking into account the verbs occurring in them, apart from their endings. The approach in (2) is only intended to show a possible methodology to extend existing theories on occurring negations. We would also like to try to help lexicographers to establish a frequency-based order of entries of possible negation forms in their dictionaries by showing them the number of respective occurrences. As with all corpus linguistic work, however, we must regard corpus evidence not as representative, but as tendencies of language use that can be detected and described. This is especially true for Sepedi, for which only few and small corpora exist. This paper also describes the resources and tools used to create the necessary corpus and also how it was annotated with part of speech and lemmas. Exploring the quality of available Sepedi part-of-speech taggers concerning verbs, negation morphemes and subject concords may be a positive side result.
Kontrastive Korpuslinguistik
(2022)