Hasil untuk "German literature"

Nina Smirnova, Daniel Dan, Philipp Mayr

Parliamentary debate constitutes a central arena of political power, shaping legislative outcomes and public discourse. Incivility within this arena signals political polarization and institutional conflict. This study presents a systematic investigation of incivility in the German Bundestag by examining calls to order (CtO; plural: CtOs) as formal indicators of norm violations. Despite their relevance, CtOs have received little systematic attention in parliamentary research. We introduce a rule-based method for detecting and annotating CtOs in parliamentary speeches and present a novel dataset of German parliamentary debates spanning 72 years that includes annotated CtO instances. Additionally, we develop the first classification system for CtO triggers and analyze the factors associated with their occurrence. Our findings show that, despite formal regulations, the issuance of CtOs is partly subjective and influenced by session presidents and parliamentary dynamics, with certain individuals disproportionately affected. An insult towards individuals is the most frequent cause of CtO. In general, male members and those belonging to opposition parties receive more calls to order than their female and coalition-party counterparts. Most CtO triggers were detected in speeches dedicated to governmental affairs and actions of the presidency. The CtO triggers dataset is available at: https://github.com/kalawinka/cto_analysis.

Julia Wunderle, Anton Ehrmanntraut, Jan Pfister et al.

Encoders remain essential for efficient German NLP and NLU scenarios despite the rise of decoder-only LLMs. This work studies two routes to high-quality German encoders under identical data and training constraints: 1) training from scratch and 2) converting decoders via LLM2Vec. We introduce two resources: ModernGBERT (134M, 1B), fully transparent German encoders in the ModernBERT style, and LLäMmleinVec (120M, 1B, 7B), decoder-to-encoder conversions trained with masked next-token prediction, both undergoing a context extension to 8.192 tokens. Across SuperGLEBer, ModernGBERT 1B sets a new state of the art (avg 0.808), surpassing GBERT Large (+4%) and the seven-times larger converted 7B model (0.787). On German MTEB after supervised fine-tuning, ModernGBERT 1B (0.551) approaches the converted 7B model (0.557). We release all models, checkpoints, datasets, and full training records, and introduce an encoder-adapted QA-NIAH evaluation. All in all, our results provide actionable guidance: when parameter efficiency and latency matter, from-scratch encoders dominate. When a pre-trained decoder exists and compute is a limited, conversion offers an effective alternative. ModernGBERT and LLäMmleinVec, including all code, data and intermediary checkpoints are published under a research-only RAIL license.

Isabelle van der Vegt, Bennett Kleinberg, Marilu Miotto et al.

This paper introduces and evaluates three translations of the Grievance Dictionary, a psycholinguistic dictionary for the analysis of violent, threatening or grievance-fuelled texts. Considering the relevance of these themes in languages beyond English, we translated the Grievance Dictionary to Dutch, German, and Italian. We describe the process of automated translation supplemented by human annotation. Psychometric analyses are performed, including internal reliability of dictionary categories and correlations with the LIWC dictionary. The Dutch and German translations perform similarly to the original English version, whereas the Italian dictionary shows low reliability for some categories. Finally, we make suggestions for further validation and application of the dictionary, as well as for future dictionary translations following a similar approach.

Mathias Angermaier, Elisabeth Hoeldrich, Jana Lasser et al.

Sociality borne by language, as is the predominant digital trace on text-based social media platforms, harbours the raw material for exploring multiple social phenomena. Distinctively, the messaging service Telegram provides functionalities that allow for socially interactive as well as one-to-many communication. Our Telegram dataset contains over 6,000 groups and channels, 40 million text messages, and over 3 million transcribed audio files, originating from a data-hoarding initiative named the ``Schwurbelarchiv'' (from German schwurbeln: speaking nonsense). This dataset publication details the structure, scope, and methodological specifics of the Schwurbelarchiv, emphasising its relevance for further research on the German-language conspiracy theory discourse. We validate its predominantly German origin by linguistic and temporal markers and situate it within the context of similar datasets. We describe process and extent of the transcription of multimedia files. Thanks to this effort the dataset uniquely supports multimodal analysis of online social dynamics and content dissemination. Researchers can employ this resource to explore societal dynamics in misinformation, political extremism, opinion adaptation, and social network structures on Telegram. The Schwurbelarchiv thus offers unprecedented opportunities for investigations into digital communication and its societal implications.

Bastian Bunzeck, Daniel Duran, Sina Zarrieß

We analyze the influence of utterance-level construction distributions in German child-directed/child-available speech on the resulting word-level, syntactic and semantic competence (and their underlying learning trajectories) in small LMs, which we train on a novel collection of developmentally plausible language data for German. We find that trajectories are surprisingly robust for markedly different distributions of constructions in the training data, which have little effect on final accuracies and almost no effect on global learning trajectories. While syntax learning benefits from more complex utterances, word-level learning culminates in better scores with more fragmentary utterances. We argue that LMs trained on developmentally plausible data can contribute to debates on how conducive different kinds of linguistic stimuli are to language learning.

Lu Kalkbrenner, Veronika Solopova, Steffen Zeiler et al.

Connectivity and message propagation are central, yet often underutilized, sources of information in misinformation detection -- especially on poorly moderated platforms such as Telegram, which has become a critical channel for misinformation dissemination, namely in the German electoral context. In this paper, we introduce Misinfo-TeleGraph, the first German-language Telegram-based graph dataset for misinformation detection. It includes over 5 million messages from public channels, enriched with metadata, channel relationships, and both weak and strong labels. These labels are derived via semantic similarity to fact-checks and news articles using M3-embeddings, as well as manual annotation. To establish reproducible baselines, we evaluate both text-only models and graph neural networks (GNNs) that incorporate message forwarding as a network structure. Our results show that GraphSAGE with LSTM aggregation significantly outperforms text-only baselines in terms of Matthews Correlation Coefficient (MCC) and F1-score. We further evaluate the impact of subscribers, view counts, and automatically versus human-created labels on performance, and highlight both the potential and challenges of weak supervision in this domain. This work provides a reproducible benchmark and open dataset for future research on misinformation detection in German-language Telegram networks and other low-moderation social platforms.

Felix Krejca, Tobias Kietreiber, Alexander Buchelt et al.

The increasing volume of online discussions requires advanced automatic content moderation to maintain responsible discourse. While hate speech detection on social media is well-studied, research on German-language newspaper forums remains limited. Existing studies often neglect platform-specific context, such as user history and article themes. This paper addresses this gap by developing and evaluating binary classification models for automatic content moderation in German newspaper forums, incorporating contextual information. Using LSTM, CNN, and ChatGPT-3.5 Turbo, and leveraging the One Million Posts Corpus from the Austrian newspaper Der Standard, we assess the impact of context-aware models. Results show that CNN and LSTM models benefit from contextual information and perform competitively with state-of-the-art approaches. In contrast, ChatGPT's zero-shot classification does not improve with added context and underperforms.

Alejandro Godinez

We present HySemRAG, a framework that combines Extract, Transform, Load (ETL) pipelines with Retrieval-Augmented Generation (RAG) to automate large-scale literature synthesis and identify methodological research gaps. The system addresses limitations in existing RAG architectures through a multi-layered approach: hybrid retrieval combining semantic search, keyword filtering, and knowledge graph traversal; an agentic self-correction framework with iterative quality assurance; and post-hoc citation verification ensuring complete traceability. Our implementation processes scholarly literature through eight integrated stages: multi-source metadata acquisition, asynchronous PDF retrieval, custom document layout analysis using modified Docling architecture, bibliographic management, LLM-based field extraction, topic modeling, semantic unification, and knowledge graph construction. The system creates dual data products - a Neo4j knowledge graph enabling complex relationship queries and Qdrant vector collections supporting semantic search - serving as foundational infrastructure for verifiable information synthesis. Evaluation across 643 observations from 60 testing sessions demonstrates structured field extraction achieving 35.1% higher semantic similarity scores (0.655 $\pm$ 0.178) compared to PDF chunking approaches (0.485 $\pm$ 0.204, p < 0.000001). The agentic quality assurance mechanism achieves 68.3% single-pass success rates with 99.0% citation accuracy in validated responses. Applied to geospatial epidemiology literature on ozone exposure and cardiovascular disease, the system identifies methodological trends and research gaps, demonstrating broad applicability across scientific domains for accelerating evidence synthesis and discovery.

Sara Herreras Martínez, Justus Mesman, Daniel Møller Sneum et al.

Abstract Background While energy communities working on electricity provision have been extensively studied, thermal energy communities (TECs) focusing on bringing district heating (DH) systems to decarbonise heat systems in buildings have been relatively under-researched. This study addresses this gap by presenting the first comprehensive examination of key factors influencing the emergence and development of TEC projects in Denmark, Germany, and the Netherlands. The study uses an established analytical framework from previous research encompassing seven dimensions: market structure, hard- and soft institutions, financing, physical infrastructure, capacity, and interactions with other stakeholders. Data are gathered through a literature review and interviews. Results TECs have emerged at different times in each country, shaped by contextual circumstances and diverse forms of institutional support. Elements that have supported the development of TECs are regulatory frameworks promoting DH growth, heat decarbonisation policies, economic incentives to use waste heat in plants, targeted financing mechanisms, and assistance to enhance the capacity of TECs. External factors such as high oil prices, seismic events, and recent rising energy prices have also spurred project initiation. TECs also rely on additional factors for success, including organisational and entrepreneurial abilities to engage with stakeholders, gain social acceptance, and secure commitment from community members. Involvement from local government, intermediary organisations, and private companies is crucial for TEC implementation. Among the studied countries, Danish TECs stand out as the most developed, benefiting from a stable policy environment, decades of experience with DH and TEC, and positive societal perceptions. Conversely, Dutch and German TECs face challenges because of the early stage of their heat transition, dealing with financial obstacles, underdeveloped policies, unfamiliarity with DH technology and with TECs, as well as the need for expensive infrastructure changes. Shared challenges across regions include capacity limitations in small projects and implementing cost-effective, local, and sustainable heat sources. Conclusions In light of the study's findings, policymakers must consider establishing stable, integral and flexible policies supporting heat decarbonisation and TECs, addressing TECs' reliance on limited capacities, involving TECs in local heat municipal plans, and facilitating high DH connection rates where DH is the most cost-effective solution from a socio-economic perspective.

Maria Rita Digilio

In ancient and medieval poetry the adjective was often treated as a conventional addition to a given noun. As a consequence, its use was not always aimed at semantic clarification and had no ambition of poetic embellishment. Moreover, the typical recourse to formulas determined the preservation of fixed expressions which in religious texts attained to a sort of sacral fixity, and this is certainly the case in Heliand. At the same time, it seems quite evident that the author of the Old Saxon poem tries to give adjectives a different claim. The use of adjectives in Heliand is to a certain extent determined by the Germanic verse tradition – based on the alliterative pattern, formulas and variation – but it very often assumes unconventional and creative traits, which are fully coherent with the general concept and the stylistic strategies of the poem.

Jan Pfister, Julia Wunderle, Andreas Hotho

We create two German-only decoder models, LLäMmlein 120M and 1B, transparently from scratch and publish them, along with the training data, for the German NLP research community to use. The model training involved several key steps, including extensive data preprocessing, the creation of a custom German tokenizer, the training itself, as well as the evaluation of the final models on various benchmarks. Throughout the training process, multiple checkpoints were saved and analyzed using the SuperGLEBer benchmark to monitor the models' learning dynamics. Compared to state-of-the-art models on the SuperGLEBer benchmark, both LLäMmlein models performed competitively, consistently matching or surpassing models with similar parameter sizes. The results show that the models' quality scales with size as expected, but performance improvements on some tasks plateaued early, offering valuable insights into resource allocation for future model development.

Markus Lepper, Baltasar Trancón Widemann

TScore is both an abstract formalism and its computer implementation to construct models of arbitrary kinds of time-related data. It is a research project about the semantics of musical notation, applying the method of computer-aided re-modelling to diverse formalisms and semantics of time-related data. Here we present the application to German tablature notation. While the current implemention is merely a proof of concept, the lean architecture of TScore allows easy adaptation and extension.

Janek Herrlein, Chia-Chien Hung, Goran Glavaš

Research on token-level reference-free hallucination detection has predominantly focused on English, primarily due to the scarcity of robust datasets in other languages. This has hindered systematic investigations into the effectiveness of cross-lingual transfer for this important NLP application. To address this gap, we introduce ANHALTEN, a new evaluation dataset that extends the English hallucination detection dataset to German. To the best of our knowledge, this is the first work that explores cross-lingual transfer for token-level reference-free hallucination detection. ANHALTEN contains gold annotations in German that are parallel (i.e., directly comparable to the original English instances). We benchmark several prominent cross-lingual transfer approaches, demonstrating that larger context length leads to better hallucination detection in German, even without succeeding context. Importantly, we show that the sample-efficient few-shot transfer is the most effective approach in most setups. This highlights the practical benefits of minimal annotation effort in the target language for reference-free hallucination detection. Aiming to catalyze future research on cross-lingual token-level reference-free hallucination detection, we make ANHALTEN publicly available: https://github.com/janekh24/anhalten

Mario Bossler, Lars Chittka, Thorsten Schank

We present the first empirical evidence on the 22 percent increase in the German minimum wage, implemented in 2022, raising it from Euro 9.82 to 10.45 in July and to Euro 12 in October. Leveraging the German Earnings Survey, a large and novel data source comprising around 8 million employee-level observations reported by employers each month, we apply a difference-in-difference-in-differences approach to analyze the policy's impact on hourly wages, monthly earnings, employment, and working hours. Our findings reveal significant positive effects on wages, affirming the policy's intended benefits for low-wage workers. Interestingly, we identify a negative effect on working hours, mainly driven by minijobbers. The hours effect results in an implied labor demand elasticity in terms of the employment volume of -0.2 which only partially offsets the monthly wage gains. We neither observe a negative effect on the individual's employment retention nor the regional employment levels.

Sandra Hofert

Abstract: This article, which continues ideas developed in the context of the Deutsche Forschungsgemeinschaft: Graduiertenkolleg 1876—215342465 (GRK1876), examines how animals are used in medieval texts to (re)present, shape, and develop the literary representation of emotions. On the basis of selected examples, it shows how diverse the literary functions of animal imagery can be and how many different poetic and aesthetic strategies can be found for staging animals, connecting them with human characters and the recipients of the tale. In this way, animals can serve as objects of cultural self-reflection and as models for philosophical orientation. by Sandra Hofert Friedrich-Alexander-University Erlangen-Nuremberg sandra.hofert@fau.de

Sarah Wingert, Christian Becker-Asano

When researching on the acceptance of robots in Human-Robot-Interaction the Uncanny Valley needs to be considered. Reusable and standardized measures for it are essential. In this paper one such questionnaire got translated into German. The translated indices got evaluated (n=140) for reliability with Cronbach's alpha. Additionally the items were tested with an exploratory and a confirmatory factor analysis for problematic correlations. The results yield a good reliability for the translated indices and showed some items that need to be further checked.

Anna Czypionka, Mariya Kharaman, Carsten Eulitz

Animacy is an intrinsic semantic property of words referring to living things. A long line of evidence shows that words with animate referents require lower processing costs during word recognition than words with inanimate referents, leading among others to a decreased N400 amplitude in reaction to animate relative to inanimate objects. In the current study, we use this animacy effect to provide evidence for access to the semantic properties of constituents in German noun-noun compounds. While morphological decomposition of noun-noun compounds is well-researched and illustrated by the robust influence of lexical constituent properties like constituent length and frequency, findings for semantic decomposition are less clear in the current literature. By manipulating the animacy of compound modifiers and heads, we are able to manipulate the relative ease of lexical access strictly due to intrinsic semantic properties of the constituents. Our results show additive effects of constituent animacy, with a higher number of animate constituents leading to gradually attenuated N400 amplitudes. We discuss the implications of our findings for current models of complex word recognition, as well as stimulus construction practices in psycho-and neurolinguistic research.

Zekun Wu, Yuan Li

Connectors are linguistic elements that link statements between textual units (Duden, 2016, p. 1083) and function as conjunctions (Heringer, 1989, p. 353). Many scholars have studied English as a Foreign Language (EFL) learners’ connector usage, but few studies have occurred in fields related to German, with deficiencies still present regarding issues such as the relatively limited classification of connectors, insufficient analytical dimensions, and lack of studies on learners at different levels. This study is based on the dynamic systems theory and uses 168 argumentative essays as its data basis, with 155 having been selected from Lerner Corpus and covering four learning stages and the rest from the FalkoEssayL1 corpus. The study uses frequency, diversity, and accuracy as three variables in the quantitative and qualitative analyses of Chinese German as a Foreign Language (GFL) learners’ acquisition of connectors. The results indicate Chinese GFL-learners’ acquisition of connectors to show a non-unidirectional, non-linear, and interactive developmental tendency. Their usage of connectors also shows certain characteristics compared to native German speakers. In addition, the phenomena of lexical plateau and fossilization are also observable among GFL-learners during certain specific learning stages. Accordingly, this study verifies and even enriches the dynamic systems theory with specific lexical data. Based on the obtained results, the influencing factors on the usage of connectors is attributable to four aspects: language input, linguistic interference, learning strategies, and thought patterns. The final section of the article discusses didactic suggestions based on these results.

Azeddine Bouabdallah, Jorge Gavilan, Jennifer Gerbl et al.

Nowadays, metadata information is often given by the authors themselves upon submission. However, a significant part of already existing research papers have missing or incomplete metadata information. German scientific papers come in a large variety of layouts which makes the extraction of metadata a non-trivial task that requires a precise way to classify the metadata extracted from the documents. In this paper, we propose a multimodal deep learning approach for metadata extraction from scientific papers in the German language. We consider multiple types of input data by combining natural language processing and image vision processing. This model aims to increase the overall accuracy of metadata extraction compared to other state-of-the-art approaches. It enables the utilization of both spatial and contextual features in order to achieve a more reliable extraction. Our model for this approach was trained on a dataset consisting of around 8800 documents and is able to obtain an overall F1-score of 0.923.

Catherine Teissier

Work as a topic has become more popular recently, after a quiet periode of almost two decades. In addition, comics and graphical novels today have gone far beyond the realm of fiction, and now also cover historical, biographical or societal subjects. After presenting the very lively context in which non-fiction comics are currently developing, particularly in France, the article examines the representation of work through sociological surveys adapted to comic and graphical form in the Sociorama series (2016-2017) and shows its strengths and limitations. As a counterpoint, it analyses the way in which the German author Birgit Weyhe elaborates a reflection on work in general, and on the status of artistic work in particular, in the comic strip Arbeit – ein eigenes Arbeitszimmer (2015).