Språkbanken Text

15th NLP4CALL in Gothenburg, Sweden

November 19-20, 2026. Gothenburg, Sweden and online

NLP4CALL logo

We are happy to announce the 15th edition of the NLP4CALL workshop on Natural Language Processing for Computer-Assisted Language Learning. This year, NLP4CALL will be a two-day hybrid event that will take place on November 19-20, 2026. Onsite participation will be possible at the Humanisten building of the University of Gothenburg (Renströmsgatan 6, Göteborg, Sweden) and a link for remote participation will be shared closer to the workshop dates. Both onsite and online participation are free of charge.

Workshop description

The workshop series on Natural Language Processing (NLP) for Computer-Assisted Language Learning (NLP4CALL) is a meeting place for researchers working on integrating Natural Language Processing and Speech Technologies in CALL systems and exploring the theoretical and methodological issues arising from this connection. The latter includes, among others, the integration of insights from Second Language Acquisition (SLA) research, and the promotion of “Computational SLA” through setting up Second Language research infrastructures.

The intersection of Natural Language Processing (or Language Technology / Computational Linguistics) and Speech Technology with Computer-Assisted Language Learning (CALL) brings “understanding” of language to CALL tools, thus making CALL intelligent. This fact has given the name for this area of research — Intelligent CALL, or short, ICALL. As the definition suggests, apart from having excellent knowledge of Natural Language Processing and/or Speech Technology, ICALL researchers need good insights into SLA theories and practices, as well as knowledge of second language pedagogy and didactics. This workshop therefore invites a wide range of ICALL-relevant research, including studies where NLP-enriched tools are used for testing SLA and pedagogical theories, and vice versa, where SLA theories, pedagogical practices or empirical data are modeled in ICALL tools. The NLP4CALL workshop series is aimed at bringing together competences from these areas for sharing experiences and brainstorming around the future of the field.

We welcome papers:

  • that describe research directly aimed at ICALL;
  • that describe the ongoing development of resources and tools with potential use in ICALL, either directly in interactive applications, or indirectly in materials, application, or curriculum development, e.g. learning material generation, assessment of learner texts and responses, individualized learning solutions, provision of feedback;
  • that discuss challenges and/or research agenda for ICALL;
  • that describe empirical studies on language learner data.

This year, the workshop has a special focus on process-oriented approaches to educational NLP, including but not limited to work related to the collection and analysis of keystroke data from langauge learners.

Important dates

  • paper submission deadline: September ~~24~~ 27, 2026 (extended)
  • notification of acceptance: October 15, 2026
  • camera-ready deadline: October 29, 2026
  • workshop: November 19-20, 2026

All deadlines are AoE.

Registration

Registrations are open until October 28! You can find the registration form here.

Program

Day 1

  • 9:00-9:10 — welcome

  • 9:10-9:20 — introduction to the special theme

  • 9:20-10:40 — session 1

    Time Title Authors
    9:20–9:40 Lexical Profiling for Dutch: From Resource to Real-Life Implementation Jasper Degraeuwe and Natalie Finlayson
    9:40–10:00 CineLexi: A corpus-based bilingual resource of evaluative vocabulary in film reviews (German–Russian) Mariia Kopylova
    10:00–10:20 Etymology Can Inform Lexical Complexity Prediction Noah-Manuel Michael, Anna Hülsing and Andrea Horbach
    10:20–10:40 Predicting Personalized Vocabulary Learning Burden from Intralexical and Interlexical Factors Mykhailo Diachenko and John P. McCrae

  • 10:40-11:00 — coffee break

  • 11:00-12:00 — session 2

    Time Title Authors
    11:00–11:15 Using LLMs for feature extraction: the case of language functions with deixis Thomas Gaillat, Xiaobin Chen and Nelly Sagirov
    11:15–11:30 Exploring the syntactic parsing of learner texts in Modern Greek: Variety trumps size, genre trumps grammaticality Christina Klironomou, Arianna Masciolini and Aleksandrs Berdicevskis
    11:30–11:45 Towards Semi-Automatic Benchmarking of Large Language Models for Self-Regulated Language Learning Emile Alexandre, Yannick Parmentier and Guillaume Nassau
    11:45–12:00 Automated L2 Speech Assessment with Audio-Based Self-Supervised Neural Networks Eli Stafford and Sylvain Coulange

  • 12:00-13:15 — lunch break

  • 13:15-14:00 — keynote 1: Prof. Andrea Horbach. A Focus on the Writing Process With and Without AI Support: Insights from the German PISA 2025 Foreign Language Assessment Data Collection

  • 14:00-14:10 — break

  • 14:10-15:10 — session 3

    Time Title Authors
    14:10–14:25 ABC Diagnostics: A Web-Based Tool for Explainable CEFR Classification Ricardo Muñoz Sánchez, Kostiantyn Kucher, Filip Kayar and Elena Volodina
    14:25–14:40 Multilingual language learning with Grammatical Framework Adrian De Lon and Inari Listenmaa
    14:40–14:55 Making Feedback Visible: Character-Level Error Localization in Handwritten CALL Input Said Yasin and Torsten Zesch
    14:55–15:10 Reading Ahead: Pace-Based Prediction for Real-Time Tracking of Children’s Oral Reading in the Browser Bohdan Podziubanchuk, Einar Meister and Tanel Alumäe

  • 15:10-15:40 — coffee break

  • 15:40-17:00 — session 4

    Time Title Authors
    15:40–16:00 How Much of Native Language Identification Is Corpus Artifact? A Cross-Corpus Audit of SVMs and LLMs on TOEFL11 and ICLEv3 Shima Rahimi
    16:00–16:20 A Tool for the Automatic Analysis of Noun Phrase Complexity in Learner Productions Saint Joannes Monengue Essengo, Bi Yoman Ghislain Aymard Toa, Thomas Gaillat, Christine Évain and Fabienne Moreau
    16:20–16:40 A Multidimensional Evaluation of LLM-Generated Learning Materials for Greek Primary School Learners Eleftherios Kalogeros, Sofia Stamou, Markella Kolocheri, Thaleia-Dionysia Kolovou, Manolis Gergatsoulis and Matthew Damigos
    16:40–17:00 A corpus of the production of written Swedish by the deaf and hard of hearing Moa Gärdenfors, Robert Östling, Lisa Rudebeck and Mats Wirén

Day 2

  • 9:00-10:20 — session 5

    Time Title Authors
    9:00–9:20 Poor Generalisation of Proficiency Classifiers Trained on Linguistic Complexity Features to New Writing Tasks Emre Gucer, Soroosh Akef and Walt Detmar Meurers
    9:20–9:40 Universal, but Grounded? Anchoring Cross-Lingual CEFR Assessment in the Framework's Own Descriptors David Alfter
    9:40–10:00 Assessing the Reliability of Automatic Speech-Processing Methods for L2 Speech: Evidence from an Automatic Pipeline, Posteriorgrams, and Whisper Probability Scores Mehak Sohail and Nicolas Ballier
    10:00–10:20 Investigating the role of revisions for learner level predictions Ahood Al Sawar, Léo Nebal and Nicolas Ballier

  • 10:20-10:50 — coffee break

  • 10:50-11:35 — keynote 2: Dr. Rianne Conijn. Moving beyond the product: Keystroke Logging for CALL

  • 11:35-13:00 — lunch break

  • 13:00-13:45 — session 6

    Time Title Authors
    13:00–13:15 Adversarial Disentanglement of Speaker Identity and L2 Phonological Errors via Raw-Waveform Architectures Amirhossein Ghanipour Amirhandeh
    13:15–13:30 Acoustic and Lexical Signals for Native Language Identification of Language Learners Ahmet Yavuz Uluslu, Mengjie Qian, Kate Knill and Mark Gales
    13:30–13:45 Translation Surprisal Predicts Learner Error in L2 French Production: A Language-Pair-Specific Difficulty Feature for ICALL Matteo Lanza

  • 13:50-14:00 — break

  • 14:10-14:50 — 2 long papers

    Time Title Authors
    14:10–14:30 Data Augmentation for L2 English Speaking Assessment using TTS Stefano Bannò, Penny Karanasou, Mengjie Qian, Kate Knill and Mark Gales
    14:30–14:50 Behind the Smokescreen: Uncovering Cross-LLM Bias and Reasoning Contradictions of LLM-as-a-judge for Generated Feedback Céline Leuzinger

  • 14:50-15:00 — closing and information of the next edition of NLP4CALL

Invited speakers

Prof. Andrea Horbach: A Focus on the Writing Process With and Without AI Support: Insights from the German PISA 2025 Foreign Language Assessment Data Collection

Bio: Andrea Horbach has been professor for Educational NLP at Kiel University and the Leibniz Institute for Science and Mathematics Education (IPN) since 2024. Her research interests include free-text scoring, argument mining, automated feedback, and the use of process data for assessment. She is particularly interested in developing explainable and fair methods that support human agency in writing and assessment processes.

Abstract: Automated essay scoring has traditionally focused on the text as the final written product. More recently, corpora that include keylogging data have made it possible to shift the focus toward writing as a process: to investigate how students write and revise their texts, and to explore whether scoring models can provide useful feedback before the final essay is completed. Moreover, AI-based writing support is becoming increasingly available, but we still know comparatively little about how students actually use such systems while writing. In this talk, I will give an overview of ongoing work based on data from the German PISA 2025 foreign language assessment. The data combine final essays and their scores with intermediate text states, keylogging information, and interaction logs from an LLM-based writing support chatbot. I will present first results on how intermediate texts and process information can be used as early evidence for automated scoring, and how students interact with AI support during writing. I will conclude by discussing what these data can tell us about timely feedback, the interpretation of process evidence, and the role of AI support in computer-assisted language learning.

Dr. Rianne Conijn: Moving beyond the product: Keystroke Logging for CALL

Bio: Rianne Conijn is an assistant professor in learning analytics in the Human-Technology Interaction group at Eindhoven University of Technology, the Netherlands. Her research focuses on developing methods to capture and model learning behavior (e.g., self-regulated learning) and writing behavior (e.g., revision processes), using fine-grained behavioral data, such as keystroke and clickstream data. In addition, she uses these methods to examine how educational technologies, such as generative AI, affect the learning process. Rianne recently received a NWO Veni grant on "Human-Centered AI in education" where she aims to improve human-AI collaboration in writing in higher education.

Abstract: This abstract has been rewritten by me and polished with AI several times, yet you only read this final version. The final product of writing reveals little of the cognitive and behavioural activities involved in producing it. Keystroke logging, e.g. using Inputlog, can provide detailed insights into how writing unfolds. However, its value depends on how we interpret and analyse these data. In this talk, I show how keystroke data can inform our understanding of students’ writing at three levels, each making different aspects of the process visible and each requiring its own analytical choices. The micro level covers individual key events and short sequences, for instance showing that pauses cannot simply be equated with planning. The meso level characterizes how a writing session unfolds over time, revealing patterns that can inform writing instruction that whole-session averages obscure. The macro level extends the analysis across a project, tracing activity across sessions, drafts, and changing task demands. Finally, I discuss how these perspectives can help us investigate writing with generative AI, drawing on ongoing research into students’ master thesis writing. Together, these perspectives show how analysing writing at different levels can deepen our understanding of how students produce texts, both with and without generative AI.

Submission information

Submissions should describe original unpublished complete or in-progress work and follow the ACL Guidelines for Generative Assistance in Authorship.

Each submitted paper will be peer-reviewed by at least two members of the program committee in a double-blind fashion. All accepted papers will be collected into a proceedings volume to be published both in the NEALT Proceeding Series and through the ACL anthology.

We accept short, long and demo papers, all of which have to adhere to the following page limits:

  • short and demo papers: between 4 and 7 pages
  • long papers: between 8 and 12 pages.

Also note that:

  • references and appendices do not count towards these page limits
  • the main body of the paper has to be self-contained, as reviewers are not required to read any appendices
  • camera-ready versions of accepted papers will be allowed an additional page to address reviewer comments.

Papers should be submitted as PDFs through EasyChair. LaTeX and Word templates are available here.

Organizers

Contact

For more information about the workshop, you are welcome to reach out to nlp4call2026@easychair.org.

Funding

This workshop is jointly supported by:

Sidansvarig: Arianna Masciolini
|
Sidan uppdaterades: 2026-11-04