---
title: "The L2 Network — paper by Ribeiro-Flucht and Chen (2026)"
description: "This design reflects that forms and functions map many-to-many. 32 references, cited by 1 paper here."
authors: ["Luisa Ribeiro-Flucht", "Xiaobin Chen"]
year: 2026
venue: "Proceedings of the Workshop on Structured Linguistic Data and Evaluation (SLiDE)"
source: "https://aclanthology.org/2026.slide-1.13/"
doi: "10.63317/38q4k9x2yvc9"
licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)"
licence_url: "https://creativecommons.org/licenses/by-nc/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/flucht-2026-l2-network-cefr-aligned-knowledge/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/flucht-2026-l2-network-cefr-aligned-knowledge/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/18/) / The L2 Network: A CEFR-Aligned Knowledge Graph for Grammar Domain Modeling

# The L2 Network: A CEFR-Aligned Knowledge Graph for Grammar Domain Modeling

**Authors:** Luisa Ribeiro-Flucht and [Xiaobin Chen](https://md.research.elamur.ai/free/en/authors/xiaobin-chen/) · **Year:** 2026 · **Venue:** [Proceedings of the Workshop on Structured Linguistic Data and Evaluation (SLiDE)](https://md.research.elamur.ai/free/en/venues/slide/) ([SLiDE 2026](https://md.research.elamur.ai/free/en/venues/slide/2026/)) · **Publisher:** [European Language Resources Association (ELRA)](https://md.research.elamur.ai/free/en/publishers/european-language-resources-association-elra/) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.slide-1.13/) · [PDF](https://aclanthology.org/2026.slide-1.13.pdf) · [DOI 10.63317/38q4k9x2yvc9](https://doi.org/10.63317/38q4k9x2yvc9) · [OpenAlex](https://openalex.org/W7166525551)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · 1 in our library, of the 149,439 papers whose bibliographies we read

## Summary
Tutoring systems built on LLMs often fail to produce output that matches a learner's proficiency level. One reason is that grammar is usually stored as a flat list of target items. The authors address this with the L2 Network, a typed, directed knowledge graph of English grammar at CEFR levels A1 and A2.

The graph has two layers. The formal layer holds grammatical patterns, taken with their level information from the English Grammar Profile. The functional layer holds communicative goals derived from CEFR can-do descriptors. Typed edges link each form to the functions it can serve and connect forms to one another through dependency and progression. This design reflects that forms and functions map many-to-many.

Expert annotators judged the proposed form–function links. The authors analyse their agreement and publish both the raw endorsements and a consensus-filtered core graph. The resource runs in Apache AGE on PostgreSQL and can be traversed with Cypher.

Its value is an explicit, inspectable domain model for learner modeling, task selection and controlled generation. The construction method can also be reused for other languages.

## Key points
- Scope: CEFR Basic User grammar (A1–A2) of English, with slot-style patterns such as SUBJ have V-en. Higher levels are left for later.
- Sources: the English Grammar Profile supplies forms and progression. Its criterial features come from a quasi-longitudinal study of the Cambridge Learner Corpus. The function layer is grounded in CEFR descriptors and the functional-notional tradition (Threshold Level 1990).
- The appendix lists 59 communicative functions, for example "Expressing obligation or necessity", "Referring to past events with present relevance" and "Requesting".
- The graph is formalised as G = (V, E), with form and function nodes and typed relations: realization, dependency and progression.
- The release is on GitHub under CC BY 4.0. It includes the schema, node and edge exports, documentation, example Cypher queries, the annotation-derived endorsement data and the consensus core.

## Interesting ideas
- Keeping grammatical forms and communicative functions as separate node types joined by many-to-many realization edges lets one pattern serve several purposes and one purpose be met by several patterns.
- Publishing the full expert endorsement data next to a consensus-filtered core lets users choose their own strictness and audit disagreements.
- Progression and dependency edges between constructions give a prerequisite structure, much like ALEKS does for maths topics.
- Describing grammar as slot patterns (e.g., SUBJ have V-en) is close to a dependency-level description and could be compiled into structural matchers.
- Graph-native queries such as "all forms realizing function X" make it easy to retrieve alternative realizations for a target meaning at a chosen level.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (frames): A comparable function layer in prag could map recognised UD constructions to speech-act-like functions, giving explicit evidence for why a clause was read as advice, a request or an offer.
- **I2** (learn): Silver data could be stored the same way, with per-witness parser votes kept beside the agreed subset, so thresholds can be re-tuned and contested cases studied later.
- **I3** (answer): The student readers could order school-course material by such prerequisites and, when a test fails, trace back to the simpler construction that is actually missing.
- **I4** (tree): Turning these patterns into UD subtree queries would let us tag sentences with CEFR grammar features and report parser accuracy per construction rather than only overall.
- **I5** (back): The generator could query such a graph to choose a construction that expresses a given function at a target difficulty, keeping the choice explainable.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [University of Tübingen](https://md.research.elamur.ai/free/en/universities/university-of-tubingen/) — University, Tübingen, Germany · 75 works · 41,081 citations

📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [💭 Psychology](https://md.research.elamur.ai/free/en/shelves/psychology/) · 1,552 works · [💭 Developmental and Educational Psychology](https://md.research.elamur.ai/free/en/shelves/psychology/developmental-and-educational-psychology/) · 499 works

🧭 Topics (1): [🕸️ Knowledge graphs & ontologies](https://md.research.elamur.ai/free/en/topics/knowledge-graphs/) · 192 works

## Cite

Cite as: Luisa Ribeiro-Flucht and Xiaobin Chen. 2026. The L2 Network: A CEFR-Aligned Knowledge Graph for Grammar Domain Modeling. Proceedings of the Workshop on Structured Linguistic Data and Evaluation (SLiDE).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{ribeiroflucht2026l2,
  title = {{The L2 Network: A CEFR-Aligned Knowledge Graph for Grammar Domain Modeling}},
  author = {Luisa Ribeiro-Flucht and Xiaobin Chen},
  booktitle = {Proceedings of the Workshop on Structured Linguistic Data and Evaluation (SLiDE)},
  year = {2026},
  doi = {10.63317/38q4k9x2yvc9},
  url = {https://doi.org/10.63317/38q4k9x2yvc9}
}
```

- [← previous The Indonesian Religiolect Corpus: Data Curation for Muslim, Protestant, and Catholic Language Varieties](https://md.research.elamur.ai/free/en/papers/sachs-2026-indonesian-religiolect-corpus-data-curation/)
- [next → The Mighty ToRR: A Benchmark for Table Reasoning and Robustness in LLMs](https://md.research.elamur.ai/free/en/papers/tahan-2026-mighty-torr-benchmark-table-reasoning/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/18/)

## References

32 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Raúl Aranovich. 2023. Modeling grammars with knowledge representation methods: Subcategorization as a test case. In European Semantic Web Conference, pages 112–116. Springer. [find](https://scholar.google.com/scholar?q=Ra%C3%BAl+Aranovich%2E+2023%2E+Modeling+grammars+with+knowledge+representation+methods%3A+Subcategorization+as+a+test+case%2E+In+European+Semantic+Web+Conference%2C+pages+112%E2%80%93116%2E+Springer%2E)
2. Mayukh Bagchi. 2021. A large scale, knowledge intensive domain development methodology. Knowledge Organization: KO, 48(1):8. [↗](https://doi.org/10.5771/0943-7444-2021-1-8)
3. Heike Behrens. 2009. Usage-based and emergentist approaches to language acquisition. [↗](https://doi.org/10.1515/ling.2009.014)
4. Luca Benedetto, Gabrielle Gaudeau, Andrew Caines, and Paula Buttery. 2025. Assessing how accurately large language models encode and apply the common european framework of reference for languages. Computers and Education: Artificial Intelligence, 8:100353. [↗](https://doi.org/10.1016/j.caeai.2024.100353)
5. Conrad Borchers and Tianze Shou. 2025. Can large language models match tutoring system adaptivity? a benchmarking study. In International Conference on Artificial Intelligence in Education, pages 407–420. Springer. [find](https://scholar.google.com/scholar?q=Conrad+Borchers+and+Tianze+Shou%2E+2025%2E+Can+large+language+models+match+tutoring+system+adaptivity%3F+a+benchmarking+study%2E+In+International+Conference+on+Artificial+Intelligence+in+Education%2C+pages+407%E2%80%93)
6. British Council. Grammar. LearnEnglish. Accessed: 2026-02-17. [find](https://scholar.google.com/scholar?q=British+Council%2E+Grammar%2E+LearnEnglish%2E+Accessed%3A+2026%2D02%2D17%2E)
7. Annette Capel. 2012. Completing the english vocabulary profile: C1 and c2 vocabulary. English Profile Journal, 3:e1. [↗](https://doi.org/10.1017/s2041536212000013)
8. Ronald Carter, Michael McCarthy, Geraldine Mark, and Anne O’Keeffe. 2016. English grammar today: An a–z of spoken and written grammar. Web adaptation in Cambridge Dictionary. Accessed: 2026-02-17. [find](https://scholar.google.com/scholar?q=Ronald+Carter%2C+Michael+McCarthy%2C+Geraldine+Mark%2C+and+Anne+O%E2%80%99Keeffe%2E+2016%2E+English+grammar+today%3A+An+a%E2%80%93z+of+spoken+and+written+grammar%2E+Web+adaptation+in+Cambridge+Dictionary%2E+Accessed%3A+2026%2D02%2D17%2E)
9. Xiaobin Chen, Elizabeth Bear, Bronson Hui, Haemanth Santhi-Ponnusamy, and Detmar Meurers. 2022. Education theories and ai affordances: Design and implementation of an intelligent computer assisted language learning system. In International Conference on Artificial Intelligence in Education, pages 582–585. Springer. [find](https://scholar.google.com/scholar?q=Xiaobin+Chen%2C+Elizabeth+Bear%2C+Bronson+Hui%2C+Haemanth+Santhi%2DPonnusamy%2C+and+Detmar+Meurers%2E+2022%2E+Education+theories+and+ai+affordances%3A+Design+and+implementation+of+an+intelligent+computer+assisted+lan)
10. Eric Cosyn, Hasan Uzun, Christopher Doble, and Jeffrey Matayoshi. 2021. A practical perspective on knowledge space theory: Aleks and its data. Journal of Mathematical Psychology, 101:102512. [↗](https://doi.org/10.1016/j.jmp.2021.102512)
11. Council of Europe. 2001. Common European framework of reference for languages: Learning, teaching, assessment. Cambridge University Press. [find](https://scholar.google.com/scholar?q=Council+of+Europe%2E+2001%2E+Common+European+framework+of+reference+for+languages%3A+Learning%2C+teaching%2C+assessment%2E+Cambridge+University+Press%2E)
12. Council of Europe. 2020. Common European Framework of Reference for Languages: Learning, teaching, assessment – Companion Volume. Cambridge University Press, Cambridge. [find](https://scholar.google.com/scholar?q=Council+of+Europe%2E+2020%2E+Common+European+Framework+of+Reference+for+Languages%3A+Learning%2C+teaching%2C+assessment+%E2%80%93+Companion+Volume%2E+Cambridge+University+Press%2C+Cambridge%2E)
13. Holger Diessel. 2019. The Grammar Network. Cambridge University Press. [↗](https://doi.org/10.1017/9781108671040)
14. Holger Diessel. 2023. The constructicon: Taxonomies and networks. Cambridge University Press. [find](https://scholar.google.com/scholar?q=Holger+Diessel%2E+2023%2E+The+constructicon%3A+Taxonomies+and+networks%2E+Cambridge+University+Press%2E)
15. Adele E Goldberg. 2003. Constructions: A new theoretical approach to language. Trends in cognitive sciences, 7(5):219–224. [↗](https://doi.org/10.1016/s1364-6613(03)00080-9)
16. Anthony Green. 2012. Language functions revisited: Theoretical and empirical bases for language construct definition across the ability range, volume 2. Cambridge University Press. [find](https://scholar.google.com/scholar?q=Anthony+Green%2E+2012%2E+Language+functions+revisited%3A+Theoretical+and+empirical+bases+for+language+construct+definition+across+the+ability+range%2C+volume+2%2E+Cambridge+University+Press%2E)
17. Julia Harrison and Fiona Barker. 2015. English profile in practice, volume 5. Cambridge University Press. [find](https://scholar.google.com/scholar?q=Julia+Harrison+and+Fiona+Barker%2E+2015%2E+English+profile+in+practice%2C+volume+5%2E+Cambridge+University+Press%2E)
18. Kristiina Jokinen. 2024. The need for grounding in llm-based dialogue systems. In Proceedings of the Workshop: Bridging Neurons and Symbols for Natural Language Processing and Knowledge Graphs Reasoning (NeusymBridge)@ LREC-COLING-2024, pages 45–52. [find](https://scholar.google.com/scholar?q=Kristiina+Jokinen%2E+2024%2E+The+need+for+grounding+in+llm%2Dbased+dialogue+systems%2E+In+Proceedings+of+the+Workshop%3A+Bridging+Neurons+and+Symbols+for+Natural+Language+Processing+and+Knowledge+Graphs+Reasoni)
19. Anisia Katinskaia. 2025. An overview of artificial intelligence in computer-assisted language learning. arXiv preprint arXiv:2505.02032. [↗](https://arxiv.org/abs/2505.02032)
20. Daniela Landert, Daria Dayter, Thomas C. Messerli, and Miriam A. Locher. 2023. Corpus Pragmatics. Cambridge Elements in Pragmatics. Cambridge University Press. [find](https://scholar.google.com/scholar?q=Daniela+Landert%2C+Daria+Dayter%2C+Thomas+C%2E+Messerli%2C+and+Miriam+A%2E+Locher%2E+2023%2E+Corpus+Pragmatics%2E+Cambridge+Elements+in+Pragmatics%2E+Cambridge+University+Press%2E)
21. Diane Larsen-Freeman. 2003. Teaching language: From grammar to grammaring. (Heinle Cengage Learning). [find](https://scholar.google.com/scholar?q=Diane+Larsen%2DFreeman%2E+2003%2E+Teaching+language%3A+From+grammar+to+grammaring%2E+%28Heinle+Cengage+Learning%29%2E)
22. Anne O’Keeffe and Geraldine Mark. 2017. The english grammar profile of learner competence: Methodology and key findings. International Journal of Corpus Linguistics, 22(4):457–489. [↗](https://doi.org/10.1075/ijcl.14086.oke)
23. Indira Padayachee et al. 2002. Intelligent tutoring systems: Architecture and characteristics. In Proceedings of the 32nd Annual SACLA Conference, pages 1–8. South African Computer Lecturers’ Association. [find](https://scholar.google.com/scholar?q=Indira+Padayachee+et+al%2E+2002%2E+Intelligent+tutoring+systems%3A+Architecture+and+characteristics%2E+In+Proceedings+of+the+32nd+Annual+SACLA+Conference%2C+pages+1%E2%80%938%2E+South+African+Computer+Lecturers%E2%80%99+Associat)
24. Ciyuan Peng, Feng Xia, Mehdi Naseriparsa, and Francesco Osborne. 2023. Knowledge graphs: Opportunities and challenges. Artificial intelligence review, 56(11):13071–13102. [↗](https://doi.org/10.1007/s10462-023-10465-9)
25. Jean-Philippe Prost. 2022. Integrating a phrase structure corpus grammar and a lexical-semantic network: the holinet knowledge graph. In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 613–622. [find](https://scholar.google.com/scholar?q=Jean%2DPhilippe+Prost%2E+2022%2E+Integrating+a+phrase+structure+corpus+grammar+and+a+lexical%2Dsemantic+network%3A+the+holinet+knowledge+graph%2E+In+Proceedings+of+the+Thirteenth+Language+Resources+and+Evaluation)
26. Giulia Rambelli, Emmanuele Chersoni, Philippe Blache, Chu-Ren Huang, and Alessandro Lenci. 2019. [Distributional semantics meets construction grammar. towards a unified usage-based model of grammar and meaning](https://md.research.elamur.ai/free/en/papers/rambelli-2019-distributional-semantics-meets-construction-grammar/). In Proceedings of the First International Workshop on Designing Meaning Representations, pages 110–120.
27. Hans-Jörg Schmid. 2015. A blueprint of the entrenchment-and-conventionalization model. Yearbook of the German cognitive linguistics association, 3(1). [↗](https://doi.org/10.1515/gcla-2015-0002)
28. Vanja Slavuj, Ana Meštrović, and Božidar Kovačić. 2017. Adaptivity in educational systems for language learning: a review. Computer Assisted Language Learning, 30(1-2):64–90. [↗](https://doi.org/10.1080/09588221.2016.1242502)
29. Jan Ate Van Ek, John Leslie Melville Trim, et al. 1991. Threshold level 1990. Council of Europe. [find](https://scholar.google.com/scholar?q=Jan+Ate+Van+Ek%2C+John+Leslie+Melville+Trim%2C+et+al%2E+1991%2E+Threshold+level+1990%2E+Council+of+Europe%2E)
30. Anh-Duc Vu, Jue Hou, Anisia Katinskaia, Ching-Fan Sheu, and Roman Yangarber. 2025. A bayesian approach to inferring prerequisite structures and topic difficulty in language learning. In BEA: Proceedings of the 20th Workshop on Innovative Use of NLP for Building Educational Applications. [find](https://scholar.google.com/scholar?q=Anh%2DDuc+Vu%2C+Jue+Hou%2C+Anisia+Katinskaia%2C+Ching%2DFan+Sheu%2C+and+Roman+Yangarber%2E+2025%2E+A+bayesian+approach+to+inferring+prerequisite+structures+and+topic+difficulty+in+language+learning%2E+In+BEA%3A+Proceedin)
31. Martin Weisser. 2014. Speech act annotation, page 84–114. Cambridge University Press. [find](https://scholar.google.com/scholar?q=Martin+Weisser%2E+2014%2E+Speech+act+annotation%2C+page+84%E2%80%93114%2E+Cambridge+University+Press%2E)
32. Fengna Zhang and Xinxin Li. 2025. Dynamic recommendation system for higher vocational english learning paths based on real-time knowledge graph update. Engineering Reports, 7(12):e70492. 9. Appendix ID Function 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Asking for advice Asking for confirmation Asking for information Comparing Comparing within a set Conveying meaning Making statements Describing Emphasizing the affected entity Expressing obligation or necessity Expressing beliefs and opinions Expressing cause Expressing certainty Expressing condition Expressing degree Expressing evaluation Expressing frequency Expressing hypotheses or inferences Expressing manner Expressing negation Expressing politeness Expressing possession Expressing possibility Expressing ability Expressing preference Expressing probability Expressing progression Expressing quantity Expressing wishes Framing or linking information Table 5: Inventory of communicative functions (n = 59). ID Function 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Giving advice Giving directions or instructions Making offers and invitations Making suggestions Providing detail Referring back to entities Referring to actions states or relations Referring to additional or alternative entities Referring to entities Referring to future actions and events Referring to impersonal events Referring to past actions and events Referring to past events with present relevance Referring to present states events and facts Referring to self Referring to self-directed actions Referring to the acting or experiencing entity Referring to the affected entity Referring to third-person subjects Reporting information Requesting Sequencing events Situating events or entities Situating entities in space Situating events in time Specifying information Pointing to entities Expressing contrast Expressing indefiniteness [↗](https://doi.org/10.1002/eng2.70492)

*Updated 1 November 2026*