---
title: "Terminology-Augmented Generation for Intangible Cultural"
description: "The paper tackles a recurring problem in machine translation of culturally loaded vocabulary: keeping LLMs consistent with… 19 references."
authors: ["Wanda Punzi Zarino", "Pilar Sánchez Gijón"]
year: 2026
venue: "Proceedings of the 2nd Workshop on Evaluating Text Difficulty in a Multilingual Context (DeTermIt! 2026)"
source: "https://aclanthology.org/2026.determit-1.8/"
doi: "10.63317/43bmxwahuzmq"
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/zarino-2026-terminology-augmented-generation-intangible-cultural/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zarino-2026-terminology-augmented-generation-intangible-cultural/)

[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/52/) / Terminology-Augmented Generation for Intangible Cultural Heritage: A Controlled LLM-Based Translation Framework

# Terminology-Augmented Generation for Intangible Cultural Heritage: A Controlled LLM-Based Translation Framework

**Authors:** Wanda Punzi Zarino and Pilar Sánchez Gijón · **Year:** 2026 · **Venue:** [Proceedings of the 2nd Workshop on Evaluating Text Difficulty in a Multilingual Context (DeTermIt! 2026)](https://md.research.elamur.ai/free/en/venues/determit/) ([Workshop on Evaluating Text Difficulty in a Multilingual Context 2026](https://md.research.elamur.ai/free/en/venues/determit/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.determit-1.8/) · [PDF](https://aclanthology.org/2026.determit-1.8.pdf) · [DOI 10.63317/43bmxwahuzmq](https://doi.org/10.63317/43bmxwahuzmq) · [OpenAlex](https://openalex.org/W7166517216)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
The paper tackles a recurring problem in machine translation of culturally loaded vocabulary: keeping LLMs consistent with curated terminology. The domain is traditional Campanian food, translated from Italian into Spanish. The authors built a concept-oriented bilingual termbase in YAML from the regional PAT register. Each entry has three layers: conceptual (definitions, production and cultural notes), linguistic (POS, gender, variants) and translational (strategy labels, preferred equivalents, genre rules). A local Model Context Protocol (MCP) server detects terms in each source text and returns the matching entries and genre policies. These are injected into the prompt, and a rule-based post-processing step then enforces the required surface forms. Mistral 7B Instruct and Gemma3 4B were compared under three conditions on 30 authentic texts from three genres. Pairing injection with deterministic enforcement tended to raise terminological compliance, but instruction-following differed noticeably between the models and across genres. The work shows how structured terminology, rather than similarity-based RAG, can steer small local models in heritage-sensitive translation.

## Key points
- The termbase covers 28 gastronomic preparations drawn from the 610 products in the Campania PAT register. Examples include *ciauliello*, *frittata di scammaro* and *cicci di Santa Lucia*.
- The MCP tools are `glossary_for_text_json`, which does boundary-aware, Unicode-safe term matching including multiword terms and variants, and `translate_usage`, which returns genre-specific policies.
- The test set has 30 real Italian texts, 10 for each genre: regulatory-administrative, promotional-touristic and narrative-cultural. The genres were chosen to follow Sabatini's cline of communicative constraint.
- Each genre has its own rule:
  - Regulatory texts keep the Italian term inside «angle quotes».
  - Promotional texts use the Spanish equivalent plus a parenthetical expansion based on the production note.
  - Narrative texts use the equivalent plus a comma-introduced gloss based on the cultural note.
- The three conditions are BASELINE (shared system prompt only), INJ (glossary injection) and INJ+ENF (injection plus deterministic enforcement). The open-weight models Mistral 7B Instruct and Gemma3 4B both ran locally.

*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

- [Universitat Autònoma de Barcelona](https://md.research.elamur.ai/free/en/universities/universitat-autonoma-de-barcelona/) — University, Cerdanyola del Vallès, Spain · 5 works · 9,880 citations
- [Parthenope University of Naples](https://md.research.elamur.ai/free/en/universities/parthenope-university-of-naples/) — University, Naples, Italy · 2 works

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works

## Cite

Cite as: Wanda Punzi Zarino and Pilar Sánchez Gijón. 2026. Terminology-Augmented Generation for Intangible Cultural Heritage: A Controlled LLM-Based Translation Framework. Proceedings of the 2nd Workshop on Evaluating Text Difficulty in a Multilingual Context (DeTermIt! 2026).

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

```bibtex
@inproceedings{zarino2026terminology,
  title = {{Terminology-Augmented Generation for Intangible Cultural Heritage: A Controlled LLM-Based Translation Framework}},
  author = {Wanda Punzi Zarino and Pilar Sánchez Gijón},
  booktitle = {Proceedings of the 2nd Workshop on Evaluating Text Difficulty in a Multilingual Context (DeTermIt! 2026)},
  year = {2026},
  doi = {10.63317/43bmxwahuzmq},
  url = {https://doi.org/10.63317/43bmxwahuzmq}
}
```

- [← previous TEN: Table Explicitization, Neurosymbolically](https://md.research.elamur.ai/free/en/papers/mehrotra-2026-ten-table-explicitization-neurosymbolically/)
- [next → Terminology-Aware Retrieval-Augmented Knowledge Distillation for Biomedical Neural Machine Translation](https://md.research.elamur.ai/free/en/papers/zafar-2026-terminology-aware-retrieval-augmented-knowledge/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/52/)

## References

19 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. P. Martín-Chozas, P. Calleja, and C. R. Limón. Terminology Enhanced Retrieval Augmented Generation for Spanish Legal Corpora. In Proceedings of the 5th Conference on Language, Data and Knowledge, pages 147–152, 2025. [find](https://scholar.google.com/scholar?q=P%2E+Mart%C3%ADn%2DChozas%2C+P%2E+Calleja%2C+and+C%2E+R%2E+Lim%C3%B3n%2E+Terminology+Enhanced+Retrieval+Augmented+Generation+for+Spanish+Legal+Corpora%2E+In+Proceedings+of+the+5th+Conference+on+Language%2C+Data+and+Knowledge%2C+page)
2. L. Buccheri. Parole del cibo in Campania. Cento voci del lessico gastronomico regionale. Franco Cesati Editore, Firenze, 2023. [find](https://scholar.google.com/scholar?q=L%2E+Buccheri%2E+Parole+del+cibo+in+Campania%2E+Cento+voci+del+lessico+gastronomico+regionale%2E+Franco+Cesati+Editore%2C+Firenze%2C+2023%2E)
3. M.-T. Cabré. La terminología. Teoría, metodología y aplicaciones. Antártida/Empúries, Barcelona, 1993. [find](https://scholar.google.com/scholar?q=M%2E%2DT%2E+Cabr%C3%A9%2E+La+terminolog%C3%ADa%2E+Teor%C3%ADa%2C+metodolog%C3%ADa+y+aplicaciones%2E+Ant%C3%A1rtida%2FEmp%C3%BAries%2C+Barcelona%2C+1993%2E)
4. L. Molina and A. Hurtado Albir. Translation techniques revisited: A dynamic and functionalist approach. Meta: Translators’ Journal, 47(4):498– 512, 2002. [find](https://scholar.google.com/scholar?q=L%2E+Molina+and+A%2E+Hurtado+Albir%2E+Translation+techniques+revisited%3A+A+dynamic+and+functionalist+approach%2E+Meta%3A+Translators%E2%80%99+Journal%2C+47%284%29%3A498%E2%80%93+512%2C+2002%2E)
5. M.-T. Cabré. Theories of terminology: Their description, prescription and explanation. Terminology. International Journal of Theoretical and Applied Issues in Specialized Communication, 9(2):163– 199, 2003. [find](https://scholar.google.com/scholar?q=M%2E%2DT%2E+Cabr%C3%A9%2E+Theories+of+terminology%3A+Their+description%2C+prescription+and+explanation%2E+Terminology%2E+International+Journal+of+Theoretical+and+Applied+Issues+in+Specialized+Communication%2C+9%282%29%3A163%E2%80%93+199%2C)
6. K. Papineni, S. Roukos, T. Ward, and W.-J. Zhu. [BLEU: A method for automatic evaluation of machine translation](https://md.research.elamur.ai/free/en/papers/papineni-2002-bleu-method-automatic-evaluation-machine/). In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pages 311–318, 2002.
7. F. Chessa, C. De Giovanni, and M. T. Zanola. La terminologia dell’agroalimentare. FrancoAngeli, Milano, 2014. [find](https://scholar.google.com/scholar?q=F%2E+Chessa%2C+C%2E+De+Giovanni%2C+and+M%2E+T%2E+Zanola%2E+La+terminologia+dell%E2%80%99agroalimentare%2E+FrancoAngeli%2C+Milano%2C+2014%2E)
8. Regione Campania. Prodotti tradizionali – elenco dei prodotti agroalimentari tradizionali. [https://agricoltura.regione.campania.it/](https://agricoltura.regione.campania.it/) tipici/prodotti_tradizionali.htm, 2025. [↗](https://agricoltura.regione.campania.it/)
9. G. M. Di Nunzio. Terminology-augmented generation (TAG): Foundations, use cases, and evaluation paths. Journal of Digital Terminology and Lexicography, 1(1):97–104, 2025. [find](https://scholar.google.com/scholar?q=G%2E+M%2E+Di+Nunzio%2E+Terminology%2Daugmented+generation+%28TAG%29%3A+Foundations%2C+use+cases%2C+and+evaluation+paths%2E+Journal+of+Digital+Terminology+and+Lexicography%2C+1%281%29%3A97%E2%80%93104%2C+2025%2E)
10. R. Rei, C. Stewart, A. C. Farinha, and A. Lavie. [COMET: A neural framework for MT evaluation](https://md.research.elamur.ai/free/en/papers/rei-2020-comet-neural-framework-mt-evaluation/). In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 2685–2702, 2020.
11. K. Fleischmann. Terminologiemanagement: Die Schlüsselkomponente für effiziente Kommunikation in Unternehmen. Information – Wissenschaft & Praxis, 76(4):169–176, 2025. [find](https://scholar.google.com/scholar?q=K%2E+Fleischmann%2E+Terminologiemanagement%3A+Die+Schl%C3%BCsselkomponente+f%C3%BCr+effiziente+Kommunikation+in+Unternehmen%2E+Information+%E2%80%93+Wissenschaft+%26+Praxis%2C+76%284%29%3A169%E2%80%93176%2C+2025%2E)
12. C. Grimaldi. Il prodotto agroalimentare campano. Tra lingua, cultura e tradizione. Aracne, Roma, 2017. [find](https://scholar.google.com/scholar?q=C%2E+Grimaldi%2E+Il+prodotto+agroalimentare+campano%2E+Tra+lingua%2C+cultura+e+tradizione%2E+Aracne%2C+Roma%2C+2017%2E)
13. F. Sabatini. Analisi del linguaggio giuridico: Il testo normativo in una tipologia generale dei testi. In M. D’Antonio, editor, Corso di studi superiori legislativi (1988–1989). [find](https://scholar.google.com/scholar?q=F%2E+Sabatini%2E+Analisi+del+linguaggio+giuridico%3A+Il+testo+normativo+in+una+tipologia+generale+dei+testi%2E+In+M%2E+D%E2%80%99Antonio%2C+editor%2C+Corso+di+studi+superiori+legislativi+%281988%E2%80%931989%29%2E)
14. S. Gupta, R. Ranjan, and S. N. Singh. A comprehensive survey of retrieval-augmented generation (RAG): Evolution, current landscape and future directions. arXiv preprint arXiv:2410.12837, 2024. — [arXiv:2410.12837](https://arxiv.org/abs/2410.12837) [find](https://scholar.google.com/scholar?q=S%2E+Gupta%2C+R%2E+Ranjan%2C+and+S%2E+N%2E+Singh%2E+A+comprehensive+survey+of+retrieval%2Daugmented+generation+%28RAG%29%3A+Evolution%2C+current+landscape+and+future+directions%2E+arXiv+preprint+arXiv%3A2410%2E12837%2C+2024%2E+%E2%80%94+%5BarXi)
15. F. Sabatini. “Rigidità-esplicitezza” vs “elasticità-implicitezza”: Possibili parametri massimi per una tipologia dei testi. In G. Skytte and F. Sabatini, editors, Linguistica testuale comparativa. Museum Tusculanum Press, Copenaghen, 1999. [find](https://scholar.google.com/scholar?q=F%2E+Sabatini%2E+%E2%80%9CRigidit%C3%A0%2Desplicitezza%E2%80%9D+vs+%E2%80%9Celasticit%C3%A0%2Dimplicitezza%E2%80%9D%3A+Possibili+parametri+massimi+per+una+tipologia+dei+testi%2E+In+G%2E+Skytte+and+F%2E+Sabatini%2C+editors%2C+Linguistica+testuale+comparativa%2E+Mus)
16. UNESCO. Convention for the Safeguarding of the Intangible Cultural Heritage. UNESCO, Paris, 2003. [find](https://scholar.google.com/scholar?q=UNESCO%2E+Convention+for+the+Safeguarding+of+the+Intangible+Cultural+Heritage%2E+UNESCO%2C+Paris%2C+2003%2E)
17. A. Lackner, A. Vega-Wilson, and C. Lang. Terminology-augmented generation: A systematic review of terminology formats for in-context learning in LLMs. In Proceedings of the 4th International Conference on Multilingual Digital Terminology Today (MDTT 2025), volume 3990 of CEUR Workshop Proceedings, 2… [find](https://scholar.google.com/scholar?q=A%2E+Lackner%2C+A%2E+Vega%2DWilson%2C+and+C%2E+Lang%2E+Terminology%2Daugmented+generation%3A+A+systematic+review+of+terminology+formats+for+in%2Dcontext+learning+in+LLMs%2E+In+Proceedings+of+the+4th+International+Conferenc)
18. A. Lackner, A. Vega-Wilson, and C. Lang. An evaluation of terminology-augmented generation (TAG) and various terminology formats for the translation use case. Journal of Digital Terminology and Lexicography, 1(2):31–47, 2025b. [find](https://scholar.google.com/scholar?q=A%2E+Lackner%2C+A%2E+Vega%2DWilson%2C+and+C%2E+Lang%2E+An+evaluation+of+terminology%2Daugmented+generation+%28TAG%29+and+various+terminology+formats+for+the+translation+use+case%2E+Journal+of+Digital+Terminology+and+Lexico)
19. P. Lewis, E. Perez, A. Piktus, F. Petroni, V. Karpukhin, N. Goyal, H. Küttler, M. Lewis, W.-t. Yih, T. Roktäschel, S. Riedel, and D. Kiela. [Retrieval-augmented generation for knowledge-intensive NLP tasks](https://md.research.elamur.ai/free/en/papers/lewis-2020-retrieval-augmented-generation-knowledge-intensive/). In Advances in Neural Information Processing Systems, volume 33, pages 9459–9474, 2020.

*Updated 2 November 2026*