--- title: "Introducing linguistic transformation to improve translation" description: "Translation memory systems in CAT tools find earlier translations through fuzzy matching, which is usually based on edit… 11 references." authors: ["Souhila Djabri", "Rocío Caro Quintana"] year: 2021 venue: "Proceedings of the Student Research Workshop Associated with RANLP 2021" source: "https://aclanthology.org/2021.ranlp-srw.7/" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/djabri-2021-introducing-linguistic-transformation-improve-translation/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/djabri-2021-introducing-linguistic-transformation-improve-translation/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2021](https://md.research.elamur.ai/free/en/papers/2021/page/9/) / Introducing linguistic transformation to improve translation memory retrieval. Results of a professional translators' survey for Spanish, French and Arabic # Introducing linguistic transformation to improve translation memory retrieval. Results of a professional translators' survey for Spanish, French and Arabic **Authors:** Souhila Djabri and Rocío Caro Quintana · **Year:** 2021 · **Venue:** [Proceedings of the Student Research Workshop Associated with RANLP 2021](https://md.research.elamur.ai/free/en/venues/ranlp/) ([RANLP 2021](https://md.research.elamur.ai/free/en/venues/ranlp/2021/)) · **Publisher:** [INCOMA Ltd.](https://md.research.elamur.ai/free/en/publishers/incoma-ltd/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2021.ranlp-srw.7/) · [PDF](https://aclanthology.org/2021.ranlp-srw.7.pdf) · [OpenAlex](https://openalex.org/W3214736189) **Citations:** 0 citations (OpenAlex, 2026-10-14) ## Summary Translation memory systems in CAT tools find earlier translations through fuzzy matching, which is usually based on edit distance. This matching works poorly when a new sentence has the same meaning as a stored one but different wording or structure, for example a voice change, reordering, a synonym or a pronoun. The authors already measured this weakness on SDL Trados and MemoQ. In this second phase, they asked professional translators to judge ten such transformations. Fifteen translators took part, five each for Spanish, French and Arabic. Each saw ten original segments from United Nations General Assembly regulation texts, each paired with a transformed version, through Google Forms questionnaires in their own language. Respondents mostly approved of the transformations. They favoured single, simple edits such as synonym swaps over edits that change word order. Most said they would apply such changes during assisted translation rather than before or after it. The findings support adding language-specific linguistic processing to TMS, possibly as an editing or automatic paraphrasing tool. The aim is to retrieve more matches and reduce translators' effort. ## Key points - The ten transformation types are: - active↔passive voice - word, phrase or clause reordering - one-synonym and two-synonym substitution - pronoun replacement - combinations of these, such as passive plus synonym or two synonyms plus reordering - Phase one compared 1500 original and transformed segments for ES-FR, FR-ES and AR-ES in SDL Trados and MemoQ. - Arabic as a source language often produced no match at all. - MemoQ handled word-order changes for Arabic especially badly. - The survey had 15 participants with 3–10 years of experience and ran 17–27 January 2021. About 60% mainly use TMS for legal or administrative texts. - On when to apply the edits, 60% chose during assisted translation, 20% pre-editing and 20% post-editing. - Respondents proposed further operations: reformulation, addition, omission, grammatical gender change and relative pronoun changes. ## Interesting ideas - A fixed list of meaning-preserving rewrites, such as voice change, reordering, synonym swaps and pronominalisation, works as a clean probe of whether a matcher sees through surface form to content. - Stacking transformations, for example passive plus synonym or two synonyms plus reordering, creates a graded difficulty ladder that shows where a system starts to fail. - Replacing a noun with a personal pronoun destroys a string match even though the content is unchanged, so recovering the referent matters as much as matching words. - Before building a tool, the authors had native-speaker professionals rate which transformations sound natural and acceptable. Users preferred single, light edits to heavy restructuring. - The paper cites alternatives to edit distance, such as overlap of syntactic parse trees and sentence-representation distance, as ways to judge segment similarity. ## For Mova How these ideas could serve Mova, our small open language model — its files, stages and tests: - **I1** (learn): Use these rewrites as invariance checks, where the frames from a sentence and its passive or reordered variant must agree; unrelated sentences serve as negative controls. - **I2** (frames): Report role-assignment accuracy per number of stacked rewrites to locate which combination breaks VerbNet role mapping. - **I3** (frames): Include pronominalised variants in coreference tests and require the resolved entity to fill the same frame slot as the original noun. - **I4** (back): When the text → graph → text path generates paraphrases, rank one-change outputs first and keep a small human naturalness check per rewrite type. - **I5** (tree): Compare UD trees, matching on lemmas and relations, to find equivalent sentences that string metrics miss, while keeping every match 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 Wolverhampton](https://md.research.elamur.ai/free/en/universities/university-of-wolverhampton/) — University, Wolverhampton, United Kingdom · 36 works · 6,331 citations - [University of Alicante](https://md.research.elamur.ai/free/en/universities/university-of-alicante/) — University, Alicante, Spain · 23 works · 1,229 citations 📚 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 🧭 Topics (3): [🔁 Machine translation](https://md.research.elamur.ai/free/en/topics/machine-translation/) · 1,105 works · [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works · [French](https://md.research.elamur.ai/free/en/topics/french/) · 83 works ## Cite Cite as: Souhila Djabri and Rocío Caro Quintana. 2021. Introducing linguistic transformation to improve translation memory retrieval. Results of a professional translators' survey for Spanish, French and Arabic. Proceedings of the Student Research Workshop Associated with RANLP 2021. The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{djabri2021introducing, title = {{Introducing linguistic transformation to improve translation memory retrieval. Results of a professional translators' survey for Spanish, French and Arabic}}, author = {Souhila Djabri and Rocío Caro Quintana}, booktitle = {Proceedings of the Student Research Workshop Associated with RANLP 2021}, year = {2021}, doi = {10.26615/issn.2603-2821.2021\_007}, url = {https://doi.org/10.26615/issn.2603-2821.2021_007} } ``` - [← previous Interactive Learning Approach for Arabic Target-Based Sentiment Analysis](https://md.research.elamur.ai/free/en/papers/balla-2021-interactive-learning-approach-arabic-target/) - [next → Issayas Tesfamariam, ትግርኛ ንዛረብ: Let’s Speak Tigrinya, Elementary Level, A First-Year Textbook, A Multidimensional Approach to the Teaching and Learning of Tigrinya As a Foreign Language](https://md.research.elamur.ai/free/en/papers/voigt-2021-issayas-tesfamariam-lets-speak-tigrinya/) - [↑ back to papers of 2021](https://md.research.elamur.ai/free/en/papers/2021/page/9/) ## References 11 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Arenas, A. G. (2019). Pre-editing and post-editing. In The Bloomsbury Companion to Language Industry Studies. https://doi.org/10.5040/9781350024960.0019 [↗](https://doi.org/10.5040/9781350024960.0019) 2. Timonera, K., & Mitkov, R. (2015). Improving Translation Memory Matching through Clause Splitting. Proceedings of the Workshop on Natural Language Processing for Translation Memories (NLP4TM), 17–23. [find](https://scholar.google.com/scholar?q=Timonera%2C+K%2E%2C+%26+Mitkov%2C+R%2E+%282015%29%2E+Improving+Translation+Memory+Matching+through+Clause+Splitting%2E+Proceedings+of+the+Workshop+on+Natural+Language+Processing+for+Translation+Memories+%28NLP4TM%29%2C+17%E2%80%9323%2E) 3. Bassam, H., Asm, M., Nadim, O., & Abeer, T. (2017). Formal Description of Arabic Syntactic Structure in the Framework of the Government and Binding Theory. 18(3), 611–625. https://doi.org/10.13053/C [↗](https://doi.org/10.13053/C) 4. Weber, E. (2014). La traducción de la voz pasiva francesa al español: ¿cuestión de lengua o cuestión de traducción? Mutatis Mutandis, 7(2), 368–385. [find](https://scholar.google.com/scholar?q=Weber%2C+E%2E+%282014%29%2E+La+traducci%C3%B3n+de+la+voz+pasiva+francesa+al+espa%C3%B1ol%3A+%C2%BFcuesti%C3%B3n+de+lengua+o+cuesti%C3%B3n+de+traducci%C3%B3n%3F+Mutatis+Mutandis%2C+7%282%29%2C+368%E2%80%93385%2E) 5. Bowker, L., & Corpas Pastor, G. (2015). Translation Technology. Oxford Handbooks Online, (November), 1–26. https://doi.org/10.1093/oxfordhb/97801995736 91.013.007 [↗](https://doi.org/10.1093/oxfordhb/97801995736) 6. Djabri, S. (2020). Análisis contrastivo de las coincidencias parciales en español, francés y árabe mediante transformaciones lingüísticas. Skopos, 11, 103–120. [find](https://scholar.google.com/scholar?q=Djabri%2C+S%2E+%282020%29%2E+An%C3%A1lisis+contrastivo+de+las+coincidencias+parciales+en+espa%C3%B1ol%2C+franc%C3%A9s+y+%C3%A1rabe+mediante+transformaciones+ling%C3%BC%C3%ADsticas%2E+Skopos%2C+11%2C+103%E2%80%93120%2E) 7. Leblanc, M. (2016). La traduction spécialisée à l ’ ère des nouvelles technologies : quel effet sur le texte de spécialité ? The impact of new translation technologies on specialized texts. 1, 77–92. https://doi.org/10.14746/strop.2016.425.006 [↗](https://doi.org/10.14746/strop.2016.425.006) 8. Levenshtein, V. I. (1966). Binary Codes Capable of Correcting Deletions, Insertions, and Reversals. Soviet Physics Doklady, 10(8), 707–710. [find](https://scholar.google.com/scholar?q=Levenshtein%2C+V%2E+I%2E+%281966%29%2E+Binary+Codes+Capable+of+Correcting+Deletions%2C+Insertions%2C+and+Reversals%2E+Soviet+Physics+Doklady%2C+10%288%29%2C+707%E2%80%93710%2E) 9. Oliver, A. (2016). La traducción asistida por ordenador. Barcelona: Oberta UOC Publishing, SL. [find](https://scholar.google.com/scholar?q=Oliver%2C+A%2E+%282016%29%2E+La+traducci%C3%B3n+asistida+por+ordenador%2E+Barcelona%3A+Oberta+UOC+Publishing%2C+SL%2E) 10. Saldanha, G., & O’Brien, S. (2014). Research Methodologies in Translation Studies. In Research Methodologies in Translation Studies. https://doi.org/10.4324/9781315760100 [↗](https://doi.org/10.4324/9781315760100) 11. Silvestre Baquero, A., & Mitkov, R. (2017). Translation Memory Systems Have a Long Way to Go. 44–51. https://doi.org/10.26615/978-954-452-042-7_006 [↗](https://doi.org/10.26615/978-954-452-042-7_006) *Updated 2 November 2026*