Legal terminology translation: humans vs algorithms

Authority vs. Automation: From Corpora to LLMs for Reliable Legal Terminology

THE TRANSLATABILITY OF MEDICAL TERMINOLOGY FROM ENGLISH TO GREEK: LINGUISTIC, CULTURAL, AND TECHNOLOGICAL DIMENSIONS

Medical terminology represents a historically rich and highly systematized domain of specialized language, shaped by the enduring influence of Greek and Latin as well as by contemporary English-dominated scientific discourse. This article examines the translatability of medical terminology from English into Greek through four complementary dimensions: linguistic, conceptual, cultural, and technological. Drawing on terminology theory, translation studies, and corpus-based approaches, the analysis highlights the morphological asymmetries between English and Greek, the semantic divergences that challenge strict term-to-term correspondences, and the cultural-ethical nuances embedded in concepts such as informed consent and palliative care. Special attention is given to the phenomenon of “reverse translation,” in which English medical terms of Greek etymology acquire new meanings that do not always align with their historical roots. The study also evaluates the role and limitations of contemporary translation technologies—terminology databases, Computer-Assisted Translation (CAT) tools, neural machine translation, and Artificial Intelligence (AI)-driven semantic systems—particularly in relation to Greek, a language underrepresented in multilingual biomedical resources. The article proposes a context-sensitive model of medical translation that integrates linguistic expertise, conceptual knowledge, cultural mediation, and critical technological literacy. The findings underscore the translator’s role as a terminological curator and interdisciplinary mediator, essential for ensuring precision, consistency, and cultural relevance in Greek medical terminology within a globalized scientific environment.

Reasoning Before Translation: Enhancing Legal Machine Translation with Structured Reasoning

The results show that the quality of small ``base'' models can be greatly enhanced, and that reinforcement learning with verifiable rewards can be applied to NMT in the legal domain and surpasses the translation quality of supervised fine-tuning.

Accuracy, Risk, and Translator Agency in AI-Assisted Specialised Translation

This article discusses the role of human–AI collaboration in specialised translation, focusing on accuracy, risk and translator agency in high-stakes legal, medical, technical, financial and institutional contexts. It critically synthesises recent research on machine translation, large language models, post-editing, terminology management, translation quality and professional agency, using a conceptual and analytical approach. The study does not present primary empirical data; rather, it attempts to establish a framework for assessing machine-generated output within specialised translation workflows. It is theoretically innovative in combining accuracy, risk and agency within a single relational model and demonstrating how each translation decision is shaped by system output, professional judgement and potential consequences. Accuracy is therefore understood as a multidimensional requirement encompassing linguistic, terminological and conceptual precision, as well as consistency, register, genre conventions and functional appropriateness. Risk refers to the potential for output to be inaccurate, incomplete or inadequately validated, resulting in errors that could affect legal interpretation, medical safety, technical procedures, institutional responsibility or end-user decisions. These concepts are presented in a six-dimensional analytical model that includes system contribution, accuracy issue, risk level, translator response, decision rationale and impact on final quality. The article argues that computational tools can be used for drafting, terminological support, reformulation, post-editing and quality assessment, but cannot serve as stand-alone authorities in specialised translation. Translator agency is thus redefined rather than eradicated: the translator must assess, edit or reject system output, explain decisions and take responsibility for the final product. The proposed framework offers a basis for future conceptual, empirical, pedagogical and professional studies.

Modelling specialized language through AI: challenges in training systems for technical domain translation

Modelling specialized language with artificial intelligence is a significant challenge for specialized translation, especially in fields where controlled terminology is essential, such as medicine, engineering, law, or the exact sciences. The performance of neural machine translation systems depends directly on the quality of the corpora used for training. However, obtaining clean, standardized, and representative corpora is difficult because access to technical documents is often limited, the structure of specialized texts varies, and terminology may be inconsistent across existing sources. As a result, unedited data can lead to semantic errors, incorrect generalizations, and conceptual confusion - issues that are particularly serious in fields where accuracy is critical. Another risk is the tendency of AI models to fill information gaps using statistical assumptions, which can lead to the use of inappropriate terms or to distorted conceptual relations. In this context, it becomes necessary to develop effective strategies for selecting, annotating, and cleaning corpora to ensure the terminological consistency required for training the models. This article proposes a hybrid AI-translator approach, in which technology and human expertise work together: AI increases the speed and volume of processing, while the translator validates terminology, checks conceptual coherence, and corrects semantic errors. By analysing current challenges and presenting practical solutions, this contribution aims to show how artificial intelligence can be integrated responsibly into specialized translation, without compromising the precision and accuracy required in technical fields

The translation strategy of unit specialized meaning in end user license agreement: a case study of Wondershare's EULA