Paradoxes in the Translation Technology Ecosystem: Annotated Corpus of Paradox-Relevant Excerpts

This dataset contains a curated analytical corpus developed for the chapter Paradoxes in the Translation Technology Ecosystem in the Routledge Handbook of Translation, Technology, and AI. It comprises verbatim excerpts extracted from academic publications. An LLM-assisted pipeline was used to identify excerpts relevant to paradoxical tensions in translation technologies and to generate initial classifications according to six sensitising paradox categories (C1–C6): Value Distribution, Professional Transformation, Empowerment–Displacement, Automation and Optimisation, Access/Justice/Standardisation, and Data Sovereignty. These excerpts, classifications, and analytical notes were then human reviewed and validated. The dataset is therefore theory-driven and interpretive in its selection and coding, rather than exhaustive or representative of the source texts. The dataset also includes analytical notes explaining the rationale for each coding decision, enabling qualitative interpretation and reuse. The dataset is intended for research reuse in translation studies, sociotechnical analysis, and AI-in-work scholarship, including secondary analysis of paradox patterns, temporal framing, and code co-occurrence. Access is embargoed until handbook publication. IMPORTANT NOTE ON QUOTE ATTRIBUTION Verbatim excerpts reproduce text as it appears in the cited source documents. In some cases, these excerpts may themselves contain quotations from other works not identified in the dataset metadata. Users are responsible for verifying the provenance of any quoted material prior to citation and should consult the original source to ensure accurate attribution.