Between Precision and Recognition : Equivalence, Language Use, and Searchability in Swedish Translation of Steam tags

Digital game distribution platforms rely on metadata to support search, filtering, and discovery in multilingual environments. Steam, the world’s largest digital game distribution platform, uses a tag-based metadata system that combines user generated tags with a controlled vocabulary. English functions as the primary source language, while tags are translated into other languages, including Swedish. This creates a multilingual information system in which semantic equivalence, language use, and searchability are central concerns. The aim of this study is to examine how Swedish translations of Steam tags correspond to actual language use in Swedish game discourse, and to discuss how this relationship may affect searchability from a Cross-Language Information Retrieval (CLIR) perspective. The study analyses a corpus of 969 professional and user-generated game reviews published on the Swedish gaming website FZ.se. An exploratory, corpus-based approach is applied, combining quantitative frequency analysis with qualitative collocation and context analysis. Nine strategically selected tag pairs are examined in depth. The results show that while many Swedish tag translations are conceptually accurate, they are often weakly established in Swedish game discourse. English terms are frequently preferred, particularly in user-generated reviews, while several Swedish translations function primarily as descriptive phrases rather than as stable genre or format labels. The analysis identifies varying degrees of equivalence between English and Swedish tags, ranging from partial equivalence to non equivalence. From a CLIR perspective, the findings indicate a tension between conceptual precision and alignment with user language. Conceptually precise translations do not necessarily correspond to the terms users employ in practice, which may reduce recall in multilingual search environments. The study highlights the importance of considering actual language use when translating metadata in multilingual information systems and contributes to library and information science research on semantic interoperability, user-generated metadata, and multilingual search in digital platforms