Using the Semantic Web to Improve Knowledge of Translations

Abstract

More than half of the almost 400 million bibliographic records in WorldCat are for languages other than English. Most of the monographs described were published only once. But a few million represent the core of our shared culture—works that have been translated into multiple languages, and sometimes translated multiple times into the same language. We learn about other cultures, and other cultures learn about ours, through these translations. As the world’s largest bibliographic database, WorldCat is positioned to provide the translation history of works, using the W3C bib extension translationOfWork to communicate the relationship of each translation to the original work. In our multilingual data enhancements project, our goal was to improve the descriptions of the most frequently published works, as they are the ones most likely to be translated and searched by users. In a database of MARC records, machine processes cannot support browsing or searching of works and their translations. Critical entities such as the title of the original work and the names of the translators are not always expressed in a machine-understandable form—and sometimes the information is missing altogether. Since a manual cleanup is not scalable, we explored the possibility of enriching MARC records with Linked Data from a third-party source, Wikidata. By integrating information from both WorldCat and Wikidata, we may be in a better position to present information about frequently-translated works in the preferred language and script of the user.