Creating and Implementing an Ontology of Texts, Documents and Works in Complex Textual Traditions 1 Peter Robinson, University of Saskatchewan Abstract This article suggests an ontology of texts, documents and works of particular relevance to the editing of complex large textual traditions, such as those of the Greek New Testament (c. 5000 witnesses), Dante's Commedia (c. 800) and Chaucer's Canterbury Tales (88). The need for this ontology is reviewed through a brief history of the Canterbury Tales project's work over three decades, with references also to the Greek New Testament and Commedia editorial projects. The central definition of the ontology is that a text is an act of communication inscribed in a document. Further, both the document and the act of communication may be represented as independent and ordered hierarchies of content objects (hence, trees), with textual nodes appearing on both trees, in different orderings and structures across the two trees. The Textual Communities environment successfully implements parts of the ontology of texts, documents and works to enable data collection, management and publication according to the needs of its current users, demonstrating the considerable advantages of this model for textual processing. However, Textual Communities does not implement the whole of this model in terms of data validation, ingestion and processing. Full exploration and implementation of the model here offered are challenges for future scholars. Successful implementation, however, would have considerable benefits, both for scholars working with complex large traditions, and also for those working with smaller but highly complex document sets, such as authorial manuscripts. 1. Background From the beginnings of what we now call digital humanities, scholars have explored the potential for computers to assist in the handling of complex textual traditions. Indeed, the work of Father Roberto Busa on the text of Aquinas stands as the starting point for this new discipline (Testori 2017). While Busa did not address the question of multiple versions of Aquinas’ text, focussing rather on lexical research into an existing established text, other scholars saw the possibility that computers might be applied to the challenging tasks inherent in the making of an edition of a work present in many witnesses. A single glance at the apparatus of a historical edition of a text in many versions, for example, of the sample page of the Nestle-Aland Greek New Testament (“Nestle Aland Novum Testamentum Graece :: Printed Editions” n.d.) calls to mind the labyrinthine patterns of information science: lists, links, hierarchies, complex and regular structures. Here is a suitable subject for the computer. In a lecture at the Clark Library in 1961, Vinton Dearing -- the textual editor of the large multi-volume University of California edition of the works of John Dryden -- compared himself as textual editor, armed with all the latest in 1 As well as the acknowledgements noted at various points in this article, I owe a special debt to Michael Sperberg- McQueen, Desmond Schmidt, and other participants in the discussion on the Humanist bulletin board in June 2020 of many points raised in this article (Humanist postings 34.89 onwards).