Allo

Ancient languages matter. They connect us to our ancestors and draw attention from many fields of study. Despite overwhelming interest, the practical study of ancient languages is usually limited to persons with access to training and resources. For introductory students, the dictionary is usually located in the appendix of the assigned grammar book. For advanced study, many ancient languages have one or more physically large, expensive, inaccessible, and difficult to use dictionaries. A systematic review of both print and digital dictionaries for eight ancient languages yielded three recurring challenges: accessibility, usability, and scope. To illustrate: the Oxford Latin Dictionary contains 2,400 pages with tiny font and weighs 9.1 pounds. The Brill Dictionary of Ancient Greek is similar; and, digital access to both requires institutional affiliation. The Chicago Assyrian Dictionary, although digitized, is comprised of over twenty volumes. In languages like Ancient Egyptian, Akkadian, and Sumerian, yet another difficulty is added. Vital dictionaries for these languages require a separate publication of sign lists to understand the script. In many cases, these resources are hand-written in French or German. While the Internet has helped with respect to distribution of some of these materials, the challenges noted above—particularly usability—remain. This is not to say that these resources are inadequate or their creators inferior; rather, there exist hitherto unexplored opportunities to leverage the Internet in order that trade-offs are mitigated. It is therefore necessary to reconsider the goals and priorities of digital dictionaries. Enter Allo–a modern dictionary platform for ancient languages. Each dictionary contains 1,000 high-frequency words. It is designed to be accessible, user-friendly, and suitable for advanced users. To accessibility: we offer it online, charge nothing, and do not require registration. To usability: the interface is optimized for tablet and mobile devices via our design-oriented approach. Because we regarded speed and search as the most critical goals, features like autocomplete, caching, and a recommendation engine are fully integrated. These features may reduce the time it takes to find a given entry by orders of magnitude; based on a small user study, up to 50 times faster than a print book and 5 times faster than a PDF file. To content scope: all entries are aggregated such that they include both a beginner and advanced resource. The federated, or aggregated, model we chose allows us to make updates when new resources are released. Each entry contains between 30–50 individual data fields. Full page images of original source materials and cursor-zooming are built in. Advanced users can find grammatical details, tables of forms, usage frequency statistics, example sentences, and a comprehensive but compact grammar from authoritative sources. We demonstrate the benefits of this approach in our first complete language: Allo Latin, a digital Classical Latin–to–English dictionary kindly provided on https://allo.conscious.ai/latin. Our experience suggests that digital dictionaries must be aggregations of information about a given entry and packaged in a clean user interface. Future innovation likely requires intense cooperation among three concerned parties: a technologist, lexicographer, and language expert.