Multimedia dictionary of American Sign Language
Sherman Wilcox, Joanne Scheibman, Doug Wood, Dennis Cokely, William C. Stokoe · 1994 · Proceedings of the First Annual ACM Conference on Assistive Technologies (Assets '94) · doi:10.1145/191028.191031
Summary
This paper describes the Multimedia Dictionary of American Sign Language (MM-DASL), a Macintosh application functioning as a bilingual ASL-English dictionary that presents signs in full-motion digital video using Apple's QuickTime technology. The project, funded by the National Institutes of Health, builds on the foundational 1965 Dictionary of American Sign Language on Linguistic Principles (DASL) by William Stokoe — the first true linguistic dictionary of ASL. The authors frame the work as developing "literacy technologies for the deaf," arguing that while writing technologies for spoken languages are ubiquitous, equivalent tools for signed languages barely exist. The MM-DASL was developed by two teams: a lexicography team (based in Maryland) consisting of deaf ASL-native consultants, bilingual consultants, and video specialists assembling approximately 3,000 signs; and an application development team (in Albuquerque) of linguists and programmers building the software from scratch in C++. The dictionary provides three search modes: English-to-ASL word search (not assuming one-to-one correspondence between English words and ASL signs), ASL formational search (looking up signs by specifying handshape, location, and movement features), and fuzzy searching in both modes.
Key findings
The ASL formational search interface was the most technically innovative and challenging component. The team created a novel "filmstrip" interface element — a scrolling strip of QuickTime video frames showing handshape, location, and movement values that users could drag and drop to specify search criteria. For dynamic handshapes (opening, closing, bending, wiggling), each filmstrip frame was itself an entire QuickTime movie showing the transition. The fuzzy search system addressed the critical problem that users often cannot precisely identify the formational features of a sign they've seen — handshapes like "S" and "A" are perceptually similar. A distance metric stored for each pair of formational values (on a 1-10 scale) enabled approximate matching, with results sorted by closeness of match. The system's modular design using QuickTime movies as interface elements (not just content) made it localizable to any signed language — developers could create dictionaries for British Sign Language or Ecuadorian Sign Language simply by creating new menu resources and movies for that language's formational values. The English search correctly handled the many-to-many mapping between English words and ASL signs (e.g., "run" maps to separate signs for "run a race," "run a business," "run for office").
Relevance
This paper is historically significant for several reasons. It represents a collaboration with William Stokoe, who established the linguistic study of sign language and whose 1965 DASL proved that ASL was a genuine language with its own phonological system — not just a collection of pantomimes. The MM-DASL applied emerging multimedia technology to a profound equity issue: deaf people whose primary language is ASL lacked the basic reference tools (dictionaries) that hearing people take for granted. The concept of searching for signs by their formational features rather than through English translation was revolutionary — it treated ASL as a first-class language deserving direct access. The fuzzy search approach anticipated modern approximate matching techniques in sign language databases. For practitioners, the paper demonstrates that accessibility tools for deaf communities must respect the linguistic independence of sign languages from spoken languages, and that video-based digital resources are essential for languages that have no widely adopted written form.
Tags: sign language · deaf accessibility · multimedia · dictionary · American Sign Language · digital video · linguistics · deaf culture · literacy