Gallaudet University Documentation of ASL (GUDA) - Documentation IS Representation

Hochgesang, J. A., R. Treviño, J. Willow, & E. Shaw. (2019, March). Gallaudet University Documentation of ASL (GUDA) – Documentation IS Representation. Presentation at Georgetown University Roundtable (GURT) 2019 – Linguistics and the Public Good, Georgetown University, DC.

We detail efforts to digitally centralize and organize existing ASL video collections at Gallaudet University. The underrepresentation of sign languages in general and of sign language corpora in North America has led to repeated misunderstandings of the project and devaluing by stakeholders and funders. However, it is through the description and subsequent analyses of this unique dataset that the community value will become apparent. The aim of our presentation is to discuss the urgency of such a project to American Deaf communities. We have access to vast video resources beginning in the early 1900s that showcase multiple discourse genres produced by deaf people from all over North America. Although not originally collected for the purpose of creating a corpus, these films have the potential to become a monitor corpus (McEnery & Hardie, 2011) that is representative of ASL communities.

Before the advent of film and even continuing well after, representations of signed languages took different forms, none of them dynamically representative of the language. Still today, no signed language has its own writing system in widespread use. English glosses have prevailed as the primary means by which linguists have accessed, analyzed and published findings about ASL. The use of English glosses without accompanying videos often leaves researchers and even ASL users wondering which sign or variant is actually meant. Glossing can also lead to a false one-to-one association of English to ASL, stripping ASL of its own unique semantics, i.e., “tyranny of glossing” (Slobin 2008).

Technological developments now allow signed language videos to be widely accessible. Although the challenges of representing a language with no written form continues, best practices of signed language corpora (e.g., Fenlon et al 2015) provide solutions that enable accessible representation of ASL through digitally shared resources that can be time-aligned to primary data (ELAN, 2018). We use ASL Signbank (Hochgesang et al, 2018), which directly links ID glosses to videos of citation forms along with grammatical information. Data is enriched by free translations and forever associating the primary data to the transcripts reduces the dependency on English. We will also collect complete metadata, essential to contextualizing the data which is necessary to make representative claims about an extremely diverse community. Here, we argue that corpus linguistics is a way to prioritize the representation of sign languages in their own right through digitally shared resources. GUDA then is the representation of ASL data through combining source videos and data processing protocols.

As concerns this project’s relationship to the public good, research on signed languages is largely inaccessible to deaf communities. The academic presentation and predominant use of English can make access to research findings difficult. This project will make it possible to easily share work online, through the use of a bilingual web platform shaped by community input and involvement. In short, it is the primary aim of this project to compile a representative, accessible and lasting ASL resource for, by, and with the Deaf ASL community furthering the ideal of representation through documentation.