African Languages in Danger of Disappearing -- Interactive Atlas

Click here to use the interactive atlas.

Open Question: What about focusing our efforts on these disappearing languages before it's too late? What can we, as computational linguists, do; should we do something?

Click here and here for more info on the interactive atlas of endangered languages.

Even with the best tools we

Even with the best tools we could create for language documentation as a last-ditch effort for the most endangered languages listed, I think the problem of our efforts being properly utilized would be three-fold: we would need willing, computer-literate documentary linguists willing to use any tool; we would need an expandable cross-linguistic tool with a very sharp learning curve so that actual annotation time would be greatly lowered; finally, we would need the time to get as much data as possible before the last of the willing informants in the speaker group dies out. Clearly, this is fraught with difficulty, and perhaps our efforts could be best directed elsewhere, proactively preparing for speaker groups that are less endangered but nonetheless decreasing steadily in population.

For those with increasing access to the information age, I personally feel that one of the strongest deterrents to transmission of languages to younger generations is its lack of use outside of a familiar context; any upwardly mobile speaker group would surely want their children to speak a language that will put them at an advantage for greater socioeconomic privilege. However, I believe this can be ameliorated by making more information technology available for these groups. Direct machine translation to an indigenous African language would not be the easiest thing in the world, but when we're looking at machine translation from, say, Swahili to Kuria, the task becomes significantly less daunting. Depending on the intermediate language and the language for which we are translating, this task could be almost totally in the realm of finite-state computational complexity.

But in the meantime, of course, I do believe that there needs to be an established repoire with those in documentary linguistics, whereby our efforts are directly in response to what they need to ease their job. Of course, it's not a one-way street; documentary linguists could certainly benefit from learning some basic computational linguistics, so that the work we put in won't be left unused because it's based in Unix or doesn't have a sleek GUI.