Luke D. Gessler

I am a first-year Ph.D. student in computational linguistics at Georgetown University, where I am a member of Corpling@GU and GUCL. Previously, I was a software engineer at Epic, and before that, I completed a B.A. in linguistics and computer science at the University of Virginia. I am interested in computational linguistics, specifically in the context of endangered language documentation and development and linguistic annotation.


Roughly half of the world's languages are facing extinction. Many of their speakers want them to be documented, but there is much more demand for documentation than there is labor to fulfill it. How can we fill this labor gap?

I think computer software is a critical part of the solution. Traditional documentation methods can benefit from software facilitating or fully automating its most time-consuming and repetitious tasks, like glossing and data management. Software may also enable new documentation methods that allow people who aren't highly trained linguists (often the "limiting reagents" in the documentation process) to contribute more than they ever could before.

So, what could this computational assistance look like? Some examples, from more to less obvious, include:

  • a comprehensive software suite for field linguists that is, in a word, good enough to convince documenters it's worth the trouble
  • quality of life features for the programs that linguists already use, e.g. a plugin for an annotation application that can correctly parse a word 90% of the time, saving the documenter much manual labor
  • an app for speakers of the target language who are not linguistically trained that somehow allows them to contribute useful documentary labor
  • an app that can deduce information about the target language through interactions its (native speaker) users have in an activity where the overt purpose is not to document the language, similar to the ESP game


lukegessler \( \mathcal{A}t \) gee male dought calm


Find my posts here.