Rethinking Accessibility through a Summer Internship in Computational Linguistics

September 19, 2021

By Caroline Gihlstorf

This is the fourteenth in a series of blog posts by participants in the 2019 ACLS Digital Extension Grant project “Ticha: advancing community-engaged digital scholarship” (PI Lillehaugen) published with the Community-based Global Learning Collaborative and the Ticha Project. Previous blog posts are available here: (1) Lillehaugen/January 2020; (2) Flores-Marical/February 2020; (3) Kawan-Hemler/March 2020; (4) Lopez/July 2020; (5) Kadlecek/1 August 2020; (6) García Guzmán/15 August 2020; (7) Park/September 2020; (8) Zarafonetis/October 2020, (9) J. Lopez/Nov 2020, (10) Velasco Vasquez/February 2021, (11) Lillehaugen/March 2021, (12) Plumb/April 2021, (13) Molina/August 2021.

I’m a student at Haverford College studying Computer Science and Linguistics. I was a research assistant on the Ticha Project during Summer 2021, funded by the Haverford College Center for Peace and Global Citizenship. My work with Ticha gave me the opportunity to apply my knowledge in both areas of study to various aspects of the project. It was also an opportunity to learn about Ticha’s goals for its resources — specifically around how to make them truly accessible. 

One of the first things I learned about the Ticha Project was that it aimed to make resources accessible. “Great!” I thought. “Free, digitized resources are certainly an example of accessibility.” I still believe this, of course, but after my work with Ticha I realized that my original notion of accessibility only scratched the surface of what the project aims to achieve. Since I was not someone for whom Ticha’s resources are designed, I didn’t have much experience trying to access or navigate them from a user’s perspective. Ticha taught me that accessibility spans many dimensions, and that it is never just enough for a resource to simply exist.

My work in Ticha spanned three main areas: encoding transcriptions of texts into XML files (XML is a markup language that assigns labels to text for more advanced software to compute), making small modifications to the Ticha website, and writing a chapter for Caseidnyeën Saën — a textbook with lessons centering around different aspects of Zapotec language, history, and culture. Each of these areas showed me new ways to determine how accessible a resource was, and what could be done to make it even easier to use.

One of the most important things I learned was to center the user, both by seeking feedback and putting myself in the user’s shoes. When I was writing the chapter on how to read a type of linguistic analysis, “accessibility” included not only the existence of the chapter itself, but the language I used to communicate my ideas. After receiving advice to move paragraphs around, reword sentences, and insert more visuals, I realized that my job was not just to write information in a way that I myself could understand, but to cater to an audience who was new to the material and likely nervous about learning it. This was a helpful reminder to consider the reader in addition to the content you want to present to them. It was recommended to me that I include sections telling the reader that although the material seemed scary, they had the ability to learn it. This helped my focus shift from simply producing content to producing content that was inviting to the apprehensive reader. 

Screenshot sharing instructions on how to read interlinear analysis.
My work on the Ticha website also helped me realize that availability does not equal accessibility. Links to the Caseidyneën Saën textbook, for example, had been available on the Ticha website, but they sat within a block of text that was uninviting and hard to read. We changed the webpage to include an image, buttons to click for the English and Spanish versions of the text, and replaced the block of text with bullet points describing the text’s content and purpose. We changed nothing about what information the page offered — it was our restructuring that made the information more accessible. 

Screenshot from Ticha webpage, demonstrating improvements discussed in narrative.My text encoding work followed a similar trajectory. As I began to encode more and more of the the Aguero Miscelaneo (the text I worked primarily with this summer), I noticed uncoded marginalia (written or typed text in the margins of the main text that are separate from and/or supplementary to it) on most of the pages. I learned that there was not yet an official way to encode marginalia, but I thought it was important to encode it along with the main text — especially the marginalia that was less legible. I proposed a system of marginalia encoding, and began implementing it on the text I was working on. The system is by no means perfect or complete, but it now exists in some form. Taking the position of the reader allowed me to notice this place for improvement.

Screenshot of the process of encoding marginalia.My work with Ticha has shown me that “accessibility” means so much more than creating a resource and making it available. Accessible resources invite users in, are easy to navigate, and can always be improved. I’ve been so fortunate to have received input and feedback on my work from my mentors at Ticha, who have helped make everything I’ve done much better. Feedback — especially from perspectives most implicated by the resources — is most valuable in ensuring that resources are both present and meet their users’ needs, and shows that there is always room to increase accessibility.


Caroline Gihlstorf, a junior at Haverford College, is an aspiring computer scientist and computational linguist who loves to read and write (both prose and code!) and is always looking for new things to learn.

  • update-img-new

    Get updates on what's new in the Campus Compact Network