UNC-based researchers developing tool to teach Cherokee language

The Cherokee language is endangered. Very few native speakers remain in America — much less in North Carolina. 

To help produce more fluent Cherokee speakers, a group of professors and students are working to create a software program that will translate English materials to Cherokee. They include Ben Frey, a professor of American Studies and  Eastern Band Cherokeean citizen, teamed up with Mohit Bansal, the John R. & Louise S. Parker associate professor in the UNC department of computer science and doctoral student Shiyue Zhang. 

The project began in 2019, according to Bansal, who is leading the computer science aspect of creating this new tool. 

He works closely with Zhang, whose role is primarily to collect data and help create the translation tool. 

Zhang said the overall goal of the project is to revitalize the endangered language and increase its day-to-day use in the community. 

“We want to increase the exposure of the Cherokee language to the general public, so people can get to know this language,” Zhang said. 

They hope that the translations that come from this software can also help improve learning for second-language learners or students in the New Kituwah Academy, an immersion school for Cherokee language, culture, traditions and history. 

“We can use the samples of good, accurate Cherokee language that appear in that book to create classroom lessons around that second-language learners could benefit from," Frey said. "We could create more activities that kids in the [New Kituwah Academy] could benefit from.”

Frey said this type of technology can help researchers learn more about the structure and rules of the Cherokee language itself by translating materials. 

Many languages have phrases that wouldn’t make sense with just the basic translation rules, Frey said, and one would have to be familiar with the language in practice to understand them.

“We can discover whether Cherokee has things like that in it as well,” Frey said. 

Frey notes that the project alone won’t save the Cherokee language, but he and colleagues are hopeful that it will pave the way for other programs like it. 

“This project’s not gonna save the language by itself, but this project can lead to all kinds of other initiatives that will prove useful in themselves,” Frey said. 

Bansal said that with the user experience and technology being developed for this project, they are hoping that they can bring that work to other researchers looking at endangered languages.

“Another thing that Shiyue is looking at is, ‘Can we relate this to other languages, and work with other communities using endangered languages?’” Bansal said. 

As of right now, the team has a demo version of the website, but there are more steps in the process before the tool becomes fully available and accurate for use. 

“We want to improve the basic translation function before we start releasing it,” Frey said.

Frey said after the tool is used to produce a translation, second-language learners will copy edit it for accuracy, Frey said. Then, a fluent speaking elder will do a final correction and sign off on the data, which is then implemented back into the software. If they come up with a bad translation, they are able to correct it and it can in turn be added to the algorithm.

The project has a long way to go before it becomes fully ready to release, but the team is working hard to get it there. 

Both Bansal and Frey emphasized that one important aspect they want to add to this project is having more people in the loop. 

With the help of fluent speakers, they will be able to make this tool ideal for second-language learners and immersion programs. 

“The next step will probably be trying to recruit more elders who are fluent speakers to use it and see if it's a pretty good tool to use with the [New Kituwah Academy],” Frey said. 

university@dailytarheel.com I elevate@dailytarheel.com