The Week in Review: August 25, 2023

Meta made quite a splash this week with the announcement of SeamlessM4T — a multimodal translation tool that allows users to undergo speech-to-speech translation in more than 100 source languages and 36 target languages.

It’s a pretty major feat, and shows that the company’s hard work with the No Language Left Behind project — which aims to develop high-quality speech recognition systems for all the world’s languages — is paying off.

Meanwhile, West Virginia University could be cutting its entire world languages department. Read up on that — plus blog posts from PoliLingua and BLEND — in this week’s recap.

Got news you’d like to share in next week’s Week in Review? Send it over to our editorial team here.

WVU’s plan to cut foreign languages, other programs draws disbelief (via The Washington Post)

College students planning to major in a foreign language might want to stray away from West Virginia University — a recent budget proposal would eliminate the school’s Department of World Languages, Literatures, and Linguistics. If the proposed budget is finalized as is, the department’s faculty would be cut and majors in Chinese, French, German, Russian, and Spanish would be no more.

The proposal isn’t finalized just yet. Faculty and staff at the university expressed outrage over the proposal in interviews with The Washington Post.

“A university is built around certain core values,” Lisa M. Di Bartolomeo, a teaching professor of Russian studies and Slavic and East European studies told the Post. “One of those core values is intercultural competency. They’re not even going to know what they’re missing. Which is absolutely tragic.”

Introducing SeamlessM4T, a Multimodal AI Model for Speech and Text Translations (via Meta)

While other tech giants like Google appear tightly focused on generative AI, Meta has recently made a breakthrough in machine translation with the recent release of SeamlessM4T. SeamlessM4T is capable of translating speech-to-text in almost 100 languages (and speech-to-speech in nearly 100 source languages, or 36 target languages). The release builds on other projects Meta’s focused on in recent years, like the company’s No Language Left Behind project.

“In keeping with our approach to open science, we’re publicly releasing SeamlessM4T under a research license to allow researchers and developers to build on this work,” the company shared in a news release. “We’re also releasing the metadata of SeamlessAlign, the biggest open multimodal translation dataset to date, totaling 270,000 hours of mined speech and text alignments.”

National Hispanic Cultural Center working to preserve a particular Spanish dialect (via KQRE)

As MultiLingual reported months ago, New Mexico is home to a distinctive dialect of Spanish that’s not spoken anywhere else — but that dialect is vanishing slowly. The National Hispanic Cultural Center has been working to preserve it in recent months. Learn more about the center’s efforts at KQRE.

“The recent archives that have been published by different parts of national newspapers saying that specific form of Spanish spoken in rural Northern New Mexico particular is set to be extinct in 18 and a half years,” Zackary Quintero, acting director of the center told KQRE, a local news channel in Albuquerque.

Can Basque, Catalan and Galician really become EU languages? (via euronews.)

Spain’s prime minister is urging the European Union (EU) to adopt a handful of new official languages: Basque, Catalan, and Galician. The EU hasn’t added a new official language to its roster since 2013, with the addition of Croatian. Analysts believe that the move is likely a way for the prime minister to rally support from speakers of Spain’s regional languages, particularly in Catalonia.