57: Making machines learn Fon and other African languages - Interview with Masakhane

Lingthusiasm - A podcast that's enthusiastic about linguistics

Gretchen McCulloch and Lauren Gawne

Science

4.8 • 743 Ratings

🗓️ 18 June 2021

⏱️ 37 minutes

🔗️ Recording | iTunes | RSS

🧾️ Download transcript

Summary

When you see something on social media in a language you don’t read, it’s really handy to have a quick and good-enough “click to translate” option. But despite the fact that 2000 of the world’s languages are African, machine translation and other language tech tools don’t yet exist for most of them. In this episode, your host Gretchen McCulloch interviews Jade Abbott and Bonaventure Dossou of Masakhane, a grassroots organisation whose mission is to strengthen and spur Natural Language Processing research in African languages, for Africans, by Africans. We talk about how they started working on language tech, Bona’s machine translator in Fon, and alternative models of participatory research and collective co-authorship. Announcements: This month’s bonus episode is about the linguistics of Pokémon names! Which sounds cuter, a Pikachu or a Charmander? Which sounds like it would be more likely to win in a fight, a Squirtle or a Blastoise? Even if you're not familiar with these pocket monsters, or if you're encountering new Pokémon you haven't heard of before, you might still have a vague sense of which names sound big or small, cuddly or powerful. This has lead to the creation of the delightful and entirely real linguistic subfield of Pokémonastics. https://www.patreon.com/lingthusiasm Join us on Patreon to learn more, and get access to 51 other bonus episodes! You’ll also get access to our Discord server, where you can chat about your favourite Pokémon names with other language nerds! https://www.patreon.com/lingthusiasm Here are the links mentioned in this episode: https://lingthusiasm.com/post/654293595359182848/lingthusiasm-episode-57-making-machines-learn-fon

Transcript

Click on a timestamp to play from that location

0:00.0	Welcome to Lingthusiasm, a podcast that's enthusiastic about linguistics. I'm Gretchen McCulloch,
0:22.3	and I'm here with Jade Abbott and Bonaventure Dosu from the Masakhani Initiative. Today, we're getting
0:27.1	enthusiastic about natural language processing research in African languages for Africans by Africans.
0:33.3	Hello, welcome to the show. Thank you for having us, Gretchen. Hi.
0:37.0	Hi. Thank you for having us. Thank youchen. Hi. Hi. Thank you for having us.
0:39.0	Thank you both for coming.
0:40.3	So I know that Masakane is a big group of people, so you are two parts of it, but there are other people who are also involved.
0:47.8	But let's start with sort of backing up the step and talking about how you both got interested in language in the first
0:56.1	place. Jade, do you want to start? Of course. So I got interested in language from a very young
1:03.3	age when I, so come from, originally from South Africa, which has a very traumatic past with a lot of division. And one of the things
1:13.6	that when I was growing up, I noticed that I was never taught to speak African languages, despite
1:19.6	the fact I was living in South Africa. And when I did try, learn the languages, I ended up with a lot of
1:26.4	anxiety in trying to get through this. And so being
1:30.7	in computer science, I turned to technology as a means to do this. And the reason that it mattered to me
1:37.6	so much was, and I don't want to be the cheesy person to quote Nelson Mandela, but I'm going to do
1:43.5	it, is he says, when you speak to quote Nelson Mandela, but I'm going to do it.
1:49.5	He says when you speak to someone in a language, they understand it goes to their head.
1:51.7	If you speak to them in their own language, it goes to their heart.
1:54.7	And this I fundamentally believe on the importance of communication.
2:00.6	And so for me, it was a matter of like bridging, like a, not only just a communication divide,
2:02.0	but also like a cultural divide and kind of bringing people together.
2:03.2	So this is why I cared a lot about language.
	...

Please login to see the full transcript.

Previous episode | Next episode

Disclaimer: The podcast and artwork embedded on this page are from Gretchen McCulloch and Lauren Gawne, and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of Gretchen McCulloch and Lauren Gawne and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.