An Interview with Dr. Dustin Crowther

In this podcast, Dr. Dustin Crowther (Assistant Professor of Second Language Studies at the University of Hawaii at Manoa), shares the findings of his recent co-authored papers on the language learning app, Duolingo. Pedagogical implications with regards to the affordances of mobile-assisted language learning (MALL) are also discussed.

Dr. Crowther’s work can be found here:


Interviewer: Hi, this is Ann from Multiʻōlelo. In this episode of the Multiʻōlelo podcast series, we’ll be discussing two recent research articles about the language learning app, Duolingo. Our guest is Dr. Dustin Crowther. He is an Assistant Professor of Second Language Studies at the University of Hawaiʻi at Mānoa. He’s areas of expertise include second language speech production and perception, Global Englishes and methodological practices in second language acquisition research. He serves as the TESOL Quarterly’s editor for research dissemination and is also an advisor for the Multiʻōlelo initiative. Today we’re very honored to have Dr. Crowther come in to discuss two of his recent co-authored papers on the mobile-assisted language learning app, Duolingo.

Interviewer: Hello, Dr. Crowther.

Dr. Crowther: Hello, thank you for having me.

Interviewer: So, I came across your studies about Duolingo and I think that’s very interesting because I read that you and your colleagues were actually the participants in these studies as well. Can you tell us a little bit about these studies?

Dr. Crowther: Sure, I can give you a little bit of framing about the study published in ReCall, a working paper study, as well as a third study that I wasn’t part of the writing process but came out of a larger project.

Dr. Crowther: So you mentioned that, in especially the ReCall study that your audience can see on the form on the screen, it came out of a class project. So there were nine of us in total plus our professor. And basically we served as researchers as participants where we were investigating our own learning through the use of Duolingo. We studied Turkish during the study. 

Dr. Crowther: I mentioned that there were a few different publications tied to this project. We have a publication in ReCall, my colleagues who worked on the project have a publication in the journal Languages, which took a more narrative perspective on the use of Duolingo. And then there’s a working paper that I put together with Dr. Kathy Kim and Dr. Shawn Loewen, in which we were sort of looking at the technological side of Duolingo, and to what extent the pedagogical approaches were informed by what we know from instructed second language acquisition research. 

Dr. Crowther: So while the three papers are distinct papers, there is definite overlap between the three of them. So what’s in one paper will almost definitely inform what are in the other papers as well. And at the center of these papers is the fact that we were the participants learning through the use of Duolingo.

Interviewer: I see. That sounds very interesting. It feels like as teachers we can try that out with our students as well. 

Interviewer: So, what were the key takeaways from these studies. Can you tell us a bit about that?

Dr. Crowther: Sure. Since the study in ReCall, the study titled “Mobile-assisted language learning: A Duolingo case study”, was the empirical study that we conducted, I can talk about the takeaways there. 

Interviewer: Okay.

Dr. Crowther: What we did in that study was we wanted to follow up on the Duolingo claim of 34 hours potentially being equivalent to a semester of university study. So we all decided to choose a language, in which none of the participants had familiarity with. So we chose Turkish. There was no one in the class who had previously studied Turkish or had any knowledge of Turkish, and we all endeavored to use Duolingo for 34 hours and see to what extent we could develop our abilities.

Dr. Crowther: In order to test this, a colleague of ours who was teaching first-year Turkish at our university actually developed a Turkish 101- I think Turkish 151 test was what we referred to in the study, to test our knowledge. And that test took into account what would generally be taught in a first-semester Turkish course as well as a certain cut-off point on the Duolingo Tree. For those not familiar, in Duolingo you typically move down a topic tree as you learn different grammatical and lexical and even phonological considerations in the language.

Dr. Crowther: What we found in performing the study was, a, we all definitely learned Turkish. We knew more Turkish than we began. And I remember we wrote this in the paper, and while it seems sort of silly to make that claim because you would expect that, it is important that we can document the fact that language learning does occur through the use of the app. 

Dr. Crowther: One of the concerns is that for most of us when we completed the study, we- I think only three of us actually- I say three of us because- I shouldn’t because I was not one of the “us” in the three, but only three members of the team actually scored high enough on the 151 test to pass the course itself. So our test scores through Duolingo learning weren’t aligned with what would be expected of a first-year university learner of Turkish based on that specific program at the university that we were at.

Dr. Crowther: One of the interpretations we had is that, and this is something that gets discussed in the working paper that’s referenced as well, is that for the most part, the activities within Duolingo are really sort of promoting more, I would say, explicit knowledge, right? Recognizing grammatical rules, looking at word order, memorizing vocabulary…

Dr. Crowther: So, when we’re being tested, it’s more like testing our explicit knowledge. And one of the considerations is to what extent in spontaneous communication can we draw upon that knowledge as opposed to what is sort of more automatically available to us. But what is sort of automatically available to us isn’t something that Duolingo is focusing on. However, in communicative-based classrooms, that we’re seeing more and more- especially in North America for the way in which languages are being addressed, you’re going to have more of a focus on that sort of spontaneous usage of the language or use of the language in communicative activities.

Dr. Crowther: So, there is a disconnect a little bit from our experiences between how well we actually did for Turkish 101 based on 34 hours that- that means that the 34 hours study that was originally published in a white paper for Duolingo focused on a test that focused on this explicit knowledge. Therefore you would expect relatively solid gains in that way. If you focus on sort of these explicit activities promoting explicit knowledge development when you’re learning, then the test that is testing that knowledge is going to help in that way. The test we took had questions in that that sort of drew upon both our explicit and our implicit abilities with the language itself. 

Interviewer: Awesome. So-

Dr. Crowther: So- Oh, sorry. Go ahead.

Interviewer: Sorry, I just- In one of the papers in Crowther et al., the working papers, you and your colleagues drew on two frameworks. One is the CALL (Computer-Assisted Language Learning) framework and the other is the MALL (Mobile-Assisted Language Learning) framework. And I was wondering if you could tell us why were these frameworks chosen, and what are they about.

Dr. Crowther: Sure. So the paper that I wrote and was published as a working paper in the Michigan State University’s Second Language Studies program, I developed with Kathy Kim, who’s now at Boston University, and Shawn Loewen, who was overseeing the project. 

Dr. Crowther: One thing to keep in mind here is, at the time that we developed this project, mobile-assisted language learning was really starting to take off. And we can say that between now, 2020, and 2017, which means this project would have been probably 2016 that we’re doing data it’s already been four years. So a lot has occurred in research since that time. I just want to sort of frame it within that timeline when the study was conducted.

Dr. Crowther: At the time, Reinders and Pegrum had a framework from MALL that was focusing very much on the affordances of mobile-assisted language learning. So what is it that being able to study a language, for example, using our smartphones, allowed us to do that we couldn’t do in the classroom. Of course, some of those affordances are, you know, you can access language learning materials basically any place, any time. 

Dr. Crowther: So if you look at the ReCall study, there’s examples of us standing in line at a Starbucks using Duolingo. One of our colleagues was on a trip to Africa in the back of a Jeep, doing Duolingo in the back of a Jeep in the middle of Africa somewhere. She might have been on Safari. I can’t remember at this point. Right?

Interviewer: Wow, so you can basically learn languages anywhere then.

Dr. Crowther: Yeah. Right. So it’s really looking at these affordances. Other affordances, any time, any place, but now you can go online. So I don’t know what a word means, well we have access to the Internet, we can go and look it up. Something cultural comes up. So this is just really, really base level of these affordances. 

Dr. Crowther: But what the Reinders and Pegrum framework didn’t do, at least in our opinion was, really look at the language learning process itself. It made references to input and a few other key constructs in second language acquisition, but it was more focused on sort of assessing to what extent the program took advantages of the affordances of MALL.

Dr. Crowther: We were interested from an ISLA perspective, where we’re looking at the systematic manipulation of the mechanisms of learning. Because, a program like Duolingo or other programs, Babbel, Busuu, Rosetta Stone — In essence, they’re making pedagogical decisions to promote our learning. So we’re curious to what extent Duolingo was actually drawing upon what we know about the language learning process in making decisions.

Dr. Crowther: So Carol Chapelle had a framework that was used for computer-assisted language learning. And while MALL in some writing has been promoted as an extension to CALL, MALL has been promoted as part of CALL. I would argue MALL is something different given the affordances. However, the CALL frameworks still gave us a shape that didn’t exist in other framework for analyzing it. So this is the case where I would say- I don’t know if I would say we used Chapelle’s framework as much as we adapted Chapelle’s framework, to analyze to what extent Duolingo was making use of what we know from an instructed second language acquisition perspective.

Interviewer: I see. Awesome. Thank you so much. I think this is gonna be my last question, which is very important. What are some of the implications for program development, whether for language learning app development or for when integration like a language learning app into foreign language instruction.

Dr. Crowther: Yeah, I think the main point- this is actually what I was gonna mention a few moments ago is, programs like Duolingo and I mentioned Babbel, Busuu, and Rosetta Stone, and I know there’s a few others, they provide these opportunities to gain quick, and in the case of Duolingo, free access to language learning materials.

Dr. Crowther: I think what’s important to recognize is that second language acquisition is a much larger process that requires not just studying of the language through the mobile app, but providing yourself opportunities to go out and use the language as a means to whether you’re trying to solidify knowledge you have, whether you’re trying to proceduralize knowledge you have, whether you’re trying to allow your innate abilities to acquire language to where- coming from these different theories, and these different perspectives on language acquisition.

Dr. Crowther: Providing opportunities and recognizing that these tools are great support tools for the process of language acquisition, they shouldn’t be- and I know Luis von Ahn [founder of Duolingo] I don’t know if I said his name right, he talked about this a little bit when he was talking about Duolingo, and like that, they’re not intended to be the be-all and end-all to learning a language, so to me what’s important to know is how do you supplement this study on an app like Duolingo with the other components of language acquisition that we know playing important role in developing your ability to use the language, specifically the ability to use the language during spontaneous communication.

Interviewer: Yeah, that is the hardest part of language learning in my opinion.

Dr. Crowther: Yes.

Interviewer: To use something spontaneously in interaction.

Dr. Crowther: Yes. Right.

Interviewer: Anything other things you would like to say before we close our interview?

Dr. Crowther: I think I would definitely say, looking at the area of mobile-assisted language learning in general and especially at these industry-based apps is, we’re seeing a really promising trend where, for example, Duolingo began very much drawing upon the knowledge of computer programmers in the app, and now we’re starting to see them hiring more and more from those with backgrounds in second language acquisition and those with backgrounds in second language testing and assessment.

Dr. Crowther: So I think there’s expectations at this point that we’re gonna keep seeing these apps available continue to evolve and improve and provide us with more and more options. And I think the challenge that we’re gonna start seeing is, how can teachers make use of these apps as part of their larger curriculum, whether you’re bringing them into K-12 contexts, using them in foreign language classrooms here in the United States, or using them in English as a foreign language contexts like Japan, China, Korea, other parts of the world…

Dr. Crowther: And then how do determine how they can use them in those contexts, and then how do we convey this knowledge to them as well, which I think very much falls in line with the goals of Multiʻōlelo in creating this dialogue between researchers and teachers in such a way that allows them sort of this harmonious communication between these different areas that are all important in this process of second language acquisition.

Interviewer: Yeah, it’s so exciting to see where technology, as it keeps on advancing, what kind of new algorithms will be available to help better support the language learning process. That’s very exciting.

Interviewer: And lastly, I just want to maybe advertise Dr. Crowther’s website here. If you have any questions, you can check out his many, many, many studies published in recent years, and-

Dr. Crowther: There’s also some wonderful scenic pictures of Vancouver, Canada. You’re welcome to look at as well on my webpage.

Interviewer: Awesome. Okay, thank you so much for giving your time to us today to tell us about Duolingo and your paper- uh research process. Thank you so much, Dr. Crowther.

Dr. Crowther: Thank you for having me.

Interviewer: Thank you.

Credits: Ann T. Choe (interviewer, post-production editor, transcriber)

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s