Does Continuing with Duolingo Mean Acquiring English?

Does Continuing with Duolingo Mean Acquiring English?

Recently, I had an opportunity to speak with a university faculty member who had attended several of my classes. I believe this person intended to express appreciation for my teaching. The person told me, “I am now studying English with Duolingo.”

For a moment, my heart sank.

It was not because I dislike new technology or apps. It was because I felt that the distinction I had repeatedly explained in my classes—the distinction between conscious learning and language acquisition—had not been fully understood.

Krashen’s Critique of Duolingo

In his 2014 paper, “Does Duolingo ‘Trump’ University-Level Language Learning?,” Stephen Krashen examined Duolingo as it existed at the time, as well as the research said to demonstrate its effectiveness.

Krashen observed that Duolingo guided learners through a sequence of tasks, many of them based on translation, and was “clearly aimed at conscious learning.”

Learners inevitably hear and read some of the target language while using Duolingo, so a certain amount of acquisition may occur. However, the central activity is not understanding messages. It is learning designated items, producing answers, and being told whether those answers are correct or incorrect.

The Claim That Duolingo Is More Effective Than University Language Classes

At the time, Duolingo claimed that it had been scientifically proven to outperform university-level language learning. This claim was based on a study by Vesselinov and Grego (2012), which was funded by Duolingo.

Krashen, however, identified serious problems with this interpretation. The researchers themselves did not conclude that Duolingo was better or worse than university language courses. Their study examined how much progress was made by self-selected participants who had responded to an advertisement on the internet and then used Duolingo.

The participants also differed substantially from typical university students.

• Their average age was 35.
• Sixty-nine percent were college graduates, and many had graduate degrees.
• Of the 156 participants who began the program, only 90 remained until the end.
• The amount of time spent using Duolingo varied enormously, from 2 to 133 hours.

The reported mean rate of improvement in test scores was 8.1 points per hour of study. The median, however, was only 3.9 points.

The mean is calculated by adding everyone’s scores and dividing the total by the number of participants. The median is the value in the middle when all the scores are arranged from lowest to highest. For example, if five values are 1, 2, 3, 4, and 30, the mean is 8, but the median is 3. One unusually high value has pulled the mean upward.

In this study, a median of 3.9 points means that half of the participants had an improvement rate of 3.9 points per hour or less. Because the mean of 8.1 points was more than twice the median, it appears that a small number of very high rates of improvement raised the overall mean. Reporting only the mean of 8.1 points could therefore make Duolingo appear more efficient than it was for a typical participant.

Furthermore, the WebCAPE test used to measure progress was a multiple-choice test focused on grammar and linguistic form. It did not measure whether learners could understand new messages in real time, read books they had previously been unable to read, or understand natural conversation. In other words, the study used a similar kind of form-focused task to measure the knowledge that Duolingo had trained.

Test Scores Are Not the Same as Acquired Language Competence

Krashen pointed out that not only Duolingo, but much conventional university foreign-language instruction as well, is based on conscious learning.

Therefore, even if Duolingo produced test scores equal to or higher than those produced by conventional university courses, this would not demonstrate that Duolingo was effective for language acquisition. Both approaches may be centered on grammar, vocabulary, translation, and practice.

Krashen stated the point clearly:

“Conscious learning does not produce true language competence.”

He also noted that approaches based on comprehensible input produce better results on tests of communicative ability and are more efficient in terms of gains per unit of time. As evidence, he cited Mason (2004), Mason (2011), and Mason, Vanata, Jander, Borsch, and Krashen (2009).

In other words, when Krashen criticized Duolingo, he used our research as part of the evidence supporting his argument.

When the Game Is Compelling but the Input Is Not

I am not saying that using Duolingo is inherently bad. It may help learners become familiar with some words and expressions and may help them maintain an interest in another language. Today’s Duolingo also includes stories, listening activities, and conversational features that were not part of the program examined in 2014.

Nevertheless, we must make an important distinction:

Being compelling as a game is not the same as offering compelling content through language.

When learners are attracted by points, levels, correct answers, and learning streaks, they may be enjoying the game system rather than becoming absorbed in a story or message.

Opening an app every day can give learners the feeling that they are studying English every day. But what matters is not how many questions they answer correctly. We need to ask:

• How much language did they understand?
• How long did they sustain their attention to meaning?
• What stories and ideas did they encounter?
• Did they become able to read books they could not read before?
• Did they become able to obtain input independently, without depending on a teacher or teaching program?

These are the outcomes we must examine.

What We Aim For

Pure Optimal Unified Input (P.O.U.I). does not aim to have learners complete one task after another.

In Story-Listening, learners listen to stories told in comprehensible, interesting, and linguistically rich language. Through Guided Self-Selected Reading, they move from listening to reading and develop the ability to choose books for themselves and continue reading.

Our goal is not to produce learners who advance to the next level inside an app.

Our goal is to develop readers who can choose books they want to read, understand them, and continue acquiring language throughout their lives.

Duolingo shows learners how many questions they completed today.

P.O.U.I. values a completely different kind of growth:

What did you understand today? What stories did you experience? How far did you travel along the path toward becoming an independent reader?