1. Forum
  2. >
  3. Topic: Duolingo
  4. >
  5. Introducing the Learning Scie…


Introducing the Learning Science team at Duolingo

Hi, my name is Bozena, and I'm the head of Learning Science at Duolingo. That means I’m in charge of the team responsible for improving how Duolingo teaches. We do a ton of research and experimentation to make Duolingo more effective. I wanted to introduce myself since you'll be hearing more from me in the future!

So who am I?

On the professional side: I have a PhD in linguistics with primary expertise in language learning. Before joining Duolingo in 2015, I worked as a researcher and lecturer at Northwestern University in Chicago. In my research I’ve been mainly interested in how people use their prior language knowledge when learning additional languages, and what kinds of pedagogical techniques lead to most effective learning. I conducted many controlled experiments on language learning in the lab, drawing inspiration from the many hours I spent teaching languages in a classroom (in addition to teaching about languages in my linguistics courses).

One of the coolest things about working at Duolingo is our scale and the data we have access to. With over 300 million users globally, Duolingo is probably the largest-scale experiment on language teaching in all of human history. This gives us unprecedented data insights into how people learn. Those insights then translate into product improvements, which in turn have real impact on people’s lives. This is a learning scientist’s dream!

On the personal side: I grew up in Krakow, Poland, which I think is the most beautiful city in the world (clearly I'm not biased at all). I love languages. When I was little, my plan was to start learning a new language every year. I didn't quite achieve that, but I did end up studying a few. My best ones are Polish, English, Spanish, and French, but I also studied German, Arabic (Modern Standard & Egyptian), Russian, Italian, Turkish, Brazilian Portuguese, and Somali. Too bad Duolingo didn’t exist when I was growing up; I would have been a super power user! Other than languages, I love dancing Argentine tango and I do Olympic weightlifting as my regular workout.

I look forward to talking more with all of you. Follow me on Twitter @BozenaPajak!

October 22, 2018



Hi neczka,

thank you for your personal and friendly post. I think many duolingo users are keen to get more in contact with the duolingo staff. Your vita is very interesting and shows what languages can effectuate. You will receive many wishes, complaints and demands here, that's for sure. :-)

i couldn't agree more when you write 'its too bad duolinge didn't exist when I was growing up' (and internet/forums etc generally). It's a good thing. Many users here in this forum don't feel 'heard' whith their wishes and suggestions so I hope your post is a step forward.

so let me express my wishes (and hopes).

'I look forward to talking more with all of you. Follow me on Twitter'

Please don't let this offer of discussion and interaction slip away into a 'foreign' platform but let it stay here "at home". Not everyone has or wants the so called 'standard social media' to follow a discussion.

Please remember the users (and members of this community) who are not so fit in English language and spread your post in as many languages (and their forums) as you can.

I'm really looking forward to your posts, best regards Angel


Very cool. I'm all for more efficient language-learning!


Welcome neczka! It seems like Duolingo and your passion for language analysis are a perfect match!


ooooohhhhhh ! This is soooooo exciting to hear about.

It certainly is a fantasticly interesting ride to continue to see all the changes and striving for improvement Duolingo has constantly developed, and is developing, so as to bring innovative and engaging language learning, and to do it with the inspiring and honorable dream to achieve this while making these resources available at no cost for the learner.

Certainly quite a balancing act, including being in the rough and tumble of a global market place.
I take my hat off to those behind the scene, such as yourself, who are involved in also developing the cutting edge of this interface, in a constantly changing environment in so many ways.

As a volunteer passionate about the value of humans to be able to do this, especially at this critical current time in the challenges that lay before all of mankind, I also applaud your stance to reach out and also connect in such an inspirational way with this public community.

It is a pleasure to meet you, and to be inspired by the work you do, Bozena.


Thanks for the introduction. I'm sure working at a place like Duolingo with so many other talented linguists must be fascinating.

I look forward to seeing your work on Duo, and how people use their prior language knowledge when learning additional languages. That's a subject I'm interested in and would love pointers to any academic papers you have on that subject.

I've been on Duo for two years now, studying Spanish and German. I'm also fluent in French and Japanese, and have spent almost two decades living and working overseas, ten of that in Japan.

Again, looking forward to hearing more about your work.


Personally I've seen a huge jump in my ability to learn language since I've been on duolingo. I used to struggle a lot in foreign language courses because my instructors would stuff grammar down my throat and grammar isn't my strong suit. I was having the same difficulty with Duolingo in the beginning, but when you switched over to the crown system I've been able to move along a lot more. My strength is in vocabulary building and the crown system has allowed me to focus on that so I can eventually build up my reading skills. After that, I plan to focus on grammar.


Hi Bozena, can you point us to some of the research that has been done on 2nd language acquisition by adults. Thanks.


There's plenty of research on this topic. I'll be periodically tweeting about new findings, and my team and I are planning to write regular blog posts about research and how it translates into the improvements we're implementing on Duolingo. Make sure to check out our blog at http://making.duolingo.com/


Hello Bozena, your language skills are quite impressive. I wished I could speak at least half of the languages that you are capable of speaking. Great job! Once in a while, I am searching for duolingo stories, especially for Italian and Japanese. Unfortunately, there are only story sets available in Spanish, Portughese, German and French. What I noticed is that these stories are all different from another. Wouldn't it be easier, if one had a source story or set in English and then quickly translate and duplicate it into other target languages? What do you think of my idea? Please let me know


This is actually (mostly) what we do. The stories might be ordered differently in different languages, but we do mostly create them first in English and then translate to other languages. So yes, this is a great idea to scale more quickly :)



Wonderful news that you have joined the Duo team! Duolingo is an amazing resource. The only other similar resource that I know is Khan Academy (math, sciences, SAT prep, etc). Their mission is similar: a world-class education, free for everyone, everywhere, forever. Go team!

As a multilingual person, with teaching experience in English and French, plus extensive experience learning in what used to be called the "language lab" environment, self-teaching, immersion learning, etc., I'd like to "put an oar in" on some of the discussion around the (maligned) Grammar Translation method, and make a suggestion.

1) Is Duo really GT? My understanding of GT is that it involves formal presentation of the grammar (usually in the learner's mother language), along with the necessity of "learning the language of learning" -- the names for parts of speech, declensions, case endings, etc etc etc. (That aspect of GT can drive learners crazy). The grammar is then reinforced by translation exercises, both from and into the target language (ie, the one you're trying to learn). It's a very, very old method (the Romans used it).

So, is DUO only GT? Not by my experience, although GT is obviously a significant component. Somewhere early on, I saw DUO described as "deductive learning." You may have noticed that DUO does not present explicit grammar rules. The courses are (more or less) carefully staged so that you use your natural deductive reasoning to figure out what is going on. We all do this in our mother languages too -- when someone is speaking too fast or slurring; when the topic is one we're not familiar with, etc. With enough trial and error (crash & burn?) we eventually figure it out.

The crucial point with deductive programs is that they must be very, very carefully structured so that there is a useful (natural?) flow to the introduction of new concepts. Comments in the DUO forums show that some courses have done this more successfully than others. The beta versions are often rough. IMHO, one cause is failure to grasp the careful structuring that is needed for deductive learning to succeed.

Learning needs to proceed on the formula of "i + 1" -- the information base already established, plus one manageable unit (yeah, I know, fuzzy language). There are examples in the forums that show when a "great leap forward" has left students hanging: the "Where did that come from?" questions show a leap too far.

The "grammar" comes out in the individual language forums, where learner questions become "teachable moments" -- again this is pretty much what happens in daily life.

Years ago, I skimmed a Language Lab method in a French program called Voix et Images de France, where the students (I was a lab assistant) had ONLY pictures and aural exercises for some time, then moved to aural/oral with pictures. Print was only introduced after several sessions. Lots of deductive learning going on, along with the immersion aspect!

2. Is Grammar-Translation useful? By my experience in 11 languages (Esperanto makes 12): YES. But it's not useful at the beginner level, where the Direct Method shines.

My teaching experience was mainly in Direct Method (I taught all levels, from absolute beginner to MBA candidates). At intermediate and advanced levels, where concepts become increasingly complex, relevant (!!) GT exercises can effectively tease out misunderstandings, refine and expand vocabulary, etc.

My Latin courses were almost entirely GT (that Roman heritage!), so I've also experienced a GT overdose. That still didn't make me dismiss GT 100%. For myself (and for my advanced students) GT was/is a useful component of a varied instructional program.

3. SUGGESTION: Language Lab component
I have many ways of ramping up the spoken component of any language I am studying: listen & repeat; read aloud; doubling the spoken sample, etc. I train my vocal apparatus the way I practice a musical instrument.

Languages are physical things -- something that's often overlooked. People assume they will have to practice the fine hand-eye-brain coordination needed for music or sports. However, they assume that the infinitely complex speech mechanisms in their bodies will just "naturally" know what to do. The brain-breathing-vocal apparatus-muscular control needed to produce speech needs training and exercise the same way that sports people (like Olympic weight lifters) need to train. And as musicians and athletes know, every instrument and every sport makes different demands on the body. Same with languages.

Language laboratories used to be a vital part of that training, enabling learners to attain native-like pronunciation and reflexive, native-like accuracy in grammar and syntax. They were very, very efficient. Every piano teacher and martial arts teacher (I did both) will tell you that repetition and practice are essential. Use of the language / instrument / movement must become reflexive, so that you can respond instantly without having to contemplate each muscular twitch. Wax on, wax off.

So my plea for DUO's tech team is: Can you (one day in the optimistic future) build a language lab component into the DUO programs? With digital technology, so much is now possible. Mango Languages uses a Language Lab component in all their courses: Listen, repeat, play back and compare, try again until you're satisfied with the match.

The same principle is used to reinforce grammar and syntax: an aural cue is given, the student responds, the "instructor voice" gives the correct response, student compares, corrects, etc.

Language lab components should be brief: never more than about 10 minutes at a stretch, because they are intensive learning.

a) Language sensitization: ("ear training" in music) -- using minimal pairs, similar utterances, etc. to train the ear to distinguish sounds of the target language. (The first week of my Russian course was entirely language lab, and ONLY at this level.) Particularly effective for tonal languages, and languages with rare phonemes. Students must be able to hear a sound before they can produce it.

b) Vocal cues to pictures-only introduction of vocabulary -- student hears an utterance, chooses the appropriate image (more language sensitization happening here, some direct method learning)

c) Grammar / syntax exercises

d) Sing-along songs, jazz chants, etc.

4. What's the rush? Before dissing a DUO course (or any other language course, including university level), it's good to take a deep breath and consider what learning a language involves.

FIRST QUESTION: Do you use your mother language perfectly?

Remember that this is a language you have "studied" (informally and formally) since the day you were born (some psychologists think even before you were born). That exposure was "24/7", except, perhaps, for periods of sleep or unconsciousness. Is your spelling perfect? Are your ideographs always accurate? Can you pronounce every word of your mother language withe the acCENT alWAYS on the correct sylLABle? Do you understand the subtle differences between every word of your language's available vocabulary? Could you write a story that would never need an editor?

I once had a French student (started as absolute beginner) who, after 16 months, 1 hour lesson per week, was angry that she had not yet attained native-speaker fluency. (The irony was, she had enough French to conduct the discussion entirely in French!)

I pointed out that, since native French speakers had the above-mentioned 24/7 lifelong learning . . . and still made mistakes, then either:

a) native francophones were verrrry slow learners, or

b) she (my French student) was a flaming genius.

The truth lies in between. Learning new languages should remind us just how much of a foundation we have in our mother languages, and should teach us . . .

Patience, grasshopper.

Enough said. I could go on for pages 'n pages.


excellent thank you indeed learning languages need a lot of practice and lot of entrances are useful< action, reflection, understanding the structure (i have a suspicion that people skip the introduction of grammar rules) and this method is a good practice of grammatical rules learned before, and they are not always introduced by DL in several languages. Thanks a lot for your contribution and good ideas about the lab.


I love the part where you elaborate on the physical nature of language. This makes sense since the motor skills and language skills are both in the same area of the brain — the frontal cortex.
Have a couple of lingots!


Cool post! Thanks for posting this! Man, you sure studied a lot of languages! :) Even though you didn't exactly achieve your goal of learning a new language very year, I think it is really neat you are relatively fluent in 4 languages! Thanks for introducing yourself!


Hi, Bozena. I've been in Krakow, I loved the city and people were so nice there. Wieliczka salt mine is so beautiful. I fell in love with Poland's nature and cities since childhood because of children's book series about Pawełek and Janeczka and their parents.
Do you have your own tips for language learning that effective and less time consuming at the same time?


Yes! I'll be posting short tips on Twitter. In addition, I hope to find time to write more forum and blog posts about this topic. Stay tuned!


I've been there, too and also loved the salt mine. I also went to see Auschwitz to learn about that part of history. As an American living in Lyon, France I've met several Polish people, even 2 from Krakow. I never read those books, but maybe in the future. I'm also glad to have read Bozena's post and want to thank him for his part in the "Duolingo Experiment". I've been focusing on the Dutch language, but occasionally I try Polish and several other languages.


Welcome! Duolingo is really lucky to have your expertise and passion. Its gonna be really fun!


Could you also see if something could be done for stories in Russian, Korean, Chinese, Arabic, and Japanese?


Welcome, Bozena. Glad to have your expertise at Duolingo. As a language enthusiast using Duolingo, I have used different methods together to learn Korean, in particular and other languages. My choices are based on overcoming several frustrations. 1. A desire to learn practical, everyday phrases and words: So I would work my way through level ones in order to reach where I could start learning numbers...may layers above Basics. 2. Writing down on paper any new words or phrases. 3. Reviewing tips and notes before starting a set. 4. Binge studying by pushing through several layers at level one. 5. Returning to much earlier completed lesson groups and doing the practice or strengthen. 6. Drawing simple pictures when none provided. 7. Biggest key of all: Repetition. 8. XP accumulation is a nice indicator but being able to write out the sentences in the Korean, Spanish, etc. or understanding podcasts or newspapers or a few words in a movie...those are the real reward for me.

I've studied languages with live teachers as an individual or in a group with Berlitz, Living Languages, and with a few university classes, and YouTube. No matter which method, if I did not study, UI did not learn.

Consider: One thing I did not realize at first is that there were tips and notes available on the web app. That's because I was studying only using the Duolingo mobile app.

Consider: Not all Duolingo language teams construct or maintain improvement in their language pair. Time, team cohesiveness, and number of contributors is important.

Consider: I've really enjoyed the podcasts in Spanish.

Consider: Pictures are helpful for nouns or short sentences.

Consider: In Korean and other non-Roman alphabet languages, I avoid the romanization altogether once I have succeeded those lessons. It is just too painful and destructive to recall the words by not only dealing with language to language but also the obstacles of romanization.

Consider: My native tongue is English. One year of high school Russian. One year of basic Japanese. Off and on over the years, I have studied Arabic, French, and Spanish. Most able in Spanish but not so much in the rest. Maybe I will come back to Russian and Hebrew in Duolingo. When I first tried those two they were just out of beta, and I could not make sense of the lessons. I'll try them another time.

Consider: the clubs now have different ways a person can try their skills. This version of the clubs in less than a year old. I hope Duolingo will give them a while to mature as more users become more skilled in their respective languages. One of my main dislikes is new things disappearing before they have a chance to prove worthy.

Consider: I opted in to the Plus subscription for several reasons: To have the downloaded lessons. To support Duolingo goal of open access regardless of subscription.

Consider: Encourage learners to write the language.


Quoting one of your points:

Consider: In Korean and other non-Roman alphabet languages, I avoid the romanization altogether once I have succeeded those lessons. It is just too painful and destructive to recall the words by not only dealing with language to language but also the obstacles of romanization.

I agree with you. My native tongue is English. Based on my small personal experience with learning Korean from English (outside of Duolingo), I agree that avoiding the romanization altogether becomes the best approach at some point in the learning process. The sooner the better. What I mean is the student should strive for this goal as soon as possible.

Quoting one of your points:

Consider: One thing I did not realize at first is that there were tips and notes available on the web app. That's because I was studying only using the Duolingo mobile app.

In regard to the Duolingo Spanish from English course, the Duolingo mobile app actually can be used to access the tips and notes that you have seen via the web app.

By the way, I have really enjoyed the Duolingo Stories in Spanish.

  • 1679

Awesome! Glad your here!!


Welcome!! That's awesome; I can't wait to hear more from you!! :D


Well welcome aboard! :)


Cześć! So happy that a Pole is in the Duo staff! I look forward to reading about your discoveries and how they are implemented in the Duolingo platform.

Now to shamelessly ask for some new Incubator additions! I would love to see some more language pairs over here that include Polish. I imagine German for Polish speakers and Polish for Germans would be quite popular for obvious economic reasons, but other pairs might also be popular. Polish is a major language in Europe, it is spoken by a lot of people, by European standards at least (45-50 million people).


I'd love to see that too, but unfortunately this is not entirely up to me ;) Maybe someday! Pozdrawiam!

[deactivated user]

    I really like duolingo. So much easier than my old curriculum, Rosetta Stone :)


    So you're very familiar with language acquisition... What do you think about if there were to be a new video function to watch short stories or videos using language acquisition like childrens' shows (Like dora where they teach you colors and small words) except with different levels, so once you acquire or already know basics, you can still use language acquisition technique for more complex concepts? I think a lot of people would be quick to try it out, and enjoy it as well! :) Much love, fellow duolinguist.


    This is definitely something we've discussed :)


    Interesting job! I'll follow you on Twitter. I work on the Economics of Education.


    I did a little work on the investments needed to provide education to every child in the world, about a billion of them, in many languages. Computers, electricity, networking, free software, free digital textbooks, school buildings and furniture...And also on the economic and social returns on those investments. @Mokurai on Twitter.


    This latest round of Duolingo Crowns concept is really very good. It makes you study harder and thus assimilate the language . I take it that you had a lot to do with it. Thank you for your passion for language education. See you on Twitter.


    Wow, you studied Somali. It's really awesome to see people wanting to learn it. Makes me feel more proud. :)


    Bozena, a pleasure to meet you. I'm so happy smart people, like yourself, are at the helm of this language learning rocket ship. Cheers!


    Heyo Bozena! I can't wait to hear more from you. You sound so friendly and interesting :)


    Thank you for letting so many participate in the grand experiment. I'm an older user. I had nearly persuaded myself to give up on language study because of my age. But something in the nature of Duolingo was capable of defeating my doubts, holding my interest and permitting me to see clear results. I'd love to get access to some of the data sets that show patterns of study and learning of which I'm unaware.


    Hi Bozena, It is great to hear from a real person. I feel as if that bird is the one in control, when I get the remarks on my tablet. I just want to say how much I appreciate Duolingo and everyone who makes it possible for us to learn languages free. I use Duolingo to take my mind off problems around me.Wishing you well.


    Hi Bozena, I would be interested in hearing your opinion on the Linkword method. I have been trying to learn French for years on and off. Maybe it's not for me.


    Awesome, sounds like this could get interesting :O


    On the professional side: I have a PhD in linguistics with primary expertise in language learning.

    So I presume you're familiar with criticism of the Grammar-Translation (GT) method, which duolingo uses (in cutesy form).

    Here are ten specific criticisms of the GT method:

    1. GT emphasizes the written language at the expense of the spoken. But being able to speak, and to understand the spoken language, are higher priorities than reading and writing for most learners

    2. GT uses a graded grammatical syllabus, and assumes that learners will progress towards mastery of the language by gradually accumulating an accurate command of each item in the syllabus. But most learners, and especially adults, want/need to start using the language straight away; they haven't got time to learn it first and only then start using it

    3. GT treats language as a stock of potential sentences: abstract grammatical frames with slots that can be filled by any vocabulary, in principle. But more recent views emphasize language as a set of tools and materials for constructing discourse, in which the sentence plays a subordinate role or in which – especially in spoken language – it isn't possible to identify sentences, as such. Recent views of language also highlight the importance of lexical 'chunks' of various kinds

    4. In GT, the prime importance attached to illustrating grammar can lead course-writers to include sentences which are unnatural, stilted, unlikely, remote from reality and so on. But recent decades have seen a rising expectation that language presented to learners should be 'authentic' and/or immediately usable for communicative purposes

    5. In GT, language is learned by conscious memorization of grammar rules and vocabulary. But nowadays memorization isn't highly regarded; internalization through exposure, experience and use is preferred. [This criticism not so relevant to duolingo]

    6. In GT, language is practised by manipulating grammar and vocabulary to write correct sentences with prescribed content – often through translation. But more recent approaches attach more value to oral practice and the expression of personal meanings

    7. In GT, the teacher and the learners speak mainly in the L1. But nowadays it's widely recommended that L2 use should be maximized

    8. In GT the teacher explains, translates, conducts practice, and corrects mistakes, and learners interact with the teacher, not with each other. But nowadays it's widely recommended that the teacher should play a less prominent role, guiding learners to make their own discoveries, eliciting language and explanations from them, encouraging them to co-operate, help and correct each other, and generally fostering learner independence

    9. In GT, translation is a basic technique. But nowadays it's widely believed that translation as a mental process interferes with learning – it's better to think in [the language being learned] – and that translation as an overt activity is only relevant for specialists who wish to work as translators or interpreters

    10. GT insists on accuracy from the start. But observation of language learning suggests that accuracy actually comes last – if ever!

    This list of criticisms (or objections) are not mine. They have been taken directly from this article:


    I've chosen to quote from that article here but criticism of the GT method is widespread in the academic literature as well as being rejected by most (all?) self-taught polyglots.

    As the head of Learning Science at the company I invite you to defend duolingo's use of the GT method and respond to these criticisms. Either here or perhaps in a separate post. Thank you.


    Yes, this criticism is completely valid. Ideally, we wouldn't use translation much (or at all) since the best way of learning a language is through immersion (although translation can still be useful for quick clarifications of the intended meaning). We're doing it mostly for practical reasons: on an app, unlike in a classroom, it's just really tough to convey the meaning of what's presented in the target language without relying on translation. But we are constantly discussing and experimenting with new features that give our users more exposure to the target language without translation. It's just not that easy to do with the constraints we are under. Stay tuned for improvements!


    I tried out the 'Duolingo Kids' app (available in Canada) a while ago and one of the things I loved about is that it didn't really on translation at all, but on illustrations and animations instead, so it was completely immersive. I would love to see this multimedia approach used in the regular duolingo app as well, as it seems like an app, with its ability to incorporate illustrations and animations, is actually really well suited to this approach. I also really enjoyed the duolingo chatbots for the same reason, they were immersive and interactive, and was dissappointed when they were removed. Right now stories are my favourite duolingo feature.


    My first exposure to German was a book called German Through Pictures, which used no English. I used to look for Roy Doty's Wordless Workshop whenever I could. He was a master.


    Is it iOS only app? No Android again?


    About animations, I wonder how much bandwidth they take and how many Duolingo users around the world (especially low-income ones) have phone plans that cost extra if they go above a certain about of data downloading/uploading per month...

    What if Duolingo had both a data-download-heavy version and a data-download-light version? :)


    Hi, I may not be in the Learning Science department of Duolingo, but I do have a degree in linguistics and I would like to address some of these criticisms. First, I am a widely opponent to the GT method. I have been saying for years that it is not an effective method, and therefore, even though Duolingo is extremely helpful to achieve an A2-level (maybe reaching B1), it is not sustainable in the long run. However, some of the points you quote are a bit unreasonable. Moreover, the limitations of Duolingo are intrinsic to its system. Using the newest methods of language learning AKA Communicative Language Teaching requires a (near-)native speaker with you all the time to speak with you, answer your question and correct your mistakes when you express yourself freely. It's unfortunately not reachable through DL, and the best way to use a similar method is to ask questions in the forums.

    1. That is entirely true. Listening and speaking are key components in language learning. That's why some exercises in DL are "Type what you hear" or "Say out loud what you read". These are clearly not enough, I agree.

    2. This criticism is not relative to the GT method, but to learning in general. As a language teacher, it's extremely frustrating when some learners are angry that we don't speak freely at the first course. Learning (whatever the method) requires time, and we should not rush it. We cannot learn and use a language at the same time in a matter of days. My point being made, I don't consider this point as a reasonable criticism of GT.

    3. That's also entirely true. However, this implies that we need to let the learners express themselves freely and then correct them. This operation cannot be automated and need to be done by a human. Thus, Duolingo cannot really achieve this.

    4. Although I do agree that "'authentic' and/or immediately usable for communicative purposes" sentences should be taught first, one has to take into account two things: 1) what is 'authentic' in a language/culture is not in another. Therefore, as a student, it can be difficult to separate what is really unnatural and what feels unnatural in one's point of view but is totally authentic. 2) unnatural sentences, if still meaningful and grammatically correct, are still sentences. What is really needed is to explain why a sentence is unnatural although it makes sense, and in that regard, I think Duolingo can be improved.

    5. As you said it yourself, "This criticism not so relevant to duolingo"

    6. Agree 100%. The only problem, as I mentioned, is the technical limitations. You should express your personal meaning and speak in your target language, but someone has to correct you and it can unfortunately not be a machine.

    7. Here again, agree 100%. I always taught that DL should use this approach, but apparently, a part of the community of learners don't agree with that.

    8. Here again, agree 100%. However, for the same reason as 6., it's not yet reachable.

    9. I tend to repeat myself now, but agree 100%. The problems are that 1) language learners still tend to believe that translation is a legit method, and don't like methods using entirely L2 and 2) this is not the premise DL chose, and even though I feel they are going towards that direction (thanks to Stories for example), I think they will still rely a lot on translation.

    10. That's a matter of preference. I am actually what is called a Grammar Nazi, and I insist on accuracy even in languages I'm not fluent in yet. I prefer speaking slowly but without mistakes, than fluently while making mistakes. It didn't prevent me to achieve fluency and a decent level in 5 languages. On the other hand, I completely understand that some people prefer the other way around since we don't all learn the same way and one method cannot be perfect for everyone.

    I don't defend DL's use of the GT method, but I understand that online language learning cannot help but use this method (or some other methods proved to be inefficient). I think it's up to us, the users, to seek for more through authentic materials and conversations with native-speakers.

    Of course, it doesn't mean that DL shouldn't be improved, but this improvement is slow.


    Thank you for this thoughtful (and comprehensive) reply Amine.

    Quoting your response to point 9:

    I tend to repeat myself now, but agree 100%. The problems are that 1) language learners still tend to believe that translation is a legit method, and don't like methods using entirely L2 and 2) this is not the premise DL chose, and even though I feel they are going towards that direction (thanks to Stories for example), I think they will still rely a lot on translation.

    I do not know if you know the history of duolingo but their original model was to self-fund free language teaching by having students crowd translate documents (which would earn revenue). So they would achieve two goals at the same time: 1) teaching languages and 2) translating pages on the web into other languages to improve access. This model is explained in their original 2 minute video:


    (Note how they promise "No ads, no hidden fees, no subscriptions").

    Now that model unfortunately didn't work out for the company and they dropped it (but I respect that they had a good go at trying to make it work first). This model also explains of course why they used the GT method and in that original context I think it made sense even with the implicit defects of that method.

    In the original model the artificial translation exercises in the language trees were offered in parallel with natural translation exercises of real documents with real world context (this was called 'Immersion"). But the natural translation side of things was dropped and this was actually the more interesting and engaging side of the translation approach. Many users were very unhappy to see it go and felt that the soul of duolingo had gone (the collaborative translation/learning environment).

    The Stories feature is good in that it moves away from the GT method, I agree. And I have a lot of respect for the person they originally hired to do it (Steve Ridout) and I think it has been very well implemented. But while it's nicely done I don't think it's anything so different from what has been done many times before - listening to/reading a story and answering questions to show comprehension.

    So that's where duolingo sits now. Their main system is an approach based on translation with the main reason for that approach now gone. And meanwhile they continue to run A/B tests to increase engagement but, from my perspective, they are only increasing addiction to a feel-good but ineffective language learning tool. And it's an addiction to a system that, again in my opinion, actually ends up retarding the users' advancement to fluency. Because it gets the users stuck in the translation mindset with the associated problems that we have discussed above.


    I know about Immersion, and I had the "chance" to use it. Actually, I didn't really like it for many reasons that I will not list here, but that's my own opinion on the matter.

    However, I never knew that Immersion was the main goal of Duolingo and I always thought of it as a by-product of the GT method used here.

    You also say

    The Stories feature is good... But while it's nicely done I don't think it's anything so different from what has been done many times before - listening to/reading a story and answering questions to show comprehension.

    and I cannot help but wonder "what else do you want to implement?". I don't know how we can improve our listening and reading abilities differently than listening/reading something (not necessarily a story) and answering (more and more complex) questions to show comprehension.

    I understand the frustration due to the lack of opportunity to express oneself in a target language, but as I said before, that requires someone with a (near-)native level to correct and guide learners. I don't think it can nowadays be done automatically, contrarily to the GT method used in DL. Maybe there are ways, and I would be glad to hear about it, but for now on, I am sceptical. Moreover, one doesn't need anyone to correct and guide learners, but a (near-)native speaker with a good understanding of the grammar, and this is not easy to find in every languages.

    So that's where duolingo sits now. Their main system is an approach based on translation with the main reason for that approach now gone. And meanwhile they continue to run A/B tests to increase engagement but, from my perspective, they are only increasing addiction to a feel-good but ineffective language learning tool. And it's an addiction to a system that, again in my opinion, actually ends up retarding the users' advancement to fluency. Because it gets the users stuck in the translation mindset with the associated problems that we have discussed above.

    I completely get the frustration that you feel, and yes, totally, A/B tests and every data-driven decisions made by DL are only here to increase addiction from learners. If you study data analytics, you'll understand pretty quickly that its main purpose is to gain and retain the maximum amount of user, and not to improve their experiences nor the product's quality.

    Nevertheless, I personally value DL for multiple other reasons. It's the only free gamified language-learning tool that offers that many language courses at such a level. It's only an A2 level, but it's still something to be able to pass an exam (Listening/Reading/Speaking/Writing) only thanks to a free app/website. And it also taught me some new words I never heard before, some grammatical construction or idioms I didn't know, and most importantly, the way to translate some sentences from one language to another when I already speak both.


    I would be delighted if we could get people who have finished a tree to provide missing Wikipedia pages or other such useful content.

    Another possibility is to go over to Quora, which supports several languages, and answer questions that you know something about, and can manage at the language level you know. We could get together on questions that are a little harder.

    I finished the earlier Spanish tree and then become a Distinguished Writer (Escritor Destacado) in Spanish Quora before coming back here and doing a later version of the Spanish tree.


    Also here, I love your ideas.

    Duolingo fully supports linking and encouraging engagement in worthwhile resources that are especially available for people to engage in and encourage people to assist in developing that are focused on providing them at no costs for the learners.

    Wiktionary is one such valuable resource.

    Also there is Duolingo Wiki , which also provides links to great resources.

    For flash card programs, there is Duolingo's Tinycards
    And also Anki and Memrise

    Also there are even more Duolingo resources in development: You can check out some of them here: Labs and now also Stories

    Especially of interest is Duolingo Events , that I highly recommend that people get involved in, in both attending events, and also hosting them. They are FANTASTIC resource and opportunity for language learning.

    If people are aware of other resources they would like to recommend, I would be love to see you sharing what you find beneficial to language learning. i.e. to share them here, for example.


    I am sure that we can improve the efficiency of DL or any other language learning Web site.

    The microwave method from Earl Stevick and others can be used almost entirely in the target language. Instead of translations and transcriptions of mostly fixed sentences, it focuses on the necessary grammatical transforms, and applies them to different sentences, proceeding from simple word substitutions to the full transformational repertoire of the language.

    I eat. I am eating. I write. I am writing.

    I have a book. I have two books. I have a mouse. I have two mice.

    I have a book. I don't have a book. I had a pencil. I didn't have a pencil.

    Then we turn that into dialog, combining a variety of transforms.

    Are you ready? Not quite. I'll be ready in a minute. Are you done? Not quite. I'll be done in a minute. Are you dressed? Not quite. I'll be dressed in a minute.

    I can make microwave lessons from DL sentences and tips. We could help a lot more people who don't know how.


    I LOVE your idea here Mokurai

    and would like to encourage you to do so, and create posts here in Duolingo, in the relevant forums.

    I would also like to assist in highlighting a resource such as this. And agree that it would help a lot more people.


    hope they'll ignore the whole first half, it's already full of phrasebook apps.


    ... the whole first half of what? You lost me.

    The post by necronudist85 is nested under the post by oskalingo. Is this the person that necronudist85 is replying to? Who is necronudist85 really replying to? Or am I replying to a prank?


    I am half Polish, but don't speak Polish. You can see my best languages. I am looking forward to Yiddish becoming available.

    I had a teacher at Weequahic High School in Newark NJ who was apparently learning a new language every year or two. He started an extra-curricular class in Swahili which was written up in a linguistics journal as the first at high-school level in the US.

    I was a Peace Corps English teacher, with some training in the Earl Stevick microwave method used for State Department and Peace Corps language training, starting with his Swahili textbook. Is that still of interest? I have the Spanish Microwave Textbook, and was using it before I started with Duolingo.

    Stevick, Earl W., John Indakwa et al. An Active Introduction to Swahili: General Conversation. Supt. of Documents, Washington, D.C., 1966.

    Stevick, Earl W. 'UHF and microwaves in transmitting language skills,' in Najam, E. (ed.) Language Learning: the Individual and the process, supplement to International Journal of American Linguistics, 1966.

    Stevick, Earl W. A Workbook in Language Teaching. Abingdon Press, Nashvile, Tenn., 1963.

    Stevick, Earl W. '"Technemes" and the rhythm of class activity,' Language Learning. 1959.

    Stevick, Earl W. 'Can we afford to be relevant?' (Foreign Language Annals).

    I have helped One Laptop Per Child and Sugar Labs with their localization projects, recruiting people for Kreyòl Ayisyen and Cambodian when OLPC was preparing to give Haiti, Kampuchea, and four other countries a thousand of our little green and white laptops each.

    I have training in Computer Science, and have managed Free Software development for education. I am also an experienced technical writer.

    I would like to find a way to volunteer with Duolingo, but none of my overtures has gotten a response.For example, your Words pages need a lot of work in both content and database functions.


    You seems to have a good way on learning the language you don't use in daily life, would you share a bit how to learn a new language without being in the environment ? It has always been easier for me to pick up a new language from daily life but not doing so well from study and books.

    Thank you!


    Thanks for your post! Very interesting bio! Learning online is a great benefit of the "computer age".


    I was very interested in Oskalingo's ten points criticizing traditional language instruction. As a 67-year-old who put off learning a second language until ten months ago I always thought immersion made more sense. But I lived 2,000 miles away from anywhere that Spanish or French was spoken. Duolingo, despite any imperfections was a revelation. But I've experimented and found additional avenues for learning French. One I have always wished I could have access to doesn't yet seem to exist in French.

    In the early days of personal computers, thirty years ago, I ran across a simple ap that mimicked a psychologist interviewing a patient. The interviewer was a very dumb computer but allowed for conversation, at least a typed conversation. I've always assumed it must have been available in French too.

    I'd love to see some software to create similar conversations in other languages. A computer would be much more relaxed about the stumbling of a new language learner and the learner would get a fatigue resistant companion. I saw something about an ap called Replika that, if configured to speak something other than English, could fit the bill.

    Has Duo investigated that possibility? It should work on this soon before hand-held translators get so good that it will discourage the hard work of learning to speak a new language.


    hello! Thank you! you have done a very good way since your youth in Poland! I wish you to arrive to improve Duolingo. dziekuje!


    cool. i love duolingo


    Hello, and THANK YOU for bringing your expertise to Duolingo. I'm finding Duolingo Irish to be very enjoyable. I wish there were a section for just truly useful Irish, as well as the 'here's how Irish works' - I love that you're teaching how to put together any sentence at all (e.g., 'the spider has water' or 'I am a man' - neither of which will I ever imagine saying). I am not a linguist, but have a computer programming background, and am helped by knowing why something (e.g., a language) works as it does. So, as a 'plus' user I'd appreciate a little section on how to recognize professions, how to recognize verbs... it would help me eventually break into the reading of books more easily.


    Hi Bozena,

    I am new to Duolingo only starting around 3 weeks ago and I feel it does a great job in teaching languages and obviously concentrates on the languages most spoken worldwide, with a few exceptions (e.g. Irish and Klingon!) and I am just learning Spanish right now. I have seen on older discussions that the daily goal settings had a much higher xp than currently and I have my daily goal set to the maximum of 50 xp but I feel if the maximum were around 250 xp per day it could be referred to as insane! For me 50 xp does not take too long to obtain I think a higher xp would give people more incentive to work hard to learning the language.

    Thanks, Josh


    behaviorism is not always a good incentive for learning and DL is using plenty of it. it is just like you were a kid and for every good score you got a picture but i am not interested now in this type of rewards. I just want to learn the languages well and if i succeed this will be the reward not the xp s or whatever.


    Hi Bozena! Krakow is indeed beautiful, though I love Wroclaw a little more. My suggestion is to bring back spaced repetition/skills decay.


    Keep up the good work! So far I've never seen a more convenient and well thought application! Android and browser based application works awesome together!


    compared with other applications DL is much better


    Hello, Bozena! Thank you for such a personal introduction. I'm a little late to the party, but wanted to share a thought.

    Many language learners (including myself) use Duolingo as one of the tools in their learning, and the range of other resources and methods people choose is very diverse. Is that something Duolingo keeps in mind or incorporates into its learning design anyhow?

    On a slightly different note, I am personally very curious how prior language knowledge affects new language acquisition. E.g. what effect does the language through which you learn another have? Does learning a new language through multiple languages you already speak help or hurt? How do you choose a language through which to learn a particular new language? If there is something brewing within Duolingo relevant to these questions, or if there are some relevant resources that spring to mind, it would be really nice to hear from you. Truly a fascinating field of expertise.

    Ok, this wasn't so much sharing a thought as pouring out a bunch of questions. Hope you do find a second to reply, and good luck with your work either way)


    Indeed the knowledge of a foreign language related to your own language is very helpful to learn another language (advanced organizer) it is more difficult for an Italian to learn Spanish but not if the Italian wants to learn Arabic or Chinese!


    there are lots of scientific discussions about learning via internet/itc, and it can be useful as an additional element in learning but not as a replacement for learning e.g. in groups with a tutor. I also see pictures in DL that do not contribute to learning, it has not any connection with the text. Also many texts are written and translated by non professionals and some of them do very difficult things like to learn Chinese, Hindi Arabic Japanese to an American and I do not think there is a lot of scientific reports about how to learn that to foreign people who do not have any connections with it.

    Learn a language in just 5 minutes a day. For free.