Weekly Incubator Update: Tracking Progress from April 26th to May 3rd

The first incubator summit in Berlin: 30-Apr-2015
The summit was a great success. It was wonderful to meet all the fellow incubator contributors and the awesome Duo team at the summit. Below are a few pictures.

The Inauguration of the summit

The session is in progress.

How many ways one can say "I love you" in Japanese?

Ethiopian dinner

Ride back from the dinner

BTW, I will be doing the first of the two AMAs on coming Saturday. Please feel free to ask your questions there. Link to follow.

Slower progress
The progress was a shade slower this week. But let's review the positives:
- New course "Spanish for Russian" has been added to the incubator
- We have a new weekly leader - the Greek team
- The Yiddish team has kicked off their journey :)

Also, please note that there has been a slight reshuffle among the top three spots. It shouldn't be a cause of concern though. There's not much difference between 99% & 100% and please remember 99.9% is reported at 99 on this report because of the round-down.

Also you will see a huge boost that the new course - Spanish for Russian - has got due to the UI translations that are already in the place from the "xyz" from Russian course(s).

Lastly, despite the delay in today's post, the data still reflects one week's progress in order to keep the week to week comparison meaningful. In other words, the numbers reported here might not match the latest ones on the Duoinc page because they were taken from yesterday.

Thought for the week
Successful Berlin Incubator Summit!

PHASE 1 Progress:

Ukrainian for English - 100% | 100% | 100% | 100% (+0)

Norwegian for English - 95% | 100% | 100% | 99% (-1)

Esperanto for English - 96% | 98% | 98% | 99% (+1)

English for Thai  - 90% | 91% | 91% | 90% (-1)

Russian for English - 74% | 75% | 77% | 81% (+4)

Hungarian for English - 79% | 80% | 80% | 80% (+0) ^

German for Italian - 69% | 70% | 71% | 71% (+0)

German for Turkish - 70% | 71% | 71% | 71% (+0)

Spanish for German - 64% | 65% | 66% | 66% (+0)

French for Italian - 58% | 60% | 60% | 61% (+1)

German for Portuguese - 49% | 55% | 59% | 60% (+1)

Spanish for Italian - 57% | 59% | 60% | 60% (+0)

German for French - 54% | 54% | 57% | 57% (+0)

Spanish for Chinese - 44% | 44% | 48% | 48% (+0)

Spanish for Russian - (New) 39% (+1)

Swedish for Russian - 39% | 39% | 39% | 39% (+0)

French for Chinese - 38% | 38% | 38% | 38% (+0)

Romanian for English - 30% | 30% | 31% | 33% (+2)

Polish for English - 27% | 29% | 31% | 32% (+1)

Vietnamese for English - 29% | 29% | 31% | 31% (+0)

Greek for English - 9% | 9% | 10% | 18% (+8) *

Hebrew for English - 7% | 7% | 10% | 11% (+1)

Klingon for English - 0% | 2% | 4% | 7% (+3)

Yiddish for English - 0% | 0% | 0% | 1% (+1)

Mean - 1.32% | 1.17% | 1.17% | 0.92% (-0.25)
Median - 1% | 1% | 1% | 1% (+0)
Mode - 0% | 0% | 0% | 0% (+0)

* This week's Leader Extraordinaire!
^ The Hungarian team's progress is as per their own calculation.

Progress Visualized:

Errata: Klingon progress should read 7% and the Mean as 0.92%

Here's what the contributing teams have said during the last week:

DespoinaTi from Team Greek said 6 days ago:

Welcome to the Greek course for English speakers!

Γεια σας!

First of all we are really sorry for being so quiet, but we are working really hard for this course! We are trying to do the best we can, and I believe the outcome is going to be awesome!

Secondly, I would like to thank the rest of the contributors/moderators for letting me be a part of this! So far it has been an amazing experience!

Now about the course: We have completed 10% of it and getting close to the second checkpoint. 222/1967 of all words and 57/100 of all word images are finished.

In case you do not hear from us any time soon, don't worry! Even if we do not share anything, we are still making progress with the course!

Sincerely, Η Ελληνική Ομάδα The Greek Team

amuzulo from Team Esperanto said 4 days ago:

Esperanto course launch estimate and more news!

When will the Esperanto course beta launch? Our best estimate at the moment is the beginning of June. Please stop asking us... it'll launch when it's ready. ;)

Some have asked for more details about the Text-to-Speech (TTS). I can confirm that we've already started recording and everyone on the team is very happy with his accent and we're all excited that we'll have natural recordings for our Esperanto course!

Another question I frequently get is whether Duolingo will later have more Esperanto courses from other languages (like learning Esperanto as a Spanish or French speaker). I talked with the community team last night and they said it could be a possibility. We just need to wait until the Esperanto course reaches phase 3 (stable) and then we can start discussing if we want to add more Esperanto courses and if so, from which language(s).

More exciting news is that I'm organizing the Duolingo Incubator Summit taking place in Berlin tomorrow where many different course volunteers will be meeting together... I can't wait! If you want up-to-date news on that, please follow @DuoIncubator on Twitter.

In any case, thank you for your patience ... or your excitement, yet lack of patience! ;) When the Esperanto course is in beta, you can read about it here first. :)

asya.000 from Team Yiddish wrote 3 days ago:


!שלום־עליכם Sholem-aleykhem! We're very excited to be working on the Yiddish course, and we have been so inspired by all of your comments and expressions of support!

Here are some of the things that have been keeping us busy:

We've been...

\1. ... working together to design a flag and a monument for Yiddish. Since Yiddish is not directly associated with a particular country, this has been a bit tricky, but we hope to have one soon!

\2. ... figuring out how to introduce the alphabet in a way that wouldn't be overwhelming to beginners.

\3. ... fine-tuning a course description.

\4. ... planning out an overall course structure, including establishing the order of introducing various vocab and grammar elements.

\5. ... learning how the incubator works and filling in the first lesson!

!זײַט געזונט Zayt gezunt! Be well!

david.aparicio from Team FR<-PT wrote 3 days ago:

Objectives of the phase 2 (@myra)

Source: #incubatorsummit Berlin 2015

Kippis from Team Japanese wrote 2 days ago:

日本語→英語コース 4月の進捗 / Progress of our course in April

[English following the Japanese text]

Guten Tag(こんにちは)、ドイツのベルリンからお送りします。昨日4月30日にここで「Duolingo インキュベーター・サミット(Duolingo Incubator Summit)」が開催されました。インキュベーター はボランティアがコースの作成や管理をするためのシステムであり、本コースも2013年10月にボランティアスタッフの手で作成が始まり、2014年3月にβ(ベータ)版として公開されています。インキュベーターで様々な言語を教えるボランティアが世界中からこのサミットに集い、Duolingo 本社のスタッフと共に様々なノウハウを交換しました。私も今回の経験を活かして本コースの質をさらに高めるよう努力したいと思います。



Guten Tag, greetings from Berlin, Germany. The Duolingo Incubator Summit was held yesterday (30th April) here. The Incubator is a platform for volunteers to contribute in creating and managing language courses; contributors began creating the EN <- JP course in October 2013 and it launched as beta in March 2014. Contributors for various languages gathered from all over the world and exchanged ideas together with the staffs from Duolingo. I would like to feedback my experience to improve this course.

This month we dealt with thousands of reports and fixed more than 2,000 sentences. Currently, we are concentrating on lessons at the top of the tree, which might be irritating for those of you in the latter half of our course. We are doing it one by one; we beg for your patience and please do not write your reports on the discussion board.

By the way, 30th of April saw another milestone in the EN <- JP incubator course.



The number of reports per 100 active users has, for the first time in the history of our course, fallen below 3. If our course becomes stable enough to maintain this level for some amount of time, we will graduate the beta, to hatch in the true sense.

Of course, this is only the end of the beginning. Our work in the Incubator shall continue.

DvirBartov frp, Team Hebrew wrote yesterday:

Slow progress...

שלום לכולם!

In the last week our progress was very slow(only 1 percent more), mainly because me and Mae were very busy..

Also, we wrote the Tips and Notes of some new skills (that's not included in the course progress) and minimized the contribution applications - so new contributors are going to join us soon!

For the people who are willing to contribute - we don't add more contributors to our team yet, simply because we haven't reached 20% yet (please don't ask us about it - when we reach 20% we'll add new contributors). If you haven't contributed yet and you want to - click here.

Furthermore, we are finally(sof sof!) on the course list in the website! (isn't it cool? :D).

That's our update for now.

Stay tuned!!

Team Hebrew <3

luke51991 from Team Norwegian wrote 17 hours ago:

Short Update

(A longer update will follow shortly) Short update, since I missed yesterday, and I know some of you might be anxious to know what's up. (I'm still in Berlin, at the Polyglot Conference). We're reviewing sentences, correcting errors, adding new words. We would like to be finished by 18. mai (one day after Norwegian Constitution Day), but we're not promising anything. Again, a longer update is to follow.

Med vennlig hilsen,

det norske laget

Selcen_Ozturk from Team Turkish wrote 9 hours ago:

We met at the incubator summit to discuss with each other and the other contributors and staff to discuss how we can improve our course. Greetings from Berlin! Alex and Selcen

TL;DR The first ever Duolingo Incubator Summit took place on Thursday, 30th April, 2015. Contributors gathered from around the world to meet and share ideas! It was simply awesome to meet the real-life folks behind these digital personas :)

The next update is expected on Sunday, 10th of May, 3:00 pm UTC, back to the regular schedule :)

Previous Update 19-Apr to 26-Apr

Last updates from the recent months: Apr, Mar, Feb, Jan.

May 4, 2015


Can't help feeling it was risky to have so many course contributors travelling in the same bus.

Only half joking...

Anyway, I'm sure it was great for you guys to meet up and be able to talk face to face. Thanks to the people who made this happen.

May 4, 2015

honestly I thought of this when we were on the bus but of course didn't talk about it :D

May 4, 2015

Not that funny if you consider how these things do happen... A few years ago a plane with many members of the Polish government crashed. Also, a plane with Russian hockey team "Lokomotiv" crashed in 2011, the whole team dying on a single day.

May 4, 2015

For Russian and Polish folks it isn't that funny :/

In this ariplane crash near Smoleńsk died:

  • President

  • Former President

  • first lady

  • former deputy prime minister (*)

  • former minister of defence

  • chiefs of staff (army, airborne and navy)

  • president of national bank

  • deputy foreign minister

  • head of national security office

  • deputy parliament speakers

  • more than dozen members of polish parliament

  • members of Katyń committee

and many more :(

May 5, 2015

So excited about Russian from English! Go team Russia!

May 4, 2015

Can someone who speaks Japanese or who was at the summit explain the logic behind the numerous ways to say 'I love you' in Japanese? Just curious.

May 4, 2015

First of all, the calculation is based on actual reports from users of the course, i.e. only the expressions suggested by users (not only by one, but by a considerable number) as a translation for "I" "love" and "you" have been used here. It's not a number made up for exaggeration (I heard people saying so) but a realistic number to cover most of the user reports. I'd also like to add that not all these forms might be actually used, but users trying to translate the phrase grammatically correct might come up with any of them.

The typical word order in Japanese is Subject - Object - Verb.

There are infinitely many ways to say "I", but the 6 major variants are わたし watashi - Most common for women, polite form for men)、わたくし watakushi - Formal variant of watashi. Uses same Chinese character but has to be a separate entry due to limits in Duolingo's system)、あたし(atashi - informal form of watashi for women, considered rather unrefined)、(ぼく boku - For men, usually young boys)、(おれ ore - informal and sometimes impolite form for men) and 自分(じぶん jibun - literally "oneself", but frequently used as first person pronoun). To Indicate that they are the subject of a sentence, the particle が (ga) or は (wa) has to be added. In a casual talk this can be omitted. However the most likely case is that the subject itself would be omitted! Thus we obtain 6×3+1.

For "you", 4 most suggested words are あなた(anata - most common translation for "you", used to address people who are not superior, and not much of an acquaintance. Contradictorily it is also used among lovers or married couples)、あんた(anta - very informal and familiar form of anata)、(kimi - to address one's inferior or those in the same level, maintaining distance)、お前(omae - quite rude, but sometimes used to show affection). Note that none of them are used to address one's superior - we would either avoid using a second person pronoun or address them with their social status.

"You" can be plural. Each of them becomes plural by adding (たち tachi)or (ra)at the end(君たち、お前ら、etc). あなた and あんた has another form, which is to add がた gata)as in あなた方. However, the Duolingo grading system does not read the Chinese character correctly so we have to add both the hiragana and the Chinese character, thus this counts as two. 5 variants including the singular for あなた and あんた, 3 for 君 and お前. 2×5+2×3. Particles attached to the object varies depending on the verb, so let us count them later.

There are three types of expression for the verb "love". The most common expression, especially when you want to confess your love to someone, is to use the adjectival noun 好き(すき suki). This actually means "like", but if you use it to a person it means something more than being a friend. However, since we teach people that "love" is stronger than "like", it is quite normal to add prefixes to intensify 好き. There are infinitely many ways to do so, but the 4 styles suggested by users are 大好き(だいすき daisuki - "greatly like")、すごく好き(sugoku suki - "very - like")、とても好き(totemo suki - ditto)、めっちゃ好き(meccha suki - ditto, but very informal). So we have 5 variations for 好き. To each of them could be added the particles だ(da - neutral)、です(desu - polite)、である(dearu - formal)、or no particle (informal). Furthermore you can insert particles なの(nano) or なん(nan)to make it assertive. The possible combinations for 好き are, thus, 1.好きだ sukida、2. 好きです sukidesu、3. 好きである sukidearu、4.好き suki、5. 好きなのだ sukinanoda、6.好きなのです sukinanodesu、7.好きなのである sukinanodearu、8.好きなの sukinano、9. 好きなんだ sukinanda、10. 好きなんです sukinandesu (I recognized that I miscounted this as 9 when I prepared for the presentation. It should be 10. So the number will increase...). For all of these forms you can add the particle (yo) at the end, which adds the nuance of appealing. Thus 5×10×2 for this form. Now to connect it with the object "you". Possible particles to be added to the second person pronouns are (ga), (wa) or (o - a bit rare but permissible). We can also insert の事(のこと no koto - "about (you)")between the pronoun and the particle. Furthermore we can add の事 and omit the particle. Thus variations are (using あなた as example) 1. あなたが好き(anata ga suki)、2.あなたは好き(anata wa suki)、3.あなたを好き(anata o suki)、4.あなたの事が好き(anata no koto ga suki)、5.あなたの事は好き(anata no koto wa suki)、6.あなたの事を好き(anata no koto o suki)、7.あなたの事好き(anata no koto suki). Thus the combination of the 好き form with the object's particle (not including the variation of the object itself) is 5×10×2×7.

The orthodox translation for "love" is the verb 愛する(あいする aisuru). It is usually used among lovers or couples, though the typical form is 愛してる(あいしてる aishiteru), a kind of progressive form, "I'm loving you right now". The variations are : 1. 愛する(aisuru - neutral)、2.愛します(aishimasu - polite)、3.愛している(aishiteiru - progressive neutral)、4.愛しています(aishiteimasu - progressive polite)、5.愛してる(aishiteru - い i omitted)、6.愛してます(aishitemasu)、7.愛す(aisu - archaic, but some users insist on this form!). All of them come with or without よ. 7×2. Particles for the second person pronoun could be or , and の事 can be added. Possible combinations are 1.あなたは愛する(anata wa aisuru)、2.あなたを愛する(anata o aisuru)、3.あなたのことは愛する(anata no koto wa aisuru)、4.あなたのことを愛する(anata no koto o aisuru). That's what makes the formula 7×2×2×2.

Another way to express "love" which can be found in any dictionary is the verb 惚れる(ほれる horeru). This actually means to "fall in love", but it's absolutely natural to use it to express your feeling. Variations are similar to 愛する, without the archaic form : 1.惚れる horeru、2.惚れます horemasu、3.惚れている horeteiru、4.惚れています horeteimasu、5.惚れてる horeteru、6.惚れてます horetemasu. With or without よ. 6×2. The good thing with this form is that the only possible particle added to the object would be (ni), so the number 6×2 is the same when you have the second person pronoun.

To put them together :

  • 6×3+1 for the subject, including the case when it is omitted
  • 2×5+2×3 for the object, not including the particle
  • 5×10×2×7 + 7×2×2×2 + 6×2 for the verb when object exists,
  • 5×10×2 + 7×2 + 6×2 for verb when object omitted

(6×3+1)× { (2×5+2×3) × (5×10×2×7+7×2×2×2+6×2) + 5×10×2+7×2+6×2} = 248,026 ways to say "I love you" in Japanese.

The number has increased due to my mistake in the presentation, but I can assure that this calculation is correct, because I'm the one who presented this in the Incubator summit!

And notice, that I have done the calculation for only one type of word order (subject-object-verb); other orders are possible, but it becomes complicated and I stopped calculating them. I have also ignored many, many, many other particles that could be added to change the sense of the sentence. I think the number could rise up to a million... if you take into consideration any kind of grammatically correct phrase.

Corrected part of the text, thanks to Shady_arc

May 4, 2015

In other words this is one of the languages that doesn't really work well with Duolingo's translation method of teaching. Much time will be wasted by contributors dealing with adding wild and unused ways of translating the exact same sentence simply because it is a "correct" way of translating it.

From this one could surmise that it would be far better if either Duolingo devised a different way of teaching Japanese without translation, or greatly reduce the translation exercises from one's mother tongue into Japanese. Either that or Duolingo needs to find a way to automatically deal with the many variations possible in such languages (or at least the reports).

May 4, 2015

Either that or Duolingo needs to find a way to automatically deal with the many variations possible in such languages (or at least the reports).

All and only the variations mentioned would be easily parseable using a context free grammar (the phrase structure grammar in linguistics is more or less the same thing) in... oooo, maybe about 10 lines. Hence my previous surprise at discovering that's not what Duolingo does.

May 4, 2015

Sure, but that still leaves me mystified as to why it wasn't used first place

My guess is that they were planning to get the incubator out as quickly as possible and so they took a lot of shortcuts. They also had to deal with legacy languages (e.g. french) and all the patches and configurations that came with it. So it wasn't likely a clean implementation dealing only with adding languages.

For example, if I had to design a software like that, I'd first cater to the lowest common denominator, the hardest languages to integrate into the system, the way to teach scripts, and then deal with simpler languages. The fact that duolingo still doesn't have a way to teach foreign scripts clearly shows that they took the opposite path, and all the drawbacks that come with it.

May 4, 2015

We briefly discussed this at the summit, since many languages have a relatively free word order.

The problems are that 1. Even if the word order is relatively free, as eg. in Japanese or German, not all combinations are possible. You cannot just allow all permutations of words in a sentence.

And 2. One of the strong points of the incubator is that it's easy to use for everyone, previous knowledge in programming/regular expressions is not required. Many people seemed to agree that we shouldn't make notations too complicated.

So~ this issue is not as easy as it seems on first glance. :/

May 7, 2015

You cannot just allow all permutations of words in a sentence.

A CFG wouldn't necessarily allow all permutations of a sentence, i.e. it could, but only if you wrote it to. Maybe a simple example would help. Say I wanted to accept [a/the/λ] boy catches [a/the/λ] ball using a CFG. (Here I'm using the standard mathematical symbol λ for the empty string). This is a CFG that would accept all of these options and only these options.

sentence: subject verb object
subject: article 'boy'
object: article 'ball'
verb: 'catches'
article: 'a' | 'the' | λ

If you wanted to accept both [a/the/λ] boy catches [a/the/λ] ball and [a/the/λ] ball is caught by [a/the/λ] boy and nothing else you could do it with

sentence: subject verb object
| object passive_verb subject
subject: article 'boy'
object: article 'ball'
verb: 'catches'
passive_verb: 'is caught by'
article: 'a' | 'the' | λ

So hopefully you can see you could get some quite powerful behaviour. But I do accept your second point, bit of a rough spot really. :/

P.S. I see you know of regular expressions. So for your edification: every set of strings recognisable by a (single) regular expression is recognisable by a CFG but not vice versa. So a CFG is a slightly more powerful grammar. It's also not quite as bad as programming but it's certainly getting up there. :P

Edit 2: I felt like a bit of a challenge so I tried to write a CFG that would accept all and only the valid Japanese sentences described by Kippis. This is what I came up with. I'm not guaranteeing that it's the best way to do it, nor that it doesn't have any errors; but it does give you an idea of how it might look (it looks a lot nicer with syntax highlighting and appropriate alignment than with Duo's verbatim environment, sorry but not something I can fix :/)

sentence: (subject | λ ) (adj_ending | vb_ending | vb_2_ending)
subject: i (subject_particle | λ)
i: 私 | わたし | わたくし | あたし | 僕 | ぼく | 俺 | おれ | 自分 | じぶん
subject_particle: が | は
you: あなた | 君 | きみ | お前 | おまえ
you_special_case: あんた (you_opt_pl | 方 | がた)
you_opt_pl: 達 | たち | ら | λ
adj_ending: (you you_opt_pl | you_special_case) (の事 | のこと | λ ) (が | は | を | λ ) adj_to_love
adj_to_love: (adj_to_love_intensifier | λ) (好き | すき) (adj_assertive adj_particle | λ ) ( よ | λ )
adj_to_love_intensifier: 大 | だい | すごく | とても | めっちゃ
adj_particle: だ | です | である
adj_assertive: なの | なん | λ
vb_ending: (you you_op_pl | you_special_case) ( は | を | λ ) ( の事 | のこと | λ ) vb_to_love ( よ | λ )
vb_to_love: 愛する | あいする | 愛します | あいします | 愛している | あいしている | 愛しています | あいしています | 愛してる | あいしてる | 愛してます | あいしてます | 愛す | あいす
vb_2_ending: (you you_op_pl | you_special_case) に vb_to_love_2 ( よ | λ )
vb_to_love_2: 惚れる | ほれる | 愛する | 惚れます | ほれます | 惚れている | ほれている | 惚れています | ほれています | 惚れてる | ほれてる | 惚れてます | ほれてます

16 lines, my 10 line guess wasn't too bad. ^_^

May 7, 2015

Chances are that they are aware of the potential for using it, considering that two staff software engineers are part of the Japanese incubator team. It is simply likely that problem is either low priority or there are bigger issues they are planning to tackle.

May 4, 2015

Sure, but that still leaves me mystified as to why it wasn't used first place. :P I mean it would've been fine, I think, for any languages on the site. There could well be reasons I'm not aware of that caused them to go with the list-and-match from all sentences but I sure don't know what they would be!

The only one I can think of is that if they always planned to have crowd sourced languages they might have reasoned that learning CFGs provided too shallow a learning curve to contributers (IMO they aren't all that complicated once you see them at work—although I did retrain myself to use DSK after working as a touch typist in QWERTY so my standard of what is OK to expect may be a little off :P).

May 4, 2015

This is what I always think whenever this topic gets raised - what's the big deal? Yes, a little bit of multiplication can get you some big numbers but we are dealing with grammar here and so of course sets of possible answers can be formulised.

My understanding of the incubator tool is that it already has what is called "short form" in that you can bracket alternative translations matching a single word e.g.

French "Il mange" would be matched to the English translation [It/he] [eats/is eating]. So with this simple example you've just given one translation formula that matches to 4 possible answers.

Having said that, I realise that I've "cheated" in using an example from two closely related languages (French and English). And I do take Dessamator's point that duolingo's "translation method" becomes somewhat more problematic for language pairs that are more dissimilar. But Turkish <> English seems to be working pretty well and those two languages are pretty different.

May 4, 2015

From what the Turkish team stated, they had a lot of difficulties getting the course ready, and had to completely discard the original template. The Hungarian (or polish) team had similar difficulties from what they said.

So I'd guess that it didn't really work that well for them either.

May 4, 2015

Thanks for explaining it one more time :)

It was great meeting you at the summit. I think we can start dreaming about a Sanskrit for English course on Duolingo!

May 4, 2015

My pleasure, not at all difficult compared to your incredible weekly work!

Someday in the future "Sanskrit for English speakers (or maybe even Hindi and Japanese speakers)" would be in the list of your weekly reports!

May 4, 2015

"आपके मुँह में घी-शक्कर" - May what you say come true.
If it entered in incubator, I will gladly spend time in making it.

May 5, 2015

I'd love to take a Sanskrit course on Duolingo! :D

May 7, 2015

Love, love, love your comment. Love Japanese. I will definitely master this language, (But it will take years :( )

私は日本語が 好きです。。。いいえ、大好きです! I hope I didn't messed anything while writing. I'm still a learner. ;) Please, be free to correct.

May 5, 2015

Thanks for your comment too. Japanese is even difficult for us natives to master, but it shouldn't be difficult to learn it up to a practical level.

Let me try to be super-strict. "のこと" in this case is used to indicate that a certain person is the object of some emotion. Thus, it's natural not to use it in case of non-creatures or abstract ideas like Japanese. So you can say 「私は日本語が好きです」or「私は日本語が大好きです」, or ... no, forget the combinations.

May 5, 2015

Ok thanks, I will correct that.

May 6, 2015

Thanks for this... extensive clarification. I was sure I'll learn something new from that:)

I think I'd call だ、です、である a sort of "copulas" rather than "prefixes". After all, prefixes go before something, which is not the case. Maybe "suffix"/"postfix" could work but I always thought these were considered separate entities from the word they attach to.

(for those who don't know Japanese: da, desu, de aru—all attached at the end— are the closest thing you have to a form of "to be", but you don't have to use them in informal speech in present tense)

May 4, 2015

Thanks for the correction, I tried to write "particles" but I guess I didn't drink enough coffee this morning...

You're absolutely right in referring to だ/です/である as copulas. However in Japanese grammar as taught in Japan, every lexeme that is neither noun, verb, adjective, adverb nor conjunction is categorized into the group of 助辞(joji - further subdivided into particles [助詞 joshi] and auxiliary verbs [助動詞 jodoushi]), so I prefer to simply call them "particles" (Here I'm ignoring the theory that だ/です/である should be dealt as part of the 形容動詞...). I'm totally ignorant about how they are taught in Japanese as a foreign language.

May 4, 2015

How they are called in teaching it as a foreign language, may depend on an explanation you read. Typically these are called particles, postpositions... something like (に) ついて、(に) よって or (の) ように might eventually be called a deverbal postposition.

Again, it might depend on how far you are gonna get. A serious learner majoring in Japanese will, probably, want to know from the very beginning that Japanese has many particle-like phrases that are formed from verbs or nouns. A beginner can just learn that に、まで、 から、で and について are sort of the same stuff :)

Personally, I am very fond of a Russian university-level textbook by Nechaeva, but she classifies some frequent particles as "case indicators" (not that Japanese really had any cases). It seemed funny at first, and then I just got used to it. Maybe it's a useful metaphor for a native speaker of Russian. Basic particles do indeed modify a noun or a phrase to get some meaning. And, just like with grammatical cases. that does not mean that only one meaning is possible or that there are no rare meanings only used in some contexts. For example, に is used both for place and for "goal, direction", and that's perfectly OK. It is also used with lots of, er, deverbal postpositions, where you can just forget about its main meaning.

May 4, 2015

Just to say a quick word of thanks. It's really interesting to hear of different approaches from different languages. I felt that when I was learning Sanskrit both from English and Japanese sources, but it's even more intriguing to think of different roads leading to my mother tongue.

May 5, 2015

Wouldn't some of those 248,026 not really occur in practice though? For example 「お前を愛しています」, mixing casual and polite forms.

May 4, 2015

「お前を愛しています」 is used! This is because, especially in older generations, it is not uncommon to talk to one's family in the polite form. And as I have written above, お前 could be used to address someone very close to you.

But you're right, there are many forms that would likely not appear: 「あたしはお前が好きなのである」 is a combination of very informal pronoun and super formal sentence-ending. The reason I've counted this is more or less answered by Shady_arc. In fact we're doing the daunting task of finding less-likely combinations, splitting the compact form into multiple forms, trying carefully not to muddle them up. And yet users do often report these unlikely combinations, as if they were testing us. And if they do, there's no good reason to reject them.

May 4, 2015

LOL. They must have learned Japanese exclusively through reading mangas about samurais.

May 5, 2015

「お前を愛しています」 is used!

Just my luck. >.<

May 4, 2015

Nevertheless, it sounds like a phrase from Taisho or early Showa periods. It can be used, but it still sounds archaic.

May 5, 2015

Yes, but there exist people who do report them. We also have users insisting that 我 (ware) for I or 汝 (nanji) for you should be accepted. Now what era do they come from?

May 5, 2015

Of course. However, from a contributor's point of view it is easier to create a few answer generator strings which accept a wide range of answers, including some weird ones.

Making the system only accept the best options (stylistically and grammatically) takes much time and isn't worth it unless you are teaching Japanese rather than English. By the way, it also leads to a large number of short-form answers, which makes further corrections HARDER — a living, breathing contributor will have to carefully read everything to see which string generates which answer.

The issue the Japanese course has is mainly with Duolingo's limit of 3000 correct answers per sentence. Certainly, adding stuff to account for all 248,026 variants is no fun, but at least it is manageable within five or ten minutes.

Consider the following sentence and its variations:

  • <A girl is sitting at the cafe, eating cookies/a cookie> → let that be the meaning of the original sentence
  • [A/The] — [girl/young lady/young woman/gal/lass] — [sits/is sitting] — [in/at] — [a/the] — [restaurant/cafe/café/coffee house/coffee shop/diner] — [/and] — [eats/eating] — [the biscuits/the cookies/cookies/biscuits/the cookie/the biscuit/a cookie/a biscuit].

Certainly, it generates "A lass is sitting in a diner eats the biscuit". Otherwise, it is a good way to accept more than 15000 answers which more or less precisely catch the meaning of the original (something about a girl in the cafe eating cookies).

If you teach English, you'd opt for a less forgiving set of answers. If the course is meant for English speakers, I wouldn't bother.

May 4, 2015

どうもありがとうございます, Kippis!

May 4, 2015

First, in Japanese the structure is slightly different because "love" might be expressed with an adjective. Which gives you a form of "to be" in a sentence.

  • you can say "I" in different ways or omit it altogether
  • you can say "you" in different ways or omit it
  • if you use an adjective for "love", then the form of "to be"(if present) may vary depending on your desired level of formality
  • you can say "love" in different ways. If you use a verb, different levels of formality are possible
  • you can add a sentence-ending particle to give a unique flavour, and also shift word order to change the emphasis

Japanese, being not an inflected language, easily combines all of the above together—and let the variables alternate almost independently of each other. Even if we just mean an exchange between lovers (not a daughter's "I love you" said to her mother).

Essentially, it boils down to this: Japanese and English are quite different languages, and you cannot really claim that any of the above alternatives are "worse" or "less precise" than the most obvious option. Less likely and more tiresome, probably :).

In the same way, the Japanese sentence "Me — soup" may translate to all kinds of "Could I have some soup, please" in English. And, in a sense, they will all mean roughly the same as the original sentence (admittedly, English does not allow for that much variation).

May 4, 2015

I know quite a bit of Japanese and I will try to explain what I think they were doing.
The words 'I love you' in English can only be said in that order. Japanese is more flexible, three words can be arranged in nine ways. Pronouns, particles and politeness markers can be both added and dropped in Japanese. The longest 'I love you' I can think of in everyday Japanese speech has about seven words in it including particles. If you count both adding and removing particles and pronouns that sentence itself could be arranged in almost 49 different orders. Japanese also has different levels of formal and humble language. If you can think of five words for love, more than a hundred words for I and you, add and remove particles and pronouns and use all language levels the maths will allow you to multiply to a very high number fast.

May 4, 2015
  • 1465

I don't know if my explanation accounts for that many, but there are kind of an absurd number of ways to say you and (especially) I, a lot of ways to say "love," as well as some grammatical changes depending on context which provides for a lot of possibilities.

May 4, 2015

I hope that the speed of the Greek Team will be like the speed of Dutch and Irish during the Itch War [?]

May 4, 2015

Good Job Course Contributors! Waiting Patiently for Greek for English Speakers but....

May 4, 2015

Have patience Baymax13 :) We are working for the best possible outcome :)

May 4, 2015

I hope you guys update frequently,

May 4, 2015

You, as well as everyone else who would like to be informed and updated about the course, are more than welcome to ask on my stream :) We do not post updates because we simply do not have anything extraordinary to say, like new skills, etc. :)

May 4, 2015

Well every little thing adds up. I think adding monthly updates would be for the best because it would stop idiots like me from bothering you about how much you have progressed.

May 4, 2015

Trust me no one is bothering us! :) Actually it is quite amazing to see that so many people are anticipating our course, and we thank you for that :D

May 4, 2015

Also do you have an estimated date as to when it will be done (The one in the incubator is unreliable, thus why it says Yiddish will be done in 15 years.)

May 5, 2015

It is still too soon to tell.. Duolingo gives the 29th of February 2016 as an estimated date. We have completed 18 skills but we still have a lot of work ahead. My personal estimation is that the course might be ready by the end of 2015, but, really, you can never tell for sure :)

May 5, 2015

I know thank you so much for all of your work!!!

May 4, 2015

You had Ethiopian food?! Why wasn't I invited?! That's not fair :(

On a more serious note, the Incubator Summit looked fun, I'm glad I'm not learning Japanese... Thanks for the insight, Jiten! :)

Also, part of me would love for the fact that they had Ethiopian food to mean something else, but it probably won't...

May 4, 2015

Adam, while we were eating delicious Ethiopian food, I caught a glimpse at one of the curtains that had a sentence written in Amharic on it and I was thinking of you, about your passion and numerous posts on Amharic. Keep fighting for this hopefully upcoming language course!

I hope that you will be part of the incubator family soon. You absolutely deserve it!

May 4, 2015

That's awesome, I wish I'd been there! I'm addicted to Ethiopian food!

Many people have responded positively to the survey I posted yesterday (with some people wanting it on Duolingo even though they wouldn't learn it), and my posts have also received a lot of positive feedback, both of which make me hopeful. :)

May 4, 2015

Very well said. Looking forward to the news! :)

May 4, 2015

Agreed, Amharic definitly should have a course on Duolingo in the near future! :) Keep up the good work Adam! :)

May 4, 2015

Perhaps it does! Have faith!

May 4, 2015

Keep believing, Thanks to your effort, everyone here knows about Amharic. It may be sooner than expected. It doesn't have that difficult (technically) script as Japanese. : )

May 5, 2015

This is my favorite update of all time <3. It was great to meet you in person Jiten ^_^

May 4, 2015

I have to say this was my favorite update of all time too! :-D Great meeting all of you!

May 5, 2015

Woohoo! Can't wait for Esperanto for English :)

May 4, 2015

Very exciting that the Greek team's progress is ramping up so quickly! Can't wait for the ever-accelerating release!

May 4, 2015

Thank you as well, Jiten. It was great to meet you and all the other guys. I'm amazed by the creativity and the personality behind each virtual face. This summit was special, I will never forget it and we should do this again in the future ;-)

It was also awesome to plan the future of our courses while sitting next to each other. I did it yesterday together with sommerlied (from our DE-EN courses) and luke51991 (from the Norwegian course).

A special thanks to Myra, Vivian, Kristine and amuzulo for organizing it and to Andre from the dev team to answer hundreds of technical questions.

Safe travels to you, Jiten! Please come back to the US safely =)

May 4, 2015

If was wonderful to meet you as well! Your city is awesome! It was fun that a few conversations went this way - you talk to someone for a few minutes and then upon hearing their Duolingo Avatar name it is like - "omg, so you are fill_in_the_avatar! :)

Hope you guys had more fun at the polyglot conference.

I am back in the US, safe and sound :)

May 4, 2015

You're welcome and I hope I get the privilege of organizing another summit sometime soon! :-D

May 5, 2015

I'm very excited for the Esperanto course and also I am very glad to see the progress of the Japanese team. ;)

May 4, 2015

The incubator summit looks like it was a lot of fun. I'm glad to see the Japanese course making progress again and I can't wait for more courses to come out.

May 4, 2015

Maybe when it graduates to beta a Japanese for English speakers will come shortly after! I would like to see that.

May 4, 2015

It isn't, they've said in a past discussion that they're having problems with the limitations of alternate translations. Apparently there is only room for 3,000 but in Japanese there are hundreds of thousands alternates for some of them. Unless they figure out a way around it, it's probably not going to happen anytime soon.

May 4, 2015

That's too bad :( Do you know if the same problem would apply with Chinese?

May 4, 2015

I would think a big problem with Chinese would be the speaking exercises, which seem very forgiving for the current languages (probably accounting for background noise, and other distractions). I imagine Chinese, as a tonal language, would need more attention to detail in the speaking, which might be difficult for a computer grader.

May 7, 2015

I don't think it does, but I don't know much about Chinese. I also don't remember off the top of my head (I can't find the original post about it).

May 5, 2015

Super excited for Russian and Greek!

May 4, 2015

Thanks for the update! :) And, by the way, how was the Ethiopian dinner?? :))

May 4, 2015

That was very delicious! We missed you at the summit!

I am looking forward to your course's beta phase :)

May 4, 2015

In any case, thank you for your patience ... or your excitement, yet lack of patience! ;)

Accurate! lol :D

Thanks, Jitengore, you rock :D

May 4, 2015

First time ever under 1% (I think)! but it was totally worth it, I'm sure the summit was amazing!

May 4, 2015

Yes! I can't wait for all of these courses to be released. I'm actually worried about whether or not my flag lineup will be able to handle three more courses by the end of this month (hopefully that's when they'll be out). It's all just so exciting.

And even though it's said a lot, it deserves to be said many times more; thank you jitengore for doing this every week. I check this update like I do an engaging webcomic and I simply cannot express my gratitude in words.

P.S. Could someone respond to this comment telling me what level they see for my Dutch flag? I thought this problem had been fixed long ago because I no longer see any discussion posts about it, but ever since the course came out the number only shows up correctly when I post a comment, but as soon as I edit it or someone replies, it shows up as level 3 (I'm level 9). Is this just the same problem and hasn't yet been fixed or am I some weird anomaly stuck in the bugs of months past?

May 4, 2015

LupoMikti, EveDance, you're welcome! Thanks for the kind words!

May 4, 2015

I see level 3

May 4, 2015
  • 1465

I'm seeing a level 9 right now for you.

May 4, 2015

I see nine!

May 4, 2015

I see level three, mine is always wrong to (I believe)

May 4, 2015

I had to reset my Dutch tree and start again. Now it shows correctly as level 6.

May 4, 2015

I saw 9 three days ago, but now it says 3.

May 7, 2015

Very informative Thank you!

May 4, 2015

Super excited for Greek! You guys are awesome! And I can't wait for your course Ukrainian Team! I know it will be great (:

May 4, 2015

Exciting news! Especially love Hebrew & Yiddish coming along! I definitely am pushing for a German/Hebrew, German/Yiddish language thing to happen as well!!!

May 4, 2015

The Ethiopian food looks fantastic

May 4, 2015

This might be off topic, but did anyone else notice that Duolingo users tend to use MacBooks over regular windows computers?

May 4, 2015

Very excited for Esperanto!

May 6, 2015
Learn a language in just 5 minutes a day. For free.