Does leveling up give you more complex exercises?
The "you" in the question is not intended to be understood in its generalized sense. I am asking about YOUR tree, and YOUR experience. This is the reason:
I was pretty excited about a tree that would give me increasingly complex exercises. But at the same time I had my doubts since I saw the confusion in the forums. Some say they have this, others see no difference and I was starting to suspect an A/B test where some users have visual crowns but no actual levels. Yesterday I just couldn't resist leveling up the Languages skill (on the row of skills 33-35) in the Chinese tree to level 4. When I tried the level 4 version I could not detect any difference. To make sure that I wasn't missing anything more subtle I did it once more, counting all the different exercise types. And I also did "Shopping 1" lesson on level three, as a comparison. (It is on the same row as "Languages".) Level 3 and 4 actually produced a similar set of exercises (in terms of exercise type) with the major exception being that level 4 didn't make me translate into Chinese at all. (This has not happened again though.)
I commented on this in another thread and the user zanzaboonda suggested that this could have something to do with the Chinese tree. So I am asking all of you, are you seeing what I am seeing, or are you seeing something different. Also, how far down the tree do you need to go before there is a real difference between your levels? If people can see the benefits on the short Japanese tree, skill 33-35 should be enough, but any input is very valuable. Trying to figure this out on my own is just wasting time on useless exercises. I have been trying out the system instead of studying the language.
The other thread mentioned: https://www.duolingo.com/comment/26876629
Interesting point to raise.
You may also be interested in reading help : What-are-Crown-Levels
And as with all the courses on Duolingo , they are constantly being worked on. And as a contributor to another course, we are also revising our tree to ensure the next update we use will also be seeking to make the most advantage of these extended possibilities.
Also , it is interesting to note with the comment :
"But until today, users have seen only part of the material the service had created, much of which it wasn’t actually presenting in the app. “I would say we were throwing out at least half of our content completely,” says Karin Tsai, the engineer who oversaw its new revamp. “Some sentences and exercises were just way too hard for users at the beginning level.” With its update, called Skill Levels, eager students have the option to do more-complex exercises, even from day one. Duolingo also announced that it’s working on additional exercises, such as reading and listening comprehension, that will roll out later this year.
ie, that not all the material course contributors had been developing for the course, have been able to be see/was not sought to be seen by learners.
This is one of the key advantages of this change, that it does encourage you to extend yourself. Though there may be a bit of a delay as we contributors, and also other back office operations continue to adjust to take full advantage of this potential.
Mind you this is is all In My Opinion, and I am not seeking to speak on behalf of Duolingo in this regard.
In theory I completely agree with the advantage of the change, but in reality these are just empty words when I look at my tree. I level up and nothing happens. It doesn't help that Karin Tsai tells us how great it is to expand the content, because I cannot see this content.
The information on the help page you linked to is not new to me. But the fact that you gave me this information about going deeper into the skills, even though the central part of my original post clearly states that this is what I am already trying to do, does confirm that we are seeing different worlds. (Basic psychology really, we interpret things according to our own experience. If you had experienced crowns without levels you probably wouldn't have given me that link.) The question is: is this because of an A/B test where we are in different groups, or because of the Chinese tree code/content?
Currently, DuoLingo is doing an A/B test about how crown levels work best, and if they work at all. That means that not everybody has the same kind of tree.
That is what I suspected. Do you have a source on this, have they made any kind of announcement? Because from what I remember reading, the official A/B test was made before launching the new system. Obviously they are always doing A/B tests, both announced and unannounced, and probably a bunch of them at the moment to tweak the new system. But anyway, if there is an official or semi official statement I would love to read it.
I thought the AB test was over and they were just rolling it out gradually now? I definitely was not in the AB test, but I got crown levels at the beginning of the month. They've been pretty unambiguous in saying that they are implementing crown levels for everyone. I think they are just rolling it out in groups so that they don't have to deal with updating all the duolingo apps out there at once.
I know, I asked them rhetorically. Only the staff could give me the full answer, but by asking about other people's trees I was hoping to get a few more clues about what is happening.
I never received any notification about the Chinese translation for the answers. May I ask if there are contributors revising the Chinese part?
So, I've gotten a lot of responses on posting revisions for English answers. I submit far fewer Chinese translation corrections, but I have recently gotten a few notifications for some of the ones that I submitted having been accepted. So they do appear to be working on it.
I also think that if two people recommend the same alternative answer, when they accept the first one, the second person doesn't get a notification.
Sending the e-mails requires the team clicking the button to do it, so there's no guarantee they do it in every instance (sometimes they deliberately choose not to). However, I think that everyone who submitted the translation for which they are clicking the button will get an e-mail (I think I saw a thread where two people worked out they'd gotten the same e-mail). If they're motivated, they can go through and send the e-mails to everybody who submits a missing answer but with typos or what have you. Might even apply to capitalization / punctuation differences, not sure.
Thinking linguistically, this points to a rather interesting way that Chinese and English may form something of a perfect storm for this kind of task. Both languages have an unusually high degree of syntactic flexibility, and at the same time, they have very little overlap. In languages like French or German or Spanish, the degree to which they overlap helps concretize a "correct" translation. Concepts have narrower ranges of correspondence, and in many cases, we are basically using the same words and grammar. With Chinese, the overlap is functionally zero.
These two factors make for a fascinating explosion of permutations. Writing out all the possible translations one way, from Chinese to English, is a large task, but then when you try to reverse it to write out all the possible Chinese back translations of the possible English translations, (since I believe sentences are used in both types of exercises) this opens some very interesting doors, because English phrases that have the same meaning will have multiple translations in Chinese that have other meanings, and this expands as a web that goes far beyond the original core concept. Duolingo seems to have worked very hard to limit that kind of translation creep, but it's a fascinating problem to think about, and might make for a good drinking game...
Yes, it seems the same sentences are used in both directions (although particularly from English into Chinese I think they can be disabled if a certain sentence would introduce too many complications for whatever reason). When contributors enter a new Chinese sentence, they naturally have to provide the English translations. Then they choose a "preferred" English translation (incidentally, the one that shows up in the sentence discussion), sometimes, perhaps, a couple, and provide the Chinese translations of that. Keeps the number of translations from growing exponentially that way ;)
Of course the numbers can still be very large. There have been posts recently from Norwegian and French contributors about how many translations are possible. Norwegian got into the several hundreds for a pretty simple sentence. One of the French examples had 1,700. I know in the Portuguese course they don't even allow all the common translations of "you" to help them no go over the 3,000 translation limit imposed by Duolingo. Apparently Hungarian and Japanese, with their flexible word orders, have received a raised limit of 10,000, and that still isn't always enough.
Compared to those two languages I think English and Chinese are relatively manageable flexibility-wise, but I don't know much about Chinese. I've just read that the word order is fairly firm.
I'd thought of the typo problems as well, but Mr.rM is talking about the Chinese part, so typos are kind of a different beast there. (It's kind of an all or nothing proposition).
Honestly, I think it is that a bunch of people suggest a fix, they accept one, and have an avalanche of work to catch up with, so they don't bother to follow up with redundant reports.
I've gotten Chinese translations acceptance notifications in the last week, but some of these were ones I filed a month ago. I think they're probably moving through things in shifts, and prioritizing adding the English translations. Honestly, I can see why. It seems a reasonable triage decisions.
I’m definitely seeing lengthier sentences, and zero one-word definition questions or matching pairs, and this is reaching level 3.
I plan to start from the top of the tree once everything has reached Level 3, and actually note differences for a more comprehensive look at higher levels, but from what I’ve noticed so far, there’s definitely variance for my account.
Just to make sure, are you talking about the Chinese course? (You don't have a Chinese flag above your post, so I just had to ask.)
I am talking about the Korean course specifically!
Edit: I would not expect these changes in possible crown variation course to course, as the crowns themselves are either ON for all courses in an account, or off, so this is my reasoning. If they are indeed conducting another A/B test based on pure repetition vs differentiation, that should be made apparent soon. :/
I have seen the introduction of a couple of new sentences here and there in the earliest Chinese skills, although not in the first skill, which is the only one I have completed to level 5 since I got Crowns a couple of days ago.
I don't know for certain, but my suspicion is the Chinese contributor team is overwhelmed. Some of them are also responsible for the 'English from Chinese' course.
In Dutch I've seen it from the very top of the tree. Level 3 is plenty of typing in Dutch, but a lot of it is write-what-you-hear. Level 4 has more straight translation. The sentences also seem to be different at level 4. It's probably easier for that to happen in a course with a greater stock of sentences. To me it seems that Chinese is limited on that count (also limited by the fact I don't think it has write-what-you-hear sentences).
I did a hiragana exercise when I saw that somebody said (which surprised me) that there were additional exercises there. I'm not sure I'd phrase it that way, but the the matching boxes exercises were harder than what I recall seeing previously: all of them having the same consonant sound. I don't know what the lower levels look like.
Did a test for Chinese. The gradation was much more subtle. I took the Nation skill from level 2 to level 3. The breakdown of the exercise types didn't differ between the two levels: 4 write in Chinese; 9 write in English; 3 match the pairs; and 4 select the character (I was surprised that these figures remained constant within a level, as well; for Dutch the counts change a bit from lesson to lesson). What did differ actually seemed to emerge within level 2: the option for answer boxes when answering in English largely went away. It seemed like it went from being there for all of them to being there for only one or two as of about the 5th of the 6 lessons in the level (my recall isn't exact; I didn't have in mind to be looking for changes w/in a level). I only did two of the nine lessons at level 3, so it's possible more differences could emerge, or at level 4, obviously.
Given that I see things changing significantly for Dutch but much less for Chinese, perhaps it is the case that, as sort of telegraphed in the original announcement thread, the system isn't yet totally set up for the languages built on the system for the new character exercises. (Of course, the fact it's working for Korean would weigh against this hypothesis.) Alternatively, the Chinese course could be stymied because it seems to have a much more limited inventory of sentences than some other courses. However, at the top of my Dutch tree the lessons were eight questions long, and in Chinese they were 20, so that would wind up with things being substantially more repetitive even given the same size stock of sentences.
I kinda think that is mostly because Chinese is a new module, and that the contributors are largely focused on getting the current sentence answer problem under control. (They've made great progress) and that means that they haven't been programming too much new content yet.
The new French tree, on the other hand, was just rolled out, and built for Crowns, and it is pretty amazing. So I don't know that these are issues with the crown system so much as the dynamics behind which languages have more active teams, and what is consuming the attention of those teams.
I'll grant you the repetitive stock sentences get old quick. And since I've only got five levels to go through, I'm planning on taking it really slow. I check back every now and then and up a relatively low level skill to see if I notice a difference, indicating stuff has been added. If it looks promising, I'll play around with the rest of the tree. I get that this is not going to be helpful for new learners, but perhaps once getting past level 3, where the new material should really kick in, new learners can take a break and use what they know to check out movies and newspapers, giving the contributors time to add new stuff at a more reasonable pace. Some of the earliest languages like French, Spanish, and German have evolved in really great ways long after one would have finished the initial tree...
This is very interesting. My tree seems to give me 6 exercises where I am supposed to match either one sound to one of several characters or one character to one of several sounds, on both level 3 and 4. (As compared to your 4.) But I get 3 match the pairs, just like you. And 10 sentences to translate, or was it eleven... (I'm quoting my notes from memory.)
Another even simpler explanation for the lack of increasing difficulty is some kind of bug in the Chinese specific code. At least it is a bit reassuring that things changed more on level 2. This means that there is hope for the future. Unfortunately I probably need to level something up in another language to compare to what extent the things I see (or do not see) is related to my account, and to what extent they have to do with the specific courses. But I don't think I want to. Since you do see a difference between Dutch and Chinese I am going to assume that the course is one of the reasons behind Crowns not working properly for me yet.
So this is a little difficult for me to tell. This is complicated by the fact that I tend to move through chinese lessons very quickly, and that I kind of wander in, do some, then wander away for a while. As a result, it is hard for me to remember if it's my subjective impression that things have gotten more advanced, or if it is definitely new.
But that said, my impression is that there are a few shifts taking place above level three, which is where the good stuff starts happening. The first is that the type of questions change. You get fewer matching questions, more translation questions. They also seem to have much longer sentences or statements. I get the feeling that they are also using some vocabulary from other lessons, but that's really hard to tell.
I've got a bit clearer picture of this from German, which is what I'm practicing on right now, and they definitely are introducing newer material at higher levels, as well as using more advanced grammar, and moving away from multiple choice and more towards "write it yourself in German."
So I think there are probably some things that had been on the cutting room floor that show up more in Chinese. They definitely seem to be less rote, and are more comprehensive exercises. But the differences I've seen so far are also pretty subtle, so that if it is new, it is still like 90% review, albeit, a more challenging type of review.
Though, I wouldn't be surprised if the Chinese team is currently really focused on updating potential answers, which has been a major effort, and don't have time yet to be developing new material, whereas older courses that have fewer kinks to work out are much quicker to deploy new Crown Level content.
I should also add, I did a lot of the Chinese skills a long time ago, and remembered them mostly being fairly short exercises. So I'm not sure if my impression that they now have more complicated exercises is from something that was added earlier, or something unique to crowns.
Thanks. If you are seeing fewer matching questions then we are indeed seeing different things, and it is a big help to know this.
Well, I should say, I think I am seeing fewer matching questions. I haven't been methodical about it like @piguy3 has. My impression from the first lessons I did when Chinese just came out was that it was like 90% matching on many of the lessons, and many times it was the same matching exercise back to back. The course was young, content was still congealing, I get it. So coming back, I felt a great relief that I wasn't doing what seemed like never-ending matching exercises. Is that because it improved since launch? Or is that crowns? Or is it just that it was never as bad as I remembered? I can't really say.
But given my experience with the other languages I'm using with crowns, I am pretty confident that eventually it will have a lot more stuff, and the differences will be less subtle. The new French tree for example, has been very rewarding above crown level 3 for me, more than it had been before I switched to crowns.
I haven't noticed a real difference in levels, in either the Welsh or the Italian course. If there is a difference, it's very subtle. I was hoping for more translation into the target language, for example. Maybe the farther down the tree I go, the more of a difference there will be? I certainly hope so. If the levels were to actually get more challenging, then I would be pretty happy with the overall crown system. This lack of increase in difficulty is honestly the only thing bothering me about this new system...
Out of curiosity, do you have the option of answer boxes in the target language for levels 2 and 3, but then not on level 4?
At low levels, I have seen no new skills, even on exercises that ask you to repeat them 60 times in total.
Today, I decided instead to try the absolute end of the tree for German with the "Culture" lessons. There were a couple new sentences at level 3, but many of the new additions seemed to be simply random sentences from earlier lessons that didn't really fit in that category at all. For example, I don't really know why a fairly simple sentence like "Das ist die Frau, die wohnt in Berlin." ["That is the woman who lives in Berlin."] would fit in the "Culture" category. (Though, to try to be positive, that was perhaps the first time I can remember seeing a relative clause appear outside the "Relative Pronouns" lesson.)
A few exercises seemed be heading in a positive direction: for example, the biggest change I noticed was more exercises than before that actually used the preterite or the perfect outside of the specific lessons that teach those tenses.
In Russian, I see that the excercises get harder at higher levels, but that is (mostly?) because of the kind of excercises they are giving us. There are no more word banks and lots of translations into the target language.
However, just like you, I have not observed this in Chinese, but the kind of excercises in Chinese have always been quite different from the European languages. So my guess is, that they don't have a concept for this yet. If I remember correctly, they stated in the original crown announcement, that crowns were coming for all languages except Chinese, Japanese and Korean, probably because of the lack of ideas (yet?) for these languages.
I find out this in Danish,if you increasing crown level of one of lessons,it becomes harder.Kinda interesting system,but not at all
I've tried maxing on the crown level for Duo topic and I just get the same exercises over and over again. So I'm kinda confused. :/