Hi, I signed up many years ago to check the platform out, but never really used it. I only recently got a subscription and I am enjoying the service a lot so far. I think it might be the most comprehensive platform for self-study I have ever encountered (so kudos for that!), but there is something that is a bit of a weak spot imho.
I am still “toying around” with some of its features and I had a look at the exchange texts feature only today.
Maybe I am missing something, but I really find it terrible.
Mostly because there is no way (as far as I see) to know how skilled someone actually is in the language you are asking for feedback on.
As an example, I checked the requests for corrections in Italian, since that’s my mother language, and most of the corrections that have been given to the existing requests are plainly, utterly wrong. They were obviously done by other students of the language, some with poor grasp of grammar and vocabulary.
I think it might be damaging for students and I doubt I will ever use that feature if I cannot trust the feedback I will get there.
Have you ever thought about a way to get the answers at least either “peer reviewed” or “evaluated” by other users? There’s other services for text exchange that do something similar and it helps understanding how reliable the feedback is.
I understand it has a lot to do with how many users use the feature, so if there’s not many it’s possibly going to be tricky, but I also believe that many students might be taken aback from using it if they don’t feel it’s reliable.
See @Bluebird14 comments about it in her profile:
I teach French to foreigners and I have a high degree from the University. (La Sorbonne). I’m sometimes appalled when I see French people here correcting texts written in French. And sometimes even some foreigners even think they can correct French texts: it’s amazing! I’m studying Ukrainian, which is a wonderful language. But I would never presume to correct a text in Ukrainian, (It would be ridiculous) you have to stay realistic and not too pretentious.
I think all of this is “community based feedbacks”.
Their values is almost none.
It’s the same thing for the dictionary provided.
From Ukrainian to French, I have seen really totally wrong translation.
Quality is uneven and most of time poor.
Concerning “Text exchange” I though they should hire teachers and then it will really have value. You need an explanation to understand correctly what is wrong.
You have pop101 sites for learning langages.
They propose to have a teacher.
I have tested it with indonesian. And for the price it’s great.
The teacher will react in one or 2 days. You have good explanation together with the corrected text.
You’re right, some translations from Ukrainian into French are average or bad.
Correcting them takes a lot of time and a lot of goodwill. I do as much as I can in Ukrainian. But I’m lucky because I can ask questions to high-level Ukrainian teachers (in addition, I speak once a week with a Ukrainian teacher who is a qualified French teacher). I am therefore privileged to have good answers on Ukrainian vocabulary and grammar.
For the French texts that I correct, I try to explain as much as possible, and I answer learners’ questions (in private)… I’m frightened and dismayed when I see French texts being corrected by people who don’t have the level to correct anything in French. But that’s how it is on LingQ, it will never replace a quality teacher, it’s only an interaction, possibly friendly (or not) … with many limits.
I think you’re all too harsh. The big value in posting texts, is that it gets you, as a student, to think about how to convey certain ideas in that language. If I want to link two concepts, which preposition should I use? Do I want to say that something happened despite of, or because of of something that happened earlier, etc? The combinations are almost endless.
So just the fact that you sit down an formulate meaning in that language, is something that I think is very beneficial. And I wouldn’t be able to do it, without some designated reader. Then, when it comes to the corrections in themself… they can be more or less helpful, more or less skilled, more or less pedantic (sometimes people even correct quotes from esteemed autours). With time you get better at evaluating them, and you learn to divide corrections into clear cut fault, perhaps non-idiomatic or not in the preferred style of the corrector/nit-picking.
LingQ has language proficiency tests, at least for some language combinations. Maybe that would be an idea to make sure that those who correct in a certain language have some proficiency. However, in the most common languages this is probably less of a problem.
I too have recognized that a lot of the corrections in German contain errors. On the other side it’s a users-help-users service and free, so the expectations one can have might be limited.
And me neither corrects English texts. However, this is mainly because there are tons of good or native speakers around here who can handle that.
Personally I just use google translator for writing something down. It’s not optimal but I get a direct response and if I am unsure I have several grammar books or can google stuff. One could also use chatGPT (I use it from time to time for good old translations).
You can always use points to pay tutors to have your writing corrected, that option is also available under the Tutors page.
Writing exchange under the Community is Free and all users can jump in here to make corrections to your writing and provide feedback.
I just checked some of the “corrections” people offer in my native language and I must agree - many of them are provided by people with very questionable command of the language they claim to know enough to be in the position of correcting others. Sometimes they even try and dare to correct things that are totally fine and they actually ruin someone else’s perfectly correct sentences. I can only imagine how confusing it must be for learners, since there’s no way of asking for further explanations and clarifications of these “corrections”. I also saw cases where someone didn’t really understand the idea and “upgraded” someone else’s text, changing the story, adding bits that were not present in the original, etc.
I guess it all boils down to “you get what you pay for” - since it’s a free feature, one shouldn’t expect anything reliable. On the other hand, it’s an educational platform after all and some of us are actually paying for it, so deliberately providing and tolerating such form of serious misinformation should be addressed somehow in my opinion.
They could implement a system were you can only correct texts in your native language and in return get points which you can use for posting requests for your tetxts.
The downside would be that the supply and demand for different languages might be to uneven.
I think it generally does quiet well with european languages. From what I’ve heard though it’s not very well performing in Korean and therefore maybe other non european languages, especially those that aren’t used by so many people.
But as there are several ai chatbots out there, others might perform better. It’s just hard to judge for a non-native.
Another issue is that some of those people might even be native speakers, but beeing a native doesn’t make you necessarely good at the language. If the person who wrote the text uses grammar points, vocabulary or formulations the corrector is unfamiliar or unsure with, he might fix it to pieces.
I know there’s paid tutoring and that is potentially a great option, but anyone can apply to be a tutor, without any check, or did I get this wrong?
As an example, I have seen someone claiming on their profile here to teach my language (which I highly doubt from what I have seen them writing in my language, or maybe they do but then… ). They don’t seem to be a tutor for the language on this platform, so no big deal imho, however that still seems like a bit of a potential weakness for the platform itself.
I hope this is not seen as some sort of an attack to the platform because that’s not my intention, I just genuinely think that being able to understand how reliable the feedback or the tutor on the other side is (e.g. being able to “vote” for corrections, see if who did it is a native speaker or a student…) would be an advantage and could get people to engage more with the feature.
I also understand the work needed to add these type of features might have been evaluated and the expected results might not be worth the effort, so I totally get it.
Just voting is not enough in my opinion. If someone wants to correct something or disagrees with someone else’s correction, they should also have a possibility (maybe even an obligation?) to explain why, what is wrong, give additional examples, etc. Otherwise there’s no difference between trusting random people’s corrections and trusting random people’s votes.
@hyacinthebesmoker Well, to be fair, those who make the corrections can already do so. You can always write why something should be done a different way if you think explanation is needed.
In my experience it does decently well for Mandarin Chinese, not perfect but for immediate feedback I think it is sufficient. It also depends on the particular model you are using, I wouldn’t recommend anything less than 4o.
Of course, it also depends on how badly you mangle the sentence.
I agree with the “you get what you pay for” sentiment. When there is no real incentive to help other than the possibility of reciprocation, you can’t expect professional-grade revisions.
If you have ever spent time correcting someone else’s writing, you know how difficult and time-consuming it can be. It’s essentially translating at times, which is a different skill than knowing a language. There have been times when I had absolutely no idea what someone was trying to say. There were times when I knew what someone wanted to say, but I was so deep in the weeds that I couldn’t think of a natural way to say it. That’s why it’s easy to end up with those “upgraded” or ruined versions. It’s very rarely just fixing conjugations, tenses, or pronouns, and it’s a lot more work than people think.
If you have ever spent time correcting someone else’s writing, you know how difficult and time-consuming it can be. It’s essentially translating at times, which is a different skill than knowing a language.
It’s very rarely just fixing conjugations, tenses, or pronouns, and it’s a lot more work than people think.
Exactly. Those “corrections” I saw are a model example of the Dunning–Kruger effect. At this point even something like Chat GPT or Perplexity AI does a much better job - and one can always ask for additional explanations and examples, if things are not clear.
You’re absolutely right! I see English and German corrections that make my hair stand on end, and the poor people being corrected think they’re doing a fantastic job (as do those uploading the correction).
Yes the utility of this feature is VERY hit or miss to put it mildly.
I decided to try it out and wrote something casual in Spanish including a stupid blart about yerba mate and someone changed “yerba” to “hierba”, and that right there was a … yeah, no, I’m done here, lol…
I have seen the same problems. When I ask for corrections to my own writing, I hope to get more than one response, so I can check them against each other. For example, if two people tell me something is wrong, I figure I can trust them. If one person tells me it’s wrong, it might or might not be wrong - someone else might have just not noticed the error. I do check to see whether the person is a native speaker, though of course that’s not a guarantee either.
It’s still useful to me get the input, while allowing for the possibility that it isn’t always correct.
). They don’t seem to be a tutor for the language on this platform, so no big deal imho, however that still seems like a bit of a potential weakness for the platform itself.
