1. Forum
  2. >
  3. Topic: Troubleshooting
  4. >
  5. Can the translation "down-vot…


Can the translation "down-vote" rules deter people from attempting to do translations?

I see that as soon as I get around the halfway-mark in a language (at least this is true for the Spanish course), DuoLingo starts suggesting that I try my hand at making translations using the "Immersion" feature, saying that I should be able to understand around 50% of the Spanish articles.

But obviously, when I attempt those translations with only 50% (or at 60% or 70% or 80% etc.) of understanding of the language, I am bound to make some mistakes in them. These mistakes in turn result in getting down-votes from the reviewers. So, if I keep attempting translations, I will collect a pretty big number of downvotes by the time I'm able to finish the course (and finally get to the 100% understanding level). Let's say I collect 50 downvotes.

Now according to the new rules, one can move to the next translation tier only if I have less than 10% downvotes - so that means, I need to get another 10x the number of upvotes = 50*10 = 500, just to move from tier 1 to tier 2.

Rather than doing all that, I will simply avoid making translations till I reach the end of the course (and get 100% understanding of articles), and will then make 100~150 translations, and get most of them right, and move to tier 2.

I understand that the whole upvote/downvote and tier mechanism has been created to ensure the quality of the translations, but I think it needs to be thought out better from the perspective of the learner. The current system just creates negative incentives for the learner to try translations.

And if top-quality of the translations is the top priority, then what is the point of DuoLingo prompting people with 50% (or 60% or 70% etc.) of course completion to try translations. Only people who have finished the course ( and have a 100% understanding) should be prompted to do translations.

April 29, 2014



Look around. The rank system has no relationship to quality of translation. You can find brilliant work in the lowest rank, and really mediocre stuff at the most absurdly high levels. Some truly bi-lingual participants have hardly started the lesson tree. At best, rank means that the person has been doing some respectable work; but a machine-translation troll only needs a couple friends of the same rank - nothing more - and he'll soon have the power to downvote just about anybody, for any silly reason that comes into his head.

Most people do not vote for anything except their own translation, anyway: up or down.

So, don't let any of this nonsense stop you. If you're here to learn you don't deserve any "downvotes". But you'll get them. The people who give them just to be mean, or to protect their version from tampering, or just because they can, are here for the wrong reason. They don't want to learn, and they don't care about teaching anything or they would simply correct your mistake. Figure out who they are, and steer clear of them. There are very few of them.


"Top-quality" of the translations is NOT the top priority. The top priority is to learn - and to encourage others in learning - a language that you don't yet know very well. You will learn in a different way by translating, than by doing the tree. Do both. Do your best. Let people show you how to do better. And learn. There is no priority above that.


Thank you Markmcopc, for your well-thought out perspective on this. It really helped motivate me to try more translations. I agree that for learners, learning is the top priority.

But I can understand that for DuoLingo (the company), the quality of the translations is the top-priority, as for them, teaching people languages is just a part of their business model, in order to get the translations done. And which can explain why they have implemented the system, the way it is.

But I agree, that as a learner, I should concentrate on learning the language, and not worry about DuoLingo's business model too much.


You'll be glad to know that Duolingo doesn't make a penny off of the translations, unless the uploader is paying for it. Chances are, you've never seen a paid translation.


I don't know if one person can convince the whole community to be a little lighter on the downvoting, but what's wrong is wrong. You should try translating some of the easier sentences, or the title of an article. As you progress and people see your higher translation tier, they will be less tempted to downvote a sentence with a small mistake. I don't really know. Just a suggestion.


What about those who wrongly downvote something as "wrong"? This teaches nothing, benefits no one, adds nothing to the translation, enforces error, punishes knowledge, and is immune to appeal. When someone with translator rank uses a downvote for frivolous reasons, it is abuse. Abuse is an absolutely convincing reason not to participate.


I agree with you, that a wrong translation is a wrong translation. And I also realize, that every correction is an opportunity for the translator/learner to improve his understanding of something that he/she did not know before.

However, the way the whole thing has been implemented practically, gives an emotionally negative feedback for making the effort of trying out translations of a greater amount of text.

I'm not saying that the community should become lighter on the error-checking or down-voting, as that is necessary for maintaining article quality. But overall, I'm saying that the design of this whole process could probably use some re-thinking by DuoLingo (the company) from the learner's/user's perspective.


Hover over the words for your native language translation when you need help. :)


Thanks for the tip, AlexisLinguist. Yes, translation suggestions help, but sometimes I understand the separate meaning of the individual words, but those words taken as a phrase, end up meaning something different from what I think they do. Those ones are the tricky ones where one trips.


For sure. That's when syntax and context become each other's best friends. :)

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