Crowdsourcing by non-native speakers

Discussion in 'General Discussion' started by Peregrinus, Jun 2, 2014.

  1. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish
    In my German studies, I use wordreference.com's German dictionary and the German dictionary of dict.cc a lot. While I rely on wordreference more, dict.cc supplies important information not found on WR, and especially allows searches by phrase, which is highly useful for idioms and phrasal usage.

    A problem with dict.cc though, is that often there are odd English translations, and also that sometimes the most frequent English translations are not that of other sources, and also not what I would expect. Especially with the odd translations, that vary from wrong (i.e. word is close but does not really exist) to less common upper register renditions (but which are higher up the list due to being sourced by more contributors), I have to think these are contributed by speakers for whom English is not their L1.

    While I hate to complain about any otherwise good and free resource, this to me shows the downside of crowdsourcing, that is often mentioned in discussions about same. Crowdsourcing is great for reporting objective facts that don't need processing or interpretation. But if there is a subjective element involved, then despite best intentions, the end result can be flawed to some degree.

    Has anyone else here ever had similar thoughts, or am I offbase?
  2. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish
    I guess no one else around here has this problem, although there aren't many of us around here yet of course.

    Since I use the dictionaries above daily, another related problem to me is when the English translations do look done by a native speaker, but in addition to a good literal translation which does not seem out of place (too stuffy/awkward), they also throw in a colloquial translation. That is fine IF the German word or phrase is itself colloquial or figural, but not otherwise. This is especially confusing with idioms. I want to learn German idioms, not English idioms translated into literal German equivalents that are not really used by Germans.

    This shows a potential danger for first time language learners who lack the experience in any L2 or the specific one being studied. This is probably only an issue with free online dictionaries, versus either paid online or traditional printed ones. While I also don't notice this problem much with Collins, I simply like wordreference and dict.cc better.
  3. wulfgar

    wulfgar New Member

    Joined:
    May 24, 2014
    Messages:
    15
    Native Language:
    English
    Intermediate Languages:
    Thai
    Yes, I have experienced this. One unfortunate example is LingQ. They crowdsource articles and accompanying audio. Most of it gets done by natives, but there are exceptions which I consider to be "dangerous" for beginners. For example, there are a lot of Mandarin lessons where the speaker is Chinese, but not a native Mandarin speaker. Not good to model your accent after theirs. Also, I saw an english text written by a Russian guy with several grammar errors. Every time I bring a situation like this up on their forum, I get told they aren't going to do anything about it..not even warn users of non-native content. So I don't bring it up any more. That's working out pretty good for me.
  4. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish
    That is disappointing to hear about LingQ, though not surprising. Obviously they and other crowd-sourced sites want to encourage people to not only contribute, but also to feel that their contributions are valued. But there has to be a balance or the quality is impacted. And when they discourage users through inaction to address the quality issue, then it can only get worse. The "wisdom" or "knowledge" of the masses often is just a race to the bottom as far as quality or accuracy is concerned. Again as you and I have noted, the main danger is for beginners who have no way of assessing the quality/accuracy of the offerings. So the tip for beginners is to use crowd-sourced sites only once having reached an at least advanced beginner level and start to have a sense of such things.

    Thanks for responding. I am glad I am not the only one who encounters and is concerned about this issue.
  5. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    Race-to-the-bottom... indeed. Word-reference.com has been my go-to dictionary for many years, because I've mostly been dealing with language combinations that they have commercial dictionaries for.

    Their forums have Always been useful for seeking out the meaning of phrases not included in the dictionaries, but it's always been down to me to work out what the answer is, because typically you find an answer by a native speaker of your target language who clearly understands the phrase, but mistranslates it into English. The surrounding discussion is always vital to working out the actual meaning, but take that away and drop it into a "crowd-sourced" dictionary, and it's just a collection of mistakes.

    Ten years ago, I saw "the wisdom of the crowds" described in terms of golf balls. You can't see the flag, but you can see lots of golf balls, so aim for the other golf balls, and you'll get close to the flag. Yeah, right. All it takes is for some idiot to fire six golf balls in the wrong direction and ever sheep follows the idiot.

    Crowdsourcing is a waste of time (see also the Emperor's Nose fallacy).
  6. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    I'm just waiting for someone to get sued to oblivion for breaking minimum wage laws.

    It is illegal in most developed countries for a for-profit company to use unpaid labour. Open-source projects aren't really a problem, because the end result is a free product anyway. But upload something to LingQ, and it's controlled by LingQ, and available only to LingQ subscribers.

    If it walks like free labour and quacks like free labour....
    luke and embici like this.
  7. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish
    The sources I mention above, i.e. wordreference.com and dict.cc are free, and those type of sources are my main interest. One doesn't expect them to be curated, but one does LingQ. And probably the only way under either model to control quality would be to somehow vet contributors as to native language and restrict translations to native speakers. Even then, the native speaker would also have to be highly proficient in the other language. So another thing to vet. But of course the sheer size of a dictionary which includes words, phrases and sentences, makes it impossible to be picky and achieve any kind of volume.

    The real different model where quality is possible is the traditional dictionary publisher, like Oxford, Collins, et al. But such partially free (online only like Collins) or subscription models may not be sustainable. I would actually use Collins as my main source if they had better example sentences (theirs tend to be pulled from only newspapers and lack a conversational focus), and their format were different. Oxford has a nice product in many language pairs, but though I don't mind paying for a dictionary that is easily viewable on my desktop, I hate the subscription model.
  8. Cainntear

    Cainntear Active Member VIP member

    Joined:
    Apr 29, 2014
    Messages:
    343
    Native Language:
    English
    Advanced Languages:
    Catalan, French, Italian, Scottish_Gaelic, Spanish
    Intermediate Languages:
    Corsican
    Basic Languages:
    Dutch, German, Irish, Polish, Russian, Welsh, Sicilian
    That they are free to the user is irrelevant -- they're ad-supported commercial operations, intended to generate a profit, and they should either invest in doing it properly, or admit that their business model is unsustainable.
  9. Chung

    Chung New Member

    Joined:
    Jun 2, 2014
    Messages:
    6
    Native Language:
    English
    Advanced Languages:
    French
    Intermediate Languages:
    Czech, German, Hungarian, Polish, Serbo_Croatian, Slovak
    Basic Languages:
    Finnish, Turkish, Ukrainian, Estonian
    My experience with crowdsourcing for learning a language is restricted to using WordReference's dictionaries and Wiktionary. However like Cainntear, I always check the entry's attestation in the former as linked in the forum's thread. It's always enlightening to see native speakers hash over meaning in response to a non-native's question. Wiktionary is my first choice for online bilingual dictionaries of Finnish and most of the Slavonic languages that I deal with since the entries often have the inflectional tables as appropriate and/or a few example sentences (especially true for Finnish entries in either the Finnish or English versions of Wiktionary).
  10. Peregrinus

    Peregrinus Active Member

    Joined:
    May 27, 2014
    Messages:
    613
    Native Language:
    English
    Intermediate Languages:
    German
    Basic Languages:
    Spanish
    I too have used WR's forum, though to a limited degree, and also agree it is much more helpful when several native speakers chime in to get it thoroughly discussed. Generational, sociolectical and dialectical differences make it unwise to rely on just one source. While I have used Wiktionary some as well, and also like its tables, I find that once one gets above the most common couple thousand words, it can start to peter out. Just as with language courses, using multiple sources is best.
  11. Fabien Snauwaert

    Fabien Snauwaert New Member

    Joined:
    Jul 4, 2014
    Messages:
    4
    I'll just comment that it seems to me like what you guys are describing here is one of the limits of free websites. Instead of paying for quality content, we end up paying with our time (which is, in the end, more precious.)

    I find this particularly true for language sites, especially the "do-it-all" sites in the sort of LiveMocha and Busuu. That's also the reason why I don't use LingQ. The idea is good but, trying to use it for Russian, it felt like a lot like going through a garbage can (like when you look for the good stuff inside of an overflowing inbox or spam folder.)

    Maybe it's just me and the fact I like things to be right (e.g.: re-reading messages several times before sending them), which is probably *not* what most people do.

    I don't find this to be an issue at all with Wikipedia though. The good far outweighs the bad. The must be better organized and more focused on quality (a "Healthy debates" vs. "Pleasing everyone" kind of deal?)

    In passing, a great book on Crowdsourcing is Cognitive Surplus by Clay Shirky, very, very inspiring, with a focus on contributing something to society.

    I've got a pretty good idea for crowdsourcing as applied to languages. I don't yet see how to make a living out of it though, and am pretty busy, but that's something I feel I should do once my business is stable in terms of income. That's something that might be worth discussing here at some point.
    Big_Dog likes this.
  12. Big_Dog

    Big_Dog Administrator Staff Member

    Joined:
    Jan 11, 2014
    Messages:
    1,039
    Native Language:
    English
    Advanced Languages:
    Spanish
    Intermediate Languages:
    French, Japanese, Mandarin, Russian, Swahili, Thai
    Basic Languages:
    Korean
    Hi Fabien. Excellent post, and glad to have you here too. I look forward to those future discussions. Like you, I also lack time, but will do my best to support crowdsourcing, or group projects, that people want to do here.
    Fabien Snauwaert likes this.

Share This Page