| Pages in topic: < [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183] > | Should “native language” claims be verified? Thread poster: XXXphxxx (X)
|
|---|
Phil Hand 
China
Local time: 00:01
Chinese to English
| Terrifying idea | Sep 10, 2012 |
José Henrique Lamensdorf wrote:
There is room on the Proz profile - as it is now - to include a picture of the translator with his/her declared native country's flag, as well as a recording of them singing their national anthem.
With my singing voice, I could kill the site dead in weeks!
no matter how hard they tested me, would Proz be willing to stand liable for that assertion? I doubt it.
IMHO it's a matter of translation honesty and outsourcer education.
I would accept that, but not the conclusion you derive from it. Many people have asked on this thread, why are you talking about nativeness and not PPPQQQ? The answer is because nativeness is where an obvious problem exists at the moment. (I regard people putting (having to put) incorrect native languages on their profiles as a problem, whoever's responsibility it is.)
Proz might well say they can't ultimately be responsible, because they're not set up to monitor us. Fine. But if they know there is a problem - we've already brought it to their attention - then the issue is no longer "we can't be expected to find this problem". It's "we don't want to do anything about this problem".
Kay Barbara wrote:
When you say "Proz is not set up to do that" when it comes to measuring quality, why is Proz then set up to measure nativeness in a given language? For both assessments Proz would need translators/natives to be the peers (or are you suggesting that Proz staff do this?).
I don't think any of the verification suggestions actually constitute measuring nativeness. Possibly Bernhard's suggestion. But for the most part, we've all been very careful to say, we're not trying to provide an absolute/correct measure here. We're just trying to weed out the egregious examples (see Lisa's very first post).
To test quality is much much much much much harder than Proz can deal with. The amount of systems you need, the liability to complaints - it's a massive headache. I certainly would not be willing to submit my translations to peer review - I'm afraid I have no faith in my peers at all. I would trust them to make a simple factual judgment - is Phil a Brit like he claims to be. But beyond that, no way.
Anyway, at the moment, because of the scale of the problem, I'm voting for a more of a nudge - something like a questionnaire that people have to fill in - to improve voluntary compliance, rather than verification.
| | | | Phil Hand
China
Local time: 00:02
Chinese to English
| That's a pretty strong argument, too | Sep 10, 2012 |
Siegfried Armbruster wrote:
If I wanted to outsource this selection/verification process I would hire a service provider...Why should I trust a company that is mainly interested in profit. That is ok, it is their business. Accepting that Proz is just a business saves you a lot of headache and allows you to use the features that are really strong.
OK, that makes sense. The reason you'd trust a company is if you're paying them; and most clients don't pay Proz, so they shouldn't be expecting Proz to look after their interests anyway. Have I understood you correctly there?
There is a counterargument to that, but I'm not sure how strong it is:
Proz is not run for clients, because clients don't pay; Proz is run for us translators, because we do pay. And we translators have an interest in presenting potential clients with an attractive database of translators here. (Even though a client can't trust Proz, clients can tell the difference between a well-ordered translator directory and the mess of Craigslist). So we paying customers of Proz are motivated to push Proz to act on our behalf. And in fact, that's exactly what's happening.
b) In my opinion testing nativeness would require a complete administrative structure that cannot be handled by Proz.
I agree with that. I've said all along that trying to get testing absolutely right was too hard. The objective here was to try to clear out some of those cases that make your eyes bleed, where they write in their profiles "Ich bin der Deutsch Mutterspracher".
| | | | Samuel Murray
Netherlands
Local time: 18:02
Member (2006)
English to Afrikaans
+ ...
| My take on native language verification, @Phil | Sep 10, 2012 |
Phil Hand wrote:
Those opposed to verification - Kay, Jose, Samuel, Siegfried. You've all talked about other things being more important than native language. But are any of those other things remotely testable? Can they be usefully measured?
Phil, I'm not opposed to verification per se, and certainly not to native language verification in principle, but I acknowledge that useful and usefully accurate native language verification is an unrealistic goal for us in the current situation, even with ten times as much resources.
While it is probably simple to determine if a language is NOT someone's native language, determining beyond reasonable doubt that a language IS his native language is far more complicated, particularly since many good translators who are fluent in their second languages would likely test positive for nativeness, if they were to submit themselves.
I know I'm repeating what has been said countless times, but I think the question isn't new either. So, in an attempt to get something new out of this discussion (with you), allow me to comment on one of your more recent posts:
Phil Hand wrote:
The point in knowing a translator's native language is that quality is inherently unknowable (the service hasn't been performed yet), and so the client is using some proxies to limit their risk.
This is actually very similar to what I said recently in what might have been regarded by some here as a cop-out post, namely that the purpose of the "native language" option in a search is not to ensure beyond all doubt that the translator is a native speaker, but to reduce the candidates to put through unnecessary pre-screening. Okay, this isn't exactly what you're saying, but it is close enough for my liking.
You are trying to use "native language" as a means to reduce risk. Risk of what? Risk of a bad translation. So what you're really interested in is a way to reduce risk of a bad translation (the native language option is just a conveniently available option that would accomplish that).
If I understand correctly (let me speculate), you would be happy with a system that prevents the 10%-20% worst offenders (who claim to be native speakers but in reality can't even use the language properly) from being included in the pool of candidates for jobs. Is that right? Some of these bad offenders have beautifully complete profile pages and professional looking web sites and résumés, but they simply aren't qualified to do the translations, even at a basic level. Some even have a string of good WWAs (since WWAs are not language-specific). If we can filter out just the bottom 20% of dishonest translators most likely to fool innocent clients, that would really be something, wouldn't it?
While most of the worst translators will fail a native language verification, the worst of the worst will certainly fail an a-native language verification too. So let's introduce an a-native language verification... to catch the worst of the worst.
Suggestion for global a-native language verification
Here's my suggestion, and perhaps we can take this to another thread if there is any support for it: Allow all translators to create profiles and use the site features, but disallow translators from bidding or being found in the search results if they haven't passed a basic language skill test.
I think I suggested something very similar to this before, but heck, I might as well try again to drum up some support for it.
If you can design a test with 500-1000 basic English language questions, and require every English target (and source?) translator to do a random selection of e.g. 50 questions from it, that'll go a along way towards screening anyone who truly can't operate in that language. Questions should be simple to answer, so that the whole process can be automated. I'm talking basic grammar, syntax, spelling, ambiguity, etc.
In fact, make this an annual test. Make the pass mark 40/50. If a translator scores less than 30 points, he fails for that year (although he can lodge an appeal for extra chances). If he scores between 30 and 40, he gets an extra 20 questions and must get 15 of them right, to pass. If a translator fails for one language but passes for another, he can bid and be found in searches for the language in which he passed.
A committee consisting of red pee members plus highly ranked translators in each popular language can come up with such a test for each language with many candidates in it.
You don't even have to have many measures to prevent cheating, because malicious cheaters will always find a way to beat the system, but people who think that they can translate but really can't are the ones that this test will catch.
Look, this may not be native language verification, but it does fit the bill -- it gets you what you've been aiming to get.
Samuel
| | | |
Phil Hand wrote:
Proz is run for us translators,.....
You are absolutely right - and Google is run to provide a search engine to its users.
Proz is run to earn money. It is a business.
[Edited at 2012-09-10 17:16 GMT]
| | |
|
|
|
Phil Hand
China
Local time: 00:02
Chinese to English
| Worth some thought | Sep 10, 2012 |
I could go with everything you've written there, with one big worry.
Samuel Murray wrote:
useful and usefully accurate native language verification is an unrealistic goal for us in the current situation, even with ten times as much resources.
While it is probably simple to determine if a language is NOT someone's native language, determining beyond reasonable doubt that a language IS his native language is far more complicated,
Yes.
the purpose of the "native language" option in a search is not to ensure beyond all doubt that the translator is a native speaker, but to reduce the candidates to put through unnecessary pre-screening. Okay, this isn't exactly what you're saying, but it is close enough for my liking.
Yes, close enough.
what you're really interested in is a way to reduce risk of a bad translation
Yes.
you would be happy with a system that prevents the 10%-20% worst offenders (who claim to be native speakers but in reality can't even use the language properly)...If we can filter out just the bottom 20% of dishonest translators most likely to fool innocent clients, that would really be something, wouldn't it?
Yes.
If you can design a test with 500-1000 basic English language questions, and require every English target (and source?) translator to do a random selection of e.g. 50 questions from it, that'll go a along way towards screening anyone who truly can't operate in that language. Questions should be simple to answer, so that the whole process can be automated. I'm talking basic grammar, syntax, spelling, ambiguity, etc.
In fact, make this an annual test.
Put that way, I quite like the idea. Like you say, it would accomplish the objective.
The problem is this: a system like that would be likely to become in practice the standard for nativeness. Once people realised that all you have to do to be allowed to call yourself native is to pass the grammar test, there would be no incentive to be honest. Anyone who passes would almost certainly put the languages they pass in up as native.
This is true of all tests, to some extent, but a repeating test is particularly susceptible. Rather than being an approximate check on a concept which people just about understand (even if they understand it differently), the test can come to replace the concept.
So that would be the challenge. How to frame the test so that it didn't end up driving out the native concept. Perhaps change the timing: when you sign up, you self declare your native language(s). Then you immediately have to take the test for those languages (within the next day or two, let's say). These tests are non-repeating, and you don't get a warning about them before you sign up.
| | | | Phil Hand
China
Local time: 00:02
Chinese to English
| I'm the paying customer | Sep 10, 2012 |
Siegfried Armbruster wrote:
Proz is run to earn money. It is a business.
My money! I'm a paying customer. They make more from memberships than they do from advertising.
| | | | Samuel Murray
Netherlands
Local time: 18:02
Member (2006)
English to Afrikaans
+ ...
| Would you be prepared to separate it from nativeness? | Sep 10, 2012 |
Phil Hand wrote:
Put that way, I quite like the idea. Like you say, it would accomplish the objective.
The problem is this: a system like that would be likely to become in practice the standard for nativeness. Once people realised that all you have to do to be allowed to call yourself native is to pass the grammar test...
I think you misunderstood me. My argument was that since the issue of native language is simply a means to an end, and not the end in itself, any mechanism that accomplishes the same end does not need to include the original means to that same end.
To put it more simply, this proposed test is not a nativeness test. Whatever ProZ.com eventually decides to do with the "native language" option is not relevant to the test idea. A pass or a fail on the test will not have any bearing on the eligibility for a "native language" credential.
Only those who pass the test are included in search results for the directory, for jobs, and possibly also for KudoZ. Those who do not pass the test are allowed to maintain a profile page, participate in the forums, use the Blue Board (if they are paying members), and participate in KudoZ (through browsing (?)), etc, but are not findable in any kind of directory search (except by name).
| | | | Phil Hand
China
Local time: 00:02
Chinese to English
Samuel Murray wrote:
To put it more simply, this proposed test is not a nativeness test. Whatever ProZ.com eventually decides to do with the "native language" option is not relevant to the test idea. A pass or a fail on the test will not have any bearing on the eligibility for a "native language" credential.
Only those who pass the test are included in search results for the directory, for jobs, and possibly also for KudoZ. Those who do not pass the test are allowed to maintain a profile page, participate in the forums, use the Blue Board (if they are paying members), and participate in KudoZ (through browsing (?)), etc, but are not findable in any kind of directory search (except by name).
So this test in language X would determine whether you're allowed to put down language X as a target language (or whether you're allowed to show up in the directory as translating into that language).
There is a missing link in there, which is that your native language would have to be constrained to be among your target languages, but that's presumably true for the vast majority of people anyway, so not a major concern.
It sounds OK to me. I think you'd run into a lot of opposition from people who don't fancy doing grammar tests, but I think this system would solve the problem I want to solve.
In my pair, it would also force a lot of people to stop offering English as a target. I have no problem with that, but I suspect they would.
| | |
|
|
|
| Isn't this what the CPN (aka PRO-tag) is all about? | Sep 10, 2012 |
Samuel Murray wrote:
I think you misunderstood me. My argument was that since the issue of native language is simply a means to an end, and not the end in itself, any mechanism that accomplishes the same end does not need to include the original means to that same end.
To put it more simply, this proposed test is not a nativeness test. Whatever ProZ.com eventually decides to do with the "native language" option is not relevant to the test idea. A pass or a fail on the test will not have any bearing on the eligibility for a "native language" credential.
Only those who pass the test are included in search results for the directory, for jobs, and possibly also for KudoZ. Those who do not pass the test are allowed to maintain a profile page, participate in the forums, use the Blue Board (if they are paying members), and participate in KudoZ (through browsing (?)), etc, but are not findable in any kind of directory search (except by name).
As I understand it, the PRO-tag program is exactly this: a translator submits a sample of their work, and gets assessed by their peers in that pair. This is certainly beyond a self-asserted native-speakerness, if taken seriously.
Maybe some false natives are hiring native Cyranos to play their role, I wouldn't know.
Yet I wonder what's the point of self-asserting on something an individual can't do. Are my ethics that naive? This will enable them to serve each client... once! And after the word spreads, they wont't have anyone willing to hire them.
Just branstorming... How about a translator profile saying, "This translator allows publishing negative LWAs as well." If they have none, nobody will care if they are natives from outer space.
Before anyone asks, yes, I would let my past clients publish negative LWAs on me, if any of them actually had any complaint on my performance that would hold water.
[Edited at 2012-09-10 18:55 GMT]
| | | | Michele Fauble
United States
Local time: 09:02
Member (2006)
Norwegian to English
+ ...
| Relevance of test | Sep 10, 2012 |
Samuel Murray wrote:
To put it more simply, this proposed test is not a nativeness test.
In other words of no relevance or help to any client searching for a target language native speaker so that the client can be assured that the translation will be into a native version of the target language.
| | | | Jeff Whittaker
United States
Local time: 12:02
Member (2002)
Spanish to English
+ ...
| Times have changed | Sep 10, 2012 |
I am old enough to remember when agencies would call translators on the phone in order to conduct an interview in your native language and sometimes in your source language(s). They were interested in finding people for the long term and not just for the project du jour. I also remember that a lot of agencies refused to consider you unless you had a minimum of ten years of experience.
| | | | Samuel Murray
Netherlands
Local time: 18:02
Member (2006)
English to Afrikaans
+ ...
Michele Fauble wrote: Samuel Murray wrote:
To put it more simply, this proposed test is not a nativeness test. In other words of no relevance or help to any client searching for a target language native speaker so that the client can be assured that the translation will be into a native version of the target language.
No, the discussion between Phil and me was a sub-thread within the larger thread to try to solve a particular problem with a particular solution.
What is a problem in some languages is not a problem in other languages, and consequently the solutions that apply to some languages do not apply to other languages. This was perhaps one value of this thread, to show and share how although most languages have a problem with native language declaration, they don't all have the same problem with it.
The suggestion on the table in this sub-thread won't help a client find a native speaker, but it will ensure that he won't find non-natives who are so terrible that they shouldn't even be translators.
The type of test I have in mind is not one designed to catch non-natives, but to catch people whose knowledge and/or skill is below a certain basic threshold. The purpose of the test is not to aid those who are religious about native language.
I have met a few translators in my own native language whose nativeness (according to the linguistic definitions that you so dearly love) is not in question, and who can speak their native language fluently, but whose written language is so bad that I would not have been surprised to be told that their translations have to be redone from scratch every time. The proposed test in this sub-thread would weed out such translators, whether they are natives or not. The only possible relevance that this has for the native language issue is that many such translators who fraudulantly claim to be natives in that language would be weeded out also.
Samuel
| | |
|
|
|
Michele Fauble
United States
Local time: 09:02
Member (2006)
Norwegian to English
+ ...
| Eliminating risk | Sep 10, 2012 |
Samuel Murray wrote:
You are trying to use "native language" as a means to reduce risk. Risk of what? Risk of a bad translation. So what you're really interested in is a way to reduce risk of a bad translation (the native language option is just a conveniently available option that would accomplish that).
No, not risk of a bad translation. Native language alone can't do that. Eliminate risk of a bad translation that is bad because the translator is unable to produce native output, i.e. written language that reads as if it was written by a native speaker.
| | | | | You are either for verification or against it | Sep 10, 2012 |
Ty Kendall wrote:
Like it or not, the native language criterion is relevant to many outsourcers, and from what can be inferred from the faqs and other snippets ProZ seems to agree and a sizeable portion of translators here agree too. Notice I said "relevant" and not "an indicator of competence or quality in translation".
I would never support having the native language filter removed.
But if the filter continues to include everyone who claims to be a native speaker at their whim, it's really time for me to leave Proz.com. It'll be a simple decision.
B
[Edited at 2012-09-10 23:42 GMT]
| | | | | 1) Numbers and 2) People getting what they've paid for | Sep 11, 2012 |
Samuel Murray wrote:
Allow all translators to create profiles and use the site features, but disallow translators from bidding or being found in the search results if they haven't passed a basic language skill test.
People who have already paid membership of the site, sometimes more than a year in advance, are probably gonna take a pretty dim view of either having part of that service withdrawn or having to jump through extra hoops to get that service, all because other people are misrepresenting themselves. It is fundamentally unfair as an approach.
In fact, make this an annual test.
Universal schemes have the intrinsic drawback of the sheer amount of effort involved, and now you want to make it annual? Even disregarding the first objection, and even allowing for the fact that some members might be quite happy not to be allowed to bid or be found, you're still likely to be administrating thousands or tests. Every year.
Whatever else we consider as options, adverse impacts on the silent, and indeed honest, majority have to be avoided, in my view, and I would bet considerably more than an annual membership fee that site management would hold a similar view.
I still feel the focus should just be on weeding out the wrong 'uns (defined by quality of language output, as opposed to translation quality itself or accident of birth/education, etc.) and that in fact, if the site just enforced its own rules, that's all we need to start. Discussion of the mechanics of the testing/verification process, within that context (e.g. various testing ideas as proposed by various people in this thread) are all well and good, and some kind of punitive measure could be considered, but in essence, agreement to enforce the "no misrepresntation" rule has to come from the top first, since confirming others' representations is what it all boils down to, does it not?
Or, as was suggested earlier, the site could agree that, in fact, it's unable or unwilling to verify anything meaningfully, change the symbols and wording as appropriate, and run the place as a free-for-all (red Ps arguably excepted, depending on your opinion of the red P, natch - but frankly, the particular phenomenon discussed here has infected the red P population too, although how the hell that happened given the peer review process is utterly beyond me, unless conducted by people who can't tell a peer review from a pier revue).
Either way, we'd know where we stand.
| | | | | Pages in topic: < [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Should “native language” claims be verified? | TM-Town | Manage your TMs and Terms ... and boost your translation business
Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.
More info » |
| | CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
