| Pages in topic: < [1 2 3 4 5 6 7 8 9 10] > | Methods for verifying "native language" claims Thread poster: psicutrinius
|
|---|
Samuel Murray 
Netherlands
Local time: 14:21
Member (2006)
English to Afrikaans
+ ...
| Agree with Neil | Aug 8, 2012 |
Neil Coffey wrote:
A linguist may well still say that the person's "native" language is still L1, and that they have undergone a process of "language attrition". Under my definition, I'm saying that for the purposes of a translation client, that's not what they have in mind by "native language".
This is a very important point. When a client seeks a native speaker translator, what he is really looking for is a tranlsator who can process the language and output a result that is consistent with *other typical native speakers*. The client has a typical (stereotypical) native speaker in mind, and wants the translator to be like that.
A client who wants a native English translator probably wants a translator whose language usage is similar to typical UK or US native speakers. Such a client's expectation has little to do with the linguistic or academic definition of "native language", and he would not consider his translator a native speaker of English unless the translator conforms to his idea about the typical English native speaker (i.e. a broadly UK/US cultural and linguistic background).
This has serious repercussions for the topic of this thread, namely how to verify it, because it would mean that one can't easily verify it with a simple set of questions to be answered truthfully.
| | | | Samuel Murray
Netherlands
Local time: 14:21
Member (2006)
English to Afrikaans
+ ...
Ambrose Li wrote: Samuel Murray wrote:
Hey, I have a great idea about HOW TO VERIFY a person's native language. ... I hope this suggestion is on-topic. According to this test I would be a native English speaker.
Actually my post was only meant to show that the accusation aimed at the poster I quoted that she had not posted on-topic was ridiculous because her post was exactly on-topic. I don't think using the quoted question as the sole method of determining a native language would work.
| | | | | This is not necessarily true -- it all depends on the person | Aug 8, 2012 |
There are many children born in the United States whose parents speak a different language at home, and at the age of 8, let's say, they can neither be considered proficient speakers of English nor Spanish, just as an example. The government is doing everything to help those children to succeed and be able to receive eduction in one of those languages. With a lot of work, after some time many of those children become able to benefit from instruction just in one language. All the testing is done ... See more There are many children born in the United States whose parents speak a different language at home, and at the age of 8, let's say, they can neither be considered proficient speakers of English nor Spanish, just as an example. The government is doing everything to help those children to succeed and be able to receive eduction in one of those languages. With a lot of work, after some time many of those children become able to benefit from instruction just in one language. All the testing is done to help those children to progress in life -- not to prejudice against them. I don't know if they will ever achieve the language proficiency level of monolingual people from highly educated backgrounds, yet they become natives speakers of the languages they might not have been considered to be before. ▲ Collapse
| | | | Samuel Murray
Netherlands
Local time: 14:21
Member (2006)
English to Afrikaans
+ ...
| Can the native claimer spot other natives himself? | Aug 8, 2012 |
G'day everyone
I think it is assumed by many here that natives can spot other natives, and that they can also spot non-natives.
Let me repeat: the assumption is that natives can spot non-natives. Well, if that is so, why not test candidates by checking whether they themselves are able to spot other natives?
Most verification methods here involve direct interaction between t... See more G'day everyone
I think it is assumed by many here that natives can spot other natives, and that they can also spot non-natives.
Let me repeat: the assumption is that natives can spot non-natives. Well, if that is so, why not test candidates by checking whether they themselves are able to spot other natives?
Most verification methods here involve direct interaction between the candidate and his examiners. In other words, the verification is done by native speakers who judge the candidate. This is a practicality obstacle. Well, this proposal of mine requires very little personal interaction, yet is far stricter than any mere self-assessment questionnaire.
All you have to do, is to ask the candidate to judge whether written texts (or transcripts of spoken text, or short sound samples) were created by native or non-native speakers. Either the candidate is able to tell which text is which, or he isn't. The pass mark does not have to be 100% for this test, but it should be high.
Here's one possible implementation of such a test: give the candidate ten texts, half of which were written by native speakers and half by non-native speakers. The test is to tell which texts were written by whom. If the candidate is unable to tell which texts were written by non-natives, or considers too many native-written texts to have been written by non-natives, he fails the test.
Or, let there be two batches of ten texts each -- one with mostly native written texts, in which the candidate has to spot the three non-native writers, and one with mostly non-native written texts, in which the candidate has to spot the three native writers.
In an audio version of such a test, the candidate can listen to ten audio samples of e.g. 3 minutes each (which is randomly selected from a database of e.g. 100 recorded samples). For the written version of the test, the texts need not be long -- 300-500 word essays on general topics would do. Also, it would have to be decided if the written samples should be translations or specifically spontaneous text.
If there is truly truth in the idea that any reasonable native can spot a non-native, then there should be no objection to such a test.
That said, I think that many who claim to be native speakers in this thread would fail this type of test, but that is because I suspect natives can't be spotted without extensive testing, or luck.
Samuel
[Edited at 2012-08-08 10:15 GMT] ▲ Collapse
| | |
|
|
|
Samuel Murray wrote:
G'day everyone
I think it is assumed by many here that natives can spot other natives, and that they can also spot non-natives.
Let me repeat: the assumption is that natives can spot non-natives. Well, if that is so, why not test candidates by checking whether they themselves are able to spot other natives?
…
If there is truly truth in the idea that any reasonable native can spot a non-native, then there should be no objection to such a test.
That said: I think that many who claim to be native speakers in this thread would fail this type of test, but that is because I suspect natives can't be spotted without extensive testing, or luck.
While I won’t comment on such a test for the English language, I doubt this will work for Chinese.
A few months ago there was a question posted here in the forums (or maybe it was Kudoz, but probably the forums) asking native Chinese speakers to check if a certain piece of text was written by a native or a non-native. IIRC, most answered essentially “maybe” and it’s not because the answerers were not native speakers; it’s that there’s so much regional variation in the Chinese language that it’s very hard to spot non-nativeness unless we’re talking about egregious errors.
If you separated the language into its regional variants then, of course, things would be much less messy. But even so dialectisms exist (that are incomprehensible to people who do not share those dialectisms). It seems unbelievable to think that Chinese is the only language with this problem.
PS: If you asked me to check if a piece of text written by a native of China is native Chinese or not I would not be able to do it. In fact a lot of words that are correct in that regional variation seem wrong to me. According to this test probably no native Chinese speaker can call themself a native speaker.
| | | | | the purpose is to verify native speakers, nothing added | Aug 8, 2012 |
Ambrose Li wrote:
The key phrase from Lisa’s original post was “from earliest childhood.” There are people who came here at a later age but are for all practical purposes indistinguishable from native-born English speakers. One of my classmates in high school came from China at eight (so during his childhood, but not earliest childhood) and by the time he was in Grade 11 he could (according to himself) speak about 8 Chinese sentences. I once met someone who came here from Hong Kong in his late teens and by the time I knew about him all he spoke was English. To say these people are not at least equivalent to native English speakers (or worse, to say they are native Chinese speakers) is to stretch the imagination.
The purpose here is verifying "native" language proficiency.
Just because all someone speaks (anymore) is English does not automatically make him/her a native speaker.
On the other hand, as long as someone learned a language during the critical period (at least beginning to do so before age 16 - which is usually a cut-off point, but definitely before adulthood), they can indeed become native speakers of that language.
Late teens? Far more difficult if not impossible.
Ambrose Li wrote:
I do not believe these are isolated cases. Even if this is not as widespread as I suspect, narrowly defining native as having acquired the language from “earliest” childhood is to automatically strike these people out from the pool of speakers who for all practical purposes are native.
Again, I wouldn't call our goal here to define native speakers as "native speakers 'for all practical purposes'; the add-on "for all practical purposes" is redundant or, worse, inviting non-native variants. A native speaker is always a native speaker, and a non-native is always a non-native.
I wouldn't take anyone off the list of candidates if they didn't acquire the language from earliest childhood on, but I would be very surprised if someone became a native speaker when they began speaking it at age 17.
B
| | | | | making it easier to prevent abuse | Aug 8, 2012 |
Here is what I would suggest:
We do away altogether with the "unverified" native language display (the grayed-out icons) for two or more languages.
Instead, you claim one native language only.
You can still base receiving the "verified icon" on a questionnaire which checks the likelihood of that claim.
Also, you must have your identity verified and then receive the "verified language" icon.
If you want to display two native languages, y... See more Here is what I would suggest:
We do away altogether with the "unverified" native language display (the grayed-out icons) for two or more languages.
Instead, you claim one native language only.
You can still base receiving the "verified icon" on a questionnaire which checks the likelihood of that claim.
Also, you must have your identity verified and then receive the "verified language" icon.
If you want to display two native languages, you must get "verified", meaning other native speakers must have judged you as a native speaker/writer.
Only if you pass, you get a second, third etc. "verified" native language icon.
That would do away with mandatory testing for two or more languages.
Advantages:
far fewer people will want to claim more than one native language
Only real native speakers will be able to bid on jobs, be found in the directory search results and on profile pages
The verification process (for a second or third etc. language) is more manageable.
less of a chance of abuse
What do you think?
B
[Edited at 2012-08-08 14:39 GMT] ▲ Collapse
| | | | | Abusive won't be prevented, just made easier | Aug 8, 2012 |
Bernhard Sulzer wrote:
Here is what I would suggest:
We do away altogether with the "unverified" native language display (the grayed-out icons) for two or more languages.
Instead, you claim one native language only.
You can still base receiving the "verified icon" on a questionnaire which checks the likelihood of that claim.
What do you think?
B
So people will claim only 1 native language and make sure they claim the one they feel is most lucrative or 'easiest' to translate into, even if it means they eliminate their real native language. It won't stop anyone from lying. It'll just stop them from listing more than one native language. Imo
| | |
|
|
|
Neil Coffey
United Kingdom
Local time: 13:21
French to English
+ ...
| "practical" definition inevitable | Aug 8, 2012 |
Bernhard Sulzer wrote:
On the other hand, as long as someone learned a language during the critical period (at least beginning to do so before age 16 - which is usually a cut-off point, but definitely before adulthood), they can indeed become native speakers of that language.
For what it's worth (but as I say below I don't think we should get too theoretical), I think that's on the very late end of what most linguists would understand by "critical period". As I say, there is even some evidence of a "critical period" for phonology of the first 6 months... at least in terms of acquiring the language in the same way as a 'canonical' native speaker.
Though...
Ambrose Li wrote:
Again, I wouldn't call our goal here to define native speakers as "native speakers 'for all practical purposes'
I think we will inevitably end up defining "native speaker" in terms of our practical purposes of translation. There is no such thing as an undisputed linguistic definition, and what linguistic definitions there are probably won't prove entirely useful in terms of the expectations of most translation clients.
[Edited at 2012-08-08 15:11 GMT]
| | | | Phil Hand
China
Local time: 20:21
Chinese to English
| Charlie's proposal is still the most parsimonious | Aug 8, 2012 |
That we just lobby Proz to enforce the existing rules.
How can we demonstrate to the site that this is what we want?
A petition? A petition thread?
| | | | | we need to find a workable solution | Aug 8, 2012 |
writeaway wrote:
So people will claim only 1 native language and make sure they claim the one they feel is most lucrative or 'easiest' to translate into, even if it means they eliminate their real native language. It won't stop anyone from lying. It'll just stop them from listing more than one native language. Imo
I don't think this would lead to more abuse. I hope the opposite is true IF you tie bestowing the "verified icon" button to a serious checklist (as Janet has suggested) AND to identity verification. And make sure that these are "sworn" statements made upon applying.
One way to do it would be to have applicants (after they qualify for the "N") sign (electronically) a statement in which they swear that they are indeed native speakers. Now, you need a definition of native language, but we've already suggested definitions as well (see Oliver's contribution).
The checklist will ask you where you grew up, where you went to school and how long etc., to allow Proz.com staff to make a decision about whether you qualify as a native speaker or not.
I also think that psychologically, a person is more likely to NOT falsely claim a language as their native language if they can only declare one language.
And, the overwhelming majority of translators claiming only one language aren't liars I would say.
If you are rejected after your application for only ONE native language, you can get "verified" before native speakers, either by way of speaking or writing or whatever is deemed necessary for that particular language or the individual.
We need to install something that is doable!!!
If that's not enough, then have everybody go through verification with native speakers before they can be listed as native speakers. And you would have to do that retroactively for all of us who have one "N". So, liars will lose their one and only "N".
I believe most of the liars would be caught by retroactively having to fill out a questionnaire.
What is your suggestion to remedy the situation?
B
[Edited at 2012-08-08 18:38 GMT]
| | | |
Phil Hand wrote:
That we just lobby Proz to enforce the existing rules.
How can we demonstrate to the site that this is what we want?
A petition? A petition thread?
Hi Phil,
I asked Ambrose that but what would you say?
Is there a sure way to "verify" a native speaker/writer of Chinese?
Ambrose thinks it's almost impossible.
B
[Edited at 2012-08-08 20:33 GMT]
| | |
|
|
|
Phil Hand
China
Local time: 20:21
Chinese to English
| I think you shouldn't ask me | Aug 8, 2012 |
Native Chinese speakers are likely to have a much better idea of this than I do...
Having said that, I have just answered this very question on the other thread, and I'll give another answer here - Ambrose, I don't know if you've had much contact with learners of Chinese. I suspect that if you had more, you'd find the question less bothersome. Despite all the publicity Mark Rowswell gets, the vast majority of us are much more spottable!
| | | | | I suggest we need to keep it smple | Aug 8, 2012 |
Samuel Murray wrote:
Here's one possible implementation of such a test: give the candidate ten texts, half of which were written by native speakers and half by non-native speakers. The test is to tell which texts were written by whom. If the candidate is unable to tell which texts were written by non-natives, or considers too many native-written texts to have been written by non-natives, he fails the test.
Or, let there be two batches of ten texts each -- one with mostly native written texts, in which the candidate has to spot the three non-native writers, and one with mostly non-native written texts, in which the candidate has to spot the three native writers.
This might be a bit complicated. What if only one text is picked wrongly? And what would be the cut-off point for "considers too many native-written texts to have been written by non-natives"?
Besides that, a highly fluent non-native speaker could possibly pick them all correctly. Does that make him/her a native speaker? No.
Samuel Murray wrote:
In an audio version of such a test, the candidate can listen to ten audio samples of e.g. 3 minutes each (which is randomly selected from a database of e.g. 100 recorded samples). For the written version of the test, the texts need not be long -- 300-500 word essays on general topics would do. Also, it would have to be decided if the written samples should be translations or specifically spontaneous text.
Would you have the texts read by a native speaker with native accent? Or would you try to trick/trip up the candidate by having a native text read by a non-native?
Why not just have the candidate decide which "narrator" sounds "native" and which "doesn't", using 2 texts? Not that I would at all support such a test!
Recognizing nativeness in texts or speech does not make the candidate a native speaker/writer.
The candidate must speak or write something him-/herself and other native speakers must judge it as non-native or native.
Are you trying to prove it's impossible to check nativeness?
Samuel Murray wrote:
If there is truly truth in the idea that any reasonable native can spot a non-native, then there should be no objection to such a test.
As far as I'm concerned, that's true.
Samuel Murray wrote:
That said, I think that many who claim to be native speakers in this thread would fail this type of test, but that is because I suspect natives can't be spotted without extensive testing, or luck.
Well, if you give me 5 non-native texts and 5 native texts, all in German, I will probably guess them all correctly. But that still doesn't prove that I am a native speaker of German. I could be a highly fluent in German, used to reading mostly "native German" texts.
I don't think extensive testing is necessary. Native speakers will spot a "speaking" or "writing" non-native, especially if the non-native is clearly non-native even to more fluent non-natives.
In case of an extraordinarily competent native speaker, it's actually less likely she/he will want to claim "nativeness". They're too professional for that.
If they still come to be verified and pass, so be it.
B
[Edited at 2012-08-08 21:28 GMT]
| | | | | getting rid of "unverified" status | Aug 8, 2012 |
This is a copy of my comment from the other thread (page 100) where it was a reply but I believe it concerns the "how" to carry out verification:
from:
http://www.proz.com/forum/prozcom_suggestions/227485-should_“native_language”_claims_be_verified-page100.html
Katalin Horváth McClure wrote:
Having a voluntary verification process and having differentiation between verified and non-verified status across the site (including the directory search) may be a (or perhaps the only) realistic and practically feasible approach. After following the debate for weeks, I think if anything, this is the approach that has a realistic chance to be implemented.
Hi Katalin,
I wouldn't be happy with a solution that keeps "unverified native languages" forever if it is decided that "verification" is optional, for two or more languages.
Verified or not verified, as long as the languages are listed as "native languages", anybody can presume that the speaker is indeed a native speaker and nobody is forced to verify two or more languages. Therefore even those who falsely claim it will continue to bid on jobs, be listed as "natives" (albeit unverified) in the directory and in their profiles.
I propose(d) to simply get rid of the "unverified native language" definition, throughout the site. Everyone should be allowed to declare one language after he/she has passed a questionnaire-like evaluation (see Janet's proposal for a checklist) and declares in an electronically signed, sworn document (possibly displayed on his/her profile page) that they are indeed a native speaker of that language. In addition, their identity must be verified. Then they get the "verified language" stamp/icon.
That would be acceptable to me.
Only if they want to (retaining the optional character of future verification of two or more languages) declare additional native languages, they must come before native speakers of that language and prove it. If they pass, they will again sign a sworn native language declaration and receive a second "verified" icon. Before they are admitted to verification, they must fill out an application that lays out strict criteria for qualifying for verification.
Katalin Horváth McClure wrote:
Now, common sense logic and the discussion in this thread suggests that perhaps ALL native language claims should be treated the same way, i.e. they would not be automatically verified whether the declared native language is the only one or there is a second one declared as well, and in order to turn the grey "N" mark to yellow, the verification process should be performed. This would be taking the issue a bit further (but not much further, in my opinion). It would involve turning off all yellow "N"s, make them grey and including them in the optional verification process. I have no doubt that it would require more resources for the verification, but maybe it can be planned as the next milestone after the first milestone (i.e. verification for two native languages) is reached.
I would simply ask single-native-language speakers who already have the "verified" icon to fill out the questionnaire and sign the sworn document. That should really be enough. If there is any doubt, the questionnaire should bring out these doubts and possibly subject them to either changing their native language or losing any right to display native languages.
I believe all these things are doable and make the whole process easier.
B
[Edited at 2012-08-08 22:53 GMT]
| | | | | Pages in topic: < [1 2 3 4 5 6 7 8 9 10] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Methods for verifying "native language" claims | 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! » |
| | Protemos translation business management system | Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!
The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
