Hacker News Reader

Comment by diskcat | Hacker News Reader

Wait what?

This doesn't sound right to me at all.

There is no arbitrary skill ceiling that can only be achieved by being a native speaker.

There is no arbitrary skill ceiling, but I think you underestimate how much long term immersion matters. I've lived in the UK fo 16 years. I started learning English at 7 and in school from I was 10, and used it daily from about 14 (I'm 41 this year).

Yet living in the UK I still regularly come across idioms and words that are second nature to people who grew up here that I haven't heard before.

E.g. I have a six year old son in primary school here, and through that now come into contact with expressions that people that grew up here experienced in primary school, and that most of them may not have used since then but are aware of, but that are totally new to me.

And I think this matters much more when translating into a target language than in the source language. While you may fail to recognise some idioms, it is easier to recognise something that sounds odd or words you don't know, than to recognise that you are missing vocabulary you didn't think was necessary (e.g. the word "mufti" was entirely new to me when my son started school; prior having him in school, if I was asked to translate anything about coming in out of school uniform, I'd have never known I was "missing" a word, and would have used a longer expression instead).

I can kind of see the logic. It's certainly possible for an adult to learn to speak another language indistinguishably from a native speaker, but in practice there's a surprisingly large gap between that and a person who can speak fluently on complex subjects with no fear of being misunderstood, but makes occasional minor errors--or even usages which are grammatically correct, but sound stilted--that betray them as non-native.

There's a big jump between 99.9% and 100%, is what I'm saying, and if you want a really professional translation job, you're not going to settle for 99.9%. And the thing is, you can't tell the difference between 99.9% fluent and 100% fluent without being 100% fluent. So if you want to be sure you're getting 100% fluency, a native speaker is your best bet.

Imagine a translator who is 50% fluent in source language and 100% fluent (native) in target language. He would often misunderstand the source text and, even though he can produce target language natively, the translation would be incorrect.

Now imagine a translator who is 100% fluent in source language and 50% fluent in target language. He completely understands the source text and, using only simple sentences, can accurately convey the meaning in the target language.

This thought experiment demonstrates that the knowledge of source language is far more important for the accuracy of translation.

i think ur parent assumed without saying that both translators were equally (well?) versed in the source language, so ur example does not compute.

Think about it this way.

Case 1: SRC->TGT (fluent) gives you a grammatically correct text that may have bits that are lost in translation. Misunderstood idioms and connotations. A reduction in nuance.

Case 2: (fluent) SRC->TGT gives you an incorrect text that contains errors of syntax and semantics, that doesn't use natural sentence construction. It'll reads like it was written by someone with less than a perfect grasp on the language, which will be true. They won't be able to accurately convey the meaning because they won't have an accurate grasp of the target language.

Also in case 1 you can consult expert quality reference texts to make up for your lack of knowledge in the source language. Not so in case 2.

Both of those people would be equally unacceptable as a professional translator. It's true that #2 would be better if those were your only two options, but that's an unrealistic limitation.

A guy who is 99% fluent in source and 100% fluent in target will generally produce a flawless translation--if he happens across a word or idiom in the source that he doesn't quite understand, it will stand out to him, and he will pause to research it before continuing.

A guy who is 100% fluent in source and 99% fluent in target will produce a perfectly clear but very slightly stilted translation, because he's not aware of the few mistakes that he makes.

If you're translating a commercial product and you have my two guys to choose between, you would want #1.

>A guy who is 100% fluent in source and 99% fluent in target will produce a perfectly clear but very slightly stilted translation,because he's not aware of the few mistakes that he makes.

Why isn't the reverse true? It is only that the lossyness of the conversion process is invisible in the result, so somebody reading it would not know that it's wrong, it still produces the same quality of translation i.e. that its not perfect (or as good as can be).

Additionally, there are so many words in english that it's not possible to know them all, and so one has to assume that all people are at most 99%. But my english alarm doesnt go off everytime a word like 'mellifluous' isn't used. So I even doubt the claim that a guy who is 99% in the target language will make a 'stilted' translation. There exists a skill level where a person can produce a text that won't read as 'wrong' but won't be as good as it could possibly be and this skill level is sufficient to not have a text feel stilted. Not to mention most people do not have the same writing skills, and a very good high school student writer writes a much more pleasing text than his schoolmate who is not very good.

Translating is often more than mechanically trying to convey the rough meaning. You want to capture the rhythm and flow and poetry (or lack of) in the original. You want to try to map idioms an word play from one language to another, and so and so forth. This requires a near-native level of understanding of the language you are translating to

I have a friend who is a translator and one of the hardest things for him in the beginning was not to act as a editor. If the original has a long, badly structured sentence with a misplaced comma, then you have to try to translate that into a long, badly structured sentence with a misplaced comma.

I think fluency in the target language is more important for the reasons PhasmaFelis and igravious say. Or to put it more bluntly, because understanding is generally an easier task than speaking/writing.

For example with my level of English (a lowish C2), I can understand all kinds of texts, including literary texts (I read lots of novels), but there is no way I can write English at a literary level. I would tell the facts of the story just fine, but the special "feeling" that good writers give to the text would be lost.

Curiously, this notion (that the best translator is one who is native or 100% fluent in the target language) is widely accepted in Western cultures, but someone told me that in China is the other way around. When they hire Chinese to English translators, they especially value their knowledge of Chinese, not English. Which may partially explain the "Engrish" phenomenon...

Well, if I wanted a Chinese text explained to me, I'd rather have a native Chinese explain it to me than somebody who is fluent in my own language and studied Chinese.

I see it all the time with translations from English to Russian. Because I know English better than most translators, I can spot all sorts of mistakes. I'm pretty sure a native English speaker, even with mediocre proficiency in Russian wouldn't have made those mistakes.

Translation also requires you to produce a style that is consistent with the source material. Professional translation is used to produce a similar quality document as the source material.

If you translate english legalese into Spanish, you expect Spanish legalese output.

Writing skills and specific domain experience is necessary. That is easier to achieve that level in your native language living in the target country.

Nowadays, translator are rarely used for "meaning translation" ( like translating journal article for the managers ) as there is always some foreigner guy speaking that language and able to "explain" rather than translate the material. Except for less common language like Hebrew, Japanese, ... or big governmental organisations.

BTW I think the requirement is reversed for interpretation. You need to be native of the source language, because you need to convey the meaning regardless of the form.

>but in practice there's a surprisingly large gap between that and a person who can speak fluently on complex subjects with no fear of being misunderstood, but makes occasional minor errors--or even usages which are grammatically correct, but sound stilted--that betray them as non-native.

Well the purpose of translation is communication, not hiding your nativeness, do I don't see the point of this.

>There's a big jump between 99.9% and 100%

Well the jump is between 0.1%

>and if you want a really professional translation job, you're not going to settle for 99.9%.

That's like saying if i wanted to hire an engineer to build me a nuclear reactor im gonna pass everybody who scored 99.9% on a test in favor of the magical unicorn who by virtue of being born in the beautiful blue glow of a reactor is the only being capable of scoring 100%.

Either the gap is actually bigger then 99.9-100 as you say, or the gap is so small as to be insignificant.

People can vary greatly in english skill if they are not native and while you can most likely figure out whether or not they were born where by their accents and their way of speaking, a sufficiently skilled speaker would give you no more trouble in communicating to you whatever complex ideas they might have in their mind.

The only reason I can think of of this restriction is protectionism, i.e. a translator union located in England would want to protect its members from competition and find it most convenient to claim a special status on nativeness without coming off as obviously anti-competitive.

You're correct, and that's not a universal restriction. But it is true that there are a lot of translation & interpretation jobs which do require that the translator only go into their native language, especially for large-scale operations that only need to communicate one-way (like simultaneous interpretation for speeches).

I used to be a DE>EN and ES>EN translator. Although there's no skill ceiling, it is still a professional etiquette among translators to translate into their native languages.

It seems odd to me too considering my native language is my weaker language. English is my second language but I flow better in it than I do Spanish... Odd. I'm better translating Spanish to English than the other way around, and yet still so many words lose me in Spanish so it gets harder for me.

I think the requirement is that ONE of the languages be the person's native language. Because language is so much more than just language, its also about culture. As another commentator pointed out , "cool" can mean so many things, which are in both similar and different at the same time. You need to understand culture to understand these differences.

Yes, English isn't my native tongue but I understand it just as well as a native would, but how many non-native languages can I have such a deep understanding of? I can translate from my native tongue to english and vice versa, but translating between two non-native languages perfectly seems like too difficult a task. Even if I am great at it, its possible that a native speaker of one of the lanugages would be better at it than me

Reply on news.ycombinator.com