This article was originally posted on WomenLearnThai.com.
Google Translate, I challenge you…
In my posts, Google Translates Documents and Email Too and Thai Google Translate: Will Crowdsourcing Work, we got to discussing how Google Translate takes online translation beyond Babel Fish by allowing users to correct bad translations.
In the comments, Jeff Gray put this bug in my ear:
…the sheer volume of people adding useful stuff to Wikipedia makes it immensely useful. Wikipedia is also inherently self correcting, because if someone writes rubbish, it will be corrected by others.
In the same way, having millions of users making minor improvements to the translation system does something that any single company could never do with internal resources only. It might be chaotic, but the sheer scale is unbeatable.
Time will tell, but the approach they’ve taken is potentially a very effective one.
Well, the time is now. From now until a year from now.
Yesterday, Aaron Handel was featured in WLTs new Interviewing Successful Thai Language Learners section.
Today, Aaron agreed to let me use a handful of his Thai sentences from his twitter account for a Google Translate Challenge.
As you can see, the results are interesting.
Be careful! There is swine flu!
Aaron Hadel: ระวังมีไข้หวัดหมู
Google: ระวัง! มีไข้สุกร!
I have swine flu already, thanks!
Aaron Hadel: ฉันมีไข้หวัดหมูแล้วค่ะ
Google: ฉันมีสุกรไข้หวัดใหญ่แล้วขอบคุณ!
I don’t know! Why do you ask?
Aaron Hadel: ไม่รู้ ซิถามทำไม
Google: ฉันไม่ทราบ!ทำไมเจ้าถาม?
Did you eat yet?
Aaron Hadel: กินข้าวแล้วหรือยัง
Google: คุณกินยัง?
Oh no! You’re a liar!
Aaron Hadel: ตายแล้วคุณเป็นคนโกหก
Google: แย่ละ!คุณเป็นคนพูดเท็จ!
I don’t want to see your face again.
Aaron Hadel: ฉันไม่อยากเห็นหน้าของคุณอีกแล้ว
Google: ฉันไม่ต้องการดูหน้าของคุณอีกครั้ง
He is busy lighting a mosquito coil.
Aaron Hadel: เขายุ่งจุดยากันยุง
Google: พระองค์คือยุ่งแสงสว่างที่ยุงม้วน
Don’t put any fish sauce on the rice. It stinks!
Aaron Hadel: อย่าใส่น้ำปลาบนข้าวมันเหม็น
Google: โปรดอย่าวางใดน้ำปลาใน ข้าว. มัน stinks!
Let the Google Translate Challenge begin!
As a sort of a control (so it does not become an issue) I have an equal number of Aaron’s sentences paired with Google Translate just waiting to go live next year.
So in one year I will put the sentences through Google Translate a second time, just to see if crowdsourcing translations really does work.
I agree with pgarcia. It doesn’t translate 100% correctly all the time, although it does do a good job – especially if one wants to translate a short excerpt of a document or something.
Andy, That is so priceless! We’ll have to start a file for Google Translate snafus.
Using Google translate to get a quick overview on the titles of announcements in the Royal Gazette, the following brilliant piece of mistranslation has to be shared:
เทศบัญญัติเทศบาลตำบลโป่ง เรื่อง การควบคุมกิจการที่เป็นอันตรายต่อสุขภาพ
Local government ordinance on the control of the nipple is a threat to health.
It should however be something like
Municipal law of subdistrict municipality Pong on the control of businesses harmful to the health
I use the translator all the time and I think it does a great job. It may not offer an exact translation but it allows me to reach my final goal ( and the final goal in almost any communication): to understand and be understood.
Jeff – Sarah Palin is such a… Ughhhh! I agree, those do come across as GT’s… She is totally disturbing and her latest appearences in the news makes me wonder just why so many people are behind her.
This is actually getting interesting as human translators are stepping up along with Google Translate – Best Translation Plugins for WordPress.
Forgive me for being off topic, but it made my day.
These two make speeches that are so bizarre, I suspect they have passed them though Google Translate via Chinese to Thai & back to English 😉
It’s even more disturbing to think they are both serious…
Yes, ‘pig’ is หมู not หมุ. It was a typographical error. I have corrected it in my twitter feed at http://twittercom/Thai_Language
Mashable has another post on Google Translate: HOW TO: Translate Your Tweets
So now we have even more ways to mess up in Thai 😀
But twitrans (no longer live) looks promising – ‘Translate Your Tweets to Any Language by Human Translators’.
Bangkok Bugle is doing a similar comparison. Seems a new Thai mag is using Google Translate for their English version, with the expected results.
Hi Jon, not yet. With Thai script being so teensy, it is easy to miss a typo.
I have a larger code to make the Thai larger, but I cannot use it with the rollovers.
And of course, on browsers we can always use keyboard shortcuts such as command + for Mac and such. I just didn’t think to check.
I agree with หมู – the ู is a longer sound (like the ‘oo’ in Rooney) where as ุ is shorter, like a ‘u’.
Catherine, did you get clarification?
This is fun. Google Translate converts ระวัง มี ไข้ หวัด หมู into ‘Look feverish cold pork’.
‘Shouldn’t ‘pig’ in Thai be หมู not หมุ?’
Let me ask Aaron as I do know pig to be หมู.
Some were so funny, I had to check them twice just to be sure 🙂
What was noticeable is that with Thai2English.com, you are given choices. For instance, with หน้า.
หน้า / nâa
[ N ] face ;
[ PREP ] in front of ; front
[ ADJ ] next ; following
[ CLAS ] page ;
[ N ] season ;
[ N ] surface ; side
[ ADJ ] top ;
But Google Translate makes this decision for us. In this case, they chose page (a classifier).
ฉัน ไม่ อยาก เห็น หน้า ของ คุณ อีก แล้ว
I do not want to see your page again.
Which should be…
I do not want to see your face again.
But if you are given a choice, you can correct something this simple by using common sense. Others, of course, will be harder. Especially when you have more than one word to fuss about.
I agree. Google Translate should not be easy to corrupt but I just didn’t want anyone to make it an issue. And since it did not mean that much more work (just hiding one set) I went with it.
That’s brilliant! As much fun as Engrish.com! I love “you are lying dead” 🙂
Hmm, has a way to go, doesn’t it?
You have quoted my reference to Wikipedia correctly. But I would point out that Google Translate isn’t as easily corrupted by malicious or incompetent entries. The statistical model looks at patterns from multiple sources. So it needs a big volume of English to Thai professionally translated text to feed it’s statistical brain.
This will be fun to see…
Shouldn’t ‘pig’ in Thai be หมู not หมุ? Or is หมุ a correct alternative? สุกร is correct for ‘swine’.