|
 |
New Google Translation languages | ||
Share with:
| ||
 Language Learning Forum : General discussion |  | |
| DaraghM Diglot Senior Member Ireland Joined 5931 days ago 1947 posts - 2923 votes  Speaks: English*, Spanish Studies: French, Russian, Hungarian | Message 1 of 9 25 February 2009 at 10:51am | IP Logged | ||
|
I noticed last week that Google seem to have added another couple of languages to their translation service at,
Google Translate They didn't have Hungarian before, but the results are not that impressive. E.g "The clothes are in the suitcase" Google Translate: "A ruhák a bőröndjét." My translation: "A ruhák a bőröndben vannak." It seems the translation engine doesn't know the Hungarian cases yet. I tried the same sentence last week, and it came up with "táska" for suitcase, so something is slowly improving. [EDIT - Corrected mistake. Thanks nó.] Edited by DaraghM on 25 February 2009 at 5:53pm 1 person has voted this message useful | |||
| Hollow Bilingual Triglot Senior Member United States luelinks.netRegistered users can see my Skype Name Joined 6319 days ago 179 posts - 186 votes Speaks: French*, English*, SpanishB2 Studies: Korean | Message 2 of 9 25 February 2009 at 11:18am | IP Logged | ||
|
WHats the difference between those two versions? I know nothing of Hungarian, just curious ^^
1 person has voted this message useful | |||
| Tyr Senior Member Sweden Joined 5562 days ago 316 posts - 384 votes Speaks: English* Studies: Swedish | Message 3 of 9 25 February 2009 at 11:26am | IP Logged | ||
|
Around my Finnish friend's place yesterday I was playing with it and shouting random things at him. None of them made any sense at all.
But then I remember 5 years back when babelfish was new and its French utterly sucked. These days you can get rather perfect French. Edited by Tyr on 25 February 2009 at 11:27am 1 person has voted this message useful | |||
 Iversen Super Polyglot Moderator Denmark berejst.dk Joined 6483 days ago 9078 posts - 16473 votes Speaks: Danish*, French, English, German, Italian, Spanish, Portuguese, Dutch, Swedish, Esperanto, Romanian, Catalan Studies: Afrikaans, Greek, Norwegian, Russian, Serbian, Icelandic, Latin, Irish, Lowland Scots, Indonesian, Polish, Croatian Personal Language Map | Message 4 of 9 25 February 2009 at 11:51am | IP Logged | ||
|
I'm looking forward also to see Latin, Esperanto, Icelandic and Afrikaans on the list.
It is amazing than anything like a translation can be made by a machine, and I would certainly like to know how it is done, partly that would shed a bit of light on the relevance - and future! - of linguistics as an academic discipline. For instance I am curious whether transformational grammar has left any mark on the translation software or not, - as far as I can see its effect on text books has been zilch nada zero, for good or for worse. The translations provided are far from being good enough to be used directly, but I sometimes use the translate facility if I am searching for a good way of expressing something and I can't find it in my dictionaries, - though I mostly have to try out several source languages and slightly different input. I also use it to get an idea about the content of homepages in languages that I don't understand. On top of that I have lately been trying a little game: I translate a whole page into Danish or English, and then I try to guess what the original version of each sentence was (it appears if you put the cursor over each sentence). It is like a small quiz, and because I do see the original sentence at the end it isn't too dangerous. With my weakest languages my guesses are sometimes quite far off the mark so there I also make a point of checking that I know all words and understand all constructions in the original version. So far it feels like a useful supplement to the use of regular bilingual texts, with the added bonus that I can choose almost any kind of content without having to search for or construct a bilingual text. However it is important to note that you have to use the translation of an original text in your target language, - the other way round you would be looking at rubbish in your target language much of the time. Edited by Iversen on 25 February 2009 at 12:44pm 1 person has voted this message useful | |||
| DaraghM Diglot Senior Member Ireland Joined 5931 days ago 1947 posts - 2923 votes Speaks: English*, Spanish Studies: French, Russian, Hungarian | Message 5 of 9 25 February 2009 at 3:17pm | IP Logged | ||
In my sentence the case ending -ben, signifies something is inside of something else, among other things. bőröndben - In the suitcase I've no idea what bőröndjét is meant to mean. Edited by DaraghM on 25 February 2009 at 3:19pm 1 person has voted this message useful | |||
| nó Tetraglot Newbie Germany Joined 5652 days ago 4 posts - 6 votes Speaks: German*, English, Czech, Hungarian Studies: Romanian | Message 6 of 9 25 February 2009 at 5:17pm | IP Logged | ||
"bőröndjét" is the accusative case of his/her/its suitcase, for example "Látom a bőröndjét", "I see his/her/its suitcase". By the way, you made a little mistake in your first translation. If I may correct you: "A ruhák a bőröndben vannak." 1 person has voted this message useful | |||
| William Camden Hexaglot Senior Member United Kingdom Joined 6052 days ago 1936 posts - 2333 votes Speaks: English*, German, Spanish, Russian, Turkish, French | Message 7 of 9 25 February 2009 at 6:24pm | IP Logged | ||
|
Turkish is an agglutinative non-Indo-European language, like Hungarian, though the two are apparently not related to each other. If I put short sentences into Google translation, it can more or less work, but you can also get it coming out as gibberish. The longer the text, the worse it is. Years ago, with translation engines, you invariably got it coming out as nonsense, so things have improved a bit but an engine that really works with languages like Turkish has not yet been developed, in my view.
1 person has voted this message useful | |||
| AlexL Diglot Senior Member United States Joined 6864 days ago 197 posts - 277 votes Speaks: English*, Spanish Studies: Italian | Message 8 of 9 25 February 2009 at 11:54pm | IP Logged | ||
|
Google translations are done using statistical machine translation. In other words, they do not handcode specific grammatical features of different languages into the program, but rather write one overarching algorithm that analyzes human-translated documents statistically and looks for patterns.
They use the same algorithm for every language; the only reason it takes a while to get a language up is because they need to acquire enough bilingual text to feed into the program. They also analyze a lot of monolingual text, for grammatical features of the language. 1 person has voted this message useful |
|  | 
Printable version
|
| Copyright 2024 FX Micheloud - All rights reserved No part of this website may be copied by any means without my written authorization. |
