|
 |
Forum Member with most language families | ||
Tags: Polyglot | ||
Share with:
| ||
 Language Learning Forum : Polyglots |  | |
| kidshomestunner Senior Member United Kingdom Joined 6403 days ago 239 posts - 285 votes  Speaks: Japanese | Message 17 of 41 28 September 2010 at 11:51pm | IP Logged | ||
They also have a great deal in common. 1 person has voted this message useful | |||
| Raчraч Ŋuɲa Triglot Senior Member New Zealand Joined 5816 days ago 154 posts - 233 votes Speaks: Bikol languages*, Tagalog, EnglishC1 Studies: Spanish, Russian, Japanese | Message 18 of 41 29 September 2010 at 9:26am | IP Logged | ||
Great! That's also my aim. I'm now studying C#, SQL and ASP at school and would like to create a translation software afterwards. The additional natlangs that I would like to learn for their special properties are Tongan (lack of noun/verb distinction) and Jingulu (non-configurational so noun and its adjectives can appear non-contiguous, like at both ends of the sentence). Out of which I would like to create a conlang that would serve as an intermediate language for translation. Good luck. Edited by Raчraч Ŋuɲa on 29 September 2010 at 10:09am 1 person has voted this message useful | |||
| Raчraч Ŋuɲa Triglot Senior Member New Zealand Joined 5816 days ago 154 posts - 233 votes Speaks: Bikol languages*, Tagalog, EnglishC1 Studies: Spanish, Russian, Japanese | Message 19 of 41 29 September 2010 at 9:54am | IP Logged | ||
Yes, but syntax difference is even wider if from different language families, that's why I am a fan of those people. I'm sort of a masochist myself coz I like to choose languages that are far different from each other (Tongan, Jingulu, Bayso, Koyukon, Tucano, Yurak) except for the big European (English, German, Spanish, Russian) and Asian (Cantonese, Tamil, Arabic, Javanese) languages which I would like to learn for their culture. I know I don't have the time and resources to learn all of them but dreams are free. Every journey begins with the first step. I think I can, I think I can.......(LOL!) Edited by Raчraч Ŋuɲa on 29 September 2010 at 9:55am 1 person has voted this message useful | |||
| Sprachprofi Nonaglot Senior Member Germany learnlangs.comRegistered users can see my Skype Name Joined 6468 days ago 2608 posts - 4866 votes Speaks: German*, English, French, Esperanto, Greek, Mandarin, Latin, Dutch, Italian Studies: Spanish, Arabic (Written), Swahili, Indonesian, Japanese, Modern Hebrew, Portuguese | Message 20 of 41 29 September 2010 at 10:46am | IP Logged | ||
Latin also has lots of non-contiguous adjectives, especially in poetry. There are no rules for word order really, you may even see prepositions in the middle of the noun phrase they define, or conjunctions 2-3 words in from the beginning of the subclause. Also, word class distinctions are quite fuzzy in Chinese, some verbs being used as prepositions and the same word being used as an adjective, noun or verb... I must say that my criteria for learning a language does not involve if it has any features that other languages don't have - the weirdest languages, like Piraha, are those that have few study materials and likely won't see a machine translation program created for them. I just try to cover a good breadth of language families and pick a language from each, whichever is the most useful from that family or whichever calls out to me the most. As for creating a new conlang as an intermediate language, I'm not sure that will work. Keep in mind that you'll also need to be able to find native speakers of all target languages to create dictionaries for your language, and it would defeat the purpose to have them create e. g. a Malayalam-English dictionary and you translating that into a Malayalam-yourlang dictionary. For Esperanto a lot of such dictionaries have been created already and we have the contact details of Esperanto-speaking collaborators for all languages we're targeting in the long run. 1 person has voted this message useful | |||
| Raчraч Ŋuɲa Triglot Senior Member New Zealand Joined 5816 days ago 154 posts - 233 votes Speaks: Bikol languages*, Tagalog, EnglishC1 Studies: Spanish, Russian, Japanese | Message 21 of 41 30 September 2010 at 10:42am | IP Logged | ||
|
Thanks for the tip. Edited by Raчraч Ŋuɲa on 03 October 2010 at 12:09am 1 person has voted this message useful | |||
| Cainntear Pentaglot Senior Member Scotland linguafrankly.blogsp Joined 6009 days ago 4399 posts - 7687 votes Speaks: Lowland Scots, English*, French, Spanish, Scottish Gaelic Studies: Catalan, Italian, German, Irish, Welsh | Message 22 of 41 01 October 2010 at 11:36am | IP Logged | ||
So is this fully procedural machine translation then? Wouldn't you be better with a machine-specialised interlingua? To me it seems that Esperanto would lose a bit too much information if you were translating between (eg) two languages with singular-dual-plural pronoun systems. 1 person has voted this message useful | |||
| Sprachprofi Nonaglot Senior Member Germany learnlangs.comRegistered users can see my Skype Name Joined 6468 days ago 2608 posts - 4866 votes Speaks: German*, English, French, Esperanto, Greek, Mandarin, Latin, Dutch, Italian Studies: Spanish, Arabic (Written), Swahili, Indonesian, Japanese, Modern Hebrew, Portuguese | Message 23 of 41 01 October 2010 at 1:32pm | IP Logged | ||
|
Using Esperanto as an intermediary language is not the same as sticking everything into
a machine translator, translating it to Esperanto and then sticking the Esperanto into a machine translator and translating everything to another language. I am keeping as much information as possible about the original in the background, so that e. g. the dual will still come out again in languages that have it, I'm just using Esperanto to add more information density if you want. For example, some Romance languages only have one word for "warm" and "hot". When translating English "hot" to such a language, the English-Esperanto dictionary gives "varmega" for "hot" vs. "varma" for "warm". "varmega" however includes the suffix -eg- which amplifies a meaning. If the Italian dictionary doesn't have an entry for "varmega", my program will use the standard transformation x-ega -> ege x and wind up translating "hot" as "very warm". Same if a language's vocabulary is not very big and we don't have a word for e. g. "forgetful". Forgetful is "forgesema" in Esperanto, forgesi (to forget) and -ema (often does), so transform that internally to "kiu emas forgesi" and translate as "who often forgets". This also helps with sentence structure. Some languages use verbs in favour of adjectives. Just enter them into the dictionary like that, e. g. the verb "beli" instead of "bela" (beautiful) and if for English we only have the entry "bela" but some language used "beli" then the program will know to transform "beli" to "esti bela" in order to come up with a translation. These kinds of shifts are sometimes possible in English and other languages, but they are not 100% reliable, not even 70% reliable, so machine translation programs couldn't draw on them yet. In response some systems have tried to create huge databases of how things relate to each other, not just grammatically but also in terms of world knowledge, but these databases can never be complete enough, and with Esperanto you don't need them. The beauty is that Esperanto can come up with words that English never even thought of, e. g. if some language's mythology involves a creature best translated as lightening-man, or a place like rain-place then this can be entered into the Esperanto dictionary without any problem, and all other dictionaries won't need to be updated with this culture-specific word, it will just be explain-translated based on the Esperanto. Edited by Sprachprofi on 01 October 2010 at 1:32pm 4 persons have voted this message useful | |||
| Doitsujin Diglot Senior Member Germany Joined 5318 days ago 1256 posts - 2363 votes Speaks: German*, English | Message 24 of 41 01 October 2010 at 2:13pm | IP Logged | ||
AFAIK, there have been several attempts to use Esperanto as an interlingua in the past and none of them lead to significant improvements in quality over traditional rule based MT systems. It seems that nowadays pretty much all major MT solution providers focus mainly on statistical MT translation solutions and provide mostly end-of-life support for existing rule based systems. I was therefore surprised to read that you're working on rule based system. Is this an academic project? 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. |
