Friday, February 26, 2010

A omne mi amicos qui cognosce solmente le anglese / To all my friends who know only English

(Languages of this post: Interlingua, English)

Io vole que omne mi amicos angloparlante qui non cognosce altere linguas que le major parte del messages de iste blog ha un version in anglese; e si vos non ha ulle interesse in studiar le altere linguas que appare in iste blog, vos pote vader directemente al fin de cata articulo, ubi vos pote trovar su version in anglese.

Le objectivo de “Interlingua multilingue” es adjuvar su lectores a apprender interlingua e su linguas fonte, e le version in anglese de iste articulos es destinate principalmente a studentes del anglese como un secunde o tertie lingua.

Le articulos de iste blog es sur themas que io trova interessante e relevante al mundo contemporanee. Io es convencite que le gente apprender linguas con un maximo de efficacitate quando illes pote travaliar con textos que es interessante e relevante a lor vitas.

Durante que le gruppo de linguas anglo-romanic ha multe differentias inter se, illos anque ha un nucleo structural e grammatic que es multo simile in omne illos. Durante que io redige textos pro “Interlingua multiliingue”, io essaya a dar emphase a iste nucleo. Si studentes de linguas pote controlar le nucleo de iste linguas, illes pote mover inter illos con un minimo de difficultate. Le comparation de mutle textos parallel equivalente es un maniera excellente de studiar linguas, e un studente pote apprender multo per le studio de iste tipo de textos.

Jean François Champolion poteva deciphrar le egyptiano ancian usante le petra Rosetta, que ha le mesme texto in hieroglyphicos, un scripto demotic, e Greco. Champolion habeva talentos exceptional como linguista e poteva apprender a leger egyptiano per un exposition limitate a un parve corpore de textos in egyptiano e Greco.

Pro studentes ordinari qui non ha le talentos exceptional de Champolion, le studio de linguas per textos parallel deveni progressivemente plus facile si le numero de textos es grande. In altere parolas, il ha un proportion directe inter le numero de textos parallel e lor utilitate pedagogic.

A iste momento, io volerea facer unes breve commentarios sur le traduction electronic e systemas informatic que pote recognoscer le linguage parlate e converter lo a in texto. Io mesme, como le major parte de nos, darea le ben venite a motores de traduction que functiona tan ben como traductores human. On poterea usar los como tutores multo patiente pro studentes human de linguas. Mesmo nunc, io ha trovate que le capacitates limitate del motor de traduction de Google es multo utile proque illos me sparnia tempore e labor durante que io redige le textos de “Interlingua multilingue”.

Nos nunc ha systemas informatic que es tan bon como stenographos commercial in le epocha ante que Internet deveniva un utilitate intellectual quasi universal. Io ha vidite demonstrationes in YouTube de Dragon Naturally Speaking, que ha versiones in anglese e espaniol que pare esser multo exacte e versatile. Con le temore, on publicara altere simile systemas pro altere linguas. Ante pauc tempore, on predeciva que tal systemas apparerea solmente post le anno 2050 a causa del grande numero de problemas que on deberea resolver ante que illos devenirea assatis potente pro uso general.

Le problemas de disveloppar systemas informatics pro traducer linguas es mesmo plus complicate. Traductores human qui es vermente bon, ultra deber dominar le vocabularios e syntaxe del linguas con le quales illes travalia, anque debe haber multe cognoscimentos cultural general durante que illes travalia pro producer traductiones que es vermente bon. A iste momento, il es difficile conciper como iste cognoscimentos pote incorporar se a in motores de traduction.

In despecto de iste difficultates tamen on non pote concluder in iste momento que disveloppar motores de traduction informatic con cognoscimentos simile es completemente impossibile. Le linguas human es systemas complicatissime, e on debera facer multe progresso in le intelligentia artificial ante que traductores electronic potera duplicar le capacitate de traductores human qui es vermente bon.

Io ha notate melioramentos progressive in le motor de traduction de Google in le menses que io lo ha usate, e io es confidente que illo continuara a meliorar se. Iste systema opera principalmente usante comparationes statistic de un ingente corpore de textos bilingue, e a illo manca un componente syntactic.

A causa de iste circumstantia, illo sovente face errores de detalio grammatic. Ma pro un persona qui ha un cognoscimento assatis bon del linguagas que ille vole traducer, le uso del motor de Google pote sparniar multe travalio, e io lo ha trovate multo utile in mi effortios pro preparar textos pro “Interlingua multilingue”.

Io es confidente que in le futuro systemas de traduction informatic essera mesmo plus utile pro linguistas qui vole preparar nove ressources pro personas qui vole apprender nove linguas e meliorar lor cognoscimentos del linguas que illes jam ha studiate. Intertanto, io es multo contente de usar le ressources que io trova utile in iste momento.

I want to say to all of my English-speaking friends who do not know other languages that most of the messages in this blog have a version in English; and if you are not interested in studying the other languages that appear in this blog, you can go directly to the end of each article, where you can find its English version.

The purpose of “Interlingua multilingue” is to help its readers learn Interlingua and its source languages, and the English version of these articles is directed primarily to students of English as a second or third language.

The articles in this blog are about subjects that I find interesting and relevant to today’s world. I am convinced that people learn languages with a maximum of effectiveness when they can work with texts that are interesting and relevant to their lives.

While the Anglo-Romance group of languages have many differences among them, they also have a structural and grammatical core that is very similar in all of them. As I edit texts for “Interlingua multilingue,” I try to emphasize this core. If language students can control the core of these languages, they can move among them with a minimum of difficulty. The comparison of a lot of equivalent or parallel texts is an excellent way of studying languages, and a student can learn a lot by the study of these kinds of texts.

Jean François Champlion was able to decipher ancient Egyption using the Rosetta stone, which has the same text in hieroglyphics, a demotic script, and Greek. Champolion had exceptional talents as a linguist and was able to learn to read Egyptian through a limited exposure to a small corpus of texts in Egyptian and Greek.

For ordinary students who do not have Champolion’s exceptional talents, the study of languages through parallel texts becomes progressively easier if the number of texts is large. In other words, there is a direct proportion between the number of parallel texts and their teaching usefulness.

Right now, I would like to comment briefly on electronic translation and computerized systems that can recognize spoken language and convert it into text. I myself, like most of us, would welcome translation engines that work as well as human translators. They could be used as very patient tutors for human students of languages. Even now I have found the limited capabilities of Google’s translation engine to be very useful because they save me time and labor as I edit the texts of “Interlingua multilingue.”

We now have computerized systems that are as good as office stenographers in the era before the Internet became a quasi-universal intellectual utility. I have seen demonstrations on YouTube of Dragon Naturally Speaking, which has versions in English and Spanish that appear to be very accurate and versatile. In time, other similar systems will be published for other languages. Not long ago, it was predicted that such systems would appear only after the year 2050 because of the great number of problems that would have to be solved before they would become powerful enough for general use.

The problems of developing computerized systems for translating languages are even more complicated. Really good human translators, besides having to dominate the vocabularies and syntax of the languages they work with, also have to have a lot of general cultural knowledge as they work to produce really good translations. At this moment, it is difficult to conceive how this knowledge can be built into translation engines.

Despite these difficulties, however, we cannot conclude right now that developing computerized translation engines with similar knowledge is completely impossible. Human languages are very complicated systems, and a lot of progress will have to be made in artificial intelligence before electronic translators will be able to duplicate the capabilities of really good human translators.

I have noted progressive improvements in Google’s translation engine in the months I have been using it, and I am sure that it will continue to improve. This system operates principally through statistical comparisons of huge corpora of bilingual texts, and it lacks a syntactic component.

Because of this circumstance, it often makes mistakes in grammatical detail. But for a person who has a pretty good knowledge of the languages he wants to translate, the use of Google’s engine can save a lot of work, and I have found it very useful in my efforts to prepare texts for “Interlingua multilingue.”

I am sure that in the future computerized translation systems will be even more useful for linguists who want to prepare new resources for people who want to learn new languages and improve their knowledge of the languages that they have already studied. In the meantime I am happy to use the resources that I find useful at this moment.

No comments: