[9fans] OT - Machine translation

[9fans] OT - Machine translation

[Andrew Simmons]

I think Boyd is being a bit harsh on machine translations. I don't know what
Google uses, but I feel that their translations into English achieve a
certain sublime poetry. As an example, the following is Google's translation
from Italian of the opening paragraph of an obituary of the English
philosopher G.E.M Anscombe. I don't see how even Mark V. Shaney could have
bettered this:

In our country it is known above all for being one of executory the
testamentary ones of Wittgenstein , of which also it has cured some of the
works. But it is not only to that G.E.M. Anscombe, than has extinguished to
ottantuno years slid five January, must its reputation in international the
philosophical community. Elisabeth Anscombe has been a philosopher
originates them, author of contributions that have influenced the
literature in various fields and that they still today continue to make to
discuss.

Re: [9fans] OT - Machine translation

[Boyd Roberts]

Andrew Simmons wrote:
> 
> I think Boyd is being a bit harsh on machine translations. I don't know what
> Google uses, but I feel that their translations into English achieve a
> certain sublime poetry.

Strangely enough google generates the same text as bablefish:

    Pourquoi est-ce que pas le babelfish d'utilisation
    (http://babelfish.altavista.com/) une sorte de version
    d'alpha d'une traduction obtiennent commencée, et vous
    pourrait la signaler, et les gens pourraient-ils s'amuser
    et soumettre alors les corrections et le document pourrait-elle
    être améliorée de là sur une base évolutionnaire?

Although it is interesting to speculate how the translation was done.

To quote from their translation FAQ:

    Unfortunately, today's most sophisticated software doesn't
    approach the fluency of a native speaker or possess the
    skill of a professional translator. Automatic translation
    is very difficult, as the meaning of words depends upon the
    context in which they are used. Because of this, accurate
    translation requires an understanding of context, as well
    as an understanding of the structure and rules of a language.
    While many engineers and linguists are working on the problem,
    it will be some time before anyone can offer a quick and seamless
    translation experience. In the interim, we hope the service
    we provide is useful for most purposes. 

I couldn't agree more.

Once the context is lost the semantic content has usually been destroyed.

Re: [9fans] OT - Machine translation

[Boyd Roberts]

plan9@itic.ca wrote:
> As an example, Russian requires a russian IT guy. Even a very youg one is ok
> !!! French requires someone that is particulary confident with both grammar
> and syntax since vocabulary is almost ``frenglish'', sometimes even english.

Yes, 'se logger' [to log in] springs to mind.

Having worked in Paris, in French, in IT for 6 years you most certainly need
to understand the peculiar set of jargon/terminology used.  It is less than
obvious.  And that is before you get to various idioms.  I would always get
a native speaker to proof my documentation to correct this.

Here's a fairly representative page:

    http://www-prima.imag.fr/legal/clone-pc.php

Or 'noyau' versus 'kernel' -- it just depends.

Get google to translate that page to get a good idea of the problem.

Re: [9fans] OT - Machine translation

[plan9]

Hi all,

Just to let you know about our experience of computer aided translations :

We have tried many of them... now we contract translators. The quality of
any translation fairly depends on 2 major factors : the first one is the
``style'' of the author. As soon as non regular syntax or idioms are used,
the translator process gets crazy. The second one is the lexical field. Some
translators are computing aware when others are not.

As an example, Russian requires a russian IT guy. Even a very youg one is ok
!!! French requires someone that is particulary confident with both grammar
and syntax since vocabulary is almost ``frenglish'', sometimes even english.
German has a very [excessively] structured syntax (software are almost good
for that) but you need a GermandictionnaryguythatspeekconstructedITwords :-)

Korean is the one I prefer :-) You need a korean guy, a korean dictionnary,
a US kbd, some korean pictures and some scotch tape... :-) The guy need to
type 4 keys to generate one Rune :-(  [and he only drinks tea]

In a very few words,

If you puting are that on you page translation were computated generates,
Not I am going to read her ...

----- Original Message -----
From: "Boyd Roberts" <boyd@strakt.com>
To: <9fans@cse.psu.edu>
Sent: Tuesday, May 14, 2002 4:35 AM
Subject: Re: [9fans] OT - Machine translation


Andrew Simmons wrote:
>
> I think Boyd is being a bit harsh on machine translations. I don't know
what
> Google uses, but I feel that their translations into English achieve a
> certain sublime poetry.

Strangely enough google generates the same text as bablefish:

    Pourquoi est-ce que pas le babelfish d'utilisation
    (http://babelfish.altavista.com/) une sorte de version
    d'alpha d'une traduction obtiennent commenc�e, et vous
    pourrait la signaler, et les gens pourraient-ils s'amuser
    et soumettre alors les corrections et le document pourrait-elle
    �tre am�lior�e de l� sur une base �volutionnaire?

Although it is interesting to speculate how the translation was done.

To quote from their translation FAQ:

    Unfortunately, today's most sophisticated software doesn't
    approach the fluency of a native speaker or possess the
    skill of a professional translator. Automatic translation
    is very difficult, as the meaning of words depends upon the
    context in which they are used. Because of this, accurate
    translation requires an understanding of context, as well
    as an understanding of the structure and rules of a language.
    While many engineers and linguists are working on the problem,
    it will be some time before anyone can offer a quick and seamless
    translation experience. In the interim, we hope the service
    we provide is useful for most purposes.

I couldn't agree more.

Once the context is lost the semantic content has usually been destroyed.

Re: [9fans] OT - Machine translation

[Andrew Simmons]

> we hope the service we provide is useful for most purposes.
>
I'm not sure the service is useful, but it certainly helps while away the
dull hours waiting for my CORBA-generated stubs or skeletons, or whatever
they are called, to compile.

Those of you with work to do should stop reading now.

Those of you who don't might enjoy the following, also from an Italian
obituary:


What remains therefore of Quine? The merit to have dissolved with reasonings
of an extraordinary thinness many philosophical distinctions of "principle"
in simple pragmatic issues of "degree" (merit that are been worth it a verbo
to its name in the divertentissimo and irriverente Philosophical Lexicon di
Daniel Dennett: " to quineare: to deny decidedly the existence or the
importance of something of real or meaningful"). Its radical naturalismo,
that it sees the task of the philosophy in continuity with the scopes of
science, that denies whichever possibility of "cosmic exile" from which
watching the world to outside of the theory that we have constructed in
order to observe it. The ability "to dissolve in ways to speak the mobilio
about our world", like a Prospero of the philosophy, declaring that "Being
is being the value of one variable" to the inside of the theory in which
impegnamo describing the world to us, and with defending one realistic
position. Its style of writing of rare elegance, the immense one corpus of
examples of enunciates to you philosophically problematic of which the
philosophy reviews are still sature. Its pragmatism and its refinement in
"ding-dramatize" the philosophical dichotomies that they made it to
conclude, in answer to the friend and master Rudolf Carnap: "the culture of
our fathers is weaving of enunciates to you. In our hands dumb it evolve and
[... ] E' one grey, black culture of facts and white woman of conventions.
But I have not found some reason substantial in order to conclude that there
are in it spins of all the black ones and others of the all white men ".