Re: New Machine Translation Project

On 4/10/06, Ph.D. <phil@...> wrote:
> Paul Bennett wrote:
> >
> > Ph.D. wrote:
> >
> > > Anyway, I thought perhaps someone here might be
> > > interested.
> > >
> > > http://unikom.org/
> >
> > I love the reason they'll succeed where others have
> > failed:
> >
> > "We'll cope with the stuff MT doesn't do well ... by ...
> > uhh ... MAKING A HUMAN DO IT! Yeah, that's the
> > ticket!"
Yes.... this bit was particularly silly:

>>UNIKOM deals with ambiguous words in the source language by
rejecting them, and requiring the substitution of unambiguous ones.
Good grammar and correct spelling are necessary.
<<

Which only works reliably if the original author of the text
to be translated is available for consultation.  Good luck.

In _Le Defi des Langues_, Claude Piron has a long section
about machine translation, in which he concludes that
the best machine translation would save labor only in the easiest
part of the human translator's work.  90% of a professional
translator's time, he says, is used in research to resolve
ambiguities or learn background knowledge that the
original author took for granted.  Usually a text to be
translated doesn't have in itself enough information
to allow translation without seeking additional information
elsewhere.


> Even better, is the way I found their web site. They
> are currently looking to hire a HUMAN to translate
> their web site into Esperanto, even though they claim
> they will soon be beta-testing their English-to-Esperanto
> software.
Yes, I have the  impression that the site author, at least,
is a beginner in Esperanto, and more enthusiastic than
knowledgable about it.   I can see the site's presentation
being convincing to someone who knows little or
nothing about Esperanto, little about linguistics, and
nothing about constructed languages other than
Esperanto.  Or maybe to someone fluent in Esperanto
but little knowledgable about linguistics and about
other constructed languages.

> > Also, one would think that the ideal medium for an
> > intermediate language would be one which is
> > specifically designed for the purpose (i.e. machine-
> > readable, self-segregating, and all the other things),
> > as opposed to one whose goal was (more or less)
> > to be readily human-learnable.
Such as, maybe, Rick Morneau's Ladekwa.  Or perhaps
more likely something machine-readable but not
human-readable?

--
Jim Henry
http://www.pobox.com/~jimhenry

Reply