[Corpora-List] Constitution

TadPiotr tadpiotr at plusnet.pl
Mon May 16 09:00:01 CEST 2005

Let me even generalize it a bit: the real problem in translation is with
collocations, to which compounds belong. Dictionaries -- and databases -- of
technical terms focus on lexicalized items, but not on the gray area of
collocations. There are very few dictionaries of technical collocations. And
that is one of the reasons why translations of technical texts are so
Tadeusz Piotrowski

> -----Original Message-----

> From: owner-corpora at lists.uib.no

> [mailto:owner-corpora at lists.uib.no] On Behalf Of John F. Sowa

> Sent: Monday, May 16, 2005 12:49 AM

> To: Lou Burnard


> Subject: Re: [Corpora-List] Constitution


> Lou,


> I was very serious when I compared the issues in EU

> terminology to the terms used in commercial businesses.


> > I hope that this is meant at least partially as a tongue

> > in cheek analogy! As a European citizen I would be >

> really depressed to learn that all the terminology in > the

> constitution was entirely constructed as a kind > of

> brand-identification.


> I wasn't thinking of advertising problems, but of the

> problems of internationalization (or i18n as they abbreviate

> it) of software programs and documentation.


> > Am I hopelessly naive in thinking that some at least of >

> the terms relate to concepts that do actually exist > in all

> the languages of the EU?


> I am certainly not an expert on EU legal terminology, but the

> most difficult challenges are not with the translation of

> single words, but of multiword compounds.


> One early Russian to English machine translation system, for

> example, would routinely produce the output "nuclear

> waterfalls" instead of "nuclear cascades".

> The physicists had no trouble interpreting the output, and

> they even joked about it.


> That system by the way was the Georgetown Automatic

> Translator (GAT), which evolved into Systran, which is still

> in use today.

> The free Babelfish service on the WWW is powered by Systran,

> whose longevity is not due to the sophistication of its

> theoretical foundation (on which research was stopped in

> 1963), but to its now very large terminologies for handling

> multiword phrases in many different fields of specialization.


> John Sowa



More information about the Corpora-archive mailing list