While there have been many variants, most MT
systems--and certainly those that have found practical
application--have parts that can be named for the
chapters in a linguistics text book.

They have lexical, morphological, syntactic, and
possibly semantic components--one for each language
being processed--for treating basic words, complex
words, sentences, and meanings.

Each one of these components transmits to the next one
a very abstract representation of the statements
produced by prior components.

Some systems make use of a so-called "interlingua" or
transition language, in which case the translation
process is divided into two parts, one translating a
statement into the interlingua and the other
translating the result of this processing into the
target language.

Automatic translation systems up to now have not been
able to make significant use of a general semantic
knowledge.  Up to a certain point, they have been able
to compensate for this inability through statistical
algorithms in artificial-intelligence engines.

---

Ben que il ha habite multe variantes, le major parte
del systemas de traduction electronic--e certemente
illos que ha trovate qualque application practic--ha
partes que pote trovar se como capitulos in un libro
de instruction sur le linguistica.

Illos ha componentes lexic, morphologic, syntactic e
possibilemente semantic--un pro le duo linguas que se
processa--, le quales tracta parolas basic, parolas
complexe, asseverationes, e significantias.

Cata uno ex iste componentes transmitte al proxime un
representation multo abstracte del asseverationes
producite per componentes anterior.

Alicun systemas usa un "interlingua", o sia un lingua
de transition, in le qual le processo de traduction se
divide in duo partes, le prime que traduce un
asseveration al interlingua e le altere que traduce le
resultato de iste processamento a in le lingua al qual
le systema traduce.

Usque nunc systemas de traduction automatic non ha
potite usar significantemente cognoscimentos semantic
general.  Usque un certe puncto, illos ha potite
compensar pro iste incapacitate usante algorithmos
statistic in motores de intelligentia artificial.

(Adaptate ex information in <http://www.lsadc.org>)


__________________________________
Do you Yahoo!?
Yahoo! Hotjobs: Enter the "Signing Bonus" Sweepstakes
http://hotjobs.sweepstakes.yahoo.com/signingbonus