[b-hebrew] The most difficult

James Christian jc.bhebrew at googlemail.com
Sat Jul 24 09:01:34 EDT 2010


Me too. This is evidently a very complex task. When people make what seem to
be subjective judgements of translation there is evidently a very complex
and objective neural process underlying their decision making process. I
would suggest that semantic fields associated with words has a lot to do
with it and that usage in certain contexts dictates which meaning of a word
was meant and therefore which translation would be judged best.

Of course, then we get into a play off with fidelity and fluency where the
two may at times come to heads and disagree. Perhaps we could be talking
about some kind of f-measure like with precision weighed against recall.

James Christian

On 24 July 2010 01:14, William Parsons <wbparsons at alum.mit.edu> wrote:

> On Fri, 23 Jul 2010 11:57:27 +0300
> James Christian <jc.bhebrew at googlemail.com> wrote:
>
> > Hi Pere,
> >
> > I've been thinking a lot recently about the meaning of words. Translation
> > has a history as long as the history of languages. Translators who have
> > defined translation dictionaries have often used their instincts guided
> by
> > word usage when choosing which word translates which word. Other than
> this
> > we seem to be lacking any kind of scientific formalism for testing our
> > translation models. Much of the way we choose to translate the bible
> rests
> > on the foundation of tradition of how the bible has been translated in
> the
> > past by those who were fluent in source and target languages.
> >
> > In my opinion, the whole field of translation needs to be formalised in
> some
> > way such that we can test translations empirically in terms of fluency
> and
> > fidelity.
>
> An interesting idea, but I find it hard to envision the means by which one
> could make "fluency" and "fidelity" of translations testable.
>
> --
> Will Parsons
>



More information about the b-hebrew mailing list