[open-humanities] TEXTUS Translations

Mon Feb 27 15:27:55 UTC 2012

Originally I had thought that there were several possible scenarios,
and we would probably want to have different kinds of relations
between editions of texts and translations of texts. Scenarios could
include:

  1. One person working on their own full translation, privately or
shared with the public
  2. A group of people working on their own translation, privately or
shared with the public
  3. An existing complete and published translation, in the public
domain and publicly available
  4. An new translation, complete and publicly available

For 1. and 2. - people may not want to share their translations in
progress with the world. Translation is an art which requires lots of
context, and my translation may change or evolve significantly,
depending on the background information I have about the text.
Similarly, users may not want to be misled by inaccurate, tentative or
unhelpfully literal or liberal translations. At a minimum I'd suggest
that it would make sense for these to be treated *more* like
annotations than authoritative sources (to be treated as if they are
on the 'same level' as originals).

For 3., several of our advisory board members explicitly expressed
reservations about poor quality translations which have entered the
public domain. However, I think this is a complicated business and
raises difficult questions about fidelity and influence. Even if a
poor quality translation is not as accurate, it may have become
important in a language or culture - affecting translation of terms,
or even the way a work is received. Some historical translations are
important texts in their own right, and don't have value only insofar
as they resonate with the original texts.

For 4. translation is an art, which means (more or less) that one
persons muck is another persons gold, and vice versa. Philosophical
translation especially requires awareness of the history, equivalence
and connotation of terms. Different scholarly communities won't
necessarily agree on the quality or value of a translation.

Suffice to say there are enough worms in the can for us not to want to
make judgements about specific texts, or assume a 'gatekeeper' role
where 'we' (the project) decide about the quality and value of
translations from any number of languages into any other number of
languages. Because we want to make the project a valuable resource in
teaching and research, we also want to avoid having a world in which
no distinction is made between source texts and doubtful translations.

So we certainly we want to distinguish between original texts and
translations. Having translations as a class of annotations is a
reasonably neat way of dealing with this that Sam and I discussed,
especially for cases 1. and 2. but I'm not sure it is adequate for 3.
and 4. I like the way that having translations as annotations
foregrounds the original texts, but it does make the translations less
readable in their own right, and it may shoehorn translations into
correspondence with the original texts in a way which is awkward or
confusing. Do we want the project to cater for, e.g. undergraduate
philosophy students interested in reading texts in their own language,
without understanding the original - or perhaps we should send them
elsewhere?

As an example, a few years ago I laid out George MacDonald's
translation of Novalis' Hymnen an die Nacht as annotations on the
original using Co-ment. It took longer than I expected. I had to
decide whether to do this paragraph by paragraph, section by section
or sentence by sentence. Sometimes bits were split or merged in the
rendering. It was quite useful, but the Co-ment interface also limited
how much of the translation you could read at once, with the effect
that the translation was more like a footnote or an afterthought.
Reading George MacDonald's text, which is a wonderful rendering, of
historical interest in its own right, was a struggle.

As Sam notes, being able to overlay multiple translations of a whole
passage would be extremely interesting. As a longer term ambition,
I've wondered about whether one could bring in dictionary tools. (As
an aside: I'd love to have something like the OED to hand by right
clicking on a word! Or even a domain-specific OED-like tool, by seeing
17th century philosophical texts which contain the word 'vernunft'.
'See this word in other texts'. You could have an advanced option to
select a date range, or to select which authors you'd like to search.)

As a compromise, what about having translations as (i) annotations
(which should adequately cover 1. and 2. for now) and (ii) translated
editions, clearly marked as such (focusing on 3. in the first
instance).

For now we could limit ourselves to *published* translated editions
for (ii), treating these as a separate class of thing, and having some
visual cue in the interface, e.g. showing primary works, THEN
translated editions, THEN annotations when doing a search, or having a
black text logo for primary materials and a grey text logo for a
translation.

We could leave new, hitherto unpublished translations - and issues
related to their quality, possible peer review / user feedback
mechanisms, etc - for a rainy day? ;-)

J.

On Sat, Feb 25, 2012 at 7:17 PM, Sam Leon <sam.leon at okfn.org> wrote:
> Hi All,
>
> One of the use cases that has been discussed for TEXTUS is translation.
>
> This was initially envisaged as side-by-side collaborative translation, much
> like side-by-side collaborative transcriptions. The end product of this
> translation process would be a new openly licensed translation discoverable
> through TEXTUS.
>
> There are various problems with this model. As translations can be
> approached in such different ways it is difficult to collaborate widely on
> them given that authors will have differing ideas about how faithful or
> transparent their translations should be. Unlike transcriptions there is no
> 'right or wrong' translation and as such it might be harder to crowd source
> them. You could easily get the scenario where any instance of TEXTUS filled
> up with lots of partial translations each taking a different approach but
> none ever getting finished.
>
> How you would manage all the many partial translations of varying quality
> would become a big issue. I'm not sure that I would want them all to be
> discoverable as separate texts through TEXTUS. I think this would dilute the
> quality of the content and risk losing the faith of the scholarly community
> a given instance of TEXTUS was targeted at.
>
> One idea that came up at the first user requirements workshop was to have a
> class of annotations that were translations. A user could highlight a
> section of text and write a translation of it that could be reviewed and
> searched by others interested in partial or full translations of a given
> work.
>
> The beauty of this would be that there would no longer be the problem of
> having to treat translations as a new class of texts. It would support
> people doing bits of translations at a time, and looking at the work of
> others, without necessarily forcing these fragmentary translations together.
> On this model TEXTUS could of course be a useful tool for those wanting to
> collaborate on full translations of a work but a new full text would not
> have to be the end product.
>
> It would also be very interesting to look at the different approaches
> translators had to a given section of text. You could pick a section of text
> and bring up all the various translation annotations, or a sub-set of them,
> that had been done on a given section. Of course, some thought would need to
> go into how translation annotations would be displayed alongside one another
> as they would be longer than many other types of annotation.
>
> It would be great to hear people's thoughts on this idea and if anyone knew
> of systems that currently do this kind of thing.
>
> Regards,
>
> Sam
>
> --
> Sam Leon
> Community Coordinator
> Open Knowledge Foundation
> http://okfn.org/
> Skype: samedleon
>
>
> _______________________________________________
> open-humanities mailing list
> open-humanities at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/open-humanities
>

-- 
Jonathan Gray
http://jonathangray.org