[Humanist] 27.210 a universal translator

Humanist Discussion Group willard.mccarty at mccarty.org.uk
Sun Jul 14 02:49:04 CEST 2013


                 Humanist Discussion Group, Vol. 27, No. 210.
            Department of Digital Humanities, King's College London
                       www.digitalhumanities.org/humanist
                Submit to: humanist at lists.digitalhumanities.org

  [1]   From:    James Rovira <jamesrovira at gmail.com>                       (2)
        Subject: Re:  27.209 A universal translator? Then what?

  [2]   From:    Daniel Allington <daniel.allington at open.ac.uk>            (52)
        Subject: Re:  27.209 A universal translator? Then what?

  [3]   From:    Mark Wolff <wolffm0 at hartwick.edu>                         (59)
        Subject: Re:  27.209 A universal translator? Then what?


--[1]------------------------------------------------------------------------
        Date: Fri, 12 Jul 2013 16:08:50 -0500
        From: James Rovira <jamesrovira at gmail.com>
        Subject: Re:  27.209 A universal translator? Then what?
        In-Reply-To: <20130712205218.0C2823A09 at digitalhumanities.org>


Differences in language are never completely removed by translation.

Jim



--[2]------------------------------------------------------------------------
        Date: Fri, 12 Jul 2013 22:13:20 +0100
        From: Daniel Allington <daniel.allington at open.ac.uk>
        Subject: Re:  27.209 A universal translator? Then what?
        In-Reply-To: <20130712205218.0C2823A09 at digitalhumanities.org>


Now difference in language signals difference in
culture. Remove the difference in language in such a way as SayHi hints
at, what would we then have?

A very interesting question. I've been thinking about this since Microsoft's demonstration of what appeared to be a working English-Chinese realtime translator last year. Human translators, as we all know, translate cultures as well as languages; automatic translators just look for equivalent words and then arrange them according to syntactic rules. So in a world where automatic translation became more prevalent, I believe that separate cultures would continue to exist - they'd just rub up against each other in a different way.

Yours,

Daniel

Dr Daniel Allington
Lecturer in English Language Studies
Centre for Language and Communication
The Open University

www.danielallington.net http://www.danielallington.net

On 12 Jul 2013, at 21:52, Humanist Discussion Group wrote:

                Humanist Discussion Group, Vol. 27, No. 209.
           Department of Digital Humanities, King's College London
                      www.digitalhumanities.org/humanist<http://www.digitalhumanities.org/humanist>
 

       Date: Sat, 13 Jul 2013 06:39:41 +1000
       From: Willard McCarty <willard.mccarty at mccarty.org.uk>
       Subject: A universal translator? Then what?

I recently stumbled across a translation app for iOS, SayHi (free) and
its big sister, TableTop ($1.99 AUS). Both of these connect to a server,
which I suspect does most to all of the heavy lifting, i.e.
voice-recognition, translation and rendering into whatever target
language. But it is quick. I don't think you would want to depend on it
for poetry, philosophical argument or other highly subtle use of
language. But the developer's stated objective, of facilitating simple
conversation in real time between, say, an English and a German speaker,
seems to have been achieved. The idea is that you put your iPad on the
table between the two of you, each taps the appropriate button in turn,
speaks and then is translated in a man's or woman's voice, depending on
a set preference.

Impressive, I think. Let us say it works (as seems) well enough that
people begin carrying their iPads around for the purpose. Let us say
that digital technologies continue to progress, devices shrink in size
and so on, let's say to the size of an earphone, or implant. All
reasonable expectations? What then? I have read too many predictions of
a universal brother/sisterhood brought about by some invention or other
-- the telephone was one of these -- to think any fundamental human
problem would be solved, but change as a result is certainly on the way.
Only a matter of time. Now difference in language signals difference in
culture. Remove the difference in language in such a way as SayHi hints
at, what would we then have?

Yours,
WM
--
Willard McCarty (www.mccarty.org.uk/ http://www.mccarty.org.uk/ )



--[3]------------------------------------------------------------------------
        Date: Sat, 13 Jul 2013 13:26:13 -0700
        From: Mark Wolff <wolffm0 at hartwick.edu>
        Subject: Re:  27.209 A universal translator? Then what?
        In-Reply-To: <20130712205218.0C2823A09 at digitalhumanities.org>

The app is interesting in that in addition to producing computer voices that speak in two language, it generates a running bilingual transcript of the conversation, so both participants can read what was said and how it was translated.  One could use the transcript as a teaching tool to show errors in translation.  I wonder about the app's underlying design, however.  There are at least 40 languages and dialects to choose from, and it is supposedly possible to translate from any one to any other.  There are therefore 40!/(2!(40-2)!) = 780 possible pairings between languages, and from a programming perspective it would be a huge task to design protocols for each pairing.  It is more likely that all communication is translated into a lingua franca first before being translated into the target language.  I would suspect the lingua franca is English (although if anyone on this list who knows something about NLP and translation can explain what's going on differently, please do).  The immediacy of the app belies the fact that all linguistic roads pass through one language.  This is not necessarily a bad thing, but anyone who has played the game Telephone knows that the more a message has to be transmitted from one point to another, the more likely it will be corrupted.  It would be interesting to see how the app works if neither participant spoke English.

mw

--
Mark B. Wolff
Associate Professor of French
Chair, Modern Languages
One Hartwick Drive
Hartwick College
Oneonta, NY  13820
(607) 431-4615

http://bumppo.hartwick.edu/~mark/





More information about the Humanist mailing list