Machine translation

[Angus B. Grieve-Smith]

On Fri, 23 Feb 2001, Franz Dotter wrote:

> Many programs which are advertised as to "translate" written into sign
> language only perform a method of producing a signed word for every
> written word. Seen objectively, this is no translation at all, at the
> best we would get a variant of Signed Dutch or English, etc. Is that
> what the deaf community wants?

        This is a good question, and I don't think anyone has done a
survey of any of the Deaf communities.  I do know that SOME Deaf people in
the US want ASL; Karen Alkoby is the only Deaf person I know of who's
actually managing one of these translation projects, and her page
(http://asl.cs.depaul.edu/) clearly focuses on ASL.

        Many of the programs that aim to translate from a spoken language
into a signed language have linguists and native signers closely involved
in the projects, like Karen's project, ViSiCAST, and my own SignSynth.
But there are many others that are created by computer scientists and 3D
graphic designers with no knowledge of Deaf culture and very little
knowledge of any signed language.  This clearly compounds the impression
that no real translation is necessary, as you point out.

        I think it's important for sign linguists to be aware of these
kinds of misunderstandings.  It sounds like there's a bunch of us
interested in this topic; I hope we'll be able to talk about it in London.

                                -Angus B. Grieve-Smith
                                Linguistics Department
                                The University of New Mexico
                                grvsmth at unm.edu