Re: [Ietf-languages] [EXTERNAL] Re: language identifiers for sign languages (incl. sgn) vs. attribute for indicating the representation of an individual language in "sign language modality"

["Doug Ewell" <[e-mail]>]

Christian Galinski wrote:

 �

> In any case, the issue of “signing” – including dialects, levels of

> performance, combinations of regional variants, etc. – is more complex

> than can be dealt with in the ISO 639 series. Furthermore, similar

> phenomena exist for drummed, whistled, haptic (and other languages and

> language modalities), which may have to be considered in the future.

> [Particularly in the field of education (and assistive technologies

> applied to support interhuman communication) there is a lot of

> development in this direction.]

 �

BCP 47, at least, doesn’t try to cover all of the modality possibilities that are covered by, for example, the ISO/TC 37/SC 2 NWIP “Identification and description of language varieties,” which was voted on in 2016 and which also attempts to cover regional spoken accents, time frames, proficiency levels, speech impairments, “motherese,” and much more.

 �

If there is a standard (or part) on the horizon that will provide a coded representation of modalities in a way that is compatible with other ISO 639 parts, and could conceivably be made to work within BCP 47, then I would argue strongly against trying to extend BCP 47 or any other ISO 639 part to accommodate this. Otherwise we will have duplicate coding, with all the confusion that duplicate coding normally brings.

 �

Note that the NWIP, as of 2016, offered no code elements, nor any clues as to how code elements would be formulated, what they would look like, or where or how the list of code elements would be maintained, making the proposal unsuitable for discussion in the context of BCP 47.

 �

> b. “sgn” attributed to a language identifier of an individual language

> (not being a sign language) could indicate one or the other kind of

> “signed language variety” – e.g. (in ISO 639) “eng-sgn” = English in

> signed language variety and – if differentiation necessary –

> “eng-UK-sgn” British English in signed language variety

> c. “sgn” attributed to a language identifier of an “individual sign

> language” could be used to make individual sign languages identifiable

> and searchable – e.g. (in ISO 639) “ase-sgn” = American Sign Language

> (ASL) and – if differentiation is necessary – “ase-CA-sgn” Canadian

> variant of American Sign Language

> �

> If this use of “sgn” is not useful/reasonable for ISO 639 principles,

> please feel free to suggest alternatives.

 �

My understanding of ISO 639 is that combination of code elements like this is not within its domain, but left to follow-on standards such as BCP 47.

 �

BCP 47 has a very well-defined syntax with clear, stable rules about which types of subtags can go where. Inserting “sgn” or any three-letter string after a region subtag, or omitting the region subtag and inverting the order of primary and extended language subtags (putting “sgn” after “ase”), is not permitted in BCP 47.

 �

“eng” is also not permitted for “English” in a BCP 47 tag. The subtag for English is “en”, derived from the ISO 639-1 code element.

 �

If this problem is one to be solved within BCP 47 (which I am increasingly convinced it may not be; see above), then the syntactical and constraints of BCP 47 must be adhered to. This is the only way parsers can know what the components of a tag are. This point is not negotiable.

 �

--

Doug Ewell | Thornton, CO, US | ewellic.org

 �