Stop requiring endonyms (Was: RFC 4645bis: making 'pes' and 'prs'extlangs/// Better use autonyms

Phillips, Addison addison at
Wed Dec 10 17:26:40 CET 2008

Please note:

1. RFC 4646 incorporates names defined by ISO 639 automatically. We don't make the initial descriptions up ourselves.

2. RFC 4646 also provides a registration mechanism whereby *any* description (including non-Latin script descriptions, autonyms, endonyms, anagrams, and gibberish) may be registered.

3. RFC 4646's proposed successor adds some additional requirements. It ensures that at least the ISO 639 Reference Name appears in the record. In the case of 'fa', if the reference name were 'Farsi' (it is not), then that name would appear and appear first. Note that this document says the following (which is very similar to text in RFC 4646):

The 'Description' field is used for identification purposes. Descriptions SHOULD contain all and only that information necessary to distinguish one subtag from others that it might be confused with. They are not intended to provide general background information, nor to provide all possible alternate names or designations. 'Description' fields don't necessarily represent the actual native name of the item in the record, nor are any of the descriptions guaranteed to be in any particular language (such as English or French, for example).

4. The important point is that people make registration requests when they want some change (addition, modification) or comment on specific proposed records when those are requested (such as additions of new records or proposed registrations). If you feel that changes should be made wholesale to the incoming ISO 639-3 records, your comments would be best directed at ltru@ and draft-ietf-ltru-4645bis. You should also probably comment on draft-ietf-ltru-4646bis (but if you are going to make comments, you should make them NOW, as we are beyond Last Call).

What I'm basically saying is: it does no good to whine about this topic. If you want something, you have to register it yourself. But then, that's what the registration process is for: requesting stuff. If you want an automagic mechanism, you need it to be written into the RFC.



Addison Phillips
Globalization Architect -- Lab126

Internationalization is not a feature.
It is an architecture.

> -----Original Message-----
> From: ietf-languages-bounces at [mailto:ietf-languages-
> bounces at] On Behalf Of Lang Gérard
> Sent: Wednesday, December 10, 2008 1:13 AM
> To: John Cowan; Stephane Bortzmeyer
> Cc: ietf-languages at
> Subject: RE: Stop requiring endonyms (Was: RFC 4645bis: making
> 'pes' and 'prs'extlangs/// Better use autonyms
> 0-In my opinion, when studying about identification and
> denomination concerning languages, the better solution is to use
> "autonyms"(i.e.:the name given to the considered language by this
> language itself), that certainly are a kind of endonym.
>  And so, I consider that ISO 639-4, that is to give the general
> methodology to build and choose Latin-alphabetic  chains of
> characters of fixed length (alpha-2 for ISO 639-2, alpha-3 for ISO
> 639-2,3 and 5), should use these autonyms along the lines of the
> following points 1 and 2.
> 1-If the considered language is a written language such that no
> valid script for this  language is written with a variant of the
> Latin alphabet,  then a romanization should be used to obtain a
> "written romanized autonym" that represents a very good basis for
> sure identification and denomination of the language, as well as a
> source for building a code element for the representation of the
> name of the considered language, or a tag to represent this
> language.
> 2-In the case where the considered language is not a written
> language, we have to use a "spoken autonym" and a phonetization
> using the International Phonetic Alphabet (IPA) should be done to
> get a representation, to eventually further simplify by
> "romanization", as a good basis for identification and source for
> building a code element for the representation of the name of the
> considered language, or a tag to represent this language.
> 3-In the specific case of Persian/Farsi:
> *ISO 639 (1988) gives the alpha-2 code element "fa" to represent
> the language name whose english version is "Persian", the french
> version is "perse" and the "original version" is "Farsi" that
> certainly is a romanization of the autonym.
> *ISO 639-1 (2002) gives the alpha-2 code element "fa" to represent
> the language name whose english versions are "Farsi, Persian", the
> frencjh versions are "farsi, perse" and the indigenous (unique)
> version is "fârsy", that is clearly another roma,nization of the
> same autonym.
> *ISO 639-2 (1998) gives the alpha-3/B code element "per" and the
> alpha-3/T code element "fas" (recommanded) to represent the
> language name whose english version is
> "Persian" and the french version is "perse".
> *ISO 6639-3 (2006) gives the alpha-3 code element "fas", considered
> as identical to ISO 639-2/T  (and equivelent to the ISO 639-2 "fa"
> and to the ISO 639-2/B "per"), to represent the macro-language name
> whose english version is "Persian", that includes the language
> names whose english version is "dari", coded as "prs", and whose
> english version is "Western farsi", coded as "pes".
>  Moreover, ISO 639-3 is also representing the language names whose
> english version is "Southwestern Fars", coded as "fay", and whose
> english version is "Northwestern Fars", coded as "faz".
> Cordialement.
> Gérard LANG
> -----Message d'origine-----
> De : ietf-languages-bounces at [mailto:ietf-languages-
> bounces at] De la part de John Cowan
> Envoyé : lundi 8 décembre 2008 16:51
> À : Stephane Bortzmeyer
> Cc : ietf-languages at
> Objet : Re: Stop requiring endonyms (Was: RFC 4645bis: making 'pes'
> and 'prs'extlangs
> Stephane Bortzmeyer scripsit:
> > I regard this trend (requiring endonyms) as a quite stupid one.
> Will
> > the british ask us to always write "London" instead of the exonym
> we
> > use ("Londres")? Will they send troops if we do not comply? If so,
> we
> > will ask the italians to stop calling our capital "Parigi" (the
> > endonym is Paris).
> Arguably the English name "Paris" is an endonym as well; in Middle
> English and Old French, the name was unsurprisingly identical, but
> sound-changes in both English and French have altered the
> pronunciation of the "a", the "r", the "i" (in English only), and
> the "s" (in French only), while leaving the orthography unchanged.
> Similarly, I suppose that the many U.S. placenames of French origin
> are pronounced as French by francophones, even though French is
> only minimally an endogenous language of the U.S. (parts of
> northern New England and Louisiana).
> In New York's Chinatown, street signs are bilingual in English and
> Chinese, but who's to say which is the exonym and which the endonym
> in that case?
> > Worse, and more on-topic for this list, will the english-speaking
> > people require that we call their language "english" while we
> always
> > used "anglais"?
> The vast majority of all names are and must be endonyms.  There are
> exonyms for Warsawa (Warsaw, Varsovie, Warschau), but none for
> Zelazowa Wola, even though it was the birthplace of Chopin (whose
> name was itself something of an exonym).
> When we deal with names across scripts, however, as in the Chinese
> and Indian cases, we are always dealing with exonyms, and then
> there is no particular advantage to having multiple exonyms,
> particularly in writing.  International postal addresses may be
> written in Latin script or the script of the destination (save for
> the country name, which must appear in the language of the source),
> and here having more than one way to write "Beijing" is nothing but
> a nuisance.
> > To me, "persan" (the french word) is an exonym, like "german"
> > ("deutsch") or "mandarin" (don't know how to write the endonym).
> Mandarin has no universal endonym; it is Putonghua 'common
> language'
> in the People's Republic, Baihua 'official language' in Taiwan.
> --
> John Cowan  cowan at
> No man is an island, entire of itself; every man is a piece of the
> continent, a part of the main.  If a clod be washed away by the sea,
> Europe is the less, as well as if a promontory were, as well as if
> a manor of thy friends or of thine own were: any man's death
> diminishes me, because I am involved in mankind, and therefore
> never send to know for whom the bell tolls; it tolls for thee.  --
> John Donne _______________________________________________
> Ietf-languages mailing list
> Ietf-languages at
> _______________________________________________
> Ietf-languages mailing list
> Ietf-languages at

More information about the Ietf-languages mailing list