Ietf-languages Digest, Vol 91, Issue 24

Anthony Aristar aristar at
Thu Jul 22 16:52:18 CEST 2010

For your information, LINGUIST sends in code-change requests to the 
639-3 authority about
once a year.  (We never concern ourselves with 639-1/2, since they are 
far too impoverished for linguistic work)
We have quite a bunch we are intending to send this year, and Vedic will 
probably be one of them.

For those who asked me privately:  the code information pages you see on 
LL are an outgrowth of
our MultiTree project ( This is a 
project to collect all linguistic
hypotheses on language relationships which have been published or have 
gained wide interest.
Obviously, this is a long-term project, and is not yet completed -- nor 
will it ever be, since new
hypotheses appear all the time.  But it is probably now the largest 
collection of languages and names
on the internet, since it includes all that Ethnologue has as well as 
all we have gleaned from the
literature.  This is why you see the odd alphanumeric codes you do:  the 
private use section of the
639-3 code-set was far too small for us.  Everything linguistic entity 
has been assigned a code in the system: the dialects
have a code based upon the  language code (e.g. hsb is Upper Sorbian, 
hsb-kam is the code for the Kamenz
dialect of  the language), and the subgroups have a four-letter code, 
e.g. West Slavic is wsla.

Obviously, in a system like this, each code can be linked to different 
names and a different tree-structure,
depending on the hypothesis it is in.  So we have put together what we 
call "composite" trees, which are
trees which are usually composed of the views of many scholars, and 
(hopefully) represent common wisdom
on language relationships.

When you request a code and its relationships, those from the composite 
tree are displayed as defaults. 
Thus the information you see on the page which has been referred to here 
is from the composite tree.


> ----------------------------------------------------------------------
> Message: 1
> Date: Wed, 21 Jul 2010 21:04:22 +0200
> From: Kent Karlsson <kent.karlsson14 at>
> To: ietflang IETF Languages Discussion <ietf-languages at>
> Subject: Re: Language Variant subtags for Sanskrit
> Message-ID: <C86D1156.15207%kent.karlsson14 at>
> Content-Type: text/plain;	charset="ISO-8859-2"
> Den 2010-07-21 01.05, skrev "Caoimhin O Donnaile" <caoimhin at>:
> ...
>> By these kind of criteria, would Vedic Sanskrit count as a
>> separate "language"?  LinguistList certainly considers it a separate
>> language rather than a dialect:
> And they count as "child languages" to Vedic Sanskrit:
>     Sanskrit; Classical Sanskrit; sanskrit (san)
>     Paisaci Prakrit; Prakrit (qpp)
>     G?ndh?ri (092)
> Note that they equate the "san" ("sa") code to Classical Sanskrit!
> ("091", "qpp", and "092" are LinguistList internal codes, while "san" is an
> ISO 639-3 code that is equivalent to the ISO 639-1 code "sa".)
>     /kent k

Anthony Aristar, Director, Institute for Language Information & Technology
  Professor of Linguistics            Moderator, LINGUIST Linguistics Program
Dept. of English                       aristar at
Eastern Michigan University            2000 Huron River Dr, Suite 104
Ypsilanti, MI 48197


More information about the Ietf-languages mailing list