Language Variant subtags for Sanskrit

[Caoimhin O Donnaile]

Peter Scharf said on Monday:

> On the grounds that a language is something that occurs in various dialects
> that are mutually understandable to each other, I think that there should
> continue to be just a single language tag for Sanskrit.  Panini,
> Wackernagel, and Whitney all described the language in single treatises and
> referred to different dialects (including Vedic) within their treatises.
>  While some modern linguists would not mind referring to Vedic as a
> different language from Classical Sanskrit, in traditional circles in India
> this would not go over well at all.

I'll just play the devil's advocate here since I know very little
about Sanskrit.

The difference between a language and a dialect is a matter of degree,
and the ISO 639-3 criteria for being a separate language are perhaps
a bit more relaxed than in some everyday contexts.  For example in
Germany:

   Hochdeutch (deu)
   Bavarian (bar)
   Kölsch (ksh)
   Limburgish (lim)
   Luxembourgeois (ltz)
   Mainfränkisch (vmf)
   Pfaelzisch (pfl)
   Plautdietsch (pdt)
   Low Saxon (nds)
   Upper Saxon (sxu)

are all considered separate languages.  In Italy the situation is similar:
   http://www.ethnologue.com/show_country.asp?name=IT

Scots is considered a different language to English.
Middle English (1100-1500) is a different language from English.
Middle French (1400-1600) is a different language from French, and 
different from Old French (842-1400).

By these kind of criteria, would Vedic Sanskrit count as a
separate "language"?  LinguistList certainly considers it a separate 
language rather than a dialect:

   http://multitree.linguistlist.org/codes/091

and LinguistList's criteria seem to match those of ISO 639-3
pretty well.

Caoimhín