"Dead" languages

[Vidhyanath Rao]

Max Wheeler <maxw at cogs.susx.ac.uk> wrote:

> The name usually used for e.g. Sanskrit, Arabic, Latin, or Hebrew
> while it had no native speakers, is "classical language". Such a
> language may well have an extremely wide range of functions, but is
> learnt formally, through instruction, by people who have some
> distinct native language.

What is the thing that is `learnt': If you mean active mastery, the above
applies to formal registers of some languages at least. If you mean
passive mastery, you need to establish that that was never acquired
without formal instruction from when there were no `native speakers'.

Let me expand on the first sentence above. There are communities in Tamil
Nadu in India of people originally from Karnataka, Maharashtra etc. At
home they speak a dialect of Kannada, Maratthi and so on. But do not
always receive formal instruction in their ``native tongue'' (but they
often do in Tamil). Their understadning of formal Tamil is similar to that
of any of their neighbors, but their understanding of the formal version
of their ``native tongue'' is quite limited. This also applies in spades
to the children of expatriates in US. It seems that what matters more to
language command is use in exchanges with peers and older children than
exchanges with parents (so much for languages learned at ``mother's
knees'').

I get the impression that the following model underlies the claims about
`classical' languages: As a language with a formal register changes, the
formal register may change (perhaps with a time lag) along with the
informal registers, or the formal register might get `stuck'. In the later
case, it becomes a `classical' language when the informal registers have
changed sufficiently.

This model fails to consider that the formal register might remain static
in some respects (as a link to the past) while changing in others (to
facilitate understanding by those without formal training). In case of
Sanskrit, the phonology and the morpheme set (but not in `hybrid'
Sanskrits, which Lars mentioned) stayed static while the syntax, and to
some extent the semantics, changed in ways influenced by the popular
speech. This can be seen simply by comparing, say, early Upanishads
(current dating makes these close to Panini, with all but the earliest
handful likely later) with drama dialogs. It must be noted that Indians
themselves were aware of this: There is a small exchange in
Padmapra:bhr.taka in which a character is ridiculed for using obsolete
syntax (aorist inj with ma:, aorist etc generally instead of past
participles). [BTW, anyone who thinks that Sanskrit was a liturgical
language only must read this play, if not the whole set of four that are
collectively called caturbha:n.i.]

    (1) Late Vedic/    (2) Dramatic   (3) Sans. cha:ya:   (4) Prakrit
    Early post-Vedic   Sanskrit        of Prakrit dialog
                           (5) Formal                          (6)
Collquial
                            Tamil                                Tamil

Conventionally, we group (1) and (2) as the same language, as we do (5)
and (6), but (4) is considered different from (2) while (3) is left out of
account. I don't see how this  can be justified on an objective basis. It
should be pointed out that recently people have been arriving at the
conclusion that medieval IA speech community had a chain of registers,
with the highest being similar to dramatic Sanskrit with the lowest being
apabram.sa (see, for example Deshpande, `Sanskrit and Prakrit:
sociolinguistic issues', espcially the chapter on vernacular Sanskrit and
the one before that.). This muddies the waters even further.