Latin Transcription of Sanskrit - one more try

[Gerard Huet]

Now that the heat on anglo-saxon imperialism has subsided, I propose
to return for a moment to the transcription discussion.
First of all, to answer Leslaw Borowski's questioning why I qualified
this discussion as "boring" - It is not at all because this question is not
important, I think we need to advance on these standardisation issues indeed.
It was just my personal feeling about Indology traffic, which I classify 
roughly in 3 categories:
30% interesting indological communication
10% tolerable information gathering (how to reach Pr X, biblio requests)
30% boring technical stuff about fonts and transliteration systems
30% noise (communalism, oppression of backward casts such as French, 
Kashmir, widow-burning, chain letters, subscribe magazines, etc) :-)

Now, returning to Francois.Voegeli's allusion to the standardised latin
transcription used say in Renou's grammar and the difference with say
ITRANS convention. Well, here again we fall in the confusion I denounced
in a previous message: what has been defined at the Indological Congress
and presumably used by Renou is a typographical convention using latin letters
augmented with diacritic marks (what I called level 3 previously), which does
not give clues on possible computer encodings in the ASCII alphabet or
other (what I called level 2, where competing conventions are ITRANS,
DEVNAG, Borowski's system, etc).

Now, let us see. I claim that discussing level 2 conventions is important,
but a bit boring, in that this question occurs repeatedly in this forum,
but is generally inconclusive, since as long as the convention is
unambiguous it is mechanically translatable into any other one.
But maybe the typographical look (level 3) may be agreed upon much more
easily, since indeed most recent indological work use a de facto standard.

So let us look at Renou's grammar for instance (or equivalently in the
Stchoupak, Nitti and Renou `Dictionnaire Sanskrit-Franc,ais').
And indeed this conforms more or less to the de facto standard, with one
serious exception: the palatal sibilant is written as a c with cedilla
(like in franc,ais), rather than the now standard s with acute accent
(\'s in TeX notation). Now this cedilla is one of these crazy peculiarities
the French (me included) relish upon. No french typewriter or AZERTY keyboard
goes without one. But I find it just confusing to use it to denote the
palatal sibilant, which corresponds phonetically to the french ch, whereas
the cedilla is to distinguish the s sound from the k sound. To my taste,
the 's is a much better notation, or its slight variant s with a dot above
as used by Monier-Williams. For the cerebral sibilant the transcriptions
sh and s with a dot under have been used, and I think the latter usage prevails.
I considered these issues when I was designing my sanskrit dictionary,
and settled on \'s and {\d s} (in TeX notation, the latter means s with a dot
down) for the two sibilants. I also considered several variants of the
anusvAra, and settled on \.m (dot above m) for the `original' one,
and {\d m} (dot under m) for the one which stands for another nasal.

My feeling on these matters is that indeed it would be rather easy to come
to an agreement on a standard on this typographical issue, possibly with a few
slight variations. This could be finalised at the Bangalore conference in january.

For historical completeness, we may mention MacDonnell's 1924 Practical
Sanskrit Dictionary, whose latin transliteration, mixing straight and
slanted letters, is remarkably unpractical. And still earlier, Emile Burnouf's
Dictionnaire classique Sanscrit-franc,ais (1866), which uses a rather
ingenious transcription system where special ligatures where designed
to represent in one letter th, dh, etc. This is called the system
of the `Ecole de Nancy', and in the enthusiasm for oriental studies that
happened at the time they went as far as casting a lead font specially
for this usage (at the `Imprimerie Orientale de Veuve Raybois, 3 rue du
faubourg Stanislas, Nancy'). If I were serious as a french imperialist
I should probably push for this tradition and build a METAFONT version
of it, except that I have a hard time distinguishing these ligatured
t's from ordinary t's without glasses, and I think this should just be put to rest
as indological archeology. Oh, I almost forgot. The Ecole de Nancy
system used x for k.s, which answers a query from Frances Pritchett
some time ago. It also used w instead of v, an interesting variation.

Enough for today
G. Huet