compound analysis in e-texts

[gail at utxvms.cc.utexas.edu]

On Mon, 2 Sep 1996, George Thompson wrote:

> Frits Staal has argued, provocatively and I think rightly, that linguistics
> in India "did not originate  *in spite of* the absence of writing but
> *because* of it" [in Arch. europ. sociol. 30, 1989; the starred passages
> reflect Staal's italics].  Perhaps it can be said that devanAgarI reflects,
> more closely than a western romanized script, *something* in the language
> itself, though I can't quite put my finger on it. Perhaps it is
> meta-language, since a writing system presupposes a conception of language.
> In any case, I think it has something to do with our problem of whether or
> not to analyze text [compounds, sandhi, etc.].
> 

Bill Poser argues that writing systems represent a phonological analysis
of the language for which they are used. This sounds to me like what you
were trying to put your finger on. Devanagari does in fact reflect a
particular phonological analysis of Sanskrit; e.g. consonant clusters are
clearly distinguished from single consonants by the use of half-letter
symbols.  Most importantly, while the roman alphabet mixes vowels among
consonants and does not organize the various Cs and Vs in its ordering of
letters in the alphabet (abcdefghij...), Devanagari has vowels clearly
distinguished from consonants and, further, organises the various sounds
according to place and manner of articulation. Stops are listed separately
from non-stops, the stops are also distinguished from each other in place
and manner of articulation by the use of rows and columns (velar stops are
written in the first row, etc.). I've noticed that the writing system does
affect the psychological view (intuitions?) that speakers have of their
language; e.g. some Hindi speakers think that the half-letters used for
consonant clusters really are half sounds!