LS> This symbol, I understand, was not attested
LS> >before 5th cent. AD (Falk, Schrift in alten Indien).

JLC quoting Gift Siromoney:
> "and such an example is found in the rock
> "inscription at An2aimalai near maturai
> "(Jebarajan Emmanuel and Gift Siromoney, 1978;
> "pp. 8-12). This third system is the system
> "referred to in the ancient grammar tolkAppiyam."
>
> Has this description been criticized?
>
> Does it contradict the statement by Falk?

The tamil puLLi, a "negative" letter to subtract out
inherent "-a" from Brahmi letters to create
"pure" consonants such as k, c, T, t, p, R, and so on
is a major achievement of the tamil grammar
encoded in tolkAppiyam. Archaeologically, it's
well attested from second century onwards
in coins and inscriptions found in rock caves.

The puLLi, a "minus" or "negative" letter,
represents possibly a substratum element
that takes time to assert itself in Tamil
orthography. Remember that many Tamil Brahmi
inscriptions are donative records to Jaina
monks coming from the North. The "negative"
character, so useful in developing non-conjunct
letters for Tamil, is in many ways comparable with
the tilaka/poTTu, or the third eye of Shiva
or, the "puLLi" (dots) representing zeros/tens
in ancient Indian mathematical representation.
Compare the puLLi letter, with the numeral for
"hundred" being represented by a symbol for
"one" with two "dots" on top.

Regards,
N. Ganesan



Iravatham Mahadevan, Occurrence of the PuLLi in the
Tamil-Brahmi script, p. 141-167,
in
Michael Lockwood, Indological essays: Commemorative
Volume II for Gift Siromoney, Dept. of Statistics,
Madras Christian College, 1992

p. 141
" Gift Siromoney was the Socrates of Tamil-Brahmi studies.
He took nothing for granted, held no authority in
reverence and went on relentlessly asking searching
questions which in my view have shed more light than
the ex-cathedra pronouncements by eminent professional
epigraphists."

" The puLLi (literally, a 'dot' or 'point') is a diacritical
mark placed over the consonant characters in the Tamil
script to indicate that the consonants are 'basic' and
do not include the so-called 'inherent' medial vowel -a.
The puLLi was also used to distinguish the short e and o
from the respective long vowels both in initial and medial
positions. The earliest description of the functions of
the puLLi is found in the TolkAppiyam (Ezuttu., 15,16).
The earliest datable inscription employing the puLLi is
the Tamil-Brahmi coin-legend occuring on the reverse of
the bi-lingual silver portrait coins of Vasishthiputra
Sri Satakarni (c. 160 A.D.). The puLLi also occurs in the
Tamil-Brahmi inscriptions engraved in the natural caverns
in Tamilnadu. "

p. 142
" The puLLi is such a familiar and indeed indispensable
feature of the Tamil script that its unusual function as
a "minus marker" is seldom recognized. Unlike all other
medial signs of the Tamil script, which serve to add a
medial vowel to a consonant to produce a consonant-vowel
(uyir-mey), the puLLi subtracts the inherent -a medial
vowel from the unmarked consonant characters. What were
the circumstances which gave rise to such an unusual device?
This question has not been adequately addressed by Tamil
scholars and epigraphists. In this brief preliminary
survey I shall attempt to compare the earliest inscriptional
evidence from the Tamil-Brahmi script with the earliest
literary evidence of the TolkAppiyam on the occurence
of the puLLi and draw some provisional conclusions to serve
as the basis for further research.


The earliest known script employed for writing in Tamil
is the Tamil-Brahmi, a regional and linguistic variant of
the Brahmi script adapted to suit the needs of the Tamil
phonetic system. All but four characters of the Tamil-Brahmi
script are virtually identical in shape with those of the
Mauryan Brahmi script. The cave inscriptions in Tamilnadu
are associated with the Jaina monks and lay devotees. According
to traditional accounts, Jainism and Buddhism spread to
South India during the reigns of Chadragupta (324-300 B.C.)
and Asoka (272-232 B.C.) respectively. The evidence is
suggestive that the Mauryan Brahmi script became known in
the Tamil country sometime during the 3rd century B.C.
and was thereafter adapted to conform to the Tamil phonetic
system. This dating receives strong confirmation from the
excavations of numerous archaeological sites in Tamilnadu,
which have yielded Tamil-Brahmi graffiti on pottery in
increasing numbers, but none datable to a period earlier than
about the 2nd century B.C."

p. 144
" In my earlier studies of the Tamil-Brahmi script (1968, 1970)
I had suggested that the TB-I and TB-II systems were successive
stages. But evidence has accumulated to indicate that this is
not the case. Some of the Tamil-Brahmi inscriptions like the
ones at Alagarmalai are found written in both styles, sometimes
within the same inscription. Tamil-Brahmi graffiti on pottery
from Uraiyur are found in either style even when they are from the
same period as indicated by stratigraphic evidence (K. V. Raman 1988).
Gift Siromoney (1982) has pointed out that the TB-II system using
the unmarked consonant with the inherent -a seems closer to Mauryan
Brahmi than to the TB-I system. On the basis of the evidence
presently available we have to conclude that the TB-I and TB-II
systems are parallel, and more or less contemporary orthographic
styles, and each developed independently from the Mauryan Brahmi
script and later evolved along characteristically different lines."

" It is quite interesting to compare the orthographic evolution of
TB-I and TB-II systems as they provide us with a glimpse into the
pre-history of the puLLi. In the Mauryan Brahmi script basic
consonants do not occur in word-final positions and are indicated in
the medial positions by conjunct-consonants. Both TB-I and TB-II
systems devised methods to depict the basic consonants and in the
process, also got rid of the cumbersome conjunct-consonants of the
Brahmi script."

"Thus the puLLi system suits the genius of the Tamil language better
and has survived to the resent day."

p. 147
" The occurence of the puLLi in the Tamil-Brahmi inscriptions was not
suspected for a long time. Venkayya (1907, 1908) assumed that as the
inscriptions were in Brahmi, the language must be Pali. Krishna Sastri
(1919) recognised the presence of a few 'apparently Tamil' words in
the inscriptions. But his approach was conditioned by the tacit
assumption that even if the language was partly Tamil, the
orthographic principles would be the same as for the Northern Brahmi.
Thus, even when he recognised that some of the consonant characters in
these inscriptions were to be regarded as basic, he chose to read them
with the inherent -a in his actual readings. He even noticed the
'dots' in the Anaimalai inscription, but considered them to be the
anusvAra of the Brahmi script. Even K.V. Subrahmanya Ayyar (1924), who
was the first scholar to establish that the language of these
inscriptions was in fact Tamil, failed to notice the occurrence of the
puLLi in the Anaimalai inscription which he discovered. However he did
recognise that the unmarked consonant characters have to be treated as
basic when required by the context, and he read them accordingly.

The uneven, rough and badly weathered rock on which the
Tamil-Brahmi inscriptions are engraved without any prior dressing of
the surface, makes the identification of the tiny puLLi marks
extremely difficult. Inked impressions (estempages), show up the
presence of pits, depressions and faults in the rock, creating a babel
of bakground 'noise' which almost drowns the 'information' of the
inscriptions. It is even more difficult to recognise the puLLi from
photographs of the impressions, considering the scale of reduction and
the quality of reproduction. The only sure way of finding the puLLi in
the rock inscriptions is to go to the caverns and look at the
inscriptions from close quarters and learn to distinguish the puLLi
made by the chisel or drill from natural faults in the rock.

The first discovery of a puLLi in a Tamil-Brahmi inscription was
made only around 1961 by Pulavar Raju at Arachchalur. In this
inscription, dated in c. 3-4 cent. A.D. on paleographic evidence, the
character e has a clear puLLi placed inside the letter indicating that
the vowel is short.

The next find of a puLLi came from a wholly unexpected source,
the bi-lingual silver portrait coins of the Andhra king Vasishthiputra
Sri Satakarni. R. Nagaswamy (1966) published a revised reading of the
Tamil-Brahmi coin legend occurring on the reverse of this coin. The
reading was further improved upon by R. Panneerselvam (1969). In this
inscription the basic consonant .t in the word Vaaci.t.ti is clearly
marked by a puLLi. This discovery is particularly important as it
provides a firm dating for the occurrence of the puLLi in the middle
of the 2nd cent. A.D.

This was soon followed by the discovery of the puLLi in the
coin-legend of the Andipatti hoard of lead coins published by P. N.
Mohandas (1968). The inscription has been edited by K. G. Krishnan
(1974). The character e in the word etiraan2 is clearly marked by a
puLLi to indicate that the vowel is short. These coins may be dated in
c. 3-4 cent. A.D. on paleographic grounds.

The next advance came when Gift Siromoney and E. Jebarajan (1978a)
published enlarged photographs of the impression from two selected
passages of the Tamil-Brahmi inscription at Anaimalai showing two
puLLi-like dots. They identified one of them as a puLLi (occuring with
.t in the word ara.t.ta), but considered that the other dot (occuring
with t in the word attuvaayi) was not a puLLi (Fig. 1). Neither is, in
any case, the anusvAra earlier presumed by H. Krishna Sastri (1919).
The importance of this finding is that it pushes back the date of
occurrence of the puLLi in Tamil-Brahmi cavern inscriptions to c. 1-2
cent. A.D., about the same period as that of the coin of Satakarni.
The finding also demonstrates the practical problem in identifying the
"white spots" in the inked impressions with the puLLi. It is only on
the basis of physical verification *in situ* that they could decide
that one was a puLLi and the other was not.

I consider the work of Siromoney and Jebarajan significant for
another reason. Their discovery of the puLLi was not accidental, but
is the result of a deliberate problem-oriented search. The problem is
best stated in their own words:

In Mahadevan's scheme of three systems of Tamil-Brahmi (TB-I),
Brahmi (TB-II), and the Tamil puLLi system (TB-III), the puLLi
does not occur in the first two systems. By careful study we have
shown the occurence of the puLLi in one inscription that was
classified as belonging to the second system. Are there any more
inscriptions - such as Alagarmalai inscriptions - of the second
system which also have the puLLi? Does the puLLi also occur in
any of the inscriptions which have been classified as belonging
to the first system? These questions show a great need for
further study in this area.

Since TB-I, II and III are orthographic styles, which are, by
definition, non-puLLi, no-puLLi and puLLi systems respectively,
therefore, in the light of the revised model of orthographic and
chronological evolution proposed by me subsequently, I would re-phrase
the question somewhat differently before proceeding to answer it. Now
that the puLLi is known to occur from at least the Middle Period (c.
1-2 cent. A.D.), does it also occur in any of the Tamil-Brahmi
inscriptions belonging to the Early Period (c. 2-1 cent. B.C.)?"