[9mm_fan]

Aravindhan,
I too could not find any thing on Nacchinakkiniyar by google. It will be nice if you can tell us about his theory sometime.

Sorry, that was a typo. The name is actually Nacchinarkiniyar. Nacchinarkiniyar was a mediaeval commentator on the Tolkappiyam. In his commentary on the first verse of the Tolkappiyam's "Eluttatikaram", Nacchinarkiniyar suggests that the forms of the letters of the ancient Tamil script were derived entirely from geometric objects, such as the square, the circle, and the cross, which were combined with each other, and modified with other lines, to form the old script which the Tolkappiyam describes. Unfortunately, my copy of the Tol. only has Ilampuranar's commentary so I can't provide an exact translation of Nacchinarkiniyar's comments, but I think this is roughly what he said.

There are certainly some resemblances in Brahmi between characters having similar sounds (the two "l"s and the two "n"s, for example), and if one takes variants into account, the core characters are quite geometrical, so one sees why he proposed this theory. If imperial Aramaic was not the source for Brahmi, it may well be worth examining the structure of each letter in the script to see if they support his theory.

This makes it likely that the dedicated long matra, too, was first introduced in a Tamil context, and that the resulting system was only later imitated in Bhattiprolu. No such Tamil inscription has however been discovered yet."
I refer to the last sentence. Does the discovery you mentioned provide this missing link?

Ah, that's an interesting question. Unfortunately, the inscription pictured in the Hindu (which, incidentally, is the second stone rather than the first) does not have any long "a"s, so one can't really say, and the transcription of the first inscription suggests it doesn't either. However, an inscription discovered in Arittapatti in late 2003 used the same system as the Bhattiprolu inscriptions - including the distinctive dedicated long "a" matra. That inscription, too, was dated to the 3rd century BC, so it's probably fairly good evidence that the long "a" matra of the Bhattiprolu inscriptions was actually introduced in a Tamil context, as Baums and Glass speculate.

To give some background for the others, the main difference between Tamil Brahmi and Asokan Brahmi is that whereas in Asokan Brahmi a consonant sign has an inherent "a" sound, in Tamil Brahmi it does not, and a matra must be added to produce an "a". If you look at the last two symbols in the photograph in the Hindu's article, you'll see they're read as "ka" and "l". The little bar on top of the cross adds the "a" to the "k". The "l", lacking the bar, is a pure consonant. In Asokan Brahmi, these symbols would have represented "kaala" rather than "kal".

The Bhattiprolu inscriptions, which are in Prakrit, use this system, and add a second feature. In Tamil Brahmi, "kaa" was written by drawing a cross (representing "k"), adding an overbar (adding a short "a"), and then writing the symbol for the independent vowel "a" next to it. The Bhattiprolu inscriptions, in contrast, have a dedicated matra to denote the long vowel "aa", which is basically the overbar for short "a" with a little vertical hook at the end. This system was thought to have been borrowed from Tamil Brahmi, but no actual Tamil inscriptions that used this system were known, until the discovery of the Arittapatti inscriptions a couple of years ago.