On Fri, Jan 15, 2016 at 10:37 AM, George Hart <glhart@berkeley.edu> wrote:
[...]
> I’d ask Suresh why Tamil, the southernmost Dravidian language, is the most conservative. If Dravidian speakers
> had entered India (from the east?) and moved through a huge territory of “Niṣādic” (and other?) speakers,
> one would expect Tamil to be far more changed from Proto-Dravidian than it is. George

Thanks George for your question.

A simpler answer could be: for whatever socio-historical reasons, Tamil became the earliest literary language among the Dravidian languages. A literate society tends to be more conservative than a non-literate. As for the question of why Kannada and Telugu languages are not found in early Brahmi inscriptions,  unlike Telugu, as Karashima hypothesizes "Buddhists who were in considerable numbers in both areas didn't care about the local language, unlike Jains (who wrote grammar and texts in the local language of Tamil)".

A contrarian answer is: it is a myth that Tamil is the most conservative among the Dravidian languages. Several of the archaic phonological, morpho-syntactic features found in North, Central and South-Central languages are lacking in Tamil. For example, phonological evidence for /q/ [x], morpho-syntactic evidence for serial verbs, relative-correlative structures without port-RC particles etc. (See Steever) are all found in North Dravidian languages. Negative past formation found in Central and South-Central languages (Ollari and Konda) are completely missing in Tamil (past negatives are not found even in Cankattamil). In fact, the negative conjugation of verbs has fallen out of use in modern Tamil, where negation is expressed through compound verbs (e.g. vara māṭṭān, pārkkavillai instead of kāṇēṉ ‘I do not see’). The copula verb man- 'to be' is replaced by ir- 'to be' in the South Dravidian languages. As you know, ir- 'to be' is not found in South-Central, Central and North Dravidian languages. (ir- most likely a local verb for 'to be'). I believe "the massive restructuring of the verbal system" (as described by Steever in Analysis to Synthesis: The Development of Complex Verb Morphology in the Dravidian Languages) in South India might be a result of rapid transformation of pre-Dravidian hunter-gatherer population through language shift as they attempt to merge with the incoming 'Dravidian' community.

The genetic research is also corroborating the theory that the high castes of Vellalas in Tamil Nadu, Vakkaligas and lingayats of Karnataka, Velama and Reddis of Andhra show external Y-DNA haplogroups, coming from North-West through Gujarat and Maharashtra, as well as through sea-faring. The following diagram of domesticated cattle in India clearly shows that the agro-pastoralism spread to south-India from North West.


​
I have more internal linguistic data that provides stronger evidence for an external origin of the Dravidian languages, which I hope to document and publish sometime this year.