The above is my minimal set of Southern Dravidian language tree. If Dravidian languages have South-West Indian origins, then I would expect them to branch as above. However, the problem is Malayalam and Toda. I suppose those two languages should exchange the places. Or, perhaps, Tulu->Malayalam->Kannada->Kodagu->Toda->Tamil must be the order.
From the above chart it's clear that Tulu is the oldest and Tamil is the youngest of South Dravidian languages.
As far as I know, language changes
1. For every 10kms
2. Over time
But the question is whether it's the accent or the words.
Accent, I would propose is a hallmark of local languages. Malayalam and Tamil accents are that of pre-Dravidian languages, which are extinct now, spoken in those regions. Dravidian speaking people expanded from South-West(most probably from present day Dakshina Kannada, Udupi and Kasaragod districts of coastal Karnataka and Kerala, also known as Tulu Nadu) of India into various directions. Northern Karnataka shows a distinct accent that could be of that of a pre-Kannada language spoken by ancestors of Kannadigas living there. Here, I'm making a grand statement that accents didn't change until the development of modern communication systems.
Ergo, Tulu accent is the closest to the proto-Dravidian language.
Words do change. However, development of literary language slows down this process by standardizing the words.
Ergo, Tamil is the least modified of all Dravidian languages.
Dravidians and Indus valley civilization:
I have already argued that Indus valley civilization was destroyed by Dravidians and Brahuis are the living proof of that event. New studies have shown that major Indian chromosomes like R,H, L have more than 10000 years of presence in the subcontinent. I suppose another Haplogroup J is younger but most of its sub groups might be 6000-8000 years old. These together constitute almost 90% of the sub-continent people( The rest 10% might again show India specific chromosomes like R and J, however, they are differentiated with their Western Eurasian female lineage which forms 10% of Indian female population).
Ergo, it's irrelevent if Dravidians and people of Indus valley civilization shared same genetic make up. After all, genes are irrelevent only culture matters. Nothing exemplifies it better than Indian society. Our misdirected philosophy hampered our growth as a civilized society with a scientific development and technological advancement. Perhaps, most recent example of knowledge becoming pervert is Hitler's "Aryan" supremacy theory. The worst sufferers of that were R1a rich East Europeans and Jews. In all probability, the word "Arya" was coined by some of that Haplogroup people who moved to Indian sub-continent(which itself was derived from Dravidian Ayya). I wonder if Hitler was R1a or R1b or I. If it's R1a, his life would be ironical. If he's R1b or I, "Aryan" doesn't make much sense I suppose.
Antiquity of the Dravidian languages:
I think IVC came to an end around 1800 BCE. Therefore, Brahuis might have reached there by now. So in all probability Dravidian languages diverged around 2000 BCE as linguists have claimed. Therefore, I think Dravidian languages started spreading around 4000 years ago. However, I don't think they replaced all the pre-existing languages in India. Most probably they drove all other linguistic families to extinct in South India. However, I suppose most of the North , West and East India was still speaking Astro-Asiatic and Semitic and some other language families that became extinct. I think except for Dravidian and Astro-Asiatic languages, Semitic and other languages were driven to extinct by the next wave of languages which was Indo-Aryan. Again, the accent could be Dravidian, Astro-Asiatic, Semitic and some extinct languages depending upon the local lingo.