The speakers of Proto-Dravidian, according to archaeological and linguistic streams of evidence, lived in the Krishna-Godavari valley in present-day north Karnataka and Andhra Pradesh. – Anirudh Kanisetti

Veteran actor Kamal Haasan courted controversy by declaring that “Kannada was born out of Tamil.” The question of which Indian language is oldest—and, by extension, most native to the soil—has been a political hot topic since the mid-20th century. Some say Sanskrit, others say Tamil. But beneath the nationalist furore, paleobotanists, historical linguists, and archaeologists have made stunning discoveries about the linguistic heritage of all Indians.

From the lost Gangetic ‘Language X’, to the possible origins of Southeast Asian languages, to the homeland of Proto-Dravidian speakers, it turns out prehistoric Indian languages were as diverse as today’s.

Where do Dravidian languages come from?

The term “Dravidian” today is often associated with India’s southern states, linked to ideas of ethnicity, culture and politics. Here I use it only in the linguistic sense. In The Dravidian Languages (2003), linguist Bhadriraju Krishnamurti writes that Dravidian languages are spoken from the tip of the peninsula deep into Central India; one isolated Dravidian language, Brahui, is spoken as far west as Balochistan in present-day Pakistan.

Anthropologist and historian Thomas Trautmann, in Dravidian Kinship (1981), also found a Dravidian substrate in many place-names in Maharashtra, and pointed out that Dravidian cultural practices—such as first-cousin and maternal uncle-niece marriages—are practiced by a few castes in Sindh and even Gujarat. Speakers of Dravidian languages, and their descendants, are extremely widespread.

Given this vast geographic range, it’s natural to ask: who were the ‘original’ Dravidian speakers? How did they spread and why? By looking at the earliest shared features of all Dravidian languages, we can assemble a hypothetical Proto-Dravidian language from which all modern Dravidian languages descend. We can figure out what plants and animals they saw, what their climate was like, and what their politics and settlements were like. Then we can look at the ecology of the subcontinent, archaeological digs, and we can see what matches.

Distinguished linguist Franklin C. Southworth, in his paper “Proto-Dravidian Agriculture” (2005), made the most rigorous attempt yet to reconstruct this lost world. Proto-Dravidian speakers had a word for “king”. They used a similar word for “hut” and “village”, suggesting small populations of related families. They knew of various agricultural and hunting tools, and a wide variety of wild animals.

Around the 3rd millennium BCE—when the Harappan civilisation was thriving on the Indus Valley—the speakers of Proto-Dravidian were also aware of many crops, such as sorghum and various types of millet and gram. They also had terms for cattle pens and domesticated sheep and goats. Finally, as archaeobotanist Dorian Q. Fuller writes in “Non-Human Genetics, Agricultural Origins and Historical Linguistics in South Asia” (2007), Proto-Dravidian speakers seem to have lived in a dry, deciduous forest environment.

One region seems a good match for all these criteria. It is a region where the ranges of the modern Dravidian language families—Northern, Central, South-Central and South—overlap, and possibly where they radiated from. This is supported by extensive archaeological findings of a “Southern Neolithic” period, with evidence of small mud homes, remains of domesticated and wild animals, and crops.

There is a 73 per cent match between Southworth’s Proto-Dravidian vocabulary of plants and those found in Southern Neolithic sites. Surprisingly, these sites are rather distant from the hotbeds of South Indian linguistic nationalism today. They are neither in south Karnataka nor in Tamil Nadu. Rather, the speakers of Proto-Dravidian, according to archaeological and linguistic streams of evidence, lived in the Krishna-Godavari valley in present-day north Karnataka and Andhra Pradesh.

The archaeology of languages

To be clear, this is not to say that all Dravidian speakers originated from the Krishna-Godavari valley. (If we are being cheeky, no human being truly ‘originated’ anywhere except Eastern Africa.) The fact is, even the Proto-Dravidian language has some words borrowed from other language families, namely Austro-Asiatic—spoken mostly in Southeast Asia today, with the Munda families of Odisha and Chhattisgarh being the Indian representatives. This may suggest that the speakers of even earlier stages of Dravidian migrated to the Krishna-Godavari valley from elsewhere, picking up influences from other languages on the way. Some genetic and linguistic theories link early Dravidian speakers to the Iranian Plateau and the Harappan civilisation, but that’s a matter for another day.

Interestingly, the Proto-Dravidian language is not a perfect match for Southern Neolithic excavations: the peoples of the Southern Neolithic practised urn burials, but there’s no vocabulary for it in Proto-Dravidian. It also doesn’t match other archaeological candidates, such as the Harappan civilisation. If their cities are anything to go by, Harappans must have had a vocabulary for engineering and geometry, but it’s practically nonexistent in Proto-Dravidian. Proto-Dravidian also doesn’t have a word for ‘rhinoceros’, which are often depicted on Harappan seals. This doesn’t mean that no Dravidian speakers lived in Harappan cities—such a vast civilisation must have been multilingual. It just means there may have been another, now-extinct early branch of Dravidian languages, which could have evolved separately from Proto-Dravidian.

Proto-Dravidian has words for some crops—especially wheat—which may be of Harappan origin, suggesting, at the very least, agricultural exchanges. The true “homeland” of the Dravidians, then, is still unclear. All we can say for certain is that around 3000 BCE, Proto-Dravidian speakers deep in the South Indian peninsula harnessed agriculture and, as their population exploded around 1100 BCE, they spread out in waves across the Indian Subcontinent.

“Broadly, the default Proto-Dravidian agricultural practice was dry farming of millets, pulses and tubers. Irrigated rice farming (alongside cash crops like cotton and sugarcane) became more important in the late 1st millennium BCE,” Dr Sureshkumar Muthukumaran, a historian, curator and lecturer at the National University of Singapore, told me. Over the centuries, Dravidian speakers traded words, animals and crops not only with North India but also with Southeast Asia. A particularly influential branch headed south, giving rise to the South Dravidian languages. Some groups, relatively isolated on the Nilgiri hills, developed languages such as Irula and Toda. Others, settling into the expansive coasts and plains, spoke the ancestors of Kannada, Tamil and Malayalam.

The language that became Tamil, according to Krishnamurti (Dravidian Languages), branched off around 600 BCE, roughly when the first cities were growing on the Gangetic Plains far to the North. Three centuries later, it had developed into Old Tamil, the first Dravidian language to have a written culture, composed in thriving new trading towns with rice-farms. Old Tamil itself was composed of many dialects, which evolved into Middle Tamil and eventually modern Tamil centuries later. Between 800–1200 CE, some Middle Tamil dialects branched off into Malayalam.

We can say with confidence that the ancestor of Kannada is not Tamil: it is a lost South Dravidian language related both to the languages of the Nilgiris and to Old Tamil. Unfortunately, the earliest written examples of Kannada date to c. 450 CE, so we don’t have a clear picture of how the language evolved in the centuries prior. Thereafter, though, many dialects of Kannada evolved, through Old Kannada into Middle and thence modern Kannada. In North Karnataka, Kannada dialects had a fertile exchange with Indo-Aryan languages such as Marathi, which in turn had a Dravidian substrate.

The mosaic of Indian languages

It is becoming increasingly clear that this complex mosaic of linguistic borrowings, evolutions, migrations, and shifts is the story of all Indians, indeed of all humanity. Rig Vedic Sanskrit provides another early example. Prof Michael Witzel, a linguist and scholar of the Vedas, writes in “Substrate Languages in Old Indo-Aryan” (1999) that already by 1500 BCE, the earliest Indo-Aryan languages had absorbed a chunk of vocabulary from now-lost Austro-Asiatic languages in Punjab—a hypothetical Harappan language called “Meluhhan” in Sindh, and a language called “Language X”, probably spoken by the earliest Neolithic farmers in the Gangetic plains. A few centuries later, c. 800 BCE, Dravidian words suddenly appear in the Vedas, possibly hinting at now-lost North Dravidian languages.

As noted above, Tamil literature and writing appeared around 300 BCE. The earliest Tamil literature is called the Sangam poetry, after assemblies of poets who compiled it. Linguists, however, generally agree that the word “Sangam” itself is borrowed from Indo-Aryan languages, while Old Tamil poets were clearly aware of Vedic mythology. Meanwhile, around the same time in North India, Prakrit literatures blossomed, overpowering the dominance of Sanskrit in religion and ritual. Krishnamurti (The Dravidian Languages) argues that Prakrits probably developed from the integration of the speakers of now-lost regional Dravidian languages into the North Indian mainstream. And, in the medieval period, starting around 600 CE, all the major Southern Dravidian languages, including both Kannada and Tamil, borrowed extensive political, grammatical, and religious terms from a revitalised Classical Sanskrit.

So, what is indigenous and what is foreign? Which language is “oldest” when all have branched off from already-diverse origins, and borrowed from or lent to each other across centuries? India’s modern linguistic diversity didn’t appear out of nowhere: all the evidence is telling us that we are the inheritors of a complex, multidimensional mixing of genes, words, technologies, and ideas across timescales of truly mind-boggling proportions. Banal statements that language A is older than language B might set social media aflame and rally nationalists to a cause. But, as is increasingly clear, patriotic oversimplifications always trample on the histories and dignities they claim to protect. – The Print, 5 June 2025

› Anirudh Kanisetti is an author and public historian.