by Tony Joseph
For example, the commonest seal in the Harappan Civilization – about 60 per cent of all seals ever recovered – has the image of a unicorn. While some believe it was a ‘state symbol’ then, much like the Ashoka Chakra is today, others consider it the emblem of one of the constituent elements of the Harappan elite, the most common assumption about the civilization being that it had a ‘ruling elite’ who shared power, rather than a solitary king or an emperor. The unicorn is often found with its head above an object that has been variously described as a brazier (a container for burning coal or wood), a filter for making some kind of ritual intoxicant and a manger (the vessel from which cattle and horses are fed). So it is obvious that there was an important mythology of some kind associated with the unicorn, though it has not survived into modern times in any recognizable form and we have no clue what it could have been.
A seal depicting a unicorn standing in front of what has been variously described as a brazier, a filter and a manger
But there are other seals that are easier to decipher. For example, there are recurring images of what looks like a deity in a peepul tree, with a kneeling worshipper in front of it. Since peepul is still considered sacred, this should cause no surprise. There is also the figure of a ‘horned deity’ that appears at several Harappan sites around the beginning of the Mature Harappan period, with what looks like a buffalo-horn headdress.
A seal from Mohenjo-daro depicting a deity wearing bangles and a horned headdress in a peepul tree, with a worshipper kneeling in front of it. Behind the worshipper is an enormous ram. At the bottom of the seal there are seven figures in procession.
Since horned figures are divinities in the Mesopotamian Civilization, some believe that this could be a shared motif. A few seals have a horned figure sitting in a striking yoga-like pose on a stool of some kind, indicating that if this was not a divinity it must at least have been a very important figure in the civilization. Another seal shows a person sitting in a tree, perhaps talking to a tiger that is turning its head to look at him. One seal shows a hero placing his foot on the horns of a water buffalo while spearing it, in front of an erect cobra.
It is clear that seals depict scenes that are evocative of religious beliefs or myths or stories that were well known to the Harappans. What these were is, of course, beyond our ken today, though we can clearly see that some concepts did pass on to later periods – from the sacredness of the peepul tree to the yoga-like pose.
A horned figure sitting in a yoga-like pose
A seal showing a man spearing a water buffalo while placing his foot on its head. Looking on is a seated figure wearing a horned headdress. This seal is different from a similar one described in the text in which a cobra is witnessing the spearing of the buffalo.
The language of the Harappans
One question is key to tracing the formation of Indian populations: what language did the Harappans speak? Most historians argue that many languages would have been spoken in the vast region of the civilization, not just one. Considering how many languages are spoken in the region today – from Baluchi, Pashto, Punjabi and Gujarati to Hindi, Brahui, Sindhi and Burushaski – this is not an unreasonable surmise.
However, it is still possible that the civilization had one predominant language family with its own dialects or subgroups. It is also likely that it had an official language that was depicted in the seals. How do we know that? When a script is used for writing different languages, the order in which the signs/letters appear usually changes in a noticeable manner. For example, in the Harappan seals that were found in Mesopotamia or the Gulf region, experts have found that the pattern or order in which the signs appear is different from the way they appear on seals found in the Harappan Civilization itself, suggesting that in the seals found abroad the Harappan script may have been used to write a different language, perhaps Akkadian or Sumerian or other languages spoken in the Gulf region. There are no such differences in sign patterns among seals found in the Harappan Civilization region itself. Therefore, it is likely that at least during the Mature Harappan period, when there was a high degree of standardization in general, there was one language that was predominantly used, perhaps for administrative, trade and legal purposes.
So what language could this be? This would be an easy question to answer once the Harappan script is deciphered. But a century of sustained and determined efforts by a wide variety of experts from all over the world have failed to crack the script. This is not necessarily surprising because the hieroglyphic script of Egypt may not have been deciphered at all had people not stumbled upon an inscription that used multiple scripts to say the same thing. In 1799 French soldiers rebuilding a fort in Egypt discovered a stone with carvings on it in a town called Rosetta in the Nile valley. This came to be called the Rosetta Stone, and it was found to have been carved in 196 BCE by a group of priests in Egypt to honour the then pharaoh, by listing all the things he had done for the people. (Propaganda, obviously, has very ancient beginnings!)
Remarkably, the Rosetta Stone had inscriptions in three scripts, all of them in use in Egypt when the stone was carved. The first was hieroglyphic, which was used for important religious documents. The second was demotic, which was then the common script of Egypt. And the third was Greek, the language of the rulers of Egypt at that time. Shortly thereafter the hieroglyphic script was deciphered by the French scholar and philologist Jean-François Champollion, who could read Greek and Coptic (a language that shares much with Demotic). Thus a script that had resisted all attempts at decipherment for centuries was finally decoded within a couple of decades of the discovery of the Rosetta Stone.
The story is similar in the case of the cuneiform script in which many Mesopotamian languages were written, including Sumerian, Akkadian and Elamite. Here too, it was the discovery of a multilingual inscription at Mount Behistun in the Kermanshah province of Iran that led to the final decipherment. The Behistun inscription was authored by Darius I, the fourth king of the Persian Achaemenid empire, sometime between 522 BCE and 486 BCE, and is an account of the king’s life, battles and victories. The inscription had the same text in three different languages written using the same cuneiform script: Old Persian, Elamite and Akkadian. For the Harappan language, however, we have not yet found an equivalent of the Rosetta Stone or the Behistun inscription and until we do, cracking the code might prove to be difficult.
Historians and archaeologists have so far overwhelmingly backed the idea that the language underlying the Harappan script was Proto-Dravidian, but the inability to break the code has left the question hanging, without a final resolution. The debate over what language the Harappans spoke has resembled a Gordian knot more than anything else – until now.
The Gordian knot is now being cut, not because the script is closer to being deciphered, but because ancient DNA findings have now joined hands with archaeology and linguistics to provide a consistent and coherent explanation for the demographic composition and the language of the Harappans. The crucial thing to keep in mind is that these three disciplines are independent of each other, with different starting points. They also use very different materials, methodologies and scientific techniques to arrive at their conclusions. Thus it is remarkable that they all arrive at the same conclusion.
Genetic and archaeological evidence
We saw in chapter 2 how new ancient DNA evidence from west Asian and the ‘Indus Periphery’ individuals showed that an Iranian agriculturist population from around the Zagros region had contributed significantly to the populations in India today. This discovery rested on two sets of ancient DNA evidence – let’s recount them briefly.
The first set of ancient DNA evidence was from the Zagros region of Iran dated to between 8000 BCE and 7000 BCE. It showed that these Zagrosians had a distinct type of west Eurasian ancestry. What differentiated them from others of the region was that they lacked the early Anatolian ancestry that the rest of them had.
The second set of evidence was ancient DNA f
rom three ‘Indus Periphery’ outlier individuals with a unique genetic composition. Between 14 and 42 per cent of their ancestry related to First Indians, and the rest to Iranian agriculturists, and none of them had any Anatolian ancestry. This was quite unlike others around them in the same region, who all had Anatolian ancestry and no ancestry related to the First Indians.
‘The Genomic Formation of South and Central Asia’ study, therefore, arrived at the inescapable conclusion that these three ancient individuals were recent migrants from the Harappan Civilization; that they represent the genetic composition of the population of the Harappan Civilization itself; and that they are a mixture of Iranian agriculturists from the Zagros region and descendants of the First Indians. The scientists also concluded that the admixture between the two populations had taken place at the latest between 4700 BCE and 3000 BCE.
Please note that at no stage of this genetic trail was any reliance made on archaeological or linguistic discoveries.
No less emphatic than the genetic evidence is the archaeological evidence. In chapter 2, we saw extensive confirmation of a connection between the Zagros region and the early farmers of Mehrgarh in Balochistan. From quadrangular houses built with narrow bricks about sixty centimetres long to circular firepits filled with burnt pebbles and sequential slab construction of pottery, the similarities were striking.
To quote Jarrige:
In spite of some obvious differences . . . the full setting of the farming economy at Mehrgarh displays evident similarities with what had been noticed in the case of the early Neolithic settlements in the hilly regions forming the eastern border of Mesopotamia [that is, the Zagros mountains of Iran] . . .
The similarities noticed between Neolithic sites from the eastern border of Mesopotamia to the western margins of the Indus Valley are significant . . . A sort of cultural continuum between sites sharing a rather similar geographical context marked with an also rather similar pattern of evolution and transformation becomes more and more evident.
In other words, archaeological evidence comes to the same conclusion as ancient DNA evidence: there is a strong connection between the Zagrosians and the people of the Harappan Civilization region, dating back to a period when agriculture was only beginning in Mehrgarh. This now brings us to the third and equally important evidence linking the two regions: language.
Proto-Zagrosian and Proto-Dravidian
What languages did the west Asians speak? We can confidently answer this question only from the time when languages were written down, beginning sometime after 3500 BCE. Many Mesopotamian languages were written in the cuneiform script, until it fell into disuse in the early centuries of the Common Era. The languages that were written down using cuneiform include isolates such as Sumerian, Elamite, Hattic, Hurrian and Urartian; Semitic languages such as Akkadian, Eblaite and Armorite; and the Indo-European languages Hittite and Luwian. All these languages became extinct long ago and are not spoken anywhere today (note that the phrase Mesopotamian Civilization covers multiple kingdoms, city states and empires over thousands of years, not just one kingdom or empire).
The ancient DNA study said that the Zagrosian population and the subcontinent’s First Indians had to have mixed at least by between 4700 BCE and 3000 BCE. The study did not mention the earliest possible date for the admixture. But we know that the Mehrgarh settlement, with its close connections to people from the eastern border of Mesopotamia, began around 7000 BCE. So if we are to arrive at a period of the migration, or the arrival of the Zagrosian people in south Asia, it would be sometime between 7000 BCE and 3000 BCE. So which of the Mesopotamian languages listed above are they likely to have brought to south Asia?
A serious candidate for this would be Elamite, the language of a people who built one of the most powerful and intermittently persistent kingdoms in the area covering central and southern Zagros mountains, as also the plains of Khuzistan and the Persian Gulf coast in today’s Iran. The first Elamite kingdom came into being sometime around 2700 BCE, but as we have seen, the migrations from Zagros to south Asia happened much earlier. The migrants, therefore, could not have been products of the urban civilization of Elam – they preceded it. They may have spoken a Proto-Elamite language. Depending on the period of their migration, it is also possible that they were not full-fledged agriculturists yet, but herders who may, nonetheless, have had some exposure to the newly emerging culture of farming.
As we saw in chapter 2, the earliest evidence for domestication of goats comes from Ganj Dareh in the central Zagros mountains of Mesopotamia, from around 7900 BCE (the same archaeological site from where the ancient DNA was collected for the genetic study). We also saw that Mesopotamians had started migrating with their animals or goats even before the full agricultural package – cultivation of barley and wheat along with some lentils and raising sheep, goats and cattle – had become common in the region. Therefore, it is quite possible that the group of Proto-Elamite-speaking people from the south or central Zagros region who migrated to Balochistan and the surrounding areas such as Mehrgarh may have been herders with some exposure to the newly emerging practice of farming.
Some of these Zagrosian migrants who reached south Asia may have remained herders to this day, like the Brahuis of Balochistan. They speak the Brahui language, which has been linguistically determined to be closely connected to Elamite. Other migrants from the Zagros may have settled down in places like Mehrgarh to become farmers, and both the herders and the farmers may have mixed with the existing population – the First Indians – at some point. The First Indians themselves had started experimenting with agriculture in places like Lahuradewa in the Middle Ganga plain by then as we saw earlier and they may have been doing so in Balochistan as well.
So the best guess we can make based on archaeological and genetic evidence is that a population of herders from the southern or central Zagros region, speakers of Proto-Elamite or a related language, migrated to south Asia sometime after 7000 BCE, mixed with the First Indians and this new, mixed population sparked an agricultural revolution in the north-western region of India and then went on to create the Harappan Civilization over the next few millennia.
If this reconstruction of history based on archaeology and genetics is correct, then Proto-Elamite must have left a significant mark on the linguistic history of the subcontinent. Could it be Brahui with its proven links to Elamite? To some extent, yes. But considering the geographical size and demographic weight of the Harappan Civilization, there should be a much larger footprint. Could there be one? Yes, there could be: Dravidian languages, spoken by nearly a fifth of the population of India. Somewhere between 250 million and 300 million people speak a Dravidian language in the world today, making it one of the top six language families. So to this we turn now: linguistic research that shows the close connection between the Dravidian languages of India and Elamite.
A paper published in 2013 titled ‘South Asia: Dravidian Linguistic History’, authored by two linguists who have specialized in Dravidian languages, Professor Franklin C. Southworth and Dr David W. McAlpin, is a good place to start. McAlpin had earlier reconstructed3 the vocabulary of Proto-Dravidian (PD), the ancient language from which all Dravidian languages – not just Tamil, Malayalam, Kannada and Telugu, but also languages such as Gondi and several others – descend, and the 2013 study uses this early vocabulary to make deductions about where and when the language may have been spoken. This is how the paper’s section on PD begins:
The reconstructed vocabulary of PD reflects a society engaged in animal husbandry, with some knowledge of agriculture. Words for sheep, goat and cattle, all inherited from Proto-Zagrosian4 [the parent language of Proto-Elamite], along with verbs referring to ‘driving’ and ‘grazing’ animals, words for ‘herd’, ‘flock’, ‘shepherd’ and several words which mean both ‘house/dwelling’ and ‘animal/stall’, indicate the importance of herding. No specific [words for] grains are reconstructible, although reconstructed agricultural terms include words for
digging and digging tools, operations such as winnowing, churning, reaping, and grinding grain, along with several words meaning ‘grain’ or ‘seed’, ‘chaff’ and ‘husk’, and possibly a word for the plough. The only reconstructible food plant names are onion/ garlic, yam, and eggplant, some of which may be later borrowings into individual Dravidian languages from local sources. Thus it is unlikely that these people were sedentary farmers at this stage. Several different land types are distinguished, including low lying land, uncultivated land, and field/open space. In general, older Elamite resembles PD, younger Elamite resembles Brahui. Thus we may posit, as a starting point, that the Proto-Dravidian community probably moved from somewhere in the south of present-day Iran.
The reconstructed Proto-Dravidian vocabulary enables us to put an upper limit on when Proto-Zagrosian and Proto-Dravidian separated – not before animals such as sheep and goat were domesticated and the early forms of agriculture had begun, since the two languages share the terms for domesticated animals and also for processing grains.
Interestingly, it also puts a lower limit on the separation: not after writing was invented. The reasoning is straightforward: the word ‘tal’ exists in both Elamite and Proto-Dravidian, but with different meanings. In Elamite, it means ‘write’; in Proto-Dravidian, it means ‘push in’. In a paper McAlpin authored in 1981,5 he explains that the original meaning of ‘tal’ was to ‘push in’, but since cuneiform writing on clay tablets involved ‘pushing in’, the word came to mean ‘writing’ in Elamite after writing was invented. The reason why Proto-Dravidian speakers retained the old meaning of ‘tal’ is that they separated from the Elamites before writing was invented, and therefore the word never acquired the meaning of writing (the word for ‘writing’ in Dravidian languages comes from ‘drawing’ or ‘paint’).