The Indus Valley and the Question of Hindu Origins

In 1928, in the brick ruins of Mohenjo-daro, a square of soapstone the size of a postage stamp came out of the ground. On it, cut in shallow lines more than four thousand years old, sat a figure. Horned, legs folded, surrounded by a tiger, an elephant, a rhinoceros and a buffalo. The man who ran the excavation, John Marshall, looked at it and saw a god he recognised. He saw Shiva.

If he was right, it would mean that the great ascetic of Hinduism was being worshipped on the banks of the Indus a thousand years before the oldest Vedic hymn was composed. It would push the tradition’s beginning back into the third millennium BCE and root it in Indian soil, older than the Sanskrit that carries it.

That is the prize. It has hung over the question of Hindu origins for almost a century. And the honest answer, the one this article will earn slowly, is that we still cannot claim it.

The civilisation that forgot to leave us its gods

Begin with what is solid, because a great deal is. The Indus or Harappan civilisation was real, vast, and for most of recorded history completely forgotten. Its mature urban phase ran from about 2600 to 1900 BCE, fixed by radiocarbon dating, which makes it a contemporary of the pyramids and of Sumer. It was the largest of the early Old World civilisations, more than a thousand known sites spread across roughly a million and a half square kilometres, with planned cities at Mohenjo-daro, Harappa, Dholavira and Rakhigarhi.

These were engineers. They laid streets on a grid. They built granaries, public wells, and a drainage system so thorough that waste water ran from private houses into covered municipal drains, a standard of urban sanitation that much of the world would not match again for thousands of years. They standardised their weights to a precise binary scale. They traded with Mesopotamia by sea.

And then there is the strangeness. For a civilisation of this scale, it is almost silent about its rulers and its gods. No great royal tombs. No palaces that announce themselves. No towering temple, no colossal idol, nothing like the god-kings of Egypt carved a hundred times life-size. What it left instead are small things. Thousands of carved seals, most no bigger than that first one. Terracotta figurines. A great watertight tank at Mohenjo-daro that the excavators named the Great Bath. And a script, appearing on the seals in lines a few signs long.

We can see their plumbing. We cannot see their prayers.

The seal that everyone wants to read

Marshall published his identification in his 1931 report, and he did not hedge. The seated posture, he argued, was a yogic one, and the lord of beasts ringed by animals could only be an early form of the great Hindu god of yoga and the wild.

“The figure is in a typical yoga attitude… and Siva was and still is regarded as a mahayogi, the prince of Yogis… he is surrounded by animals, and Siva is par excellence the Lord of Animals, Pasupati.”

It was a magnificent leap, and it was contagious. Around that one seal a whole architecture of continuity was built, and it is worth stating in its strongest form, because the strongest form is what has to be answered. The figurines of broad-hipped women became an early cult of the Mother Goddess. Certain conical and ring-shaped stones became early forms of the liṅga and yoni, the aniconic symbols of Shiva and his consort. The Great Bath became ritual bathing, the ancestor of the pilgrim stepping into the Ganges. The pipal tree, sacred in later Hinduism, appears on the seals. Laid end to end, these resemblances seem to draw a single unbroken line from the Indus drain to the temple tank, four thousand years without a gap.

The tradition itself often tells the story this way. It is a satisfying story. It may even be partly true.

The one fact that governs everything

Here is the fact that the satisfying story has to walk around. We cannot read a single Indus sentence.

The script has more than four hundred signs. It has resisted more than a century of decipherment by gifted people, and there is no scholarly consensus on what language it even records. The reasons are brutal and simple. The inscriptions are extremely short, often four or five signs, the length of a name on a luggage tag. There is no bilingual text, no Rosetta Stone pairing the unknown script with a known one. And the underlying language is unknown, so there is nothing to sound the signs against.

This is not a footnote. It is the hinge of the entire question. Because we cannot read the Indus people in their own words, every claim about what they believed is an inference from objects. And every one of those inferences reaches across a gap of one or two thousand years to compare an Indus image with a Hindu practice attested only much later. The method is a chain of resemblances, and a chain of resemblances is exactly as strong as its assumption that things which look alike are connected.

Pull that assumption and watch the architecture sway.

What the skeptics saw

The doubts did not come from people hostile to the tradition. They came from people looking harder at the seal.

In the 1970s the art historian Doris Srinivasan examined the famous figure again and argued that the “three faces” Marshall saw are better read as the ears and muzzle of a bull, that the image may be a divine buffalo-man rather than a proto-Shiva at all. Others questioned whether the figure is even male. The seated posture, read as yoga, is also simply how a person sits on the ground without a chair. The Indologist Wendy Doniger, surveying the claim, noted dryly that the figure may be no more than someone sitting cross-legged with a prominent loincloth knot. The Finnish scholar Asko Parpola, who has given his life to the script, reads the figure instead against Proto-Elamite seated-bull imagery and a buffalo-horned deity.

And the other resemblances soften under the same pressure. The “Mother Goddess” figurines may be toys, or household charms, or ancestors, the way clay figures are used in countless cultures with no temple cult behind them. The “linga” stones may be gaming pieces or weights. The Great Bath may be a civic bath. Each individual reading might be right. The problem is that not one of them can be confirmed while the script stays shut, and a tower built entirely of unconfirmed readings is not yet history.

Resemblance is not descent. Descent is not the same as shared meaning. A symbol can survive five thousand years and mean something entirely new at the end of it.

The honest map

So where does that leave the question we started with. It leaves it open, and open is a real answer, not an evasion.

Some threads probably do run from the Indus into later Indian religion. Reverence for certain trees, the sanctity of bathing and water, perhaps some early form of a goddess: these are plausible, and they are unprovable on present evidence. They are the kind of thing a careful person can believe and cannot demonstrate. The larger claim, that the Indus people were already Hindus, worshipping Shiva, practising yoga, reading their own liṅga the way a priest at Kashi reads one today, runs far ahead of anything the seals can carry while they cannot be read.

This is the ordinary condition of deep history. The evidence is partial. The honest historian says which parts are firm and which are hope, and keeps the two from bleeding into each other. To insist on certainty here, in either direction, the nationalist who needs the Indus to be Vedic or the skeptic who needs it to be alien, is to ask the stones to say something they have not yet said.

The tradition does not need the Indus cities to be old. It is old enough without them, demonstrably, from the moment the Vedic hymns begin and we can finally read the words. What the Indus offers is not a proven cradle. It is a magnificent, locked first room, standing at the threshold of the story, full of objects we can describe forever and cannot yet hear.

A key may come. A decipherment, if it is ever achieved, could change this chapter overnight and settle in a week what a century has not. Until then the right posture is the patient one.

The seals can be described. They cannot yet be read.