The Lost Civilization of the Indus Valley

In 1920, a railway engineer in Punjab named Daya Ram Sahni was looking for ballast to use under tracks. He went to a mound the locals called Harappa, where ancient brick had been quarried for centuries to support the British rail network. He pulled out a brick. It was kiln-fired, perfectly proportioned, and clearly very old. He pulled out another. And another.

By the time the British colonial archaeologists arrived in earnest, hundreds of miles of railway had been built on bricks recycled from a civilization no one knew existed. What Sahni had stumbled across was the second-largest urban culture of the Bronze Age — larger in geographical extent than Egypt and Mesopotamia combined — and it had been entirely forgotten for nearly four thousand years.

A civilization built on a grid

The Indus Valley Civilization (IVC) flourished from roughly 2600 BCE to 1900 BCE across what is now Pakistan and northwestern India. At its peak, it included over a thousand settlements and at least five major cities — Harappa, Mohenjo-daro, Dholavira, Lothal, Rakhigarhi — connected by trade and shared culture.

The cities were planned. Streets ran in a grid pattern, oriented to the cardinal points, with main thoroughfares as wide as ten meters. Houses had multiple storeys, baked-brick foundations, and — most strikingly — connected to a citywide drainage system. Every house had a private bathroom or latrine, with water flowing into covered drains under the street, draining into soakaways outside the city walls. No other Bronze Age civilization built urban sanitation on this scale. Rome would not match it for another two thousand years.

The Great Bath at Mohenjo-daro, a forty-square-meter brick-lined pool sealed against leakage with bitumen, suggests ritual or communal bathing on a scale that has no parallel in the era. Whether it was religious, hygienic, civic, or all three, we do not know. We have no temples that look like Mesopotamian or Egyptian temples. We have no royal tombs. We do not even know what their gods looked like.

The mystery of the missing king

This is what makes the IVC strange. Every other Bronze Age civilization left enormous monuments to its rulers. The Egyptians built pyramids. The Sumerians built ziggurats inscribed with kings' names. The Chinese built Erlitou with palace platforms. The Indus people built — granaries, drainage systems, public baths, and surprisingly modest houses, almost all the same size.

There are no royal palaces in the IVC. No giant statues of rulers. No tombs filled with treasure. The Mohenjo-daro "Priest-King" statue, the most famous IVC artifact, is only nineteen centimeters tall and was probably not a priest or a king at all — that name is a 1920s archaeologist's guess. The artifact's actual function is unknown.

One reading of the evidence: the IVC was unusually egalitarian, governed by some form of council or guild rather than a king. Another reading: the elites simply expressed status differently than we are used to recognizing. We do not know which is right because we cannot read their writing.

The undeciphered script

The Indus people wrote. They wrote a lot. Over four thousand inscribed objects have been found — seal impressions, copper tablets, pottery shards, even small stone signboards. The script has roughly four hundred distinct signs, suggesting a logo-syllabic writing system somewhere between an alphabet (too few signs) and a pure logographic system (too many).

And we cannot read it.

We have no Rosetta Stone. We have no parallel inscription in a known language. We have no idea what language family the script encodes — Dravidian and Indo-European are both candidates, but the evidence is not strong either way. Most inscriptions are short, just five or six signs, which makes statistical decipherment hard. Modern computational linguists have applied entropy analysis, n-gram frequency, comparison with known undeciphered scripts, and machine learning. We have made some progress — the script almost certainly encodes a language, not just commercial tallies — but not enough to read it.

This is the great unsolved mystery of South Asian prehistory. An entire civilization, one of the most populous of its era, has left behind tens of thousands of words and we cannot extract a single sentence.

Trade across the ancient world

What we can read are the routes. Indus seals have been found in Mesopotamian cities — Ur, Kish, Susa — and Mesopotamian beads in Indus sites. Sumerian texts mention a place called Meluhha, somewhere east, that exported lapis lazuli, carnelian, copper, gold, and ivory. Most scholars now believe Meluhha was the Indus Valley.

The trade was sophisticated. The IVC produced standardized weights — small cubes in a binary system, accurate to fractions of a gram — used across the entire civilization for over half a millennium. They produced cotton textiles thousands of years before any other civilization. They produced exquisite carnelian beads using a drilling technique so refined that modern researchers initially thought it was impossible without diamond drills.

The Lothal site in Gujarat has what looks like the world's oldest dockyard, a brick-lined basin connected to a tributary of the Sabarmati river, dated to around 2400 BCE. From Lothal, ships sailed to the Arabian Peninsula and beyond.

The puzzle of the collapse

Around 1900 BCE, the cities began to empty. The drainage systems stopped being maintained. The trade networks frayed. By 1700 BCE, urban life in the Indus Valley had ended. Smaller villages persisted, but the great cities were abandoned. Thousands of years later, when farmers cleared fields and kicked up bricks, no one remembered who had made them.

The cause is debated. The old hypothesis — Aryan invasion — has been largely discarded; the genetic and archaeological evidence does not support a violent overthrow. The current consensus points toward climate change. The Indian summer monsoon weakened around 2000 BCE. The Sarasvati-Ghaggar river system, on whose banks many IVC settlements sat, dried up. Agriculture became unsustainable.

People did not vanish. They moved east, into the Ganges plain, where the monsoon was still reliable. The cultural memory of the cities seems to have faded within a generation or two. The IVC simply dissolved.

What it means

The Indus Valley is a reminder of how much of the deep past we have lost. We know the Egyptians and Sumerians intimately because their literate traditions survived through Greek and Roman scholars. The IVC had no such inheritor. Its language was forgotten. Its writing went unread. Its cities became brick quarries.

And yet what they built — the grid streets, the sewers, the standardized weights, the long-distance trade — was not less than what their better-remembered contemporaries built. In some respects it was more. The Indus Valley shows that elaborate, peaceful, mercantile urban civilization is older and more widely distributed than the textbook narrative suggests. It also shows that being remembered by history is not the same as having mattered. There were probably many such civilizations. We have lost most of them.

Every now and then a new tablet comes out of the dust, or a satellite survey catches the outline of an unexcavated city under a wheat field, and the dotted line on the map gets a little more solid. The work continues. Maybe the script will yield. Maybe a new site will hold the bilingual inscription that finally lets us read four thousand years of silence.

Or maybe it will stay quiet. The brick remembers; the speech does not.