In September 1924 the archaeologist John Marshall told the readers of a London magazine that excavators in British India had found the remains of a civilization nobody knew had existed. Digs at Harappa in Punjab, begun in 1921 by Daya Ram Sahni, and at Mohenjo-daro in Sindh, begun in 1922 by R. D. Banerji, had turned up brick cities, carved seals and an unknown script. The first excavated cities, Harappa and Mohenjo-daro, lie in what is now Pakistan, and later finds stretched the map from Afghanistan to Gujarat and the outskirts of Delhi.

Cities on a grid

Archaeologists divide the civilization into early, mature and late phases. The mature, fully urban phase lasted from roughly 2600 to 1900 BCE, making it a contemporary of Old Kingdom Egypt and the city states of Sumer. Over a thousand settlements are known, and the area they cover is larger than Egypt and Mesopotamia combined. Mohenjo-daro may have held tens of thousands of people. Its streets ran roughly north to south and east to west, houses were built of fired bricks in standard proportions, and many homes had bathing platforms connected to covered drains that ran along the streets.

On a raised mound at Mohenjo-daro stands the most famous structure of all: the Great Bath, a watertight brick tank sealed with a layer of bitumen, reached by steps at each end. Most scholars think it served some ritual purpose, perhaps purification, though nobody can be sure. Other sites have their own surprises. Dholavira in Gujarat built huge reservoirs to catch scarce rain, and Lothal has a large brick basin that some read as a dockyard and others as a water tank.

Trade, weights and seals

Indus merchants worked with remarkable consistency across hundreds of miles. They used standardized cube-shaped stone weights that followed a fixed series of ratios, so a trader in Punjab and one in Gujarat could measure goods the same way. Their beads of carnelian, shell and other materials turn up far to the west. Mesopotamian texts mention a distant land called Meluhha, which most scholars identify with the Indus region, and an Akkadian cylinder seal names an interpreter of the Meluhhan language, a hint that the two peoples dealt with each other directly.

Thousands of small square seals, often showing an animal such as a bull or a one-horned beast together with a short line of signs, were probably used to mark goods or show authority. The signs are the great unsolved puzzle. No one has deciphered the Indus script: most inscriptions are only a few signs long, no bilingual inscription like the Rosetta Stone has turned up, and we do not know what language it records. Some researchers even doubt it writes a full language at all, though most think it does.

Who was in charge?

In Egypt and Mesopotamia, rulers left palaces, tombs and boasting inscriptions. The Indus cities are different. No grand palaces, royal tombs or monuments to kings have been found, and weapons are relatively scarce. Some scholars imagine rule by merchant elites or councils, others by priests, and others suggest powerful rulers simply left less trace. The uniformity of bricks, weights and town plans shows that some form of coordination existed. What form it took remains open, and it is worth noticing that a large society may have organized itself without the obvious marks of a god king.

Decline

After about 1900 BCE the big cities shrank, the script and standard weights fell out of use, and people moved east and south into smaller villages. In the mid twentieth century the archaeologist Mortimer Wheeler blamed invading Aryans, pointing to skeletons in the streets of Mohenjo-daro. Later study showed those bodies came from different periods and showed no signs of a massacre. Most archaeologists now link the decline to climate change, weaker monsoon rains and shifting rivers, which would have strained farming and trade. A river system called the Ghaggar-Hakra, once dotted with settlements, appears to have dried up. Decline was gradual and uneven, and many village traditions carried on into later Indian life.

Worked example: the "Priest-King"

One of the best known Indus objects is a small stone bust of a bearded man wearing a robe decorated with trefoil patterns, found at Mohenjo-daro in 1927. Excavators named it the "Priest-King", and the label stuck in textbooks and museum cases. Yet nothing about the object tells us the man was a priest, a king or both. The name came from the expectations of its finders, who knew that other ancient civilizations had priest kings and assumed this one did too.

Look at how the label shapes thought. Once a statue is called a priest king, it becomes "evidence" that the Indus cities had priest kings, even though the reasoning runs in a circle. A careful historian separates what the object shows (a man of some status, dressed in fine cloth, carved with skill) from what we bring to it. The same caution applies to the "unicorn" seals, the "Great Bath" and the "granary" at Harappa, whose use is still debated. The habit is useful far beyond archaeology: when a name does your thinking for you, ask who chose it and what they were assuming.