The Indus Valley: a great civilization we still cannot read

The Indus Valley civilization has no known name for itself. It has no known kings, no known cities' names, and no story told in its own words. We have its streets, its drains, its bricks, its beads, and thousands of small stone seals covered in signs. We cannot read those signs. That fact shapes everything we know about this civilization, and everything we do not. Start with a word that matters for the whole year. A civilization, in the way historians use the term, is a society with cities, specialized work, and complex organization. The Indus civilization fits that definition well. It is also called the Harappan civilization, after Harappa, the first of its cities to be excavated. Both names are modern. The people who lived there left us no record of what they called themselves. The story of how we learned about them begins with accident and neglect. In the 1820s and 1830s, a British soldier turned explorer named Charles Masson traveled through the Punjab, in what is now Pakistan. He described large mounds of earth and broken brick near a village called Harappa. He did not know what they were. Later, Alexander Cunningham, the first Director General of the Archaeological Survey of India, visited the site several times between the 1850s and 1870s. Cunningham was a serious scholar. In 1875 he published a small stone stamp seal from Harappa. It showed a bull-like animal with a single horn and a line of strange signs above it. Cunningham did not understand how old it was. He thought it might be foreign, or from a much later period. He had in his hands a piece of a civilization more than four thousand years old, and he could not place it. Meanwhile, the site was being destroyed. A source here is not an ancient text but a historical fact about the nineteenth century. When the British built the railway line between Lahore and Multan in the 1850s, contractors needed ballast. Ballast is the layer of crushed stone or brick laid under railway tracks to hold them steady. The ancient baked bricks of Harappa were ideal. They were hard, uniform, and plentiful. Workers carried them off in huge quantities. Part of the ancient city of Harappa now lies under a railway line. Historians still regret this loss, because it removed much of the evidence we would most like to have. Systematic work came in the twentieth century. John Marshall became Director General of the Archaeological Survey of India and pushed for proper excavation. Two Indian archaeologists did much of the first hard work. Daya Ram Sahni began digging at Harappa in 1920 and 1921. R. D. Banerji excavated at Mohenjo-daro, in the province of Sindh, between 1919 and 1922. Mohenjo-daro means something like "mound of the dead" in the Sindhi language, though that is the modern local name and not the ancient one. Banerji found seals like the one Cunningham had published. Soon it became clear that the two sites, about 600 kilometers apart, shared the same type of objects, the same style of bricks, and the same script. In September 1924, Marshall announced the discovery in the Illustrated London News, a popular weekly magazine. He asked scholars to help. The reply came quickly. Specialists in Mesopotamia, the land between the Tigris and Euphrates rivers in modern Iraq, recognized similar seals from their own excavations. Those Mesopotamian finds could be dated, because Mesopotamia had written records and layered sites. The comparison showed that the Indus culture belonged to the third millennium BCE. A millennium is a period of one thousand years. So the Indus civilization was roughly as old as the great early cities of Mesopotamia and the pyramid age of Egypt. Until 1924, almost no one had guessed this. Many people had assumed that India's first cities arose much later. Work continued through the century. Ernest Mackay dug at Mohenjo-daro. Mortimer Wheeler, who later became one of Britain's best-known archaeologists, worked at both Harappa and Mohenjo-daro in the 1940s. He was a strong organizer and introduced more careful methods of recording layers in the soil. Then came Partition in 1947, when British India was divided into India and Pakistan. The border cut through the ancient civilization. Mohenjo-daro and Harappa fell in Pakistan. But many other sites lay in India, and Indian archaeologists set out to find them. They excavated Lothal in Gujarat, Kalibangan in Rajasthan, and Rakhigarhi in Haryana. R. S. Bisht began digging at Dholavira, on an island in the Gulf of Kutch, in 1990. In Pakistan, teams worked at Ganweriwala and elsewhere. At Harappa itself, modern excavations were led by George Dales, then by Richard Meadow and J. Mark Kenoyer. Each generation has added detail, and the picture keeps changing. Now the dates. Archaeologists divide the story into phases. A phase is a stretch of time defined by typical styles of pottery, buildings, and tools. The Mature Harappan phase, when the cities were at their height, is generally dated from about 2600 to 1900 BCE. Before that came the Early Harappan phases, when villages and small towns grew and trade networks formed. Farming in the region goes back much further. At Mehrgarh, in the hills of Baluchistan in Pakistan, farmers were growing wheat and barley and herding animals from about 7000 BCE. After the Mature phase comes the Late Harappan phase, when the large cities faded and life changed. We will look at that change in the paired lesson today. Here, keep in mind that these dates are estimates based mainly on radiocarbon dating and on comparison with other regions. Radiocarbon dating measures a type of carbon in once-living material, like charcoal or bone, to find its age. It gives a range, not an exact year. The size of the civilization is striking. Its sites spread across much of modern Pakistan and northwest India. There was even a distant outpost at Shortugai in northern Afghanistan, probably linked to the trade in a blue stone called lapis lazuli. Together the area covered perhaps a million square kilometers. That makes it one of the largest cultural zones of the Bronze Age. The Bronze Age is the period when people made tools mainly from bronze, an alloy, meaning a mixture, of copper and tin. Neither ancient Egypt nor Mesopotamia in its early centuries covered so wide a region with such a uniform material culture. Many sites lie along the Indus River and its tributaries. Many others lie along a river system that is now dry, called the Ghaggar-Hakra. In the Bronze Age it may have carried water in some seasons, though how much and for how long is still debated. This matters because rivers were the lifeblood of these societies. They gave water, fertile silt, and routes for transport. The major urban centers include Mohenjo-daro, Harappa, Dholavira, Ganweriwala, and Rakhigarhi. How many people lived in them is uncertain. Estimates for Mohenjo-daro have ranged widely, from tens of thousands downward, and recent scholars tend to prefer lower figures. The honest answer is that we do not know. The cities were large for their time, but we should avoid inflated numbers. Next, the cities themselves. The first thing that strikes any visitor to Mohenjo-daro is order. Streets were laid out in a grid-like pattern, with main roads running in straight lines and smaller lanes branching off. The builders used baked bricks of standard sizes. The proportions were often close to 1:2:4, meaning a brick's thickness, width, and length stood in that ratio. Such standardization across hundreds of kilometers suggests shared rules or shared craft traditions. Baked brick is also durable in a wet climate, and the Indus floods were a real danger. Many buildings stood on raised platforms, perhaps for protection against high water. Then there is the water. Houses commonly had a bathing area, with a floor of brick, and a drain that led to a covered channel under the street. Wells were common, and in Mohenjo-daro there were hundreds. Drains ran along the main streets, with inspection holes for cleaning. For a Bronze Age city, this is unusual. Many ancient cities had far less planned drainage. We should be careful, though. Having drains does not tell us who organized them, or whether a central authority did. That question belongs to the next lesson. On the western side of Mohenjo-daro stands a raised area called the citadel. A citadel is a fortified or elevated part of a city, often holding its most important buildings. On this mound lies the structure known as the Great Bath. It is a large rectangular pool, about 12 meters long, lined with fine brick and made watertight with a layer of natural tar called bitumen. Stairs lead down into it, and rooms surround it. Many scholars think it was used for ritual washing, like water tanks later found beside Indian temples. That is a reasonable guess, but it is only a guess. No text tells us what it was for. Dholavira shows another side of Indus engineering. It lies in a dry, harsh region, on an island in the Gulf of Kutch. The city's builders created a system of large reservoirs, cut into rock and lined with stone, to catch and store rainwater. This was a response to a hard environment. It shows that Indus people were not simply farmers using easy river land. They solved local problems with local technology. Now the crafts, which are among the most beautiful things the civilization left. Indus workers made beads from carnelian, a reddish stone. Some beads were long and tubular and drilled with great skill. Some were decorated with white designs made by etching, using an alkali treatment and heat. They made faience, a glassy material formed from ground quartz and fired to produce bright blue or green surfaces. They carved shell into bangles and inlay. They made tools and vessels of copper and bronze, and clay figurines of animals, carts, and human figures. They made distinctive pottery, often red with black painted patterns. Two famous objects deserve a caution. One is a small carved stone bust from Mohenjo-daro, with a beard and a patterned cloak, known as the "Priest-King." The other is a small bronze statuette of a slim young woman with her hand on her hip, called the "Dancing Girl." Both names are modern inventions. We do not know that the bearded man was a priest or a king, or that the bronze figure was a dancer. These names show how easily we fill silence with our own stories. The Indus economy rested on farming. People grew wheat and barley, and in some regions millet and rice. Cotton was cultivated, and Indus people were among the earliest to use it for cloth. They kept cattle, sheep, goats, and water buffalo. The production of surplus food, meaning more than the farmers themselves needed, made it possible for others to become full-time craftworkers and traders. Evidence for regulated exchange comes from small stone weights. These are cubes, often of chert, a hard stone, and they come in a graded series, with each weight a simple multiple of a smaller one. They appear across the whole region. Such uniformity suggests that merchants and officials agreed on how to measure goods. Again, though, we know the method but not the institutions behind it. Trade reached far beyond the valley. Indus carnelian beads and Indus seals have been found in Mesopotamia, in the Gulf region, and in Oman. This is hard physical evidence of contact across the Arabian Sea and the Gulf. Written evidence comes from the other side. Mesopotamian texts of the Akkadian period, around 2300 BCE, and later, mention a distant land called Meluhha. These texts record ships and goods coming from there, including carnelian, and later sources mention a community of Meluhhan speakers. Most scholars identify Meluhha with the Indus region. Not all do, and the identification rests on matching the goods and the geography, not on any Indus text naming itself. At Lothal in Gujarat, archaeologists found a large brick basin connected to a channel. Some have called it a dockyard for ships. Others argue it was a water reservoir or irrigation tank. The debate continues. Now we reach the heart of the matter, the script. Around four to five thousand inscribed objects are known. Most are stamp seals, small squares of a soft stone called steatite. Each usually shows an animal, carved with great skill, and above it a short line of signs. The animal is often the one-horned creature that puzzled Cunningham, which scholars sometimes call the "unicorn," though it may just be a bull shown in profile. Other seals show bulls with humps, elephants, and rhinoceroses. There are also small tablets, graffiti scratched on pottery, and inscribed copper tablets. At Dholavira, archaeologists found what they call a "signboard," a set of ten large signs that may once have been mounted above a gateway. It is the largest Indus inscription in terms of sign size. Here is the first great problem. The inscriptions are extremely short. The average is about five signs. The longest on a single surface has about seventeen. Compare this with Mesopotamian cuneiform, the wedge-shaped writing pressed into clay, which gave us long legal contracts, letters, and myths. Or with Egyptian hieroglyphs, which covered temple walls. The Indus script gives us nothing of that kind. No long text, no letter, no story has been found. The number of different signs is about four hundred, though counts vary depending on how scholars decide whether two similar signs are really the same. Iravatham Mahadevan, an Indian scholar, published a concordance in 1977. A concordance is a systematic list showing every sign and every place it occurs. His listed about 417 signs. That number is large. It is too large for an alphabet, where each sign stands for a sound, and there are usually fewer than fifty signs. It is too small for a pure picture-writing system with a different symbol for every word. It fits best with a mixed system, where some signs stand for whole words and some for syllables. That is the kind of system we see in early Mesopotamian and Egyptian writing. But this is an inference, not a proof. We do know the direction. On many seals, signs are crowded at the left end, as if the writer ran out of space. In some cases the signs are written over each other. That pattern suggests writing ran from right to left. Note that on a seal, the carved image is reversed when pressed into clay. So scholars have to take care about whether they are looking at the seal or at its impression. Why can we not read it? Four reasons stand out. The first is that there is no bilingual text. The Rosetta Stone, found in Egypt in 1799, carried the same decree in Greek, which scholars could read, and in Egyptian scripts. That key allowed Jean-François Champollion to decipher hieroglyphs in the 1820s. Nothing like it has been found for the Indus script. The second reason is that the texts are very short. Decipherment relies on patterns, and a few signs per object give few patterns to test. The third reason is that we do not know the language. Even if we could say what sound a sign stood for, we would not know what words those sounds made. The fourth reason is that no long continuous text has been found. These four problems together make decipherment extremely hard. Many have tried. The most widely discussed idea is that the language belonged to the Dravidian family. Dravidian languages are Tamil, Telugu, Kannada, Malayalam, and others spoken mostly in southern India today, plus Brahui, a small language spoken in Baluchistan. Iravatham Mahadevan argued for this, as did the Finnish scholar Asko Parpola and some Soviet researchers in the 1960s and 1970s. Parpola proposed readings using the rebus principle. A rebus uses a picture of one thing to stand for a word that sounds the same. In English, a drawing of a bee and a leaf could stand for "belief." Parpola noted that in early Dravidian, the word for "fish" sounded like the word for "star." So a fish sign might refer to a star, perhaps a god or a name. This is clever, but it depends on reconstructing a language from four thousand years ago, and the results cannot be tested against any known text. A second idea is that the language was an early form of Indo-Aryan, the family that includes Sanskrit. A few scholars, including N. S. Rajaram and N. Jha, have argued this. Mainstream scholars reject their claims. The reasons are technical. For example, the horse is central to early Sanskrit texts like the Rigveda, but there is no clear evidence for the domestic horse in Mature Harappan sites, and the readings proposed depend on choices that other experts do not accept. This topic is sometimes tied to modern political arguments about the origins of Indian culture. A good historian separates the evidence from the politics, and the evidence does not currently support these decipherments. A third possibility is that the language belonged to a family that has completely vanished, leaving no descendants. That is not unusual in history, and it may be the most cautious position. Several people have announced full decipherments, including Rao and others, but none has won acceptance. A real decipherment would need to produce readings that other scholars can check and that give consistent sense across many inscriptions. None has done so. There is also a deeper debate: does the script encode language at all? In 2004, Steve Farmer, Richard Sproat, and Michael Witzel published an article arguing that the signs may be non-linguistic. That would mean they were symbols, like heraldic emblems or religious marks, that carried meaning but did not record speech. They pointed to the short length of the texts, the lack of long inscriptions, and the lack of signs that look like a writing tradition developing over time. In 2009, Rajesh Rao and colleagues replied with a study using conditional entropy. Entropy here measures how predictable one sign is given the sign before it. They found that the Indus sequences had a degree of order similar to that of known languages, and different from random sequences or rigid systems. Sproat disputed their method and conclusions. The argument remains unresolved. If the signs are not language, then there is nothing to decipher in the usual sense. If they are, we have no key yet. John Marshall and others also wondered whether the Indus signs were related to Brahmi, the script used in India from around the third century BCE, from which most later Indian scripts developed. A gap of well over a thousand years separates the two, and no connecting evidence has been found. The link has not been established. This silence limits what we can say about Indus beliefs. Some objects look religious, and writers often claim more than they should. A famous seal shows a figure seated in a cross-legged posture, wearing a horned headdress and surrounded by animals. Marshall called it "proto-Shiva," linking it to the later Hindu god. That is suggestive but unproven. Many female terracotta figurines have been called mother goddesses. At Kalibangan, some brick-lined pits with ash have been called fire altars. Each of these ideas may be right, or partly right. None can be confirmed without texts. What, then, is the value of studying a civilization we cannot read? It forces us to be honest about evidence. We can describe what people built, made, and traded. We cannot say what they believed, how they ruled, what stories they told, or what they called themselves. A historian must hold both facts at once: much is known, and much is not. That is a different kind of history from the one we get when written records survive. It depends more on archaeology, the study of the material remains of past societies, and on careful, modest reasoning. The silence also reminds us that "civilization" does not depend on writing that we can read. The Indus people built cities and managed water on a great scale. They traded across seas and standardized weights and bricks across a vast area. They did all this and left us only their objects. Whether they chose not to write long texts, or wrote on perishable materials like palm leaf or cloth that have rotted away, is another open question. Both are possible. This leads directly to the next lesson. We have seen what the Indus cities looked like and what they left. We have not asked who ran them, or why the great cities faded after about 1900 BCE. Without readable texts, those questions must be answered from the ground itself: from the layout of houses, the absence of obvious palaces, and the changing climate and rivers. In the second lesson today, we will weigh that evidence and see what Harappa suggests about power, and how historians explain the end of a civilization that never told us its own story.