Not so forgotten, and not a sudden end
The Indus civilization is often called forgotten, and in a literal sense it was: for roughly three thousand years nobody could read its seals or name its cities, and it was identified by archaeologists only in the 1920s. Today it is among the best-studied Bronze Age civilizations. The same caution applies to its end. The cities did not simply vanish overnight. Across several centuries after about 1900 BCE, large urban centres shrank or were abandoned, trade and craft specialisation weakened, and the writing on seals dropped out of use. People continued to live in the region, mostly in smaller villages.
Scholars therefore tend to speak of transformation or deurbanisation rather than collapse. No single explanation is accepted by everyone, and the evidence differs from place to place.
What changed
The archaeological record from about 1900 BCE, often called the beginning of the Late Harappan period, shows several trends:
- Many large cities, including Mohenjo-Daro, show decline and then abandonment, though the exact timing differs.
- The seals and standardised weights of the mature period become scarce.
- Settlement shifts eastwards toward the Ganges plain and southwards toward Gujarat and beyond.
- Long-distance trade, including contacts with Mesopotamia, appears to have reduced.
Scholars lack a fine-grained chronology, and radiocarbon dating gives ranges of decades or centuries, so it is hard to tell whether changes happened in a single generation or over many.
Climate and monsoon change
A leading idea is that the climate became drier. Evidence from lake sediments, cave deposits and other records suggests that the summer monsoon weakened during the later third and early second millennium BCE, though the dates and regional effects are debated. Some researchers connect this with a widely discussed dry episode around 2200 BCE, sometimes called the 4.2 kiloyear event, although its dating, extent and effect on South Asia are contested.
A drier climate would have reduced harvests in the zones dependent on seasonal rain or flooding, and could have pushed farmers toward better-watered areas. Critics point out that the Indus economy was diverse and included summer and winter crops, which may have made it more resilient than a single-crop system. Climate is therefore usually treated as a stress rather than a sole cause.
Rivers that moved
Another set of proposals concerns the rivers. The Ghaggar-Hakra system, a seasonal river channel in north-west India and Pakistan, is associated with a dense cluster of Indus-period settlements. Some researchers argue that this river lost water over time, perhaps because tributaries shifted away or because monsoon flow decreased, which would have undermined nearby settlements. The Indus itself is also known for changing course and flooding.
Not all scholars agree about the Ghaggar-Hakra story. Whether the channel was ever a large, perennial river during the mature Harappan period is disputed, and the identification with the Sarasvati of later texts is a separate and contested question. Shifts in course and flooding certainly caused local damage, but how far they explain a region-wide change is uncertain.
Trade, economy and social change
Some scholars look to economics. Mesopotamian demand for Indus goods may have weakened in the early second millennium BCE, hurting coastal and craft centres. Others suggest that the integrating institutions, whatever they were, lost influence, and that people turned to regional networks and smaller communities. These ideas are hard to test, because we do not know how the Indus economy was run, and because readable written records are missing.
Disease has also been proposed, for example epidemics in crowded towns, but direct evidence is limited and the idea remains speculative. Similarly, some mention environmental harm from deforestation or overuse of land, but this is difficult to prove for a specific time and place.
The invasion idea and why it was dropped
In the middle of the twentieth century, Mortimer Wheeler proposed that the cities were attacked and destroyed by Indo-Aryan invaders, and he pointed to references to forts in the oldest Vedic hymns, and to human remains in the upper levels of Mohenjo-Daro. This idea became popular and appeared in textbooks.
It has since been largely abandoned for several reasons:
- The skeletons at Mohenjo-Daro come from different levels and times, and do not form a single massacre layer. George Dales pointed this out in the 1960s.
- There is no archaeological sign of a large invading army, burnt cities across the region, or a sudden cultural replacement.
- The Vedic period chronology puts the composition of the hymns later than the decline, and the texts themselves are not records of conquering an Indus city.
- Decline appears gradual and spread over many places and centuries.
Questions about migration and language remain separate and are discussed by scholars using linguistics, archaeology and ancient DNA. The relationship between the Indus people and later speakers of Indo-Aryan languages is debated, but nothing supports a simple story of destruction by invaders. The Vedas and Upanishads belong to a later cultural world, and linking them to Indus cities requires care.
A combination of causes
Most researchers now prefer a multi-factor explanation: a weakening monsoon, shifting rivers and local flooding, changes in trade, and social choices, together leading people to move away from the largest cities and towards smaller rural communities. The balance among these probably varied between regions, and that is part of why a single answer is elusive.
What remains uncertain
Without a deciphered script, we cannot read any account of what residents thought. Deeper study of settlement patterns, climate records and DNA continues to refine the picture, and future results may favour some explanations over others. For now, the safest summary is that the Indus cities declined slowly from about 1900 BCE through a mix of environmental and social change, that the people and many of their practices did not disappear, and that the invasion story should be treated as a discredited chapter in the history of the subject.