The Saraswati river and its desiccation are described in the ancient Indian works. In the RgVeda Samhita, the oldest, it is a mighty river. By the time of the Mahabharata, it has dried up.
The conventional theory is that the RgVeda Samhita was supposedly compiled by the newly arrived horse-and-chariot, Indo-European language bearing invaders/immigrants who imposed themselves and their language (Sanskrit) tracelessly on the indigenous population. The RgVeda Samhita as per the Harvard Sanskritist and invasionist Michael Witzel has at most 4% of its 10,000 word vocabulary of non-Indo-European origin. The comparable figure for ancient Greek is more than 30% - and the Indo-Aryans entered (and supposedly came to dominate) a much larger geographical area with a much larger population than ancient Greece.
The evidence however is that the Saraswati dried up well before this migration/invasion took place!
Michel Danino has a newspaper article about the latest state of research.
Maps (from elsewhere):
Main archaeological sites of the Harappan civilization:
Old Saraswati:
Reproducing the article here because the Daily Pioneer articles tend to vanish.
The now dried up Saraswati river holds the key to many riddles of ancient Indian history — from the fate of the Harappans to the identity of the Vedic people. A convergence of archaeological, geological and climatic studies may soon provide us some answers
The riddle of the Saraswati river never goes long out of public view. The fascination the lost river has exerted on Indian minds is understandable: Praised in the Rig Veda’s hymns as a “mighty” river flowing “from the mountain to the sea” somewhere between the Yamuna and the Sutlej, it is reported a few centuries later by the Brahmanas (commentaries on the Vedas) as disappearing in the desert at a point called Vinashana, which was then a highly revered pilgrimage site. The Mahabharata, whose great war is waged in the region of Kurukshetra watered by the Saraswati and its tributaries, paints a similar picture, adding some details about the broken-up westward course of the river all the way to Prabhasa on the Arabian Sea. The river went on dwindling down, eventually becoming ‘mythical’, finally relocated at the confluence between Ganga and Yamuna as an ‘invisible’ river — a convenient device to remember it.
A modern myth is that satellite imagery ‘rediscovered’ the river in the 1970s. Actually, it only confirmed what had been known for over two centuries: As early as in 1760, a map from The Library Atlas published by Bryce, Collier & Schmitz showed the Saraswati (spelt ‘Soorsuty’) joining the Ghaggar (‘Guggur’) in Punjab; indeed, even today a small stream called ‘Sarsuti’ seasonally flows there. In 1778, James Rennell, a noted English geographer and cartographer, published a Map of Hindoostan or the Mogul Empire with similar details. In the early 19th century, several topographers surveyed the bed of the Ghaggar, a seasonal river that flows down from the Shivalik hills, and found it much too wide for the paltry waters it carried during monsoons; the first scholar to propose that the Ghaggar-Saraswati combine was the relic of the Vedic Saraswati was the French geographer Louis Vivien de Saint-Martin, who authored in 1855 a massive Geography of India’s North-West According to the Vedic Hymns. Subsequently, nearly all Indologists, from Max Müller to Monier-Williams or Macdonell (and later Louis Renou) accepted this thesis. Geologists such as RD Oldham (1886) joined in, followed by geographers such as the Indian Shamsul Islam Siddiqi (1944) or the German Herbert Wilhelmy (1969).
The Indus Civilisation
The story of the Saraswati’s rediscovery would thus have ended long ago if archaeology had not sprung a major surprise by redefining its role in antiquity. In the 1920s, cities of the Bronze Age like Mohenjodaro and Harappa came to light; initial findings were limited to the Indus Valley and Baluchistan, but in 1941, the intrepid explorer Sanskritist Marc Aurel Stein conducted an expedition in the then Bahawalpur State — today’s Cholistan, a very arid region of Pakistan which is technically part of the Thar desert. The Ghaggar’s dry bed continues there under the name of ‘Hakra’, and had long been known to be dotted with numerous ruined settlements. Stein’s contribution, encapsulated in his paper titled ‘A Survey of Ancient Sites along the Lost Saraswati River’, was to show that some of those sites went back to Harappan times. So the Saraswati, too, had nurtured the ‘Indus civilisation’, which prompted a few archaeologists to propose the broader term of ‘Indus-Saraswati civilisation’.
Indeed, decades of further explorations both in India and Pakistan have established that the Saraswati basin was home to about 360 sites of the Mature Harappan Phase (the urban phase that saw cities thrive, from about 2600 to 1900 BCE). This includes settlements such as Bhirrana, Rakhigarhi, Kunal or Banawali (all in Haryana), Kalibangan (Rajasthan) or Ganweriwala (Cholistan) — altogether, almost a third of all known urban Harappan sites. (Gujarat was also host to over 300 of them, another indication that the term ‘Indus civilisation’ is something of a misnomer.)
Again, that the Ghaggar-Hakra was the Saraswati’s relic was accepted by most archaeologists, including Mortimer Wheeler, Raymond Allchin (both from Britain), Gregory Possehl, JM Kenoyer (both from the US), Jean-Marie Casal (France), AH Dani (Pakistan), BB Lal, SP Gupta, VN Misra or Dilip Chakrabarti (India).
The Aryan Issue
Despite the broad consensus, scholars such as Romila Thapar, Irfan Habib and the late RS Sharma started questioning this identification in the 1980s. What prompted this rather late reaction? It was a new development: A study of the evolution of the pattern of Harappan settlements in the Saraswati basin now revealed that in its central part — roughly southwest Haryana, southern Punjab and northern Rajasthan — most or all Harappan sites were abandoned sometime around 1900 BCE, a period coinciding with the end of the urban phase of the Indus civilisation. Clearly, the river system collapsed — which archaeologists now saw as a factor contributing to the end of the brilliant Indus civilisation.
Why was this a problem? We must remember that the Saraswati is lavishly praised both as a river and a Goddess in the Rig Veda, a collection of hymns which mainstream Indology says was composed by Indo-Aryans shortly after their migration to India around 1500 BCE. However, by that time, the Saraswati had been reduced to a minor seasonal stream: How could the said Aryans praise it as a ‘mighty river’, the ‘best of rivers’, ‘mother of waters’, etc? There is a chronological impossibility.
Hence, the objectors asserted, the Ghaggar-Hakra was not, after all, the Saraswati extolled in the Rig Veda. While some (Rajesh Kochhar) tried to relocate the river in Afghanistan, others (Irfan Habib) decided that the Saraswati was not a particular river but “the river in the abstract, the River Goddess”; but both theses ran against the Rig Veda’s own testimony that the river flowed between the Yamuna and the Sutlej.
However, what should have remained a scholarly issue now turned into an ideological and often acrimonious battle: On the one hand, those who stuck to the identity between the Saraswati and the Ghaggar-Hakra concluded that the composers of the Rig Veda must have lived in the region during the third millennium BCE at the latest — but as the only settlements known of that period were Harappan ones, they often held that the Harappans were part of the Vedic people; cultural evidence such as a Harappan swastika, yogic postures, figurines in namaste and more was pressed into service to bridge the Harappan and the Vedic worlds. On the other hand, scholars who continued to swear by an Aryan immigration in the mid-second millennium BCE, and therefore a pre-Vedic Harappan civilisation, accused the former of ‘chauvinism’, ‘jingoism’ or worse, conveniently forgetting that dozens of Western scholars had, for a century-and-a-half, accepted the same location for the Saraswati river.
New Research
Leaving aside the controversy, we now have scientific research combining geology and river studies. Satellite imagery is another useful tool, but cannot by itself date the numerous buried palaeo-channels (ancient waterways) it has brought to light; anyone can today access websites such as Google Earth and view the well-marked bed of the Ghaggar, but when did a perennial river last flow through it, and where did it draw its waters from?
Several recent studies have thrown new light on the ancient river, though sometimes with contradictory findings. Thus, in an article of April 2011 published in the noted magazine Science,
A Lawler claimed that “the Ghaggar-Hakra was at most a modest seasonal stream... from 2500 BCE to 1900 BCE”, that is, at the height of the Harappan civilisation. This ran against the notion of a mighty, or simply perennial, Saraswati flowing during mature Harappan times. Lawler based himself on recent independent studies piloted by geologists Sanjeev Gupta, Peter Clift (both from the UK), and Hideaki Maemoku (Japan), which suggested that the river had largely dried up long before Harappan times.
But Clift had, in a paper of September 2009 in Geoscientist, found that “between 2000 and 3000 BCE, flow along a presently dried up course known as the Ghaggur-Hakkra river ceased, probably driven by the weakening monsoon and possibly also because of headwater capture into the adjacent Yamuna and Sutlej rivers”.
Clift’s multi-national team, using sophisticated methods to date zircon sand grains and identify their provenance, published in the journal Geology of 2012 a paper which showed that the Yamuna once flowed into the Ghaggar-Hakra, but switched eastward tens of thousands of years ago; the Sutlej also contributed to the Ghaggar system but abandoned it 10,000 years ago or earlier. But the paper remained non-committal as regards the precise time for the drying of the Ghaggar itself.
More recently, in March 2012, a similar team of geoscientists published in Proceedings of National Academy of Science a paper entitled ‘Fluvial landscapes of the Harappan civilisation’ (its lead author was Liviu Giosan, with Clift as second author). The team disagreed that “large glacier-fed Himalayan river watered the Harappan heartland on the interfluve between the Indus and Ganges basins”; rather, “only monsoonal-fed rivers were active there during the Holocene” (that is, the last 10,000 years or so). In particular, “rivers were undoubtedly active in this region during the Urban Harappan Phase”. Indeed, the geoscientists found “sandy fluvial deposits approximately 5,400 (years) old at Fort Abbas in Pakistan, and recent work on the upper Ghaggar-Hakra interfluve in India also documented Holocene channel sands that are approximately 4,300 (years) old”.
In other words, the Ghaggar-Hakra was active during the mature Harappan period, although not fed by glacial sources; it was a monsoon-fed river, like rivers of central or southern India: “Reliable monsoon rains were able to sustain perennial rivers earlier during the Holocene, (which) explains why Harappan settlements flourished along the entire Ghaggar-Hakra system without access to a glacier-fed river.”
While this conclusion of a perennial but monsoon-fed Saraswati in Harappan times may be provisionally accepted, further studies surveying larger areas may slightly alter it, since we know from a 15th century Islamic chronicle that the Sutlej and Ghaggar systems were still connected in medieval times, and therefore sands of Himalayan provenance carried by the Sutlej should be identifiable in the Ghaggar’s central and lower basin.
But that is, after all, a detail: What matters is the acknowledgement of a perennial Ghaggar’s role in sustaining numerous Harappan urban settlements, and the coincidence between its dwindling down and the withdrawal of Harappan sites from its central basin. This is further supported by another 2012 study, directed by Indian geologist Rajiv Sinha and published in Quaternary International, which mapped palaeo-river sedimentary bodies in the subsurface by measuring their electrical resistivity (water-bearing sediments having a lower resistivity than dry ones). The study offered “the first stratigraphic evidence that a palaeochannel exists in the sub-surface alluvium in the Ghaggar valley. The fact that the major urban sites of Kalibangan and Kunal lie adjacent to the newly discovered subsurface fluvial channel body suggests that there may be a spatial relationship between the Ghaggar-Hakra palaeochannel and Harappan site distribution”.
Such a conclusion had been reached by archaeologists long ago, since Kalibangan, for instance, shows no evidence of independent water supply; unlike Mohenjodaro, it had very few wells, and unlike Dholavira, no reservoirs, yet it was continually occupied for several centuries: For its water supply through the year, it must therefore have depended on the Ghaggar, on whose left bank it lay (with entries into its fortified areas facing the riverbed).
A convergence of archaeological, geological and climatic studies is thus on the horizon, and we may soon be in a position to better understand the reasons for the decline of the Indus civilisation. As regards the Saraswati river, allowing for some metaphorical inflation in the Vedic hymns, nothing in the recent research contradicts the river’s break-up and gradual extinction as depicted in India’s ancient literature.
We are thus back to the original problem: If we accept the Vedic hymns’ description of a river flowing from the mountain to the sea and located between the Yamuna and the Sutlej, the Ghaggar remains the sole candidate; but as we now know, this description can only apply to the third millennium BCE or earlier, an epoch that does not fit with the conventional scenario of a second millennium Aryan migration into India. We still have to wait for the last word on India’s protohistory.
The writer is the author of The Lost River:
On the Trail of the Sarasvati (Penguin, 2010) and a long-time student of Indian protohistory; he is currently guest professor at IIT Gandhinagar and visiting professor at IIM Ranchi
PS: Science: Fluvial landscapes of the Harappan civilization
The conventional theory is that the RgVeda Samhita was supposedly compiled by the newly arrived horse-and-chariot, Indo-European language bearing invaders/immigrants who imposed themselves and their language (Sanskrit) tracelessly on the indigenous population. The RgVeda Samhita as per the Harvard Sanskritist and invasionist Michael Witzel has at most 4% of its 10,000 word vocabulary of non-Indo-European origin. The comparable figure for ancient Greek is more than 30% - and the Indo-Aryans entered (and supposedly came to dominate) a much larger geographical area with a much larger population than ancient Greece.
The evidence however is that the Saraswati dried up well before this migration/invasion took place!
Michel Danino has a newspaper article about the latest state of research.
Maps (from elsewhere):
Main archaeological sites of the Harappan civilization:
Old Saraswati:
Reproducing the article here because the Daily Pioneer articles tend to vanish.
The now dried up Saraswati river holds the key to many riddles of ancient Indian history — from the fate of the Harappans to the identity of the Vedic people. A convergence of archaeological, geological and climatic studies may soon provide us some answers
The riddle of the Saraswati river never goes long out of public view. The fascination the lost river has exerted on Indian minds is understandable: Praised in the Rig Veda’s hymns as a “mighty” river flowing “from the mountain to the sea” somewhere between the Yamuna and the Sutlej, it is reported a few centuries later by the Brahmanas (commentaries on the Vedas) as disappearing in the desert at a point called Vinashana, which was then a highly revered pilgrimage site. The Mahabharata, whose great war is waged in the region of Kurukshetra watered by the Saraswati and its tributaries, paints a similar picture, adding some details about the broken-up westward course of the river all the way to Prabhasa on the Arabian Sea. The river went on dwindling down, eventually becoming ‘mythical’, finally relocated at the confluence between Ganga and Yamuna as an ‘invisible’ river — a convenient device to remember it.
A modern myth is that satellite imagery ‘rediscovered’ the river in the 1970s. Actually, it only confirmed what had been known for over two centuries: As early as in 1760, a map from The Library Atlas published by Bryce, Collier & Schmitz showed the Saraswati (spelt ‘Soorsuty’) joining the Ghaggar (‘Guggur’) in Punjab; indeed, even today a small stream called ‘Sarsuti’ seasonally flows there. In 1778, James Rennell, a noted English geographer and cartographer, published a Map of Hindoostan or the Mogul Empire with similar details. In the early 19th century, several topographers surveyed the bed of the Ghaggar, a seasonal river that flows down from the Shivalik hills, and found it much too wide for the paltry waters it carried during monsoons; the first scholar to propose that the Ghaggar-Saraswati combine was the relic of the Vedic Saraswati was the French geographer Louis Vivien de Saint-Martin, who authored in 1855 a massive Geography of India’s North-West According to the Vedic Hymns. Subsequently, nearly all Indologists, from Max Müller to Monier-Williams or Macdonell (and later Louis Renou) accepted this thesis. Geologists such as RD Oldham (1886) joined in, followed by geographers such as the Indian Shamsul Islam Siddiqi (1944) or the German Herbert Wilhelmy (1969).
The Indus Civilisation
The story of the Saraswati’s rediscovery would thus have ended long ago if archaeology had not sprung a major surprise by redefining its role in antiquity. In the 1920s, cities of the Bronze Age like Mohenjodaro and Harappa came to light; initial findings were limited to the Indus Valley and Baluchistan, but in 1941, the intrepid explorer Sanskritist Marc Aurel Stein conducted an expedition in the then Bahawalpur State — today’s Cholistan, a very arid region of Pakistan which is technically part of the Thar desert. The Ghaggar’s dry bed continues there under the name of ‘Hakra’, and had long been known to be dotted with numerous ruined settlements. Stein’s contribution, encapsulated in his paper titled ‘A Survey of Ancient Sites along the Lost Saraswati River’, was to show that some of those sites went back to Harappan times. So the Saraswati, too, had nurtured the ‘Indus civilisation’, which prompted a few archaeologists to propose the broader term of ‘Indus-Saraswati civilisation’.
Indeed, decades of further explorations both in India and Pakistan have established that the Saraswati basin was home to about 360 sites of the Mature Harappan Phase (the urban phase that saw cities thrive, from about 2600 to 1900 BCE). This includes settlements such as Bhirrana, Rakhigarhi, Kunal or Banawali (all in Haryana), Kalibangan (Rajasthan) or Ganweriwala (Cholistan) — altogether, almost a third of all known urban Harappan sites. (Gujarat was also host to over 300 of them, another indication that the term ‘Indus civilisation’ is something of a misnomer.)
Again, that the Ghaggar-Hakra was the Saraswati’s relic was accepted by most archaeologists, including Mortimer Wheeler, Raymond Allchin (both from Britain), Gregory Possehl, JM Kenoyer (both from the US), Jean-Marie Casal (France), AH Dani (Pakistan), BB Lal, SP Gupta, VN Misra or Dilip Chakrabarti (India).
The Aryan Issue
Despite the broad consensus, scholars such as Romila Thapar, Irfan Habib and the late RS Sharma started questioning this identification in the 1980s. What prompted this rather late reaction? It was a new development: A study of the evolution of the pattern of Harappan settlements in the Saraswati basin now revealed that in its central part — roughly southwest Haryana, southern Punjab and northern Rajasthan — most or all Harappan sites were abandoned sometime around 1900 BCE, a period coinciding with the end of the urban phase of the Indus civilisation. Clearly, the river system collapsed — which archaeologists now saw as a factor contributing to the end of the brilliant Indus civilisation.
Why was this a problem? We must remember that the Saraswati is lavishly praised both as a river and a Goddess in the Rig Veda, a collection of hymns which mainstream Indology says was composed by Indo-Aryans shortly after their migration to India around 1500 BCE. However, by that time, the Saraswati had been reduced to a minor seasonal stream: How could the said Aryans praise it as a ‘mighty river’, the ‘best of rivers’, ‘mother of waters’, etc? There is a chronological impossibility.
Hence, the objectors asserted, the Ghaggar-Hakra was not, after all, the Saraswati extolled in the Rig Veda. While some (Rajesh Kochhar) tried to relocate the river in Afghanistan, others (Irfan Habib) decided that the Saraswati was not a particular river but “the river in the abstract, the River Goddess”; but both theses ran against the Rig Veda’s own testimony that the river flowed between the Yamuna and the Sutlej.
However, what should have remained a scholarly issue now turned into an ideological and often acrimonious battle: On the one hand, those who stuck to the identity between the Saraswati and the Ghaggar-Hakra concluded that the composers of the Rig Veda must have lived in the region during the third millennium BCE at the latest — but as the only settlements known of that period were Harappan ones, they often held that the Harappans were part of the Vedic people; cultural evidence such as a Harappan swastika, yogic postures, figurines in namaste and more was pressed into service to bridge the Harappan and the Vedic worlds. On the other hand, scholars who continued to swear by an Aryan immigration in the mid-second millennium BCE, and therefore a pre-Vedic Harappan civilisation, accused the former of ‘chauvinism’, ‘jingoism’ or worse, conveniently forgetting that dozens of Western scholars had, for a century-and-a-half, accepted the same location for the Saraswati river.
New Research
Leaving aside the controversy, we now have scientific research combining geology and river studies. Satellite imagery is another useful tool, but cannot by itself date the numerous buried palaeo-channels (ancient waterways) it has brought to light; anyone can today access websites such as Google Earth and view the well-marked bed of the Ghaggar, but when did a perennial river last flow through it, and where did it draw its waters from?
Several recent studies have thrown new light on the ancient river, though sometimes with contradictory findings. Thus, in an article of April 2011 published in the noted magazine Science,
A Lawler claimed that “the Ghaggar-Hakra was at most a modest seasonal stream... from 2500 BCE to 1900 BCE”, that is, at the height of the Harappan civilisation. This ran against the notion of a mighty, or simply perennial, Saraswati flowing during mature Harappan times. Lawler based himself on recent independent studies piloted by geologists Sanjeev Gupta, Peter Clift (both from the UK), and Hideaki Maemoku (Japan), which suggested that the river had largely dried up long before Harappan times.
But Clift had, in a paper of September 2009 in Geoscientist, found that “between 2000 and 3000 BCE, flow along a presently dried up course known as the Ghaggur-Hakkra river ceased, probably driven by the weakening monsoon and possibly also because of headwater capture into the adjacent Yamuna and Sutlej rivers”.
Clift’s multi-national team, using sophisticated methods to date zircon sand grains and identify their provenance, published in the journal Geology of 2012 a paper which showed that the Yamuna once flowed into the Ghaggar-Hakra, but switched eastward tens of thousands of years ago; the Sutlej also contributed to the Ghaggar system but abandoned it 10,000 years ago or earlier. But the paper remained non-committal as regards the precise time for the drying of the Ghaggar itself.
More recently, in March 2012, a similar team of geoscientists published in Proceedings of National Academy of Science a paper entitled ‘Fluvial landscapes of the Harappan civilisation’ (its lead author was Liviu Giosan, with Clift as second author). The team disagreed that “large glacier-fed Himalayan river watered the Harappan heartland on the interfluve between the Indus and Ganges basins”; rather, “only monsoonal-fed rivers were active there during the Holocene” (that is, the last 10,000 years or so). In particular, “rivers were undoubtedly active in this region during the Urban Harappan Phase”. Indeed, the geoscientists found “sandy fluvial deposits approximately 5,400 (years) old at Fort Abbas in Pakistan, and recent work on the upper Ghaggar-Hakra interfluve in India also documented Holocene channel sands that are approximately 4,300 (years) old”.
In other words, the Ghaggar-Hakra was active during the mature Harappan period, although not fed by glacial sources; it was a monsoon-fed river, like rivers of central or southern India: “Reliable monsoon rains were able to sustain perennial rivers earlier during the Holocene, (which) explains why Harappan settlements flourished along the entire Ghaggar-Hakra system without access to a glacier-fed river.”
While this conclusion of a perennial but monsoon-fed Saraswati in Harappan times may be provisionally accepted, further studies surveying larger areas may slightly alter it, since we know from a 15th century Islamic chronicle that the Sutlej and Ghaggar systems were still connected in medieval times, and therefore sands of Himalayan provenance carried by the Sutlej should be identifiable in the Ghaggar’s central and lower basin.
But that is, after all, a detail: What matters is the acknowledgement of a perennial Ghaggar’s role in sustaining numerous Harappan urban settlements, and the coincidence between its dwindling down and the withdrawal of Harappan sites from its central basin. This is further supported by another 2012 study, directed by Indian geologist Rajiv Sinha and published in Quaternary International, which mapped palaeo-river sedimentary bodies in the subsurface by measuring their electrical resistivity (water-bearing sediments having a lower resistivity than dry ones). The study offered “the first stratigraphic evidence that a palaeochannel exists in the sub-surface alluvium in the Ghaggar valley. The fact that the major urban sites of Kalibangan and Kunal lie adjacent to the newly discovered subsurface fluvial channel body suggests that there may be a spatial relationship between the Ghaggar-Hakra palaeochannel and Harappan site distribution”.
Such a conclusion had been reached by archaeologists long ago, since Kalibangan, for instance, shows no evidence of independent water supply; unlike Mohenjodaro, it had very few wells, and unlike Dholavira, no reservoirs, yet it was continually occupied for several centuries: For its water supply through the year, it must therefore have depended on the Ghaggar, on whose left bank it lay (with entries into its fortified areas facing the riverbed).
A convergence of archaeological, geological and climatic studies is thus on the horizon, and we may soon be in a position to better understand the reasons for the decline of the Indus civilisation. As regards the Saraswati river, allowing for some metaphorical inflation in the Vedic hymns, nothing in the recent research contradicts the river’s break-up and gradual extinction as depicted in India’s ancient literature.
We are thus back to the original problem: If we accept the Vedic hymns’ description of a river flowing from the mountain to the sea and located between the Yamuna and the Sutlej, the Ghaggar remains the sole candidate; but as we now know, this description can only apply to the third millennium BCE or earlier, an epoch that does not fit with the conventional scenario of a second millennium Aryan migration into India. We still have to wait for the last word on India’s protohistory.
The writer is the author of The Lost River:
On the Trail of the Sarasvati (Penguin, 2010) and a long-time student of Indian protohistory; he is currently guest professor at IIT Gandhinagar and visiting professor at IIM Ranchi
PS: Science: Fluvial landscapes of the Harappan civilization
