Habib then argues a case of mistaken identity: a better candidate than the Ghaggar would be the Sirsa river,† which runs from Kalka in a northwesterly direction to meet the Sutlej above Rupar (marked in Fig. 3.4). But while the Sirsa does derive its name from ‘Sarasvatī’ (one more clue to local memories of the river), its modest stature as a minor tributary of the Sutlej—which, because of its well-defined valley, it would have held in ancient times too—cannot match the textual descriptions. Moreover, the Sirsa was never ‘lost’ and would render meaningless the whole subsequent tradition of Vinashana, the place where the Sarasvatī disappears in the sands of the desert. Finally, the Sirsa does not flow ‘from the mountain to the sea’.
Habib’s next point is a curious one: ‘The Ghaghar and Sarasvati were separate rivers and did not join each other at all before Firoz Tughluq forced such a junction’ in the fourteenth century by digging a canal ‘from the Ghaggar . . . to the fort of Sarsati’. Sarsati (one more relic of the ancient river’s name) is today’s city of Sirsa (see Fig. 3.1), in Haryana’s westernmost district. If the Sarsuti did not meet the Ghaggar, where then did it flow? Habib proposes that ‘the Haryana Sarasvatī [was] a small isolated river, probably drying up near Sirsa . . . It had no natural connection with any other river, and so could not have been at the heart of any system.’ The point is untenable: the Sarsuti (rather the combined Sarsuti-Markanda, since, as I pointed out earlier, the Sarsuti is technically a tributary of the Markanda) did flow into the Ghaggar, and the point of confluence was not around Sirsa, but some 120 km upstream! This was, as pointed out as early as 1893 by C.F. Oldham,69 near Shatrana, and all subsequent surveys have endorsed the location: it is well marked on several maps reproduced in this book (Figs 2.3, 3.1, 3.4 and 3.8). What flowed through Sirsa, therefore, was the combined Ghaggar-Markanda-Sarsuti: the Sarsuti was never an ‘isolated’ or ‘separate’ river. Tughluq’s canal might have been intended to stabilize the flow of the Ghaggar or divert some of its waters; either way, it is irrelevant to the issue of the lost river.
The argument that follows is more challenging; it can be summarized thus : in the Nadīstuti hymn we are familiar with the fact that ‘the Yamuna, Sarasvati and Sutudri [Sutlej] are recognised as three distinct rivers’ (see p. 38). Experts who have argued that the Yamunā flowed into the Sarasvatī (Figs 3.4 and 3.8) have proposed several channels for the purpose, some close to the Shivaliks, and others further south that are part of the Chautang system identified with the Vedic Drishadvatī. In the first case, ‘the Yamuna should have surely retained its own name; and the Vedic Sarasvati, then, could not have existed at all’. In the second, not only would the Drishadvatī have had no existence of its own, but also the Ghaggar north of it would still be carrying very little water and therefore could hardly be called a ‘mighty river’. Similarly, if the Sutlej flowed from Rupar into the Ghaggar, ‘it would then be the Sutudri that the Ghaghar would be called, and not Sarasvati, an insignificant rain-fed seasonal stream on its own’. Although the paradox is well turned, it rests on at least three unstated assumptions : that all those rivers had well-defined single beds; that they alone were present in the area and no other; and that the Vedic Sarasvatī was today’s Sarsuti. In the synthesis that I will propose in the next chapter, none of these assumptions is accepted.
Habib then lists four canals dug to divert waters from the Yamuna into the Chautang between the fourteenth and seventeenth centuries, and regards the Yamuna’s failure to change its course westward by enlarging those canals as proof that the river always flowed eastward. However, this ignores the possibility of a tectonic uplift of the Sarasvatī basin or subsidence of the Yamunā’s, which, as we saw earlier, was proposed by several experts as a likely explanation for the Yamunā’s desertion of the Sarasvatī basin; in such an event, it follows that the river could not have ‘climbed back’ into the now raised watershed.
In the end, if, as Habib concludes, ‘All claims built upon the greatness of River Sarasvati are, accordingly, nothing but castles in the air, however much froth may be blown over them,’ how do we explain the primeval froth whipped up in the Rig Veda? Should we assume that the rishis, having perhaps consumed an overdose of Soma, hallucinated on the banks of a skimpy Sarsuti? Habib offers a more generous explanation: ‘Sarasvati in most of the references to it in the Rigveda is not a particular river, as in a few undoubted cases it is, but the river in the abstract, the River Goddess. When the poet priest here sings of Sarasvati, he sings not of a particular river he sees, not a particular river named Sarasvati, nor any of the Sapta Sindhavah (Avestan, Hapta Hindu), but a mighty sister of these rivers, or, alternatively, one containing all of them . . .’
In other words, the Sarasvatī is a generic river, not a specific one—a pure mythological creation, somewhat like an inverse image of the Styx in ancient Greece. Certainly, the Sarasvatī is mythologized and deified in the Rig and Yajur Vedas, and more so in later literature, but this cannot be the whole explanation. If Sarasvatī were ‘the river in the abstract’, why place it specifically after the Gangā and the Yamunā and before the Sutlej, or in company with the Drishadvatī and the Āpayā?70 Surely, all those other rivers cannot be ‘abstract’ ones, too? Also, some of the Sarasvatī’s descriptions in the Rig Veda (such as ‘breaking through the ridges of the mountains’, ‘flowing from the mountain to the sea’, the river’s ‘two grassy banks’) evoke physical rather than abstract traits. And what can one say of specific instructions found in the Shatapatha Brāhmana,71 for instance, where one is asked to collect Sarasvatī water first among several others and sprinkle the combined waters on a king for his consecration? Are these abstract waters, too?
There is no need to dwell on Habib’s criticism of the rare scholars who assert, probably in a flush of overenthusiasm at the river’s rediscovery, that the Sarasvatī was ‘mightier than the Indus’, for I agree with him that such a statement receives justification neither in the Rig Veda nor in archaeology. But it is also not legitimate to suggest that the proponents of the Ghaggar-Sarasvatī identity are guilty of ‘false patriotism’, as he does at the conclusion of his paper. Were we to accept the charge, we should equally lay it at the door of C.F. Oldham, Pargiter, Aurel Stein, Mortimer Wheeler,72 Jean-Marie Casal,73 Asko Parpola,74 the Allchins, Gregory Possehl, Jane McIntosh and the late Pakistani archaeologist Ahmad Hasan Dani75—all of whom supported the same identity.
Indeed, in his conclusion, Habib himself can be seen injecting ideology into the issue: to him, claims about the ‘once “mighty” Sarasvati’ amount to the ‘taking away of the Indus (“Sarasvati”) culture from the Dravidians and non-Aryans’—a throwback to nineteenth-century notions of such racial entities, notions that have thankfully been abandoned by today’s anthropologists and geneticists.
To locate the Sarasvatī in the Ghaggar system is not a matter of patriotism true or false, but rather, to borrow from Habib’s subtitle, a matter of commonsense: centuries after the Rig Veda’s hymns, why should a few Brāhmanas state that the same river now ‘vanishes’ at a particular point? Could the Mahābhārata indulge in long descriptions of the Sarasvatī’s course, including ashrams and tīrthas on its banks, if there was no river at all? ‘Commonsense’, it seems to me, dictates that rather than ascribe an overactive imagination to our early poets, we should acknowledge that the numerous references to the Sarasvatī across the literary corpus build a remarkably consistent picture—and one that happens to match the regression of the Ghaggar-Hakra system.
DID THE SARASVATĪ FLOW TO THE SEA?
I have saved two points presented by Irfan Habib for separate treatment. In the first, he argues that the Ghaggar-Hakra could not have flowed to the sea and must have ended in Cholistan. Curiously, after denying both the Yamunā and the Sutlej to the Ghaggar-Hakra, Habib now explains that the Hakra must have had enough water to sustain ‘the Harappan (as well as pre-Harappan and post-Harappan) sites [that] seem to run in a large belt along the Hakra river, and are especially numerous around Derawa
r’—so where did the Hakra get its waters from? He proposes that ‘given the earlier natural conditions, the Desert River could come down to the Bahawalpur Cholistan, fed by its own rain-fed Siwalik and Terai tributaries.’ Habib does not seem to realize that this nullifies his whole line of argument so far: if, as he now admits, there was in the end a river, rain-fed or otherwise, flowing from the Shivalik to Cholistan, and sustaining some 170 settlements there, it cannot have been a puny stream, for the distance involved is no less than 1000 km! In that case, what exactly is the objection to naming this ‘Desert River’ the Sarasvatī?
Let us, however, proceed with the point at issue. Habib offers the absence of Harappan sites beyond Cholistan as proof that the ‘Desert River’, however it may have been named at the time, vanished in the sands of Cholistan. In fact, in recent years, at least five Harappan sites appear to have been spotted in the Hakra Valley beyond Cholistan,77 but even if their existence is confirmed, we will admittedly be very far from the concentration found in Cholistan. Following Aurel Stein, who had already described the Derawar region as ‘deltaic’,78 Possehl also proposes that the Sarasvatī (as he calls the river) ended in an ‘inland delta’.79 But other archaeologists take the contrary view of nineteenth-century surveyors such as Oldham (p. 32), Raverty (p. 26) and Sivewright (p. 27) that the river did flow on to the sea
through the valley of the Nara: Louis Flam, for instance, sees a continuity between the Hakra and the Eastern Nara, whose valley ends at the northern fringe of the Rann of Kachchh:
There is little doubt and little disagreement that the Hakra-Nara Nadi was a seasonal river with perennial characteristics during the fourth and third millennia B.C. . . . Southwest of Fort Derawar . . . the Hakra Course becomes increasingly unclear and intermittently becomes ‘lost’ beneath the sand dunes which have encroached upon the area. Remnants of the Hakra’s course re-emerge where dunes are less numerous. From Fort Derawar to the south, the Hakra can be aligned with the Raini and Wahinda remnants, which subsequently connect with and blend into the Nara channel.80
Flam painstakingly builds a synthetic picture that includes a lower course of the Indus, east of where it flows today: the river joined the Nara at the latter’s estuary in the Rann of Kachchh (Fig. 11.1). As regards the Nara’s upstream course,
The Nara Nadi‡ originated in the Hakra River of Cholistan . . . During at least part of the third millennium B.C. the Sutlej River was a tributary of the Hakra River in Cholistan . . . not a tributary of the Indus as it is today . . . There is clear archaeological and geomorphological evidence that sometime near the end of the third millennium B.C. the Hakra River was captured by the Gangetic system and the Sutlej River discharge was diverted to the Indus system.81
M. Rafique Mughal agrees with Flam. I quoted earlier his conviction that ‘in ancient times the Ghaggar-Hakra was a mighty river, flowing independently [of the Indus] along the fringes of the Rann of Kutch’. Another clue is the fact that the lower course, the Nara, is also known as ‘Hakro’, according to Mughal.82 Wilhelmy, too, is quite explicit on this point:
In the local usage, the lower course [i.e. the Nara] is also given the name Hakra. The local people were probably aware of the continuity of the entire line of valleys and that a single river flowed here once upon a time.83
And more recently, Bridget Allchin drew a map of the lower Indus in which the Nara’s lower course down to the sea is labelled ‘Hakra’.84 Clearly, then, the Hakra’s continuity with the Nara cannot be rejected out of hand; scientific surveys and archaeological explorations of the short stretch between the Hakra and the Nara can alone provide a definitive answer.
NO SARASVATĪ IN HARAPPAN TIMES?
Irfan Habib quotes a geological study of the region by a joint Indo-French team. I cited earlier (p. 62) Marie-Agnès Courty’s study of sediments in the Ghaggar-Chautang region: she pointed out the difference between ‘true grey sands’ of clearly Himalayan origin and some 7 or 8 m of alluvium covering them—the latter consisted mainly of alternating layers of loamy sand and silty clay, and could only have originated from weaker streams flowing from the Shivaliks, and not from the inner Himalayas. Considering that it would have taken several millennia for this 8-m layer to accumulate, Courty naturally concluded that ‘Yamuna-like rivers, rising from the Himalayas, stopped flowing in the study area well before the Protohistoric [i.e., Harappan] period’.
In other words, no Sarasvatī of Himalayan origin could have flowed in Harappan times, if such a river ever existed at all. ‘This should be definitive,’ comments Habib, and we can understand his sense of triumph. Indeed Henri-Paul Francfort, director of the French side of the team, cautioned against ‘the illusory existence of protohistoric settlements concentrated along the banks of an immense perennial river, the ancient Sarasvati’, and added, ‘We know now, thanks to the fieldwork of the Indo-French expedition, that when the protohistoric peoples [i.e., the Harappans] settled in this area no large perennial river had flowed there for a long time.’85
What an anticlimax to our long journey! The mythical river was just that in the end: a myth. With a view to mislead us, sacred texts made up a non-existent mighty river and narrated its disappearance; local traditions of a ‘lost river’ signified a wishful concoction; early explorers, geologists, geographers such as Wilhelmy, and countless archaeologists including Aurel Stein and A. Ghosh fell victims to a mirage created by the treacherous desert. Before we turn the page on what, it now appears, was no more than a lovely legend, let us take a closer look at the implications of the radical French thesis.
Francfort is well aware of the numerous settlements in the Ghaggar-Chautang region. Where did they find water for agriculture if there was no ‘large perennial river’ and if, moreover, the region’s climate, according to Courty’s analysis, was already well into an arid phase? Francfort proposes that those settlements would have used ‘vast irrigation systems’,86 and his colleague Pierre Gentelle envisages ‘the existence of long canals for irrigation’,87 bringing water from the Yamunā to the Chautang region. What that involved on the part of the Harappans was, in Francfort’s words, ‘the construction, maintenance and management of canals more than 200 km long, and therefore quite a remarkable stage of social development for the civilization’.88 In reality, if water was to be diverted from the Yamunā all the way to Kalibangan, the canal would have to be not 200 km in length but twice as long; and if we include ramifications needed to reach all major sites, the total length of the irrigation network would be well over 1000 km. Yet, despite suggestions by Francfort’s team, such protohistoric canals have not come to light so far.
Instead, what their digital processing of satellite photographs did bring out is a network of ‘hydrographic fossil systems’ some 300-500 m wide. Francfort calls them ‘canals’, but explains in the next sentence that they ‘mark the courses of ancient natural waterways which were used and perhaps, in some places, rerouted by man. These traces of small river channels, which were completely unknown until now, appear to have reached all of the archaeological sites’.89 ‘River channels’ and ‘natural waterways’ are not exactly ‘canals’, and even admitting that the Harappans used a few of these waterways for agriculture, it is hardly credible that they could have diverted enough water from the Yamunā to fill this vast network—and prevented its regular siltation.
Also, if there were no rivers flowing west of the Yamunā, why did they take so much trouble routing and maintaining those canals when they could have quite simply gone and settled closer to the Yamunā and its fertile banks? And there’s the rub: as the maps show eloquently, there were no Early Harappan sites close to the Yamunā (Fig. 6.7) and very few sites of the Mature phase (Fig. 6.8).
There must be a less artificial explanation for these networks of ‘natural waterways’ in the Ghaggar-Chautang region. I suggest that it partly lies in the high rainfall that the Shivaliks were experiencing at the time. We saw in Chapter 8 some studies contradicting the thesis that the Harappan milieu had entered
a phase of aridity. In particular, Netajirao Phadtare (p. 177) found evidence of ‘a warm, humid climate, with highest monsoon intensity’ in the Garhwal Himalayas from about 4000 to 2500 BCE (which takes us to the Mature Harappan phase), with a ‘sharp decrease’ after 2000 BCE. Along the same lines, Rita Wright and her colleagues (p. 179) found waters in the Beas to increase around 3500 BCE until around 2100 BCE, when ‘the river flow begins to fall’; the last date also marked the start of ‘a 600-year period of reduced rainfall’ around Harappa.
This may not be the last word on the question, but let us note that the regions researched by Phadtare and Wright are on either side of the upper catchment area of the Sarasvatī basin. If their results are confirmed, precipitation must have been high there also (though not necessarily in the plains below), from which two consequences follow. The first is that we can do away with ‘canals’ bringing water all the way from the Yamunā: Francfort’s ‘natural waterways’ would simply have carried waters streaming down from the Shivaliks. Conceding that a small part might have been diverted to irrigation, their excess flow would ultimately collect in the main trunk of the Ghaggar. That it did hold water in Mature Harappan times is proved by the location of important settlements close to, or on, the main channels, as a look at Fig. 6.8 makes clear. Why should the Harappans have built Kalibangan on the very edge of the Ghaggar if the river was dry? Banawali and Rakhigarhi are also located along important channels of the Ghaggar and the Chautang respectively: here as in Punjab and Sind, river communication was vital for the larger settlements.
The second consequence of a more intense monsoon in the catchments is that the sedimentation rate for the accumulation of the 8-m layer of alluvium above the grey sand of Himalayan origin would have been much faster than was estimated by Courty, who assumed an arid regime for the region. Even if the plains were arid or gradually becoming so, a high annual precipitation in the Shivaliks and above would alter the picture, as it would create many perennial streams in the foothills—which would inevitably collect in the Ghaggar.
The Lost River: On The Trail of Saraswati Page 23
