The Bharata War
Let us review the three main Indian traditions regarding the time of the Bharata War.
1. The Puranic Evidence
To examine this tradition we depend on the collation of data by Pargiter. According to the Puranas, a total of 1,500 years (in certain texts 1,015, 1,050, or even 1,115 years) (Vayu 99.415; Matsya 73.36 etc) elapsed between the birth of King Pariksit and the accession of Mahapadma Nanda. The king lists for this period add up to 1,498 or 1,500 years in the most reliable records It appears that the correct elapsed duration is 1,500 years as it tallies with the detailed count.
Based on his collation, Pargiter suggested an important emendation as follows:19The Great Bear (the riksas or the Seven Sages or Saptarsi) was situated equally with regard to the lunar constellation Pusya while Pratipa was king. At the end of the Andhras, who will be in the 27th century afterwards, the cycle repeats itself. In the circle of the lunar constellations, wherein the Great Bear revolves, and which contains 27 constellations in its circumference, the Great Bear remains 100 years in (i.e. conjoined with) each in turn
This implies a period of 2,700 years from a few generations before the War to the middle of the third century AD. Support for this reading comes from the following statement that has often been misinterpreted: The Saptarsi were in Magha at the time of Yudhisthira but had shifted to Purvasadha (ten naksatra on) at the time of Nanda and Satabhisaj (a further four naksatras) at the end of the reign of the Andhras (Vayu P. 99.423). This astronomical evidence would point to a gap of about 1,000 years between Pariksit and Nanda and another 400 years between Nanda and the end of the Andhras Considering that Pratipa was only seven generations before Pariksit, or about 150 years earlier, this gives a total interval of about one-half the interval of 2,700 years mentioned above. But we do know that the gap between Nanda and the end of the Andhras was more than 800 years It is clear that this second reference counts two hundred years for each naksatra. This may have had something to do with the Jain tradition that counted a total of 54 naksatras and to the number stated one had to add a like number for a correct count.
As for the duration of reigns, Vayu Purana 99.416 speaks of a gap of 829 years between Nanda and the end of Andhras. Elsewhere this gap is given to be 836 years Adding the dynastic lists with 100 years to the Nandas, 137 years to the Mauryas, 112 years to the Sungas, 45 years to the Kanvas, and 460 years to the Andhras one gets a total of 854 years
The Puranas also assign one hundred years to Mahapadma Nanda and his eight sons Furthermore, in Magadha 22 Barhadrathas, 5 Pradyotas and 10 Si ́sunagas are assigned for the period between the Bharata War and the inauguration of Mahapadma Nanda for a total of (967+138 +346) 1,451 years The historian of astronomy P.C. Sengupta argues that to the Pradyotas one should add another 52 years, giving a total of 1,503 years Over the same period are said to have ruled 30 Paurava kings and 29 Aiksvakus It is also stated that when Mahapadma Nanda defeated the ksatriyas, there had reigned since the Bharata War 24 Aiksvakus, 27 Pancalas, 24 Ka ́sis, 28 Haihayas, 32 Kalingas, and so on.
Assuming that the lists are complete and that the year assignments are wrong, various suggestions have been made for the duration of the average reign On the other hand, using the statement that ten centennials (ten naksatras) had passed between the time of Pariksit and Nanda, one gets approximately 1,100 years upto Candragupta, which yields circa 1420 BC for the Wari20
Considering that Candragupta became king about 324 BC the direct reference to the years elapsed (counting 1500 years of the Puranic statement and 100 years of the Nandas) leads to the date of is 1924 BC. But clearly the average reigns for the kings are too long, unless these lists are incomplete and the names are the most prominent ones, in which case there would have been other kings who ruled for very short intervals
If the naksatra reckoning was for some reason actually being done per each two centuries as the gap of 829 years for four naksatras indicates, then there should be about 2,000 years between Pariksit and Nanda. This would take the Bharata battle to around the middle of the third millennium BC. We will show later that this takes us to 2449 BC.
2. The Kaliyuga Tradition
According to the famous astronomer Aryabhat.a (c. 500 AD) the Kaliyuga began in 3102 BC, which the Mahabharata says happened thirty-five years after the conclusion of the battle. This implies the date of 3137 BC for the War if we assume with the tradition that the Kaliyuga era began 35 years after the War. But there are other accounts, such as that of Kalhana in his Rajatarangini 1.51, where it is stated that 653 years of the Kaliyuga had passed when the Kurus and the Pand.avas lived on the earth.
3. Varahamihira’s Statement
Varahamihira (550 AD) claims that according to the earlier tradition of the astronomer Vriddha Garga, the Pand.ava king Yudhisthira was ruling 2,526 years before the commencement of the Saka era (Brihatsam. hita 13.3). This amounts to 2449 BC for the War and 2414 BC for the beginning of the Kali era.
There is no reference to the Kaliyuga era in texts before Aryabhat.a, and so it has been claimed that this era was devised by Aryabhat.a or his contemporaries. The first inscriptional reference to this era is in the Aihole inscription of 633/634 C.E.
After analyzing the astronomical evidence, P.C. Sengupta spoke in favour of the date of 2449 BC. We will examine these conflicting accounts and see if they can be compared considering independent evidence. Here we will use the king lists of the epics and the Puranas, the Greek evidence, and contemporary archaeological insights.
Analysis of the Literary Evidence
The Puranic Evidence
We have seen that the Puranic data has been interpreted variously to yield dates for the Bharata War that range from the latest of 1424 BC to the earliest of late-fourth millennium BC.21 Each of these will be separately examined.
This date is suggested by the mention in some Puranic manuscripts of the interval of 1,050 years between Pariksit and Nanda. This date is too late by about 500 years when compared to the totals of the reigns in the Puranas On the other hand, it does bring the average reign period to the realm of possibility, as it reduces to about 27 years, assuming of course that the lists are complete. The fact that a submerged temple at Dvaraka dating to the middle of the second millennium BC has been discovered has been taken as the evidence of the destruction of that city soon after the Bharata War. However, we do not know if this temple is the one that was lost to the sea soon after the Bharata War.
There is no archaeological evidence suggesting a flowering around 1500 BC. For this epoch for the War, one would expect evidence for the tremendous literary activity of the arrangement of the Vedas and the composition of the other texts. The second millennium BC is archaeologically the lesser age or the dark age.
We must reject this date if we consider the evidence related to the Sarasvati river, which was supposed to be a major river during the time of the Bharata WaRiSince this river dried up around 1900 BC, the figure of 1424 BC for the War is too late. The rapid decline around 1900 BC of cities, such as Kalibangan in the mid-course of the Sarasvati, makes it impossible for us to assume that the river could have somehow been called “major” when it ceased to flow all the way to the ocean 1924 BC
This date is a result of the stated interval of 1,500 years between Pariksit and Nanda, and the count obtained by adding up the durations of the reigns. This appears to be the original interval of the Puranas that became corrupted. Pargiter has suggested that the Puranas, as living bardic material, were transcribed into Sanskrit sometime between the reigns of the Sungas and the Guptas from the then form in Prakrit. This translation often used ambiguous constructions which is how the figure of 1,500 was read wrongly at some places According to Lalit Mohan Kar,22 “If a comparative estimate is desired between the totals, as given by the different Puranas (vis, 1015, 1050 and 1115 years), and the sum total found by calculation of the details [1500 years], the scale must turn in favour of the latter, as a corruption, or at least a variation, depends on the mutation of two or three letters of the alphabet, as is evident from there being those different versions of the total period, while the details are more definite.”
If the Bharata War story was a metaphor for the natural catastrophe that occurred in India around 1900 BC, then this is the correct date. On the other hand, if the War did take place (although it was remembered in an embellished form), then the natural catastrophe may have contributed to it by causing a breakdown of the old order.
This is the date mentioned by Varahamihira. The Puranas may be interpreted to point to this date, and also this date may be correct if the genealogies represent only the chief kings.
It is indirectly supported by the archaeological evidence. Since a great deal of literary output of Vedic times was produced and arranged during the centuries after the War, one would expect that such efforts would have been supported by kings and that one would find a correlation with prosperity in the land. The archaeological evidence indicates that the Harappan era represents a period of great prosperity.
This date implies that the Harappan phase of the Sindhu-Sarasvati tradition is essentially post-Vedic. But this date also implies that the genealogical lists are hopelessly incomplete which is plausible if a great catastrophe, such as the drying up of the Sarasvati, caused the tradition to be interrupted.
The problem with this date is that the Puranic evidence does not support it. On the other hand, some scholars have suggested that the Sarasvati river went through two phases of diminution: first, around 3000 BC, after which the river ceased to flow all the way to the sea; second, 1900 BC, when due to further shrinkage the river was unable to support the water needs of the communities around it, ending the most prosperous phase of the Harappan era. Since the Rigveda describes the Sarasvati as sea-going so, going by this theory, the Rigveda must be prior to 3000 BC.
This date could be reconciled with the Puranic accounts only if we take it to define the last phase of the RIgveda and assume that the Bharata War was wrongly transferred to this earlier era when the last major assessment of ancient Indian eras and history was done during the early Siddhantic period of Indian astronomy in early centuries AD.
The Saptarsi Era and the Greek Notices
The Indian tradition of the seven risis, the stars of Ursa Major, is an ancient one which goes back to the RIgveda. The Satapatha Brahmana speaks of a marriage between the risis and the naksatras; specifically it mentions that the risis were married to the Krittikas In the Puranas, this notion of marriage is elaborated when it is clearly stated that the risis remain for a hundred years in each naksatra. This Puranic account implies a centennial reckoning system with a cycle of 2,700 years Such a system has been in use in parts of India since centuries before Christ, and it is called the Saptarsi era. Each cycle of 2,700 years was called a cakra, or cycle. By current reckoning in Kashmir, in use at least from the time of Kalhana (1150 AD), Saptarsi era began in 3076 BC, and there is evidence that, originally, it started in 6676 BC.23It appears that it is the beginning of this era that is quoted by the Greek historians Pliny and Arrian:
From Father Liber to Alexander the Great, they reckon the number of their kings to have been 154, and they reckon (the time as) 6,451 years and 3 months [Pliny, Naturalis Historia, 6.59-60] From Dionysos to Sandrocottos (Candragupta) the Indians count 153 kings, and more than 6,042 years; and during this time, thrice for liberty … this for 300 years, the other for 120 years [Arrian, Indica, 9.9]
These two traditions, perhaps derived from the same source, can be reconciled if the Arrian years are all added up, which gives (6,042+ 300+ 120) or 6,462 years, which is only 11 years different from the other account. These eleven years might represent the gap between the time of Alexander and the Greek embassy to Candragupta Maurya. If one takes the year 314 BC for the embassy to Candragupta, one gets 6776 BC as the beginning of the Indian calendar in use at that time. This is just one centennial removed from the epoch of 6676 BC suggested by its current beginning of 3076 BC, together with an additional 3,600 years
As to the count of 153 or 154 kings, it accords quite closely if one follows up the list until the Bharata War, with the kings of the Magadhan line together with the ten kings of the Barhadrathas, whose names the Puranas tell us are lost. This total up to Candragupta is 143, which is only ten or eleven less than the Greek total. This close accord tells us that the king lists of the fourth century BC are about the same as those now, excepting that the current lists have dropped a few names This loss of about ten kings from the lists in a span of five or six hundred years, when the current versions of some of the Puranas became fixed, suggests that a similar loss might have occurred before, and it supports the view that the genealogies are incomplete.
It has been argued that the Kaliyuga and the Varahamihira traditions about the Bharata War can be reconciled if it is assumed that a change in reckoning from a system of 28 naksatras to that of 27 naksatras took place sometime after the time of Candragupta. It is also suggested that the Kaliyuga tradition might be authentic and the Varahamihira tradition was derived from it.
But the evidence from the Rigveda supports the notion that the original system of naksatras was 27 and that it was modified to 28 lateRiThe notion of 27 naksatras can also be found in the Taittiriya Samhita.
It is significant that the epoch of 6676 BC is exactly 3,600 years earlier than the starting point of 3076 BC for the Saptarsi era, as accepted now. Since it is clear that at the time of the Mauryas, the cycles of the Saptarsi era were counted back to 6676 BC, it appears that the new count that goes back to 3076 BC was started later to make it as close to the start of the Kali era as possible.
There exists another plausible explanation for how the tradition of the starting point of 6776 BC arose. By the time of the Greeks, the naksatras were listed starting with A ́svini (as in Surya Siddhanta 8.9). As Magha is the tenth naksatra in a count beginning with A ́svini, one needs to add 900 years to find the epoch for the beginning of the cycle. This takes one to 3976 BC. One more complete Saptarsi cycle of 2,700 years before that brings us to 6676 BC.
Although the limitations and ambiguities of the Puranic evidence have been much debated, it should be realized that much old criticism has lost its weight in view of the new archaeological discoveries indicating continuity in Indian culture. Thus the calendrical framework described above is perfectly consistent with the other evidence, although one would take it to have been confirmed only after its details are corroborated independently.
Relative Chronology of the Texts
Our examination of the evidence leaves us with three choices for the Bharata War: 1924 BC, 2449 BC, and 3137 BC. One might wish to speak of a High Chronology and a Low Chronology to indicate the limits within which one might safely place the War based on the current evidence. If we anchor our dates to the catastrophic events of 1900 BC and see the Mahabharata story as the mapping of a geological disaster into a human one, then one must place the RIgvedic era somewhat before 2000 BC. The tradition that the Bharata War began about 1,500 years before the Nandas would agree completely with this view.
The Brahmanas and the Aranyakas would then belong to the early or mid-2nd millennium BC, the forest age between the two early urbanizations of India.
Since the earliest Vedic literature, as in the Sam. hitas, is encyclopaedic, the longer time-spans over which it developed allow us to narrow the gap between the three choices We don’t wish to depend on literary tradition alone, and therefore take the physical event of the drying up of the Sarasvati river to help determine the period of the texts
Thus, since the Rigveda mentions a Sarasvati flowing all the way down to the sea, this text should be earlier than 1900 BC. How much earlier, we cannot say. Indeed, if the theory that the Sarasvati river ceased to reach the sea about 3000 BC is true, then the Rigveda should be prior to this early epoch. But wishing to be as conservative as possible, we take the latest possible date for the drying up of the Sarasvati, and this has the virtue of being the about same as the Puranic date of 1924 BC. This has further support from the reference in the Brahmanas about the migration east from the Sarasvati area due to heat and, presumably, famine.
Analyzing the astronomical evidence alone, Sengupta in 1947 came up with the following chronology for the references in the texts: the Vedic Sam. hitas, 4000-2500 BC; Brahmanas, 2500-1000 BC; Baudhayana Srauta Sutra, 900 BC; and so on My own analysis of the astronomy gives three phases:24
Rigvedic astronomy: 4000 – 2000 BC The astronomy of the Brahmanas: 2000 – 1000 BC Early Siddhantic and early Puranic astronomy: 1000 BC – 500 AD The date of Vedanga Jyotisa of Lagadha is 1300 BC, thus placing it in the Brahmana age.
Much of the early Sutra literature can be expected to belong to the first half of the first millennium BC, which may also be the age of the Bhagavad Gita.
The Development of Ideas
Indian culture, as depicted by its texts and its art, has unique features. For example, the ancient Indian rock art, which is believed to be several tens of thousands years old, has tessellations that are unique in the ancient world.25 Some have suggested that these designs may represent “mystical” experience. The Vedic texts are mystical, and they themselves say so when they assert that words have limitations
The Sindhu-Sarasvati cultural tradition has characteristics that indicate a social and political organization, and hence a world-view, different from the other traditions of West Asia. There is very little monumental architecture and it appears that the elites were a religious aristocracy.
The Harappan art includes motifs that could very well represent the goddess imagery of the Puranas. One image is a cylinder seal from Kalibangan that shows a goddess holding back two warriors; here, using a very clever, representational style, the goddess is also shown separately merging into a tiger, suggesting that the tiger is the mount of the goddess Durga as Mahisasura-Mardini is depicted in the Puranas as riding a lion or a tiger.
A significant building at Mohenjo-Daro has been identified as a fire temple. The building has a central courtyard and a symmetric arrangement of rooms. Every alternate room has a low brick platform and one of the rooms has a staircase leading to an upper floor. It appears that a fire altar was placed in the central courtyard.
This fire temple has symmetric features that have much in common with the architectural mand.alas discovered in North Afghanistan,26 which have been dated to 2000 BC. Since textual evidence suggests that such mandalas came to be employed long after the Rigvedic age, this evidence provides a useful chronological marker. Apart from the textual evidence, one would expect that an artistic representation of the abstract yantric concept would take centuries to develop.
The notion of the yantra and the mythology of the goddess represent a mature stage in the evolution of Indian religious imagination. Their existence in the 3rd millennium calls for a drastic revision of the academic chronology for these ideas
Libation vessels made of the conch shell turbinella pyrum have been found at Mohenjo-Daro. One of these has vermillion filled incised lines We know such conch vessels were used in the Vedic ritual and for administering sacred water or medicine to patients
The Vedic altars had an astronomical basis In the basic scheme, the circle represented the earth, while the square represented the heavens or the deity. But the altar or the temple, as a representation of the dynamism of the universe, required a breaking of the symmetry of the square. As seen clearly in the agnicayana and other altar constructions, this was done in a variety of ways Although the main altar might be square or its derivative, the overall sacred area was taken to be a departure from this shape. In particular, the temples of the goddess were drawn on a rectangular plan. The dynamism is expressed by a doubling of the square to a rectangle or the ratio 1:2, where the garbhagriha is built in the geometrical centre.
The constructions of the Harappan period appear to be according to the same principles. The dynamic ratio of 1:2:4 is the most commonly encountered size of rooms of houses, in the overall plan of houses and the construction of large public buildings This ratio is also reflected in the overall plan of the large walled sector at Mohenjo-Daro called the “citadel mound”.
If the Harappan iconography expresses the ideas of the original Purana, we are quite close to the traditional chronology of Indian history.
New findings are leading to a new view of ancient India, revealing substantial convergence between the archaeological record and the literary tradition. To be as conservative as possible within the parameters of the new archaeological and astronomical evidence, we think it prudent to consider 2000 BC as the divide between the early Vedic and the later Vedic literature.
The new paradigm is of the greatest significance in understanding the development of philosophical ideas in India. As the Harappan record becomes more accessible, we will be able to provide material evidence of innovations that had their parallels, or inspiration, in philosophical thought.
- The participants included an archaeologist (Greg Possehl, Univ of Pennsylvania), three linguists (Madhav Deshpande, Univ of Michigan, Andree Sjöberg, Univ of Texas, and Michael Witzel, Harvard Univ), and a historian of science (Subhash Kak); Lonnie Kliever of SMU served as the moderatoRiThe participants looked at both the idea of invasions and that of a more peaceful process of immigration
- E.g. Robb (1993). Basically, the proposition is that the ancient world was much more complex than supposed in the 19th century models This complexity viewed within the Indian context is examined in Kak (1994b). Even the idea of the neat centum/satem split geographically has been undermined by the discovery of Bangani, a centum language in India.
- Wakankar (1992).
- Shaffer and Lichtenstein (1998).
- Kennedy (1995).
- Lal (1997).
- Shaffer and Lichtenstein (1995).
- Seidenberg (1978).
- For example, see Kak (1994a, 1995a,b, 1996a, 1998b,c).
- Kak (1988).
- Allchin (1995), pages 176-179.
- See Sharma (1995); for new evidence on the domestication of the horse several thousand years before the older postulated period of the second millennium BC, see Anthony et al. (1991).
- Kak (1994b).
- Sengupta (1947).
- Kak (1998b, 1998d).
- Napier (1986, 1998).
- Kak (1998d); see also Alvarez (1978) and Taylor (1992).
- Pargiter (1922); see Bhargava (1971), Frawley (1991), and Klostermaier (1994, 1998).
- Pargiter (1922).
- This date has been quite popular with scholars for some time but has much evidence going against it.
- Another date of 950 BC was proposed to fit in with the theory of the Aryan invasions But this date has nothing to commend it. For a critique see Kak (1994a).
- Kar (1916). Also note that in Sengupta (1947; page 55) the date is given as 1921 BC.
- See Kak (1994a) for a further discussion
- Kak (1998c). For another attempt to construct a new chronology of the texts, see Feuerstein (1998).
- See, for example, Rao and Kak (1998). For further details on the rest of this section, see Kak (1998a, 1998d) and Feuerstein et al (1995).
- Kak (1994a), pages 43-46.
Allchin, F.RI(ed.), 1995. The Archaeology of Early Historic South Asia. Cambridge.
Alvarez, O. 1978. Celestial Brides: A Study in Mythology and Archaeology. Stockbridge.
Anthony, D., Telegin, D.Y., Brown, D. 1991. The origin of horseback riding. Scientific American, December, 94-100.
Bhargava, P.L. 1971. India in the Vedic Age. Lucknow.
Feuerstein, G. 1998. The Yoga Tradition Prescott.
Feuerstein, G. Kak, S. and Frawley, D. 1995. In Search of the Cradle of Civilization Wheaton
Frawley, D. 1991. Gods, Sages, and Kings Salt Lake City.
Kak, S. 1988. “A frequency analysis of the Indus script.” Cryptologia 12: 129-143.
———- 1994a. The Astronomical Code of the RIgveda. New Delhi.
———- 1994b. “On the classification of Indic languages” Annals of the Bhandarkar Oriental Institute 75, 185-195.
———- 1995a. “The astronomy of the age of geometric altars” Quarterly Journal of the Royal Astronomical Society 36: 385-396.
———- 1995b. “From Vedic science to Vedanta.” Adyar Library Bulletin 59: 1-36.
———- 1996a. “Knowledge of planets in the third millennium BC.” Quarterly Journal of the Royal Astronomical Society 37: 709-715.
———- 1998a. “Mind, immortality and art.” Presented at the International Seminar on Mind, Man and Mask, Indira Gandhi National Centre for the Arts, New Delhi, Feb 24-28, 1998.
———- 1998b. “Early theories on the distance to the sun” Indian Journal of History of Science 33: 93-100.
———- 1998c. “The sun’s orbit in the Brahmanas” Indian Journal of History of Science 33: 175-191.
———- 1998d. “Indic ideas in the Graeco-Roman world.” to appear in Indian Historical Review.
Kar, L.M. 1916. “Introduction” to The Matsya Puranam. Allahabad.
Kennedy, K.A.RI1995. “Have Aryans been identified in the prehistoric skeletal record from South Asia? Biological anthropology and concepts of ancient races” In The Indo-Aryans of Ancient South Asia., G. Erdosy (ed.). 32-66. Berlin
Klostermaier, K. 1994. A Survey of Hinduism. Albany.
———- 1998. A Short Introduction to Hinduism. Oxford.
Lal, B.B. 1997. The Earliest Civilization of South Asia. Delhi.
Napier, A. David. 1986. Masks, Transformation, and Paradox. Berkeley.
———- 1998. “Masks and metaphysics in the ancient world: an anthropo- logical view.” Presented at the International Seminar on Mind, Man and Mask, Indira Gandhi National Centre for the Arts, New Delhi, Feb 24-28, 1998.
Pargiter, F.E. 1922. Ancient Indian Historical Tradition London
Rao, T.R.N and Kak, S. (eds) 1998. Computing Science in Ancient India. Lafayette.
Robb, J. 1993. A social prehistory of European languages Antiquity, 67, 747-760.
Seidenberg, A. 1978. The origin of mathematics Archive for History of Exact Sciences 18: 301-342.
Sengupta, P.C. 1947. Ancient Indian Chronology. Calcutta: University of Calcutta Press
Shaffer, J.G. and Lichtenstein, D.A. 1995. “The concept of ‘cultural tradi- tion’ and ‘palaeoethnicity’ in South Asian archaeology.” In The Indo- Aryans of Ancient South Asia., G. Erdosy (ed.). 126-154. Berlin
———- 1998. “Migration, philology and South Asian archaeology.” In Aryan and Non-Aryan in South Asia: Evidence, Interpretation and Ideology, J. Bronkhorst and M. Deshpande (eds). In press, Ann ArboRi
Sharma, A.K., 1993. “The Harappan horse was buried under the dunes” Puratattva 23: 30-34.
Taylor, T. 1992. “The Gundestrup cauldron” Scientific American 266(March): 84-9.
Wakankar, V.S. 1992. “Rock painting in India.” In Rock Art in the Old World, M. Lorblanchet (ed.). 319-336. New Delhi.
This article has been republished with permission from the author.
Featured Image: ResearchGate