The Mahabharata is, as Moriz Winternitz expressed in 1908, a literarisches Unding, an ungeheuerliches Chaos. This premise has allowed scholars of different persuasions and agenda to compose a Mahabharata of convenience through which to nurture – with more bombast than debate, and with scarcely anything that could be called cumulative results – their own contradictory notions of origins and their equally contradictory development theories. Much of their underpinnings derive from a concurrence of “historically effected consciousnesses,” as Hans Georg Gadamer would say. On the one side, Westerns use a default paradigm of the Iliad and see the Mahabharata as chaotic, fantastic and incomplete. On the other hand, some Indian scholars seek to fashion an ancient and glorious Indian epic period and their dates for the occurrence of the Mahabharata wander beyond credulity, with little or no archaeological evidence and inconclusive astronomical evidence. In any case, as the debate rages on, at times with ad hominem arguments, between Michael Witzel and his colleagues on one side and David Frawley, NS Rajaram, and their colleagues on the other, it might be prudent to remember that neither side has been able to solve the riddle of India yet. Witzel has been, in my opinion, proven wrong in his assertion that the Indus seals are not a script by some fantastic work by Asko Parpola, and Natwar Jha and Rajaram are yet to conclusively prove their claim that they have deciphered the script. There are many questions still as to in which direction the script was written and whether the Indus script represents an early form of Vedic Sanskrit or some proto-Dravidian tongue. Genetic evidence seems to disprove any idea of a migration into India from outside – if so, how is it that Brahui, a language supposedly built on Dravidian structure and spoken in Baluchistan, exists completely cut off from southern India?
I have already written on the genetic debate earlier (Re-Evaluating the Saraswati-Sindhu Civilisation on January 26, 2011) and so I will not repeat those points again. Instead, I wish to evaluate the astronomical evidence that so much has been made of. As far as I can tell, based on several research papers on this topic and numerous astronomical simulations, the scientific evidence is not conclusive either. It must be remembered that only roughly 10% of Mohenjo-Daro has been excavated and no undersea excavations have taken place at Dholavira yet – these sites may prove to be a goldmine of information in the coming century.
The Mahabharata is replete with astronomical observations, the most famous being the solar eclipse on the 14th day of the War, resulting in the death of Jayadratha. Another quite famous incident is the occurrence of two eclipses within a span of thirteen days. There are approximately another 140 astronomical observations in the text, and the premise of archaeoastronomy is that the combination of all these observations can happen only once in the ten thousand years or so that civilisation of any kind has existed in the Indian subcontinent (Mehrgarh has been dated to around about 7,500 BCE). This sounds like a fairly straight-forward proposition. However, translation of the text is one issue (which I will come back to) that can skew our calculations. Another problem is that the Mahabharata was part of an oral tradition and it was not written down completely for a while. In the passing on of the text from one generation to another, some errors may have crept in. Furthermore, it is widely accepted that the Mahabharata in its entirety was not the work of one poet (the style of writing and the linguistic form strongly indicates the evolution of the Mahabharata over a period of a couple of centuries) – it started off as a short 8,800-verse poem called Jaya. Later poets added to the kernel and the poem expanded to about 30,000 verses known as the Bharata and finally it came into the form we know it today, with about 100,000 verses known as the Mahabharata. In the addition of all the extraneous material, some things may have inadvertently been dropped or some obscurities, exaggerations and contradictions pertaining to religious, cultural and sociological themes could have crept into the text due to vested interests (the famous example is in the Bhishma Parva when Shani is said to be near Rohini as well as near Visakha). On the other hand, it is expected that descriptions of natural phenomena such as an evening sky or a flood or an earthquake would not have been distorted intentionally. Similar would be the case, after due allowances are made for a poetic language, with planetary positions and observations of eclipses. The Mahabharata contains descriptions of several celestial phenomena as the story unfolds. Our discussion should ascertain the internal consistency of these descriptions – obviously, a fair degree of consistency among the various statements available in the text is a prerequisite for them to be treated as real observations, worthy of being taken as historical facts. Finally, it is important to know what kind of calendar the ancients used before we can make any sense of their observations.
It is stated in the Adi Parva (2.13) that the war between the armies of the Kurus and the Pandavas took place at Samantapanchaka during the transition between the Dvapara Yuga and the Kali Yuga. This gives us the place from where some of the observations were made. Since the key evidence used in these astronomical quests is the description of eclipses, let us look at how they are defined in the Mahabharata. In the Bhishma Parva (12:40-47), it states that the diameter of Rahu is 12,000 yojanas, whereas the diameters of the moon and the sun are 11,000 and 10,000 yojanas respectively. The circumference of Rahu, the moon and the sun work out to be 36000, 33,000 and 30,000 yojanas respectively. Hence, being bigger, Rahu at appropriate times covers the moon and the sun. Indirectly, the value of the irrational number Pi (3.1415926536…) has been taken as 3.00 in the text. Nevertheless, the statements are interesting in that a physical reason is given for eclipses instead of the usual mythological stories.
There are eight places in the text where solar eclipses are mentioned, but not all of them are credible as actual observations. The first is in Sabha Parva after the Pandavas are banished to the forest. In answer to a question on how Pandavas started their journey, Vidura gives a graphic description of the various events and incidentally mentions, “In the cloudless sky there was lightning; the earth shook and Rahu caught the sun, but not on the fifteenth day.” (Sabha Parva 79.29)
Several portents are described in the text starting with the Rajasuya sacrifice and ending with the gambling episode. Hence the above eclipse might not have occurred exactly at the time of Pandavas leaving Hastinapura. Dhritarashtra confirms this eclipse much later as he broods over the sad happenings, in the last chapter of the Sabha Parva: “Meteorites are falling in daytime, and Rahu covered Sun on an odd day causing great fear among people.” (Sabha Parva 80:23)
The place of observation of this eclipse was Hastinapura, not far away from Kurukshetra. The second mention of a solar eclipse is in Udyoga Parva when Kama tells Krishna: “The brightness of the moon is covered up and Rahu is approaching Sun.” (Udyoga Parva 142:11)
This is not an observation of an eclipse but only an expectation. The alternate reading of this verse is similar and the text is very clear that the conversation between Krishna and Kama. Since prediction of eclipses is not mentioned anywhere in the text, even as a possibility, the above can at best be an observation before the month of Kartik in the same year.
The third mention of a solar eclipse is again in Udyoga Parva, describing an old battle between Bhishma and his teacher Bhargava. Bhishma himself says that on the fourth day of their 23-day battle at Kurukshetra, “Svarbhanu suddenly approached the brightened Sun.” (Udyoga Parva 183:22) The language of the text is realistic in that the eclipse was unexpected. This should have been some 50 or 60 years before the War, when Bhishma was in his youth.
The fourth solar eclipse is in Bhishma Parva when Vyasa tells Dhritarashtra: “The earth shakes often, similarly Rahu caught up the sun.” (Bhishma Parva 3:11) This statement is in past tense. Thus some time before the war all versions indicate occurrence of a solar eclipse. But a difficulty arises due to the mention of another eclipse a few verses later in the same chapter. This fifth mention of a solar eclipse in the Mahabharata text is: “Moon and Sun were eclipsed in the same month at thirteen days interval (Bhishma Parva 3:29).
The sixth mention of an eclipse appears in Salya Parva, on the last day of the war, before the duel between Bheema and Duryodhana. Among the several bad omens an eclipse finds mention as: “Rahu caught up Sun at an odd time. The earth shook along with trees and forests.” (Salya Parva 56:10)
All the editions carry this verse. However this event is not credible as an eclipse. This follows from the previous statements in Bhishma Parva that the eighteen-day war started the day after a Kartika full moon. The seventh solar eclipse, which is again not reliable, is in the Asvamedhika Parva (77.15, 77.18), during a fight between Arjuna and the Saindhavas, when Rahu catches the sun and the moon together. It would look quite interesting for the sun and moon to be together on a new moon day anyway, but in the next verse it is said that a part of the moon fell out, breaking the sky. This implies that the moon was visible. Thus, this event is not reliable. The eighth and final solar eclipse is in Mausala Parva, in the thirty-sixth year after the War: The fourteenth day has been made into the fifteenth day again by Rahu. Krishna understood that the 36th year and Gandhari’s curse had arrived. (2:19-20)
This eclipse is mentioned in all the editions of the text. Also this event finds prominent mention in the Prabhasa Khanda of Skanda Purana, while describing the last days of Krishna. In contrast with solar eclipses, mention of lunar eclipses in the Mahabharata is almost absent. In Udyoga Parva, Kama during his conversation with Krishna mentions that the light of the moon is circumscribed. (141:10) In the Bhishma Parva, during his discourse, Vyasa says that the Kartika full moon became invisible and devoid of light, turning fiery in the lotus-coloured sky. (2:23) This may indicate an eclipse or just describe an optical anomaly. The only clear lunar eclipse is the one associated with a solar eclipse in the same month, already mentioned above.
There is general agreement among all existing editions of the epic, and so we may proceed with a more detailed analysis of the occurrences. Three solar eclipses in particular, one each from Sabha Parva, Bhishma Parva, and Mausala Parva, have to be found satisfying a time sequence. The first or the second or both could have succeeded or preceded a lunar eclipse in the same month, and while they should have been observable at Kurukshetra, the third one should have been observed at Dwaraka 36 years after the War. The interval between the first and second eclipse is not stated in the text. But from other considerations it is clear that fourteen years should have elapsed between the events at Jayanta commencement of the War. The text is not clear on how long Pandavas stayed at Upaplavya (Udyoga Parva) after their year incognito before the start of the War, but it can be inferred that they did not fight in the same year, within a few months, after their thirteen year exile was completed. This we see, in all editions, in the statement of Krishna during his speech in the conclave of kings at Upaplavya: “How, about a year ago, after the cattle stealing episode, even when Bhishma pleaded for peace…” (Udyoga Parva 88:19-20)
In the Sabha Parva, before going to the forest, both Bheema and Arjuna say that they will come back and fight in the fourteenth year that is, after their exile of thirteen years. Indeed, there are other indications – Vyaasa announces after the events of Jayanta that the end of the Kauravas is foretold fourteen years from that date. But the above statement of Krishna should be taken to indicate that the Pandavas waited for slightly over a year. In any case, this does not resolve when solar eclipse occurred. Hence we are forced to take a sliding window of 13-15 years as the possible interval between first and second solar eclipses.
This dilemma is resolved by the mention of a planetary position in Udyoga Parva 142:8-10 (Shani is near star Rohini; Mars, in retrograde, is approaching Anuradha from Jyestha. There is a planet near Chitra) and Bhishma Parva 2:32 (Shani is staying near Rohini).” These are unambiguous about Shani being near star Rohini in the war year and give a strong hint as to the position of Shani in the year the War took place. However, two conflicting statements appear later. First, in Bhishma parva 3:14-19, Vyaasa observes, “Mangal is in Magha and Brihaspathi is in Shravana. Shani is afflicting Purva-Phalgunl. Shukra, previously getting up in star Purvabhadra and having circled in the North is looking up, with company. The dark planet blazing with smoke and fire is with Jyestha. Brightened Dhruva is positioning itself anti clockwise. A rough planet is seen between Chitra and Swati. The fire coloured planet has gone retrograde in Sravana.” The red-bodied planet is in Brahma rashi.” In Bhishma Parva 3:23-29, we are told that “Two blazing planets have reduced the brightness of Saptarsis. Brihaspathi and Shani, being stationary for a year are near Visakha. There is a sharp planet with the first star of Krittika, like a cornet. In the three stars preceding this, Budha is seen often. I know instances of amavasya falling on the fourteenth, fifteenth and sixteenth day of the fortnight, but not on the thirteenth day like now. The moon and the sun both got eclipsed in the same month, oddly in a thirteen-day interval.” Now either these two positions occurred at some other time or errors have crept into the text. The corresponding verses might have come into their present positions due to some recording or copying errors.
Scholars have speculated that when Yudhishtra asks for clarification from Vyasa regarding the three types of portents previously mentioned to him by Narada after the Rajasuya yaaga, the text sounds incomplete. What were the celestial and terrestrial anomalies? Could the double eclipse be one of them? There is a conversation between Vyasa and Dhritarashtra in Sabha Parva, which is too short by the standards of the epic. Is it possible some of the verses of Bhishma Parva really belong to Sabha Parva? Alternatively, is it possible that what Vidura mentions and Dhritarashtra later muses, as an eclipse at an odd time at the end of Sabha Parva could be the one with Shani near Visakha?
[DIVERSON 1: Wrong sequencing of texts in ancient manuscripts, based on which the present day printed versions are made available, is not unusual. Even smaller compositions such as Dharma Sutras have needed special efforts in arranging them in a sequence. For example Suryanarayana arranged the available versions of the Apastamba Dharma Sutra in the traditionally understood order in 1933. Similarly, Lagadha’s Vedanga Jyotisha text was edited and arranged sequentially, still more recently, by Kuppanna Sastri in 1984. In a text as large as Mahabharata it is no wonder the chronological ordering of the events and observations might have mixed up due to transmission errors.]
In terms of limits for their search, all scholars agree that the Mahabharata should have been before 900 BCE (the few who don’t, like Wendy Doniger and Romila Thapar, are not worth engaging seriously given the fundamental nature of the flaws in their understanding of Ancient India). As an extra precaution, since Krishna is accepted in all traditions to be earlier than the historical personality of Gautama Buddha (563-483 BCE), 501 BCE is selected as the lower limit. The upper limit is based on the limitation of the astronomical software and its database. The location for the eclipse search has been taken as Kurukshetra at 76°49’E and 20°59’N. Initially each year in the above time interval is searched for occurrence of solar eclipses. This is followed by a search for lunar eclipses only in years with solar eclipses. It is found that during 501-1000 BCE, 177 solar eclipses were possible at Kurukshetra. The numbers for 1001-2000 BCE and 2001-3000 BCE were 369 and 345 respectively. These are further scanned for preparing a list of double eclipses, that is, a solar eclipse succeeding or preceding a lunar eclipse at a fortnight’s interval. It is found that in the period under consideration 247 such double eclipses were possible. Among the 247 double eclipses, our interest is initially restricted to those during which Shani was near Rohini. Among these 247 double eclipses visible at Kurukshetra, a third eclipse visible from Dwaraka ((22° N, 69° E) 36 years hence should be entered into consideration. The number of possibilities is now brought down to eight.
Thus, a detailed search and analysis of 2,500 years of data produces eight sequences of three eclipses that are compatible with the statements in the Mahabharata: (520, 505, 469/70); (724, 711, 676); (811, 798, 763); (843, 830, 795/96); (1493/91, 1478, 1443/44); (2638, 2624, 2588/89); (2758/57/56, 2743, 2708); (2759/58/57, 2744, 2709). It would be prudent to remember at this point that we have already guessed and approximated some of our data, such as Shani being near Rohini. Therefore, we are not positive these dates are fully valid. If we consider yet another astronomical observation, that Mangal is between Anuradha and Jyestha, we can close in on 1478 BCE as a probable date for the Mahabharata. Unfortunately, the sky charts show that not Mangal but Shani was in retrograde at the desired time. To make this date fit with our calculations, it would have to have been Budha that was close to Chitra. A further problem with forcing this date is that the Saraswati, mentioned several times in the epic as a huge river, had dried up about 400 – 700 years earlier.
As most things are in the subcontinent, the Saraswati is fraught with its own problems. According to Rigvedic tradition, the river flowed from the Himalayan foothills to the Arabian Sea, forming a delta in the Rann of Kutch region. Around 2200 BCE, seismic activity caused the waters of the river’s two main sources to change course. The Sutlej moved course westward and became a tributary of the Indus River. The Yamuna moved course eastward and became a tributary of the Ganges. The tremendous loss of water which resulted from these movements caused the once mighty river to become sluggish and dry up in the Thar Desert. Recently, archaeologists using satellite images have been able to trace the course of the river. The river is mentioned 72 times in the Rigveda and is listed in the Nadistuti sukta. Thus, although the river is most definitely not mythical, many scholars think that references to the Saraswati referred to the Ghaggar-Hakra River. Such an assumption does not explain, except as poetic license, the description of the Saraswati as eight kilometres across in some areas.
Another approach to calculating the date the Mahabharata took place is from other records. There exist plenty of records in the Sanskrit and Buddhist tradition to allow us to venture that many of the characters of the epic were historical even if not all the stories are true. Aryabhata (476 – 550), the Indian mathematician, postulated that the Kali Yuga began on February 17/18, 3102 BCE. This was based on the premise that the sun, moon, and the five main planets would be in a particular alignment at the beginning of a yuga. Embarrassingly, this is not accurate – one scholar has shown that the actual date when the seven celestial bodies were in alignment was January 10, 3104 BCE. This date points to the occurrence of the War in 3137 BCE (because, according to the text, the Kali Yuga starts about 35 years after the War). Around 500 CE, a major review of the Indian calendar was attempted. Aryabhata, Varahamihira and others used the naksatra references that the Saptarsi were in Magha at the time of the Mahabharata war to determine its epoch. Aryabhata declared the war to have occurred in 3137 BCE, and Varahamihira assigned it 2449 BCE. This discrepancy arose perhaps from the different assumptions regarding the naksatras each astronomer made.
Evidence from the Puranas suggest another date. The Puranas are known for their exaggeration and poetic license, and in some cases, downright fictionalisation (There is no evidence anywhere of Radha, Krishna’s true sweetheart, except in the Puranas. Other figures such as Subhadra, Satyabhama, and Rukmini have multiple sources that corroborate their historicity). However, the Puranas maintain lists of kings of various dynasties in power. If we take the reign of Chandragupta Maurya (324 BCE) as a marker, the puranic lists take us back to about 1900 BCE. Michael Witzel has pointed out, however, that these lists may not be strictly chronological and some of the dynasties may overlap in years, as has been the case sometimes in other cultures.
This leaves us with three plausible dates (3139 BCE, 2449 BCE, 1900 BCE) and one less plausible date (1478 BCE). Presently, there is not enough evidence for us to pick one from these possibilities as the most likely candidate for the Mahabharata War. A lot of archaeological work remains to be done for any of these dates – the civilisations of Ancient Egypt and Mesopotamia left us with umpteen examples of their seals, texts, pottery, weapons, science, and engineering. Excavations in India have uncovered only a miniscule amount of material evidence for such an advanced ancient culture. The Mahabharata is filled with stories of various kings and the elaborate social structures existing in India. The text speaks of wonderful palaces, particularly at Indraprastha and at Jayanta – unless any trace of these was destroyed and removed by marauding armies later on, where are the remains of these buildings? The descriptions make it clear that stone was used, as were precious stones and marble. It is unlikely that not even a piece of it is left. It is also unlikely that these sites would be recycled for building material by later Hindu kings because the Mahabharata holds a quasi-religious position in the Hindu canon. Even if Muslim invaders sacked the ruins, there would be at least some record of the material being brought from the ruins to newer projects.
There are many questions that remain unanswered, and the first place to look for clues is with the Archaeological Survey of India and their sloth. And the jingoism of Indian nationalists does not help either – it is HIGHLY unlikely that nuclear weapons, missiles, or aircraft existed in the epic era. At best, the ability to imagine something does not prove its existence – or historians a thousand years from now might be talking about the fabulous teleportation devices and invisibility cloaks that existed in the 20th century! What is required, not only for India and her sense of history but for the richness of human experience, is systematic, back-breaking excavation and analysis, based on concrete evidence and not sleight of hand interpretations and translational tricks with the epic. To that end, most scholars have been found wanting.