Part II. The Author in Context
8. An Evolving Story
In this part of my dissertation I have presented, analysed and compared stories linked to the Mudrārākṣasa, preserved in various textual traditions. My attempt to do so is not the first. TRAUTMANN (1971:10–67) focussed on the legend of Cāṇakya, which meets on many points with my present interest, but does not cover all facets of the story from the point of view of the Mudrārākṣasa. He reconstructs a “primitive Cāṇakya-Candragupta-kathā” (ibid. 45–48) which, he contends, served as the basis of the Buddhist and the Jaina traditions and was also, at a later time, integrated into the Kashmiri Bṛhatkathā versions in a distorted form, made subordinate to the tale of Śakaṭāla. He lists the minimum elements of this primitive version as follows:
1. The prophecy involving Cāṇakya’s teeth and their ensuing breaking;
2. (probably) The nagging wife who sends him to get money;
3. His humiliation and vow (probably including the untying of his hair tuft);
4. His wanderings in the countryside;
5. The lesson of the hot gruel;
6. The alliance with Parvataka and the resulting victory;
7. The pacification of the kingdom;
8. The birth of Bindusāra; and
9. (probably) Cāṇakya’s rivalry with Subandhu.
I find Trautmann’s argument convincing and his conclusions plausible and in gen-eral corroborated by the slightly larger textual corpus I have examined. I would add that what he calls the “primitive Cāṇakya-Candragupta-kathā” already seems to be a mature tale and posit that there had been an even more primitive version. The text of the Purāṇas and the Mahāvaṃsa may conserve this version, though it is also possible that they are heavily abridged summaries of the “later primitive” tale. In any case, the fabulous elements in Trautmann’s primitive story indicate that this version has been influenced by previously existing tales about despotic rulers, crafty ministers and magical brāhmaṇs. In particular, the story of Cāṇakya’s humiliation and revenge probably arose from that of Śakaṭāla’s (or Virocana’s) Revenge. Again agreeing with Trautmann, I hold that the full tale of Cāṇakya’s Revenge (the one Trautmann calls primitive) interbred with Śakaṭāla’s story to give birth to the stories of Śakaṭāla’s Revenge by Cāṇakya.
However, we have yet to find a place for the Mudrārākṣasa in this network of leg-ends. TRAUTMANN (1971:46) says that this play “is a consciously artistic creation and as such freely has recourse to invention,” but his interest lies in the story (and history) of Cāṇakya, so for the sake of parsimony he did not examine the Mudrārākṣasa in depth. RUBEN (1956:150–200) had set out precisely with this aim. He was mainly interested in the other versions of the legend in order to point out what Viśākhadatta borrowed from other sources and what in the drama was his own innovation.
Ruben, however, did not trouble himself with what I consider the most important aspect of my present work: the examination of what Viśākhadatta presumed his audience would know. Moreover, Ruben’s analysis seems to start out from two implicitly accepted axioms. One of these is that Viśākhadatta was a contemporary of Candragupta II Vikramāditya, which, though possible, is far from certain. The second is that the extant story versions are of the same age as the textual witnesses preserving them. While it is true that a story cannot be proven to be older than the oldest text that records it, I main-tain that it is safe to accept (or, to put it more strongly, rash to reject) the notion that texts which purport to preserve ancient tradition do in fact preserve ancient tradition even if they had altered slightly in oral transmission before they were first committed to writ-ing.315 Ruben’s contention (RUBEN 1956:176–177) that elements of the Buddhist and Jaina traditions (such as the poison damsel and Candragupta’s unquestioning obedience to Cāṇakya) can be traced back to the Mudrārākṣasa is in my opinion untenable.
In order to further explore the connections between the Mudrārākṣasa and the Nanda legends, I have analysed the textual traditions posterior to and based on the play (which had already been scrutinised by both RUBEN [1956:192–201] and RAGHAVAN [1973:44–
54]). Among various partial agreements, there is a central detail on which—in spite of dis-agreeing on a number of other points— the epitomic tradition (all of whose witnesses hail from the far south) tallies with some representatives of the Mudrārākṣasa scholiastic tra-dition (represented only by northern witnesses316). This detail is the origin of the Nine Nandas and Candragupta, who are represented as the nine legitimate and one illegitimate (or at least, lower-born) sons of the elder Nanda in these works.317 This setup appears in none of the legends independent of the Mudrārākṣasa and provides a feasible background to the play, yet the implications of the play are not sufficiently explicit to make it likely that this setup was thought up independently by several later authors on the basis of the Viśākhadatta’s work.
Another intriguing motif that recurs persistently in the Mudrārākṣasa tradition, yet is entirely absent from the independent traditions as well as the play itself, is the as-sociation of various foodstuffs with Cāṇakya’s eradication of grass. In Ananta’s version there is a rationalising explanation attached: he uses honey (mākṣika) to attract ants which in turn are supposed to destroy the grass. This might be possible but sounds like quite an outlandish way to go about the task and would presumably work even less with gruel (Bikaner MS) or buttermilk (Jagaddhara and the Bengal edition). The two epitomes have a much more convincing version which I believe to be the origin of the above versions: he
315 Thus BONGARD-LEVIN (2000:115) remarks that the Mahāvaṃsa-ṭīkā draws on material “predominantly based
… on the Indian tradition” and may considered as be “one of the early Indian sources” about Candragupta and Cāṇakya. He also notes (ibid.) that Hemacandra “relied on very early Jaina writings.”
316 With the exception of the Trivandrum MS (note 240 on page 137). RAGHAVAN (1973:47–48) says nothing about the origin, age or script of this MS, so it may be northern or fairly modern.
317 See page 136ff. and particularly Table 7 on page 142.
does not destroy the grass with food, but annihilates it by making food out of it and con-suming it:318 a powerful, almost cannibalistic image of revenge.
Finally, an episode found only in the Cāṇakyakathā may shed light on a minor but nagging problem in the Mudrārākṣasa and could thus be old in origin in spite of the fact that it does not appear anywhere else in the tradition. When Cāṇakya goes to Parvataka to secure his alliance, Candragupta in this tale remains behind in Pāṭaliputra. The young bas-tard prince does not, however, wait idly for Cāṇakya’s return with an army, but with Jīvasiddhi’s help procures another one. The Jaina monk, now a trusted member of the Nanda cabinet, exiles him (v. 173), then begins persecuting Śabaras (central Indian barbar-ians) on the pretext that they are sheltering Cāṇakya (v. 175–176). Candragupta, using the fortune he inherited from his father,319 recruits the Śabaras, and when the time comes, uses them to besiege the city from another quarter.320 If such an episode had been part of an early story known to Viśākhadatta, it might explain why the list of tribes at the first siege of Pāṭaliputra (the one mounted by Candragupta and Cāṇakya against the Nandas) is different from those at the second siege (by Malayaketu against Candragupta).321
In the light of such details I believe that some representatives of the Mudrārākṣasa tradition preserve traces of a very old story version that probably pre-dates Viśākhadatta.
What we know for certain is that the extant texts of the traditions independent of the Mudrārākṣasa do not fully provide the background that Viśākhadatta took for granted, nor do they contain all the widely attested elements of the Mudrārākṣasa tradition that cannot originate from the drama itself. The inevitable conclusion is that there must once have existed an influential text or tradition that accounts for both. The alternative would be to postulate two lost texts, one known to Viśākhadatta and one known to the transmitters of the tradition, but the presently available evidence shows nothing in favour of adopting this more complex hypothesis. Therefore my conclusion is that in Viśākhadatta’s days there was one particular widely prevalent “mystery text” about the accession of Candra-gupta. This text gradually dwindled in popularity over the first half of the second millen-nium, perhaps being superseded by alternative accounts of the same historical episode, such as that of the Kathāsaritsāgara. This created a new demand for a prequel to the Mudrā-rākṣasa that was more consistent with the story of the play than the tales current at this later time. Scholiasts obliged and supplied diverse preliminary stories,322 some elements
318 In the MRNK he grinds the grass into a pap and eats it, while in the Cāṇakyakathā he first burns it, then drinks the ashes (presumably mixed with water).
319 See note 210 on page 132.
320 Cāṇakyakathā 194, pratyūṣa eva ca mahāpṛtanānuyātau cāṇakyaparvatanṛpau samam ekabhāge| mauryaḥ kirātapatisainyayuto ’nyabhāge ruddhākhilāśam atha puṣpapurīm arundhām|| The word kirāta seems to be used here as a synonym for śabara.
321 See Table 6 on page 52 for details and the surrounding text for a discussion. Śabaras only appear as a variant reading in the list of participants in the second siege, so the connection is tenuous.
322 EHRMAN 1959:133 also points out that mediaeval commentators felt a need to add preambles to the Mudrā-rākṣasa, and observes that these preambles are not satisfactory. However, Ehrman ascribes this simply to the “extreme intrigue” characteristic of the play, “Драма Вишакхадатты отличается чрезвычайной сложностью интриги.”
of which may have been based on actual knowledge of the mystery text, while some were based on then-popular story versions and the imagination and conjecture of the scholiasts.
A Missing Link
The two pillars of my attempt to find out more about this lost tradition are the two longest and most coherent representatives of the Mudrārākṣasa tradition: the pream-ble written by Ananta and the Cāṇakyakathā of Ravinartaka. Ananta’s tale is probably re-lated in some degree to all the northern scholiastic prologues discussed above, and it—or a story version not directly derived from it yet closely related—seems to have lived on in north Indian paṇḍit lore at least into the twentieth century, as shown by the prologue in the Bengali edition, as well as the two prologues in Bhaṭṭācāryya’s (likewise Bengali) edi-tion (see note 173 on page 125). In turn, the Cāṇakyakathā (or an earlier work of which it was a transcreation) appears to have been very popular in the south, as evidenced by its retellings by Mahādeva and Ḍhuṇḍhirāja. Thus we basically have a northern and a south-ern tradition, which go back to a common root from which their shared motifs originate.
The northern tradition has one more curious feature. While all the versions dis-cussed above end where a decent Mudrārākṣasa prequel should, that is, right before the drama begins, the paraphrase published by WILFORD (1799:263–267) does not. It continues with an “alternative Mudrārākṣasa” condensed into the space of two paragraphs. The gist of this story is that Cāṇakya and Candanadāsa (the latter under duress) both threaten to immolate themselves unless Rākṣasa will consent to come over to Candragupta, which he at length does. While it is possible, as I hypothesised in the Prolegomena (page 26ff.), that this is just a garbled summary of the Mudrārākṣasa hastily regurgitated to Wilford by a paṇḍit, another possibility is that this alternative ending originates from the elusive lost text, that it comprised part of some versions Ananta’s “prologue”323 still extant at the turn of the 19th century, and that it has been excised from the versions now extant because it contradicts the well-known Mudrārākṣasa.
Even my brief examination of Ananta’s and Ravinartaka’s texts has showed that both abound in lively dialogue. Also, while the former is in prose and the latter in ślokas, both are studded with numerous verses in longer lyrical metres. Some (probably most) of these would have been composed by the respective authors of these texts, while some give the impression of being popular subhāṣitas. A few, however, could well be relics cited ver-batim from the mystery text. Note in particular that in Ananta’s version of Cāṇakya’s vow324 the brāhmaṇ describes himself as “having been anointed as the foremost of all scholars in this very hall.” However, no such episode has been described in Ananta’s text, though it does happen in some of the Buddhist versions of the legend (see note 75 on page
323 Which in this case was not a prologue to the Mudrārākṣasa at all but an epitome of my mystery text.
324 Recounted in a śārdūlavikrīḍita stanza, see note 185 on page 128.
105). This inconsistency—and I believe others could be uncovered by a thorough study of both this text and the Cāṇakyakathā—suggests that the verse has been lifted directly from an older text which was summarised with omissions in the surrounding prose (or śloka) passages.325
Because of the use of dialogue and lyrical verses in the epitomes, I think it likely that the lost text we are looking for was a drama, and believe there is a fair chance that it was the play called Pratijñācāṇakya, ascribed to a certain mahākavi named Bhīma. This iden-tification is admittedly a long shot, but I know of no presently available evidence that would exclude it, and hope that one day something might surface from the dank and dark depths of manuscript collections to corroborate or disprove it.
The existence of the Pratijñācāṇakya is known from two references by Abhinava-gupta in his commentary on the Nāṭyaśāstra.326 One is concerned with the stock character called śakāra, a despicable, comical figure who speaks with a lisping accent. Abhinava re-marks that low birth is not a requirement, for even a king may be portrayed as a śakāra, as for example King Vindhyaketu is portrayed in the Pratijñācāṇakya of Bhīma. He then adds that there are no famous śakāras born in Aryan lands.327 In the second reference the title is Pratibhācāṇakya328 and the author is not named, but the likelihood that this is a different play is negligible. Here the context is about introductory scenes (praveśaka), which may be used to let the audience know about events that could not feasibly have been presented on stage, as for example in the Pratijñācāṇakya, the Mudrārākṣasa and the Tāpasavatsarāja (another play with a political theme, about King Udayana).329
The latter reference reveals little about the play, but it does to some extent cor-roborate what the title also implies: that it is a political drama about Cāṇakya’s vow. The former, however, is all the more intriguing. If the Pratijñācāṇakya featured a king called Vindhyaketu, represented as a śakāra and thus in all likelihood a barbarian, then we have
325 It must be noted at this point that the Cāṇakyakathā seems to contain no verses cited from the Mudrā-rākṣasa. The Mudrārākṣasanāṭakakathā, on the other hand, does (see also page 129 and note 192 there), though it contains no metrical verses at all in the section preliminary to the events of the drama. The most likely explanation is that Mahādeva cited these verses from his copy of the Mudrārākṣasa (rather than finding them already there in whatever text he based his work on), but this does not exclude the possibility that some of the stanzas in the Cāṇakyakathā are from the mystery text.
326 See WARDER 1988:402. DHRUVA (1930:xxxiii n1) remarks that according to the editors of the drama
Kundamālā, the Pratijñācāṇakya is “referred to in works on Alaṃkāra in the South.” This, however, seems to be a false trail. The preface to the edition of the Kundamālā referenced by Dhruva does mention the Pratijñācāṇakya, but only as one of a list of plays about whose existence we know from a list of alaṃkāra works including the Abhinavabhāratī (KAVI &SASTRI 1923:i). DHRUVA (ibid.) also notes that the Bhāvaprakāśana of Śāradātanaya mentions a play (without a title) in which Nanda and Candragupta appear as characters, as an example for the bhāsvara variety of nāṭaka. This may be a reference to the Pratijñācāṇakya.
327 Abhinavabhāratī ad NŚ 12.148, pratijñācāṇakye tan mahākavinā bhīmena rājāpi vindhyaketuḥ śakāra iti bhūyasā vyavahṛtaḥ … na cāryadeśajātiḥ śakāraḥ kaścit prasiddhaḥ.
328 Pratijñācāṇakya could be rendered as “Cāṇakya’s Vow” or “A Play about Cāṇakya Involving a Vow,” while Pratibhācāṇakya would be “The Genius of Cāṇakya”—appropriate but less convincing than the former variant.
329 Abhinavabhāratī ad NŚ 18.35, tathābhūtābhidhānayuktaś ca praveśako bāhulyena tāpasavatsarāja-pratibhācāṇakya-mudrārākṣasādiṣu.
a notable connection to Malayaketu of the Mudrārākṣasa.330 Interestingly, scholars who have written about this connection all seem to take for granted that Vindhyaketu is anal-ogous to Malayaketu.331 It is in my opinion beyond doubt that the Vindhyaketu of the Pratijñācāṇakya must correspond to Parvataka of the Mudrārākṣasa and the Cāṇakya tradi-tion. Bhīma’s play is named after Cāṇakya’s vow, so in all likelihood it dramatises the events leading up to its fulfilment, probably including some preceding events, but surely not the events after it was fulfilled, which form the subject of the Mudrārākṣasa. Malayaketu (or an analogue of his) would have had no considerable role in these earlier events. Nota-bly, he does nothing at all in any of the texts witnessing the Cāṇakya tradition except the Mudrārākṣasa, and there too, Viśākhadatta feels the need to introduce him to the audience as the son of Parvataka. Finally, there is even a slight bit of positive evidence linking the name Vindhyaketu to the Cāṇakya traditions: the analogue of Parvataka in the Bengali edition’s preamble is named Viśvaketu, which sounds reasonably similar.332
RAMAMURTI (1930:80–81) mentions that anthologies cite stanzas ascribed to Bhīma, and some of these “seem to have been taken from a drama.” Unfortunately the only an-thology to which he refers by name is the Sūktimuktāvalī.333 This collection has three cita-tions ascribed to Bhīma, none of which seem connected to the story of Cāṇakya. All that is known about the date of mahākavi Bhīma is that he must predate Abhinavagupta (ca. 1000 CE), who refers to his play. RAMAMURTI (1930) proposes to identify him with Bhīmaṭa, a king of Kālañjara who probably lived in the 9th century and is known from other sources to have composed dramas. WARDER (1988:401–402) approves of this identification but does not con-sider it certain. Both Ramamurti and Warder assume implicitly that the Pratijñācāṇakya must be based on the Mudrārākṣasa and thus later than it, but I see no reason why this should be so, and the identification of Bhīma with Bhīmaṭa is far from convincing.334 Thus, while there is no positive proof that the Pratijñācāṇakya predates the Mudrārākṣasa and tells a story which became widely known by Viśākhadatta’s time, this assumption is in no way less feasible than that the Pratijñācāṇakya is a ninth-century reworking of the Mudrā-rākṣasa.
As an interesting aside, it might be noted that the Cāṇakya tradition lives on.
There is a contemporary Sanskrit drama bearing the title Pratijñākauṭilya,335 which does precisely what I assume the Pratijñācāṇakya did, dramatising the events from the birth of
330 Malayaketu’s name means “Flame of Malaya” or “Banner of Malaya.” Malaya is a famous mountain in South India (see also page 59); Vindhya is the name of another famous mountain (range) that separates South India from North India.
331 Thus e.g. RAMAURTI 1930:81, RAGHAVAN 1973:54 and WARDER 1988:402.
332 And would look even more similar in many Indic scripts, where the n of Vindhyaketu could be represented by an easy-to-lose anusvāra, and dh is often indistinguishable from v.
333 See page 190.
334 Against this identification it might be noted that the Sūktimuktāvalī contains a stanza ascribed to Bhīmaṭa beside the three ascribed to Bhīma mentioned above.
335 VAKULABHUSHANAM 1968. The English foreword to the play was written by V. Raghavan.
the nine Nandas to the coronation of Candragupta. The plot seems to be based on the southern epitome tradition, and the cast of characters includes among others the old king Sarvārthasiddhi and his lowborn son Maurya, as well as Parvateśvara336 and his son Malayaketu. The Candrābhiṣeka of Bāṇeśvara337 may also be an early modern witness to this tradition.
The Advanced Cāṇakya-Candragupta-kathā
Since it appears that Viśākhadatta expected his audience to be familiar with a particular version of the Cāṇakya legend which cannot have been identical to any of the stories handed down to us along lines of transmission not associated with the Mudrā-rākṣasa, I now attempt to reconstruct an “advanced Cāṇakya-Candragupta-kathā” that could have served as the background for the Mudrārākṣasa. This hypothetical ur-text may have been the Pratijñācāṇakya itself or may have served as its basis. The grounds for my recon-struction include on the one hand explicit allusions in the Mudrārākṣasa to details also de-scribed in one or more independent traditions, and on the other hand vague allusions in the Mudrārākṣasa to details consistently found in several witnesses of the tradition associ-ated with the play. In the following paragraphs bold emphasis marks statements that are quite certain, while the more dubious parts are printed in plain letters.
An old king named Nanda ruled in Pāṭaliputra. He may have been a good king and a positive character in the narrative, but my intuition is that he was probably presented as a neutral or negative figure. He had a clever minister whose name was probably either Rākṣasa or Subandhu (or both). He may have attained his position by solving riddles. He too, like Nanda, may have been portrayed as positive or neutral, but I believe he was a fiendish figure and was named Rākṣasa to highlight this. A member of the royal family was called Sarvārthasiddhi. This may have been the old king himself or, more probably, his brother or uncle.
The Elder Nanda had nine legitimate sons and an illegitimate one. The nine legit-imate sons may have had the same mother (probably called Sunandā) or may have been half-brothers by three different mothers. The illegitimate son’s mother was called Murā, may or may not have been married to the king, but was certainly inferior in rank to the mother(s) of the legitimate sons. The legitimate sons were referred to collectively as the Nine Nandas, and the illegitimate son was known as Maurya, as well as probably by his personal name Candragupta. He was a very talented child and may well have been the firstborn son of the elder Nanda.
The Elder Nanda died or retired, handing the kingdom over to his nine legitimate sons who probably devised (with the collaboration of Rākṣasa) a scheme to rule in turns.
They became despotic rulers and they treated Maurya badly. He may have served for some
336 Described as a Turuṣka, i.e. Turkish king.
337 See note 14 on page 28.
time as the general of the armies, but the Nandas imprisoned him on false charges of trea-son, probably with Rākṣasa as their accomplice. He was later released (or, less likely, he died in prison but his son Candragupta was released) because his help was needed in solv-ing a politically important riddle, and was thereafter assigned to a job of no consequence, probably becoming the overseer of the soup kitchen.
One day he encountered Cāṇakya, a brāhmaṇ who had come to receive a donation (probably just a free lunch) from the Nandas and was at the moment busy liquidating grass, probably by uprooting, grinding and ingesting it. He explained that he was doing so because the grass had pricked his foot. Maurya befriended Cāṇakya and invited him to the palace, expecting a quarrel between the arrogant kings and the arrogant brāhmaṇ.
Cāṇakya took a seat at high table and the Nandas, ineluctably, ousted him, possibly giving the foremost seat to Rākṣasa. He unknotted his tuft and swore not to tie it again until he has uprooted the Nandas like so much grass.
He offered the throne to Maurya (or his son Candragupta) and left the capital, with or without the bastard prince. Cāṇakya sought alliance with Parvataka, the barbarian ruler of a mountainous region, promising him half the kingdom in return for his armed forces. Parvataka was probably a reluctant vassal of the Nandas, and his personal name may have been Vindhyaketu. Maurya/Candragupta may have accompanied Cāṇakya or stayed behind in Pāṭaliputra to make another alliance with a barbarian chieftain living closer to the mainland. They besieged and captured Pāṭaliputra, and Cāṇakya anointed Candragupta as king. The Nandas died in the battle, but their minister survived and pre-tended to accept the new king. As a token of his submission he sent him a poison damsel.
Cāṇakya found out what she was and persuaded Candragupta to present her to Parvataka, thus eliminating him and his claim of half the kingdom. The story may have ended with a brief epilogue in which Cāṇakya persuaded or forced Rākṣasa to become Candragupta’s minister in truth, then retired to the forest.