Epic Threads - John Brockington on the Sanskrit Epics
edited by Greg Bailey & Mary Brockington, 366 pages. New Delhi: Oxford Univ. Press 2000
Where the Indian family is concerned, Ramayana is by far the more comfortable of the two great epics. When Annie Besant wrote her adaptations of these, she found that in Valmiki's composition black and white are quite unambiguously contraposed and Rama is so obviously the ideal hero that the petit-bourgeois householder is not confronted with any awkward moral issues. Mahabharata, on the other hand, pitilessly holds up a mirror to the greed, anger and lust that are the stuff of human existence, raising far too many knotty questions around its heroes. Rama is more remote, 'safe' and readily worship-able as a divine incarnation than Krishna. But was Ramayana always the paean of praise to the Vishnu avatara whom Tulsidas brought into every village home of the Hindi heartland as maryada purushottam Ramachandra?
In Righteous Rama [OUP 1984] Brockington had made his first attempt to trace the evolution of the epic, subjecting it to linguistic and stylistic examination to find five distinct layers the text went through to establish Rama as the moral ideal of righteousness:
- Orally transmitted c. 5th ' 4th century BC - all shloka stanzas of books 2-6, i.e. 37% of the text.
- Orally transmitted c.3rd ' 1st century BC - all verses in longer metres in books 2-6 (4.27% of the text) and 54 sargas of book 2, 23 of book 3, 19 of book 4, 28 of book 5, 36 of book 6.
- Written c.1st ' c.3rd centuries AD - books 1 and 7 (24.57% of the text) and possibly parts of the end of book 6.
- Written c.4th - c.12th centuries AD - passages marked with (*) or Appendix I of the Critical Edition with good manuscript support.
- Written c.12th century AD onwards - passages marked with (*) or Appendix I with poor manuscript support.
He also made a very important point regarding the difference between the two great epics of India.
Mahabharata, as it evolves, depicts a steady tendency to shift from the dramatic to the ethical viewpoint, while the evolution of Ramayana shows a change from the poetry of action to the poetry of feeling, emphasizing the emotional and lyrical at the expense of the heroic. Frank Whaling's 1980 study1showed that around 4th century BC Valmiki is known simply as belonging to Koshala ignorant of Central and South Indian geography. This changed by the 1st century BC in the Balakanda to a famous sage whose ashrama was near Gliga. By the 2nd century AD in the Uttarakanda Valmiki has become a friend of Dasharatha and a Bhargava. In the latest texts, he is a robber belonging to a low caste. According to Whaling, a ballad of 12000 verses split into adhyayas developed into cantos (sargas) with tag verses of variant metres. Then regional variants sprang up interpolating geographical cantos and supernatural material. Finally, under Brahmanical influence, came the legendary part of the Balakanda and the entire seventh book with the mythic origin of the epic through divine inspiration.
Epic Threads is a selection of Brockington's studies that is possibly one of the supreme examples of 'lower criticism.' The Western obsession with an 'ur epic' is grimly pursued through lexical microscopic analysis to determine the accretions, perhaps missing the point Whaling made that the significance of the epic lies not in the cause but in the effect, the end, that is what Rama developed into for people as a symbol of meaning from Rama the man through Rama the successor of Indra to Rama the avatara of Vishnu, though not yet worshipped. The papers span the period 1969-1995 and are arranged in four sections: linguistic evidence of stages of composition in six papers during the period 1969-1982; manuscript studies in two papers of 1986 and 1991; four papers setting forth the implications of such analysis (1976-7, 1995-97); finally three papers on the relationship between the two epics (1978, 1985, 1986). An extremely useful table has been added from Righteous Rama plotting the stages of composition of the Ramayana (reproduced above). Brockington's thesis is that it is not just the 'growth' of a martial ballad through padding, such as the Bala and the Uttara kandas, into a classical epic. It is also, as Greg Bailey perceptively notes in his learned introduction, a thematic and theological development mirroring changes in the socio-cultural and religious context. Brockington furnishes exhaustive statistical and typological surveys of linguistic features from the Ayodhya, Aranya and Kishkindha kandas, analyzing even particular manuscripts within recensions to reveal peculiarities of syntactic structures, using raw linguistic data of the numbers and statistical frequency of occurrence of words, phrases, and grammatical forms. The focus is non-narrative and the plot is referred to only for explaining particular linguistic features.
The first two papers analyze the verbal and nominal systems, inferring that 'the high frequency of periphrastic futures in the first part of the Ayodhya kanda is due to interpolation or rewriting.' This corrects the misconception propagated by Louis Renou that such usage is not found in this epic. Paninian rules of using different past tenses according to situations are found not to be followed with any regularity while rules regarding compounds and use of the suffix tak with han are also contravened in the Ayodhya and Aranya kandas. By inference, these precede Panini. Further, whereas in Ramayana the perfect is the most common past tense, in the Udyogaparva of Mahabharata it is the imperfect that is more frequent. Again, at least 600 of the approximately 5700 long compounds in the former occur in the latter, showing their stereotyped nature. These are used in dramatic or emotional passages and occur frequently in the later parts of both epics, especially in the war portions (as much as 30 to 40% of all shloka stanzas). Refrains are most common in Ayodhya kanda, personal epithets and stock padas in Aranya kanda and repeated padas and long compounds in Kishkindha kanda. Unfortunately, Brockington does not spell out the implications. In terms of the nominal system, Ramayana, like Mahabharata's older portions, falls in-between the Brahmanas-Sutras and classical literature. Important findings for dating the epic are the infrequent use of a gerundive as a prior part of a compound, whereas it is characteristic of Buddhist Sanskrit, and the use of multi-member dvandvas and 'dhara compounds that are absent from the early Upanishads.
The eight books of Ayodhya and Aranya kandas that are abnormal in other respects also show a heavy incidence of vriddhied derivatives, thus arguing for considerable interpolation (e.g. Bharadvaja entertaining Bharata's army, Bharata's return to Ayodhya, Lakshmana's description of winter, Sita's dissuasion of Rama from unprovoked killing, Rama's visit to Agastya's hermitage). Brockington points out that the significant increase in the proportion of long compounds from the Ayodhya to the Kishkindha kanda argues for interpolated material retained even in the Critical Edition. Lakshmana's entry into Sugriva's palace, the entry of the Vanaras into Rikshabila, Hanumana's views of Lanka and Ravana's palace, his emotions on seeing Sita, the killing of Aksha - all show an unusually high proportion of long compounds. Brockington even isolates a particular class of compounds beginning with krodha peculiar to the Bala kanda and with taptakancana in the Aranya kanda that are wholly absent from Ayodhya kanda. His conclusion is that none of the compounds of more than eight syllables can belong to the genuine epic. The syntax of cases is simple, regular with infrequent use of prepositions or their substitutes and of periphrases, all indicating that 'the genuine epic dialect belongs to the older strata of the language.'
By studying stereotyped expressions, Brockington concludes that the epic is the work of a conscious artist working 'within the limits, and in the spirit, of a living epic tradition.' It is revelation to find that only Rama is called satyaparakramahin Ayodhya kanda frequently, and that this recurs just once in Aranya and Sundara kandas, is absent in Kishkindha kanda but becomes frequent in Yuddha kanda. In Aranya and Yuddha kandas, he is most often described as son of Dasharatha, which is absent in Ayodhya kanda. Aklishtakarman is restricted to Rama in the epic but is applied to both Krishna and Partha in the Mahabharata. Stock epithets for Sita are only three in Ayodhyakanda, but increase in Aranyakanda and later, all relating to her being Janaka's daughter and belonging to Mithila and Videha. In the forest she is most of all Vaidehi and slender-waisted. A common phrase she shares with Draupadi is dharmapatni yashasvini. Ravana is most commonly described as ravano rakshasadhipah, Hanuman as marutatmajah.
The implications of these findings are spelled out in the paper 'The names of Rama' where he infers that this is due to the shift in emphasis to Rama as a morally righteous hero. Brockington cites no reason for concluding that, 'the story of Surabhi at 2.68 is undoubtedly borrowed from the Mahabharata' (p.122) despite the admission that there is only one really close parallel in language. In the last paper of the book he summarizes this study by dividing these formulaic expressions into four main groups: those frequently occurring in both epics and indicating a common inherited tradition; those occurring only in Mahabharata or borrowed from it in later parts of Ramayana; those peculiar to Ramayana or only in the late passages of Mahabharata; and those occurring only in later parts of both epics showing interlinkages with the Puranas. The analysis suggests that 'redactors of the Northern recension were more familiar with the Mahabharata than those of the Southern' and that the two traditions had merged by the time of the later parts of the epics. The fourth variety has a broad religious import reflecting the altered area of interest and could be a consequence of epic transmission having been taken over by Brahmins. The parallels or borrowings are concentrated in certain books of each epic, which helps to unravel their textual history.
Figures of speech are analyzed to reveal that the epic is in the early stages of formation of the corpus of standard imagery that characterizes classical Sanskrit literature. Thus, there are few similes referring to the lotus, in contrast to its profusion in later literature. Similes are the commonest figure of speech and others are used very sparingly only at dramatic points of the narrative towards the later parts of the epic. This is where the suspect passages differ strikingly, as they amplify a detail or episode by using a figure of speech. Finding that one in six of similes in Ayodhya kanda, one in four in Aranya kanda and one in five in Kishkindha kanda have exact or similar parallels in Mahabharata, Brockington concludes that there was a common stock of similes shared between the epics and displaying the homogeneity of the epic tradition. He also points out the greater frequency and sophistication in using figures of speech in Ramayana than in Mahabharata, but does not realize that this is what sets apart the kavya from the itihasa.
An important finding is that Indra predominates among the gods mentioned in the similes, with no mention of Vishnu and Shiva in Ayodhya kanda. The suspect passages abound in similes referring to Vishnu. Brockington identifies four similes of Shiva and two of Vishnu as interpolations retained in the Critical Edition besides 2.85 (Bharata's army being entertained) and 2.88-89 (the beauties of Chitrakuta). The fact that similes referring to cattle reveal no veneration of the cow indicates the antiquity of the epic. The elephant accounts for the maximum number of similes referring to animals, followed closely by snakes (frequent in Sundarakanda). The natural world predominates in Ayodhya and Kishkindha kanda similes, the gods in the Aranya kanda. The ocean occurs frequently in Ayodhya kanda but is absent in Kishkindha kanda (where mountains are much more) and rare in Aranya kanda. Ayodhya kanda has ten similes involving the night that do not recur elsewhere. Brockington concludes that 'the similes reflect a pattern of society and culture characteristic of a fairly early period' and the usage bears out the dating of the original epic as having been composed by the fourth century BC.
In the earliest stage of the text, proverbial expressions are few, accounting for just a quarter of the total and many occur for the first time in the Ramayana. Over 40% occur in the reworked or added portions; some revealing a didactic element that is so obvious in the Shanti and Anushasana parvas of Mahabharata. Brockington improves upon Sternbach's specialisation in this field by noting 72 proverbs of which 52 recur in the later epic as well (Sternbach and Hopkins mention 20). Over a third of these do not occur in other Sanskrit texts. Many of the proverbs recurring in Mahabharata relate to the evils of a king not protecting his people and to a kingless state. He points out that in the Northern recension, proverbs are twice as many as in the Southern, and the alamkaras are also more frequent, reflecting the more polished text of the former.
Studying the syntax of the epic, Brockington finds that the connective particle uta, frequent in the later epic, is rare in Valmiki and restricted to added passages. The relative system is simple with infrequent use of the double relative, restriction of a causal sense to yad, rare occurrence of pronominal adjectives and similar forms (even yatah is seldom used). Sentences are connected by a connective particle often stressed by use of anaphora through repeating the verbal idea of the earlier sentence by an absolutive or participle. He concludes what is surely axiomatic, viz. 'the earlier portions were written in what is basically a very simple, straightforward style' complex constructions becoming more frequent later.
A major contribution by Brockington is his research showing that several assertions in the Critical Edition are mistaken and the simple contrast posited between Northern and Southern recensions 'does not adequately represent the complexities of the chain of transmission involved.' He finds that the Southern recension is not as uniformly consistent as presumed, that the interrelationship between various recensions is far more complex than recognized so far and that Kerala had a definite alternative tradition. His textual study of Malayalam manuscripts of the Ayodhya kanda shows that those selected for the Critical Edition do not represent fully the Malayalam subrecension. He found around 125 manuscripts that the editors did not use until the Uttara kanda. Only 29 mss were used for the Ayodhya, Aranya and Sundara kandas against 41 for the Uttara. He also discovered that the manuscript evidence in the Critical Edition is incomplete. No Oriya mss have been used and only one Maithili. He found that attention had not been paid to checking the mss for their alignment that can alter not just from one kanda to another but even within a single kanda. He points out that the Varadaraja commentary is the oldest extant one on the text and would contain valuable clues. His findings warrant reopening the 'closed' status of the Critical Edition for taking full account of the wealth of available material. He makes a path-breaking suggestion that needs to be taken up by editors of the epic: instead of the Critical Edition's system of assigning mss to recensions or to script versions, which is too limiting in either slotting a particular mss to a version or dismissing it as contaminated, a better model would be the Venn diagram of mathematics (T.S. Eliot would have condemned this as a typical instance of 'jargon'). This model (p.204) depicts the different originals that have formed a specific mss while avoiding the impossible task of reconstructing the entire chain of transmission.
After such intensely demanding reading, the remaining papers show a welcome move into the area of 'higher criticism', an exemplary instance of which can be found in Sri Aurobindo's writing in the early years of the twentieth century:
'The longer speeches in the Ramayana, those even which have most the appearance of set, argumentative oration, proceed straight from the heart, the thoughts, words, reasonings come welling up from the dominant emotion or conflicting feeling of the speaker; they palpitate and are alive with the vital force from which they have sprung - they have, like his (Homer) the large utterance - of the primal emotions - Valmiki, when giving utterance to a mood or passion simple or complex, surcharges every line, every phrase, turn of words or movement of verse with it; there are no lightning flashes but a great depth of emotion swelling steadily, inexhaustibly and increasingly in a wonder of sustained feeling, like a continually rising wave with low crests of foam.'
Brockington cannot scale such heights, but examining the religious attitudes in the epic by drawing upon his study of similes he finds that the religious pattern is more archaic than has been generally recognised. The elaborations of Brahmanic literature are absent, leading him to propose 'a definite dichotomy between Brahman and Kshatriya in the immediate post-Vedic period.' The pantheon is markedly Vedic (even Garuda occurs more often than Vishnu, with whom he has hardly any association in the epic; neither has Shri). Indra is the most important deity and commissions Rama's exploits. Whaling showed that Rama's heroic exploits follow the paradigm of Indra's victory over Vritra and he actually uses Indra's weapon and chariot. Indra even says that with his help Rama will defeat Ravana whose son has conquered him. Thus, Rama is his successor in the battle against the demons. Brockington's findings are that Varuna has faded in importance, his twin Mitra is absent, and his pasha (noose) has been taken over by Yama. There is no serious thought regarding life after death, or of a previous existence. Destiny is assigned a limited role. Brahma becomes prominent in the second stage as the creator (but not four-headed) and replaces Indra as boon-bestower. It is he, and not Vishnu, who as a boar raises up the earth from the waters. While in Ramayana it is Indra who restores the dead warriors to life at Rama's request, in the Ramopakhyana of Mahabharata this is changed to Brahma. Ganesha is absent, but Skanda is mentioned as Kartikeya and Guha. Shiva is not regarded as superior to the gods and is usually referred to as Rudra along with his wife Agrajaputri and mount Nandi. References to images are totally absent till the third or fourth stage. Buddhism and Jainism are unknown. There is little sign of cow-veneration (Bharadvaja offers Rama beef at 2.48.16 and Rama refers to the sage Kandu killing a cow at his father's command). The Asuras are characterised more by power than by demonic traits. The meagre detail regarding sacrifices mostly relates not to the sages but to the Rakshasas who are associated with caityas (cult spots).
Both epics make no reference to any fixed or constructed place of worship. Religion is projected more as social duty, matching the Kshatriya background of the epic, and less formal aspects of worship receive emphasis. The ashvamedha description comes in the later Bala and Uttara kandas, and it is interesting to see the similarity between the description in the Bala kanda and Yudhishthira's ashvamedha (Vyasa cites the example of Rama when urging Yudhishthira to perform this sacrifice).
Brockington challenges Sheldon Pollock's argument that the divinity of Rama is integral to the epic from the earliest stages and was suppressed from the Ayodhya to the Yuddha kandas by pointing out that many interpolations make no mention of such divinity while others do. In that case, how can Pollock maintain that Rama's divinity is part of the un-interpolated part of the epic? Moreover, the two passages on which Pollock depends are in sargas that differ markedly in style from the epic core. Brockington points out that Rama is wholly human in the earliest stage. Hanuman denies identification with Vishnu twice and explicitly calls Rama human in 5.48.11 and 49.26. In the second stage, Rama begins to display divine qualities and is compared particularly with Indra (60 times at least), while Lakshmana is compared to Vishnu, in keeping with the early tradition of Vishnu being the youngest of the Adityas. Indra remains the standard comparison for the warrior king as the Kshatriya's ideal. Indra loses ground in Bala and Uttara kandas to Vishnu and Shiva who gain status considerably. The latest parts subordinate Shiva to Vishnu. In the third stage, around the second century AD (Bala and Uttara kandas and sarga 105 of Yuddha kanda) Rama is identified with Vishnu. This is consolidated in the fourth and fifth stages.
Another important contribution of Brockington is showing that Valmiki's Rama is quite different from the hero of the vernacular versions of the epic. He draws heavily upon his papers on religious attitudes and stereotyped expressions to show how the stereotyped expression ramo dharmabhritam varah changed in meaning from 'a pillar of the correct social order' to 'best of upholders of righteousness.' It is interesting to see how the same conclusion was arrived at without lexical analysis on thematic grounds by Whaling who points out that Rama upholds the ideal of a transcendent dharma which cannot be pinned down in terms of Dharmashastra but which is operative in the practical world in books II to VI. Brockington ignores Whaling's perceptive insight that Valmiki stresses Rama's Aryan ancestry while Vyasa is not concerned about Krishna's, and that Rama has none of Krishna's philosophical brilliance. However, it is interesting that in book VI after the fire ordeal Indra and Brahma refer to Rama as Krishna!3He is a dynastic hero promoting the idea of raja dharma and presiding over the ideal kingdom as an ideal individual, setting norms of dharma, of human relationships and of heroism.
Whaling also made the telling point that the critical century of the Ramayana tradition was not that of its composition in the fourth century BC but when people came to believe in him as God (Tulsi and Kabir in the 15th century AD). Brockington points out that while filial duty is the mainspring of the first part of the plot, it is devotion to Sita in the second. Moral scruples creep in at the second stage (Sita advising Rama against killing Rakshasas, Rama justifying killing Vali, the fire-ordeal). Later still Sita has to prove her chastity by appealing to the earth and Rama's character has evolved from the martial to the moral, from the hero to the avatara. The analysis of 'the names of Rama' reveals the relative lack of stereotyping in epithets stressing his strength, prowess, and pugnacity, suggesting that the martial aspect was not particularly important in the conceptualization of his character. Moreover, the popular image of Rama as the bowman has no support in the epic where this appellative is applied more often to Lakshmana and as often to others, including Rakshasas. Surprisingly, allusion to Rama's wisdom (ramasya dhimatah) is not noticeably frequent except in the fourth stage. Brockington points out that the reference to Rama's wisdom in the context of Sita's banishment 'indicates the (later) poet's explicit approval of the attitudes involved.' The terms indicating Rama's parentage decline in frequency while the more general dynastic terms raghava and kakutstha increase, with a shift to the elaborate raghunandana in the later stages - the form preferred from Kalidasa to Tulsidas.
A parallel examination of 'the names of Sita' is fascinating in revealing that she is depicted as a much more independent and forceful figure by Valmiki than she is made into in later versions and in the popular conception. She is ready to argue with Rama and it is her wishes that prevail at two crucial junctures of the plot (the exile and the golden deer). Even in the third stage, the same spiritedness is displayed in undergoing the fire-ordeal and calling upon the earth to vindicate her. The commonest epithets applied to her are patronymics - Vaidehi, Maithili, Janakatmaja, Janaki, in that order. All such epithets allude to her as Janaka's actual daughter, the most striking being 5.31.12 where she refers to herself as duhita janakasyaham and none allude to her miraculous birth. Surasutopama, comparing her to a god's progeny, is exclusively applied to Sita in a way that cannot be paralleled among the stereotyped epithets applied to Rama. Adjectives denoting ornamentation are not part of her basic image (sarvangashobhana stresses her inherent beauty).
In an extremely important comparative study of the epic and its summary (Ramopakhyana) appearing in Mahabharata, Brockington is able to place the latter to the end of the second stage of Ramayana and prior to the composition of the Bala and Uttara kandas. He guesses that Ravana's genealogy was included in Ramopakhayana earlier than in Ramayana because otherwise more of the Uttara kanda would have found place in it. He rules out 'reverse borrowing' because the stereotyped expressions typical of Ramayana occur in Ramopakhyana in sequence with the narrative, while those characteristic of Mahabharata occur more randomly. He even shows that Ramopakhyana is based on an older form of the Northern recension (best represented in the extant NE recension) that had not diverged as far from the Southern as now, thus silencing dissent against Sukthankar's conclusions, and encouraging taking them further.
The vexed question of dating the epic receives an extremely valuable input from the study of four passages in Mahabharata and Harivamsa, two of which treat Rama as Vishnu's avatara and the other two include him among 16 ancient kings. Brockington shows their common dependence on the end of the Yuddha kanda, with the Shantiparva passage drawing directly on it being the earliest, treating Rama as a mortal hero. The Harivamsa version is familiar with the Uttara kanda's divinisation of Rama and is expanded in the Sabha parva passage. The Dronaparva passage is familiar with Rama as mortal hero and as avatara. Brockington concludes that such evidence corroborates dating the final version of Ramayana well ahead of Mahabharata's final redaction, sometime in the middle of the later epic's growth.
The conclusions that emerge from such a rigorous presentation of evidence can assuredly constitute the foundations of what one may describe as the 'higher criticism' of the epic. Brockington's work encourages the completion of a complete lexical study that should no longer be regarded as a Sisyphusian task today with digitized versions of the epics available for computerized analysis. It is fascinating to see how 'lower criticism' has moved through its microscopic analysis to a finding posited long back in Sri Aurobindo's 'higher criticism': 'His (Valmiki's) picture of an ideal imperialism is sound and noble and the spirit of the Koshalan Ikshwakus that monarchy must be broad-based on the people's will and yet broad-based on justice, truth and good government is admirably developed as an undertone of the poem.' The difference is that, precisely because of his preoccupation with minutiae and 'righteous Rama,' Brockington loses a sense of proportion and does not realize that unlike Vyasa for whom in Mahabharata good government is the 'uppermost and weightiest drift'. He (Valmiki) is a poet who makes occasional use of public affairs as part of his wide human subject' (Sri Aurobindo).